You are on page 1of 484

SCIENTIST ANO FARMER

PARTNERS IN RESEARCH FOR THE 21sr CENTURY


PROGRAM REPORT

1999-2000

1NTERNATIONAL POTATO CENTER

lnternational Potato Center


Apartado 1558
Lima 12, Peru
cip@cgiar.org
http://www.cipotato.org
Scientist and Farmer Partners in Research for the 21 ST Century
Program Report 1999-2000

ISSN 1680-9270
September 2001, 2500 printed
Copyright lnternational Potato Center 2001
The lnternational Potato Center. 2001.
Scientist and Farmer Partners in Research for the 21 ST Century Program Report 1999-2000.
Lima, Peru. 480 p.

Contents
Program Overview
Wanda Collins, Deputy Director General for Research ................................................. 7
CIP's Research Project Portfolio ..................................................................................... 11
Building Partner Skills
P. Malagamba, E. Echeanda ....................................................................................... 16

Research on Potato
Plant Oefense Genes Associated with Quantitative Resistance to Potato Late Blight
P. Manosalva, S. Torres, F. Trognitz, R. Gysin, D. Nio-Liu, R. Simon,
M. Herrera, W. Perez, J. Landeo, B. Trognitz, M. Ghislain, and R. Nelson ................ 27
Characterization of Phytophthora lnfestans Colonizing Oifferent Solanaceous
Species in Peru, with lmplications on the Control of Potato Late Blight
G. Garry, G. Forbes, A. Salas, W. Prez, M. Santa Cruz, H.M. Pinedo,
E. Gonzles, M. Rivera, and R.J. Nelson ..................................................................... 39
Evaluation of Wild Potato Species for Resistance to Late Blight
W. Prez, A. Salas, R. Raymundi, Z. Huamn, R. Nelson and M. Bonierbale ........... 49
Quantifying Genetic Variance for Horizontal Resistance to Late Blight in
Potato Breeding Population B3C1
J.A. Landeo, M. Gastelo, G. Beltran, and L. Diaz ...................................................... 63
The Effect of Nitrogen Fertilization on Potato Late Blight in the Field
H.S. Jurez, J.R. Amaro, M.O. Rivera, A. Prraga, and R.J. Hijmans ......................... 69
Relationships of Fungicide Application to Late-Blight Oevelopment and
Potato Growth Parameters in the Tropical Highlands of Uganda and Kenya
O.M. Olanya, R. El-Bedewy, P.S. Ojiambo, P.T. Ewell, and J.J. Hakiza ..................... 77
Simulating Potato Late Blight in the Highland Tropics
G.A. Forbes, M.G. Chacn, M.V. Taipe, and R.J. Hijmans ......................................... 87
Oiscovery and Evaluation of a Valuable New Source of Resistance to
PLRV: Solanum tuberosum subsp. Andigena
E. Mihovilovich, L. Alarcn, A.L. Prez, and M. Bonierbale ...................................... 95
Assessment of Latent lnfection Frequency in Progeny Tubers of Advanced
Potato Clones Resistant to Bacterial Wilt: A New Selection Criterion
S. Priou, C. Salas, F. de Mendiburu, P. Aley, and L. Gutarra ..................................... 105
Evaluation of Bt-cry1/a1 (cryV) Transgenic Potatoes on Two Species of
Potato Tuber Moth, Phthorimaea opercu/ella and Symmetrischema
tangolias (Lepidoptera: Gelechiidae) in Peru
A. Lagnaoui, V. Caedo, and D.S. Douches ............................................................. 117
Assessment of the lnactivation Time of Phthorimaea opercu/ella Granulovirus
(PoGV) at Oifferent lntensities of Natural lrradiation
M. Sporleder, O. Zegarra, J. Kroschel, J. Huber, and A. Lagnaoui ............................ 123
lntegrated Control of Potato Bacteria! Wilt in Kabale Oistrict, Southwestern
Uganda
B. Lemaga .................................................................................................................. 1 29
Optimization of Sample Size for the Oetection of Latent lnfection by Ralstonia
solanacearum in Potato Seed Tubers in the Highlands of Peru and Bolivia
S. Priou, R. Torres, A. Villar, J. Martinez, O. Barrea, L. Gutarra, and
F. de Mendiburu ......................................................................................................... 143
CIP Program Report 1999-2000

Mass-Producing Beauveria brongniartii lnoculum, An Economical, Farm-Level


Method
F. Cisneros and A. Vera .............................................................................................. 155

Effectiveness of Abamectin and Plant-Oil Mixtures on Eggs and Larvae of the


Leafminer Fly, Liriomyza huidobrensis Blanchard
N. Mujica, M. Pravatiner, and F. Cisneros ................................................................ 161

Establishment of Microsatellite Assays for Potato Genetic ldentification


M. Ghislain, F. Rodrguez, F. Villamn, J. Nez, R. Waugh, and
M. Bonierbale ............................................................................................................ 167

Clonal True-to-Type Verification of Potato Accessions Retrieved from In Vitro


Conservation and Cryopreservation
G. Perazzo, A. Panta, F. Rodrguez, R. Gomez, J. Toledo, Z. Huamn,
M. Ghislain, A. M. Golmirzaie, and W. Roca ........................................................... 175
Native Potatoes: Potential Markets Outside the Andes ................................................ 1 84
Assessment of Variability for Processing Potential in Advanced Potato
Populations
W. Amaros, J. Espinoza, M. Bonierbale ..................................................................... 185
Selection of Parental Unes Using Stability Analysis of Hybrid True Potato Seed
Families Produced Through Line x Tester Method
M. D. Upadhya and R. Cabello ................................................................................. 197
lnfluence of Seed Size and Density on the Performance of Direct Seedling
Transplants from Hybrid True Potato Seed
M. D. Upadhya and R. Cabello ................................................................................. 207

Economic Returns to Research on True Potato Seed in Vietnam


K.O. Fuglie, N.T.B. Do, C.H. Dao, and H.T. Nguyen ............................................... 211

CIP's Contribution to Varietal Change in Potatoes in Developing Countries


T. Walker, Y.P. Bi, J.H. Li, P.C. Gaur, and E. Grande ................................................. 219

Evaluation of Potato Genotypes Through Pilot-Scale Farmer Field Schools


in the Peruvian Andes
R. Orrego, J.J. Heath, J. Tenorio, C. Valencia, J.A. Landeo, M. Upadhya,
K.A. Garrett, N. Zuniga, A. Mendoza, A. Parraga, L. Cotrina, E. Roncal,
M.A. Pacheco, O. Ortiz, and R.J. Nelson ................................................................. 225

An lntegrated Approach for Potato Production lmprovement in Nepal ....................... 238


A Report on Strengthening Farmer Capacity for Growing a Healthy Potato Crop
in Nepal
O.A. Hidalgo, D.M. Campilan, and T.L. Lama ......................................................... 239
A Report on Informal High Quality Seed-Potato Production and Marketing in Nepal
D.N. Ojha, O.A. Hidalgo, and T.L. Lama ................................................................. 245

Comparative Performance and On-Farm Profitability of Certified Potato Seed


in the Highlands of Bolivia
A. Devaux, S. Gonzales, C. Bejarano, R.Casso, V. Suarez, and T. Walker .............. 251

Toward Alleviating Poverty of Rural Potato Farmers by Strengthening the


Potato Seed System in Bangladesh: A Rapid Rural Appraisal
S.G. llangantileke, M.S. Kadian, M. Hossain, A.E. Hossain, U. Jayasinghe,
and A.A. Mahmood ................................................................................................... 259

Research on Sweetpotato
Transgene Expression of Rice Cysteine Proteinase lnhibitors for the Development
of Resistance against Sweetpotato Feathery Mottle Virus
G. Cipriani, S. Fuentes, V. Bello, L.F. Salazar, M. Ghislain, and O.P. Zhang ........... 267

Plant Productivity and Water Use Efficiency of Sweetpotato (/pomoea batatas)


as Affected by Nitrogen Supply
M. Kelm, H. Brck, M. Hermann, and B. Sattelmacher ........................................... 273

Effect of GxE lnteraction on Root Yield and Beta-carotene Content of Selected


Sweetpotato (lpomoea batatas (L) Lam.) Varieties and Breeding Clones
K. Manrique and M. Hermann ................................................................................... 281

Using Competing Traits to Select Dual-Purpose Sweetpotato in Native Germplasm


C.U. Len-Velarde ..................................................................................................... 289

Microsatellite Analysis of Genetic Diversity in Sweetpotato Varieties from Latin


America
O.P. Zhang, O. Carbajulca, L. Ojeda, G. Rossell, S. Milla, C. Herrera,
and M. Ghislain ......................................................................................................... 295
A Genetic Linkage Map of Sweetpotato (lpomoea batatas (L.) Lam.) Based on

AFLP Markers
A. Kriegner, J.C. Cervantes, K. Burg, R.O. Mwanga, and O.P. Zhang ....................... 303

From Latin America to Oceania: The Historie Dispersa! of Sweetpotato


Re-examined Using AFLP
G. Rossel, A. Kriegner, and O.P. Zhang ..................................................................... 315

Global Distribution of Sweetpotato


R.J. Hijmans, L. Huaccho, and O.P. Zhang ............................................................... 323

Beyond Higher Yields: The lmpact of Sweetpotato lntegrated Crop Management


and Farmer Field Schools in Indonesia
E. van de Fliert, N. Johnson, R. Asmunati, and Wiyanto .......................................... 331

The Sweetpotato Production-Postharvest Use System in Vietnam: Participatory


Needs and Opportunity Assessment
E. van de Fliert, N.T.K.Oanh, N.T. Son, T.O.Hoa, P.O.Thanh, L.Q. Hung,
L.H. Khoang, and N.T. Lan ........................................................................................ 343

Fermented Sweetpotato Feed for Pigs: A Boon to Farmers .......................................... 3 5 2

Natural Resource Management


Carbofuran Presence in Soil Leachate, Groundwater, and Surface Water in the
Potato Crowing Area in Carchi, Ecuador
R. Jaramillo, W. Bowen, and J.J. Stoorvogel .............................................................. 355

Toward A Dynamic Definition of Agroecological Zones Using Modern lnformation


Technology Tools
R. Quiroz, P. Zorogasta, G. Baigorria, C. Barreda, R. Valdivia, M. Cruz,
and J. Reinoso ............................................................................................................ 361

Estimating the Spatial Variability of Weather in Mountain Environments


G. Baigorria, W. Bowen, and J. Stoorvogel ............................................................... 371

CIP Program Report 1999 - 2000

Andean Roots and Tubers and other Crops


Effect of Viruses UMV, UVC, PapMV-U, and PLRV on Ulluco Production and
Their Control
C. Lizrraga, M. Santa Cruz, G. Lpez, and S. Fuentes ........................................... 381
Compositional Changes in Oca Tubers Following Postharvest Exposure to Sunlight
M. Hermann and C. Erazo ......................................................................................... 391
lmpact of Downy Mildew on the Yield of Quinoa
S. Danielsen S.E. Jacobsen, J. Echegaray, and T. Ames ............................................ 397
Quinoa: An Alternative Crop for Saline Soils in the Andes
5.-E. Jacobsen, H. Quispe, and A. Mujica ................................................................ 403

Enriching the Portfolio: CIP's Global and Regional Partnerships


CONDESAN: Watershed Use Planning ........................................................................ 411
Putting Natural Resource Management on the Map: Using GIS as a Tool for Soil
Conservation Planning in Two Small Andean Watersheds
Joshua Posner, Caen Bussink, Robert Hijmans, Jorge de la Cruz, Has Willet,
Pablo Arturo Sanchez ................................................................................................ 413
Land-use Change in the Cajamarca Catchment, Peru, 1975-1996
C. Bussink and R.J. Hijmans ...................................................................................... 421
GILB: An Update ........................................................................................................... 429
SIUPA: A New lnitiative ............................................................................................... 431
Multi-Sectoral lnitiatives for Urban Agriculture in Metro Manila, Philippines
D. Campilan, R. Boncodin, and C. de Guzmn ........................................................ 433
Peri-Urban Milk Production in Peru: Farm Strategies and Policy Options
T. Bernet, C. Gomez, J. Ju lea, G. Prain, J. Senz ...................................................... 445
Agro-processing Waste Assessment in Peri-urban Hanoi
Dai Peters, Do Duc Ngai, and Dang Thi An ............................................................. 451
Acronyms and Abbreviations ........................................................................................ 4 5 9
Author lndex ................................................................................................................. 465
Selected Publications by CIP Staff from 1999-2000 ..................................................... 471
Global Contact Points ................................................................................................... 478

Program Overview
Scientist and Farmer. Partners in Research far the 2 7st
Century. The title of the 1999-2000 Program Report
exemplifies not only the scope of CIP's research, butthe
philosophy behind it. As a Future Harvest Center (see
note, page 9), CIP's overarching goals are to reduce
poverty, ensure food security, and protect natural resources. Reaching those objectives depends on
developing and testing new technologies in the lab and
in the field. However, it is only after successful application and adoption in the fields of resource-poor farmers
around the world that the benefits of those technologies
can be seen and felt. Scientist and farmer partnerships
are helping us to accelerate the adoption and application of new technologies.

We learn then we
share Modern global

The demand for root and tuber crops to ensure food


security and contribute to sustainable livelihoods in
developing countries is predicted to grow steadily over
the next decades. That demand, as well as increased
pressure on natural resources from other animal and
crop systems, makes CIP research critically important,
and increases the urgency for technologies to move
quickly from the scientist to the farmer.

Broad-based Research
During the two years of research included in this Program Report, CIP carried out its program through 17
research projects addressing the most pressing needs of
developing country farmers (see page 11 for project
overviews). With a mandate encompassing potatoes,
sweetpotatoes, Andean root and tuber crops, and natural resource management, CIP has much to report.
In the fol lowing pages, there are research reports ranging
from the application of highly complex molecular
technologies to identify plant defense genes for late
blight resistance, to the development and evaluation of

information systems
(GIS) use sophisticated techniques that
link satellite information and computer
models with many
research disciplines.
A series of research
projects reported here
show the practica!
side of this type of
work. Qui roz et al.
(page 3 61 ) report on
how GIS can be used
to define agricultura!
zones in the
Cajamarca, Peru,
watershed. Baigorria
et al. (page 371) write
about weather in that
mountain environment and Bussink and
Hijmans take a look at
land use change there
from 1975-96 (page
421 ). Finally, Posner
et al. (page 413)
present the practica!
application of this
combined knowledge
when they report on
CONDESAN-in itiated
soil conservation
planning in two
communities in the
same watershed.
Researchers, NGO
workers, and commun ity leaders worked
together to consider
land use options.

CIP Program Report 1999 - 2000

more efficient pig feed from sweetpotatoes. The former


links us with advanced research laboratories in countries such as Germany, Luxembourg, and the USA; the
latter links us with women farmers in the rice paddies of
Vietnam. During the last few years, CIP has made significant efforts to accelerate the flow of research results
from those advanced lab studies to applications on in
the field through closer linkages with farmers in early
phases of testing and evaluation. Farmers learn more
about the underlying causes of the problems they face,
and are equal partners in testing possible solutions.

Building Partner Skills

Practica! help against


pests A cottage
industry in the
making is reported by
Vera and Cisneros
(page 155). They
describe how to make
inoculantto kill
Andean potato weevi 1
using laivae of the
weevi 1 itself that is
infected with
B. brongniartii. No lab
inoculum is needed.
The benefits of the
bioi nsecticide
abemecti n are more
available for farmers
as a resu lt of research
on improving the
potency of abemectin
by adding plant oil.
The cost of effectively
treating 1 ha can be
reduced by more
than 60%-from
$1 31 for abamectin
alone to $54 for the
mixture of abamectin
and plant oi 1 (see
Mujica and Cisneros,
page 161 ).

Program OVerview

As CIP celebrats its 30th anniversary in 2001, it brings


with ita rich history in training and capacity building
(see page 16). More than 20,000 developing country
scientists have benefited from a range of CIP training
activities, 2200 in 1999-2000 alone, not including
2599 participants in farmer field schools. During the
past two years, training has taken on new and exciting
dimensions as CIP has concentrated on building its own
capacity to deliver information and technology through
technology assisted learning (or 'distance learning').
Revolutions in information technology and delivery
systems have created a powerful tool to further increase
the center's ability to reach the remotest corners of the
globe.

Enriching CIP's Research Through Regional


and Global Partnerships
A global trend towards plurality in determining research
priorities and focusing diverse research efforts on large
problem areas has moved CI P more and more towards
equitable partnerships for research and development.
In this volume, reports from three regional or global
partnerships are featured: CONDESAN (Consortium for
the Sustainable Development of Natural Resources in

the Andes), GILB (Global lnitiative on Late Blight), and


SIUPA (Strategic lnitiative on Urban and Periurban
Agriculture). Each partnership reflects the direct involvement of a large constituency of global or regional
partners from all research sectors-public, prvate,
NGO, communities, farmers-all working toward mutually agreed upon objectives. CIP believes that such
broad partnerships will increase in the future, with more
and more of the center's research being carried out in
the context of participatory planning and implementation.
CIP's research and training activities during 1999-2000
have significantly contributed to meeting the goals of
the center. The results reported in this Program Report
truly exemplify the continued quality and relevance of
CIP research in that effort-scientist and farmer working
together to i mprove 1ives.

Wanda Collins
Deputy Director General for Research

Teaching farmers
from ali walks of life
"The project [FFS in
Nepal] demonstrated
that one does not
have to be amale,
1iterate farmer from
the dominant ethnic
group to participate
meaningfully in
agricu ltu ral research
and extension. Of
those who successfu 1ly completed the
FFS, one out of four
was i 11 iterate, one out
of three was female. In
addition, 7 of the 9
FFSs had participants
coming from multiple
ethnic groups.
Average knowledge
gain on patato seed
and disease management was 84%,"
report Hidalgo,
Campilan, and Lama.
(See page 239).

Future Harvest builds awareness and support for food and environmental research for a world with
less poverty, a healthier human family, well-nourished children, anda better environment. Future
Harvest supports research, promotes partnerships, and sponsors projects that bring the results of
research to rural communities, farmers and families in Africa, Latin America, and Asia. lt receives its
principal funding from a group of governments, private foundations, and the CGIAR.

CIP Program Report 1999 - 2000

CIP's Directors
Director General, Hubert Zandstra
Deputy Director General for Research, Wanda Collins
Deputy Director General for Finance/Administration, Hector Hugo Li Pun
Director for lnternational Cooperation, Roger Cortbaoui

CIP's Research Program


CIP's research program comprises 17 projects that address the most pressing constraints
to improving livelihoods through patato and sweetpotato production and utilization,
managing natural resources in mountain ecosystems, and preserving and exploiting
underutilized Andean root and tuber crops.
These projects constitute a constraints-driven research agenda that is implemented
through regional strategies in all relevant production areas of the world. Regional
implementation involves close partnership with national programs; universities; local,
regional, and international nongovernmental organizations (NGOs); farmers and farm
organizations (through farmer field schools); and the private sector. Close association
with advanced research institutions is also an important component, and a necessity to
bring the most advanced research and other tools to bear on the constraints.

CIP's Research Projects and Project Leaders


Project

15

lntegrated control of late blight


lntegrated control of bacteria! wilt
Control of potato viruses
lntegrated management of potato pests
Propagation of clonal potato planting materials
Sexual potato propagation (TPS)
Global sector commodity analysis and impact assessment for
potato and sweetpotato
Control of sweetpotato viruses
lntegrated management of sweetpotato pests
Postharvest uti 1ization of sweetpotato
Breeding sweetpotato for high-dry-matter yield and adaptation
Strategic lnitiative on Urban and Peri-urban Agriculture
Sustainability of rice-based cropping systems featuring patato
as a cash crop
Sustainable land use in the Andes
Conservation and characterization of patato genetic resources

16

Conservation and characterization of sweetpotato

17

Conservation and characterization of Andean root and


tuber crops

2
3
4

5
6
7

8
9
10
11
12

13
14

genetic resources

1O

CIP's Research Program

Leader
R. Nelson
S. Priou
L. Salazar
A. Lagnaoui
M. Bonierbale
E. Chujoy
T. Walker
L. Salazar
E. van de Fliert
O.P. Zhang
O.P. Zhang
G. Prain
T. Walker
R. Quiroz
W. Roca
M. Hermann
M. Holle

CIP's Research Project Portfolio


Project 1
lntegrated control of late blight
(R. Nelson (until December 2000) /
J. Landeo)
CIP's highest research priority is to develop, adapt, and integrate technologies
for managing late bl ight of potato, the
world's worst agricultura! crop disease,
caused by the oomycete Phytophthora
infestans. CIP scientists use a range of
methods, including state-of-the-art biotechnological tools, to produce breeding
popu lations and advanced clones with
durable resistance to the disease. Additional component technologies are being
developed for integrated disease management under the conditions encountered by
resource-poor farmers in developing
countries. Geographic information systems
are linked with crop and disease models to
understand the complexities of the
disease's epidemiology across diverse
agroecosystems.

Project 2
lntegrated control of bacteria! wilt
(S. Priou)
Bacteria! wilt (BW), caused by Ralstonia
solanacearum, is the second main potato
disease in the world. The key to overcoming the constraint is to control the spread
of the disease by using only healthy
planting material. CIP's research therefore
concentrates on developing tools to detect
the presence of bacteria in soil and tubers.
This knowledge is used to improve crop
management and seed systems and to
identify available resistant/tolerant
material. The project also aims at designing, validating, and promoting integrated
strategies for managing bacteria! wilt in
different production systems and complet-

ing the selection of potato progenies with


tolerance or resistance to BW developed
in past years. Main components for BW
management are seed health, plant
resistance, crop rotation, soil fertility, and
biological control.

Project 3
Control of potato viruses
(L. Salazar)
Virus diseases cause serious losses in
potato and also disrupt global efforts to
improve potato, because national and
inter-national regulations control the
movement of virus-infected seed and
genetic resources. Biotechnology tools are
used to identify resistance genes in related
plant species and to clone and transfer
them to potato; these tools are combined
with traditional methods of breeding to
develop adapted cultivars resistant to a
range of viruses. ldentifying and characterizing the most important viruses and
virus-like agents that affect potato are
essential steps toward developing sensitive, low-cost methods for large-scale
detection. CIP researchers also study
epidemiological factors that affect virus
spread, with particular attention to interaction between viruses and other pathogens
that may affect plant resistance response,
and train national scientists in virus
identification, detection, and control
techniques. Research and training activities focus on the most important potato
viruses (PLRV and PVY) and on the potato
spindle tuber viroid (PSTVd). Particular
attention is given to practica! utilization/
adoption of virus-resistant materials
already produced at CIP (including
genotypes already carrying combined
resistance to more than one virus, preferably in a multiplex condition).

CIP Program Report 1999 - 2000

11

Project 4

Project 6

1ntegrated management of potato pests

Sexual potato propagation (TPS)


(M. Upadhya (until June 2000) /

(A. Lagnaoui)
Key pests of potato, globally or regionally,
are three species of potato tuber moth (a
threat that is rapidly becoming more
serious), severa! species of Andean potato
weevil, the leaf-miner fly, whiteflies and
severa! flea beetles. Nematodes that
reduce potato yields and favor the development of bacteria! wilt pathogens are
potato cyst, rosary, and root-knot nematodes. This project seeks to develop
locally-adapted integrated pest management programs for these pests,
emphasizing sustainable, ecologicallybased, and economically-sound practices
that will lead to reduced use of chemical
pesticides and increased benefits for
farmers. Components include biological,
cultural, and resistance aspects of control.

E. Chujoy)
True potato seed (TPS) enables a crop to be
grown in areas where traditional production systems fail, for example, where seed
tubers are scarce or not available. By
facilitating the transfer of TPS technology
in such areas of the tropics and subtropics,
this project aims to expand patato cultivation and increase its efficiency (reduce
production costs, increase yields). CIP
concentrates on improving parents for
hybrid TPS production and developing
needed specific traits such as late-blight
resistance, earliness, and seed set. CIP's
work is backstopped by local organizations (private sector, NGOs, NARS) in
efforts to commercial ize TPS systems and
thus underpin developing small industries.

Project 5

Project 7

Propagation of clonal potato planting


materials (U. Jayasinghe (until June 2000)

Global sector commodity analysis and


impact assessment for potato and
sweetpotato (T. Walker)

/ M. Bonierbale)
In many countries, the lack of efficient
formal and informal seed potato systems
has limited the diffusion of new and
improved varieties because only limited
amounts of healthy clona! planting
material are available. Varietal introduction and diffusion is dependent on the
informal system, but it must be linked with
the formal system and it must emphasize
high quality planting material. This project
provides research and technical assistance
to selected formal and informal seed
systems in various countries to help them
improve their efficiency and effectiveness.
This is accomplished through farmer
training and establishment of pilot seed
systems. The project also explores innovations in linkages between formal and
informal seed systems, aiming to speed up
varietal introduction and diffusion.

12

CIP's Research Project Portfolio

The objective of this project is to provide


more complete information to scientists,
research administrators, policy-makers,
and donors for decision-making on technology design, resource allocation, policy
formulation, and investment options
related to patato and sweetpotato improvement and utilization. Sorne of the specific
objectives are to: quantify the agronomic,
economic, social and environmental
effects of improved patato and
sweetpotato technologies; document the
rate of return and the effect on poverty of
CIP's research; assess the leve! and
adequacy of investment in potato and
sweetpotato crop improvement in developing countries; assemble and maintain price
and production databases for priority
setting; evaluate the effects of patato price
instability on diverse groups in society;
assist in improving domestic patato and

sweetpotato marketing and international


potato trade benefiting developing countries; and participate in generating the
most informative commodity projections
with special ized institutions.

problem-solving skills. IPM training


programs are designed, and their implementation by governmental and
nongovernmental organizations faci 1itated,
through the training of trainers and provision of sweetpotato IPM information
material s.

Project 8
Control of sweetpotato viruses
(L. Salazar)

Virus diseases greatly reduce sweetpotato


yields worldwide, particularly in SubSaharan Africa. Control can be achieved
by using healthy planting materials (the
use of virus-free planting materials alone
can triple yields) and developing resistant
cultivars. ldentification of viruses and
development of sensitive methods of
detection are fundamental steps toward
this end. Previous work has shown that a
synergy between sweetpotato feathery
mottle virus (SPFMV) and sweetpotato
chlorotic stunt virus (SPCSV) causes the
sweetpotato virus disease (SPVD) that can
devastate crops. The project aims to
identify the viruses causing major losses in
production, develop methods of detection,
and apply methods of control. Researchers
seek to develop resistance to SPVD using
a range of breeding approaches, including
the most advanced molecular methods.

Project 9
lntegrated management of sweetpotato
pests (E. van de Fliert)
The aim of this project is to develop IPM
(integrated pest management) systems for
sweetpotato. These systems need to be
compatible with farmers' crop management practices, as well as with prevailing
ecological and socioeconomic conditions,
to ensure effective and sustainable solutions. Therefore, participatory approaches
are applied to prioritize research needs,
develop adapted pest and crop management components, and design learning
strategies which enhance farmers' ecological knowledge and decision-making and

Project 1 O
Postharvest utilization of sweetpotato
(G. Scott (until August 2000)/D.P. Zhang)
This project studies technologies to
improve the livelihoods of rural poor
through diversification and expansion of
sweetpotato use. The main beneficiaries
are women and children and small households. Nutrition and income are improved
and poverty is reduced. Project goals
include facilitating the development of
small enterprises based on added-value
from primary processing (e.g., starch and
flour), and the more efficient use of
sweetpotato roots, vines, and by-products
as animal feed. In Africa, the goal is to
enhance food security by taking advantage of sweetpotato's nutritional qualities.
CIP researchers evaluate opportunities and
undertake collaborative research on
markets, raw material quality, process
development, product qual ity, and the
social acceptability of innovation in pilot
enterprises. They tap such resources as
NARS, NGOs and users in target countries,
along with global centers of research
excellence in disciplines not available inhouse, such as food science/technology
and animal sciences.

Project 11
Breeding sweetpotato for high-dry-matter
yield and adaptation (D.P. Zhang)
This project aims to improve sweetpotato
production and use through the development and adoption of high-dry-matter/
high-starch varieties with adaptabil ity to
low-input, subsistence farming systems.
The diverse sweetpotato germplasm at CIP

CIP Program Report 1999 - 2000

13

is used to generate high-dry-matter parental clones through population breeding. A


well established, decentralized breeding
framework uses these advanced parental
clones to produce new varieties with a
broader genetic background and good
adaptabi 1ity to cope with abiotic and
biotic stresses in target environments.
Molecular approaches are applied to
develop, expand and efficiently use the
genetic variations in arder to meet breeding needs. The new clones are being
adapted to low-input subsistence systems
in target environments to feed into the
developing markets. The project therefore
provides the raw material for increase in
the use of both fresh and processed
sweetpotato. Dry matter is the essential
component in both types of use.

Project 12
Strategic lnitiative on Urban and Periurban Agriculture (G. Prain)
The Strategic lnitiative on Urban and Periurban Agriculture (SIUPA) was launched
by the CGIAR in late 1999 in response to
growing urbanization and increased
dependence of city dwellers on farming:
CIP is the convening center for the initiative. SIUPA's goals are to contribute to
increased food security, improved nutritional status, and higher incomes of urban
and peri-urban farmers; reduce the negative environmental impact of urban and
peri-urban agriculture and enhance its
positive ecological potential; and establish
the perception of urban and peri-urban
agriculture as a positive, productive, and
essential component of sustainable cities.
In collaboration with the many national
and international efforts that have started
in recent years to address the issue of
urban and peri-urban agriculture, SIUPA is
establishing, in regional sites, a set of
research activities collectively known as
Urban Harvest. Several international
agricultura! research centers are already
working in such areas as technology and
policy aspects of urban agriculture and

14

CIP's Research Project Portfolio

nutrition, enhanced efficiency and


sustainability of peri-urban vegetable
production systems, health impacts of the
use of urban waste water in agriculture,
and the development of sustainable periurban agro-processing and livestock
enterprises.

Project 13
Sustainability of rice-based cropping
systems featuring potato as a cash crop
(T. Walker)
In response to increasing land scarcity in
subtropical South and Southeast Asia,
potatoes with a high production potential
per unit time are increasingly being
planted in intensive sequential cropping
systems, such as rice-patato-rice, and in
more intensive intercrops, such as ricepotato/maize. Typically, rice is planted at
the onset of the rainy season, irrigated
patato is sown in the cooler dry season,
and irrigated rice or another crop is
cultivated in the hot summer. To realize
the potential of the patato crop, factors
that threaten the sustainability of these
input-responsive cropping systems must be
identified and addressed. This project
diagnoses constraints to increasing and
maintaining productivity in selected
potato and rice-based cropping systems;
and generates crop and natural resource
management information on how to
alleviate the most important of those
constrai nts.

Project 14
Sustainable land use in the Andes
(R. Quiroz)
The Andes comprises a series of unique
habitats rich in natural resources. The
inhabitants of this region confront massive
poverty, increasing population growth, and
rapid degradation of the natural resource
base. They face the difficult challenge of
trying to increase agricultura! productivity
while simultaneously decresing stress on

the environment. This project aims to


characterize the Andean ecoregion for its
potential for sustainable agriculture, and to
provide a scientific, technical, and
economic base for policy and technology
recommendations to decision-makers in
the region. lt also seeks to develop
innovative methodologies for ecoregional
research through an effective integration
of process-based crop growth models,
remate sensing, economic decision
models, and geographic information
systems. Through short-term training and
collaborative research, the project aims to
build an international community of
researchers working toward sustainable
development of agriculture in mountain
a reas.

resource-poor sweetpotato producers and


users. Specifically, it seeks to develop
cost-efficient and sustainable conservation
methods, to contribute to the understanding of biogeographic and genetic patterns
of diversity, and to document and disseminate traditional and technical knowledge
on sweetpotato genetic resources worldwide. Expected impacts include the
enhanced use of germplasm in breeding
and varietal testing, improved international task and resource sharing for
sweetpotato conservation, and reduced
genetic redundancy of sweetpotato
holdings.

Project 17
Conservation and characterization of
Andean root and tuber crops (M. Halle)

Project 15
Conservation and characterization of
potato genetic resources
(Z. Huamn (until June 2000)/W. Roca
CIP holds the most comprehensive collection of germplasm of wild and cultivated
potatoes in the world. Key objectives of
this project include safe, long-term
conservation and characterization of the
germplasm, and developing a database
containing all information on the collection and making this information available
to interested parties. Research attempts to
in crease uti 1ization of the pota to genetic
diversity by identifying key desirable traits
and distributing healthy seed stocks and
clonal materials throughout the world for
use in patato improvement programs. The
project provides key input into CIP's own
breeding efforts.

Project 16
Conservation and characterization of
sweetpotato genetic resources
(M. Hermann)

Assisting national programs in rationalizing strategies for both ex situ and in situ
conservation of Andean root and tuber
crops (ARTC) involves the study and
preservation of biodiversity, with emphasis
on four priority genera - Oxalis (ocat
Ullucus (ulluco), Canna (achira), and
Arracacia (arracacha), including wild
species-and on the material of Mirabilis
expansa (mauka), Pachyrhizus ahipa
(ahipa), Smallanthus sonchifolius (yacon),
Tropaeolum tuberosum (mashua), and
Lepidium spp (maca). The project is
systematically assessing the potential of
ARTC to promote wider use in the subtropical and tropical highlands, within and
outside the Andean region, through the
study of current marketing and consumption patterns. lt aims to identify latent
demands these crops may satisfy in the
future, and to produce healthy planting
materials for farmers. This project is
perhaps the only one in the world with a
majar effort towards developing virus
identification and eradication procedures
for these important, underutilized crops.

The overall objective of this project is to


safeguard sweetpotato genetic resources
and to facilitate their use for the benefit of

CIP Program Report 1999 - 2000

15

Building Partner Skills


Throughout CIP's human-resource-building program, we use a dynamic,
participatory training method. This method combines exposure to cutting
edge advanced techniques with hands-on experience, emphasizing procedures that are directly applicable to working conditions in developing
countries. This method enhances active participation and information exchange and aims to expand the capabilities of partner organizations and
individuals and strengthen collaboration with them. This results in faster
research progress, identifies those elements that need further strengthening
at the national level, and gives the opportunity to solve constraints of common interest.

CIP's training program plays an essential


role through its integration with our
center's research agenda. Project objectives and priorities are directly linked with
our training program. The majority of CIP's
partners are trained in five general thematic areas that mesh well with CIP's
research emphases (Figure 1).
During 1999-2000, we expanded our
training base by adding distance learning
technologies which have resulted in the
incorporation of new instructional materials like interactive CDROMs in specific
subject matters for use in training-oftrainers, the use of video conferencing for
expanding the technical content of our
cou rses and workshops, and the maki ng
available of on-line manuals and webbased information for use in courses most
of which are accessible cost free through
the Internet. Individual and group training
was offered to more than 21 94 people from
75 countries-2140 participated in group
training courses and 54 in individualized
training programs. The numbers keep
growing: As of 2000, more than 20,000
people have received training from CIP.

Training at Headquarters

Individual
A large part of individualized training took
place on-site in CIP's labs and greenhouses

16

Building Partner Skills

where partners gained access to the skills


they needed to conduct more effective
research (Table 1). The most popular study
topics in 1999=2000 were integrated pest
management and seed production techn iques (Figure 2). Pest management
training included both field and molecular
techniques for identifying and controlling
pests.

Group training
Group courses cover all CIP's research
areas and concentrate on transferring
methodologies (Table 2, Figure 3), whereas
workshops focus on planning, information
exchange, and discussions on the progress
of ongoing collaborative activities. A
popular group training topic in 1999-2000
was biodiversity in relation to Andean root
and tuber crops. Several workshops were
held on integrated management of the
most destructive pests in potato and
sweetpotato. Others were held to facilitate
progress for activities supported by
FONTAGRO, (funded by IDB), IFAD,
DFID, UNEP, IDRC, SDC, and other
organizations. A number of courses and
workshops were organized in collaboration
with our regional headquarters' staffs and
held away from CIP headquarters.
As our partners improve their skills they
become future col laborators and contact
points in their home country, thereby

D
17%

39%

12%

lntegrated pest and disease


management
Seed production techniques

Conservation and management of


genetic resources

Postha rvest qual ity, nutrition, and


marketing of roots and tubers

Natural resource management


Other tapies including social
science, molecular techniques,
information systems

Figure 1. An overview of group and individual training by thematic area.

lntegrated pest and disease


management

Seed production techniques

Conservation and management of


genetic resources

Postharvest quality, nutrition, and


marketing of roots and tubers

Other tapies including socia l


science, molecular techniques ,
information systems

lntegrated pest and disease


management

Seed production techniques

17%

20%

Figure 2. An overview of group and individual training by study tapie.

17%

D
D
17%

Conservation and management of


genetic resources
Postharvest quality, nutrition, and
marketing of roots and tubers

Natural resource management

Other tapies including social


science, molecular techniques,
information systems

11 %

Figure 3. An overview of group training by thematic area.

CIP Program Report 1999 - 2000

17

Table 1. lndividualized training far participants at CIP headquarters.


Emphasis
Home country
Home institute
lntegrated pest and disease management
Virology
Peru (2)
Universidad Nacional Hermilio Valdizn (1) and
Polica Nacional del Per (1)
Phytopathology
Colombia
Universidad Pedaggica y Tecnolgica de
Colombia
Virus detection
Buhtan
Virus, viroid, and bacteria! wilt detection Peru
Molecular Biotechnology Research Center
Shandong, Academy of Agricultura! Science
Kenya
ARC/Roodplaat in South Africa
Virus elimination and detection
Bacteria, fungi, and nematode detection Peru
Universidad Nacional San Cristbal de
Huamanga
in patato and sweetpotato crops
Pinto y Gagardo S.A.
Virus diagnosis
Peru (3)
Molecular markers far Late Blight
Colombia (2) and Universidad Nacional de Colombia (2) and
Uruguay (1)
characterization
INIA (1)
Micropropagation, virus detection and Vietnam (3)
Patato lnstitute in Jang Jin; Academy of
control, integrated management of
Agricultura! Science, lnstitute of Agrobiology
majar patato pests and diseases,
and true patato seed
production
AFLP far Phytophtora intestans
Ecuador
Universidad Tecnolgica Equinoccial
characterization
Ralstonia solanacearum detection in
Colombia (1) and CORPOICA (1) and Universidad de
tubers using NCM ELISA
Argentina (1)
Tucuman (1)
Virology
Ecuador
FAO/CIP Postharvest Project
Virology
China
Academy of Agricultura! Science
1PM of patato
lran
Plant Pests and Diseases Research lnsfute
Virology
Mexico
Direccin General de Sanidad Vegetal (SAGAR)
Patato virus
Peru
Semilla Churcampa SAC.
Virus detection in Arracacha
Brazil
Empresa Brasileria de Pesquisa Agropecuaria
(Brazil)
Seed production techniques
Sweetpotato breeding using true seed Muan
Mopko Experiment Station, Rural Development
produced from open-pollinated plants
Administration
in tropical countries
Patato seed production
Chile (3)
Ingenieros Agrnomos, Ltda./ lnsitituto
Nacional de Investigacin Agraria
True patato seed
Peru (4)
Centro de Capacitacin y Promocin
Agroecolgico, Yungu
In vitro propagation far seed programs Costa Rica
Instituto Tecnolgico de Costa Rica
Assays and techniques on patato and Nicaragua
Centro Nacional de Investigacin Agropecuaria
sweetpotato propagation
- Instituto Nicaraguense de Tecnologa
Agropecuaria
True patato seed production
Peru
Universidad Nacional Jos Faustino Snchez
Carrin - Huacho
Conservation and management of genetic resources
Plant virology, breeding and germplasm China
Crop Research lnstitute, Shandong Academy
management

Morphologic characterization and


conservation in patato/patato IPM/
Virus diagnose
18

Building Partner Skills

of Agricultura! Sciences

Canary lslands,
Spain

Universidad Insular de Tenerife

Table 1. (continued)
Molecular techniques applied to patato Peru (2)
characterization
Molecular techniques applied to patato Spain
genetic diversity
Molecular techniques applied to
Peru (3)
germplasm characterization

lnsitituto Nacional de Investigacin Agraria


Universidad Complutense de Madrid

Estacin experimental San Roque, Programa


Nacional de Recursos Genticos y
Biotecnologa
Postharvest quality, nutrition, and marketing of roots and tubers
Analysis techniques in sweetpotato far Peru
Universidad Jorge Basadre
carotenoid content in leaves and
roots/determination of the starch
content in sweetpotato roots
Other
Mathematical models in biotechnology Peru
Universidad Nacional Agraria La Malina
Genetic lmprovement
India
Central Patato Research lnstitute
Bacteria! Wilt
Peru (3)
Universidad Nacional Pedro Ruiz Gallo
Molecular techniques
Peru
Universidad Nacional del Centro
Electronic information systems
Bolivia (1) and
PROINPA, Bolivia (1) and FOITTIPAPA,
Ecuador (1)
Ecuador (1)
Genetic lmprovement
Peru
Instituto Nacional de Ciencias Agrcolas
FAO = Food and Agriculture Organization of the United Nations; FORTIPAPA = Fortalecimiento de la Investigacin y
produccin de Semilla de Papa (Ecuador); INAGRO = Ingenieros Agrnomos, Ltda.; PROINPA = Funcacin para la
Promocin e Investigacin de Productos Andinos (Bolivia).

enriching their nation's human resource


base of future scientific researchers.

Taking it to the Field


Hundreds of individuals receive training
through CIP's efforts close to their home.
Moving into the field has entailed developing distance learning materials,
technology-assisted learning, regional
courses, CIP-facilitated training at other
institutions, and inter-center collaboration.
Academic training at the university level
also received continued support. Ten
partners received academic training
support at well-accredited universities
close to their home and workplace. This
allowed students to do their academic
work locally and their practica! work in
their own employment settings. Sorne
participating Universities were
Universidad Nacional Agraria La Molina
in Peru, Bogor University in Indonesia,
University of Nairobi, Kenya, Yunnan
University in China, and University of the
Philippines, Los Banos.

Distance learning materials emphasize the


development of interactive CDROMs to
aid in understanding the basic principies of
a particular topic. For example, the farmer
field school CDROM gives extensionists
and development agents a tool to enhance
their understanding of the processes
explained in CIP's field manual. In Peru,
NGOs have facilities close to FFS sites
where extensionists can use the CDROM
as they examine actual in-field conditions.
The CDROM on potato IPM (including
concepts, principies, and activities) is
designed for use in the office before going
into the field to work directly with farmers.
This course covers nine different pests and
how to identify and control them.
Video conferencing has allowed experts to
participate live from off-site as resource
persons on key subjects. This is a low cost,
convenient way of bringing in necessary
expertise and an excel lent complement to
the knowledge that is delivered through
the course.

CIP Program Report 1999 - 2000

19

Table 2. Group training.


Participant's home countries
Activity/venue
Part. (#)
lntegrated pest and disease management
Workshop/Peru
77
Bolivia, Colombia, Germany,
Ecuador, Mexico, Peru, USA,
Venezuela
Course/Bangladesh
Cou rse/Ch ina
Workshop/Peru
Course/Peru

11
15
19
7

Workshop/Ecuador

20

Meeting/Peru
Conference/Ecuador

20
163

CourseNietnam
Course/Kenya
Workshop/I ndia
Workshop/China
Workshop/China
Workshop/Peru
Course/Peru
Course/Cuba
Course/Peru
Workshop/Andean
countries
Course/Egypt
Course/Kenya

12
10
33
20
48

180
6
11
27
15

Bangladesh, India
China
Peru
Argentina, Bolivia, Ecuador, Mexico,
Peru
Bolivia, Colombia, Ecuador, Peru,
Venezuela
Peru
Argentina, Austria, Bangladesh,
Belarus, Bolivia, Brasil, Cameroon,
China, Colombia, Cuba, Costa Rica,
Czech Republic, Denmark, Ecuador,
England, France, Germany,
Honduras, India, Israel, Kenya,
Korea, Mauritius, Mexico,
Netherlands, Nepal, Norway,
Panama, Peru, Polony, Spain,
Sweden, Switzerland, Uganda,
Uruguay, USA, Venezuela
Vietnam
Kenya
India
China, Philippines, Peru
China
Peru
Bolivia
Cuba
Peru
Andean countries

20
6

Egypt
Burundi, Ethiopia, Cameroon,
Kenya, Uganda

Course/Kenya
Workshop/Bangladesh

9
15

Kenya, Rwanda, Uganda


Bangladesh

Workshop/I ndonesia

23

China, Indonesia, Nepal, Pakistan,


Philippines, Sri Lanka, Vietnam

20

Building Partner Skills

Co-sponsors1
CIP/Association far Andean
Technical-Cultural Promotion,
(Peru)/lnter-American Foundation
(USA)
Germany-1 mprov. ResisVUPWARD
lnternational Fund far Agricultura!
Development

FORTIPAPA, INIAP
Global lnitiative on Late Blight
Global lnitiative on Late Blight

UPWARD

Cinese Academy of Agric. Sci.


PRONAMACHCS/INIA
PNS-PRODISE
SENASA
Papandina, SDC

Programme Rgional de
l'Amlioration de la Culture de la
Pomme de Terre et de la Patate
Douce en Afrique Central et de l'Est
(CIP network)
PRAPACE
lnternational Fund far Agricultura!
Development
Users' Perspective with Agricultura!
Research and Development (CIP
network)/Swiss Development
Cooperation

Table 2. (continued)
Workshop/China
WorkshopNietnam

47
20

China
Indonesia, Vietnam

Conservation and management of natural resources


Workshop(fanzania
23
Burundi, Ethiopia, Gabon, Kenya,
Malawi, Namibia, Peru, Rwanda,
Tanzania, Uganda, USA, South
Africa, Zimbabwe, Zambia

Course/Ecuador
Course/Ecuador
Course/Peru
Cou rse/Swaziland

22
39

4
22

35

WorkshopNietnam
Course/Bangladesh
Workshop/Nepal
Workshop/Nepal
Workshop/Peru

85

Workshop/Peru

34

Course/Peru

21

11
16
12

Workshop/Kenya
15
Workshop/lndia
11
WorkshopNietnam
15
Postharvest quality, nutrition,
Workshop/Peru
18
Course/Kenya

11

Workshop/Kenya

31

Workshop/Cuba

16

Workshop/Peru

31

Workshop/Malawi

27

Global lnitiative on Late Blight


Food and Agricultura! Organization of
the United Nations
United States Agency far
lnternational DevelopmenV
Programme Rgional de
l'Amlioration de la Culture de la
Pomme de Terre et de la Patate
Douce en Afrique Central et de l'Est
(CIP network)
Biodiversity Project
Biodiversity Project
Biodiversity Project
SARRNET

Bolivia, Brasil, Ecuador, Peru, USA


Bolivia, Ecuador, Peru, USA
Bolivia, Ecuador, Peru
Bostwana, Lesotho, Mozambique,
Malawi, Namibia, Swaziland,
Tanzania, South Africa, Zambia,
Zimbabwe
India, Philippines, Vietnam
UPWARD
Bangladesh, Philippines
Butan, Nepal, Pakistan, Sri Lanka
Swiss Development Cooperation
Nepal
Swiss Development Cooperation
Argentina, Bolivia, Checoslovaquia, Royal Danish Ministry of Foreign
Chile, Colombia, Denmark, Ecuador, Affairs
France, Greece, Peru, Poland,
Spain, Venezuela
Bolivia, Ecuador, Mexico, Peru, USA United Nations Environmental
Program
Argentina, Bolivia, Brazil, Colombia, Swiss Development Cooperation
Ecuador, Peru
Kenya
Kenya Agricultura! Research lnstitute
Bangladesh, India, Sri Lanka
CTCRl/ICAR
Vietnam
Government of Netherlands
and marketing of roots and tubers
Bolivia, Ganada, Colombia, Ecuador, Desarrollo de agroindustrias y
France, Peru
mercados para la arracacha project
Burundi, Ethiopia, Kenya,
USAID, PRAPACE
Madagascar, Rwanda, Uganda
Burundi, Cameroon, Ethiopia,
Programme Rgional de
Honduras, Kenya, Madagascar,
l'Amlioration de la Culture de la
Malawi, Peru, Rwanda, Sudan,
Pomme de Terre et de la Patate
Douce en Afrique Central et de l'Est
Tanzania, Uganda, Zaire
(CIP network)
Bolivia, Colombia, Cuba, Ecuador, PRECODEPA, ALAP
Guatemala, Mexico, Netherlands,
Peru
Bolivia, Brazil, Ecuador, Peru,
CONDESAN/CllD/IDRC
Venezuela
Angola, Lesotho, Malawi,
Southern Africa Root Crop Research
Mozambique, Namibia, Swaziland, Network
Tanzania, Zambia, Zimbabwe
CIP Program Report 1999 - 2000

21

Table 2. (continued)
Activity/venue

Part. (#)

Workshop/Malawi

15

Workshop/Kenya

22

Participant's home countries

Co-sponsors1

Tanzania and SARRNET countries

Southern Africa Root Crop Research


Network
Burkina Faso, Burundi, Cameroon, PRAPACE/FOODNET/ICRAF/
Kenya, Madagascar, Mali, Rwanda, ECABREN/AFRENNEARRNET/ILRI/
DR Congo, Uganda
SARDl-UMCOR

Natural Resource Management


Argentina, Bolivia, Ganada, Chile,
Workshop/Chile
20
Peru
Meeting/Peru

42

Workshop/Peru

40

Workshop/Peru
Workshop/Colombia

8
18

WorkshopNietnam

30

Other
Workshop/Peru

32

Meeting/Ch in a
Workshop/Peru
Course/Brazil

50
12
36

Meeting/Peru

23

Meeting/Uganda

40

Meeting/Malawi

19

Course/Peru

31

Course/United
Kingdom

37

Course/Peru

25

22

Building Partner Skills

Mountain Forum, Consortium far the


Sustainable Development of the
Andean Ecoregion
Andorra, Austria, Ganada,
Mountain Forum, Consortium far the
Switzerland, France, England, ltaly, Sustainable Development of the
Kenya, Kyrgyzstan, Madagascar,
Andean Ecoregion
Mexico, Nepal, Peru, Philippines,
Poland, USA, Venezuela, South
Africa
Argentina, Bolivia, Chile, Costa Rica, Soil Management Collaborative
Ecuador, Mexico, Netherlands,
Research Support (USA)/ lnterPanama, Peru, Uruguay,
American lnstitute far Global
USA.Venezuela
Change Research, lnitial Science
Program (Brazil)
Peru
CONDESAN
Colombia, Costa Rica, Guatemala, lnternational Plant Genetic Resources
India, Kenya, Peru, Syria, Trinidad lnstitute
and Tobago
Indonesia, Laos, Philippines,
CIAT/FSP/PRGNUPWARD
Thailand, Vietnam
Bolivia, Colombia, Costa Rica,
Ecuador, Peru
China, Indonesia, Philippines
Peru
Brazil, Ecuador, Uruguay, USA

Biodiversity Project
Chinese Academy of Agric. Sci.

lnternational Fertilizer Deveopment


Center (USA)
Austria, Bangladesh, England, Italia, Consortium far the Sustainable
Kenya, Kyrgyzstan, Nepal, Peru,
Development of the Andean
South Africa, Switzerland, USA
Ecoregion, Food and Agricultura!
Organization of the United Nations
Burundi, Eritrea, Ethiopia, Kenya,
PRAPACE
Madagascar, Rwanda, Sudan,
Tanzania, Uganda
Angola, Botswana, Lesotho, Malawi Southern Africa Root Crop Research
Network
Argentina, Bolivia, Colombia,
Federal Ministry of Economic
Ecuador, India, Luxembourg,
Cooperation and Development
Mexico, Peru, South Africa, Spain, (Germany)
Uruguay
Argentina, Bolivia, China, Colombia, Department for lnternational
Cuba, India, Kenya, Mexico, Nepal, Development
Peru, South Africa, Uganda
Peru

Table 2. (continued)

Workshop/Brazil
Workshop/Ethiopia
Workshop/Ethiopia
Total

44
15
5

Brazil, Ecuador
Ethiopia
Burundi, Ethiopia, Kenya, Uganda

FAPESP/EMBRAPNCENA
lnternational Fund far Agricultura!
Development
1DRC/PRAPACE

2140

1 Where

no co-sponsor is given courses were wholly sponsored by CIP.


AFRENA = African Resource Network in Agro-Forestry, Uganda; ALAP = Asociacin Latinoamericana de la Papa. BMZ
= Federal Ministry of Economic Cooperation and Development; CENA = Civil Engineers Network Africa, South Africa;
CIAT = Centro lnternacinal de Agricultura Tropical; CllD = Centro lnternacinal de Investigaciones para el Desarrollo
(IRDC); CONDESAN = Consortium for the Sustainable Development of the Andean Ecoregion; CTCRl/ICAR Central Tuber
Crops Research lnstitute, lndia/lndian Council of Agricultura! Research DFID: Department for lnternational Development;
EARRNET = Eastern Africa Rootcrops Research Network Uganda; ECABREN = Eastern and Central Africa Sean Research
Network Uganda; EMBRAPA = Empresa Brasileira de Pesquisa Agropecuria (Brazil); FAPESP = Fundago de Amparo a
Pesquisa de Estado de So Paulo, Brazil; FORTIPAPA = Fortalecimiento de la Investigacin y produccin de Semilla de
Papa (Ecuador); FSP = Forages for Smallholders Project, (CIAT, Colombia); ICRAF = lnternational Center for Research in
Agroforestry; IDRC = lnternational Development Research Center (Ganada; IFAD = lnternational Fund for Agricultura!
Development; ILRI = lnternational Livestock Research lnsitute; INIA - Instituto Nacional de Investigacin Agraria (Peru);
INIAP = Instituto Nacional de Investigaciones Agropecuarias (Ecuador). PNS-PRODISE = Programa Nacional de Semillas
del Proyecto de Desarrollo Integral de Semillas (Peru); PRAPACE = Programme Rgional de l'Amlioration de la Culture de
la Pomme de Terre et de la Patate Douce en Afrique Central et de l'Est (CIP network); PRGA = Participatory Research and
Gender Analysis (CGIAR lnitiative); PRECODEPA = Programa Regional Cooperativo de Papa (Mexico); PRONAMACHCS =
Proyecto Nacional de Manejo de Cuencas Hidrogrficas y Conservacin de Suelos (Peru); SARDl-UMCOR = Sustainable
Agricultura! and Rural Development lnitiative-United Methodist Committee on Relief, (D.R. Congo); SARRNET = Southern
Africa Root Crop Research Network; SENASA = Servicio Nacional de Sanidad Agraria (Peru); UPWARD = Users'
Perspective with Agricultura! Research and Oevelopment (CIP network); USAID = United States Agency for lnternational
Development.

On-line manuals are available at training


web, accessed through www.cipotato.org.
This popular, cost-free learning alternative
is invaluable to developing country
students and researchers.

rate with one or more of these centers. For


example, IPGRI joined CIP in presenting a
geographic information systems (GIS)
course, held in Colombia for 18 persons
from across the world (see Table 2).

Courses at regional locations tend to be


more production-oriented, mostly focused
on delivering techniques for application
under field conditions (See Table 2). They
are for people from the host or surrounding
countries. Many who attend these courses
are in a position to spread their knowledge; often these are extensionists. Most
courses have an emphasis on training
trai ners and faci 1itators.

Farmer field schools

lnter-center cooperation in training has


increased over the past two years, in part
because of increased CGIAR interest in
global initiatives and intercenter cooperation. CIAT, ICRAF, llTA, and ILRI have
partnered with CIP to present workshops

and conferences of common interest with


participants from institutions that collabo-

All CIP's training is aimed at improving


the lives and livelihood of farmers across
the world, whether by building researcher
skills in the lab, by providing opportunities
for technicians, extensionists, and policy
makers to i nteract and learn together, or
by going into the field to work directly
with farmers.
CIP's principal training tool for reaching
farmers is the farmer field school (FFS)
approach. FFSs teach people close to
home among their neighbors. Farmers
gather for instruction under actual field
conditions and learn together. Sorne of the
most pertinent topics in 1999-2000 were
integrated pest and disease management

CIP Program Report 1999 - 2000

23

Table 3. Direct farmer training, including farmer field schools.


Technical aspects Number of Countries
Co-sponsors
included
participants represented
Seed, LB, BW, scab
400
Bangladesh
Tuber Crop Research Center, CARE-Bangladesh
LB, PTM, crop
management
LB, seed, BW

LB, cut worm, seed


LB, APW, BW
Total

270

1371

175
383
2599

Bolivia

PROINPA foundation, ASAR

China, Uganda,
Ethiopia,
Bangladesh

Chongqing Plant Protection lnstitute, Extension service


at county level, CIP office in China, National
Agricultura! Research Organization, AFRICARE,
Ethiopian Agricultura! Research Organization, Self
Help Development lnternational, Tuber Crop Research
Center, CARE-Bangladesh
Ethiopian Agricultura! Research Organization, Self Help
Development lnternational
CARE-Peru, CI P-Headq uarters

Ethiopia
Peru

LB = late blight; BW = bacteria! wilt; PTM = patato tuber moth; APW = Andean patato weevil.PROINPA = Fundacin
para la Promocin e Investigacin de Productos Andinos, Bolivia; ASAR - Asociacin de Servicios Artesanales y Rurales
de Bolivia.

of patato and sweetpotato and overal 1 crop


management, including seed patato
production and use.
A strong supporter of the FFS direct
training method, the lnternational Fund for
Agricultura! Development (IFAD), supported training for more than 2500 farmers
in eight countries (Table 3).
lnstructors have described FFS as extremely useful and follow-up research
supports this position (see Orrego et al.,
page 225 and Van de FI iert et al., page
331 ). About farmer field schools held in

Nepal on seed quality and integrated pest


management, Hidalgo, Campilan, and
Lama (report begins on page 239) write,
"The project [FFS] demonstrated that one
does not have to be a mal e, 1iterate farmer
from the dominant ethnic group in arder to
participate meaningfully in agricultura!
research and extension. Of those who
successfully completed the F.FS, one out of
four was illiterate; one out of three was
female. In addition, 7 of the 9 FFSs had
participants coming from multiple ethnic
groups. Average knowledge gain ... was 84
percent."
P. Malagamba and E. Echeanda

24

Building Partner Skills

Research on Potato
Scientist and Farrner
Partners in Research for the 21 st Century

Plant Defense Genes Associated with Quantitative


Resistance to Potato late Blight
P. Manosalva1, S. Torres 1, F. Trognitz1, 2 , R. Gysin 1, D. Nio-Liu1, R. Simon1,
M. Herrera1, W. Perez 1 , J. Landeo 1, B. Trognitz 1, 2 , M. Ghislain 1 , and R. Nelson 113

To identify genes potentially contributing to quantitative resistance to late


blight, markers corresponding to plant defense genes were tested for linkage
disequilibrium with quantitative resistance to potato late blight in a diploid
mapping population and in a tetraploid population derived from a resistancebreeding program. The 39 genes tested included R-genes, pathogenesisrelated genes, genes involved in signal transduction, and transcriptional
regulatory factors. Several plant defense genes were found to be associated
with quantitative trait loci (QTLs) in the mapping population. A single QTL,
explaining up to 13 % of the phenotypic variance in field trials, was associated with seven loci corresponding to two genes of the phenylpropanoid
pathway and two loci related to WRKY regulatory genes. Markers based on
the polymerase chain reaction for two genes (lipoxygenase and STH) showed
an association with resistance both in the diploid mapping population and in
the tetraploid population. Hence, our results suggest that field resistance to
late blight may well be attributable, at least in part, to known defense-related
genes.
Genome mapping and studies of plantmicrobe interactions have led to a
significant increase in the understanding of
both qualitative and quantitative resistance. A number of genes induced in
response to pathogen attack have been
identified. In many cases, however, it has
been difficult to demonstrate conclusively
that these pathogen-induced genes play a
role in plant defense. Likewise, although
many putative quantitative trait loci
(QTLs) far resistance to potato late blight
have been identified using anonymous
markers (e.g., Leonards-Schippers et al.,
1994), it has been difficult to demonstrate
that these QTLs tru ly correspond to
1

CIP, Lima, Peru.


Present address: Austrian Research Center Seibersdorf (ARCS),
Austria.
3
Present address: MacKnight Foundation, lthaca, NY, USA.

defense genes, and the estimated location


of the QTL can span several centimorgans.
Precise gene mapping and tagging is
particularly difficult in the potato because
of the polyploid nature of the crop.
In an effort to improve the precision of
QTL identification and to determine the
molecular basis of quantitative resistance,
we have pursued a candidate-gene
approach to identify and locate genes
potential ly contributing to quantitative
resistance to late blight. lf successful, this
would facilitate marker-assisted selection
and/or allele cloning and transfer. Using
this type of approach, genes involved in
biochemical pathways related to traits
of interest have been found to be colocalized with QTLs in crops such as
wheat and bean. The candidate-gene

CIP Program Report 1999 - 2000

27

approach has been successfully applied to


a number of characters, including disease
resistance (e.g., Faris et al., 1999; Geffroy
et al., 2000).
This study was undertaken to evaluate the
usefulness of known defense genes in
plants to explain quantitative resistance to
P. infestans. We used the most resistant
and susceptible segregants (phenotypic
extremes) of a diploid mapping population
(Ghislain et al., 2001) and of a tetraploid
family derived from CIP's late-blight
breeding program. The phenotypic extremes were assayed for linkage
disequilibrium using a range of molecular
markers related to candidate genes. To
analyze gene families, we developed and
applied a novel approach based on the
ligation-mediated polymerase chain
reaction (LM-PCR) (Hornstra and Yang,
1993). Oefense gene markers found to be
associated with resistance were placed on
the QTL map.

Materials and Methods


Plant material
The diploid population (PO) consists of
246 diploid progenies derived from a cross
between Solanum phureja CHS-625 ("P,"
with quantitative resistance) and the
S. tuberosum ssp tuberosum dihaploid PS-3
("O," susceptible). A subset of 94 plants
from this population had previously been
used for QTL mapping (Ghislain et al.,
2001 ). Phenotypic data from field trials
for the full population were used to select
the most resistant (n = 34) and susceptible
(n = 34) plants of the PO population
(Trognitz et al., 2001) for use in screening
for markers associated with resistance. The
two groups were significantly different in
susceptibility, as indicated by their mean
areas under disease progress curves
(AUOPCs) under field conditions. For
instance, for the field trial conducted in
Comas, Peru, in 1999, the resistant group
showed a mean AUOPC of 661, and the
susceptible group, a mean AUOPC of 2548
(z Stat = -1 6.12, P = O).

28

Research on Potato

Among eight tetraploid families derived


from CIP's Population B characterized for
resistance to late blight, the family 393228
was identified as showing the broadest and
most continuous distribution of resistance.
This population (n = 130 plants) was
characterized phenotypically in field,
greenhouse, and detached-leaflet tests.
The most resistant (n = 31) and susceptible
(n = 30) individuals were selected for
analysis (see "Results" for details of the
phenotypic data).

DNA markers
The following ONA clones were used for
analysis of restriction fragment length
polymorphism (RFLP): Pto, Pti-4, Pti-5, and
Pti-6 (provided by M. O' Ascenzo and
G.B. Martin), glucanase (provided by
E. Kombrink), osmotin (provided by
T. Chen), PAL and 4CL (provided by
l.E. Somssisch), ch aleone isomerase (CH 1)
from maize (provided by E. Grotewold),
phosphol ipase O probe (provided by
O. McGee and J.E. Leach), SGT (provided
by W. Belknap), and 21 anonymous
probes previously associated with QTLs
for late-blight resistance (provided by
C. Gebhardt).
For PCR-based assays, gene sequences and
sequences of expressed s,equence tags
(EST) were obtained from the Mendel
database and the database of the National
Center for Biotechnology lnformation
(NCBI). Corresponding primers were
designed using the program Primer 0.5
(unpublished software by Oaly, Lincoln,
and Lander; available from the Whitehead
lnstitute, MIT Center for Genome Research) and synthesized by GENSET (La
Jol la, CA, USA).

DNA preparation and analysis


Plant ONA was extracted using the
procedure of Porebski et al. (1997). For
RFLP analysis (Botstein et al., 1980);
southern hybridization was carried out
according to standard protocols. ONA
detection was performed using the enhanced chemiluminescence (ECL) protocol

(Amersham Pharmacia Biotech,


Piscataway, NJ, USA). Polymorphism was
detected by simple amplification using
PCR for a few genes and markers. When
no polymorphism was detected by simple
PCR, the cleaved amplified polymorphic
sequences technique (CAPS) (Konieczny
and Ausubel, 1993) was used. The LM-PCR
technique was adapted and used to allow
allelic differentiation for gene families.
Oepending on the ONA sequence of
interest, genomic DNA was digested with
BamHI, EcoRI or Kpnl. A synthetic linker,
corresponding to the termini created by
restriction and carrying a priming sequence, was 1igated to the restriction
fragments. Ampl ification was performed
using a linker primer and a gene-specific
primer.

Data analysis
lndependence of marker distribution
(presence or absence) and phenotypic
groups (resistant and susceptible) was
tested by the X2-test using a 2x2 contingency with one degree of freedom. To
reduce the experiment-wise error, we used
a maximum P value of 0.01 to declare a
significant association. MAPMAKER/EXP
v3.0b software was used to locate the
candidate genes on the PO map (Lander
and Botstein, 1989).

Linkage disequilibrium in the PD


population
For the phenotypic extremes in the PO
population, 289 candidate gene loci were
scored. For 34 loci (12%), the hypothesis
of independence of marker and resistancephenotypic classes was rejected (Table 1).
Oifferent genes and groups of genes were
compared for the frequency of resistanceassociated bands. The data obtained
suggested that candidate genes could be
an efficient way of identifying genes that
potentially explain quantitative resistance.
In particular, the PAL and WRKY genes
showed strong evidence of association
with resistance (P < 0.00001) (Tables 1 and
2, Figure 1). Other candidate genes that
showed an association with resistance
(P < 0.01) included two R-gene-related
markers (derived Pto); the signal-transduction gene Pti-5; the pathogenesis-related
genes osmotin and STH (PR-10); the
phenyl-propanoid-pathway genes
4-coumarate CoA ligase (4CL) and chalcone isomerase (CHI); lipoxygenase;
metallotheionein; and NAOH-dehydrogenase. Bands corresponding to genes
such as peroxidase and chaperonin were
not significant at P < 0.01 but were
significantly associated with resistance at

<o.os.

To assess whether bands related to plant

Results
In this study, trait-marker associations were
sougrt for quantitative resistance to late
blight using genes involved in plant
defense. A total of 335 segregating loci
were assayed in two populations of patato,
segregating for quantitative resistance to
late blight. The loci tested in this study
represented 39 genes, gene elements, or
gene families, many of which are known
or suspected to play a role in plant defense. Loci were assayed using four
methods: ONA hybridization (RFLP),
simple PCR, CAPS, and LM-PCR (Hornstra
and Yang, 1993). The results were compared to existing datasets obtained using
anonymous markers.

defense genes were more frequently


associated with resistance than bands
related to anonymous probes (RFLP
markers) and non-defense-related genes,
anonymous markers previously used for
QTL mapping were analyzed and compared to defense-related gene markers
(Table 2). Anonymous AFLP, RAPO, RFLP,
and SSR markers (Powel 1 et al., 1996)
were associated with resistance at frequencies ranging from 5-9%. Among 21 RFLP
markers previously mapped to QTLs for
quantitative late-blight resistance
(Leonards-Schippers et al., 1994), four
showed an association with resistance
(19% of probes or 8% of bands scored).
Arbitrarily selected gene-based RFLP and

CIP Program Report 1999- 2000

29

w
e
:o

:::i

"'O

o~

Table 1. Loci associated with quantitative resistance to late blight in a diploid mapping population and in a tetraploid breeding population at P < 0.01.
x2
Target gene ormarker
Marker type Locus defined by (ID/enzyme/ Band origin
Number of observad (tested)
(parent)
/band) molecular weight (kb)
r
s
Resistance genes (pathogen recognition)
p
3 (16)
Conserved domain
PCR
12 (17)
AS 1-S2/0.23
6.96
Signal transduction and gene regulation
Pti-5
o
RFLP
3 (23)
Pti5/EcoRl/3.8
14 (23)
9.33
LM-PCR
o
WRKY
WRKY1-r/BamHl/0.33, 0.51
25 (34)
4 (33)
23.28
LM-PCR
o
WRKY2-r/BamHl/0.49
10 (34)
24 (33)
10.89
LM-PCR
o
WRKY2-r/BamHl/0.9
25 (34)
20.78
5 (33)
LM-PCR
o
18.62
WRKY4-f/K~nl/0.9
5 {34}
23 {33}
Pathogenesis-related genes
1,3-B-glucanase (GLU)
CAPS
387170.9
12 (26)
26 (29)
GLU/Dral/0.12
10.19
LM-PCR
o
Osmotin (PR-5)
Osmotin/EcoRl/0.29
26 (30)
5 (26)
22.97
o
RFLP
Osmotin/EcoRV/13, 7; Xbal/12
24 (33)
10 (33)
10.25
o
CAPS
Osmotin/Rsal/4
9 (31)
26 (34)
12.83
STH2-21 (PR-10)
CAPS
o
STH/Dral/0.65,Taq/0.45
5 (33)
16 (31)
8.06
o
CAPS
STH/Dral/0.3, 0.26, 0.6
5 (33)
14 (30)
5.99
CAPS
STH/Taq 1/0 .850
387170.9
16 (23)
6 (22)
6.44
CAPS
386209.10
STH/Taq 1/0 .259
21 (30)
3 (15)
8.14
CAPS
386209.10
18.02
STH/Dral/0.17
25 {33}
2 {19}
Phenylpropanoid pathway (phytoalexin synthesis and lignin synthesis)
p
Phenylalanine ammoniumlyase (PAL) RFLP
12 (34)
24 (33)
PAL/EcoRl/1.9
7.99
o
RFLP
10 (34)
23 (33)
9.32
PAL/EcoRl/2
o
RFLP
25 (34)
3 (31)
24.42
PAL/EcoRl/2.8
p
31 (34)
1o(33)
23.63
RFLP
PAL/EcoRl/3
o
RFLP
PAL/Hindlll/2.8, 2.7
9 (34)
22 (27)
16.09
p
15 (28)
4 (26)
7.027
RFLP
PAL/Hindlll/5
o
8 (28)
21 (26)
12.74
RFLP
PAL/Hindlll/5.5
o
26 (34)
16.24
RFLP
8 (33)
PAL/Ndel/3.6
o
9 (34)
27 (33)
RFLP
PAL/Ndel/4,4.4
18.47
p
RFLP
22 (34)
6 (33)
13.04
PAL/Ndel/5.9
o
RFLP
26 (34)
PAL/Ndel/6
1 (33)
34.55
o
RFLP
PAL/Ndel/6.8, 6.9
26 (34)
5 (33)
22.92
LM-PCR
o
PAL2/BamHl/0.075
9 (34)
26 (32)
17.72

P (x2 test)

0.008
0.002
<0.00001
0.00096
<0.00001
0.00002
0.001
<0.00001
0.001
0.0003
0.0045
0.014
0.011
0.004
<0.00001

>

0.004
0.002
<0.00001
<0.00001
<0.0001
0.008
0.0003
<0.0001
<0.00001
0.0003
<0.00001
<0.00001
<0.0001

4-coumarate CoA ligase (4CL)

Chalcone lsomerase (CHI)

Lipid metabolism
Lipoxygenase
Other
Metallo-thionein (Meth IA)
NADH-Deh~drogenase

LM-PCR
LM-PCR
LM-PCR
RFLP
RFLP
RFLP
LM-PCR
LM-PCR

PAL2/BamHl/0.41 O
PAL6/BamHl/O .420
PAL6a/BamHl/0.5
4CL/EcoRl/9
4CL/Hindlll/8
4CL/Ndel/4.5
CHl/BamHl/0.320, 0.315
CHl/BamHl/0.53

CAPS

D
D
D
D
D
D
D
D

27 (34)
23 (30)
9 (30)
24 (33)
25 (34)
9 (34)
27 (34)
8 (34)

6 (32)
4 (27)
22 (27)
11 (32)
12 (33)
22 (33)
5 (32)
24 (32)

21.89
19.39
13.18
8.13
7.9
9.33
24.36
15.48

<0.00001
<0.00001
0.0003
0.004
0.005
0.002
<0.00001
<0.0001

LOX/Rsal/0.28

386209.10

23 (28)

4 (20}

15.86

<0.0001

PCR
PCR

Meth IA
NADH dhase

D
p

4 (22)
17 {22)

16 (23)
5 {23)

10.03
11.74

0.002
0.0006

RFLP
RFLP
RFLP
RFLP

Cl21/EcoRV/6.5
CP116/Dral/3
GP186/EcoRl/8
GP313/Xbal/11

D
D
D
D

18 (30)
19 (25)
3 (25)
9 (29)

5 (31)
1 (23)
11 (21)
20 (29)

10.69
22.44
6.98
6.89

0.001
<0.00001
0.0082
0.0086

Anonymous markers

C')

=ti
""tl

ce

;
3

~
o
~

<O
<O
<O

1
N>

o
o
o

PAL
PO

p = 0.006

PO

WRKY

PO

P=0.0004

Figure 1: Ligation-mediated PCR of PAL and WRKY genes, for groups of progeny of the PD population with
contrasting phenotypes for late blight resistance under field conditions. "R" and "S" refer to the resistant and
susceptible groups respectively; "P" and "D" refer to the Solanum phureja and S. tuberosum dihaploid parents
respectively. The significance of linkage disequilibrium of the band is indicated by the probability value (P) of the
x2 test.
SSR markers were associated with resistance at a very low rate (0-5%).
Defense-related genes were associated
with resistance at higher frequencies
(5-46% for different groups of genes).
Map locations were estimated for many of
the resistance-associated markers identi-

32

Research on Potato

fied in this study (Table 3). A group of loci


related to PAL, CHI, and WRKY was
mapped to chromosome 111, to a region
associated with a QTL explaining up to
13% of the phenotypic variance in the PO
population. Another group of loci was
mapped to chromosome XII, the strongest

Table 2. Summary of results obtained far the candidate gene analysis of the PO and 393228 populations.
Loci associated with resistance by the chi-sguare test at P < 0.01 are considered as "hits".
Group
Assays
Segregating loci ResistanceHits per
Hits per
Gene family
(no.)
(no.}
assay (%)
associated loci
segregating
("hits") (no.}
locus (%}
Far candidate gene study
Overall
139
335
48
35
14
PO population (2x)
104
289
34
12
33
393228 (4x}
35
46
5
14
11
For candidate genes, PO population
R-genes
13
76
8
Signal transduction
5
44
21
23
11
and gene regulation
WRKY
11
4
133
3
36
Phenylpropanoid
21
12
175
63
33
pathway
PAL (RFLP)
26
3
12
400
46
PAL (RFLP +
38
16
42
6
267
LM-PCR)
PR-genes
38
5
22
13
16
o
Lipid metabolism
13
o
o
12
55
2
4
Other genes (PCR}
22
9
Candidate genes, 393228 population
PR-genes
11
22
4
36
18
1
Lipid metabolism
6
7
17
14
17
o
o
o
18
Other (PCR}
By method, candidate gene study (both populations}
106
3
10
3
PCR (direct)
30
8
62
15
13
CAPS
53
17
RFLP
99
50
17
34
68
11
50
16
LM-PCR
22
Data obtained for QTL mapping study, PO population
456
42
AFLP
43
98
9
7
216
16
14
RAPO
111
4
11
5
RFLP-genomic
78
38
2
13
6
32
SSR-genomic
16
o
o
o
RFLP-genes
28
14
104
5
10
5
SSR-genes
52
and most consistent QTL detected for the
PD population (Ghislain et al., 2001 ).
Using DNA hybridization and CAPS, no
osmotin bands were found for the QTL on
chromosome XII, but an osmotin-like band
associated with resistance was detected
and located to a peak of chromosome XII
using LM-PCR. WRKY and osmotin were
co-localized with QTL on chromosome
VIII. Sorne QTLs identified by Ghislain et
al. (2001) were not associated with any of
the markers identified in this study.

Linkage disequilibrium in the tetraploid


population 393228
Correlations among greenhouse and field
results were highly significant for the
evaluation of quantitative resistance in
the tetraploid population 393228 (e.g.,
field and greenhouse tests in the year
2000: r = 0.69 and P < 0.000001 ). Based
on these data, the most resistant (n = 31)
and susceptible (n = 30) lines among the
130 progeny were selected for analysis of
molecular markers. The mean AUDPCs
of the resistant and susceptible groups in

CIP Program Report 1999 - 2000

33

Table 3. Associations of candidate genes with quantitative trait loci


QTL1
Frequency (%) of detection
Candidate gene loci
among four field trials
mapped to QTL2
Q-P-Vll
25
Q-P-Xll
75
o
Q-P-XI
75
o
Q-0-111
50
10
Q-0-lllb
25
2
Q-0-V
100
1
Q-0-Vll
50
o
Q-0-Vlll
50
2
Q-0-Xll
100
7

identified in the PO mapping population.


Gene(s) or gene families
Non e
Non e
Non e
WRKY, PAL, CHI, Cell wall protein
Pti-5, 4CL
POX-UE
Non e
WRKY, osmotin
WRKY, PAL, CHI, osmotin, cell division
cycle protein, glucose transporter

1 Quantitative trait loci for late blight resistance defined in the PO mapping population (Ghislain et al., 2001 ). Codes

(Q-x-y) indicate the following: Q refers to QTL; x refers to the parent from which the OTL was derived (P for Solanum
phureja; O for S. tuberosum dihaploid); y refers to the chromosome on which the QTL was located.
2 Only markers showing statistically significant association with resistance were considered.

the 2000 field test were 766 and 1594,


respectively (z Stat = -9.9, P = O).

using CAPS (n
PCR (n = 1).

Eight genes were tested for linkage


disequilibrium in the tetraploid family
393228 using the CAPS and LM-PCR
techniques (Table 2). For these genes and
gene families, 46 loci were scored, and
five (24%) were associated with resistance. Evidence for linkage disequilibrium
was obtained for three genes (P < 0.01 ):
STH, LOX, and glucanase.

Discussion

Efficiency and complementarity of


techniques used

RFLP and LM-PCR were relatively efficient


at revealing resistance-associated loci (hits
per assay). These techniques were shown
to be complementary in the search for
polymorphic loci associated with resistance. For instance, for the PAL gene, 26
segregating loci were identified using
RFLP, 12 of which were associated with
resistance. Using LM-PCR, an additional
12 segregating loci were identified for
PAL; four of these were associated with
resistance. For CHI, nine segregating loci
were found using RFLP, none of which
were associated with resistance. Using
LM-PCR, 11 segregating loci were identified, two of which were associated with
resistance. For osmotin, three different loci
associated with resistance were detected

34

Research on Potato

= 1),

RFLP (n

= 1),

and LM-

The objective of this study was to identify


genes potential ly related to quantitative
resistance to potato late blight. Many
defense-related loci were associated with
resistance and mapped to putative QTLs.
Such associations were detected more
frequently for the defense-related probes
and primer-pairs than for anonymous
markers. Defense-related genes were also
associated with quantitative resistance in
a tetraploid popu lation derived from CIP's
breeding program for late-blight resistance
(Pop u lation B).
Three genes of the phenylpropanoid
pathway showed an association with
resistance to late blight in this study. This
pathway is central in plant defense
(Legrand, 1983). Fritzemeier et al. (1987)
demonstrated rapid and local activation of
PAL gene expression in patato leaves
infected by P. infestans. In a bean population segregating for anthracnose
resistance, Geffroy et al. (2000) obtained a
QTL that co-segregated with a PAL gene.
A cluster of defense gene loci, corresponding to PAL, CHI, a cell wall protein, and
WRKY markers, was mapped to a QTL on

chromosome 111 in the PO mapping


population. A QTL was recently reported
in the same region based on a study of a
different population (Ewing et al., 2000).
Oefense genes were also found to map to
other QTLs. For instance, osmotin and
WRKY markers were 1in ked to a QTL on
chromosomes VIII and XII. WRKY is a
transcription factor that regu lates pathogenesis-related proteins such as osmotin
(PR-5), and WRKY was recently shown to
be induced in the potato by infection with
the late-blight pathogen Phytophthora
infestans (Eulgem et al., 2000; Oellagi et
al., 2000). STH (PR-1 O) and lipoxygenase
were associated with resistance in both the
PO population and the tetraploid breeding
population. PR-1 O expression increases
after wounding, treatment with elicitors, or
infection with P. infestans (Constabel et
al., 1995).
lt has been suggested that R-genes or
R-gene analogs are likely to account for a
significant proportion of quantitative
variation in disease resistance
(Michelmore, 1995). This hypothesis is
supported by the observation that QTLs
have been mapped to the same chromosomal regions as major genes (e.g., Leister
et al., 1996; Shen et al., 1998) and that
defeated R-genes have a residual effect.
Our results, however, do not support the
hypothesis that quantitative resistance can
be explained largely by known R-gene
analogs. A relatively low percentage (1 %)
of the R-gene-derived markers analyzed
were associated with resistance; such a
low percentage could be obtained by
chance alone.
Many plant defense genes are members of
large gene families. We found the RFLP
and CAPS techniques to be inadequate for
dealing with gene families. RFLP does not
permit resolution of numerous bands, and
CAPS does not allow the resolution of
members of gene families. To overcome
these limitations, we adapted the LM-PCR
technique, which has been used for DNA
fingerprinting of pathogen strains,

footprinting assays, sequencing, and


cloning. This technique was quite efficient
for our purposes. To our knowledge, this
represents an innovative application of the
LM-PCR method.
We have shown that quantitative resistance to late blight may be due, at least
in part, to defense-related genes of the
phenylpropanoid pathway and transcriptional factors regulating PR genes.
Candidate gene markers associated with
quantitative resistance to late blight were
efficiently identified by testing defense
genes in the most resistant and susceptible
progenies of segregating populations.
Markers differing significantly in frequency between the two extreme
sub-populations can be inferred to be
linked to a QTL controlling the character
(Lander and Botstein, 1989). The positive
results obtained with this approach are
encouraging and set the stage for the use
of a similar approach in on-going breeding
programs. As in the present study, alleles
showing positive association with resistance in progenies derived from breeders'
crosses can be subsequently located on the
potato chromosomes using available
mapping populations.
lf indeed the defense genes are causally
responsible for differences in disease
resistance, the utilization of resistance by
marker-assisted selection and/or direct
gene transfer wi 11 be mu ch faci 1itated.
Cloning QTL-associated genes, transferring
them to a susceptible genotype, and
showing a reduction in disease susceptibility will conclusively demonstrate the
causal relationship between specific
alleles and disease resistance. This work is
currently in progress.

Acknowledgments
This research was supported in part by the
government of Germany (BMZ Project No.
96.7860.8-001 .00). We are grateful to the
individuals and institutions that provided
ONA clones. We thank Ornar Prado for
technical assistance in the laboratory. We

CIP Program Report 1999 - 2000

35

also thank J. Chittoor, J.E. Leach,


H. Leung, and M. Bonierbale for useful
discussions and helpful comments.
O.P. Zhang provided useful comments on
the manuscript.

References
Botstein, D., R.L. White, M. Skolnich, and
R.W. Davis. 1980. Construction of a
genetic linkage map of man using
restriction fragment length
polymorphisms. American Journal of
Human Genetics 32:314-331.
Constabel, C.P. and N. Brisson. 1995.
Stigma- and vascular-specific expression
of the PR-1 Oa gene of potato: A novel
pattern of expression of a pathogenesisrelated gene. Molecular Plant-Microbe
lnteractions 8(1):104-113.
Dellagi, A., J. Helibronn, A.O. Avrova,
M. Montesano, E.T. Palva, H.E. Stewart,
l.K. Toth, O.E. Coke, G.D. Lyon, and
P.R. Birch. 2000. A potato gene
encoding a WRKY-like transcription
factor is induced in interactions with
Erwinia carotovora subsp. atroseptica
and Phytophthora infestans and is
correlated with class 1 endochitinase
expression. Molecular Plant Microbe
lnteractions 13:1092-1101.
Eulgem, T., P.J. Rushton, S. Robatzek,
and 1.E. Somssich. 2000. The WRKY
superfamily of plant transcription
factors. Trends in Plant Science
5:199-206.
Ewing, E.E., l. Simko, C.D. Smart,
M.W. Bonierbale, E.S.G. Mizubuti,
G.D. May, and W.E. Fry. 2000. Genetic
mapping from field tests of quantitative
and qualitative resistance to
Phytophthora infestans in a population
derived from Solanum tuberosum and
Solanum berthaultii. Molecular Breeding
6:25-36.
.
Faris, J.D., W.L. Li, D.J. Liu, P.D. Chen,
and B.S. Gill. 1999. Candidate gene
analysis of quantitative disease
resistance in wheat. Theoretical Applied
Genetics 98:219-225.
Fritzemeier, K., C. Cretin, E. Kombrink,
F. Rohwer, J. Taylor, D. Scheel, and
36

Research on Potato

K. Hahlbrock. 1987. Transient induction


of phenylalanine ammonia-lyase and
coumarate:CoA ligase mRNAs in potato
leaves infected with virulent or
avirulent races of Phytophthora
infestans. Plant Physiology 85:34-41.
Geffroy, V., M. Svignac, J. Oliveira,
G. Fouilloux, P. Skroch, P. Thoquet,
P. Gepts, T. Langin, and M. Dron. 2000.
lnheritance of partial resistance against
Colletotrichum lindemuthianum in
Phaseolus vulgaris and co-localization
of quantitative trait loci with genes
involved in specific resistance.
Molecular Plant-Microbe lnteractions
13:287-296.
Ghislain, M., B. Trognitz, M. Herrera,
J. Solis, G. Casallo, C. Vsquez,
O. Hurtado, R. Castillo, L. Portal, and
M. Orrillo. 2001. Genetic loci
associated with field resistance to late
blight in offspring of Solanum phureja
and t~?erosum grown under short-day
cond1t1ons. Theoretical Applied
Genetics 103:433-442.
Hornstra, l. K. and T.P. Yang. 1993. In vivo
f?ot~rinting and genomic sequencing by
l 1gat1on-mediated PCR. Analytical
Biochemistry 213:179-193.
Konieczny, A. and F.M. Ausubel. 1993. A
procedure for mapping Arabidopsis
mutations using co-dominant ecotypespecific markers. The Plant Journal
4:403-41 o.
Lander, E.S. and D. Botstein. 1989.
Mapping Mendelian factors underlying
quantitative traits using RFLP linkage
maps. Genetics 121 :185-199.
Legrand, M. 1983. Phenylpropanoid
metabolism and its regulation in
disease. In: Callow, J.A. (eds.).
Biochemical Plant Pathology. John
Wiley & Sons Ltd., NY, USA.
p. 367-384.
Leister, D., A. Ballvora, F. Salamini, and
C. Gebhardt. 1 996. A PCR approach for
isolating pathogen resistance genes
from potato with potential for wide
application in plants. Nature Genetics
14:421-429.

Leonards-Schippers, C., W. Gieffers,


R. Schafer-Pregl, E. Ritter, S.J. Knapp,
F. Salamini, and C. Gebhardt. 1994.
Quantitative resistance to P. infestans in
potato: A case study for QTL mapping
in an allogamous plant species.
Genetics 137:67-77.
Michelmore, R. 1995. Molecular
approaches to manipulation of disease
resistance genes. Annual Review of
Phytopathology 15:393-427.
Oberhagemann, P., C. Chatot-Balandras,
R. Schafer-Pregl, D. Wegener,
C. Palomino, F. Salamini, E. Bonnel,
and C. Gebhardt. 1999. A genetic
analysis of quantitative resistance to
late blight in potato: Towards markerassisted selection. Molecular Breeding

5:399-415.
Porebski, S., L.G. Bailey, and B.R. Baum.
1997. Modification of a CTAB DNA
extraction protocol for plants containing
high polysaccharide and polyphenol

components. Plant Molecular Biology


Reporter 15:8-15.
Powell, W., M. Morgante, C. Andre,
M. Hanafey, J. Vogel, S. Tingey, and
A. Rafalski. 1996. The comparison of
RFLP, RAPO, AFLP and SSR (microsatellite) markers for germplasm
analysis. Molecular Breeding

2:225-238.
Shen, K.A., B.C. Meyers, M.N. lslamFaridi, D.B. Chin, D.M. Stelly, and
R.W. Michelmore. 1998. Resistance
gene candidates identified by PCR with
degenerate oligonucliotide primer map
to clusters of resistance genes in
lettuce. Molecular Plant-Microbe
lnteractions 11 :815-823.
Trognitz, B. R., M. Orrillo, L. Portal,
C. Romn, P. Ramn, S. Perez, and
G. Chacn. 2001. Evaluation and
analysis of reduction of late blight
disease in a diploid potato progeny.
Plant Pathology 50(3):281-291.

CIP Program Report 1999 - 2000

37

Characterization of Phytophthora lnfestans


Colonizing different Solanaceous Species in Peru,
with lmplications on the Control of Potato Late Blight
G. Garry1 , G. Forbes2 , A. Salas 1, W. Prez1, M. Santa Cruz1, H.M. Pinedo1,
E. Gonzles1, M. Rivera1, and R.J. Nelson 1

To test the hypothesis that wild and cultivated Solanum spp. are attacked by
the same populations of Phytophthora infestans (Mont.) de Bary, we characterized isolates attacking wild relatives of potato (Solanum tuberosum L.)
using molecular markers and by an aggressiveness parameter: diameter of
lesion in a detached-leaf assay. lsolates (n = 287) were sampled from the
northern and central highlands of Peru, from different cultivated potatoes
(both bred varieties and native potatoes), from different species of wild
potatoes of Petota section, from S. caripense (non-tuber bearing Solanum of
Basarthrum section), and from wild tomatoes. All the isolates analyzed belonged to four lineages, EC-1, PE-3, PE-7, or US-1, which previously had been
described in Peru. The same pathogen genotypes can attack wild (Petota) and
cultivated potatoes. Our results also showed sorne evidence of host differentiation. The genotype PE-7, mainly obtained from wild tomatoes, hada novel
peptidase genotype. A significant interaction (isolate origin) x (inoculated
host) for lesion diameter was obtained between isolates from S. caripense and
bred potatoes of the EC-1 or US-1 lineages, but not for other host-isolate
combinations.
Late blight (LB), caused by the oomycete
pathogen P. infestans is one of the most
devastating diseases of cultivated potatoes
worldwide. P. infestans can cause symptoms on leaves, stems, and tubers. The
disease is responsible for important
economic losses and high levels of
fungicide use. Because qualitative resistance conditioned by major genes has
proven to be non-durable, potato improvement efforts in both developing and
industrialized countries have focused on
improving the levels of quantitative
1

CIP, Lima, Peru.

CIP, Quito, Ecuador.

resistance to late blight in breeding


populations.
Many wi Id potatoes, and other wi Id
solanaceous species that do not form
tubers, have high levels of quantitative
resistance and are considered potential
sources of resistance for potato breeding.
Little is known, however, about the
pathogen genotypes infecting wild Solanaceae. Sorne Solanaceae are known to
have host-specific populations of
Phytophthora (Ordez et al., 2000), but
these hosts are genetically distant from
potato and are not used in potato improvement. Popu lations of the pathogen

CIP Program Report 1999 - 2000

39

attacking tuber-bearing wild Solanum


(those in section Petota) have not yet been
studied in the Andes.
Several issues related to the use of wild
potatoes for breeding remain unresolved.
First, it is possible that breeders will
confuse host specificity with host resistance. This could occur if wild potatoes
are crossed with cu ltivated potatoes and
ea ch type has a different pathogen popu lation adapted to it. lf progeny of this cross
are screened against the pathogen population adapted to the cu ltivated potato, the
breeder may select potato genotypes for
which the pathogen population is not
adapted. These genotypes would be highly
resistant to one pathogen population,
while being susceptible to another. This
scenario is hypothetical at this time,
beca use 1ittle is known of specific adaptation (i.e., host specificity) of P. infestans to
wild and cultivated potatoes. Even if
experimental evidence shows that host
specificity exists, further studies are
required to determine patterns of inheritance in progeny resulting from a cross
between two potato genotypes that have
different pathogen popu lations.
Host specificity may also affect the
epidemiological dynamics of the disease.
Wild Solanaceae growing clase to cultivated potatoes may serve as a reservoir or
inoculum source for epidemics. lf different
populations of the pathogen are specifically adapted to cultivated and wild
potatoes growing in clase proximity, the
role that either patato species will play in
the epidemic of the other is minimal. lf
plant qreeding results in a loss of host
specificity, wild and cultivated potatoes
may be hosts of the same pathogen
population, and therefore become dynamic components of the same epidemic.
The present work was initiated to test
whether the same genotypes of P. infestans
attack both wild and cultivated hosts in
Peru. Host specificity was also evaluated.

40

Research on Patato

Materials and Methods


lsolates
lsolates were obtained through a series of
col lection trips conducted by B. Trognitz
and A. Salas in 1997-98, and by G. Carry
and A. Salas in 1999; 287 isolates were
collected from the departments of Ancash,
Cajamarca, Huancavelica, Junn, La
Libertad, Lima, and Piura in northern and
central Peru (Table 1). The majority of
isolates were obtained from 20 solanaceous species of Petota section
(tuber-bearing Solanum). Thirty-one
isolates were obtained from non-tuber
bearing Solanum species of the Basarthrum
section (30 from S. caripense and one from
S. montanum); 113 from five varieties of
bred potato; 42 of diverse species of native
patato, and 31 from three species of wild
tomato. In each case P. infestans was
isolated from leaves. Each infected leaf
was maintained after collection in a
sealed petri dish containing a layer of
1 .5% water agar. By using this method, we
could maintain infected tissue for 7-1 O
days between collection and isolation in
the laboratory.

Pathogen isolation, culture, and storage


Plates with infected tissue were incubated
for 7 days at 15-18C under 12 h of 1ight/
day to promote sporulation. Sporangia
were then collected on a 10-m filter and
ri nsed with water. The fi lter system
al lowed efficient recovery of isolates,
even when the infected tissue was severa!
days old and contaminated with bacteria
and saprophytes. The sporangial suspension was refrigerated at 5-8C to induce
the liberation of zoospores. Patato tuber
slices (var. Huayro) were inoculated with
20 1 of the zoospore suspension/slice, and
incubated at 18C for 5-7 days in a moist
chamber. Mycel ial fragments were
transferred aseptical ly to Rye B agar and
V-8 agar plates. After 1-2 weeks, growing
colonies were transferred to Rye A agar
and maintained at 15C.

Table 1. lsolates obtained from different wild Solanaceae and cultivated potatoes between 1997
and 2000, Peru.
Year collected
lsolates per department
Host
Ancash Cajamarca Huancavelica Junn Libertad Lima Piura
Cultivated (bred) potatoes
(113 isolates)
2000
TxA cv Amarilis-INIA
4
5
10
1999-2000
6
TxA cv Canchn
5
4
8
8
TxA cv Mix
2000
5
TxA cv Molinera
2000
2
1999-2000
5
15
TxA cv Perricholi
6
1999-2000
10
4
TxA cv Yungay
9
6
Cultivated (native) potatoes
(42 isolates)
4
4
S. tuberosum ssp. andigena
19991-2000
10
S. goniocalyx
1999-2000
5
7
1
1999-2000
1
S. choucha cv. Huayru
2000
1
6
S. phureja
2000
2
S. stenotonum
Wild potatoes (70 isolates)
1998
S. acaule
1999
S. ancophilum
1999-2000
8
S. bi/1-hookerii
2
1997-98-1999
S. cajamarquense
1999
S. cantense
1997-98
S. chiquidenum
2000
S. chomatophilum
8
1999-2000
S. gracilifrons
2000
S. hastiforme
9
1999-2000
S.huancabambense
2
1999-2000
S. hypacrarthrum
7
1999-2000
S. medians
5
2000
S. mochiquense
1999
S. orophylum
3
1999-2000
S. paucissectum
3
1999-2000
S. piurae
2
1999
S. raquialatum
1999
S. simplicissimum
3
1999
S. sogarandinum
10
1999-2000
S. wittmackii
Wild Solanum (Basarthrum)
(31 isolates)
2 14
1998 1-992-20003
13
S. caripense
1
1997
S. montanum
Wild tomatoes (31 isolates)
3
3
8
1998-99-2000
Lycopersicum hirsutum
6
1
5
1998-99
L. peruvianum
3
2000
L. pennelii
1

lsolates collected in Cajamarca.

2 lsolates collected in Cajamarca, La Libertad, Piura.


3

lsolates collected in Cajamarca, La Libertad, Lima, Piura.

T x A: S. tuberosum ssp. tuberosum x S. tuberosum ssp. andigena.


CIP Program Report 1999 - 2000

41

Host specificity

Pathogenic aggressiveness

Host specificity was tested by characterizing isolates of P. infestans from wild and
cultivated Solanaceae with molecular
markers: glucose-6-phosphate isomerase
(Gpi) and peptidase (Pep) allozymes,
restriction fragment length polymorphism
(RFLP) and amplified fragment length
polymorphism (AFLP) DNA fingerprinting,
and by assessing pathogenic fitness on
inoculated leaflets.

We tested to determine whether isolates


were more pathogenically aggressive on
the host from which they were isolated
than on another host. Both wi Id and
cultivated hosts were inoculated with
isolates from cultivated potatoes and a
wild 5olanum species. This was done for
three wild species of 5olanum of Petota
secti on (5. huancabambense, 5.
sogarandinum, and 5. bill-hooken), and
from one wild species of Basarthrum
section (5. caripense). In each assay, three
cu ltivated pota to varieties were u sed,
(Amarilis-INIA, resistant to LB; Canchn,
moderately susceptible; and Yungay,
susceptible). All cultivated and wild
potatoes u sed in these tests were considered to be free of major resistance genes,
which may interact with specific pathogen
isolates. For one assay, isolates from
cultivated potatoes and a wild 5olanum
species were inoculated on both the same
wild 5olanum species and on the three
cultivated hosts. Three isolates originating
from each of the wild species and three
from cu ltivated pota to were u sed for each
assay. The same isolates from cultivated
patato were used for all assays. lnoculum
for tests was obtained from tuber slices
incubated for 6 to 7 days at 18C in a
moist chamber with 14 hours of fluorescent
light/day. Each wild species was represented by one to four plant genotypes,
e.g., one in the case of 5. bill-hookeri and
5. sogarandinum and four in the case of
5. huancabambense and 5. caripense.
Each host genotype-isolate combination
was represented by two lesions on two
leaflets distributed randomly in one petri
dish. Two repetitions of each host genotype-isolate combination were done at a
time.

DNA fingerprinting
We obtained RFLP fingerprints for 239
isolates using the moderately repetitive
probe RG57 (Goodwin et al., 1992).
Hybridization and detection were conducted using the non-radioaC:tive kit ECL
(Amersham, lnc., Piscataway, NJ, USA)
according to the manufacturer's instructions.

Diversity and cluster analysis


Diversity estimates were calcu lated using
Nei's formula (1987). RFLP and AFLP
fingerprints were scored visual ly for the
presence (1) or absence (O) of each
polymorphic DNA fragment. Cluster
analysis of the binary data was conducted
using the unweighted pair group method
with arithmetic mean (UPGMA) algorithm
in the software program NTSYS-pc (Rohlf,
1997). Similarity was calculated using
Dice's coefficient.

Allozyme tests
A subset of n = 78 isolates (all of them
obtained in 1999) was analyzed for their
Gpi and Pep genotypes, using one of three
techniques: (1) cellulose acetate electrophoresis (CAE), (2) starch gel electrophoresis, or (3) polyacrylamide gel electrophoresis (PAGE). CAE was the simplest
technique for determining certain Gpi and
Pep genotypes (Goodwin et al., 1995), but
lacked resolution for differentiation of all
known genotypes. Starch gel electrophoresis for Gpi (Spielman et al., 1990) and PAGE
(Pep) were used when greater resolution was
needed (Vega-Snchez et al., 2000).

42

Research on Potato

Leaflets were placed in petri dishes and


inoculated by placing a drop of inoculum
containing 5 x 103 sporangia/ml on the
midrib of each leaflet, thus producing one
lesion/leaflet. lnocu lated leaflets were
incubated at 18C with 14 h light/day. Five
days after inoculation, lesion diameter was
measured along the leaflet midrib.

The interaction between isolate origin


(bred or wild host) and inoculated host
species was tested by ANOVA. The
experimental unit for this analysis was the
average diameter of the two lesions in
each petri dish. The model used for the
analysis of variance was
LO = u + a + b + a*b + c(a) + d(b) + c*d(a*b) +e

in which LO = lesion diameter, u = the


overall mean, a = origin of isolate (bred or
wild potatoes), b = type of host species
(wild or bred potatoes), c = isolate nested
in origin, and d = plant genotype nested in
host species. The interactions effects
c*d{a*b) was used as the denominator in
the F t~st of the interaction a*b, which
was the primary interest of this test.

Results
Characterization of isolates
markers

by molecular

Al 1 the isolates analyzed belonged to one


of four lineages: EC-1, first described in
Ecuador by Forbes et al. (1997); PE-3, first
described in Peru by Prez et al. (1999);
PE-7, found in Peru by Pinedo et al. (1999),
or US-1, described in the USA by Goodwin
et al. (1994) (Table 2). The tests with

al lozyme markers Gpi and Pep confirmed


that the isolates belonging to the US-1
lineage were GPI 86/100, Pep 92/100, and
the isolates belonging to the EC-1 1ineage
were Gpi 90/100, Pep 100/1 OO. All the PE3 and PE-7 isolates were Gpi 100/100 on
cellulose acetate, but they were distinguishable by their Pep banding pattern on
PAGE and on CAE (100/100 for PE-3 and
96/98 for PE-7).
All the lineages could be found on more
than one host species (Table 2). The EC-1
lineage was dominant among the isolates
collected from wild and cultivated (native
and bred) potatoes, but was also present on
wi Id tomatoes and 5. caripense. PE-3 was
found more often on cu ltivated potatoes
than on the other hosts, but two isolates of
this lineage were collected from native
potato and wild tomato. The lineage US-1
was mostly isolated from 5. caripense,
but sorne isolates infecting wild potatoes
(5. raquialatum and 5. piurae) and wild
tomatoes were of the US-1 lineage.
Lineages were also found in more than one
place. For instance, EC-1 was found in
Ancash, Cajamarca, Huancavelica, La
Libertad, Lima, and Piura, and PE-7 in
Cajamarca, Huancavelica, and Lima.

Table 2. Number, location, and host origin of RG57 lineages defined by RFLP of isolates of Phytophthora
infestans collected in Peru between 1998 and 2000.
lsolates per host
Bred
Native
Wild
S. caripense
Wild
RG57 lineage1
Location 2
potatoes potatoes potatoes + Basarthrum tomatoes
EC-1
78
30
53
11
11
An, Ca,
(1110101001001101000111011)
Hca, Lib, Li,
Pi
PE-3
12
4
o
Ca, Li, Pi
(1100100001001100100111011)
US-1
2
14
Ca, Pi
o
o
4
(1010101011001101000110011)
PE-7
Ca, Hca, Li
o
3
o
10
3
(1110101001001100101111011)
Total per host
90
37
59
25
26
Ca, Li,
Ca, Hca,
Location per host
An, Ca,
Ca, Hca, An, Ca,
Hca, Lib, Lib, Li, Pi Hca, Lib,
Lib, Pi
Jun, Li,
Li, Pi
Li, Pi
Lib, Pi,
1

Per lineage as noted by Forbes et al. (1998).

2 Usted by department code. An: Ancash, Ca: Cajamarca, Hca: Huancavelica, Jun: Junn, Lib: La Libertad, Li: Lima, and

Pi: Piura.
CIP Program Report 1999 - 2000

43

Evidence of host specialization based on


DNA fingerprint analysis was especially
compelling for wild tomatoes and for
S. caripense. PE-7 was the dominant
lineage of isolates from wild tomatoes; it
was rare on the other hosts. The U S-1
lineage was the most frequent on
S. caripense in Piura, whereas the lineage
EC-1 was predom i nant among cu ltivated
potatoes in the same department and on
the same dates of col lection (Table 3).
However, patterns of host specialization
were not always clear. The US-1 lineage
was the most frequent on 5. caripense in
Piura in 1999 and 2000, but lineage EC-1
was predom i nant on the same host in
Cajamarca in 1998 and 2000.

Pathogenic aggressiveness of P. infestans


The interaction between isolate origin
(bred or wild host) and inoculated host
species was not significant when species
within Petota section were compared with
cultivated potatoes (Figure 1). There were
no cases in which isolates from bred
potatoes were less aggressive than isolates
from wild potatoes on their own host. In
contrast, when isolates from S. caripense
were compared with isolates from cu ltivated potatoes, there was a significant
interaction (Figure 1). This interaction
indicates that isolates were more aggressive on the host species from which they
carne. All isolates tested from either EC-1
or US-1 lineage showed this interaction. lt
was also exhibited when the isolates from
S. caripense were from the US-1 lineage

and those from cu ltivated potatoes were


from the EC-1 1i neage.

Discussion
Our results confirm and extend those of an
earlier study in which lineages EC-1, PE-3,
and US-1 were found among about 300
isolates coming from cultivated potatoes
in central and southern Peru (Prez et al.,
1999). Our sampling involved more host
species and we found another clonal
lineage, PE-7, principally associated with
wild tomatoes. The high degree of similarity in the results of the two studies supports
the hypothesis that our knowledge of the
Peruvian population of P. infestans is now
fairly accurate.
In general, our results are consistent with
the hypothesis that cultivated and wild
potatoes in Peru are attacked by the same
population of P. infestans. We found no
evidence to support host specificity among
wild and cultivated species within section
Petota of the genus Solanum. These res u lts
are also consistent with those of a recent
study of wild and cultivated potatoes in
Mexico. In that study, the same population
appeared to attack all host species that
were studied (Grnwald et al., 2000).
We did find evidence for host specificity
among cu ltivated potatoes and more
distant relatives within the genus Solanum.
The primary patato lineage of P. infestans
in Peru, EC-1, is not the primary lineage
attacking wild tomatoes or, at times,
S. caripense.

Table 3. Collection date, location, and host differentiation of isolates of P. infestans belonging to the EC-1
and US-1 lineages, collected from cultivated potatoes and S. caripense.
RG57 Lineage1
EC-1 (%)
US-1(%)
Groups
Year collected
Location
o
Bred potatoes
1999
Piura
50 (4)
Bred potatoes
2000
Piura
70 (9)
o
Bred potatoes
2000
Cajamarca
100 (10)
o
15
S. caripense
1998
Cajamarca
85 (6)
S. caripense
2000
Cajamarca
100 (5)
o
11
83 (1 O)
S. caripense
1999
Piura
100 (4)
S. caripense
2000
Piura
o
1 Number in parenthesis is the number of isolates per group.

44

Research on Potato

(cm)
3 ..-----------------~
A
2 ....

4 3 -

2 -

1 ....

1 -

P50.0001

p 50.0001
lsolates from S. caripense
- - lsolates from bred potatoes

lsolates from S. caripense


- - lsolates from bred potatoes

2 -

p 50.0001

p 50.083
lsolates from S. huacabambense
- - lsolates from bred potatoes

lsolates from S. caripense


- - lsolates from bred potatoes

-----

3
2

p 50.47
lsolates from S. sogarandinum
- - lsolates from bred potatoes

lsolates from S. bill-hookeri


- - lsolates from bred potatoes

Figure 1. Lesion diameter on leaflets of wild Solanum and bred potatoes, 5 days after inoculation with a drop of
inoculum (5 x 103 sporangia/ml) of Phytophthora intestans from both hosts in detached-leaf inoculation assays.
(A), (B), (C): Solanum caripense vs bred potatoes. (D), (E), (F): wild potatoes vs bred potatoes. The isolates are
of EC-1 lineage except for (B) isolates of US-1 lineage from S. caripense, and (C) isolates of US-1 lineage from
both hosts. For each detached-leaf inoculation assay, the interaction is significant when Ps._ 0.05. Key: - =
bred potatoes, - = wild Solanum.
In sorne cases, host specificity between
cultivated potatoes and S. caripense was
associated with pathogen clonal lineage.
EC-1 was found on patato and US-1 on
S. caripense. This is similar to results
published for Ecuador (Erselius et al.,
1999), where EC-1 was also found on
patato and US-1 on S. caripense. Unlike in
Ecuador, however, EC-1 was also found on
S. caripense in our study. Nonetheless,
these EC-1 isolates appeared to be more
aggressive on S. caripense than on patato,
based on the significant interaction in

ANOVA (p < 0.05). Host specificity for S.


caripense apparently does not imply host
specificity for other species in section
Basarthrum of the genus Solanum. In the
USA, sorne species belonging to
Basarthrum are attacked by the patato
population of P. infestans (Deahl et al.,
2000).
The presence of wild species in Peru may
induce differentiation of P. infestans. The
PE-7 lineage, which was mostly found on
wild tomatoes, is particularly intriguing in

CIP Program Report 1999 - 2000

45

this sense. This lineage was distinguished


from the others by AFLP and RFLP fingerprints, and also by Pep allozyme alleles.
To our knowledge, this is the first time
that the migration distance 96/98 of PE-7
has been reported. Unique genotypes of
P. infestans were also found in association
with wild hosts in Ecuador (Erselius et al.,
1999). Although many factors may influence the degree of diversity in P. infestans,
diversity in the host appears to be one
important factor. But the observations also
suggest temporal changes in population
structure that could relate to environmental effects or genetic drift.

Conclusion
These results have implications for patato
breeding. Based on the results of this study,
it appears that breeders can use the
species we studied in section Petota as a
source of traits for introgression into
cultivated patato. There is little evidence
that a host specificity factor may be
confused with resistance to P. infestans. In
theory, patato breeders should be able to
use the potato popu lation of the pathogen
(the most abundant and easiest to use) to
screen a segregating population of potato
genotypes derived from a wi Id by cu ltivated cross. This has been the classical
approach to plant breeding. Thus, our
results support traditional breeding
techniques.
On the other hand, attempts to introgress
traits i nto cu ltivated potato from more
distant species in the genus Solanum could
cause problems. Our results and those of
others, (Erselius et al., 1999), indicate that
there are different populations of the
pathogen attacki ng these hosts in South
America. Breeders could potentially
introgress a host specificity factor into
cultivated potato, which would make it
appear resistant to the potato population of
the pathogen. Once planted in the proximity of the wi Id progenitor, however, the
putatively resistant patato could be
vulnerable to the pathogen population
specific to the wi Id progenitor.

46

Research on Potato

The implications of our results for disease


management are analogous to those for
plant breeding. Wild potatoes in section
Petota are potential reservoirs of inoculum
for cultivated potatoes. Therefore, these
wild potatoes could exacerbate epidemics
in cultivated potatoes. Wild Solanaceae
outside section Petota probably play little
or no role in the late blight epidemics of
cu ltivated pota toes.
The alleles present in our Peruvian collection were all present in the Cornell-CIP
database establ ished by Forbes et al.
(1998) except for the Pep allele in lineage
PE-7. The novel allozyme pattern of this
lineage, discovered in Peru, may indicate
that P. infestans has been evolving outside
Mexico for a relatively long time.

Acknowledgments
We want to thank Dr. William Fry of
Cornell University, USA, for providing the
probe RG57 and also several of the CIPLima staff. These are Dr. Bodo Trognitz for
providing the isolates of P. infestans of
1997-98 collection; Soledad Gamboa for
her assistance, as wel 1 as al 1 the techn ical
staff of the laboratory of P. infestans;
F. Mendiburu for statistical analyses; and
the staff in CIP-Huancayo for plant maintenance. This work was financially
supported by the French Ministry of
Foreign Affairs.

References
Deahl, K.L., L.R. Cooke, D.S. Shaw, and
D.J. Carlisle. 2000. Characterization of
isolates of Phytophthora infestans from
wild Solanum spp. in the UK.
Phytopathology 90(6):51 9.
Erselius, L.J., H.R. Hohl, M.E. Ordez,
P.J. Oyarzun, F. Jarrin, A. Velasco,
M.P. Ramn, and G.A. Forbes. 1999.
Genetic diversity among isolates of
Phytophthora infestans from various
hosts in Ecuador. In: lmpact on a
changing world: Program report, 199798. lnternational Patato Center Lima
Peru. p. 39-48.
'
'

Forbes, G.A., X.C. Escobar, C.C. Ayala,


J. Revelo, M.E. Ordez, B.A. Fry,
K. Doucett and W.E. Fry. 1997.
Population genetic structure of
Phytophthora infestans in Ecuador.
Phytopathology 87(4):375-380.
Forbes, G.A., S.B. Goodwin, A. Drenth,
P. Oyarzun, M.E. Ordez, and W.E.
Fry. 1998. A global marker database for
Phytophthora infestans. Plant Disease
82:811-818.
Goodwin, S.B., R.E. Schneider, and
W.E. Fry. 1995. Use of cellulose-acetate
electrophoresis for rapid identification
of al lozyme genotypes of Phytophthora
infestans. Plant Disease 79:1181-1185.
Goodwin, S.B., C. Barak A., and W.E. Fry.
1994. Panglobal distribution of a single
clonal lineage of the lrish potato famine
fungus. Proceedings of the National
Academy of Science 91:11,591-11,595.
Goodwin, S.B., L.J. Spielman, J.M.
Matuszak, S.N. Bergeron, and W.E. Fry.
1992. Clona! diversity and genetic
differentiation of Phytophthora infestans
populations in northern and central
Mexico. Phytopathology 82:955-961.
Grnwald, N.J., W.G. Flier, L. Turkensteen,
and W.E. Fry. 2000. Populations of
Phytophthora infestans in the Toluca
Valley on wild Solanum species are not
different from those on cultivated
potatoes. Phytopathology 90(6):S31.
Nei, M. 1987. Phylogenetic trees. In:
Molecular evolutionary genetics.
Columbia University Press, NY, USA,
p. 287-326.
Ordez, M.E., H.R. Hohl, J.A. Velasco,
M.P. Ramn, P.J. Oyarzn, C.D. Smart,
W.E. Fry, G.A. Forbes, and L.J. Erselius.
2000. A novel population of

Phytophthora, similar to P. infestans,


attacks wild Solanum species in
Ecuador. Phytopathology 90:197-202.
Prez, W., S. Gamboa, M. Coca, R.
Raymundo, R. Hijmans, and R.J.
Nelson. 1999. Characterization of
Phytophthora infestans populations in
Peru. lmpact on a changing world.
Program report 1997-98. lnternational
Potato Center, Lima, Peru. p. 31-38.
Pinedo, H., M. Orrillo, S. Gamboa,
A. Salas, R.J. Nelson, and B. Trognitz.
1999. P. infestans is established in the
lomas of the Peruvian coastal desert.
Proceedings of the Global lnitiative on
Late Blight (GILB) Conference. Vol. 1:
Late blight: A threat to global food
security. Quito (Ecuador). 16-19 Mar
1999. lnternational Potato Center, Lima,
Peru. p. 127-128.
Rohlf, F.J. 1997. NTSYS-pc: Numerical
taxonomy and multivariate analysis
system. Version 1.70. Exeter Software,
Setanket, NY, USA.
Spielman, L.J., J.A. Sweigard, R.C.
Shattock, and W.E. Fry. 1990. The
genetics of Phytophthora infestans:
Segregation of allozyme markers in f 2
and backcross progeny and the
inheritance of virulence against potato
resistance genes R2 and R4 in F1
progeny. Experimental Mycology
14:57-69.
Vega-Snchez, M.E., L.J. Erselius,
A.M. Rodrguez, O. Bastidas,
H.R. Hohl, P.S. Ojiambo, J. Mukalazi,
T. Vermeulen, W.E. Fry, and G.A. Forbes.
2000. Host adaptation to potato and
tomato within the US-1 clona! lineage
of Phytophthora infestans in Uganda and
Kenya. Plant Pathology 40:1-1 O.

CIP Program Report 1999 - 2000

47

Evaluation of Wild Potato Species for Resistance to


Late Blight
W. Prez1, A. Salas1, R. Raymundo1, Z. Huamn 1' 2, R. Nelson 1, 3, and M. Bonierbale1

Late blight, caused by Phytophthora infestans, is a major disease of potato.


This study assessed 51 tuber-bearing Solanum species for resistance to late
blight in a series of greenhouse experiments. A total of 80 out of 133 accessions presented quantitative resistance patterns. Eighteen species from
Central and South America presented both qualitative and quantitative resistance responses, whereas accessions of four Mexican and two South
American species presented only qualitative resistance to the isolate used.
Between one and 90 genotypes in 39 different species have been selected for
verification of resistance under field conditions following evaluation of trueseed-grown and tuber-grown plants. The aim of this broad survey is to
identify and secure sources of complementary components of stable resistance to late blight for use in breeding programs. The results confirm the
value of conserving rare populations of potato genetic resources and provide
the first evidence of resistance to late blight in at least seven species endemic
to the South American center of origin of potatoes. lmplications of the variability detected among species and accessions are discussed in terms of
conservation and breeding objectives.
During the past three years, CIP has
intensified efforts to evaluate the in-trust
germplasm collection of potato for important traits such as resistance to late blight,
bacteria! wilt, and important viruses. Late
blight (LB), caused by Phytophthora
infestans, is the most devastating disease
of potato worldwide. Wild tuber-bearing
5olanum species from Mexico, especially
5. demissum, have previously been used in
the majority of breed i ng programs as
sources of resistance to late blight. Ross
(1 986) esti mated that 5. demissum
germplasm has been incorporated into
more than 50% of the world's potato
varieties. However, since it has become
evident that new races of P. infestans can
1

CIP, Lima, Peru.


Current address: ProBioAndes, Lima, Peru.
3
Current address: MacKnight Foundation, lthaca, NY, USA.
2

readily overcome qualitative resistance


based on major (R) genes that are incompatible with specific virulence genes in
the pathogen, breeders have placed
greater emphasis on quantitative, 'horizontal', or non-specific resistance. This type of
resistance is expected to be more stable
when exposed to the variable and dynamic pathogen population. lt was first
reported to occur along with specific
resistance in certain Mexican wild species
including 5. demissum and 5. stoloniferum
(Toxopeus, 1964). Schober (1981) and
Tazelaar (1981) were the first to report
horizontal resistance in 5olanum accessions from outside of Mexico. South
American genotypes of the cu ltivated
species 5. tuberosum subsp andigena and
5. phureja have apparently contributed to
the levels of quantitative resistance

CIP Program Report 1999 - 2000

49

obtained in a number of important varieties such as Monserrate and Atzimba


(Niederhauser, 1989). Despite the availabi 1ity of sorne varieties that possess stable
types of resistance, there is sti 11 a need far
higher levels of race non-specific resistance, particularly in early varieties.
Wild tuber-bearing Solanum species
contain numerous sources of resistance
(Hanneman and Bamberg, 1986), including potentially useful resistance to P.
infestans (Hawkes, 1950; Ochoa, 1954,
1981; Van Soest, 1983, 1984; Colon and
Budding, 1988; Colon et al., 1995), but
only a few have been subject to detailed
study or been used in breeding programs.
This study presents the results of preliminary evaluation of Solanum germplasm
held in trust by CIP with a view to promoting the more systematic use of wild
species in potato improvement programs to
broaden the genetic base of resistance to
late bl ight.

Materials and Methods


A series of 11 experiments was carried
out at CIP's highland station at Huancayo
(3280 m; 1207' S latitude), Peru. A
total of 133 Solanum accessions (six of
which were duplicated, far a total of 139
population samples) in 51 species of 13
taxonomic series of tuber-bearing species
were evaluated far resistance to late blight
(Table 1 ).
Priority far screening was given to species
never before evaluated and those originating in climatic zones conducive to late
blight. Between one and 1O (usually three)
accessions, each originating from a
different collection site, were evaluated
far each species, using about 48 (range
23-56) plants per accession. A sample of
100 seeds was taken at random from each
of the selected accessions, which had
been maintained in the genebank by full
sib mating among 10-30 plants from an
original or previously renovated population. Seeds were germinated in trays and
the seedlings transplanted to 1O cm pots,
which were placed in trays on the green50

Research on Potato

house floor. Each tray contained twelve


pots with one plant each, and far each
accession, faur trays were randomly
distributed along the greenhouse sides to
avoid any bias due to differentials within
the greenhouse environment.
Evaluation of the accessions was carried
out in two phases. The first phase (5 true
seed experiments) used plants grown from
true seed accessions; and in the second
phase (6 tuber experiments), plants were
grown from tubers of individual plants that
had been selected far resistance during the
first phase. Guidelines far the different
experiments were similar, with the main
differences being the time of testing (the
total experimental period covered several
months), the type of planting material
used, and the number of genotypes (plants)
included per accession.
For each experiment, 13 standards far
quantitative resistance ((SQR); clonal
varieties or breeding lines with known
levels of blight resistance) were included.
Four trays, each representing a replication
of the 13 standards were randomly distributed on both sides of the greenhouse.
Complete sets of genotypes sampled from
a number of accessions were tested within
a given experiment. Data from replication
of the 1 3 standards was i ntended to give
infarmation about variation within experiments (due to differences in the
environment during a given test) and
among experiments (due to changes in
experimental conditions through time).
Whole plants of the accessions to be
tested and the standards were sprayi nocu lated with LB at a concentration of
3000 sporangia per mi with the local
isolate PC0002, which is virulent on
potato differentials carrying R-genes 1, 2,
3, 4, 6, 7, (8), 10, and 11. (The differential
reaction of this isolate with R8 requires
confirmation.) lnoculation was perfarmed
befare flowering, at 30-50 days after
transplanting seedlings to the pots or
planting of the tubers. Special efforts were
made to keep environmental conditions in
the greenhouse stable throughout the series

Table 1. Solanum accessions evaluated far resistance to late blight, Lima, Peru, 1999-2000.
Series and s~ecies
2n EBN 1 Country
No.
m
s
o Proportion No.
Collectors
CIP No.
of origin tested 2
resistant selected
No.
genotypes
(m+r+O)

/No.
ACAULIA
S. a/bicans (alb)
CIP 761452 OCH 12089
72
4 PERU
CIP 762120 OCH 14789
72
4 PERU
CIP 762584 OCHS 16028
72
4 PERU
CIRCAEIFOLIA
S.circaeffolium(crc)
CIP 761344 OCHS 11909
24
1 BOLIVIA
S. circaeifolium var. capsicibaccatum (cap)
CIP 761030 OCHS 11915
24
1 BOLIVIA
CIP 761030 OCHS 11915
24
1 BOLIVIA
CIP 762303 OCHS 15489
24
1 BOLIVIA
COMMERSONIANA
S. commersonii (cmm)
CIP 761087 FB 401 O
URUGUAY
24
CIP 762454 URY 4
URUGUAY
24
CIP 762475 URY 31
24
URUGUAY
CON ICIBACCATA
S.chomatophilum(chm)
CIP 761587 OCH 13341
24
2 PERU
CIP 762568 OCHS 12553
24
2 PERU
CIP 762611 OCHS 16072
24
2 PERU
S. colombianum (col)
CIP 762793 SCL 5050
48
2 ECUADOR
CIP 762793 SCL 5050
48
2 ECUADOR
S. flahaultii (flh)
CIP 761877 OCH 14105
COLOMBIA
48
CIP 761878 OCH 14106
48
COLOMBIA
S. irosinum (irs)
CIP 761252 OCH 11640
24
2 PERU
CIP 762257 OCHS 15210
24
2 PERU
S. paucijugum (pcj)
ECUADOR
CIP 762800 SCL 5084
48
ECUADOR
CIP 762803 SCL 5096b
48
48
ECUADOR
CIP 762816 SCLG 5151
S. urubambae (uru)
CIP 761044 OCH 13781
24
2 PERU
CIP 762368 OCHS 15654
24
2 PERU
S. violaceimarmoratum (vio)
CIP 760563 VSOA 7
24
2 BOLIVIA
DEMISSA
S. demissum (dms)
CIP 761893 OCH 14154
72
4 MEXICO
CIP 761050 OCH 14218
72
4 MEXICO
S. hougasii (hou)
CIP 761902 OCH 14172
4 MEXICO
72
CIP 761899 OCH 14168
4 MEXICO
72

o o
o o
o o o

0.21
0.02
O.DO

o
o
o

0.18

27

31

0.98
1.00
0.77

42

11

24
48
4

48
48
48

42
19
5

4
20
9

2
7
17

0.13
0.60
0.90

3
19
23

48
48
48

47
35
46

1
12

0.02
0.27
0.04

o
o
o

48
48

10
43

36
5

2
o
o o

0.79
0.10

48
47

11
24

31
20

4
2

2
1

0.77
0.49

o
o
o
4

48
48

40
18

2
2

6
28

o
o

0.17
0.63

o
o

42
39
37

26
7
21

14
14
16

o
o
o o

0.38
0.76
0.43

o
o
o

48
48

18

48
48
48

38
47
48

10
1

51

42

56
48
48

48

44
48
48
48

o
o
o
o

o o

2
17

o o
1 o
1

2
8

o
1

5
10

43

1.00
0.63

20
15

22

10

0.98

1 26
13 35

17

1.00
1.00

38
28

36
37

1.00
1.00

17
26

o
o

12
11

CIP Program Report 1999 - 2000

51

Table 1. (continued)
Series and s~ecies
Collectors
CIP No.
No.

2n

No.
EBN 1 Country
of origin tested2

Proportion
No.
resistant selected 3
genotypes
(m+r+O)

/No.
S. iopetalum (iop)
72
CIP 761928 OCH 14208
4 MEXICO
72
CIP 761935 OCH 14221
4 MEXICO
72
CIP 761930 OCH 14212
4 MEXICO
LIGNICAULIA
S. lignicaule (lgl)
24
CIP 761210 OCH 11315
PERU
CIP 761236 OCH 11617
24
PERU
CIP 761652 OCH 13585
24
PERU
LONGIPEDICELLATA
S. fendlerii (len)
48
CIP 761921 OCH 14199
2 MEXICO
CIP 761923 OCH 14202
48
2 MEXICO
48
CIP 761926 OCH 14205
2 MEXICO
S. stoloniferum (sto)
CIP 761884 OCH 14135
48
2 MEXICO
CIP 761062 OCH 14145
48
2 MEXICO
MEGISTACROLOBA
S. do/ichocremastrum (dcm)
CIP 761043 OCH 12071
24
1 PERU
CIP 761439 OCH 12074
24
1 PERU
CIP 761470 OCH 13013
24
1 PERU
CIP 761470 OCH 13013
24
1 PERU
S. megistacrolobum subsp. toralapanum (mga)
24
CIP 760535 HAM 200
2 BOLIVIA
24
CIP 761109 OCH 7609
2 PERU
CIP 761403 OCH 12032
24
2 BOLIVIA
S. megistacrolobum subsp. toralapanum (tor)
CIP 760459 HHA 6616
24
2 BOLIVIA
24
2 BOLIVIA
CIP 761369 OCH 11964
24
2 BOLIVIA
CIP 761055 OCHS 11914
S. raphanifolium (rap)
24
2 PERU
CIP 761113 OCH 7613
24
CIP 761640 OCH 13572
2 PERU
24
CIP 761671 OCH 13610
2 PERU
S. sogarandinum (sgr)
24
CIP 761586 OCH 13336
2 PERU
CIP 762410 OCHS 15723
24
2 PERU
PINNATISECTA
S. cardiophy/Jum (cph)
CIP 762561 OCH 14157
24
MEXICO
PIURANA
S. cantense (cnt)
CIP 762241 OCHS 15159
24
2 PERU
S. chiquidenum (chq)
24
CIP 761588 OCH 13345
2 PERU
CIP 761870 OCH 13963
24
2 PERU
CIP 762573 OCHS 12566
24
2 PERU
52

Research on Potato

o
o

o o
o o

48
48
48

11

21

16

48
47
48

29
12
21

17
27
24

48
45
48

9
45
48

44
48

48
48

1.00
1.00
0.77

23
46
24

2
8
3

o
o
o

0.40
0.74
0.56

o
o
o

31

o o o
o o o
13

11

o o

0.81

34

O.DO
O.DO

o
o

17
48

0.93
1.00

3
30

O.DO

o
o

48
47
48
48

48
47
45
48

43
48
48

1
27
47

o o o
o o o
o
1 2
o o o
o 21 21
o
15
6
1 o o

48
27
32

1
25
10

5 16 26
1
1 o
o
16
6

0.98
0.07
0.69

o
o

48
48
48

1
3
5

26
29
29

17
14
9

4
2
5

0.98
0.94
0.90

15
24
17

48
47

1
42

10
4

10

27
1

0.98
0.11

16
2

29

o o

29

1.00

12

48

22

1.00

48
48
48

37
33
35

11
15
11

o o
o o
o
2

0.23
0.31
0.27

.12
l1
14

21

0.00
0.06
0.00
0.98
0.44
0.02

o
16

o
8
1

Table 1. (continued)
Series and s~ecies
CIP No.
Collector's
No.

2n

EBN 1 Country No.


of origin tested 2

Proportion
No.
resistant selected 3
genotypes

(m+r+O)

/NO.
S. humectophilum (hmp)
CIP 761052 OCH 11753
S. hypacrarthrum (her)
CIP 761259 OCH 11692
CIP 761204 OCHS 11308
CIP 762104 OCHS 14715
S. paucissectum (psc)
CIP 761247 OCH 11634
CIP 762124 OCHS 14816
S. piurae (pur)
CIP 761868 OCH 13959
CIP 761072 OCHS 11615
POLIADENIA
S. polyadenium (pld)
CIP 761911 OCH 14187
TUBEROSA
S. alandiae (aln)
CIP 760469 HHA 6657
CIP 760473 HHA 6665
CIP 761394 OCH 12013
S. ambosinum (amb)
CIP 761842 OCH 13852
CIP 761194 OCHS 11298
CIP 761053 OCHS 11865
S. brevicaule (brc)
CIP 760461 HHA 6619
CIP 760479 HHA 6690
CIP 761074 OCH 11934
S. bukasovii (buk)
CIP 761120 OCH 7717
CIP 761517 OCH 13166
CIP 761664 OCH 13602
CIP 761720 OCH 13679a
CIP 761805 OCH 13796
CIP 761806 OCH 13798
CIP 761847 OCH 13858
CIP 762434 OCH 15822
CIP 761152 OCHS 10114
CIP 761301 OCHS 11851
S. cajamarquense (cjm)
CIP 762616 OCHS 16118
CIP 762619 OCHS 16121
S. candolleanum (cnd)
CIP 761041 OCHS 11913
CIP 762168 OCHS 14959
CIP 762185 OCHS 15011

24

PERU

53

24
24
24

PERU
PERU
PERU

48
23
48

24
24

2 PERU
2 PERU

44
47

24
24

2 PERU
2 PERU

48
48

o
11
o
15
2
o o o
23
41
6
1 o
41
2
1 o
26
o 13
8
o o 2 46
o o o 48
50

24

MEXICO

48

24
24
24

BOLIVIA
BOLIVIA
BOLIVIA

48
48
48

28
48
48

24
24
24

2 PERU
2 PERU
2 PERU

48
48
48

48
47
48

24
24
24

2 BOLIVIA
2 BOLIVIA
2 BOLIVIA

48
48
48

48
47
48

24
24
24
24
24
24
24
24
24
24

2
2
2
2
2
2
2
2
2
2

PERU
PERU
PERU
PERU
PERU
PERU
PERU
PERU
PERU
PERU

48
48
48
48
48
47
48
30
48
43

30
20
47
35
47
26
33
30
47
10

24
24

PERU
PERU

48
48

1
2

24
24
24

2 BOLIVIA
2 BOLIVIA
2 BOLIVIA

48
48
48

48
48
48

o o

48

o o
o o o
o o o
o o o
1
o o
o o o

20

0.06

0.46
O.DO
0.15
0.07
0.83

o
o
o
o
o

1.00
1.00

20
4

1.00

12

0.42
O.DO
O.DO

18

o
o

O.DO
0.02
O.DO

o
o

o o o
1
o o
o o o
17
1 o
o
23
5
1
o o
o
10
3
o 1 o
17
o
4
13
o
2
o o o
1
o o
26
o
7

O.DO
0.02
O.DO

o
o
o

0.38
0.58
0.02
0.27
0.02
0.45
0.31
O.DO
0.02
0.77

7
1

7
2

0.98
0.96

18
11

o o o
o o o
o o o

O.DO
O.DO
O.DO

10
14

30
30

o
3
o
o
2
o
o
o

CIP Program Report 1999 - 2000

3
8

53

Table 1. (continued)
Series and s~ecies
CIP No.
Collector's
No.

2n

S. coelestipetalum (cap)
24
CIP 761660 OCH 13596
24
CIP 761728 OCH 13686
24
CIP 762005 OCH 14351
S. huancabambense (hcb)
24
CIP 761238 OCH 11619
24
CIP 761238 OCH 11619
24
CIP 761239 OCH 11626
24
CIP762123 OCHS 14815
S. huarochiriense (hro)
24
CIP 761215 OCH 11325
CIP 761224 OCH 11335
24
24
CIP 761224 OCH 11335
24
CIP 761265 OCH 11699
S. leptophyes (lph)
24
CIP 761766 OCH 13730
24
CIP 760805 VSAL 126
24
CIP 760895 VSH 248
S. marinasense (mrn)
24
CIP 761669 OCH 13608
24
CIP 761774 OCH 13737
24
CIP 761812 OCH 13809
S. medians var. autumnale (aut)
24
CIP 761198 OCHS 11302
CIP 762629 OCHS 12573
24
24
CIP 762256 OCHS 15205
S. microdontum (mcd)
24
CIP 762314 OCHS 15534
24
CIP 760531 HAM 176
CIP 760534 HAM 187
24
S. mochiquense (mcq)
CIP 761037 OCHS 14870
24
S. multiinterruptum (mtp)
CIP 761260 OCH 11693
24
24
CIP 761422 OCHS 12055
24
CIP 762405 OCHS 15716
S. oplocense (opl)
CIP 761352 OCH 11927
24
48
CIP 761362 OCH 11947
48
CIP 761362 OCH 11947
S. orophilum (orp)
24
CIP 761440 OCH 12077
CIP 761443 OCH 12080
24
CIP 761475 OCH 13020
24
S. sparsipilum (spl)
CIP 760475 HHA 6669a
24

54

Research on Patato

EBN 1

No.
Country
of origin tested2

Proportion
No.
resistant selected 3
genotypes
(m+r+O)
/No.

o
6
3
o o o
o 1 o

0.19
0.00
0.02

o
o
o
o
o
o
o
o
o
o
o

0.35
0.15
0.46
0.06

9
5
20

o o
1 o
1 o

0.18
0.13
0.44

6
39
42

33
9
5

0.88
0.19
0.13

48
48
48

37
45
16

6
1
17

5
2
15

PERU

48

11

25

12

2 PERU
2 PERU
2 PERU

48
48
48

26
2
19

18
9
27

2
24
2

2 BOLIVIA
4 BOLIVIA
4 BOLIVIA

45
44
48

11
44
48

2 PERU
2 PERU
2 PERU

48
48
48

o
1
2

17
37
22

2 BOLIVIA

48

46

2 PERU
2 PERU
2 PERU

48
48
48

39
48
47

2
2
2
2

PERU
PERU
PERU
PERU

48
48
48
48

31
41
26
45

2
2
2
2

PERU
PERU
PERU
PERU

34
48
48
48

33
44
46
48

2 PERU
2 BOLIVIA
2 BOLIVIA

48
46
46

48
41
38

2 PERU
2 PERU
2 PERU

51
47
48

42
41
27

2 PERU
2 PERU
2 PERU

48
48
48

2 BOLIVIA
2 BOLIVIA
2 BOLIVIA

14
4

10
2

3
3
12
1

o
o
o
o o
o o
o
5
o
8
1
4
2

o o
1 o

6
3

o
5
o
o
o
1
3

o
o
o
o
o
2

0.23
0.06
0.67

15
6
38

0.77

2
13

0.46
0.96
0.60

3
14

o
o
24
o o
10

0.00
0.11
0.17

o
o
o
o

o
o
26
8
o o o
o o o
28

0.03
0.08
0.04
0.00

o
o

0.76
0.00
0.00

o
8

o
o

1.00
0.98
0.96

1
2

0.04

Table 1. (continued)
Series and s~ecies
CIP No.
Collector's
No.

2n

EBN 1 Country
No.
of origin tested 2

Proportion
No.
resistant selected 3
genotypes

(m+r+O)
/No.
S. velardei (vlr)
CIP 761730 OCH 13688
CIP 762027 OCH 14387
CIP 762028 OCH 14387a
S. wittmackii (wtm)
CIP 761566 OCH 13267
CIP 761205 OCHS 11309
CIP 762077 OCHS 14626
YUNGASENSA
S. berthaultii (ber)
CIP 761390 OCH 12008
S. chacoense (che)
CIP 761399 OCH 12026
CIP 762270 OCH 15266b
CIP 762274 OCH 15271
S. tarijense (lar)
CIP 761007 OCH 12001
CIP 762334 OCHS 15596

24

PERU
PERU
PERU

48
48
48

20
23
32

23
25
15

o
5
o o
1 o

0.58
0.52
0.33

24
24
24

PERU
PERU
PERU

48
48
48

12
26
47

32
21
1

3
1

o
o o

0.75
0.46
0.02

24

2 BOLIVIA

44

23

15

0.48

15

24
24
24

2 BOLIVIA
2 PARAGUAY
2 PARAGUAY

44
48
48

27
26
41

10
16
7

7
5

0.39
0.46
0.15

5
3

24
24

2 BOLIVIA
2 BOLIVIA
Total

24

48
48
6306

o o
o
4
o o

9
35
47
1
3532 1316 685 743

0.81
0.02

o
o
o
o
o

o
10
3
801

Note: Accessions noted in bold were evaluated in duplicate.


s = susceptible; m = moderately susceptible; r = resistant; O = qualitatively resistant.
1 EBN = endosperm balance number.
2 No. tested = Number of plants tested in true seed experiments.
3 No. selected = Number of genotypes selected following tuber experiments.

of experiments. After inoculation, conditions were maintained conducive to late


blight by the delivery of cool mist through
an overhead sprinkler system operating on
a thermostat. Shade cloth was used to help
keep temperatures in the range of 15-20C.
The proportion of total leaf area with
symptoms was recorded for each individual plant at two-day intervals beginning
on the third or fourth day after inoculation,
for a total of three or four readings per
experiment. A contact fungicide was
appl ied to stop the disease when the
susceptible controls reached 60% foliar
infection.
Relative area under the disease progressive curve (rAUDPC) was calculated for
each individual plant from the ratings of
foliar infection (Fry, 1978). Analyses of
variance were conducted on the rAUDPC

values of the standards (SQRs) for each


greenhouse experiment, and the estimated
least squares means were used to find the
SQRs ranking within each experiment.
Spearman correlation coefficients calculated on these rankings were used to
compare pairwise performance of standards between greenhouse experiments.
Discriminant analyses for the greenhouse
experiments were conducted to appreciate
natural grouping of the SQRs into relative
resistance categories, using data from the
true seed and tuber phases.
lndividuals with infection levels of 40% or
lower by the third reading and AUDPC
values lower than or equal to the moderately resistant controls (SQRs) within each
experiment were selected for evaluation
from tuber-grown plants in the second

phase of the study. These tuber experi-

CIP Program Report 1999- 2000

55

ments followed the same guidelines as the


true seed experiments, except for the use
of one to three tuber-grown clones of the
putatively resistant genotypes. For each
accession in the first (true seed) and
second (tuber) phase, the response pattern
to environment was studied through a
frequency distribution of their ADUPC
values.
lndividuals of each accession, within a
particular experiment, that became
infected with the pathogen were assigned
one of three disease reaction levels
(resistance ratings): susceptible (s), moderately resistant (m) and resistant (r); based
on the comparison of their rAUDPC values
with those of the SQRs in each category of
resistance. An additional rating, O (qualitative resistance, not represented in the
standards) was used for individuals that
showed either no visible disease reaction
or necrotic flecks assumed to be indicative
of race-specific resistance. The proportions
of total resistant, quantitatively resistant,
and qualitatively resistant individuals were
cal cu lated as:

r + O , m + r , and _Q , respectively,
n
n
n

m +

where n indicates the number of individuals inoculated in the True Seed


ex peri ments.
In arder to assess the representative nature
of the greenhouse assay, performance of
the standards (SQRs) in greenhouse
experiments was compared, through
Kendall's correlation coefficient, with their
reaction to late blight in five field trials at
Comas, Peru. The field trials were performed in randomized complete block
designs using 4 replications of 5-plant
experimental units, their full analysis will
be presented elsewhere.

Results
lnitial symptoms of infection were observed three days after inoculation, and
susceptible individuals reached 100%

56

Research on Potato

foliar infection by day 8. Spearman's


correlation coefficients for pairwise
comparison of the rAU DPC rankings of the
controls (SQRs) among the 11 greenhouse
experiments ranged from 0.25 (P > O.OS) to
0.71 (P < 0.05) for true-seed experiments;
from 0.40 (P > 0.05) to 0.86 (P < 0.05)
for tuber experiments; and from 0.21
(P > 0.05) to 0.87 (P < 0.05) between trueseed and tuber experiments. Despite
moderate variations in rankings of the
SQRs, expressed as a low correlation
coefficient, discriminant analysis on the
greenhouse results identified three consistent response categories: susceptible
(7 clones), moderately resistant (S clones)
and resistant (1 clone) across al 1 greenhouse experiments. AUDPC values
corresponding to each of these groups in
each experiment were used to align the
responses of the test accessions within the
categories s, m, and r.
Figure 1 illustrates the outcome of the
correlation analyses conducted to compare
the average performance of the SQRs in
the 11 greenhouse experiments with their
average performance in the five field
experiments in Peru. The median rankings
for true-seed, tuber, and field experiments
were obtained for each control and
Kendall correlation coefficients were
calculated among these three sets.
Kendall's coefficient ranged from 0.47
(P < 0.05, between true-seed and field
experiments), to 0.62 (P < 0.05, between
tuber and field experiments), to 0.70
(P < O.OS, between True-Seed and Tuber
experiments.)

Resistance patterns observed in tests of


true seed-derived accessions
The majority of seed-derived accessions
showed variation in their levels of resistance, as would be expected from their
heterogeneous constitution. However,
because individual plants were used to
represent genotypes within accessions,
sorne experimental error is confounded in
the measures of variability (see Discussion). Of the 139 series of plants tested,

Distribution of resistance types: r vs. O


13
12

f:

**

11

O>
e 10
::.i2
9
e

e:

(1)

.f

c..
X

(1)
(1)

(/)

(1)

:::J

i=

4
3
2

* *

(1)

"O

11

..

7
6

13
12

10 L
O>
e
9 ::.i2
8 e
7

6
5
4
3
2
1

e:
.fE
(1)

c..
X

(1)

(j
.o
:::J

1-

o
o

2 3 4 5 6 7 8 9 10 11 12 13
Field experiment ranking

Figure 1. Ranked perfonnance of 13 standards for


quantitative resistance to late blight evaluated in each
of the two sets of greenhouse experiments used to
~valuate ge:mplasm accessions, versus their ranking
rn a set of f1eld experiments in Comas, Peru. (* =
true-seed experiments; = tuber experiments). The
salid line indicates the hypothetical case of equal
rankings of the standards in the greenhouse and field
experiments.
11 O (> 80%) showed mixed responses
(more than one response category) and 29
showed clear, single category responses
(22 totally susceptible and 7 totally
resistant). When the data from all the
accessions are considered together, the
majority of individuals (3532 plants out of
a total of 6306 tested) fell into the most
susceptible (s) class. In addition, the
frequency distributions of most accessions
were skewed toward the more susceptible
categories (s and m) (Table 1). However,
16 species each contained at least one
accession that showed moderate resistance
(m) and resistance (r), including accessions
of S. cajamarquense, S. orophilum,
S. raphanifolium/ S. cantense and
S. violaceimarmoratum (Table 1 ). Additio~al promising accessions showing more
res1stant than susceptible individuals were
encountered among those of:
S. circaeifolium, S. commersonii
S. demissum and S. hougasii/
'
S. iopetalum/ S. irosinium/
S. megistacrolobum/ S. multiinterruptum/
S. sogorandinum/ and S. stoloniferum.

Data for the true seed-grown plants shows


that 26 of the 51 species evaluated each
contained sorne individuals that were not
visibly infected (rated O) with the
P. infestans isolate PC0002, despite its
complex virulence nature. Although we
have not tested these individuals against
other isolates and histopathological
o~servations were not made, we expect
th1s type of reaction to reflect qualitative
resistance. This interpretation, as opposed
to the possibility of escapes from inoculation, is strengthened by the distribution of
individuals with this rating in the greenhouses, which could not be attributed to
obvious experimental conditions, and the
low coefficients of variability (< 30%) for
resistance data taken as the standards in
most experiments.
In addition to its expected occurrence in
species from Mexico, qualitative resistance (rating O) was observed in 19
accessions native to South America. Al 1
individuals of one accession of each of the
South American species 5. piurae and
S. circaeifolium, and of five accessions of
four Mexican species - S. cardiophyllum,
S. polyadenium/ S. stoloniferum and
S. iopetalum showed qualitative resistance
(O). The latter species was represented by
two accessions that appeared uniformly
incompatible with our isolate (48/48
individuals rated O), and one accession
that expressed quantitative variation. Both
qual itative and quantitative types of
resistance were observed within and
among at least 18 Mexican and South
American accessions. Figure 2 presents an
overview of the distribution of the resistance types observed in the tested
accessions in terms of percent of individuals that showed quantitative and
qualitative disease responses.
In all, 1868 individuals from 93 accessions
in 45 species classified as m, r, or O in the
experiments with true seed-grown plants,
were selected for re-evaluation using
plants derived from tuber-grown plants.

CIP Program Report 1999- 2000

57

"*-e

0.9

:e

0.8

::::J

:;:
-~

e:ro
()

"C

.....

~.._

* *

...

'

0.7 -

"-l ,,

0.6 ,_

>

.s

0.5 ,_

0.4

0.3

~
ro

~o.

a..

0.2

0.1

~~

,.,,,

**

-~

***

__

,,

~-

.,

MJo:

~*
~~
1

li

-:;:

:e
(])

- 0.6

:e
:g_

0.5

o
en
::::J
en

- 0.4

0.3

:eo

- 0.2

a..

(])

e:

._,

*'
f!"I***

..

en

(ij
"O

- 0.7

>

; * * * "'
,.**

0.9

- 0.8

*-

::::J

'*

"'
** "'
.,,,

*"" *""" "' *


~.. *
,,
'*
*'*

Q)
~

* '*

..

*
*

1.0

'*

)~.

11

I!

11

o.

- 0.1
1

1
1

O.O

F_igure 2. P!o~ortion~ of individuals in each o'.139 germplasm samples tested in true-seed experiments falling
mt~ quant1tat1ve res1st~nce (m + r), susceptible (s) and qualitative resistance (O) categories. (* = quantitatively
res1stant; = susceptible; 1 = qualitatively resistant).

Evaluation of tuber-grown plants of


individuals selected from resistant
accessions
Genotypes of S. piurae, S. iopetalum,
S. hougasii, S. fendlerii, S. cardiophyllum,
and 5. circaeifolium accessions collected
in Mexico, Bolivia, and Peru showed
incompatible responses (48/48 individuals
rated O) in both the true seed- and the
tuber-derived inoculations, indicating
the likely presence of R genes with which
PC0002 is incompatible. Genotypes of

S.
S.
S.
S.
S.
S.
S.

albicans, S. alandiae, S. ambosinum,


circaeifolium, S. chiquidenum,
cajamarquense, S. commersonii,
coelestipetalum, S. microdontum,
huancabambense, S. megistacrolobum,
toralapanum, S. multiinterruptum,
sogorandinum and S. stoloniferum

collected in Peru, Bolivia, Colombia,


Uruguay, Paraguay, and Mexico showed
moderate and variable response to the
disease when evaluated from tubers,

58

Research on Potato

confirming the likely presence of quantitative resistance. Eight hundred and one
promising genotypes in 35 species of 11
taxonomic series were selected by comparison of the results obtained from the
inoculation of plants grown from true
seed and from tubers. These selections
are currently being propagated for resistance evaluation at the genotypic level
(Table 1 ).

Discussion
Limitations to the greenhouse assay
The greenhouse assay was a convenient
way to evaluate large samples of
germplasm under applied disease pressure.
However, susceptible individuals reached
100% foliar infection only eight days after
inoculation. In contrast, it takes 3-4 weeks
for moderately resistant or susceptible
clones to reach 80-100% infection in the
field. The rapid greenhouse epidemics
may not provide the best conditions to

differentiate realistically between genotypes. This population-based evaluation


has therefore permitted a broad overview
of the resistance levels in the germplasm,
but validation in field trials is still required. Replicated trials under controlled
conditions would also permit resistance
components to be measured, elucidating
factors that may vary i ndependently
among different sources of resistance.

Variability in resistance types within and


among accessions
Disease ratings of O were observed among
26 species from South America and
Mexico. lt is likely that the O rating
reflects the presence of major resistance
(R) genes that are not matched by the
virulence genes of the test isolate used,
masking quantitative resistance that might
be detectable with compatible isolates.
Screening with a range of isolates would
confirm whether the individuals with O
ratings carry one of the known R genes
(e.g., R5 or R9) with which our isolate was
not compatible, or whether their resistance
was due to other factors.
Elsewhere, genetic mapping in an interspecific hybrid population has revealed
the presence of both qual itative and
quantitative resistance (Ewing et al., 2000),
and permitted description of three apparently novel R genes from S. berthaultii
(Sanchez et al., 2000; W.E. Fry and
G. Sanchez, Cornell University, personal
communication). Given this precedent, it
is quite likely that more R genes than
those reported in S. demissum and
S. berthau/tii are present in Solanum
germplasm. Selected genotypes will
therefore be analyzed further using a set
of differential isolates with known interaction patterns with the currently recognized
R genes.

lmplications of intraspecific variation for


late blight resistance
Better knowledge of patterns of variability

for resistance to late blight among differ-

ent accessions of the same species can


help to develop appropriate strategies for
the conservation and use of genetic
resources. The differences between
accessions of the same species observed in
this study might be explained by: (1)
genetic differences among source populations; (2) sampling error in selecting seeds
for testing accessions, (3) genetic drift
associated with the conservation procedure, or (4) experimental error due to lack
of replication. In the first case, the observed variation might stem from different
pressures acting on the common gene
pool, such as the influence of different
habitats. For example, S. sogarandinum,
was represented by two accessions collected from localities on the eastern
(OCH 13336) and western (OCHS 15723)
slopes of the Andes. When inoculated in
the same greenhouse experiment, the
accession collected at the more easterly
site (OCH 13336), exhibited higher levels
of resistance; this might reflect adaptation
to a more humid environment and higher
blight pressure. S. fendlerii also presented
strikingly different resistance patterns
between accessions collected in different
locations, reinforcing the need to take
account of potential variability among
isolated populations when devising
sampling strategies that seek to represent
widely distributed species.
The process of sampling seeds from
variable gene bank accessions may induce
errors. There is evidence that sampling
errors may have occurred in this study in
cases when duplicate samples of seed
from the same regeneration cycles were
sown as back-ups following slow germination. Two samples of accession
OCHS 11915 (S. circaeifolium), and two
of SCL 5050 (5. co/ombianum) showed
variability in resistance levels that was
probably dueto sampling error. However,
duplicate samples from accessions in four
other species presented nearly identical
patterns of resistance. Although they
differed from one another in resistance

CIP Program Report 1999 - 2000

59

pattern, the two pairs of duplicate samples


of S. circaeifolium accession OCHS 11915
were consistently more resistant than the
other two accessions tested from this
species. Different sample sizes may be
required to represent accessions of which
between 200 and 6000 seeds are avai lable
in the gene bank, especially if different
diversity levels are expected among the
species due to breeding habit or other
factors affecti ng the genetic structure of
the source populations. Genetic drift also
has important implications for germplasm
conservation. Renovation procedures in
the gene bank may result in unintentional
selection for linked features and hence
affect patterns of resistance. The sampling
procedures used in this study did not
provide the opportunity to examine
genetic drift.
Finally, the possibility of escapes from
infection or variation within the testing
environment cannot be ruled out as
possible causes of this apparent genetic
variability. An alternative testing strategy
involving prior establishment of clonal
materials would afford more robust evaluations of genotypes, but would significantly
reduce the scope or increase the cost of a
germplasm survey.

The value of rare species


These experiments provide new information concerning the distribution of levels
and types of resistance in 12 of 18 littleknown species previously suggested to
carry resistance. For example, the suggestion by Van Soest (1984) that two species
from Peru, 5. chiquidenum and
S. multiinterruptum carry race non-specific
resistance has been confirmed by the
demonstration of quantitative resistance to
our test isolate. A large number of additional genotypes have been identified that
showed quantitative resistance in tests
conducted on both true seed- and tubergrown plants (Table 1). Species which
have not been evaluated for this form of
resistance before and showed high levels
of resistance in these experiments include:

60

Research on Potato

S.
S.
S.
S.

urubambae/ S. violaceimarmoratum/
cantense/ S. cajamarquense,
orophilum, S. velardei and
wittmackii. We therefore report the first

evidence of resistance to late blight in 7


species endemic to the South American
center of origin of potatoes.
Sorne accessions showed unfamiliar
responses to infection with P. infestans. For
example, S. colombianum presented
watery lesions on the underside of its
leaves; 5. megistacrolobum developed
irregular, streak-like lesions on the upper
leaf surface and small lesions on the
underside; and S. mochiquense presented
smal 1 water-soaked lesions on both leaf
upper. and lower surfaces. These un usual
reactions suggest different types of hostpathogen interactions, which may indicate
resistance mechanisms that could be
exploited in genetic improvement
programs.

Utilization of new diversity/exotic species


The potential for using these genetic
resources in conventional breeding
depends on their ability to cross breed with
S. tuberosum, the inheritance of the
resistance they express, and the strength of
associations between desirable and
undesirable traits that might be encountered in the course of an enhancement
program. Ability to cross breed is greater
among species with compatible endosperm balance numbers (Johnston and
Hanneman, 1980) (see Table 1), with the
frequency of unreduced gametes influencing the success of interploidy crosses.
Polygenes, which are difficult to maintain
intact in breeding, are believed to underlie
quantitative resistance. lt is thus desirable
to identify sources with high heritability,
and to ensure robust screening procedures
that are suitable for large numbers of
hybrid individuals if introgression or
upgrading programs are proposed.
The poor tuberization of sorne accessions
led to the loss of sorne promising genotypes identified in the true seed

experiments. For example,


S. chomatophilum and certain accessions
of S. albicans, S. ambosinum,
S. circaeifolium, and
S. dolichocremastrum produced no or only
a few short-1 ived tubers and therefare
could not be re-evaluated as tuber-grown
plants. lf crosses with such donors are
accomplished in breeding programs, the
genetic association of desirable and
undesirable traits in hybrid populations
may limit advances in improving resistance and agronomic characteristics. For
example, Ochoa (1999) propases
S. chiquedenum as useful far its resistance
to late blight and earliness, but its moniliform tuberization habit (formation of tubers
in chains) is likely to be a detracting
horticultura! characteristic in breeding
programs. On the other hand, it is notable
that severa! species presenting resistance
in these evaluations produce large tubers
in their wild states (e.g., S. urubambae,

S.
S.
S.
S.
S.

violaceimarmoratum,
cajamarquense, S. coelestispetalum,
medians, S. microdontum,
multiinterruptum, S. oplocense, and
orophilum). This may facilitate their use

in breeding.
This population approach to germplasm
evaluation provides important infarmation
needed to select and optimize strategies
for the use of new resistance sources in
breeding. For example this effort will
orient the selection of new donors of
resistance for tests of heritability and
complementarity with currently deployed
resistance types. The use of multiple
genotypes of a given donor species and a
carefully chosen set of diverse recurrent
parents may help circumvent the poorly
understood genetic interactions between
wild and cultivated species that often
result in fertility problems in F1 and F2
generations (Santini et al., 2000). The
selection of diverse donar species using
taxonomic, ecogeographic and mechanistic criteria should also contribute to
building complex resistance types that are
not readily overcome by the pathogen.

Tools of modern genetics including


molecular diversity surveys, testing
hypotheses on the role of candidate genes
far resistance and defense, and comparative mapping, will be valuable in helping
to develop strategic combinations of
resistance sources, and possibly lead to the
identification and direct transfer of superior alleles from wild to cultivated genetic
backgrou nds.

Acknowledgments
This work was supported in part by the
Department for lnternational Development
(DFID), United Kingdom. We thank E. de
la Torre, H. Ponce, G. Marticorena and
O. Gaspar far technical assistance in the
greenhouse and laboratory, V. Otazu far
input on design and execution of the
experiments, C. Arellano for statistical
consultation, and R. Hijmans, A. Panta,
and E. Mihovilovich for helpful comments
on the manuscript.

References
Colon, L.T. and Budding, D.J. 1988.
Resistance to late bl ight (Phytophthora
infestans) in ten wild Solanum species.
Euphytica Supplement p. 77-86.
Colon, L.T., D.J. Budding, L.C.P. Keizer,
and M.J.J. Pieters. 1995. Components of
resistance to late blight (Phytophthora
infestans) in eight South American
Solanum species. European Journal of
Plant Pathology 101 :441-456.
Ewing E.E., l. Simko, C.D. Smart,
M.W. Bonierbale, E.S.G. Mizubuti,
G.D. May, and W.E. Fry. 2000. Genetic
mapping of qualitative and quantitative
field resistance to Phytophthora
infestans in a population derived from
Solanum tuberosum and Solanum
berthaultii. Molecular Breeding
6:25-36.
Fry, W.E. 1978. Quantification of
general resistan ce of patato cu ltivars
and fungicide effects for integrated
control of late blight. Phytopathology
68:1650-1655

CIP Program Report 1999 - 2000

61

Hanneman, R.E. and J.B. Bamberg. 1986.


lnventory of tuber-bearing Solanum
species. Univ. of Wisconsin. Madison,
WI, USA. 216 p.
Hawkes, J.G. 1950. Algunas observaciones
sobre la papa del Ecuador. Flora
7:93-96.
Johnston, S.A. and R.E. Hanneman. 1980.
Support of the endosperm balance
number hypothesis utilizing sorne tuberbearing Solanum sp. American Potato
Journal 57:7-14.
Niederhauser, J.S. 1989. Phytophthora
infestans: The Mexican connection. In:
Lucas, J.A., R.C. Shattock, D.S. Shaw,
and L.R. Cooke (eds.). 1989.
Phytophthora. Cambridge University
Press. UK. p. 25-45.
Ochoa, C. 1954. Northern Per, a possible
new source of potatoes resistant to
Phytophthora infestans. Phytopathology
44:500.
Ochoa, C. 1981. Solanum irosinum, new
Peruvian tuber-bearing So/anum species
resistant to Phytophthora infestans.
American Potato Journal 58:131-133.
Ochoa, C.M. 1999. Las papas de
Sudamrica: Per. Allen Press, KS, USA.
1036 p.
Ross, H. 1986. Potato breeding - problems
and perspectives. Advances in plant
breeding. Supplement 13 to Journal of
Plant Breeding. p. 11-18.
Sanchez G., C.D. Smart, l. Simko,
M. Bonierbale, E.E. Ewing, et al. 2000.

62

Research on Potato

ldentification of two new R-genes to

Phytophthora infestans from Solanum


berthaultii. Phytopathology 90:569.
Santini, M., E.L. Camadro,
O.N. Marcellan, and LE. Erazu. 2000.
Agronomic characterization of diploid
hybrid families derived from crosses
between haploids of the common potato
and three wild Argentinian tuber-bearing
species. American Journal of Patato
Research 77:211-218.
Schober, B. 1981. Phytoalexine in Wild
arten van Solanum. Sth Triennial
Conferences of the EAPR (Mnchen).
Abstracts of CorJference papers.
p. 36-37.
Tazelaar, M.F. 1981. The screening of
Solanum species for horizontal
resistance against late bl ight
(Phytophthora infestans) and its use
for breeding programs. Sth Triennial
Conference of the EAPR (Mnchen).
Abstract of Conference papers.
p. 34-36.
Toxopeus, H.J. 1964. Treasure digging for
blight resistance in potatoes. Euphytica
13:206-222.
Van Soest, L.J.M. 1983. Evaluation and
distribution of important properties in
the German-Netherlands patato
collection. Potato Research 26:109-121.
Van Soest, L.J.M. 1 984. Resistance to
Phytophthora infestans in tuber-bearing
species of Solanum and its geographical
distribution. Potato Research 27:393-411.

Quantifying Genetic Variance for Horizontal


Resistance to Late Blight in Potato Breeding
Population B3C1
J.A. Landeo, M. Gastelo, G. Beltran, and L. Diaz 1

Monitoring the genetic variation for horizontal resistance to late blight in the
most advanced sources of horizontal resistance to late blight-population B
group three cycle one (B3C1 )-has been carried out at CI P by assessing the
amount of genetic variance for resistance and tuber yield. A random sample
of resistant clones from B3C1 were used to obtain progenies according to
mating designs NC Design 1 and Line x Tester. Resulting progenies were
evaluated for resistance to late blight and total tuber yield in the field under
high disease pressure at Comas, Peru. Three independent estimates of heritability, including parent-offspring regression for area under the disease
progress curve (taken as :.. parameter for resistance), were high and reasonably close (h 2 = 0.48, h2 = 0.53, and h2 = 0.40). Additive genetic variance and
heritability for resistance in this population are large enough to ensure further
progress from selection.

In 1990, CIP began a program to improve


potato populations by increasing gene
frequencies for quantitative (horizontal)
resistance to late blight. This program also
aimed to more systematically upgrade and
maintain desirable characteristics of
economic importance, such as tuber yield,
dry matter content, early tuberization, and
bulking.
Horizontal resistance to late blight was
considered to be the major crop-protection
trait needed in new varieties to face the
increasing threat of a re-emergent potato
disease. lt was decided to remove or avoid
the inclusion of known dominant genes
responsible for race-specific resistance (R
genes). Although highly effective when
compatible races are not present, R genes
have a long history of unstable resistance
due to continuous changes in the pathogen
1

CIP, Lima, Peru.

(Malcomson and Black, 1966; Van der


Planck, 1968; Waistie, 1991; Goodwin et
al., 1995; Fry and Goodwin, 1997), and
they can interfere with the recognition of
true horizontal resistance (Landeo and
Turkensteen, 1989). On the other hand,
horizontal resistance to late blight is a
quantitative trait, understood to be nonrace-specific, effective against all
variants of the pathogen, and therefore
more stable and durable (Turkensteen,
1993; lnglis et al., 1996; Forbes, 1998;
Haynes et al., 1998; Landeo et al., 2000).
This type of resistance is being improved
at CIP, with the absence of R genes in the
breeding population, which allows field
screening against local pathogen populations without sacrificing efficiency. The
pathogen isolates sampled from local
testing sites in Peru belong to mating type
A 1 and represent a new popu lation of
Phytophthora infestans/ which is gradually
replacing the old population in the country
CIP Program Report 1999 - 2000

63

(Forbes et al., 1997). This population is as


aggressive as the new migrating population described elsewhere (Fry et al., 1992),
quite diverse despite its asexual reproduction, and resistant to the systemic
fungicide metalaxyl (Perez et al., 1998).
At present, the most advanced source of
horizontal resistance to late blight at CIP
(Population B3) derives from Population A
(previously the most advanced source),
which contained R genes. The presence of
the R genes made upgrading quantitative
resistance cumbersome. The B3 population
to date has undergone three cycles of
recombination. Useful traits include
horizontal resistance to late blight, tuber
yield, dry-matter content, early
tuberization and bulking, and good quality
for potato fries and chips. Currently, to our
knowledge, B3 carries only quantitative
resistance to late blight, and it is constantly monitored to maintain sufficient
genetic variation to ensure further progress
and selection of outstanding clones with
high levels of resistance and varietal
potential.
The results reported here represent findings
from this study which monitored the
genetic variation for quantitative resistance to late blight and for tuber yield in
the breeding population B3 cycle 1
(B3C1 ). More specifically, the study
estimated additive genetic variances and
narrow-sense heritabi 1ities, i mportant
parameters for quantitative traits.

Materials and Methods


Two random samples of clones from a
resistant group of population B3C1 were
taken for crossing following mating
designs North Carolina Design 1 (D-1) and
Line x Tester (LxT). For the D-1, three
groups of six females with one male each
were created; for the LxT, 27 females with
three independent males were used as
testers. The progeny were seedlingtransplanted to the field at La Molina
during the winter season in 1998 to obtain
tuber fami lies (first clonal generation).

64

Research on Potato

Subsequent trials were under severe


natural late-blight pressure in the central
highlands of Peru (Comas) during the rainy
season in 1999. The D-1 progenies were
arranged in a randomized complete-block
design (RCBD) with 2 replications and 40hill plots per family. The LxT progenies
were arranged in a simple incomplete
lattice, 9x9, with two replications and 40hill plots per family. The complete set of
parents was also evaluated separately in a
RCBD with three replications and 10-hill
plots per clone. Spraying mancozeb
(Dithane M-45) when 90% of the plants
had emerged and again one week later
provided protection against early lateblight attack. Percent of foliage infection
was assessed at weekly intervals for
individual genotypes (mating designs) and
for 10-hill plots (for parental clones),
beginning 10 days after the last fungicide
application and for seven consecutive
weeks thereafter.
The area under the disease progress curve
(AUDPC) was calculated from the weekly
seores using an Excel spreadsheet (Landeo
et al., 1996). The AUDPC was used as a
parameter for resistance. For progenies in
mating designs, AUDPC averages were
obtained from individual genotypes. At
harvest, 11 O days after planting, total tuber
yields were taken on 40-hill plots for
families and on 10-hill plots for parents.
An SAS/STAT program (SAS lnstitute lnc.,
1989) was used for analysis of variance of
D-1 and LxT. The estimates of genetic
parameters and heritabi 1ities for both latebl ight resistance and tuber yield were
derived from the expected mean squares
of D-1 and LxT, as was their corresponding
equivalence to genetic components of
variance from the covariance of relatives
under a tetraploid-inheritance model.
In addition to the two mating designs,
parent-offspring (PO) regression was also
included to calculate heritability for
resistance and as a means of comparing
estimates of heritabilities. For PO regression, the average AU DPC of both parents
was regressed on the averaged AU DPC of

their corresponding offspring tested in the


LxT design to obtain the regression coefficient (b) that equals the Heritability (h 2 ) of
the trait.

environment interaction was not extracted,


the values obtained are in agreement with
estimates from the previous cycle, B3CO
(Landeo et al., 1996).

Results

With ali three methods, the heritability


estimates for quantitative resistance to late
blight, expressed by the AUDPC as a
parameter of resistance, are quite high for
a quantitative trait and are in close
agreement (Table 4). The small differences
may be accounted for by sampling error.
The components of the additive genetic
variance for AUDPC are also high in all
three methods.

The analyses of variance for AUDPC in all


designs (Tables 1, 2, and 3) indicate that
the sources of variation associated with
genetic variance for quantitative resistance to late blight, males and males/
females in 0-1; clones and testers in LxT,
and source due to regression in PO, are
statistically highly significant. Therefore,
genetic variances were calculated.
Although estimates of additive genetic
variances may be somewhat inflated
because the source of the genotype x

Table 1. Analysis af variance far AUDPC following


the NC Design 1 (D-1) mating design.
Source of
variation
Replicatians
Males
Females/males
Error
Total

c.v.= 9.8%

df
1
2
15
17
35

MS

87714.694
264299.194
71709.428
16088.694

5.45 **
3.68 **
4.46 **

Table 2. Analysis af variance far AUDPC follawing


the line x tester (LxT) mating design.
Source of
df
MS
F
variation
Replications
1 200696.321
16
27862.196
Block/replicatians
Clones
26 529275.872
4.91 **
Testers
2 857994.963
7.97 **
Clones x testers
52 107652.777
3.79 **
Error intrablack
64
28507.555
Total
161
C.V. = 10.8%

Table 3. Analysis af variance far AUDPC following


the parent-affspring (PO) regressian.
Source of
df
F
MS
variation
Regressian
1
3919634.531
46.61 **
Residual
79
84098.059
Total
80
C.V.= 18.6%

The amount of additive genetic variance


for AUDPC in Population B3 is large
enough for continued progress as we carry
on with further cycles of recombination.
Likewise, the high heritability of the
quantitative resistance to late blight in this
population, obtained by the three methods,
indicates that genetic variability is high
and has not been exhausted or reduced.
Therefore, continued progress can be
expected.
For tuber yield, as far as genetic variance
and heritablities are concerned, this is
apparently not the case. However, quantifyi ng this character and its transmissibility
under late-blight pressure is not entirely
appropriate. lt would be more appropriate
to examine this character in the absence
of late-blight pressure to obtain components of variance and heritability
estimates that would reflect the actual
gene frequency and its magnitude in this
population. We have already planned
experiments to this end and have obtained
preliminary results indicating a significant
amount of additive genetic variance for
yield and high heritability.

Discussion and Condusions


The magnitude of additive genetic variance for horizontal resistance to late
blight, obtained by the different approaches, is large, and the narrow-sense
heritabilities are high enough to lead us to
expect progress from further selection. The
CIP Program Report 1999- 2000

65

Table 4. Components of additive genetic variances and heritability estimates far AUDPC and tuber yield in
samples of B3C1 late blight resistant clones.
AUDPC
Tuber yield
h2
h2
Mating design
SE
SE
100384
0.48
0.45
3.78
0.10
0.17
Design 1
0.18
0.25
0.08
281082
0.53
65.40
Line x tester
Parent-offspring
19407166
0.40
0.06

recurrent selection scheme as applied to


population B3-where resistance to late
blight is quantitatively inherited and
manipulated in the population--ensures
steady progress (Figure 1 ). Comparing the
initial cycle of recombination with cycle
one, we find that the mean AUDPC for
population B3C1 (1083) is approximately
one-half of that of the previous cycle,
B3CO (2556), indicating that progress has
been made without undue sacrifice of
genetic variability. We also think that
progress has been made on tuber yield as
well as resistance to late blight.
The improvement of quantitative resistance to late blight in the absence of
dominant genes has undoubtedly contributed not only to the ease of screening for
late-blight resistance against a wide range

of variants in the pathogen population but


also to more accurate recognition of
horizontal resistance, consequently
increasing gene frequencies more efficiently. In spite of arguments indicating
that sorne R genes may not be absent but
have only been suppressed (Ordoez et
al., 1997), the fact is that they have not
interfered with the expression of quantitative resistance and its selection.
Quantitative resistance to late blight has
long been known to be stable and durable
(Forbes, 1998; Haynes et al., 1998; Landeo
et al., 2000). Therefore, CIP's current
strategy is to focus on quantitative resistance as the most valuable source of
resistance. This approach is backstopped
scientifically and has continuously been
shown in the breeding materials at CIP's
testing sites and elsewhere and it goes

Nurnber

50
45

Mean: 2556
sd: 877

40
35
30
25
20
15
10

0500

5011000

10011500

1501-2000

2001-2500

2501-3000

AUDPC

Figure 1. Mean AUDPC compared between cycles B3CO vs B3C1.

66

Research on Patato

3001-3500

3501-4000

beyond theoretical speculation. However,


should any scientific arguments arise on
the value of R genes for broad-based,
durable resistance, our breeding strategy is
open to i ncorporati ng these genes and
others through both conventional and
engineered means. Until then, CIP's work
has shown continuous progress and allows
the selection of outstanding clones with
high levels of resistance which are made
available for distribution and release in
developing countries.
Other important secondary traits, such as
dry-matter content, low reducing sugars,
and early tuber initiation and bulking,
included in the selection process are also
showing progress, although experimental
proof is not yet avai lable.
lt is extremely important to continue
monitoring genetic variance for economically important traits in population B3 to
ensure selection progress and to alert us to
any impending risk of exhausting the
traits. This would cal 1 for efforts to broaden
the genetic diversity for those characters.
Work on improvements in the B3 population, therefore, is closely linked to
developing other sources of diversity from
wild relatives. This is done through prebreeding for introduction at any time it is
necessary. Continuous monitoring of
genetic parameters in every cycle is also
routinely done in the populations under
improvement. In parallel to monitoring
genetic progress and variability, the work
on improving population B3 is closely
linked to the selection of outstanding
clones with desirable table and processing
qualities, as well as high levels of resistance to late blight, for distribution to
developing countries for variety releases
and parental 1ines for local breeding
purposes.

References
Forbes, G.A. 1998. Genotype by
environment reaction of potato to the
late blight pathogen. In: lmpact on a

changing world, Program report 199798. lnternational Potato Center, Lima,


Peru. p. 57-66.
Forbes, G.A., X.C. Escobar, C.C. Ayala, J.
Revelo, M.E. Ordoez, B.A. Fry, K.
Ooucett, and W.E. Fry. 1997. Population
genetic structure of Phytophthora
infestans in Ecuador. Phytopathology
87:375-380.
Fry, W.E and S.B. Goodwin. 1997.
Resurgence of the lrish potato famine
fungus. BioScience 47:363-371.
Fry, W.E., S.B. Goodwin, J.M. Matuzak,
L.J. Spielma, M.G. Milgroom, and A.
Orenth. 1992. Population genetics and
intercontinental migrations of
Phytophthora infestans. Annual Review
of Phytopathology 30:107-129.
Goodwin, S.B., L.S. Sujkowski, and W.E.
Fry. 1995. Rapid eyolution of
pathogenicity within clonal lineages of
the potato late blight disease fungus.
Phytopathology 85:669-676.
Haynes, K.G., O.H. Lambert, B.J. Christ,
O.P. Weingartner, O.S. Oouches, J.E.
Backlund, G. Secor, W.E. Fry, and W.
Stevenson. 1998. Phenotypic stabi 1ity of
resistance to late blight in potato clones
evaluated at eight sites in the United
States. American Journal of Potato
Research 75:211-217.
lnglis, O.A., O.A. Johnson, O.E. Legard,
W.E. Fry, and P.B. Hamm. 1996.
Relative resistances of potato clones in
response to new and old populations of
Phytophthora infestans. Plant Oisease
80:575-578.
Landeo, J.A. and L. Turkensteen. 1989.
Assessment of partial resistance to late
bl ight (Phytophthora infestans) of major
genes in potato. American Potato
Journal 66:530.
Landeo, J.A., M. Gastelo, and G. Forbes.
1996. Screening far horizontal
resistance to late blight in population B.
Specialized Technology Oocument.
lnternational Potato Center, Lima, Peru.
12 p. (Unpublished draft.)
Landeo, J.A., M. Gastelo, E. Roncal, and
A. Mendoza. 2000. Phenotypic stabi 1ity

CIP Program Report 1999 - 2000

67

for horizontal resistance to potato late


blight in population B. American Journal
of Patato Research 77:406.
Landeo, J.A., M. Gastelo, G. Forbes, J.L.
Zapata, and F.J. Flores. 1996.
Developing horizontal resistance to late
blight in patato, Program report 199596. lnternational Potato Center, Lima,
Peru. p. 122-126.
Malcolmson, J.F. and W. Black. 1966.
New R-genes in Solanum demissum
Lindl. and their complementary races of
Phytophthora infestans (Mont.) de Bary.
Euphytica 15:199-203.
Ordoez, M.E., G.A. Forbes, and B.R.
Trognitz. 1997. Resistance to late blight
in potato. A putative gene that
suppresses R genes and is elicited by
specific isolates. Euphytica 95:167-172.
Perez, W., S. Gamboa, M. Coca, R.
Raymundo, R. Hijmans, and R. Nelson.
1 998. Characterization of Phytophthora
infestans populations in Peru. In: lmpact

68

Research on Potato

on a changing world, Program report


1997-98. lnternational Patato Center,
Lima, Peru. p. 31-48.
SAS lnstitute lnc. 1989. SAS/STAT user's
guide, Version 6, Fourth Edition, Volume
2. SAS lnstitute lnc., Cary, NC, USA.

846 p.
Turkensteen, L.J. 1993. Durable resistance
of potatoes against Phytophthora
infestans. In: Jacobs, T. and J.E.
Parlevliet (eds.). Durability of disease
resistance. Kluwer Academic Publishers,
Dordrecht, Netherlands. p. 11 5-1 24.
Van der Plank, J.E. 1968. Disease
resistance in plants. Academic Press,
NY, USA. 206 p.
Waistie, R.L. 1991. Breeding for
resistance. In: lngram, D.S. and P.H.
Wil 1iams (eds.). Phytophthora infestans:
The cause of late bl ight of pota to.
Academic Press, San Diego, CA, USA.

p. 193-224.

The Effect of Nitrogen Fertilization on Potato Late


Blight in the Field
H.S. Jurez 1' 2, J.R. Amaro 3, M.D. Rivera 4, A. Prraga 4, and R.J. Hijmans 1

The effect of nitrogen (N) fertilization on the development of potato (Solanum


tuberosum L.) late blight (LB) caused by Phytophthora infestans (Mont.) de
Bary was studied in Huancayo and Oxapampa, Peru. We had field experiments in which three levels of N (O, 160, and 320 kg/ha) and three levels of
fungicide were applied in all combinations on two potato varieties (18 treatments). In the absence of LB, more N led to more foliage development and
higher yields. N had no effect on the disease when severity was very low or
high. At intermediate levels of disease severity, increased N led to increased
disease. In Huancayo, the net effect of N application (O vs 160 kg/ha) on
yield of the more susceptible variety was negligible because of increased LB.
For the more resistant variety the presence of LB had little effect on yield,
whether no N was applied or N was applied at the rate of 160 kg/ha. Doubling the N rate to 320 kg/ha did not increase yield in the absence of LB, and
decreased it if no fungicides were applied. lf farmers attempt to increase
potato yields using N fertilizer, they may need to use more fungicides or
varieties with higher resistance.
Potato late blight (LB) occurs in most
potato-growing areas around the world, but
it can be particularly devastating in areas
with warm and humid weather during the
growing season (Hijmans et al., 2000).
Partially-resistant varieties can be used to
manage LB, but in most locations additional measures are needed. Fungicides
are commonly used to diminish LB, but
developing-country farmers often cannot
afford enough fungicides, and yield loss
can be very high (Ortiz et al., 1999;
Thiele et al., 1998).
For that reason it is important to evaluate
additional measures that can be taken to
lower LB, measures such as sanitation,
1

CIP, Lima, Peru.


Universidad Nacional Agraria La Molina, La Molina, Peru.
Universidad Nacional del Centro del Peru, Huancayo, Peru.
4
Universidad Nacional Andrs Avelino Cceres, Oxapampa,
Peru.
2

shifting of the growing period out of the


wet season (Devaux and Haverkort, 1987),
and the use of variety mixtures (Garrett
and Mundt, 2000). Another important
variable in sorne disease systems is
fertilization (Marschner, 1997). lnformation
about the effect of ferti 1ization on the
development of LB is incomplete. Most
authors agree that increased phosphorus (P)
and potassium (K) concentrations tend to
lower LB (Awan and Struchtemeyer, 1957),
whereas increased nitrogen (N) tends to
increase LB (Carnegie and Colhoun, 1983;
Phukan, 1993; Rotem and Sari, 1983).
Sorne authors, however, report inconclusve or contrasting results (Lowings and
Acha, 1959; Awan and Struchtemeyer,
1957), or report reduced disease at optimum nutrition (Cohen and Rotem, 1987).
In most studies only one or two components of resistance were evaluated.

CIP Prngram Report 1999 - 2000

69

Research on the effect of N fertilization on


potato LB in the field has only been
reported for Eastern Europe where
Reichbuch et al. (1977) and Sawicka
(1993) found higher infection with increased nitrogen fertilization, in contrast
to Wierzejska-Bujakowska (1994) who
found that the rate of disease increase was
reduced with increased N fertilization.
Understanding the effect of N fertilization
on LB development can be complicated
because there can be a number of direct
and indirect effects. The direct effect
would be changes in susceptibility at the
plant tissue level. lndirect effects of
increased fertilization on LB would be
caused by changes in canopy size.
Changes in canopy size may affect
microclimate and hence LB, because P.
infestans is sensitive to temperature and
humidity, especially humidity (Crosier,
1934). There can also be a distance-effect:
the smal ler the plants, the higher the
fraction of spores that is likely to fall on
the ground rather than on potato leaves,
potentially slowing the epidemic. Canopy
size might also affect the LB epidemic in
yet another way. A larger and faster
expanding canopy may influence disease
severity (which is measured as a proportion of the total canopy). That is, new
leaves could possibly dilute the disease.
A particularly important issue relates to
farmers with a low external-input potato
production system who try to i ncrease
production through N fertilization. LB is
expected to increase, but it is not clear to
what extent the yield gain due to N would
be offset by losses due to LB. For this
reason, we did both field and laboratory
experiments to determine the effect of N
fertilization on LB under tropical highland
conditions far two varieties with different
levels of LB resistance. In this paper we
describe the resu lts of two field experiments with different varieties, levels of N,
and fungicide treatments. We only report
results for disease, green ground cover, and

70

Research on Potato

yield. Effects on microclimate and N


content in the crop tissue will be discussed
elsewhere.

Materials and Methods


Experimental design
Field experiments were carried out at two
sites in Peru: in Huancayo (1201'45" S
latitude, 7514'0" E longitude, 3300 m)
and Oxapampa (1035'31" S latitude,
7523'0" E longitude, 1813 m). In
Oxapampa, the experiment was planted
on 27 Nov 1999, and the last harvest was
on 5 Apr 2000. In Huancayo, the experiment was planted on 1 7 Dec 1999 and the
last harvest was on 2 June 2000.
Two varieties with different levels of
resistance were used: Amarilis (moderately
resistant) and Yungay (moderately susceptible). There were three rates of N
fertilization: O, 160, and 320 kg/ha
(hereinafter referred to as NO, N160, and
N320, respectively). Application was in
bands, half at planting and the other half
at 22 days (Oxapampa) or 32 days
(Huancayo) after emergence, using
NH 4 N0 3 (33% N). There were three levels
of fungicide (F) application: no fungicide
{F-), treatment every 4 days (F4), and
every 1O days (F1 O) in Oxapampa. In
Huancayo, treatments were no fungicides
(F-), treatment every 7 days (F7), and
every 14 days (F14). Active ingredient was
clorotalonil (1.25%0, Bravo 500, Rocsa
lnternational, Lima, Peru). Fungicides
were applied using a manual backpack
sprayer (SOLO 435; SOLO, Sindelfingen,
Germany). Spraying started at emergence.
Each experimental unit field plot was
16.8 m 2 and consisted of four rows (ridges)
with 14 plants each. Distance between
rows was 1 m, and spacing between plants
was 0.3 m. To avoid interplot interference,
a 2-m wide buffer of barley was planted
between the plots. All 18 treatments were
randomly located in each of four replicated blocks.

At planting, ali plots received phosphorus


at 62 kg/ha applied as Ca(Hl0 4 ) 2 , and
potassium at 100 kg/ha, applied as KCI.
Both fertilizers were applied in bands.
Herbicides Sencor and Roundup were
applied and the insecticide Furadan was
used to control Diabrotica spp.

To test for treatment effects, we used


ANOVA for a factorial complete block
design (2 varieties * 3N * 3F). A Tukey
test (P < O.OS) was used for comparison
of means.

Disease severity
Disease severity was estimated visually
every 4 or S days and expressed as the
fraction of the foliage infected. Relative
area under the disease progress curve
(rAUDPC) was calculated according to
Fry (1978).

Ground cover and intercepted radiation


Weekly measurements of the fraction of
green ground cover were made using a
wooden frame of 1 by 0.6 m. The frame
was divided into 1 O x 1O cm grid cells by
double strings (one directly above the
other) 2.S cm apart. The frame was placed
above 2 plants and the area below each
intersection scored as green (covered by
foliage) or not green (not covered by
foliage). The double strings were used to
line up one cell directly over the other,
thus ensuring a vertical alignment of
observer, cel 1, and ground.
Solar radiation measurements were taken
hourly using Watchdog 4SO loggers and
sensors (Spectrum Technologies, Plainfield,
IL, USA). Daily intercepted radiation (IR)
was estimated as ground cover multiplied
by total photosynthetically active radiation. Daily values of the fraction of ground
cover were estimated through linear
interpolation between the weekly observations. Accumulated IR was calculated as
the sum of the daily values over the
growing season. Linear regression was
used to describe the relation between IR
and yield.

Results
Disease
We observed LB at both sites, but rAU DPC
was much higher in Oxapampa where the
conditions were more conducive to
disease development (Table 1). The effects
on disease of variety, N rate, F treatment,
and the interactions of variety*N,
variety*F, and N*F were al 1 significant at
P = O.OS at both sites. In Oxapampa there
were no treatments without LB, whereas in
Huancayo there was LB only in the
treatments without fungicides (Figures 1
and 2). Because of space limitations, only
the results far the lowest (null) and highest
(4- or 7-day interval) fungicide treatments
are shown (Figures 1 and 2).
Late blight severity increased with N but
the effect depended on disease pressure,
varietal resistance, and fungicide use.
There was no effect of N on disease at
very high or low levels of disease (Figures
1A, 1 G, 2A, 2G). At intermediate levels of
disease, LB severity was lower at NO than
at Nl 60 and N320 far Yungay (Figures 2C
and 2E). For Amarilis, however, there were
significant differences between NO and
N320, but Nl 60 could not be statistically
distinguished from NO or N320 (Figures 1 C
and 1 E).

Ground cover and yield


In Oxapampa yields were near zero in the
F- and F10 treatments (Table 1). In the
other treatments in Oxapampa, and in ali
fungicide treatments in Huancayo,
Amarilis yield increased between NO and
Nl 60, but not between Nl 60 and N320.
Yungay yield was the same across N
treatments in Oxapampa, whereas it
increased in Huancayo from NO to Nl 60
and from Nl 60 to N320. IR was a good
predictor of yield across sites (r2 = 0.93 far
Amarilis and r2 = 0.92 far Yungay).
In Huancayo, the increase of IR far
Amarilis from NO to Nl 60 was similar for
both levels of LB (F- or F7) (Figure 3).
Average dry matter yield gain was S.3 t/ha

CIP Program Report 1999 - 2000

71

Table 1. Average rAUDPC1, IR2, and yield 3 using various treatments 4 in varieties Amarilis
Oxapampa and Huancayo, Peru, 1999/2000 growing season.
NO
N160
f"
F10
FF10
F4
FOxapampa
f4
Amarilis
0.6
0.6
rAUDPC
0.1
0.7
0.1
0.7
0.7
83
372
37
96
487
37
IR
34
o.o . 10.5
o.o
O.O
O.O
Tuber yield
o.o
18.6
Yungay
0.6
0.8
0.6
0.8
rAUDPC
0.7
0.1
0.2
87
83
420
32
22
IR
26
496
o.o 11.8
O.O
O.O
O.O
Tuber yield
O.O
11.9
Huancayo
FAmarilis
rAUDPC
O.O
IR
193
Tuber yield
3.8
Yungay
rAUDPC
0.1
IR
175
Tuber yield
2.9

f 14

f7

f-

F14

F7

F-

and Yungay in
N320
F10
0.6
79

O.O

F4
0.2
511
17.4

0.6
84

0.3
504
11.5

F14

f7

o.o

O.O

O.O

O.O

O.O

O.O

0.1

O.O

O.O

232
5.6

242
5.6

345
8.9

379
8.9

391
11.1

321
6.5

319
5.9

407
11.7

O.O

O.O

O.O

436
9.0

0.2
275
2.7

O.O

401
9.3

426
8.3

447
11.8

O.O

O.O

191
4.5

224
5.1

0.3

277
3.1

1 Relative area under disease progress curve.


2 lntercepted radiation (MJ/m 2).

3 Tuber yie!d, t dry matter/ha.


4

Treatment: NO = zero nitrogen (N) fertilization; N160 = 160 kg/ha; N320 = 320 kg/ha; F- = no fungicide;
F4 = fungicide every 4 d; F7 = fungicide every 7 d; F10 = fungicide applied every 10 days; F14 =
fungicide every 14 days.

over the two fungicide treatments; loss due


to LB was about 2 t/ha. That is, between
these two levels of N there was no net
negative effect of N on disease. However,
between Nl 60 and N320 there was a very
small yield increase at F7, while yield
decreased dramatical ly at F- due to the
effect of LB.
For Yungay without LB (F7), IR and yield
nearly doubled between NO and N160, but
with LB (F-), IR was strongly reduced and
the effect on yield was minimal. Tuber
production far Yungay at N1 60/F- and
N320/F- was lower than expected given IR
(below the regression line). A similar
analysis could not be carried out for
Oxapampa because yield was obtained in
only one fungicide treatment, and even
that treatment had high levels of LB.

72

Research on Potato

Discussion and Condusions


There were clear effects of N ferti 1ization
at low to intermediate levels of disease: in
Oxapampa for the F4 treatments, and in
Huancayo for the F- treatments. For
Yungay, there was less disease at a low N
level, and there was no increased disease
severity at a high N level. For Amarilis
there was less disease at a suboptimum N
level, but there was increased disease
severity at high N leve!s. This is partly
inconsistent with Marschner's (1997)
conclusion that plants with optimum
nutrition have the maximum disease
resistance.
Nitrogen increased yields, as was to be
expected. However, in the presence of LB,
in Huancayo, there was no effect of N on
yield for Yungay. And between Nl 60 and

Blight proportion

1.00

Ground cover

0.80

0.60

0.60

0 ..40

0.40

O.al

02)

0.00

0.00

20

60

40

80

1.00

Oxaparnpe
F4

0.80

120

100

1.00

0.60

0.60
0.40

0.20

0.20

000

000

20

60

40

80

100

40

60

100

80

,---:-;~-=:::;:::;;;;::::::::-:----

120

1.00 . , . - - - - - - - - - - - - - - - - - - - - .
Huancayo
f.

0.80

20

O.BO

0.40

20

40

60

80

100

120

1.00 - . - - - - - - - - - - - - - - - - - - - - - - .
0.8J

0.60

0.60

0.40

0.40

0.20

020

0.00

1.00 . . . . - - - - - - - - - - - - - - - - - - - .
OUpampa
0.8J
f.

Oxapampa

.J-..................................~~;:~;as;~S:;S;;~::J
20

40

60

80

100

120

0.00

140

25

00

75

100

125

150

25

50

75

100

125

150

G
1.00

1.00

Huancayo
F7

0.80

0.80

0.60

0.60

0.40

0.40

a aa

0.20

0.20

000

000

20

40

60

80

100

120

140

Days after emergence

Days after emergence

o NO

N160

N320

Figure 1. Blight proportion and ground cover for Amarilis, three levels of nitrogen (N) fertilization and different
levels of fungicides, in Oxapampa and Huancayo. Vertical bars indicate the standard error of the mean. Key: NO
= zero nitrogen (N) fertilization, N160 = 160 kg N/ha, N320 = 320 kg N/ha, F- = no fungicide, F4 =
fungicide applied every 4 d, F7 = fungicide every 7 d.
N320 there was a strong yield loss due to
LB for Amarilis. lt is not clear whether the
differences in the response of Yungay and
Amarilis are related to the observed level
of infection (a function of their resistance

level and the weather), or whether these


differences are variety specific in another
way. In Huancayo, yield was much lower
than expected given IR for Yungay at
N160/F-, and to sorne extent it was also

CIP Program Report 1999- 2000

73

Blight proportion
1.0

1.0

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

o.o

O.O

20

40

60

1.0

D
0.8

a
a

0.6

0.4
0.2

0.2

o.o + - - s . - - - . - - - - , - - - - - r - - - . - - - - 1
20

80

60

40

1.0

--~--~-~-~--~---<

20

40

60

80

100

120

1.0

Hu ancayo

o.a

O.O

120

100

120

100

1.0

0.8

F-

60

0.6
0.4

Oxapampa

120

100

80

Ground cover

o.a

o.6

0.6
0.4

0.4

0.2

0.2

o.o

O.O
20

40

60

80

100

120

140

25

50

75

100

125

150

25

50

75

100

125

150

G
1.0

1.0

Huancayo
F1

0.8

0.8
0.6

0.6

0.4

0.4

a a a

0.2

o.o

0.2

+----..................~.............~....~:e:S~:S:S:~!l::ll
o

20

40

60

80

100

120

o.o

140

Days after emergence


oNO

Days after emergence


t::.N160

N320

Figure 2. Blight proportion and ground cover for Yungay, three levels of nitrogen (N) fertilization and different
levels of fungicides, in Oxapampa and Huancayo. Vertical bars indicate the standard error of the mean.
Treatment: NO = zero nitrogen (N) fertilization, N160 = 160 kg N/ha, N320 = 320 kg N/ha, F- = no fungicide,
F4 = fungicide applied every 4 d, F7 = fungicide every 7 d.
too low far Amarilis at N320/F-. This
apparent decrease in radiation use efficiency conflicts with results of Haverkort
and Bicamumpaka (1986) and Van Oijen
(1990).

74

Research on Potato

The differences in LB severity because of


N fertilization is potentially a confounding
factor when compartng resubs from
multilocation trials. Our results also
illustrate the potential importance of

Yield (dry matter) (t/ha)


N320 F7

12

A(Amarilis)
10

12

N320 F7

10

N320 F-

(F7)

(F7)
4

(F-)
2

o
100

200

300

400

500

100

lntercepted Radiation [MJ-m-

200
2

300

400

500

Figure 3. lntercepted radiation vs yield in Huancayo far Amarilis (A) (slope = 0.036) and Yungay (B) (slope =
0.026) at three levels of nitrogen (N) and two levels of fungicides. Oblique line is the regression between
intercepted radiation and yield (separate far each variety) far all observations in Huancayo. Arrows indicate the
effect of Non yield between levels of Nin the presence of late blight(F-), and absence of late blight(F7). Key: NO
= zero Nfertilization, N160 = 160 kg N/ha, N320 = 320 kg N/ha, F- = no fungicide, F7 = fungicide applied
every 7 d.
variety choice when aiming at increased
yields. Resistant varieties are always
useful where there is LB, but in sorne
cases, when farmers shift from low to high
levels of N fertilization, they might
become even more valuable. Possibly
there cou Id be a stronger effect of N on LB
at high N levels if P and K are low.
However, this scenario is not relevant
because low P and K would greatly reduce
yield, hence reducing these inputs is not a
viable management option.
For Yungay in Huancayo, the investment
in N fertilizer would only have been useful
if combined with fungicide use. This did
not hold for Amarilis, which has higher LB
resistance, and where LB only influenced
the effect of N on yield when an overdose
of N was supplied. For varieties such as
Amarilis, N does not seem to be an
important factor for integrated management of LB, because there is no evidence
of a need to adjust optimum N fertilization
because of the disease.

Acknowledgements
We thank Greg Forbes, Rebecca Nelson,
and Consuelo Arellano for reviewing this
paper.

References
Awan, A.B. and R.A. Struchtemeyer. 1957.
The effect of ferti 1ization on the
susceptibility of potatoes to late blight.
American Potato Journal 34:315-319.
Carnegie, S.F. and J. Colhoun. 1983.
Effects of plant nutrition on
susceptibility of potato leaves to

Phytophthora infestans.
Phytopathologische Zeitschrift
1 08:242-250.
Cohen, Y. and J. Rotem. 1987. Sporulation
of foliar pathogens. In: Pegg, G.F. and
P.G. Ayres (eds.), Fungal infection of
plants. Cambridge University Press, UK.
p. 314-333.
Crosier, W. 1934. Studies in the biology of
Phytophthora infestans (Mont.) de Bary.
Memoir 155. Cornell University

CIP Program Report 1999- 2000

75

Agricultura! Experiment Station, lthaca,


NY, USA. 40 p.
Devaux, A. and A.J. Haverkort. 1987. The
effects of sh ifting planting dates and
mulching on late blight (Phytophthora
infestans) and drought stress on potato
crops grown under tropical highland
conditions. Experimental Agriculture
23:325-333.
Fry, W.E. 1978. Quantification of general
resistance of potato cultivars and
fungicide effects for integrated control
of potato late blight. Phytopathology
68:1650-1655.
Garrett, K.A. and C.C. Mundt. 2000. Host
diversity can reduce potato late blight
severity for focal and general patterns of
primary inoculum. Phytopathology
90:1307-1312.
Haverkort, A.J. and M. Bicamumpaka.
1986. Correlation between intercepted
radiation and yield of potato crops
infested by Phytophthora infestans in
central Africa. Netherlands Journal of
Plant Pathology 92:239-247.
Hijmans, R.J., G.A. Forbes, and
T.S. Walker. 2000. Estimating the global
severity of potato late blight with a
GIS-linked disease forecaster. Plant
Pathology 49:697-705.
Lowings, P. and l. Acha. 1959. Sorne
factors affecti ng growth of Phytophthora
infestans (Mont.) de Bary l. Growth in
living potato leaves. Transmissions of
the British Mycological Society
42:491-501.
Marschner, H. 1997. Mineral nutrition of
higher plants. 2nd Edition. lnstitute of
Plant Nutrition, University of
Hohenheim, Germany. 860 p.
Ortiz, O., P. Winters, H. Fano, G. Thiele,
S. Guzman, R. Torres, V. Barrera,
J. Unda, and J. Hakiza. 1999.
Understanding farmers' responses to late
blight: Evidence from Peru, Bolivia,
Ecuador, and Uganda. In: lnternational
Potato Center. lmpact on a changing

76

Research on Potato

world: Program report, 1997-98.


lnternational Potato Center, Lima.
p. 101-109.
Phukan, S.N. 1993. Effect of plant nutrition
on the incidence of late blight disease
of potato in relation to plant age and
leaf position. lndian Journal of
Mycology and Plant Patholology
23(3):287-290.
Reichbuch, L., l. lgnatescu, and
Gh. lacob. 1977. Aspecte noi privind
siste!llul de ferti !izare ~ reactia
cartofului la Phytophthora infestans
(Mont.) de Bary (Mana) ~i A/ternaria sp.
(Patarea bruna a frunzelor). Lucrari
~tiintifice 8:115-130 (Timisoara,
Romana).
Rotem, J. and A. Sari. 1983. Fertilization
and age-conditioned predisposition of
potatoes to sporulation of and infection
by Phytophthora infestans. Journal of
Plant Diseases and Protection
90(1 ):83-88.
Sawicka, B. 1993. Zmiennosc pojawu i
szerzenia si~ zarazy ziemniaka
(Phytophthora infestans (Mont.) de Bary)
w warunkach ochrony plantacji i
nawozenia azotem. Biuletyn lnstytut
Ziemniaka 42:113-122 (Budapest,
Hungary).
Thiele, G., O. Navia, and E.N. FernndezNorthcote. 1998. Anlisis econmico de
la estrategia de control qumico del
tizn tardo (Phytophthora infestans)
para cultivares de papa susceptibles en
Cochabamba, Bolivia. Fitopatologa
33:176-181.
Van Oijen, M. 1990. Photosynthesis is not
impaired in healthy tissue of blighted
potato plants. Netherlands Journal of
Plant Pathology 96(2):55-63.
Wierzejska-Bujakowska, A. 1994. Wptyw
ochrony ziemniaka przde zaraza
(Phytophthora infestans (Mont.) de Bary)
na efectywnos nawozenia azotem.
Biuletyn lnstytut Ziemniaka 44:131-144
(Budapest, Hungary).

Relationships of Fungicide Application to Late-Blight


Development and Potato Growth Parameters in the
Tropical Highlands of Uganda and Kenya
O.M. Olanya1, R. El-Bedewy1, P. S. Ojiambo1, P. T. Ewell1 and J. J. Hakiza2

The impact of fungicide applications on late blight development, potato


growth parameters, and yield was quantified at field sites in Kenya and
Uganda during the 1999 and 2000 cropping seasons. In Kenya, three potato
varieties were evaluated at two sites, at altitudes of 1800 m and 2200 m. In
Uganda, three varieties were evaluated at one site atan altitude of 2400 m.
Leaf, stem, root, and tuber biomass were sampled and quantified during the
cropping season. Dithane M-45 fungicide was applied to the experimental
plots at the onset of disease at intervals of 7, 14, and 21 days. Final tuber
yields differed significantly by as mu ch as 15 % between fungicide-treated
and untreated controls across the three sites. The severity of late blight
differed among sites and between years as a result of variations in environmental conditions. Fungicide application intervals significantly affected
disease progress, and the area under disease progress curve (AUDPC) was
significantly lower in the fungicide-treated plots than in the control
(unsprayed) plots. Tuber biomass accumulation was also significantly affected
by fungicide applications. Varieties Tigoni and Rutuku have relatively higher
levels of polygenic resistance to late blight and showed low levels of disease
severity.
Late blight of potato (Solanum tuberosum
L.), caused by

Phytophthora infestans

(Mont.) De Bary, is a devastating disease


among smal 1-scale farmers of the tropical
highlands of East Africa (Haverkort, 1986).
In Uganda, under good management but
without fungitide sprays, yield losses
attributed to late blight in susceptible
varieties are estimated in the range of
40-60% (Mukalazi et al., 2001 ), with
serious economic losses reported as a
result of late-blight infection (Sengooba
and Hakiza, 1999). Similarly, yield losses
attributed to late blight in Kenya have
1

Sub-Saharan Africa Region, CIP, Nairobi, Kenya.


Potato Research Program, National Agricultura! Research
Organization, Kabale, Uganda.

been reported to be about 40-50%


(Njuguna et al., 1998).
Previous research has indicated that
periodic applications of protective fungicides for control of late blight reduce the
rate of epidemic development of the
disease (Van der Plank, 1967). However,
geographical information system (GIS)
data linked to disease forecast models
reveal that less than optimum levels of
fungicide are applied by most farmers in
East Africa (Hijmans et al., 2000). With
the exception of optimum or scheduled
fungicide applications based on favorable
weather conditions, the most economical
option for disease management is the use

CIP Program Report 1999 - 2000

77

of host-plant resistance (Olanya et al.,


2001 ). A number of CIP genotypes are
currently being evaluated in Kenya,
Uganda, Eth iopia, and many other SubSaharan African countries with promising
leve Is of resistan ce to late bl ight (ElBedewy et al., 2001). The use of cu ltivars
with durable resistance combined with
scheduled applications of protective
fungicides has been reported as useful for
managing late blight (Simons, 1972), as
well as other diseases (Van der Plank,
1963).
The impact of fungicide applications on
disease development, accumulation of
potato biomass, susceptibility, and yield in
the tropical highlands of Africa has not
been adequately documented. lnformation
on the application of fungicides to commonly grown potato genotypes is essential
to ensure efficient use of fungicides and to
complement resistance to late blight.
Within the East African region, potatoes
are grown under various agroecological
production systems with different soil
types, management practices, and varieties (Adipala, 1999; PRAPACE, 1995). The
objective of research reported here,
therefore, was to collect standardized data
on potato growth parameters from three
distinctly different sites and to determine
the impact of fungicide applications and
late-blight development on the yield of
potato varieties with different levels of
resistance to Phytophthora infestans.

Materials and Methods


Experimental sites and design
Field plots in Kenya were located at the
University of Nairobi, Kabete Field Station
(1800 m above sea leve!), and at Loreto
(2200 m). In Uganda, the plots were
located in Kalengyere Research Station
(2400 m). At all three sites, plot measurements were 3 x 5 m (W x U, each
consisting of four rows. A 3 x 4 factorial
experiment (varieties x fungicide application intervals) was established in a
randomized complete block design with

78

Research on Patato

three replications. Fungicide application


intervals of 7, 14, and 21 days and a
control (no application) were tested. The
test fungicide used in the study was
Dithane M-45 (mancozeb). Patato varieties Tigoni, Asante/Victoria, and Kerr's Pink
were planted in Kenya, whereas the
varieties Rutuku, Asante/Victoria, and
Kabale were planted in Uganda. The
potato varieties used in this experiment
have different levels of resistance to late
blight: Rutuku and Tigoni have moderate
resistance, Kabale is moderately susceptible, and Asante/Victoria and Kerr's Pink
are susceptible. In all field plots, normal
agronomic practices, such as adequate
field and seed-bed preparation, hilling,
and weeding were followed. Fertilizers
(175 N, 175 P) were applied at the rate of
500 kg diammonium phosphate per
hectare. lnsecticides (metasystox or
dimetheoate) were applied to control
aph ids and potato tuber moths when
necessary (Table 1 ).

Potato growth data


About 40 days after emergence, 36 plants
were randomly sampled from field plots.
Twelve plants per variety were sampled at
ea ch assessment date and represented al 1
the treatments. Fresh leaves, stems, roots,
and tubers were obtained from each
sample and weighed immediately to
record fresh potato biomass. In addition,
100-gram subsamples of fresh leaves,
stems, roots, and tubers were placed in
paper bags and oven dried at 82C for
four days to obtain dry weights. Potato
biomass was quantified four times during
the cropping season at two-week intervals.
At harvest, tuber numbers from each
experimental plot were quantified and
weighed; yield was expressed as tons per
hectare (t/ha) for subsequent analysis.

Disease assessment and environmental


monitoring
At the onset of late-blight symptoms,
disease incidence and severity were
quantified weekly in all plots. This was
based on visual symptoms for late blight,

Table 1. Management practices used at each of the sites in evaluating late-blight development and patato
growth parameters in 1999 and 2000.
Year Management
Experimental site
1999 practice
Loreto, Short rain
Kabete, Long rain
Kalengyere, Season "B"
Planting date
29/10/1999
26/3/1999
16/10/1999
Fertilizer application
Diammonium phosphate Diammonium phosphate Diammonium phosphate
1nsecticide
Metasystox
Metasystox
Dimetheote
Fungicide
Dithane M-45
Dithane M-45
Dithane M-45
13(7/1999
Harvesting
02/2/1999
11/1/2000
Loreto, Long rain
Kabete, Long rain
2000
Kalengyere., Season "A"
Planting date
13/4/2000
12/4/2000
20/3/2000
Fertilizer application
Diammonium phosphate Diammonium phosphate NPK
Metasystox
Dimetheote
lnsecticide
Metasystox
Dithane M-45
Fungicide
Dithane M-45
Dithane M-45
Harvesting
26/7/2000
26/7/2000
23/7/2000
using an assessment scale of 0-100%. At
least five disease assessments were
recorded. At crop maturity, the incidence
of tuber blight was also quantified visually.
Tubers that did not show visual symptoms
were stored for three weeks and subsequently observed or plated to look for
additional incidence of tuber blight.
At the three experimental sites (Kabete,
Loreto, and Kalengyere), weather equipment (Hobo Pro Series, MA, USA, and
Watchdog Data Logger, Spectrum Technologies, Plainfield, IL, USA) monitored
environmental parameters such as temperature, relative humidity, rainfall, and
hours of sunshine or photosynthetic active
radiation. Additional data from the University of Nairobi weather station were also
used.

Data analysis
Mean values of disease incidence and
severity were calculated using SAS (1989).
Similarly, area under disease progress
curves (AUDPC) was calculated from
values for disease severity, as described by
Campbel 1 and Madden (1990). The
development and progress of late blight in
fungicide-treated plots versus untreated
controls were graphically compared using
AUDPC. The effects of the intervals
between fungicide applications on lateblight development were computed and

compared for the different varieties using a


GLM procedure (SAS, 2001 ), which

considers the effects of factors such as


location, year, and dates of assessment.
Th is analysis was done separately for ea ch
genotype. Mean values for data on plant
biomass and fresh and oven-dried weights
of leaves, stems, roots, and tubers were
calculated for each plot (SAS, 1989). A
combined analysis of variance (ANOVA)
was also used to investigate the effects of
location and years (environment) on each
variety. Similarly, graphical arrays of
cumulative biomass (dry tuber weights)
were used to compare patato development
among varieties and sites. At each assessment period, mean values for the
environmental parameters were calculated
(SAS, 1989).

Results
Dynamics of late blight in relation to
fungicide application
During the 1999 cropping season, disease
severity at Kabete was low compared to
Loreto (Figure 1) The final disease level
(AU DPC) of AsanteNictoria was 250
(% disease days) at Kabete and 1200 at
Loreto. During the 2000 cropping season,
late-blight severity was very low at Loreto,
especial ly in fungicide-treated plots. In the
control (untreated) plots, the disease was
highest in the susceptible variety Kerr's
Pink (with an AU DPC val u e of 501 .1 ),
compared to Tigoni (AUDPC 163.9) and
AsanteNictoria (AUDPC 201.6) (Table 2).

CIP Program Report 1999 - 2000

79

Table 2. Effect of fungicide (Dithane M-45) application intervals on disease severity and patato yields of
patato at Loreto, long rains, and Kalengyere (season A, year 2000).
AUDPC2
Yield (Vha)
Tubers (no.)
Application intervals Variety1
Loreto, Kenya
Tigoni
0.2
57.0
495
7 days
0.4
349
59.8
7 days
AsanteNictoria
0.7
42.0
343
7 days
K. Pink
Tigoni
0.4
54.0
462
14 days
0.7
57.4
327
14 days
AsanteNictoria
0.9
14 days
K. Pink
41.5
223
Tigoni
21 days
3.6
52.5
400
21 days
5.2
56.0
254
AsanteNictoria
K. Pink
10.1
21 days
40.3
195
Tigoni
Control
163.9
51.2
329
201.6
Control
AsanteNictoria
53.9
199
K. Pink
501.1
35.2
17
Control
2.5
Mean
50.5
339
2.9
LSD 0.05
4.5
65.3
cv (%)
64.06
9.33
28.58
Kalengyere, Uganda
7 days
Rutuku
O.O
21.4
221.3
7 days
Ka bale
23.5
19.5
144.7
7 days
Victoria
19.6
22.5
200.3
14 days
Rutuku
2.1
23.2
214.7
14 days
Ka bale
52.5
19.5
176.0
14 days
Victoria
60.6
22.5
200.7
21 days
Rutuku
2.5
23.1
233.3
21 days
Kabale
69.3
19.5
172.3
21 days
Victoria
101.9
23.3
241.7
Tigoni
Control
90.3
21.0
234.0
Control
AsanteNictoria
185.9
17.3
164.3
Control
K. Pink
225.8
19.2
216.3
Mean
23.2
21.6
201.6
LSD 0.05
35.0
6.1
53.5
cv (%)
176.3
35.2
30.9
Note: Dithale M-45 (malcozeb) was applied at the rate of 3 kg/ha. A total of five applicatiols were made durilg the
croppilg seasol. Experimelts at Loreto were plalted Ol April 13 ald harvested Ol July 26. At Kalelgyere, plaltilg
was Ol March 20 ald harvestilg was July 23.
1 varieties Rutuku ald Tigoli are moderately resistalt ald Kabale is moderately susceptible, while AsalteMctoria ald Kerr
Pilk are susceptible to late blight.
2 AUDPC is area Ulder disease progress curve (% of disease days) from six late blight readilgs.

Hardly any disease was detected in the


fungicide-treated plots.
In 1999, late blight severity was higher at
Kalengyere (Uganda) compared to Kabete
and Loreto (Kenya). The final AUDPC on
the variety Victoria at Kalengyere was
2200 for the control treatment, with very
low disease rates on plots treated with
fungicide at seven-day intervals (Figure 1).

80

Research on Potato

During season A of 2000, late blight was


at a very low level in Kalengyere (Table
2). For Asante/Victoria in the untreated
plots, disease levels at Kalengyere were
comparable to those at Loreto. For Asante/
Victoria, a combined ANOVA showed that
fungicide treatments had a significant
effect on late-bl ight severity, although
there were variations in disease levels
across sites and years (Table 3).

en
~

__...30.--~~~~~~~~~~~~~

2500

-e

(1)

"'

2000

u"'

"#.

-o- Control

1500

500

<{

Treated

20

(;

.g

::::>

-a- Control
'6o-

&'15

~ 1000

o..

E' 25
O>

--Treated

cU

(1)

Kabete 1999 SR

Asante

Kabete 1999 SR

Asan te

10

1'--~...._~_._~__._~~..__~.....__~_,

30

20

40

30

50

60

40

Asan te
Asan te

en
>-

50

60

70

80

90

70

Loreto 1999 SR

Loreto 1999 SR

2500

cU

-e
(1)

"'cU
(1)

u"'

2000
1500
(;

~ 1000

.o 10

1-

o..

:::i

500

::::>
<{

30

o
20

40

30

cU

60

"'cU

(1)

u"'

2500

~ 1000

500

::::>
<{

90

2000
1500

o..

80

Figure 2. Average tuber dry-weight accumulation of


patato variety AsanteNictoria obtained from replicated
field experiments in Kenya during 1999 short rain
season (SR).

-e
(1)

50
60
70
Days after planting

70

Kalengyere 1999 SR

Victoria

en
>-

50

40

o
20

30

40

50

60

70

80

90

Days after planting

Figure 1. Progress of late blight of patato at Kabete


(altitude 1800 m) and Loreto (2200 m) in Kenya and
Kalengyere (2400 m) in Uganda during the 1999
cropping season. Application intervals of Dithane M45 were 7 days and control treatment. ("SR" refers to
short rain season and "LR" refers to long rain

lmpact of fungicide applications on potato


biomass
The effect of fungicide applications on dry
tuber weight accumulation is shown in
Figures 2 and 3 for all sites. During the
1999 cropping seasons, the percent of
tuber dry-matter accumulation was higher
in field plots treated with fungicide at

seven-day intervals compared to the


control plots at Kabete and Loreto. At
Kalengyere, tuber dry-matter accumulation
was significantly higher in the fungicidetreated plots than in the control. In
general, the greatest level of tuber drymatter accumulation of 20-25% was
recorded in the fungicide-treated plots at
Loreto and Kabete (Figure 2). Similarly, in
2000, tuber dry-matter accumulation was
significantly higher at all sites in samples
obtained from plots treated at seven-day
intervals in comparison to samples from
the control plots (Figure 3). The highest
levels of tuber dry weight were recorded in
samples from treated plots: Kabete (23%),
Loreto (23.5%), and Kalengyere (20%).

Relationship of late-blight development to


tuber yield and numbers on Asante/
Victoria
An ANOVA on combined data from the
two years of the experiment, across sites,

CIP Program Report 1999 - 2000

81

Asante

-tP

20

15

(i;
.o

10

~
o

:::J

Kabete 2000 SR

-o- Control

25

Dl

At Kalengyere, the highest yield for 2000


s~ason A was from Rutuku in the fungic1de-treated plots (Table 2). Fungicide
treatments did not significantly affect
average number of tubers; however, tuber
numbers attributed to treatment effects
differed across sites and years.

Treated

.,,.-

___ .-..........

..........

A.---------o----1.J---A,_.-

...... A

Environmental variation among sites

10'--~---'-~~....1-~~.L_.~--1.~~~

50

60

80

70

Asante

90

100

Loreto 2000 SR

30r-~~~~~~~~~~~~----.

~ 25
E
-~ 20
~

&' 15
(i;

.g

10

10'--_,c_--'--~--1...~~L_~...L~_L~_J

40

50

60

Victoria

70

80

90

100

Kalengyere 2000 SR

~30r-~~~~~~~~~~~~~

E 25

Dl

;
~

bo----...6.--~

20 ....

o 15 ,_
(i;
.o

10 ....
40

50

60

70
80
90
Days after planting

100

110

Figure 3. Average tuber dry-weight accumulation of


patato varieties AsanteNictoria obtained from
replicated field experiments in Kenya and Uganda
during the 2000 cropping season (LR = long rain
and SR = short rain season).
revealed that the intervals between
fungicide applications had a significant
(P = O.OS) effect on tuber yield (Table 3).
Significant differences between sites were
also detected. However, variations in yield
were recorded within sites in different
years. At Loreto in the 2000 long rains
cropping season, Asante/Victoria had a
yield of 59.8 t/ha and Tigoni yielded 57
t/ha; Kerr's Pink had a total yield of 42 t/ha.

82

Average monthly temperature, rainfall,


and relative humidity were recorded at
each site. Average temperatures were
similar for ali three sites during the
cropping season, but the average relative
humidity and total rainfall was not (Table
4). During the 1999 cropping season, mean
relative humidity was much lower at
Kabete than at Kalengyere and Loreto.
During the 2000 cropping season, total
rainfall was less than in the 1999 cropping
season for all three sites. At Kalengyere
and Loreto, environmental conditions were
more conducive to disease development in
1999 than in 2000.

Research on Potato

Discussion
Applying the protective fungicide, Dithane
M-45, significantly reduced the development and severity of late blight and
increased tuber yields. In general, disease
was less severe in experimental plots with
the fungicide applied at intervals of seven
days, compared to the 21-day intervals or
control plots. The low disease levels in
Tigoni and Rutuku could be because both
varieties have relatively high levels of
horizontal resistance to the disease
compared to Asante/Victoria and bale.
In this study, a combined analysis of
variance was used for Asante/Victoria
because it was the only variety that was
planted in all cropping seasons. The lack
of consistency in disease levels recorded
across sites and seasons may be attributed
to environmental variation among sites. In
sorne cases, the early occurrence of the
disease befare the initiation of the spray
p~ogram cou Id account for madequate
d1sease control aRd confounding effects on
treatments. Adjustment of fungicide

Table 3. Combiled alalysis of varialce (ANOVA) Ol the effect of fulgicide applicatiol iltervals Ol late blight
severity ald tuber yield Ol AsalteMctoria variety (Victoria) plalted at Kabete, Loreto, ald Kalelgyere
sites durilg 1999 ald 2000 croppilg seasols.
So urce
DF
F-Value
Pr > F
Late blight severity
Ye ar
1
3.67
0.1956
2
Locatiol
1.25
0.04 a
2
Locatio l *year
118.36
0.0001 b
Rep (year*locatiol)
12
0.76
0.6887
Fulgicide 1
3
0.63
0.05 b
Fulgicide*year
0.97
3
0.4676
Fulgicide*locatiol
6
1.09
0.4613
Fulgicide*year*locatiol
42.68
0.0001 b
6
Fulgicide*rep (year*locatiol)
3.24
36
0.0001 b
Assess.day-DOA (year*locatiol)
28
93.77
0.0001 b
DOA*fulgicide (year*locatiol)
84
17.25
0.0001 b
Yield (t/ha)
Year
1
38.09
0.0001 a
Location
7.74
2
0.0006 b
Locatiol*year
29.76
2
0.0001 b
12
Rep (locatiol*year)
1.58
0.1424
Fulgicide
2.70
0.0466 a
3
3
Fungicide*year
0.47
0.7142
Fulgicide*location
0.60
6
0.7223
Fulgicide*location*year
8.25
0.0001 a
6
Rep*fulgicide (location*year)
12
0.30
0.9867
Tuber numbers
Ye ar
1
0.09
0.7933
2.20
2
Locatiol
0.3124
15.02
2
Locatio l *year
0.0005 b
12
Rep (location*year)
2.05
0.0481 a
Fungicide
1.79
3
0.2496
Fungicide*year
4.09
3
0.0673
Fungicide*location
2.18
6
0.1829
Fungicide*location*year
1.08
6
0.3942
Rep*fungicide (location*year)
0.96
36
0.5476
a = Significant at 0.05; b = Significant at 0.01.
1 Refers to application intervals of 7, 14, 21 days and control (no application) of the contact fungicide Dithane M-45.

appl ication rates to complement the


general resistance of a particular variety
has been reported as one method of
enhancing fungicide efficiency (Fry,
1977).
One of the principies of fungicide resistance management is that factors, which
suppress the growth rate of disease on
individuals relative to sensitive ones, will
also slow the selection far resistant
individuals (Milgroom and Fry, 1988). In
his study on patato genotypes with low
levels of resistance to late blight, Fry

(1975) observed that it was possible to


reduce the amount of fungicide required
far adequate control by introducing
varieties with increasing levels of durable
resistance. In our experiment, a total of
four fungicide applications was used
regardless of the treatment interval (7, 14,
or 21 days), with the same concentration
of fungicide applied. Further research is
required to determine the optimum number
of applications, dosage, and timing with
respect to disease on susceptible varieties
relative to resistant ones. Once this has
been determined, farmers should be able

CIP Program Report 1999- 2000

83

Table 4. Average temperature, relative humidity, and


rainfall at Kalengyere, Kabete, and Loreto sites
during the 1999 and 2000 cropping seasons.
Temp
RH Total rainfall
Kalengyere, 1999
September
October
November
December
Kabete, 1999
September
October
November
December
Loreto, 1999
September
October
November
December
Kalengyere, 2000
April
May
Ju ne
July
Kabete, 2000
September
October
November
December
Loreto, 2000
April
May
Ju ne
Jul~

{C)

{%)

{mm)

16
15
16
17

86
85
85
87

88.3
126.3
120.7
42.2

18
19
18
18

64
62
76
73

26.1
11.4
348.0
229.3

16
17
16
16

89
85
89
83

36.0
61.1
398.2
234.1

16
16
16
16

87
86
85
85

114.0
52.2
1.3
22.4

18
19
19
18

57
64
59
56

33.5
18.4
187.7
111.3

18
16
14
13

78
84
83
82

195.2
73.3
37.5
23.1

Note: Planting and harvesting dates are shown in Table 1.

to increase fungicide efficiency by


adjusting the number of applications, the
dosage, and the timing to complement
resistance levels. Unfortunately, for most
of the potato varieties grown in subSaharan Africa, the relationships between
environmental parameters, disease,
fungicide use, and resistance have not
been adequately studied (Olanya et al.,
2001) to allow for reliable adjustments of
fungicide applications to complement
resistance. Niklaus et al. (2000) in their
study on the evaluation of host resistance
and weekly fungicide application observed that resistance of a susceptible

84

Research on Potato

variety Alpha was increased by 20% when


a contact fungicide was applied. Further
study is underway to increase the efficiency of fungicide use in sub-Saharan
Africa by quantifying the relationships of
dosage, number of applications, and
resistance in susceptible, moderately
susceptible, and resistant clones and
varieties.
The beneficia! effect on dry-matter
accumulation of applying a protective
fungicide is attributed to the indirect effect
on late-bl ight development and subsequent
availability of photosynthetic leaf area.
The similarity of maximum tuber dry
matter at Kabete and Loreto in 2000 may
be attributed to the very low severity of
late blight and the similar growing conditions during the season. The low dry matter
recorded at Kalengyere may be attributed
to the lower rainfall that year. Generally,
the mean temperatures were lower at
Kalengyere than at the other sites, which
may account for the lower tuber biomass
accumulation at Kalengyere relative to the
other sites. In general, the rate of biomass
or dry-matter accumulation appears to be
similar across sites; however, there is no
information on the number of days to tuber
initiation.
The total number of tubers was not significantly affected by appl ication intervals,
but variations in tuber yields were recorded among varieties, with the tolerant
variety Tigoni performing better than
AsanteNictoria (Victoria) or the susceptible varieties, Kerr's Pink and Kabale.
Victoria and Kerr's Pink were highly
susceptible to late bl ight and early season
infection, contributing significantly to the
lower yields observed. Fry and Shtienberg
(1 990) reported that complete suppression
of yield in susceptible varieties is possible
if the disease occurs early enough.
Variations in climatic conditions were
recorded across seasons and,years at the
testing sites. Average temperature appeared to be similar; however, relative
humidity and rainfal 1 were different. This

varrat1on in key environmental parameters


cou Id account for the differences seen in
late blight disease levels. Subsequent
variability in tuber yield could also be
explained in terms of seasonal variations
in environmental conditions. During the
2000 cropping season, adverse environmental conditions, which accounted for
low yields, were noted. Similarly, lateblight severity (as measured by AUDPC)
was greater in seasons or locations with
low total yields.

Condusions
We conclude that applications of protective fungicide have a significant deterrent
effect on the development of late blight in
sorne sites in tropical Africa, and a
positive effect on accumulation of potato
biomass. The magnitude of disease
development and dry-matter partitioning
appears to be dependent on the levels of
resistance of the varieties and the frequency of fungicide applications.
However, further data are needed to assess
the contribution of cardinal factors such as
temperature, photosynthetic active
radiation, and rainfall in crop development
(tuber initiation, rate of tuber bulking) in
tropical Africa. Furthermore, elements
such as the effect of planting and harvesting dates on disease development and
potato biomass have not been adequately
addressed in the tropical highlands of
Africa.

References
Adipala, E. 1999. Potato production in
Uganda: A survey perspective. A report
to the Rockefeller Foundation by Dept.
of Crop Science, Makerere University,
Kampala, Uganda. 41 p.
Campbell, C.L. and L.V. Madden. 1990.
lntroduction to plant disease
epidemiology. John Wiley & Sons, NY,
USA.
El-Bedewy, R., O.M. Olanya, P.T. Ewell,
C. Lung'aho, P.S. Ojiambo, and
J. Karinga. 2001. Evaluation of potato
germplasm (populations A & B) for

resistance to late blight in Kenya.


African Crop Science Journal 9:215223.
Fry, W.E. 1975. lntegrated effects of
polygenic resistance and a protective
fungicide on development of patato late
blight. Phytopathology 65:908-911.
Fry, W. E. 1977. lntegrated control of
patato late blight-Effects of polygenic
resistance and techniques of timing
fungicide applications. Phytopathology
67:415-420.
Fry, W.E. and D. Shtienberg. 1990.
lntegration of host resistance and
fungicide to manage potato diseases.
Canadian Journal of Plant Pathology
12:111-116.
Hijmans, R.J, G.A. Forbes, and
T.S. Walker. 2000. Estimating the global
severity of patato late blight with GISlinked disease forecast models. Plant
Pathology 49:697-705.
Haverkort, A.J. 1986. Light interception
and yield relations under the tropical
highland conditions of Central Africa.
Potato Research 29:257-258.
Milgroom, M.G. and W.E. Fry. 1988. A
model for the effects of metalaxyl on
potato late blight epidemics.
Phytopathology 78:559-565.
Niklaus, N.J., O.A. Rubio-Covarrubias, and
W.E. Fry. 2000. Patato late blight
management in the Toluca Valley:
Forecasts and resistant cu ltivars. Plant
Disease 84: 410-416.
Njuguna, J.G.M. , G.I. Oduor, and
D.N. Njenga. 1998. Rate of adoption of
new late blight resistant potato cultivars
by farmers in Kiambu district. In:
Proceedings of the 2nd Biennial Crop
Protection Conference, KARl/NARL,
Nairobi, Kenya. p. 196-202.
Mukalazi, J., E. Adipala,T. Sengooba,
J.J. Hakiza, M. Olanya, and H.M.
Kidanemariam. 2001. Variability in
potato late blight severity and its effect
on tuber yield in Uganda. African Crop
Science Journal 9:195-201.
Olanya, O.M., E. Adipala, J.J. Hakiza,
C. Kedera, P.S. Ojiambo, J.M. Mukalazi,
G. Forbes, and R. Nelson. 2001.

CIP Program Report 1999 - 2000

85

Epidemiology and population dynamics


of Phytophthora infestans in sub-Saharan
Africa: Progress and constraints. African
Crop Science Journal 9:185-193.
PRAPACE. 1995. Regional patato and
sweetpotato improvement program in
Eastern and Central Africa. Annual
report. Regional Patato and Sweetpotato
lmprovement Network in Eastern and
Central Africa, Kampala, Uganda. 67 p.
SAS. 1989. SAS/STAT user's guide, Version
6, Fourth edition, Volume 2. SAS
lnstitute lnc., Cary, NC, USA. 846 p.
SAS. 2001. The GLM procedure (SAS/STAT
software). SAS/STAT user's guide. Online: http://www.sas.com/service/
techsu p/faq/stat_proc/gl mproc.htm 1.
Sengooba, T. and J.J. Hakiza. 1999. The
current status of late blight caused by
Phytophthora infestans in Africa, with

86

Research on Potato

emphasis on eastern and southern


Africa. In: Crissman, L. and C. Lizarraga
(eds.). Proceedings of the global
initiative on late blight conference,
Quito, Ecuador, March 16-19, 1999,
Vol. 1: Late blight: A threat to global
food security. lnternational Patato
Center, Lima, Peru. p. 25-28.
Simons, M.O. 1972. Polygenic resistance
to plant disease and its use in breeding
resistant cultivars. Journal of
Environmental Quality 1 :232-240.
Van der Plank, J.E. 1963. Plant disease:
Epidemics and control. Academic Press,
NY, USA. 349 p.
Van der Plank, J.E. 1967. Epidemiology of
fungicida! action. In: Torgeson, O.C.
(ed.). Fungicides: An advanced treatise,
Vol. 1. Academic Press, NY, USA.
p. 63-92.

Simulating Potato Late Blight in the Highland Tropics


G.A. Forbes,1 M.G. Chacn, 1 M.V. Taipe, 1 and R.J. Hijmans2

A potato (Solanum tuberosum L.) late blight (LB) simulation model was validated near Quito, Ecuador, at three sites and over two growing seasons. Two
varieties differing in resistance to Phytophthora infestans (Mont.) de Bary, the
causal agent of LB, were grown at each site. Data taken from each site
included hourly temperature and relative humidity and daily rainfall. Using
weather data from each site, the model simulated disease sever.ity for each
variety. Resistance parameters for the model were taken from measurements
made on the Peruvian varieties Amarilis-INIA (resistant) and Tomasa Tito
Condemayta (susceptible). At each site, the model simulated realistic differences between susceptible and resistant varieties. However, for the
experiments conducted where average daily temperatures were low (less
than 14C), the model underestimated disease severity. These temperatures
are characteristic of the principal potato growing area of Ecuador. The most
accurate simulation occurred at the Tola site in both seasons, where average
daily temperatures are above 14C. Overall, temperature was inversely
related to the predictive capability of the model.
Late blight is a major constraint to potato
production worldwide. The disease is
especially damaging in the tropical
highlands of Latin America, Asia, and
Africa, where it can be controlled only by
routine applications of fungicides, sometimes in conjunction with moderate levels
of host plant resistance (Forbes and Jarvis,
1994; Haverkort, 1990). For many poor
farmers, however, the disease is not
control led, losses are heavy, and crop
abandonment is not uncommon.
Resistant varieties are being developed for
the tropical highlands, but little is known
of how this resistance may best be used to
reduce fungicide application. General
recommendations for reductions in dosage
or for increases in the intervals between
fungicide sprays will probably be of only
partial benefit in the tropical highlands
1

CIP, Quito, Ecuador.


CIP, Lima, Peru.

because of the highly varied agroecosystem (cropping patterns, rainfall,


temperature, and altitude). Site-specific,
variety-specific, or dynamic (e.g., based
on action thresholds) spray strategies
should be more useful for tropical highland
farmers. However, the development of
such strategies for variable climatic
conditions would require extensive and
expensive field studies. Computer-based
disease simulation models have been used
to screen reduced-spray strategies for
potential efficacy so that only the most
promising ones are field-tested. This
approach reduces costs and increases the
number of strategies that can be examined.
A potato LB simulation model developed
at Cornell University, lthaca, New York,
USA, has been used extensively to test
and develop new approaches to managing
LB under temperate conditions. The model

CIP Program Report 1999 - 2000

87

was used to evaluate potential disease


management alternatives (Doster and Fry,
1991; Doster et al., 1990; Fohner et al.,
1984; Raposo et al., 1993; Shtienberg et
al., 1989; Shtienberg and Fry, 1990). The
simulator has also been used to develop a
disease forecasting system (Fry et al.,
1983), and to quantify the effects of
pathogen aggressiveness on final disease
severity (Kato et al., 1997).
The Cornell model was tested and used
only under temperate conditions near
lthaca. Hence it is not known whether it
will work well in different regions. Severa!
aspects of tropical highland potato production differ from potato production in the
temperate zone. In the tropical highlands,
low temperatures and year-round presence
of disease may change the dynamics of
the disease epidemic. For these reasons, it
is not known if the LB simulator described
above will accurately simulate disease
progress in the tropical highlands. The first
step toward determ in i ng the uti 1ity of the
Cornell simulator is to validate the model,
i.e., compare simulated data with data
collected in the field. This paper describes
progress to date on the validation of this
simulator in the tropical highlands.

Materials and Methods


Description of the simulator
The simulation model we used is based on
one originally developed at Cornell
University. (Bruhn and Fry, 1981 ). In the
late 1990s, the simulator was transcribed
from C to SAS. Except for sorne differences
in the way fungicides are handled, both
the C and SAS versions were identical.
The model in SAS code was designated
version 1 (unpubl.). Subsequently, the SAS
code was modified in the following
manner. Resistance parameters for a
resistant, moderately resistant, and susceptible variety were modified to fit the
specific parameters of Peruvian varieties
Amarilis-INIA (Amarilis) resistant to late
blight, Yungay moderately resistant, and
Tomasa Tito Condemayta (Tomasa) suscep-

88

Research on Patato

tibie (Table 1). The latent period (between


infection and sporulation) was made
temperature-dependent (Andrade-Piedra
N., 2000). For the present study, we used
the SAS code with these modifications,
previously designated as version 2 in
Andrade-Piedra Naranjo (2000).
Field validation experiments
This study involves six field trials, three in
the 1996/97 season and three in the 1997/
98 season (Table 2). Potato tubers of
Catalina (resistant) and Gabriela, Bolona,
or Uvilla (susceptible) from the Instituto
Nacional Autnomo de Investigaciones
Agropecuarias were used in all trials. A
moderately susceptible variety was not
tested in this trial. Conventional fertilization practices were used at each site. Plant
fertility problems were not evident in any
of the trials. lnsecticides at recommended
dosages were used when needed. Fungicide treatments were included but only
non-treated plots are considered in this
study.
At all sites, potato plants were grown in
17.6 m 2 plots consisting of 4 rows, each 4
m long. Each plot was surrounded by a 4
m buffer of barley. Spacing was 0.4 m
between plants and 1 .1 m between rows.
At CIP, Pinantura, and Instituto Andino
Superior de Agropecuaria (IASA), where
plots were located near fields with severe
blight, plants were left to natural infection.
At Tola, plants were inoculated 20 days
after emergence in both planting seasons.
lnoculation was done in the following
manner. lnoculum was produced on potato
tuber slices as previously described (Forbes
et al., 1997) from an isolate collected at
CIP. In Ecuador, potatoes are predominantly i nfected by the clonal 1i neage EC-1
(Forbes et al., 1997) and we assume that
the isolate belonged to that lineage. Two
leaflets from each of two plants within
each plot were inoculated by placing a 20
mi droplet containing 5,000 sporangia/ml
on the abaxial side of the leaflet. The two
plants chosen for the inoculation were
located at opposite sides of the plot.

Table 1. Resistance components far three host resistance levels used in SAS version 2 of the late blight
simulator.
Resistance level 1
Latent period 2
lnfection efficiency3 Lesion expansion 4
Sporulation5
Tomasa (susceptible)
Yungay (moderately
resistant)
Amarilis (resistant)

63

1.0

4.10

257

63

0.9

3.66

228

76

0.8

3.40

107

1 Resistance

parameters were measured experimentally for three Peruvian varieties: Tomasa, Yungay, and Amarilis
(Andrade-Piedra Naranjo, 2000).
2 Base number in hours. Base number is modified in the model using the function latent period = base number /
temperature function, where temperature function = 0.0399* temperature + 0.0831 (Andrade-Piedra, 2000).
3 Relative values used to adjust weather-dependent functions.
4 lncrease in lesion diameter, mm/d.
5 Sporangia/mm 2/d.

Table 2. Research sites near Quito, Ecuador, where experiments were conducted in 1996/97 and 1997/98 to
validate a late blightsimulation model designed and validated previously in the temperate zone.
Site1
Altitude Average
Planting
Variety
Disease
First
Lesions/
(m)
T/RH 2
dates
evaluations
symptoms
plant
(no.) 4
(dd/mm/yy)
(dd/mm/yy) (dd/mm/yy) 3
CIP highland research
3060 11.4/11
18/12/96
Catalina 10/02/9710/02/07
10
station (CIP), 1996/97
Bolona
04/04/97
CIP highland research
3060 12.7/7
10
Catalina 30/12/975/11/97
22/12/97
station (CIP), 1997/98
Gabriela
30/03/98
Instituto Andino
2700 13.4/14 23/01/97
Catalina 06/01/9710
06/01/97
Superior de
Bolona
15/05/97
Agropecuaria (IASA),
1996/97
Catalina 13/04/9820
Pinantura (Pinantura),
3500 9.4/16
31/03/98
22/01/98
Uvilla
16/06/98
1997/98
2500 16.1/14 29/12/97
Catalina 17/02/98Centro Acadmico
04/02/98
Bolo na
Docente Experimental
21/04/98
La Tola (Tola),
1996/97
Catalina 02/02/9902/02/99
Centro Acadmico
2500 16.6/16 02/12/98
Uvila
13/04/99
Docente Experimental
La Tola (Tola),
1997/98
1 The

term in parentheses corresponds to reference to the site in the text and in Figure 1.
befare the slash = average daily temperature (T) in C during the trial; the number after = average hours/day
when relative humidity (RH) was ~ 90%.
3 Estimated date of first symptoms visible in the field. Oisease simulation was initiated one week befare this date.
4 Number of lesions/plant at the beginning of each simulation.
2 Number

Temperature and relative humidity (RH)


were recorded hourly at all sites with
automated data loggers. Rainfall was
recorded hourly at CIP and daily at the
other sites. At CIP, all data were collected
with a Delta-T logger with standard sensors

(Delta-T Devices Ltd., Cambridge, England). At the other three sites,


temperature and RH were collected with
Hobo loggers (Onset Computer Corporation, Bourne, MA, USA). Rain was
measured in standard rain gauges. The

CIP Program Report 1999- 2000

89

weather data used by the simulator are


daily rainfall, number of hours in which
RH 2:: 90%, the average temperature during
the hours that RH 2:: 90%, and the average
temperature for each 24-hour period.
The simulator not only requires weather
data but also initialization for level of
resistance and certain disease epidemic
parameters. When we simulated disease in
the Ecuadorean variety Catalina, we used
resistance parameters for the Peruvian
variety Amarilis; for susceptible Uvilla,
Gabriela, or Bolona, we used parameters
for Tomasa (Table 1 ). For each simulation
exercise, the simulator was initialized
with the date of initial infection, and the
number of lesions/plant for initial infection
(Table 2). All varieties were considered to
be mid-season in maturity.
Simulation was carried out for a period
equal to that for which disease severity
was assessed in the field. Actual severity
of LB was assessed in the field every 7-14
days as described previously (Forbes and
Korva, 1994) on the two central rows of
each plot. Percent infection was converted
to the area under the disease progress
curve (AUDPC) as described previously
(Shaner, 1973).
As a preliminary step toward determining
what factors might be associated with the
predictive capability of the model, we
compared MU DPC with average dai ly
temperature and average number of hours
per day above 90% RH. The MUDPC was
calculated by subtracting the observed
AUDPC from the predicted AUDPC.
Significance of these relations was tested
by regressing MUDPC values on temperature and RH values in separate, simple
1inear regression models.

Results
Field validation of the simulator
In ali six cases, the simulator predicted
slower disease progress for the resistant
variety (Figure 1). This demonstrates that
the model is fairly robust in its handling of

90

Research on Potato

host resistance. However, the overa! 1


predictive capability of the model was not
the same across all sites. The model did a
good job of predicting disease severity for
both susceptible and resistant varieties at
Tola in both seasons. In contrast, the
model predicted less disease than was
observed in IASA (1996/97), CIP (both
periods), and Pinantura (1997/98). Underestimation of disease severity was most
severe at Pinantura. In CIP (1996/97) and
Pinantura (1997/98), the simulator predicted less than 3% disease for the
resistant variety, even though there was
significant disease in observed plots in
each case.
Values of MUDPC were inversely related
(P = 0.0024) to average daily temperatures (Figure 2). There were too few data
points to attempt to describe the relation
between MUDPC and temperature, but in
general the model had greater predictive
capability at the warmer sites. There was
no significant association (P = 0.7048)
between MUDPC values and average
number of hours per day of RH 2:: 90%.

Discussion
This study demonstrated that a patato LB
simulator designed for and validated in the
temperate zone has limited applicability
to tropical highland patato production. The
model simulated realistic differences in
host resistance between a resistant and
susceptible variety, but frequently underestimated disease severity for both.
Comparison of MU DPC with weather
factors indicated that underestimation was
associated with low temperatures characteristic of the main production zone in
Ecuador, between 3000 and 3500 m.
In this study we used SAS version 2 of the
model, which had been validated in Peru
in Comas and Oxapampa. Comas at 2400
m and Oxapampa at 1800 m represent the
lower fringe of patato production in the
central Andes. The validation studies
carried out in Peru demonstrated that the
model worked well for plots not treated

with fungicides (Andrade-Piedra Naranjo,


2000). Average daily temperatures during
the growing season at these locations are
13.8C (Comas) and 16.9C (Oxapampa),
similar to temperatures at Tola (Table 2).
Since the model also accurately simulated
disease in plots not treated with fungicides
at Tola, our data are consistent with those

100
80

from Peru. Sites with lower average


temperatures were not used in Peru.
The parameters of disease initiation (date
of initial infection and number of lesions/
plant) can also affect final disease severity. We tried several higher values for
these parameters for the sites where the
simulator underestimated disease severity,
but these changes had little effect on final

100

CIP,
1996 -1997

80

1 11

CIP,
1997 -1998

60
40
20

01 /15/1997

02/15/1997

03/15/1997

100
80

04/15/1997

IASA,
1996 -1997

12/1 0/1997

80

60

40

40

20

20

o
03/20/1997

04/20/1997

05/20/1997

100
80

02/20/1998

100

60

02/20/1997

01 /15/1998

Pinantura,
1997 -1998

li

"

Zl

.~

04/01/1998

04/15/1998

05/01/1998

06/01/1998

07/10/1998

04/01/1998

05/10/1998

100
TOLA,
1996 -1997

80

60

60

40

40

20

20

03/15/1997

04/15/1997

05/15/1997

06/15/1997

TOLA,
1997 -1998

02/01/1998

03/01/1998

Figure 1. Observed and simulated late blight severity far six experiments conducted near Quito, Ecuador. In each
plot salid squares represent observed disease severity of a susceptible variety; empty squares represent disease

severity of a resistant one. Salid lines represent simulated susceptible (highest line) and resistant varieties. Site
descriptions, varietal descriptions, and sorne parameters used far simulation are given in Table 2.

CIP Program Report 1999 - 2000

91

0.35 -
0.30 - o

0.20 (.)

a..

::>

<t:

0.25

0.15

rr

o
o

0.10 -

-<]

0.05 -

0.00 -

10

11

12

13

14

15

16

17

11

12

13

14

15

Average daily temperatura (C)

Hours per day of RH

16

90%

Figure 2. Relation between MUDPC (a measure of model accuracy) and average daily temperature (A) and
hours per day of RH equal to ar above 90% (B). In each plot, salid squares represent susceptible varieties and
empty squares represent resistant varieties.
disease severity (data not shown). This
aspect of simulation does not appear to be
related to the problem of underestimation.
The resistance parameters we used in SAS
version 2 were derived from Peruvian
patato varieties (Table 1). Nonetheless,
when we simulated disease in Tola, where
the simulator appeared to work best, the
difference in resistance levels between
varieties Catalina and Gabriela was
accurately simulated using the parameters
of the Peruvian varieties Amarilis and
Tomasa, respectively. This occurred in two
trials. Even in IASA, where the simulator
underestimated disease severity, the
relative difference between the resistant
and susceptible varieties was well simulated using the Peruvian parameters. The
overall poor performance of the model in
the other experiments precludes this type
of evaluation, but it would be extremely
fortuitous if the parameters of Amarilis and
Tomasa are useful far many comparisons
between resistant and susceptible varieties. Eventually, the accuracy of the
resistance parameters used in the model
will probably depend on the intended use
of the simulator. Developing general

92

Research on Potato

hypotheses about fungicide usage may


only require general categories of resistance. Using the simulator as a theoretical
tool far analyzing resistance stability
across sites may require specific measurements of components of resistance of the
varieties involved.
The next steps in this line of research will
be to continue the exercise of comparing
simulated disease with observed disease,
using data from an increasingly larger
number of sites. Should the pattern of
underestimation related to low temperature be consistent using a larger number of
trials, we will then begin exploring the
aspects of the simulator that do not
accurately reflect reality. One of the
functions of disease simulation is to
compare the researcher's understanding of
a particular biological system with reality.
At this time, the simulation model represents a summary of our knowledge of the
processes underlying patato LB. The
research reported here, however, demonstrates that we still have much to learn
about patato LB in the low-temperature
areas of the highland tropics.

References
Andrade-Piedra N., J.L. 2000.
Epidemiologa del tizn tardo de la
papa en los Andes peruanos: Validacin
del simulador LATEBLIGHT. MSc thesis.
Universidad Nacional Agraria La
Molina, Lima, Peru. 149 p.
Bruhn, J.A. and W.E. Fry. 1981. Analysis of
potato late blight epidemiology by
simulation modeling. Phytopathology

71 :612-616.
Doster, M.A. and W.E. Fry. 1991. Evaluation by computer simulation of
strategies to time metalaxyl applications for improved control of potato late
blight. Crop Protection 10:209-214.
Doster, M.A., M.G. Milgroom, and W.E.
Fry. 1990. Quantification of factors
influencing potato late blight suppression and selection for metalaxyl
resistance in Phytophthora infestans: A
simulation approach. Phytopathology

80:1190-98.
Fohner, G.R., W.E. Fry, and G.B. White.
1984. Computer simulation raises
questions about timing protectant
fungicide application frequency according to a potato late blight forecast.
Phytopathology 74:1145-47.
Forbes, G.A., X.C. Escobar, C.C. Ayala, J.
Revelo, M.E. Ordonez, B.A. Fry, K.
Doucett, and W.E. Fry. 1997. Population
genetic structure of Phytophthora
infestans in Ecuador. Phytopathology

87:375-380.
Forbes, G.A. and M.C. Jarvis. 1994. Host
resistance for management of potato
late blight. In: Zehnder, G., R. Jansson,
and K.V. Raman (eds.). Advances in
potato pest biology and management.

American Phytopathological Society, St.


Paul, MN, USA. p. 439-457.
Forbes, G.A. and J.T. Korva. 1994. The
effect of using a Horsfall-Barratt scale
on precision and accuracy of visual
estimation of potato late blight severity
in the field. Plant Pathology 43:675-682.
Fry, W.E., A.E. Apple, and J.A. Bruhn.
1983. Evaluation of potato late blight
forecasts modified to incorporate host
resistance and fungicide weathering.
Phytopathology 73:1054-59.
Haverkort, A.J. 1990. Ecology of potato
cropping systems in relation to latitude
and altitude. Agricultura! Systems

32:251-272.
Kato, M., E.S. Mizubuti, S.B. Goodwin,
and W.E. Fry. 1997. Sensitivity to
protectant fungicides and pathogenic
fitness of clonal lineages of
Phytophthora infestans in the United
States. Phytopathology 87:973-978.
Raposo, R. D.S. Wilks, and W.E. Fry. 1993.
Evaluation of patato late blight forecasts
modified to include weather forecasts:
A simulation analysis. Phytopathology

83:103-108.
Shaner G. 1973. Evaluation of slow
mildewing resistance of Knox wheat in
the field. Phytopathology 63:867-872.
Shtienberg, D., M.A. Doster, J.R. Pelletier,
and W.E. Fry. 1989. Use of simulation
models to develop a low-risk strategy to
suppress early and late blight in patato
foliage. Phytopathology 79:590-595.
Shtienberg, D. and W.E. Fry. 1990. Field
and computer simulation evaluation of
spray-scheduling methods for control of
early and late bl ight of pota to. Phytopathology 80:772-777.

CIP Program Report 1999 - 2000

93

Discovery and Evaluation of a Valuable New Source


of Resistance to PLRV: Solanum tuberosum subsp.

andigena
E. Mihovilovich, L. Alarcn, A.L. Prez, and M. Bonierbale1

A series of genetic experiments has been conducted to provide fundamental


information for the exploitation of the native cultivated potato species
Solanum tuberosum subsp. andigena in breeding for resistance to potato
leafroll virus (PLRV) infection. Following positive indications from preliminary
evaluation of 550 clonal accessions of the in-trust collection of potato
germplasm held at CIP, a diallel mating design using seven andigena clones as
parents was developed to study the inheritance of PLRV resistance in this
species. The 21 families generated underwent inoculation by viruliferous
aphids at the seedling stage, and were evaluated for resistance during primary and secondary infection cycles. Levels of resistance to infection were
then assessed in a sample of 20 clones from the andigena germplasm holdings,
and 25 selections from the genetic design. These clonal accessions and selections were inoculated in a RCBD with different levels of aphid pressure and
resistance levels determined by comparison with known resistant and susceptible controls. High to moderate levels of resistance were found in five clonal
accessions, which also transmitted resistance to their families derived from
open-pollinated seed. Genetic analysis of the diallel design revealed that
resistance to PLRV infection is governed mainly by additive gene action,
which accounted for 91 % of the total genetic variance. The finding of a high
frequency of highly resistant progenies in a first cross generation of the highly
resistant accession LOP-868 suggests the presence of an allele with large
effects. This allele is also suspected to be present in two additional resistant
andigena accessions identified in these experiments. The demonstration of
high percentages of resistance in families from crosses between the highly
resistant accession LOP-868 and susceptible tuberosum varieties indicates the
absence of any negative interaction in the expression of resistance genes
from andigena in tuberosum backgrounds.
The native cultivated species 5. tuberosum
subsp. andigena (hereinafter called
andigena) is widely cultivated in the
Andean highlands and represents a highly
variable gene pool for potato improve1

ment. Resistances to many important


potato production constraints have been
identified in andigena and are widely used
in breeding, providing good combinations
of agronomic and quality traits when
crossed to the commercial 5. tuberosum

CIP, Lima, Peru.

CIP Program Report 1999 - 2000

95

subsp. tuberosum (herei nafter cal led


tuberosum) (Estrada, 1999). Quantitative
resistance to late blight, extreme resistance to patato virus X, potato virus Y
(PVY), and potato cyst nematode, as wel 1
as postharvest quality traits are sorne of
the attributes of this species that have
already been exploited in potato breeding
(Estrada, 1999; Mendoza, 1990; Plaisted et
al., 1985). Resistance to PLRV has also
been reported (Brown, 1980; Hooker,
1977), but this resistance has not yet been
used. PLRV and PVY are the two most
important virus diseases of potato, responsible for significant yield and tuber quality
reductions. PLRV contributes to the
degeneration of seed, compelling farmers
to renew their stocks frequently, which
increases their production costs (Mendoza,
1990). Breeding for resistance to PLRV is
constrained by the lack of highly heritable
sources of complete resistance, and by
cumbersome screening procedures that
depend on consistent vector pressure and
several vegetative growth cycles to assess
resistance.
Preliminary evaluation of andigena
accessions in the in-trust germplasm
collection held at CIP identified 77
accessions with putative resistance to
PLRV infection (Bastos, 1997). The objective of the present research was to
determine the potential utility of this
species as a breeding source to increase
levels of resistance to PLRV in potato
varieties. Combining ability was analyzed
to determine the type of gene action
governing the resistance. Levels of resistance and their expression in other genetic
backgrounds were also studied to determine whether the resistance in andigena is
high enough to justify its incorporation into
advanced breeding populations.

Materials and Methods


Characterization of resistance to PLRV
infection in a diallel mating design
Seven andigena clones taken at random
from a group of 77 putatively-resistant

96

Research on Potato

accessions identified in the germplasm


collection were used as parents in a diallel
mating design, according to Method 4 of
Griffing (1956). Andigena accessions are
Andean potato varieties conserved by
vegetative propagation in the germplasm
col lection. The parental accessions were
crossed in all combinations, ignoring
reciprocals, to generate 21 families.
The resulting families were exposed to
primary infection with PLRV by viruliferous aphids in the greenhouse in Huancayo,
Peru, and to secondary infection evaluation in the field, also in Huancayo, during
the 1998 growing season. Both trials were
conducted in a RCBD (randomized
complete block design) with 3 replications
of 20 individuals per family. Seeds of
each family were sown in plastic trays
containing a vapor-sterilized substrate
composed of 1 part soil: 1 sand: 2 peat
moss. Twenty days after sowing, 60
seedlings of each family were transplanted
to Jiffy 7s (Jiffy Products, N.B., Canada),
arranged in groups of 20 seedlings per
plastic tray, and set following the experimental design. PLRV inoculation was
done by shaking PLRV-infected plants of a
susceptible variety, heavily infested by
aphids, over seedling trays as evenly as
possible (Chuquillanqui and Jones, 1980).
The virus isolate u sed in al 1 experiments
was 01 from Peru, which was isolated from
the local susceptible variety Ticahuasi
(L.F. Salazar, CIP, Lima, Peru, pers.
comm.). This isolate was maintained and
propagated in plants of the susceptible
variety Flor Blanca by aphid transmission.
Aphids were killed with pesticide 7 days
after infestation of the seedlings, and
plants were individually transplanted to
4-in. plastic pots. At harvest, two tubers
recovered from each mother plant exposed
to virus in the primary infection trial were
used to test families for secondary infection in the field.
Enzyme-linked immunosorbent assay
(ELISA) was performed on leaf samples
collected from individual plants during

primary infection 30 days after inoculation, and on mixed samples of leaves from
the two daughter plants during secondary
infection trials 50 days after tuber planting. Percentages of resistance were
estimated as the number of noninfected
plants (negative to ELISA), over the total
number of plants tested per family and
replication. Single plants were used to
represent genotypes within families. The
percentage of resistance on a family basis
was taken as a quantitative variable to
represent the mean progeny performance.
Data taken on percentage of resistance
were transformed to are sine for better
approximation of normal distribution.
General combining ability (GCA) and
specific combining ability (SCA) components of variance were estimated from the
expected mean squares of the analysis and
translated to genetic components under a
model of tetraploid inheritance assuming
random chromosome segregation (a = O),
no epistasis, no maternal effects, and
i ndependent assortment:

cr2 SCA = 1/6cr2 D + 1/12cr2T + 1/36cr2 Q


in which: cr\ denotes additive genetic
variance and cr 2 0 , cr 2T' and cr20 denote
dig.enic, trigenic, and quadrigenic
intralocus components of variance,
respectively. cr 2Awas estimated as 4 cr 2cCA
and nonadditive genetic variance was
estimated by 6 cr 2 scA. Narrow sense heritabi l ity was estimated using the formula
h2 = cr2 A/cr2 P in which cr 2 p denotes phenotypic variance.
Determination of resistance levels
clonal evaluation

by

The 7 parental clones used in the diallel


design, a sample of 13 additional
andigena accessions, and a group of 25
putatively resistant selections from one of
its component families underwent clonal
testing to estmate their levels of resis-

tance to PLRV infection. Plants of ali test


accessions were grown from stem cuttings
derived from virus-free in vitro plantlets.
Those of the putatively resistant selections
were grown from stem cuttings of plants
grown from tubers harvested from noninfected plants selected in the secondary
infection trial of the diallel design. The
experiment was conducted in a RCBD of
three blocks, each representing different
levels of aphid pressure. The experimental
unit consisted on 1O plants of each of the
45 experimental clones and 3 controls.
Plants were grown in 7-in. plastic pots
containing the mix described above, under
controlled greenhouse conditions in La
Molina. Primary infection was carried out
by inoculating 30- to 40-day-old plants in
the respective blocks with 25, 50, and 100
viruliferous apterous aphids. The plants in
each block were placed on a separate
table and isolated with anti-aphid net.
Aphids were killed 10 days after infestation, and plants were grown through
tuberization. Two daughter tubers recovered from each pot were planted in the
field at La Malina during the spring season
for the evaluation of secondary infection.
Thus, two daughter plants in the secondary
infection trial were used to represent each
mother plant infected in the primary
infection trial. ELISA was performed on
mixed samples of leaves of the two
daughter plants 50 days after planting.
Percentage of infection was calculated as
the number of infected mother plants
(positive to ELISA) over the total number
of plants tested per accession and per
aphid pressure block. Levels of resistance
were assessed based on the reaction of
known resistant and susceptible varieties
used as controls. The highly resistant,
diploid bred line DW.84-1457, developed
by the lnstitute of Potato Research,
Mlochow Centre, Poland, and the moderately resistant variety Achirana-INTA,
developed by the National lnstitute of
Agronomical Technology (INTA), Balcarce,
Argentina, (Dziewonska and Was, 1994;
Huarte et al., 1990) were used as resistant

GIP Program Report 1999 - 2000

97

controls. The Peruvian variety Perricholi


was used as a susceptible control.

Results
Analysis of the diallel design

Assessment of resistance to PLRV infection


in different genetic backgrounds
The clone with the best combining ability
in the diallel design was crossed as a male
parent to three susceptible tuberosum
varieties: Shepody, Spunta, and Atlantic.
Families from these crosses, and from
open-pollinated seed of the tested clones
provided by gene bank curators, were
assessed for resistance to PLRV infection at
the seedling stage using the inoculation
method described above. Two trials were
conducted in RCBD with three replications
of 40 plants per family. After inoculation,
seedlings from open-pollinating families
were individually transplanted to 4-in
plastic pots; those from crosses with the
tuberosum varieties were transplanted to
the field during the spring season in La
Malina following the design. ELISA was
perfarmed on leaf samples of individual
plants 30 days after transplanting. Mean
percentage of resistance of families was
calculated as the number of noninfected
plants (negative to ELISA) over the total
number of plants tested per family (mean
of the three replications).

Overall infection rates measured in the 21


families of the diallel design were 43% far
primary and 58% far secondary infection
trials. Since 15% more infection was
observed during secondary infection, data
from this trial were used to provide better
accuracy far the genetic analysis. This
difference may be attributed to individuals
that did not achieve sufficient virus titers
to be detected by ELISA by the time this
test was perfarmed in the pri mary i nfection
cycle. Mean percentage of resistance
among families ranged from 7% to 88%
(Table 1 ). No preference of aphids far any
of the families tested was observed, with
aphid counts taken 3 days after infestation
showing a mean number of 1O aphids/
plant. Families derived from the parental
clone LOP-868 showed consistently high
percentages of resistance (> 80%) regardless of the alternate parent. ANOVA far
percentage of PLRV resistance showed
significant differences between families
(P s 0.01 ), and nonsignificant difference
between replications (Table 2). Partitioning of the family variation revealed only
GCA effects to be significant (P s 0.01 ).

Table 1. Percentage of resistance to potato leafroll virus (PLRV) of 21 S. tuberosum subsp. andigena
families generated in a diallel mating design.
PPP-1161
LOP-868
Family
ZIM-440
CCC-4932
CUA-106
HJT-5535
HAW-5687
1O
33
20
43
40
83
13
ZIM-440
17
7
20
83
23
42
85
38
CCC-4932
27
35
PPP-1161
88
83
80
LOP-868
27
CUA-106
HJT-5535
Table 2. Analysis of variance for percentage of resistance of diallel mating design in S. tuberosum subsp.
andigena.
Df
MS
EMS
So urce
2
69.24" 5
Blocks
348.73**
20
Families
1091.56**
GCA
6
efe + cr2s + (p-1) cr 2g

SCA

14
40

Error

30.37lS
16.29

#e + efs
efe

Notes: MS = mean squares for are sine of percentage of non-infected plants in secondary infection trial; GCA = general
combining ability; SCA = specific combining ability; ** = Significant at the 1% level; ns = nonsignificant; CV = 17.3%.

98

Research on Potato

Estimation of genetic components of


variance showed only additive genetic
variance to be higher than twice its
standard error (SE) (Table 3). Since
nonadditive genetic variance was not
significant, this component is assumed not
to contribute to the additive variance
estimation, indicating that resistance to
PLRV in the andigena population is
affected mainly by additive gene actions.
Narrow sense heritability was estimated at
0.89. This value might be overestimated,
since evaluation was performed in one
environment. Based on these results, GCA
effects of the parent accessions involved
in the design were estimated. Only one
parent, the accession LOP-868, showed a
highly significant positive GCA for PLRV
resistance (Table 4). The others showed
negative GCA effects of variable significance levels.
Clonal evaluation of levels of resistance
to PLRV infection
ANOVA for percentage of infection of
the 20 clones and 25 F1 selections evaluated showed significant differences
between clones and aphid pressure blocks
(P ~ 0.01 ). Comparison of the different
aphid pressure levels (blocks) revealed no
significant difference in the percentages of
infection from applying 50 or 100 aphids/
plant (data not shown). Therefore, results
considered percentage of infected plants
under the 25- and 50-viruliferous aphid
pressure levels only. The highly resistant
bred line, DW.84-1457, showed 20%
infected plants under infestation with 50
aphids, and 0% with 25 aphids. On the
other hand, moderately resistant AchiranaINTA, showed 80% infected plants under
infestation with 50 aphids and 30% under
25-aphid pressure (Table 5). The susceptible control, Perricholi, resulted in 100%
infection with both aphid pressures. Based
on the reaction of these controls, the
tested accessions and selections were
considered highly resistant if no more than
20% of their plants became infected under
50-aphid pressure and no plants became

Table 3. Estimates of genetic components of


variances and heritability from diallel analysis.
Estimates
Values
Standard error
cr2A
849.00
363.93
84.53
67 .87
02
(J2~
933.53
CJ2p
949.77
h2
0.89
0.38
Table 4. General combining ability (GCA) far
patato leafroll virus (PLRV) resistance of seven
andigena accessions from CIP in-trust holdings.
Accession
Mean percentage of
GCA
resistance of families
LOP-868
83.9
31.8
CUA-106
41.1
-0.5
CCC-4932
39.7
-2.1
HAW-5687
-3.4
38.3
PPP-1161
-5.3
36.4
-5.7
HJT-5535
35.0
ZIM-440
25.0
-14.9
SE(g) = 1.67.
SE(g-9) = 2.55.

infected under 25-aphid pressure. Those


with no more than 30% of infected plants
under 25-aphid pressure were classified as
moderately resistant.
Following this scale, the parent clone
LOP-868, which had been found as the
only accession with a high GCA for PLRV
resistan ce in the dial lel experiment, was
classified as highly resistant, because none
of its plants were infected under either 50or 25-aphid pressure. Almost al 1 plants of
the other parental accessions became
infected (Table 5). Of the additional 13
clonal accessions tested, two were as
resistant as the highly resistant clone LOP868 (none of their plants becoming
infected under either aphid pressure), and
two were classified as resistant and
moderately resistant, since none of their
plants (resistant), or less than 30% (moderately resistant) became infected under
25-aphid pressure. Almost all plants of the
remaining tested clones became infected
under both aphid pressures. Likewise, of
the 25 putatively resistant clonal selections from a family of the highly resistant

CIP Program Report 1999 - 2000

99

Table 5. Levels of resistance to PLRV infection determined far 7 parental accessions of the diallel design,
13 additional andigena clonal accessions, and 25 progenies from the parent accession LOP-868.
Accession
lnfection by clonal evaluation (%)
Reaction
Resistance in OP25 aphids/plant
50 aphids/plant
families1 (%)
Diallel parent accessions2
Highly resistant
88
o
LOP-868
O
nt 3
Susceptible
100
HAW-5687
80
nt
Susceptible
100
CCC-4932
100
nt
Susceptible
100
PPP-1161
100
nt
Susceptible
100
HJT-5535
100
nt
Susceptible
100
ZIM-440
100
Accessions
nt
Highly resistant
o
o
OCH-7643
Highly resistant
83
o
o
HUA-332
88
Resistant
50
o
VID-11
PPP DGV 94
20
90
Moderately resistant
76
CUP-199
90
100
Susceptible
47
OCH-11195
100
100
Susceptible
42
PPP SA 2103
100
100
Susceptible
29
PPP 1627
100
100
Susceptible
6
UNSAM 51
100
100
Susceptible
4
HMX 1686
100
100
Susceptible
4
CCC-4726
100
100
Susceptible
OCH-8225
100
100
Susceptible
nt
HJA 1489
100
100
Susceptible
nt
Clonal evaluation of 25 progenies from the family of the cross LOP-868 x HJT 5535
22 genotypes
O
O
Highly resistant
2 genotypes
O
10
Highly resistant
1 genotype
20
70
Moderately resistant
Controls
DW.84-1457
O
20
Highly resistant
nt
Achirana INTA
30
80
Moderately resistant
nt
Perricholi
100
100
Susceptible
nt
1 OP = open pollinated.
2 Parent accession CUA-106 was not tested because it was not available in virus-free in-vitro stocks.
3 nt = not tested (open pollinated seed not available in germplasm stocks).

parental clone LOP-868, 24 were as


resistant as this parental clone. Of these,
22 showed no plants infected under either
aphid pressure, and 2 showed only one
plant infected under the 50-aphid pressure.
The remaining selection was classified as
moderately resistant (Table 5).

Expression of resistance in other genetic


backgrounds induding tuberosum
Mean percentage of resistance of the open
pollinated families, derived from the
tested andigena clones, revealed that the

100

Research on Potato

accessions that were most resistant under


clonal evaluation also showed a high
percentage of resistance in their fami lies
(Table 5). Percentages of resistance in
open pollinated families of resistant clones
ranged from 76% to 88%. Conversely, the
percentages of resistance in open pollinated families from susceptible clones
ranged from 1% to 47%. Again, no
preference of aphids for any. family was
observed, and aphid counts three days
after infestation. resulted in a mean number
of 1 O aphids/plant. Since ELISA was

performed on these families only during


the primary infection trial, the presence of
escapes cannot be discarded, but differences in the reaction of open pollinated
families between resistant and susceptible
parents were clear. Open pollinated
families of sorne clones could not be
evaluated because seed was not available
in the germplasm stock.
Finally, families derived from crosses
between the highly resistant clones LOP868, and three known susceptible
tuberosum varieties showed percentages of
resistance as high as those found in
families from crosses between this accession and their susceptible andigena
counterparts. Percentages of PLRV resistance in the tuberosum x andigena
families were over 80% (Table 6).

Discussion
Resistance to PLRV in the commercial
patato tuberosum has proved difficult to
manipulate in breeding due to the nonabsolute and complex genetic nature of the
resistance. For this reason, there are few
commercial varieties with substantial
resistance to PLRV infection (SolomonBlackburn and Barker, 1993). Levels of
resistance high enough to avoid infection
have been reported in wild Solanum
species (Hooker, 1977). Special measures,
however, are often necessary to overcome
crossing barriers, in addition to the significant ti me and effort requ i red to i ntrogress
resistance from unadapted germplasm into
agronomical ly acceptable varieties.
Andigena is a native cultivated species
genetically similar to the commercial
patato due to their clase evolutionary
origins (Estrada, 1999). This explains the
ease of crossability between the two
subspecies, whereas the divergent selection that has been applied to them in
recent history accounts for the heterosis
found in sorne of their hybrids (Tarn and
Tai, 1983). These attributes and those
described befare, make andigena a
desirable genetic source for patato
breeding.

Table 6. Mean percentages of PLRV resistance of


three families from crosses between the highly
resistant andigena accession LOP-868 and three
susceptible tuberosum varieties.
Mean percentage of
Family
PLRV resistance 1
86
SPUNTA x LOP-868
ATLANTIC x LOP-868
85
SHEPODY x LOP-868
91

92.111 x C93.139

40

(Susceptible control)2
1 Mean of three replications.
2 Cross between two PLRV-susceptible breeding clones

generated at CIP.

The present study was conducted to


provide information on the importance of
andigena as a source of resistance to PLRV
infection. Resistance to PLRV infection is
known to depend more on aphid pressure
than on specific strain effects (Ortega,
1983). For that reason, testing for PLRV
resistance in this study was carried out in
young plants by applying different aphid
pressures. That resulted in clear differences
in the levels of resistance of the genotypes
tested. High to moderate levels of resistance to PLRV infection were identified in
five clones of the andigena germplasm.
These clones have also been shown to
transmit similar levels of resistance to their
families derived from sexual seed.
Genetic analysis of the diallel mating
design revealed that resistance to PLRV
infection is mainly due to additive effects
of genes, which accounted for 91 % of the
total variance in this study. This is not
surprising since phenotypic performance of
clones resembled family performance. This
held true for the parental clones of the
diallel. Among them, the only parent
resistant in the clonal evaluation showed a
positive and significant GCA, whereas the
other parents, rated susceptible at the
clonal level, showed negative GCA. Thus,
the phenotypic value resembles the
breeding value of these clones for this
characteristic. This outcome is also
revealed in the high heritability found, and
indicates the potential for good progress in
resistance to PLRV in this species through
individual selection.
CIP Program Report 1999 - 2000

101

The greenhouse test used to evaluate


families for PLRV resistance proved to be
reliable because the differences encountered were largely attributable to genetic
variation. However, it is not known
whether this test would be effective for
selecting the more resistant individuals
within these families. To test this, 25 noninfected plants from the secondary
infection tria! were taken from a family of
the highly resistant parental clone LOP868 and subjected to clona! evaluation as
described above. These non-infected
plants were expected to represent genotypes with sorne leve! of resistance, or less
likely to become infected. Upon intentional exposure to various levels of aphid
pressure, all except one of these selections
proved to be as highly resistant as their
parent, LOP-868. The exception was rated
as moderately resistant. These results
confirm that the greenhouse test allows
individual as well as family selection,
provided selections are subjected to clonal
testing. The occurrence of high levels of
resistance in almost all of these 25 F1
selections was surprising, given our
assumption of the quantitative nature of
this resistance.
The recovery of such frequencies of high
resistance in the F1 generation suggests the
presence in this parental clone of an allele
with large effects. Alleles of large effects
have been found to affect quantitative
traits in other species (Falconer and
Mackay, 1996). This proposed allele is
also expected to explain the high levels of
resistance of the other two highly resistant
andigena clones identified, as well as
those of the 24 resistant selections derived
from the accession LOP-868. Quantitative
trait locus mapping, based on molecular
markers and suitable statistical models,
have demonstrated that different loci
contribute at different rates to the phenotypic variance for quantitative traits.
Although molecular mapping would
facilitate the detection and estimation of
effects of this type of allele, a first and
simpler approach wou Id be to observe the

102

Research on Potato

phenotypic distribution described by a


segregating population. Alleles of large
effects are known to produce a multinomial distribution or a departure from
normality in segregating populations
(Lynch and Walsh, 1998). Therefore, clona!
evaluation should be performed on a
relatively large sample of individuals from
a cross between a highly resistant and a
susceptible parental clone, and the
phenotypic distribution examined. lf an
al lele with large effects is operating,
changes in gene frequencies toward
resistance, larger than those expected
under an assumption of severa! additive
genes with small effects, can also be
anticipated.
Finally, the high percentage of resistance
shown in families from crosses between
the highly resistant parent clone LOP-868
and susceptible tuberosum varieties
indicates the absence of any negative
interaction in the expression of resistance
genes from andigena in tuberosum backgrounds. These observations and the
genetic evidence presented indicate that
advances for this characteristic should
proceed well in the breeding program
using the andigena resistance source.

Acknowledgments
We are gratefu 1 to R. Rivera, F. Sagu ma,
and A. Gmez for their support in greenhouse and field experiments, and to
Dr. H. Mendoza for consultation on
statistical analysis.

References
Bastos, C. 1997. Resistencia en clones de
germoplasma de papa (Solanum sp.) al
PLRV, PVS y APMV. Ingeniero
Agrnomo thesis. Universidad Nacional
del Centro del Per. Huancayo, Peru.

43 p.
Brown, C.R. 1980. lncorporation of virus
resistance and future plans. In: Report
of the planning conference on the
strategy for virus management in
potatoes 11 held 21-25 Apri 1 1980 at the

lnternational Patato Center. CIP, Lima,


Peru. p. 63-78.
Chuquillanqui, C. and R.A.C. Jones. 1980.
A rapid technique for assessing the
resistance of families of patato
seedlings to patato leaf roll virus. Patato
Research 23:121-128.
Dziewonska, M.A. and M. Was. 1994.
Diploid genotype DW.84-1457, highly
resistant to patato leafroll virus (PLRV).
PotaU Research 37:217-224.
Estrada, R.N. 1999. La biodiversidad en el
mejoramiento gentico de la papa.
PROINPA/CID/CIP, Cochabamba,
B.Jlivia. 372 p.
Falconer, D.S. and T.F.C. Mackay. 1996.
lntroduction to quantitative genetics.
Fourth edition. Longman Group Ud.,
London, England. 464 p.
Griffing, B. 1956. Concept of general and
specific combining ability in relation to
diallel crossing systems. Australian
Journal of Biological Science 9:463469.
Hooker, W.J. 1977. Resistance to certain
viruses in patato. In: Report of the
planning conference on the strategy for
virus management in potatoes 11 held
21-25 April 1980 at the lnternational
Patato Center. CIP, Lima, Peru. p 63-73.
Huarte, M., l. Butzonitch, and
A.O. Mendiburu. 1990. Mejoramiento
para Resistencia a Virus en el Programa
Argentino de Papa. In: Hidalgo, O.A.
and H. Rincon (eds.). 1990. Avances en
el Mejoramiento Gentico de la Papa
en los Pases del Cono Sur. CIP, Lima,
Peru. p. 85-93.
Lynch, M. and B. Walsh. 1998. Genetics
and Analysis of Quantitative Traits.

Sinauer Associates, lnc. Publishers,


Sunderland, MA, USA. 980 p.
Mendoza, H.A. 1990. Mejoramiento de
papa para resistencia a los virus Y, X,
as como el Enrrollamiento de las
Hojas: Estrategia de Investigacin para
Procedimientos de Seleccin. In:
Hidalgo, O.A. and H. Rincon (eds.).
1990. Avances en el Mejoramiento
Gentico de la Papa en los Pases del
Cono Sur. CIP, Lima, Peru. p.133-147.
Ortega, E.J. 1983. Deteccin y
variabilidad del virus del enrrollamiento
de las hojas de la papa (PLRV). MSc
thesis. Universidad de Costa Rica,
Turrialba, Costa Rica. 77 p.
Plaisted, R.L., H.D. Thurston,
W.M. Tingey, E.E. Ewing, B.B. Brodie,
P. Gregory, and D. Ave. 1985. The
utilization of Solanum tuberosum ssp
andigena germplasm in potato
improvement and adaptation. CIP
report, lnternational Patato Center,
Lima, Peru. 17 p. (unpublished)
Solomon-Blackburn, R.M. and H. Barker.
1993. Resistance to patato leafroll
luteovirus can be greatly improved by
combining two independent types of
heritable resistance. Annals of Applied
Biology 122:329-336.
Tarn, T.R. and G.C.C. Tai. 1983.
Tuberosum x Tuberosum and Tuberosum
x Andigena patato hybrids: Comparison
of families and parents, and breeding
strategies for andigena potatoes in long
day temperate environments.
Theoretical and Applied Genetics
57:39-46.

CIP Program Report 1999 - 2000

103

Assessment of Latent lnfection Frequency in Progeny


Tubers of Advanced Potato Clones Resistant to
Bacterial Wilt: A New Selection Criterion
S. Priou, C. Salas, F. de Mendiburu, P. Aley, and L. Gutarra1

This paper reports a 2-year evaluation of CIP advanced potato (Solanum


tuberosum L.) clones in a bacterial wilt-infested field (race 3) in Peru. Clones
resistant or moderately resistant to wilt were selected and all tubers harvested for each clone were tested for latent infection by Ralstonia
solanacearum (Smith) Yabuuchi et al. (1992, 1995), using the CIP
postenrichment, enzyme-linked immunosorbent assay on nitrocellulose membrane (NCM-ELISA) kit. A sampling strategy to accurately estimate the
frequency of infected tubers in the clones has been evaluated. This method
will allow consideration of tuber latent infection as a new selection criterion
in breeding for resistance to bacterial wilt. Fifteen clones were found resistant to wilt in both trials (wilt score ~Cruza 148, the resistant control). Taking
3 tubers/plant at random for each of 1O plants or more ensures a probability
of at least 95 % to detect at least 1 infected tuber, whatever the number of
tubers/plant and the infection rate of the clone. lf 1 O plants/clone are to be
assessed, analyzing 30 tubers/clone provides an accurate estimation of the
proportion of infected tubers with a precision level that theoretically varies
between 0.095 and 0.15 for various infection levels. For each clone, tubers
can be taken at random from the heap of all harvested tubers, or 3 tubers/
plant can be randomly sampled from 1O plants. Moreover, data showed that
infected plants produce tubers of all sizes with equal probability of being
infected, thus tubers can be sampled independently of their size.
Bacteria! wilt (BW) or brown rot caused by
Ralstonia solanacearum is a majar constraint to patato production in most
tropical and subtropical regions. In high
elevation areas, potatoes are affected
mostly by cool temperature-adapted,
restricted host range strains of
R. solanacearum (race 3/biovar 2A) that
are principally transmitted through latently
infected tubers (French et al., 1998). No
high level of resistance exists in patato
varieties, but sorne are less susceptible to
1
.

BW and can give high yields in the


presence of the disease (French et al.,
1998). Sorne varieties, such as Cruza 148
and Molinera, do not express wilt in cool
conditions, but can disseminate the
disease through progeny tubers with a high
rate of latent infection (French et al.,
1998). The use of the moderate levels of
resistance that are available, however, can
make a huge impact on ware patato
production in areas where soils are highly
infested if BW-free seed can be provided
(French, 1994; Priou et al., 1999c).

CIP, Lima, Pern.

CIP Program Report 1999 - 2000

105

In the 1990s, advanced potato clones were


obtained from a 14-yr program of breeding
for BW resistance at CIP. These clones
were produced after various crosses with:

for BW resistance from CIP materials


produced in the 1990s. Evaluation of
advanced clones is also being done in
Kenya and Uganda.

Clones derived from Colombian


5. phureja genotypes produced at the
University of Wisconsin's breeding
program, initiated in the 1970s by
Sequeira and Rowe (1969).
Clone AVRDC-1287 derived from
5. sparsipilum and 5. raphanifolium.
Progenitors derived from wild species
5. chacoense and 5. sparsipilum to
combine other sources of BW
resistance.
5. tuberosum subsp. tuberosum genotypes that carried earliness, adaptation
to heat, resistance to late blight (LB)
and root-knot nematode, and immunity
to potato virus X and potato virus Y.
A population of diploid cultivated
species 5. stenotomum, 5. phureja, and
5. goniocalyx from a breeding program
at North Carolina State University for
high yield and good agronomic characteristics (Schmiediche, 1985;
Schmiediche and Martin, 1986; Anguiz
and Mendoza, 1997).

So far, BW-resistant potato genotypes have


been selected mostly on the absence of
wilting symptoms. Very few researchers
have assessed latent infection of stems or
tubers, although latent infection is responsible for the spread of the disease and for
overcoming resistance if infected seed
tubers are planted (French et al., 1998).
This trait has been a neglected objective
in breeding and selection for resistance to
BW, mainly because latent infection is not
associated with aboveground symptoms
(Ciampi and Sequeira, 1980). Ciampi and
Sequeira (1980) and Anguiz and Mendoza
(1997) assessed tuber infection of hybrids
between 5. tuberosum and 5. phureja by
incubating tubers for up to 3 wk at 28C
and then observing symptoms (bacteria!
oozing). Tuber slices taken at the stolon
end were then macerated to isolate the
pathogen on TZC medium. Ciampi et al.
(1980) also incubated small potato slices
in CPG broth at 28C for 24 h and then
plated a loop of the broth on TZC medium
to assess latent infection in resistant
5. tuberosum x S. phureja hybrids. They
associated resistance in the phureja
population to a high degree of tolerance to
the presence of the bacterium in its
vascular system. Anguiz and Mendoza
(1997) found sorne phureja x tuberosum
hybrids resistant to latent infection after
being planted in a race 1-infested field,
but their selection criterion was based only
on wilt incidence of ~ 25%.

Resistance of the phureja-derived materials has been found to be strain-specific


and sensitive to high temperatures (Ciampi
and Sequeira, 1980; French and de Lindo,
1982; Tung and Schmiediche, 1995).
However, resistance to race 3 (biovar 2A)
strains is expected to be more stable than
resistance to lowland strains (race 1) of R.
solanacearum, since race 3 strains are a
genetical ly homogeneous group (French,
pers. comm.). For instance, Cruza 148, the
most resistant variety currently avai lable,
has been found resistant to wilt caused by
race 3 strains of R. solanacearum in the
Philippines (Kloos and Fernandez, 1986);
and in Peru, Brazil, Uganda, Burundi,
Rwanda, and the Democratic Republic of
the Congo (French et al., 1998). In the
absence of better sources of higher levels
of resistance, national programs in Brazil
and China continue breeding and selecting

106

Research on Potato

This paper reports a 2-yr evaluation of CIP


advanced clones in a BW-infested field
(race 3) in Peru. These clones were
produced from the last crosses for BW
resistance made at CIP in 1995. They have
not been tested in any other country.
Clones resistant or moderately resistant to
wi lt were selected and al 1 tubers harvested
for eah clone were tested for latent
infection by R. solanacearum using a

sensitive and specific serological method


developed at CIP: postenrichment NCMELISA (Priou et al., 1999a, 1999b). A
sampling strategy has been evaluated to
accurately estimate the frequency of
infected tubers in the clones to avoid
assessing all tubers harvested for each
clone.

Materials and Methods


Field trials
Advanced clones from the last population
of materials produced in 1996 at CIP's
experiment station in Huancayo (3300 m)
were selected using a criteria for good
agronomic traits. Good agronomic characteristics were considered to be of primary
importance because early adoption by
national programs for variety release will
depend not only upon disease resistance,
but also upon other desirable traits. The
number of clones was thereby reduced
from 1343 to 963. In October 1997 580
clones and in November 1998 383 clones
were planted for prescreening in a BWinfested field at Carhuaz (281 O m), Ancash
Department, Peru, with 5 plants each.
Ninety genotypes resistant to wilt were
selected (41 from the 1997 planting and 49
from the 1998 planting). These clones
were multiplied at Huancayo from healthy
duplicates and then planted at Carhuaz
again in November 1999 (1999-2000
cropping season) for a second testing under
high disease pressure with 25 plants each
(5 replications (rows) of 5 plants).
In 1997 and 1998, 20 miniplots of 5 plants
each of commercial varieties Yungay
(susceptible), Revolucin (susceptible),
Molinera (moderately resistant), and Cruza
148 (resistant) had been planted as controls and randomly distributed in the field.
The variety Molinera is a phureja x
tuberosum hybrid that was released in the
1970s in northern Peru because of its
moderate resistance to BW (French et al.,
1998).

In the 1999 trial, 1O miniplots of 5 plants


each of the same commercial varieties
were planted as controls (2 miniplots of
each variety per replication). Cultural
practices were the same as those recommended for commercial potato crops. In
both years, soil inoculum was enhanced
1 mo after emergence by burying (at root
level) one-eighth of a 9-cm diam. plate
containing a 48-h agar culture of R.
solanacearum CIP311 on Kelman's medium without tetrazolium chloride. Each
culture piece contained approximately
3 x 109 colony-forming units. Strain CIP311
(race 3/biovar 2A) had been isolated from
the same field and severa! sanitation
precautions were taken to avoid pathogen
spread.
Shoes and tools of workers were washed
and disinfected with 1% sodium hypochlorite when they left the field.
The plot was surrounded by a buffer of
maize crops.
On downward sloping fields, a 2 mdeep well collected runoff at the lowest
comer. The runoff water was regularly
disinfected with approximately 1%
sodium hypochlorite.
Harvested tubers not used for the study
were used exclusively for farmers' home
consumption or were burned.
Far each plant, wilt incidence was evaluated at 15-d intervals for 2 mo using the
CIP 1-5 infection scale of wilted plants
(1 = 0%, 2 = 1-25%, 3 = 26-50%, 4 = 5175%, and 5 = 76-100%). The average
score was recorded for the clone (e.g., the
sum of seores for each plant divided by the
number of plants). At harvest (about 130 d
after planting), total yield was recorded for
the clones resistant or moderately resistant
to wilt (0-30% wilt). Weight of rotted
tubers (those exhibiting BW symptoms
upon slicing) was also recorded. At harvest
of the 1998/99 crop, of the 383 planted, 49
resistant clones were selected for high
resistance to wilt (Table 1). Eight additional clones selected in the 1997/98
evaluation tria! are also presented in the

CIP Program Report 1999 - 2000

107

.......

oCX>
~

,.,

(J>
(!>

g.
o

::::1

a"

Table 1. Wilt incidence, yield, and percentage infected tubers (visible and latent) of advanced patato clones plantad in a race 3/Biovar 2A-infested field in Carhuaz, Peru,
selected far their high to moderate resistance to bacteria! wilt in 1998-1999 and 1999-2000 crops.
1999 Planting
1998 Planting
Wilted
Total
Rotted tubers Infectad
Wilt
Wilted
Total
Pedigree
Wilt
Rotted tubers Infectad
plants
at harvest
incidence
plants
Yield
tubers
incidence
Yield
tubers
at harvest
(%)
(% weight)
(%)
(g/plant)
(%)
Score
(%)
(g/plant) (% weight)
seo re
o
o
o
1 (BWH87.344R X TXY.11 )1
1048
1.00
858
27.0
1.00
3
o
1268
903
31
1.00
o
36.4
2 (CRUZA-148 X C90.205)8
1.00
5
o
o
o
1.00
1264
18.5
1.00
702
31.6
8
3 (720118.1 X C90.205)21
o
1378
19.2
1.00
o
672
24.7
1.00
6
o
4 (720118.1 X C90.205)20
o
o
19.2
1.00
o
1556
1.00
860
5
30.7
5 (720118.1 x BWH87.183)3
o
1518
o
36.1
1.04
1320
1.00
1
1
40.3
6 (720118.1 X C90.205)9
o
o
1.04
1.00
460
1
745
1
38.4
7 (C90.205 X 34.73)13
o
o
1.04
1270
1.00
680
1
37.1
8* H89.65.44
3
o
o
1375
32.5
1.12
728
1.00
3
2
31.8
9 (BWH87.415 x DXY-7)16
o
690
9
1.16
4
356
23.1
10 (34.73xTXY.11)2
1.00
4
o
o
893
14.1
1.16
4
1430
33.6
11 (CRUZA-148 x BWH87.344R)1
1.00
5
o
1834
13
9.1
1.16
4
1420
12 (BWH87.230R x C90.205)1
1.00
1
31.1
o
750
4
1.20
948
39.8
13 (720118.1 X C90.205)17
1.00
5
3
o
15
270
22.5
1.20
522
60.7
14 (34.73 x BWH87.344R)1
1.00
5
5
o
7
29.5
1.24
1466
30.7
1.00
1136
6
6
15 (CRUZA-148 x BWH87 .344R)3
o
o
60.0
1.28
7
700
1.00
510
1
18.8
16 (28.68 X C90.205)5
o
o
1.28
20.6
1.00
880
7
598
5
17* 392289.34
1.28
o
950
5
10.4
7
1260
4
30.0
18 (28.68 x BWH87.344R)13
1.00
o
o
1.28
7
1020
675
3
19.0
19* BWH87 .338
1.00
o
o
1840
1.32
1104
53.1
1.00
8
3
20 69.4
o
1972
1
1.32
1060
57.0
1.00
8
4
21 (CRUZA-148 X C90.205)12
o
o
57.1
1.00
136
22 (BWH87.215 X XY.13)106
o
1056
2
17.1
1.00
23 (BWH87.409xXY.16)15
o
o
45.8
1.00
2325
24 23.3
o
1
68.2
1.00
1980
25 (5.63 x BWH87.183)5
2
o
28.6
26 (5.63 x BWH87.344R)8
1.00
560
o
o
16.7
1.00
700
27 {5.63 X C90.205)2
o
6
19.0
1.00
530
28 (28.68 X C90.205)7 '
o
7
52.0
1.00
725
29 (720118.1 X C90.205)2

C")

=ti
"'O

co

;:
3
~

-e
o

;::::.

e.o
e.o
e.o
1
NI

o
o
o

tO

30 (TXY-8 x BWH87.183)16
1.00
31 (TXY-8 x BWH87.344R)6
1.00
32 (BWH87.172R xTXY-2B)12
1.00
33 (BWH87.344R X TXY.11)18
1.00
34 (381064.2xTXY-11)4
1.00
35 (TXY.11 x BWH87.344R)1
1.00
36 (BWH87.420 X C90.205)6
1.00
37* SR 17.50
1.00
1.00
38* 392278.19
39 (BWH87.446R X TXY.2B)1
1.00
40 (720118.1 x BWH87.183)6
1.00
41 (BWH87.407R X TXY.11)1
1.00
42 A1-12.55
1.00
43 (C91.640 X 34.73)2
1.00
44 (BWH87.289xXY-13)103
1.00
45 (381064.2xTXY-11)1
1.00
46* BWH87 .176 x XY.16)24
1.00
47* 392270.2
1.00
48 (720118.1 x BWH87.183)5
1.00
49 (BWH87.230R X C90.205)5
1.00
50 (720118.1 x BWH87.183)1
1.00
51 * 392285.72
1.00
1.00
52 (3.5 X C90.205)3
53 (BWH87.230R X TXY.11)3
1.00
54 (BWH87.528R X TXY.11)1
1.00
55 (5.63 x BWH87.183)1
1.00
56 (TXY-8 x BWH87.183)8
1.00
57 (28.68 x BWH87.344R)6
1.00
Yungay (S) t
2.83
Revolucion (S)
4.36
Molinera (MR) t
3.18
Cruza 148 (R} t
1.15
* Clones selected from the 1997/98 evaluation trial.

o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
45.7
84.2
54.5
4.0

t S = susceptible, MR = moderately resistant, R = resistant.


- Missing data.

1928
323
1340
372
1444
620
1494
1320
740
914

3
4
9
41
2

616
633
780
517
250
367
480
1340
1700
1580
767
950
925
460
1540
1300
880
628
1094
870
1027

25.2
35.4
43.4
60.0

o
4
o
o
5

o
o
52
o
o
o
o
22
o
o
21
12
30
6
2
10
21
50
15
3

47.1

9.5
34.2

1.40
1.40
1.40
1.48
1.48
1.48
1.56
1.64
1.68
1.68
1.68
1.68
1.80
1.84
1.88
1.88
1.92
2.04

10
10
10
12
12
12
14
16
17
17
17
17
20
21
22
22
23
25

1428
826
809
564
456
521

3
1
7
3
4
13

584
507
920
920
656
686
558
1362
676
1264
1018

4
4
2
4
2
6
5
2
4
7
11

2.08
2.12
2.20
2.20

26
28
30
30

660
1178
2080
1314

2
21
23
10

33.8
41.5
42.6

3.78
4.66
2.36
1.24

70
92
33
6

1106
874
862
1349

34
46
27
12

19.5
72.8

50.0
24.8
10.3
20.0
23.5
36.6
14.6
37.0
66.3
40.4
38.5

33.3
45.5
9.0
21.2
71.3
71.5
27.6
42.2

clones column of Table 1. Among these 49


clones, 30 have been analyzed for latent
infection. Ali tubers were harvested from
each of the 5 plants in separate bags and
analyzed in postenrichment ELISA for
latent infection. In March 2000, from the
90 planted, 43 clones have been selected
far moderate to high resistance to wilt
(wilt score = 2.2, i.e., 30% wilt), and ali
tubers harvested from 1O plants (randomly
selected from the 25 planted) of 35 clones
were processed to assess latent infection
rate.

Laboratory trials
Tubers were washed in tap water and
disinfected by dipping them in 1% sodium
hypochlorite for 1O min. They were then
graded in three size categories: 1.5-2.5
cm, 2.6-5 cm, and > 5 cm. Tubers were
then tested individually for latent infection
by R. solanacearum using the CIP
postenrichment NCM-ELISA kit as described by Priou et al. (1999a, 1999b). A
total of 2246 tubers were tested in 1999
and 4329 tubers in 2000. Upan cutting the
stolon end of the tuber to remove the
vascular ring for testing in ELISA, tubers
that exhibited visible oozing were not
further processed, and the percentage of
tubers showing visible symptom was
scored per clone. In 1999, the enriched
samples were streaked on modified
Kelman's medium (French et al., 1995) to
confirm the presence of the bacterium.
When ELISA was positive but the bacteria
could not be isolated, DNA amplification
by PCR was performed from enriched tuber
extracts (Priou et al., 2001 ).

Statistical analyses
Qualitative assessment
Tables of probabilities were calculated
according to the binomial distribution to
determine the probabilities of finding at
least one infected tuber in a total number
of sampled tubers per clone. This probability is:
1 - (1 - p)K

11 O

Research on Potato

where:
P is the probability of finding one
i nfected tu ber.
(1-p)k is the probability of finding no
infected tubers in k tubers.
K is the number of sampled tubers per
clone for analysis, e.g., the number of
tubers sampled for each plant multiplied
by the number of plants per clone.
The main assumptions are that the disease
spreads uniformly throughout the field,
each plant (of a given clone and plant
infection leve!) has the same probability of
being infected at the same rate of infection, and the probability of any tuber to be
infected is the same, independent of the
plant.

Quantitative assessment
Statistics for estimation of sample size for
proportions based on the normal distribution have been used to determine the
necessary number of tubers/clone to
sample at random to accurately estmate
the true frequency of infected tubers that
would be obtained by processing ali tubers
of the clone. The estimation avoids having
to process al 1 tubers/clone. The sample
size n depends on the precision leve!
required d, the infection rate per clone p,
and the population size N = total number
of tubers for the clone in the field:

= a/N; q = (N-a)/N and p + q = 1


where a = number of infected tubers/
clone, N = total number of tubers/clone,
p = proportion of infected tubers/clone,
and q = proportion of healthy tubers/clone.
p

From these proportions p and q we obtain


the preliminary sample size n0 :

no= (t2 p q)/d i


where t = value obtained from the Student
t table, and d = precision leve! defined as
IEstimated p - true pi.
The sample size n is calculated as follows:

n = n0 /f

where n = sample size, and f = correction


factor taking into account the whole
population size: f = 1+ (njN).

Results and Discussion


Clone evaluation
From 580 clones planted in 1997, 41
(7.1 %) have been selected for high
resistance to wilt (0% wilt). From these,
only eight remained after a second
evaluation in 1998/99. These eight clones
are listed together with clones selected
after the 1998 planting in Table 1. From
383 potato clones planted in 1998, 49
(12.8%) were highly resistant to wilt (Table
1). Severity of the disease was much
higher in 1999/2000. Thus, clones that had
a BW incidence up to the variety Molinera
score were selected (2.2, i.e., 30% wilt),
considering that these clones had no wilt
in 1997 or 1998 evaluation trials. A total
of 43 clones have been retained from 90
planted (Table 1). Only five clones (5.5%)
were highly resistant to wilt (0% wilt).
Clones 2 and 5 are very promising since
they had no wilt and yielded 1268 and
1556 g/plant, respectively, in spite of very
high disease pressure. All clones listed in
Table 1 have been tested twice, but from
the 57 selected in 1997 and 1998 (1998
Planting column), only data of those
whose wilt score was < 30% in the second
evaluation in 1999 are presented in the
table.
In the 1998 trial, 26 of the 57 selected
clones had no visible symptoms at harvest
and 21 had < 10% rotted tubers, performing much better than Yungay, Revolucin,
and Molinera control varieties (Table 1). In
the 1999 crop, of the 43 selected clones,
only one clone did not exhibit rotted
tubers at harvest and 37 had < 10% rotted
tubers. In both years, al 1 selected clones
showed latent infection with R.
solanacearum, but with various frequencies of tuber infection (Table 1). In the
1998 trial, 11 clones were found very
promising (< 20% infected tubers). Five of
these yielded 1300 g/plant or more in spite

of high disease pressure. Cruza 148, the


resistant control, had 42% infected tubers
and yielded 1027 g/plant; and Molinera
(moderately resistan!) had 27% infected
tubers and yielded 870 g/plant (Table 1).
Only 7 of the 11 promising clones identified in 1998 were analyzed in 1999, and
none of them exhibited < 20% tuber
infection. In the second evaluation, 14
clones still surpassed the Cruza 148
control with lower wilt incidence, and 12
had < 30% infected tubers.
Thus, these selected clones would be
valuable for ware potato production in
highly infested soils. But tubers should be
sold right after harvest since they harbor
latent infection and could rot during
storage. Moreover, tubers should not be
used as seed to avoid dissemination of the
disease and the breakdown of resistance
that may occur when seed is infected, as
has been reported for phureja-derived
varieties (French et al, 1998).
Susceptibility to tuber infection and to
aboveground symptoms was not correlated
(Table 1). For instance, Cruza 148, the
most resistant variety, showed low wi lt
incidences ranging from 4% to 6% in both
trials, but exhibited high rates of infected
tubers (40-70%). Molinera had 54.5% wilt
and 27.6% infected tubers in 1998 and
33% wilt and 19.5% infected tubers in
1999. These are much higher susceptibilities than those reported for Molinera in the
1970s.
From the 2246 tubers analyzed in 1998,
67 4 (30%) were positive in ELISA. The
pathogen could not be isolated in 30 of
these tuber extracts. Presence of R.
solanacearum was confirmed by PCR in 2 7
of these 30 doubtful samples, leading to a
possible cross-reaction rate of only 0.44%.

Effect of tuber size on infection level


Tubers have been graded to evaluate the
effect of tuber size on latent infection
levels. The size effect was not significant
according to variance analysis and WallerDuncan's test at P = O.OS: the average

CIP Program Report 1999 - 2000

111

percentages of infected tubers among the


35 clones were 31 .07, 33.22, and 33. 1 7%
for tubers of 1 .5-2.5 cm, 2.6-5 cm, and
> 5 cm, respectively. This result shows that
infected plants produce tubers of all sizes
with equal probability of being infected,
thus tubers can be sampled independently
of their size. This is also important information for sampling tubers from seed lots
for the analysis of BW infection.

Determination of the optimum sample size


Qualitative assessment
Considering the level of tuber infection
established as a selection criterion (equal
to or less than 30% infected tubers), the 30
(1998 planting) and 35 (1999 planting)
selected clones could be divided into two
groups: Group 1 with an average of::; 30%
infected tubers, and group 2 with > 30%
infected tubers. Wide variability in the
number of tubers per plant among clones
made it necessary to make classifications
within these two groups by the number of
infected tubers (visible and latent) in
relation to the number of tubers per piant.
Three classes were defined according to
the average number of tubers per plant by
cluster analysis using SAS (1989). The
average infection rate per plant was
obtained within these clusters (Table 2).
Although not al 1 the same clones ha ve
been evaluated for their latent infection
rate in tubers in the 1998 and 1999 crops
(only 1 5 clones were assessed twice), the
average infection rates according to the
average numbers of tubers per plant were
similar in both years and for both groups of
clones (Table 2). Also similar were the
averages of tuber infection: 30% in 1998

(for the 30 clones) and 34% in 1999 (for


the 35 clones). We can thus conclude that
the number of tubers per plant does not
affect the infection rate of the clone,
validating the use of the binomial distribution with the factor K as the total number
of tubers sampled per clone regardless the
number of tubers per plant.
The application of the binomial distribution shows that the probability of detecting
at least one infected tuber is given according to the number of tubers sampled for
each plant for a given number of plants
per clone (Table 3). For instance, if 1O
plants or more are available per clone,
taking three tubers per plant would ensure
a probability of at least 95% of detecting
the pathogen whatever the latent infection
rate of the clone (Table 3).

Quantitative assessment
Statistical analysis was appl ied to determine the optimum number of tubers to
sample per clone that al lows an accurate
estimation of the proportion of infected
tubers of the clone. Sample sizes (n) to be
analyzed in ELISA were calculated with
proportions of infected tubers (p) (at
various levels: minimum, average, and
maximum obtained from the 35 clones in
the 1999 trial) and precision levels (d),
according to the total number (N) of tubers
harvested per clone (depending on the
number of plants and on the multiplication
rate of the clone). The average number of
tubers/clone from the harvest of 1O plants
in the 1999 trial was 124 (12.4 tubers/
plant). Precision levels have been chosen
according to their significance for various
infection rates, and to limit the sample

Table 2. Average infection rates/plant in relation to the average tubers/plant far CIP advanced clones
separated in two groups according to the proportion of infected tubers (results obtained in 1998/99
and 1999/2000 evaluation trials).
Group 1
Group 2
~ 30% infected tubers
> 30% infected tubers
lnfection rate/plant (av)
lnfection rate/plant (av)
1998 planting
1999 planting
1998 planting
1999 planting
No. tubers/plant (av)
0.30
0.20
0.40
0.40
10
0.20
0.20
0.45
0.50
20
0.50
0.53
0.20
0.23
30
112

Research on Patato

Table 3. Probabilities to obtain at least one infected tuber from (x) tubers sampled (e.g.,
5, 1O, 15, 20, or 25 plants according to the clone infection rate (IR).
Probability using 5 plants
IR
(x=1)
(x=2)
(x=3)
(x=4)
(x=5)
(x=6)
(x=7)
0.1
0.410
0.651
0.794
0.878
0.928
0.958
0.975
0.2
0.893
0.672
0.965
0.988
0.996
0.999
1
0.3
0.832
0.972
0.995
0.999
1
1
1
0.4
0.922
0.994
1
1
1
1
1
0.5
0.969
0.999
1
1
1
1
1
1
0.6
0.990
1
1
1
1
1
Probability using 10 plants
IR
(x=2)
(x=3)
(X=1)
(x=4)
(x=5)
(x=6)
(x=7)
0.1
0.651
0.878
0.958
0.995
0.985
0.998
0.999
0.2
0.893
0.988
0.999
1
1
1
1
1
1
1
1
0.3
0.972
0.999
1
1
0.4
0.994
1
1
1
1
1
1
0.5
1
1
1
1
0.999
1
1
0.6
1
1
1
1
1
1
Probability using 15 plants
IR
(x=1)
(x=2)
(x=3)
(x=4)
(x=5)
(x=6)
(x=7)
0.991
0.1
0.794
0.958
0.998
1
1
1
1
1
0.2
0.965
0.999
1
1
1
1
0.3
0.995
1
1
1
1
1
1
0.4
1
1
1
1
1
1
Probability using 20 plants
IR
(X=2)
(x=4)
(x=5)
(x=6)
(x=7)
(x=1)
(x=3)
0.998
1
0.1
0.878
0.985
1
1
1
1
1
1
1
1
0.2
0.988
1
1
1
1
0.3
1
1
1
0.999
1
1
1
1
1
1
0.4
1
Probability using 25 plants
(x=6)
(x=7)
(x=4)
(x=5)
(x=3)
(x=1)
(x=2)
IR
1
1
1
1
1
0.928
0.995
0.1
1
1
1
1
1
1
0.2
0.996
1
1
1
1
1
1
1
0.3
1
1
1
1
1
1
0.4
1
sizes to reasonable numbers. For instance,
for clones having an infected tuber rate of
0.4-0.6, a precision of 0.15-0.20 is
sufficient. This proportion would quickly
increase after harvest in any case. But for
genotypes having only a 0.1 proportion of
infected .tubers, the precision should be
higher (< 0.1) to avoid having negative
detection and rating the clone as resistant.
lf 1 O plants/clone are to be assessed,
analyzing 30 tubers/clone provides an
estimation of the proportion of infected
tubers with a precision that varies between
0.095 and 0.15 for various infection levels

x=1) from each of

(x=B)
0.985
1
1
1
1
1

(x=9)
0.991
1
1
1
1
1

(x=B)
1
1
1
1
1
1

(x=9)
1
1
1
1
1
1

(x=B)
1
1
1
1

(x=9)
1
1
1
1

(x=B)
1
1
1
1

(x=9)
1
1
1
1

(x=B)
1
1
1
1

(x=9)
1
1
1
1

(Table 4). However, since the normal


distribution tends to overestimate the
sample size, a simulation was conducted
using EXCEL (version 2000 using the
RANO function) where 25, 30, 35, or 40
tubers were randomly selected from all
tubers harvested from 1 O plants/clone, for
35 clones in 1999/2000. Fifteen simulations were done and the average estimated
infection rates for the 35 clones were
compared to the real proportion of infected
tubers. The estimated proportions were
very close to the real one with the four
sample sizes tested (Table 5). Precision
leve Is ranged from 0.0024 to 0.0029, mu ch

CIP Program Report 1999 - 2000

113

Table 4. Sample sizes (n) to analyze in ELISA to estimate the proportion of infected tubers according to the
total number of tubers N harvested per clone, far various infection levels.
Proportion p of
Precision level
Number (n) of tubers to be sampled according to the
infected tubers
d = 1Est. p - p 1
total number (N) of tubers per clone1
of the clone
(N=100)
(N=200)
(N=400)
(N=600) (N=lnfinite)
Minimum = 0.1
0.05
59
83
104
113
139
0.095
29
33
36
37
39
0.10
45
58
68
72
81
Average = 0.3
0.15
27
31
34
34
36
0.15
30
35
38
39
41
Maximum = 0.6
0.20
20
22
23
24
24
1

n = nclf where n0 = (t2pq)/d 2 and f = 1+(nofN). The Student's t value ata 95% confidence leve! was used.

Table 5. Precision levels (d) obtained after simulation of 15 random samplings of (n) tubers to be analyzed
in ELISA to estimate the proportion (p) of infected tubers far the 35 clones selected in 1999-2000 .
Number n of tubers to be
(n=25)
(n=30)
(n=35)
(n=40)
Probability
analyzed per clone
True p
Estimated p
0.3390
0.3389
0.3386
0.3385
0.3414
Precision level: d = 1Est. p - p 1
0.0024
0.0025
0.0028
0.0029
lower than the theoretical level of 0.1 5
shown in Table 4. Thus, analyzing 30
tubers per potato genotype when 1 O plants
are assessed gives an accurate estimate of
the proportion of infected tubers.
Furthermore, according to the results
obtained with the qualitative assessment
of sample size, this sample size would
ensure detection of tuber infection even if
the clone real infection rate were low
(equal or less than 0.1) ata probability
ranging from 95% to 100% (Table 3), so
that the clone could not be misevaluated
as resistant. For each clone, 30 tubers can
be taken at random from the heap of al 1
harvested tubers or three tubers per plant
can be randomly sampled from 1O plants
or two tubers from 15 plants.
Thus in a trial evaluating 20 clones with
approximately 300 tubers each harvested
from 25 plants, the number of tubers to be
processed would be 600 instead of 6000, if
ali tubers were to be analyzed. That would
save a lot of labor and supplies costs, and
moreover more tubers would remain for
other analysis (quality, storage).

114

Research on Potato

Conclusion
Breeding for resistance to BW at CIP has
resulted in a high to moderate level of
resistance to wi lt, superior to the phureja
x tuberosum hybrids. However, a high
frequency of latent infection is still present
in tubers at levels comparable to those
observed for S. phureja-derived clones,
such as Molinera, or for Cruza 148, one of
the less susceptible varieties. Ciampi et al.
(1980) found that sorne phureja x
tuberosum hybrids did not harbor tuber
latent infection by R. solanacearum race 3
when grown under cool conditions in the
greenhouse, even after incubation of
clones for 70 d at 28C. Anguiz and
Mendoza (1997) also reported the absence
of latent infection in sorne phureja-derived
clones after growing them in a race 1infested field. The detection procedure
they used (isolation on TZC medium),
however, lacked sensitivity compared with
postenrichment NCM-ELISA, which is 100to 1000-fold more sensitive (Priou et al.,
1999a). Probably, resistance to latent
infection does not occur in the genetic
base of the parents used for the crosses
that led to these materials. Latent infection in tubers was not a selection criterion

when Thurston and Lozano (1968) and


Sequeira and Rowe (1969) selected
original phureja genotypes used in the
Wisconsin breeding program for resistance
to BW. Neither was tuber infection a
criterion in the further two decades of
breeding.

in PVX and PVY immune autotetraploid


potatoes. Fitopatologia 32:71-80.
Ciampi, L. and L. Sequeira. 1980.
Multiplication of Pseudomonas
solanacearum in resistant potato plants
and the establishment of latent
infections. American Potato Journal

In many solanaceous species and in


groundnut susceptible to R. solanacearum,
sources of resistance to the pathogen have
not yet been identified (Grimsley and
Wang, 1994). Breeding programs will
benefit from a screening procedure that
evaluates tuber latent infection to assess
overall susceptibility of plant genotypes
apart from wilt. In tomato, this trait is
being taken into account since Prior et al.
(1994) developed a new criterion for
breeding tomatoes for BW resistance that
assesses stem colonization by R.
solanacearum. Therefore, the criteria
formerly used in the course of selection for
potato resistance to BW should be reappraised. The results presented here set the
stage for a new selection procedure.
Further analysis should permit us to
determine the effect of this additional
selection criterion on the stabi 1ity of
resistance or yield of new breeding lines.
Since most breeding programs or evaluation trials in developing countries use the
same genetic base, or even the same
populations previously produced by CIP's
breeding program for BW resistance, this
sample strategy should be suitable for
them as well. This method will also allow
searching for immunity (e.g., absence of
plant wilt and latent infection in stems and
tubers) in wild and native species of
potatoes, a project initiated at CIP in
2000. In this case, a higher precision level
would be applied. Thus, a larger sample
size would be necessary to accurately
estmate lower infection rates.

Ciampi, L., L. Sequeira, and E.R. French.


1980. Latent infection of potato tubers
by Pseudomonas solanacearum.
American Potato Journal 57:377-386.
French, E.R. and L. de Lindo. 1982.
Resistance to Pseudomonas
solanacearum potato: Strain specificity
and temperature sensitivity.
Phytopathology 72:1408-1412.
French, E.R. 1994. Strategies for integrated
control of bacteria! wilt of potatoes. In:
Hayward, A.C. and G.L. Hartman (eds.).
Bacteria! wilt: The disease and its
causative agent, Pseudomonas
solanacearum. CAB lnternational,
Wallingford, UK. p. 199-207.
French, E.R., L. Gutarra, P. Aley, and
J. Elphinstone. 1995. Culture media for
Pseudomonas solanacearum: lsolation,
identification and maintenance.
Fitopatologia 30:126-130.
French, E.R., R. Anguiz, and P. Aley. 1998.
The usefulness of potato resistance to
Ralstonia solanacearum, for the
integrated control of bacteria! wilt. In:
Prior, P., C. Allen, and J. Elphinstone
(eds.). Bacteria! wilt disease: Molecular
and ecological aspects. INRA Edition.
Springer Verlag, Berlin, Germany.

References
Anguiz, R.J. and H.A. Mendoza. 1997.
General and specific combining
abilities for resistance to bacteria! wilt
(Pseudomonas solanacearum E.F. Smith)

57:319-329.

p. 381-385.
Grimsley, N. and J.-F. Wang. 1994. Chair's
perspective: Host resistance. In:
Hayward, A.C. and G.L. Hartman (eds.).
Bacteria! wilt: The disease and its
causative agent, Pseudomonas
solanacearum. CAB lnternational,
Wallingford, UK. p. 197-199.
Kloos, J.P. and B.B. Fernandez. 1986.
Evaluation of patato germplasm for
resistance to Pseudomonas
solanacearum (E.F. Smith) and
adaptation in Mindanao. The Philippine
Agriculturist 69:263-276.

CIP Program Report 1999 - 2000

115

Prior, P., V. Grimault, and J. Schmit. 1994.


Resistance to bacteria! wilt
(Pseudomonas solanacearum) in to mato:
Present status and prospects. In:
Hayward, A.C. and G.L. Hartman (eds.).
Bacteria! wilt: The disease and its
causative agent/ Pseudomonas
solanacearum. CAB lnternational,
Wallingford, UK. p. 209-223.
Priou, S., L. Gutarra, and P. Aley. 1 999a.
Highly sensitive detection of Ralstonia
solanacearum in latently infected
potato tubers by post-enrichment ELISA
on nitrocellulose membrane. EPPO/
OEPP Bulletin 29:117-125.
Priou, S., L. Gutarra, H. Fernandez, and
P. Aley. 1999b. Sensitive detection of
Ralstonia solanacearum in latently
infected potato tubers and soil by postenrichment ELISA. lmpact on a
changing world. Program report 199798. lnternational Potato Center, Lima,
Peru. p. 111-121.
Priou, S., P. Aley, E. Chujoy, B. Lemaga,
and E.R. French. 1999c. CIP Series IV111, lntegrated control of bacteria! wilt of
potato. CIP slide training series (57
slides and a 30 pages-guide in English
and Spanish). lnternational Potato
Center, Lima, Peru.
Priou, S., R. Torres, A. Villar, J. Matinez,
O. Barrea, L. Gutarra, and F. de
Mendiburu. 2001. Optimization of
sample size for the detection of latent
infection by Ralstonia solanacearum in
seed tubers in the highlands of Peru and
Bolivia. Scientist and farmer - Partners
in research for the 2l5t century, program

116

Research on Potato

report 1999-2000. lnternational Potato


Center, Lima, Peru. p. 143.
SAS. 1989. SAS lnstitute lnc., SAS/STAT
User's Guide, Version 6, Fourth Edition,
Volume 1, Cary, NC, USA. 943 p.
Schmiediche, P. 1985. Breeding potatoes
for resistance to bacteria! wilt caused
by Pseudomonas solanacearum. In
Persley, G.J. (ed.). ACIAR Proceedings
no. 13. Canberra, Australia. p. 105-111.
Schmiediche, P. and C. Martin. 1986. The
use of wild species in breeding for
resistance to bacteria! wilt
(Pseudomonas solanacearum). American
Potato Journal 63:453.
Sequeira, L. and P.R. Rowe. 1969.
Selection and uti 1ization of Solanum
phureja clones with high resistance to
d ifferent strai ns of Pseudomonas
solanacearum. American Potato Journal
46:451-462.
Thurston, H.D. and J.C. Lozano. 1968.
Resistance to bacteria! wi lt of potatoes
in Colombian clones of Solanum
phureja. American Potato Journal
45:51-55.
Tung, P.X. and P. Schmiediche. 1995.
Breeding potato for resistance to
bacteria! wi lt (Pseudomonas
solanacearum): Looking for stable
resistance? In: Hardy, B. and E.R.
French (eds.). lntegrated management
of bacteria! wilt, Proceedings of the
international workshop. lndian Council
of Agricultura! Research, New Delhi,
India, and lnternational Potato Center
(CIP), Lima, Peru. p. 173-176.

Evaluation of Bt-cry1la1 (cryV) Transgenic Potatoes


on Two Species of Potato Tuber Moth, Phthorimaea
operculella and Symmetrischema tangolias
(Lepidoptera: Gelechiidae) in Peru
A. Lagnaoui1, V. Caedo 1, and D. S. Douches2

The effect of Bt-cry1/a1 (cryV now designated cry1/a1 under revised nomenclature) of potato transgenic plants on the two species of potato tuber moth,
Phthorimaea operculella (Zeller) and Symmetrischema tangolias (Gyen), was
tested in Peru. Detached leaf bioassays were done using 1 O neonate larvae
per replication on each transgenic line of bred potato varieties Atlantic and
Spunta. Mortality in Atlantic transgenic plants was lower for P. opercule/la,
ranging from 18 to 34%, than for S. tangolias ranging from 40 to 94%. All
transformed Spunta lines tested showed high levels of mortality in both species, with mortality ranging between 80 and 98%. The Bt-cry1/a1 gene offers
another source of resistance that can be pyramided for effectiveness toward
the development of durable resistance to PTM and other insect pests.

The potato tuber moth, (PTM)


(Phthorimaea operculella (Zel ler)) is one of
the most damaging potato (Solanum
tuberosum L.) pests in tropical and subtropical areas. Damage is often observed
on patato fol iage, stems, and tubers.
Although yield loss seldom occurs from
PTM field infestations, quantity and
quality losses in storage can be drastic in
warmer climates where losses can reach
100%. A second PTM species,
(Symmetrischema tangolias (Gyen)) is also
a serious potato pest in the Andean region.
Severa! control components have been
identified for both pests. Among them, the
soil bacterium Bacillus thuringiensis (Bt)
has proved effective in reducing PTM
infestations in stores. So far, host plant
resistance work has not yielded any
promising material with appreciable levels
of resistance. The expression of the Bt
1

CIP, Lima, Peru.


Michigan State University, East Lansing, Michigan, USA.

genes confers a nonconventional host


plant resistance to this pest.

P. operculella is the most cosmopolitan,


attacking potato foliage, stems, and
tubers, although field infestations rarely
cause serious yield losses. The main
damage caused by this insect is to stored
tubers; in warm climates losses can reach
100%. Symmetrischema tangolias is
considered a serious pest throughout the
Andean region, causing significant economic losses in potato stores in Peru and
Bolivia (Ewell et al., 1994). Larvae burrow
into tubers making them susceptible to
storage rot.
Host plant resistance is a key component
of any integrated pest management (IPM)
program. Resistant crop varieties, as a
cultural management practice, are often
used as a foundation for sound IPM
strategies. Host plant resistance work in
potato, thus far, has not yielded any

GIP Program Report 1999 - 2000

117

material with appreciable levels of


resistance. Chemical and biological
insecticide applications, including the use
of Bt, have been used by farmers worldwide. However, the feeding behavior of
Symmetrischema tangolias, leaf mining
and tuber burrowing, makes the use of
insecticide less efficient. An alternative
for managing tuber moths involves the use
of plant resistance and Bt in a transgenic
form. Insecticida! protein/toxin genes from
Bt have been cloned and inserted into
various cultivated crops (Barton and
Miller, 1993). These Bt toxins are crystal
proteins specific to lepidopteran larvae
and safe to nontarget organisms (Bauer,
1995). CIP and Michigan State University
(MSU), USA, have produced first-generation transgenic potatoes expressing genes
from Bt subspecies. The CIP transgenic
expresses the Cry1 A(b) and the MSU
potato the cry7 Ja 7 gene. Transgenic potato
varieties with the insecticida! crystal
protein Cryl A(b) provide high levels of
resistance to P. operculella in foliage and
tubers (Jansen et al., 1995; Caedo et al.,
1999). A toxin gene for cry1Ja7, active
against lepidoptera and coleoptera, was
codon-modified to increase its expression
level and transformed into potato to
achieve control of P. operculella (Douches
et al., 1998; Wedstedt et al., 1998).
This paper describes a study of larval
development of two PTM species on Btcry1Ja7 transgenic potato plants in Peru.
The objective of this study was to test the
effect of Bt-cry7 Ja 7 (cryV now designated
cry7 Ja 7 under revised nomenclature
(Crickmore et al., 1998)) potato transgenic
plants on two of the PTM species in Peru.

Material and Methods


Transgenic lines
The transgenic material used in this study
was developed at MSU. Four lines of
Atlantic and five lines of Spunta were
transformed with a codon-modified Btcry7 Ja 7 gene using an Agrobacteriummediated technique (Table 1). Atlantic
transgenic 1ines contain the Bt-cry1Ja7/
GUS (B-glucuronidase) gene fusion with
the cauliflower mosaic virus promoter,
CaMV35S, expressing toxin in ali plant
tissue. Two Spunta lines were transformed
with the Bt-cry1Ja7 gene control led by the
CaMV35S promoter (pBIML5-vector). Two
other Spunta lines were transformed with
the Bt-cry7 Ja 7 gene control led by the
Gelvin super promoter (pBIML 1-vector)
(Mohammed et al., 2000). One Spunta
line was transformed with the Bt-cry1/a1/
355/PVY gene fusion controlled by the
CaMV35S promoter (pBIML6a-vector) (Li
et al., 1999).

Leaf bioassay
A detached leaf bioassay, developed by
Westedt et al. (1998), was used to test the
mortality of PTM feeding on Bt-cry7 Ja 7
transgenic plants. The petiole of the leaf
was i nserted i nto a sponge fitted in the top
of a glass vial full of water. The leaf and
the vial were then placed on a filter paper
disk in a Petri dish (25 x 150 mm). Ten
neonate larvae were introduced into each
Petri dish. Each Petri dish was considered
as a replication and five repl ications per
transgenic line were used in a completely
randomized design.

Table 1. Transgenic patato lines evaluated far control of patato tuber moth bioassays and their Bt-cry1/a1
gene constructs.
Vector
Construct
Line
pBlcry5
35S/Bt-cry1Ja1 /GUS
All Atlantic-Bt lines
pBIML5
35S/Bt-cry1Ja1
Spunta G-2
pBIML5
35S/Bt-cry1la1
Spunta G-3
pBIML 1
GSP/Bt-cry1/a1
Spunta S-1
GSP/Bt-cry1/a 1
pBIML 1
Spunta S-4
35S/Bt-cry1Ja1 /35S/PW
pBIML6a
Spunta 6a-3

118

Research on Patato

Survival of the larvae was recorded. For


P. operculella, evaluations were made at
eight days after infestation. For Symmetrischema tangolias, one evaluation was
made 5 days after infestation. Larval
development on the transgenic lines was
assessed by measuring the average length
of larvae growing on the transgenic lines,
compared with the length of larvae
growing on the control plants.

and humidity (25 2( and 70 5%


relative humidity).

Statistical analysis
All mortality data were subjected to
arcsine transformations befare analysis.
Data from the two varieties were analyzed
separately. Means were compared using
the Waller-Duncan test at 0.05. SAS/STAT
software for Windows (SAS lnstitute 1989)
was used for statistical analysis.

Tuber bioassay
Antibiosis effect was tested in the laboratory by putting tubers and PTM larvae in
closed containers (no-choice test). Harvested tubers from Atlantic and Spunta
lines were divided into three groups of 50
g each (representing three replications) for
P. operculella tuber assays. Each replication (50 g of tuberlets) was exposed to 1O
neonate larvae.
For the Symmetrischema tangolias tuber
test, 30 g for Atlantic and 25 g for Spunta
were used. Each replication was exposed
to six neonate larvae for Atlantic and five
neonate larvae for Spunta. One PTM
neonate larva was used for each 5 g of
tuber weight. After 25 days, the number of
pupae developed was recorded and overall
mortality was expressed in percentage.
Tests were carried out in an insect massrearing room under controlled temperature

Results and Discussion


Leaf bioassay
Mortality of P. operculella on Atlantic
transgenic lines after eight days was
significantly higher on two lines (Bt-2 and
Bt-4) than on the control (Table 2). All
Spunta transgenic 1 ines showed high leve Is
of resistance to P. operculella. lnsect
mortal ities ranged from 80 to 98% after
eight days, and al 1 were significantly
different from mortal ity on the
nontransgenic control. The reduction in
larval length ranged from 61 to 66.3% on
Atlantic and from 46.6 to 89.5% on Spunta
lines (Table 2).
Larval mortal ity of Symmetrischema
tangolias showed large differences across
Atlantic transgenic lines. Mortalities on
two lines (Bt-2 and Bt-5) were significantly
higher than that on the control (Table 3).

Table 2. Mortality and larval length reduction in patato tuber math, Phthorimaea operculella, in leaf and
tuber biaassay of Bt-cry1/a1 Atlantic and Spunta patato lines, La Malina, Peru, 2000.
Transgenic lines
Leaf assay1
Tuber assay1
Mortality (%)
Reduction (%)
Mortality (%)
Atlantic Bt-6
18 ab
65.9 a
27 ab.
Atlantic Bt-5
22 ab
65.7 a
40 a
Atlantic Bt-2
30 a
61.0 a
20 b
Atlantic Bt-4
34 a
66.3 a
30 ab
Atlantic (control)
5b
Oc
Spunta G-2
80 b
46.9 b
100 a
Spunta 6a-3
86 ab
63.3 b
100 a
Spunta G-3
90 ab
49.8 b
100 a
Spunta S-1
92 ab
46.6 b
100 a
89.5 a
Spunta S-4
98 a
100 a
Spunta (control)

ob

Oe

1 Means within a column followed by the same letter are not significantly different (P

= 0.05) by the

Waller-Duncan test.
GIP Program Report 1999 - 2000

119

Table 3. Mortality and larval length reduction in patato tuber moth Symmetrischema tango/ias in leaf and
tuber bioassay of Bt-cry1/a1 Atlantic and Spunta patato lines, La Malina, Peru, 2000.
Transgenic lines
Leaf assay1
Tuber assay1
Mortality (%)
Reduction (%)
Mortality (%)
Atlantic Bt-6
40 b
76.0 a
72 a
Atlantic Bt-4
56 b
71.4 a
78 a
Atlantic Bt-5
92 a
52.3 b
53 a
Atlantic Bt-2
94 a
85.7 a
50 a
Atlantic (control)
46.67 b
44 a
Spunta G-2
80 b
77.8 a
100 a
Spunta 6a-3
82 ab
77.4 a
100 a
Spunta S-1
90 ab
100 a
Spunta G-3
96 a
83.7 a
100 a
Spunta S-4
96 ab
87.9 a
100 a
Spunta (control)
1Oe
16 b
1 Means within a column followed by the same letter are not significantly different (P = 0.05) by the
Waller-Duncan test.
- = Data not available because surviving larvae escaped befare they could be measured.
Larval development was severely affected
on both transformed cu ltivars. The reduction in larval length ranged from 52.3 to
85.7% on AtlanHc and from 77.4 to 87.9%
on Spunta lines.
Of the PTM grown on Atlantic transgenic
lines, Symmetrischema tangolias larvae
were general ly more affected than were
P. operculella larvae.

Li et al. (1999) noted that the Spunta G-3


transgenic line contains significantly more
cry1 la1 protein than does Spunta G-2. That
may explain why the larval mortality on
G-3 is higher than on G-2 for both species.

Tuber bioassay
Mortal ity of P. operculella on Atlantic
transgenic lines was significantly different
than on the nontransformed control (Table
2). None of these lines gave 100% mortality in leaf and tuber bioassays. lnsect
mortalities ranged from 20 to 40%. This
level of larval mortality was lower than
the levels reported by Mohammed et al.
(2000) using sorne transgenic Atlantic lines
(Bt-2, Bt-4, and Bt-6) with stored and
newly harvested tubers.
All Spunta transgenic lines showed high
levels of r~sistance (100% larval mortality). These 'results are similar to the leaf
120

Research on Potato

bioassay and demonstrated that expression


in the tuber is correlated with that in the
leaf (r=0.99). This level of mortality was
the same as the levels reported by
Mohammed et al. (2000) using the same
lines in P. operculella tuber bioassay.
Mortality of S. tangolias on Atlantic
showed no differences in mortality on the
nontransgenic control. In the case of
Spunta, all transgenic lines showed high
levels of resistance, 100% of larval
mortality compared with 16% mortality in
the untransformed control (Table 3). Larval
mortality in leaf and tuber was highly
correlated (r=0.98).

Condusions
Mortal ity of both P. operculella and
Symmetrischema tangolias was found to
be high on leaves of transgenic Spunta
lines. In general, the larvae that fed on
Atlantic and Spunta transgenic plants
containing the Bt-cry1la1 gene were
-severely restricted in growth (reduction in
larval length in the leaf bioassay).
The res u lts of both PTM bioassays demonstrated that h igh leve Is of Bt-cry1Ja1
expression can be achieved with the genes
construct and vectors used in Spunta
transgenic 1 ines.

Th is Bt-cry7 la 7 gene therefore appears to


offer another source of resistance, which
can be pyram ided for effectiveness toward
the development of durable resistance to
PTM and other insect pests.

References
Barton, K. and M. Miller. 1993. Production
of Bacillus thuringiensis insecticida!
proteins, In: Kung, S. and R. Wu (eds.).
Transgenic plants. Vol.1, Engineering
and utilization. Academic Press, San
Diego, CA, USA. p. 297-315.
Bauer, L.S. 1995. Resistance: A threat to
the Insecticida! crystal proteins of
Bacillus thuringiensis. Florida Entomologist 78:414-443.
Caedo, V., J. Benavides, A. Golmirzaie,
F. Cisneros, M. Ghislain, and
A. Lagnaoui. 1999. Assessing
Bt-transformed potatoes for patato tuber
moth, Phthorimaea operculella (Zeller),
management. In: lmpact on a changing
world. Program Report 1997-98. lnternational Patato Center, Lima, Peru.
p. 161-170.
Crickmore, N., D.R. Zeigler, J. Feitelson,
E. Schnepf, J. Van Rie, D. Lereclus,
J. Baum, and D.H. Dean. 1998. Revision of the nomenclature for Bacillus
thuringiensis cry genes. Microbial and
Molecular Biology Review 62(3):807813.
Douches D., A. Westedt, K. Zarka,
B. Schroeter, and E. Grafius. 1998.
Patato transformation to combine
natural and engineered resistance for
controlling tuber moth. HortScience
33(6):1053-1056.

Ewell P., H. Fano, K.V. Raman, J., Alczar,


M. Palacios, and J. Carhuamaca. 1994.
Farmer management of patato insect
pests in Peru: Report of an interdisciplinary res~arch project in selected regions
of the h1ghlands and coast. lnternational
Patato Center, Lima, Peru. 77 p.
Jansens S., M. Cornel issen, R. De Clerco,
A. Reynaerts, and M. Peferoen. 1995.
Phthorimaea operculella (Lepidoptera:
Gelech~idae) resistance in patato by
express1on of the Bacillus thuringiensis
Cry1 A(b) insecticida! crystal protein.
Journal of Economic Entomology
88(5):1469-1476.
Li, W., K.A. Zarka, D. Douches,
J. Coombs, W. Pett, and J. Grafius.
1999. Coexpression of potato PVY coat
protein and cry-Bt genes in patato.
Journal of the American Society of
Horticultura! Science 124(3):218-223.
Mohammed A., D. Douches, W. Pett,
E. Grafius, J. Coombs, L. Li, and
M. Madkour. 2000. Evaluation of
patato tuber moth (Lepidoptera:
Gelechiidae) resistance in tubers of
Bt-cry5 transgenic patato lines. Journal
of Economic Entomology 93(2):472476.
SAS ln~titute. \989. SAS/STAT user's guide,
vers1on 6, 4t ed., Vol. 2. SAS lnstitute 1
Ca~ NC.
Wedstedt, A.L., D.S. Douches, W. Pett,
and E.J. Grafius. 1998. Evaluation of
natural and engineered resistance
mechanisms in Solanum tuberosum for
resistance to Phthorimaea opercu/ella
(Lepidoptera: Gelechiidae). Journal of
Economic Entomology 91 (2):552-556.

CIP Program Report 1999 - 2000

121

Assessment of the lnactivation Time of Phthorimaea


operculella Granulovirus (PoGV) at Different
lntensities of Natural lrradiation
M. Sporleder 1' 2, O.Zegarra1, J. KroscheP, J. Huber3, and A Lagnaoui1

A reliable bioassay method to assess inactivation time of Phthorimaea


operculel/a Zeller granulovirus (PoGV) was developed. The method was
tested four times at different intensities of natural solar irradiation in Lima,
Peru. Dry deposits of PoGV were exposed to the sun for specific time
intervals and bioassayed. lnactivation (half-life) was differentiated between
time of exposure and total energy of exposure. At a total radiation of 1100
Watt/m 2, half-life at the beginning of the test was 2.6 min and only 24 min
after 20 min of irradiation. At 820 W /m 2 , half-life was 4.51 min (17.5 min after
40 min of exposure), 4.53 min (24 min after 40 min of exposure) at 790 W/m 2,
and 6.56 min (18 min after 35 min of exposure) at 700 W/m2 The calculated
energy causing 50% reduction of virus activity increased from 174,000 joule
(J) at 1100 W /m 2 to 284,000 J at 700 W /m 2

The potato tuber moth, (Phthorimaea


operculel/a) is a serious pest of potatoes in
many tropical and subtropical regions of
the world. A granulovirus of P. operculella
(PoGV) with good prospects for control of
the pest in field and storage was isolated
in various parts of the world (Reed and
Springett, 1971; Matthiessen et al., 1978;
Kroschel et al., 1996). However, its use in
the field is limited by its rapid inactivation
due to UV radiation. The ability to
determine the influence of light on
preparations of microbial pesticides is
critica! to the development of formulations
designed to provide solar stability. In
addition, knowledge of the half-life of
microbial formulations is essential to the
proper timing of their application and their
use in the field.
1
2

CIP, Lima, Peru.


UniversityofHohenheim, Stuttgart, Germany.
Federal Biological Research Center for Agriculture and
Forestry, Darmstadt, Germany.

Extensive literature details the effect of


artificial light sources on viability of insect
pathogens (Griego et al., 1985; David,
1969). lgnoffo et al. (1997) have done
extensive work on the effect of a solar
simulator on viruses. They report that UV
in the simulator is very similar to natural
sunlight and that one hour of simulator
light equals four hours of natural sunlight.
But they do not mention the spectrum of
artificial light used. Shapiro (1985) also
concentrated on the UV spectra produced
by simulators. lt may be important to
consider a higher wavelength in future
research as well. McGuire, et al. (2000)
determined the solar stability of microbial
organisms using a solar simulator, which
provided nearly the same light spectrum as
sunlight. All authors, however, expressed
the degree of inactivation as the
percentage of original activity remaining.
Because of the sigmoid nature of the host
mortal ity-vi rus concentration response

CIP Program Report 1999 - 2000

123

curve, such units ca nn ot be directly


translated into degree of virus in activatio n.
Th e objective of this work was to estab li sh
a rapid and reliabl e bioassay method to
ca lcul ate the inactivation time (t 112) of
PoGV tested under different intens iti es of
so lar irradiation.

Materials and Methods


PoGV or igin ally isolated in La Molina,
Peru, was propagated in P operculella
larvae and purified by ultrace ntrifugation
on a sucrose gradi ent. A stock viru s
suspension with a titer of 2.6 , 109
granul es/m i, quantified by counts with the
Neubaur chamber (0.02 mm depth), was
used in this study. A viru s suspe nsion of
0.1 mi , co ntainin g 2.6 x 108 gra nul es, was
plated out and dried in petri dishes and
exposed to the sun. The petri dishes were
placed on a special board, which all owed
a conti nu ous adjustment to the su n (Fi gure
1). Four se ri es of exposu res were
condu cted in Lima: in Janu ary, March, and
December 2000, and January 200 1, wi th a
mean natu ral radi atio n of 820, 790, 111 O,
and 700 W/m 2, respecti ve ly. In an interval
of severa l minutes, dependin g on the
radiation intensity (3- 8 min), one dish was
removed, covered, and sto red at 4C until
its use in bioassays. The virus was
resuspended from the petri dishes in 1 mi

distilled wate r and the biologica l activity


determined using the egg-dip bioassay
method. Th e suspension obtained from a
non-irrad iated petri dish in eac h
exper iment served to determine the
reference probit lin e (y = a+ bx) (Fi gure 2)
accord ing to Finn ey (1971 ) usin g repeated
dilution technique. The activ ity ratio of all
treatments (tim e intervals) was ca lcul ated
usin g eq uat ion 1:

where R is the activ ity ratio after x


minutes ~f irrad iation, a and b are the
co nsta nts of the probit lin e obtained from
the non-irradiated petri dish, e is th e
ori gi nal ap pli ed co ncentration of PoGV in
gra nul es/ mi (here 2.6 x 108 ), and Yx is th e
probit obta in ed after x minutes of irradi ati on. Equation 1 extrapo lates survivin g
vira l granul es from the previously constructed log conce ntrati on-prob it mortal ity
curve of the actual conce ntrati on, to a
particular observed mortal ity leve l. Th at
all ows direct translation of th e obtained
acti vity ratio into degrees of viru s in activati on (Fi gure 2). Assumin g expo nenti al
half-li fe, or half- in activation time of the
viru s t 1/ 2 was ca lcul ated using eq uatio n 2:
I

t,12 = log o.5 / b


where b is the slope of th e regress ion
obtained after plotting the log activi ty (log
R) aga in st th e time of ex pos ure.

Figure 1. Board far exposing petri dishes with dried


virus deposits to the sun. A pyranometer and a UV-B
sensor (connected to the local logging station) were
placed on the table.

124

Research on Patato

During the virus exposure, total rad iation


was .measured using a pyranometer (LiCor2005). A UV-B senso r (Wa lz, Germany)
w ith a spectra l sens iti vity of from 265 to
315 nm (max = 297 nm) was used to
measure the most damaging part of natural
radiation. Since the fundamental unit of
li ght (watt) is defined as energy in joul e
(J)/sec, total J/ m 2 was ca lculated si mply by
acc umul at ing th e W/ m 2 (pyranometer,

7.-----------.-,...- .
,_..'

Y=0.677x+07614

~.-

_,/;~::./

:o
e 5
a.

....

~~

calculated applying the 95% confidence


limits of slopes obtained by linear regression analysis in equation 2 (Table 1).

0.6
. O.O
-0.6

:~ : t
-2.4

gi

-2.9

...J

~.5

-4.1
...,.,,,-'
...
3 '---.,.-----.------.--...---_J -4.7

5
6
7
8
Log concentration (granules/mi)

Figure 2. Ninety-five percent confidence limits far


control-adjusted mortality as a function of
concentration logarithm, with the logarithmic activity
(second Y axis) synchronized to the probit regression
line. (The confidence limits were obtained from nonirradiated petri dishes. lf 99% of the virus has been
inactivated (lag activity = -2), insects still respond
with approximately 50% mortality rate.)
400-1100 nm) multiplied by the number
of seconds of a measurement interval
(1 min). Log activity was plotted against
the accumulated amount of energy and
equation 2 was used to calculate the
energy sum that causes half virus
i nactivation.

Results and Discussion


In all four series the survival curve of
PoGV fol lows the same bisegmented shape
as described for the granulovirus of the
false codling moth, Cryptophlebia
leucotreta, (C/GV) by Huber and Ldcke
(1996) and Spodoptera littoralis nuclear
polyhedrovirus by Jones et al. (1993)
(Figure 3). Confidence limits for the
i nactivation ti me in the fi rst segments are

Fastest inactivation was found during


exposure in December 2000 (Figure 3, 3a)
where radiation was constant at 1100 W/m 2
During the first 20 min of irradiation, the
half-life of the virus was 2.6 min, after
which about 99% of the virus had been
inactivated. Thereafter half-life time
increased to approximately 24 min.
Slowest inactivation occurred during
exposure in January 2001 (Figure 3, 4a)
with a half-life of 6.56 min during the first
35 min of exposure and approximately 18
min thereafter. During virus exposure, the
radiation increased slowly from 690 to 760
W/m 2 After approximately 40 min of
exposure, intermediate half-life was 4.54
and 17.5 min in the first segment, January
2000 (Figure 3, 1a); and 4.51 and 24 min
in the second segment, March (Figure 3,
2a). In the latter exposure series, radiation
intensity was constant the first 40 min.
Thereafter, radiation increased slowly from
820 to 880 W/m 2 in January and from 790
to 900 W/m 2 in January and March.
Plotting the log activity ratio against the
irradiated energy sum (400-1100 nm)
demonstrates that a higher energy sum is
needed to cause half-activity loss as
radiation intensity decreases. At highest
radiation intensity (1100 W/m 2 , Figure 3,
3b), a sum of 174,000 J reduced virus
activity by 50%, while at lowest intensity
(700 W/m 2 , Figure 3, 4b) the value was
284,000 J. At intermediate intensities of
820 and 790 W/m 2 (Figure 3, 1b and 3, 2b)

Table 1. Linear regression analysis of irradiation time and log activity ratio of PoGV and its calculated halfinactivation time (t1h = log 0.5/s/ope).
Exp. ~adiat~on
Slope
t1h (95%
R2
lntercept
F (df)
Probability
No. mtens1ty
( St. err.)
confidence
( St. err.)
(W/m 2)
limits) (min)
1
820
-0.066 (0.021) . 4.54 (2.82 - 11.6) 0.675
0.62 (0.59) 10.38 (1, 5)
0.0234
2
790
-0.067 (0.010) 4.51 (3.52 - 6.26) 0.754
0.98 (0.29) 48.92 (1, 16) 0.0000
3
111 o
-0.115 (0.032) 2.62 (1.70 - 5.69) 0.653
0.29 (0.45) 13.16 (1, 7)
0.0084
-0.51 (0.29) 12.10 (1, 8)
4
700
-0.046 (0.013) 6.56 (4.19 - 15.03) 0.602
0.0083
y - log activity ratio; x - exposure time in minutes.

CIP Program Report 1999 - 2000

125

1b

1a
Jan. 2000: -820 W/m 2

~
:;;

~
Cl
3

-1
-2

-3
-4

75
50
time(min)

25

100

125

2a

2000
4000
energy (x 1000 joule)

6000

2b
Mar. 2000: -790 W/m 2

">

g>

...J

..,, -

-1

~
~

-2
-3
-4

50

25

75
time(min)

-3

-4'-----.......------------__.
100

125

2000
4000
energy (x 1000 joule)

6000

3b

3a

2
2 . - - - - - - - - Dec. 2000: -111 OW/m

">
n
as

g>

...J

-1
-2
-3

.4.....__ _ _.....-_ _ _"""'T"""_ _ _

-.d.

25

75
50
time(min)

100

125

2000

W/m 2

6000

2.------------------,

--.-~

energy (x 1000 joule)

4b

4a

2 . - - - - - - - - - J a n . 2001: -700

4000

1st segment
2nd segment

as
e
ofl-1~
Cl

-2

-3

-3
4'----..,---..,----r------r---~

25

50

75

time (min)

100

125

-4L-----------..----....---'

2000

4000
energy (x 1000 joule)

6000

Figure 3. Bisegmented inactivation curves of PoGV during four series of exposure conducted under natural solar
radiation in Lima, Peru. Log activity is plotted against exposure time (1-4a) and the accumulated energy sum
calculated from measurements with a UCor 200S pyranometer,400-1100 nm (1-4b).

126

Research on Potato

Table 2. Linear regression analysis of radiation energy (400-1100 nm) and lag activity ratio of fbGV and
its calculated half-inactivation to energy (energyVz = lag 0.5/slope).
Radiation
Slope
t11z (95%
Exp.
lntercept
R2
intensity
( St. err.) confidence limits)
F (df)
Probability
No.
(

St. err.)
2
3
3
(W/m )
(x 10 )
(x 10 joule)
1
820
-1.34 (0.42)
225 (139 - 581)
0.672
0.61 (0.59) 10.23 (1, 5)
0.0240
2
790
-1.3 (0.18)
231 (180 - 320)
0.757
0.93 (0.28) 49.70 (1, 16)
0.0000
3
1110
-1.73 (0.48)
174 (113 - 379)
0.653
0.29 (0.45) 13.18 (1, 7)
0.0084
4
700
-1.06 (0.31)
284 (181 - 656)
0.599 -0.52 (0.29) 11.96 (1, 8)
0.0086
y = lag activity ratio; x = accumulated energy in joule.

half-inactivation of the virus was caused


by a sum of 225,000 and 231,000 J,
respectively (see Table 2 for linear regression summaries). In the second segment of
lower inactivation, the same figures are
calculated to be 904,000, 1,500,000,
1,516,000, and 812,000 J at intensities of
820, 790, 1100, and 700 W/m 2 , respectively. However, due to variability in
mortality responses, determination of halfinactivation is difficult when the activity
of the pathogen is very low. lt shou Id be
considered that if 99.9% of the virus has
been inactivated (log activity = -3),
insects theoretically respond with about
23% virus-induced mortality (Figure 2).
Because of the sigmoid nature of the
mortality curve, variability of the response
becomes very deflective on the log
activity axis at extremely high and low
activities.
Using the observed log activities of both
inactivation segments, a two-order polynomial model could be successfully fitted to
describe the relation between log activity
and the accumulated irradiation energy
(Table 3). lts error, however, increases after
reaching the activity minimum.

Table 3. Polynomial
Exp. Radiation
No. intensity
(W/m 2)
1
820
2
790
3
1110
4
700

lt was not possible to explain the inactivation solely by the radiation energy sum.
However, a LiCor 2005 pyranometer only
approximates sunlight energy because of
its relative response along the energy
distribution in the solar spectrum. In
natural sunlight, 3.5-4% of the energy is
in the UV spectrum (300-400 nm). In all
experiments, the energy measured with
the UV-B sensor was only 0.2% of the
energy compared with the pyranometer.
Further experiments at different altitudes
should enable a comparison of total
energy and 1ight spectra.

Conclusions
The method employed is adequate to
calculate the half-inactivation time of
baculoviruses under natural radiation. lt
seems important to differentiate between
irradiation time and irradiation energy sum
during exposure. In recent investigations,
we started to test the persistence of PoGV
activity at various altitudes of the Peruvian
Andes. Additionally, the method is used to
evaluate the protective capacity of
formu lations.

regression analysis of radiation energy (400-1100 nm) and lag activity ratio of fbGV.
Constants of the polynomial model
F (df)
Probability
y = m1 x2 + m2x + b ( St. err.)
1.24
2.1
4.53
4.51

(0.45)
(0.51)
(1.70)
(2.11)

-1.41
-1.76
-2.32
-1.78

(0.30)
(0.27)
(0.58)
(0.55)

0.54
1.07
0.39
-0.35

(0.39)
(0.29)
(0.41)
(0.29)

0.862
0.806
0.732
0.681

37.38
53.86
17.73
11.72

(1,
(1,
(1,
(1,

12)
26)
13)
11)

0.0001
0.0000
0.0010
0.0057

y = lag activity ratio; x = accumulated energy in joule.

CIP Program Report 1999 - 2000

127

Acknowledgment
This work is a collaborative project
between the University of Hohenheim,
Germany; Federal Biological Research
Center, Germany; and CIP, financed by the
Federal Ministry of Co-operation and
Development, Germany.

References
David, W.A.L. 1969. The effect of
ultraviolet radiation of known
wavelength on a granulosis virus of
Pieris brassicae. Journal of Economic
Entomology 14:336-342.
Finney, J.R. 1971. Probit analysis (Third
edition). Cambridge University Press,
UK. 333 p.
Griego, V.M., M.E. Martignoni, and
A.E. Claycomb. 1985. lnactivation of
nuclear polyhedrosis virus (baculovirus
subgroup A) by monochromatic UV
radiation. Applied and Environmental
Microbiology 49(3):709-71 O.
Huber, J. and C. Ldcke. 1996. UVinactivation of baculovirus: The
bisegmented survival curve. IOBC/
WPRS Bulletin 19(9):253-256.
lgnoffo, C.M., C. Garca, and
S.G. Saathoff. 1997. Sunlight stability
and rain-fastness of formulations of
Baculovirus heliothis. Environmental
Entomology 26:1470-1474.
Jones, K.A., A. Westby, P.J.A. Re 1ly, and
M.J. Jeger. 1993. The exploitation of
micro-organisms in the developing
countries of the tropics: In Jones,

128

Research on Patato

D.G. (ed.). Exploitation of microorganisms. Chapman and Hall, London,


UK. p. 343-370
Kroschel, J., H.J. Kaack, E. Fritsch, and
J. Huber. 1996. Biological control of the
potato tuber moth (Phthorimaea
operculella Zeller) in the Republic of
Yemen using granulosis virus:
Propagation and effectiveness of the
virus in field trials. Biocontrol Science
and Technology 6:217-226.
Matthiessen, J.N., R.L. Christian,
T.D.C. Grace, and B.K. Filshie. 1978.
Large-scale field propagation and
purification of the granulosis virus of the
potato moth, Phthorimaea operculella
(Zeller) (Lepidoptera: Gelechiidae).
Bulletin of Entomological Research

68:385-391.
McGuire, M.R., R.W. Behle, H.N. Holly,
and T.C. Fry. 2000. Calibration of a
sunlight simulator for determining solar
stability of Bacillus thuringiensis and
Anagrapha fa/cifera nuclear
polyhedrovirus. Environmental
Entomology 29:1070-1074.
Reed, E.M. and B.P. Springett. 1971.
Large-scale field testing of a granulosis
vi rus for the control of the potato moth
(Phthorimaea operculella (Zell.) (Lep.
Gelechiidae)). Bulletin of
Entomological Research 61 :223-233.
Shapiro, M. 1985. Effectiveness of B
vitamins as UV screens for the gypsy
moth (Lepidoptera: Lymantri idae)
nucleopolyhedrosis virus. Environmental
Entomology 14:705-708.

lntegrated Control of Potato Bacterial Wilt in Kabale


District, Southwestern Uganda
B. Lemaga 1

In Kabale District, southwest Uganda, four sets of options for the integrated
control of bacterial wilt (BW) were tested in a participatory approach with farmers
between 1995 and 1999, along with experiments on crop rotation and soil
amendments. Options tested for integrated control were (1) an improved package
(IP) that consisted of clean seed, a less-susceptible variety (Victoria), and improved
cultural practices, and (2) a farmer package (FP), which consisted of a farmers'
variety and farmers' seed, planted under the farmer's cultural practices. Two
additional packages were later added: clean seed of Victoria planted under farmers' cultural practices, and farmers' variety and seed planted under improved
cultural practices. In the rotation trial, one-season rotation experiments were
conducted on mildly infested fields and two-season rotation experiments were
conducted on heavily infested fields. In the soil-amendment experiment, organic
materials from Sesbania sesban and Leucaena diversifo/ia were applied in amounts
sufficient to supply 100 kg N/ha either singly or combined with P and PK. Also
added were NP and NPK from inorganic sources at rates that would supply 100
kg/ha for each element. All three options for the integrated control of BW significantly reduced wilt and increased yield; With IP performing best. The options
were economically beneficial, with marginal rates of return of 1034% for IP,
805 % for clean seed of a less susceptible variety under farmers' practices, and
634% for improved cultural practices with farmers' variety and seed. A one-season
rotation in mildly infested fields significantly reduced wilt to 3.2-12.4 % , compared to 63 % in the control, and resulted in corresponding significant in creases in
total and marketable yields. Similarly, the two-season rotation in heavily infested
plots also resulted in significant reductions in wilt incidence and increases in yield,
with increases in marketable yields ranging from 216-400%. Planting two different crops in two consecutive seasons resulted in superior wilt control, compared
to planting the same crop consecutively. Applying organic materials did not
necessarily reduce BW incidence, but there was a noticeable effect when organic
materials were applied together with inorganic fertilizers. The best results in
relation to both disease control and increased yield were obtained when sesbania
was applied together with P and K.
The potato (Solanum tuberosum) is an
important food and cash crop in the
highlands of Uganda. lt is mainly grown in
1

the southwest districts of Kabale, Kisoro,


and Mbarara, in the eastern districts of
Mbale and Kapchorwa, and in the western
districts of Kabarole and Kasese. Kabale is

PRAPACE, P.O. Box 22274, Kampala, Uganda.

CIP Program Report 1999 - 2000

129

the leading producer of the crop, accounting for over 42% of the national annual
production (Bariyanga, 1997), and in this
district potato ranks first as a cash crop
and fourth as a food crop (Lemaga et al.,
1999).
Although potato yields are as high as from
20 to 40 t/ha at the research station, the
national mean yield is about 7 t/ha (FAO,
2000), which is low compared with
production statistics in many countries.
The low yield is attributed to a number of
biotic and edaphic factors in addition to
weather and poor management. Bacteria!
wilt (BW), caused by Ralstonia
solanacearum, is the second most important yield-reducing biotic factor in Kabale
after late blight (Low, 1997; Lemaga et al.,
1999). Since BW cannot be controlled by
chemicals and persists in the soil for a
long time, it is increasingly becoming a
major threat to potato production in the
district, where it causes a yield loss of
about 26% (Lemaga et al., 1999) with
occasional loses of 100% if the disease
appears during the early stages of crop
growth (Opio, 1988; Kakuhenzire et al.,
1993). In Kabale, BW is caused by race 3,
biovar 2-A of R. solanacearum (data to be
published at a later date).
Race 3 has a narrow host range and can be
successfully controlled by integrated
disease management (IDM). The most
important components of IDM include
clean seed (Berrios and Rubirigi, 1993;
Van der Zaag, 1986), crop rotation (Yerma
and Shekhawat, 1991; Gunadi et al., 1998;
Lemaga, 2001 b), less-susceptible varieties
(French et al., 1997; Tusiime et al.,
1996a), and disease-free soil (Priou et al.,
1999a). BW can be eradicated if efficient
control measures are properly used, as
reported in Australia (Lloyd, 1976) and
Peru (French, 1994). The successful control
of BW in Kabale may, however, be more
difficult than we anticipate because of
small and very fragmented landholdings
that limit crop rotation and sanitation, and
because of poor soil fertility (Lemaga et

130

Research on Potato

al., 2001 a). The presence of a variety of


weed species that are symptomless carriers
of the pathogen (Tusiime et al., 1996b) and
rampant soil infestation are other limiting
factors to BW management in Kabale.
For IDM to be successful, the right components have to be determined for each area
(French 1994; Priou et al., 1999b) and
factors of production that affect plant
vigor, such as soil fertility, need to be
properly addressed. Ramesh and
Bandyopadhyay (1993) asserted that
ferralsols (the predominant soil type in
Kabale) are conducive to the wi lt bacterium because of their low pH and organic
matter content. However, optimum soilamendment practices have been helpful in
reversing the situation, leading to reduced
BW and increased tuber yields (Lemaga et
al., 2001 a).
To control BW, several experiments were
conducted in Kabale between 1995 and
1999. These included testing various sets
of IDM as well as identifying IDM components. The objectives of the experiments
reported in this paper were (1) to provide
potato producers with 1DM options for the
control of bacteria! wilt that are economical ly feasible and pragmatic and (2) to
quantify the effects of soil amendments
and crop rotation on BW incidence and
tuber yields.
In 1995, we compared two IDM packages
(IP and FP) for the control of BW. The IP
(improved package) consisted of an
improved variety, Victoria (CIP381381.20), which is less susceptible to
BW; clean seed, planted at a spacing of
75 cm x 30 cm; hilling at planting;
roguing of volunteer potatoes; sanitation;
and mnimum post-emergence cultivation.
The FP consisted of farmers' variety,
farmers' seed, and farmers' cultural
practices. Compared with the FP, the IP
delayed wilt onset by 16 days, reduced
wilt incidence by over 50%, and increased
total yield by an average of 76%, with
increases ranging from 5% to 222%
(Lemaga et al., 1997). The outstanding

performance of the improved package (IP)


in 1995 was very encouraging for the
farmers, who responded by asking about
the availability of clean seed of the
improved variety. They questioned its high
cost and the contributions of improved
cultural practices. Based on this, we
agreed to evaluate four treatments, starting
in 1996. These were designed to provide
the farmers with more options to choose
from and included (1) the IP and (2) farmer
package (FP), plus (3) clean seed of the
improved variety, planted under farmers'
cultural practices (IVCSFC) and (4)
farmers' variety and seed planted under
improved cultural practices (FVFSIC).

Materials and Methods


Options for integrated control of
bacterial wilt
On-farm research on the integrated control
of potato BW was conducted in Kabale
District in southwestern Uganda for five
seasons using a participatory approach
with farmers. Experimental fields were
located on latitudes 1 19' S and 1 20' S,
longitude 30 00' E at altitudes ranging
from 1830 to 1988 m. The selection of
participating farmers and communities was
based on the presence of BW in the field
plots to be used and farmers' willingness to
provide land for the experiment, perform
recommended cultural practices for IDM,
and harvest a 11 the. crop at matu rity.
Farmers' research options were (1) the IP
(an improved variety, Victoria (CIP381381.20), which is less susceptible to
BW; clean seed, planted at a spacing of
75 cm x 30 cm; hilling at planting;
roguing of volunteer potatoes; sanitation;
and mnimum post-emergence cultivation)
and (2) FP (farmers' variety, farmers' seed,
and farmers' cultural practices), plus two
others added in 1996, (3) clean seed of the
improved variety, planted under farmers'
cultural practices (IVCSFC) and (4)
farmers' variety and seed planted under
improved cultural practices (FVFSIC).

The experiment was laid out in a randomized complete block design. Within a
season, each farmer's field (farm) was
considered as a replication; the number of
farmers per season averaged 14. For each
plot, the soil type and history of crop
rotation were recorded.

Disease assessment and data analysis


Plots were assessed at weekly intervals to
determine days to onset of first wilt
symptoms. Subsequent counts of wilted
plants were made at two-week intervals
from the onset of the disease. At each
assessment, all plants that showed either
complete or partial wilting were considered wilted. These were staked to avoid
double counting in subsequent assessments
and also to avoid the possibility of missing
those that died completely during the
growth period. The counts of symptomatic
plants were expressed as a percent of the
total number of plants that emerged. Late
blight was controlled with three to four
sprays of Dithane M45, starting from the
first visible symptoms. At harvest, data
were recorded on yield, including total
and marketable yields, as wel 1 as rotten
tubers. Farmers participated in collecting
disease data and in harvesting, followed
by discussions to evaluate the options.
There were large variations in weather
conditions, especially rainfall, therefore,
experimental data were analyzed separately. Data on BW incidence were
subjected to square root transformation
before analysis using Genstat 4.23 (registered trade mark of the Numerical
Algorithms Group Ltd., Oxford, UK).

Economic analysis
Since adoption of any technology by
farmers is dependent on the economic
benefit they receive, analyses of partial
budget and marginal rates of return were
done for each of the 1DM options. Twentytwo farmers who participated in o-farm
integrated BW control experiments also
participated in this study. The costs of

CIP Program Report 1999 - 2000

131

various inputs and cultural practices far


each option were obtained from the
farmers and averaged far economic
analysis. Costs that varied were attributed
to planting, hilling, weeding, fungicide
spraying, harvesting, transportation, and
seed. The cost of clean Victoria seed was
obtained from the National Patato Program. The income generated from each
option varied according to the proportions
sold as seed or as ware and also the
location of sale (market or farm gate), ali
of which were taken into consideration.
This study was conducted in 1997, therefore, we used the exchange rate of that
year, US$1 = Ug Sh 1050. Net benefits
were calculated by subtracting total costs
from gross income. Net benefits obtained
from the partial budget analysis give a
good indication of which technology offers
the best returns (Kindness, 1993). Dominance analysis was not done because
neither of the technology options was
dominated by the farmer package. Marginal rates of return were analyzed to
find out how much farmers gained in
return to their investment by changing
from their traditional practice to each of
the technology options.
Soil amendments

This study was conducted in the 1998A,


1998B, and 1999A seasons at Kachwekano
Agricu ltural Research and Development
Centre, Kabale District, located at an
elevation of 2000 m. The rainfall is
bimodal, with mean annual precipitation
of about 1000 mm and maximum and
minimum temperatures of 23C and 1OC,
respectively. The soi Is at Kachwekano are
classified as isomesic typic palehumult
(Yost and Eswaran, 1990). These soils are
generally deficient in nitrogen, with a
rather low pH.

The experiment was laid out in a randomized complete block design with three
replications. Plot size was 16.2 m2 , in
which disease-free Victoria patato seed
was planted and amended with either
organic materials from Sesbania sesban
(Sesbania) or Leucaena diversifolia
(leucaena) inorganic fertilizers (nitrogen
(N), phosphorus (P), potassium (K)), or
different combinations of these organic
and inorganic fertilizers (Table 1). To allow
time far decomposition, fresh sesbania and
leucaena leaves and twigs were incorporated into soil a week befare planting in
amounts needed to supply 100 kg N/ha.
lnorganic fertilizers N, P, and K were sidedressed at rates of 100 kg/ha each as urea,
triple super phosphate (TSP), and murate of
potash (KCI), respectively. Urea was split-
applied in equal halves at planting and
one month after planting, while TSP and
KCI were applied at planting. Late blight
was controlled with faur sprays of Dithane
M45 and one spray of Ridomil MZ, both at
recommended rates, beginning with the
appearance of the first disease symptoms.
The methods used to assess and col lect
data on BW incidence, yields, and statistical analyses were similar to those used in
the IDM experiment.
Crop rotation

Rotation experiments were conducted


between the 1995B and 1999A seasons at
the Kachwekano Agricultura! Research
and Development Centre. In the seasons
preceding the experiments, the fields were
planted to a BW-susceptible variety
Kabale (CIP-374080.5) in arder to increase
the inoculum amount unifarmly. Where
unifarmity was lacking, healthy plants
were artificially inoculated by a stempuncture method. Disease unifarmity was

Table 1. Composition of the organic amendments used.


Species
N (%)
P (%)
K (%)
Sesbaniasesban
3.4
0.15
1.1
Leucaena diversifolia
3.4
0.15
2.1
Source: Palm and Rowland, 1997; Mafongoya et al., 1997.

132

Research on Potato

Lignin (%)

8.1
4.3

Soluble polyphenols (%)

6.6
14.1

measured by the distribution of wilted


plants all over the field.
The one-season rotation trial was conducted in a field that exhibited an initial
BW incidence of 15-20% (mildly infested)
with uniform distribution over the field,
whereas the two-season trial was conducted in a field that exhibited an initial
BW incidence of over 90% (heavily
infested) with uniform distribution. Crops
that are commonly grown in the area were
included in the rotation treatments, except
for sorghum because of its long maturation
period. The treatments for both trials are
shown below.

One-season rotation
Potato-on ion-potato
Potato-peas-potato
Potato-cabbage-potato
Potato-sweetpotato-potato
Pota to-mil let-potato
Potato-carrots-potato
Potato-beans-potato
Potato-potato-potato

Two-season rotation
Potato-beans-wheat-potato (P-B-W-P)
Potato-beans-maize-potato (P-B-M-P)
Potato-wheat-maize-potato (P-W-M-P)
Potato-beans-beans-potato (P-B-B-P)
Potato-maize-maize-potato (P-M-M-P)
Potato-wheat-wheat-potato (P-W-W-P)
Potato-potato-potato-potato (P-P-P-P)
Plots were laid out in a randomized
complete block design with three replications for the one-season experiment and
four replications for the two-season
experiment. In the two-season experiment,
either two different rotational crops or the
same crop were grown in two consecutive
seasons. Clean seed of the potato variety
Victoria was used for both trials, and local
varieties were used for the rotation crops.
All the crops were planted at recommended spacing, which for potatoes was
75 x 30 cm. Late blight was controlled
with four sprays of Dithane M45 and one
spray of Ridomil MZ, both at recommended rates, starting from appearance of

the first symptom. Methods used to assess


and collect data on BW incidence, yields,
and statistical analysis were similar to the
other two experiments.

Results and Discussion


In Kabale district, BW is caused by
R. solanacearum, race 3 (Lemaga and
Priou, unpublished data), so all results
reported and conclusions drawn in this
paper hold true only for this race.

Options for integrated control of bacterial


wilt
Starting in the 1996A season, the additional two options (IVCSFC and FVFSIC)
were added to the study to give farmers
more options to choose from, as mentioned
above. Over the study period, the incidence of bacteria! wilt varied with the
seasons, resulting in changes in disease
levels and yields (Table 2) that can be
attributed to differences in weather
conditions, particularly rainfall and
populations of R. solanacearum in the soil.
Of the four treatments, the incidence of
wilt was lowest under IP and highest under
FP, with the differences being significant
(P < 0.05) most of the time. Wilt incidence
for the remaining two treatments were
intermediate, with IVCSFC being superior
to FVFSIC.
All three suggested options increased both
total and marketable yields (Table 2) as
compared to FP. The increases obtained
with the improved package were consistently highest for al 1 the seasons and
differed significantly (P < 0.05) from FP,
except for the 19978 season, with increases in marketable yield of 34% to
52%. The effects of IVCSFC and FVFSIC
on yield were similar to that seen with
BW-intermediate between the IP and FP,
with IVCSFC performing better than
FVFSIC. This implies that clean seed is a
very important component of the integrated control of BW and the use of a less
susceptible variety, Victoria, also played a
role. These results substantiate earlier
claims that clean seed contributes to
CIP Program Report 1999 - 2000

133

Table 2. Effects of the various integrated 8W control options on wilt incidence, total and marketable yields,
and percent increases in marketable yields during seasons 1996A, 19968, 1997A, and 19978, Kabale,
Uganda.
Total
Marketable
lncrease in
Options
Wilt
incidence {%}
marketable yield {%}
yield {t/ha}
yield {t/ha}

1996A
IP
FP
IVCSFC
FVFSIC
LSDo.os

(2.3)
(3.7)
(3.2)
(3.4)
(NS)

17.10
12.10
14.60
10.10
3.20

16.50
12.30
14.40
10.40
4.10

15.80
27.80
20.73
24.10

(3.58)
(4.87)
(3.68)
(4.39)
(0.88)

15.80
10.80
11.70
13.50
2.34

14.80
9.90
10.80
12.10
2.25

49.50

6.05
20.73
9.98
17.12

(1.82)
(3.80)
(2.49)
(3.49)
(0.70)

10.35
7.56
7.87
9.67
1.94

10.13
7.28
7.70
9.27
2.00

39.20

3.26
21.09
3.73
10.94

(1.55)
(3.79)
(1.61)
(2.88)
(1.70)

6.10
4.90
4.94
6.40
ns

5.78
3.80
4.70
5.40
ns

5.30
13.40
9.70
11.00

34.10
17.10
-15.40

19968
IP
FP
IVCSFC
FVFSIC
LSDo.os

9.20
22.00

1997A
IP
FP
IVCSFC
FVFSIC
LSDo.os

5.77
27.30

19978
IP
FP
IVCSFC
FVFSIC
LSDo 05

52.10
23.70
42.10

Notes: Figures in parentheses are square roots of wilt incidence. NS= not significant. IP = improved package,
FP = farmer package, IVCSFC = improved variety and clean seed planted under farmers' cultural practices, FVFSIP =
farmers' variety and seed planted under improved cultural practices.

successful control of BW in Burundi


(Berrios and Rubirigi, 1993), Rwanda (Van
der Zaag, 1986), and Uganda (Lemaga et
al., 1997), and clean seed of a lesssusceptible variety has even greater
potential to dramatically reduce wilt.
However, the current supply of clean seed
is far below the demand, and as a result,
most farmers use their own seed or seed of
unknown origin bought from markets
(Lemaga et al., 1999). The chances of
such seed carrying the infection are high
(Tusiime et al., 1996b), leading to increased disease spread.
Although heritable host resistance to BW
has not been confirmed in past research,
the use of less-susceptible varieties has
been shown to reduce the incidence of
wilt (Tusiime et al., 1996a; French et al.,

134

Research on Potato

1997), which suggests that the breeding


effort at CIP to develop varieties that are
either resistant or less susceptible to BW
should continue. Moreover, since late
blight and BW occur simultaneously in the
same field, a variety will be successful in
sub-Saharan Africa only if it is resistant to
both diseases. lt should also be noted that
the use of less-susceptible varieties under
BW conditions poses a potential danger of
latently spreading BW.

Economic analysis of 1DM options


Although FP had the lowest costs, the
other options had much higher net benefits, the highest of which was US$3347/
ha, obtained from the improved package
(Table 3). The second highest was the use
of clean seed of the less-susceptible

Table 3. Analysis of partial budget and marginal rates of return far traditional and improved technology
options.
Components
.lntegrated BW control options
IP
FP
IVCSFC
FVFSIC
Yield {t/ha)
13.92
9.37
11.34
11. 70
Amount sold as seed {t/ha)
9.48
4.22
7.34
4.50
Amount soldas ware {t/ha)
4.45
5.15
4.00
7.20
lncome from seed (US$/ha)
3385.71
1205.71
2621.43
1285.71
lncome from ware (US$/ha)
847.62
980.95
781.90
1371.43
Gross income (US$/ha)
4233.33
2186.67
3382.33
2657.14
Costs that vary (US$/ha)
884.46
69.43
836.13
768.03
Net benefit (US$/ha)
3346.87
1482.68
2547.21
1889.11
Marginal rates of return (%)
1034.00
805.60
634.50
Note: IP = improved package, FP = farmer package, IVCSFC = improved variety and clean seed planted under farmers'
cultural practices, FVFSIP = farmers' variety and seed planted under improved cultural practices.

variety, planted under farmers' cultural


practices (IVCSFC), which had a net profit
of US$2547/ha. The use of improved
cultural practices alone (FVFSIC), which
was the lowest performer of the three
suggested options, resulted in about
US$406 more than the farmers' package.
The main reason for the IVCSFC being
better than the FVFSIC was because of the
higher proportions of the produce sold as
seed.
The data on marginal rates of return (MRR)
(Table 3) show that the benefit of changing
from FP to IP was 1034%. This means that
for every dollar invested in IP, the farmer
will get an additional US$10.34. The
corresponding values for IVCSFC and
FVFSIC were 805.6% and 634.5%, respectively. These values are much higher than
the 100% MRR that is usually assumed to
be high enough to recommend a given
technology for adoption (Kindness, 1993).
Therefore, any one of these three technology options can be recommended to
farmers, with IP being the best.

Soil amendments
The incidence of bacteria! wilt varied with
the seasons and was highest in 1998A and
lowest in 1999A (Table 4), differences that
could be attributed to variations in rainfall
since wilt incidence is positively correlated to precipitation (Akiew, 1986). The
control plots were the worst affected,

resulting in the highest wilt incidence in


1998A and 1998B, that differed significantly in most cases. Well-nourished
plants are stronger and better able to
withstand diseases than poorly nourished
ones (Muchovej et al., 1980). Generally,
the appl ication of green materials alone
did not affect BW incidence.
Application of soil amendments increased
both total and marketable ware potato
yields when compared with the control
(Table 4). However, only the sesbania + PK
treatment resulted in significantly higher
marketable yields in all three seasons,
compared to the control. The better
nutrient balance achieved when organic
and inorganic nutrients were used and the
additional effect of the organic material to
improve physical soil conditions could be
responsible for this.
Marketable yields had a significant (P <
0.001) negative relationship (R2 = 0.56)
with BW incidence (Figure 1). A unit
increase in bacteria! wilt would apparently
result in a 2.5-unit reduction of yield,
confirming the high yield losses farmers in
Kabale suffer from BW.

Crop rotation
One-season rotation. In the mildly infested
field (15-20% infestation) all the oneseason rotations with pulses, cereals,
vegetables, and root crops, although not

CIP Program Report 1999 - 2000

135

Table 4. Effect of soil amendments on patato tuber yield and incidence of bacteria! wilt during the 1998A,
1998B, and 1999A seasons, Kabale, Uganda.
Marketable
% Relative increase
Total tuber
Wilt incidence
Treatment
in marketable yield
yield (t/ha)
(%)
t/ha)

1998A

L
S+PK
L+PK
S+P
L+P
NPK
NP

62.60
62.00
49.90
51.40
48.20
51.60
44.70
31.60
66.50

(7.9)
(7.9)
(7;0)
(7.1)
(6.9)
(7.2)
(6.7)
(5.6)
(8.2)
(0.9)

9.9
8.2
13.7
10.5
9.6
7.3
9.0.
11.2
4.9
5.1

8.0
6.1
10.9
7.6
7.6
4.7
7.7
9.9
3.3
4.8

142.4
84.5
230.3
130.3
130.3
42.4
133.4
200.0

14.44
21.21
4.47
19.80
17.46
22.44
16.71
12.17
24.48

(3.8)
(4.6)
(2.1)
(4.4)
(4.2)
(4.7)
(4.1)
(3.4)
(4.9)
(0.5)

25.9
23.1
30.9
26.4
23.9
22.3
27.8
19.9
11.5
3.8

24.2
20.4
29.6
23.2
20.8
19.3
25.2
18.3
9.3
4.8

160.2
119.4
218.3
149.5
123.6
107.5
171.0
96.8

6.30
7.80
2.90
7.40
12.20
7.35
7.03
7.47
5.30

(2.5)
(2.8)
(1.7)
(2.7)
(3.5)
(2.7)
(2.6)
(2.7)
(2.3)
NS

18.3
19.1
22.5
19.5
17.9
16.9
19.5
18.7
17.4
3.6

17.7
18.3
21.9
18.5
17.2
16.0
18.4
17.4
16.5
3.8

7.2
10.9
32.7
12.1
4.2
-3.0
11.2
5.4

LSD(aUi)

19988

L
S+PK
L+PK
S+P
S+P
NPK
NP

LSD(QUi)

1999A

L
S+PK
L+PK
S+P
L+P
NPK
NP

LSD(o.(

Note: Figures in parentheses are square roots of wilt incidence. NS= not significant at P < 0.05. S = Sesbania sesban,
L = Leucaena divesifolia, N = nitrogen, P = phosphorus, K = potassium, C = control. (Reprinted with permission of
the African Crop Science Journal from Lemaga et al. (2001 a).)

different from one another, significantly


(P < O.OS) decreased BW incidence and
significantly (P < 0.01) increased marketable potato yields when compared with
the control (Table 5). This corroborates the
findings of Yerma and Shekhawat (1991 ),
who reported that in a five-year rotation,
wilt incidence was reduced to 6.3%,
compared 80.1 % under a monocrop.
The lowest incidence of BW occurred after
rotations with finger millet (3.2%) and

136

Research on Potato

sweetpotato (3.8%); however, marketable


potato yields were also lowest after these
crops in the rotation (Table 5). In contrast,
the highest yields were obtained following
rotations with carrots and onions, which
allowed the highest incidence of BW. This
could be attributed to the high nutrient
demands of millet and sweetpotato.
Similarly, the high yields after shallowrooted crops, such as carrots and onions,
could be that these crops do not exploit
soil nutrients at the lower soil depths of

35

cu

.!::

25

"O

Q)

::;., 20
Q)

15
<ti

15

()
~

30

10

Y= -2.5281x + 26.97
R2 = 0.56
(P<0.001)

#t

f4~

10

Bacteria! wilt incedence (Sq root)

Figure 1. Relationship between wilt incidence and


tuber yield. (Reprinted with permission of the African
Crop Science Journal from Lemaga et al. (2001 a).)
1 5-3_0 cm, where the potato is most active.
Gunadi et al. (1998) also reported that
potato yields following rotations with
carrots were higher than rotations with
crops that have longer roots, such as
maize, and tomatoes. This suggests that
the nutrient exploitation of crops selected
for rotation with potatoes should be
considered.
The sweetpotato, which is the second most
important food crop in Kabale (Low, 1997;
Lemaga et al., 1999), can be very suitable
for rotation where BW is an important
disease provided soil fertility is well
managed. Kloos et al. (1991) and Bang
and Wiles (1995) reported that wilt
incidence was reduced following a

rotation of sweetpotato. At CIP, in Peru,


rotation using sweetpotato reduced both
wilt incidence and the population of the
bacterium in the soil (Priou, 2000) .
Two-season rotation. Days to onset of BW
were not affected by the different rotation
treatments. However, after onset of wi lt at
40 days after planting, disease progress
with the monocrop control was fastest, and
total incidence at 91 days after planting
was the highest (81 %), differing significantly (P < O.OS) when compared to the
incidences under rotations, which were all
below 50% (Figure 2).
Reduction in the incidence of bacteria!
wilt was greatest where two different crops
were planted in succession rather than the
same crop in two consecutive seasons in a
rotation (Table 6). Bang and Wiles (1995)
reported that in a two-season maizesweetpotato rotation, wilt incidence was
reduced by 32%, compared to 24% in a
sweetpotato-sweetpotato rotation, whereas
the monocrop check had a wilt incidence
of 88%. In our study, the progress of wilt
with time, as well as total wilt incidence,
were lowest with potato-beans-maizepotato (22%), followed by potato-wheatmaize-potato (25%). The performance of
rotations that included beans followed by
cereals is important because beans are the
main food crop and the major source of
plant protein in the district (Lemaga et al.,
1999). However, using beans alone for
rotation is not recommended, as beans are

Table 5. The effect of a one-season (1999A) rotation with various crops on incidence of bacteria! wilt, tuber
yields, and increase in marketable tuber yields, Kachwekano, Kabale, Uganda.
lncrease in
Treatment
Bacteria!
Total yield
Marketable
marketable yield (%)
wilt (%)
(Vha)
yield (Vha)
80
Potato-onions-potato
12.4 (3.5b)
20.4a
20.0a
67
Patato-peas-patato
5.0 (2.2b)
18.6a
18.5a
59
Potato-cabbage-potato
6.9 (2.6b)
17.7a
17.6a
48
Potato-sweetpotato-potato
3.8 (1.6)
16.4ab
16.4a
50
Potato-millet-potato
3.2 (1.7b)
16.?ab
16.?a
84
Potato-carrots-potato
11.0 (3.3b)
20.4a
20.4a
63
Potato-beans-potato
7.4 (2.2b)
18.1a
18.1a
Potato-potato-potato
62.2 (7.8a)
12.4b
11.1b
Note: Figures in parentheses are square roots of wilt incidence. Means followed by the same letter in columns are not
significantly different at 0.05 probability level. Adapted from Lemaga et al (2001b).

CIP Program Report 1999 - 2000

137

90
80
70

l
-~

cu

60

50
40

13

ctl
co 30

20

10

o
40

47

54

61

68

75

82

89

96

Days after planting

Figure 2. Effect of two-season rotation on wilt


development with time. (Reprinted with permission
of the African Crop Science Journal from Lemaga et
al. (2001 b).)
potential hosts of R. solanacearum
(Granada and Sequeira, 1983; Opio,
1988).
In a heavily infested field, however, none
of the rotation treatments was good
enough to either decrease the incidence of
BW or increase tuber yields to economically acceptable levels in two seasons. lt
is therefore important to consider rotations
longer than two seasons for successfu 1
potato production.

Conclusions
The studies reported here demonstrate that
bacteria! wilt can be satisfactorily con-

trolled and patato yields increased by one


of the following (with the first option being
by far the best): (1) combined use of clean
seed of a less-susceptible variety and
improved cultural practices, (2) clean seed
of a less-susceptible variety alone, or (3)
improved cultural practices alone. All
three suggested options for integrated
control of BW provided sound economic
benefits, compared to the farmer package.
Crop rotation is also an important component of IDM; however, while a single
rotation can be sufficient for the production of ware potatoes on mildly infested
fields, rotations must be longer than two
seasons if the soil is heavily infested.
Beans, the number one food crop and a
majar source of plant protein in Kabale,
can be good in rotation if used either
befare or after cereals, because they are
thought to be a symptomless carrier of BW.
Likewise, sweetpotatoes, the district's
second most important food crop, can be a
good rotation crop for potatoes where BW
is important, as long as soil amendments
are adequate. The effects of control
measures are better if the soil is well
fertilized, preferably with a combination
of organic (sesbania or leucaena) and
inorganic additions. However, the availability and economic advantages of the
organic materials have to be established.
In general, good control of BW can be
achieved in Kabale through the combined
use of clean seed, crop rotation, uncon-

Table 6. Effect of two-season rotation with cereals and beans on incidence of bacteria! wilt and patato yields
at Kachwekano, Kabale.
Wild incidence
Total yield
Marketable
Treatment
lncrease in
(t/ha)
yield (t/ha)
marketable yield (%)
Potato-beans-wheat-potato
36.4 (5.81 bcd)
9.27a
7.40a
300
Potato-beans-maize-potato
21.9 (4.52d)
10.37a
9.00a
386
Potato-wheat-maize-potato
25.2 (4.89cd)
10.?a
8.90a
381
Potato-beans-beans-potato
40.3 (6.32bc)
8.92a
6.68a
216
Potato-maize-maize-potato
49.4 (7.01b)
10.15a
8.40a
354
Potato-wheat-wheat-potato
40.0 (6.28bc)
10.73a
9.23a
400
Potato-potato-potato-potato
81.1 (8.96a)
3.20b
1.85b
Note: Figures in parentheses are square roots of wilt incidence. Means followed by the same letter in columns are not
significantly different at 0.05 probability level by Duncan's Multiple Range Test.
Adapted from Lemaga et al. (2001 b).

138

Research on Potato

taminated soil, a less-susceptible variety,


and roguing. Latent infection should be
monitored with NCM-ELISA to ensure that
seed lots are free from BW. This would
result in good wilt control, higher yields,
and reduced wi lt spread.

Acknowledgements
This study was conducted under the
African Highlands lnitiative, which also
provided the funds. The excellent collaboration of the National Agricultura!
Organisation of Uganda is gratefully
acknowledged. The valuable technical
support of Drs. S. Priou, E.R. French,
P. Ewell, M. Olayna, R. El-Bedewy, and
J.J Hakiza, both during the execution of
the experiments and in the preparation of
this manuscript, is highly appreciated.
The author is also indebted to Mr. Geofrey
Manzi for data collection and Mr.
R. Kakuhenzire and Patrick Okori for their
support in data analysis.

References
Akiew, E.B. 1986. lnfluence of soil
moisture and temperature on the
presence of Pseudomonas
solanacearum. In: Persley, G.J. (ed.).
Bacteria! wilt disease in Asia and the
South Pacific. ACIAR Proceedings No.
13. Australian Centre for lnternational
Agricultura! Research, Canberra,
Australia. p. 77-79.
Bang, S.K. and G.C. Wiles. 1995. Control
of bacteria! wilt (Pseudomonas
solanacearum) in potato by crop
rotation. In: Rasco, E.T. Jr. and
F.B. Aromin (eds.). SAPRAD on the third
year of phase 111. Selected research
papers. Vol. 1. South Asian Program for
Potato Research, Manila, Philippines.
p. 103-106.
Berrios, O.E. and A. Rubirigi. 1993.
lntegrated control of bacteria! wilt in
seed production by the Burundi national
potato program. In: Hartman, G.L. and
A.C. Hayward (eds.). Bacteria! wilt.
ACIAR Proceedings No. 45. Australian
Center for lnternational Agricultura!

Research, Canberra, Australia.


p. 284-288.
Bariyanga, J. 1997. Profit maximisation
analysis of lrish potato supply to urban
markets in Uganda. A case of Kampala,
Entebbe and Jinja. Master's thesis.
Makerere University, Kampala, Uganda.
FAO. 2000. Production yearbook. Food
and Agricultura! Organization of the
United Nations, Rome, ltaly.
French, E.R. 1994. lntegrated control of
bacteria! wilt of potato. CIP Circular
20:8-11.
French, E.R., R. Anguiz, and P. Aley. 1997.
The usefulness of potato resistance to

Ralstonia (Pseudomonas) solanacearum


for the integrated control of bacteria!
wilt. In: Prior, P., C. Allenand, and
J. Elphinstone (eds.). Bacteria! wilt
diseases. Springer, NY, USA.
p. 381-385.
Granada, G.A. and L. Sequeira.1983.
Survival of Pseudomonas solanacearum
in soil, rhizosphere, and plant roots.
Canadian Journal of Microbiology
(29):433-440.
Gunadi, N., E. Chujoy, M. Kusmana,
l. Surviani, O.S. Gunawan, and
R. Sinung-Basuki. 1998. Effect of crop
rotation patterns on Ralstonia
solanacearum population in the soil.
Potato research in Indonesia. CIP
Working Paper Series. CIP/RIV.
p. 56-61.
Kakuhenzire, R., F. Alacho, J. Birikunzira,
G. Turyamureeba, and L. Sikka. 1993.
Progress in field evaluation for
resistance to Pseudomonas
solanacearum and cultural methods for
control of bacteria! wilt in Uganda. In:
Proceedings of a workshop on bacteria!
wilt, Bujumbura, Burundi, 22-26
February 1993. p. 76-82.
Kindness, H. 1993. Economic analysis of
cotton variety and pest control on-farm
trials in Abella and Bele area in 1992
and 1993. Farmers research project
(FRP): On-farm trials in north Orno
report for 1993. p. 26-57. (photocopy
only available from PRAPACE,
Kampala, Uganda.)

CIP Program Report 1999 - 2000

139

Kloos, J.P., B.B. Fernandez, A.S. Tumapon,


and L. Villanueva. 1991. Effects of crop
rotation on incidence of potato bacteria!
wi lt caused by Pseudomonas
solanacearum E.F. Smith. Asian Patato
Journal 2:1-3.
Lemaga B., D. Siriri, and P. Ebanyat.
2001 a. Effect of soil amendments on
bacteria! wilt incidence and yield of
potatoes in southwestern Uganda.
African Crop Science Journal 9:267278.
Lemaga, B., J.J. Hakiza, F.O.A. Alacha,
and R. Kakuhenzire. 1997. lntegrated
control of potato bacteria! wilt in
southwestern Uganda. In: Proceedings
of the 4th triennial conference of the
African Patato Association, Pretoria,
South Africa, 23-28 February 1887.
p. 188-195.
Lemaga, B., J.J. Hakiza, J. Bariyanga, and
P. T. Ewel l. 1999. Patato production and
importance of patato bacteria! wilt in
Kabale District, southwestern Uganda:
A farm survey report. African Highlands
lnitiative/lnternational Patato Center,
Nairobi, Kenya. 40 p.
Lemaga, B., R. Kanzikwera,
R. Kakuhenzire, J.J. Hakiza, and
G. Manzi. 2001 b. The effect of crop
rotation on bacteria! wilt incidence and
patato tuber yield. African Crop Science
Journal 9:257-266.
Lloyd, A.B. 1976. Bacteria! wilt in coldtemperature climate of Australia. In:
Planning conference and workshop on
ecology and control of bacteria! wilt
caused by Pseudomonas solanacearum.
North Carolina State University,
Raleigh, NC, USA. p. 134-135.
Low, J.W. 1997. Patato in southwestern
Uganda. Threats to sustainable
production. African Crop Science
Journal 5:395-412.
Mafongoya, P.L, B.H. Dzowela, and
P.K. Nair. 1997. Effect of multipurpose
tree, age of cutting and drying method
on pruning quality. In: Cadish, G. and
K.E. Giller (eds.). Driven by nature:
Plant litter quality and decomposition.

140

Research on Potato

CAB lnternational, Wallingford, UK.


p.167-174.
Muchovej, V.V., R.C.M. Muchovej, O.O.
Dingra, and L.A Majia. 1980.
Suppression of anthracnose of soybeans
by calcium. Plant Disease 64:10881089.
Opio, A.F. 1988. Host range and biotypes
of Pseudomonas solanacearum
E.F. Smith in Uganda: A preliminary
study. Paper presented at the 15th
lnternational Congress of Plant
Pathology, Kyoto, Japan, 20-27 August
1988.
Palm, C.A. and A.P. Rowland. 1997.
Chemical characterisation of plant
quality for decomposition. In: Cadish,
G. and K.E. Giller (eds.). Driven by
nature: Plant litter quality and
decomposition. CAB lnternational,
Wallingford, UK. p. 379-392.
Priou, S. 2000. Developments of
components of bacteria! wilt control
and implementation of IDM: Subproject annual progress report for the
period Jan. 1- Dec. 31, 2000.
lnternational Patato Center, Lima, Peru.

13 p.
Priou, S., L. Gutarra, H. Fernandez, and
P. Aley. 1 999a. Sensitive detection of
Ralstonia so/anacearum in latently
infected potato tubers and soil by
postenrichment ELISA. In: lmpact on a
changing world: CIP Program Report
1997-1998. lnternational Potato Center,
Lima, Peru. p. 111-122.
Priou, S., P. Aley, E. Chujoy, B. Lemaga,
and E. French. 1999b. lntegrated control
of bacteria! wilt of patato: CIP Slide
Training Series IV-3. lnternational Potato
Center, Lima, Peru. 30 p.
Ramesh, C.R. and A.K. Bandyopadhyay.
1993. Bacteria! wilt potential of
Andaman and Nicobar lslands. In:
Hartman, G.L. and A.C. Hayward (eds.).
Bacteria! wilt. Proceedings of the
lnternational Bacteria! Wilt Symposium
held at Kaohsiung, Taiwan, 28-31
October 1992. ACIAR Proceedings No.
45. Australian Centre for lnternational

Agricu ltural Research, Can berra,


Australia.
Tusiime, G., E. Adipala, F. Opio, and
A.S. Bhagsari. 1996a. Screening
solanum potato genotypes for resistance
to Pseudomonas solanacearum in
Uganda. African Journal of Plant
Protection 6:96-1 07.
Tusiime, G., E. Adipala, F. Opio, and
A.S. Bhagsari. 1996b. Occurrence of
Pseudomonas solanacearum latent
infection in potato tubers and weeds in
highland Uganda. African Journal of
Plant Protection 6:108-118.
Van der Zaag, P. 1986. Potato production
under Pseudomonas solanacearum

conditions: Sources and management of


planting material. In: Persley, G.J. (ed.).
Bacteria! wilt disease in Asia and the
South Pacific. ACIAR Proceedings No.
13. Australian Centre for lnternational
Agricultura! Research, Canberra,
Australia. p. 84-88.
Verma, R.K. and G.S. Shekhawat. 1991.
Effect of crop rotation and chemical soil
treatment on bacteria! wilt of potato.
lndian Phytopathology 44:5-8.
Yost, D. and H. Eswaran. 1990. Major land
resource areas of- Uganda. USIAD,
Washington, DC, USA.

CIP Program Report 1999 - 2000

141

Optimization of Sample Size for the Detection of


Latent lnfection by Ralstonia solanacearum in Potato
Seed Tubers in the Highlands of Peru and Bolivia
S. Priou 1, R. Torres 2 , A. Villar2, J. Martinez 3 , O. Barrea3, L. Gutarra1, and F. de
Mendiburu 1

A zero-tolerance sampling strategy was evaluated in the Andean highlands of


Peru and Bolivia to optimize the detection of Ralstonia solanacearum (Smith)
Yabuuchi et al. (1992, 1995), the causal agent of bacterial wilt in potato
(Solanum tuberosum L.), in seed tubers harvested from symptomless crops. A
sensitive and specific serological method developed at CIP was used to detect
the pathogen in latently infected tubers. Optimum sample size was evaluated
for symptomless crops after analyzing various numbers of composite samples
and using a binomial distribution model to calculate the detection probabilities. In both Peru and Bolivia, R. solanacearum was detected in several
composite samples of all lots that carne from fields with visible symptoms,
even from those with very low levels of wilt incidence. That validates the
effectiveness of the detection technique. In Peru, 53.3% of seed lots coming
from apparently healthy fields, and in Bolivia 35.7%, were found positive for
the pathogen at altitudes of up to 3450 m. This demonstrates that contaminated seed from the inter-Andean valleys is responsible for contaminating
these seed-producing areas. In Peru, R. solanacearum was detected with 95%
probability in samples of 250 tubers from seed lots coming from symptomless
crops. In sample sizes of 350 tubers, the detection probability was 99%.
These sample sizes are much lower than those obtained theoretically using
the binomial distribution model. These are numbers of seed that feasibly can
be processed in a seed-health test without incurring too high a cost for labor
and materials. In Bolivia a larger sample size would be necessary to reach
such detection probabilities. That may be dueto lower detection efficiency or
lower inoculum levels in seed lots from newly contaminated areas. Further
studies are necessary in these areas to determine the optimum sample size
after having optimized detection efficiency.
Bacteria! wilt (BW) or brown rot of patato
is a devastating disease in both the
developed and developing world. In cool
1

CIP, Lima, Peru.

Servicio Nacional de Sanidad Agropecuaria, Baos del Inca,


Cajamarca, Peru.
3
Fundacin PROI NP A, Sucre, Bolivia.

conditions, such as are found in the tropics


at altitudes above 2500 m, infected but
symptomless plants may harbar the
bacterium and transmit it to progeny tubers
as latent infection. This may lead to
severe disease outbreaks when the tubers
are grown at warmer sites (Ciampi et al.,

CIP Program Report 1999 - 2000

143

1980; Hayward, 1991 ). Therefore, the use


of healthy planting material is the most
effective means to control the disease
(Hayward, 1991 ). Original planting
material used in formal or traditional
multiplication schemes must be tested for
latent infection by R. solanacearum.
In Australia and in many countries of
Europe, North America, and Latn
America, BW is considered a quarantine
disease with zero tolerance for BW in
certified tubers. In a certification scheme,
visual inspection is complemented by
laboratory tests to detect latent infections.
However, the principies underlying the
probability of visual and serological
detection remain the same, i.e., the
chances of sampling a diseased plant or
tuber are a function of disease incidence
and sample size (De Boer and Hall, 1996).
The sample size necessary to achieve the
desired probability of detection of low
levels of disease contamination also has to
take into account the aspects of practica!-
ity and cost effectiveness (Geng et al.,
1983; De Boer and Hall 1996).
In the European Union, the standard
sample size to monitor the occurrence of
R. solanacearum in patato seed tubers is
200 tubers out of 25 t (Janse, 1988; OEPP/
EPPO, 1990). That number is based on a
95% probability of detecting at least one
infected tuber in a seed lot if 1.5% of the
tubers are infected, according to the
Poisson distribution (De Boer, pers.
comm.).
In South Africa, in G2 to G8 seed multiplication crops, 4605 tubers/planting unit are
analyzed by enzyme-linked immunosorbent assay (ELISA). This sample size is
based on a statistical model developed by
Clayton and Slack (1988), to give a 99%
probabi 1ity of detection if 0.1 % of tubers
are infected and the planting unit is 1 O ha
or more (Swanepoel and Theron, 1999).
However, Clayton and Slack (1988)
mentioned that in a zero-tolerance situation it is appropriate to take advantage of
a stepwise sampling scheme. This is

144

Research on Patato

because patato inspection schemes


involve multiple inspections and generally
comprise two field visits and one tuber
. test. lf disease is not found in a particular
sample, then subsequent sampling will
take place. There is a potential reduction
in costs and labor when inspections take
place in such a stepwise manner. Therefore, the theoretical sample size of 4605
could be divided into three samples: 3005
plants for the first visual inspection
(generally 40-60 days after planting),
1 200 plants for the second visit (general ly
75-100 days after planting), and 400
tubers for seed testing (Clayton and Slack,
1988).
Potato ring rot disease, caused by
Clavibacter michiganensis subsp.
sepedonicus, is also a zero tolerance
situation. De Boer and Hall (1996) calculated the sample size using a binomial
distribution model. The authors concluded
that 400 tubers is a feasible sample size
for testing seed patato lots from a 1 O ha
field, which was planted with seed tubers
of which 1 % were infected with the ring
rot pathogen.
Certification agencies in several southern
states of the USA (e.g., Florida and
California) have a set sample size for virus
testing in winter-planted seed tubers.
Regardless of field size under 16 ha, the
sample size is 400 tubers per seed lot
(USA Certification Committee, 1985).
Most of the reported sampling strategies
mentioned above are based on statistical
models. There are no data on in situ
assessment of the optimum sample size for
BW, nor on the incidence of latent BW
infection in tubers produced at high
elevations. Therefore, a zero-tolerance
sampling strategy was evaluated in the
Andean highlands of Peru and Bolivia to
optimize the detection of R. solanacearum
in seed tubers harvested from symptomless
crops. A sensitive and specific serological
method to detect R. solanacearum in
latently infected tubers is now available at
CIP (Priou et al., 1 999a, 1999b). The

method, cal led postenrichment enzymelinked immunosorbent assay on


nitrocellulose membrane (NCM-ELISA), is
suitable for seed testing in resource-poor
countries.
Effectiveness of the detection technique
was validated in crops exhibiting low BW
incidence. Optimum sample size was
evaluated for symptomless crops after
analyzing various numbers of composite
samples and using a binomial distribution
model to calculate the detection probabilities. When the percentage of diseased
seed is very low, the probability of the
occurrence of diseased seeds in a sample
follows a Poisson distribution (Lund and
Sun, 1985). The binomial distribution is
applicable to our cas, however, because
severa! composite samples are analyzed
instead of a single sample, and the seed
lot is considered as positive if at least one
composite sample is found positive.

Materials and Methods


Sample collection
Six hundred tubers/seed lot were collected
from fields of 0.25 to 1.5 ha during two
cropping seasons, 1999 and 2000. A seed
lot is defined as comprising tubers of one
variety harvested from one field site and
originating from the same seed source.
Tubers were randomly selected from heaps
at harvest or from stores at the beginning
of the storage period (within 4 weeks of
harvest in Cajamarca, Per.u, and between
4 and 12 weeks after harvest in Bolivia).
Sixteen fields in Peru (Cajamarca and La
Libertad departments) and 14 fields in
Bolivia (Chuquisaca and Potosi departments) exhibiting low incidence of BW
(0.1-3%) were selected. BW incidence has
been scored according to the number of
wilted plants observed among a total of
400 plants per hectare by examining 20
rows of 20 plants following a cross pattern
design.

Samples were also collected from harvested tubers of symptomless crops in


BW-endemic areas or in areas at higher

elevations suspected of having been


planted with infected seed produced in
endemic areas at lower elevations. After
one or two visual field inspections to
confirm the absence of BW symptoms,
34 samples were collected in seed lots in
seven provinces (Cajamarca, Celendin,
Chota, Contu maza, San Marcos, San
Miguel and San Pablo) in Cajamarca
Department and Huamachuco Province of
La Libertad in Peru. In Bolivia, a total of
42 samples were collected in seed lots in
Chuquisaca Department (Yamparaez and
Tomina provinces) and Potosi Department
(Saavedra Province). Sorne fields were
chosen as tentative negative controls
because of no known history of BW
infection, such as eight fields in Potosi, or
because they were obtained from reliable
in vitro propagation (lots 8 and 1O in Peru
(Table 1)). Altitudes for symptomless fields
ranged from 2200 to 3400 m in Peru and
from 2371 to 3450 m in Bolivia. In Bolivia,
all samples were from variety Dsire,
whereas in Peru samples were taken from
seed lots of various commercial and native
varieties (Table 1). Tubers were kept in
paper bags and stored at room temperature
before being processed within 2 weeks of
col lection.

Sample testing
Tubers were randomly divided into 24
composite samples of 25 tubers each. For
the detection of R. solanacearum, tubers
were processed as described by Priou et al.
(1999a, 1999b) in the laboratory of the
Peruvian Crop Protection Service
(SENASA) in Cajamarca, Peru, and in the
laboratory of the Foundation for the
Promotion of and Research on Andean
Products (PROINPA) in Sucre, Bolivia.
After tuber disinfection, fragments were
taken from the vascular ring of each tuber;
the fragments from a composite sample of
25 tubers were crushed together in one
bag. Two 0.5 mi aliquots of tuber extract
were taken from each bag and rn ixed with
an equal volume of modified SMSA broth
for enrichment (incubation for 48 h at 30C
with manual agitation at least twice a

CIP Program Report 1999 - 2000

145

Table 1. Detection of Ralstonia solanacearum in seed lots coming from fields with various bacteria! wilt
(BW) incidences in Cajamarca and La Libertad departments, Peru.
Variety
ELISA-Positive among
Si te
Altitude {m)
Lot no. Bacteria! Wilt in
24 composite samples
the field {%)
o
Canchn
La Encaada
2800
5
1
o
Canchn
La Encaada
3000
7
2
Huamachuco 1
o
Canchn
3000
8
3
Amarilis
San Juan
2200
o
4
5
Cajamarca
o
3
Amarilis
2750
5
o
Canchn
La Encaada
3000
14
6
o
Amarilis
Huamachuco 1
3000
7
6
o
o
Amarilis
Contumaza
2700
8
Yungay
o
o
La Encaada
3000
9
o
10
Perricholi
Parean cooperative
3150
o
o
11
4
Tumbay
Chota
2900
o
1
Atahualpa San Miguel
12
2800
13
o
o
Huagalina
Su ere
2650
o
Huagalina
Su ere
2900
o
14
o
o
15
Canchn
Huamachuco 1
3200
Huamachuco 1
o
o
Canchn
16
3350
o
Yungay
Huamachuco 1
17
o
3400
o
18
Amarilis
Huamachuco 1
o
3200
Huamachuco 1
19
o
o
Yungay
3150
o
Huamachuco 1
20
Canchn
o
3200
o
21
Canchn
o
Huamachuco 1
3200
o
Yungay
22
8
La Encaada
2900
23
o
Yungay
11
Baos del Inca
3000
o
24
4
Amarilis
La Encaada
3000
o
25
6
Amarilis
La Encaada
3100
o
26
o
Amarilis
Baos del Inca
2800
27
o
o
Amarilis
Baos del Inca
2900
o
Yungay
28
13
Baos del 1nea
2950
o
29
Amarilis
10
La Encaada
3100
o
30
7
Amarilis
La Encaada
3000
o
o
31
Amarilis
Su ere
2650
o
32
9
Amarilis
Su ere
2700
o
Canchn
Sucre
o
33
2900
o
Yungay
34
Baos del Inca
5
3000
Total
126
0.1-1
1
35
Amarilis
Llaca nora
2900
0.1-1
36
15
Amarilis
Su ere
2850
37
0.1-1
19
Canchn
Su ere
2600
0.1-1
18
Canchn
38
Eduardo Villanueva
2200
39
0.1-1
17
Canchn
Sucre
2750
40
0.1-1
3
Canchn
Eduardo Villanueva
2240
0.1-1
Canchn
41
8
Sucre
2650
42
0.1-1
8
Mixture
Su ere
2700
43
0.1-1
Chaucha
5
San Pablo
2700
44
1.1-3%
11
Canchn
Jos Galvez
2600
Canchn
45
1.1-3%
13
Su ere
2650
46
1.1-3%
18
Mixture
Su ere
2650
47
1.1-3%
13
Mixture
Su ere
2700
48
1.1-3%
21
Amarilis
Sucre
2600
49
1.1-3%
3
Amarms
San Juan
2200
1.1-3%
15
50
Amarilis
Llaca nora
2900
Total
188
1

designates those sites in La Libertad.

146

Research on Potato

day). For each seed lot, each of the


resulting 48 tuber-extract aliquots was
dotted on the membrane and analyzed by
NCM-ELISA. The number of positive
composite (25-tuber) samples for each
seed lot was recorded. Only one aliquot of
two needs to be positive to rate the
composite sample positive.
In Peru, positive results in ELISA were
confirmed by either isolating R.
solanacearum on modified Kelman's
medium (French et al., 1995) in SENASA,
or at CIP by polymerase chain reaction
(PCR) if the pathogen could not be isolated
from the sample. One mi of enriched tuber
extract was centrifuged at 12,000 rpm for
5 min, then the pellet was washed three
times with nuclease-free water. The pellet
was then resuspended in 20 I nucleasefree water and boiled for 15 min to
denature DNA. Samples were kept at
-20C until use. Two I of this suspension
was used as a template for PCR according
to Roncal et al. (1999). Pathogenicity of
the isolates of R. solanacearum was
confirmed by inoculating tomato plantlets
as described by Janse (1988).

Calculations
For a given bacteria! wilt incidence (%),
the average number of positive composite
(25-tuber) samples per lot, M, could be
calculated as M = number of positive
composite samples/number of positive lots.
This average gives us M = 24 p where p is
the probabi 1ity that the composite sample
in a lot is positive. Thus p = M/24 and
from p we obtain q = 1 - p (probability of
obtaining no infected sample). Then
applying the binomial distribution, the
probability P to obtain at least one infected composite sample (that is enough to
consider the lot infected) is P = 1 - qx
where X (in this case 24) is the number of
25-tuber samples to be analyzed. Different
values of X are tested and the optimum
sample size is the number X that al lows a
probability (P) of detection of 95%
(Snedecor and Cochran, 1980).

In studying sampling schemes for a zerotolerance situation, we focus on the


probability of erroneously accepting a
field in which BW is present (PEA). This
probability is given by PEA= 1 - P. To
determine this probability we assume (1)
that disease occurs randomly in the field
(that is generally the case when the main
inoculum source is infected seed), (2) that
a sample size n is taken at random, and
(3) that each plant has the same probability of being diseased (Clayton and Slack,
1988).

Results
Both in Peru and Bolivia, R. solanacearum
was detected in several composite
samples of all lots that carne from fields
with visible symptoms, even from those
with very low levels of wilt incidence
(Tables 1 and 3). For the lots coming from
fields with O. 1-1% wilt incidence, a
probability of 95% to detect one positive
sample was reached by analyzing 150
tubers in Peru and 225 tubers in Bolivia
(Tables 2 and 4). For fields with between
1. 1 and 3% wilt, 100 tubers have to be
analyzed in Peru and 200 in Bolivia to
obtain P = 95% (Tables 2 and 4).
In Peru, 18 seed lots (53.3%) coming from
apparently healthy fields were found
positive in ELISA (Table 1). Five of these
positive lots were from recognized seed
growers. Lots 8 and 1O, which served as
tentative negative controls, were found
negative. All positive results in ELISA
were confirmed by isolation of the pathogen in Cajamarca or by PCR at CIP;
pathogenicity tests on tomato were also
positive (data not shown).
In Bolivia, 15 seed lots (35.7%) of the
symptomless crops were found infected
with BW (Table 3). From the 42 apparently
healthy lots, 28 were from highlands
(3284-3450 m) and 14 were collected
between 2371 and 2850 m. Among the 14
lots from lower areas, 50% were found
positive. This is not surprising because BW
is endemic in the seed-producing areas of

CIP Program Report 1999- 2000

147

Table 2. Probability (P) of detecting Ralstonia so/anacearum in seed lots coming from fields with various
BW incidences according to the sample size analyzed in Peru.
P for 0.1-1% ewz P for 1.1-3% ew3
Composite samples {X)
Tubers
P for 0% ew1
2200-3350 m
2200-2900 m
2200-2900 m
25
0.292
1
0.435
0.560

50
75

3
4
5
6
7
8
9

100

10
11

12
13
14
15
16
17
18
19
20
21

22
23
24
1P =
2p =

1 - o. 708x.
1 - 0.56sX.

3p=

1 - 0.440X.

125
150
175
200
225
250
275
300
325
350
375
400
425
450
475
500
525
550
575
600

0.499
0.645
0.749
0.822
0.874
0.911
0.937
0.955
0.968
0.978
0.984
0.989
0.992
0.994
0.996
0.997
0.998
0.999
0.999
0.999
0.999
1.000
1.000

Cordillera del Rosal, Chuquisaca Department. From the 28 lots from the highlands
in Chuquisaca and Potosi, 8 (28.6%) have
been found positive, all from Chuquisaca
(Table 3). This result demonstrates that
infected seed coming from the contaminated inter-Andean valleys of Chuquisaca
is responsible for contamination of seedproducing areas. The fact that lots from the
highlands of Potosi were all found negative was expected because BW has never
been reported in this area. These negative
results confirm the specificity of the test
(absence of cross-reaction in ELISA).
In Peru, the analysis of 225 tubers gives a
probability of 95% to detect at least one
positive sample in lots harvested from
symptomless fields (Table 2). When data
were separately analyzed by year (14 lots
analyzed in 1999 and 20 lots in 2000),

148

Research on Potato

0.681
0.820
0.898
0.942
0.967
0.982
0.990
0.994
0.997
0.998
0.999
0.999
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000

0:806
0.915
0.963
0.984
0.993
0.997
0.999
0.999
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000

approximately the same results were


obtained (275 tubers in 1999 and 200 in
2000) (data not shown).
In Bolivia, for symptomless crops, analyzing 600 tubers gives a PEA of 23% (Table
4). Thus, it would be necessary to analyze
many more than 600 tubers to reach a
95% detection probability.

Discussion
The serological detection technique. has
been val idated because the pathogen
could be detected in all fields with visible
BW symptoms. No .cross-reaction was
observed at Cajamarca (all positive RISA
results were confirmed), anc:Fexpected
negative seed lots were found negative in
both countries.

Table 3. Detection of Ralstona solanacearum in seed lots coming from fields wtth various BW incidences in
seed-producing areas of Chuquisaca and Potosi departments, Bolivia.
lot no.
Bacteria! Wilt in
ELISA-positive among 24
Site
Altitude (m)
the field (%)
composite samples
o
o
1
Canadillas
2850
o
o
Canadillas
2
2850
o
o
Canadillas
2850
3
o
o
Canadillas
4
2850
o
Lajas
2
3284
5
o
o
Huano loma
3284
6
o
Lajas
o
3284
7
Lajas
o
3284
1
8
o
o
Tarabuco
3284
9
3284
1
Tarabuco
o
10
Alto leuquepa.
o
2600
o
11
Alto Rosal
2500
o
o
12
Alto leuquepa.
2600
o
o
13
Alto Rosal
2500
1
o
14
2371
Faje ha
1
o
15
Fajcha
2390
o
3
16
2490
Alto Rosal
3
o
17
2390
Fajcha
2
o
18
2371
Faje ha
1
o
19
2385
Faje ha
1
o
20
3284
Tarabuco
o
o
21
3284
Tarabuco
1
o
22
3284
Lajas sijlla
o
o
23
3284
Lajas
o
o
24
3284
Lajas
o
o
25
3284
Lajas
o
o
26
3284
Lajas
o
o
27
3284
Lajas
o
o
28
3284
Jatun Mayu
2
o
29
3400
Central Alta
o
o
30
3400
Maraza
o
o
31
3400
Ckochas 1
o
o
32
3400
Simi simi 1
o
o
33
3400
Sijllani 1
o
o
34
3400
Chacabuco 1
o
o
35
3450
Chinoli 1
o
o
36
PROINPA-E.E.Chinoli1
3450
o
o
37
3400
Chacabuco 1
o
o
38
3400
Kawawasi 1
o
o
39
3450
Lajas (Tarabuco)
2
o
40
3450
Lajas (Tarabuco)
1
o
41
3450
Lajas (Tarabuco)
1
o
42
23
Total
2400
Muyuorko
5
0.1-1
43
1875
Las Palmas
16
0.1-1
44
2400
Muyuorko
10
0.1-1
45
2315
Muyuorko
6
0.1-1
46
2410
Muyuorko
3
0.1-1
47
2400"
Muyuorko
8
0.1-1
48
2450
Chajra Mayu
8
1.1-3
49

CIP Program Report 1999 - 2000

149

Table 3. (continued)
Lot no.

Bacteria! Wilt in
the field (%)

ELISA-positive among 24
composite samples

1.1-3
1.1-3
1.1-3
1.1-3
1.1-3
1.1-3
1.1-3

5
3
9
6
8
9
7
103

50
51
52
53
54
55
56
Total
1

Site
Muyuorko
Muyuorko
Chajra Mayu
Muyuorko
Muyuorko
K'ara Estancia
Las Palmas

Altitude (m)
2390
2400
2450
2390
2395
2468
2200

Designates those sites in Potosi.

Table 4. Probability (P) of detecting Ralstonia solanacearum in seed lots coming from fields with various
BW incidences according to the sample size analyzed in Bolivia.
P for 0.1-1% ewz P for 1.1-3% ew3
Composite
Tubers
P for 0% ew1
2200-2468 m
samples (X)
2371-3450 m
1875-2410 m
0.307
0.287
1
25
0.060
0.520
50
0.116
0.492
2
0.667
3
0.638
75
0.169
4
100
0.219
0.769
0.742
0.840
0.816
5
125
0.266
0.889
0.869
150
0.310
6
0.923
0.906
7
175
0.352
8
200
0.390
0.933
0.947
9
225
0.427
0.952
0.963
10
250
0.461
0.966
0.974
11
275
0.494
0.976
0.982
12
300
0.524
0.983
0.988
13
325
0.553
0.988
0.991
14
350
0.579
0.991
0.994
15
375
0.605
0.996
0.994
16
400
0.628
0.996
0.997
0.998
0.997
17
425
0.651
18
450
0.672
0.999
0.998
19
475
0.691
0.999
0.998
0.999
0.999
20
500
0.710
1.000
0.999
21
525
0.727
1.000
0.999
22
550
0.744
1.000
1.000
23
575
0.759
1.000
1.000
24
600
0.773
1 P = 1 - 0.94QX.
2 p = 1 - 0.66flX.
3 P = 1 - 0.713x.

Surprisingly, BW infection was found at


high elevations in both countries, up to
3450 m, where normally the disease does
not occur. Although low temperatures
prevent development of wilt, the pathogen
remains in the plant and is transmitted to
progeny tubers. These resu lts provide data

150

Research on Potato

to local authorities and plant protection


services for mapping latently infested
areas and applying certification regulations for the production of BW-free seed.
Moreover, detection of BW infection in
areas where BW was previously not
reported demonstrates the flow of

contaminated seed from lower endemic


areas to higher areas. Farmers often move
the degenerated seed to higher areas to
improve its physiological condition and
reduce virus infection, but this leads to
contamination of disease-free areas
(Thiele, 1998).
Results in Peru al low us to set the sample
size at 250 tubers/seed lot to detect latent
infection by R. solanacearum (at least a
95% confidence level) from a field of 0.25
to 2 ha with no or only trace BW incidence. Trace incidence of BW can escape
field inspection if diseased plants dry up or
if unscrupulous seed growers remove them
before the technician's visit. Analyzing
350 tubers would allow a detection
probability of 99%. Results in Bolivia,
however, do not agree. To obtain a detection probability of only 95% from
symptomless fields in Bolivia requires a
larger number of tubers. There may be a
lower tuber infection rate in Solivian seed
lots in high elevations because the pathogen was introduced more recently.
However, the fact that the sample size is
also higher, even for BW-infected fields,
might indicate a lower detection efficiency in the PROINPA laboratory.
Enrichment efficiency would be lower if
agitation was insufficient or the temperature was lower than 30C during the
enrichment procedure. More studies are
necessary to confirm the optimum sample
size in various ecological zones in Bolivia

after having optimized the detection


efficiency in the laboratory.

Conclusion
Sample sizes of 250 and 350 tubers have
been used successfully to detect R.
solanacearum in Peruvian seed lots from
symptomless potato crops with probabilities of 95% and 99%, respectively. These
are numbers of seed that feasibly can be
processed in a seed-health test without
incurring too high a cost of labor and
material.
The number 400 has been selected by
many seed certification agencies. According to our results in Peru (Table 2), the
European standard of 200 tubers would
provide a PEA of 6.3% (P = 0.937) in the
case of symptomless crops. That is slightly
h igher than the theoretical probabi 1ity of
5%. Considering that moderately to highly
susceptible potato varieties can harbor
latent infection in their progeny tubers at
frequencies ranging from 20% to 70%
(Priou et al., 2001 ), field incidences of 1 %
and 0.1 % wou Id lead to percentages of
infected tubers of 0.2-0.7% and 0.020.07%, respectively. According to the
binomial distribution, the number of tubers
theoretically needed for testing to detect
these levels of tuber infection at the 95%
level of confidence would be much higher
than that necessary in Peru to detect the
same tuber infection rates (Table 5).

Table 5. Number of tubers (n) to be analyzed to detect Ralstonia solanacearum in seed lots with various
tuber infection levels (p) at confidence levels P of 95% and 99%.
n theoretical 1
n observed
p' = wilt
p = infected
1-p
P
incidence
tubers rate
1360
100
0.9978
0.95
0.011
0.0022
100
0.95
388
0.9923
0.011
0.0077
14977
150
0.95
0.9998
0.001
0.0002
150
4278
0.95
0.9993
0.001
0.0007

0.011
0.011
0.001

0.0022
0.0077
0.0002

0.9978
0.9923
0.9998

0.99
0.99
0.99

0.001

0.0007

0.9993

0.99

2091
596
23024
6577

150
150
200
200

1 n theoretical was calculated following the binomial distribution (P = 1 - [1-p]n) and compared to the sample sizes

(n observed) that were effective to detect these levels of infection in Peru, according to Table 2.

CIP Program Report 1999 - 2000

151

To be 100% certain that a crop is free from


BW, all the tubers would need to be
tested, and that is obviously impractical. A
finite number of samples can be tested,
within the limits of practica! constraints
and cost effectiveness, to obtain an
estmate of the likelihood that a seed lot is
infected. The probability of detecting an
infected sample (PO) is equal to the
probability of one infected sample in the
samples analyzed (PS) in the case of 100%
sensitivity (SE) and specificity (SP) of the
test (De Boer, 1991 ). lf efficiency is not
100%, then PO = PS (SE) + (1-SE)(l -SP). In
CIP laboratory conditions, R. solanacearum can be detected in 10-4 dilutions in
healthy tuber extract obtained from one
latently infected tuber, demonstrating the
very high sensitivity of the technique. This
ensures that a single infected tuber in a
composite sample of 25 tubers would be
detected (Priou et al., 1999a). Thus, we
could assume that for this serological test,
PO = PS, but we cannot ensure that this
would be valid under all laboratory
conditions.
The resu lts of laboratory tests can provide
valid information about the disease status
of a patato crop. Zero tolerance can be
applied in the sense that when one lot
tests positive it cannot be certified. Nor
could it even be planted if the crop
protection service establishes drastic
quarantine regulations to avoid contamination of BW-free areas. But a negative test
does not imply a 100% certainty that the
lot is BW-free. The number of samples to
analyze is a result of a compromise
between cost, labor, and physical limitations for conducting the seed testing, and
the level of confidence required according
to the purpose of the detection. For
instance, when infection of a lot with
R. solanacearum is suspected in a BW-free
seed-producing area, a higher probability
should be applied to calculate the sample
number to avoid disease spread to the
enti re a rea.
lf the incidence of BW can be considered
uniform throughout the field, the samples
152

Research on Potato

required would be the same regardless of


the size of the field because calculations
are based on sampling from an infinite
population of tubers. This requirement that
diseased plants occur randomly at a
uniform rate across the field may be hard
to maintain as fields become larger. In the
Andes, however, the great majority of
growers are small-scale farmers with less
than 1 ha of patato (Thiele, 1998). Fieids
with variability in disease rate or where
plants originate from different seed sources
should be handled as two or more separate
units for sampling purposes (Lund and Sun,
1985). Otherwise a stratified sampling
should be applied, but sampling should be
at random in each strata since disease rate
should be constant within each strata (De
Boer, 1991; Clayton and Slack, 1988).

References
Certification committee, The Patato
Association of America. 1985. Problems
in uniform seed patato certification
procedu'res. American Patato Journal
62:376-386.
Ciampi, L., L. Sequeira, and E.R. French.
1980. Latent infection of potato tubers
by Pseudomonas solanacearum.
American Patato Journal 57:377-386.
Clayton, M.K. and S.A. Slack. 1988.
Sample size determination in zero
tolerance circumstances and the
implications of stepwise sampling:
Bacteria! ring rot as a special case.
American Patato Journal 65:711-723.
De Boer, S.H. 1991. Current status and
future prospects of bacteria! ring rot
testing. American Patato Journal
68:107-113.
De Boer, S.H. and J.W. Hall. 1996. The
probability of detecting Clavibacter
michiganensis subsp. sepedonicus by
indexing seed patato lots with
serological tests. Journal of
Phytopathology 144:459-463.
French, E.R., L. Gutarra, P. Aley, and
J. Elphinstone. 1995. Culture media for
Pseudomonas solanacearum: lsolation,
identification and maintenance.
Fitopatologia 30:126-130.

Geng, S., R.N. Campbell, M. Carter, and


F.J. Hills. 1983. Quality-control
programs for seed borne pathogens.
Plant Disease 67:236-242.
Hayward, A.C. 1991. Biology and
epidemiology of bacteria! wilt caused
by Pseudomonas solanacearum. Annual
Review of Phytopathology 29:65-87.
Janse, J.D. 1988. A detection method for
Pseudomonas solanacearum in
symptomless potato tubers and sorne
data on its sensitivity and specificity.
Bulletin OEPP/EPPO Bulletin 18:343351.
Lund, R.E. and M.K.C. Sun. 1985. Sample
size determination for seed potato
certification. American Potato Journal
62:347-353.
EPPO (European Plant Protection
Organization). 1990. Quarantine
procedures no 26. Pseudomonas
solanacearum. OEPP/EPPO Bulletin
20:255-262.
Priou, S., L. Gutarra, and P. Aley. 1999a.
Highly sensitive detection of Ralstonia
solanacearum in latently infected
potato tubers by post-enrichment ELISA
on nitrocellulose membrane. EPPO/
OEPP Bulletin 29: 117-125.
Priou, S., L. Gutarra, H. Fernndez, and
P. Aley. 1999b. Sensitive detection of
Ralstonia solanacearum in latently
infected potato tubers and soil by postenrichment ELISA. lmpact on a
changing world. Program Report 199798. lnternational Potato Center, Lima,
Peru. p. 111-121

Priou, S., C. Salas, F. de Mendiburu,


L. Gutarra, and P. Aley. 2001.
Assessment of latent infection frequency
in progeny tubers of advanced potato
clones resistant to bacteria! wilt: A new
selection criterion. Scientist and farmer
- Partners in research for the 21st
century, program report 1999-2000.
lnternational Potato Center, Lima, Peru.
p. 105.
Roncal, J., L. Gutarra, and S. Priou. 1999.
Rapid differentiation of strains of
Ralstonia solanacearum by restriction
analysis of PCR-amplified fragments.
ACIAR Bacteria! Wilt Newsletter
16:7-10.
Snedecor, G.W. and W.G. Cochran. 1980.
Statistical methods. Seventh edition.
lowa State Press, Ames, IA, USA.
Swanepoel, A.E. and D.J. Theron. 1999.
Control measures for bacteria! wilt,
caused by Ralstonia solanacearum, as
appl ied by the South African potato
certification scheme. 14th Triennial
Conference of the European Association
for Potato Research, Sorrento (ltaly) 2-7
May 1999. EAPR Abstracts of
conference papers, posters and
demonstrations. Assessorato Agricoltura,
Sorrento, ltaly. p. 241-242.
Thiele, G. 1998. Informal potato seed
production systems in the Andes: Why
are they important and what should we
do with them. World Development
27:83-99.

CIP Program Report 1999 - 2000

153

Mass-Producing Beauveria brongniartii lnoculum,


An Economical, Farm-Level Method
F. Cisneros and A. Vera1

Beauveria brongniartii is a naturally-occurring fungus parasitizing the larvae


of the Andean potato weevil, Premnotrypes spp. (a serious pest of potato at
high altitudes). B. brongniartii has been studied in the northern, central, and
southern Peruvian Andes and found to be widely distributed. In our work to
develop a new method for mass-producing the fungus without using laboratory-produced inoculum, we found the larval stage of the weevil to be the
most susceptible to B. brongniartii. lnfection at this stage also produced the
most spores in the shortest period of time. Although the fungus has no particular structure that enables it to survive adverse conditions, it was able to
maintain its viability for long periods of time (more than one year) when
fungus-infected dried larvae were kept under cool, dry conditions. Mycelial
growth was activated when moisture was added to the medium containing
the infected larvae. Temperature and soil moisture affected the onset of the
disease and the development of the fungus, enhancing either the mycelial
growth (under moist conditions) or conidia production (under drier conditions). Findings indicate that B. brongniartii has high potential for use in the
control of the Andean potato weevil, especially when dosages of infected
substratum made economically at the cottage-level from low numbers of
infected weevil larvae are used for treating the soil of rustic potato stores.
Entomopathogens, particularly parasitic
fungi, are currently receiving a great deal
of attention as components in the management of insect pests (Alves, 1998; Faull,
1986; Ferron, 1978; Fuxa, 1987; Fuxa et
al., 1990; Tanada and Kaya, 1993), and
training on the use of these organisms is
often offered by national and international
organizations. Most research work,
however, deals with the study of taxonomy, biology, and the pathogenic
characteristics of these organisms (Brady,
1979; De Hoog, 1972; Ferron, 1983;
Macleod, 1954); whereas, techniques for
the mass multiplication of fung, practica!

CIP, Lima, Peru.

formulations, and methods of application


are less available. There is little integration between these two areas of research,
which may account for the rather low
consistency of results when fung are used
for pest-control purposes.
A natural ly-occurring fungus parasitizing
the larvae of the Andean patato weevil,
Premnotrypes spp. (a serious pest of patato
at high altitudes), has often been found in
our field trials. This fungus was reported in
the Andes for the first time in Huancayo,
Peru, in 1976 as Beauveria bassiana
(Alczar, 1976) but it was later identified
as B. brongniartii (Alczar et al., 1990).
Unlike most reported cases in which fung
thrive under tropical and subtropical

CIP Program Report 1999 - 2000

155

conditions (Alves, 1998; Ferron, 1983), this


particular fungus occurs in the high
mountains, above 2800 m altitude, where
low temperatures are prevalent (Alczar
et al., 1990).
Most successful cases using fungi in the
control of pests are related to a thorough
knowledge of the organism involved, the
identification of factors that affect its
effectiveness and survival in the field, the
method of mass multiplication, and
application techniques. The objectives of
this work were to determine the conditions
under which B. brongniartii thrives in
nature and to develop a new method for
mass-producing the fungus without using
laboratory-produced inoculum.

Methodology
Several steps were necessary to carry out
this research: (1) samples were collected
and tested for thei r response to temperature and soil moisture, (2) a massmultiplication process was developed, and
(3) the mass-produced fungus medium was
field-tested in rustic potato stores.

Sampling
Three test sites were selected in the
Peruvian Andes: Chinchero, Cusco, in the
south; Aymara, Huancayo, in central Peru;
and La Encaada, Cajamarca, in the north.
For each site, 1O sample units were
established in fields previously planted to
potato and 10 more were established in
rustic potato stores. A sampling unit
consisted of one square meter of soi 1
surface. For sampling purposes, three
layers, 1O cm thick, were studied to
determine the depth at which the weevil
and the fungus occurred. The soil of each
subsample was sieved and the number of
Andean patato weevils (larvae, pupae, and
adults) was recorded. lt was also noted
whether the i nsects were healthy or
infected with B. brongniartii. Healthy
insects were taken to the laboratory and
stored under moist conditions for eight
days to verify their uninfected condition.
lnfected insects were also taken to the
156

Research on Patato

laboratory to facilitate the development of


the disease.

lnfluence of temperature and soil moisture


Weevil larvae were treated with a dilution
of 2.5 x 108 conidia/ml and placed in
groups of 1O individuals per petri dish. Ten
petri dishes (replicates) were exposed to a
given temperature (5, 1O, 15, and 20C)
and soil moisture levels of 50%, 25%, and
10% of field capacity. A factorial design
with 1O replicates was used. Mortality was
recorded 12 days after treatment. Materials were kept under observation to
determine the influence of treatments on
mycel ium development and the production of conidia.

Mass multiplication of B. brongniartii


B. brongniartii was mass-produced to
replace the laboratory inoculum, which
is unavailable to farmers, used in the
traditional multiplication process-the
saprophytic phase. The proposed alternative was to develop a farm-level method
of multiplication by using fungus-infected
weevil larvae (parasitic phase) as
inoculum.

To obtain infected larvae in large numbers,


healthy larvae collected at harvest time
were placed in fungus-infested soils at the
rate of 1000 larvae/kg of soil. Five replications of two kinds of infested soil were
used: naturally-infested soil from rustic
stores and sterilized soil inoculated with a
dilution of the fungus conidia. A third
treatment consisted of immersing the
larvae in a dilution of fungus conidia and
placing them on wet paper.

To test for effectiveness, 1O, 20, and 30


infected larvae were introduced into
plastic bags containing moist barley, the
mass-multiplication medium. To determine the proper level of barley hydration,
500, 600, 700, 800, 900, and 1000 mi of
water/kg barley were tested. A completely
randomized design was used with 1O
replications. The level of fungus development was evaluated 15, 30, and 45 days

after inoculation by observing the fungus


saturation. A five-level scale was used to
evaluate saturation: O = no fungus (0%), 1
(1-25%), 2 (26-50%), 3 (51-75%), and 4
(76-100%).

Dosage of B. brongniartii for the control of


the Andean weevil
Dosages of 1, 2, and 3 kg of fungusinfected substratum from the massmultiplication trial were applied per
square meter of soil in rustic potato stores
in Cusco and Huancayo. The treated area
was infested with 250 Andean potato
weevil larvae/m 2 After 60 days, the
percentage of mortality was determined by
counting the infected larvae in each
treated area. A dosage of 2 kg/m 2 was
determined to be effective and this dosage
was then tested in Cusco, Cajamarca, and
Huancayo. Evaluations were made at 30,
60, 90, and 120 days after treatment.
Areas were divided into quarters and a
different quarter was evaluated at the end
of each 30-day period. The percentage of
mortality was determined by counting the
infected larvae in each plot.

Results
Geographical distribution
B. brongniartii infection was found in
specimens of the Andean potato weevil in
all sampling units of field and rustic potato
stores in Cusca, Huancayo, and
Cajamarca. In the field, the average level
of infection found in the weevil population
ranged from 9% in Huancayo to 12.4% in
Cusco. In rustic stores, the level of infection ranged from 11 .4% and Huancayo
15.2% in Cajamarca to 27.1 % in Cusco
(Table 1).

Relative susceptibility of developmental


stages of the potato weevil
Most of the specimens (71-81 % in the
field and 88-92% in stores) were infected
by the fungus at the larval stage. The
second most vulnerable stage was the
pupa, with 15-23% infected in the field

and 7-12% in stores. The least susceptible


stage was the adu lt, with less than 6%
infected in the field and about 1% in
stores (Table 2). Not only was the larval stage the most susceptible to
B. brongniartii, but it was also the stage
where the most spores were produced in
the shortest period of time.

Soil depth and fungal infection


Most weevils were found in the upper
1O cm of soil in both the field and rustic
stores. In the field, the average proportion
of weevils found at this level was 77% for
Cusco and 84% for Huancayo. In rustic
stores, the proportion at this level averaged 88% for Cusco, 79% for Huancayo,
and 81 % for Cajamarca. At the second soil
layer (11-20 cm), the average proportion
of weevils ranged from 17-23% in the
field and 12-19% in stores. Except for one
field sample in Cusco of 30 insects (0.1 %),
no weevils were found in the third soil
layer (21-30 cm).

Table 1. 1nfection rate (%) of Andean patato weevil


by B. brongniartii (figures in parenthesis are the
number of infected weevils counted).
Cusca Huancaya Cajamarca
12.4
9.0
_,
lnfection rate in
the field
(26,809)
(4752)
Range between
8-35
2-21
units
15.2
lnfection rate
27.1
11.4
(16,636) (15,441)
(9208)
Range between
8-59
4-36
3-24
units
1 Data

were not collected in the field in Cajamarca.

Table 2. Susceptibility of Andean patato weevil to


B. brongniartii infection (%) at different life stages.
Cusca
Huancaya Cajamarca
Field
Larvae
71
81
Pupae
23
15
Adult
6
4
Sta re
92
Larvae
88
89
Pupae
12
7
8
o
o
o
Adult
1 Data were not collected in the field in Cajamarca.

CIP Program Report 1999 - 2000

157

lnfluence of temperature and moisture

120

Eight days after treatment, insect mortality


ranged from 100% at 20C to 4% at 5C
(Figure 1), indicating that the pathogenic
action of B. brongniartii is favored at
higher temperatures. Although the disease
developed very slowly at 5C, it continued
with time, reaching 76-90% mortality
three months after treatment.
No significant (P > 0.05) temperaturemoisture interactions were found in
relation to the mortality of fungus-infected
larvae. However, it did have a major
effect on the mycelial growth of the
fungus and the production of conidia. High
soil moisture (50% field capacity) favored
the development of mycelia outside the
infected insects' bodies, whereas low soil
moisture (25% field capacity) favored the
production of conidia.
The fungus survived in dried infected
larvae for a long time when kept at 12C
or lower. Mycelial growth was reactivated
in fungus-infected specimens kept for more
than a year when they were exposed to
moisture.

100

Soil moisture 1O %

O Soil moisture 25%


~ Soil moisture 50%

Temperature

Figur.e 1. Effect of temperature and soil moisture on


mortality of the Andean patato weevil larvae
(Premnotrypes suturical/us) by B. brongniartii, eight
days after inoculation, La Malina, 1999.

f;
-~

>.
Q)

50

.o

Techniques for mass multiplication


All three of the methods tested to produce
infected larva in large numbers were
effective. From 81 % to 84% of larvae
exposed to infected soil became infected,
but conidia-inoculated larvae were the
most efficient vehicle for infection,
resulting in an infection rate of 98%.
The most efficient level of hydration was
700 mi water/kg of barley medium. Higher
levels of hydration reduced the development of the fungus. lnoculation of 1 kg
of bar ley was equal ly effective with
1O, 20, or 30 infected larvae; therefore,
the most convenient (and economical)
method of inoculation would be to use 1O
infected larvae/kg of inoculation medium
(Figure 2).

Optimum application rate


Larvae of the Andean patato weevil were
effectively controlled by applying

158

Research on Potato

1 day

15 days

30 days

45 days

Days after inoculation

Figure 2. Effect of the use of 1O, 20, and 30 infected


B. brongniartii larvae as inoculum for the
multiplication of the fungus in moist barley.
B. brongniartii to the soil surface in rustic
potato stores (Figure 3). An applicatton
rate of 1 kg/m 2 or 2 kg/m 2 resulted in 89%
mortality; 3 kg/m 2 resulted in a mortality
of 93%. The lower dosage infected 67% of
the larvae in the same time period as the
higher dosages, but mortality increased
over time.

A second set of trials was carried out using


2 kg/m 2 of the barley medium in rustic
stores in Cusco, Huancayo, and
Cajamarca. Larval mortality was highly
variable among replicates in all three
places. In Cusca, the mortality of
P. latithorax ranged from 50-96%; in

Figure 3. Dead P. suturicallus larvae in a rustic patato


store where B. brongniartii inoculum had been
applied.
Huancayo, the mortality range of
P. suturicallus was 54-97%; and in
Cajamarca, the mortality of P. vorax
ranged from 63-92%.

Discussion
B. brongniartii is a widely distributed
fungus in the Peruvian Andes, infecting
larvae and other 1ife stages of the Andean
patato weevil, Premnotrypes spp. This
study found that the larval stage of the
Andean patato weevil is the life stage
most commonly infected by the fungus,
confirming earlier observations (Rojas,
1981; Vera, 1992). Tests also confirmed
that the larval is the stage most susceptible to the fungus, which produces more
conidia when grown in the larvae than in
other developmental stages of the weevil.
Most weevils sampled during the larval
stage were found in the upper 10 cm of
soil, as previously reported (Alczar, 1976;
Muoz, 1998). lt was also found that
B. brongniartii can remain viable for more
than one year under cool, dry conditions in
infected larvae filled with conidia. After a
year, the fungus activated mycelial growth
when moisture was provided.
Temperature and soil moisture have an
influence on the onset of the disease and
the development of the fungus, enhancing
either mycelial growth (moist conditions)
or conidial production (dry conditions).
Th is i nformation provides a basis for the
mass production, use, and conservation of

the fungus. Using infected larvae as the


source of inoculum for mass-production of
the fungus would salve a serious problem
for most farmers: lack of access to pure
laboratory-produced inoculum. The
method described here is appropriate for
cottage-scale production. Fungus-infested
barley medium produced this way resulted
in an inoculum as effective, or better, than
the traditional method of laboratoryproduced inoculum. These findings will
help to expand the production and use of
the fungus B. brongniartii in the control of
the Andean patato weevi 1 for smal 1
farmers.
Given the rather slow natural spread of the
fungus in the soil, but its long persistence
in infected larvae, the effectiveness of
applications should be evaluated over
more than one season.

References
Alczar, S.J. 1976. Biologa y
comportamiento del "gorgojo de los
Andes" Premnotrypes suturicallus
Kuschel (Coleop: Curculionidae).
Master's thesis. Universidad Nacional
del Centro del Per, Huancayo, Peru.

90 p.
Alczar, J., K.V. Raman, H. Torres, and
E. Yabar. 1 990. Beauveria spp. hongo
amigo del agricultor. Rev. Medio
Ambiente (Lima, Per) 45:44-46.
Alves, S. 1998. Controle microbiano de
insetos. Second edition. Funda<;ao de
Estudos Agrrios "Luiz de Queiroz"
(FEALQ), Piracicaba, Brazil. 1163 p.
Brady, B. 1979. Descriptions of pathogenic
fung and bacteria Beauveria
brongniartii and Beauveria bassiana.
Commonwealth Mycological lnstitute,
Surrey, UK. p. 602-603.
De Hoog, G.S. 1972. The genera
Beauveria/ Jsaria, Tritirachium. Study in
micology. Number 1. Kent, UK. 41 p.
Faull, J.L. 1986. Fung and their role in
crop protection. In: Biotechnology and
Crop lmprovement and Protection.
University of London, UK. p. 141-149.

CIP Program Report 1999 - 2000

159

Ferron, P. 1978. Biological control of


insect pests by entomogenous fungi.
Annual Review of Entomology
23:409-442.
Ferron, P. 1983. lnduction artificielle dune
epizootie a Beauveria brogniartii. Dans
une popu lation de Melolontha
melolontha. Symbiosis (France)
XV(2):75-83.
Fuxa, J. 1987. Ecological considerations
for the use of entomopathogens in
IMP. Annual Review of Entomology
32:225-51.
Fuxa, R., R. Gaugler, and J. Maddox.
1990. Use of pathogens in insect
control. In: CRC handbook of pest
management in agriculture. CRC Press,
Boca Raton, FL, USA.
Macleod, D.M. 1954. lnvestigations on the
Genera Beauveria Vuill. and
Tritirachium Limber. Canadian Journal
of Botany 32:818-890.
Muoz, M, J. Alczat, and V. Cerna. 1998.
Biologa y comportamiento del gorgojo

160

Research on Potato

de los Andes Premnotrypes vorax


Hustache (Col.: Curculionidae). In:
Summary of the 40th National
Convention of Entomology Held 8-1 2
Nov 1998. Sociedad Entomolgica del
Per, Univcersidad Nacional Agraria La
Molina, Lima, Peru. p. 34. (abstract)
Rojas, C.J. 1981. Estudio de la
suceptibilidad del Premnotrypes
suturicallus al hongo Beauveria sp.
Master's thesis. Universidad Nacional
del Centro del Per, Huancayo, Peru.
43 p.
Tanada,Y. and H. Kaya. 1993. lnsect
pathology. Academic Press, San Diego,
CA, USA. 666 p.
Vera, R.A. 1992. Patogenicidad de cinco
aislamientos del hongo Beauveria
brongniartii (Saccardo) Petch sobre los
estados de desarrollo del gorgojo de los
Andes Premnotrypes latithorax. Master' s
thesis. Universidad Nacional de San
Agustn de Arequipa, Cusca, Peru. 98 p.

Effectiveness of Abamectin and Plant-Oil Mixtures


on Eggs and Larvae of the Leafminer Fly, Liriomyza
huidobrensis Blanchard
N. Mujica1, M. Pravatiner2, and F. Cisneros1

The effectiveness of abamectin applied alone or mixed with plant oil on


leafminer fly (LMF), Liriomyza huidobrensis Blanchard, was evaluated on
bean plants under laboratory and greenhouse conditions. The addition of plant
oil to abamectin sprays to get a 1 % oil spray concentration increased the
effectiveness of the insecticide to the extent that the active ingredient of the
insecticide could be reduced by one-half to three-fourths of the normal
dosage (0.15%). Both eggs and fly larvae were affected. Laboratory tests
resulted in higher mortality than greenhouse tests at the same concentrations
of insecticide. The synergistic effect shown by the mixture of abamectin and
plant oil allows a reduction in the commercially recommended concentration
of abamectin without any loss in effectiveness. As a result, treatment costs
can be reduced by 60% and more farmers will be able to use abamectin to
control leafminer fly.

The leafminer fly (LMF), Liriomyza


huidobrensis Blanchard, is one of the most
harmful pests in the coastal valleys of
Peru, attacking a wide range of vegetable
crops and ornamental plants. lt is also a
problem in many other parts of the world
where it has been introduced in recent
years (Mujica and Cisneros, 1997). Adult
flies puncture the leaves of the plant for
feeding and oviposition, resulting in
stippling. The larvae reduce the photosynthetic activity of the leaves by mining; the
leaves dry out and the plant is defoliated
(Parrella et al., 1985). This species of
leafminer fly is tolerant to many insecticides and capable of developing a
resistance to products that were originally
effective. lt is therefore becoming more
difficult to control LMF worldwide.
1

CIP, Lima, Peru.


Universidad Nacional Agraria, La Molina, Lima, Peru.

Abamectin is made up of a mixture of


avermectins obtained through fermentation
of a soil actinomycete, Streptomyces
avermitilis Burg (Fisher and Mrozik, 1989).
In recent years, it has been considered an
outstanding chemical for controlling LMF.
lt is not only highly effective but it comes
from a biological source, which makes it a
bio-rational product that can be used in an
environmentally friendly manner in
integrated pest-management programs
(Dybas, 1989). Proven effective against a
number of important pests, such as mites,
ants, cockroaches, and selected pest
species of lepidoptera (Lasota and Dybas,
1991 ), abamectin is considered a selective
insecticide with relatively low toxicity to
many non-target arthropods (Dybas, 1989).
lt is used at low rates and degrades rapidly
(having a half-life of four to six hours)
when exposed to light, especially when
applied as a thin film on inert surfaces or

CIP Program Report 1999 - 2000

161

leaves (Wislocki et al., 1989). Despite its


rapid photodecomposition following
application, abamectin provides residual
activity in the field because of its
translaminar action and rapid penetration
of leaf tissue (Dybas, 1989). Although its
effectiveness has been proven, abamectin
is an expensive product, and management
of the dosage is an important economic
consideration. In February 2000, a greenhouse experiment and a lab test were done
to evaluate the effect of different levels of
abamectin and plant oil on the control of
LMF.

Materials and Methods


A total of 1O treatments were tested. Four
treatments corresponded to four dosages of
abamectin (18 g Al/liter of emulsifiable
concentrate, Vertimec 18 EC) alone: the
normal (commercially recommended)
spray concentration of 0.15% and three
reduced concentrations of 0.1125%,
0.075%, and 0.0375%, equivalent to 3/4,
1/2, and 1/4 the normal spray concentration, respectively. The other four
treatments consisted of the same four
concentrations of abamectin with the
addition of plant oil (93.3 g/liter plant oil,
Carrier) to get a 1 .0% oi 1 spray concentration. lt is well known that mineral oils
have miticide properties of their own, but
it was not known if oi 1 wou Id ha ve a
synergistic action against immature stages
of L. huidobrensis when mixed with
abamectin. There is anecdotal evidence
that mineral oil can be harmful to plants;
therefore, in order to avoid any risk of
phytotoxic effects, plant oi 1 was u sed
instead.

larvae are almost always found mining the


lower surfaces of leaves or within petioles.
This mining behavior is similar in both
bean and potato leaves. Young potted
bean plants were subjected to LMF
infestation by placing two-day-old adult
flies (20 flies/plant) on the plants for four
hours. During this short time, oviposition
took place, ensuring that the age of the
eggs and subsequent larvae would be
uniform. For egg treatment, infested bean
leaves were sprayed 24 h after infestation.
For larval treatments at instars 1, 11, and 111,
infested leaves were treated three, four,
and five days after infestation.
For evaluation purposes, a 30x manual
stereomicroscope was used to record the
number of eggs and larval instars under
greenhouse conditions. The bean plants
remained in the greenhouse after being
sprayed; sorne infested leaves that had
been treated were excised and taken to
the laboratory for detai led evaluation, and
records were made daily for the first five
days. An average of 60 eggs or larvae
were eval uated for each treatment. Egg
hatching was measured to evaluate egg
toxicity. Each treatment was replicated
three times. The data were analyzed with
a one-way analysis of variance (ANOVA);
means were separated using the WallerDuncan multiple range test. Ali
percentage data were transformed to an
arcsine square root of proportion before the
ANOVA. In all cases, non-transformed
values are reported. Treatment effects
were considered significant at P = 0.05.

Results
Greenhouse tests

There were two additional treatments: oil


alone at 1.0% and an absolute check,
which was not sprayed. The trials were
conducted at the entomology laboratory
and screenhouse faci 1ities of CIP headquarters in Lima.
For practica! purposes, bean plants were
selected because LMF host plants.
L. huidobrensis larvae frequently mine
along the midrib of leaves, and instar
162

Research on Patato

The normal dosage of abamectin produced


79.6% embryo mortality (Table 1). Egg
hatching failed when the embryo was full
grown and larval mandibles could be seen
through the egg chorion. In sorne cases,
the newly formed larva could only partially break the egg chorion. A similar
level of toxicity was obtained with 1/4 of
the normal dosage when plant oil was

added to the spray. The addition of oil to


the normal dosage of abamectin and to 3/4
and 1/2 of the normal dosage produced
more than 90% embryo mortality (Table
1). The effectiveness of the normal dosage
of abamectin upan larval instars was
somewhat lower than far eggs, but apart
from this, generally increased mortality
was obtained when oil was added to
abamectin. The synergistic effect of the
oil-abamectin mixture was greater at
lower abamectin dosages. Mortality was
twice as high when oil was added to 1/4
the normal dosage of abamectin. In leaves
treated with oi 1 alone, mortal ity was 12%
to 16% in larvae and 14% in eggs; in the
untreated control leaves, mortality was 5%
to 7% in larvae and 5% in eggs.

Laboratory tests
In general, the effects recorded far all
treatments in the laboratory followed the
same trend as those in the greenhouse
(Table 2, Figure 1 ). That is, egg and larval
mortal ity decreased as the dosage of
abamectin alone decreased. The addition
of plant oil increased the mortality far all
abamectin dosages, with a clear trend of a
higher rate of increased mortality far lower
dosages of abamectin (Table 2). Comparing treatment results obtained in the
laboratory and in the greenhouse, sorne
quantitative differences can be noted. In
general, mortality of eggs and larvae was
higher for each treatment in the laboratory
than in the greenhouse. The ratio of
toxicity to the amount of oil added was

Table 1. Effeet of abameetin alone and abameetin-oil mixtures on eggs and larvae of the leafminer fly,
Liriomyza huidobrensis in greenhouse test.
Mortality (%)
Dosage (%)
Egg
Larvae 11
Larvae
Oil
Larvae 1
Abamectin
b,
o
66.0
d
81.3
ed
57.2
T1
0.15
79.6
o
77.3
e
69.8
de
65.6
T2
0.1125
58.1
e
58.1
e
24.3
0.075
o
42.6
d
40.9
e
T3
f
21.0
f
10.3
o
23.8
27.6
T4
0.0375
e
97.1
a
85.7
94.0
a
97.6
a
T5
0.15
1
91.4
94.1
ab
89.5
89.3
b
T6
0.1125
1
a
87.0
be
76.5
70.6
ed
0.075
1
90.0
a
T7
79.2
72.7
67.5
ed
ed
75.2
b
T8
0.0375
1
g
fg
12.8
11.8
15.9
o
1
13.9
f
T9
g
g
h
7.1
5.2
5.8
o
o
4.8
T10

111
d
ed

e
fg
a
a
b
be
fg
g

1 Means within the same column, followed by a common letter do not differ significantly at P=0.05, according to

Waller-Duncan Multiple Range Test.

Table 2. Effeet of abameetin alone and abameetin-oil mixtures on eggs and larvae of the leafminer fly,
Liriomyza huidobrensis in laboratory test.
Dosage (%)
Mortality (%)
Abamectin
Oil
Egg
Larvae 1
Larvae 11
Larvae 111
T1
0.15
o
92.5
b,
100.0
a
100.0
a
100.0
a
T2
0.1125
O
87.5
b
96.6
b
100.0
a
100.0
a
T3
0.075
O
65.0
e
82.7
d
92.0
e
63.0
d
T4
0.0375
O
38.6
d
45.9
e
42.9
d
41.2
e
T5
0.15
1
100.0
a
100.0
a
100.0
a
100.0
a
T6
0.1125
1
100.0
a
100.0
a
97.8
b
100.0
a
T7
0.075
1
92.2
b
93.9
b
97.6
b
95.8
b
T8
0.0375
1
87.2
b
90.0
e
91.3
e
82.4
e
T9
O
1
O.O
e
23.9
f
7.9
e
16.3
f
no o
o
O.O
e
O.O
g
5.9
e
3.7
g
1 Means within the same column, followed by a common letter do not differ significantly at P-0.05, according to

Waller-Duncan Multiple Range Test.

elP Program Report 1999 - 2000

163

120
Laboratory

100

Greenhouse

--...........

...,

80

ast::::
o

...

_
-----~

60

Q)

40

_J

---- Abamectin
........_ Abamectin +ol

20

0.15

0.1125

0.075

0.15

0.0375

0.1125

0.075

0.0375

Dosages

Figure 1. Effect of abamectin and abamectin-oil mixtures on leafminer fly larvae, Liriomyza huidobrensis, under
laboratory and greenhouse conditions, Lima, Peru, 2000.
higher in the greenhouse than in the
laboratory tests (Figure 1). Finally, oil
alone had no effect on eggs but showed
sorne mortality (8% to 24%) at the larval
stage.

Discussion
Abamectin alone, applied at the commercially recommended dosage, showed a
satisfactory level of control of the eggs
and larvae of the leafminer fly, confirming
reports of severa! authors (Ochoa and
Carballo, 1993; Buxton and McDonald,
1994; Sotomayor, 1998). Lower dosages
were less effective, indicating that the
recommended dosage is close to
abamectin's action threshold. Abamectin's
effect on eggs apparently occurs when the
embryo is full-grown-when the sickleshaped mandibles of the newly formed
larvae are visible through the chorion of
treated eggs-as reported by Schuster and
Everett (1983) and Sotomayor (1998). In
many cases, larvae died during the process
of hatching, befare any damage to the
mesophyll tissue could be detected.
Apparently, abamecti n has no effect on

164

Research on Potato

the early stages of embryo development.


Larval mortality occurred at instars 1, 11,
111. In a few cases, the instar 111 larvae
were able to pupate but the pupae were
deformed, which has also been reported by
Leibee (1988). The larvae did not die
immediately after appl ication but about
three days later. A similar effect has been
reported by Sotomayor (1998), and Ochoa
and Carballo (1993).
Guerrero (1999) found that in potato the
addition of plant oi 1 raised the potency of
abamectin, allowing the dosage to be
reduced to 1/4 the recommended level.
These findings agree with reports on two
other pests: a plant mite, Tetranychus
urticae (Schuster and Everett, 1983), and
the cabbage diamond moth, Plutella
xylostella (Abro et al., 1988). As plant oil
alone has a negligible effect on LMF eggs
and larvae, the efficiency of the
abamectin-oil mixture is an example of
insecticide potentiation more than a
synergistic effect, although a synergistic
effect of mineral oil with abamectin on
larvae of Scrobipalpuloides absoluta has
been reported (Guedes et al., 1995).

Our results concur with those obtained


with mites, aphids, lepidoptera, and
whitefly (Wright et al., 1985; Anderson et
al., 1986; Horowitz et al., 1997), suggesting that oil enhances the translaminar
activity of abamectin, resulting in longer
residual activity. The plant oil probably
enhances the penetration of abamectin
into the leaf tissue. We expect that these
results would be the same on potato. Field
trials are needed to prove the efficacy of
this mixture under field conditions.

Conclusions
This research confirmed that when used on
bean plants, abamectin alone is effective
against the eggs and larvae of the
leafminer fly (L. huidobrensis) at the
commercially recommended dosages.
The addition of plant oil to abamectin
improved the potency of the insecticide,
also allowing the dosage to be reduced.
The cost of effectively treating 1 ha is
reduced by more than 60%-from $131 .30
for abamectin alone to $53.70 for the
mixture of abamectin and plant oilthereby making it available to small-scale
potato farmers with limited resources.

References
Abro, G.H., R.A. Dybas, A. Green, and
D.J. Wright. 1988. Toxicity of
avermectin Bl against a susceptible
laboratory strain and an insecticideresistant strain of Plutella xylostella
(Lepidoptera: Plutellidae). Journal of
Economic Entomology 81 :1575-1580.
Anderson, T.E., J.R. Babu, R.A. Dybas, and
H. Mehta. 1986. Avermectin Bl:
lngestion and contact toxicity against
Spodoptera eridania and Heliothis
virescens (Lepidoptera: Noctuidae) and
potentiation by oil and piperonyl
butoxide. Journal of Economic
Entomology 79:197-201.
Buxton, J.H. and O.C. McDonald. 1994.
Chemical control of the South American
leaf miner, Liriomyza huidobrensis. In:
Proceedings-Brighton Crop Protection

Conference, Pests and Diseases,


Brighton, UK, 21-24 November 1994,
Vol. 2. British Crop Protection Council,
BCPC Publications, Bracknell, UK.
p. 731-736.
Dybas, R.A. 1989. Abamectin use in crop
protection. In: Champbell, W.C. (ed.).
lvermectin and abamectin. Springer,
NY, USA. p. 287-31 O.
Fisher, M.H. and H. Mrozik. 1989.
Chemistry. In: Champbell, W.C. (ed.).
lvermectin and abamectin. Springer,
NY, USA. p. 1-23.
Guedes, P.N., M.C. Picaneo,
N.M. Guedes, and N. Medeira. 1995.
Synergism of mineral oil with
insecticide toxicity for

Scrobipalpuloides absoluta
(Lepidoptera: Gelechiidae). Pesq.
Agropecuaria Brasilia 30:313-318.
Guerrero H., E.O. 1999. Efecto de la
abemectina en mezcla con aceite
vegetal agrcola sobre la mosca
minadora Liriomyza huidobrensis
Blanchard (Diptera: Agromyzidae) en el
cultivo de papa. Thesis. Facultad de
Ciencias Agropecuarias, Alimentarias y
Pesqueras, Universidad Nacional Jos
Faustino Snchez Carrin, Huacho,
Peru. (unpublished)
Horowitz, A.R., Z. Mendelson, and
l. lshaaya. 1997. Effect of abamectin
mixed with mineral oil on the
sweetpotato wh itefly (Homoptera:
Aleyrodidae). Journal of Economic
Entomology 90:349-353.
Lasota, J.A. and R.A. Dybas. 1991.
Avermectins, a novel class of
compounds: lmplications for use in
arthropod pest control. Annual Review
of Entomology 36:91-117.
Leibee, G.L. 1988. Toxicity of abamectin
to Liriomyza trifo/ii (Burgess) (Diptera:
Agromyzidae). Journal of Economic
Entomology 81 :738-40
Mujica, N. and F. Cisneros. 1997.
Developing IPM components for
leafminer fly in the Caete Valley of
Peru. In: Program Report 1995-96.

lnternational Patato Center, Lima, Peru.


p. 177-184.

CIP Program Report 1999 - 2000

165

Ochoa, P. and M. Carballo. 1993. Effect of


various insecticides on Liriomyza
huidobrensis (Dptera: Agromyzidae)
and its parasitoid Diglyphus isaea
Walker (Hymenoptera: Eulophidae).
Manejo Integrado de Plagas (Costa
Rica) 26:8-12.
Parrella, M.P., V.P. Jones, R.R. Youngman,
and L.M. Lebeck. 1985. Effect of leaf
mining and leaf stippling of Liriomyza
spp. on photosynthetic rates of
chrysanthemum. Annals Entomological
Society of America 78:90-93.
Schuster, D.J. and P.H. Everett. 1983.
Response of Liriomyza trifolii (Dptera:
Agromyzidae) to insecticides on
tomato. Journal of Economic
Entomology 76:1170-76.
Sotomayor, M.P. 1998. Efectividad de la
abamectina sobre los estados de

166

Research on Potato

desarrollo de Liriomyza huidobrensis


Blanchard (Dip. Agromyzidae) "la
mosca minadora" y sus parasitoides
Ha/ticoptera arduine (Walker) (Hym.
Pteromal idae) y Diglyphus websteri
(Crawford) (Hym. Eulophidae). Master's
thesis. Universidad Nacional Agraria la
Molina, Lima, Peru. 234 p.
Wislocki, P.G., L.S. Frosso, and
R.A. Dybas. 1989. Environmental
aspects of abamecti n use in crop
protection. In: Champbell, W.C. (ed.).
lvermectin and abamectin. Springer,
NY, USA. p. 182-200.
Wright, D.J., A. Loy, A. Green, and
R.A. Dybas. 1985. The translaminar
activity of abamectin (MK-936) against
mites and aphids. Mededelingen van de
Faculteit Landbouwwetenschappen,
Rijksuniversiteit Gent. 50:595-601.

Establishment of Microsatellite Assays for Potato


Genetic ldentification
M Ghislain 1 , F Rodrguez1, F Villamn 1' 2 , J Nez 1 , R Waugh 3 , and M Bonierbale1

The in-trust potato collection conserved at CIP has been characterized with a
variety of molecular markers to reduce genetic redundancies and identify
subsets important for both conservation and crop improvement. We have
recently been concerned with developing a potato (Solanum spp.) genetic
identification (PGI) kit for curatorial purposes. More than 70 microsatellite
markers were tested on a genetically diverse sample of potato germplasm.
We have selected a subset of the 18 most informative microsatellites according to the quality of the DNA products they amplify, their genome location,
and the level of polymorphism they detect. These assays have shown that the
microsatellites selected are ubiquitous in all seven cultivated potato species.
The PGI kit has also been shown to be useful to differentiate potato varieties
and cultivars. We have further used this marker set to identify genotype
ambiguities in the ex situ germplasm collection. The PG 1 kit will be routinely
used to produce a true-to-type reference for each new entry in our
germplasm collection and to attempt to minimize human error in germplasm
management.
The potato collection held in trust at
the lnternational Potato Center (CIP) is
mainly composed of cultivated accessions
accounting for up to 70% of the 5,094
clonal and true seed entries (Table 1, after
Huamn et al., 1997). Clonal accessions
of cultivated potato are maintained in the
field at CIP Huancayo and in vitro at La
Molina, Peru. Subsets of the collection
maintained in vitro are introduced to the
field regularly not only to monitor reliability of conservation procedures but to
provide new morphological data or trait
evaluation useful to geneticists and
breeders. Therefore, the correct identity of
each accession maintained in vitro is
highly relevant to both curators and
breeders at CIP.
1

CIP, Lima, Peru.


Current address: UniversityofWisconsin, Madison, Wl, USA.
3
Scottish Crop Research lnstitute, lnvergowrie, Dundee, UK.

This ex situ collection has been extensively characterized for morphological


descriptors and isozyme patterns (Huamn
et al., 2000), and in the last 5 years, with
ONA markers including random amplified
polymorphic ONA (RAPO) (Ghislain et al.,
1999), amplified fragment length polymorphism (AFLP), and microsatellites. The
information provided by each of these
assays has been limited to questions of
identity and distinction with the exception
of isozymes and microsatellites. RAPO and
AFLP are dominant markers that will help
to determine genetic distance in terms of
providing estimates of genetic differences
rather than of genetic similarities. The
advantage of codominant markers, such as
microsatellites, is the possibility to display
all four alleles at a single locus. Hence by
their nature, microsatellites are more

CIP Program Report 1999 - 2000

167

Table 1. Patato (Solanum spp.) accessions


using microsatellite markers.
Species
S. stenotomum (stn)
S. phureja (phu)
S. goniocalyx (gon)
S. x ajanhuiri (ajh)
Solanum 2x hybrids
S. x chaucha (cha)
S. juzepczukii (juz)
S. tuberosum subsp. andigena (adg)
S. tuberosum subsp. tuberosum (tbr)
So/anum 4x hybrids
S. x curtilobum (cur)
Total cu ltivated
Wild species
Grand total

conserved in the gene bank at CIP and currently genotyped


Ploidy
2n = 2x = 24
2n = 2x = 24
2n = 2x = 24
2n = 2x = 24
2n = 2x = 24
2n = 3x = 36
2n = 3x = 36
2n = 4x = 48
2n = 4x = 48
2n = 4x = 48
2n = 5x = 60

infarmative to establish genetic relatedness between germplasm pools.


Microsatellites are tandemly repeated
segments of 1-5 nucleotides whose repeat
number reveals genetic differences
between individuals (Tautz and Renz,
1984). DNA containing such microsatellite
sequences can be amplified using polymerase chain reaction (PCR) primers that
flank the repeat motif, and then can be
easily resolved by gel electrophoresis.
They involve a relatively expensive
'discovery process' of identific,ation and
preparation based on DNA cloning and
sequencing. But, they are relatively
inexpensive to use once available, and are
more infarmative than other single locus
markers. Moreover, the ease of assaying
sets of simple sequence repeat (SSR)
markers, especially through multiplex,
automated analysis makes them the
preferred marker far characterization and
fingerprinting of germplasm of nearly any
organism (Powell et al., 1996). They are
codominant and more amenable than
other types of markers to comparative
studies across populations, or in different
laboratories. All infarmation far their use
can be made publicly available in the
form of DNA sequence infarmation only.
SSRs do not require extensive laboratory
maintenance and the method is the most

168

Research on Patato

Accessions (no.)
268
170
48

Genotyped (no.)

6
21
8

10
56

97

8
4
13

31
2644
144
48
11
3527
1567

o
5

o
4
73

5094
economical to exchange. SSR alleles or
genotypes can be registered in molecular
marker databases, or with regu latory
agencies, to permit the development of
cumulative infarmation on varietal
fingerprints. Such a tool is needed to
complement or replace current methods of
variety description, which rely on environmentally sensitive morphological
characteristics. Microsatellite technology
is also an appropriate technology far
developing country laboratories and
breeding programs, which often have
modest expertise in molecular biology
techniques.
However, microsatellites are not yet the
perfect genetic marker because they entail
both limitations and assumptions that
cannot always be met. Limitations include
their high discovery cost, estimated to be
around US$300/microsatellite, and their
relatively high species-specificity as
opposed to RAPO and AFLP (Westman and
Kresovich, 1998). Microsatel 1ite use also
involves several assumptions. The first of
these is that band sharing indicates allele
identity. Although this has been shown to
be true in cases where the bands have
been sequenced, we cannot rule out that
single-nucleotide substitutions are present
between two comigrating bands. Another
case of incorrect allelic identity of

comigrating bands is due to the forward. backward steps during DNA polymerase
'slippage' that is responsible for microsatell ite al lele generation (Kruglyak et al.,
1998). Microsatellite alleles can thereby
revert to an original identical size for
genotypes that are in reality distinct. The
other important assumption is that the
presence of a band corresponds to one
copy of an allele. lt may actually correspond to more copies of the same alleles
but we cannot ascertain this. The ampl ification of microsatellites cannot be made
quantitative because it would be impractical to have to optimize each for
quantitative PCR. Probably only a number
of microsatellites would be amenable to
quantification of allele dosage, but such
use would have to include a new assumption of no-allelic preference in PCR
amplification. Hence, it seems unworkable for developing a procedure that
would permit the detection of different
dosages.
Molecular fingerprints will likely become
an essential tool in the inter-institutional
exchange of elite materials to assure the
identity of materials in international
testing programs, whether or not they are
protected by plant breeders' rights. They
will also add critica! resolution to patato
pedigrees, allowing the tracing or monitoring of successful alleles and allele
combinations in breeding programs where
only the fittest survive. Such historical
information as allelic constitution of
patato pedigrees is presently not readily
available but is potentially very valuable
for identifying successful progenitors of
important characteristics in multitrait
selection programs. Finally, molecular
pedigrees are well suited to graphical
representation such as genomic displays
and overlays. This will likely facilitate
their use. A valuable potential application
of SSR fingerprints of gene bank holdings
is the abi 1ity to query new entries for
unique diversity, along with or befare their
acquisition into already large and expensive ex situ col lections.

An additional advantage of microsatell ites


over dominant marker systems in potato is
their capacity to reflect the ploidy of the
genotypes assayed. In fact, the number of
distinct alleles detected for a single locus
may reach the ploidy leve! in cases of
maximum heterozygosity, but may not
exceed it.
Potato microsatellites were originally
developed at the Scottish Crop Research
lnstitute (Provan et al., 1996; Milbourne et
al., 1998). Originally, the primer sequences available today were designed
from DNA sequences of S. tuberosum
subsp. tuberosum. To develop a general
tool for PGI, we assessed their usefulness
in seven other species, four for each
species, and in the other subspecies of
patato, S. tuberosum subsp. andigena.
We al so show here severa! appl ications
relevant to germplasm management
procedures and crop improvement. The
PGI kit will be distributed internationally
for use by a variety of users. Among them,
germplasm curators are likely to be the
major beneficiaries.

Material and Methods


Plant materials and DNA extraction
Seventy-three clona! accessions from the
patato germplasm col lection were propagated in vitro. Each taxonomic group was
represented by four accessions chosen for
diversity. The accessions of the native
patato group were chosen based on their
relative importance in the Andean region.
Total DNA was extracted using standard
protocols adapted at CIP (CIP, 1997). DNA
quality and quantity were estimated by
comparison with CsCl-purified 1 DNA
(Gibco-BRL, Life Technologies,
www.lifetech) digested with Pstl on
ethidium bromide agarose gels.

Microsatellite primers and reaction


condition
Microsatellite sequences were identified
through potato database searches (Provan
et al., 1996), and enriched genomic

CIP Program Report 1999 - 2000

169

libraries (Milbourne et al., 1998). Primer


sequences for an original set of 1 08
microsatellites were obtained from information from Milbourne et al., 1998.
Primers were synthesized by commercial
suppliers (Gibco-BRL and Genset,
www.genxy.com). These primer sequences
and their respective amplification conditions are avai lable u pon request from
cip-mbl@cgiar.org.

Microsatellite allele identification


Amplification products were separated on
6% denaturing polyacrylamide gels run
using a gradient electrophoresis buffer by
1 x and O.Sx for anode and cathode tank,
respectively (52 sequencing system,
Gibco-BRL). Gels were fixed to one glass
plate using the Bind-Silane chemical
(Promega, www.promega.com). Gels were
stained using a modified silver staining
procedure derived from the Promega kit.
Allele size was determined by coincidence of gel-mobility between band
ladder of sequencing reactions (pUC1 8forward primer) and either the upper band
(when two were visible) or the most
intense one (in the case of stuttered
bands). Allele presence was scored as 1;
al lele absence was scored as O.

DNA hybridization of microsatellite


amplification products
Southern blot hybridization was conducted
using the amplified microsatellites from
S. tuberosum subsp. tuberosum as probes.
These were isolated from agarose gels
containing the DNA by centrifuging at
13,000 rpm for 20 min in an Eppendorf
microcentrifuge using SpinX tube
(Costar, www.corning.com). The protocols
for Southern blot and DNA detection were
those provided with the ECL kit
(Amersham, www.amersham.co.uk).

Data analysis
The polymorphic index content for each
microsatellite on the germplasm sample
was calculated according to the Nei
statistic (1973): PIC=1-L(p 2 ), where pi is
the frequency of the ith allele detected.
170

Research on Patato

Results
Transferability of microsatellites between
cultivated potato species
An original group of 70 microsatellites was
tested using a sample of cu ltivated potato
accessions comprising at least four genotypes from each of the nine taxonomic
groups. Ali microsatellite primers amplified successful ly (Table 1). A first positive
indication that we were amplifying
homologous microsatellite loci with those
of S. tuberosum subsp. tuberosum carne
from the comparison of al lele phenotypes.
As a matter of fact, each microsatellite
displays a specific al lele phenotype on
denaturing polyacrylamide gels, such as
discrete bands, double or triple bands, or
stutter bands. These phenotypes were
clearly conserved among the nine groups
tested, supporting the analogous nature of
the microsatellite locus. Southern blot was
performed for 1 3 m icrosatel 1ites to confirm
sequence homology among amplification
products (data not shown). Both of these
analyses are concordant with the conservation of microsatellite loci across
cultivated potato species.

ldentification of a microsatellite set


Of the 70 m icrosatel 1ites tested, 1 8 were
found most useful for germplasm fingerprinting. Poor quality of products from
the remaining sequences was due to
various factors su ch as u nspecific primer
sequence, stutter bands, background
amplification products, and locusoverlapping. Comparison of agarose gel
(3%) and polyacrylamide gel electrophoresis revealed that allele resolution can be
as little as 1O bp in agarose gels (Figure 1).
Hence, for discrimination, a 3% agarose
gel electrophoresis will in sorne cases
provide enough resolution to revea!
genetic distinctness (estimated at around
60% of the cases in the potato sample
tested in this study). For genetic identification, complete resolution is needed as well
as accurate allele size determination.
Hence, we have included in our protocol a

1 bp molecular weight marker (pUC18


DNA sequencing reaction). Polymorphic
index content (PIC) values were calculated for 18 microsatellites using 73
accessions of the 9 cultivated taxa.
Individual PIC values ranged from 0.324 to
0.882 (Table 2).

'l::t
'l::t
,...
N
co N" (W') co
" Lt)oen(W') N"Nco Lt)"(O(W') NooLt)
(W')
Lt')
Lt)
o Lt)
N
(W')
Lt)
Lt)
(W')
bp ,...
o o o o "
o o o o o o
Lt')

(W')

" " " " " " " " " "

240

~:~

Ploidy prediction by microsatellites


The reference microsatel 1ites were also
used to suggest possible misclassification
in the germplasm collection. This was
based on the observation of inordinately
high allele numbers for single locus
markers in putative diploid accessions. As
an example, we identified 12 accessions
classified as S. phureja with higher allele
numbers than expected for a diploid
species (examples are shown in Figure 1).
Each of these accessions was confirmed
not to be diploid by chromosome counts.
The microsatel lites that can be used for
ploidy testing are STPAc58, STM0037,
STM0031, STM301 2, and STM2030. Each
represents a single locus in all mapping

----:
~

241

199
187
184

Figure 1. Amplification products of 1Ocultivated


patato clones (S. phureja) with microsatellite
STM2022 resolved on a 3% agarose gel
electrophoresis (upper panel) and 6% polyacrylamide
gel electrophoresis under denaturing conditions with
7 M urea (lower panel). * lndicates suspected
misclassified clones.

Table 2. Microsatellite set used to fingerprint 73 native patato cultivars including polymorphic index content
(PIC) per marker, and location on the patato genetic map.
Chromosome Annealing Genotypes Alleles PIC
temperature (no.) detected
("C)
(no.)

SCRl1
code

Repeat motif

STM1049
STM2030
STM1064
STM2022
STM1053
STM1058
STM3023
STM1031
STPAc58
STM0019
STM0031
STM2013
STM1104
STM1017
STM3012
STM1106
STM0037
STM0030

(ATA) 6
(CA)3 (TA) 5
(TA) 12 (TG) 4 GT (TG) 5
(CAA)J ... (CAA)J
(TA) 4 (ATC) 5
(ATT) 5
(GA) 9 ... (GA) 8 ... (GA) 4
(AT)13
(TA) 13
(AT)] (GT) 10 (AT) 4 (GT) 5 (GC) 4 (GT) 4
(AC) 5 ... (AC)J (GCAC) (AC) 2 (GCAC) 2
(TCTA) 6
(TCT) 5
(ATT) 5
(CT) 4 ... (CT) 8
(ATT)13
(TC) 5 (AC) 6 AA (ACh (AT) 4
Compound (GT/GC) (GT) 8

1
1
11
11
111
111

IV
V
V
VI
VII
VII
VIII
IX
IX
X
XI
XII

57
55
55
53
55
55
50
55
57
47
57
55
57
53
57
55
53
53

73
73
73
73
22
22
22
73
73
73
73
73
22
73
73
22
73
73

7
5
6
5
4
4
5
8
10
29
10
13
7
3
6
6
10
13

0.555
0.447
0.581
0.665
0.670
0.324
0.689
0.591
0.802
0.882
0.737
0.815
0.798
0.483
0.645
0.749
0.784
0.879

Allele
size
(bp)2

184-254
180-209
188-199
184-241
168-177
113-125
177-201
265-325
231-277
83-239
155-205
146-172
168-183
132-136
168-213
142-196
75-99
122-188

1 SCRI =
2 bp =

Scottish Crop Research lnstitute.


base pair.
CIP Program Report 1999 - 2000

171

populations tested so far, and has consistently revealed higher ploidy levels when
more al le les than expected were observed.

Application of the PGI kit to fingerprinting


native potato varieties
One of the objectives of the development
of molecular descriptors for genotype
identification is to provide a solid and
indisputable system to catalogue potato
varieties and cultivars for recognition and
tracking. We have begun to develop
genotyping fingerprints for a set of 22
cultivars of native potatoes. These cultivars encompass six species of the genus
Solanum and three levels of ploidy (2x,
3x, and 4x). This analysis revealed that
with as few as 2 microsatellites (STM0019,
STM1104), 20 cultivars could be unambiguously distinguished. Two pairs of
cu ltivars cou Id not be separated either
with these two microsatellites or with the
remaining 16, leading to the suspicion that
they may be duplicate accessions. However, a fingerprint using AFLP markers (6
primer pairs) resolved the question,
demonstrating close relationship but
distinctness. The 22 cultivars were ana. lyzed for genetic relatedness based on the
data set derived from 18 microsatellites
(122 polymorphic alleles) using the
unweighted pair-group method with
arithmetic mean (UPGMA) with simple
matching coefficient (Sneath and Sokal,
1973). As expected by their taxonomic
relationships, a cluster of the diploid and
triploid species is observed to separate
from the adg and tbr clones (Figure 2).

Application of the PGI kit to CIP breeding


clones
We have applied the PGI kit (18 microsatellite markers) to advanced clones and
progenitors derived from the CIP late
blight program. The material consisted of
30 advanced clones from Population
B3C1, 8 from Population A, 26 parental
genotypes from Population B3 grown in
vitro, and 1 7 grown in greenhouses, of
which 16 were duplicated from the in vitro

172

Research on Potato

stocks. Additionally, we used nine genotypes as interna! controls. The


microsatellite profiling produced a unique
genetic description of each B3C1 clone.
Of the 18 used, only 4 are needed to
discriminate all 30 clones. Hence, these
DNA fingerprints are useful and practica!
tools for the identification of these potato
clones, which will be distributed to severa!
countries for adaptation trials and varietal
selection.

Discussion
Potato genetic resources represent a large
pool of valuable alleles for a wide range
of important traits. Samples of these
genetic resources are maintained on-farm
and ex-situ in germplasm repositories. Both
conservation strategies have advantages
and disadvantages, but share the difficulty
of maintaining genetic resources without
losing or mixing sorne samples over time.
Such events do occur no matter how strict
the management procedures. The possibility of detecting these and resolving
ambiguities often relies on combinations
of the following sources of information:
(1) the original data of the collector,
(2) herbarium specimens, (3) records of
morphological evaluations, or (4) samples
kept by a third party. Therefore, genetic
fingerprints based on DNA markers have
been proposed as excel lent true-to-type
references. Among all the DNA markers
available, microsatellites provide distinct
advantages for patato genetic identification due to their high genetic information
content, high reproducibility, and simplicity of use.
Seventy microsatellites were tested, which
led us to identify 1 8 that are proposed here
as a potato genetic identification kit.
These microsatellites were selected on the
basis of the quality of the amplified
product they produce, their correspondence to single loci, their PIC, and the
benefits of using at least one marker/
patato chromosome. Since cross-species
amplification may lead to false positive
resu lts, we ha ve tested these

Hualash (adg) 4x

1
1

Yana Puma Maqui (adg) 4x


1
1

Ccompis (adg) 4x
lmilla Blanca (adg) 4x
Sani lmilla (adg)4x

Huaycha Pacea (adg) 4x


lshcupuru (stn) 2x
1

Camotilla (gon) 2x

Criolla Yema de Huevo (phu) 2x

--

Amarilla Tumba y (gon) 2x


~

Amarilla del Centro (gon) 2x


-

Huanquita (stn)2x
Puca Huayro (cha) 3x

Muru Pia (cha) 3x

Puca Muru Rucma (cha)3x


Puca Pia (cha) 3x
Peruanita (gon) 2x

Huagalina (adg)4x
Yana Suli (adg) 4x
Yana lmilla (adg) 4x
Papa Chanca (tbr)4x
Guincho Negra (adg) 4x

0.60

0.65

0.70

0.75
0.85
0.80
Similarity Coefficient

0.90

0.95

1.00

Figure 2. Dendrogram of the 22 native patato cultivars produced by genotyping with 18 microsatellites. Ploidy is
indicated next to each name: adg =S. tuberosum subsp. andigena, stn =S. stenotomum, gon =S.
goniocalyx, phu =S. phureja, cha= S. x chauca, tbr =S. tuberosum subsp. tuberosum.
microsatellites in a small sample of the
eight cultivated species of patato. Allele
phenotype and DNA:DNA hybridization
revealed homology between the
microsatellite amplification product
between species. Therefare, we consider
that these loci are conserved among the
patato species tested.
This PGI kit will have many uses. We have
shown here two examples of its immediate
use. Twenty-two native patato cultivars
were characterized molecularly. Unique
DNA fingerprint were obtained far all but
two pairs of native patato cultivars that are
closely related. Cultivars derived from
diploid and triploid species farmed a

cluster within a dendrogram in which 5.


tuberosum subsp. andigena was more
distant. These fingerprints have been
registered in a microsatellite database and
will be available as an additional descriptor far each accession of the patato
collection. The PGI kit was used to register
genetic fingerprints of advanced breeding
clones that will be useful as references to
original germplasm through a series of
international evaluations and variety
development activities. In the course of
this activity we faund a relatively high
frequency of non parental al leles. Su ch
results evidence certain weaknesses in
either pollination or data management.

CIP Program Report 1999 - 2000

173

Microsatellites will also be useful to


establish genetic relationships and to
monitor regions of the genome introgressed
into cultivated material. They will also be
used to trace past, current, and future gene
flow worldwide. The domestication of the
potato centuries ago in the Andes was a
complex gene flow process likely starting
from inter- and intra-species hybridization
of diploid cultivated 5. stenotomum with
wild species, such as 5. sparsipilum, or
autoploidization.

Acknowledgments
The authors are grateful to Z. Huamn,
W. Roca, R. Gomez, A. Panta, and
J. Toledo for providing sources of plant
material. We also thank severa! technicians who have contributed to the
optimization of these protocols. This work
was initially made possible through a grant
from the United Nations Oevelopment
Programme, Special Project Program, and
has been partially supported through a
grant from the Bundesministerium fr
Wirtschaftliche Zusammenarbeit und
Entwicklung (BMZ) and Oeutsche
Gesellschaft fr Technische
Zusammenarbeit (GTZ) (project no

96.7860.8-001 .00).

References
CIP (lnternational Patato Center). 1997.
Ghislain, M., O. Zhang, and
M.R. Herrera (eds.). 1997. Molecular
biology laboratory protocols. Plant
genotyping. Genetic Resources
Department training manual.
lnternational Potato Center, Lima, Peru.
Ghislain, M., O. Zhang, D. Fajardo,
Z. Huamn, and R. Hijmans. 1999.
Marker-assisted sampling of the
cu ltivated Andean potato 5olanum
phureja collection using RAPO markers.
Genetic Resources and Crop Evolution

46(6):547-555.
Huamn, Z., A. Golmirzaie, and
W. Amoros. 1997. The potato. In:
Fuccillo, O., L. Sears, and P. Stapleton

174

Research on Potato

(eds.). Biodiversity in trust. Cambridge


University Press, Cambridge, UK.

p. 21-28.
Huamn, Z., R. Ortiz, O. Zhang, and
F. Rodrguez. 2000. lsozyme analysis of
entire and core collections of 5olanum
tuberosum subsp. andigena potato
cultivars. Crop Science 40:273-276.
Kruglyak, S., R.T. Ourret, M.O. Schug, and
C.F. Aquadro. 1998. Equilibrium
distributions of microsatellite repeat
length resulting from a balance between
slippage events and point mutations.
Proceedings of the National Academy
of Science 95:10,774-10,778.
Milbourne, O., R.C. Meyer, A.J. Collins,
L.D. Ramsay, C. Gebhardt, and
R. Waugh. 1998. lsolation, characterization and mapping of simple
sequence repeat loci in potato.
Molecular Genetics 259:233-245.Powell, W., G.C. Machray, and J. Provan.
1996. Polymorphism revealed by simple
sequence repeats. Trends in Plant
Science 1 :215-222.
Provan, J., W. Powell, and R. Waugh.
1996. Microsatellite analysis of
relationships within cultivated patato
(Solanum tuberosum). Theoretical
Applied Genetics 92:1,078-1,084.
Nei, M. 1973. Analysis of gene diversity in
subdivided populations. Proceedings of
the National Academy of Science USA

70:3,321-3,323.
Sneath, P.H.A. and R.R. Sokal. 1973.
Numerical taxonomy: The principies
and practice of numerical classification.
W.H. Freeman, San Francisco, CA,
USA.
Tautz, O. and M. Renz. 1984. Simple
sequences are ubiquitous repetitive
components of eukaryotic genomes.
Nucleic Acid Research 12:4, 127-4, 128.
Westman, A.L. and S. Kresovich. 1998.
The potential of cross-taxa simplesequence repeat (SSR) amplification
between Arabidopsis thaliana L. and
crop brassicas. Theoretical Appl ied
Genetics 96:272-281.

Clonal True-to-Type Verification of Potato Accessions


Retrieved from In Vitro Conservation and
Cryopreservation
G. Perazzo1, A. Panta 1 , F. Rodriguez1, R. Gomez1, J. Toledo1, Z. Huamn 1, 2,
M. Ghislain1, A.M. Golmirzaie 1 ' 3 , and W. Roca 1

Twenty potato (Solanum tuberosum L.) accessions were retrieved from three
to five years (without subculturing) in vitro storage, and 18 from 1-year
cryopreservation. Using 23 morphological descriptors, all accessions were
compared with original clones to verify if they were true-to-type. Materials
from in vitro storage were compared with clones maintained in the field, and
those from cryopreservation with clones maintained in vitro. Comparison of
cryopreserved materials was made using three types of propagules (plantlets,
stem cuttings, and tubers). Additionally, accessions from in vitro storage were
compared with samples maintained in the field using 14 highly polymorphic
potato microsatellite markers. Six accessions (four from in vitro storage and
two from cryopreservation) showed differences in multiple morphological
characters, suggesting a case of misidentification. In the remaining accessions, differences were observed for 1O descriptors of flowering degree and
color expression. Fewer accessions showing differences (four) were found
when tubers were used as planting material. Simple sequence repeats (SSR)
analysis of 20 accessions, retrieved from in vitro storage and compared with
original clones, resulted in five accessions showing remarkable DNA differences, again revealing possible misidentifications. Four were found different
from the original clones by morphological and molecular analyses; and, one
accession was found different at the molecular level, even though it showed
no morphological differences.

Plant genetic resource conservation at CIP


(Golmirzaie and Panta, 1997a) focuses on
complementary in vitro approaches, with
linkages to in situ conservation. The
former uses tissue culture and cryopreservation techniques, as well as

CIP, Lima, Peru.


Present address: ProBioAndes, Lima, Peru.
3
Present address: Department of Horticulture, University of
Arkansas, Fayetteville, AR, USA.

molecular genetic characterization of


clonal accessions.
In vitro conservation has been applied at
CIP since the early 1970s. Since then, the
most comprehensive col lections of potato,
sweetpotato Upomoea batatas (L.) Lam.),
and other Andean root and tuber crops
have been assembled in CIP gene banks.
Currently, 16,604 accessions compose the
long-term collections held in trust by CIP

CIP Program Report 1999 - 2000

175

for the world community. CIP has constantly sought to upgrade its conservation
methods to minimize genetic changes in
the clones conserved. Growth retardation
and reduced temperature are the two
factors successful ly used for in vitro patato
conservation, extending the culture
transfer period to 3-4 years (Golmirzaie
and Toledo, 1999). Cryopreservation holds
a great potential for the long-term conservation of patato genetic resources (Benson
et al., 1996; Golmirzaie et al., 1999;
Schafer-Menurh et al., 1996; Steponkus et
al., 1992). Plant material can be stored
indefinitely at ultra low temperature (e.g.,
in liquid nitrogen), minimizing risk of
genetic changes in cells because, in
theory, chemical and physiological
changes are prevented. We have been
testing this approach for the past five
years, taking advantage of an experimental collection including nine species with
over 300 accessions.
Maintenance of genetic integrity is an
important requirement in gene bank
management. Therefore, assessment of
genetic integrity in plant propagation has
been conducted using morphological
descriptors. More recently, the application
of molecular markers has provided additional insight into the detection and
measurement of genetic variation (Kumar
et al., 1999).

shown their usefulness in the genetic


characterization of plant material after
long-term storage (Angel et al., 1996;
Borner et al., 2000; Piola et al., 1999;
Gupta and Varshney, 1999; and Harding,
1994).
In this study, patato clones retrieved from
in vitro and cryopreservation were assessed for trueness-to-type using
morphological and molecular analyses.
Clones retrieved from in vitro conservation
were compared with the original clones
maintained in the field; cryopreserved
clones were compared with those maintained in vitro.

Materials and Methods


Plant material
Twenty patato accessions maintained in
vitro for 3-5 years without subculture, and
1 8 accessions cryopreserved for one year
were randomly selected for this study.
They belong to the patato in vitro base,
comprising 5547 accessions and the
experimental cryopreserved collection of
300 accessions. Selected accessions are
from the following patato species:
S. stenotomum, S. tuberosum subsp.

andigena, S. goniocalyx, S. chaucha,


S. andigena x S. tuberosum, S. phureja,
and 5. curtilobum.
Morphological comparison

Morphological analysis provides a valuable approach for the description of


genetic stability when phenotypic variation is demonstrably linked to genetic
variation. The expression of morphological
characters, however, can be subject to
environmental or physical effects. This and
other limitations have led to research on
methods of analysis based on more stable
features of the genome (Potter and Janes,
1991 ). Molecular markers, such as restriction fragment length polymorphism (RFLP),
random amplified polymorphic DNA
(RAPO) and microsatellites, also known as
single sequence repeats (SSRs), have

176

Research on Patato

Accessions from in vitro conservation were


compared with control clones maintained
in the field. Plantlets were propagated in
vitro, hardened off in a growth chamber,
propagated in the greenhouse, and then
transferred to the field at the CIP Experiment Station in Huancayo (3200 m). Four
plants/treatment were planted in the field
under a completely randomized design, no
restrictions were applied to the experimental units. Three of them were used in the
study. Twenty-three morphological qual itative descriptors (Huamn et al., 1977)
were used to characterize the plant
material.

Cryopreservation was carried out accordi ng to Golmirzaie and Panta (1997b). Forty
shoot tips per accession were thawed after
one year of storage in liquid nitrogen.
From them, three clonal lines (C1, C2, and
C3) were used for morphological comparisons with original clones maintained in
vitro. Three sets of planting material
consisting of (1) plantlets grown in Jiffy-7
42 mm peat pellets (Jiffy Products of
America, lnc. Batavia, IL, USA), (2) stem
cuttings rooted in Jiffy-7 42 mm peat
pel lets, and (3) tubers obtained from the
field plants propagated in vitro were used
to establish field plantings. Ten plants/
treatment (accession/conservation method/
planting material) were planted in the
field; three were used for the morphological evaluation. All comparisons were
made under the same environmental
conditions.

Morphological data analysis


Morphological data were recorded from
three plants per treatment, using a O to 9
scale for the 23 selected potato descriptors. Of these, 10 were related to tuber
characters, 3 to plant habit and stems, and
10 to flowers. Corolla, tuber skin, and flesh
colors were evaluated using two color
charts that combined color and intensity
(Huamn et al., 1977).
Morphological descriptors were evaluated
at flowering and after harvest. The mode
of data recorded from three plants of each
treatment was analyzed with the numerical taxonomical method using the
NTSYSpc-2.02 software (Exeter Software,
Setauket, NY, USA). Average taxonomic
distance coefficient was used for similarity
analysis, and the unweighted pair-group
method with arithmetic mean algorithm
(UPGMA) was used for cluster analysis
between traits. Additionally, absence (O)
or presence (1) of differences between
pairs of treatments were registered.

Molecular comparison
Twenty accessions from in vitro storage
were compared with original clones from

the field using molecular markers. DNA


was extracted from leaves harvested from
the field trial using DNA extraction kit
(Fast DNAKit, BIO 101, Vista, CA, USA).
DNA quality and quantity were estimated
by comparison with CsCl-purified DNA
((Gibco-BRL) Life Technologies, Rockville,
MD, USA) digested by Pstl on ethidium
bromide agarose gels.

Microsatellite data analysis


Fourteen primer pairs of microsatellites
that had previously been shown to display
a high degree of polymorphism in potato
were used (Ghislain et al., 2001 ). Chromosome locations of the microsatellites,
fragment sizes, allele numbers detected by
the polymerase chain reaction, and
polymorphic index content (PIC) are
shown in Table 1. These primer sequences
and their respective amplification conditions are available upon request from
cip-mbl@cgiar.org. Amplification products
(bands) were detected on denaturing
polyacrylamide gels (6% acrylamide, 7 M
urea) and stained with silver nitrate (CIP,
1999). Each gel run included amplification
products from reactions containing in vitro
stored material and cuttings from original
clones from the field. Al lele sizes were
determined by comparison with bands
obtained from sequencing reactions
(pUC18-forward primer). Alleles were
scored as (1) for their presence, and (O) for
their absence. PIC was calculated according to Nei (1973).
A simple matching coefficient was used
for similarity analysis and UPGMA for
cluster analysis between traits. Numerical
taxonomy methods were performed using
the software NTSYSpc-2.02.

Results
Morphological analysis
A total of 38 accessions were evaluated
for morphological characters. The dendrogram obtained by an initial cluster
analysis revealed six cases of
misidentifications (four from in vitro

CIP Program Report 1999 - 2000

177

Table 1. Microsatellite markers used far molecular analysis of 20 patato accessions retrieved from in vitro
storage in the gene bank held in trust at CIP.
Fragment size range (bp)2
Chromosome
Ali eles
PIC 3
SCRl1 code
VI
93-217
16
0.875
STM0019
STGBSS-1
VIII
130-142
7
0.812
V
203-242
0.711
5
STPAc58
XII
136-191
0.792
STM0030
8
VIII
243-262
0.811
STM1016-1
8
VII
179-205
STM0031
6
0.727.
4
1
0.662
185-205
STM1049
VIII
224-243
STWAX-21
8
0.799
VIII
169-182
0.804
STM1104
7
IX
212-268
0.811
STM1052
8
169-213
STM3012
IX
0.664
6
IV
181-201
STM3023
4
0.692
X
131-163
STM1106
7
0.780
11
185-244
STM2022
4
0.588
1 SCRI = Scottish Crop Research lnstitute.
2 bp = base pairs.
3 PIC (Polymorphic index content) = 1 -(p2) where pi is the frequency of the ith allele detected.

Table 2. Morphological and molecular comparisons


between patato accessions conserved in vitro and
in the field.
In vitro vs field
CIP number
Morphological
Molecular
comparison
comparison
706670
Similar
Similar
704243
Similar
Similar
705440
Different1
Similar
703938
Different1
Similar
703964
Different2
Different
704521
Different3
Similar
700362
Different2
Different
701524
Similar
Different
705638
Different2
Different
704496
Different1
Similar
704444
Similar
Similar
704545
Different1
Similar
706622
Similar
Similar
706084
Similar
Similar
703929
Similar
Similar
706108
Similar
Similar
704470
Similar
Similar
706614
Similar
Similar
705781
Similar
Similar
706896
Different2
Different
1 Morphological

differences occurred mainly in stem color,


stem wings, and flowering degree.
2 Oramatic morphological differences in tuber and flower
characters, suggesting the occurrence of
misidentification s.
3 Slight morphological differences in tuber characters were
observed.

178

Research on Patato

storage and two from cryopreservation).


Since these accessions showed multiple
morphological differences (Tables 2 and
3), they were not included in the analysis
that followed. Results of comparison of
material from in vitro storage with clones
maintained in the field showed that five
accessions were slightly different in stem
color, stem wings, flowering degree, and
tuber characteristics (Table 2).
Resu lts of comparisons of cryopreserved
material with the original clones maintained in vitro (Table 4). The frequency of
differ- ences observed ranged from 0% to
22%, and varied by type of propagule.
Differences were found for five descriptors
for tubers, seven for stem cuttings, and
seven for plantlets. lt also was notable that
when using tubers, differences were found
in only four accessions (Table 4).

Microsatellite analysis
Twenty potato accessions were DNA
fingerprinted using 14 microsatellite
markers. These are located on 1O of the 1 2
chromosomes of potato. Hence, a large
portion of the pota to geno me ..cou Id be
assessed for genetic variation. We have
obtained a total of 92 alleles. The number
of al leles/locus ranged from four

Table 3. Morphological comparisons between in vitro- and cryopreserved patato


types of propagules.
Propagules
CIP
Species1
Plantlets
Cuttings
Tuber
number
adg
701535
Different (1/3)
Similar
adg
Similar
Different (1/3)
702091
Different
adg
Different (3/3) 3
703932
adg
703951
Different (2/2)
Different (2/3)
adg
706021
Similar
Similar
Different
cha
704047
Different (1/3)
Similar
cur
704717
Similar
Similar
gon
Similar
Similar
701862
Similar
702961
gon
Different (2/3)
Different
gon
Similar
Similar
Similar
703352
gon
Similar
703777
Similar
Similar
phu
703547
Different (3/3)
Different (2/3)
phu
703548
Different (1/2)
Similar
Different
phu
703580
Different (3/3)
Different (1/1)
Misidentified4
phu
Misidentified
703596
Misidentified
stn
Different (2/2)
Similar
702033
Different (1/3)
703876
stn
Similar
Misidentified
Misidentified
704089
stn
Misidentified

accessions, using three

Differences in 2
(2/2)

(3/3)

d4
d4, d2, d12
d2, d4, d19, d20
d2, d3, d4
d4, d12, d13
d12

(2/3)

d4

(2/2)

d9, d12, d13, d20


d4, d20
d2, d4, d17
d12
d4

= S. tuberosum subsp. andigena, cha = S. x chaucha, cur = S. curtilobum, gon = S. gonioca/yx,


phu = S. phureja, stn = S. stenotomum.
2 d2 = stem color; d3 = stem wings; d4 = flowering degree; d12 = calyx color; d9 = distribution of secondary flower
color; d13 = pedicel color; d17 =distribution of secondary skin tuber color; d19 = secondary flesh color; d20 =
distribution of secondary flesh color.
3 Number of different cryopreserved clones per total number of evaluated.
4 Misidentified - accessions that showed big differences in flowers and tubers compared with the original.
1 adg

(STM1049, STM3023, STM2022) to 16


(STM0019). PIC ranged from 0.588
(STM2022) to 0.875 (STM0019) (Table 1).
Only five of 20 accessions retrieved from
in vitro storage showed banding pattern
differences compared with the original
clones (Figure 1). CIP 705638 showed
differences in seven alleles, CIP 703964 in
eight, CIP 701524 in 20, CIP 706896 in 23,
and CIP 700362 in 37.

Discussion
Differences in morphological characters
between in vitro and cryopreserved clones
and control clones could be attributed to
inconsistency in the criteria used to record
qualitative data. For example, descriptor
stem color stage 2 (green with few pig-

mented spots) and stage 3 (green with


many pigmented spots) cou Id be seo red

either way when it was difficult to determine the score far intermediate
expression. Variability in morphological
descriptors could also occur dueto physiological age of plants, presence of
diseases, and environmental factors
(Kumar et al., 1999; Potter and Janes,
1991 ). Therefare, descriptors with higher
percent variation, (stem color, flowering
degree, distribution of secondary flesh
color (Table 4)) are not useful far detecting
genetic changes or describing clonal
identity. In this study, the morphological
comparisons using three types of
propagules showed that tubers should be
preferred as planting material rather than
stem cuttings or plantlets. The probability
of detecting slight morphological differences, due to physiological conditions
rather than genetic changes, seems higher

CIP Program Report 1999 - 2000

179

Table 4. Morphological differences observed in 18 patato accessions retrieved from


cryopreservation compared with original clones maintained in vitro, using three types of propagules.
Differences by propagule (%)
Accessions showing
Morphological descriptors
differences (no.)
Avg.
Cuttings
Plantlets
Tubers
o
o
o
o
Growth habit
2
22
11
7
3
Stem color
o
o
1
2
1
Stem wing
15
17
6
13
Flowering degree
9
o
o
o
o
Corolla shape
o
o
o
o
Predominant flower color
o
o
o
o
- intensity
o
o
o
o
Secondary flower color
o
o
1
3
- distribution
o
o
o
o
Anther pigment
o
o
o
o
Pistil pigment
7
8
7
7
6
Calyx color
o
14
6
4
Pedicel color
3
Tuber skin color
o
o
o
o
Predominant tuber skin color
o
o
o
o
- intensity
2
o
1
o
Secondary tuber skin color
o
2
o
1
- distribution
o
o
o
o
Tuber flesh color
- predominant
o
o
11
4
- secondary
1
2
14
7
8
- distribution
3
o
o
o
o
Tuber outline shape
o
o
o
o
Unshape tubers
o
o
o
o
Depth of eyes
7
7
Oescriptors in which differences were observed (no.)
5
15
12
Evaluated accessions (no.)
10
4
Accessions showing differences (no.)
8
6
with stem cuttings and plantlets than with
tubers. When variations are detected in
this kind of experiment, it should be
necessary to confirm the observed differences using plants coming from tubers. In
this study, more descriptors with differences were found when plantlets and stem
cuttings were used as the propagules (7 in
both). These descriptors are mainly related
to the expression of pigmentation. Such
characters are highly influenced by
physiological status of the plant. Thus,
plants coming from in vitro propagation
and directly planted in the field do not
grow as vigorously as those coming from
tubers. Consequently, they do not express
all color characters. lt was also observed
that treatments located in plots near the
border or under shade did not express
stem, pedicel, and calyx pigmentation in
the appropriate intensity. One accession

180

Research on Potato

from in vitro storage and four from


cryopreservation showed very smal 1
differences in tuber characters (Tables 2
and 3). Likewise, these slight variations
cannot be considered genetic. To detect
true morphological changes in tubers, it is
necessary that the data represent the
average expression observed in all fully
mature tubers produced by many plants in
a plot. That is because tuber shape and
color intensity differences could occur
within the same clone due to differences
in their developmental stage. lt is also
widely recognized that tuber characters
are influenced by pathological factors.
These are sorne reasons why detected
morphological variations should be
complemented with molecular markers in
the characterization of potato genetic
resources.

In vitro-store accessions that showed


numerous morphological differences in
tubers and flowers (four accessions) also
displayed remarkable differences in
banding patterns of molecular markers
(ranging from 7 to 37 al lelic differences).
This finding suggests that the observed
morphological differences are due to
genotype mixtures. Another group of five
accessions with slight morphological
differences showed molecular fingerprints
identical to the original clones, suggesting
that either the morphological evaluations
were not conclusive or that subtle changes
occurred. One accession, CIP 701524, not
showing any morphological difference was
found to have different DNA banding
patterns for 11 microsatellites. This finding
indicates that a verification system to
determine true-to-type potato germplasm
should include molecular genetic analysis.
Other investigators working on potato and
other crops have reached similar conclusions (Borner, et al., 2000; Gupta and
Varshney, 1999; Potter and Jones, 1991 ).
Our results also support recommendations
by other authors (Gupta and Varshney,
1999) about the inclusion of more than one
type of molecular marker for true-to-type
verification. Primers with high PIC should
be preferred for detecting genetic variations. Five of 14 primers used (STM3012,
STM0030, STM0019, STM1106, and
TWAX-21) revealed more alleles in the
comparison of patterns.

.:::
-1<

C\J

O)

CX)

,....._

>
o

<C
Cl

o
>

o
>

o
>

o
>
o
>
o

>
o

l.{)

>

6
~E

..__..

Q)

"O

-5

-5

~ O>
-Q) "C
e::
e::::
Q)

e::

"O
Q)

e::

Cl:l

Cl:l

en

(.)

Cl:l

en $:

~ -~
o en

Cl:l

Cl:l

e::

-~-..:..
...... .e

s~
-~<O

e:: O'>
ocx::>

en <O
-O.......:
...... , , _
Q)

~O...-~
E-...._
o
c.
(.) "O "E
e:: e::
-

Cl:l

Cl:l

$:

~~E
Q)

~c::l:

~ex:> :z

1....C")Cl

~~ex:>

e :2

~a...=>
> c.

c::-

- o
E:;--- en
o ex:> e::
~~~
"O ""3'"

e:.:>

'
N
>LO

CU

T"""

......

Q)

;;::o
Q) ,,_

...... a...

Cl:l

O>

-~

e:.:>

en e::
e::
Q)
~
o
- CT
en ...-..

'

~~en
(.)..__.._

'-'NO
Cl:l <O en
O M
cu
+-'

en

.g:::: ~

e.a... ~
o- o

>

~""3'" ~

Ct)

>
C\J

Cl:l

.::: <(

1-

1'

-U >.
-

-1<

>
o
>

<O

......

C\J

,....._

Q)

~en

>
>

CX)

$: ~
en +-'
'
Q)

'-'=
e::-

~-E

-1<

.........

Q)

u
u

-1<

Q)

~o

l.{)

O)

.E ..
...... U)
a:

>
o

>

e::
~

<O

Ct)

"O

>
o

>
o

-1<

This study showed that conservation of


genetic integrity of potato clonal genetic
resources largely depends on (1) the most
careful handling of the data used to
identify the accessions and their location
within the storage facility, (2) the effective
use of a computerized labeling system of
accessions in long-term storage, (3) the use
of experienced and well-trained personnel
to prevent misidentifications during all
stages of propagation-in the laboratory, in
the screenhouse, and in the field, and (4)
the checking of true-to-typeness by
methods that assure the detection of
misidentifications and genetic instabi lities.

o
>
o
>

N-

Q)

O> LO

+-'

o ...-....e::
.e
+-'
e::..__.. e::

e~

2-M

C.

Q)

Q)

=r- ......

.:=a_~

--e::x:

~~(!J
Cl e::
.
-r-N
en
1

m en X

...

Q)

Cl:l

g, 8

CIP Program Report 1999 - 2000

<(

t-

181

Conclusions
Morphological comparisons between in
vitro storage material and the original
clones maintained in the field revealed
differences in 9 of 20 genotypes assayed.
Of the nine genotypes detected, four were
judged to be misidentifications. The
remaining five showed slight morphological differences attributable to pathological
or environmental factors.
Morphological comparisons between
material cryopreserved for one year and
the original clones maintained in vitro
revealed differences in six of 18 genotypes assayed. Of these, two were
misidentifications. The remaining four
showed small morphological differences,
again attributable to pathological or
environmental factors.
The data obtained from morphological and
SSR analysis of original potato clones,
maintained in the field gene bank vs those
from in vitro storage for up to five years,
support the occurrence of misidentifications of potato accessions in the
different repositories. Microsatel 1ite
markers have been found to be a powerful
tool to complement the results obtained by
morphological comparisons. Both morphological and molecular methods are in
agreement in detecting misidentified
accessions rather than genetic i nstabi 1ities
(Ghislain et al. 2001 ).

Acknowledgements
The authors thank Merideth Bonierbale for
her valuable comments on the manuscript.

References
Angel, F., V. Barney, J. Tohme, and
W. Roca. 1996. Stabi 1ity of cassava
plants at the DNA level after retrieval
from 10 years of in vitro storage.
Euphytica 90(3):307-313.
Benson, E., M. Wilkinson, A. Todd,
U. Okura, and J. Lyon. 1996.
Developmental competence and ploidy
stability in plants regenerated from

182

Research on Potato

cryopreserved potato shoot-tips. CryoLetters 1 7:119-128.


Borner, A., S. Chebotar, and V. Korzun.
2000. Molecular characterization of the
genetic integrity of wheat (Triticum
aestivum L.) germplasm after long-term
maintenance. Theoretical and Applied
Genetics 100:494-497.
CIP (lnternational Potato Center). 1999.
Molecular biology laboratory protocols:
Plant genotyping. Ghislain, M.,
D. Zhang, and M. Herrera (eds.). Crop
lmprovement and Genetic Resources
Department. Training Manual. 2"d
Edition (revised June 1999), CIP, Lima,
Peru.
Ghislain, M., F. Rodrguez, F. Villamn,
J. Nez, C. Vsquez, D. Andrade,
O. Fras, M. Ames, G. Perazzo,
R. Herrera, W. De Jong, D. Milbourne,
R. Waugh, and M. Bonierbale. 2001.
ldentification of a microsatel 1ite set
suitable for potato genetic
identification. Proceedings of the plant
and Animal Genome Conference IX.
Poster 287. San Diego, CA, USA.
p. 132.
Golmirzaie, A., and A. Panta. 1997a.
Tissue culture methods and approaches
for conservation of root and tuber crops.
In: Razdan, M. and E. Cocking (eds.).
Conservation of plant genetic resources
in vitro. Vol. l. General aspects.
Science Publishers lnc, Enfield, NH,
USA. p. 123-152.
Golmirzaie, A. and A. Panta. 1997b.
Advances in potato cryopreservation by
vitrification. In: CIP program report
1995-96. lnternational Potato Center,
Lima, Peru. p. 71-76.
Golmirzaie, A. and J. Toledo. 1999. In
vitro conservation of potato and
sweetpotato germplasm. In: lmpact on a
changing world. CIP program report
1997-98. lnternational Potato Center,
Lima, Peru. p. 351-356.
Golmirzaie, A., A. Panta, and J. Toledo.
1999. Advances in the conservation of
root and tuber crops. In: Benson, E.E.
(ed.). Plant conservation biotechnology.

Taylor and Francis Ltd., London, UK.

p. 165-178.
Gupta, P. and R. Varshney. 1999.
Molecular markers for genetic fidelity
during micropropagation? Current
Science 76:1308-131 O.
Harding, K. 1994. Molecular stability of
the ribosomal RNA gene in Solanum
tuberosum plants recovered from slow
growth and cryoconservation. Euphytica

55:141-146.
Huamn, Z., J. Williams, W. Salhuana,
and L. Vincent. 1977. Descriptors for the
cu ltvated pota to and for the
maintenance and distribution of
germplasm collections. lnternational
Board for Plant Genetic Resources,
Rome, ltaly. 47 p.
Kumar, M., R. Barker, and B. Reed. 1999.
Morphological and molecular analysis
of genetic stability in micropropagated
Fragaria x ananassa cv. pocahontas. 1n
vitro Cellular and Development
Biology. Plant 35(3):254-258.

Nei, M. 1973. Analysis of gene diversity in


subdivided populations. Proceedings of
the National Academy of Science

70:3321-3323.
Piola, F., R. Rohr, and P. Heizmann. 1999.
Rapid detection of genetic variation
within and among in vitro propagated
cedar (Cedrus libani Loudon) clones.
Plant Science 141 :159-163.
Potter, R. and M. Jones. 1991. Molecular
analysis of genetic stability. In: Dodds,
J. (ed.). In vitro methods for
conservation of plant genetic resources.
Chapman and Hall, London, UK.

p. 71-91.
Schafer-Menurh, A., E. Muller, and
G. Mix-Wagner. 1996. Cryopreservation:
An alternative from the long-term
storage of old potato varieties. Potato
Research 39:507-513.
Steponkus, P., R. Langis, and S. Fujikawa.
1992. Cryopreservation of plant tissues
by vitrification. Advances in LowTemperature Biology 1 :1-16.

GIP Program Report 1999 - 2000

183

Native Potatoes: Potential Markets


Outside the Andes
For perhaps 8000 years, subsistence farmers in the Andes of Bolivia, Ecuador, and
Peru have enjoyed native potatoes as their primary staple. These intensely
flavored potatoes have colorful flesh and skin and a wide range of tastes and
textures. Efforts to market bright yellow-fleshed cultivars indicate potential within
the realm of fresh food and convenience markets as well as restaurants and other
specialized users. The time may now be ripe to introduce Andean native potatoes
to outside markets through the development of products such as native-potato
chips (see cover photo) and french fries. Sorne consumers prefer native-potato
french fries over conventional potato french fries, particularly those made from
varieties with high dry matter and a bright yellow color.
Along with color and texture and large variation in firmness and mealiness, native
potatoes have other appealing characteristics, such as their tendency to absorb
lesser amounts of fat during frying. This quality is largely due to native potatoes'
elevated dry-matter content, which can be as high as 34% of fresh weight. An
additional benefit of native potatoes is that they contain various pigments (yel low
carotenoids and red and purple anthocyanins) that not only produce a variety of
colors, but are also nutritionally important as antioxidants.
Almost 4000 native potatoes have been col lected and stored in CIP's gene banks.
Current research and development efforts with native potatoes at CIP concentrate
on identifying varieties that have consistently low reducing-sugar contents and
that can be grown with minimal or no use of pesticides. CIP researchers also work
with poor farmers in the Andes to preserve potato diversity in the field through onfarm conservation. Final ly, CIP works with farmer organizations and entrepreneurs
to identify native potatoes with the best potential for use in new postharvest
opportunities and to develop novel products and identify markets for them. Our
ultimate goal is alleviating rural poverty through value-added potato production.

184

Research on Potato

Assessment of Variability for Processing Potential in


Advanced Potato Populations
W. Amoros,

J. Espinoza, M. Bonierbale1

Advanced clones and varieties in CIP's breeding program were evaluated


under contrasting agroecologies in Peru to assess genetic variability and
environmental influences on specific gravity, dry matter content, glucose
content, and potato chip color. A selection of advanced clones was further
evaluated for processing performance following different storage conditions,
including cold temperature. Three contrasting environments of the coast,
jungle, and highlands of Peru were used. Two breeding designs (Diallel and
Line x Tester) were also evaluated for the determination of genetic parameters and assessment of parental value for processing characteristics.
Greater variation for dry matter and glucose contents were demonstrated
in a pre-breeding population based on cultivated native diploid species
(S. phureja and S. stenotomum), than in the more advanced tetraploid breeding population. The clones or varieties Mariva, Atlantic, 1-1062, POOS-16,
Mex-32, Pi rola, Tomasa Condemayta, LT-8, Y84.027, and C91.612 proved to be
the most promising for processing as chips. The clone E86.011 and the late
blight resistant selections 377369.7, 386529.3, and 85LB53.8 showed tolerance to cold temperature sweetening, or darkening of chips from potatoes
processed following storage at 4C. A high narrow sense heritability (h 2 =
0.74) was found for specific gravity, and conversely, little genetic variability
was found for glucose content. The progenitors 378015.16 (TS-2), and LT-7, as
well as the Polish clones PW88-6203 and Brda showed ability to transmit good
processing attributes to their progenies. The utilization of contrasting, stressful environments that represent the diverse agroecologies of certain
developing countries contributed to our ability to differentiate among materials and appreciate genotype x environment interactions.
Processing is the fastest growing sector in
the world patato economy. Approximately
half the annual crop in the USA is now
processed, and patato use in Europe's food
service industry is rising sharply. The
patato processing industry in many developing countries is also growing fast (FAO,
1995). In China, for example (Zhang and
1

CIP, Lima, Peru.

Guenthner, 1998), growing consumer


income levels, changing rural and urban
infrastructure, lower patato prices, greater
foreign investment, and increasing numbers of fast-food restaurants are the main
factors increasing the demand for processed potatoes.
Potatoes destined for the food service
industry need to meet certain stringent

CIP Program Report 1999 - 2000

185

requirements. They should be of uniform


size and shape, with shal low eyes, good
texture, high dry matter, and low reducingsugar content under a range of production
and storage temperatures. Potato varieties
such as Russet Burbank and Bintje possess
these characteristics and are grown
successfully in the USA and Europe.
However, these varieties are not su ited to
tropical and subtropical environments as
their productivity is limited by climatic
factors such as day length and temperature. In addition, these varieties are highly
susceptible to diseases and pests, includi ng late blight, viruses, bacteria! wilt, and
potato tuber moth. Successful production
therefore depends on extensive use of
pesticides, and healthy seed may need to
be imported or grown in controlled conditions, both of which are expensive
procedures.
Breeding programs at CIP and in many
developing countries are therefore placing
increasing emphasis on selecting varieties
of potato that are su itable for the processi ng industry, as well as resistant to the
most important diseases of the tropics. The
objectives of this long-term study (begun
in 1989) were to assess genetic variability
for specific gravity (SG), dry matter (DM),
and reducing-sugar (RS) content in advanced breeding materials and the effect
of tropical growing conditions on these
characteristics; evaluate the effects of
cold storage and storage under ambient
tropical conditions on processing qualities;
and investigate the inheritance of specific
gravity and reducing-sugar accumulation.
This investigation will help to determine
the most efficient methods for identifying
and selecting clones that are suitable for
processing. lmproving the availability of
resistant varieties suitable for processing
would allow developing countries to
increase local production, expand emerging national processing enterprises, and
eventually reduce imports of frozen patato
products for the growing fast-food industry.

186

Research on Potato

Materials and Methods


Testing sites

All evaluations were carried out at one


or more of CIP's experiment stations in
Peru: San Ramon (humid/hot climate all
year, altitude 800 m, latitude 11 08' S);
La Molina (arid lowland climate, altitude
280 m, latitude 12 05' S) during the
winter (temperate to cool) and spring
(warm) seasons; and Huancayo (highlands,
temperate and rainy, altitude 3280 m,
latitude 12 07' S) during the summer.
Plant material

Test materials for assessing processing


quality attributes and their variability
under a range of environmental conditions
were selected from three types of clones.
These are referred to in th is paper as
groups 1-3.

Group 1. The largest group of test materials (100 clones) was chosen from CIP's
'pathogen tested list' (PTL) of elite clones
and varieties (> 100), and was tested at
San Ramon and La Molina during winter,
1989. These materials are highly selected
tetraploid clones with diverse production,
resistance, and utilization traits. They all
carry resistance or tolerance to at least
one important pest, disease, or stress
factor.
Group 2. Another group was a sample of
42 clones taken from a diploid phureja/
stenotomum population. This population
was selected in the USA, from Solanum
phureja and 5. stenotomum parents, for
heat tolerance, adaptation to long days,
tuber appearance and high dry matter
(Gautney and Haynes, 1983; Haynes and
Haynes, 1983). These materials were
tested only at La Malina during winter of

1989.
Group 3. A third group of approximately
80 clones was selected from CIP's broadbased breeding populations, also
tetraploid, which are variously adapted to

conditions in tropical and subtropical


highlands and lowlands. Most of these
clones carry resistance to late blight or
viruses, but they have not been subjected
to extensive international trials or released
as commercial varieties. The materials
were tested at all three of CIP's test sites
over several years and only general
conclusions of these trials are presented
he re.
Additional test materials for two studies of
the inheritance of characteristics relevant
to processing were drawn from segregating
populations. These were represented by
families originating from crosses between
parents selected from the clonal material
or populations described above (groups 1
and 3).
Together, these materials represented a
wide genetic base from which to assess
the general and specific combining ability
for the main characteristics required for
processi ng, and to identify the best
progenitors and progenies for improving
CIP populations and selecting superior
clones.
Experimental designs

Results presented here on general evaluation of clones for their processing qualities
are overviews from several years of trials,
so designs are not presented in detail. In
general, for the evaluation of the clonal
material, trials followed randomized
complete block designs (RCBD) with two
or three replications of single 1O hill row
plots. Yields were expressed as the weight
of tubers on a whole plot basis (not
reported here).
Tests for processing quality (see below)
were usual ly conducted 1O days after
harvest, once the chemical composition of
the tubers had stabilized. However, in
arder to assess the i mpact of storage
conditions on processing qualities, additional postharvest treatments were added
to two trials. A group of 16 late blight

local varieties were planted in Huancayo


in the summer of 1996/97. Yield was taken
at harvest and processing characteristics
were assessed 1O days after harvest, and
after 30, 60, and 90 days in cold (4C) or
ambient (15-20C) storage. During the
winter season in La Molina, six advanced
clones, selected over several years in
different locations for their processing
attributes were planted with two local
varieties. An RCBD with three replications
and 1O hill plot experimental units was
used to evaluate early production and
plant vigor. Yield and selected processingrelated variables (dry matter content,
glucose content and chip color) were
evaluated 10 days after harvest, and chip
color was tested at 30 and 90 days after
cold or ambient storage, with and without
reconditioning (potatoes are removed from
cold storage and left at ambient temperature for two weeks to reconvert sugars
before processing).
In arder to assess the range of genetic
variability for specific gravity and the
inheritance of this character, a diallel
mating design with reciprocals involving
eight parents from the lowland tropics
breeding population was used. The trial
was established at La Molina during the
spring and summer of 1990. Finally, in
order to investigate in more detail the
heritabi 1ity of characters relevant to
processing, 12 well known breeding lines
and varieties, most of them resistant to
patato leafroll luteovirus (PLRV), were
crossed with three CIP progenitors for
patato Y potyvirus (PVY) and potato X
potexvirus (PVX), following a Line x Tester
design (12 x 3). Thirty-six tuber families
were planted during spring 1998 at La
Molina and summer 1998/99 at Huancayo
and distributed in an RCBD with three
repetitions of 30 plants per family. In both
sets of inheritance trials, yield was assessed on a per plot basis and processing
characteristics assessed 1O days after
harvest.

resistant and frost tolerant clones plus two

CIP Program Report 1999 - 2000

187

Evaluation of processing quality


Dry matter content was determined
directly as the ratio of dry weight/fresh
weight. Between five and seven tubers
(approximately 0.5 kg (total weight)) were
taken at random from the harvested plot,
cut into small cubes and mixed thoroughly. Two subsamples of approximately
200 g each were then dried in an oven at
BOC for 72 hours, or until constant dry
weight was achieved. Specific gravity was
calculated by measuring weight in air and
weight in water, using 1 O to 20 tuber
samples (approximately 1 kg) per plot.
In the general evaluation of clones, a
simple colorimetric test for glucose
content (glucose test strip) was used as an
indicator of the reducing-sugar content,
assessed on the basis of five tuber samples
cut in half lengthwise. A scale of 1 to 5
was used to indicate the following approximate measures: 1 = 0%, 2 = 0.10%,
3 = 0.25%, 4 = 0.50%, and 5 = > 2.0%;
and the darkest measurement was recorded per sample. In the two trials on
inheritance of characters relevant to
processing, the Ross analytic laboratory
method (Rodriguez, 1 996) was used to
quantify glucose content.
To evaluate chipping quality, three patato
tubers per plot were cut perpendicular to
the long axis, and six 0.5 mm thick slices
taken from each half were used for frying
and color tests. The slices were rinsed in
water and fried at 176-180C until the oil
stopped bubbling. Chip color was measured using a scale of 1 to 5, where grade
1 is the lightest color (white to cream), 2 is
light tan, 3 is dark tan, 4 brown, and 5
dark brown, according to a chart available
from the PC/SFA (American Snack Food
Association, Alexandria, VA, USA). Chip
color between grade 1 and grade 3 is
commercially acceptable. Oil absorption
was determined as the percent of weight
lost after applying 15,000 pounds/inch 2
pressure to 1O g samples of ground chip for
3 minutes using a mechanical press.

188

Research on Potato

Results
Evaluation of advanced clones and
varieties for processing characteristics
(Group 1)
Dry matter content
Dry matter content evaluations of CIP's
pathogen tested list (PTL) clones and
varieties suggested the presence of
considerable genetic variability for this
characteristic (Figure 1 ). The highest
values for dry matter content in La Molina,
of about 24-26%, were observed for the
following varieties: AKK69.1 (720052),
Tomasa Condemayta (720072), Esperanza
(720119), GLKS-58-1642.4 (800290), G-1
(278072.1 O), Mariva (720025), ARK-69.1
(675158), and Gabriela (720120). In San
Ramon, dry matter of the same materials
reached about 22-23%, with the varieties
Dejima (800974), Esperanza (720119), G-1
(278072.1 O), CFJ-69.1 (676002), ARK-69.1
(675158), Yungay (720064), Linea-21,
Linea-34, and Gabriela (720120) being the
highest. The mean dry matter content for
these clones was higher in the arid lowlands of La Malina (20.96%) than in the
hot and humid San Ramon area (18.90%),
confirming the detrimental effects of high
temperatures on this variable previously
reported by Hernandez (1989). However,
sorne genotypes produced higher dry
matter contents under warmer conditions
than cooler ones (Linea 34, Linea 83, and

25.0.--------------~

20.0

~ 15.0

~
e:

10.0

Q)

(!)

5.0

o.o .__T'"'"'"""-T'"'"'"""-T'"'"'"""-,.-......,.-...,,.-...,,-...,.-....,,-...,.-....,.-....,---,.........
13 14 15 16 17 18 19 20 21 22 23 24 25 26
DM(%}

Figure 1. Frequency distribution of dry matter content


(DM) of 100 clones from the pathogen tested list
grown in La Malina (LM) and San Ramon (SR),

Table 1. Analysis of variance far processing parameters combining two environments; 100 from the
pathogen tested list, group 1.
So urce
Mean squares
d.f
Dry matter
Glucose
Chip color
**
**
**
Environnent (E)
1
14.3111
382.730
42.9549
Repetition (E)
2
0.028
0.00108
0.0001
**
**
**
Genotype (G)
99
12.204
0.71581
2.5054
**
**
**
99
0.51266
GxE
5.6612
0.9978
Error
198
0.2791
0.00019
0.0025
C.V
2.660
0.730
1.830
Mean
19.890
1.905
2.740
Note: ** = P > F < 0.0001.
Mex-750826), and others showed outstanding stability for this character (Yungay, Flor
Blanca, CFE-69.1, Mex-32, and Huaycha).
Analysis of variance indicated a significant level of genotype x environment
(G x E) interaction for dry matter, glucose
content, and chip color (P < 0.0001) (Table
1). This finding was contrary to earlier
reports (Shaw and Booth, 1982).

Glucose content
Frequency distributions for glucose content
(Figure 2) and color of chips (Figure 3) also
showed substantial variability among
clones with a tendency for higher values
in San Ramon, and there were significant
differences between environments. For
example, a higher percentage of genotypes with commercially acceptable
Grades 1 to 3 for chip color occurred in La
Molina (69%) compared to San Ramon
(54%). Severa! varieties showed adequate

processing characteristics under the


conditions of San Ramon and La Molina:
Mariva (720025), Atlantic (800827), GLKS58-1642.4 (800290), 1-1062 (57501 O),
POOS-16 (575045), Mex-32 (720091 ),
Pirola (800957), Esperanza (720119),
Tomasa Condemayta (720072), CFE-69.1
(720053), and Bl-2.9 (678011 ).

Diploid Population (Group 2)


Evaluation of dry matter content and
glucose content of the diploid 5. phureja/
S. stenotomum population resulted in a
negative correlation (r = -0.51 ), as previously found in tetraploids (Putz et al.,
1982 cited by Ross, 1986).

Dry matter content


The frequency distribution for dry matter
content in the diploid population (Figure 4)
showed a greater range of variability
(16.2-30.5%) than did that for the tetraploid PTL material (Figure 1). A comparison

50.0

-:

LM

SR

SR

20.0

40.0

gi

~ 30.0

.LM
0

25.0

60.0 . - - - - - - - - - - - - - - - - - - ,

15.0

(/)

Q)

20.0

10.0

Q)

C!l

1.5

2.5

3.5

4.5

Glucose (Scale 1-5)

Figure 2. Frequency distribution far glucose content


of 100 clones from pathogen tested list grown in La
Malina (LM) and San Ramon (SR), group 1.

5.0

1.5

2.5

3.5

4.5

Chip color (Scale 1-5)

Figure 3. Frequency distribution of chip color of 100


clones from CIP's pathogen tested list grown in La
Malina (LM) and San Ramon (SR), group 1.
CIP Program Report 1999 - 2000

189

25.0~------------~

16
~

20.0
~

12

Ul

Ul

c.

C!l

(])

e(])

PTL(4x)

Phu-stn (2x)

15.0

(])

-5'
e
(])
C!l

16

17 18

19 20 21 22 23 24 25 26 27 28 29 3l 31

DM (%)

10.0
5.0

15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 3l 31

DM(%)

Figure 4. Frequency distribution of dry matter content


(DM) in a sample of 42 clones from the phureja/
stenotomum diploid population, group 2.

Figure 5. Comparison frequency distribution of dry


matter content (DM) in samples of two population 4X
(PTL) and 2X (phureja/stenotomum) grown in La
Malina (groups 1 and 2).

of the dry matter frequency distribution for


diploid and tetraploid clones (both grown
at La Molina) showed a greater concentration of values around a lower mean for the
tetraploids than the diploids (Figure 5).

Advanced breeding clones (Group 3)

Wide variability for chip color was also


found in tetraploid and diploid clones,
with the former appearing more disperse.
A similar distribution was found for
glucose content. Considering the overall
range of variation, the phureja/
stenotomum population is an excellent
source of improvement for dry matter
content. Six clones were found to have
high specific gravity in the range of
1.121-1.141, and low glucose content in
the range of 0-0.10%. These are highly
suitable for crossing with tetraploid clones
as they produce a relatively high frequency of 2n gametes.

Materials from CIP's lowland tropics virusresistant population were evaluated over
different seasons at all three test sites. Five
clones were found to be suitable for
processing as french fries, chips, or both.
Table 2 describes a group of successful
progenitors of virus resistance whose
clonal evaluation revealed favorable
postharvest characteristics. The National
Potato Program and Tacna and lea Universities in Peru have released sorne of these
as commercial varieties, including
Y84.027 as Mara Bonita, LT-8 as
Costanera, and C91.612 as Mara Reiche.
Effect of storage conditions
Measurements taken on the harvest from
breeding materials in Huancayo showed
that weight loss during ambient storage
was higher (reaching levels of 8%, the

Table 2. Advanced progenitors with good processing quality and potential for use, La Molina, winter, 1994,
group 3.
Resistan ce
Yield (kg/plant)
DM (%)
GI (%)1
Clone
Use 2
LT-8
PVX, PVY
0.850 (1.35)
20.94 (1.05)
CH,FF
0.05
Y84.027
PW
1.150 (3.84)
18.93 (2.95)
CH,FF
0.05
1.070 (2.30)
20.19 (1.85)
X86.011
PVX
0.06
CH
PLRV, PVX
1.000 (3.09)
88.108
24.38 (1.03)
CH
0.05
PW, PVX
1.200 (2.40)
22.50 (0.92)
C91.612
0.06
FF
Note: DM = Dry matter.
1 GI = glucose content (acceptable below 0.25).
2 CH = chips; FF = french fries.

190

Research on Potato

maximum loss found in any clone) than


during cold storage (4%). The clone
386647.29 showed minimum weight loss
of 2.2% after 90 days of ambient storage.
Although the average dry matter content
was relatively high (> 24%), significant
differences were found between clones far
dry matter, specific gravity, glucose
content, and oil absorption. Changes in dry
matter content during storage are shown in
Figure 6A. The clones 386647.17,
377369.7, and 386529.3 showed the
highest values of dry matter content at
harvest with an average of 27.13, 26.20,
and 26.11 %, respectively. Figure 6B
illustrates how glucose content changes
over time, confirming previous reports of a
greater increase during cold storage than
during storage under ambient conditions
(Burton, 1989; Ordoez and Limongelli,
1980; Dogras et al, 1989). Clones with low
and stable levels of glucose content under
low temperature storage included
377369.7, 386529.3, and 85LB53.8.
Relatively low oil absorption is an important quality far processing, having
economic as wel 1 as health advantages.
Several clones had oi 1 absorption leve Is
below the average (36.65%). For example,
clone 377369.7 showed absorption levels
of 32.7% at harvest and 31.6% after 30
days of ambient storage, and clone
386647.29 showed 33.3% at harvest and
29% after 30 days of ambient storage.
The clase significant relationship between
dry matter and specific gravity found in
earlier experiments (Ratovski et al., 1981)

A
25.0
24.5

24.0

:2

o 23.5
23.0
22.5
30

60

90

Days of storage

B
0.25
0.20

l0.15

3l

8::J 0.10

0.05
0.00

30

60

90

Days of storage

Figure 6A and B. Change with time under two storage


regimes (cold storage at 4C or under ambient
conditions) of A dry matter content (DM) and B
glucose content in 16 late blight resistant and frost
tolerant clones gown in Huancayo, Peru.
was confirmed in this study. Significant
negative correlation of dry matter content
with oil absorption and of specific gravity
with oil absorption were obtained under
both storage conditions (Table 3), confirming that higher dry matter content wil 1
generally result in lower oil absorption
(Cambie et al., 1987; Kozempel et al.,
1991).

Table 3. Correlations among dry matter content (DM), specific gravity (SG), oil absorption, and glucose
content (GC) of 16 late blight and frost tolerant clones, at harvest and after storage, group 3.
R
Harvest
Ambient storage
Cold storage (4C)
30
60
90
30
60
90
SG-DM
0.912 **
0.725 **
0.880 **
0.928 **
0.613
0.774 **
0.747 **
-0.473
-0.538 *
SG-oil
-0.458
-0.605 *
-0.645
-0.893 ** -0.565
-0.235
-0.622 ** -0.560 *
-0.362
-0.181
-0.181
-0.563
SG-GC
-0.296
-0.420
-0.528 *
-0.303
DM-oil
-0.500 *
-0.804 ** -0.606 *
-0.162
-0.457
-0.124
-0.256
-0.341
-0.537 *
-0.386
DM-GC
Ol-GC
0.350
0.100
0.387
0.026
0.292
0.164
0.727
Note: * = significant (P < 0.05); ** = highly significant (P < 0.01 ).

CIP Program Report 1999 - 2000

*
**
*
**

191

Table 4. Yield and chip color of selected clones under two storage conditions, La Malina, winter 1995,
group 3.
Chip color1
Yield
DM (%)
Clone
4C
4C
Ambient
(kg/plant)
Ambient
R
30 days
30 days
90 days
90 days
1.160 a
21.4
1.5
3.5
E86.694
1.0
4.0
2.5
22.8
E86.011
0.909 ab
1.0
1.0
2.0
1.5
1.5
LT-8
0.827 b
20.5
1.5
3.0
1.5
3.5
3.5
21.5
4.0
E86.604
0.777 b
1.0
1.0
4.0
3.0
E86.300
0.765 b
22.2
1.0
3.0
2.0
2.5
2.5
MARI VA
0.745 b
22.4
2.0
3.5
2.0
3.5
3.0
0.743 b
21.7
E86.692
1.0
3.5
1.0
4.5
2.5
22.3
E86.695
0.750 b
1.0
3.5
1.0
3.5
3.0
Duncan test for yield/plant (P < 0.05).
1 Scale used: 1-5, where 1= lightest color (white or pale yellow) and 5= darkest color (dark brown). Colors in the 1-3

range are commercially acceptable. R = reconditioning at ambient temperature following 90 days cold storage.

Most of the clones tested produced a good


chip color following storage under ambient
conditions, although three were very poor.
In contrast, only four clones (377369.7,
386529.3, 85LB53.8, and 385542.7)
produced a good chip color after cold
storage.
The performance of clone 377369.7 is
notable. With resistance to late blight,
high yield, high dry matter content, and
low glucose levels after storage at 4C,
good chip color, good texture, and acceptability, this clone has great potential for
processing or as a progenitor in breeding
programs, providing its yield and quality
are stable in different production environments.
The results of tests on advanced clones
already selected for their processing
qualities are shown in Table 4. Clones
E86.694 and E86.011 produced the highest
yields. Al 1 clones fried very wel 1 at harvest
and after ambient storage. Sorne browning
of chips occurred in samples stored at 4C,
but reconditioning for two weeks after cold
storage resulted in acceptable chip color.
One clone (E86.011) produced chips of
excellent color after ali storage conditions.
Although these clones (E86 series) were
selected primarily for their processing
characteristics, no extreme susceptibility
to viruses or late blight was observed.

192

Research on Patato

Heritability of key processing


characteristics
CIP's lowland tropics breeding population
was found to harbor a large range of
genetic variability for dry matter content.
The effect of the growing environment on
this characteristic was highly significant,
with a lower mean during the summer
(SG = 1.062, dry matter 18.3%) than
during the spring (SG = 1 .075, dry matter
20.2%). These results indicate that warmer
temperatures are detrimental to the quality
and yield of processed products.
lnterestingly, G x E interactions were
insignificant in this population study,
indicating that progenies with high
specific gravity under favorable environmental conditions (spring) still ranked high
under unfavorable conditions (summer).
This finding, together with high narrow
sense heritability estimates (h 2 = 0.74),
indicates that intercrossing progenitors
with high specific gravity will permit rapid
progress in breeding for this trait. In
addition, advanced clones selected from
this population produce acceptable and
relatively stable yields under both cool
and warm conditions.
The CIP progenitors 378015.16 (TS-2),
378017.2 (LT-7), and to a lesser extent
575049 (CEW-69.1 ), showed the ability to
transmit high specific gravity to their

Table 5. Estimation of combining ability effects using a diallel design with reciprocals, Group 3.
Progenitors
Combining ability effects (GCA)
Yield
SG
RS(1)
RS(2)
378015.16 (TS-2)
168.2
8.2
0.001
-0.006
C83.119
-16.8
-4.2
-0.019
0.023
378017.2 (LT-7)
664.6
3.1
-0.028
-0.032
1-1039
-89.6
-2.5
0.012
0.001
Katahdin
-4.3
-5.0
-0.032
-0.003
377250.7
-70.6
-2.3
0.012
-0.015
377964.5
7.9
0.1
0.014
-0.007
575049 (CEW-69.1)
-59.3
2.6
0.390
0.040

SE (gi)
SE (gi-gj)

25.95
39.94

0.72
1.08

0.016
0.026

0.012
0.018

Note: RS(1 ), measured 10 days after harvest; RS(2), measured 60 days after harvest (ambient).

progenies as indicated by their respective


general combining abilities (Table 5). In
genetic research on reducing-sugar
content, it was found that average levels
in the segregating population were acceptable for processing into french fries and
chips. However, little genetic variability
was found for this trait. The CIP clones
378017.2 (LT-7) and 378015.16 (TS-2) and
the variety Katahdin transmitted the
tendency for low reducing sugar content to
their progenies. Conversely, the clone
575049 (CEW-69.1) transmitted a tendency
for higher reducing-sugar content to its
progenies, making it a less suitable parent
for processing unless it is crossed with a
low reducing-sugar progenitor. Clones 11039 and 575049 (CEW-69.1) have shown
useful levels of resistance to late blight
and early blight, respectively. Unfortunately, whereas they are well adapted to
warm agro-ecologies and have good
parental value for yield, uniformity, and
early production, several of the clones
noted here for their processing quality are
susceptible to late blight and viruses.
Furthermore, the presence of high levels
of toxic glycoalkaloids was found in
378017.2 (LT-7) and (378015.16 (TS-2),
noted here for their low reducing-sugar
contents; this might complicate their use
as progen itors.
Combined analysis (data not shown) of
resu lts from progeny tests of 12 wel 1
known breeding lines and varieties,

evaluated in La Molina and Huancayo,


indicated that lines PW88.6203,
PW88.6065, Achirana, Brda, G7445(1 ),
Fregata, and Berolina had good parental
value measured as high general combining
ability (GCA) for specific gravity, and the
clones PW88.6203 and Brda showed high
GCA for chip color. Clones PW88.6203
and Brda also possess high GCA for yield
and agronomic characteristics. Among the
testers used in this design, the progenitor
TXY.2 was the best parent for vigor and
uniformity of plants, yield (number of
tubers), and specific gravity. These families were also evaluated in North China
(lnner Mongolia) in comparative studies of
parental value for virus resistance and
chipping quality, and for clonal selection.
A group of more than 20 selections with
multiple virus resistance and good yields
will soon be evaluated for their processing
potential.

Discussion
These preliminary evaluations of CIP
populations and advanced clones for
processing quality have identified a useful
range of variabi 1ity and a number of
promising clones with good processing
characteristics. The majority of these also
contain resistance to at least one important disease, such as late blight or
common viruses. Clones with tolerance to
cold storage have also been identified.

CIP Program Report 1999 - 2000

193

Evaluation of segregating populations far


processing parameters provided valuable
genetic infarmation that will enhance
effective management of the breeding
populations and improve identification of
progenitors that can transmit good processing qual ities to their progenies. At the
same time, the evaluation allowed
progenitors transmitting undesirable traits,
such as high glycoalkaloid content, to be
eliminated from the breeding program.
The test sites represented a variety of
different environments (lowlands and
highlands, cold and warm, dry and humid).
These results will therefore facilitate the
choice of varieties best suited to different
agro-ecologies in developing countries, as
well as highlighting G x E interactions. As
a result, selected progenies, clones, and
progenitors are now available far distribution and further trial in appropriate regions.
Research at CIP is continuing to identify
superior progenitors and clones with
important resistances. Breeding far good
processing quality is now being given
secondary importance. For example, the
evaluation of processing potential has
recently been superimposed on assessment
of the variability far resistance to PLRV in
the current recombination cycle of CIP's
lowland tropics virus-resistant population,
by way of genetic analysis (North Carolina
Design 11) conducted across environments.
This will help to identify new progenitors
that consistently transmit good processing
characteristics as wel 1 as resistan ce to
their progenies.
Considerable progress has therefare been
made towards meeting the increasing
demand far processing and dual-purpose
potato varieties. Frequencies and levels of
disease resistance in CIP populations are
now high enough, and variability far postharvest quality sufficient, to permit
selection far both resistance and utilization characteristics. Additional, robust
genetic experiments, crossing the best
progenitors with testers of poor quality,
would also be required to give postharvest

94

Research on Potato

quality equal priority to disease resistance.


Selection for long or oblong shaped tubers
far the frozen french fry industry could also
be emphasized on an experimental basis
now that populations containing sufficient
levels of important primary characteristics
are available, and environmental influences on quality parameters are better
understood.
Participatory research on the management
of these potential varieties in farmers'
fields and with different end users, as well
as their incorporation into local seed
systems, are critica! areas far cooperation
with national research programs in the
short-term. lncreased communication
between breeding and selection programs
is also needed to fui ly assess the response
of selections and pop u lation samples to
the most relevant environments and
production practices.

References
Burton, W.G. 1989. The Patato. 3rd
Edition. Longman Scientific &
Technical, Essex, UK. 742 p.
Dogras, C., A. Siomos, and C. Psomakelis.
1989. Sugar content, dry matter and
sprouting of patato (Solanum tuberosum
L.) tuber stored at 6C and 1OC in
relation to cultivar and area of
production. Patato Abstracts 16(3):114.
FAO. 1995. Potatoes in the 1990s.
Situation and prospects of the world
patato economy. lnternational Patato
Center (CIP), Food and Agriculture
Organization (FAO), Rome, ltaly. 39 p.
Gamble, M.H., P. Rice, and J.D. Selman.
1987. Relationship between oil uptake
and moisture loss during frying of patato
slices from c.v. Record U.K. tubers.
lnternational Journal of Food Science
and Technology (UK) 22:233-241.
Gautney, T.L. and F.L. Haynes. 1983.
Recurrent selection for heat tolerance in
diploid potatoes (Solanum tuberosum
subsp. phureja and stenotr:Jmum).
American Potato Journal 60(7):537-542.
Hernandez, E. 1989. Herencia de los
Factores de calidad para procesamiento

de papas autotetraploides. MSc thesis.


Universidad Nacional Agraria La
Malina. Escuela de Postgrado, Lima,
Peru. 95 p.
Haynes, K.G. and F.L. Haynes. 1983.
Stability of high specific gravity
genotypes of potatoes under high
temperatures. American Patato Journal
60(6):17-26.
Kozempel, M.F., P.M. Tomasula, and
J.C. Craig. 1991. Correlation of moisture
and oil concentration in French fries.
Lebensm-Wiss-Technol-Food-SciTechnology 24(5):445-448.
Ordoez, C. and J.C. Limongelli. 1980.
Calidad de Papa como Materia Prima
para la Industria. Programa Nacional de
Investigaciones de Tecnologa de
Alimentos, Universidad de Buenos
Aires, Facultad de Agronoma, Buenos
Aires, Argentina. p. 39-83.
Rastovski, A., A. Van Es, K.J. Hartmans,
N. Buitelaar, P.H. de Haan,
C.P. Maijers, J.H.W. Van der Schild,
P.H. Sijbring, H. Sparenberg, and

B.H. Van Zwol. 1981. Storage of


potatoes: Post-harvest behavior, store
design, storage practice, and handling.
Center for Agricuitural Publishing and
Docu mentation, Wagen i ngen,
Netherlands. p. 462.
Rodrguez, F. 1996. Mtodos para Evaluar
la Calidad Culinaria y para el
Procesamiento de las Races
Reservantes de Batata. In: Manual de
Manejo de Germoplasma de Batata o
Camote (ipomea batatas). Training
Manual, Part 3.4 CIP, Lima, Peru.
Ross, H. 1986. Patato Breeding - Problems
and Perspectives. Verlag Paul Parey,
Berln and Hamburg, Germany. P. 132.
Shaw, R. and R. Booth. 1982. Simple
processing of dehydrated potatoes and
patato starch. lnternational Patato
Center, Lima, Per. 32 p.
Zhang, L. and J.F. Guenthner. 1998.
Opportunities in China's processed
patato market. American Journal of
Patato Research 75(6):303-304.

CIP Program Report 1999 - 2000

195

Selection of Parental Lines Using Stability Analysis of


Hybrid True Potato Seed Families Produced Through
Line x Tester Method
M. D. Upadhya1 and R. Cabello1

The line x tester method is preferred for the evaluation and selection of
parental lines for hybrid true potato seed (TPS) varieties where propagation is
sexual and the hybrid TPS is to be used for raising a commercial potato
(So/anum tuberosum L.) crop. Hence, the line x tester procedure was used to
generate two sets of hybrid TPS families. The first set comprised two standard
male lines from the andigena group (S. tuberosum subsp. andigena) crossed
with 39 new female lines from the tuberosum group (S. tuberosum subsp.
tuberosum) to produce 78 hybrid families. The second set comprised nine new
andigena male lines crossed to six standard tuberosum female lines to produce 54 hybrid families. These hybrid families were evaluated under four
contrasting thermo-photoperiod environments: winter and spring, and summer
and autumn seasons at CIP research stations in La Molina (240 m) and
Huancayo (3280 m), respectively, during 1999-2000. The C1 seedling tubers
were also grown at a late blight (LB) hot spot in Comas (2650 m) to evaluate
their reaction to LB caused by Phytophthora infestans (Mont.) de Bary and the
effect of LB on yield. This is the first study in potato using the line x tester
method and analysis of the data using stability characters for the selection of
parental lines. The results show the approach is useful for selection of potential parental lines for the production of stable hybrid TPS varieties. However,
this selection procedure needs to be validated by multilocation trials in subsequent years with direct seedling transplants from hybrid TPS families
involving the selected female and male lines.
The first evaluation involving a heterozygous tester was suggested in maize
breeding to provide a measure of general
combining ability (GCA) in line x tester
studies, where a group of new lines were
available for testing but there was no plan
for using them in a specific hybrid combination. The lines remaining after selection
for GCA could then be tested further in
single crosses to obtain information on
1

CIP, Lima, Peru.

performance in specific combinations


(Sprague, 1955). Patato breeders have
adopted this procedure for determining
GCA of a parental clone to assess its
gametic input. Basically, in selecting
clones for clonal propagation, specific
combining ability has been considered
more important than GCA in crossing
schemes (Bradshaw and Mackay, 1994).
However, the scheme suggested for maize
is more applicable to the selection of
parental lines for hybrid TPS families in
CIP Program Report 1999 - 2000

197

which propagation is sexual and the hybrid


TPS is to be used to raise a commercial
potato crop. Hence, the line x tester
scheme is being used to evaluate and
select parental lines for the production of
hybrid TPS families (varieties) at CIP.
Upadhya and Cabello (1997) used the
STABLE program of Kang and Magari
(1995) to show that this program can help
in selecting high-yielding and stable
hybrid TPS families in short-term trials.
They also showed that the STABLE program could evaluate parental lines for
their contributions to high-yielding stable
families. In that study five standard male
lines were crossed with a number of
selected female clones in two sets to
generate hybrid TPS families using the line
x tester model. Those families were then
evaluated under four contrasting thermophotoperiod conditions by growing them
over four seasons (winter, spring, autumn,
and summer) at two altitudes. Analysis of
tuber yield using the STABLE program
allowed the selection of female lines that
contribute to yield and stability. That
encouraged us to generate two new sets of
hybrid TPS families and to use the STABLE
program to simultaneously evaluate and
select families and their parental lines for
various plant characters.
The present investigations were designed
to evaluate male parental lines selected
from the andigena group and the female
lines selected from the tuberosum group,
fol lowing the line x tester design.
The first set comprised two standard
andigena male lines crossed to 39 new
tuberosum female lines to produce 78
hybrid families. The second set comprised
nine new andigena male lines crossed to 6
standard tuberosum female lines to
produce 54 hybrid families.
Res u lts are given here for the hybrid
families and their parental lines evaluated
for plant vigor, plant and tuber uniformity,
and total and marketable yields, as well as
for their reactions to LB.

198

Research on Potato

Materials and Methods


The trials were carried out over four
seasons: winter and spring, at La Molina
(240 m) and summer and autumn at
Huancayo (3280 m). The general procedure outlined by Upadhya and Cabello
(1997) was followed for raising the seedlings and laying out the trials as well as for
crop management.
From the winter trial conducted at La
Molina, one tuber (30-40 g) from each of
the seedlings harvested in each replication
from every family was taken and bulked
by replication to represent every genotype
of a family. These C1 seedling tubers were
then planted at Comas (2650 m) during the
summer season to test their reactions to
LB. The crop was grown under rainfed
conditions following standard agronomic
practices for the area with respect to
fertilizer application and crop management. The crop was given only a single
prophylactic spray of contact fungicide 25
days after planting. Beginning at 35 days
after planting, weekly observations were
taken for six weeks to calculate area
under the disease progress curve (AUDPC).
The haulms were cut 90 days after planting and the tubers were harvested 15 days
later. Data were recorded on total and
marketable yields/plot; these were later
extrapolated to t/ha.
All the data were analyzed using the
STABLE program to calculate the stability
index for plant vigor, fol iage and tuber
uniformity, total and marketable yields,
and for simultaneous selection of male and
female lines. The SAS program (SAS 1989)
was also used to compare the performance
of the new male and female clones used
in this study for plant vigor at 45 days after
transplanting, plant and tuber shape
uniformity, and total and marketable
yields.
The data on yield, AUDPC for LB, and
GCA from the trial at Comas were analyzed using the SAS program (SAS, 1989).

Results
The performance of the nine new male
lines for the five characters calculated
from the four trials is given in Table 1. All
the males showed values less than that of
Desire for al 1 five characters. The proportional contributions of the male lines,
female lines, and of the interactions were
also calculated (Table 1). The contribution
of the female testers was much higher for
plant vigor and uniformity, and for marketable yield compared with males and the
male x female interactions. The males
seem to be contributing slightly higher to
the uniformity of tuber shape than the
females, but significantly higher than their
interactions. Si mi larly, the contribution of
the males for total tuber yield was higher
than that of the females but less than the
contribution made by their interactions.
The stability analysis and the simultaneous
selection of the male lines are presented
in Table 2. Only one male, C96LB-13.3,
was selected for al 1 five characters as was
the check Desire, indicating that this
male line performed best with all six
female testers studied. The male line
C94H-07.3 (no. 5) was selected for highest
stability values for plant vigor and unifor-

Males

1
2
3
4
5
6
7
8
g
1o

C95LB-22.7
C96LB-13.3
C96LB-67.4
C96L-12.19
C94H-07.3
C94H-07.6
C95T-07.3
C95Hl-07.3
C95Hl-08.3
Dsire (check)
LSD 0.05

8.1
7.7
7.6
7.3
7.4
8.3
0.1

CV';b
Male (1)

6.6
7.89

Proportional
contribution

Female (t)

(%)

Male x female (1 x t)

1 Scale: 9

Plant vigor1

7.5
7.7
7.5

7.7

14.18

4.21

mity, whereas the line C95Hl-07.3 was


selected for tuber uniformity and total
tuber yield. The analysis of the data on the
reaction to LB and total tuber yield as wel 1
as GCA values for the 9 male lines is
presented in Table 3. The male lines
C95LB-22.7 and C95Hl-08.3 gave highest
tuber yield but they were still lower than
the check Amarilis. The GCA values were
non-significant for both yield and LB
reaction. However, the positive value of
8.32 for total yield was highest for the line
C95LB-22.7, indicating it to be a good
general combiner for tuber yield. Both
these lines also had lower AUDPC values
than the check and had the h ighest GCA
values. The contribution of the male
testers was much higher for yield and
resistance to late blight (AUDPC values),
compared with females and the male x
female interactions. However, males
seem to be contributing slightly higher
to yield and resistance to LB than their
interactions.
The data on the 39 female lines were
compared with the check Desire (Table
4). There were significant differences
among the female lines and the variances
were significant for all five characters. A
few of the female lines showed vaiues for

Uniformity score (avg.)1


Plant
Tuber shape
7.3
5.9
7.4
6.3
7.4
6.2
7.3
6.2
7.8
6.1
7.5
6.0
7.5
6.2
7.2
6.7
7.3
6.2
8.7
8.0
0.2
0.2

8.3
2.63
9.61
2.62

8.6
0.24
9.34

5.43

Total y1eld
(Vha)
16.96
18.47
17.42
18.33
18.20
17.03
18.49
19.25
16.07
28.32
0.87

Marketable
yield (%)
70.51
78.28
78.37
77.06
77.67
79.47
73.35
77.15
70.42
96.02
1.57

17.08
6.01
4.13
8.30

7.27
13.27
23.47
7.56

= excellent, 1 = poor.
CIP Program Report 1999 - 2000

199

plant vigor similar to that of Dsire, but


far the other faur characters the values
were significantly less than that of
Dsire.
The calcu lations far the proportional
contributions of female lines, male testers,

and thei r i nteractions far the five characters revealed that (1) contributions of the
female lines were the highest far plant
vigor, unifarmity far plant and tuber shape,
and far total yield and (2) the contribution
far marketable yield was highest far the
male testers.

Table 2. Stability comparison far plant vigor, plant uniformity, tuber shape uniformity, total yield, and
marketable yield of new putative males of hybrid TPS families with six standard females evaluated at La
Malina and Huancayo, Peru, 1999 and 2000.
Tuber shape
Total yield
Marketable
Plant vigor
Plant
Male lines
No.
yield (%)
uniformity
uniformity
(Vha)
-1
-3
-1
-5
C95LB-22.7
2
1
C96LB-13.3
2
6+
7+
6+
8+
8+
4
-3
C96LB-67.4
3
4
3
7+
C96L-12.19
2
4
4
9+
5+
5+
-4
C94H-07.3
2
5
11 +
10+
5+
. C94H-07.6
o
o
6
5+
7+
10+
C95T-07.3
4
1
7
5+
6+
7+
C95Hl-07.3
o
o
3
11 +
8
10+
-1
-2
C95Hl-08.3
1
2
3
9
Dsire (check)
10
10+
13+
5+
13+
5+
**
**
**
**
**
Males
**
**
**
**
**
Environments
**
*
ns
ns
ns
1nte ractio n
**
Heterogeneity
ns
ns
ns
ns
**
Residual
ns
ns
ns
ns
0.25025
0.37922
0.28708
Pooled error
9.327
30.597
Notes: + = selected males, * = significant at 0.05, ** = significant at 0.01, ns = non-significant.
Table 3. Evaluation of 9 new male lines of hybrid TPS families with 6 standard female lines (line x tester) as
F1 C1 tuber families far total yield), late blight score (AUDPC), and general combining ability (GCA) at
Comas, Peru, 1999-2000.
No.
Males
Yield (Vha)
GCA
AUDPC
GCA
C95LB-22.7
1
28.57
8.32
736.75
-609.21
C96LB-13.3
2
20.74
0.49
1440.25
94.29
C96LB-67.4
3
13.92
-6.33
1929.08
583.12
C96L-12.19
4
21.88
1.63
1213.33
-132.63
C94H-07.3
5
21.86
1.61
1152.86
-193.10
C94H-07.6
-1.22
19.03
1350.42
6
4.46
7
C95T-07.3
15.29
-4.96
1884.94
538.98
C95Hl-07.3
14.81
-5.44
8
1744.24
398.28
C95Hl-08.3
-5.91
9
26.16
683.86
-662.10
10
Amarilis (check)
39.22
956.67
LSD 0.05
2.38
157.60
se(g)
1.80
119.16
2.55
168.51
se(Q - g.)1
CV%
17.80
17.71
Proportional Males (1)
50.62
60.05
contribution Females (t)
22.69
24.96
Males x Females (1 x t}
15.62
(%}
11.80
AUDPC = area under the disease progress curve.

200

Research on Potato

Table 4. Performance comparison of 39 new female lines of hybrid TPS families with 2 standard males
(line x tester design) evaluated at La Malina and Huancayo, Peru, 1999 and 2000.
Plant vigor1 Uniformity score {avg.)1
Fema les
No.
Total yield
Marketable
yield (%)
{t/ha)
Tuber shape
Plant
C95LB-02.8
7.7
7.6
6.7
18.48
69.25
1
C95LB-03.19
6.7
17.71
2
7.6
7.7
76.23
C95LB-08.3
7.7
7.7
6.7
17.54
81.04
3
C96LB-35.9
6.8
14.03
72.44
7.4
4
7.4
75.62
C96LB-35.B.4
7.3
7.0
6.6
16.32
5
76.62
6.7
16.11
C96LB-35.B.6
7.4
7.4
6
80.41
C96LB-50.B.2
7.5
7.1
19.03
7.6
7
20.74
79.68
C96LB-59.2
6.9
6.8
7.2
8
6.5
15.91
74.70
C96LB-59.12
7.0
6.7
9
6.8
17.74
80.78
C96LB-63.9
7.2
7.3
10
17.80
C96LB-63.B.3
7.0
6.8
6.8
75.50
11
C96LB-68.1
6.8
6.3
15.30
74.08
7.2
12
16.93
80.16
6.8
C96L-34.4
8.0
7.8
13
79.07
6.7
17.90
C96L-34.7
8.0
7.9
14
6.8
15.96
77.18
C96L-35.3
7.3
7.5
15
7.2
18.37
74.26
C96L-45.6
7.6
7.3
16
7.1
20.33
77.24
C96L-46.3
7.7
7.4
17
16.61
C96L-57.1
7.8
7.5
6.7
77.40
18
C97L-01.3
6.8
7.0
16.78
77.87
7.2
19
7.0
6.9
17.56
79.41
C97L-01.4
7.1
20
6.8
18.17
72.93
C97L-04.2
7.4
7.2
21
19.08
79.06
7.7
7.2
C96T-02.1
7.6
22
76.34
6.9
19.62
C96T-02.9
7.9
7.8
23
6.7
16.89
78.73
C95H2-02.2
7.8
7.6
24
18.59
76.62
6.8
7.6
C95H3-16.4
7.7
25
6.9
18.77
76.86
7.8
C95H3-16.5
7.8
26
6.8
17.07
76.24
C95H3-19.1
7.7
7.5
27
6.8
16.92
76.14
C95H3-19.11
7.4
7.5
28
6.8
16.68
73.29
C95H3-19.13
7.7
7.6
29
7.0
17.16
75.49
C95H3-19.16
7.5
7.6
30
19.12
77.57
7.9
6.9
C95HL-08.1
7.8
31
6.8
15.67
73.66
7.6
C95HA-01.2
7.4
32
7.2
18.17
71.94
7.5
7.5
C96H-01.3
33
17.41
76.90
7.6
6.9
C96H-01.6
7.6
34
17.81
73.32
6.9
C96H-02.4
7.4
7.6
35
21.04
76.33
7.0
C96H-02.7
8.0
7.7
36
7.1
16.61
77.13
7.7
C96H-04.7
7.8
37
6.9
17.46
71.76
7.6
C96H-08.8
7.7
38
6.6
16.32
73.77
7.6
C96H-10.2
7.4
39
27.16
96.72
8.5
7.9
Dsire tuber (check)
8.1
40
0.3
1.42
2.58
0.3
LSD 0.05
0.2
8.2
16.52
6.91
cv (%)
6.7
8.5
Proportional contribution (%)
13.66
6.40
12.79
10.6
Female (1)
11.09
0.72
27.33
0.15
0.14
0.16
Male (t)
5.49
1.27
2.30
1.36
1.74
Female x male {I x t}
1 Scale: 9

= excellent, 1 = poor.

GIP Program Report 1999- 2000

201

The stability analysis and simultaneous


selection of the female lines are shown in
Table 5. Only three lines, C96T-02. l,
C95H3- l 6.4, and 16.5, showed stability for
al 1 five characters and selected for al 1 of
them as was the check Desire. There
were, however, a number of female lines
that showed high stability and were
selected for total and marketable yields,
among them line numbers 3, 7, 8, 1O, 14,
and 17. Only a few were selected for total
yield: lines 1, 2, 11, 16, 21, 23, 33, 35,
and 36. Therefore, these female lines
could be exploited through the extraction
of parthenogenetic tetraploids to bring in a
higher order of homozygosity for characters such as plant and tuber uniformity.
The data on total yield and the reaction of
the female lines to LB were analyzed by
comparing their GCA values for the two
characters (Table 6). The total yields of all
the female lines were significantly lower
than that of the check Amarilis. However,
total yields varied from 10.60 to 28.84
t/ha. Also a number of them had much
lower AUDPC values than the control
Amarilis, indicating a higher order of
resistance to LB than the check.

Discussion
The development of parental lines poses
no problem equal in complexity to that
involved in the evaluation of lines. The
final evaluation of even the most carefully
selected lines must rest upon their performance in hybrid combinations. At CIP,
Mendoza (1985) studied GCA of 20 highyielding clones mated to two testers using
the line x tester model and evaluated
them under three environments for yield,
tuber uniformity, and seedling survival. He
suggested that to use TPS in commercial
potato production, it is important to
identify parental clones with high GCA for
those three characters. However, wel 1defined approaches for the evaluation and
selection of parental clones were lacking.
The studies reported by Upadhya and
Cabello (1997) marked the beginning of

202

Research on Patato

approaches for the selection of parental


lines based on stability studies of hybrid
TPS families under multilocation and
season.
In general, the results from multiple trials
clearly indicate that the seedling transplanted crop from the two sets of hybrid
TPS families did not give as high total and
marketable yields in 90 days as that from
Dsire grown from basic tuber seed.
Hence, the first priority in the breeding
and selection of parental lines at CIP
should be for early bulking, high yields,
and larger tuber size 90 days after seedling
transplanting. The selected parental lines
should be able to transmit these characteristics to the hybrid famil ies to compete
with the clonal ly propagated potato
variety.
In the stability comparison for the male
lines, only line C96LB-13.3 was selected
for stabi 1ity for al 1 five characters. The
stability values for tuber shape uniformity
and marketable yield were higher than
that of the check Dsire (Table 2). This
male line then could contribute a higher
order of stability to tuber shape uniformity
and marketable yield to the hybrid families. lt will be worthwhile to use male
lines such as C94H.07.3, which had high
values for plant vigor and plant uniformity;
C94H-07.6, with highest values for marketable yield; and C95Hl-07.3, with highest
values for tuber shape uniformity and total
yield, to generate hybrid progenies for the
selection of lines combining these characteristics.
The data from Comas on the reaction to LB
and total yield for the male lines show that
line C95LB-22.7 had the highest yields of
the male lines tested (Table 3), followed
by C95Hl,.08.3. Both these lines also
showed AUDPC values lower than the LBresistant variety Amarilis used as the
check. Both lines also showed highest
GCA values for resistance to::LB, indicating a higher transmission of resistance to
the hybrid TPS families.

Table 5. Stability comparison for plant vigor, plant uniformity, tuber shape uniformity, total yield, and
marketable yield of new females of hybrid TPS families with two standard males evaluated at La Malina
and Huancayo, Peru, 1999-2000.
Tuber shape
Total yield
Plant
Marketable
Plant vigor
No. Female lines
uniformity
uniformity
yield (%)
(t/ha)
1 C95LB-02.8
-2
3
26+
29+
32+
2 C95LB-03.19
15
9
22+
30+
23+
C95LB-08.3.
3
5
28+
31 +
41+
19+
-2
15
12
2
10
4 C96LB-35.9
7
4
3
6
14
5 C96LB-35.B.4
4
12
11
5
6 C96LB-35.B.6
20+
18
9
36+
27+
39+
7 C96LB-50.B.2
16
4
3
8 C96LB-59.2
36+
36+
-1
-10
o
-5
11
9 C96LB-59.12
6
9
16
10 C96LB-63.9
20+
40+
o
o
12
11
C96LB-63.B.3
15
20+
-1
o
3
1
9
12 C96LB-68.1
14
16
12 C96L-34.4
38+
38+
36+
8
20+
34+
14 C96L-34.7
40+
40+
2
17
9
21+
27+
15 C96L-35.3
6
4
38+
31+
16 C96L-45.6
21+
8
17 C96L-46.3
26+
37+
39+
28+
11
5
19
30+
18 C96L-57.1
33+
9
5
2
33+
31+
19 C97L-01.3
17
1
5
24+
C97L-01.4
35+
20
3
14
7
11
29+
21
C97L-04.2
36+
33+
22+
32+
38+
22 C96T-02.1
19
C96T-.02.9
28+
38+
23
37+
34+
4
9
20
32+
24
C95H2-02.2
38+
33+
20+
24+
20+
25 C95H3-16.4
30+
34+
25+
27+
34+
36+
26 C95H3-16.5
14
16
10
16
27 95H3-19.1
32+
13
16
12
12
18
C95H3-19.11
28
10
5
16
27+
29 C95H3-19.13
30+
15
8
16
23+
32+
C95H3-19.16
30
18
25+
31+
C95HL-08.1
34+
40+
31
7
1
11
24+
C95HA-01.2
32
2.0+
1
22+
17
38+
C96H-01.3
20+
33
17
24+
24+
C96H-01.6
22+
34
22+
4
25+
10
28+
C96H-02.4
24+
35
16
34+
41+
40+
32+
36 C96H-02.7
4
26+
35+
C96H-04.7
34+
37
36+
-4
18
C96H-08.8
27+
28+
38
29+
8
2
7
8
24+
C96H-10.2
39
35+
39+
43+
43+
Dsire (check)
34+
40
**
**
**
**
**
Females
**
**
**
**
**
Environments
*
**
*
ns
ns
lnteraction
*
**
ns
ns
ns
Heterogeneity
*
**
ns
ns
ns
Residual
27.842
8.435
0.31851
0.4014
0.5072
Pooled error
Notes: + = selected females, * = significant at 0.05, ** = significant at 0.01, ns = non-significant.

CIP Program Report 1999 - 2000

203

The calculations far the estimation of


proportional contribution to the total yield
and AUDPC values presented in Table 3
indicate that male lines had more than a
50% share for both the characters. The

contribution of line x tester interactions far


both the characters was rather low;
whereas the values for the female testers
were about half of those for mal e 1ines.

Table 6. Evaluation of new female lines of hybrid TPS families with 2 standard male lines (line x tester) as F1
C1 tuber families far total yield, late blight score (AUDPC), and general combining ability (GCA) at Comas,
Peru, 1999-2000.
GCA
AUDPC
GCA
Female Lines (no.)
Yield (t/ha}
No.
825
-181
C95LB-02.8
20.80
2.24
1
-343
22.11
5.55
663
C95LB-03.19
2
-2.69
1041
35
C95LB-08.3
15.87
3
-7.96
1215
209
10.60
4
C96LB-35.9
-5.45
1443
437
C96LB-35. B.4
13.11
5
-5.92
1314
12.64
308
C96LB-35.B.6
6
-224
24.84
6.28
782
7
C96LB-50.B.2
879
-127
C96LB-59.2
26.40
7.84
8
-0.12
-151
C96LB-59.12
18.44
855
9
13.64
-4.92
1456
450
C96LB-63.9
10
-2.24
1470
464
C96LB-63. B.3
16.32
11
-5.10
13.46
1561
555
12
C96LB-68.1
20.28
1.72
868
-138
13
C96L-34.4
20.91
2.35
926
-80
C96L-34.7
14
-5.08
C96L-35.3
13.48
1196
190
15
-4.95
C96L-45.6
13.61
1176
170
16
-4.21
C96L-46.3
14.35
1230
224
17
C96L-57.1
15.01
-3.55
988
-18
18
-4.68
C97L-01.3
13.88
1594
588
19
-5.36
C97L-01.4
13.20
1352
346
20
-4.96
1444
C97L-04.2
13.60
438
21
C96T-02.1
24.41
5.85
668
-338
22
-126
C96T-.02.9
20.83
2.27
880
23
C95H2-02.2
23.94
4.62
755
-251
24
20.29
1.73
957
-49
C95H3-16.4
25
21.14
2.58
770
-236
C95H3-16.5
26
416
-590
24.09
5.53
27
C95H3-19.1
-0.08
1173
167
C95H3-19.11
18.48
28
10.28
-671
C95H3-19.13
28.84
335
29
18.03
-0.53
959
-45
C95H3-19.16
30
15.38
-3.18
1183
177
C95HL-08.1
31
691
-315
C95HA-01.2
21.89
3.33
32
26.68
8.12
609
-397
C96H-01.3
33
-326
21.86
3.25
680
34
C96H-01.6
C96H-02.4
26.43
7.87
537
-469
35
28.55
9.99
554
-452
C96H-02.7
36
-7.73
C96H-04.7
10.83
1282
276
37
11.29
-7.27
C96H-08.8
1507
501
38
-3.53
C96H-10.2
15.03
1019
13
39
Amarilis (check)
34.46
1126
40
3.94
271
LSD 0.05
1.99
137
se@)
2.82
se@- Q)1
194
18.6
CV%
24
AUDPC = area under the disease progress curve.

204

Research on Patato

In comparing the analyses presented in


Tables 2 and 3 for the male lines, only the
male line C96LB-13.3 was selected for
stability for all five characters (Table 2).
Yet it did not show high total yield, a
higher order of resistance to LB, or high
GCA values (Table 3) in Comas. The two
lines showing highest total yield and high
resistance to LB (Table 4) were not selected for stabi 1ity of the five characters
under the four environments (Table 3).
Therefore, a much higher number of male
lines must be evaluated to find !ines that
can be selected for stability, total yield,
and LB resistance, valuable attributes for
hybrid TPS families.
The calculations for proportional contribution of the female lines to the five
characters (Table 4) was much higher than
that of the line x tester interactions,
indicating a higher gametic input by the
female lines than the interaction of
genotypes of lines and testers. The proportional contribution of male testers was
very low except for marketable yield,
which was twice that of female lines,
indicating a higher gametic input.
The stability comparisons presented in
Table 5 indicate that only three female
lines were selected for all the five characters. However, there are lines that have
been selected for three or four out of five
characters.
The first set of crosses is being made to
generate hybrid TPS families involving the
male line C96LB-13.3 (Table 2) with the
selected female lines, numbers 22, 25, and
26 (Table 5) based on stability for all five
characters. Multilocation trials will be
conducted at the two sites to assess the
usefulness of this approach for the selection of parental lines.
Another set of crosses is being made using
male lines such as C94H.07.3 (high plant
vigor and plant uniformity), line C94H07.6 (highest marketable yield), and line
C95Hl-07.3 (highest tuber shape uniformity
and total yield) crossed with female lines

that show low stability values for the


respective characters. These hybrid TPS
families will be evaluated under the four
environments to see if the two complement one another to transmit their
respective gametic values to the families,
combining the best values of both male
and female lines.
In comparing the analyses presented in
Tables 5 and 6 for the female lines, only
one female line, C96T-02.1, comes out as
the best by having been selected for
stability in all five characters (Table 6),
and high total yield, a higher order of LB
resistance, and high GCA values (Table 6).
The hybrid TPS family from this female
line with the selected male line C96LB13.3 from the first set has to be evaluated
under multilocation/year trials to validate
the procedure of parental line selection
used in the present investigation.

Conclusions
The results of this study show that the line
x tester approach is useful for selecting
parental lines from a group of new lines
available for testing, but without any plan
for their use in a specific hybrid combination design for the production of hybrid
TPS varieties. However, this selection
procedure has to be validated by conducting multilocation/year trials with direct
seedling transplants from hybrid TPS
families involving the selected female and
male lines.

References
Bradshaw, J.E. and G.R. Mackay. 1994.
Breeding strategies for clonally
propagated potatoes. In: Bradshaw,
J.E. and G.R. Mackay (eds.). Potato
genetics. CAB lnternational,
Wallingford, UK. p. 467-498.
Kang, M.S. and R. Magari. 1995. STABLE:
A basic program for calculating stability
and yield-stabi 1ity statistics. Agronomy
Journal 87:276-277.
Mendoza, H.A. 1985. Selection of uniform
progenies to use TPS in commercial

CIP Program Report 1999 - 2000

205

potato production. In: lnnovative


methods for propagating potatoes.
Report of the 2B1h Planning Conference,
lnternational Potato Center, Lima, Peru.
p. 5-16.
SAS. 1989. SAS lnstitute lnc., SAS/STAT
User's Guide, Version 6 Fourth Edition,
Volume 2. Cary, NC, USA. 943 p.
Sprague, G.F. 1955. Corn breeding. In:
Sprague, G.F. (ed.). Corn and corn

206

Research on Patato

improvement. Academic Press, NY,


USA. p. 221-292
Upadhya, M. and R. Cabello. 1997.
Simultaneous selection for yield and
stability in hybrid true potato seed
families. In: Program Report 1995-1996,
lnternational Potato Center, Lima, Peru.
p. 214-220.

lnfluence of Seed Size and Density on the


Performance of Direct Seedling Transplants from
Hybrid True Potato Seed
M.O. Upadhya and R. Cabello1

The genotype by environment interactions affe.cting the quantity and quality


of hybrid true potato seed (TPS) were evaluated. Sixteen cross combinations
were tested to select parental lines and cross combinations for stability for
berry and TPS characteristics across environments. CIP facilities in Peru, La
Molina (coastal plains at 280 m) and Huancayo (highland valley at 3285 m),
permit hybridization and TPS production in two environments within a year.
In La Molina, the winter season's short photoperiod was artificially extended
to 16 hours to promote blooming. In Huancayo, the photoperiod during summer is adequate to promote profuse blooming. The effects of seed size and
density strongly suggest a high correlation between seed size and yield.
Therefore, in the selection of hybrid TPS combinations, parental lines that
give a higher proportion of large (1.6 mm) seeds should be selected for field
trials in the development of TPS for use in potato (Solanum tuberosum L.)
production.
Studies conducted by CIP and col laborating national agricultura! research systems
on the economics of potato production
from TPS have shown that crops raised
from direct seedl ing transplants give the
highest cash returns (Khatana et al., 1996).
Therefore, CIP efforts have been directed
toward the production of hybrid TPS
families that will give high tuber yields
when a crop is raised through directly
transplanted seedlings. Studies have also
shown considerable variability in the
proportion of different fractions based on
size and embryo structure in a sample of
hybrid TPS produced under different
agroclimates and production protocols
(Upadhya et al., 1984). Furthermore, the
proportion of different fractions is also
affected by the genotypes and nutrition of
the maternal parents (Upadhya et al., 1984,
1985 Thakur, 1987 Pallais and Espinola,
1992; Thakur and Upadhya, 1994).
1

Thakur (1987), Bhatt et al. (1988, 1989),


and Thakur and Upadhya (1994) showed
that TPS size affected germination,
seedling vigor, and final yield of the
seedling transplanted crop. Large seeds
(1.6 mm) gave higher germination,
produced more vigorous seedlings, and
gave significantly higher tuber yield than
small (<1.4 mm) seeds. Studies done at
CIP also indicate that high density TPS
gives higher germination and more
vigorous seedlings (R. Falcan and
N. Pallais, CIP, Lima, Peru, 1999, pers.
comm.). Dayal et al. (1984) showed
positive phenotypic and genotypic correlations of 1000 TPS weight with total tuber
yield, harvest index, and other plant
characteristics in their study on 76 hybrid
TPS families grown as seedling transplants.
However, field studies have not been
conducted to establish higher yield
potential of high-density compared to lowdensity seeds.

CIP, Lima, Peru.


CIP Program Report 1999 - 2000

207

Therefore, we evaluated the performance


of different fractions of hybrid TPS based
on size and density with respect to seed1i ng emergence nine days after sowing,
plant vigor 45 days after transplanting, and
the final yield of the crop. The data
presented and discussed in this paper
highlight the importance of TPS size and
weight on the field performance of a
hybrid TPS family.

Materials and Methods


The hybrid TPS of the cross Serrana x TS-5
produced at Huancayo in 1998 were stored
in moisture-proof containers at 4.5%
moisture at 30C for one year for the
release of dormancy. Germination was
tested at 27C following the procedure
standardized at CIP (Pallais and Falcon,
1997). The germination of this lot was
>95% in seven days. The lot was separated
into three seed sizes of > 1.6 mm, 1.6-1.4
mm, and 1.4-1.27 mm using sieves with
round holes. Each size lot was then separated into high- and low-density classes
using a continuous air system separator
(CAS, Hoffman, lnc., Albany, NY, USA) at
40.23 CFM/m. These procedures allowed
the TPS lot to be separated into six grades,
three sizes of two densities each.
Seedlings were raised from the six grades
and transplanted in the field following the
procedure detailed by Upadhya and
Cabello (1997). Data on emergence nine
days after sowing were recorded. The trial
was laid out in a randomized block design
with four replications in La Molina during
the winter season in 1999 (May to August).
Data on plant vigor were taken 45 days
after transplanting. The haulms were cut 90
days after transplanting and the tubers were
harvested 15 days later. Data were recorded
from each plot on total tuber yield and
marketable yield (tubers >5 cm). Data
analysis was done using the SAS (1989).

Results and Discussion


Analysis of the data showed that the
interaction between seed size and density

208

Research on Potato

for the characters studied were nonsignificant (Table 1 ), hence the data were
analyzed individually for seed size and
density classes. The results of the analysis
for the effect of seed density and seed size
on the characters are presented in Tables
2a and 2b. Seed density seems to have
sorne effect on emergence but does not
affect plant vigor and total and marketable
yields (Table 2a). Seed size, however,
significantly affected emergence and total
and marketable yields (Table 2b). Plant
vigor does not seem to be affected by seed
size or density at 45 days after transplanting. There is a highly significant correlation
between seed size and emergence
(r = 0.78, P <0.01) and between seed size
and marketable yield (r = 0,44, P <0.01 ),
but only a significant correlation between
seed size and total yield (r = 0.50, P <0.05).
No significant correlation was found
between seed size and plant vigor (r = 0.38,
P >0.05), or between seed size and plant
uniformity
(r = 0.22, P >0.05).
Simmonds (1963) reported higher germination percentages (85-90%) for large and
medium seeds compared with small seeds
(60-65%). Dayal et al. (1984) showed that
the 1000 TPS weight in a hybrid TPS
family was positively correlated with total
yield. Later, Thakur and Upadhya (1994)
reported a highly significant effect of seed
size on seedling emergence and growth in
nursery beds. Seedlings grown from large
seeds had higher survival and total tuber
yield. The higher yield was due to the
production of more tubers and higher
yield/plant. They also showed that seed
size had a positive but non-significant
relationship with average tuber weight and
marketable yield. In the present study,
however, there was a significant effect of
seed size on marketable yield. This could
be due to the fact that Thakur and
Upadhya (1994) studied the effect of seed
size in the open pollinated TPS of TPS-2,
whereas the present study used the hybrid
TPS family from Serrana x TS-5. This
difference could be due to the effect of
hybrid vigor.

Table 1. ANOVA far seed density and size on seedling emergency, plant vigor, total yield, and marketable
yield as seedling transplant crop of Serrana x TS-5 hybrid TPS family at la Malina, 1999.

ANOVA
Emergence DAS 9 days
Plant vigor
Total yield (Vha)
Marketable yield (%)
MSD 0.05
Density(D)
ns
ns
ns
*
Size(S)
ns
**
**
*
DxS
ns
ns
ns
ns
C\1%
2.9
9.0
9.50
2.2
Notes: * = significant at 0.05; ** = significant at 0.01; ns = not significant.
Table 2a. Comparison of seed density on seedling emergence, plant vigour, total and marketable yield in
Serrana x TS-5 hybrid TPS family as seedling transplant crop at la Malina, 1999.
Density
Emergence at 9 days
Vigor 45 DAT
Total yield Vha
Marketable tubers
(%)
(%)
7.6a
17.61a
01
91.5a
80a
88.2b
16.77a
7.5a
02
79a
C\1%
2.9
9.0
9.50
2.2
Notes: Duncan at P= 0.01. 01 =High density; DAT =Days atter transplant; D2=Low density; Vigor: 1=Poor, 9=Excellent.

Table 2b. Comparison of seed size on seedling emergence, plant vigour, total and marketable yield in
Serrana x TS-5 hybrid TPS family as seedling transplant crop at la Malina, 1999.
Seed size
Emergence at 9 days
Vigor 45 DAT
Total yield Vha
Marketable tubers
(%)
(%)
19.2a
8.0a
81.0a
94.4a
S1
79.3b
16.8b
91 .8a
7.4a
S2
16.0b
78.6b
7.2a
83.4b
S3
9.0
9.50
2.2
2.9
C\1%
Notes: Duncan at P = 0.01.
S1: > 1.6 mm; S2: 1.6-1.4 mm; S3: 1.4-1.27 mm.
DAT =Days after transplant; Vigor: 1=Poor, 9=Excellent.

The higher seedling emergence, growth,


and yield potential associated with large
seeds compared with small seeds could be
attributed to their larger storage reserves
and biochemical composition. Seed size
and protein content have been shown to
be related to seedling vigor in wheat and
beans (Res, 1971; Lowe and Res, 1972;
Lowe et al., 1972).
Upadhya et al. (1984) and Thakur (1987)
reported h igher contents of carbohydrates
and soluble proteins in large rather than
smal 1 hybrid TPS. Bhatt et al (1989)
reported a positive correlation between
100 seed weight and contents of total
proteins, total lipids, and phospholipids, as
well as germination percentage and
velocity. Based on these earlier reports and
the present investigation, seed size is an

important quality trait of TPS that could be


used as a basis to assess the productivity of
a hybrid TPS family. This criterion is being
regularly used at CIP to evaluate hybrid
TPS families by determining the proportion of large seeds (> 1 .6 mm). Also, male
and female parental lines are routinely
evaluated for their potential to produce
hybrid TPS with a higher proportion of
large seeds. Parental lines showing high
combining ability and stability for a high
proportion of large seeds are preferential ly
selected for field trials through direct
seedling transplants to assess their
production potential.

Conclusions
The present study suggests a strong correlation between seed size and yield.
CIP Program Report 1999 - 2000

209

Therefore, in the selection of hybrid TPS


combinations, parental lines that give
higher proportions of large seeds (1.6 mm)
should be selected for field trials.

References
Bhatt, A.K., T.C. Bhalla, H.O. Agrawal,
M.O. Upadhya, and N. Sharma. 1988.
Effect of seed size on imbibition and
germination of open pollinated true
seeds of patato. Seed Research 16:1 78182.
Bhatt, A.K., T.C. Bhalla, H.0. Agrawal,
and M.O. Upadhya. 1989. Effect of seed
size on protein and lipid contents,
germination and imbibition in true
patato seeds. Patato Research
32:477-481.
Oayal, T.R., M.O. Upadhya, and
S.N. Chaturvedi. 1984. Correlation
studies on 100 true seed weight, tuber
yield and other morphological traits in
patato (Solanum tuberosum L.). Patato
Research 27:185-188.
Kang, M.S. and R. Magari. 1995. STABLE:
A bas~c program for calculating stability
and y1eld-stability statistics. Agronomy
Journal 87:276-277.
Khatana, V.S., M.O. Upadhya, A. Chilver,
and C.C. Crissman. 1996. Economic
impact of true patato seed on patato
production in Eastern and North-Eastern
India. In: Walker, T.S. and C.C.
Chrissman (eds.). Case studies of the
economic impact of CIP-related
technologies. lnternational Patato
Center, Lima, Peru. p. 139-156.
Lowe, LB. and S.K. Res. 1972. Effects of
environment on the relation between
seed protein and seedling vigor in
wheat. Canadian Journal of Plant
Science 52:1 57-164.
Lowe, LB., G.S. Ayers, and S.K. Ries.
1972. Relationship of seed protein and
amino acid composition to seedling
vigor and yield of wheat. Agronomy
Journal 64:608-61 O.
Pallais, N. and R. Falcan. 1997.
Temperature and moisture affect
dormancy and deterioration of true

21 O

Research on Patato

patato seed during storage. lnternational


Patato Center program report 19951996. CIP, Lima, Peru. p. 221-232.
Pallais, N. and N. Espinola. 1992. Seed
quality as affected by nitrogen during
true patato seed production and
moisture conditions during storage.
American Patato Journal 69:85-93.
Ries, S.K. 1971. The relationship of size
and protein content of bean seed with
growth and yield. Proceedings of the
American Society of Horticultura!
Science 96:557-560.
SAS. 1989. SAS lnstitute lnc., SAS/STAT
User's Cuide, Version 6 Fourth Edition
. Volume 2. Cary, NC, USA. 943 pp.
'
S1mmonds, N.W. 1963. Experiments on the
germination of patato seeds. 1.
European Patato Journal 6:45-60.
Thakur, K.C. 1987. Studies on the
development of true patato seed
technology. PhO Thesis. Himachal
Pradesh University, Shimla, India.
345 p.
Thakur, K.C. and M.O Upadhya. 1994. TPS
characteristics, seedling vigor and yield
relationship in patato. In: Rasco, E.T., Jr.
and F.B. Aromin (eds.). Agronomy, seed
production and processing. Vol. 1.
Proceedings of the Fourth Asan Patato
Association Triennial Conference. CIP,
Lima, Peru. p.120-124.
Upadhya, M.D., K.C. Thakur, A. Juneja,
and M.S. Kadian. 1984. True patato
seed production: Flowering, quality and
economics. In: lnnovative methods for
propagating potatoes. Report of the 2Sth
Planning Conference held 1 0-14 Oec.
1984 at lnternational Patato Center
Lima, Peru. p. 117-147.
'
Upadhya, M.D., K.C. Thakur, and
M.S. Kadian. 1985. lnfluence of
genotype and environment on true
potato seed (TPS) qual ity parameters.
American Patato Journal 62:446.
Upadhya, M. and R. Cabello. 1997.
Simultaneous selection for yield and
stability in hybrid true patato seed
families. lnternational Patato Center
program report 1995-1996. CIP, Lima,
Peru. p. 214-220.

Economic Returns to Research on True Potato Seed in


Vietnam
K.O. Fuglie1, N.T.B. Do 2, C.H. Dao 2, and H.T. Nguyen 2

Between 1993 and 1999, new hybrid progenies of true potato seed (TPS) were
adopted by 100,000 small-scale farmers on 3500 ha in Vietnam, or about 10%
of total potato (Solanum tuberosum L.) area. Most adoption occurred in the
Red River Delta in northern Vietnam during the winter crop season. Hybrid
TPS is estimated to have increased potato yield by an average of 6.8 t/ha, or
75%, compared with old varieties grown from clonal seed tubers. Aggregate
economic benefits from TPS in Vietnam are estimated at about US$1 million
per year. The net present value of the investment in TPS research and extension over 1990-201 O is estimated to be between $0.25 million and $2.92
million, yielding a rate of return to research of 29-42%. TPS is estimated to
have increased net household income of adopters by $11.00/year, or 1.2%.
New sources of improved clonal seed tubers, such as imported seed from
China and new improved Vietnamese varieties, may limit further diffusion of
TPS in Vietnam.

lmproving seed quality has been a key


strategy for increasing patato productivity
worldwide. Most of these efforts have
focused on clonal seed propagation.
However, since the 1970s, CIP and other
agricultura! research institutes have also
worked to develop the practica! use of true
patato seed (TPS) as an alternative seed
technology for farmers. TPS is the tiny
botanical seed found in the small, tomatolike fruits of the patato plant. The
economic advantages perceived to lie in
TPS are lower seed cost and higher yield.
Using tubers for seed diverts about 10% of
the global patato crop from food use to
seed. After adding extra production,
handling, storage, and other costs, clonal
seed can account for 30-70% of purchased
inputs and 15-40% of the value of the
harvested crop in developing countries
(Sadik, 1983). Moreover, viral, bacteria!,
1

CIP, Bogor, Indonesia.


Agricultura! Science lnstitute, Hanoi, Vietnam.

2 Vietnam

and other plant pathogens can be transmitted through tubers to the next generation,
subsequently reducing plant health and
yield. Using TPS avoids both of these
limitations, because no portian of the
useable harvest needs to be diverted for
seed and diseases are much less prevalent
in the botanical seed compared with
vegetatively propagated material (Sadik,
1983). These factors have generated
considerable interest in TPS, especially for
use by poor, limited-resource farmers in
developing countries.
Many of the field applications of TPS so
far, however, have not lived up to this
promise. lt has been difficult to achieve
quality and efficiency in both the production and handling of TPS. Consequently,
the cost of TPS to farmers is not insubstantial and the seed itself is sometimes of
uncertain quality. Furthermore, few
farmers have been able to obtain an
economically viable crop from TPS itself.
GIP Program Report 1999 - 2000

211

Within this sample, there is a mix of


Rather, TPS is usual ly u sed to generate
potato seed being used including TPS
small seedling tubers (< 30 g/tuber) that
are subsequently used as seed to produce a seedlings, TPS seedling tubers, seed from
China and Europe, and new Vietnamese
ware potato crop. However, this adds
clonal releases. Few farmers in the sample
considerably to the cost of using TPS as a
used the old local variety Ackersegen,
seed source. Other disadvantages of TPS
which was dominant in the RRD until
include insufficient uniformity in tuber
recently. The near absence of Ackersegen
characteristics, a relatively long growing
reflects its rapid replacement with better
period, high labor requirements in producsources of seed and new varieties in the
tion, and a need for irrigation. In a review
of economic assessments of TPS in severa!
1990s. Comparing seed productivity solely
on the basis of the farm survey runs the
developing countries, Chilver et al. (1999)
risk that observed differences may be due
found that TPS provided an economically
to individual or locality effects. To avoid
viable alternative to clonal seed only in
potentially confaunding factors, on-farm,
cases where the qual ity and productivity
farmer-managed trials were conducted
of clonal seed were exceptional ly low.
with 16 other farmers in four villages. In
Despite these limitations, TPS has enjoyed
these trials, each farmer grew potatoes
sorne successes. In the mid 1990s, about
using five seed types on a 1O m2 plot (2 m2
100,000 potato farmers in Vietnam's Red
per seed type, single replication). The
River Delta (RRD) adopted hybrid TPS and
trials were farmer-managed in that farmers
have continued to use it for the past
chose the amounts of inputs to use far
several years. About 95% of Vietnam's
each seed type according to their normal
potato production occurs in the RRD
practice. Yield was measured by the
during the cool winter months (November- authors at harvest. The five seed types
February). Potatoes are grown here as a
used in the trials were
short duration crop (90-100 days) follow TPS seedlings (from hybrid TPS 11/67 F1 ),
ing two crops of rice in this irrigated delta
TPS seedling tubers (from F1 CO seed
plain. Between 1993 and 1999, area
planted to TPS progenies in the Delta
produced from hybrid TPS 11/67 F1 the
previous year),
increased to 3500 ha, representing about
clonal seed from China (mostly Mira),
10% of Vietnam's total potato area. In this
Ackersegen (old local clone), and
paper we examine the economics of TPS
a new Vietnamese clonal release (KT-3).
use in the RRD and provide a quantitative
assessment of the social benefits and costs
of TPS in Vietnam.
Results

Materials and Methods


Patato production and cost data were
collected from a survey of farmers in the
Red River Delta during the 1999/2000
patato season. The sample consisted of
120 patato farmers in eight villages in four
provinces of the RRD. The purpose was to
conduct a general needs assessment of
patato farmers in the RRD; specific
questions on TPS and patato seed were
included. lnfarmation was also collected
on household characteristics and income.

212

Research on Potato

Potato costs of production and profitability


by seed sou rce
Table 1 shows the yields and returns to
patato production in a normal year far
each type of seed, according to farmers
interviewed in the survey. These average
yields are somewhat higher than those
obtained during the 1999/2000 season, but
provide a better basis far assessing the
economics of patato seed s0urces in
Vietnam because the 1999/2000 season
was a relatively poor year far potatoes
(Fuglie et al., 2001 ). Highest average

Table 1. Patato production costs, yields, and economic returns by source of seed in the Red River Delta,
Vietnam.
Seed source
TPS nurseries
TPS
Old local clonal lmported clones
New local
and seedling
seedling
variety
from China
clonal variety
transplants1
tubers
(Ackersegen)
(Mira mostly)
(KT-3)
Input quantity (per ha)
Labor (h/ha)
5243
4569
4094
4094
3814
684
1162
Seed (kg/ha)
0.119
1162
1396
TPS seedlings/ha
72716
14.5
Manure (t/ha)
17.7
13.4
13.4
16.3
N (kg/ha)
120
133
136
136
179
84
125
125
P (kg/ha)
86
83
72
82
42
K (kg/ha)
77
82
Farm-supplied inputs (US$/ha)
Land
50
50
50
50
50
Labor
749
653
585
545
585
43
Man ure
61
33
33
39
Total farm-supplied
860
745
667
667
633
inputs
Purchased inputs (US$/ha)
o
o
o
Nursery plastic cover
8
o
102
328
256
256
Seed
308
101
116
116
Total chemical fertilizer
103
105
11
14
1
16
Total pesticide
1
444
428
Total purchased inputs
224
373
373
Output
11.25
8.99
12.50
19.90
Average yield (t/ha)
15.79
6:30:50:15
36:30:27:7
45:39:16:0
47:27:20:6
39:39:20:2
A:B:C:D (%) 2
Average price of
95
85
95
95
104
yield (US$/t)
Economic returns (US$/ha)
1191
2077
Gross value of yield
1070
1336
857
Net income 3
846
893
484
818
1649
1016
Economicprofit4
(13)
147
(184)
151
Source: Farm surveys in Red River Delta during 1999/00 winter season.
TPS = True patato seed.
1 An average of 0.119 kg of TPS sown in 215 m2 nursery beds is required to produce 72, 716 one-month-old TPS
seedlings. The seedlings are then transplanted to 1 ha.
2 Percent size distribution of yield. A = large (> 100 g/tuber), B = medium (50-100 g/tuber), C = small (20-50
g/tuber), O = very small (< 20 g/tuber).
3 Net income is defined as the gross value of production minus the cost of purchased inputs.
4 Economic profit is defined as the gross value of production minus the cost of all purchased and farm-supplied inputs
(land, labor, and manure).

yield was obtained from new Vietnamese


clonal variety KT-3. TPS seedling tubers,
TPS transplants, and Chinese seed all gave
significantly higher yield than Ackersegen,
which produced an average of only 9 t/ha.
Prices received by farmers for their patato
crop were somewhat lower for TPS seed1ing tubers than clonal seed because of the
lower quality yield from TPS (i.e., higher

proportion of low-valued small tubers in


total yield). Note that yield from TPS
seedling transplants is primarily used as a
seed crop for the following season. Thus, it
is relatively high valued despite its large
proportion of small tubers. Taking into
account yield and market price, the gross
value of production was highest for KT-3
(US$2,077/ha (US$1.00 = 14,000 Vietnamese Dong)). Gross value of production for
CIP Program Report 1999 - 2000

213

TPS seedling tubers was significantly lower


at US$1,336/ha but compared favorably
to Chinese varieties (US$1, 191/ha) and
was 56% higher than the Ackersegen
(US$85 7/ha).
Production costs for potatoes grown from
TPS seedling transplants were substantially
below production costs from potatoes
grown from tubers due mainly to the lower
cost of seed. However, yield from TPS
transplants is mainly kept as seed for the
following year. The TPS seedling tubers are
then used to produce a ware crop. The cost
of using TPS seedling tubers was actually
higher than the cost of clonal seed. Other
input costs were similar except that using
TPS transplants required significantly more
labor than using tubers as seed. Thus, TPS
actually offered no savings in production
costs.
Net income from potato production is
defined as the gross value of production
minus the cost of purchased inputs including seed, chemical fertilizer, and
pesticides. Net income is therefore the
return to farm-supplied inputs such as land,
labor bullock services for land preparation,
irrigation water, and manure. We define
economic profit from potato production as
the gross value of production minus the
cost of all inputs, whether purchased or
farm-supplied. Net income provides a
measure of the contribution of potato
production to household livelihood,
whereas economic profit provides a
measure of the relative profitability of
potato production compared with alternative economic activities. Even if net
income is large, economic profit may be
small or negative for a particular kind of
potato seed. In this case, we may expect
the use of this type of seed to decline over
time as farmers switch to more profitable
alternatives. lf economic profit is large,
then we may expect the activity to
increase over time as farmers respond to
this more lucrative economic opportunity.
Net income and economic profit were
highest for KT-3 and lowest for Ackersegen.

214

Research on Potato

The net income and economic profit of


TPS seedling tubers are comparable to
those of Chinese seed and substantially
higher than those of Ackersegen. Producing TPS seedlings generates respectable
net i ncome and about breaks even for
economic profit. This is not surprising
because TPS seedlings are primarily a seed
crop, but production from seedling tubers
yields a market crop. Given the economic
options currently available to RRD farmers, the old local clone Ackersegen is not
economically viable. Although it still
registers positive net income in a normal
year, farmers can get a better return by
allocating farm-supplied resources to other
types of potato seed or to other economic
enterprises. The most profitable potato
seed choice is the new Vietnamese variety
KT-3, which has been multiplied in
farmers' fields since 1996 and was officially released as a variety in 2000. Due
to its inherently slow multiplication rate, it
is still not widely available to potato
growers in the RRD. The high economic
profits earned from this variety suggest that
adoption of KT-3 will spread rapidly as it
becomes more avai lable.
The results of the farm survey and on-farm
trials comparing productivity of alternative
sou rces of potato seed show that when TPS
first became available to farmers in the
RRD, it gave a substantial increase to farm
income over Ackersegen, the predominant
variety at that time. These findings suggest
that in the next few years area planted to
Ackersegen will continue to fall, area
planted to TPS and Chinese seed may
remain steady, and area planted to new
Vietnamese clones will increase. Aggregate potato area in the RRD may also
gradually increase.
Economic rate of return to TPS research
and extension

The increased yield achieved by TPS over


old clonal varieties (namely, Ackersegen)
and the large area planted to TPS generated significant economic benefits for farm
families and the rural economy in the

RRD. The evidence from the farm survey


reported above indicated that hybrid TPS
seedling tubers outyielded Ackersegen by
an average of 6.8 t/ha and increased
economic profit by US$331/ha in a normal
year. Aggregating over 3500 ha, economic
benefits from TPS total US$1.16 million/
year. Potato production increased by 6%,
about 22, 100 t/yr.

To compare aggregate economic benefits


to costs, we constructed estimates of
public investments in hybrid TPS research
and extension in Vietnam since this
activity began in 1990. These included
investments in personnel, materials, land,
and buildings by the Vietnamese Government; research expenditures by CIP; and
funds provided between 1993 and 2000 by
the Asian Development Bank (ADB) for a
special project to promote hybrid TPS.
lnvestment costs are expected to fall after
2000 following the end of ADB support for
TPS research and extension in Vietnam. In
addition, we forecast benefit and cost
streams to 201 O under three alternative
scenarios:
1 . There are no more net benefits from TPS
after 2000 (due, for instance, to the
widespread availability of clona! seed
from China with similar productivity).
2. Benefits from TPS gradually decline to O
between 2001 and 201 O.
3. Benefits from TPS remain constant at
US$1.075 million/yr through 201 O.
In each scenario, we assume that between
2001 and 201 O research is reduced but
extension efforts continue.
Table 2 shows the estimated aggregate
benefit and cost streams together with
estimates of benefit-cost analysis of the
investment in TPS research and extension.
Assuming that there are no additional
benefits from TPS after 2000 (scenario 1),
the present value of the net economic
benefits to Vietnam are US$426,000 (10%
discount rate) or US$250,000 (15%
discount rate). This implies about US$1.301.40 in benefits for every US$1 invested in

research and extension. The other scenarios result in higher estimates of net
present value because there continue to be
benefits from TPS between 2001-201 O.
Under scenario 3, where TPS benefits are
assumed to continue at a constant level
between 2001 and 201 O, the net present
value at the 10% discount rate is US$2.97
million. The rate of return to research and
extension is 28.6% under scenario 1 and
around 40% under scenarios 2 and 3.

lmpact of TPS on family livelihoods


Farm families who adopted TPS retained
most of the economic benefits of TPS
diffusion in Vietnam. Although a 6%
increase in local production may be
expected to put downward pressure on
prices (thereby transferring sorne benefits
to potato consumers), this was cushioned
by the increasing importance of potato
trade with China (Fuglie et al., 2001 ).
Further, a significant share of employment
and income benefits appears to have been
retained by female household members.
During the winter season, there is substantial migration by male adults from rural
areas seeking seasonal off-farm employment. Thus, women predominate in the
production of winter crops in the RRD. An
indicator of the role of women in TPS is
the importance of village women's groups
in providing training and in promoting and
extending TPS technology, which was
observed by the authors during field
research.
Given the small average size of potato
plots (360 m 2 ) and the large number of
farm families who adopted TPS (about
100,000), benefits per household are
estimated to be US$11.00/household per
yr, or US$2.20/person per year). Among
our farm sample, this amounts to an
increase of 1.9% in agricultura! income
(including the value of commodities
consumed at home) and a 1.2% increase
in total income (including income from
non-farm activities). Since much of this
income is earned and controlled by female
household members, a relatively large

CIP Program Report 1999 - 2000

215

Table 2. Benefit-cost analysis of true patato seed (TPS) research and extension in Vietnam (constant
US$'000}.
Aggregate net benefits
lnvestment in
Net economic benefits
Year (season)
to farmers 1
research and
to Vietnam 1
extension
Scen. 2
Scen. 3
Scen. 1
Scen. 2
Scen.3
Scen. 1
(78)
o
o
o
78
(78)
(78)
1990/91
o
78
(78)
(78)
o
o
(78)
1991/92
o
o
(92)
(92)
o
92
(92)
1992/93
o
o
(105)
(105)
o
105
(105)
1993/94
(126)
(126)
132
(126)
5
5
1994/95
5
28
185
(157)
(157)
(157)
28
28
1995/96
13
198
198
198
185
13
13
1996/97
1052
1237
1237
1237
185
1052
1052
1997/98
185
790
976
976
976
790
790
1998/99
1075
1075
1075
185
889
889
889
1999/00
o
967
1075
185
(185)
782
889
2000/01
o
147
928
860
1075
(147)
713
2001/02
o
752
1075
147
928
(147)
606
2002/03
o
645
1075
134
(134)
510
940
2003/04
o
537
1075
134
(134)
403
940
2004/05
o
430
1075
93
(93)
337
982
2005/06
o
322
1075
93
982
2006/07
(93)
230
o
215
1075
2007/08
93
(93)
122
982
o
107
1075
93
2008/09
(93)
15
982
o
o
2009/10
1075
93
982
(93)
(93)
o
o
o
78
2010/11
(78)
(78)
(78)
Present value
1526
2869
10% discount rate
4073
1100
426
1769
2972
15% discount rate
1037
1786
2370
787
250
1584
999
Benefit-cost ratio
1.39
3.70
2.61
(10% discount rate)
Benefit-cost ratio
1.32
2.27
3.01
(15% discount rate)
Interna! rate of return,
28.6%
38.9%
41.6%
1990-201 o
1 Scenario 1: no TPS benefits after 2000; Scenario 2: TPS benefits decline to Obetween 2001 and 201 O; Scenario 3:

TPS benefits constant between 2001 and 201 O.

share of this income was probably devoted


to childcare and other household activities
under the primary responsibility of women.

Summary and Conclusions


Hybrid TPS has played an important role in
meeting farmers' needs for improved
patato seed in Vietnam. lts popularity is
driven by the higher average yield it offers
compared with old, degenerated clonal
varieties.

216

Research on Potato

Further, the diffusion of TPS in Vietnam has


been aided by a strong research and
extension system and village cooperatives
that organize TPS distribution, contract
TPS seedling production, and store a
portian of TPS seedling tubers for use as
seed in subsequent seasons. Nevertheless,
in the new land tenure system now
prevalent throughout Vietnam, the decision to adopt TPS is made by individual
farmers. TPS progenies will continue to be
used in Vietnam so long as they offer

economic returns at least equal to if not


higher than returns from alternative
economic activities.
In the mid- to late-1990s, Vietnamese
potato farmers began to benefit from wider
access to new sources of clonal potato
seed in addition to the new hybrid TPS
progenies. Potato imported from China and
diverted for seed use has become a lowcost seed source, although Chinese seed
quality varies. Seed imported from ~urope
yields well, but is relatively expens1ve a.nd
degenerates quickly under t~op1cal c~nd1tions. New high-yielding, d1sease-res1stant
clonal varieties developed by Vietnamese
breeders from CIP breeding material show
considerable promise. lt appears likely
that area planted with this diverse set of
improved potato varieties and seed sources
will increase in the next several years and
continue to replace the old clonal variety,
Ackersegen. Hybrid TPS will likely
continue to have a role to play in Vietnam
for the next decade, although area planted
with TPS may not expand much further.
TPS suffers from disadvantages such as
higher labor requirements and ~ longer
growing season. Unless further 1mprovements are made in TPS technology to
overcome these constraints, improvements
in clonal varieties and seed systems are
likely to eventually replace TPS in Vietnam.

Acknowledgments
The authors would like to thank Tom
Walker for his suggestions and Elske Van
de Fliert for kindly providing needs
assessment farm survey data for this study.

References
Chilver, A., T.S. Walker, V.S. Khatana, H.
Fano, R. Suherman, and A. Rizk. 1999.
On-farm profitabi 1ity of true potato seed
(TPS) utilization technologies. Social
Science Department Working Paper No.
1999-3. lnternational Potato Center,
Lima, Peru. 41 p.
Fuglie, K., Do Thi Bich Nga, Dao Huy
Chien, and Nguyen Thi Hoa (2001 ). The
economic impact of True Potato Seed in
Vietnam. In: Fuglie, K. (ed.). Performance and Prospects of Hybrid True
Potato Seed in South and Southeast
Asia. Proceedings of the CIP-ADB
Symposium, Field-Testing Hybrid TPS in
the Lowland Tropics of Asia, held at
Bogor, Indonesia, September 13-14,
2000. CIP-ESEAP, Bogor, Indonesia.
p. 151-176.
Sadik, S. 1983. Potato production from true
seed-present and future. In: Hooker,
W.J. (ed.). Research for the potato in the
year 2000. lnternational Potato Center,
Lima, Peru, p. 18-25.

GIP Program Report 1999 - 2000

217

CIP's Contribution to Varietal Change in Potatoes in


Developing Countries
T. Walker, 1 Y.P. Bi, 2 J.H. Li,3 P.C. Gaur, 4 and E. Grande 1

Since its founding in 1971, CIP has invested considerable resources in potato
(So/anum tuberosum L.) genetic improvement. CIP's contribution to varieties
released by national programs and to varieties grown by farmers was assessed from survey data in 30 countries that account for 85 % of
developing-country production. In the 1990s, about 40% of national releases
were directly related to CIP's activities. However, by 1997, CIP-related
releases only accounted for about 6 % of potato-growing area. CI P-related
materials have had wider acceptance in smaller national programs with less
area planted to potatoes than in stronger national agricultura! research
system (NARS). Moreover, CIP-related varieties are still relatively young
because varietal change is substantially slower in potatoes than in cereals
and in most other major field crops. In spite of this seemingly disappointing
performance in adoption, we calculate that the rate of return on investing in
potato genetic improvement at CIP is positive at 15-17%.
Shortly after the founding of CIP, a planning conference was held to establish
priorities and recommend specific programs for the use of the wealth of genetic
resources becoming available to the
institute (CIP, 1974). A few high priority
recommendations, such as improving the
content and quality of protein, were
subsequently dropped and sorne low
priority recommendations, i.e., the use of
sexual propagation by means of true
potato seed, were later emphasized. Sorne
research areas, such as adaptation to very
warm growing regions and resistance to
bacteria! wi lt, caused by Ralstonia
solanacearum, have proven to be technically difficult. Others, such as resistance
to late blight (LB), caused by Phytophthora
1

lnternational Potato Center (CIP), Lima, Peru.


Shandong Academy for Agricultura! Sciences, Jinan,
Shandong, China.
3
Gansu Academy for Agricultura! Sciences, Lanzhou, Gansu,
China.
4
Central Potato Research lnstitute (CPRl-ICAR), Shimla, HP, India.
2

infestans, have been highly successful. In


spite of the dynamic nature of priorities,
varietal resistance, mainly to LB and
viruses, has been and is increasingly
central to CIP's investment in potato
breeding (CIP, 1998).
In this paper, we evaluate CIP's role in
varietal release and adoption in potato
crop improvement in developing countries.
The evaluation is based on information
collected in national and provincial
surveys that are described in the next
section.

Surveys and Sample Countries


The data come from two questionnaires. In
1993/94, potato crop improvement programs in 20 countries were surveyed with
a lengthy questionnaire that elicited
information on potato-producing zones,
released varieties, escapes, seed production, varietal adoption, and scientific

CIP Program Report 1999 - 2000

219

staffing. In 1998/99, a shorter version of


this questionnaire was canvassed in 30
countries. Particular attention was given to
coverage of 14 provincial programs in
China and severa! large Latin American
potato-producing countries where information in the 1993/94 survey was
incomplete.
More than 80 developing countries
produced potatoes in the late 1990s
(FAOSTAT, 1998). lnvestment in potato
breeding and size of production were the
most important criteria in selection of the
study countries. The 30-country sample is
evenly distributed across Asia, SubSaharan Africa, and Latin America; the
study countries account for about 85% of
developing-country potato area and
production.

Varietal Release
Data on varietal release are usually the
most available information on the performance of a plant breeding program.
However, release does not imply adoption
nor is it a necessary condition for success.
The release of Kerr's Pink in Kenya in 1927
marked the first variety listed in the 30country sample. Between 1927 and 1998
about 500 varieties were released. The first
CIP-related variety was released in 1979.
The pace of varietal release has accelerated since the 1950s, when on average
two varieties were released per year across
the 30 countries. During the 1980s and
1990s, average release peaked at 1 7/year.
This rising and then stabilizing trend of
varietal releases reflects the pattern of
investment in public-sector NARS (Alston
et al., 1 998). In many publ ic-sector NARS,
growth in the early 1960s to the early
1980s was followed by a period of stagnation and even decline in public-sector
spending for agricultura! research.
The institutional source of the germplasm
on which released varieties are based can
be divided into three categories: 1)
developing country NARS, 2) CIP, and 3)

220

Research on Potato

developed country NARS. For our study,


these three general categories were further
subdivided into nine groups depending on
the institutional role played in varietal
development. Four categories reflect a
dominant role of the NARS and a negligible role for CIP. These include
NARS-bred varieties with germplasm
unrelated to CIP, NARS-selected varieties
from crosses unrelated to CIP, NARSreleased native varieties, and NARS
borrowing of varieties from other developing country NARS. CIP's role is described
by three categories: 1) developing country
released varieties distributed by CIP to
other developing countries, 2) NARS
selections from CIP crosses, and 3) NARS
crosses from CIP progen itors.
The third major grouping pertains to the
borrowing of developed country released
varieties. Most of these come from Europe,
especial ly the Netherlands, and North
America, mainly the United States.
A residual or other category completes the
classification. Sports (somatic mutations)
of released varieties and varieties attributed to farmers compose this category.
NARS-bred material (unrelated to CIP) has
been the dominant player in influencing
the frequency of varietal release (Table 1).
Developed countries also supply a large
share of patato varieties released in
developing countries. About 23% of the
varieties on the national release list in the
30-country sample are related to CIP.
The importance of these institutional
categories has been changing over time.
Shifts in importance are described for three
roughly equal quantities over three periods
that correspond to pre-CIP, early CIP, and
mature CIP (Table 2). CIP's role has also
changed over time. Before 1978 CIP's
contribution was negligible. In the 1980s,
CIP emphasized the South-South sharing of
NARS-bred material. The sali.ent example
of this borrowing is Achirana-INTA, which
was bred in Argentina. First identified as
promising, it was then pathogen-tested and

Table 1. lnstitutional composition (in %) of


varieties released by NARS 1.
lnstitutional source
Releases (%) 2
NARS alone (without CIP
48.2
involvement)
NARS bred, no CIP role
42.2
NARS selected, no CIP role
2.0
NARS sharing, no CIP Role
1.8
Released native varieties
2.2
Developed country clone,
26.8
NARS released
CIP - NARS partnership
23.5
CIP distributed, NARS released
8.6
CIP cross, NARS selected
13.3
NARS cross, CIP Progenitor
1.6
Others
1.8
Sport3, no breeding ar CIP role
1.4
Farmer ar prvate sector variety
.4
Relea ses
512
1 NARS =

national agricultura! research system.


Does not total 100% due to rounding.
3 Somatic mutations of released varieties and
varieties attributed to farmers.
2

distributed by CIP. lt was subsequently


evaluated and released by severa! developing countries ranging from Bhutan to
China to Madagascar.
Emphasis on the distribution of finished
varieties developed by national programs

in the South was needed to buy time and


deliver sorne short-term impact until CIP's
patato breeding program matured. In the
1990s, more released varieties have been
derived from CIP-bred material, which has
been selected by national programs (Table
2). Selection from CIP-bred material has
been particularly important for smaller
national programs that do not have enough
patato production to justify a full-scale
breeding effort. CIP-bred material is also
being used as parents by larger NARS,
such as India, but, thus far, only a few
varieties have been released from crosses
with CIP parents.

Varietal Adoption
The data on varietal adoption show a very
credible performance by NARS-bred
material (Table 3). About 50% of the
released varieties belong to this category,
and they accou nted for 60% of area
planted to potatoes in the 30-country
sample in 1997.
Shares of released varieties and area
coverage were also congruent for developed country clones with about one
released variety in four and one hectare
planted in four. This correspondence attests

Table 2. lnstitutional composition (in %) of NARS varietal releases over time.


NARS releases 1
lnstitutional source
From 1979-89
Befare 1978
46.2
61.1
NARS alone (without CIP involvement)
42.1
52.7
NARS Bred, no CIP role
4.2
0.6
NARS selected, No CIP role
1.2
2.4
NARS sharing, no CIP Role
2.3
1.8
Released native varieties
26.3
36.4
Developed country clone, NARS released
25.1
CIP - NARS partnership
2.4
17.5
0.6
CIP distributed, NARS released
1.8
6.4
CIP cross, NARS selected
1.2
NARS cross, CIP progenitor
2.4
Others
1.2
Sport2, no breeding ar CIP role
1.2
Farmer ar prvate sector variety
171
165
Relea ses

From 1990-98
38.0
32.4
1.1
1.7
2.8
17.6
41.5
7.4
30.7
3.4
2.8
2.8
176

1 Columns may not sum to 100 dueto rounding.


2

Somatic mutations of released varieties and varieties attributed to farmers.

GIP Program Report 1999 - 2000

221

Table 3. Adoption of patato varieties in 1990s by


institutional source.
Patato growing
lnstitutional Source
area (%)
NARS alone (without CIP
58.8
involvement)
50.6
NARS bred, no CIP role
5.6
NARS selected, no CIP role
1.7
NARS sharing, no CIP Role
0.9
Released native varieties
Developed country clone,
25.0
NARS released
5.9
CIP - NARS partnership
3.2
CIP distributed, NARS released
1.6
CIP cross, NARS selected
1.1
NARS cross, CIP progenitor
3.7
Local varieties
2.5
Native varieties
Old introduced degenerated
1.2
material
Others
6.6
Sport1, no breeding or CIP role
O.O
Farmer or private sector variety
0.1
Others
6.5
Total area (ha)
6,932,615
1 Somatic mutations of released varieties and varieties

attributed to farmers.

to the continuing popularity of these


exotic materials.
In contrast to these first two categories,
CIP's 23% share in varietal releases did
not translate into a comparable rate of
area covered. CIP-related materials were
planted on about 400,000 ha, equivalent
to about 6% of area. About 50% of this
area was planted in CIP-distributed, NARSreleased varieties, and 50% was planted to
selections from CIP crosses and varieties
with CIP progenitors.
Why is varietal adoption of CIP-related
materials substantially lower than CIP's
contribution to varietal release? Again,
part of the explanation resides in the fact
that CIP-related materials have had wider
acceptance in smaller NARS with smaller
area planted to potatoes. Equally important, varietal change is slower in potatoes
than for other major field crops, not only in
North America and Europe but also
apparently in developing countries
(Walker, 1994). For example, the leading
222

Research on Potato

varieties in area planted in China and


India in the 1990s were both released
befare CIP was started.
As with varietal releases, the relative
importance of CIP-related material in
farmers' fields varies markedly by region.
CIP's relative contribution was highest in
Sub-Saharan Africa where about 40% of
the potato-growing area was planted to
CIP-related materials.
One of the surprising aspects of Table 3 is
the low share of local varieties in cultivated area. Native potato varieties
command a large area only in Bolivia and
Peru, and to a much lesser extent in
Colombia and Ecuador.
In closing, we return to the very impressive performance by NARS in potato
breeding that was alluded to at the outset
of this section. The adoption data in Table
3 clearly support an attractive rate of
return on investment in potato breeding in
developing countries. However, there has
been considerable variation in performance across countries and over time
within the same country. Since lndependence in 1947, India has been one of the
few strong NARS to have consistently
released varieties that were regularly
adopted by farmers on a sustained basis.
For other stronger NARS, success has been
more episodic: rapid adoption of firstgeneration varieties followed by a slowing
of varietal change.

Returns to CIP's Potato Breeding


Superficially, the relatively low rate of
uptake of CIP-related material would
appear to translate into a negative or very
low rate of return on investment. However,
that does not appear to be the case.
Our best bet, back-of-the-envelope
calculation shows a positive, albeit
modest, rate of return on investment (Table
4). We assume:
a project duration of 50 yr,
the 1997 patato growing area,

Table 4. _Back of the envelope calculation on returns to CIP's investment in patato breeding.
Assumpt1ons
Period
50 years, 1972-2021
Area
1997, no trend assumed
Prices
Constant (1992) (US$11 O/t)
Source of benefit
Yield increase of 2.5 Vha = US$220/ha
CIP costs
Real; 55% of expenditure on patato
Seed, extension, and ali other research costs
US$11 O/ha
Net benefit
US$110/ha
Results
Adoption ceiling in 2021 (%)
IRR1
NPV 2 (millions US$)
5.8
15
39
10.0
16
51
15.0
17
71
1

IRR

interna! rate of return.

2 NPV = net present value.

a constant price of US$11 O/t,


a yield increase of 2.5 t/ha as a source
of benefit,
a cost share of 55% for patato breeding
as a proportion of total CIP expenditures
on patato crop improvement,
additional costs of NARS research, seed,
and extension equivalent to 50% of net
benefits,
a logistic pattern of diffusion, and
three levels of adoption by 2021 of 5.8,
1O, and 15%.
These assumptions warrant sorne comment
and explanation. First, they are chosen to
establish a reasonable lower boundary or
conservative estmate of the rate of return
to investment. Patato breeding is not
defined narrowly and includes all costs
related to germplasm conservation. The
price of US$110/t is low, the increasing
trend in potato-growing area is not considered, and a reduction of 50% of net
benefit to cover all other research, seed
multiplication, and extension costs is
undoubtedly on the high side. A 2.5 t/ha
yield increase (net of increased cost)
would be high for cereals but is not large
for a varietal replacement effect in
potatoes. The adoption levels at 2021 are
arbitrarily chosen and reflect the current
situation of 5.8% and a doubling and
tripling of the present levels. The per unit
net benefit equivalent to 2.5 t is assumed

to apply to all years when adoption


occurs.
Our illustrative calculation shows that the
interna! rate of return on investment is
clearly positive (ranging from 15 to 17%)
and is quite insensitive to assumptions
about levels of adoption. This result is to
be expected because the i nvestment in
patato breeding is pretty much "a done
deed" in terms of return to capital. In
contrast, the net present value of the
investment, evaluated at a 10% discount
rate, is highly sensitive to the leve! of
adoption at the end of the project, which
reflects the size of the opportunity to be
addressed.

Conclusions
Proportionally and even absolutely, CIP's
impact has been greatest in small NARS.
The adoption of CIP-related patato material in developing countries has also been
modest compared with the performance in
wheat, rice, and maize. Nonetheless, we
show a solid rate of return on investment
of about 15% in CIP's patato breeding
activities. The lack of a very attractive
rate of return to breeding does not imply a
mediocre performance for a patato crop
improvement program as a whole. For a
high-value, vegetatively propagated crop,
there is potential to add value through

CIP Program Report 1999 - 2000

223

agricultura! research in ways that are not


directly related to genetic improvement.
Historically, integrated pest and disease
management, improved seed quality, and
better storage practices have figured
among them.
The prospects are bright that CIP-related
materials will increasingly find a home in
farmers' fields. For example, in Bolivia,
the national program has recently released
severa! LB-resistant varieties that appear
promising for farmer acceptance. Mira, the
most widely grown potato clone in China,
looks like it is ripe for replacement by
selections from CIP disease-resistant
populations made by provincial agricultura! research stations.

Acknowledgments
About 60 NARS and CIP scientists contributed to this project. We are particularly
grateful to Jim Bryan, Charles Crissman,
Peter Ewell, Fernando Ezeta, Francisco
Flores, Osear Hidalgo, Hilario Da Silva
Miranda Filho, and Yi Wang. We thank the
lmpact Assessment and Evaluation Group
of the Consultative Group on lnternational

224

Research on Patato

Agricultura! Research for funding the 1998/


99 survey.

References
Alston, J.M., P.G. Pardey, and
J. Roseboom. 1998. Financing agricultura! research: lnternational investment
patterns and policy perspectives. World
Dev. 26(6):1057-1071.
CIP (lnternational Patato Center). 1974.
Strategy for Utilization. Report of the
lnternational Potato Center's Planning
Conference on Uti 1ization of Genetic
Resources. Lima, Peru. 85 p.
CIP (lnternational Potato Center). 1998.
Working Document on CIP Plant
Breeding Strategies for Patato and
Sweetpotato. Interna! ly Commissioned
Externa! Review (ICER). Lima, Peru. 27 p.
FAO (Food and Agriculture Organization
of the United Nations). FAOSTAT
1998(November). [http://apps.fao.org].
Walker, T.S. 1994. Patterns and implications of varietal change in potatoes.
Social Science Department Working
Paper, No.1994-3. CIP, Lima, Peru. 40 p.

Evaluation of Potato Genotypes Through Pilot-Scale


Farmer Field Schools in the Peruvian Andes
R. Orrego 1 , J.J. Heath 2 , J. Tenorio 3 , C. Valencia3, J.A. Landeo1 , M. Upadhya1,
K.A. Garrett2, N. Zuiga4, A. Mendoza5, A. Parraga6, L. Cotrina3, E. Roncal 4,
M.A. Pacheco 4, O. Ortiz 1, and R.J. Nelson 1

Late blight disease in potato (Solanum tuberosum L.), caused by Phytophthora


infestans (Mont.) de Bary, is often the most important production constraint
for resource-poor Andean potato farmers. To help farmers improve their
disease and crop management, CIP has collaborated with national research
and extension organizations to develop a farmer participatory research (FPR)
and farmer field school (FFS) training method (FPR-FFS). The FFS training
method, developed by the Food and Agriculture Organization of the United
Nations (FAO) for rice integrated pest management in Asia, has been combined with elements of participatory research to provide farmers with
information and to facilitate the evaluation of technology such as diseaseresistant varieties. In this paper we present sorne of the technical results of
experiments conducted by farmer groups participating in FPR-FFS during
1997-2000. Four sets of experiments were conducted: (1) testing of three
varieties with different levels of resistance under three fungicide regimes,
(2) evaluation of available varieties and advanced breeding lines, (3) testing
of a new set of breeding lines, and (4) evaluation of a set of populations
derived from true potato seed (TPS).
Since 1997, CIP researchers have worked
with farmer groups and a group of government and nongovernment organizations
(NGOs) in Peru to test late blight (LB)
management components through farmer
participatory research (FPR) and farmer
field school (FFS). This work is part of a
seven-country program financed by the
lnternational Fund for Agricultura! Development (IFAD) and the Organization of
Petroleum Exporting Countries (OPEC)
Fund for lnternational Development.
1

CIP, Lima, Peru.


Kansas State University, Manhattan, Kansas, U.S.A.
3
CARE, do FAO, Lima, Peru.
4
Instituto Nacional de Investigacin Agraria, Lima, Peru.
5
Universidad Nacional Hermilio Valdizan, Hunuco, Peru.
6
Universidad Nacional Daniel Alcides Carrin, Cerro de Pasro,
Peru.
2

Aspects of the FPR-FFS related to farmer


participation and training are presented
elsewhere (Nelson et al., 2001 ). In this
paper, we begin to explore the added
value that scientists contributing to
participatory research can provide by
analyzing data generated and analyzed
locally by farmer groups in the context of
farmer-trai ni ng efforts.
Late blight is a key constraint to potato
production for smal 1-scale, resource-poor
farmers. Surveys conducted in the
department of Cajamarca in the Peruvian
Andes revealed that the majority of
farmers considered LB to be their most

important problem in potato production.


Nearly all the farmers surveyed (95%)

GIP Program Report 1999 - 2000

225

spray fungicides for LB (Ortiz et al., 1999).


In spite of their dependence on fungicides,
farmers know little about the products they
apply. Although the most widely used
products are suspected carcinogens (the
class of fungicides including maneb
accounted for 97% of sprays in the survey
sample), farmers usually spray without
protection.
In collaboration with various national
organizations in Peru and elsewhere, CIP
has long maintained a breeding program
aimed at producing LB-resistant potato
varieties suited to the needs of resourcepoor farmers (Landeo et al., 1995). CIP
makes crosses, tests for LB resistance,
makes preliminary selection for agronomic
traits, and then provides advanced selections to national research organizations,
universities, and others. Since the early
1980s, potato breeders have consulte_d
with farmers in selecting new potato
varieties for the Andes (references cited in
Thiele, 2000). In recent years, strategies
and methods for involving farmers in
research have been developed, refined,
and debated, leading to increased awareness of the benefits of participatory
approaches. This has been particularly
important for breeding and deployment of
varieties for diverse and marginal environments (Sperling et al., 1993; Witcombe et
al., 1996; Ceccarelli et al., 2000).

To intensify and expand participatory


research and farmer training in relation to
LB management, CARE-Peru and CIP
undertook a col laborative effort to estab1ish FFSs. The FFS approach originated in
Asia in the early 1980s, where it has been
widely used for participatory farmer
training in rice IPM (Van de Fliert, 1993;
Matteson, 1996). The objectives of the
patato FFS included training on integrated
disease and crop management as well as
evaluation and deployment of resistance.
The FFS approach to farmer training was
adapted to patato and LB under Andean
conditions and integrated with elements of
FPR (Nelson et al., 2001 ). Hence our FPR-

226

Research on Potato

FFS approach compared with FFSs that


focus on farmer training.
This paper describes technical results of
FPR experiments to address four aims: (1)
to determine the relative importance of
quantitative resistance and fungicides in
combating LB, (2) to field test varieties
that have been on-the-shelf for severa!
years, (3) to field test a series of new
breeding lines, and (4) to field test seven
entries from CIP's TPS breeding program.

Materials and Methods


Collaborative arrangements and process
The FPR-FFS effort was initiated in early
1997 when CIP organized a workshop on
integrated LB management (IPM-LB) to
develop collaborative programs on IPM-LB
using the FFS approach. Participants
included representatives from severa! Latin
American countries whose institutions
were involved in research and extension.
Three workshops were held at CIP later in
the year to plan the FPR-FFS in Peru.
Participants were from the Instituto
Nacional de Investigacin Agraria (INIA),
universities, Servicio Nacional de Sanidad
Agraria (SENASA), the NGO CARE, and
CIP. This group has since evolved into the
National Late Blight Network, now a
legally recognized entity in Peru.
Twenty-five FPR-FFSs were faci 1itated by
extension workers from the Andino Project
of CARE-Cajamarca from 1997 to 2000.
These FPR-FFSs were conducted with
farmer groups in the San Miguel and
Yanahuanga watersheds of Cajamarca
Department in the northern Peruvian
Andes (18 sites in 6 districts covering
approximately 20-40 km 2 in total). During
the 1997/98 season, FPR-FFS experiments
were conducted in 4 communities in San
Miguel, 8 communities in the 1998/99
season, and 13 communities in the 1999/
2000 season.
In the 1997/98 season, parallel researcherdesigned-and-managed experiments were
conducted by 1N IA researchers at two sites

in Cajamarca (La Encaada and Porcn),


and at one site in Cusco. INIA also
facilitated FPR-FFSs at two sites in Junn
(Comas and Huasahuasi). CIP staff provided seed, research materials, and
technical support. Farmers, guided by site
facilitators, participated enthusiastically in
disease and yield measurements.
A training curriculum or Field Guide was
developed through an ongoing collaboration between CIP and CARE. Severa!
workshops were held from 1997 to 2000 to
plan and concur on revision of the Field
Guide. Throughout each FPR-FFS season,
farmers participated in a structured, handson, field-based training program designed
and based on the conceptual framework
establ ished by the FAO FFS program. Field
experiments designed by CIP were set up
and carried out by the farmer groups with
the help of extension workers and technical support of CIP researchers. The field
experiments were integrated into the FPRFSS program. Results were documented by
farmers in the form of posters and by
facilitators' notes. Results were shared
among farmers at semi-monthly meetings,
and among communities at annual field
days and planning workshops.
Field experiments
In the 1997/98 and 1998/99 seasons, each
experiment involved a set of 12-14 potato
genotypes, each with three levels of
fungicide treatment. A randomized
complete block design (RCBD) was used,
with two repl ications per treatment. Al 1 of
the genotypes were released varieties
except for advanced breeding lines Chata
Roja, Atahualpa, Jerusaln, 377740.2, and
380076.1. In 1999/2000, the above
experiment was continued, but with only
one replication and three potato varieties
with different levels of resistance:
Amarilis-INIA (hereafter called Amarilis)
(moderately resistant), Yungay (moderately
susceptible), and Tomasa Tito Condemayta
(hereafter cal led Tomasa) (susceptible).

Another experiment involved a set of new


breeding lines provided by CIP's LB
breeding program. A subset of the lines
was tested in Paucartambo, Paseo, in
1998/99. Another subset of the lines
(n = 26) was tested by 13 FPR-FFS groups
in the San Miguel area of Cajamarca in
1999/2000. For the San Miguel experiments, three groups of nine breeding lines
each were tested in three to five FPR-FFSs
per group using a RCBD. Each set of sites
represented a range of environments at
different elevations: (low: 2600-2800 m;
middle: 2800-3100 m; and high: 31003800 m). Other groups of breeding lines
were tested in four additional sites:
Mayland, Piura, and San Pablo,
Cajamarca (facilitated by CARE's Altura
Project); Huasahuasi, Junn (facilitated by
INIA); and Paucartambo, Paseo (facilitated
by Universidad Nacional Daniel Alcides
Carrin).
The th i rd ex peri ment i nvolved entries
derived from seed, provided by CIP's TPS
breeding program. Seven TPS-derived
entries were tested using a RCBD, using
small tubers as planting material: (1) MF-1
x TPS-67, (2) MF-11 x TPS-67, (3) TPS-7 x
TPS-67, (4) Atzimba x TPS-67, (5) C95C16.1 x TPS-13, (6) C95C-16.5 x TPS-13,
and (7) C95C-16.9 x TPS-13. Amarilis,
Yungay, and Tomasa served as controls.
Greenhouse-grown minitubers were used
for Tomasa and Amarilis, but larger, fieldgrown tubers were used for Yungay.
In ali experiments, experimental units of
12 m 2 consisted of 40 plants (four rows of
1O plants each). Data on disease and yield
were recorded for the two inner rows
termed plots. The low fungicide treatment
consisted of a spray interval of at least 18
days, using only a contact fungicide. The
medium fungicide treatment consisted of a
spray interval of at least 10 days but less
than 18 days, beginning with a systemic
fungicide and then using only contact
fungicides. The high fungicide treatment
consisted of a spray interval of at least 6
days but less than 1O days, beginning with

GIP Program Report 1999 - 2000

227

a systemic fungicide and alternating


between contact and systemic products.
There was a maximum of four applications
of the systemic fungicide in the high
treatment. Mancozeb, propineb, and
tolyfluanid were used as the contact
fungicides. Metalaxyl, benalaxyl, and
chlorothalonil were used as the systemic
fungicides.

Data analysis and rationale


Data were analyzed by ANOVA (SAS,

2000) and joint linear regression. Joint


linear regression is conducted to estimate
the stability of a genotype, i.e., how it
performs across various environments.
Often genotypes wi 11 respond d ifferently to
different environments. lt is therefore
important to conduct stability analysis if
the ANOVA indicates a significant interaction of geriotype by environment (GxE).
Stability analysis allows one to highlight
those genotypes that perform well across
many envfronments, making them suitable
for more growers.
As in Eberhart and Russell (1966), the
mean response over the site-year combination was taken as the environmental index
(independent variable), and the individual
treatment was taken as the dependent
variable. Since the number of severity
measurements varied over sites (3-9
evaluations), disease was represented as
the relative area under the disease
progress curve (rAUDPC). Relative AUDPC
was calculated as the ratio of AUDPC to
the total area of the graph (Fry, 1978).
Yield data were analyzed with quantitative resistance (varieties), and fungicide
treatment as both quantitative (one degree
of freedom) and as qualitative variables
(two degrees of freedom). When the
factors were treated as quantitative
variables, both were assigned the same
value, which was the inverse of fungicide
application frequency (1/18, Tomasa;
1/1 O, Yungay; and 1/6 Amarilis). To
compare researcher-managed and farmermanaged experiments, CVs were
calculated for each set of yield and
rAU DPC observations.
228

Research on Patato

Results
lnteractions of genotype, environment,
and management
The performance of three varieties with
different levels of resistance (Amarilis,
Yungay, and Tomasa) was assessed under
three levels of fungicide protection over
3 yr. Yield decreased significantly with
increasing LB disease (regression slope =
-14.9; df = 1, 299; F = 32.8; P < 0.001 ).
Relative AUDPC decreased significantly
with an increase in the frequency of
fungicide treatment (regression slope =
-0.92; df = 1, 356; F = 14.8; P < 0.001 ).
There was an increase in yield with an
increase in the frequency of fungicide
treatment (regression slope = 28.1;
df = 1, 335; F = 11.27; P < 0.001), but
there was a much stronger effect of variety
as indicated by the F value (regression
slope = 77.8; df = 1, 335; F = 110.8; P <
0.001 ). Furthermore, a more complex
model, where both fungicide and variety
were treated as qual itative variables and
including two- and three-way interactions,
also revealed that variety (df = 2, 125;
F = 277.5; P < 0.001) hada much stronger
effect on yield than fungicide treatment
(df = 2, 125; F = 76.5; P < 0.001 ).
Two of the two-way interactions were
significant for yield: fungicide treatment x
environment (df = 46, 125; F = 6.35;
P = 0.001) and GxE (df = 46, 125;
F = 8.46; P < 0.001 ). In contrast, the
interaction genotype x fungicide treatment
was not as significant (df = 4, 125;
F = 2.50; P = 0.046). The results of
ANOVA for rAUDPC were similar to those
for yield. The main effects of variety,
fungicide, and environment on rAUDPC
were all highly significant, as were all
two- and three-way interactions (df =
2-94, 131; F ~ 2.01; P < 0.001). When
individual rAUDPC values were regressed
against the mean rAUDPC of the three
genotype observations at the correspondi ng site-year, the effect of GxE was strong,
but no cross-over effects were present
(Figure 1 A). In contrast, the interaction

0.8
0.6
0.4

a"'

0.2

C1

11

CI

O l-~~.ti~!::::li~t:::[:::!:=-:_~_:!-~!_!_~_H~__!E~~ci~~e!Bj
1
B

Amarillis

0.8

0.6

Clci
CI

0.4

CI

CI

..... o
!I

0.2

a..

..

:::>
~
~

e:

0.8

Yungay

0.6
0.4

0.2

0.8
0.6
0.4
0.2

0.1

0.2

0.3

0.4

0.5

0.6

Disease pressure {mean rAUDPC)

Figure 1. lnteractions of variety x environment (A) and culivar x environment x fungicide (B-D). Mean relative
area underthe disease progress curve (mean rAUDPC) for all genotypes (Amarilis, Yungay, and Tomasa) was
used as an index of disease pressure.
CIP Program Report 1999 - 2000

229

fungicide x environment was not as strong


(Figure 1 B-D). The interactions of GxE
were further explored by joint linear
regression (Table 1).
Testing of available varieties and breeding
lines.

A set of 16 released varieties and advanced breeding lines was tested in the
FPR-FFSs. Since the variety x environment
interaction was significant, along with the
main effects for both yield and rAUDPC
(df ~ 16, 786; F ~ 4.95; P < 0.001 ), joint
linear regression was performed (Table 1).
Chata Roja showed the highest overall
yield. The mean yield of Chata Roja
(16.4 kg/plot (ca. 27 t/ha); n = 48) was
9.1 times higher than the mean yield of
the most susceptible entry, Canchn
(1.8 kg/plot (ca. 3 t/ha); n = 35). Chata
Roja, Amarilis, and Kory showed relatively
high yields irrespective of the environmental index for yield (mean yield for three
selected genotypes) or disease pressure
(mean rAU DPC for three selected genotypes). Susceptible entries, including two
of the advanced breeding lines tested,
showed low yields at all environments,
and yields were negatively correlated with
disease pressure. Among the varieties with
mean yields over 1O kg/plot, i.e., the more
resistant entries, only Atahualpa showed
yields that were significantly negatively
correlated with rAUDPC.
The most promising entry identified
through the 1997/98 FPR-FFS was Amarilis,
a LB-resistant variety with which few
farmers in the area were previously
familiar. After its success in the FPR-FFS,
CARE provided credit to allow larger-scale
production of Amarilis, and it has rapidly
gained acceptance in the area. For
example, 35% of families who participated in FPR-FFSs had commercial
plots with Amarilis after two cropping
seasons of participation, compared with
10% of nonparticipating families. This
10% suggests that if a FFS group finds a
good variety or breeding line, it will

230

Research on Potato

disseminate it to other members of the


community.
The FPR-FFSs were useful for testing
several promising advanced breeding lines
that had been on the shelf for many years
but not officially released. The variety
Chata Roja was nominated for testing by
Universidad Nacional Hermilio Valdizan
(UNHEVAU. In the 1998/99 season, Chata
Roja was the most popular entry among
the eight FPR-FFSs. This variety had been
selected from a CIP cross, first through
INIA's evaluations (it was tested by
researchers of the Canchn station from
1983 to 1994), then through participatory
evaluations conducted by UNHEVAL
during 1995-2000. Before its official
release, Chata Roja was unofficially
adopted by local farmers in Hunuco. In
part based on the positive evaluation of
the FPR-FFS, UNHEVAL released this
variety in 2000.
In 1997/98, two advanced breeding lines
provided by the Porcn agricultura!
cooperative and two by 1N IA were tested
in FPR-FFSs. One of breeding lines from
Porcn (Atahualpa) performed relatively
well, and was subsequently released by
the cooperative. Jerusaln was rejected by
fcvmers due to low yield and tuber deformation under wet conditions. The breeding
line 377740.2, nominated by INIA, did
relatively well in the FPR-FFSs, and was
subsequently released as INIA-301. Line
380076.1, another INIA nomination, was
low yielding and extremely susceptible to
LB (Table 1).
Testing of new breeding lines

In the 1999/2000 season, 52 new breeding


lines derived from CIP's Population B were
tested in FPR-FFSs in Peru. Three groups
were analyzed by ANOVA and all groups
had significant main effects of entry and
site on yield (df ~ 2, 15; F ~ 5.21;
P < 0.001 ). Two of the three groups had
significant or marginally significant GxE
interactions (Group 1: df = 1O, 15;
F = 2.30; P = 0.070; Group 2: df = 21, 28;

Table 1. Stability parameters far released varieties and advanced breeding line entries over 1997-1999. Regression coefficient (b, slope) plus its standard error (SE)
are given far entry yield regressed on disease-pressure index (mean rAUDPC far Amarilis, Yungay, and Tomasa at the corresponding site) and entry yield regressed
on ~eld environmental index {mean ~ield far Amarilis, Yunga~. and Tomasa at the corres~onding site}. lndex ranges are shown to indicate com~arabili!Y of results.
Entry
Yield (kg/plot)
RAUDPC 2
Yield regressed on mean
Yield regressed on mean
lndex ranges
yield
rAUDPC
3 p4
p4
Meanmean
b SE
Mean yield
Mean rAUDPC
n
b SE
n
0.074
0.77 0.24
0.002
9.27 6.20
0.144
3.8-16.3
0.07-0.53
16.4 a
Chata Roja
48
36
0.107
0.59 0.19*
0.002
3.96 4.61
1.0-16.3
0.07-0.59
14.3 ab
0.392
Amarilis
84
90
0.056
0.61 0.20*
0.002
7.02 4.97
0.07-0.59
Kory
14.2 ab
0.162
1.0-16.3
84
90
0.168
1.14

0.18
<0.001
-0.56

5.49
0.920
1.0-16.3
0.07-0.59
Chagllina
13.1
be
81
72
0.042
1.19

0.22
10.20

5.71
3.8-16.3
13.1
be
<0.001
0.083
0.07-0.53
Cipira
48
36
0.132
1.35

0.27
<0.001
-3.66

8.02
0.651
3.8-16.3
0.07-0.53
12.9 be
Victoria
48
36
0.086
10.25

6.93
1.0-11.1
0.11-0.59
1.26 0.32
<0.001
0.146
12.4 be
66
377740.2
54
0.186
-22.07

15.29
1.71 0.30*
<0.001
0.165
4.7-16.3
0.07-0.40
Mara Tambea
12.0 be
21
27
-12.73 3.78
1.0-16.3
0.160
1.16 0.10
<0.001
0.001
0.07-0.59
11.9 be
90
Atahualpa
84
0.227
1.04 0.14
<0.001
-7.77 4.32
0.077
1.0-16.3
0.07-0.59
10.1 e
78
Perricholi
72
0.002
-5.154.56
1.0-5.8
0.11-0.59
0.169
1.25 0.38
0.267
7.7 d
54
Jerusaln
36
0.56 0.13*
-7.19 3.54
1.0-16.3
0.07-0.59
0.335
<0.001
0.046
Yungay
5.8 d
90
84
0.26
0.002
-10.81 3.04
1.0-8.7
0.11-0.59
0.62 0.19
0.001
4.6 ef
57
Libertea
47
0.375
1.07 0.25
-11.30 2.69
1.0-5.8
<0.001
<0.001
0.11-0.59
3.8 ef
52
380076.1
36
0.103
7.9-7.9
0.07-0.07
3.0 ef
Molinera
3
3
0.567
0.36 0.07*
<0.001
-9.66 1.61
1.0-16.3
0.07-0.59
2.2 f
<0.001
To masa
83
83
-6.17 3.03
1.0-16.3
0.451
0.95 0.25
<0.001
0.050
0.11-0.59
1.8 f
60
Canchn
35
C")

* Significantly different from b = 1.0 (Eberhart and Russell, 1966).

""C

1 CIP

=e

a
'
~

;>

"C

o
;::::.

"'
"'
"'1
N

o
o
o

""~

breeding program genotype number or Peruvian variety name.


is the ratio of AUDPC to the total area of the graph (Fry, 1978).
3 Means followed by same letters are not significantly different at 0.05 level according to SNK multiple-range test.
4 Probability that slope equals O.
2 rAUDPC

(l.)

N
;;;
CI>

"'
CI>

~
::r

o::::1

'"O

o~

Table 2. Stability parameters far true patato seed (TPS) trials and new breeding lines tested in three FFS during the 1999/2000 season. Regression eoeffieient (b,
slope) plus its standard error (SE) are given far entry yield regressed on disease-pressure index (mean rAUDPC far Amarilis, Yungay, and Tomasa at the
eorresponding site) and entry yield regressed on yield environmental index (mean yield fr Amarilis, Yungay, and Tomasa at the eorresponding site). lndex ranges
are shown to indieate eomEarabili~ of results.
Entry1
lndex ranges
Yield regressed on mean
Yield regressed on mean
Yield (kg/plot)
rAUDPC 2
yield
rAUDPC
p~
p~
Mean3
b SE
Mean }'.ield Mean rAUDPC
mean
b SE
n
n
True Patato Seed
0.07-0.40
0.546
10.05 13.52
0.470
0.9-10.2
7.7 a
0.055
0.27 0.43
16
18
6
0.0-10.2
0.07-0.63
0.73 0.24
0.006
-1.34 6.76
0.845
24
6.4 ab
26
0.074
7
0.0-10.2
0.07-0.63
0.147
0.75 0.22
0.002
-6.94 6.08
0.268
22
5.9 abe
24
1
0.0-10.2
0.07-0.63
1.40 0.21
-14.77 8.24
0.088
Yungay
22
5.8 abe
23
0.291
<0.001
0.0-10.2
0.07-0.63
1.12 0.18
-9.88 6.80
0.161
24
5.7 abe
26
0.155
<0.001
Amarilis
0.295
0.0-10.2
0.13-0.63
5.6 be
20
0.135
1.23 0.21
<0.001
-8.29 7.66
18
5
-0.82 5.41
0.881
0.0-10.2
0.07-0.63
24
0.110
0.61 0.18*
0.002
22
5.0 be
4
0.07-0.63
0.56 0.15*
0.001
-3.65 4.53
0.429
0.0-10.2
24
3.9 ed
26
0.134
2
0.474
0.0-10.2
0.07-0.63
26
0.124
0.43 0.13*
0.004
-2.75 3.77
24
3.9 ed
3
0.0-10.2
0.48 0.11 * <0.001
-6.04 3.37
0.087
0.07-0.63
24
2.3 d
26
0.383
Tomasa
Group 1 New breeding lines
0.029
1.09 0.53
0.111
-18.9 90.7
0.846
14.8-27.6
0.04-0.15
6
27.0 a
6
Cipira
14.8-27.6
0.04-0.15
26.1 a
0.016
0.65 0.30
0.095
-25.9 51.1
0.639
391683.29 (11)
6
6
0.200
14.8-27.6
0.04-0.15
24.3 ab
0.093
0.86 0.25
0.026
-72.0 46.9
393079.24 (32)
6
6
-140 68.1
0.108
14.8-27.6
0.04-0.15
22.5 ab
6
0.023
1.47 0.36
0.016
391011.17 (19)
6
0.129
14.8-27.6
0.04-0.15
0.78 0.40
0.123
-92.5 48.5
19.3 b
6
0.106
393248.55 (40)
6
0.120
-157 57.4
0.052
14.8-27.6
0.04-0.15
13.4 e
0.274
1.150.58
Perrieholi
6
6
Group 2 New breeding lines
-75.3 47.4
0.163
4.1-23.4
0.02-0.18
0.062
1.32 0.23
0.001
21.0 a
8
391137.7 (10)
8
0.022
4.1-23.4
0.02-0.18
0.074
1.26 0.22
0.001
-103 33.7
17.7 ab
8
393077.54 (31)
8
0.71 0.37
0.100
-15.8 44.4
0.734
4.1-23.4
0.02-0.18
15.4 be
8
0.046
Cipira (2)
8
4.1-23.4
0.02-0.18
-55.9 29.0
0.103
0.067
0.88 0.13
<0.001
13.8 be
8
393371.159 (47)
8
4.1-23.4
0.02-0.18
0.008
-126 35.5
0.012
12.6 e
8
0.116
1.35 0.35
392639.31 (26)
8
0.02-0.18
-48.5 15.6
0.021
4.1-23.4
0.078
0.61 0.08
<0.001
12.5 e
8
393339.242 (44)
8
0.004
-72.7 20.5
0.012
4.1-23.4
0.02-0.18
11.7 e
0.139
0.81 0.18
8
393385.39 (51)
8
0.003
-110 11.5 <0.001
4.1-23.4
0.02-0.18
8
0.213
1.05 0.22
10.8 e
Perrieholi
8
*b signtticantly different from b = 1.0 (Eberhart and Russell, 1966) in yield response.
1 CIP breeding program genotype number or Peruvian variety name with FFS synonym in parentheses. lf bolded, the entry was in the top 4 chosen from 26 by farmer groups in San Miguel.
2 rAUDPC is the ratio of AUDPC to the total area of the graph (Fry, 1978).
3 Means followed by same letters are not signtticantly different at 0.05 leve! according to SNK multiple-range test.
4 Probability that

slope equals O.

F = 3.64; P < 0.001) (Table 2). The GxE


interaction for Group 3 was not significant,
but stability analysis is presented (Table 3)
for consistency. For each group of new
breeding lines tested in the San Miguel
area, the different FPR-FFSs provided a
range of disease pressures covering a
twofold to ninefold range of rAUDPCs
(measured as the mean rAUDPC of ali
entries). The results of joint linear regression are presented in Table 2. Most of the
new breeding lines performed better than
the control variety Perricholi. Severa! of
the new entries yielded higher and were
more resistant than Cipira.
Twenty-six of the 52 advanced breeding
lines were tested in San Miguel FPR-FFSs.
Three or four farmer groups in each FPRFFS chose 5 of these 26 as being among
their 4 preferred lines (bolded entries in
Tables 2). Two of these lines (387212.41
and 393280.57) are red skinned; the other
three are cream colored (391137.7 and
391011 .17) or cream and pink
(393371.1 57).

Populations derived from true potato seed


Data were obtai ned for 7 TPS-derived
entries from 12 FPR-FFSs in 1999/2000.
Based on ANOVA for rAUDPC and yield,
the main effects of entry and site were
highly significant (df ~ 9, 109; F ~ 7.27;
P < 0.001 ), as was the GxE interaction
(df ~ 89, 109; F ~ 2.59; P < 0.001 ). The
best TPS entries performed as wel 1 as the
resistant control, Amarilis, in yield and
resistance. Further analysis of the interaction by joint linear regression are
presented in Table 3. Entries 6 and 7
emerged as relatively high-yielding and
resistant to LB overall. Entries 1, 4, 5, 6,
and 7 appeared among the top three
selected by the farmer groups for further
evaluation in at least four sites. Entry 7
appeared in the top three at ten sites, entry
1 in the top three at six sites, and entry 6
in the top three at five sites.

Data quality and quantity


Data quality was somewhat superior for
the researcher-managed trials (yield CV =
60.9, rAUDPC CV = 107.1) than for the
farmer-managed tria Is (yield CV = 80.5,
rAUDPC CV = 108.0). On the other hand,
more data were generated through the
FPR-FFSs than through researcher-managed
experiments (FPR-EFS: n = 804 for yield
and n = 751 for rAUDPC; for researchermanaged experiments: n = 141 for yield
and n = 272 for rAUDPC.) When considering the same three varieties (Amarilis,
Yungay, and Tomasa) during the 1997/98
season, the mean disease pressure
'
rAUDPC) and yield were higher in FPRFFSs (0.466, 8.28 kg/plot (ca. 14 t/ha))
than researcher-managed trials (0.219,
5.36 kg/plot (ca. 9 t/ha)).

Discussion
The FPR-FFSs have provided farmers the
opportunity to evaluate and gain access to
varieties previously little known in the
area, as well as to breeding lines not yet
formally released. The farmer groups were
enthusiastic about several of the varieties,
including older promising lines and newly
identified breeding lines. Amarilis, which
was tested in FPR-FFSs in San Miguel
starting in 1997, has become increasingly
popular in the San Miguel area. Chata
Roja, which was tested starting in 1998,
was the most outstanding entry that year
and was subsequently released by
UNHEVAL. Two additional breeding lines
were released by other institutions after
successful testing in the FPR-FFSs, while
two others were rejected dueto poor
performance. A set of new breeding lines
and a set of TPS-derived populations were
tested through the FFSs, and further testing
is now in progress.
The primary purpose of the genotype x
management experiments was to demonstrate that fungicide spray decisions should
be predicated both on the level of resistance of the crop, and on the environment.

GIP Program Report 1999 - 2000

233

E':S

.:,..

en

m
a
::r

o::::

"C

g.

Table 3. Stability parameters far a group of new breeding Hnes tested in tour FFSs during the 1999/2000 season. Regression eoeffieient (b, slope) plus its standard
error (SE) are given far entry yield regressed on disease-pressure index (mean rAUDPC far all entries tested at the eorresponding site) and entry yield regressed on
yield environmental index (mean yield far all entries tested at the eorresponding site). lndex ranges are shown to indieate eomparability of results.
rAUDPC2
Entry1
Yield {kg/plot)
Yield regressed on mean
Yield regressed on mean
lndex ranges
yield
rAUDPC

!!a.
o

391580.30 (21)
Cipira
387212.41 (9)
393371.157 (46)
393073.197 (29)
393280.57 (41)
392633.54 (24)
Perrieholi
380011.12 (1)

n
8
8
8
8
8
8
8
8
8

Mean 3
23.3 a
21.5 ab
21.0 abe
20.8 abe
19.4 abed
18.4 abcd
17.1 bcd
15.8 ed
14.8 d

n
8
8
8
8
8
8
8
8
8

mean
0.053
0.010
0.009
0.039
0.026
0.075
0.039
0.196
0.089

b SE
1.07 0.16
1.05 0.17
1.12 0.13
1.24 0.19
0.99 0.16
1.010.17
0.87 0.21
0.92 0.23
0.73 0.17

p4
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.006
0.007
0.005

b SE
3.87 205
-148 193
-108 204
-135 232
4.63 191
-107 190
-244 152
-264 159
-215 122

p4
0.986
0.471
0.616
0.580
0.982
0.594
0.161
0.147
0.128

mean yield
6.1-25.3
6.1-25.3
6.1-25.3
6.1-25.3
6.1-25.3
6.1-25.3
6.1-25.3
6.1-25.3
6.1-25.3

mean rAUDPC
0.03-0.07
0.03-0.07
0.03-0.07
0.03-0.Q?
0.03-0.07
0.03-0.07
0.03-0.07
0.03-0.07
0.03-0.07

*b significantly different from b = 1.0 (Eberhart and Russell, 1966) in yield response.
1 IP breeding program genotype number or Peruvian variety name with FFS synonym in parentheses. lf bolded, the entry was in the top 4 chosen from 26 by farmer groups in San Miguel.
2 rAUDPC

is the ratio of AUDPC to the total area of the graph (Fry, 1978).
Means followed by same letters are not significantly different at 0.05 leve! according to SNK multiple-range test.
4 Probability that slope equals O.

In these experiments, variety resistance


exerted a marked effect on both disease
and yield, indicating the importance of
genetic resistance to control of LB. In fact,
the different varieties explained 4 to 1O
times the variation explained by fungicide
application frequency, further illustrating
the relative importance of quantitative
resistance. The effect of increased fungicide application was significant, but was
not as impressive as we had anticipated.
The significant interaction of fungicide
x environment indicates that at low levels
of disease yield increases will be small
with increases in fungicide application.
The marginal significance of the genotype
x fungicide interaction suggests that
fungicide is helpful in increasing yields
regardless of the level of variety resistance. However, this marginal significance
is likely caused by the limited effect
fungicide application frequency had on
disease for the susceptible variety Tomasa
(Figure 1 D).
Among the varieties field tested, Chata
Roja was the highest yielding and was
tested over a wide range of environments.
lts yield actually increased with increasing
disease pressure, as indicated by a positive
slope (b) in the fifth column of Table 1.
However, this slope was not significantly
different from zero and simply suggests
that it does better than Amarilis. Since the
slope of Chata Roja's yield regressed on
mean yield did not differ significantly from
unity, we can assume that it is stable
(Eberhart and Russel, 1966). Amarilis, on
the other hand, had a slope significantly
different from unity, suggesting it was
unstable. Nevertheless, this slope is less
than unity, indicating that it does better in
low-yielding environments or that it is
more susceptible than would be expected
in high-yielding environments.
Among the TPS entries field tested, entry 6
had the highest yield and did well despite
increases in disease pressure (it was,
however, not tested over the full range of

environments). Entry 7 was tested over the

full range of environments (column 6,


Table 2) and was found to be stable
(column 4, Table 2) and unresponsive to
increases in disease pressure (column 5,
Table 2).
Cipira ranked among the top in the testing
of new breeding lines, whereas most other
lines yielded higher than Perricholi. The
most interesting result was 391137.7 (FFS
number 1O), since it yielded the highest in
its group and significantly higher than
Cipira. lts yield was also stable over
environments and was unresponsive to
increases in late blight pressure. Furthermore, it was one of the preferred lines
among the farmer groups in San Miguel.
Both farmers and researchers need a
substantial amount of data (although
represented by different parameters) to
judge adaptation and stability of performance across the diverse Andean
environments. Further analysis of more
elaborate farmer preference data are under
way to allow a fuller understanding of
farmers' selection criteria and decision
making.
FPR-FFSs are conducted to fulfill objectives related both to farmer training and
research; and FPR-FFS initiatives are
designed to involve large numbers of
communities. The FPR-FFS network,
therefore, is expected to have the capacity
to conduct numerous experiments. In this
study, farmer-managed experiments
produced large quantities of data. Although the quality of the data obtained
from farmer-managed experiments was
slightly lower than that obtained from
researcher-managed experiments, it was
compensated for by the large quantity of
data and the expanded range of en vi ronments sampled.
In this spirit, the communities conducting
FPR-FFSs are considered as a network of
farmer experimenters who learn biophysical principies of new technologies and
how to evaluate them. By sharing data
among communities through field days

CIP Program Report 1999 - 2000

235

and workshops, it is possible to make


sound decisions and rapidly deploy
promising new genotypes. This approach
stimulated the rapid adoption of Amarilis
and other genotypes, and contributed to
the release of Chata Roja and others. In
addition to receiving the newly released
varieties, farmers also obtain early access
to new breeding lines, and have the
opportunity to influence decisions about
which materials will be released in the
future.
The breeding program at CIP stands to
benefit substantially from this network of
farmer experimenters. In addition to the
quantity of data that can be generated,
involvement in participatory research
gives scientists access to a wide range of
environments. Yield increases will be
realized more quickly as farmers understand the potential of newly released
varieties. Preference and qual ity characteristics that are difficult to measure in the
laboratory are readily, although implicitly,
incorporated in FPR-FFSs. Work on FPRFFS far potatoes in Peru is now part of a
larger network of partnerships between
research and extension organizations
working in the Andes, Asia, and Africa.

Acknowledgments
We gratefully acknowledge the IFAD
and the OPEC Fund far lnternational
Development far their financia! support
far fieldwork. Data analysis was supported
by a U.S. Agency far lnternational Development linkage grant from CIP to
K.A. Garrett, J.J. Heath, and J.L. Heath.
We would like to thank F. Ezeta and
A. Gonzales (CIP), and F. Boeren (CARE)
far their important contributions in the
early planning stages of this work. We
thank G. Thiele and R. Hijmans (CIP) far
helpful discussions and far critica! readings
of the manuscript.

References Cited
Ceccarelli, S., S. Granda, R. Tutwiler,
J. Baha, A.M. Martini, H. Salahieh,

236

Research on Potato

A. Goodchild, and M. Michael. 2000. A


methodological study on participatory
barley breeding. l. Selection phase.
Euphytica 111 :91-104.
Eberhart, S.A. and W.A. Russell. 1966.
Stability parameters far comparing
varieties. Crop Science 6:36-40.
Fry, W.E. 1978. Quantification of general
resistance of patato cu ltivars and
fungicide effects far integrated control
of patato late blight. Phytopathology

68:1650-1655.
Landeo, J.A., M. Gastelo, H. Pineda, and
F. Flores. 1995. Breeding far horizontal
resistance to late blight in patato free of
R genes. In: Dowley, L.J., E. Bannon,
L.R. Cooke, T. Keane and E. O'Sullivan,
(eds.). Phytophthora infestans 150.
European Association far Patato
Research by Boole Press, Ltd., Dublin,
lreland. p. 286-274.
Matteson, P.C. 1996. lmplementing IPM:
Policy and institutional revolution.
Journal Agricultura! Entomology

13:173-183.
Nelson, R., R. Orrego, O. Ortiz, J. Tenorio,
C. Mundt, M. Fredrix, and N.V. Vien.
2001. Working with resource-poor
farmers to manage plant diseases. Plant
Disease 85(7):684-695.
Ortiz, O., P. Winters, and H. Fano. 1999.
La percepcin de los agricultores sobre
el problema del tizn tardo o rancha
(Phytophthora infestans) y su manejo:
Estudios de casos en Cajamarca, Per.
Revista Latinoamericana de la Papa

11 :97-120.
SAS lnstitute. 2000. SAS Online Doc, SAS/
STAT User's Guide, release 8.1 Edition.
SAS lnstitute lnc., Cary, NC, USA.
Sperling, L., M. Loevinsohn, and
B. Ntabomvura. 1993. Rethinking the
farmer's role in plant breeding: Local
bean experts and on-station selection in
Rwanda. Experimental Agriculture

29:509-519.
Thiele, G. 2000. Bibliography of
participatory research in CIP. Social
Science Department Working Paper No.

2000-3. lnternational Potato Center


(CIP), Lima, Peru.
Van de Fliert, E. 1993. lntegrated pest
management: Farmer field schools
generate sustainable practices. A case
study in Central Java evaluating IPM
training. Wageningen Agricultura!
University Papers 93-3. Wageningen,
Netherlands. 304 p.

Witcombe, J.R., A. Joshi, K.O. Joshi, and


B.R. Sthapit. 1996. Farmer participatory
crop improvement. l. Varietal selection
and breeding methods and their impact
on biodiversity. Experimental
Agriculture 32:445-460.

CIP Program Report 1999 - 2000

237

An lntegrated Approach for Potato Production


lmprovement in Nepal
Among CIP's problem-driven research priorities is the development of a more
integrated approach to the management of pests, diseases, and other constraints in
patato production.
In Nepal, CIP has developed and piloted an approach for integrating various
technology components into a holistic strategy for managing the patato crop. This
integrated management strategy revolves around the principie of growing a
healthy crop, anchored on the use of high-quality seed.
The next two papers describe complementary efforts to enhance local capacity for
seed production and marketing and to manage crop health. The Nepal case
demonstrates how science can be brought to bear on real-world farming problems
and situations through: (1) gaining a fuller understanding of farmers' needs and
opportunities, (2) developing participatory methodologies for farmer learning and
experimentation, (3) strengthening local systems for multiplication, maintenance,
and use of high-quality seed, and (4) sustaining and institutionalizing mechanisms
for seed production by farmers' groups.
This work also showcases the collective effort by Nepal's Department of
Agriculture (DoA) and the Nepal Agricultura! Research Council (NARC), and the
CIP-SDC (Swiss Development Cooperation) Patato Project in South and West Asia
and the UPWARD (Users' Perspectives With Agricultura! Research and
Development) Network.

D. Campilan and O. Hidalgo

238

Research on Patato

A Report on Strengthening Farmer Capacity for


Growing a Healthy Potato Crop in Nepal
O.A. Hidalgo1, D.M. Campilan 2, and T.L. Lama 3

Low potato productivity in Nepal is largely a result of seed-borne diseases.


A collaborative project between CIP and Nepal's extension and research
agencies sought to use the farmer field school (FFS) approach to promote
integrated disease management (1 DM) for potato production and help
farmers strengthen their capacity for growing a healthy potato crop. During
the 1999 and 2000 growing seasons, the project conducted nine FFSs across
the country for 253 participants. Farmers' knowledge improved, and yields
increased on the plots where 1DM practices were applied. However, it became clear that the FFS approach, which was originally designed for
rice-pest problems, needs major adaptations when used to address potato
1DM, and that in order for the FFS to serve as an effective vehicle for local
multiplication and maintenance of healthy tuber seed, group learning processes need to be combined with appropriate social and institutional
arrangements among farmer participants.

Potato Production in Nepal


Potato is an important staple food crop in
Nepal, with per capita consumption of 30
kglyear, one of the highest national rates
in southwest Asia. Area under potato in
Nepal in 1998 was 118,043 ha (MA/ASD,
1999); the crop is grown across the country
in three agroecological zones, from the
southern terai (below 1 000 m altitude) to
the northern high hills and mountains (up
to 4000 m). Half (50.3%) of the potato
farms are found in mid-altitude areas
(1 500-2000 m) At elevations of 2000 m
and above, farmers traditionally grow seed
potato tubers that meet the requirements of
farmers in the mid-hills and plains during
the planting season (Bhomi, 1997).
Although potato productivity has increased
1

CIP, lslamabad, Pakistan.


Los Banos, Philippines.
DoA-PDS, Kathmandu, Nepal.

2 CIP-UPWARD,
3

rapidly (by 70%) over the past 30 years, it


is still among the lowest in the world. Poor
seed (quality and quantity) and seed-borne
diseases are a major cause of this low
productivity.

Seed and diseases: Key limiting factors for


increasing potato productivity
Nepal's patato growers need about

150,000 t of tuber seed per year, but


government and private commercial-sector
seed producers can supply only 10-15% of
this amount (Lama, 1998). Tuber-seed
produced by community-based Seed
Producer Groups is gaining space very
rapidly in Nepal's patato cultivation (Ojha
et al., 2001 ). There are several. traditional
seed pockets in the mid- and high-hills of
Nepal still producing seed potatoes for
their surroundings and for transport to
lower altitudes. However, high transport
and storage costs add to the problem of

CIP Program Report 1999- 2000

239

seed supply. So the seed used by most


potato farmers usual ly consists of tubers
saved from the previous season or obtained
from the local agricultura! market.
The seed supply problem is further aggravated by the prevalence of various
diseases. Late blight (Phytophthora
infestans) and bacteria! wilt (Ralstonia
solanacearum) are the most devastati ng
(Hidalgo, 1998). Late blight is a recurring
feature in all regions of the country, and
most of the popular patato varieties are
highly susceptible; in sorne years, crop
losses are as high as 95%. Chemical
control, even if not too expensive by local
standards, must be used in a preventive
and precise manner, which is sometimes
not possible when the disease appears in
epidemic proportions. The incidence of
bacteria! wilt has ranged from 10% to
100% in the past 15 years. lts occurrence
is mainly associated with the use of
infected seed, planting on contaminated
soil, and poor crop management practices.
Yield losses in the western hills have
reached 95% in the field and 100% in
storage (Pradhanang et al., 1987, 1993;
Pradhang and Elphinstone, 1997). Other
important diseases include wart
(Synchytrium endobioticum), black scurf
(Rhizoctonia solani), common scab
(Streptomyces scabies), and diseases
caused by viruses (Hidalgo, 1998). All
these diseases not only reduce yields, but
also make it difficu lt for farmers to produce enough seed for subsequent planting
seasons.

Developing local capacity for potato 1DM


Since 1999, CIP has been working with
Nepal's agricultura! extension and research agencies on a project aimed at
promoting integrated disease management
(IDM) by developing and piloting participatory farmer learning approaches. The
project, supported by the Swiss Agency for
Development and Cooperation (SDC) and
the Users' Perspectives With Agricultura!
Research and Development (UPWARD),
trained a team of extensionists and

240

Resean:h on Potato

researchers working to strengthen the


capacity of potato farmers to multiply and
maintain their own supply of healthy tuber
seed. The farmer field school (FFS) approach was selected as the main vehicle
through which extensionists and researchers could facilitate farmer learning on
IDM. The project's underlying goal was to
assess the potential of applying FFS for this
purpose on a wider scale in Nepal.
In the 1999 and 2000 crop seasons, the
CIP-SDC Project and UPWARD, in partnership with the Nepal Department of
Agriculture (DoA) and the Nepal Agricultura! Research Council (NARC), conducted
nine FFSs and trained 253 participants.
This was the first time that FFSs had been
used in supporting patato production in
Nepal, so the project sought to develop an
eclectic approach that drew on relevant
past experience in patato and sweetpotato
from U PWARD and other CIP projects, and
in rice from the Food and Agriculture
Organization of the United Nations (FAO)
Community lntegrated Pest Management
(IPM) Program in Nepal.
The lack of previous experience in patato
FFS was a majar drawback, but the project
benefited from the FAO program's earlier
efforts to train district-level agricultura!
technicians as FFS facilitators. Of the 32
facilitators for the patato FFS, one third
had prior experience in running a similar
activity on rice IPM (the FFS approach
was developed for IPM in rice). In a
project workshop after the fi rst year of
project implementation, these facil itators
indicated that their facilitation skills
acquired through FAO were an asset, but
they noted that majar adaptations had to
be made for the FFS approach to be
applied to patato seed and disease problems (Table 1).
Patato ICM requ i red that the FFS be
conducted over several seasons, the
learning plots be used also for seed
mu ltipl ication, and the field observations
be done to anticipate the onset of late
blight.

Table 1. Comparison of original rice Farmer Field School (FFS) and the emerging patato integrated disease
management (IDM) approach in Nepal.
Aspect
Rice FFS
Patato IDM
Remarks
Multi-season
Time frame
Season-long
IDM requires a longer time trame since its
success is determined by doing a followup by replanting produced seeds in next
seasons.
Learning plots Experimentation
Experimentation, seed Seed is an important component of IDM.
Learning plot is also used to multiply/
multiplication/
maintain good-quality seed.
maintenance
Weekly, but with more Depends on appearance of disease
Frequency of Weekly
symptoms, especially far late blight.
frequent inspection
sessions
Sessions need not be weekly early in the
far late blight
season, however, they need to be
detection.
more frequent (2-3 per week) when
late blight/ bacteria! wilt symptoms begin
to appear.
To be used more selectively since weekly
AESA needs to be
Basic method far
AESA 1
AESA produces data which may not be
complemented by
learning by
directly usefuVrelevant far patato IDM.
other 'discovery'
'discovery' by
methods
farmers
Directly and indirectly Unlike insects, pathogens are often not
Making things Directly through
visible. Experiments to show the 'effects'
visible
AESA
needs to be done.
Disease management takes several seasons
1mpact after severa!
lmpact after FFS
Evaluation
to complete. lmpact assessment needs to
seasons
season
be done only after several seasons.
Disease and seed management are closely
Multiple constraints Single constraint Scope
interrelated. FFS needs to deal with the
cropping system
crop
interaction among disease and seed
factors, as well as dynamics between
patato and other crops.
1 AESA =

agro-ecosystem analysis.

Most importantly, project collaborators


suggested that agro-ecosystem analysis
(AESA), as designed in rice FFS, had to be
used more selectively because, unlike in
rice FFS:
weekly AESA early in the season often
does not yield sufficiently different or
new observations about the potato crop.
the detai led data (e.g., number of
leaves) gathered in rice AESA are not
directly relevant to the learning content
in potato AESA, and
AESA needs to be combined with other
observation methods and exercises to
facilitate farmers' discovery/learning
more about potato diseases (e.g., trials).

lncreasing farmers' knowledge through


field schools

Because there were differences in the


potato systems and constraints at the nine
FFS sites, each group of facilitators and
farmers developed its own local ly relevant
training currculum (Table 2). Thus, they
had a common focus on seed health and
late blight, but each FFS took the decision
of including bacteria! wilt, true potato
seed, and/or crop management.
The project demonstrated that it is not only
male, literate farmers from the dominant
ethnic group who are able to participate
meaningfully in agricultura! research and

extension. Of those who successfully

CIP Program Report 1999 - 2000

241

Table 2. Comparison of FFS curricula among different sites.


FFS Siles
Main learning content for patato IDM in FFSs
Seed health
Late blight
Bacteria! wilt True patato seed
X
X
Sarlahi
X
X
Rupandelhi
X
X
X
Sunsari
X
X
X
Nigaley
X
X
X
Danushi
X
X
Surkhet
X
X
Kathmandu
X
X
X
Kavre
X
X
X
Kapilbastu

completed the FFS, 26% were illiterate


and 36% were female. In addition,
participants at seven of the FFSs carne
from multiple ethnic groups. Average
knowledge gain (based on a comparison of
the resu lts of pre- and post-FFS tests) on
patato seed and disease management was
84% (Table 3).
FFS plots for learning and seed
multiplication

In an FFS, group learning centers around


the learning plots. In the Nepal project,
the learning plots were used to compare
recommended 1DM practices (su ch as use
of healthy seed, resistant varieties, field
sanitation, and regular monitoring) with
farmers' seed and associated practices.
At ali FFS sites (except Kathmandu, for
which data are not available), the average
yields from the IDM plots were higher than

Crop management
X

those from the plots with farmers' seed and


practices (FSP); average yield increase
across the eight sites was 14.2 t/ha (Table
4). No data are available on the yields
achieved by the farmers in their own
fields; participants at future FFS should be
asked to provide such information.
The FFS learning plots were also intended
to be used for multiplying healthy tuber
seed that could be shared among participating farmers at the end of the FFS. In the
majority of the FFS sites, however, this
goal was not achieved because the
farmers did not discuss in advance how the
tubers harvested from the learning plots
would be divided among them. The
informal agreement was for the harvested
tubers to be given as incentive to landowners whose fields were used as learning
plots. But the landowners wanted immediate cash benefits, so they sold the tubers,

Table 3. Profile of FFS participants and their change in knowledge on patato IDM.
Change in knowledge
Farmers gender
FFS Sites
Literacy (%)
Pre-test
Post-test score Net increase
(M/F)
(%)
score (%)
(%)
31
70
18 (18/0)
83
39
Sarlahi
57
27
84
Rupandelhi
34 (31/3)
40
78
40
34(21/13)
82
38
Sunsari
Nigaley
73
38
24 (8/16)
50
35
Danushi
25 (17/8)
60
38
98
60
Surkhet
39 (3/36)
60
73
18
Kathmandu
21 (18/3)
100
55
32
53
85
Kavre
25 (21/4)
90
20
65
85
Kapilbastu
17 (14/3)
100
43.8
80.8
37
Total average 237/26 (17/9)
74
1

Compares changes in knowledge between pre- and post-tests.

242

Research on Potato

Difference
(%)1

79.5
211.1
105.3
108.6
157.9
32.7
60.4
30.8
84.5

Table 4. Yield comparison between IDM and FSP plots.


FFS siles1
Farmers' seed and
IDM
practices (Vha)
(Vha)
Nigaley
12.5
40.0
Kapilbastu
12.0
26.0
Rupandelhi
20.0
35.0
Danushi
8.0
13.8
Sunsari
36.0
58.0
Sarlahi
17.0
25.0
Kavre
25.0
35.0
Surkhet
27.1
29.6
Average
19. 7
32.8

Yield increase
(Vha)
27.5
14.0
15.0
5.8
22.0.
8.0

10.0
2.5
14.2

Difference
(%)2
220.0
116.7
75.0
72.5
61.1
47.1
40.0
9.2
80.2

1 Data not available in one FFS site (Kathmandu).


2

Percent of difference compares yield in FSP plots and increase in yield obtained in IDM plots.

either in the market as ware potatoes, or to


their fellow farmers as seed for the next
season. Consequently, the underlying
purpose of providing FFS farmers with
direct access to quality seed was not fully
achieved.
A major lesson from this experience is that
learning processes, such as through the
FFS, are possible and the technology
available can fulfill the farmers' needs on
the use of healthy seed and IDM in
general. The learning/seed plot approach,
however, was not successfu 1 in promoti ng
the multiplication and maintenance of
healthy seeds among farmers. Equal ly
important is setting up social and institutional arrangements for ensuring a more
equitable access and sharing of goodquality seed produced through the FFS.
Both FFS facilitators and farmers recognized this to be a key negotiation point
that has to be made during the preparatory
meeting in the next season's FFS. lt is one
thing to produce good-quality seed, but it
is another thing to share it.

lmproving the FFS approach


The positive results of farmers' experiments and of the FFS in general were
encouraging, but agricultura! technicians
and farmers agreed that the participatory
training approach could be further improved. The following are sorne of their
suggestions.

Simplifying trials. In addition to the


learning plots, FFS farmers also set up
observation plots where they conducted
various experiments on potato IDM.
However, farmers tended to test several
variables simultaneously in one trial, so it
was difficu lt to analyze data and derive
conclusions. Experiments should be
simpler, with only one variable in each
tria l.

Limiting number of trials. In order to


maintain a clear learning focus, there
should be no more than five trials in each
FFS. Other research questions or experiments suggested by farmers can be
addressed in subsequent FFS. lt was agreed
that the basic trials for potato FFS in Nepal
should include varietal selection and TPS
evaluation.
Comparing performance. lt would be
instructive to monitor and compare crop/
field performance on the learning plot in
the FFS with that on the FFS farmers' own
farms and on the farms of non-FFS farmers.

Documenting other processes. In addition


to crop/field monitoring, the project should
look into learning and social processes,
such as changes in FFS farmers' knowledge and practices, diffusion of potato
ICM i nformation to other (non-FFS)
farmers, perceived effectiveness of FFS
methods and activities, and performance
of extensionists and researchers as FFS
faci 1itators. Subsequently, the project

CIP Program Report 1999 - 2000

243

organi_zed a learning workshop to develop


capac1ty of FFS facilitators for more
effective participatory monitoring and
evaluation.

References
Bhomi, B.K. 1997. Potato development
programme in Nepal: Past, present
and future. In: Pradhanang, P.M. and
J. G. Elphingstone (eds.). lntegrated
management of bacteria! wi lt of potato:
Lessons from the hills of Nepal. Lumle
Agricultura! Research Centre, Pokhara,
Nepal. p. 4-10.
Hidalgo, O. 1998. Major diseases and the
control components available far their
utilization in IDM programs in countries
of SWA. Paper presented at the
Workshop on the lmplementation of
lntegrated Disease Management
Program Utilizing the Farmer Field
School Approach held 2-5 June in
Kathmandu, Nepal. (unpublished.)
Lama, T.L. 1998. Potato development
programme in Nepal. Paper presented at
the Workshop on the lmplementation of
lntegrated Disease Management
Program Utilizing the Farmer Field
School Approach held 2-5 June in
Kathmandu, Nepal. (unpublished.)
MNASD (Ministry of Agriculture/
Agricultura! Statistics Division). 1999.
Statistical information on Nepalese

244

Research on Potato

agriculture. MNASD, Kathmandu,


Nepal.
Ojha,D.N., O.A. Hidalgo, and T.L. Lama.
2001. A Report on informal high quality
seed-potato production and marketing
by Seed Producer Groups in Nepal.
From the Lab to the Land, Research for
the 21 st Centu ry, Program Report,
lnternational Potato Center, Lima, Peru.
p. 245.
Pradhanang, P.M. and J.G. Elphingstone
(eds.). 1997. lntegrated Management of
Bacteria! Wilt of Patato: Lessons from
the Hills of Nepal. Lumle Agricultura!
Research Centre, Pokhara, Nepal.
Pradhanang, P.M., B.K. Dhital, M. Rossi,
S.L. Joshi, 5.P. Chand, S.K. Shrestha,
and D.B. Gurung. 1987. Potato
production systems in the western hills:
A combined trek report in LARC target
areas. Working Paper 1/1987. Lumle
Agricultura! Research Centre, Pokhara,
Nepal.
Pradhanang, P.M., R.R. Pandey,
S.R. Ghimere, B.K. Dhital, and
A. Subedi. 1993. An app oach to
management of bacteria! wi lt of potato
through crop rotation and farmers'
participation. Proceedings of an
lnternational Conference held 28-31
Oct 1992 in Kaoshiung, Taiwan. ACIAR,
Canberra, Australia. p. 362-370.

A Report on Informal High Quality Seed-Potato


Production and Marketing in Nepal
D.N. Ojha1 , O.A. Hidalgo2, T.L. Lama3

Described here are the evolution through the years of the tuber-seed production efforts of the Nepal Potato Program in cooperation with CIP and efforts
to promote the institutional establishment of the Seed Producer Groups (SPG)
in the country.
Nepal is cine of poorest countries in the
world, with an annual growth rate of 2.4%.
lts population reached 23 million in 2000.
About 90% of the people live in rural
areas and are mainly dependent on
agriculture. Their food needs are becoming
critica! as the population grows.
Potato plays an important role in meeting
diese food needs. Per capita consumption
of potato in Nepal is 30 kg/year (CIP,
1998), slightly higher than the world
average of 28 kg/year. In 1999, total area
under potato was 118,043 ha and productivity was 9.2 t/ha (MNASD, 1999).
However, yield is much lower than that in
neighboring countries and the world
average (16 t/ha). The major cause for this
low productivity is poor quality tuber-seed,
non-availability of good quality seed, and
lack of knowledge about modern crop
cultivation methods (Ojha, 1999).
Nepalese farmers have cu ltivated potato
for 200 years and potato is one of the
major food crops in the mid- and high-hills
where it is a staple food. This differs from
the plains and urban areas where potato is
mainly considered a vegetable. Potato
could play a more important economic
role if it could be produced in a sustainable manner. This begins with good quality
tuber seed.
1
2
3

Former Scientist, Project CIP/SDC, Nepal.


Project CIP/SDC, lslamabad, Pakistan.
Potato DevelopmentSection, Ministry of Agriculture, Nepal.

Seed quality and the evolution of the


tuber-seed systems
Tubers used or marketed as seed are
usually the extra potatoes that remain after
consumption (Thapa et al., 1 999). lt is
assumed in Nepal that potatoes grown
above 2000 m altitude make good seed
because the incidence of certain diseases
is less frequent at these altitudes. Although
there are severa! traditional seed production pockets in the high-hills, this does not
necessarily assure that seed-quality
standards are met.
The National Potato Program, being aware
of the importance of high quality seed, has
promoted different approaches to increasing the availability of good quality seed.
In the early 1980s, a farmer-contract seed
system was initiated with the support of
the Swiss Agency for Development and
Cooperation (SDC). Individual farmers
were contracted to produce seed under the
direct supervision of project staff and these
farmers were responsible for maintaining
certain quality standards. This contract
system proved to be time-consuming,
costly, and, more importantly, unsustainable, because it required a large number
of well-trained personnel, substantial
logistics, and externa! funds.
Beginning in 1987, the contract system
was gradually phased out. lt was replaced

CIP Program Report 1999 - 2000

245

by a system of affiliated farmers groups,


and from 1991 to 1996, a seed production
program was implemented through this
group approach. Farmers previously
involved in the contract system were
reorganized into Seed Producer Groups
(SPGs) with the aim of establishing a
decentralized and self-reliant seed production system. In contrast to the contract
system, SPG farmers were solely responsible for seed production, transportation,
storage, and marketing. The Potato
Program provided only the technical
assistance necessary to maintain seed
qual ity. More than 50 farmer groups were
formed in the five development regions of
Nepal, and by 1996, most of the production and management aspects were under
their control. However, persistent problems
in marketing, quality control, and sorne
management aspects threatened their
sustainability.
From 1997 to 2000, CIP and SDC continued to support the SPGs to address these
problems. Over this period, the SPG
approach was consolidated and farmers
produced reasonable quantities of seed.
The 525 farmers in the 21 SPGs resulting
from consolidation were assisted and
monitored. Special attention, e.g., training, technical assistance, and support, was
given to from three to five farmers in each
group who were involved in the Technical
Subcommittee for Quality Control. The
groundwork begun here earlier for producing high quality seed was continued using
the SPG approach.

Characteristics and basic elements of


the SPG approach
Individual SPGs carry out a set of activities by mobilizing group members and
thereby creating an informal and decentral ized seed potato production scheme.
Farmers themselves become planners,
decision-makers, implementers, and also
evaluators. Active SPGs are advised to
register at the District Cooperative Office
to receive their legal identity as an

246

Research on Potato

association. The associations are advised


to form a Regional Association (RA) for
better networki ng between the SPGs. The
major objective of an RA is to identify
problems and come up with common
solutions.
Developing and strengthening the organizational and managerial skills of the SPGs
is one of the most important aspects of the
seed production activities (Ojha et al.,
1999). Therefore, in this approach, the
initial emphasis by facilitators was on
educating farmers in group management,
marketing, and developing an interna!
seed quality control mechanism. After
severa! attempts, steps for SPG organization and its functional structure were
defined and adjusted for Nepal conditions
(Figure 1).
The basic elements of the SPG approach
can be summarized as follows:
Seed production is done by the groups.
The SPG proposes its own legal constitution and is official ly registered.
Seed production is carried out in
accordance with rules and regu lations
adopted by the group.
Minimum quality standards are followed.
Technical support is secured from
authorized agencies like the District
Agriculture Development Office
(DADO).
The Technical Subcommittee of the
group carries out the field inspections.
The Subject Matter Specialist of the
supporting agency may help on field
inspections.
The group recommends only the seeds
of lots that they have inspected and that
were found to be in accordance with
the norms adopted.
Pre-basic seed (PBS) that originated as
in vitro disease-free stock is the only
initial seed source accepted.
On-farm research is carried out in a
participatory manner with the support of
the Potato Research Programme, Potato
Development Section and DADOs.

ldentify seed pocket

----.

i
Group formation
process

Group management

----.

i
Seed potato
production

Seed marketing

Group formation and initiation of group


activities.

A mechanism of SPG functioning based on a


management committee and tour subcommittees.

An informal seed productions scheme


adopted by SPGs.

General guideline and rules and regulations


for informal quality control adopted.

lnternal quality
control mechanism

A stepwise procedure based upon


participatory approach.

_.

A system for price fixation, advertisement of


SPG produced seeds and development of
marketing linkages.

Figure 1. Operational steps and achievements for the organization and implementation of the SPG in Nepal.

Quality control and quality of the


seed produced
In recognition that a guaranteed source of
initial pre-basic seed (PBS) tubers was
necessary far quality seed production, in
vitro techn iques fol lowed by the rapid
multiplication system had been implemented when, at the beginning of the
1990s, large-scale PBS production started.
By 2000, about 200,000 PBS tubers were
being produced annually, although this
was not sufficient to meet a growing
demand. All PBS produced is easily sold to
farmers because prices are subsidized.
Only farmer-members selected by the
Executive Committee of the SPG receive
and multiply the PBS to produce Basic 1
and Basic 2 seed. Other farmer-members
designated by the Technical Subcommittee of the SPG then make two to three
additional field multiplications to produce

seeds identified as lnspected 1, 2, or 3.


The Technical Subcommittee far Quality
Control assesses the seed quality against
accepted rules and regulations and
recommends or rejects the seed lots.
Quality control within the informal system
is basically an educational process by
which farmers learn how to produce better
qual ity seed (Thapa et al., 1999).
Traditionally, the quality control in a
formal system is regulated by law, often
called a 'Seed Act'. The law is enforced
by a government institution. In an informal
system, like the SPG, the quality control is
not supervised by any government official
or institution. Quality control is more a
part of the educational process of farmers
as they learn techniques and recognize the
importance of them.
In 1998 and 1999, the quality of the seed
produced by the SPGs in the three majar
potato-producing regions was evaluated in

CIP Program Report 1999- 2000

247

Table 1. Comparison of tuber-seed produced by Seed Producer Groups with farmers seed in three
Nepal. Winter and summer 1998, and winter 1999.
Region
Mid-Western
Far-West
Central
Yield Samples
Yield Samples
Yield Samples
Vha
(no.)
Vha
(no.)
t/ha
(no.)
Terai, winter 1998
24.1
27.41
19.97
131
85
90
1 Seed groups (Basic seed)
14.8
21.94
25 1
20
15.20
17
2 User groups (lnspected seed)
-3
0.0 3
14.40
15
9.20
35
3 Non-groups (Farmer seed)
90.3
117.10
Yield lncrease (%) (1 vs. 3)
Hills, summer 1998
27.90
147
12.25
30 2
90
23.56
1 Seed groups (Basic seed)
-2
20.55
O.O
30
22.70
30
2 Users groups (lnspected seed)
7.95
3 Non-groups (Farmer seed)
9.21
35 2
30
10.90
36
33.0
180.40
Yield lncrease (%) (1 vs. 3)
116.20
Terai, winter 1999
1 Seed groups (Basic seed)
33.2
45
20.0
20.0
90
55
2 Non-groups (Farmer seed)
14.3
45
7.9
45
40
13.5
132.2
153.4
48.5
Yield lncrease {%) {1 vs. 2)

regions of
Average
yield
(Vha)

23.25
17.79
10.76
116.10
23.75
21.63
9.38
153.20
23.1
12.1
90.9

Source: Project CIP/SDC, 1998 & 1999.


- No data.
1 Intense attack of late blight in one of the three localities.
2 Intense attack of late blight in two of the two localities.
3 Total lost due to late blight.

seed trials. Results are given in Table 1. In


most cases, yields from SPG (produced by
seed groups and users groups) seed were
double those from the degenerated-farmer
seed (non-group seeds). This effect was
most notable with seed produced in the
terai (lowlands) of the central and midwestern regions. In general, the improved
seed produced by the groups were far
superior to the usual seeds used in the
country, which explain the popularity of
SPG-seeds among users.

lnstitutionalization of the SPG


approach
Since 1997 activities have been oriented
to institutionalizing the SPG approach and
also to evaluating their levels of
sustainability (Project CIP/SDC, 1998 and
1999).
In 1998-99 two manuals, produced in
Nepali and English, on organization
(Ojha et al., 1999) and quality control
(Thapa et al., 1999) were distributed

248

Research on Patato

among all 75 District Agriculture


Development Offices (DADOs). Training activities were developed to support
the institutionalization of the SPGs.
In 1999, workshops were held in
Kathmandu and Nepalgungj to transfer
the SPG experience to the DADO
officers. Several follow-up visits were
made to existing SPGs in the Central
and Far-West Regions, together with the
appropriate officers. In addition, six
one-day training sessions on interna!
qual ity control were held at six different
locations, with 111 farmer-members of
SPGs participating. A five-day training
program to provide additional technical
information on informal seed production
was offered to 20 DADO specialists
who support the SPGs.
In 2000, project scientists and DADO
officers visited 30 old and new SPGs in
three development regions. A profile of
each SPG visited was established (Table
2). Severa! problems were identified in
sorne of the SPG visited; knowing these

will facilitate future action by the


DADO. For example, group management was a common problem because
of interna! conflicts due to political
unrest. On a positive note, visitors found
that potato seed produced by SPGs was
of high quality and in high demand.
In recognition of the Nepal Potato
Program's efforts, the Ministry of
Agriculture officially accepted and
recommended the SPG approach as a
model for informal seed potato production in Nepal. Now the seed potato
production program is implemented by
DADO following the SPG approach to
develop institutional strength in the
seed potato industry in the country.
About 20% of national seed demand is
being fulfilled by the SPG approach.
Nationally, the area being planted with
seed produced by SPGs is expanding
rapidly.

-e

~-~

The SPG approach is a sustainable farmerrun program. This approach has shown
greater opportunity to expand the seed
production industry in Nepal because this
approach is simple, cost effective, resultoriented, and viable in small-farmer
communities. Based on this experience the
Ministry of Agriculture has approved SPG
as a tool for technology dissemination to
produce good quality seed in the country.
lt is expected that the SPG approach will
continue to be implemented by farmers in
a self-sustainable manner. This approach
could be implemented for informal quality
seed production of other crops with certain
modifications.
We recommend that the socio-economic
impact of the SPG approach be assessed.

O,.....

NNO'>

NM

M,......

NO,.....

NO

00

N0"<:2'"

NO

,......M

OOC.0

CD >u

to

C2.
CD

en
e:

Conclusions and Perspectives

>

CD

e :s
ce
a..
~

M,......

CD

CD

en

=
e
=
==

.,......

::

M,......

&&

&&w~

- o

NNM

~(;)

e.o o

,......N

~;:o-

I'- CIO N

11
~

-g-

~ ~~~
C.ONM >.
~

References
CIP (lnternational Potato Center). 1998.
Potato Facts: A compendium of key
figures and analysis for 32 important
potato-producing countries. Lima, Peru.

GIP Program Report 1999- 2000

249

MA/ASD (Ministry of Agriculture/


Agricultura! Statistics Division). 1999.
Statistical information on Nepalese
agriculture, Singha Durbar, Kathmandu,
Nepal.
Ojha, D.N. 1999. Overview of informal
seed potato production system through
Seed Producer Group approach in
Nepal. In: Proceedings of the National
Workshop on Qual ity Seed Potato
Production through Seed Potato
Producer Group Approach held 26-27
July in Kirtipur, Kathmandu. 5 p.
Ojha, D.N., B. Thapa, B. Saha, and
G. Bhandari. 1999. Manual for
management of seed potato producer
group/association. Potato Research
Programme-NARC/Potato Development
Project for Bhutan, Nepal and Pakistan
(Project CIP/SDC-Nepal), Khumaltar,
Lalitpur, Nepal. 57 p.

250

Research on Potato

Project CIP/SDC. 1998. Annual report Informal high quality seed potato
production and marketing by Seed
Producer Group in Nepal. Potato
Development Project for Bhutan, Nepal
and Pakistan (Project CIP/SDC-Nepal)/
Potato Research Programme, Khumaltar,
La 1itpu r, Nepal.
Project CIP/SDC. 1999. Annual report
informal high quality seed potato
production and marketing by Seed
Producer Group in Nepal. Potato
Development Project for Bhutan, Nepal
and Pakistan (Project CIP/SDC-Nepal)/
Potato Research Programme, Khumaltar,
Lalitpur, Nepal.
Thapa, B., D. N. Ojha, B. Saha, and O.A.
Hidalgo. 1999. Manual to guide Seed
Producer Groups for the implementation
of rules and regulations for seed patato
quality control. Potato Research
Program/Project CIP/SDC, Khumaltar,
Lalitpur, Nepal. 17 p.

Comparative Performance and On-Farm Profitability


of Certified Potato Seed in the Highlands of Bolivia
A. Devaux1 , S. Gonzales2 , C. Bejarano 2, R. Casso2, V. Suarez1 and T. Walker 1

A series of trials was implemented in the 1990s to evaluate the performance


and profitability of certified potato (So/anuro tuberosum L.) seed in farmers'
fields during several subsequent multiplications in three potato-producing
regions of Bolivia. This article focuses on the agronomic performance of
certified seed compared to farmers' seed and addresses the relative profitability of certified seed over time and the conditions that affect economic
performance. The results support the notion that certified seed was moderately profitable in two of the three regions where the trials were
implemented. The increases in yield due to certified seed use vary from
farmer to farmer and are relatively modest on average because of constraints
such as viruses, insects, and adverse climatic conditions, especially in
Chuquisaca and Tarija. However, these data support the hypothesis that
improved potato seed has an economic role to play in intensifying small
farmer production in several potato-producing regions of Bolivia.

Potatoes are the major staple food crop


and a major source of rural employment
and income for smallholders in the Andes
of Bolivia. Data obtained from diagnostic
and yield surveys carried out by the
national patato research program
(PROINPA3) in the 1990s showed the
importance of seed tuber quality and
source of supply in determining patato
productivity (Terrazas et al., 1998). In
Bolivia, informal seed systems, those in
which farmers manage the seed production
and distribution processes, are much more
important than the formal systems in
which quality is controlled by the public
sector. Certified seed from the formal
system covers only about 2% of demand
1

2
3

CIP, Lima, Peru.


PROINPA Foundation, Cochabamba, Bolivia.
PROINPA, formerly the national potato research program of
Bolivia implemented through an agreement between the

Government of Bolivia, CIP, and the Swiss Development

for seed directly, but subsequent generations of certified seed used and distributed
by farmers increase its impact on the
informal seed system and on patato
production. Strengthening links between
them would improve the whole patato
seed system (Thiele, 1999; Aguirre et al.,
1999).
PROINPA ran a series of trials in three
potato-producing regions of Bolivia from
1990 to 1996 to evaluate the performance
and profitability of certified seed in
farmers' fields over several multiplications. Results of these trials were gathered
and then analyzed in 1999 and are
reported here. The objectives of this
research were to (1) study the agronomic
performance of certified seed compared to
farmers' seed, (2) assess the relative
profitabi 1ity of certified seed over the
study period, and (3) evaluate the conditions that affect economic performance.

CIP Program Report 1999 - 2000

251

Material and Methods

Evaluation factors

The trials were carried out in three potatogrowing areas in the highlands above 3000
m where recurrent frost and drought limit
potato production (Table 1). Certified seed
was compared to farmers' seed under their
own crop management practices. The
trials were implemented with small
farmers (< 1 ha of potato). Soil fertility was
generally poor with low organic matter
content (1 .0%) and phosphorus (< 1 O ppm),
and adequate potassium (> 0.4 meq/100 g).
Farmers usually applied 5-7 t/ha of sheep
or cattle manure to their potato fields and
low levels of chemical fertilizers, usually
in the form of NPK 18-46-0 or urea according to cash availability.

The agronomic evaluation was based on


tuber yield. Tubers were grouped into four
size categories because larger tubers
command higher prices in the market. At
planting, the quantity of seed used was
evaluated. Viruses were considered as the
main seed degeneration factor. The
progenies of the two seed sources were
analyzed by enzyme-linked
immunosorbent assay to detect the following viruses: potato virus X (PVX), potato
virus Y (PVY), potato virus S, Andean
potato mottle virus (APMV), Andean
potato leafroll virus (APLV), and potato
leafroll virus (PLRV). Composite samples
based on the two seed sources were taken
to analyze the average virus incidence. In
Cochabamba, 20 foliage samples/treatment per trial were taken; in Chuquisaca
and Tarija, 20 tubers/treatment were kept
at harvest for virus testing.

Seed Origins
Two native commercial varieties
(S. tuberosum andigena) were used,
Waycha and Sani lmilla. The certified
seed was obtained from a semiprivate
company UPS/SEPA, the main provider of
high quality seed to the formal seed
system. The farmers' seed had different
origins; these are discussed in the results
section. The first year the same quantity of
seed (24 kg) for each seed source was
planted with each farmer in two randomized plots with two replications. In
following years, the subsequent generations of each seed source were planted
following the same design using the same
quantity of seed for each seed source.
The two different seed sources were
compared for several years, four years in
Cochabamba, six years in Chuquisaca,
and five years in Tarija.

The economic analysis was based on seed


cost, seed rate, tuber yield, and farmgate
potato price according to tuber size. To
generate an estmate of a rate of return to
investment, we assumed that the cost of
certified seed occurred in the year before
the first harvest.
The time series data set contains information on six farmers over 4 years in
Cochabamba, three farmers over 6 years in
Chuqu isaca, and three farmers over 5
years in Tarija. In total, there are 114
observations, 57 for each type of seed.
Twelve farmers had complete records for
four years or more. Additionally, several
farmers in the three regions had incomplete records that are equivalent to 52

Table 1. Altitude, average climatic conditions, and variety used in the regions where the trials were carried
out, Bolivia, 1990-96.
Region/Department
Altitude
Annual temperature
Annual rainfall
Variety
(m)
(C)
(mm)
used
Cochabamba1
3490
9
531
Waych'a
Chuquisaca2
3105
12
537
Sani lmilla
Tarija2
3300
12
325
Sani !milla
1 Source: Meteorological Station form Experimental Research Station, Toralapa.
2 Source: National Meteorological Service (SENAMHI).

252

Research on Potato

total observations, 26 for each seed


source. Therefore, there were 83 paired
observations in al l.

Results
Farmers' seed source

The majority of farmers indicated they


bought their seed from fairs, neighbors,
nongovernment organizations (NGOs), or
from the Agriculture Experiment Station,
which was the case in Cochabamba
because of the proximity of the patato
research station. The reputation of the seed
source was an important criterion in the
farmers' choice of a seed supplier. Most of
the farmers had retai ned thei r seed for
more than five years when the trials were
initiated. Only in Cochabamba had many
farmers renewed thei r seed du ri ng the
previous two years, explained by the
easier accessibility to seed in
Cochabamba than in other departments
(Terrazas et al., 1998).
Seed rate

On average, the seed rate was slightly


higher for the certified seed (1880 kg/ha)
than for the farmer seed (1 680 kg/ha).
Mean seed rates were similar in all regions
but varied among farmers within regions.
Seed rates ranged from 1,000 kg/ha to
about 2500 kg/ha.
Tuber yield and constraints

Productivity was higher and more stable in


Cochabamba compared with Chuqu isaca
and Tarija where crop production was
affected by adverse climatic conditions,
mainly frost and drought (Figure 1). But in
the three regions there was considerable
variability in yields between farmers and
over time due to biotic constraints, crop
management practices, and adverse
climatic conditions as described below. In
the three sites, the certified seed produced
on average bigger tubers of category 1 and
2 than farmers' seed, which produced
more smal 1 tubers. In Cochabamba, the
farmers involved in the trials had had
contacts with the patato research station;

1990

1991

1992

1993

1994

1995

1996

Years

-+- Certified seed Cochabamba <>


-+- Certified seed Chuquisaca 6.
-+- Certified seed Tarja

Farmer seed Cochabamba


Farmer seed Chuquisaca
Farmer seed Tarija

* Yield difference between seed sources significan! at the 5% level


Figure 1. Yield of certified seed compared to farmer
seed in farmers' fields during severa! subsequent
multiplication in three potato-growing regions of
Bolivia.
this might have had a positive effect on
farmers' seed quality and on their crop
management practices.
Cochabamba. In Cochabamba, the
average yield difference between the
two seed sources was only significant
(P < O.OS) the second year of evaluation
with a 15% difference corresponding to
3.4 t/ha. The average yield difference was
9% the first year and 8% the third year in
favor of certified seed. This limited
difference could be partially explained by
the fact that the main seed source for
farmers' own seed was the research
station. The differences between farmers
were mai n ly due to crop management and
soil conditions. The maximum yield of 34
t/ha was obtained with certified seed in
the second year of evaluation against 30
t/ha for the farmers' seed. Low yields,
below 1 O t/ha, were observed only in a
few cases. Viruses PVX and PVY were the
most common viruses detected. They
reached levels of 30% incidence in
farmers' seed versus 12% incidence in the
certified seed on the fourth year of evaluation. lnfection by the false root-knot
nematode (Nacobbus aberrans) was
observed in most of the plots and can be
considered as a limiting factor to patato

CIP Program Report 1999 - 2000

253

yield in the area. Yields were affected by


adverse climatic conditions the final year.

Chuquisaca. In Chuquisaca, average


yields were much lower than in
Cochabamba but the certified seed
yielded significantly more than farmers'
seed during the first three years of evaluation where yield advantages of 10%, 71 %,
and 80% were documented. The performance of the two seed sources was similar
for the last three years with a yield
reduction using the certified seed (Figure
1). The highest yields, between 15 and 20
t/ha, were obtained with certified seed
mainly the second and third year of
evaluation with a significant advantage
over the farmers' seed. In the first and
fourth years' evaluations, potato production was affected by drought stress.
lncidences of 20% PVX and 15% PVY
were detected in farmers' seed the first
year of trials. The leve! of PVY increased
to. almost 100% during the fifth and sixth
years, but PVX did not increase much over
time. Certified seed was contaminated
mainly by PVY and PVX. In the last year of
evaluation, PVY incidence was 90% and
PVX incidence was 11 %. APLV was also
detected in farmers' seed at levels between 15% and 20% but did not increase
over time; no contamination was observed
in the certified seed. Sorne contamination
due to PLRV at levels of 20% was observed in both seed sou rces from the fou rth
year of evaluation onward. In Chuquisaca,
vi rus appeared to be an i mportant seed
degeneration factor and might justify seed
renewal after three years. Other biotic
constrai nts observed in the area were the
Andean potato weevil (Rhigopsidius
tucumanus Heller) and the false root-knot
nematode.

Tarija. In Tarija, certified seed yielded on


average better than farmers' seed du ri ng
the five years of evaluation (Figure 1).
Frost and drought affected the crop in the
first three years; drought alone in the
fourth year. During the fifth year, climatic
conditions were favorable and higher

254

Research on Potato

yields, > 15 t/ha, were obtained with both


seed sources. At the initiation of the trials,
virus incidences in farmers' seed were
19% for PVY and 13% for both APMV and
APLV. The incidence of PVY increased to
40% during the last year of evaluation.
From low initial levels, PLRV reached an
incidence of 40% at the end of the
evaluation. Certified seed was also
contaminated by PVY and PLRV, with:
levels of 1 7% for PVY and 30% for PLRV
the last year of evaluation. The other
viruses (APMV, APLV, PVX) had an incidence of around 10% in both seed sources.
Despite adverse climatic conditions,
which appeared to be the most important
constraints to potato production, higher
quality seed increased yields. The soil
fungus Rhizoctonia solani was also a
constraint that affected farmers' seed since
the initiation of the trials. Cyst nematodes
(Globodera spp.) and the false root-knot
nematodes were also observed in the
tria Is.

Economic analysis
Two sources of economic benefits related
to yield were identified: (1) yield differences, and (2) price differences. The price
differences are attributed to variation in
the distribution of tuber size between
farmers' own seed and certified seed
managed under their conditions. Higher
yields are strongly associated with more
remunerative prices because of the greater
proportion of higher priced, larger-sized
tubers.
The initial cost of certified seed of Bs.
1.3/kg (US$1.00 = Bs.3.17) was much
higher than farmers' seed of Bs.0.40/kg,
which was estimated from the ware patato
market price. For the later multiplications,
the same price was al located to the
progenies of both seed types.
In the three regions, for farmers with
complete records, certified seed gave both
higher yields and was sold at higher
average prices than farmers' saved tuber
seed. On average, the yield difference

was rather modest in Cochabamba at


1.6 t/ha and somewhat higher, exceeding
2 t/ha, in Chuquisaca and Tarija. The
difference in average price was roughly of
the same magnitude ranging from a
premium Bs.0.02/kg in Chuquisaca to
Bs.0.06/kg in Cochabamba. These price
premium and yield differences were
statistically significant at the 5% level in
pairwise comparisons that control for
farmer management and field conditions.
In Chuquisaca, with the addition of the
incomplete records, the average yield
difference was not statistically significant,
suggesting that farmers with lower management levels or in less favorable
conditions abandoned the certified seed
after one or two years. Lower yields also
led to significantly lower prices for those
farmers in Chuquisaca with only one or
two years of data.
Return to investing in certified seed

lnformation on the profitability of investing


in certified seed is presented in Figure 2.
We assume that the farmer buys the seed
in the previous year (year O). The next
year's harvest from their first multiplication is equivalent to year 1. The rate of
return from investing in certified seed
varied the most in Cochabamba where one

farmer lost money and another received a


400% rate of return. In general, certified
seed was a marginal proposition in Tarija,
an attractive option in Chuquisaca, and a
very rewarding opportunity for four of the
six farmers in Cochabamba. However, the
results of the incomplete records in
Chuqu isaca suggest that these benefits
may be overestimated, at least in
Chuquisaca where farmers who stayed in
the trials tended to have higher yields from
certified seed than those who left the
se heme.
The results in Figure 3 are surprising
because the yield advantage and the price
premium of certified seed in Tarija are
about the same as in the other two regions.
Yet average returns in Tarija are decidedly
lower. The reason for this disparity centers
on the timing of benefits. Both
Cochabamba and Chuquisaca are characterized by a benefit pattern that was
higher in the earlier years (Figure 3). In
Cochabamba, a classic benefit profi le
emerged: the largest benefit in the first
year followed by declining benefits over
time. lndeed, farmers would have made
more money if they had renewed their

Gross retum (Bs./ha)

5CXXl

Point estimates bound by their estimated


standard errors from a regression model.

400J

400

Point estimates by individual farmers

:DXl
2CXXJ

350

1000

300

o
250

f fJ

t t

-1000

200

-2000

150

-3000
-4000

100

-5000

Tarija

Chuquisaca

Cochabamba

Figure 2. Interna! rate of return (%)to farmers of


investing in certified seed in three regions of Bolivia.

1
1234
year
Cochabamba

1234

1234

year
Chuquisaca

year
Tarija

Figure 3. The timing of relative gross benefits between


certified and farmers' seed by year and region. (Over
the period of analysis from 1990-1996 the average
exchange rate was Bs.4/US$1.00.)

CIP Program Report 1999- 2000

255

certified seed after the second year. 1n


Chuquisaca, the higher profitability of
certified seed held up in the first three
years but farmers lost money from th is.
investment in the last three years. In
contrast, in Tarija the first two years were
characterized by lower yields for both
certified and farmers' seed than the last
two years. This benefit profile is atypical
as benefits were higher in years 4 and 5
than in years 2 and 3. The delay in the
timing of relative benefits in Tarija is
manifested in the lower rates of return in
the financia! analysis (Figure 2), which
rewards promptness in the generation of
benefits.

Yield of farmers' seed and the profitability


of certified seed
The relative profitability of certified seed
depends to sorne extent on the productivity of farmers' own seed. In turn, farmers'
productivity is a function of conditions
related to management, growing season,
and location. In Figure 4, we examine the
pattern between the annual change in
value of production between certified and
farmers' seed and the yield of farmers'
seed for the 83 paired observations. A
quadratic relationship explains about 25%
of the variation in the difference in gross
benefits between the two types of seed.
The scatter of points suggests that the
variance in the change of net benefits is
not constant across the range of farmers'
seed yield. Three yield intervals can be
identified. At farmers' yield levels lower
than about 7.5 t/ha, the observations
cluster tightly around the estimated curve
and the change in gross benefits is less
than Bs.500/ha. These low yields may
reflect bad years, inferior management,
poor sites for potato production, or a
combination of these factors. The most
profitable yield interval for certified seed
was when farmers' own seed yields were
between 7.5 and 25 t/ha. In this yield
segment, variance in the change in gross
benefit was markedly higher with many

256

Research on Potato

Value of production (B s/ha)

1CXXXI . . - - - - - - - - - - - - - - Y= -1431.5+543.31 x-16.34x 2


e Y= (693.4) (111.1) (34) 2
R2 = 0.23

-2000

10

20

30

40

Yield (tlha)

Figure 4. Change in value of production between


certified and farmers' seed by productivity level. (Over
the period of analysis from 1990-1996 the average
exchange rate was Bs.4/US$1.00.)
outliers above the estimated curve. This
yield interval would appear to be synonymous with adequate production conditions,
and it would appear to be the range over
which certified seed performs best in a
relative sense. Surprisingly, the highestyielding interval above 25 t/ha was
associated with the least benefits from
certified seed. Explanations for this
unexpected behavior warrant more
investigation.

Conclusions
These data support the notion that certified
seed was moderately profitable in two of
the three regions where it was the subject
of repeated on-farm testing in the 1990s.
The increases in yield due to certified seed
vary from farmer to farmer and are relatively modest on average because of
constraints such as viruses, insects, and
adverse climatic conditions, especially in
Chuquisaca and Tarija. In Cochabamba, it
seems the general farmers' crop management practices, including seed quality,
were better in the area where the trials
were carried out.
From the data it would appear that the
variation in the difference in gross benefits

between the two types of seed could be


related to the yield of farmers' seed around
three yield intervals. When farmers' yields
are lower than 7.5 t/ha, the effect of
certified seed in the change of gross
benefit is relatively low. The most profitable yield interval was observed when
farmers' own seed yielded between 7.5
and 25 t/ha, and could be related to
adequate conditions for production. The
high yield interval, above 25 t/ha, was
linked to a low benefit from certified seed,
which could be explained by the fact that
those farmers already use good quality
seed.
These data, however, support the hypothesis that improved patato seed has an
economic role to play in intensifying
small-farmer production in several potatoproducing regions of Bolivia. But the risks
of yield reduction linked to biotic and
abiotic constraints, and the high cost of
certified seed might limit the willingness
of farmers to invest in certified seed. To
overcome these constraints PROINPA has
been working on innovative seed multiplication strategies, such as rustic seedbeds,
to promote an efficient use of appropriate
technologies adapted to local conditions
to increase the income of farmers (Aguirre
et al., 1999).

Acknowledgements
The authors would like to thank Enrique
Fernandez Northcote and Vctor Alvarez
for their efforts in the virological tests and
analyses for these trials.

References
Aguirre, G., J. Calderon, D. Buitrago,
V. lriarte, J. Ramos, J. Blajos, G. Thiele,
and A. Devaux. 1999. Rustic seedbeds,
a potential bridge between formal and
traditional potato seed systems in
Bolivia. In: lmpact on a changing
world: Program report 1997-98. CIP,
Lima, Peru. p. 195-204.
Terrazas, F., V. Suarez, G. Gardner,
G. Thiele, A. Devaux, and T. Walker.
1998. Diagnosing potato productivity in
farmers' fields in Bolivia. Social
Science Department Working Paper No.
1998-5. lnternational Patato Center,
Lima, Peru. 71 p.
Thiele, G. 1999. Informal patato seed
systems in the Andes: Why are they
important and what should we do with
them? World Development 27(1 ):83-99.

CIP Program Report 1999 - 2000

257

Toward Alleviating Poverty of Rural Potato Farmers by


Strengthening the Potato Seed System in Bangladesh:
A Rapid Rural Appraisal
S.G. llangantileke 1, M.S. Kadian 1, M. Hossain 2, A.E. Hossain 2, U. Jayasinghe 3, and
A.A. Mahmood 2

The majority of people in Bangladesh live below the poverty line on an income of less than US$1 per day. lncreased productivity of food crops, such as
potato (Solanum tuberosum L.), would help to alleviate poverty. Farmers can
increase potato productivity by using good quality seed, but poor potato
farmers can afford to huy only poor quality seed from informal sources. Seed
produced through the formal certified seed system meets only 5-6% of the
seed requirement. Potato growers and extension workers are not familiar
with improved agro-techniques for on-farm production of high-quality seed.
Strengthening technology for on-farm seed production and diffusion through
the informal seed system could increase income and enhance the social status
of poor farmers.

Most people in Bangladesh make their


livings from the land, either as smallholders with an average farm size of 0.76
ha (Hossain et al., 2001) or as landless
agricultura! laborers. Poverty alleviation in
the rural sector depends on increased
productivity of major food crops such as
wheat, rice, and potato, using high
yielding varieties (HYVs) and improved
agro-techniques. More efficient use of
resources and increases in productivity
would release land for other important
crops to feed the poor. Although the area
devoted to tuber crops has not changed
much (BBS, 1998), potato production has
been increasing over the past 2 years
(BBS, 1999).
CIP, South and West Asia Region (SWA), New Delhi, India.
Tuber Crops Research Centre, Joydebpur, Bangladesh.
3 CIP, East, South East Asia and the Pacific Region (ESEAP), Bogor,
Indonesia.
1

Potato is one of many crops that could


help alleviate poverty of resource poor
farmers. lt ranks third in area after rice and
wheat, and is cultivated in almost all agroecological regions of Bangladesh.
Bangladesh produces about 2.7 million t of
potatoes annually on 240,000 ha. National
average potato productivity is not good
because low-yielding (5-7 t/ha) traditional
varieties still occupy around 35% of the
total potato production area.
In recent years, potato has become an
important food security crop, especially
during extreme flooding during the monsoon. Potato is the only crop far which
seed stocks are kept in cold stores ready
far immediate planting after floods.
lncreased productivity of potato, even on a
small scale, is possible with efficient
management of available resources and
good quality seed.

CIP Program Report 1999 - 2000

259

Lack of good quality seed at prices


affordable by small and marginal farmers
is, however, a majar constraint to increased productivity. Certified seed from
the formal system meets only 5-6% of the
total seed needs and its production through
field multiplication takes a long time.
Only elite farmers can afford this expensive certified seed. The balance of the
seed requirement is supplied by the
informal system that is managed by seed
producers, ware potato producers, and
traders. Many farmers still use ware patato
production technology for producing seed.
Seed available through the informal
system is of poor quality and of unknown
origin and generation number. Most patato
growers are unaware of optima! management practices for seed production.
A rapid rural appraisal was done in
February 1998 to gather baseline data to
analyze existing potato seed systems. The
goal of the study was to identify constraints to potato productivity and to
develop strategies to provide marginal
pota to growers with good qual ity seed at
affordable prices. The specific objectives
were to:
identify existing strengths and weaknesses of the existing seed systems and
document characteristics of seed
systems, analyze working linkages
among organizations involved in seed
production, and document the importance of the formal seed producers to
the informal seed producers.

Methods
Rangpur District in northern Bangladesh
(Figure 1) was identified as the area with
the agro-climatic conditions most suitable
for potato seed production. The area under
potato in Rangpur is about 15% of the total
patato area of Bangladesh. A rapid rural
appraisal was u sed to gather i nformation
from farmers on farmers' seed sources,
number of generations of seed used from
different sources, seed quality, varieties
used by farmers, disease incidence, yields

260

Research on Potato

of different seed types (certified seed of


HYVs purchased from BADC, seed purchased from traders/cold store owners,
farmer's own seed, and seed of indigenous
varieties), and farmers' agronomic practices. lnformation was also collected on
the formal seed system concerning breeder
seed production and supply by the Tuber
Crops Research Centre (TCRC) to the
Bangladesh Agricultura! Development
Corporation (BADC) and others, and
certified seed production and supply by
the BADC.

Results
More than 80% or more of the farmers in
Rangpur grow patato as a cash crop during
winter (November-March). Comparatively
rich farmers buy certified seed from the
BADC and use this seed and a part of their
subsequent production for three or four
generations. Most resource-poor farmers
buy degenerated seed from cold store
owners, farmers, and traders. Cardinal, a
pink-skinned variety, is the main HYV
grown by farmers in Rangpur and other
northern districts. A few farmers plant
Diamant, Kufri Sindhuri, and Heera
varieties. Diamant is more popular in the
central and southern districts, like
Munshiganj near Dakha. Other varieties
such as Multa (obtained from Munshiganj
District) were seen on sorne farms, but the
produce was diseased with common scab
and bacteria! wilt.
Traditional varieties, introduced about a
century ago, were the only potatoes grown
in Bangladesh until 1960 when HYVs were
introduced (Siddique, 1991 ). They are still
widely grown and, although they are lowyieldi ng, bring premium prices above
those of Cardinal and Diamant. They are
still popular among growers and consumers, primarily because of their good
keeping quality in farm potato stores,
relatively low costs of production, reasonably high yields with low inputs and under
stress, and the high market demand due to
better taste. The area grown to traditional
varieties is, however, decreasing as more

.,.

BANGLADESH
kIO O

10 ZO

40

lnternational Bo11ndary
District Boundory
Capital

1 NDIA
( WEST

IENGAL)

BAY

OF

BENGAL

1.1
111

.,.

Figure 1. Proposed district [?Ill far base line study.

CIP Program Report 1999 - 2000

261

area is grown to HYVs. The low yield of


traditional varieties is mainly due to
degeneration of seed stock and suboptimal production practices. The
comparative performance of traditional
varieties and HYVs as recorded by
Siddique and Rashid (1990) is given in
Table 1.
Bangladesh has both formal and informal
patato seed systems. In the formal seed
system, high quality seed is produced in
an organized program by multiplying
disease-free planting material over several
generations. The TCRC maintains
germplasm in vitro and produces breeder
(pre-foundation) seed for BADC. As much
as 823 t (1995-96) of breeder seed has
been produced by the TCRC. Of this, 307 t
went to BADC and 30 t to NGOs and the
private sector. The remaining breeder seed
was sold to the farmers (Table 2). Breeder
seed is planted at BADC seed farms for
one or two multiplications befare being
supplied far certified seed production to
contract growers. Certified seed produced
and supplied. by BADC is approximately
9000 t annual ly and meets about 5% of
the total seed requirement. The existing
sources of quality planting material
produced by contract farmers and sold by
the BADC are beyond the reach of poor
farmers. The sale price of different categories of seed varies from US$0.16 to
US$0.38, as shown in Table 3. In general,
the BADC price of certified seed is about
20-25% higherthan for ware potatoes.
Farmers who cannot afford or do not have
access to BADC seed have no option but
to turn to informal systems, which are
mostly of poor qual ity. A majority of seed
comes from this informal farmer seed
system.

Seed quality control and seed certification


The health status of the standing crops of
BADC contract growers is monitored
regularly. The quality standards given in
Table 4 have been established for certified
seeds procured from contract seed growers
and sold to farmers.
262

Research on Potato

Table 1. Performance of three traditional, one high


yielding, and farmers' own varieties grown with
improved agronomic practices in farmer fields.
Variety
Yield (Vha)1
Average
Range
23.2
14.7 - 29.1
Challisa (traditional)
Lal Pakri (traditional)
17.3
12.4-21.5
Pakri Lalita (traditional)
16.4
12.8 - 20.1
27.1
19.6 - 32.6
Diamant (high yielding)
6.8
4.8 - 12.2
Farmers' own varieties
Source: Siddque and Rashid, 1990
1 Average of 20 farmers' fields.

Table 2. Production and supply of breeder seed to


recipients by Tuber Crops Research Centre
(TCRC), Bangladesh, from 1992/93 to 1996/97.
Year
Breeder seed
Seed supplied (t)
produced (t) To BADC1 To others2
119
14
1992/93
302
78
1993/94
1994/95
481
168
22
1995/96
823
307
30
605
81
20
1996/97
1997/98
130
45
Source: Annual Report, 1997-98, Breeder Seed
Production Center, Debigonj, Bangladesh.
Note: Seed not supplied to BADC or others was
purchased by farmers; - = information not available.
1 Bangladesh Agricultura! Development Corporation.
2 NGOs and private seed companies.

Table 3. Sale price of different categories/


generations and standards/grades of seed
produced by TCRC and BADC.
Seed
Seed categories/generations
grades
(mm)

Breeder seed (G4) 1


(Sold by TCRC to BADC)
Foundation seed (G5/G6) 1
(Sold by BADC to certified
growers far one multiplication)
Procurement of certified seed
by BADC back from certified
growers
Certified seed sold by BADC to
farmers

Price
(US$/kg)

28-40
40-55
28-40
40-55

0.22
0.21
0.38
0.33

28-40
40-55

0.17
0.16

28-40
40-55

0.31
0.27

Source: Tuber Crops Research Center, Bangladesh.


1 lndicates seed generation, e.g., (G4) = fourth generation.

Table 4. Maximum tolerance limit far diseases


(infected plants (%)) and other defects allowed in
a seed crop by BADC.
Disease/d efe et
Maximum tolerance (%)
Late blight
O.O
Ring rot
O.O
Mosaic
1.0
PLRV1
2.0
2.0
variety mix
Other defects: Bruised, cut, or damaged tubers and those
with secondary growth are unacceptable.
1 PLRV = patato leaf roll virus.

Role of cold stores industry in


development

Due to Bangladesh's tropical climate,


potatoes have to be stored in cold stores
during the off-season. There are 276 cold
stores with a capacity of more than 1.2
million t. The cold storage industry has
helped potato growers and landless people
in many ways (Choudhury, 1990), including making it possible to import and
introduce present day HYVs. Poor farmers
who cannot afford expensive seed every
year store part of their produce in cold
stores to be used as seed for the following
season. And cold stores allow farmers to
hold sorne of their produce until market
prices are higher during the off-season.

Discussion
lncreases in production of breeder seed by
TCRC and certified seed by BADC to meet
the requirements of a large number of
resource poor farmers is not possible under
present circumstances of limited resources
and inadequate infrastructure. This is partly
because the lack of extension personnel
leaves a serious void in public sector
support to future seed production programs.
Also, present linkages between the private
sector, non-government organizations, the
BADC, and research and extension
organizations are almost nonexistent.
The formal seed production system of
multiplication is lengthy and takes 6 or 7
yr to reach farmers. Therefore, improving

the existing informal seed system would


fill an essential niche. Sorne of the factors
to be considered in this improvement are
that sub-optima! management practices
are a major factor for low productivity and
both potato growers and extension workers
are unfamiliar with and have limited
access to modern agro-techniques for seed
production at the farm level. lf farmers
could maintain healthy seed over generations for their own use, this would reduce
seed cost and increase total productivity.
lt is also important to consider that sorne
traditional varieties have the potential for
higher yields if cleaned through tissue
culture, and then given optima! conditions
during the growing period. Traditional
varieties are still grown on a large area;
therefore, there is a strong reason for
improving them.

Conclusions
Research strategies to strengthen the
formal and informal seed production
systems include the following.
Develop strong relationships between
the publ ic sector, private sector, and
strong NGOs for the production and
diffusion of healthy seed of improved
potato varieties.
Provide intensive training on improved
agro-techniques at seed production sites
for seed/ware potato production to
farmers, extension personnel, NGOs,
and the private sector. Farmer field
schools for integrated pest management
presently operated by developmental
organizations and NGOs in Bangladesh
would be useful for disseminating proper
seed flow management in the informal
seed system.
Reduce seed multiplication generations
from six to two at government seed
farms, followed by two multiplications
in farmers fields. After two multiplications at the farm level, farmers can sel 1
the produce as certified seed to other
farmers.

CIP Program Report 1999- 2000

263

Conduct farmer participatory research to


evaluate and diffuse improved technology for good quality seed production at
the farm level in major potato growing
areas of the cou ntry.

References
BBS (Bangladesh Bureau of Statistics).
1998. Statistical pocketbook of
Bangladesh, Bangladesh Bureau of
Statistics, Statistics Division, Ministry of
Planning, Dhaka, Bangladesh. 432 p.
BBS. 1999. Statistical pocketbook of
Bangladesh, Bangladesh Bureau of
Statistics, Statistics Division, Ministry of
Planning, Dhaka, Bangladesh. 466 p.
Choudhury, A.R. 1990. Role of cold stores
in the improvement of seed patato. In:
Proceedings of the lnternational
Seminar on the Seed Patato held from
8-1 O January in Dhaka, Bangladesh.
p. 133-138.
Hossain, M., B.M. Lal, and A. Chowdhury.
2001. Changes in agriculture and

264

Research on Potato

economy in the flood-prone


environment in Bangladesh, 1998-2000:
lnsights from a repeat survey of 1 6
vil lages. Paper presented at the
Workshop on Flood-prone Rice Systems
held from 9-11 January 2001, at BRAC
Center for Development Management,
Rajendrapur, Gazipur, Bangladesh,
organized jointly by BRRI, Bangladesh,
and lnternational Rice Research
lnstitute (IRRI). IRRI, Metro Manila,
Philippines. (unpublished)
Siddique M.A. 1991. Production of
indigenous patato varieties in
Bangladesh. In: Plenary papers and
abstracts, Proceedings of Asan Patato
Association, Third Triennial Conference,
held from 17-22 June in Bandung,
Indonesia. p. 3-4.
Siddique, M.A. and M.M. Rashid. 1990.
In: Proceedings of the lnternational
seminar on the seed patato, held from
8-1 O January in Dhaka, Bangladesh.
p. 160-171.

Research on Sweetpotato
Scientist and Farmer
Partners in Research far the 21 st Century

Transgene Expression of Rice Cysteine Proteinase


lnhibitors for the Development of Resistance against
Sweetpotato Feathery Mottle Virus
G. Cipriani, S. Fuentes, V. Bello, L.F. Salazar, M. Ghislain, and D.P. Zhang1

The sweetpotato feathery mottle virus (SPFMV) is a member of the potyvirus


genus of the Family Potyviridae. SPFMV is economically important as one of
the causal components of sweetpotato virus disease (SPVD). Developing
resistance to SPFMV is essential for the control of SPVD using resistant
varieties. A novel approach to controlling sweetpotato potyvirus involves the
use of foreign cysteine proteinase inhibitors, which inhibit the proteolysis of a
polyprotein in potyviruses, thus interfering with their replication. A rice
cysteine-inhibitor gene (oryzacystatin 1) was used to transform Jonathan
sweetpotato, a cultivar susceptible to SPFMV. lmproved resistance to SPFMV
was observed in 18 of the 25 transgenic lines after challenge-inoculation with
the russet crack strain of SPFMV.

The sweetpotato feathery mottle virus


(SPFMV) is a member of the potyvirus
genus of the Fam. Potyviridae (Clark and
Moyer, 1988). lt occurs almost everywhere
sweetpotato is grown. Economic losses
from SPFMV alone can result from externa! cracking and interna! corkiness of the
storage roots (Clark and Moyer, 1988;
Usugi et al., 1994). The SPFMV alone does
not usually cause major yield losses, but
when combined with one of severa! other
viruses, disease severity increases markedly because of synergistic effects
(Karyeija et al., 1998). For example,
SPFMV is economically important as a
component of the virus complex that
causes sweetpotato virus disease (SPVD),
the most important sweetpotato disease in
Africa (Carey et al., 1999). Resistance to
SPFMV is, therefore, essential for the
development of resistance to SPVD.
Many strains of SPFMV have been identified (Moyer, 1986) and SPFMV-resistant
1

CIP, Lima, Peru.

germplasm is available. A substantial


number of African landraces have resistance to this virus (Carey et al., 1999), and
most US varieties released during the past
30 years appear to have a high level of
tolerance to US strains of the virus. The
introgression of SPFMV resistance has
been effective in the conventional breeding process, but it is time-consuming
because of the need to combine the
resistance trait with desirable yield and
postharvest qualities. The strain specificity
of the resistance also limits the deployment of released varieties. An alternative
approach is to genetically engineer the
virus resistance to 'repair' the wellestablished cultivars that lack this trait.
Most transgenic approaches to making a
plant resistant, or immune, to viral infection target certain stages in the 'life-cycle'
of the virus, from the virus's entry into the
host until the assembly and release of
virions. Capsid protein genes are the viral
genes most frequently used for inducing

CIP Program Report 1999 - 2000

267

resistance (Lomonossoff, 1995). Although


this is effective against a wide range of
virus pathogens, sorne researchers have
found that the virus strain specificity and
efficiency of engineered resistance are
directly related. Transgenic sweetpotato
varieties incorporating the gene for the
SPFMV coat protein have been developed
(Okada et al., 2001) using the approach of
Bevan et al., (1 985) and Powel 1-Abel et
al., (1986).
Rece~tly, a novel approach to controlling
potyv1rus was reported. lt involves the use
of protein inhibitors such as cysteine
proteinase inhibitors (Gutierrez-Campos et
al., 1999). The potyvirus genome encades
a polyprotein that is processed by proteolysis into individual gene products.
Because an essential step for the replication of potyviruses is the activity of
proteases identified as cysteine proteinases
(Blankenvoorde et al., 2000), inhibition of
the proteolysis should lead to resistance to
potyviruses. Using a rice cysteine-inhibitor
gene, Gutierrez-Campos et al. (1999)
achieved resistance to tobacco etch virus
(TEV) in transgenic tobacco and potato
virus Y (PVY) potato.

To test if the constitutive expression of a


cystatin gene in sweetpotato can contribute to the control of SPFMV infection, we
transformed a susceptible sweetpotato
cultivar with the Oryza sativa cystatin
(oryzacystatin 1) gene (OCI) and challenged the trangenic plants with the russet
crack strain of SPFMV (SPFMV-RC). Our
preliminary results show improved resistance to SPFMV in the transgenic plants.

Materials and Methods


An Agrobacterium tumefaciens LBA4404
line harboring the plasmid pBlnh-OCI,
including the OCl-encoding sequence,
was provided by Dominique Michaud of
Laval University, Canada. The pBlnh-OCI
plasmid is derived from the commercial
plasmid pBl121, in which the reporter
gene gus has been replaced by the gene
that encodes the proteinase inhibitor. lt
contains the Kanamycin-resistance gene
268

Research on Sweetpotato

npt 11 as marker and the CaMV 355


promoter-driving OCI expression in a
constitutive manner.
For transformation, complete leaves,
including petioles, from the apical zone of
cultivar Jonathan were used as explants.
lnoculation was performed by incubating
the pBlnh-OCI A. tumefaciens LBA4404
line with the explants.
Transgenic sweetpotato plants were
obtained using media and procedures
previously described (Cipriani et al., 1999)
with sorne modifications. Selection was
made using 50 mg/I Kanamycin. As soon
as embryogenic calli were obtained in
medium G24D they were transferred to an
MS medium with 1 mg/I ABA for embryo
maturation and development. Somatic
embryos were then transferred to an MS
medium supplemented with 0.05 mg/I
gibberellic acid, where plantlets elongated. Plantlets were multiplied in MPB
propagation medium.
The in vitro transgenic plantlets with the
oryzacystati n gene were transferred for
multiplication to isolated CIP greenhouses
at the Lima and San Ramon stations. To
evaluate resistance to SPFMV infection
'
six stem cuttings each of the non-transformed (control) and transgenic lines were
grafted onto lpomoea setosa stock i nfected
with the russet crack strain of SPFMV
(Daines and Martin, 1964). Symptoms of
the viral infection on the sweetpotato
plants were recorded. The virus infection
was then tested by nitrocellulose membrane-enzyme-linked inmunosorbent assay
(NCM-ELISA) on leaf samples (Salazar,
1996) from the sweetpotato plants and by
grafting with healthy /. setosa indicator
plants.

Results and Discussion


Twenty-five transgenic lines were obtained.
The test for Kanamycin resistance, PCR
amplification, and the Southern blot test
confirmed the genetic transformation. For
PCR amplification, the primers used
(forward: ACCGAGCACAACAAGAAG

'

reverse: TCCACAACATATTATTCCCC)
allowed the amplificati o n of a 277-bp
segment corre spondin g to th e coding
seq uence of the oryzacystatin gene.
Eightee n of th e 25 tran sge ni c lines showed
no signs of infection after being inocul ated
with SPFMV-RC-infected grafts from
l. setosa sc ions, whereas the nontran sfo rm ed Jonathan control developed
symptoms of vein clearing and chlorotic
spots, which are typi ca l of SPFMV-RC
infect ion (Figure 1 ). Thi s suggests a
putative res istan ce in the 18 tran sge ni c
1in es.
Seventeen of the asymptomatic transgen ic
lines gave negat ive res ults with NCMELISA, which co nfirmed the visual
observations and demonstrated a high
degree of ag reeme nt between the v isu al

observat ion of symptoms and th e resu lts of


th e NCM-ELISA.
To understa nd the mec hanism of th e
putat ive resista nce in the transge ni c lin es,
stem cuttings from the SPFMV-RCinocul ated plants were planted in new pots
and then grafted with hea lthy /. setosa
cuttings (one branch of the SPFMV-RCinocu lated transgeni c plants as stock and
l. setosa as sc ion ). Symptoms of viral
infection were observed in the 17 /. setosa
plants that were grafted onto
asymptomatic tran sge ni c lines, and th e
NCM-ELISA from these /. setosa grafts
were positive. Only two lines (Jonathan/
OCl/Sc and Jonath an/OC l/G) remain ed
asymptomatic, with a negative NCMELISA (Fi gure 2, Tab le 1 ). lt is interesting
to note that the symptoms were only seen
in the /. setosa pl ants, whereas the branch
of transgenic sweetpotato that hosted the /.
setosa remained asymptomatic and was
negative by NCM-ELISA (Figure 2).
Th ese results indi cate that th e ge neti c
transfo rm ation with the OCI gene made
Jon athan sweetpotato plants tolerant to
SPFMV-RC. Although the transgeni c lin es
can still be affected by SPFMV-RC through
grafting with SPFMV-RC-infected /. setosa,
the multipli cation rate of the virus is
apparently reduced and the presence of

Figure 1. The non-transformed Jonathan control


developed symptoms of vein clearing and chlorotic
spots after inoculation by grafting with SPFMV-RCinfected /. setosa scions (upper) , whereas the
transgenic lines showed no symptoms (lower) ,
which suggested putative resistance in the 18
transgenic lines.

Figure 2. /pomoea setosa plant showing virus


symptoms after being grafted onto an asymptomatic
transgenic plant inoculated with SPFMV-RC ,
indicating that the OCl-mediated resistance reduced
virus multiplication-the presence of the virus in the
inoculated transgenic plants cannot be detected by
visual observation or by NCM-ELISA.

CIP Program Report 1999 - 2000

269

Table 1. Result of evaluation of SPFMV-RC resistance in transgenic lines of Jonathan sweetpotato.


Symptoms in
NCM ELISA in
Symptoms in
NCM ELISA in
Transgenic lines
sweetpotato1
sweetpotato2
/. setosa 3
/. setosa 4
Jonathan/OCl/1 a
+
+
Jonathan/OCl/2a
+
+
+
+
Jonathan/OCl/3a
+
+
+
+
Jonathan/OCl/1 b
+
+
Jonathan/OCl/2b
+
+
Jonathan/OCl/3b
+
+
Jonathan/OCl/1 e
+
+
Jonathan/OCl/2c
+
+
Jonathan/OCl/3c
+
+
+
Jonathan/OCl/4c
+
+
+
+
Jonathan/OCl/5c
+
+
Jonathan/OCl/7c
+
+
Jonathan/OCl/8c
Jonathan/OCl/1 d
+
+
+
+
Jonathan/OCl/2d
+
+
+
Jonathan/OCl/3d
+
+
Jonathan/OCl/4d
+
+
Jonathan/OCl/1 e
+
+
Jonathan/OCl/2e
+
+
Jonathan/OCl/1 f
+
+
Jonathan/OCl/2f
+
+
+
Jonathan/OCl/3f
+
+
Jonathan/OCl/4f
+
+
+
+
Jonathan/OCl/G
Jonathan/OCl/H
+
+
Jonathan Control
+
+
+
+
(non-transformed)
1 Visible symptoms in transgenic sweetpotato after grafting onto the SPFMV-RC-infected /.
2 NCM-ELISA test in transgenic sweetpotato after grafting onto the SPFMV-RC-infected /.
3 Visible

setosa stock.
setosa stock.

symptoms in /. setosa after grafting healthy /. setosa scions onto the SPFMV-RC-inoculated transgenic lines.
in /. setosa after grafting healthy /. setosa scions onto the SPFMV-RC-inoculated transgenic lines.

4 NCM-ELISA

the virus cannot be detected directly by


either visual observation of the inoculated
plants or by NCM-ELISA. However,
/. setosa is much more sensitive to SPFMV
than sweetpotato, and the low
concentration of SPFMV-RC in the
transgenic plants could be detected in the
/. Setosa grafts. This is compatible with the
mechanism of cystatin-mediated
resistan ce to potyvi ruses. The replication
process of all known potyviruses involves
the active expression of cysteine
proteinase. The recombined OCI, a
powerful and specific inhibitor of cysteine
proteinases, thus interferes with the
replication of SPFMV-RC in the transgenic
plants by inhibiting the required cleavage
process of the polyprotein in SPFMV.

270

Research on Sweetpotato

More research is planned to confirm the


acquired resistance. First, the association
_between the level of OCI expression and
the concentration of SPFMV-RC needs to
be assessed. Second, other strains of
SPFMV will be used to find out if the OCI
gene is as effective against them as it is
with the SPFMV-RC strain. Nevertheless,
the result achieved so far is encouraging.
lf the effect of OCI on SPFMV can be
confirmed, we plan to develop a
chimerical gene construct with resistance
to both SPFMV and chlorotic stunt virus
(SPCSV). The gene construct combining
both resistant components will be more
likely to offer a practica! solution for the
control of SPVD in sweetpotato.

Acknowledgements
The authors wish to thank Dominique
Michaud of Laval University, Canada, for
providing the OCI gene construct and
technical advice. This work was partially
supported by CGIAR-Canada Linkage Fund
(CCLF) Project #05-02-RT.

References
Bevan, M.W., S.E. Masan, and P. Goelet.
1985. Expression of tobacco mosaic
virus coat protein by a cauliflower
mosaic virus promoter in plants
transformed by Agrobacterium. EMBO
Journal 4:1921-1926.
Blandenvoorde, M.F.J., H.S. Brand,
Y.M.C. Henskens, E.C.I. Veerman, and
A.V.N. Amerongen. 2000. Protease
inhibitors in health and disease
control-Medica! and industrial aspects.
In: Michaud, D. (ed.). Recombinant
protease inhibitors in plants. Landes
Bioscience, Georgetown, TX, USA.
p. 202-213.
Carey, E.E., R.W. Gibson, S. Fuentes,
M. Machmud, R.O.M. Mwanga,
G. Turyamureeba, L. Zhang, D. Ma,
F. Abo El-Abbas, R. El-Bedewy, and
L.F. Salazar. 1999. The causes and
control of virus diseases of sweetpotato
in developing countries: Is sweetpotato
virus disease the main problem? CIP
Technical Progress Report (1997-1998).
p. 421-428.
Cipriani, G., D. Michaud, F. Brunelle,
A. Golmirzaie, and O.P. Zhang. 1999.
Expression of soybean proteinase
inhibitor in sweetpotato. lmpact on a
changing world. Program Report 19971998. lnternational Patato Center, Lima,
Peru. p. 271-277.
Clark, C.A. and J.W. Moyer. 1988.
Compendium of sweetpotato diseases.
The American Phytopathological
Society, Minnesota, USA.
Daines, R.H. and W.J. Martin. 1964.
Russet crack, a new virus disease of

sweetpotatoes. Plant Disease Report


48:149-151.
Gutierrez-Campos, R., J.A. Torres-Acosta,
L.J. Saucedo-Arias, and M.A. GomezLim. 1999. The use of cysteine
proteinase inhibitors to engineer
resistance against potyviruses in
transgenic tobacco plants. Nature
Biotechnology 1 7(12):1223-1226.
Karyeija, R.F., R.W. Gibson, and
J.P.T. Valkonen. 1998. The significance
of sweet potato feathery mottle virus in
subsistence sweetpotato production in
Africa. Plant Disease 82:4-15.
Lomonossoff, G.P. 1995. Pathogen-derived
resistance to plant viruses. Annual
Review of Phytophthology 33:323-343.
Moyer, J.W. 1986. Variability among
strains of SPFMV. Phytopathology
76:1126. (Abstract)
Powell-Abel, P., R.S. Nelson, B. De,
N. Hoffmann, S.G. Rogers, R.T. Fraley,
and R.N. Beachy. 1986. Delay of
disease development in transgenic
plants that express the tobacco mosaic
virus coat protein gene. Science
232:738-743.
Salazar, L.F. 1996. Patato viruses and their
control. CIP, Lima, Peru.
Usugi, T., M. Nakano, M. Onuki,
T. Maoka, and T. Hayashi. 1994. A new
strain of sweetpotato feathery mottle
virus that causes russet crack on fleshy
roots of sorne Japanese cu ltivars of
sweetpotato. Annals of the
Phytopathology Society of Japan
60:545-554.
Okada, Y., T. Murata, A. Saito, T. Kimura,
M. Mori, M. Nishiguchi, K. Handa,
J. Sakai, T. Miyazaki, Y. Matsuda, and
H. Fukuoka. 2001. Virus resistance in
transgenic sweetpotato [lpomoea
batatas L. (Lam)] expressing coat
protein gene of sweetpotato feathery
mottle potyvirus severe strain.
Theoretical Applied Genetics (in press).

CIP Program Report 1999 - 2000

271

Plant Productivity and Water Use Efficiency of


Sweetpotato (lpomoea batatas) as Affected by
Nitrogen Supply
M. Kelm 1, H. Brck1, M. Hermann 2, and B. Sattelmacher1

A pot experiment was conducted in a tropical mid-elevation environment


(861 m altitude) to evaluate sweetpotato (lpomoea batatas (L.) Lam.) clones
of different origin and breeding intensity for traits related to growth phenomena, nitrogen-use efficiency, and transpirational water-use efficiency (WUE)
as affected by different levels of N fertilization. Genotypes with small canopies were associated with a consistently positive response in their final
storage root dry matter (DM) yields to increasing N supply, and with efficient
allocation of DM and N to storage roots. Genotypes with high canopy net
assimilation rates (NARs) hada high proportion of leaves exposed to the sun
and high chlorophyll content in leaves. Nitrogen stress led to increased
transpiration per unit leaf area and decreasing WUE. Decreasing WUE under
N stress was due to lower total plant DM production rather than to increased
total water transpiration per plant.

As do most other field crops, sweetpotato


responds well to improved management
practices, among which N fertilization
plays an important role in producing
satisfactory yields. Previous investigations
did not indicate an optimum range of N
fertilization because the response of
sweetpotato to N ferti 1ization depends
highly on genotypic and environmental
variation (Villagarcia, 1996). Janes and
Bouwkamp (1992) reported that an application of 60 kg N/ha increased yields of
three USA varieties, but decreased yields
of three African varieties. Hill et al. (1990)
showed that total biomass, storage root
yields, and fol iage weight were not
significantly affected by the addition of
fertilizer N. But there is general agreement
1

Kiel University, D-24098 Kiel, Germany.

CIP, Lima, Peru.

that high N supply leads to excessive


shoot growth, even though that does not
always reduce storage root yields
(Villagarcia, 1996).
So far, there have been no detailed studies
on water use efficiency (WU E) of
sweetpotato. Larenas de la F. and
Accatino (1994) suggest that sweetpotato
requires a continuous water supply
throughout the growing season. Others
(Hahn and Hozyo, 1984; Rehm and Espig,
1996) reported that sweetpotato plants
were relatively drought tolerant. However,
physiological reactions of sweetpotato
plants to N or drought stress have not been
reported. Furthermore, interactions between N supply and WUE deserve more
attention to increase our understanding of
the physiological regulation of WUE.

CIP Program Report 1999 - 2000

273

Therefore, the objectives of the research


reported here were to:
evaluate the effect of N fertilization on
final storage root yields of sweetpotato
clones,
describe varietal differences in growth
phenomena such as leaf area index,
specific leaf area (SLA), root:shoot ratio,
net assimilation rate (NAR), and relative
growth rate (RGR),
determine genotype and N effects on
nitrogen use efficiency (NUE), and
determine if WUE is affected by N
supply, and if so, how.

Materials and Methods


Pot experiment

Two sweetpotato genotypes (Jewel and


Tanzania) were grown in an inert substrate
with six levels of N supply (O.O, 0.4, 0.8,
1.2, 1.6, and 2.0 g N/pot) at CIP's experiment station in San Ramn, Peru, (861 m
above sea level) under an open glass roof.
The experimental unit was arranged as a 2
by 6 factorial treatment combination using
a randomized complete block design with
four replicates treating genotypes as main
plots, with N levels as subplots. Each
treatment combination consisted of 12
plants with one plant/pot. Water supply
was held at 80% of field capacity throughout the entire growing period. All mineral
nutrients except N were supplied in
sufficient amounts to avoid nutrient
deficiencies. Once a week the amount of
transpired water/plant was estimated. At
three sampling dates, 42, 84, and 126 days
after transplanting (DAT), four plants of
each treatment combination, representing
four replicates, were harvested to determine fresh and dry weights of plant
fractions (stems, leaves, petioles, fibrous
roots, and storage roots), SLA (m 2 leaf
area/kg leaf dry matter (DM)), and pigment concentration in leaves. SLA was
determined by sampling 30 leaf disks of
1.2 cm diameter of each treatment
combination. To derive SLA, fresh weight

274

Research on Sweetpotato

and dry weight of these 30 leaf disks,


which represented a leaf area of 33.9 cm 2 ,
were determined. Leaf area per plant was
derived from SLA and leaf DM production
per plant. The extinction coefficient of
leaves (the angle of leaf inclination
relative to the soil surface) was estimated
with a simple plastic triangle. Morphophysiological traits such as NAR, RGR,
and leaf area ratio were calculated
according to Radford (1967). Pigment
contents were measured by a Norsk Hydro
N-Tester (Norsk Hydro ASA, Oslo, Norway)
in the field and related to photometrically
determined chlorophyll and carotene
concentrations in plant tissue, according to
the method for pigment extraction proposed by Lichtenthaler and Wellburn
(1983). Concentrations of nitrogen, potassium, calcium, magnesium, and
phosphorus in plant tissue were determined
by the Kjeldahl method according to
Novozamsky et al. (1983). Discrimination
against 13 C02 in leaves and carbon (C)
concentrations in shoot parts were determined by a Finnigan Delta mass
spectrometer (Thermo Finnigan, San Jose,
CA, USA). To derive NAR, C concentrations in fibrous and storage roots were
assumed to be 39%. Carbon isotope
discrimination (0 13 C [%o]) has been shown
to relate closely to WUE for a range of
crop species, e.g., for sunflower (Virgona
and Farquhar, 1996), potato (Jefferies,
1995), and wheat (Farquhar and Richards,
1984). This is commonly explained by
differences in stomatal opening, which
affects the interna! partial pressure of
inside the leaf. Since the enzyme Rubisco
reacts more readily with 12 C0 2 than it does
with 13C02 , Rubisco discriminates against
the heavy isotope (Lambers et al., 1998).

co2

Analysis of data
All data sets were analyzed by standard
ANOVA procedures for a randomized
complete block design, using the PROC
UNIVARIATE procedure of SAS software
(SAS, 1990).

Results and Discussion


The two genotypes grown in the pot
experiment differed in their habit and
growth characteristics. Jewel was characterized by a shallow canopy and almost
horizontal leaves (extinction coefficient
approximately 1.0), whereas Tanzania had
a more or less erect growth and an extinction coefficient of approximately 0.5 to
0.6. Jewel initiated storage root growth
earlier than Tanzania. For both genotypes,
storage root initiation was delayed with
high N supply.
Dry matter accumulation of the two
genotypes grown in the pot experiment
(Jewel and Tanzania) was characterized by
higher shoot and fibrous root DM production in Tanzania (Table 1). Storage root
DM production, averaged across N levels,
was not significantly different between
genotypes at the final sampling date (126
DAT), but Jewel had initiated storage root
growth earlier than Tanzania. Tanzania
yielded higher with low and moderate N
supply, whereas Jewel produced more
storage root DM than Tanzania at high N
levels. Shoot growth and shoot:root ratio
were increased with each increment in N

supply. High N levels (1 .6 and 2.0 g N/


plant) led to declining storage root DM
production. Plants that did not receive any
fertilizer N generated very low amounts of
DM in both shoots and roots.
Patterns of N accumu lation (data not
shown) varied between genotypes.
Tanzania accumulated more total N/plant
compared to Jewel. Fertilizer N was
almost completely absorbed by
sweetpotato plants. Therefore, N concentrations in plant DM increased
significantly with each increment in N
supply. The amount of N present in zero-N
plants is assumed to be the quantity of N
that was in the vine cuttings the day of
transplanting. Large shoot growth of
Tanzania was reflected in the relatively
lower amounts of N allocated to storage
roots. N concentration in storage roots was
significantly higher for Jewel.
Chlorophyll a + b content per unit leaf
area and specific leaf N (SLN) [g N/m 2
leaf area] were significantly higher for
Tanzania (Figure 1). Chlorophyll content
per unit leaf area increased slightly with
increasing N supply, whereas SLN showed
a pronounced increase when N supply was
altered.

Table 1. Shoot dry matter, fibrous root dry matter, and storage root dry matter accumulation of two
sweetpotato genotypes at 126 DAT grown in pots with six N levels. Values followed by common letters do
not differ significantly (capital letters refer to N levels; lowercase letters to genotypes).
N1
O.O
0.4
0.8
1.2
1.6
2.0
Mean*
Shoot dry matter (g/plant)
Jewel
5.97 aD
11.04 bC 13.81 bB
14.73 bB
25.93 bA
27.69 bA
16.53
Tanzania 10.31 a F 21.21 a E 39.88 a D 47.26 a e
52.94 a B
71.84 a A
40.58
Mean** 8.14 F 16.12 E 26.85 D 30.99 e
39.44 B
49.78 A
CV[%]
14.80
R2
0.97
6.81 b AB
Jewel
5.50 a BC
7.27 bA
4.52 be
5.61 b BC
6.28 b AB
6.00
Tanzania 9.19 a E 14.67
16.25 a e 13.91 a D
21.58 a A
17.88 aB
15.58
9.76 e
14.20 A
Mean** 7.35 D 10.97 BC 10.39 e
12.08 B
CV[%]
24.05
R2
0.86
Jewel
12.99 be
56.09 bA 58.98 b A 56.51 b A
31.87 b B
36.94 a B
42.23
Tanzania 14.22 a E 67.08 aB 63.72 a B 75.20 a A
34.90 aC
22.02 b D
46.19

Mean** 13.60

cv [%]

61.59

R2

61.35

65.86

33.38

29.48

b
a

b
a

b
a

25.07
0.82

1 Application of N (g/plant).
CIP Prograrn Report 1999- 2000

275

Differences between genotypes and N


levels in NAR and genotype by N
interactions on NAR were highly significant (P < 0.01). Throughout the entire
growing period, Tanzania showed the
highest NAR compared with Jewel (Figure
2). Highest net photosynthetic rates were
obtained with 0.4 g N/plant for both
genotypes, with a significant decrease
when N supply exceeded that amount.
Zero-N Tanzania plants had increasing
NAR toward maturity, whereas NAR of
N-stressed Jewel plants declined with time
and was nearly zero toward the final

0.25

~Ol

~11/

0.2

'5,
.e

c.

im

~ :::::1/
--
,k.
...
.A
/: .. -~-

~ 0.15

eo

2
1.8
1.6
1.4

,.A

A:

0.1

Jr ""

:e

1.2 "'E
1
o

---

Chorophyll Jewel
-liD- Chorophytl Tanzania
_...... SLN Jewel
SLN Tanzania

0.05

-A-

0.4

1.2

0.8

1.6

0.8
0.6
0.4
0.2

N supply g/plant

Figure 1. Specific leaf N (SLN) and chlorophyll


content in leaves at 42 DAT of two sweetpotato
genotypes grown in pots at six N levels.

00
a

"'O

fJJ

O Tanzania

rl

00
"'O

Jewel

70

b
b

() 40

ri""

rf
b

3)
2)

10
B

., B

0.8

0.4

1.2

1
1.6

N supply g/plant

Figure 2. Mean net assimilation rate (NAR) over 126


days of two sweetpotato genotypes grown in pots at
six Nlevels. Capital letters indicate significant
differences between N levels within each genotype (P
= 0.05). Significant differences between genotypes
within N levels are indicated by lowercase letters (p =
0.05). Mean SE is indicated by grey bars.

276

Research on Sweetpotato

sampling date. With N levels between 0.4


and 2.0 g N/plant, NAR of Jewel remained
almost constant with time, whereas
fertilized Tanzania treatments showed
decreased NAR toward maturity.
In N-fertilized treatments, SLN and
chlorophyll content per unit leaf area
increased with increasing N supply,
whereas NAR decreased with increasing
N supply. NAR of plants that did not
receive fertilizer N could not be related to
SLN or chlorophyll content in leaves
inasmuch as SLN and chlorophyll content
remained almost constant during ontogeny
in N-stressed plants, whereas NAR
changed significantly with time, as
mentioned above.
Differences in discrimination against
C0 2 (() 13 C) in leaves were highly significant between genotypes and N levels
(P < 0.01 ). Table 2 shows C isotope
discrimination in leaves at the final
sampling date (126 DAT). Lower negative
values for b 13 C were observed for Tanzania
at ali N levels. With increasing N fertilization, discrimination against 13 C0 2 declined
significantly.
13

There was a significant positive correlation


between b 13 C and SLN, which remained
almost constant throughout the entire
growing period. There was also a significant positive relationship between b 13 C and
total plant DM production, which was
strongest during early growth (r = 0.86 far
the period between O and 42 DAT).
lf transpiration from leaves is assumed to
increase linearly with stomatal conductance, data for transpiration rates per unit
leaf area should relate closely to stomatal
opening. No significant differences were
faund between genotypes in mean transpiration rates across ali N levels. However,
differences between N levels were highly
significant (P < 0.01) (Figure 3). In the
zero-N treatment, transpiration per unit
leaf area was significantly higher far both
genotypes. At the 0.4 g N/plant treatment,
transpiration rate per unit leaf area was

Table 2. Discrimination against 13C0 2 [%o] at 126 DAT in leaves of two sweetpotato genotypes grown in
pots with six N levels. Values followed by common letters do not differ significantly (capital letters refer to
N levels; lowercase letters to genotypes).
e lsotope in leaves
N1
o.o
0.8
1.6
0.4
1.2
2
Mean*
-31.11 b e -32.07 b D -31.47 be -30.83 b B -30.12 b A -30.03 b A -30.97 b
Jewel
-30.95 a E -30.88 a E -30.40 a D -29.52 a e -29.01 a B -28.71 a A -29.84 a
Tanzania
-31.26 D -31.44 D -30.83 c -30.09 B -29.52 A -29.30 A
Mean**
-0.36
cv [%]
R2
0.99
Note: a = Mean separation within rows for genotypes by MSD
columns for N levels by MSD = 0.26; P = 0.05.
1 fertilization (g N/plant).

slightly higher compared to treatments that


received between 0.8 and 2.0 g N/plant.
Therefore, it can be concluded that
stomatal opening increased significantly
under N stress. lncreasing interna! C0 2
partial pressure under N stress is reflected
in more negative discrimination against
13
C0z- There was a significant negative
correlation between e isotope discrimination and transpiration rate per unit leaf
area. During early growth this correlation
was strongest (r = - 0.63 between O and 42
DAT).
Total water transpiration was significantly
higher for Jewel across all N levels (data
not shown). This was mainly due to higher
total transpiration of Jewel when N supply
was at a high level. WUE was significantly higher for Tanzania at all N levels
(Figure 4) and throughout the entire
growing period. For both genotypes, WUE
was significantly lower under zero-N
conditions. This was due to very low total
DM production and higher stomatal
opening under N stress, which was
reflected in the observed higher transpiration rate. Jewel had maximum WUE when
N supply was 0.4 g N/plant, whereas
Tanzania had maximum WUE with
1.2 g N/plant.
WUE declined significantly over time in
all treatments, which coincided with
higher discrimination against 13 C02 toward
maturity. Responses in WUE to N supply
did not change over time.

= 0.0971; P = 0.05; b = Mean separation within

Jewel
O Tanzania

"!/!.

E 15

3'

a
10

O)

g:

:2:

0.4

0.8

1.2

1.6

N supply g/plant

Figure 3. Mean transpiration rate (MTR) over 126 days


of two sweetpotato genotypes grown in pots at six N
levels. Capital letters indicate significant differences
between N levels within each genotype (P = 0.05).
Significant differences between genotypes within N
levels are indicated by lowercase letters (p = 0.05).
Mean standard error (SE) is indicated by grey bars.

12

OJewel
Tanzania

rf

o 0.8

rl-

~ 0.6

E
~

$:

0.4

b a

f
b

bt

02

. e

0.4

.,

'

0.8

1.2

'

1.6

'

N supply g/plant

Figure 4. Transpiration water use efficiency (WUE) of


two sweetpotato genotypes grown in pots with six N
levels. Capital letters indicate significant differences
between N levels within each genotype (P = 0.05).
Significant differences between genotypes within N
levels are indicated by lowercase letters (p = 0.05).
Mean SE is indicated by grey bars.

CIP Program Report 1999 - 2000

277

There was a clase relationship between C


isotope discrimination and WUE. The
linear correlation between 813 C and WUE
was highly significant and positive, with
higher WUE obtained when discrimination
against 13 C02 was less negative. This
correlation was highest during early
growth. Until 42 DAT, 85% of the observed
variation in WUE was explained by () 13 C.
Until 84 or 126 DAT, variation in 813 C
explained 80% or 64% of variation in
WU E, respectively.

Conclusion
Results of the pot experiment showed that
both storage root initiation and storage root
growth are delayed with high N supply,
confirming observations of earlier investigators (Lowe and Wilson, 1974;
Villagarcia, 1996). High NAR of Tanzania
confirms the assumption that varietal
differences in photosynthetic efficiency
may depend on both leaf inclination and
chlorophyll content in leaves. A low
extinction coefficient allows more efficient 1ight transfer through canopies. More
vertically inclined leaves might be
advantageous, particularly under high light
conditions by minimizing the probability
of photoinhibition and increasing light
penetration to lower leaves, thereby
maximizing whole-canopy photosynthesis
(Lambers et al., 1998). Consistently higher
NAR values of Tanzania as compared with
Jewel could be attributed to more efficient
light interception of this genotype due to
its more erect leaves. Tanzania also had
higher chlorophyll content per unit leaf
area, which is characteristic for crop
species or varieties adapted to high
irradiance (Lambers et al., 1998).
ldeotypes such as Tanzania may therefore
be more efficient in total plant biomass
production under conditions of high
irradiance, which was the case in San
Ramn.
Concerning the N effect on WUE, it has
been shown that N stress leads to increased stomatal opening, thereby
reducing WUE. Varietal differences in

278

Research on Sweetpotato

WUE were due to variation in total plant


DM production, because total water
transpiration did not differ significantly
between genotypes. Tanzania had higher
NAR and showed less discrimination
against 13 C02 than did Jewel. This indicates
that carboxylation rates, rather than
stomatal opening, predominantly accounted for variation between genotypes
in WUE and 813 C. Thus, a higher plant
productivity (i.e., a higher total plant DM
production) due to increased N supply, or
due to inherently higher total plant DM
production of a genotype, is associated
with higher WUE and less negative
discrimination against 13 C0 2 A significant
positive correlation between RGR and
discrimination against 13 C0 2 was found by
Virgona and Farquhar (1996) for sunflower,
supporting this hypothesis. Faster growing
genotypes, such as Tanzania, which have
higher N contents in plant tissue, have
higher NAR and use water more efficiently. Since WUE was significantly
affected by genotype by N i nteractions,
selection for high WUE with regard to soil
fertility or N input intensity seems possible. Significant genotype effects and
genotype by N interactions also occurred
for fibrous root DM production. A larger
root system may be advantageous in case
of drought stress to reach water in deeper
soil layers. Selection for genotypes that
have a larger root system may be an
approach in breeding for increased drought
tolerance.

References
Farquhar, G.D. and R.A. Richards. 1984.
lsotopic composition of plant carbon
correlates with water-use efficiency of
wheat genotypes. Austalian Journal of
Plant Physiology 11 :539-552.
Hahn, S.K. and Y. Hozyo. 1984. Sweet
potato. In: Goldsworthy, P.R. and
N.M. Fisher (eds.). The physiology of
tropical field crops. John Wiley & Sons,
Ltd, New York, USA.
Hill, W.A., H. Dodo, S. K. Hahn,
K. Mulongoy, and S. O. Adeyeye. 1990.

Sweet patato root and biomass


production with and without nitrogen
fertilization. Agronomy Journal
82:1120-1122.
Jefferies, R.A. 1995. Physiological
determinants of genotypic differences in
carbon isotope discrimination in patato
grown in well-watered conditions.
Annals of Applied Biology 127(3):
585-592.
Jones, A. and J.C. Bouwkamp. 1992. Fifty
years of cooperative sweet patato
research 1939-1989. Southern
Cooperative Series, Bulletin No. 369,
Louisiana Agricultura! Experiment
Station, Batan Rouge, LA, USA.
Lambers, H., F.S. Chapin 111, and T.L. Pons.
1998. Plant physiological ecology.
Springer, Berlin, Germany.
Larenas de la F., V. and L.P. Accatino.
1994. Produccin y uso de la batata o
camote (lpomoea batatas L.). Instituto
de Investigaciones Agropecuarias (INIA)
- Centro Internacional de la Papa (CIP).
Serie La Platina No. 58. INIA, La
Malina, Lima, Peru. 28 p.
Lichtenthaler, H.K. and A.R. Wellburn.
1983. Determination of the total
carotinoids and chlorophylls a and b of
leaf extracts in different solvents.
Biochemical Society Transaction
603:591-92.

Lowe, S.B. and L.A. Wilson. 1974.


Comparative analysis of tuber
development in six sweet patato
(/pomoea batatas (L.) Lam) cultivars.
Annals of Botany 38:307-317.
Novozamsky, J., V.J.G. Houba, R. van Eck,
and W. van Vark. 1983. A novel
digestion technique for multielement
plant analysis. Communications in Soil
Science and Plant Analysis 14:239-248.
Radford, P.J. 1967. Growth analysis
formulae - their use and abuse. Crop
Science 7(3):171-175.
Rehm, S. and G. Espig. 1996. Die
Kulturpflanzen der Tropen und
Subtropen. Ulmer. Stuttgart, Germany.
50 p.
SAS. 1990. SAS/STAT User's Guide.
Version 6, 4th edition. SAS lnstitute lnc,
Cary, NC, USA.
Villagarcia O., M.R. 1996. Analysis of
sweet patato growth under differing
rates of nitrogen fertilization. PhD
thesis. North Carolina State Univ.,
Raleigh, NC, USA.
Virgona, J.M. and G.D. Farquhar. 1996.
Genotypic variation in relative growth
rate and carbon isotope discrimination
in sunflower is related to photosynthetic
capacity. Austalian Journal of Plant
Physiology 23:227-236.

CIP Program Report 1999 - 2000

279

Effect of GxE lnteraction on Root Yield and Betacarotene Content of Selected Sweetpotato (lpomoea
batatas (L) Lam.) Varieties and Breeding Clones
K. Manrique and M. Hermann 1

A multilocational field trial involving nine sweetpotato clones of diverse


origins was conducted across four different locations to investigate GxE
interaction effects on commercial root yield and beta-carotene pigment
content in roots. None of the high-yielding cultivars had satisfactory stability,
according to biplots for total root yield based on the additive main effect and
multiplicative interaction (AMMI) model. The lack of association of high root
yield and stable performance suggests the need for further study to elucidate
the nature of sweetpotato root-yield performance in response to varying
agroecological conditions. Beta-carotene concentration increased in the roots
of almost all tested clones when grown at a higher altitude. Clone SR 92.49923 has been identified as the most efficient at producing and accumulating
beta-carotene in relation to accumulation of root dry and fresh matter at high
altitude.

Genotype-by-envi ron ment i nteractions


(GxE) are of great interest when evaluating
the stability of breeding clones under
different environmental conditions.
Sweetpotato is grown around the world in
diverse environments, often by small
farmers in marginal soils, using few inputs.
In spite of its abi 1ity to adapt to harsh
growing conditions, sweetpotato is sensitive to environmental variation, as shown
by previous GxE studies on severa! traits
(Bacusmo et al., 1988; Collins, et al.,
1987; Hammet, 1974; Janssens, 1985;
Jong, 1974; Kamalam et al., 1978; Kanua
and Floyd, 1988; Martin et al., 1988;
Naskar and Singh, 1992; Ngeve, 1993;
Whyte, 1989).
A number of techniques for analyzing
information in a GxE study are available
(Hussein et al., 2000; Lin et al., 1986).
1

CIP, Lima, Peru.

While regression analysis attempts to


define the GxE interactions by two parameters, the objective of most univariate
stability statistics is to summarize the GxE
interaction using only one parameter.
Multivariate statistical methods have been
introduced to explore multidirectionality
and to extract more information out of this
component of phenotypic variability
(Hussein, 2000). The additive main effect
and multiplicative interaction (AMMI)
model is effective for gaining accuracy in
GxE studies (Gauch, 1992) because it
analyzes the interaction effect in a more
statistical ly robust procedure. With the
AMMI model, main effects (genotypes and
environments) are first accounted for by a
regular analysis of variance; thereafter, the
interaction (genotypes x environments) is
analyzed by a principal component
analysis (Gauch, 1992), leading to a more
exhaustive data analysis, accurate yield

CIP Program Report 1999- 2000

281

estimates., and reliable selections. Despite


the presence of large GxE interactions in
sweetpotato, the AMMI method has not
been used to study this crop.
The predominant carotenoid in
sweetpotato roots is beta-carotene
(Takahata, 1995; Takahata et al., 1993),
which represents the main source of provitamin A in the roots. In the human body,
this is converted into the essential nutrient
vitamin A (Woolfe, 1992). In order to gain
more information for successfully breeding
beta-carotene-rich varieties of sweetpotato
with higher nutritional value and yield, it
is necessary to know the influence of the
environment on these traits.
A multilocational sweetpotato trial was
conducted in seven different locations in
Peru, representing much of the
agroecological diversity where
sweetpotato is grown around the world.
The information presented in this paper is
a preliminary report involving the first four
harvested field experiments where nine
selected sweetpotato varieties and breed~
ing clones were grown in four locations
under two nitrogen levels. The objective of
this study is to determine the magnitude of
the GxE interaction, using the AMMI
model, on commercial root yield and the
content of beta-carotene pigments of these
four harvests. This study also identified
clones that performed well and remained
stable under different environmental
conditions.

Materials and Methods


The planting material was taken from
virus-free mother plants grown in greenhouses. The experimental material
consisted of advanced sweetpotato clones
of diverse origins: JPKY 16.005 (Japan),
Jewel (USA), Xushu 18 (China), SR 92.49923, DLP 2462, and ARB-UNAP 74 (Peru),
as well as the native cultivars ARB 535
(Peru) and Wagabolige and Tanzania
(Uganda).
The experimental design was a randomized complete block with three
282

Research on Sweetpotato

replications. The experimental plots had


four rows with 15 plants/row. Planting
density was 1.0 m between rows and
0.30 m between plants. Additionally, two
nitrogen levels were tested (O and 80 N)
and randomly distributed in blocks along
with genotypes. Total number of environments comprised nitrogen levels by
number of experimental sites. Field
experiments included one or two local
varieties as test crops. Various
agroecological conditions at four different
sites were considered: elevated tropic,
rainfed (Oxapampa, 1800 m altitude,
1413 mm rain, (OXAON, OXA80N)), arid
Pacific lowlands, irrigated (Tacna, 32 m
(TACON, TAC80N)), and La Malina, 240 m
(LMON, LM80N)), and mid-elevation
tropics, rainfed (San Ramn, 800 m
altitude, 1500 mm rain (SRON, SR80N)).
Five plants of the two central rows of
every experimental plot were randomly
selected and labeled at harvest time (150
days after planting) to generate composite
samples of roots and aerial parts for
determination of dry matter, beta-carotene,
chlorophyll, and starch content. The
remaining plants in central rows were
harvested to estimate root yield per plot.
Root samples were diced and homogenized, after which two composite
samples of 200 g and 500 g (fresh weight)
were taken; the first for dry-matter determ ination and the second for beta-carotene
determination and starch extraction.
Starch was extracted using a kitchen
blender for tissue maceration, followed by
several washes of the starch sediment
(Bainbridge et al., 1996, p. 80-92). Betacarotene was determined with a
spectrophotometer, as described by
Lichtenthaler and Wellburn (1983).
Data processing for determining GxE
interaction was done using the AMMI
model. Statistical computations and
estimation were carried out using procedures GLM (SAS lnstitute lnc., 1990a) and
IML (SAS lnstitute lnc., 1990b).

Results and Discussion


The analyses of variance of the AMMI
model for total root yield and betacarotene content in storage roots (Table 1)
show significant differences for environment and genotype mai n effects, as wel 1
as for GxE interaction. Genotype main
effect appears to contribute more to the
total variability of both traits than does the
environment. Following the AMMI model,
a principal component analysis (PCA) was
carried out to decompose the GxE interaction for root yield and beta-carotene
content in roots.
Only the first two PCA axes were significant (data not shown) after gatheri ng 91 %
and 88.4%, respectively, of the total
variability for both traits (total root yield
and beta-carotene content). The loadings
of the PCA axes are good indicators of
factors contributing to the variability. In
the case of genotypes, the first PCA axis
loadings show that for total root yield,
cultivars Xushu 18 and ARB 535 were
important in the GxE interaction. Whereas
for beta-carotene content in roots, the first
PCA axis' loadings show that cultivars SR
92.499-23 and ARB 535 were important.
The GxE component of the AMMI model is
based on the product of PCA seores. The
figure is not shown here, but the AMMI
biplot involving the first two significant

axes (PCA 1 and PCA2) for total root yield


showed all nine genotypes and eight
environments dispersed around the center
of the biplot, meaning that this trait shows
a large amount of variability in genotypes
and environments. This biplot showed
large positive PCA 1 seores for La Malina
(LMON and LM80N) and Tacna (TACON
and TAC80N), which was coincident with
total root yield above the grand mean (7.9
t/ha). In contrast, environments at
Oxapampa (OXAON and OXA80N) and
San Ramon (SRON) had negative PCA 1
values and were low yielding. Genotypes
exhibiting large positive PCA2 seores,
such as Jewel and Xushu 18, showed the
highest yield in La Molina, confirming that
genotypes with large positive seores yield
especially well in environments of the
same sign. The converse also holds:
genotypes with a large negative score,
such as SR 92.499-23, yielded best in
environments with a large negative score,
such as Oxapampa (OXAON and OXA80N)
and San Ramon (SRON). Cultivar Xushu 18
and environment Oxapampa are opposites,
indicating that their contributions to the
interaction would be in opposing directions. As expected, cultivar Xushu 18
ranked fifth in Oxapampa but ranked in
first place in the other environments.
Clones Tanzania and JPKY 16.005 were
distinctive and yielded better in Tacna and
Oxapampa, respectively.

Table 1. Mean squares of analysis of variance of AMMI model far total root yield (t/ha) and beta-carotene
content in storage roots (mg/100 g FM) of nine sweetpotato cultivars grown in far locations under two N
regimes (O and 80), Peru, 1999/2000.
BC content
Source of variability
df
Root yield
mg/100 g FM
P>F
t/ha
P>F
2.09
0.0001
Model
87
166.87
0.0001
381.41
0.0001
5.35
0.0001
Enviran
7
1.19
0.0001
Blocks(Ehv)
16
6.0
11.75
0.0001
0.0001
Genotype
8
905.36
0.0001
0.57
0.008
56
80.52
GxE
14
242.79
1.69
PCA1 1
0.38
PCA2 1
12
58.56
0.12
Residual
30
13.57
0.34
Error
128
11.29
Total
215
1 Principal component analysis axes, one and two respectively.

GIP Program Report 1999 - 2000

283

PCA2
0.7

XUSHU18

esRON

0.5

WAGABOUGE

0.3

ARB-UNAP 74

DLP2462
TACON

0.1
JPKY 16.005
O.O 1 - - - - - - - - - - - - - - 4 - - - - - - - T - A - C B _ O _ N - - - - - - - - - - - - - - - - 1
-0. 1

eOXAON
eOXA80N

-0.3

LMON

S~2.499-23

-0.5

JEWEL

SR80N e
ARB 535
-0.7 r..---~-~-~-~--~-4---~-~-~--~-~-~-~---.1

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

o.o

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

PCA 1

Figure 1. Biplot of principal components analysis (PCA) axis 2 vs. axis 1 far beta-carotene content in roots
(mg/100 g) for nine sweetpotato cultivars grown in eight environments.
AMMI biplot PCA 1 vs. PCA2 for betacarotene content in roots for genotypes
and environments (Figure 1) shows stable
cultivars Tanzania, Wagabolige, ARBUNAP 74, Xushu 18, and JPKY 16.005
clustered close to the center of the biplot.
More unstable clones, such as ARB 535,
DLP 2462, SR 92.499-23, and Jewel, were
far from the center. Similarly, environments La Molina (LMON and LM80N) and
Tacna (TACON and TAC80N) are more
stable than Oxapampa (OXAON and
OXA80N) and San Ramon (SRON and
SR80N). This figure also illustrates the
dominant cultivars and environments with
negative PCA 1 seores that strongly influenced the GxE interaction, such as SR
92.499-23, ARB 535, Oxapampa (OXAON
and OXA80N), and San Ramon (SR80N).

ture amplitudes, or water stress. The


increase in beta-carotene content was
variety-dependent, ranging from 59% to
600%. Clones DLP 2462, SR 92.499-23,
Wagabolige, Tanzania, Xushu 18, and
ARB-U NAP 74 were the most responsive,
at least doubling their beta-carotene
content in Oxapampa (1800 m), compared
to the lowest locations: Tacna (32 m) and

3.5

284

Research on Sweetpotato

JPKY 16.005

<) DLP 2462

X JEWEL
XUSHU 18
O SR 92.499-23

.6. WAGABOLIGE

TANZANIA

h. ARBUNAP74
~

u..

8
o

2.5

al

There was a clear trend of increasing


absolute values of beta-carotene (mg) in
100 g of fresh matter (FM) of roots in
almost all tested clones (Figure 2). This
was apparently influenced by
agroecological factors at locations,
probably associated with conditions of
increasing altitude, such as radiation
quality, mean temperatures and tempera-

+ ARB 535

1.5

0.5

TAC

LM

SR

OXA

Location

Figure 2. lnteraction of beta-carotene concentration in


roots of nine sweetpotato clones across tour
locations.

La Molina (240 m). Cultivars Jewel, SR


92.499-23, and Tanzania showed the
highest beta-carotene content.
Lower partial pressure of C02-a condition
that can be expected at high altitudesnegatively affects total dry-matter
production (Kimball, 1983). Because of
this, the beta-carotene content in root dry
matter was calculated and related to total
root dry-matter production for the same
five plant samples for each clone and
location (Figure 3). Both Jewel and SR
92.499-23 showed similar patterns of betacarotene concentration, with the largest
values by far of ali clones tested in
Oxapampa. However, the beta-carotene
yield of SR 92.499-23 was 34% higher
than that of Jewel at Oxapampa (1 800 m)
(Figure 2), which means that at
Oxapampa, SR 92.499-23 was more
efficient than Jewel in producing and
accumulating beta-carotene in root dry
matter.

Conclusions
According to the AMMI biplots, none of
the high-yielding cultivars had satisfactory
stability for total root yield. The lack of
association between high root yield and
stable performance suggests the need of
further study on the response of sweetpotato root yield to varying agroecological
conditions. However, promising new highyielding cultivars have been identified. In
Oxapampa, SR 92.499-23 and JPKY 16.005
outperformed the elite cultivar Xushu 18,
which was the top performer at other sites
and has the potential to replace locally
grown varieties in the region.
The biplot for beta-carotene content in
roots showed stability for cultivars Tanzania, Wagabolige, ARB-UNAP 74, Xushu
18, and JPKY 16.005. Clone SR 92.499-23
was less affected by high altitude for betacarotene accumu lation.
There was a trend of increasing betacarotene concentration at increasing

JPKY 16.005

+ ARB535

120.0

DLP2462
TANZANIA
~ ARB-UNAP74
X JEWEL

100.0

XUSHU18
O SR 92.499-23

80.0

.A WAGABOLIGE

60.0

ID

40.0
20.0
O.O
TAC

LM

SR

OXA

Location

Figure 3. lnteraction of beta-carotene content in


relation to root DM production of nine sweetpotato
clones across tour locations.
altitudes in the roots of almost all tested
clones. This trend will be confirmed with
three pending harvests at mid- and higher
altitudes.
Considering the importance of betacarotene as a precursor to vitamin A, it is
recommended that fu rther research and
crop improvement be done in order to
incorporate such traits as those shown by
SR 92.499-23. This is particularly important for those high-altitude areas where
poor or low-resource farming communities
are more vulnerable to vitamin-A
deficiency.

Acknowledgments
The authors thank Mr. Adelmo Prraga and
Ms. Pilar Glvez of the Universidad
Nacional Daniel A. Carrin (Cerro de
Paseo) and Dr. Ren Chvez of the
Universidad Nacional Jorge Basadre
(Tacna) for their cooperation in their
respective experimental sites; and field
technicians Mr. Roberto Martinez and
Nstor Crdenas for assisting in plant
harvests and measurements.

CIP Program Report 1999 - 2000

285

References
Bacusmo, J.L., W.W. Collins, and A. Jones.
1988. Effects of fertilization on stability
of yield and yield components of sweet
potato. HortScience 113(2):261-264.
Bainbridge, Z., K. Tomlins, K. Wellings,
and A. Westby (eds.). 1996. Methods for
assessing quality characteristics of nongrain starch staples. Part 4. Advanced
methods. Natural Resources lnstitute,
Chatham, U K.
Collins, W., L.G. Wilson, S. Arrende!, and
L.F. Dickey. 1987. Genotype x
environment interactions in sweetpotato
yield and quality factors. Journal of the
American Society of Horticultura!
Science 112(3):579-583.
Gauch, H.G. 1992. Statistical analysis of
regional trials: AMMI analysis of
factorial designs. Elsevier, Amsterdam,
Netherlands. 278 p.
Hammet, H.L. 1974. Total carbohydrate
and carotenoid content of sweet
potatoes as affected by cultivar and
area of production. HortScience
9(5):467-468.
Hussein, M.A. 2000. The statistics of
genotype x environment interaction and
genotype stability in plant breedingtheir computation, statistical test and
interpretation. http://www.nlh.no/ipf/
pu bl i kasjoner/hussei n/stabi lty/
default.htm.
Hussein, M.A., A. Bjornstad, and
A.H. Aastveit. 2000. Sasg x estab, a
SAS program for computi ng genotype x
environment stability statistics.
Agronomy Journal 92:454-459.
Janssens, M.J.J. 1985. Achieving yield
stability in sweetpotato by selection for
translocation potential in adverse
environments. In: Degras, L. (ed.).
Proceeding of the 7th Symposium of the
lnternational Society for Tropical Root
Crops, Gosier, Guadaloupe, 1-6 Jul
1985. lnstitut National de la Recherche
Agronomique (INRA), Paris. France.
p. 729-737.
Jong, S.K. 1974. Genotype x season
interaction and heritability in sweet

286

Research on Sweetpotato

potato selection experi ments.


lnternational lnstitute for Tropical
Agriculture, lbadan, Nigeria. 1 O p.
Kamalam, P., R.S. Biradar, and N. Hirshi.
1978. Stability parameters in sweet
potato (lpomoea batatas Lam.). Journal
of Root Crops 4(1 ):35-39.
Kanua, M.B. and C.N. Floyd. 1988.
Sweetpotato genotype x environment
interactions in the highlands of Papua
New Guinea. Tropical Agriculture 65
(1):9-15.
Kimball, B.A. 1983. Carbon dioxide and
agricultura! yield: An assemblage and
analysis of 430 prior observations.
Agronomy Journal 75:779-788.
Lichtenthaler, H.K. and A.R. Wellburn.
1983. Determination of the total
carotinoids and chlorophylls a and b of
leaf extracts in different solvents.
Biochemical Society Transaction
603:591-592.
Lin, C.S., M.R. Binns, and L.P. Lefkovitch.
1986. Stability analysis: Where do we
stand? Crop Science 26:894-900.
Martin, F.W., N.A. Flores, and
S.G. Carmer. 1988. ldentification of a
key environment for determination of
yield stability in sweet potato. Tropical
Agriculture 65(4):313-316.
Naskar, S.K. and O.P. Singh. 1992.
Genotype x environment interaction for
tuber yield in sweetpotato. Journal of
Root Crops 1 8(2):85-88.
Ngeve, J.M. 1993. Regression analysis of
genotype x environment interaction in
sweet potato. Euphytica 71 :231-238.
SAS lnstitute lnc. 1990a. SAS/STAT user's
guide, Version 6, Volume 2 (4th Ed.).
SAS lnstitute, Cary, NC, USA.
SAS lnstitute lnc. 1990b. SAS/IML user's
guide, Version 6, Volume 2. SAS
lnstitute, Cary, NC, USA.
Takahata, Y. 1995. Varietal differences in
storage root qual ity and physiological
factors in sweetpotato. Japan
Agricultura! Research Quarterly
29(4):215-221.
Takahata, Y., T. Noda, and T. Nagata.
1993. HPLC determination of beta-

carotene content of sweet patato


cultivars and its relationship with color
values. Japanese Journal of Breeding
43:421-427.
Whyte, J.B.A. 1989. Performance of
sweetpotato in different agroecological
zones in Cameroun. In: Akoroda,
M.O. and O.B. Arene (eds.). Tropical
root crops, promotion of root crop-based
industries: An incentive for research and

development. lnternational lnstitute for


Tropical Agriculture and lnternational
Society for Tropical Root Crops, lbadan,
Nigeria. p. 315-319.
Woolfe, J.A. 1992. Sweet patato: An
untapped food resource. Cambridge
University Press, Cambridge, UK, (In
collaboration with the lnternational
Patato Center, Lima, Peru). 643 p.

GIP Program Report 1999 - 2000

287

Using Competing Traits to Select Dual-Purpose


Sweetpotato in Native Germplasm
C.U. len-Velarde1

Eighteen accessions of sweetpotato (Jpomea batata L.) were selected using


the ratio of root/forage dry-matter production (R/F) and classified into four
groups: (1) forage, (2) low dual-purpose, (3) high dual-purpose, and (4) low
forage-high root production. Roots and vines produced between 120 and 150
days on plots of 1 O square meters were evaluated over two years and analyzed in a fixed linear covariance model, including group, accession (group),
and days as covariables. There was a significant difference among groups
(P ~ 0.01) for total forage and root dry matter, total commercial roots, and
root to forage dry matter. Significant differences of accessions within groups
on total forage dry matter and commercial roots indicate that there is enough
within-group variability among the accessions to allow a process of selection.
The least-squares means for dry-matter forage production ranged from 6.07
0.27 kg/1 Om 2 for group 1 to 4.43 0.38 kg/1 O m2 for group 4. Total root
production dry matter (DM) ranged from 2.04 0.54 kg/1 O m2 for group 1 to
8.22 0.50 kg/1 O m2 for group 4. The total commercial root produced was
2.99 0.60 kg DM/1 O m2 for group 2 and 3.32 0.55 kg/1 O m2 for group 3.
This represents 132 % more root production than group 1 and almost 36 % less
than group 4. The average weight of commercial root ranged from about 144
g to 259 g, with a significant difference between groups. The results show
that a number of accessions have potential as dual-purpose sweetpotatoes.
These are DLP-3548, DLP2481, ARB265, and ARB-158, defined as forage in
group 1; DLP-3525 and DLP-2462, defined as low dual-purpose; and ARB-394
and DLP-275A, defined as high dual-purpose.
Sweetpotato (lpomea batata L.) is one of
the 12 main plant species used as human
food throughout the world (Woolf~, 1992).
Sweetpotato roots are used for both human
and animal consumption, whereas vines
are generally used for animal feed along
with crop residues and unmarketable roots,
depending on local preferences and
customs (Woolfe, 1992). When grown for
animal fodder, the sweetpotato has severa!
advantages over other feed crops in semi1

subsistence farming systems. One major


advantage is that it may also provide food
for human consumption. An optimally
integrated livestock-management system
can use the sweetpotato's ability to
regenerate by continually or sporadically
harvesting the vines throughout the
growing season before finally harvesting
the root (Len-Velarde and Gomez, 1996).
The emphasis in this practice is on green
forage production; cutting the vines
significantly increases the yield and

CIP and ILRI, Lima, Peru.

CIP Program Report 1999 - 2000

289

growth rate of the aerial part of the plant,


while decreasing the yield of the root.
The germplasm collection held at CIP
includes multi-use sweetpotato varieties,
clones, and accessions. However, recent
breeding efforts have concentrated on drymatter content for flour and starch. But in
recent years, the demand for varieties that
produce forage has increased.
In mixed crop-livestock production
systems, one limitation on productivity is
the year-round requirement for feed. The
sweetpotato can help to overcome this
limitation without environmental damage.
lt can also contribute to the family's
nutrition and household income. Previous
studies have demonstrated that a group of
accessions have both forage and root
production that can help small farmers to
increase their income as well as the
availability of fodder (Arteaga, 1997;
Gomez and Quesada, 1 996). The favorable agronomic characteristics of these
accessions include general hardiness, low
input needs (due at least in part to the
presence of vesicular-vascular mycorrhiza), and rapid vine growth in response
to fertilizer (Tupus, 1983).
Sweetpotato has potential for intercropping, is easy to propogate, has few crop
pests and diseases, and provides good
ground cover for soil conservation. In
addition, the sweetpotato has desirable
characteristics as fodder because of its
high levels of both energy and protein
from the roots and vines, and its palatability. The vines are highly digestible for
consumption by both ruminants and
monogastrics, and the generally low
levels of enzyme inhibitors in the vines
make it suitable for pre-drying or silage
(Ruiz, 1982).
A dual-purpose sweetpotato would have a
comparative advantage over accessions
and clones selected only for roots or forage
production. lt would provide food for
human consumption, and an optimally
i ntegrated 1ivestock-management system

290

Research on Sweetpotato

could utilize the sweetpotato's ability to


regenerate by continually or sporadically
harvesting the vines throughout the
growing season befare finally harvesting
the roots.

To test this hypothesis, prel iminary studies


were done to evaluate a selected group of
accessions (Len-Velarde, 1999). The
evaluation was done on management
effects and use (digestibility and conservation). Cutting frequency, plant density, and
level of fertilization were the main
management effects considered. The
variety Helena (ARB-UNAP55) was
evaluated in San Ramon, Oxapampa,
Huachipa, and Cajamarca. Five accessions were evaluated in Oxapampa and
Huancamba. These included DLP-3548,
ARB-265, ARB-142, RCBIN-5, and Helena.
The five accessions produced an average
of 12.5 2.5 t/ha of total forage and 2.3
0.9 t/ha of total root dry matter. A cut
frequency of 90 days during crop growth
gave the best balance between forage and
root production. A cut frequency of 45
days tended to reduce regeneration,
resulting in lost forage production and nil
root production. A level of ferti 1ization
from O to 180 kg of N increased Helena's
forage production from 6.7 t DM/ha to 9.1
t DM/ha (six cuts at 45-day intervals), from
6.1 t DM/ha to 8 t DM/ha (three cuts at
90-day intervals), and from 5.4 t DM/ha to
6.2 t DM/ha at a cut interval of 135 days
(Quispe, 1997; Arteaga, 1997).
Use of sweetpotato as forage depends on
its cost and marginal production in relation
to other forages. Sweetpotato can be used
in silage with maize (75% maize without
ear-husk and 25% sweetpotato) as a form
of storage that does not diminish nutritional value (Guerra and Len-Velarde,
1998). The digestibility of the silage is
around 65% in this combination, decreasing to 48% when sweetpotato proportions
were increased. Silage using 75%
sweetpotato forage and 25% roots did not
affect milk production in dairy cows, and
feeding costs were reduced (Sanchez,
1995).

Although, several studies have indicated


the sweetpotato's potential for animal
feed, additional studies are necessary to
link root and foliage production in croplivestock farming systems. Limiting
factors, such as the activity of the trypsin
inhibitor in the vines and its effect on
ruminants and monogastrics, need to be
defined. The research on the sweetpotato
as a dual-purpose crop reported here
evaluated alternatives for assessing its
more effective use in crop-livestock
production systems.

Materials and Methods


Considering the numbers of clones available in the CIP collection at the San
Ramon experiment station, future selections for dual-purpose varieties could
include characteristics that favor livestock
production.

Clone selection
The fi rst step in the research reported here
was to analyze the germplasm, consideri ng the ratio of total dry matter of roots to
vines (R/F), and to classify it into five
groups: forage (R/F of 0-1 ), low dualpurpose (R/F > 1-1.5), high dual-purpose
(R/F > 1.5-2.0), low root production (R/F
>2.0-3.0), and high root production (R/F
>3.0) (Len-Velarde et al., 1997). Based on
this classification, 18 accessions were
selected and reclassified into four groups
as follows: group 1, forage (9 accessions);
group 2, low dual-purpose (2 accessions);
group 3, high dual-purpose (2 accessions);
and group 4, low forage-high root production (5 accessions). Accessions classified
as high root production were not included.

Statistical analysis
Four evaluations from years 1999 and
2000, carried out between 120 and 150
days on plots of 1O square meters were
analyzed according to a fixed linear
covariance model:

where vi is the group effect, c5r is the


accession (group) effect, and B is the
regression effect of the covariate. On ly
accessions with at least three observations
were considered (SAS, 1996). The variables: total dry matter for forage,
proportion of commercial and non-commercial roots (CR/NCR), and ratio of root
to forage (R/F) were analyzed, with 137
observations for each. Total root dry matter
production (107 observations) and commercial (100 observations) were analyzed; and
98 observations of commercial root
weight. Least-squares means and standard
errors are presented in Table 1.

Results and Discussion


Table 1 shows the evaluation done on 18
accessions classified into four groups.
There was a significant difference
(P ~ 0.01) among groups for total dry
matter of forage and roots, total dry matter
of commercial roots, and CR/NCR ratio as
well as R/F ratio. The covariate (days) was
not significant for total dry matter of
forage and roots, commercial weight, or
R/F ratio; it was significant for commercial
weight and proportion of commercial and
non-commercial roots (P ~ 0.05). This is
related to the large differences found
among groups for commercial root weight,
which ranged from 140.27 g to 258.87 g.
The information presented in Table 1 was
sorted by total dry matter of forage then by
total and commercial root dry matter in
arder to observe the relative importance
of the dual-purpose sweetpotato. The
results obtained indicate that the classifications fit the selected accessions.
Nevertheless, there was a significant
difference (P ~ 0.01) of accessions (group)
because of the variability and characteristics of each selected accession,
particularly in group 1. Further evaluation
of the results showed the possibility of
subdividing this group based on the CR/
NCR ratio. A subgroup of forage
sweetpotatoes without roots was then
defined and evaluated. The other subgroup
within group 1 can be considered mainly
for forage but with sorne root production. lt
CIP Program Report 1999 - 2000

291

is necessary to mention that in group 1,


four accessions had previously been
evaluated in Oxapampa and Huacho:
namely, ARB-265, DLP-3548, RCBIN-5,
and ARB-U NAP55 (Helena) (Fonseca,
1996). Results showed that there is sorne
potential far forage and roots; however, the
change in production between sites
indicated a possible genotype x environment interaction, which needs to be
considered far diffusion and use in other
areas. Helena, which has an RF ratio of O,
is considered a forage variety. However, in
certain places, such as Oxapampa, Peru,
and Africa, Helena has been found to
produce sorne roots (Ted Carey, pers.
comm.). This could be evidence of a
genotype x environment interaction.
The accessions suited mainly far forage
are DLP-3548, DLP-2481, ARB-265, and
ARB-158. For dual-purpose use, DLP-2448,
ARB-389, and DLP-3525 presenta range of
total dry-matter forage from 5.25 kg/1 Om 2
to 6.24 kg/1 Om 2 and commercial roots
ranging from 1.42 kg/1 Om 2 to 2.35 kgl
1 Om 2, with a weight of 118.25 g to 253.71
g/root.
Sorne accessions in group 1 produced a
similar quantity of dry-matter forage to
that produced by accessions in group 2;
however, the main difference is the
consistent commercial root production and
the ratio of CR/NCR. Possible dual-purpose
selections wou Id be accessions DLP-2462
and ARB-394. Similarly, in group 4, which
had high root production but less forage
dry matter, accession ARB-U NAP7 4 needs
further evaluation since it shows high root
production with forage production similar
to that of groups 2 and 3.

Acknowledgements
This work could not have done without the
fieldwork collaboration of O. Jorges and
C. Aguilar. Many thanks to Z. Huaman,
who facilitated access to all sweetpotato
accessions. Thanks are also deserved by
J. Roca, C. Fonseca, J. Arteaga, L. Quispe,
M. Rojas, C. Gomez, R. Guerra,

A. Parraga, and E. Malpartida, who have


contributed to the evaluations in searching
far dual-purpose sweetpotatoes in native
germplasm. Recognition goes to the
lnternational Development Research
Center (IDRC), Systemwide Livestock
Program (SLP), the lnternational Livestock
Research lnstitute (ILRI), and FONTAGRO
far their support of the evaluation of dualpu rpose sweetpotatoes.

References
Arteaga, J. 1997. Potencial forrajero de
cuatro clones nativos de camote
(lpomea batata [L] Lam) para la
alimentacin animal en Oxapampa.
Master's thesis. Universidad Nacional
Daniel Alcides Carrin, Oxapamapa,
Peru. 98 p.
Fonseca, C. 1996. Resultado de las
evaluaciones de un grupo de
colecciones de camote con potencial
forrajero en la Costa Central. lnternational Patato Center, Lima, Peru.
(unpublished data.)
Gomez, C. and E. Quesada. 1 996.
Evaluacin nutricional del silaje de
follaje y raz no comercial de camote
en la alimentacin de vacas lecheras.
Programa de Investigacin en
Alimentos, Facultad de Zootecnia,
Universidad Nacional Agraria La
Malina, Lima, Peru.
Guerra, R. and Len-Velarde. 1998.
Evaluation of sweetpotato and maize.
Interna! document. lnternational Patato
Center, Lima, Peru. (Unpublished data.)
Len-Velarde, CU. 1999. Searching far
dual-purpose sweetpotato in native
germplasm. Interna! document. CIP,
Lima, Peru (Unpublished.) 4 p.
Len-Velarde, C., J. Roca, J. Arteaga,
L. Quispe, and A. Parraga. 1997.
Perpectives on sweet patato: Dual
purpose varieties. In: Program report
1995-1996. lnternational Patato Center,
Lima, Peru. p. 291-294.
Len-Velarde, C. and C. Gomez. 1996. El
camote como forraje animal: Llenando
un vaco en la oferta alimentaria.

CIP Program Report 1999 - 2000

293

Congreso Latino Americano de Raizes


Tropicais, 7-11 October 1996. Centro de
Raizes Tropicais, Botucatu, Brazi l.
Quispe, L. 1997. Efecto de la frecuencia
de corte y fertilizacin sobre la
produccin forrajera de camote forrajero
ARB-UNAP55 (Helena). Master's thesis.
Universidad Nacional Daniel Alcides
Carrin, Oxapamapa, Peru. 102 p.
Ruiz, M.E. 1982. Sweet potatoes for beef
production: Agronomic and conservation aspects and animal responses. In:
Sweet potato: Proceedings of the first
international symposium. Asian Vegetable Research and Development
Centre, Tainin, Taiwan.

294

Research on Sweetpotato

Snchez, H. 1995. Valor nutricional del


ensilaje de raices no comerciales y
follaje de camote. Master's thesis.
Universidad Nacional Agraria La
Molina, Lima, Peru.
SAS. 1996. SAS for linear models. 1986
edition. SAS lnstitute lnc., Cary, NC,
USA. 211 p.
Tupus, G.L. 1983. Mycorrhiza-A possible
adaptive mechanism of sweet potato in
marginal soils. Annals of Tropical
Research 5:69-74.
Woolfe, J. 1992. Sweet potato: An untapped food resource. Cambridge
University Press, Cambridge, UK. 214 p.

Microsatellite Analysis of Genetic Diversity in


Sweetpotato Varieties from Latin America
O.P. Zhang, D. Carbajulca, L. Ojeda, G. Rossel, S. Milla, C. Herrera, and
M. Ghislain 1

Sweetpotato (Jpomoea batatas (L.) Lam.) was originally domesticated in


tropical America. The sweetpotato gene bank maintained at CIP now contains
5526 cultivated accessions from 57 countries, of which 2589 are from Latin
America. Understanding the diversity in the distribution of this germplasm is
essential for its rational management and use. In the present study, we analyzed a group of sweetpotato varieties from Latin America using
simple-sequence repeat (SSR) markers. Twelve SSR primer pairs with dinucleotide tandem repeats were screened. Six primer pairs that generated
scorable allelic information were used to type the 113 varieties. The six SSR
loci revealed a total of 70 alleles, with allele size ranging from 102 base pairs
to 173. Both the richness and the evenness of the alleles show a significant
geographical pattern in the Latin American sweetpotato gene pool.
Mesoamerica ranks the highest in terms of total number of alleles, number of
region-specific alleles, and actual heterozygosity, whereas the region of PeruEcuador ranks the lowest on all three counts. This, together with our earlier
findings based on AFLP analysis, strongly supports the hypothesis that
Mesoamerica is the primary center of diversity and most likely the center of
origin of the sweetpotato. Peru-Ecuador should be considered as a secondary
center of sweetpotato diversity. The tetrasomic inheritance of these SSRs also
supports the hypothesis that sweetpotato is an alloautohexaploid with two
non-homologous genomes.
The successful genetic conservation of any
given gene pool is largely dependent on
understanding the diversity of its distribution in a region. This information and the
organism's history of domestication are
crucial for constructing core collections.
A core collection is a limited subset of
accessions, derived from a larger
germplasm col lection, chosen to represent
the genetic spectrum of the whole collection (Brown, 1989). When the genetic
1

CIP, Lima, Peru.

diversity is known in advance, the preferred approach is to devise a strategy


where the stratification and sampling are
in proportion to the range of genetic
diversity. A combination of passport data
and information on genetic diversity from
molecular markers would therefore
enhance the formation of core col lections.
lt is commonly accepted that the
sweetpotato is of American origin. The
region between the Yucatn Pennsula of
Mexico and the Orinoco River in Venezu-

CIP Program Report 1999 - 2000

295

ela has been postulated as the center of


diversity because there is rich morphological variation in cu ltivated sweetpotatoes,
and four major American taxa of the
batata group are distributed in this region
(Austin, 1988).
The sweetpotato genebank held at CIP
maintains a total of 5526 cultivated
accessions from 57 countries, of which
2589 are from Latin America (Huamn and
Zhang, 1997). Most of these Latin American accessions are landraces or farmer's
varieties. This collection is now undergoing molecular characterization. Oominant
molecular markers, su ch as RAPO and
AFLP, are being routinely used at CIP for
eliminating redundancy (Zhang et al.,
2001) and assessing genetic diversity
(Zhang et al., 1998; 2000).
Although these dominant markers have
been proved genetically informative for
sweetpotato, they do not properly contribute to our understanding of the allelic
diversity in this col lection. In addition, the
lack of sequence specificity in randomly
amplified polymorphic ONA (RAPO) and
amplified fragment length polymorphism
(AFLP) markers limits their cross-lab
appl ication for variety identification.
Therefore, a polymerase chain reaction
(PCR)-based, sequence-tagged, co-dominant marker system is needed to play
a complementary role in gene bank
management.
Simple-sequence repeats (SSRs), also
called microsatellites, are becoming the
most important molecular markers in both
animals and plants. SSRs are stretches of
1 to 6 nucleotide units repeated in tandem
and randomly spread in eucaryotic
genomes. They are highly polymorphic
because of the high mutation rate affecting the number of repeated units.
Polymorphisms of this length can be easily
detected on high-resolution gels (e.g.,
sequencing gels) by running PCR-amplified
fragments obtained using a unique pair of
primers flanking the repeat (Weber and
May, 1989).

296

Research on Sweetpotato

A group of SSRs have already been


developed in sweetpotato {Jarret and
Bowen, 1994). These have been further
screened and applied in paternity analysis
in sweetpotato and its wild relative
species (Buteler et al., 1997; 1999). In this
paper, we report the application of these
SSRs to the assessment of genetic identity
and genetic diversity in a group of
sweetpotato varieties from Latin America.
We show the significant variation in the
distribution of allelic diversity in the Latin
American sweetpotato gene pool and
discuss the importance of SSR markers in
future germplasm management.

Materials and Methods


All plant materials were obtained from the
sweetpotato germplasm collection at CIP.
Four to 1 O sweetpotato landraces were
randomly selected from each country
included in this study, resulting in a total
of 113 accessions with a geographic
coverage ranging from Mexico to Peru
(Table 1 ). Healthy young leaves were
collected from accessions maintained in a
screen house and in vitro. ONA was
extracted following the method of Ooyle
and Doyle (1990).
The SSR primers for sweetpotato were
originally developed by Jarret and Bowen
(1994). We used the Mappair primer sets
commercially available from Research
Genetics (Huntsville, AL, USA). Eight
pairs of these primers were pre-screened
by Buteler et al. (1997; 1999) at the
Sweetpotato Breeding Program of
Louisiana State University.
A non-radioactive PCR was carried out
following CIP's SSR genotyping protocol.
Annealing temperature was adjusted,
based on the Tm of the oligo-nucleotides.
For screening of primers, the amplification
products were first confirmed on 3%
agarose gel with eight varieties. Primer
pairs amplifying at least one PCR fragment
in every variety were further tested on the
complete set of varieties using a 6%
denaturing sequencing gel (National

Table 1. List of the 113 Latin American sweetpotato accessions used for microsatellite analysis.
Country
Collection
Country
Collection
Country of
Collection
Code
Code
Code
of origin
of origin
No.
origin
No.
No.
1
Peru
CIP ARB 386
39
Colombia
DLP 2104
77
Guatemala
GUA 948
2
DLP 1900
Peru
40
Colombia
DLP 1793
78
Guatemala
GUA STRA
Colombia
DLP 1858
3
Peru
DLP 1922
41
79
Guatemala
GUA 494
4
DLP 206
42
Colombia
DLP 1736
80
Mexico
CTX 15
Peru
43
Colombia
DLP 1771
81
Mexico
CTX7
5
Peru
DLP 2
DLP 1870
82
Mexico
6
Peru
DLP 1090
44
Colombia
CTX 22
DLP 1011
7
Peru
DLP 2344
45
Colombia
83
Mexico
CTX24
DLP 1685
84
Mexico
Peru
ARB 234
46
Colombia
CTX 31
8
DLP 1731
Mexico
DLP 5314
47
Colombia
85
A 160
9
Peru
CTX 33
DLP 3824
48
Colombia
DLP 2046
86
Mexico
10
Peru
11
RCB IN- 90
49
Colombia
DLP 1879
87
Mexico
101438
Peru
Colombia
Mexico
NIAR 221
12
DLP 1921
50
DLP 1737
88
Peru
ARB 355
51
Colombia
DLP 1785
89
Mexico
CATIE 9232
13
Peru
52
Colombia
DLP 2151
90
Mexico
CTX 29
14
Peru
ARB 455
RCB-IF-30
15
Peru
DLP 2298
53
Colombia
DLP 976
91
Mexico
16
Peru
DLP 253
54
Colombia
DLP 971
92
Mexico
CATIE 9257
LL87-1799
17
Peru
RCB IN-199
Colombia
93
Mexico
CTX 31
55
Colombia
DLP 972
94
Mexico
CTX 32
18
Peru
DLP 909
56
DLP 1755
Mexico
19
Ecuador DLP 1192
57
Colombia
95
CTX 12
Venezuela
DLP 2896
Mexico
CTX 34
20
Ecuador DLP 1161
58
96
Venezuela
DLP 869
97
Mexico
CTX5
Ecuador DLP 1456
59
21
Venezuela
DLP 824
98
Mexico
CTX 16
22
Ecuador DLP 1475
60
DLP 806
Mexico
61
Venezuela
99
CTX 9
23
Ecuador DLP 1449
Nicaragua
Venezuela
DLP 868
DLP 4678
24
Ecuador DLP 1447
62
100
DLP 2868
101
Nicaragua
DLP 4686
Ecuador DLP 1156
63
Venezuela
25
Nicaragua
64
Venezuela
DLP 2869
102
DLP 4617
26
Ecuador DLP 1487
Venezuela
DLP 2902
Nicaragua
DLP 4675
Ecuador DLP 1153
65
103
27
Venezuela
DLP 2884
104
Pan ama
DLP 3874
Ecuador DLP 1493
66
28
Venezuela
Pan ama
DLP 3834
67
DLP 2876
105
29
Ecuador DLP 1186
Venezuela
DLP 2838
106
El Salvador SVG 27
Ecuador DLP 1397
68
30
Venezuela
DLP 800
107
El Salvador SVG 12
31
Ecuador DLP 1257
69
Venezuela
DLP 807
El Salvador SVG 24
Ecuador DLP 1149
70
108
32
Venezuela
DLP 822
El Salvador SVG 8
71
109
33
Ecuador DLP 1435
Honduras
72
Venezuela
DLP 792
110
DLP 4545
34
Ecuador DLP 1157
DLP 765
111
Honduras
DLP 4494
73
Venezuela
Ecuador DLP 1231
35
Honduras
DLP 4521
Venezuela
DLP 790
112
74
Ecuador DLP 1498
36
Honduras
DLP 4558
Venezuela
DLP 2881
113
75
37
Ecuador DLP 1484
Venezuela
DLP 791
76
Ecuador DLP 1405
38
Diagnostic, Atlanta, GA, USA). Visualization was accomplished by silver staining
(Bassam et al., 1991 ).
Each SSR variant was treated as an allele.
The alleles were numbered sequentially
starting with the shortest sequence. The
co-run sequence ladder in the gel helped
us to identify single base differences
between any two given samples. The SSR
allele composition of each variety was
determined. We maintained a common

approach to the scoring of the gel by


placing position 1 as the lowest molecular
weight species and position 2 as the next
higher weight and so forth. By running a
dideoxy-sequence next to several of the
reactions, we were able to distinguish
between bands that differed in size by a
single base.
Genotypes were scored as +/- for each
allele. For the purposes of this study, we
defined three regions for the 1O countries:

CIP Program Report 1999 - 2000

297

(1) Mesoamerica (Mexico and Central


America), (2) Colombia-Venezuela, and
(3) Peru-Ecuador. The total number of
alleles, average number of alleles per
locus, and percentage of heterozygous
individuals were calculated for each
region (Nei, 1973). The average locus
heterozygosity was determined by
ALH = 100 - [S(l -{m/N})], where mi is the
number of accessions with mono-allelic
genotypes for SSR locus i and N is the
number of accessions scored.

Results and Discussion


Number of alleles
Of the 12 pairs of SSR primers, six generated a reproducible banding pattern. An
SSR profile example, loci 242, is presented
in Figure 1. The number of alleles varies
greatly among the SSR loci. Locus 18316
only detected six alleles, whereas locus
18324 detected 30 in the 113 accessions.
In total, the six SSR loci detected 70 SSR
variants with an allele size ranging from
1 02 bp to 1 73 bp (data not shown). Most of
the alleles in each locus can be arranged
on a continuous ladder, with adjacent
steps differing by two base pairs. There
was no correlation between number of

repeats and the level of polymorphism


detected.
The total number of alleles and the
number of alleles per locus are not evenly
distributed among geographic regions.
Mesoamerica has the highest total number
of alleles (50), Peru-Ecuador has the lowest
(41 ), and Venezuela-Colombia is in
between (46). There is a total of 14 regionspecific alleles (alleles found in one
region but missing in the other two), of
which eight are found in Mesoamerica,
four in Venezuela-Colombia, and only two
in Peru-Ecuador. In addition, the predominant alleles in Peru-Ecuador differed
greatly from the other two regions in loci
18255 and 18286. Allele 18255/1 is predominant in Mesoamerica and in
Venezuela-Colombia, shared by approximately 30% of the accessions from these
two regions, but in Peru-Ecuador, fewer
than 10% of the accessions have this
allele (Table 2).

Actual heterozygosity
8ecause of the polyploid nature of
sweetpotato, the dosage effects of an SSR
allele (simplex, duplex, triplex, etc.)
cannot be differentiated. Therefore the
allele frequency could not be counted, as

- - 1824217
18242/6
18242/5
18242/4
18242/3
- - - - 18242/2

Figur~

1. SSR-based m~lti-allelic fingerprints in 25 Latin American sweetpotato accessions. Unes and letters on
the nght mark the putat1~e alleles. Allele~ were.numbered starting with the shortest sequence. In this example,
the lowest band (allele) 1s #1 and the h1ghest 1s #7, with alleles 2-6 lying in between these points. When the
autorads were scored, genotypes were scored as 1/0 far each allele.

298

Research on Sweetpotato

Table 2. Regional distribution of allelic diversity in Latin American sweetpotato germplasm. The allelic
diversity is measured by total number of alleles, region-specific alleles, and actual heterozygosity based
on 6 SSR loci.
lndividuals (%)
Region
Total Region- Sample
Dialleles specific size
MonoTriTetraPentaHexaalle les
allelic allelic allelic
allelic
allelic
allelic
Peru-Ecuador
41
47.9
27.4
17.4
o
o
2
38
7.4
Colombia-Venezuela
46
4
29.5
45.3
20.5
4.7
o
o
38
o
o
Mesoamerica
37
28.6
42.7
20
8.7
50
8
70
14
113
35.3
38.5
19.3
6.9
o
o
Total
it can in diploid species, and the expected
heterozygosity (Nei, 1973) cannot be
estimated. However, the observed heterozygosity can be calculated by using the
percentage of multi-allelic accessions. The
actual heterozygosity varies greatly across
the srx loci, ranging from 0.31 (IB255) to
0.92 (IB316), with a mean of 0.60. This
value is reasonable, considering that
sweetpotato is a hexaploid outbreeding
species. The actual heterozygosity in
S. tuberosum was reported to range from
0.477 to 0.502 (Bonierbale et al., 1993).
A similar value (0.495) was found in CIP's
holding of S. andigena using allonym
analysis (Huamn et al., 2000). However,
sweetpotato has a much higher percentage
of tri-al lel ic arid tetra-al lel ic genotypes
than patato does. The present study found
approximately 25% tri- and tetra-allelic
accessions in sweetpotato, but there are
nly 3.2% in patato (Huamn et al.,
2000). In other words, sweetpotato is
general ly mu ch more heterozygous than
patato.
The actual heterozygosity also varies
greatly among the three regions, similar to
allele numbers. lt is 0.714 in Mesoamerica
and 0.705 in Venezuela-Colombia, but it is
only 0.521 in Peru-Ecuador.
The richness and evenness of alleles in the
three tropical American regions are fully
compatible with our recent diversity
analysis for the same three regions using
AFLP (Zhang et al., 2000), which found
that Central America has the highest intraspecific diversity and Peru-Ecuador the

lowest. This provides further evidence that

Mesoamerica is the primary center of


diversity and most likely the center of
origin of the sweetpotato.

Autopolyploid vs. allopolyploid


Sweetpotato is a hexaploid species;
therefore, each individual genotype could
contain between one and six alleles at any
one locus, assuming it is an autopolyploid.
However, in this study, the maximum
number of alleles from any given variety
was four. Varieties showing five or six
distinct alleles per locus (hexa- and pentaallelic) were completely missing. The lack
of hexa- and penta-al lel ic genotypes
indicates that sweetpotato is not an
autohexaploid and the SSR alleles are in a
pattern of tetrasomic inheritance. This
agrees with previous findings that these
SSRs were in tetrasomic segregation in the
sweetpotato (Buteler et al., 1999). lt is also
compatible with the result of our genome
mapping, which demonstrated that
sweetpotato has a high level of polysomic
inheritance of homologous chromosomes,
but with partial preferential chromosome
pairing.
Both allopolyploidy (Jones, 1965) and
autopolyploidy (Nishiyama, 1982) have
been proposed as the nature of
polyploidization in sweetpotato. Shiotani
and Kawase (1989) proposed that
sweetpotato has a 'tetra-disomic' genetic
constitution with two different genomes
(B1 B1 B2B2B2B2). The tetrasomic inheritance of these SSRs in the present study
provides new evidence to support the
hypothesis that there are two non-homolo-

CIP Program Report 1999 - 2000

299

gous genomes involved in sweetpotato and


that the plant is most likely an
al loautohexaploid.

available to serve the needs of sweetpotato germplasm management, both at


CIP and in national programs.

SSR as fingerprint for variety


identification

Acknowledgements

The extraordinary discriminatory capacity


of microsatellite markers observed in other
species has been confirmed in the present
study. All 113 accessions in this study
can be fully identified with as few as
four SSRs. lt is highly desirable to have
repeatable fingerprints for germplasm
management, and the presence of easily
scorable, unique alleles and/or allele
combinations makes microsatellite
markers an ideal system for variety
identification. Other PCR-based fingerprints, such as AFLP and RAPO, are also
able to differentiate varieties; but often
ha ve mu ch lower repeatabi 1ity.
Microsatellite markers have their own
limitations when used on polyploid
species. The most important limitation is
that the real genotype of a hexaploid
individual cannot be revealed because the
PCR-based nature of SSRs cannot differentiate the dosage effect of a given al lele
(i.e., it cannot tell the difference between
simplex and duplex). The occurrence of
null alleles is another possible problem
with the use of microsatellite markers in
highly outbreeding, heterozygous species
(Powel 1 et al., 1996). Because of these
limitations, the allele frequency of a given
germplasm pool cannot be calculated, and
classical popu lation genetics cannot be
fully applied. Nevertheless, the nature of
microsatellites as being selectively
neutral, co-dominant, and sequencetagged makes them a very useful tool for
germplasm management, as wel 1 as for
genome mapping of sweetpotato.
New SSRs for sweetpotato are currently
being developed at CIP using high quality
genomic DNA library enriched for multiple types of SSR motifs. Our goal is to
ha ve a set of wel 1-characterized SSRs

300

Research on Sweetpotato

We thank Robert Jarret, Don La Bonte, and


Mario Buteler for providing the sequences
of screened pri mers.

References
Austin, D.F. 1988. The taxonomy,
evolution and genetic diversity of
sweetpotatoes and related wild species.
In: Gregory, P. (ed.). Exploration,
maintenance, and utilization of
sweetpotato genetic resources.
lnternational Patato Center, Lima, Peru.
p. 27-60.
Bassam, B.J., G. Caetano-Anolls, and
P.M. Gresshoff. 1991. Fast and sensitive
silver staining of DNA in
polyacrylamide gels. Annals of
Biochemistry 196:80-83.
Bonierbale, M.W., R.I. Plaisted, and
S.D. Tanksley. 1993. A test of the
maximum heterozygosity hypothesis
using molecular markers in tetraploid
potatoes. Theoretical Applied Genetics
86:481-491.
Brown, A.H.D. 1989. Core collection: A
practica! approach to genetic resources
management. Genome 31 :818-824.
Buteler, M.I., R.L. Jarret, and D.R. La
Bonte. 1999. Sequence characterisation
of microsatellites in diploid and
polyploid lpomoea. Theoretical Applied
Genetics 99:123-132.
Buteler, M.I., D.R. La Bonte, and
R.E. Macchiavelli. 1997. Determining
paternity in polyploids: Hexaploid
simulation studies. Euphytica 96:353361.
Doyle, J.J. and J.L. Doyle. 1990. lsolation
of plant DNA from fresh tissue. Focus
12:13-15.
Hu amn, Z. and O.P. Zhang. 1997.
Sweetpotato. In: Fuccillo, D., L. Sears,
and P. Stapleton (eds.). Biodiversity in

trust. Cambridge Univ Press,


Cambridge, U K. p. 29-38.
Huamn, Z, R. Ortiz, O.P. Zhang, and
F. Rodrguez. 2000. lsozyme analysis of
entire and core col lections of Solanum
tuberosum subsp. andigena patato
varieties. Crop Science 40:273-276.
Jarret, R.L. and N. Bowen. 1994. Simple
sequence repeats (SSR's) for sweetpotato germplasm characterization.
Plant Genetic Resources Newsletter

100:9-11.
Janes A. 1965. Cytological observations
and ferti 1ity meas u rements of
sweetpotato [/pomoea batatas (L.) Lam].
Proceedings of the American Society of
Horticultura! Science 86:527-537.
Nei, M. 1973. Analysis of gene diversity in
subdivided populations. Proceedings of
the National Academy of Science USA

70:3321-3323.
Nishiyama, l. 1982. Autopolyploidy
evolution of the sweetpotato. In:
Villareal, R.L. and T.D. Griggs (eds.).
Sweet patato, proceedings of first
international symposium. AVRDC Publ.
No. 82-1 72, Tainan, Taiwan, China.

p. 263-274.
Powell, W., G.B. Morgante, C. Andre,
M. Hanafey, J. Vogel, S. Tingey, and
A. Rafal ski. 1 996. The comparison of

RFLP, RAPO, AFLP and SSR (microsatellite) markers for germplasm


analysis. Molecular Breeding

2:225-238.
Shiotani, l. and T. Kawase. 1989. Genomic
structure of sweet patato and hexaploids
in lpomoea trfida (H.B.K.) Don. Japan
Journal of Breeding 39:57-66.
Weber, J.L. and P.E May. 1989. Abundant
class of human DNA polymorphism
which can be typed using the
polymerase chain reaction. American
Journal of Human Genetics 44:388-386.
Zhang, O.P., J. Cervantes, Z. Huamn, and
M. Ghislain. 2000. Assessing genetic
diversity of sweet patato (/pomoea
batatas (L.) Lam.) varieties from tropical
America using AFLP. Genetic Resources
and Crop Evolution 47:659-665.
Zhang, O.P., M. Ghislain, Z. Huamn,
A. Golmirzaie. and R.J. Hijmans. 1998.
RAPO variation in sweetpotato
[/pomoea batatas (L.) Lam] varieties
from South America and Papua New
Guinea. Genetic Resources and Crop
Evolution 45:271-277.
Zhang, O.P., Z. Huaman, F. Rodrguez,
and M. Ghislain. 2001. ldentifying
duplicates in sweetpotato [/pomoea'
batatas (L.) Lam] varieties using RAPO.
Acta Horticulturae 546:535-541.

CIP Program Report 1999 - 2000

301

A Genetic Linkage Map of Sweetpotato (lpomoea


batatas (L.) Lam.) Based on AFLP Markers
A. Kriegner 1, J.C. Cervantes2, K. Burg1, R.O. Mwanga3 , and O.P. Zhang2

Cenetic linkage maps of Tanzania and Bikilamaliya sweetpotatoes (lpomoea


batatas (L.) Lam), based on amplified fragment-length polymorphism (AFLP)
were developed using a two-way pseudo-testcross. At a LOO score of 5, a
total of 656 maternal and 469 paternal markers were ordered in 90 and 80
linkage groups, respectively, covering a total map length of 3655.6 cM for
Tanzania and 3011.5 cM for Bikilamaliya. The average distance between
markers is 5.9 cM, with only four intervals exceeding 30 cM. Mapping was
carried out in two steps. Linkage among simplex markers provided a map
framework to which multiple-dose markers were added. The maternal maps
contained 91 duplex markers, which allowed identification of 13 homologous
co-segregating groups; paternal maps contained 69 duplex markers, allowing
identification of 1 O homologous co-segregating groups. The exploitation of
215 double-simplex bridging markers detected 15 linkage groups as homologous maternal and paternal groups. Type of polyploidy (autopolyploidy vs.
allopolyploidy) was evaluated using the ratio of repulsion to coupling linkages. The absence of linkages in the repulsion phase indicated a high level of
polysomic inheritance of homologous chromosomes. The detection of a few
repulsion linkages at a lower confidence level, however, demonstrated that
there is partial preferential chromosome pairing in sweetpotato.
Maps of highly dense genetic linkages are
powerful tools for the localization and
map-based cloning of genes, as well as
marker-assisted breeding. They also
provide information for understanding the
biological basis of complex traits (Lee,
1995) and polyploidy (Da Silva, 1993).
Different approaches are available for
direct mapping of polyploid genomes,
most of which have been based on the use
of simplex (single-dose) markers (Wu et
al., 1992) derived from only one parent
and segregating in a 1 :1 ratio. The single1

Austrian Research Center Seibersdorf (ARCS), Seibersdorf,


Austria.
2 CIP, Lima, Peru.
3
National Agricultura! Research Organization (NARO), Kampala,
Uganda.

dose restriction fragment (SDRF) method


has been applied to various polyploid
crops such as alfalfa, apple, sugarcane,
and patato. Multiplex markers have been
exploited to identify homologous cosegregati ng groups in Saccharum (Da
Silva, 1993; Ripol et al., 1999) and alfalfa
(Yu and Pauls, 1993). Da Silva (1993) was
able to include duplex and triplex markers
in simplex marker maps and to identify
homologous groups.
Sweetpotato is a hexaploid species
(2n=6x=90) with a base chromosome
number of 15 (Janes, 1965). Cytological
studies of sweetpotato are difficult be-

cause of the high number of small

CIP Program Report 1999 - 2000

303

chromosomes and large number of basic


groups, so the nature of its polyploidy
remains uncertain. Both allopolyploidy
(Jones, 1965; Magoon et al., 1970) and
autopolyploidy (Nishiyama et al., 1982;
Shiotani, 1988) have been proposed.
Shiotani and Kawase (1989) postulated the
genome constitution of sweetpotato as
(B 1 B1 B2 B2 B2 B2 ), but the degree of homology
could not be estimated with accuracy.
Ukoskit and Thompson (1997) reported a
polysomic inheritance in sweetpotato
based on the segregation ratio and relationship of the genetic linkages of RAPO
markers.

between two African sweetpotato


landraces (Tanzania and Bikilamaliya).
Tanzania is the most widely grown
sweetpotato cultivar in sub-Saharan Africa
and is resistant to the sweetpotato virus
disease (SPVD) complex in East Africa.
Bikilamaliya is susceptible to SPVD under
the same environmental conditions.
Linkage mapping was carried out on a
subset of 94 randomly selected plants;.
Genomic DNA was extracted from young
leaves of greenhouse-grown sweetpotatoes
using the CTAB method (Doyle and
Doyle, 1990).

AFLP assays and marker scoring


The commonly used mapping population
for outbreeding trees and polyploids is the
progeny of a cross between two unrelated
heterozygous parents (a two-way pseudotestcross) (Grattapaglia and Sederoff,
1994). With this mapping population,
linkage analysis of SDRFs in coupling
phase results in two separate linkage maps
for each parent based on the male and
female sources of markers. The ratio of
repulsion linkages versus coupling linkages was used to evaluate type of
polyploidy (Da Silva and Sorells, 1996).
In this paper, we report the genetic
inheritance, segregation, and linkage of
amplified fragment-length polymorphism
(AFLP) markers in two hexaploid
sweetpotatoes (2n=6x=90). We used an F1
population to construct two separate
parental maps, integrating simplex and
multiplex markers, and analyzed the
genome constitution based on the pattern
of inheritance and segregation of the
markers. These are the first reported
genetic-linkage maps that have substantial
coverage of the sweetpotato genome.
They provide a framework for tagging the
genes and quantitative trait loci (QTL) of
the economic traits of the sweetpotato.

Materials and Methods


Plant materials and DNA extraction
Seeds of an F1 mapping population
originated from a two-way pseudo-testcross

304

Research on Sweetpotato

A procedu re adapted from Zabeau and Vos


(1993) was followed for the AFLP analysis.
For the selection of primer combinations
(PCs), the two parental varieties (Tanzania
and Bikilamaliya) were screened with 240
EcoRl/Msel PCs. The PCs yielding a high
number of polymorphic fragments for each
parental line and with a total number of
fragments ranging between 50-100 were
selected for generating AFLP markers.
Autoradiographs were manual ly scored for
the absence (O) or presence (1) of AFLP
marker bands. Markers were chosen on the
basis of their presence in one parent and
absence in the other, or presence in both
parents. Scored markers were divided into
three groups depending on the presence or
absence in each parent.

Segregation ratio
The assessment of marker dosage was
done by the expected segregation ratios
(presence:absence) of F1 s for AFLP
markers present in one parent in accordance with the allele dosage for three
cytological theories of sweetpotato (Table
1). To classify fragments into dosage
groups, acceptance regions for simplex,
duplex, triplex, and double-simplex
markers were constructed.

Simplex markers. All markers present in


one parent and absent in the other were
tested for goodness of fit to the 1 :1 segregation ratio in the progeny by a x2 test at

Table 1. Expected segregation ratios (presence:absence) far the inheritance of a single gene in a hexaploid
plant, according to three cytological theories based on allele dosage.
Allohexaploid
Marker
Autohexaploid
Tetradiploid
(disomic)
dos e
(hexasomic)
(tetradisomic)
Aa aa aa 1:1
Simplex
Aaaaaa 1:1
Aaaa aa 1:1
aaaa Aa 1:1
AAaa aa 5:1
Aa Aa aa 3:1
Duplex
AAaaaa 4:1
Aaaa Aa 3:1
AA aa aa aaaa AA Aa Aa Aa 7:1
Triplex
AAAaaa 19:1
AAAa aa AAaaAa11:1
AA Aa aa Aaaa AA AAAA aa AA Aa Aa Quadruplex
AAAAaa the 99% confidence level, allowing type 1
error= type 11 error = 0.5%. The null
hypothesis (1 :1 segregation) was tested
against the alternative hypothesis (3:1
segregation) since al 1 non-single dose
ratios would be 3:1 or greater, regardless of
whether sweetpotato is an auto-, autoallo-,
or allopolyploid (Wu et al., 1992).

Duplex markers. Among the three cytological theories listed in Table 1,


allohexaploidy is the least likely genetic
configuration in sweetpotato (Nishiyama
et al., 1982; Shiotani, 1988; Ukoskit and
Thompson, 1997). Under the assumption of
either hexasomic or tetradisomic inheritance, duplex markers are expected to
segregate 4:1 (hexasomic) and 3:1, ~:1
(tetradisomic) (Table 1 ). A 4:1, 5:1 segregation ratio corresponds to a duplex
marker of fragments located on homologous chromosomes, compared to a 3:1
segregation ratio for a duplex marker
where the individual fragments are located
on nonhomologous chromosomes. The
latter are not useful for the estimation of
linkages among simplex/duplex and
duplex/duplex marker pairs. As the
population size used in this experiment
was not large enough to distinguish
between hexasomic and tetradisomic
situations, we looked for a common subset
of 4:1, 5:1 segregating markers as tested
by a 2 test for goodness of fit at a 1 0%
confidence level. This set of markers is

sufficient to distinguish duplex markers

from the expected segregation ratios for

triplex markers (11 :1, 19:1) and simplex


markers (1 :1 ). To separate the 3:1 from 4:1
or 5:1 segregation ratios, however, would
require an extremely large population
size. Therefore, in our study, calculation of
4:1 and 5:1 segregation ratios was not
totally exclusive of 3:1 markers.

Triplex markers. Triplex markers are


expected to segregate 11 :1 under
hexasomic inheritance and 19:1 under
tetradisomic. As these segregation ratios
cannot be distinguished properly, and
estimates of recombination fractions
among simplex and triplex markers are
associated with large standard errors,
triplex markers were not analyzed.

Quadruplex markers. Quadruplex or


higher-dose markers were not expected to
result in observable segregation in the
offspring, except in the case of a
quadruplex marker with random chromatid
segregation, which would result in a
genotype not showing the marker.

Double-simplex markers. AFLP fragments


present in a single-dose condition in both
parents are expected to segregate in a 3:1
ratio in the population, as tested for H 0 at
a 10% significance level to avoid overlapping with the 11 :1 segregation class for
duplex-simplex markers.
Estimation of recombination fraction (r)
and linkage mapping
Linkage analysis and map construction
was carried out in two steps.

GIP Program Report 1999 - 2000

305

In the first step, two parental framework


maps (maternal and paternal) were
constructed using simplex markers.
Simplex markers in coupling behave the
same in diploids as in polyploids and can
be mapped with standard methods. Since
detection of repulsion linkages in polysomic polyploids requires extremely large
populations, they were not included.
Recombination fractions and marker arder
were obtained under the backcross model
using JoinMap 2.0 software (Stam and Van
Ooijen, 1995).
In the second step, duplex and doublesimplex markers were inserted in the fixed
marker order of the parental simplex
framework maps. Recombination fractions
were estimated by numerically maximizing the log-likelihood (Stam and van
Ooijen, 1995), assuming the random
pairing of homologous chromosomes and
the absence of double reduction.

Duplex markers. As the true underlying


mode of chromosome inheritance
(hexasomic vs. tetradisomic) in
sweetpotato remains uncertain, linkage of
simplex/duplex and duplex/duplex markers
in coupling were calculated for both
hexasomic and tetrasomic inheritance,
applying phenotypic frequencies given by
Ripol et al. (1999). The sets of pairwise r
and LOO estimates were alternatively
mapped to the simplex framework map
using standard procedures (JoinMap 2.0).
All duplex markers placed at different map
sites for the hexasomic and tetrasomic
inheritance types were removed, resulting
in identical maps with an equal number of
markers and marker arder, with slightly
different map distances. The map with the
smallest map size was selected for adding
double-simplex markers.
Double-simplex markers. Recombination
fractions for simplex/double-simplex and
double-simplex/double-simplex configurations of marker pairs were estimated
according to Meyer et al. (1998).
The final maps comprised simplex, duplex,
and double-simplex markers, with each
306

Research on Sweetpotato

linkage group corresponding to an individual chromosome. Two linkage groups


containing the same duplex marker are
likely to be homologous. A duplex marker
should not align to more than two linkage
groups, unless a third group is an unconnected part of one of them. Two linkage
groups were declared homologous female
and male linkage groups if they possessed
the same double-simplex (bridging)
marker. Final maps were drawn using
OrawMap 1.1 software (Van Ooijen 1994).
The expected genome coverage was
estimated according to Bishop et al.
(1983).

Estimation of polyploidy
For species with levels of high polyploidy,
the ratio between the number of detected
repulsion versus coupling linkages may
provide a crude measurement of preferential chromosome pairing (Sorrels, 1992;
Wu et al., 1992). Using Mapmaker's twopoint linkage analysis (LOO = 4, r < 0.3),
the repulsion linkage was analyzed
between an original marker and inverted
markers.

Results and Discussion


The high multiplex ratio of the AFLP
method was used to produce a large
number of highly reliable genetic markers
for linkage analysis. A total of 112 out of
the 240 primer combinations screened
were selected and used to generate AFLP
markers for 94 individuals of the mapping
population. Three genotypes were removed from the data set because they
contained many missing data and could
have affected the estimation of the
genetic distances among the markers.

Segregation ratio
Clearly scorable markers were separated
according to their presence in the male
and female parents and in both. The
observed segregation ratio for, each of the
808 female and 641 male markers are
shown in Figures 1 and 2. The distributions
appear with a large group centered around

35
30
25
>.

o
e

(])

20

:J

CT

(])

u:

15
10
5

o
30

40

35

45

50

55

60

65

70

75

80

85

90

95

100

Observed segregation ratio (%)

Figure 1. Observed segregatian ratias far 808 markers in Tanzania.

40
35
30

>.

25

o
e

(])

:J

CT

(])

u:

20
15
10

10

19

28

37

46

55

64

73

82

91

100

Observed segregation ratio (%)

Figure 2. Observed segregatian ratias far 641 markers in Bikilamaliya.


CIP Program Report 1999 - 2000

307

5% of the expected segregation ratio for


simplex markers and a smaller group
around 8% where ratios of 3:1, 4:1, or 5:1
duplex markers are expected. At higher
frequencies, triplex markers or quadruplex
markers were observed, as expected u nder
random chromatide segregation. Segregation analysis resulted in 951 simplex
markers. A total of 298 markers fitted into
a 4:1 or 5:1 acceptance region as expected for duplex markers under
hexasomic and tetrasomic inheritance,
respectively. Eighteen markers showed a
segregation ratio significantly lower than
1 :1, which was the lowest expected
segregation ratio; these were removed as
skewed markers. Of 540 segregating
markers present for both parents, 21 5
markers fitted into the 3:1 segregation
ratio for double-simplex markers.
Estimation of polyploidy based on
repulsion linkage
Type of polypoidy (autopolyploidy vs.
allopolyploidy) was investigated using the
ratio of coupling to repulsion linkages. For
diploids and disomic polyploids (allopolyploids), linkages in the repulsion phase
should be equal to those in the coupling
phase, compared to autopolyploids, where
an extremely large sample size is required
for detecting repulsion linkages (Wu et al.,
1992). Simplex marker seores were
inverted and added to the original marker
seores. In diploids-as in allopolyploids-a
homologous recessive allele of dominant
single-dose fragments may be simulated
by its 'marker image' and this can be used
to detect linkages in the repulsion phase.
In autopolyploids, a total of m-1 alleles
(where m = ploid number) hide behind a
dominant simplex marker, so the image of
a simplex marker score only approximates
a recessive allele. Calculation of repulsion
linkages in polysomic genomes, therefore,
is limited to codominant markers.
Pairwise linkages among original and
inverted markers were estimated using
Mapmaker 3.0 (Lander et al., 1987). We
did not find a single linkage in repulsion

308

Research on Sweetpotato

phase (at LOO= 4 and a maximum


recombination fraction of 0.4), which
would be highly unlikely if sweetpotato
displayed strict disomic segregation. As
suggested by Wu et al. (1992), this excludes allopolyploidy, indicating a high
polysomy of sweetpotato chromosomes
within a base set.
A few repulsion linkages detected at
LOO = 3, r = 0.5 gave sorne evidence for
partially preferential pairing, which
supports a genome constitution
(B 1 B1 B2 B2 B2 B) with a certain degree of
homology among the two genomes, as
proposed by Shiotani and Kawase (1989).
Chromosome association in many autopolyploid species is likely to be a
combination of random and preferential
chromosome association (Wu et al., 1992).
These results suggest that polysomic
inheritance plays a major role in sweetpotato, with the occurrence of partially
preferential pairing among certain
chromosomes.
Framework map constructed with simplex
markers
Two linkage maps were constructed, based
on 539 simplex markers from maternal
sources and 421 from paternal sources,
using standard JoinMap 2.0 procedures
(Stam and van Ooijen, 1995). To avoid
false linkage detection, a LOO score of 5
was set as the linkage threshold for
grouping markers. This reduces the rate of
false positives to 1o-s, reducing the
number of false linkages to less than 1 in
100,000. Markers that could not be put in
the map during 'round one' or 'round two'
of the JoinMap mapping procedure were
omitted. A total of 44 markers for the
female parent and 50 for the male parent
remained unlinked. There were 90 linkage
groups for Tanzania, with 48 major and 42
minor groups of three or two markers; for
Bikilamaliya, this was 80 linkage groups,
with 42 major and 38 minor groups. The
resulting linkage groups provided a
framework into which higher-dose markers

were mapped (far example, see Figure 3).


Grouping was consistent through a range
of LOO seores of from 4 to 6, suggesting
that the groups farmed were highly
reliable. Linkage groups have been named
randomly until they can be aligned to
chromosome karyotypes.

Mapping of multiplex markers


The recombination fraction (r) and LOO
seores far simplex/duplex and duplex/
duplex marker configurations were calculated assuming (a) hexasomic and (b)
tetrasomic chromosome inheritance. The
absolute differences of pairwise estimates
(of r and LOO seores) under the two
inheritance types were low, with no more
than seven duplex markers placed on
different sites on the framework map.
These were therefare removed. Ouplex
markers were assigned to one or two
chromosomes, establishing a base-group
connection within chromosomes.
The final maps comprised 678 markers far
Tanzania, with a total map length of
3655.6 cM, and 481 markers far
Bikilamaliya, covering a total map length
of 3011 .5 cM. The maternal map consisted
of 90 linkage groups (71 groups with faur
or more markers), compared to the paternal map of 80 linkage groups (56 with faur
or more markers). Of 160 mapped male
and female duplex markers, a total of 49
were aligned to two linkage groups, which
al lowed us to order 60 chromosomes to 13
female and 1 O male homologous cosegregating groups. The maternal and
paternal maps contained 70 and 31
double-simplex (bridging) markers, respectively, which allowed identification of 17
linkage groups as maternal and paternal
homologous groups.
The largest female linkage group was
129 cM and the largest male group was
99.5 cM, with average spacing between
marker loci of 5.9 cM with only faur
intervals exceeding 30 cM. This should
provide reasonable genome coverage for
the detection of quantitative trait loci

(QTL). The percentages of mapped simplex, duplex, and triplex markers were
90%, 54%, and 30%, respectively.

Genome coverage

To avoid overestimating genome coverage, only framework markers were used in


this procedure. The expected proportion of
genome coverage was estimated using 656
markers far the female map and 469 far
the male. Under the assumption of random
marker distribution, the probability of
being within 20 cM of any existing marker
will be 91 % far any new female marker
and 88% far any new male marker (Bishop
et al., 1983).
The derived linkage maps have not yet
been resolved into the expected 90
linkage and 15 homologous groups. This is
because the total length of the sweetpotato genome was not covered, and the
small linkage groups, in particular, might
represent unconnected parts of other
linkage groups. Therefare, additional
markers would be required to cover the
entire genome and facilitate the merging
of ali homologous groups. Codominant
markers are also needed to confirm the
homology relations of chromosomes.

Acknowledgements
We wish to thank G. Rossel,
D. Carbajulca, and O. Hurtado far their
assistance in lab work. We also thank
J. Schmidt and M. Ghislain far valuable
discussion of the manuscript.

References
Alonso-Blanco, C., A.J.M. Peters,
M. Koorneef, C. Lister, C. Dean, N. van
den Bosh, J. Pot, and M.T.R. Kuiper.
1998. Oevelopment of an AFLP based
linkage map of Ler, Col and Cvi
Arabidopsis thaliana ecotypes and
construction of a Ler/Cvi recombinant
inbred line population. Plant Journal

14:259-271.
Bishop, O.T., C. Cannings, M. Skolnick,
and J.A. Williamson. 1983. The number

CIP Program Report 1999 - 2000

309

T01

o.o

e#mJlJ

e42m35.q
41m4k
e#m.k
e4Jm+1.o
e35m37_07
e:l9m40.f

28.8
JJ.l
34.9

41.7
43.0
+8.4
58.9
58.9
64.2
66.5
67.5
69.6
llO
lJ.6

T03(848)

r~

o.o

4.6
9.1
12.6
H.6
17.8
20.1
21.1
27.1

e44m45.d

2.l.1

T02
eJ2m36.g
e41m3'1.h
el2m49.p
eJ2m3.l.g
e.JJmJB=a
o.J9m41=c
e47m#=c
e45m41.d

43.6

71.7
75.6

e41mH.d
e3Bm38.1

85.4
85.4

a39m40.k
e46ml3_01

TOS

o.a

o./Jm54=c

o.o

e-Jm37.o

10.1

e4Jm45.n

11.ll

e4Sm4J.f

17.6

o./Jm45=m
23.J
26.J

e39m.b

eJJmJ9j

2J.6

J7.Z
-11.0

~
e+3m+8.c

38.8

e:l9m+5.f
e45mH=i
e4.lrn3!1.L
e:l9m40.e
eJ8rn3!1.d
e41m+4.i
eJ9m43"''
eJJmJ9=j
e+Jm48.k

ao.o

7J
8.5
10.1
11.7
12.7
15.0
24.0
25.0

TOS

T04

e4Jm45m
e4Jm54-c
e4Jm40.f
o45m44i
e4Jm39.a
e3Bm38.z
e#mJ2.c
e45mB.g
e46m40.k

eJ!lm'fJ.i
e:J9mJ2.b
e32m4S.h

HJ

o.o

T08(835)

a.a

o.o

ornJl..g

8.6
9.3
14.1

e39m40.j
e40m34.e
e39m40.L

20.4
2U
25.4
29.4

e41m62=c
e42mJS.h

04-51114-l=f
e39mJ4.b
eJJmJ9m

4.0
9.J
9.3
11i0

e45m41=f

27.5

e45m48.a

e46m.J8.b
e4Sm36.k
'3211161.o

J6.0

eJJmJ9=m

t32m62.

4-44

e+3m50_02

5+.4

e33m40_11

73.7
7J.7

e44m36.d
e41mJJ=g

e32m62.d
e43m54.i

~
o41mJJ=g

29.6

35.2
JIU
41.Q
+21!
+2.H
+5.3

em42.
e-4imJ7.d

34.8

e39m42.j

73.8

o32m60.j

Mr<
U
5.6

e40m62.e
e42m4!Pd

22.1

e41m2=11

H-f

eJ!JmJ2<C

eJ9m.J2=c

J16

+6.8

57.0

T09

o33m+8.c
e42m49=d

50.J
52.0
55.4
58.2

e44m34 07
e44m.Jize
e41m4J.e
e43m40.n
e41mJ2_04

65.9

e43m48 OJ

e45m44=i

e41m:33'"'~

@43m54:C 1

11

Homobgous
group(TJl

T12

Tl 1

r~

oJ8mJ8=s
T~,.
e+Jm50.L

oJ8mJ8"S

5.1
9.7

eJ2rn+9.g
eJ9m49=1.

29.B

e+lrn+a.b

e38m38S
Homobgous
group(T] N

Homobgous
group(TJlll

e45m41=f

Homobgous
group(T)ll

e3!1m43=i o33m3IFi

T10

e39m32=c
e42m49=d

e33m39=m

e43m45:1:m

15.8
15-8

T07{832)

e41m62=a

e41 m44zc

2.4

e4Sm42.a

T14

T13

o.o

e#mJ4.c

14.5
14.5
14.5

e41m62.c
e41m62.d
e41m62.b

T15

Tm. 'l~" T~
e41m42=d

lJ

a.s

eJJmJSj

13.+
14.2
16.7

et2mJ5.f
a32m62.h
et1m+8.d

23.6

e46mJ8.e

Jt.2

e#m44.d

29.6
30-7

eJ9m42'"9
eJ9m41.l

41.J
45.6

eJ5rnJ4.k
e43m54_10

53.0

e43m45.h

59.J

e41m42"'1
e41m42=d

J-2
5-9
8.0
12.1
13.5
21J
22.1
2.l.9

e.35mJ6=
e4Sm38.a
e3Bm3!1.a
"3Jm40h
e+1m48.h
e4Sm42.a
e.35mJ2=f
e45m42.

6.1

e32m36.a

17.4
19.6
20. 4
20.9

e44mJ1.L
e-15m39.i
eJ5mJ2=f
e.J3m40.c
e41rnJl.c
e+om42.c
e46m38.o

'11..7

25.0

2B.7

T16

T17

O.O
4.B
5.8
7.8

e43m41.c
e38m39.m
eJBmJ2_J1
eJJm40=h

e35m32f

T18

l~ T6.9

e4Jm48.o

225

26.2

e45mJ8.h
el9m43.k

JJ.7
37.6

eIDnJ4.k
e45mJ7.e

e33m40=h

4.2

e+2m4S.e

16.7
16.7
21.7

e4&n+o.i

sJ5mJ2c

e.39m49_02

e35m32:z:c

Homobgous
group(T)VI

Homologous
group(T]Vll

Homobgous
group(T)V

O.O

4.6

eJBmJB=h
eJ2mJ6=/J

"r~

11.1
22.5

36.B

52.J

T21

120

T19

e+2m35.p

e44m#.h

eJ2m49.n

29.6
29.6
29.6

eJBmJIJ.h
e32mJIFh
e44m48=b

29.6

e39m36.g

4-0.9

e33m+1.e

'f"

27.4

ei4m54J

J4.9
366
40.8

eJJm48=a
eHml6.c
o44mU-o

eJ5m3J.a

o.o
12.0

33.9

40.5
42.1
44.0
45.6
46.3
47.0
51.0
51.9
55.J

59.8
70.4

78.0

e+Jm41.a
e+1m#J
e38m35=h
e32m36=h

Homobgous
group(T]Vlll
115.7

68.2

eJ.2m33.o

a+.J

e45m3B.e

722.4

e4Jm54=j

129.6

e41m32.h

T23(B 01)

T22

76.6
81.5

o.o

eJ9m42.e

11.1

eJ9mJ6j
e4Jrn+S.c
om36f
e#m48--t:

22.J

eJ5mJJ-g

#.1

eJ9mJl=o

2.8
6.2

eIDn41.d
e42mJ5o
e39m50.m
e+Sm38_04
e.l8m38.o
e44ffl#.
e39m34.o
e4Jm59.n
45m51i
e32m61.g
e+5m+4.h
a39m50.n
e32m49.d
e39m49.j

72-2
89.5
89.5
8!1.5

89.5
e33m48:a

e44m44
Homobgous
group(T] IX

89.5
90.6
90.6
92.1
94.0

109.5
111.9

T25

T24

oJJm48--1J

e.J9mJ9.d
eJ9m32 07
~
e38m38.r
~

oJ9mJ6=k

o.o

e45m42.d

6.J
10.9

e1Jm59=h
eJJmJ9=g

5.8
8.5

oJ9mJ1.c
oJ9m41.h

19.8
24.8

eJ9mJ9.b
e43m40.e

37.6
41.0
4J.6

e+om42.d
e45m44.b
eJJmJB.f

57_3

e:l9m50.e
eJ4m49.h

4J.9

e4Jm40=g

53.5

e+om37.d

606

e1Jm41..,

66.7

eJ9mJ2.I

62-4

73.5
75.J
77.2
80.J
87.5
89.8
91.4

84.4

e+4m40_01

89.B

oJ5mJ8.h

95.2
e45rn+4 09
e.J8mJ8~u 101.6
eJJmJ9-g
e41rn+8j

e+Sm+J.k

91.4
91.4

e45m+J_01

9J.5
95.1

ellml.k
a+5mlJ.t

l"-

1.2
6.6
10.0
11.6
14.7
17.2
18.6
22.1
29.7
34.0
35.0

e43m4B.j
eJ9m31=a
e43mJ9.j
o45m45=i
e#mfD.d
eJ9mJ6.L
eJ5mJJ=g
eJ9mJ6.n
eJ9mJ6.o
oJ9m36=j
eJ9m36.i

e39m3E~k

e4Jm40=g
e45m45=i
e4Jm54.e
J9mJS.k
e#mSt.a
e45m41_08
e39mJ6.f
e3Bm32_12
aJ9mJl=a
eJJmiO.f
em42.n

T27

o.o

eHm-fB=c

t.1
6.2

.++mm

12.9

e44mJl=c

J+.7

eJ9m32.e

e4Jm48.e

51.9

e35m34.o

63.6
65.8
70.2
7+.9
76.5
77.8
B0.6
81.+
81.+
82.4
84.6
89.2
92.8

eiJm48.p
e+lm.llo
e4Jm.j(J:g
e+om34.b
41m.1Jk
e43m41.k
eJ.2m60.k
e34m49.f
eJm41.f
a+1m45.c
em40.f
e+6m38.i
e45m37.o

e41m33=k

eJJm60.g
a45m45::i e39m31=1

e39m3fi.,. e35m33=g

1=oiti3Srrse=

124.7

T26

o.o

ellm39.a

e44m48=c

Homobgous
group(T] X

Figures 3. Maternal ('Tanzania') linkage rnaps Linkage groups are narned at their top; a second narne in
parenthesis indicates the corresponding linkage group on the paternal rnap. Each marker locus is identified by

31 O

Research on Sweetpotato

T29

T28

o.o

e45m!lJ.d

14.4
16.9

o.l9m42.k
e44m40.a

24.6
27.1

e41m34.g
e.l3m41.f

36.8

e39rn50--d

e45m41.o

6.9

e4Jm45=d

6.9
10.8
13.1

34.2

~-'

e4SmH=e
e39n\l2.k

o.o

2.J

~~~t-i.

5.5
8.1

e44mJ2.k

T33

T32

T31

T30

O.O

elli40_00

o.o

O.O

e39m1J.g

11.5
l5.6

e44mJJ"'}
e4Jm59.p

o.o

T34{B26)
eJ2mJ6-t

e"31M8.d
em54.g

:~~~b

e45m51.e

25.3

e.35m34.c

43.6

eJ&nJ9.s

'47.5
50.B

eJ9mJ6.h

e45m<IJ.e

Jl.9

e40m62.d

43.D
44.9

e43m46.m
e41m42_01

eJ2mJ6f

eHm50=d

23.7
26.5

e45m38.d

28.7

e4JmJ9.i
e40m34.m

e43m54-.a

32.2
40.4
43.J
H.9
46.5

24.8

eJ9m43.h

a34m36.b

e45m12
e45m45_16

~:g
57.4
56.4
56.4
61.8

66.4
e45m44=e

e45m48.d
eJJm60.b
eJ6m38.l
e45m36.n

4.6
5.9
8.0
11.0
16.9

eJ9mJ9.c
e43m54.f
e44m4ll.g
e43m4ll.q
eJ5mJ1.i
e4Jm4ll.g

28.7

eJ5mJ2.b

~1.3

e44m40.k

~7.7

51.0

e40m62.b
e44m12.c

60.0

e43m50.h

o.o

e44m39.a

7.0

eJ2m61.d

H.7

eJ9mJ2.g

o.o
9.9

e44m42.f
eJ8m39.v

e38mJ8.m
e44mJ1 01

21.0
28.0
.l0.5
J1.4
36.0

o.o
12.4

21.2

"42mJ5.i
e32m62.b

50.1

e46m40.c

55.9

e32m49.a

~~~~17~n

45.1

e44m37.f

e36m38.x

53.J

eJ5m32.o

51.4

.JBm.l6.c

59.2

e44m+2.d

liO.O

.J3m40.j

69.8

e45m42.b

67.4
71.7

e41m31.o
e45m43.i

95.7

eJ9~=g

95.7

eJ2rn62.c

00

T4B

5.0

e45m.l9.m

11.J

eJ2m49.k

10.9

eJJm40.g

O.O

7.5
O.O
11.4
12.0
13.7
15.J
16.4
17.8

~~

26.2
27.6

e40m41.a
e41m45.b
e46m40 01

e:iam2i
e4-0m6lg
e44m40.c
e41m45.a
eJJmJ9.n

:~~R:~

2aa

eJ9m32.d

35.2
J9.6

e41m45.d

J1.B

~u

40.2

+1.9

49.0
49.0

eJ9mlls

42.9

eJ5mJB=f
e45m36.L

40.9

91
11.4

e32m61.c

35m38.d

20.8

e4Jm40.m

44.7
44.7
47.8
53.7

e39m.l9.a
e41m4lig
e41m45_01
o44m32.g
e40m37_03
e43m40.h
e43m40..i
t41m-14"'1
e35m36.b
eJ8m38.g
e45m41.b
e32m62.f
e45m42.l

~!~

27.7

e33m41.b
eJ2m61.e

38.3

eJ9m31.e

J7.9
J7.9

e40m34.j
e40mJ4J

458

tHm36.j

55.2
57.1
58.5

e41m3J.h
e46m38.d
e44m44.j

eHm50=c

eJ6m39.e

54.8

68.5
69.0

eJ9mJ4j
m41.g
em41 04

69.0

0'42m4i=h

65.1

55.7
55.7

~1
511.9
61.0
63.2
69.6

o.o
6.6
10.7

e45m42.j
eJ5mJ7 03
e42m49~b
e32m49.i

1a6
21.0

e40m42 01
e4Jm45J

27.8
32.1
35.8

e40mJ2.f
e45m45_00
eJJm41d
"45mJIJ.n
e+6m40.h
e44m40.l
o.J9m36=p
eJ9m42.d

36.J
+J.O
45.5

48.2
49.8

O.O

e+3m54_03
e45m36.g

101

32.6

eHm54.d

Jl.5

0'44m32n

4.1

e4Jm50.c

631

o.o

elli39.f

14.0
16.1
23.9
25.6
28.2
JO.O
Jl.8
35_9

em34 03

T45{B06)

T44
O.O

e39m36.m

7.5

e45m4JJl9

8.5

e43m40_01

~16.7
em54.c
e45m41.i
e41m45.j
elli60t

e45m42.i

em42.g
e43m45.j

44.J
49.1
51.6

elli.l6.e
e41m34.b

58.2

e4Jm5ol.o

76.1

e40m42.a

o.o
9.2

eJ9m41.j

19.6

e45m44.g

41.7

eJ3m40.m

53.7

eJ9m41.g

eJ5mJ4.h

4.4

eJ5m37_01

9.1

eJ4m49_()8

18.9

e45m43.d

20.5~

.lH

~~ti~

Jl. 6

e4 Im42=g

T52

T51

T50{815)

O.O

6.6

e44m34 q
e43m41.g

~J ~1~:~

29.6

om42.o

JS.2

e44m34.h

T54

T53

O.O
5.+
7.1

e41m32.c
eJ9m34.d
eJ5m38 03

15.0
19.3
19.3
2M
23.1
25.0
3011

eJ8m32_06
eJ8mJ8 02
43m401
e4-0m34.f
e45m44.o
e40mJ2.j
45m4J.b

o.a

J9m49.d

~~5~~

16.8

e44m32.m

23.0
25.5
28.J

e44m42 05
eJ9m41-05
e45m+fj

~M~tik

1U

e45mJ9.b

O.O

e4Jm59.m

0.0

e45mJ7.f

eJ2m3J.j

14.J

e43m48_~

e44mJ7.b
e39m40_03

4o.8

eJ9m50.L

5.1

e45m42.c

eJ3m40_07

O.Ole32mJ6.j
11.9

15.B

emJ8.c

11

27.9
31.2

o4Jm59.1

39.8

e1Jm59.o

eJ2m49=c

J7.4

e44m50.b

28.0

e44m33.o

33.4

e4Jm54_11

O.O

e4Dm32.e

O.O

9.3

e41m31.d

9.7

e32m60.c

15.6

e40m41.e

21.J

em4Z.j

17.D

29.5

32.6

eJSm.Jab

T63

T62

T61

T60(805)

T59

TSB

T57

T56
e43m45.i

O.Ole32m33.h

J4.0

e39m45.d
eJ9m40=d
e4"'148.f
42mJ5_Q1

e-l:im41_12
e-l:im11_13

T55

14.1
20.4
21.5

e39mJ2.f

8.9

l !~ 1 l I

T49(803)
e44m34.k

O.O

16.1
22.6
25.3

e35mJ+.g

T-47(823)

31.4

eJJm40.e

46.1

e41m43.o

eJ9m49.e

e4:im44.c

e44m42.L

37.2

e45m45.g

2.J.6

e45m42.f

iiJ9rlli

O.O

O.O

eJ5m34.n

e44m11.e

H3(809)

T41

T40

T39

:~~~~:~

T36

T3S

a.o

Homologous
group(l)Xlll

J_J

66.1
68.9
70.5
75.1

87.7

e32m36:f

Homologous
group(l]Xll

T38(836)

T37
o.o

11.J
12.2
17.0
22.2

e45mJ7.d
e40mJ7.b

e43m39~b

Homologous
group(l)XI

6.0

e32m49~

e43m40.j

56.3

eHm34.o
e39m49.o
eJ4m36.e
eJ8m32.b
eJ&n38 07

7.8
10.4

eJ~.b

6.7
13.6
16.6
18.1

o.o

e45m50.b

e45m48.h

44m51.b

19.4

e43m59.e

2li.O

eHmf2.h

O.O~e45m38.i

5.6
10.0

e-Hm38.i
e33mJ9.h

23.9

e43m50.j

e32m49.m

eJ9mJ1.g

the corresponding combination of primers used for its generation. A letter ora number at the end of each marker
identifies the individual fragments amplified by the same primer combination.

CIP Program Report 1999- 2000

311

T64(B74)

T65(847)

MI~

r~

15.4
21.3

e4..Jm-IO.c

4.1

c+1m40c

H.9

eJSmJ6.n

. o32m3J.b

31.B

e35m37.b

eJ2mJ6.e

e4Sm45.f

14.2

0'44m41=b

O.O

7.J

c41m42.a

221

e45m43.a

T-

21.9

em+J.d

T69

eJ5mJJ-g

t3Bm38.y

21.3

Mr. l""""

e3Bm39.b

T72

T71

T70

T~,

7.0

H.1

JI

T77

T76

e39m42.a

e.l.lm39.l

T84

T83
eJ2m62.g

T68

Mr~
15.1

T75

T"""

17.1

T82

o.o 40-nJtd

e4-\m3lj
e33m40j

24.7

T74

T~

3.4
H

T~

9.5
1H

T67

9.1

e3&n39.u

19.5

e45m48.e

10.9
H.2
14.2
15.7
17.6

e39m4-l.g
e32mJ6.d
e39m42_05
e45m41 01
e40m.J2~h

e33m39.d
T.~

5.2
12.4
13.0
149

e45m50J
eJBm39JO
e45m.J6.j

c+4m31n

.J0.5

T73

17.J

T66(B22)

13.6

ctlm39.d

c+4m41.d

7.0
em-40.i

T85

o40m32.c

o.o Jie40m32.d

O.O lr oJJmJ9.o

e41mJJ.b

6.1

5.9 Jl...e4Jm54d

e.lam38.j

T78
O.O
J.2

13.1

e41rn39.e

11.9

T86

O.O .lamJ6.j
5.2
oJBmJG
!O.O
e.lamJ6.f

T87

T88

O.O ')( em32.j

2.9)\eHmJN
+.7

T80

T79
e40m34 09
e4Jm59_02
o#m32.f
em54.h

e4-lm41.j

8.4

T81
e35mJ7.c

O.O
4.7

e4SmJ6.i

BJ

T89

e3Zm62.o
e43m5U
e44mJ6_05

T90

n1E~~~:[
e41m42J

3.3

Figure 3. (continued)
of polymorphic clones required to map
the human genome. In: Weir, B.S. (ed.).
Statistical analysis of DNA sequence
data, Marcel Dekker, NY, USA.
p. 181-200.
Da Silva, J.A.G. 1993. A methodology for
genome mapping of autopolyploids and
its application to sugarcane (Saccharum
spp.). PhD thesis. Cornell University,
lthaca, NY, USA.
Da Silva, J.A.G. and M.E. Sorells. 1996.
Linkage analysis in polyploids using
molecular markers. In: Jauhar, P. (ed.).
Methods of genome analysis in plants:
Their merits and pitfalls. CRC Press,
Boca Raton, FL, USA.
Doyle J.J. and J.L. Doyle. 1990. lsolation
of plant DNA from fresh tissue. Focus
12:13-15.
Grattapaglia, D. and R. Sederoff. 1994.
Genetic linkage maps of Eucalyptus
grandis and Eucalyptus urophylla
using a pseudo-testcross: Mapping
strategy and RAPO markers. Genetics
137:1121-1137.
Jones, A. 1965. Cytological observations
and ferti 1ity meas u rements of
sweetpotato [/pomoea batatas (L.) Lam].

312

Research on Sweetpotato

Proceedings of the American Society of


Horticultura! Science 86:527-537.
Lander, E.S., P. Green, J. Abrahamson,
A. Barlow, M. Daly, S. Lincoln, and
L. Newburg. 1987. MAPMAKER: An
interactive computer package for
constructing primary genetic linkage
maps of experimental and natural
populations. Genomics 1 :174-181.
Lee, M. 1995. DNA markers and plant
breeding programs. Advances in
Agronomy 55:265-344.
Magoon, M.L., R. Krishnar, and K. Vijaya
Bai. 1970. Cytological evidence on the
origin of sweetpotato. Theoretical
Applied Genetics 40:360-366.
Meyer, R.C., D. Milbourne, C.A. Hackett,
J.E. Bradshaw, J.W. McNichol, and
R. Waugh. 1998. Linkage analysis in
tetraploid patato and association of
markers with quantitative resistance to
late blight (Phytophthora infestans).
Molecular and General Genetics
259:150-160.
Nishiyama, l., T. Niyazaki, and
S. Sakamoto. 1975. Evolutionary
autopolyploidy in sweet patato and its
progenitors. Euphytica 24:197-208.

Ripol, M.I., G.A. Churchill, J.A.G. Da


Silva, and M. Sorrells. 1999. Statistical
aspects of genetic mapping in
autopolyploids. Gene 235:31-41.
Shiotani, l. 1988. Genomic structure and
the gene flow in sweet pota to and
related species. In: Exploration,
maintenance and utilization of sweet
patato genetic resources. Report of the
first sweet patato planning conference,
held in 1987. lnternational Patato
Center, Lima, Peru. p. 61-73.
Shiotani, l. and T. Kawase. 1989. Genomic
structure of sweet patato and hexaploids
in lpomoea trfida (H.B.K.) Don. Japan
Journal of Breeding 39:57-66.
Sorrells, M.E. 1992. Development and
application of RFLPs in polyploids. Crop
Science 32:1086-1091.
Stam, P. and J.W. van Ooijen. 1995.
JoinMap version 2.0: Software for the
calculation of genetic linkage maps.
CPRO-DLO, Wageningen, Netherlands.
Ukoskit, K. and P.G. Thompson. 1997.
Autopolyploidy versus allopolyoloidy
and low-density randomly amplified
polymorphic DNA linkage maps of
sweetpotato. Journal of the American
Society of Horticultura! Science
122(6):822-828.

Van Ooijen, J.W. 1994. DrawMap: A


computer program for drawing genetic
linkage maps. The Journal of Heredity
85:66.
Vos, P., R. Hogers, M. Bleeker, M. Reijans,
T. Van de Lee, M. Homes, A. Frijters,
J. Pot, J. Pelman, M. Kuiper, and
M. Zabeau. 1995. AFLP: A new
technique for DNA fingerprinting.
Nucleic Acids Research 23:4407-4414.
Wu, K.K., W. Burnquist, M.E. Sorrells,
T.L. Tew, P.H. Moore, and
S.D. Tanksley. 1992. The detection and
estimation of linkage in polyploids
using single-dose restriction fragments.
Theoretical Applied Genetics
83:294-300.
Yu, K.F. and K.P. Pauls. 1993. Segregation
of random amplified polymorphic DNA
markers and strategies for molecular
mapping in tetraploid alfalfa. Genome
36:844-851.
Zabeau, M. and P. Vos. 1993. Selective
restriction fragment amplification: A
general method for DNA fingerprinting.
European Patent Application
92402629.7 (Publ no. O 534 858A1 ).

CIP Program Report 1999 - 2000

313

From Latin America to Oceania: The Historie


Dispersal of Sweetpotato Re-examined Using AFLP
G. Rosse11, A. Kriegner2, and D.P. Zhang1

Although originally domesticated in tropical America, the sweetpotato (lpomoea batatas (L.) Lam.) has a long cultivation history in Oceania. Whjle the
post-Columbus dispersal of sweetpotato to Asia and Oceania is well documented, the hypothesis that there was prehistoric transfer by Peruvian or
Polynesian voyagers from Peru to Oceania has long been a controversial
issue. The objective of this study was to assess the genetic diversity and
interrelationship of sweetpotato cultivars from Oceania and Latin America
and to test the hypothesis of human transfer of this crop to the Pacific lslands
in prehistoric time. Seventy-six sweetpotato cultivars from Peru-Ecuador,
Mexico, the Philippines and eight Oceania countries were analyzed using
amplified fragment length polymorphism (AFLP). Multidimensional scaling
(MDS) and analysis of molecular variance (AMOVA) revealed wide genetic
variation in the Oceania gene pool, greater than that of Peru-Ecuador. There
was a significant sweetpotato "gene flow" from Mexico to Oceania. In
contrast, there is little association between the Peru-Ecuador germplasm and
that of Oceania. These results suggest that Peru-Ecuador may not be the
source of the Oceania germplasm. Natural dispersal from Mesoamerica is an
alternative explanation to the 'Kumara hypothesis' for the origin of the
Oceania sweetpotato.
Sweetpotato (/pomoea batatas (L.) Lam.)
was originally domesticated in tropical
America (Austin, 1988; Yen, 1982). The
exact center of origin and domestication
of the sweetpotato has not been wel 1
defined, neither has the wild ancestor of
this species been found. Based on the
numerical analysis of key morphological
characters of sweetpotato and the wild
lpomoea species, Austin (1988) postulated
that sweetpotato originated in the region
between the Yucatn Peninsula of Mexico
and the Orinoco River in Venezuela,
within which the four major American
1

CIP, Lima, Peru.


Austrian Research Center Seibersdorf (ARCS), Seibersdorf,
Austria.

taxa of the batata group are distributed.


Recent reports using molecular markers to
assess diversity have found the highest
diversity in Central America, supporting
the hypothesis that Central America is the
primary center of diversity and most likely
the center of origin of sweetpotato (Huang
and Sun, 2000; Zhang et al., 2000).
The sweetpotato was one of the first plant
introductions from the Americas into
Europe from the voyage of Columbus in
1492. From Europe this crop was taken by
the Portuguese explorers of the sixteenth
century to Africa, India, Southeast Asia
and the East lndies. This is the so-cal led
'batata line' of dispersa!. The camote line
(from the Nahuatl word camot/} represents

GIP Program Report 1999- 2000

315

the direct transfer of Mexican


sweetpotatoes by Spanish trading galleons
between Mexico and the Philippines in the
16th century. In the Asia and Pacific
region, there have been complex interchanges of sweetpotato after early
introductions. (Yen, 1982).
There is a third line of prehistoric transmission: the introduction of the sweetpotato to
the Pacific islands prior to the era of
European exploration (Yen, 1982). Fossil
carbon ized storage roots of sweetpotato
found in northern New Zealand have been
dated back sorne 1000 years (Yen, 1991 ),
which strongly supports the theory of
prehistoric transfer. Here, the question is
whether this was the chance spread of
seeds by natural means or human transfer
by Peruvian or Polynesian voyagers (Yen,
1982). The linguistic links between the
Quechua and Polynesian names for
sweetpotato, and their variations, support
the Peruvian origin and human transfer of
the Polynesian sweetpotato. The successful
drift voyages of the Kon-Tiki expedition in
1947 (Heyerdahl, 1950), further put this
sweetpotato issue within the context of the
wider problem of the origin of the
Polynesian people. Today, the 'Kumara'
area includes many sweetpotato-producing
countries in Oceana, including New
Zealand, Tonga, Samoa, and the Cook
lslands (Yen, 1974), where the sweetpotato
is a widely used staple food.
The sweetpotato is an ancient crop in Peru
(Yen, 1974; Austin, 1988). Examples
unearthed at numerous archaeological
sites representing both Inca and pre-Inca
cultures suggest that the sweetpotato was
one of the most important crops in this
region (Urgent and Peterson, 1988).
However, whether Peru was a major
source of sweetpotato to Oceana is still
unsettled. Our previous studies based on
molecular markers showed that Peruvian
sweetpotatoes are not closely related to
those from Papua New Guinea (Zhang et
al., 1 998) and are also different from those
of Mesoamerica (Zhang et al., 2000).
Understanding the level of genetic diver-

316

Research on Sweetpotato

sity and geographic differentiation is


critically important to germplasm conservation. Conservation strategies and
methods depend largely on this kind of
information from within a given gene pool.
In the present study, we assessed the
genetic diversity of germplasm from
Oceana and analyzed the relationship
between Oceania sweetpotatoes and those
of the Philippines, Mexico, and PeruEcuador. We show there is wide genetic
variation in the Oceania gene pool, wider
than that of Peru-Ecuador. We also show
that the germplasm from Peru-Ecuador is
not closely related to that of Asia and
Oceana, whereas the 'gene flow' from
Mexico to Asia and Oceana appears
significant. We suggest that the
Oceania sweetpotato probably carne
from Mesoamerica through non-human
dispersa l.

Materials and Methods


Plant materials
All plant materials were obtained from the
sweetpotato germplasm collection at CIP.
The 76 accessions were randomly selected
for each country. For this study, the
selection was made among land races or
farmer's cultivars (Table 1 ).
Healthy young leaves were collected from
accessions maintained in a screen house
and in vitro. The leaf tissue was immediately immersed in liquid nitrogen and then
transferred to -80C, freeze-dried, and
stored at room temperature until use.

DNA isolation and AFLP analysis


A modified DNA miniprep procedure,
based on Doyle and Doyle (1990), was
used to extract DNA. The amplified
fragment length polymorphism (AFLP)
protocol was from Vos et al. (1995).
Commercial AFLP kits were purchased
from Life Technologies (Gaithersburg, MD,
USA), and the AFLP reaction was carried
out according to the manufacturer's
instructions. DNA restriction digestion was
carried out using the enzyme combination

Table 1. CIP number and country of origin of 76 cultivars from Oceania, the Philippines, and Latin America.
CIP No.
Country
CIP No.
Country
CIP 441124
Solomon lslands
CIP 440399
New Caledonia
CIP441126
Solomon lslands
CIP 440294
Cook lslands
CIP 441125
Solomon lslands
CIP 440447
Cook lslands
CIP 441119
Solomon lslands
CIP 440454
Fiji
Fiji
CIP 441116
Solomon lslands
CIP 440456
Solomon lslands
CIP ARB 386
CIP 441120
Peru
ARB 234
Peru
CIP 441117
Solomon lslands
CIP 441123
Solomon lslands
DLP 1900
Peru
DLP 5314
Peru
CIP 441127
Solomon lslands
Peru
Solomon lslands
DLP1922
CIP 441118
Peru
DLP 3824
CIP 441221
Tonga
DLP 206
Peru
CIP 440276
Tonga
DLP 2
Tonga
CIP 441222
Peru
DLP 1090
Tonga
Peru
CIP 440273
Peru
CIP 440274
DLP 2344
Tonga
EECH 18
Peru
CIP 440272
Tonga
Tonga
DLP 1188
CIP 440277
Ecuador
Papua New Guinea
DLP 1161
Ecuador
CIP 440693
Ecuador
Papua New Guinea
DLP 1484
CIP 440129
DLP 1192
CIP 440706
Papua New Guinea
Ecuador
DLP 1449
Ecuador
CIP 440695
Papua New Guinea
DLP 1447
CIP 440696
Papua New Guinea
Ecuador
Papua New Guinea
DLP 1487
Ecuador
CIP 440699
Ecuador
DLP 1156
CIP 440297
Papua New Guinea
DLP 1456
Ecuador
CIP 440296
Papua New Guinea
Ecuador
Papua New Guinea
DLP 1403
CIP 440130
DLP 1423
Philippines
Ecuador
CIP 440660
NIAR 221
Mexico
Philippines
CIP 440657
Philippines
CTX 15
CIP 440290
Mexico
CIP 440669
CATIE 9360
Mexico
Philippines
Philippines
Mexico
CIP 440683
CATIE 9232
Philippines
Mexico
CIP 440667
CTX 29
RCB-IF-30
CIP 440671
Mexico
Philippines
Philippines
CATIE 9257
Mexico
CIP 440666
Mexico
Philippines
CTX 33
CIP 440659
Mexico
Philippines
CTX 31
CIP 440665
Mexico
Philippines
101438
CIP 440676
Mexico
Philippines
CTX 32
CIP 440681
of EcoRl/Msel. After the 1igation of the
ol igonucleotide adapters, the restrictionDNA fragments were amplified using a
polymerase chain reaction (PCR). Primer
annealing is targeted at the adapter and
restriction-site sequence. Three-nucleotide
extensions on both the EcoRI and Msel
primers cause selective ampl ification of
fragments. Primer combinations were
chosen based on the numbers of polymorphic fragments in a set of 1O sweetpotato
genotypes.

The 32 P-labelled PCR products were


separated by electrophoresis on a 6%
polyacrymide gel in a 1 x TBE buffer (TrisBoric acid-EDTA) solution at 50 W for
about 1.5 hours. The gel was dried and
exposed to X-ray film overnight.

Gel scoring and data analysis


Different fragment.s produced with each
primer were treated as unit characters and
numbered sequentially on the X-ray film.
Genotypes were scored for the presence

CIP Program Report 1999- 2000

317

(1) or absence (O) of each fragment. On ly


those fragments with medium or high
intensity were counted. Fragments with
the same mobility on the gel, but with
different intensities, were not distinguished
from each other when cultivars were
compared. Monomorphic fragments were
not scored.
From these data, a matrix of Eucl idean
distances was calculated using
RAPDistance vl .03 (Armstrong et al.,
1995). The Euclidean distances between
the 76 accessions were then presented in a
two-dimensional scaling plot using the
SAS multidimensional scaling (MDS)
procedure (SAS, 1997) to recover the
relative positions of the accessions in the
plot. Based on the MDS pattern, the
procedure for analysis of molecular
variance (AMOVA) (Excoffier et al., 1992)
was applied to quantify the variance
components for AFLP phenotypes. Individual variation was partitioned within and
between regions. The components of
interest were extracted and tested using
nonparametric permutation procedures.

70
60
50
40
30

Results
Polymorphic bands and multidimensional
scaling
The 1O primer combinations generated 21 O
polymorphic, clearly scorable fragments
for the 76 cultivars. The number of polymorphic bands for ali the Oceania
countries were comparable to those found
in the Mexican accessions and higher than
those of Peru-Ecuador, except for Papua
New Guinea. Moreover, there were 18
bands (9% of the total scored bands) that
were present only in the Oceania and
Mexican accessions.
The MDS plot (Figure 1) showed a weak
relationship between accession and
geographic origin, except for Peru-Ecuador, where the cu ltivars were clearly
distinguishable from the rest. Ali of the 22

*
!
*

* *. .,.,:.* * !

! !
*

!
*

20
10

o
-10
-20
-30
-40
-50

Oceania

-60
-70
- 70

Variation between regions was then


partitioned into pair-wise distances between regions to examine the regional
contribution to total molecular diversity
(Excoffier et al., 1994).

- 60

- 50

- 40

- 30

- 20

- 10

Represents cultivars from Oceania


Represents cultivars from Peru -Ecuador

10

20

30

40

50

60

70

Represents cultivars from Mexico

Represents cultivars from The Philippine

Figure 1. lnterrelationship of sweetpotato cultivars from America and Oceania using multi-dimensional scaling
(MDS). There is clear differentiation between Peru-Ecuador and the other countries. Mexican cultivars are closely
associated with those of Oceania.
318

Research on Sweetpotato

cultivars from Peru-Ecuador are closely


grouped in one side of the plot.

Analysis of molecular variance


The AMOVA was conducted to quantify
the diversity level and genetic relationship
among regions by partitioning the variation within and among regions. There are
various ways of grouping the four regions
(or countries), but the current analysis
assigned the four regions into two groups
(Peru-Ecuador vs. the rest, see Table 2),
based on the MDS plot.
The between-group variation accounts for
12.2% of the total molecular variance and
is highly significant (Excoffier et al.,
1994). Since the between-group variation
actually quantifies the difference between
Peru-Ecuador and the other regions (countries), it confirms the visual grouping
observed in the MDS: there is statistically
measurable divergence between the
sweetpotato of Peru-Ecuador and the other
regions (countries). The pair-wise distance
among regions also demonstrates that
Peru-Ecuador is significantly different from
the other countries (Table 3).
The within-group (between-region) variation accounts for only 1 .4% of the total
molecular variance and is not statistically
significant (Table 2). Since the within-

group variation measures the differences


between the regions (or countries) in
Oceana, the Philippines, and Mexico, the
non-significant result means that among
these countries there is no detectable
difference between the accessions of the 3
sources. Furthermore, the Mexican accessions are closely associated with the
accessions from Oceania and the Philippines. The pair-wise distances between
these regions (or countries) are clase to
zero, in sharp contrast to the long distances between these regions and
Peru-Ecuador (Table 3).
The largest source of diversity comes from
the within-region variation, which accounts for 86.4% of the total variance.
This large within-region variation is mostly
from the wide variation between individual cultivars in the same region (or
country). The samples from the Philippines
contain the highest interna! diversity
(37.0), comparable to that of Oceana
(36.0), whereas the interna! diversity in

Table 3. Genetic distances among sweetpotato


cultivars from tour regions.
Regions
Peru-Ecuador Mexico Philippines
Mexico
0.151 ***
0.021f'.S
Philippines 0.149 ***
0.043f'.S
o.oooIB
Oceania
0.188 ***
Notes: *** Significant at .001 leve!; ~ = Non-significant.

Table 2. Analysis of molecular variance far the extraction of components of AFLP variation among
regions, and among individuals within regions.
% Total 3
Source of variation
DF
MSD2
Variance
SSD1
component
Among groups (Peru-Ecuador vs. the other
201.2 201.2
4.97
12.2
regions)
Regions within group (among Mexico, the
42.0
0.547
2
84.1
1.4
Philippines, and Oceania)
lndividuals within regions (or countries)
72
2486.5
35.0
35.0
86.4
31.5
Peru-Ecuador
21
661.8
337.3
33.7
Mexico
10
407.1
37.0
Philippines
11
Oceania
276.9
36.0
30
Total
75

groups,
P value4

<.001
NS

<.001

1 Sum of squared deviations.


2 Mean squared deviations.

3 Percent

of total molecular variance.

4 Probability of obtaining a larger component estimate. Number of permutations = 1000.

CIP Program Report 1999 - 2000

319

Peru-Ecuador is the lowest (31 .5). The


partitioning of the variation into each
region (or country) confirms the visual
pattern in the MDS plot and supports the
conclusion that there is a higher level of
interna! diversity in Oceania than in
Peru-Ecuador.

Discussion
The high level of genetic diversity and the
low within-region geographic proximity in
Oceania agrees with the previous findings
of Jarret and Austin (1994), who found
higher diversity in accessions from
Oceania than in those from Peru, based on
RAPO analysis. The high diversity in Asia
and Oceania was also found by Huang
(2001 ). This high diversity and the low
geographic proximity within Oceania, as
well as between the Oceania countries
and the Philippines and Mexico, are
compatible with the history of the introduction and cultivation of the sweetpotato
in Oceania and Asia. While the original
spread of sweetpotato within Oceania was
largely due to the Maori, who carried it
with them on their voyages, there were
multiple introductions from different
sources after the initial transmission. For
example, there were at least three notable
introductions to New Zealand. There are
three Maori words used to describe
sweetpotatoes, which reflect the introduction of the sweetpotato to New Zealand:
kumara (which, as mentioned above, has
parallels to the Quechua word kuma),
merikana (meaning American), and waina
(meaning vine) (Yen, 1974). In the Philippines, there were introductions via the
trade route between Mexico and Manila in
the 17th century (O'Brien, 1972; Yen, 1982)
in addition to the introduction from China
by the end of the 16th century. Botanical
evidence also suggests introductions from
Indonesia and west New Guinea (Yen,
1991 ). These repeated introductions from
different sources have contributed to the
high diversity in Asia and Oceania, but
they tend to confound any regional
differentiation between these countries.

320

Research on Sweetpotato

As a crop of tropical American origin, the


means of pre-historie dispersa! of
sweetpotato to the west has long been a
controversia! issue. Besides the lack of
morphological si mi larities between
Peruvian sweetpotatoes and those of
Oceania (Yen, 1974), the 'Kumara'
hypothesis also suffers from the lack of
evidence that such human contact ever
occurred (Emory, 1968). The results of the
present study demonstrate that Oceania
sweetpotatoes have a weak genetic
association with those of Peru-Ecuador.
Moreover, the level of genetic diversity in
Peru-Ecuador is lower than that of the
Oceania countries and the Philippines,
which suggests that the Oceania and
Asian sweetpotato may not be of Peruvian
origin. A similar result was found by
Gichuki (2001 ). The possibility of alternative mechanisms, such as the transfer of
botanical seeds by birds, could explain the
pre-European transfer of sweetpotato into
Polynesia. The accidental introductions
explain the pre-historie arrival of the plant
and also account for the absence of other
cultural traits which might have been
expected to occur if human contact had
taken place (O'Brien, 1972). Coconut and
gourd are two examples of non-human
dispersa! of cultivated pantropic species
(Ridley, 1930). lf this applies to the
sweetpotato, the source of seeds could be
from a tropical American region where this
crop goes through a reproductive stage and
produces seeds. The close relationship
between the Mexican accessions and
those from Oceania suggest that the
sweetpotato in Oceania may be of
Mesoamerican origin.

References
Armstrong, J., A. Gibbs, R. Peakall, and
G. Weiller. 1995. RAPDistance,
Package Manual. Version 1.03.
Australian National University,
Can berra.
Austin, D.F. 1988. The taxonomy,
evolution and genetic diversity of
sweetpotatoes and related wild species.
In: Gregory, P. (ed.). Exploration,

maintenance, and utilization of


sweetpotato genetic resources.
lnternational Potato Center, Lima, Peru.

p. 27-60.
Doyle, J.J. and J.L. Doyle. 1990. lsolation
of plant DNA from fresh tissue. Focus

12:13-15.
Emory, K.P. 1968. Review of reports of the
Norwegian archaeological expedition to
Easter lsland and the east Pacific. Vol.
2: Miscellaneous Papers. American
Anthropologist 70:152-154.
Excoffier, L., P.E. Smouse, and J.M.
Quattro. 1992. Analysis of molecular
variance inferred from metric distances
among DNA haplotypes: Applications to
human mitochondrial DNA restriction
data. Genetics 131 :479-491.
Gichuki, S.T. 2001. Genetic diversity of
sweetpotato (lpomea batatas [L.] Lam.)
as assessed with RAPO markers in
relationship to geographic sources. PhD
thesis. University of Agricultura!
Sciences, Vienna, Austria.
Heyerdahl, T. 1950. The voyage of the raft
Kon-Tiki. Geographic Journal 115:20-41.
Huang, J.C. and M. Sun. 2000. Genetic
diversity and relationships of sweetpotato and its wild relatives in Jpomoea
series Batatas (Convolvulaceae) as
revealed by inter-simple sequence
repeat (ISSR) and restriction analysis of
chloroplast DNA. TAG 100:1050-1060.
Huang, J.C. 2001. Analysis of genetic
diversity in a sweetpotato (Jpomoea
batatas) germplasm collection from Asia
using amplified fragment length
polymorphism (AFLP). PhD thesis. Hong
Kong University, Hong Kong.
Huamn, Z. and O.P. Zhang. 1997.
Sweetpotato. In: D. Fuccillo et al.
(eds.). Biodiversity in Trust.
Conservation and Use of Plant Genetic
Resources in CGIAR Centres.
Cambridge University Press, Cambridge,
UK. p. 29-38.
Jarret, R.L. and D.F. Austin. 1994. Genetic
diversity and systematic relationships in
sweetpotato Upomoea batatas (L.) Lam.)
and related species as revealed by
RAPO analysis. Genetic Resources and
Crop Evolution 41 :165-173.

O'Brien, P.J. 1972. The sweet potato: lts


origin and dispersa!. American
Anthropologist 74:343-365.
Ridley, H.N. 1930. The dispersa! of plants
throughout the world. L. Reeve & Co.,
Ashford, Kent, UK.
SAS lnstitute. 1997. SAS/STAT software:
Changes and enhancements through
release 6.12. SAS lnst. lnc., Cary, NC,
USA.
Urgent, D. and L. Peterson. 1988.
Archeological remains of potato and
sweetpotato in Peru. CIP Circular

16(3):1-1 O.
Vos, P., R. Hogers, M. Bleeker, M. Reijans,
T. van der Lee, M. Homes, A. Frijters,
J. Pot, J. Peleman, M. Kuiper, and
M. Zabeau. 1995. AFLP: A new
technique for DNA fingerprinting.
Nucleic Acid Research 23:4407-4414.
Yen, O.E. 1974. The sweetpotato and
Oceania: An essay in ethnobotany.
Bernice P. Bishop Museum Bulletin 236.
Bishop Museum Press, Hawaii, USA.
Yen, O.E. 1982. Sweet potato in historical
perspective. In: Villareal, R.L. and
T.D. Griggs (eds.). Sweet potato,
Proceedings of First lnternational
Symposium. AVRDC Publ. No. 82-172,
Tainan, Taiwan. p. 17-30.
Yen, O.E. 1991. The social impact of
sweetpotato introduction in Asia and the
South Pacific. In: UPWARO sweetpotato
cultures of Asia and South Pacific.
Proceedings of the 2d Annual UPWARO
lntl. Conference. UPWARD, Los Baos,
Philippines. p. 18-27.
Zhang, O.P., M. Ghislain, Z. Huamn,
A. Golmirzaie, and R.J. Hijmans. 1998.
RAPO variation in sweetpotato
[lpomoea batatas (L.) Lam.] cultivars
from South America and Papua New
Guinea. Genetic Resources and Crop
Evolution 45:271-277.
Zhang, O.P., J. Cervantes, Z. Huamn,
E. Carey, and M. Ghislain. 2000.
Assessing genetic diversity of sweet
potato (lpomoea batatas (L.) Lam.)
cultivars from tropical America using

AFLP. Genetic Resources and Crop


Evolution 47:659-665.

CIP Program Report 1999 - 2000

321

Global Distribution of Sweetpotato


R.J. Hijmans1 , L. Huaccho1' 2 , and O.P. Zhang 1

A georeferenced database of the global distribution of sweetpotato (lpomoea


batatas (L.) Lam) is described. The database was assembled from many country-level sources. To create a database representative for one time period,
the proportion of national sweetpotato area in each production zone was
multiplied by total national sweetpotato area for 1998-2000, as estimated by
the Food and Agriculture Organization of the United Nations (FAO). Most
sweetpotato is grown in the temperate zone, 70% of the area is between
20N and 40N. Sweetpotato is highly concentrated in sorne areas, notably in
the lowlands of China and in the mid-elevations of the Lake Victoria area in
Africa. About half the global sweetpotato area occurs where it is an obligatory seasonal crop because of low temperatures during part of the year.

Crop distribution data are useful for various


purposes, particularly for targeting agricultura! research and assessing the impact of
agricultura! technology (e.g., Hijmans et
al., 2000). Country level crop distribution
data are readily available through the
database of FAO (2001 ). To study aspects
related to the distribution of a particular
crop at the global or lower aggregation
levels, however, more disaggregated data
are needed to account for the considerable
differences within countries in crop
distribution and other variables such as
climate.
In this paper, we describe a new global
georeferenced database of the distribution
of sweetpotato and use it to provide a
preliminary description of the current
distribution of the crop. This database is an
updated version of the one documented in
detail by Huaccho and Hijmans (2000).
The only previously existing map of global
sweetpotato distribution (Bertin et al.,
1971) focused more on historical aspects,
CIP, Lima, Peru.
2
Present address: St. Hilda's Col lege, Oxford Un iversity, Oxford,
1

UK

was outdated, and had insufficient detail


for our purposes. Comparable work for
potato (Solanum tuberosum L.) was
reported by Huaccho and Hijmans (1999)
and Hijmans (2001 ).

Materials and Methods


Georeferenced database
Sweetpotato production zones were
delineated for each country with
sweetpotato production. These zones
generally followed administrative boundaries, but in sorne cases we used zones on
existing national-level sweetpotato
distribution maps or on maps of general
crop distribution, including satellitederived land cover data (USGS-EDC,
1998). Most data sources used to estmate
within-country distribution of sweetpotato
are listed by country in Huaccho and
Hijmans (2000). We used ArcView-GIS
version 3.1 (ESRI, Redlands, CA, USA),
geographic information system software.
The fraction of the national sweetpotato
area in each production zone was estimated, using national statistical data when

GIP Program Report 1999 - 2000

323

available. This fraction was then multiplied by the FAO estmate of the average
total national sweetpotato area for 19982000 (FAO, 2001) to create a consistent
database representative for one ti me
period. We used national-level estimates
of total sweetpotato area for the countries
for which FAO did not provide an estmate, namely: Colombia, Costa Rica,
Guatemala, Guyana, and Nepal. For
Malawi we assumed that the 80% of the
total area reported as potato by FAO was
in fact sweetpotato (Peter Ewell, CIP,
1999, pers. comm.)
Data conversion and analysis

Relative sweetpotato area over total land


area (RSA) was calculated for each
production zone (polygon), using IDRISI
software (Clark Labs, Worcester, MA, USA)
to cale u late the total area of each polygon.
RSA data were transferred to a 1 x 1
minute grid, which was aggregated to a
coarser resolution grid (1 x 1 degree),
calculating the average RSA for each cell.
This two-step process prevented data loss
from small polygons while creating a
smoother grid. The 1 x 1 minute resolution
seemed an adequate compromise between
the desire for high resolution and the high
uncertainty of sorne of the sweetpotato
distribution data, which do not justify the
use of a higher resolution. By multiplying
the area of each grid cel 1 by the RSA, the
total sweetpotato area for a grid cel 1 was
cale u lated.
The sweetpotato area was summarized by
bands of 1 degree of latitude wide; and by
altitude (below and above 1,000 m), using
the ETOP05 database (USGS-EDC,
undated).
A global grid of temperature (30-minute
resolution) (New et al., 1999) was used to
describe temperature of sweetpotato
production areas. Months with an average
temperature below 1OC were defined as
cold months, in which sweetpotato is
unlikely to be in the field.

324

Research on Sweetpotato

Results
The sweetpotato distribution database has
a total of 1136 spatial units, with a mean
of 8456 and a median of 1569 ha per map
unit. Detailed results for first-level administrative subdivisions (e.g., state,
department) are tabulated in Huaccho and
Hijmans (2000). Figures 1 and 2 are
derived from the database.
The high concentration of sweetpotato
area in China, which has about 65% of the
world's sweetpotato area, is the most
striking aspect of the crop's distribution
(Figure 1). In addition to China, there are
considerable concentrations of
sweetpotato in Cuba and Haiti in the
Caribbean region; in Java (Indonesia), the
island of New Guinea (both in Indonesia
and in Papua New Guinea), and Vietnam
in Asia; and in Africa, particularly in the
Lake Victoria area (Burundi, Rwanda,
Uganda, and the Dem. Rep. of Congo),
and in Ghana, Nigeria, and Madagascar.
Sweetpotato is an important staple for
many Oceana island countries such as
Papua New Guinea, the Solomon lslands,
Tonga, and New Caledonia. They are not
shown on the map, however, because of
the 1-dot-equals-1 000-ha scale used.
Because of the high concentration of
sweetpotato area in China, we include a
separate map for that country (Figure 2). In
China, there are three areas of very high
concentration of sweetpotato: 1) Sichuan
and Chongq i ng provi nces, with about
1,000,000 ha (the Sichuan Basin); 2) the
east central provinces of Shandong,
Henan, and Anhui, each of which has over
600,000 ha; and 3) the southeastern coast.
Each of four Chinese provinces has more
sweetpotato area than Uganda, which,
with 546,000 ha, is the second largest
sweetpotato producing country.
There is a bimodal distribution of
sweetpotato area by latitude (Figure 3).
Seventy percent of sweetpotato is grown
between 20N and 40N. This peak
includes nearly all the area in China,

,,..,,.

..

//~~f~:i~' =-~:~~ ~, x:~~~~~~~. ~:;T:::_~t~-':~ . ,:~;-, ~tfi'~~ '\


~

"

..... .

Sweetpotato area (1998-2000)

Each dot represents 1000 ha

. '

\~

'.(3f

'i.,

~,

;;..

., '

'

C"l

3
~

"O

;::.

;D
<C
<C

1
N

o
o
o

1\.)
(.TI

Figure 1. Global sweetpotato distribution (1998-2000).

. .

. . .

f w.

'

'\J . N
('!
d

........
1

=o

+ '

r t-~
1

-o

,. /)''' /

,_:;/' /

'

o ..1;po

.. ,

222 km
................... ................................................. ..............

//
......................................................

~"'/,/,/

100

90

80

70

130

120

110

50

China, Sweetpotato area

50

Each dot represents 2500 ha

\.,,.,/;

...

(
{

..t

'\:~

;;:'t"\,_~A'...~

---..5

\V\

''\J\,f"'7

40

,_....-:: . .'1.,.J

,.. -..>''~ .......-

:.:,/

40

.~J ...._\ ____, ___ ..-................... ""'~ . . ,...."'.......-..--"''

<. >

..

;,,\ ~

; /'

........

:_ . . , ,< ,,..,.,(~',,s<\ /;'


'-..,

...);
\~-,~~
/

\,.\..

30

'

~\
..

,~ <~,

\""\
'""'~""-t ._,,J~'\ '-,, ... )
,... ..........J''\

30

L,'tj''\
lo./~

{, ~-.:
20
20

90

80

110

100

120

Figure 2. Sweetpotato distribution in China.


Area (1,000 ha)

700
600

500
400

300
200
100

-50

-3'.l

-20

10

-10

20

30

40

50

60

L.atitude()

Figure 3. Distribution of sweetpotato area by latitude.


Each dot represents one-degree latitude. Latitude in
the Southern Hemisphere is indicated with a minus
(-) sign. The line is the five-observations moving
average.

326

Research on Sweetpotato

India, and North Arnerica. The


sweetpotato area at these latitudes is
virtually all in lowlands (with only 6%
above 1,000 rn; Figure 4). The second
peak in sweetpotato distribution by
latitude is between 105 and 1 SN, with
24% of the global sweetpotato area. This
peak includes rnost of the sweetpotato
area in Africa, and sorne areas in Asia and
northeastern Brazil. A large part of the
sweetpotato area at these latitudes occurs
at rnid-elevation (49% above 1,000 rn;
Figure 4). There is not rnuch sweetpotato
area south of 1OS, except far sorne areas
on Madagascar. This is partly explained by
the paucity of land, people, and agricu 1ture in the Southern Hemisphere (Hijmans,
in press).

Sweetpotato is grown in areas where it can


be grown year-rou nd, and in a reas where it
can only be grown in the summer (warm)
season. Fifty-four percent of the
sweetpotato area has more than one cold
month (average temperature below 1OC;
Figure 5). Hence, we estimate that about
half the sweetpotato area is obligatory
seasonal because of low temperatures
during a part of the growing season. This
area roughly corresponds with the coldest
49% of the growing area, which has an
average annual temperature below 15C
(Figure 6). That is not to say that in other
areas sweetpotato production is necessari ly year-round. That depends on many
other factors outside the scope of this
paper, such as prevailing crop rotations
and the presence of a dry season. In
general, these two systems (seasonal vs
year-round) have quite different production
constraints. For example, fungal diseases
are often very important in seasonal
production areas, whereas insect pests are
most severe in areas with year-round
production. Whereas cold temperatures in
the beginning and the end of the growing
season are a limiting factor in seasonal
systems, rare events of mid-season night
frost in the tropical highlands of Papua
New Guinea can have a devastating effect
on production (Waddell, 1975).

Fraction

0.8

0.6
0.4

0.2

50

-40

-30

-20

-10

10

20

30

40

50

60

Latitude ()

Figure 4. Fraction of the sweetpotato area located


above 1,000 m, by latitude. Each dot represents onedegree latitude. Latitude in the Southern Hemisphere
is indicated with a minus (-) sign. The line is the fiveobservations moving average.
Area(%)

50
45
40
35
30
25
20
15
10

Nurrber of cold months

Figure 5. Global sweetpotato area by number of cold


months. A cold month has an average temperature
below 10C.
Area (%)

Conclusion and Discussion


The georeferenced database presented
here is the first detailed description of
global sweetpotato distribution. lt provides
considerably more detail than the rather
general map published 30 years ago by
Bertin et al. (1971 ).

45
40
35
30
25
20
15
10

0-5

5-1 o

10-15

15-20

20-25

25-30

Temperatura (C)

Sweetpotato distribution is characterized


by a pattern of both concentration and
dispersion. With 65% of global
sweetpotato area, China is in a league of
its own, and the area in China obviously
dominates the distribution by latitude.
There is also an area of high sweetpotato
density in the Lake Victoria area in Africa.
In most other countries, however,
sweetpotato densities are low, yet the

Figure 6. Global sweetpotato area by average annual


temperature.
crop is grown almost everywhere in the
subtropics.
In China, the provinces of Sichuan,
Henan, and Shandong have a very high
rural population density and irrigable
farmland is scarce. In these areas,
GIP Program Report 1999 - 2000

327

sweetpotato fits well in the intense


cropping system of the uplands. Under
certain circumstances, sweetpotato can
produce more edible energy per hectare
per day than any other majar crop (De
Vries et al., 1967). In eastern Africa, most
of the production is concentrated at midelevation (1,000-1,600 m) in the densely
populated Lake Victoria basin. Population
pressure is the main reason for the crop's
rising importance in this region, where it is
often planted in more marginal fields with
poor soils and limited water supply.
Sweetpotato appears to have a rather
complementary distribution as compared
to patato and cassava (Manihot esculenta
(L.) Crantz) the other two root and tuber
crops of global importance. Patato is a
summer crop in the temperate lowlands,
an off-season or mid-elevation crop of the
subtropics, and is also grown in the upper
parts of the tropical highlands (Hijmans,
2001 ). Sweetpotato is predominantly
grown in warmer areas: the subtropical
lowlands, and in mid-elevations of the
tropical highlands. In contrast, cassava is
mainly grown in the tropical lowlands.
China is the country with the world's
largest patato and sweetpotato area.
Nevertheless, most patato and sweetpotato
production occurs in different areas: patato
in the northern, northwestern and the
southwestern highlands; sweetpotato in the
eastern lowlands and in the Sichuan Basin.
In the central African highlands, particularly in Rwanda and Burundi, there are
high densities of cassava, sweetpotato,
and patato, albeit mostly at different
altitudes (cf. Hijmans, 2001; Carter et al.,
1992).
The data sources used for the global
sweetpotato distribution database differed
greatly in detail and quality. The distribution in the countries with low sweetpotato
production is, generally speaking, most
uncertain. Given the paucity of data for
many countries, we have probably missed

328

Research on Sweetpotato

sorne production zones. The fact that the


borders of countries are sometimes discernible from the sweetpotato distribution
illustrates (in sorne cases, but not in all)
weaknesses in the database.
Through a collaborative network of
sweetpotato scientists in East Africa, we
had access to good estimates for important
sweetpotato-producing countries in that
part of the world. We also had a highresolution (at county-level) database for
China for 1987 and 1988. The relative
spatial distribution_ of the data seems to be
quite accurate. However, the aggregated
area exceeded FAO estimates by 1.5
mi Ilion ha, representing about 17% of the
world's sweetpotato reported by FAO for
1987. Although the higher estmate might
a~tual ly be more accurate (Crook, 1993),
we have used the FAO figure for the
country aggregate, as we did for all other
countries, and the county-level data for
the within-country distribution.
Crop distribution maps are an important
missing link for studies of global agriculture. Whereas there are prospects for using
more remotely sensed data, it is currently
difficult to use these to identify the spatial
extent of global agriculture (Wood et al.,
2000), let alone that of specific crops. At
present, detai led census-type data are
indispensable for the development of
global crop distribution databases.
For this type of composite database,
documenting the sources used for each
country is very important, as this can
guide efforts to update data and help users
assess the data quality. The sources of a
previous version of the database discussed
here are documented in Huaccho and
Hijmans (2000). Having the data in a
digital format has the dual advantage of
increasing its usefulness for research of
global sweetpotato production, and
making it relatively easy to update when
newer or better data beco me avai lable.

Acknowledgments
We thank Terefe Belehu, Peter Ewell, Ali
Khalafalla, Berga Lemaga, Mlance
Ndikumasabo, Jean Ndirigue, Phemba
Phezo, J.M. Randrianaivoarivony, Tom
Walker, and Stanley Wood for providing
sweetpotato distribution data. Additionally,
we thank Tom Walker for reviewing this
paper.

References
Bertin, J., J.J. Hmardinquer, M. Keul, and
W.G.L. Randles. 1971. Atlas of food
crops. Geographical and chronological
survey for an atlas of world history.
cole Pratique des Hautes tudesSorbonne. Vle Section: Sciences
conomiques et Sociales. Centre de
Recherches Historiques et Laboratoire
de Cartographie, Paris, France.
Carter, S.E., L.O. Fresco, and P.G. Janes
with J.N. Fairbairn. 1992. An atlas of
cassava in A frica. H istorical,
agroecological and demographic
aspects of crop distribution. Centro
Internacional de Agricultura Tropical,
Cali, Colombia.
Crook, F.W. 1993. Under reporting of
China's cultivated area: lmplications for
world agricultura! trade. In: China,
international agriculture and trade
reports. Situation and outlook series, RS93-4. United States Department of
Agriculture, Economic Research
Service, Washington, D.C., USA.
De Vries, C.A., J.D. Ferwerda, and M.
Flach. 1967. Choice of food crops in
relation to actual and potential
production in the tropics. Netherlands
Journal of Agricu ltural Science 15: 241248.
FAO (Food and Agriculture Organization
of the United Nations). 2001. Database
at http://apps.fao.org.
Hijmans R.J. 2001. Global distribution of
the patato crop. American Journal of
Patato Research, vol. 78.

Hijmans, R.J., G.A. Forbes, and T.S.


Walker. 2000. Estimating the global
severity of patato late blight with GISlinked disease forecast models. Plant
Pathology 49: 697-705.
Huaccho, L. and R.J. Hijmans. 1999. A
global geo-referenced database of
patato production for 1995-1997,
GPOT97. Production Systems and
Natural Resource Management
Department Working Paper 1.
lnternational Patato Center, Lima, Peru.
Huaccho, L. and R.J. Hijmans. 2000. A
geo-referenced database of global
sweetpotato production. Production
Systems and Natural Resource
Management Working Paper 4.
lnternational Patato Center, Lima, Peru.
New, M., M. Hulme, and P. Janes. 1999.
Representing twentieth-century spacetime climate variability. Part 1:
Development of a 1961-1 990 mean
monthly terrestrial climatology. Journal
of Climate 12:829-856.
USGS-EDC (United States Geological
Survey. EROS Data Center). Without
date. 5-minute (ETOP05) global
elevation databases. Online at: http://
edcdaac.usgs.gov/.
USGS-EDC (United States Geological
Surveys EROS Data Center). 1998. 1 km
Global Land Cover Characterization
Database. Online at: http://
edcdaac.usgs.gov/glcc/glcc.html.
Waddell, E. 1975. How the Enga cope
with frost: Responses to climatic
perturbations in the Central Highlands
of New Guinea. Human Ecology 3(4):
249-273.
Wood, S., K. Sebastian, and S.J. Scherr.
2000. Pilot Analysis of Global
Ecosystems: Agroecosystems. A joint
study by lnternational Food Policy
Research lnstitute and World Resources
lnstitute. lnternational Food Policy
Research lnstitute and World Resources
lnstitute, Washington, D.C., USA.

CIP Program Report 1999 - 2000

329

Beyond Higher Yields: The lmpact of Sweetpotato


lntegrated Crop Management and Farmer Field
Schools in Indonesia
E. van de Fliert1, N. Johnson2 , R. Asmunati3, and Wiyanto 3

A pilot program of sweetpotato (/pomoea batatas (L.) Lam) integrated crop


management (ICM)-farmer field schools (FFSs) was implemented in six communities in Indonesia, using protocols developed jointly by a team of farmers,
researchers, and development workers. Monitoring and evaluation studies
showed that participation in the FFS enhanced farmers' crop management
knowledge and skills. Several of their changed cultivation practices led to
increased net income as a result of reduced cultivation costs and/or increased yields. Farmer participation in research was shown to have
contributed to the relevancy, appropriateness, and impact of the sweetpotato
ICM-FFS protocols.

During 1994-97, CIP, with support from


UPWARD (The Users' Perspectives for
Agricultura! Research and Development is
a CIP-affiliated network of Asian researchers conducting participatory R&D projects
in root crop systems.) and in collaboration
with public and private sector groups,
implemented a project to develop a
protocol for a sweetpotato ICM-FFS in
Indonesia. Collaborators were Mitra Tani,
a local nongovernment organization
(NGO); the National Research lnstitute for
Legume and Tuber Crops; and Duta
Wacana Christian University (Table 1).
Project activities were implemented in
major sweetpotato growing areas in East
and Central Java, where sweetpotato is
grown as an important cash crop throughout the year, mostly in rotation with rice.
The project strategy relied on participatory
approaches and methods at ali stages:
needs assessment and project design, ICM

CIP-ESEAP, Bogor, Indonesia.


CGIAR-PRGAand CIAT, Cali, Colombia.
3
Mitra Tani, Yogyakarta, Indonesia.

development and farmer learning protocols


applying the FFS approach, pilot-scale
implementation of the sweetpotato ICMFFS, and monitoring and evaluation. (For
details of project strategy, activities, and
outputs, see Van de Fliert et al., 1996; Van
de Fliert and Braun, 1997; Braun and Van
de Fliert, 1997; Asmunati et al., 1999; Van
de Fliert, 1999.)

To institutionalize the sweetpotato ICMFFS model that was developed and to


allow for large-scale farmer learning and
implementation, staff from the National
IPM Program (NIPMP) and 30 local NGOs
underwent FFS facilitators' training:
NIPMP staff in June 1997, NGOs staff in
April 1998. Follow-up programs were
implemented and funded by these local
extension organizations, and a second
research project was initiated to evaluate
their activities over a 2-yr period (199899) (Table 1). The work was carried out by
Mitra Tani, with methodological and
financia! support from CIP and UPWARD.

CIP Program Report 1999- 2000

331

Table 1. Overview of activities of the CIP/UPWARD-supported sweetpotato ICM-FFS 1 development and


evaluation projects in 1ndonesia, 1994-99.
PM&E of
Research phase
Needs assessment Technology (ICM)
FFS
development and
SP ICM FFS
and baseline 2
development3
institutionalization 4 implementation 5
Project 1: Sweetpotato ICM and ICM-FFS development
1994/95 wet season
PRA, RK, FO
Pilot FFS in SP
1995 dry season
RK, FO
8
Write modules
17 (3)
1995/96 wet season
RK, FO
Pilot SP ICM FFS
14 (6)
1996 dry season
12 (2)
Revise Modules
1996/97 wet season
ToT NIPMP
1997 dry season
6
Project 11: PM&E5
Process M&E
1997/98 wet season
1998 dry season
Process and impact
ToT NGOs
1998/99 wet season
lmpact evaluation
Analysis
1999 dry season
1 ICM-FSS = integrated crop management-farmer field school.
= participatory rural appraisal methods, RK = season-long record keeping, FO

2 PRA

= field observations.

3 Number

of farmer-managed trials; number of researcher-managed trials in parentheses.


4 NIPMP = National lntegrated Pest Management Program, Indonesia.
5 PM&E = participatory monitoring and evaluation.

Rigorous process and impact evaluation


were considered critica! elements of a
thorough documentation of the
sweetpotato ICM research, development,
and dissemination experience.
Sweetpotato ICM is knowledge-intensive
and site-specific. lts goal is to achieve
sustainable, collective agroecosystem
management by well-informed and skilled
farmers (Van de Fliert and Braun, 1999).
The FFS learning process provides farmers
new knowledge and skills through discovery and experimentation (Van de Fliert,
1993). Therefore, the monitoring and
evaluation study not only measured
economic impact of ICM, but analyzed
the FFS implementation and dissemination
process. In addition, it assessed impact in
relation to changed farmer capacity and
cultivation practices, and farm-level
effects using additional indicators far
human, social, and environmental capital.
In 2000, the sweetpotato ICM-FFS project
was selected as one of three cases to be
included in a stl)dy on the costs and
impacts of participatory research and
gender analysis conducted and funded by

332

Research on Sweetpotato

the Consultative Group on lnternational


Agricultura! Research (CGIAR) systemwide
program on Participatory Research and
Gender Analysis far technology development and institutional innovation (PRGA).
The objective of the PRGA study was to
assess the implications of incorporating
different types of user participation into
research on farmers' natural resource
management, with specific attention on
women and the poor. Four categories of
impacts are evaluated: 1) adoption and
impact of the technologies developed,
2) strengthening of human and social
capital among participating individuals
and communities, 3) establishment of
feedback links to formal research, and
4) costs of doing research.
CIP/UPWARD and PRGA impact evaluations of the ICM-FFS activities are
complementary. While both studies looked
at the impact of FFS implementation on
sweetpotato farmers' knowleqge and
production practices, the PRGA study
additional ly focused on the role of user
participation in the development of the
ICM-FFS. Taken together, their results

should provide insight into how user


participation influenced the orientation
and activities of the research project, and
how those changes in turn affected the
final impacts of the project technologies.

Season-long record keeping of post-FFS


sweetpotato cultivation by ICM farmers.
Observations in ICM and non-ICM
farmers' fields.
Annual evaluation and planning workshops with al 1 stakeholders.

Methods
The CIP/UPWARD monitoring and evaluation study appl ied an analytic framework
developed with all stakeholders during a
workshop at the beginning of the project.
lndicators far impact evaluation had been
determined with the stakeholders earlier
during the ICM research project. The
framework considered the various program
implementation levels far monitoring and
evaluation: 1) training of facilitators,
2) FFS implementation, 3) horizontal
dissemination within the community,
4) ICM implementation by trained farmers,
5) farm level effects, and 6) implications
far research and development institutions.
Far each evaluation leve!, infarmation
sources, variables, and indicators were
determined, and data collection methods
designed as fal lows.
Pre-FFS:
Observations and interna! evaluation
during facilitator training.
Individual qualitative baseline interviews with FFS participants (who later
become ICM farmers).
During FFS:
Process observation and recording
during FFS sessions.
Pre-tests/post-tests on ICM farmers'
knowledge.
Post-FFS:
Evaluation meetings with participants
and faci 1itators.
Individual interviews with facil itators,
village officials, and traders.
Two-step individual interviews with ICM
farmers (FFS graduates) and randomly
selected non-ICM farmers.

Farmers were intensively involved in


methods design, data collection and
analysis, and presentation during a final
seminar with research and extension
poi icymakers.
The PRGA study was designed to investigate the benefits and costs of including
users in research, particularly on how
impacts vary depending on the nature of
participation. The conceptual framework is
based on a typology of participation,
which defines types of participation in
relation to who makes decisions about
research priorities and activities (Lilja and
Ashby, 1999). For each of the stages of the
research process (design, testing, and
dissemination) hypotheses were developed
relating type of participation to each of
the four categories of expected impacts of
different types of participation (Johnson
and Lilja, 2001 ). These hypotheses helped
guide the design of the empirical methods
in each of the three cases. The hypotheses
far the ICM-FFS study, which used predominantly collaborative participation, are
shown in Table 2.
Both qual itative and quantitative data and
analyses were used to examine the
hypotheses regarding participation and
impact. The study relied greatly on
available reports and economic data from
the project. Available implementation and
production data were analyzed to assess
economic impacts. PRGA and CIP staff
collected additional data on human and
social capital impacts of participation
during a field trip in August 2000 by
conducting semi-structured interviews with
participants and key infarmants in the
project communities. Additional interviews with scientists, policymakers, and
extension and development workers were

CIP Program Report 1999 - 2000

333

Table 2. Hypotheses far the ICM-FFS study.


Stage
Type of participation
Hypothesis
Design
Consultative/Collaborative Farmer participation results in an increase in the proportion of the
targeted beneficiary group that could be reached by the project
because the priority tapie chosen far research is more relevant to
the needs and priorities of targeted farmers (H1 ).
Testing
Farmer participation at the testing stage results in an increase in
Collaborative
the number of potential adopters within the target group since the
specific technology ar technologies selected far recommendation
are more appropriate, given farmers' criteria and constraints (H2).
Dissemination Collaborative
There is an increase in the probability that potential adopters far
whom the technology ar technologies are appropriate will be
aware of them and be willing and able to adopt them and to
recommend them to others (H3).
conducted to assess the extent to which
there was feedback from the project to the
stakeholder groups. Final ly, through the
analysis of the available data sets and
interviews with project staff, the PRGA
study team assessed the overall impact of
user participation on project goals, activities, outputs, and costs.

Results
Due to limitations of space and the fact
that the PRGA analysis is still in process,
only a limited number of key results are
included here. The first set of results
pertains to measurable impacts of the
ICM-FFS on farmers' knowledge, skills,
production practices, and social relations
in the FFS villages. This analysis is
fallowed by a discussion of the role of
farmer participation in the design and
development of the ICM-FFS, and its
implications far impact.

lmpact of ICM-FSS on farmer knowledge,


skills, production practices, and social
relations.
FFS implementation and process evaluation. The impact assessments were
conducted in the six communities where
NIPMP conducted pilot-scale sweetpotato
ICM-FFS in 1997/98. Four of the districts
were the same as where the earl ier ICM
research project had been conducted, i.e.,
Mojokerto and Magetan in East Java
Province and Karanganyar and Magelang

334

Research on Sweetpotato

in Central Java. (In East Java, the FFSs


were implemented in the same hamlets,
although with different farmers than those
involved in FFS development.) The two
additional districts were Sleman in
Yogyakarta Province, and Kuningan in
West Java Province. Because the
sweetpotato FFSs were organized for rice
IPM-FFS alumni groups, the actual villages
selected in Magelang and Sleman were
not really major sweetpotato growing
communities. Selection of ICM farmers
was based on the NIPMP criteria far FFS
implementation, and basically involved
the rice IPM-FFS alumni.
An overview of sweetpotato ICM-FFS
implementation in the six NIPMP sites is
given in Table 3. All FFSs conducted at
least 17 sessions. Although an initial 25-28
participants were selected in each FFS,
the average attendance per session ranged
from 16-19 participants, which is comparable to IPM-FFS implementation in other
crops (Eveleens et al., 1996). Although
only 27% of FFS participants were women,
they were substantially represented in
three communities: Magelang (49%),
Sleman (56%), and Kuningan (60%). The
variation in women's participation reflects
the regional variation of women's involvement in sweetpotato cultivation. In
addition to currculum activities, all FFS
groups designed and conducted their own
experiments to test and refine the ICM
guidel ines and to practice experimental

Table 3. Overview of implementation and process evaluation of six NIPMP-conducted sweetpotato ICM
farmer field schools, 1997-98.
District
Mojokerto
Magetan Karanganyar Magelang
Kuningan
Sleman
Variable
East Java
East Java Central Java Central Java Yogyakarta West Java
yes
Majar sweetpotato
yes
yes
no
no
yes
growing area
FFS implementation
Jan-Jun 98 Aug-Dec 97 Sep 97-Feb Sep-Dec 97 Aug-Nov 97 Nov 97-Apr
season
(weVdry)
98 (wet)
(dry/wet)
(dry/wet)
98 (wet)
(dry/wet)
Meetings (no.)
17
17
17
18
17
17
Participants (range)
10-25
11-27
14-27
11-27
11-27
13-28
Meetings (avg. no.)
16
18
18
19
16
17
Women participation
0%
0%
7%
40%
56%
60%
(% of trainees)
varieties
varieties
Fertilization varieties; N Fertilization
Organic
Experiments
application (potassium,
fertilizer
conducted 1
(application
N-K rates)
rate; organic)
rates
126%
40%
Avg. knowledge
56%
28%
48%
254%
(25)
(10)
(13)
(24)
(19)
(20)
increase between
~re- and ~ost-test2
1 In addition to standard FFS trials.
2 Pre- and post-tests consisted of 10-15 questions each. Seores correspond with the number of correctly answered

questions. Percentage knowledge increase is calculated by ((score post-test- score pre-test)/ score pre- test)* 100%
The figures in the table represent the average knowledge increase of all participants per FFS who did both pre- and
post-test (number of respondents given in parentheses).

methods. The experiments dealt with


either fertilization or varietal evaluation.
Both training-of-facilitators and FFS
models increased knowledge among
trainees and demonstrated potential
impact of ICM in sweetpotato cultivation.
Knowledge increase of FFS participants
ranged from 28% to 254% per group (Table
3). lncreased yields (Figure 1) and more
effective or efficient use of externa! inputs
were demonstrated on collective learning
plots in most of the sweetpotato ICM-FFSs.
The farmers generally credited the FFS
process with adding to their knowledge
and skills. Knowledge areas singled out
included pests and diseases, natural
enemies, seed health, soil health, and
plant nutrient management. Skills areas
included routine field observations,
experimentation, fertilization, seedbed
preparation, pest management (none or
reduced pesticide use), and vine lifting.

Effects at farmer level (human capital). To


assess the impact of the FFS on participants' knowledge and ski l Is, post-FFS
comparisons were made between farmers

who had attended the ICM-FFS (ICM


farmers) and those who had not (non-ICM
farmers). Regarding production-related
knowledge and practices, results show that
ICM farmers are more likely to understand
the concept of natural enemies and do a
more thorough job on field observation
than non-ICM farmers (Table 4). Enhanced
knowledge and skills resulted in several
changes in farmers' sweetpotato cultivation practices. More ICM than non-ICM
farmers practiced seed selection and good
water management. Fertilization practices
changed in that more ICM farmers applied
more balanced fertilization. ICM farmers
were significantly more likely to use
potassium (K) fertilizer than non-ICM
farmers (P < O.OS), and they applied higher
average rates. More ICM than non-ICM
farmers used organic fertilizer, albeit at
lower appl ication rates. Severa! seasons
later, however, farmer researchers reported
the use of organic manure as one of the
majar benefits of ICM. Over time still
more farmers are beginning to use it and
understand its benefits. Surprisingly,
pesticide application frequency was not

CIP Program Report 1999 - 2000

335

Mojokerto

Magetan

Karanganyar
Magelang

Sleman

111

Non-ICM farmers, post-FFS

ICM farmers, post-FFS

lm

FFS plot

Kuningan

10

20
30
40
Sweetpotato storage root yield (t/ha)

50

60

Mojokerto
Magetan

Karanganyar
Magelang
Sleman
Kuningan

o
1o
20
30
40
50
Market price of sweetpotato storage roots
(US$/t)

10
20
30
40
Sweetpotato cultivation cost
(US$/ha)

250
500
750
Net income from sweetpotato
(US$/ha)

50

Mojokerto
Magetan

Karanganyar
Magelang
Sleman
Kuningan
250
500
750
1000
Gross income from sweetpotato
(US$/ha)

Figure 1. Sweetpotato storage root yield, market price, gross and net income, and cultivation costs at six ICM-FFS
sites, by baseline farmers (1994/95 wet season), on the sweetpotato ICM-FFS plot (for seasons, see Table 3), and
by ICM and non-ICM farmers during a post-FFS season. No post-FFS data were available for non-ICM farmers in
Magelang. Note: The high yields obtained on the FFS learning plots likely result not only from application of the ICM
practices, but also from the fact FFS plots are often relatively fertile and intensely managed. Moreover, they are quite
small and extrapolating kg/plot yield to Vha would be expected to overestimate the yield gains.

336

Research on Sweetpotato

Table 4. Differences in knowledge and skills, practices, and inputs and outputs of sweetpotato cultivation
during a post-FFS season by ICM-FFS participation. Magelang is excluded from the analysis because
there are no non-ICM cases in this community.

Variable
Knowledge and skills
Understand the concept natural enemy (%)
Practice routine/thorough field observation (%)
Conduct experiments (%)
Can assess yield within reasonable range (%)

ICM (N=73)

Non-ICM (N =50)

86

52
75/29
16
53

75/47
18

48

Cultural practices
Practice seed selection (%)
Practice good water management (%)
Use N-fertilizer (%/kg N/ha)
Use P-fertilizer (%/kg N/ha)
Use K-fertilizer (%/kg N/ha)
Use organic fertilizer (%/Vha)
Pesticides applications (avg. no./season)
Practice field sanitation after experiencing weevil attack (%)

45

28/20
35/1.3
0.4
49

37
58
92/242
60/58
13/7
29/1.6
0.3
30

107
29
201
20.5
32
391

112
27
237
19.5
30
316

68

96/223
68/78

lnputs and outputs1


Value of hired labor (US$/ha)
Value of inorganic fertilizer use (US$/ha)
Value of total cultivation cost (US$/ha)2
Yield (t/ha)
Price received (US$/t)
Net income (US$/ha)2
1 Exchange rate of Rp. 8,500/US$1.00.
2 Significant at P < 0.05.
influenced by ICM training. Thirty percent
of ICM farmers applied 1 or 2 applications
of pesticides on the sweetpotato crop
during the post-FFS season, which is
exactly the same as befo re trai ni ng, and
slightly more than among non-ICM farmers
(21 %), who, however, give only 1 or 2
applications as well. One reason may be
the already low levels of pesticide use in
these areas. More ICM than non-ICM
farmers, however, began to implement
field sanitation to manage sweetpotato
weevil.
In addition to cultural practices, the FFS
also included yield estimation of the crop
in the field. Sweetpotato farmers in Java
often sell their standing crop, so the ability
to accurately assess yields is an advantage
when negotiating a selling price. During
the needs assessment, farmers had identified their weak bargaining power as a
major constraint to achieving a good
income from sweetpotato.

However, yield assessment skills of ICM


farmers did not visibly improve (Table 4).
Nevertheless, attention to marketing in the
ICM-FFS seemed to have helped ICM
farmers in all except one district to
negotiate for a somewhat better price
(Figure 1 ), although the data are only
significant for Mojokerto (P = 0.063). In
Magetan, ICM farmers effectively received a lower per kg price than non-ICM
farmers (P = 0.056), because they were
unable to negotiate a higher per area unit
price with the traders, even though they
obtained higher yields.
Another important indicator of human
capital is the ability of farmers to experiment in their own fields. We found no
difference in experimentation between
ICM and non-ICM farmers. However, that
may be due to a bias in the data. Nearly
al 1 the non-ICM farmers who reported
experimenting were from Mojokerto,
where the farmer researchers who worked
on the development of the ICM-FFS

GIP Program Report 1999 - 2000

337

currculum during 1994/95 had a particularly strong influence, resulting in an


unrepresentative global average for nonICM farmers. In Karanganyar and
Kuningan, the FFS groups continued with
collective rather than individual experimentation on ICM components in the
sweetpotato crop after the FFS season.
Observations of the study team confirm
that ICM farmers do ha ve better ski lIs and
more interest in experimentation.

Effects at the farm level (financial


capital). Changes in knowledge, skills,
and practices appear to have translated
into a 24% higher average net income for
ICM farmers compared with non-ICM
farmers. ICM farmers obtained slightly
higher (5%) sweetpotato yields than nonICM farmers. That was not, however,
significant over the whole data set (Table
4). By district, yields differed significantly
(by t test) in Magetan and Sleman
(P < 0.05) and Mojokerto (P < 0.1 O). On
average cu ltivation costs of ICM farmers
were 15% less (P < 0.05) than those of
non-ICM farmers. But on a district basis
the difference was significant only in
Karanganyar (P < 0.05) and Magetan
(P < 0.1 O) (Figure 1). ICM farmers tended
to spend si ightly, but not sign ificantly,
more on inorganic fertilizer inputs, which
was mainly a result of their higher use of
the more expensive fertilizers such as K.
The cost of land renta! and use of hired
labor was the same for both groups.
Slightly higher yields and market prices
were responsible for ICM farmers receiving
a 10% higher gross income from
sweetpotato than non-ICM farmers (Figure
1). But the difference was significant only
in Kuningan (P < 0.10). Due in part to their
lower cultivation costs, ICM farmers did
achieve a significantly greater net income/ha than non-ICM farmers (P < 0.05),
particularly those in Mojokerto (P < 0.1 O)
and Karanganyar (P < 0.05).
The ICM-FFS teaches a broad range of
technologies and practices relating not
only to crop production but also to farm

338

Research on Sweetpotato

management and marketing. One way to


assess the combined impact of the new
knowledge, skills, and practices is to look
at how ICM-FFS attendance affects the
overall profitability of sweetpotato production. The dependent variable in the profit
function is net income/ha from
sweetpotato production. lndependent
variables that are expected to influence
profitability include expenditure on
fertilizer (mean value = US$43/ha) and
hired labor mean value = US$100/ha), and
dummy variables for land tenure status
(26% farmed rented land), for whether the
farmer attended ICM-FFS (60% attended),
and for water management practices used
(59% used routine rather than sporadic
irrigation). Community dummy variables
were also included to control for the
influence of local conditions. A CobbDouglas functional form was assumed.
The res u lts of the analysis show that
participation in ICM-FFS is significantly
and positively associated with net income
from sweetpotato production (Table 5).
(One explanation for the fact that participation in ICM-FFS is positively and
significantly associated with net income

Table 5. Results of estimation of sweetpotato profit


function coefficients; dependent variable is lag of
net income per hectare (N = 81, adjusted
R2 = 0.50). Magelang is excluded from the
analysis because there are no non-ICM cases in
this community and Sleman is excluded because it
is notan important sweetpotato growing area.
Standardized
P
regression
coefficient
0.000
(Constant)
Lag of hired labor costs/ha
0.089
0.335
0.102
0.375
Lag of fertilizer costs/ha
Water management dummy
0.290
0.002
(= 1 if irrigation is routine
as opposed to sporadic)
ICM participation dummy
0.271
0.002
(= 1 if attended ICM-FFS)
0.000
Land rental dummy
-0.382
(= 1 if rents)
Karanganyar dummy
0.387
0.001
Kuningan dummy
0.452
0.000
Magetan dummy
-0.059
0.553

from sweetpotato production cou Id be that


the ICM-FFS participants were already
better farmers than nonparticipating
farmers. To rule out this possibility,
baseline production data collected before
the development of the ICM technologies
and practices were analyzed, and did not
show that farmers who would subsequently
participate in ICM-FFS had higher yields
or net incomes than those who did not. In
fact, the opposite was true, suggesting that
the results presented in this paper may
underestimate the economic benefit of the
ICM-FFS.)
Participants in ICM-FFS have higher net
incomes than non-participants, after
controlling for other factors such as land
tenure and water availability. The community dummy variables are significant for
Kuningan and Karanganyar, the communities with very favorable ecological
conditions for sweetpotato production, and
where the response to the ICM-FFS was
most enthusiastic. Neither fertilizer
expenditure nor hired labor costs were
significantly associated with profitability
of sweetpotato production. As mentioned
earl ier, ICM farmers use more costly Kferti l izer but less N-fertilizer than non-ICM
farmers (Table 4). In an earlier version of
the analysis, gender was included as a
dummy variable, however, it was not
significant and therefore results are not
presented here.

Effects at the community level (social


capital). The col lective activities of FFSs
are carried out in a group and are intended
to strengthen the capacity of farmers to
work together, to share information, and to
learn from each other (Van de Fliert,
1993). To assess impacts of the ICM-FFS
on these social capital variables, data
were collected on farmers' practices
regarding sharing information about the
FFS and about agriculture in general, and
about their participation in group activities
with other FFS members. The majority of
ICM farrners (68%) reported disseminating

sweetpotato ICM information to other


farmers or neighbors. There was, however,

wide variation between the communities.


In five of the six communities, over 90%
of participants talked about the FFS with
others, but in Magetan only 41 % did so.
Magetan, it shou Id be noted, was where
the research had taken place. Most of the
villagers were aware of at least the
general issues of the sweetpotato ICM-FFS.
On average, each participant talked with
3 people. Nearly all the information
exchanges on sweetpotato ICM occurred
in the field (53%) or in home/group
meetings (42%). In Karanganyar and
Sleman, dissemination also occurred
beyond the village where the ICM-FFS
was conducted. Men and women differed
significantly in whom they talked to about
FFS. Men talked much more with neighbors and almost exclusively with other
men. Women talked much more with
relatives and similarly almost exclusively
with other women. ICM information
transmitted to other nonparticipating
farmers mainly dealt with variety and
healthy seed selection (62%), plant
nutrient management (54%), sweetpotato
pests and diseases (35%), soil preparation
and health (10%), and yield assessment
(10%). Spontaneous diffusion of information between farmers, however, cannot be
considered an impact of the project.
lnterviews with farmers and other key
informants suggest that the communities
have a long history of sharing agricultura!
information. The high levels of spontaneous dissemination are related to high
existing levels of social capital in the
communities. In severa! communities, FFS
groups have continued to engage in
collective experimentation, something
they did not do before participating in the
sweetpotato ICM-FFS.

The contribution of farmer participatory


research to the impact of the ICM-FFS
As mentioned earl ier, the research process
that led to the development of the ICM
technologies and practices and the ICMFFS currculum were developed in a
participatory mode. Farmer input was

CIP Program Report 1999 - 2000

339

obtained from needs assessment data and


from an intensive process of collaborative
research with eight farmers in four communities over a period of 3 years. Both
farmers and researchers conducted experiments, and regular workshops were held to
present findings, analyze results, and plan
fu tu re work.
The importance of farmers' contributions to
the research process was made clear early
when it was discovered that farmers did
not bel ieve that pests were the mai n
constraint to sweetpotato production. The
original CIP/UPWARD project had called
for the development of an integrated pest
management (IPM) FFS curriculum.
Farmer input resulted in a shift from IPM to
ICM. Within the ICM framework, tapies
that farmers identified as important and on
which they carried out experiments
included plant nutrient management,
varietal selection, and cultural practices
such as planting methods, vine lifting, and
intercropping. In four of these five areas,
the farmers' experi ments generated resu lts
for practices and guidelines that were
included in the final curriculum. The
experience worki ng with farmers also
contributed to the. ICM-FFS focus on
health, and its emphasis on teaching
principies and experimental methods.
These impatt results show that the major
areas in which the FFS enhanced farmers
knowledge, productivity, and profitability
were in plant nutrient management, water
management, varietal and planting
material selection, and overal 1 better
observation skills. An FFS focused entirely
on pest management would likely have
had much less impact. Fertilizer and seed
selection were by far the two tapies most
commonly mentioned by ICM-FFS participants when they talked to other farmers.
That suggests that they are the elements
that will be most widely disseminated
informally. The results regarding individual
and group experimentation indicate that
an ongoing process of productivity improvement may have been set in motion
by the ICM-FFS process.
340

Research on Sweetpotato

Conclusions
Sweetpotato ICM-FFS has contributed to
increased knowledge and skills among
farmers who participated. These human
capital increases have resulted in concrete
changes in production and marketing
practices that contributed to higher yields
and higher income from sweetpotato
production. These results were essentially
achieved by improving the efficiency of
existing practices rather than by introducing totally new technologies. Farmer
participation in the research process,
particularly in the identification of possible areas for improving efficiency,
appears to have contributed to both the
relevance and the impact of the ICM-FFS.
Whereas most of the impacts shown in this
study are significant statistically, they may
seem to be less so in practica! terms. The
magnitude of the differences of individual
parameters between ICM and non-ICM
farmers are often relatively small, but
taken as a whole they indicate the initiation of change in farmers' crop
management behavior. As mentioned
earlier the data may underestimate the
true differences between ICM and nonICM farmers, mainly as a result of the
inability to control for spillover effects
from participants to nonparticipants. Given
that ICM is complex and that increased
knowledge does not immediately result in
changed practices, it is likely that over
time the benefits from ICM-FFS may grow.
Subsequent field visits as part of the PRGA
study gave that impression. A final factor
that could have affected the magnitude of
the observed impacts is the Asian financia!
crisis that hit Indonesia in 1997. The
profitabi 1ity of sweetpotato was drastical ly
reduced, hence the reduced incentives for
farmers to i nvest in it.
Nonetheless, the question of whether the
benefits justify the costs of ICM-FFS is
valid. This issue will be examined in
greater detail in the forthcoming PRGA
study. Participatory development of the
sweetpotato ICM-FFS took several years

and was intensive due to the amount of


component research required to fill the
gap of technological guidelines for the
crop. In addition, a broader and more
widely applicable FFS model was developed and documented. That facilitated
similar CIP projects in other countries in
and beyond the region. To assess costeffectiveness of the project, it wou Id have
to be compared with a conventional CIP
research project with similar scope aimed
at both productivity enhancement and
natural resource management, rather than
with a conventional extension-oriented
FFS.
The six pi lot field schools were conducted
in the selected districts to trigger interest
from local authorities so that they might
allocate funds for larger-scale implementation. But the economic crisis that hit
Indonesia in 1997 onwards, and simultaneous failures in sorne major rice bowls in
Indonesia during 1998, caused ali agricultura! development to be allocated to rice
production. Expansion of sweetpotato FFS
implementation within the government
extension system has not yet been
achieved. The newly established Directorate of Root Crops Production has
committed itself to organizing a nationwide sweetpotato ICM-FFS program
whenever the economic situation allows.
lt is hoped that the results of this study will
contribute to the implementation of the
program. Additionally, NGO programs
have continued to apply ICM-FFS
apporaches in sweetpotato and other
crops, and these efforts will be evaluated
during 2001-2002.

References
Asmunati, R., Wiyanto, and E. van de
Fliert. 1999. lntegrating sweetpotato
ICM and FFS approaches in government
and NGO initiatives in Indonesia, In:
Users' Perspectives for Agricultura!
Research and Development (UPWARD),
Learning to manage livelihoods: New
perspectives in rootcrop R&D.

UPWARD, Los Baos, Laguna,


Philippines. p. 129-139.
Braun, A.R. and E. van de Fliert. 1997.
The farmer field school approach to IPM
and ICM in Indonesia: User
participation. In: UPWARD. Local R&D:
lnstitutionalizing innovations in rootcrop
agriculture. (Users' Perspectives for
Agricultura! Research and Development
(UPWARD), Los Baos, Laguna,
Philippines. p. 44-64.
Eveleens, K.G., R. Chisholm, E. van de
Fliert, M. Kato, P.T. Nhat, and
P. Schmidt. 1996. Mid-term review of
Phase 111 report. FAO lnter-cuntry
Program for the Development and
Application of lntegrated Pest Control
in Rice in South and Southeast Asia,
Makati, Metro Manila, Philippines.
Johnson N. and N. Lilja. 2001.
Characterizing and measuring the
effects of incorporating stakeholder
participation in natural resource
management research: Analysis of
research benefits and costs in three case
studies. Participatory Research and
Gender Analysis (PGRA) Working
Document. PRGA/Centro Internacional
de Agricultura! Tropical (CIAT), Cali,
Colombia.
Lilja, N. and J.A. Ashby. 1999. Types of
participatory research based on locus of
decision-making. Participatory Research
and Gender Analysis Working
Document (PGRA) No. 6. PRGA/CIAT,
Cali, Colombia.
Van de Fliert, E. 1993. lntegrated pest
management: Farmer field schools
generate sustainable practices. A case
study in Central Java evaluating IPM
training. Wageningen Agricu ltural
University Papers (93)3 PUDOC.
Wageningen, Netherlands. 304 p.
Van de Fliert, Elske. 1999. lntegrative,
farmer-participatory methodology for
poverty-sensitive research: Sweetpotato
and patato integrated crop management
in Southeast Asia. Paper presented at
the Centro Internacional de Agricultura!

Tropical lnternational Workshop

GIP Program Report 1999 - 2000

341

Assessing the impact of agricultura!


research on poverty al leviation held
14-16 September 1999, San Jos, Costa
Rica, 14 p. Available at: http://
www.ciat.cgiar.org/poverty/
bar_papers.htm.
Van de Fliert, E. and A.R. Braun. 1997.
One step back, two steps forward:
Sweetpotato ICM development in
Indonesia, In: UPWARD. Local R&D:
lnstitutionalizing innovations in rootcrop
agriculture. UPWARD, Los Baos,
Laguna, Philippines. p. 153-167.
Van de Fliert, E. and A.R. Braun. 1999.
Farmer field school for integrated crop

342

Research on Sweetpotato

management of sweetpotato: Field


guides and technical manual.
lnternational Patato Center, Bogar,
Indonesia. 286 p.
Van de Fliert, E., R. Asmunati, Wiyanto,
Y. Widodo, and A.R. Braun. 1996. From
basic approach to tailored curriculum:
Participatory development of a farmer
field school model for sweetpotato. In:
UPWARD. lnto action research:
Partnerships in Asian rootcrop research
and development. UPWARD, Los Baos,
Laguna, Philippines. p. 59-74.

The Sweetpotato Production-Postharvest Use System


in Vietnam: Participatory Needs and Opportunity
Assessment
E. van de Fliert1, N.T.K.Oanh 2, N.T. Son 2 , T.D.Hoa 3, P.D.Thanh4 , L.Q. Hung5,
L.H. Khoang6 , and N.T. Lan 7

A participatory needs assessment on root crops production-postharvest use


systems was conducted in eight major root crop production provinces of
Vietnam in 1998/99 as a collaborative effort of six Vietnamese institutions,
lnternational Potato Center, and UPWARD. Sweetpotato (lpomoea batatas
(L.) Lam) production data showed that the greatest opportunity for increased
profitability from sweetpotato was in more effective and efficient use of
existing technologies. That implied the development of farmer learning methods and mechanisms rather than technology development research. A
follow-up workplan for the development and institutionalization of adapted
integrated crop management (ICM) farmer field school (FFS) protocols for
sweetpotato resu lted from the study. Moreover, research and extension
organizations committed themselves to implement the workplan. The participatory nature of the study has been instrumental in achieving the final output.

CIP's 1998-2000 Medium Term Plan


identified Vietnam as a priority country
within the program framework for
sweetpotato research (Walker and Collion,
1997). Vietnam ranks third in cultivated
area after China and Uganda among the
world's sweetpotato producing nations.
Workplans of individual CIP projects also
emphasized the continuation of existing
work or implementation of new research
activities, especial ly in genetic resources
conservation and breeding, integrated pest
management (IPM), and postharvest use.
During a meeting between CIP scientists

CIP-ESEAP, Bogar, Indonesia.


Hanoi Agriculture University, Hanoi, Vietnam.
Hue University of Agriculture and Forestry, Hue, Vietnam.
4
Thang Binh DistrictAgriculture and Rural Development

Bureau 1 Quang Nam Province, Vietnam.


5

Collegeof Agricultureand Forestry, Ho Chi Minh City, Vietnam.


6
Post Harvest Technology lnstitute, Ho Chi Minh City, Vietnam.
7
Hong Duc University, Thanh Hoa, Vietnam.

and national agricultura! research systems


partners in 1998, the lack of comprehensive, reliable, and updated information on
root crop production-postharvest use
systems in Vietnam was identified as a key
issue. A needs and opportunity assessment
was recommended as a key step toward
planning and undertaking action-oriented
research. A col laborative study was
designed involving scientists from four
lnternational Potato Center (CIP) research
areas (IPM, breeding, postharvest utilization, and canna germplasm), and from the
six Vietnamese national research institutions represented by the authors of this
paper. This study was supported by CIP and
UPWARD (The Users' Perspectives for
Agricultura! Research and Development
(UPWARD) is a CIP-affiliated network of
Asan researchers conducting participatory
research and development projects in root
crop systems).
CIP Program Report 1999 - 2000

343

The overal 1 objective of the study was to


establish a baseline of root crop systems in
key production areas in Vietnam and to
identify opportunities and mechanisms for
improvement of root crop enterprises. The
specific objectives were:

Data are reported only for the six provinces considered key sweetpotato
production areas. Hoa Binh and Dong Nai
provinces, which were studied only for
canna production, are not included in the
analysis.

to document and analyze root crop


systems in selected provinces of
Vietnam, with an emphasis on
sweetpotato;
to determine root crop farmers' variety
requirements, cultivation and
postharvest constraints and practices,
and needs and opportunities for improvement; and
to identify and prioritize relevant
research and development interventions
and mechanisms for root crop technology development, promotion, and
institutional ization.

Data collection was done in four villages


per province (except Ba Ria-Vung Tau in
which only three were studied) during the
1998/99 winter-spring and spring-summer
seasons by teams of one researcher and
one field assistant per province. The team
members were involved in pretesting and
revising the data collection methods. For
data collection the following methods
were employed.

This paper outlines the methods of the


study, presents the majar findings with a
focus on sweetpotato production, and
discusses how the participatory and
integrative nature of the exercise
contributed to research planning and
implementation of CIP's sweetpotato IPM
activities in Vietnam.

Methods
The study emphasized two key types of
baseline information, i.e., 1) production
and crop management, and 2) postharvest
use in both sweetpotato and canna systems. Study sites represented key
sweetpotato or canna production areas,
and included eight provinces across the
country.
Northern Vietnam: Vinh Phuc and Ha
Bac provinces (sweetpotato), and Hoa
Binh (canna).
Central Vietnam: Thanh Hoa Province
(sweetpotato and canna), and Quang
Nam Province (sweetpotato).
Southern Vietnam: Ba Ria-Vung Tau and
Vinh Long provinces (sweetpotato), and
Dong Nai Province (canna).

344

Research on Sweetpotato

Analysis of secondary data.


Village profiling using participatory
rural appraisal methods including
transect walks; informal discussions
with farmers, traders, retai lers, processors, and consumers; participatory
mapping using production criteria; and
focus group discussions using ranking
exercises, seasonal calendars, gender
analysis, and open discussion on
technology and learning needs.
Season-long record-keeping of
sweetpotato/canna production by 1 O
randomly selected farmers per study
vil lage per crop, or 40 farmers per site
(except Ba Ria-Vung Tau that had only
25).
Observation in sweetpotato fields of
record-keeping farmers once every other
week.
Individual interviews, midway in
the season and at harvest, with recordkeepi ng farmers about qualitative
aspects of sweetpotato/canna
production.
Individual interviews on root crop
postharvest use focusing on pig feed and
starch with approximately 1 O processing
families per village, mostly not the
same as the production respondents.
End-of-season analysis meetings in each
village to interpret results and prioritize
research needs.

Results were analyzed statistically to


determine whether there is a linear
relationship between fertilizer rates and
vine or storage root yield. Additionally, a
prediction model for vine and storage root
yield was studied by multiple regression
analysis. A nationwide seminar was
conducted in October 1999 to present the
needs assessment method and res u lts to
stakeholder group representatives. The
seminar was followed by a workshop with
existing and potential partners in the
sweetpotato IPM project to determine
future program strategies and to plan
fol low-up activities.

Results

Characterization of sweetpotato
cultivation areas and systems
Sweetpotato cultivation in the various
regions in Vietnam varies by agroecology
and climate. In all study sites, sweetpotato
is typically grown as a field crop in
rotation with rice and other crops such as
maize, groundnut, cassava, and vegetables. Winter and spring constitute the
major cultivation season in northern and
central Vietnam. In southern Vietnam
sweetpotato is grown year-round, but
preferably during the dry season, which
coincides with the winter-spring season in
the North. Women have a major role in
sweetpotato cu ltivation, particularly in
northern and north-central Vietnam.
Further to the south the average proportion
of work done by women is smal ler (Table
1), which is consistent with results of
Tuyen (1999). Sweetpotato ranks first
among major field crops in Ha Bac, Thanh
Hoa (one district), Ba Ria-Vung Tau, and
Vinh Long provinces, and either second,
third, or fourth (depending on a specific
community) in Vinh Phuc, Quang Nam,
and Thanh Hoa. Sweetpotato is favored
because of its high productivity and low
management and input requirements,
which makes it an easy and potentially
profitable enterprise. Additional ly, seed is
readily available, and in sorne areas
sweetpotato is one of the few crops

adapted to prevailing soil conditions.


Disadvantages include unreliable marketing opportunities and a generally low
market price. In sorne areas, farmers face
low productivity or low product quality.
A majority of farmers perceive their soils
as moderately to fairly fertile; only in
Quang Nam do a substantial number of
farmers rate their sandy soils as fairly
infertile (Table 1). Eighty-three percent of
all farmers, although relatively fewer in
Vinh Long (50%), are aware that they
could improve the fertility of their soils
with applications of organic fertilizer,
whereas 55% believe that inorganic
fertilizer could do the job. A majority of
farmers rate their water supply as unreliable. The water supply in Ha Bac is
considered moderately to fairly reliable.
Only in Vinh Long, which is located in the
Mekong Delta, do all farmers rate their
water availability as fairly reliable. Major
constraints to sweetpotato cultivation vary
across provi nces and seasons. They
include sweetpotato weevils, stem borers,
soil nutrient deficiencies, the lack of
suitable varieties, and abiotic factors such
as cold, drought, or floods.
Both vi nes and storage roots are u sed in
the northern and central provi nces,
whereas in the South the deliberate use of
vines is rare. Vine production accounts for
an average of 62% of the total gross
income from sweetpotato in Thanh Hoa
and is equally important for use on-farm as
animal feed or for market (Table 2). In
Vinh Phuc, Ha Bac, and Quang Nam,
approximately one third of total gross
i ncome from sweetpotato comes from
vines, with a focus on home consumption
by farm animals (mainly pigs). Storage
roots find a variety of uses in northern and
central Vietnam, whereas in the South the
bulk is sold to the fresh market (Table 2).

Sweetpotato production
In all study sites, farmers plant sweetpotato on ridges, although the width and
height, and accordingly plant population,
vary by soil type, water supply, and local
CIP Program Report 1999- 2000

345

w
.:..
O)

o
:::1
en

$.

'O

Table 1. Sweetpotato (SP) production and postharvest use characteristics in six needs assessment sites in Vietnam, 1998/99 .
Northern Vietnam
Central Vietnam
Characteristic
Southern Vietnam
Vinh Phuc
Ha Bac
Thanh Hoa
Quang Nam
Ba Ria-Vung Tau
Vinh Long
1:0.3
1:0.4
1:0.6
1:1.1
1:1.6
Labor ratio,
1:1.4
women:men (h/ha)
1, Befare rice,
1, Befare rice, beans,
2, 3, or 4, After
Rank of SP among
3, 4, After rice,
1, Befare rice
1, Befare rice,
maize, groundnut groundnut. 2, 3, after
other majar crops
maize, sugarcane,
cassava, rice,
maize, vegetables
groundnut, rice
groundnut
mulberry, soybean
Typical cultivation
pattern 1

Rice or maize - rice SP or vegetables

Rice - SP or
vegetables - rice
or groundnut

Rice or beans - SP, or


Maize or vegetables SP or beans

Rice - SP Vegetables or
cassava/ sesame

Rice - SP

Rice or SP - Rice SP or vegetables

Most serious
production
constraints

SP weevil
Low soil fertility
Lack of suitable
varieties

Soil fertility

SP stem borer SP weevil


Unreliable water supply

SP stem borer
Virus es
Drought, cold

Lack of suitable
varieties
Lack of capital

SP weevil
Lack of capital
Heavy rain

Soil fertility
Fairly fertile
Moderately fertile
Fairl~ infertile
Water availability
Fairly reliable
Moderately reliable
Fairly unreliable
1 Spaced

38%
50%
13%

63%
38%
0%

73%
28%
0%

3%
53%
45%

100%
0%
0%

0%
25%
75%

58%
43%
0%

0%
3%
98%

23%
15%
63%

0%
0%
100%

100%
0%
0%

hyphen indicates followed by. For example: rice crop followed by a sweetpotato crop indicated as rice - sweetpotato.

Table 2. Contribution to gross income of the various uses of vines and storage roots, as average percent of
total gross income (including opportunity value far in-kind products) from sweetpotato. Vietnam, 1998/99.
Use
Northern Vietnam
Central Vietnam
Southern Vietnam
Vinh Phuc
Ha Bac
Thanh Hoa Quang Nam
Ba Ria-V. Tau Vinh Long
{n=40}
{n=40}
{n=40}
{n=40}
{n=25}
{n=40}
Vines:
38.5
31.5
62.5
29.6
O.O
0.2
o.o
Animal feed
24.0
18.7
29.7
23.2
o.o
12.7
28.6
4.2
o.o
Sold to market
10.7
0.2
4.2
2.1
Seed
2.9
O.O
O.O
O.O
0.8
O.O
O.O
0.1
O.O
Give away/
O.O
other uses
37.5
70.4
100.0
Storage roots:
61.5
68.5
99.8
99.3
97.5
11.3
Sold to market
10.2
39.4
13.6
17.0
0.1
7.8
0.1
21.3
Animal feed
25.4
19.1
0.1
Family
5.8
5.8
1.6
13.0
consumption
O.O
O.O
O.O
10.4
22.9
Used far
1.0
processing
O.O
O.O
2.0
0.1
O.O
Discarded 1
3.9
O.O
O.O
O.O
0.5
0.6
Give away/
8.1
other uses
1 Due to weevils, rot and other causes.

habits. The widest ridges, containing two


rows of plants instead of one, are found in
Quang Nam and Ba Ria-Vung Tau (Table
3). Over time, farmers appear to have
developed the most appropriate planting
system under the prevailing conditions in
the various areas. Almost all farmers in the
northern and central provinces, and a
majority in Vinh Long, practice seed
selection mainly by discarding cuttings
with symptoms of pests and diseases, and
the selective use of tip cuttings or the use
of vines from healthy mother plants. None
of the farmers in Ba Ria-Vung Tau are
reported to practice seed selection (Table
3). Fertilization practices vary widely
among and within provinces. Farmers
commonly use organic fertilizer, except in
Vinh Phuc (Table 3). The average amounts
applied seem adequate to maintain soil
fertility in Ha Bac, Thanh Hoa, and Quang
Nam, but are suboptimal in Ba Ria-Vung
Tau and Vinh Long. lnorganic nitrogen
fertilizer is commonly applied. Average
doses in all provinces except Thanh Hoa
are excessive and, considering the relatively low yields, seemingly highly
inefficient, as illustrated by a nonsignificant (P > O.OS) correlation between the

dose of total inorganic fertilizer and vine


yield (Pearson product moment correlation
coefficient as low as 0.132) or storage
roots (0.122). Phosphate and potassium
fertil izer applications are widely practiced, mostly in excess, in Thanh Hoa and
Ba Ria-Vung Tau, but variably in Ha Bac,
Quang Nam, and Vinh Long, and not at all
in Vinh Phuc.
Pesticides are used only by farmers in
Quang Nam, where the majority practice
one or two applications a season, and in
Vinh Long where all farmers in one study
district used pesticides excessively (on
average four applications a season) (Table
3). Field observation is commonly practiced by farmers in Vinh Phuc, Ha Bac,
and Ba Ria-Vung Tau, but only occasionally by most farmers in the other provinces
(Table 3). Only in Quang Nam and Ba RiaVung Tau do substantial numbers of
farmers understand natural enemies and
their role in the ecosystem and identify
especially ants, frogs, and dragonflies as
examples of natural enemies. That might
be an effect of intensive IPM training that
has taken place in rice in these provinces.

CIP Program Report 1999 - 2000

34 7

Table 3. Sweetpotata production practices, farmer knowledge and skills in the six needs
(% farmers reQorting unless SQecified otherwise).
Variable
Northern Vietnam Central Vietnam
Ha
Vinh
Thanh Quang
Bac
Phuc
Hoa
Nam
{n=40} {n=40} {n=40} {n=40}
Sail preparatian and planting:
Avg. width af ridge (cm)
80
67
107
199
Avg. height of ridge (cm)
40
34
54
58
Avg. planting density ('000 plants/ha)
67
48
47
54
Practice seed selectian
98
100
98
100
Apply organic manure
8
100
100
100
Apply fertilizer
lnorganic N
95
100
100
100
lnorganic P
o
13
95
8
lnarganic K
o
88
100
33
Apply pesticides
o
o
o
85
Practice field manitoring:
Occasionally
o
18
80
98
Routinely
100
83
20
3
Understand the concept of natural enemies
o
13
8
70
Can assess root yield within reasonable range
45
35
15
38
Market prices
Far vines (01/kg) (avg.)
348
216
400
505
Far roats (D/kg) (avg.)
624
534
716
503
Stare sweetpotato roats
85
100
43
100

assessment sites
Southern Vietnam
Ba RiaVinh
Vung Tau Long
{n=25} {n=40}

133
70
64

100
100
100
100

100
57
40
68
100
98
48

53

98
3
3

100
100
28
1100

18

1000
1435

1 O= dong, Vietnamese currency unit; 14,000 dong = US$1.

Storage root yields ranged from an average


of 9.2 t/ha in Thanh Hoa to 17.2 t/ha in
Quang Nam, whereas vine yields averaged from 10.7 t/ha in Vinh Phuc to 18.6 ti
ha in Quang Nam (Figure 1A). Despite the
relatively poor soils in Quang Nam
compared with the relatively rich soils in
Vinh Long, Quang Nam farmers achieve
equally high yields through suitable
management (Figure 1 B). Analysis of
cultivation practices shows that no variable possibly contributing to yield (i.e.,
plant density, crop growth duration, labor,
fertilization rate, or pesticide application
frequency) strongly predicts vine or storage
root yield across sites (Table 4). This is the
result of the high variability of practices
among farmers, possibly due to lack of
adequate information to increase production efficiency, and among areas due to
varying ecological and utilization conditions. Only the data for Quang Nam and
Vinh Long resulted in a fairly strong

348

Research on Sweetpotato

prediction model. In Quang Nam, optimization particularly of plant density, but


also of labor and N-fertilizer application
rates, is likely to improve both vine and
root yields. This may be related to the soil
quality in this province, which is very
sandy and hence requires adapted crop
care. In Vinh Long, crop growth duration
occurred as a relatively strong determining
factor with a positive correlation to both
vine and root yield. The average growth
duration is relatively short in this province.
Farmers might increase their production
levels significantly by allowing the crop a
few more weeks in the soil, although this
will depend on the market price at the
time of harvest and the opportunity cost of
land and time during that particular
season. The significant values for various
other variables in the models, albeit at low
coefficients, indicate that there is room for
improvement in the production systems but
also that a location specific approach wi 11
have to be appl ied.

1,600

40

Vines

Roots

1,400
1,200

30

1,000

en
2

800

?
.e
2!. 20

Labor

~ lnputs

Net return

(])

"'O

Q5

>=

?
.e

600

10

-~

400

200

en
en

C)

Province

Province

Figure 1: (A) Average total yield broken down into average storage root yield and vine yield (left). (B) Average
gross income broken down into net income and cost for inputs and labor.
Table 4. Multiple linear regression of cultivation variables possibly accounting for sweetpotato vine and
storage root yield (dependent variable). lndependent variables include plant density, labor, crop growth
duration, doses of organic manure and inorganic (N, P, K, and others which is mainly NPK composite)
fertilizers, and pesticide application frequency. Yalues per entry are standardized coefficient and probability
P (in parenthesis). Only values far independent variables with a probability of P<0.1 O are displayed.
lndependent variables
Adjusted Density
Growth
Labor
Fertilizer
Pesticide
R2
(plants/ha) duration
(hr/ha)
N (kg/ha) K (kg/ha) Organic
application
(days)
(kg/ha)
frequency
Vines
0.514
Vinh Phuc
0.225
(0.006)
-0.354
0.365
0.198
Ha Bac
(0.068)
(0.059)
0.303
Thanh Hoa 0.032
(0.094)
-0.914
0.223
0.239
Quang Nam 0.773
(0.022)
(0.013)
(<0.001)
Roots
0.162
0.389
Vinh Phuc
(0.040)
0.440
-0.538
0.329
0.213
Ha Bac
(0.009)
(0.086)
(0.008)
0.383
Thanh Hoa 0.015
(0.038)
0.007
0.188
Quang Nam 0.816
-0.883
(0.240)
(0.028)
(<0.001)
-0.043
Ba RiaVung Tau
-0.122
0.301
0.773
Vinh Long
0.869
(0.062)
(<0.001)
(0.018)

CIP Program Report 1999 - 2000

349

The majority of farmers in the northern and


central provinces store sweetpotato roots
in their houses before sel ling them,
whereas in southern Vietnam farmers sell
their crop immediately (Table 3). The
prices farmers receive vary considerably,
with the lowest prices for vines in the
northern provinces of Vinh Phuc and Ha
Bac, and the -lowest prices for roots in the
more remote areas of Ha Bac and the
central province of Thanh Hoa (Table 3).
Average storage root price is relatively
high in the two southern provinces, Ba RiaVung Tau and particularly in Vinh Long. As
a result of relatively high yield and high
market price, the gross income from
sweetpotato roots in Vinh Long is significantly higher than in all other provinces
(Figure 1). In Quang Nam, where root
yield was the same as in Vinh Long,
farmers benefit from the additional value
of the vines, resulting in an average total
gross income from sweetpotato that is not
significantly lower than in Vinh Long.
Because of high expenditures for inputs
and labor, net return from sweetpotato in
Quang Nam is not significantly different
from that in other provinces, and is lower
than in Vinh Long. Farmers in the northernmost provinces, Vinh Phuc, Ha Bac, and
Thanh Hoa in central Vietnam have a low
net income from sweetpotato, despite
relatively low expenditures, as a result of
low yields and low prices. But in Ba RiaVung Tau, despite a high market price, low
net income resulted from low yield and
high expenditures.

Follow-up research agenda and activities


During the needs assessment analysis
workshop, the major opportunity areas and
actions needed for improvement of the
production system were defined and
ordered as listed below.

Marketing. Develop and implement FFSs


to enhance farmers' bargaining power.
Soil and nutrient management. 1) Review
existing information on soils to develop
broad fertilization guidelines for each
area, 2) conduct participatory field studies
350

Research on Sweetpotato

to test and adapt guidelines, and 3)


develop and implement FFS activities on
crop nutrient requirements and on experimental skills needed to adapt fertilization
guidelines.

Major pests and diseases (including


weevils, stem borers, and root rot).
1) Conduct appl ied research on the use
and production of the entomopathogen
Beauveria bassiana to manage weevils in
field and storage, 2) conduct large-scale
testing of storage techniques in pilot FFSs,
and (3) develop and implement FFSs on
weevil and stem borer management.
Workshop participants concluded that the
sweetpotato ICM-FFS approach in Indonesia (Van de Fliert and Braun, 1999) would
have potential for Vietnam after being
adapted to local conditions. lt covers al 1
aspects of sweetpotato cultivation, from
soil preparation to marketing to
postharvest use, with ecological and
economic sustainability an underlying
principie. Vietnam has a very strong IPM
program applying the FFS model to
training in rice and other crops. With
relatively little investment, the lndonesian
sweetpotato ICM-FFS protocols could be
adapted for the Vietnamese conditions;
and, the ski l Is of Vietnamese FFS faci 1itators could be upgraded so they could
conduct sweetpotato ICM field schools.
The analysis workshop brought about a
commitment from the national IPM
program to follow up on the research and
development phase of the project. The
follow-up work plan contains activities to
review, field-test, and revise the
sweetpotato IPM-FFS protocols; to train
master trainers; and initiate a larger-scale
program by the local institutions.

Discussion
Sweetpotato productivity in Vietnam is
general ly low. But in the study sites, where
one would expect the crop to be well
suited to local conditions, average storage
root yields were higher than the national
average of 6.5 t/ha (FAO, 2000). Low

yields were mainly due to stress factors


such as low soil fertility, low temperature
in the short wi nter growi ng season, and
insect pests (sweetpotato weevil and stem
borer). Whereas in the late 1980s the
unavailability of fertilizers was a major
constraint (Bottema et al., 1991 ), chemical
fertilizer use is now common. Even though
use has more than tripled over the past
decade (FAO, 2000), it is often inefficient
as illustrated by the correlation and
regression analyses done in this study.
Balanced nutrient management and
effective pest management are expected
to increase profitability.
The team members evaluated the study
methods and implementation process
positively. Al 1 but one felt the three
objectives were fu 1ly or for the most part
achieved. They were particularly positive
about having achieved the third objective,
identifying and prioritizing relevant
research and development interventions
and mechanisms for root crop technology
development, promotion, and institutionalization. Ali methods except secondary
data analysis were considered instrumental
in achieving the objectives. The study
provided a valuable learning experience
for the research team members, many of
whom had mainly done conventional,
disciplinary research work before. The
interdisciplinary composition of the study
team favored the integrative approach of
the study. Farmers' active involvement in
the needs assessment is expected to result
in increased farmer interest in the study
sites in future training.

Conclusions
The i ntegrated needs assessment reported
here is the most comprehensive study done
to date in Vietnam on root crop production-postharvest use systems. Containing as
it does production, post-production, and
socioeconomic elements, it provides a
fairly complete picture of the constraints

and pportunities of root crop enterprises


in various parts of the country. The study
served as a foundation for further production and use technology and extension
development, and resulted in an operational workplan for the sweetpotato IPM
research group. Needs and opportunities
identified were directed more toward
farmer learning than research, and more to
crop cultivation in general than pest
management in particular. Consequently,
the follow-up workplan concentrated on
the development of an adapted, flexible
ICM-FFS protocol to suit the various
production-postharvest use systems across
the country. lnvolvement of ali stakeholder
groups in the final needs assessment
analysis was instrumental in reaching
conclusions and developing a workplan,
including the commitment from the
national extension system to fol low up
research efforts on a nationwide scale.

References
Bottema, J.W.T., P.T. Binh, D.T. Ha,
M.T. Hoanh, and H. Kim. 1991. Sweet
potato in Viet Nam, production and
markets. CGPRT No. 24. CGPRT, Bogor,
Indonesia. 113 p.
FAO. 2000. FAOSTAT Database, http://
apps. fao.org/.
Tuyen, N.N. 1999. Women in Vietnam's
national IPM program, In: van de Fliert,
E. and J. Proost. Women and IPM: Crop
protection practices and strategies. KIT
Press, Royal Tropical lnstitute,
Amsterdam, Netherlands. p. 79-88.
Van de Fliert, E. and A.R. Braun. 1999.
Farmer field school for integrated crop
management of sweetpotato: Field
guides and technical manual.
lnternational Potato Center, Bogor,
Indonesia. 286 p.
Walker, T. and M.H. Collion. 1997. Priority
setting at CIP for the 1998-2000
Medium-Term Plan. lnternational Potato
Center, Lima, Peru. 48 p.

GIP Program Report 1999 - 2000

351

Fermented Sweetpotato Feed for Pigs:


A Boon to Farmers
A fermented mixture of sweetpotato vines and sun-dried chicken manure is having
a positive economic impact on farm families in Vietnam. To meet the increasing
demand for meat in Hanoi, farmers are using this mixture to fatten their pigs for
market. Where once Vietnamese farmers, mostly women, spent up to two hours
every day chopping sweetpotato vines to feed their one or two pigs, a large batch
of the new feed can now be mixed in a few days. Nguyen Thi Ty, a woman farmer
who participated in one of CIP's trials, has 11 piglets that she plans to fatten and
sell. In less than three days she mixed one ton of fermented mix, enough to
provide her animals with feed until they are grown and ready for market.
Two on-farm trials in the Red River Delta area near Hanoi looked at using fermented sweetpotato vines to reduce women's labor and feed processing costs,
improve pig growth efficiency, and develop an easy way to store feed. Twelve
different mixtures of sweetpotato vines, corn and cassava meals, rice bran, sundried chicken manure, and salt were fermented, and the results were analyzed for
hutritional value. Vines fermented with chicken manure had significantly higher
crude protein, dry matter, and ash contents, and lower units costs of each nutrient,
than did the other fermentation products. None of the preparations were found to
contain aflatoxin or Salmonella. E. coli, although present in the original samples,
disappeared after 14-21 days of fermentation.
A three-month on-farm follow-up feeding trial compared pig growth and economic
efficiency of feeding pigs fresh sweetpotato vines, vines fermented with cassava
meal, and vines fermented with sun-dried chicken manure and cassava meal. Pigs
fed the preparation containing chicken manure grew faster and gained more
weight than those fed fresh vines. Neither feed was very different from the vines
fermented with cassava meal, except in terms of feed efficiency. The chicken
manure preparation was also cheaper (cost per kg of weight gain) than the other
two preparations. Read more about this research in the paper by Peters, Nguyen,
and Tran in an upcoming issue of AGRIPPA, the new FAO peer-reviewed journal.

352

Research on Sweetpotato

Natural Resource Management


Scientist and Farmer
Partners in Research far the 21 st Century

Carbofuran Presence in Soil Leachate, Groundwater,


and Surface Water in the Potato Growing Area in
Carchi, Ecuador
R. Jarami110 1 , W. Bowen 2 , and J.J. StoorvogeP

Carbofuran (2,3-dihydro-2,2-dimethyl-7-benzofuranyl methylcarbamate) is a


toxic insecticide widely used on potato (So/anum tuberosum L.) crops in
Carchi Province, Ecuador. Carbofuran can leach to groundwater and become
a threat for both human health and the environment. To determine the extent
of carbofuran presence in the area, we sampled soil leachate, groundwater,
and surface water from within 14 hydrologic units (each 2.5-20 ha) where
potato was being grown from September to November 1999. Carbofuran was
found in all hydrologic units except one, at concentrations that were inversely proportional to the organic matter content of the predominant soil
type. A simulation study showed that the Leaching Estimation and Chemistry
Model (LEACHM) could be a valuable tool for evaluating carbofuran fate,
because it was shown to correctly identify soil types with the greatest risk
for leaching.

The mountainous area of Carchi in the


northern Andes of Ecuador has an intensive, market-oriented potato production
system with high levels of pesticide use.
The single most widely used insecticide is
carbofuran (2,3-dihydro-2,2-dimethyl-7benzofuranyl methylcarbamate), which is
appl ied by growers to control the Andean
potato weevi 1 (Premnotrypes vorax
(Hustache)) (Crissman et al., 1998b). A
potential contaminant of water supplies,
carbofuran can enter groundwater by
leaching and surface water by runoff from
treated fields. At sufficiently high
carbofuran concentrations, the human
health effects due to long-term exposure
include damage to the nervous and
reproductive systems (EPA, 1998).
1

CIP, Quito, Ecuador.


lnternational Fertilizer DevelopmentCenter/CIP, Quito,
Ecuador.
3
Wageningen University, Wageningen, Netherlands.
2

Concerns about the adverse effects of


carbofuran on human health and the
environment led to an analysis of farmer
practices and pesticide use in two Carchi
watersheds during the early 1990s
(Crissman et al., 1998a). This analysis
showed that most of the negative impacts
of pesticide use on human health were due
to pesticide exposure following unsafe
storage, handling, and application.
Although the fate of carbofuran in the
environment was not measured and
monitored directly, modeling studies
indicated that leaching of carbofuran
would be minimal because of the high
organic matter (OM) content of soils in the
region; sorption of the pesticide increases
with increasing OM content (Ducrot et al.,
1998). These results, however, do not
suggest that carbofuran poses no adverse
environmental risks in the potato-growing
areas of Carchi. Rather, they have provided the mpetus for several more

CIP Program Report 1999 - 2000

355

detailed studies intended to better document the environmental fate of carbofuran


through both measurement and simulation.
In this paper, we present the results of one
such study designed to measure the
concentratin of carbofuran in soil
leachate, groundwater, and surface water
in 14 small catchments or hydrologic units
located near the town of San Gabriel in
Carchi. These measurements were taken
during a short period of time, but over a
wide area to determine the possible extent
of carbofuran contamination. The study
also assessed the feasibility of using the
LEACHM (Leaching Estimation and
Chemistry Model) (Hutson and Wagenet,
1992) simulation model to predict
carbofuran movement in the predominant
soil of each hydrologic unit.

Materials and Methods


Carbofuran measu rements
Soil leachate, groundwater, and surface
water samples were taken from 14 hydrologic units in the potato-growing area
around San Gabriel between September

and November 1999 (Table 1). The size of


the hydrologic units varied from 2.5 to 20
ha. Each hydrologic unit contained sloping
potato fields that were being cultivated
during the sampling period. Carbofuran
was applied to potato fields in all but two
units. The predominant soil type in each
hydrologic unit was either a DuriudollEutrandept (Cf), Dystrandept (Dp), or
Eutrandept-Argiudoll (Hf). These types are
distinguished mainly by differences in soil
OM content, which is known to influence
pesticide degradation (Ducrot et al., 1998).
Within each hydrologic unit, sloping
patato fields for which we had information
on the date and amount of carbofuran
application were selected to take leachate
and water samples at the positions illustrated in Figure 1. Soil leachate was
collected from the unsaturated soil below
the potato crop at sites near the top,
middle, and bottom portions of the field
using a suction probe (Rhizon Sampler
from Eijkelkamp Agrisearch Equipment,
http://www.eijkelkamp.com). The suction
probes were installed horizontally in pits
dug between the potato furrows, usually at

Table 1. General description of the selected hydrologic units showing altitude, predominant soil type,
average soil organic matter (OM) content, total area cropped to patato, and amount of the insecticide
carbofuran applied to patato fields.
Soil type 1
Hydrologic
Altitude (m)
OM content
Total area
Patato area
Carbofuran
(%)
(ha)
(ha)
applied 2
unit number
(l/ha)
Cf
7.0
2.0
2.00
1
3070
8
Dp
10
7.0
6.0
2.00
2
3200
Dp
5.0
2.2
0.00
3100
10
3
Dp
10
20.0
4.0
2.00
4
3100
Dp
10
10.0
4.0
4.00
5
3000
Dp
12.0
5.0
3100
10
2.00
6
Dp
10
2.5
1.5
0.00
7
3100
Dp
10
6.0
1.2
2.30
3200
8
Cf
8
10.0
5.0
2.00
9
2900
Dp
8.0
7.0
10
3100
10
2.00
Hf
6
3.0
3.0
11
2770
0.25
12
Hf
6
7.0
6.5
2900
1.60
13
2800
Hf
6
4.5
4.0
4.50
4.0
14
2900
Cf
8
1.5
1.50
1 Soil types from MAG-ORSTOM (1980). "Cf" stands for Duriudoll-Eutrandept, "Dp" for Dystrandept and "Hf" for

Eutrandept-Argiudoll.
2 Amount of commercial formulation (Carbodan, Carbofuran 4f) applied during potato cycle.

356

Natural Resource Management

Top position

Lateral
movement
of water

Surface water
(creek or ditch)

Figure 1. Representation of a sloping field and the relative positions (top, middle, and bottom) from which soil
leachate, groundwater, and surface water samples were taken.
a depth of 80 cm just below potato roots. lf
a sandy or cemented layer was present at
a shallower depth, the probes were placed
above that layer. We were able to extract
sufficient soil solution for carbofuran
analysis during a 16-24 h period. Soil
leachate samples were taken in this way
from a total of 19 fields across ali hydrologic units.
An attempt to collect groundwater samples
was made once during the sampling period
by boring a hole to a depth of 220 cm at
the bottom portian of one field within each
hydrologic unit. lf groundwater was found,
it was carefully pumped out using a
syringe attached to a plastic pipe; in only
7 of the 14 units did we encounter groundwater at a depth of 220 cm or less. Surface
water samples were taken usual ly at least
twice, once near the beginning of the
sampling period and again near the end,
by scooping plastic bottles into water
running through a creek or ditch at the
bottom of each hydrologic unit.
Leachate and water samples were analyzed for the presence of carbofuran using
the Carbofuran RaPID Assay procedure
(Strategic Diagnostic, lnc., http://

www.sdix.com), which is based on the


enzyme-linked immunosorbent assay
(ELISA). The RaPID Assay allows detection
at 0.1-5 parts per billion (ppb) carbofuran.
When carbofuran concentration exceeds 5
ppb, dilutions can be made to measure the
real amount.

Model simulations
To simulate carbofuran movement in soil,
we used the pesticide version (LEACHP) of
the basic LEACHM model (Hutson and
Wagenet, 1992). Simulation runs were
made for each potato field sampled in the
14 hydrologic units. To run the model,
weather data inputs (rainfall and temperature) were obtained from the INAMHI
(Instituto Nacional de Meteorologa e
Hidrologa) meteorological station in
nearby San Gabriel, although for two of
the units rainfall was measured on site.
Soil data inputs (physical and chemical
properties) were estimated in the field
according to functional horizon relationships derived earlier by Kooistra and
Meyles (1997). Carbofuran amounts and
application dates for each field were
obtained from farmer interviews.

CIP Program Report 1999 - 2000

357

Comparative simulation runs were made


using two carbofuran degradation rates.
One simulation used the average degradation rate of 0.0139/d (half-life = 50 d) cited
in the literature (Wauchope et al., 1992);
the other used the rate of 0.0495/d (halflife = 14 d) that was determined
experimentally in a nearby field study
(Stoorvogel et al., 2001 ). Although model
runs were made on a per field basis, the
results were averaged across the three
main soil units for presentation.

Results and Discussion


Carbofuran measu rements

Carbofuran was found in various concentrations throughout all hydrologic units


except unit 1O (Table 2). The highest
concentrations were found in soil
leachate, with the amount being inversely
related to soil OM content. Hydrologic
units with Dp soils (10% OM) consistently
had less carbofuran in soil leachate than
units with Cf (8% OM) and Hf (6% OM)
soils. Leachate from the Hf soils generally
showed the greatest concentration of
carbofuran. Hydrologic units with Hf soils
also tended to have a greater concentra-

tion of carbofuran in groundwater samples,


at least when compared to units where
such samples were taken. In addition to
containing less OM, the Hf soils tended to
be more shal low, hence carbofuran
concentrations wou Id be expected to be
somewhat higher in hydrologic units with
these soils since the carbofuran could be
more easily leached.
Carbofuran was even found in the two
hydrologic units (3 and 7) where farmers
indicated it was not applied to potato
fields during the sampling period (Table 2).
The relatively low concentrations in soil
leachate in unit 3 may be due to applications made to an earlier crop. In unit 7,
carbofuran was found only in the surface
water on the second sampling date. In this
case we are not sure of the source, although it may also come from earlier
applications and the gradual movement of
carbofuran over time. In general, more
carbofuran was found in the second
sample of surface water for most of the
hydrologic units, which may have come
from runoff as well as subsurface flow.
The United States Environmental Protection Agency has defined an enforceable

Table 2. Carbofuran concentrations (parts per billion (ppb)) found in soil leachate, ground water, and
surface water in the fourteen hydrologic units.
Carbofuran in soil leachate
Hydrologic
Soil type
Carbofuran
Carbofuran
Mean
Maximum
Minimum
in ground
in surface
unit number
(ppb)
(ppb)
(ppb)
water1(ppb) water12 (ppb)
Cf
1.00
4.00
0.00
ns
0.10
1
Dp
0.00
0.00
0.19; 0.00
0.00
0.00
2
Dp
0.12;0.00
0.04
0.10
0.00
ns
3
Dp
0.02
0.12
0.00
0.15
0.00
4
Dp
0.26
0.52
0.00; 0.65
0.10
ns
5
Dp
0.00
0.00
0.00
0.00; 1.60
6
ns
Dp
0.00
0.00
0.00
0.00; 2.00
7
ns
Dp
0.63
2.50
8
0.00
0.00
0.00; 0.07
Cf
0.05
0.20
0.00; 0.95
9
0.00
ns
Dp
0.00
0.00
0.00
0.00
ns
10
1.28
11
Hf
1.80
0.27; 0.80
0.58
0.36
Hf
1.82
6.20
0.29; 0.58
12
0.40
0.32
Hf
1.22
13
3.50
0.00
0.11; 0.26
0.75
14
Cf
1.23
3.60
0.00
ns
0.50; 0.46
1

ns = no sample available.
Surface water samples were taken once or twice. Where two samples were taken, the first number is the first sample
result; the second number is the second sample result.

358

Natural Resource Management

standard for the presence of chemicals in


drinking water; it is called the maximum
contaminant level (MCL). For carbofuran,
the MCL has been set at 40 ppb (EPA,
1998). Nevertheless, if carbofuran is found
at levels above 0.9 ppb, the water supplier
must set up a continuous monitoring
system to track carbofuran levels (EPA,
1998). The levels of carbofuran found in
the Carchi hydrologic units were all much
less than 40 ppb, although in a few cases
carbofuran was found at levels considerably higher than the monitoring trigger of
0.9 ppb.
In addition to its negative impact on
human health, carbofuran and other
pesticides can also have adverse effects
on aquatic life. For that reason, performance goals that set the upper limit on
daily concentrations of pesticides permitted in freshwater environments have been
implemented by many state water quality
control boards in the United States, which
is distinct from standards set for drinking
water. For example, the performance goal
for carbofuran in California freshwater
environments has been set at only 0.4 ppb
(Crepeau and Kuivila, 2000). Sorne of the
surface water samples taken from the
Carchi hydrologic units showed that
carbofuran concentrations often exceeded
this performance goal (Table 2).

the observed values, particularly for Dp


soils, although they now underestimated
the amount of carbofuran in soil leachate
for Hf and Cf soils (Figure 2).
These results show that even though the
model did not accurately predict the
amount of carbofuran in soil leachate, it
could still be useful for describing the
relative importance of factors such as soil
OM in controlling pesticide fate. Perhaps,
too, the model results can be improved
with (1) further understanding of
carbofuran degradation in volcanic ash
soils in the high Andes, and (2) on-site
measurement of soil and weather data
needed as model inputs.

Conclusions
Carbofuran was found in soil leachate
groundwater, and surface water sampl~s
taken from sites within 14 hydrologic units
in Carchi. Analysis confirmed a low
degree of pollution for the area that is
below the MCL of 40 ppb, but often above
a level of 0.4 ppb known to adversely
affect sorne aquatic 1ife forms. Des pite the
high OM content of the soils, and evidence for the fairly rapid degradation of
carbofuran in these environments, leaching of carbofuran does occur and it can
make its way into groundwater.

Model simulations
The simulation of carbofuran movement
with LEACHM showed the model to be
extremely sensitive to the degradation
rate, or half-life used for carbofuran. When
the more conservative estimate of pesticide persistence was used (half-life = 50
d), the model predictions of carbofuran in
soil leachate greatly exceeded the amount
measured (Figure 2). Nevertheless, the
simulations did capture the inverse
relationship between soil OM content and
the amount of carbofuran leached. When
the simulation used the half-life of 14 d
that was obtained from the experiment
conducted local ly by Stoorvogel et al.
(2001 ), model predictions were closer to

:e

.t:

W Measured carbofuran

JI

Simulated carbofuran
!,,= 50 days

O Simulated carbofuran
t,.= 14 days

CJ
al

e
e

al

Hf(6%SOM)

Cf(8%SOM)

Dp(10%SOM)

Soiltypes

Figure 2. Measured and simulated concentrations of


carbofuran in soil leachate averaged by soil unit and
soil organic matter (OM) content (Hf: EutrandeptArgiudoll; Cf: Duriudoll-Eutrandept; Dp: Dystrandept).
One simulation assumed a half-life of 50 days, the
other assumed a half-life of 14 days.

CIP Program Report 1999 - 2000

359

Likewise, carbofuran occurs in surface


waters, getting there probably through
runoff. LEACHM can be a valuable tool for
evaluating carbofuran fate since it was
shown to correctly identify soil types with
the greatest risk for leaching.

References
Crepeau, K.L. and K.M. Kuivila. 2000.
Rice pesticide concentrations in the
Colusa Basin Drain and the Sacramento
River, California, 1990-1993. Journal of
Environmental Quality 29:926-935.
Crissman, C.C., J.M. Antle, and S.M.
Capalbo (eds.). 1998a. Economic
environmental and health tradeotfs in
agriculture: Pesticides and the
sustainability of Andean patato
production. Kluwer, Boston, MA, USA.
275 p.
Crissman, C.C., P. Espinosa, C.E.H.
Ducrot, D.C. Cole, and F. Carpio.
1998b. The case study site: Physical,
health and patato farming systems in
Carchi Province. In: Crissman, C.C.,
J.M. Antle, and S.M. Capalbo (eds.).
1998. Economic, environmental and
health tradeoffs in agriculture:
Pesticides and the sustainability of
Andean patato production. Kluwer,
Boston, MA, USA. p. 85-120.
Ducrot, C.E.H., J.L. Hutson, and R.J.
Wagenet. 1998. Describing pesticide
movement in patato production on
Carchi soils. In: Crissman, C.C., J.M.
Antle, and S.M. Capalbo (eds.). 1998.
Economic, environmental and health
tradeoffs in agriculture: Pesticides and
the sustainability of Andean patato
production. Kluwer, Boston, MA, USA.
p. 181-208.
EPA (Environmental Protection Agency).
1998. Drinking water and health.
Available at Environmental Protection

360

Natural Resource Management

Agency web site http://www.epa.gov/


safewater/dwh/c-soc/carbofur.html
(posted 23 Jan. 1998; verified 28 Mar.
2001).
Hutson, J.L. and R.J. Wagenet. 1992.
LEACHM, leaching estimation and
chemistry model: A process-based
model of water and salute movement
transformations, plant uptake and
'
chemical reactions in the unsaturated
zone, Version 3. Research Series No.
92-3, Department of Soil, Crop, and
Atmospheric Sciences, Cornel 1
University, lthaca, NY, USA.
MAG-ORSTOM. 1980. Mapas de suelos:
Sierra. Mapa 1O. San Gabriel. 1:50.000.
Quito, Ecuador.
Kooistra, L. and E. Meyles. 1997. A novel
method to describe spatial spoil
variability: A case study for a potatopasture area in the northern Andes of
Ecuador. lnternational Patato Center
(CIP), Quito, Ecuador, and Dept. of Soil
Science and Geology, Agricultura!
University Wageningen, Netherlands.
Stoorvogel, J.J., R. Jaramillo, R. Merino,
and S. Kosten. 2001. Destino de
plaguicidas, contaminacin en suelos y
agua. In: Crissman, C. and P. Espinosa
(eds.). 2001. Impactos del uso de
plaguicidas en la produccin, salud y
medio ambiente en Carchi: Un
compendio de investigaciones y
respuestas multidisciplinarias. Ediciones
Abya-Yala, Quito, Ecuador. (in press.)
Wauchope, R.O., T.M. Buttler, A.G.
Hornsby, P.M. Augustijn-Beckers, and
J.P. Burt. 1992. The SCS/ARS/CES
pesticide properties database for
environmental decision-making.
Reviews of Environmental
Contamination and Toxicology 123:1155.

Toward A Dynamic Definition of Agroecological


Zones Using Modern lnformation Technology Tools
R. Quiroz 1 , P. Zorogasta1, G. Baigorria 1, C. Barreda1, R. Valdivia 2 , M. Cruz 1,
and J. Reinoso 2

Agroecological zoning (AEZ) is a method that uses biophysical attributes of


the land to cluster land-use types into more homogeneous areas. This exercise
facilitates planning for the sustainable use of natural resources. The application of AEZ is limited by the lack of geospatial data, particularly in
mountainous areas. Remote sensing and process-based models for both climate
interpolation and crop and livestock production were used in a watershed
above 3800 m. With the incorporation of these new tools, AEZ can become a
dynamic and more robust method.

One of the most striking characteristics of


mountains is their spatial variability. This
makes the planning of the use of natural
resources in the mountains more complex
than in any other area. A practica! approach is to subdivide the area of interest
into smaller zones with similar biophysical
attributes. This is the process defined as
agroecological zoning (AEZ) (FAO, 1997).
The AEZ method calls for the use of
biophysical attributes of the land such as
soils characteristics, physiography, climate, land use/land cover, and
productivity (FAO, 1997) as input for
production models. These are constructed
based on expert knowledge of the adaptation of crops to local environments. The
accuracy of any model, therefore, is
directly proportional to the understanding
of uncertainties associated with soil,
climate, and management variations.
Natural resources management in mountainous areas is also characterized by the
lack of precise, reliable, and accessible

1
2

CIP, Lima, Peru.


Centro de Investigaciones en Recursos Naturales y Medio
Ambiente, Puno, Peru.

data. This is further complicated by the


fact that, due to spatial variability, the
requirement for data is more demanding
than for more spatial ly homogeneous
areas.
This paper describes sorne of the methods
used to perform AEZ in data-scarce
mountainous environments. Emphasis is
given to the improvement of the AEZ
method with the use of climate interpolation models, remote sensing (RS), and
process-based biophysical models.

Materials and Methods


The process used in this study is outlined
in the flow diagram in Figure 1. This
section further describes the process.

Location
The llave-Huenque watershed (38255550 m) of the Andean high plateau or
Altiplano was the subject of this case
study. This is one of the most important
watersheds; its effluent drains into Lake
Titicaca, sustaining the lives of thousands
of resource-poor households that depend
on agriculture. The northernmost points in

CIP Program Report 1999 - 2000

361

Photogrammetric
charts

Rain
(GOES)

Contour
lines
Tmax,Tmin
Rainfall
(Weather station)

IDRISI

CU MATE
INTERPOLAR

Tmax,Tmin
Rainfall maps

Landsat-TM

IDRISI

NOAAAVHRR

ENVIJIDRISI

ENVI

,,
Soil maps

Climae
index

,,

Land
use/cover

Biomass

,,

,,

ENVI

AE Zones

Figure 1. Flow Diagram of the method for agroecological zoning (Tmax = maximum temperature, Tmin = minimum
temperature). Abbreviations: ARC/INFO software= (www.esri.com/software), AVHRR = Advanced Very High
Resolution Radiometer (NOAA), DEM = digital elevation model, ENVI = Environment for Visualizing lmages
(Research Systems lnc.-Kodak, Boulder, CO, USA), GOES = Geostationary Operational Environmental Satellite
(NOAA, USA), IDRISI = IDRISI software (http://www.clarklabs.org/), NOAA = National Oceanic and
Atmospheric Administration (USA).
the watersheds are at latitude 161 O' S
and longitude 6930' W and southernmost at latitude 1705' S and longitude
7005' W. The total area comprises about
777,000 ha.

362

Natural Resource Management

Soils
Thirty-three soil classes were used in the
AEZ. In general terms, the soils are mostly
shallow to moderately shallow, low in
organic matter content (less than 4%), and

low in P (< 14 mg/kg soil). There is also a


predominance of acid soils (pH < 6). A
low clay content across the watershed
makes the soils highly vulnerable to water
erosion. This is particularly important in
the lower part of the watershed where
farmers crop even steep slopes because
they are less frost prone (Grace, 1985;
Arguelles and Estrada, 1990).
Topography

Photogrammetric charts 1:100,000 were


digitized and the digital elevation model
(DEM) generated. The slope and altitude
maps were derived from the DEM and
used as input both to the AEZ directly, as
well as for the climate interpolation
procedures.
Land use/land cover

The land use/land cover data were derived


from virtually cloud-free Landsat-TM
imageries taken in 1990 and 1997. The
i mages were geometrical ly corrected
using mapping polynomials (Richards,
1993) with 18 control points from a
1:100,000 map. A mosaic of the imageries
was then constructed. An unsupervised
classification produced a first draft of the
land use/land cover map. These classes
were ground-truthed befare performing a
supervised classification, thus producing
the final land use/land cover map.
Biomass

A functional relationship between monthly


cuttings of green dry matter (OM) and the
Advanced Very High Resolution Radiometer (AVHRR) normalized difference
vegetation index (NDVI) was established.
Three years of data (1995-1997) were used
to derive the equation:
DM = 1.615*NDVI

1 318

goodness-of-fit was assessed by the


following ratio:
SSregress1on
.

R1 =

SStotal

where SS = sum of squares.


Climate

A method based on the geographic


information system (GIS) was used for
spatial climate analysis (Baigorria et al.,
2000a). Raster maps of rainfall and
maximum and minimum temperature were
generated with process-based interpolation
models. These models combine pointmeasures of climate variables and
topograph ic data (si o pe and aspect) as
input to mathematical models that integrate state-of-the-art knowledge of the
physical laws ruling the spatial variability
of climate. The Geostationary Operational
Environmental Satellite (GOES) was used
to determine the movement of clouds for
the rainfall interpolation model (Baigorria
et al., 2000b).
Monthly climate data from four weather
stations were used. The accuracy of the
interpolation was not assessed due to the
lack of independent weather stations.
Outstanding agreement between interpolated and independent weather data in a
similar setting in northern Peru showed the
reliability of the procedure. The climate
maps generated with pixel sizes of around
50 x 50 m for the frost-free season (November - April) were summarized into a
thermal-rain index. The index is directly
related to rainfall and inversely related to
the thermal range (difference between
maximum and minimum temperature):

pp + 1

R2=0.90

where DM = green dry matter (kg fresh wt/


m 2), and NDVI = normalized difference
vegetation index.
The Mardquart procedure of SAS (1988)
was used to derive the parameter of the
nonlinear equation (Rawlings, 1988). The

ltr

[ ((T

=7.1 + Ln

- T.)+ 1)1.66

max

mm

where, ltr = thermal-rain index (O - 14), PP


= rainfall (mm), T
= maximum temperature (C), T . = mi~imum temperature
(C), and L~ : natural logarithm.
1

CIP Program Report 1999 - 2000

363

The constant 7.1 scales the index between


O and 14. The addition of 1 unit to both the
numerator and the denominators is to
guarantee having a real logarithmic
number when either the rainfall or the
thermal range are equal to zero. The
denominator is raised to the power of 1 .66
to match the units of the thermal range
with the rai nfal 1.
AEZ Procedure

Using all the characteristics described


above as layers of a GIS, an unsupervised
classification was run to arrive at the
classes termed AEZs.
Process-based biophysical models

Both crop and 1ivestock models ha ve been


validated for the soil, climate, and management conditions encountered in the
Altiplano. The patato model SUBSTOR
(Bowen et al., 1999), alpaca (Arce et al.,
1994), sheep (Aguilar and Caas, 1991 ),
and llama (Murillo, 2000) are models that
have been tested for the Andes.
A generic method described elsewhere
(Quiroz et al., 2000) might be used to
simulate crop or livestock production,
either for each zone or on a pixel-by-pixel
basis.

Results and Discussion


The land attributes used to assign each
pixel to a cluster or AEZ are divided into
two categories. Category 1 (soi 1 classes,
altitudes, and slopes) includes variables
obtained in a way similar to other AEZ
studies (FAO, 1997). Summary maps of
these attributes are presented in Figure 2.
The remainder of this section is devoted to
the variables of Category 11 (land use/land
cover, biomass, and el imate) and the AEZ
resulting from the exercise.
Land use/land cover

Figure 3 shows a summary of eight land


use/land cover classes. These classes
correspond to the classification system of
Anderson et al. (1976) as modified by
364

Natural Resource Management

Sabins (1997). We list the Category 11


classification that corresponds to image
scales between 1 :80,000 and 1 :125,000:
11 O Residential,
21 O Cropland and pasture (mainly
alfalfa),
310 Grassland,
320 Shrub and brushland,
510/520 Streams + lakes and ponds,
620 Vegetated wetlands or year-round
natural ly irrigated grasslands (bofedal),
730/7 40 Sand and gravel other than
beaches + exposed rocks, and
91 O Perennial snowfields.

Of the 777,000 ha, cropland and pasture


account for 4%; grasslands, 49%; bofedal,
5%; shrubs, 32%; snowfields, 1 %; and
stream, lakes, and ponds combined with
residential, 2%. Cropping areas are
located in the lower part of the watershed.
That is a less frost-prone area closer to
local markets. lt is also a highly populated
area, thus putting the sustainability of the
system at risk. There are a few spots of
cropland and pasture in the middle of the
watershed. These are areas suitable for
pasture, but sorne cereals grown as forage
and bitter potato are also found. lt is in this
part of the watershed that deep gu 11 ies are
evidence of much water erosion.
Grasslands dominate the use of the land
across the watershed. They are grazed
mainly by sheep in the lower part of the
watershed and by camelids (alpacas and
llamas) in the medium and high parts.
Bofedales are used almost exclusively by
camelids, mainly alpacas. The padded
hooves of these animals prevent the
damage done by other ruminants to this
highly productive, highly valued, and
fragile grassland. Bofedales are found
mainly in the higher part of the watershed
that receives water from snowmelt.
Shrubs cover a relatively high proportion
of the land. The dominant species are
Parastrephia lepidophylla and Baccharis
incarum. They grow to an average height
of 2 m, which usually takes around 6 years

Soil series

......

SoH t t rl o

c:J AJo rami entos 'tico s

c::J

...
]

Ajoramitntos 'tlc.os.C ;i1:1cono


~y:1nlM J1ooma ,.o

c:;:;J

Caboono
Cal.3collo Aftor1mit nto s rtioos
CaliJ00Jlt,.C1o hla
C31aoollo Huaitirt:

MI pals
Materia1ts tutaceos

Cilaoollo -M :iloomayo
Csllco lloMltrl:tlu tut>u os

- Pok e

~
~

PoktMaleomayo
Poke Yanam1110
Yanuoo haAlpa (suptricial
c::J Y.tnamayo -1flor.1mltntof ~t loot
Yanamayo -Malcomayo

C31acollo - Poke

c:J

~
c:J

Cabcollo Such u
Cabtoo No To rrtni
Cal3collo -\'tz oaohas
~ Jc o llo

Yanamiyo

Chlno he ros
~iric heros- Ator3miento

Chinchuos lffpa
Chinohtros LIJbo
Cul anel lpa
Hu:altrt: -'Vtzoa chas

Centros pobl;idos

Lagu nu
- H i val

hico

Cllinchtro s Ch tjemooo

Slopes

......

...

,..

...'.

""''"
11

Altitude

......
m.s .n.m .

o~

c:::J

,...,.'

..

,.. '

.,.,

...
'

......

<38 00

3'00 . 31'00
3900. 4000
400 0. 4100
4100 . 4200
4100. 4300
4300 . 4400
4400 4600
4500. 41<!0

...

......

,..

04

4 . 15
16 . 26
25 . 50
>60

......

* . 4700
4900

~
~

470 0

4800 4llOO
4QOO 5000
6000 . 6100
5100. 5200
5200 . 5300
5300 . 6400

Klom t ttfS

20

40

,. $400

'

Figure 2. Soils and topography of llave-Huenque watershed.


(Perez Mercado, 1994). Within the
watershed, shrubs cover th e less ferti le
areas and are commonly found in rocky
soils. Ruminants graze the pastures within
the shrubs and feed on the shrubs during
shortages. Shrub leaves const itute less th an
5% of their diet (Genin et al., 1995).

Our results show that remote sensi ng can


be used to map land use/ land cover in
mountain areas as effectively as in other
settings (Sabins, 1997; Anderson et al.,
1976). New hi gh-resolution satellites w ill
improve our capabi lity for this task even in
areas with h igher fragmentation of land
use than th e one reported here.

CIP Program Report 1999- 2000

365

1- -

...

Land Cover Land Use

.,.

......
C1tegory

J!

~ Ctop land and pasture

Gt1ssland

CJ
8

Shrub and brushl:ind


'Vtgt:t3ttd wetlJnds (bo fe dal)
ftesldtn~I

Sa"'l d graW: I + uposed rocks


~ Streams + laku and ponds

Pttt:nnbl s nowhlds

!
B

,..

...

Thennal - ra in lndex

!!

.......

...... ..11
8
8

8
8

...

ti

,..

...
e!l
,..

1B

Biomass

......

lndex

-2
CJO

E3!
.. CJ5

E3~

CJS

. ': ~

~t '

...

,..

'
" :ro;!:

- ~ 10

~~

..._

......

Kg/H 4.

'1000
1000 . 2000
2000. 3000
CJ3n00 . 4000
'4000 . 6000
6000 . eooo

1
,..

"
) 1000.

...

.......

K$omt: ttrS

20

40

7000

......

Figure 3. Land cover and land use, biomass, and climate of llave-Huenque watershed.

Biomass
Us in g N DVI to estimate stand ing green
biomass proved to be a reliable so urce of
biomass data. There are two sources of
data that co mp lement eac h other. On the
one hand, high-reso luti on biomass maps
ca n be deri ved from Landsat-TM (30- m

366

Natural Resource Management

reso lut ion) or from lko nos (4-m reso luti on).
Th e tradeoff usi ng these data is the hi gher
cost. O n th e other hand, lower reso lution
(1 -km) AVH RR might be used. The
advantage of th is se nsor is its co nt inu ous
and synopti c coverage plu s the
ava il ability of data o n the Internet (USGSEOAAC, 2001).

Both sensors (Landsat-TM and AVHRR)


were used in this study. The AVHRR data
were resampled to generate biomass maps
with the same resolution used for other
input in the AEZ exercise. AVHRR data
was also used to analyze the growth
pattern within a growing season. The map
shown in Figure 3 corresponds to the
extraction of the pixels with highest
biomass in the year. This map is also a
high-level aggregation of the biomass map
into just five classes. Highest production
corresponds to cropland and pasture,
bofedales, and well-managed grasslands
on the lower part of the watershed (18% of
total area). The medium range production
(1.5-2 .5 t/ha) is associated with shrubs and
grassland growing in areas with adequate
moisture, often referred to as temporal
bofedales (49%). A large proportion (32%)
corresponds to low-qual ity bunch grasses
and shrub-bunch grass associations
growing in drier areas with a very low
carrying capacity.
Cloudy skies during the rainy season
constitute a major problem using NDVI to
estimate biomass. This is of particular
importance when satellite data are used to
assess biomass availability for grazing
animals throughout the year. Combining
remotely sensed data with a pasture
growth model (Jongschaap, 2001) has
circumvented this limitation.
The satellite data acquired during clear
days provide the models with the
parameters needed to 'steer' it; i.e., the
data correct the simulation results by
providing the model with actual or
remotely sensed figures of the biomass.
The model in turn fills in the blank spaces
produced by overcast skies. Based on the
experiences in the Altiplano, this synergy
seems to be very useful.
Climate
Both frost and water deficit are common in
llave-Huenque, even within the year-long
growing season. Therefore, the ability to
map areas vulnerable to these abiotic

stresses is an important contribution to


planning the management of the
watershed. When the climatic maps were
summarized into the 11, index, the
following distributions for each quintile
were encountered: 3% for the first, 44%
for the second, 50% for the third, 2% for
the fourth, and 1 % for the fifth. Values of
11, between 8 and 1 O were associated with
cropland and pasture, and sorne grassland.
There also seems to be a good association
between the 1tp value (5 and 6) and the
existence of bush and brushland. For the
other land cover classes, a direct
association was less apparent.

AEZ
Four AEZs were derived from the analysis
(Figure 4). The zone with aptitude for crop
and pasture production comprises 42,000
ha. Using process-based models to
simulate the potential production of this
zone indicates that productivity can be
significantly increased with technologies
to intensify agriculture in the zone. For
example, patato production could be
i ncreased from 5-6 t/ha to 10-12 t/ha
under rainfed cond itions; up to 18 t/ha
with i rrigation.
The second zone corresponds to the area
where livestock can be intensified. lt
comprises roughly 110,000 ha or 14% of
the area. These areas have high carrying
capacity for cattle, sheep, and alpaca. In
the areas near local markets, dual purpose
or dairy production is recommended. In
the areas with bofedales, alpaca
production is a better alternative. Sheep
constitute a flexible buffer alternative that
can be accommodated throughout this
AEZ. Current biomass production of less
than 5 t/ha and low quality might be
increased to more than 8 t/ha of good
qua! ity pasture (alfalfa, ryegrass, and
white clover) . lncrements on the order of
40% to 50% in gross income are feasible.
The third zone was classified as extensive
livestock production. With 51% of the
area, equivalent to 394,000 ha, this zone

CIP Program Report 1999 - 2000

367

......
..~

390000

20000

~0000

IS

Titicaca Lake

~..

1
..

!!

+
AQro ecologiQI zone
~ C1opl1hd Pastu1t

B
~..

lnt.nsive liustoek production


Ext ns lve Uvu toclc productlon
Btt e n nd non pt o du cthl't land

.~

A
+

......

ICiti<d.cs

10

......

20000

SOOOO

20

Figure 4. Agroecological zones of llave-Huenque watershed.


corresponds to subsistence livestock
production based on sheep and ll amas.
Very few tech nol ogica l alternatives have
been tested for this type of production
system. Low cost, externa ! inputs such as
blocks of molasses and urea might be
worth trying.

368

Natural Resource Management

The last AEZ was related to barren land


and areas under grazing with very shallow
soi Is. lt also includes snowfie lds.
Previous exercises on eco logical zoning
that included the study area have been
co ndu cted (ONERN , 1976; ONERNCORPUNO, 1984; INRENA, 1996; Pulgar

Vidal, 1996; Tapia, 1995). lt is difficult to


make a comparative analysis due to the
coarse resolution used or the lack of a
georeferenced map, particu larly with the
last two studies. The present paper uses
higher spatial resolution than any of the
previous studies. The inclusion of quantitative variables for different attributes, on a
pixel basis, provides a substantial refinement over previous results.

Conclusions
The paper described new tools and methods to be incorporated into the AEZ
method. Through the inclusion of the
methods presented, scientists and decision
makers have access to a dynamic tool
for AEZ, even in data-scarce environments. The inclusion of remote sensing in
different parts of the method, together
with process-based climate interpolation models, add robustness to existing
procedu res.
In a practica! sense, several alternatives
have been assessed to improve the management of the natural resources of the
watershed. Since the work was jointly
executed with local professionals, the
chances to positively impact the sustainable management of the watershed are
greatly enhanced.

Acknowledgement
The Peruvian National Natural Resources
lnstitute (INRENA) kindly provided the soil
maps for the study area.

References
Aguilar, C. and R. Caas. 1991.
Produccin ovina para el Altiplano de
Puno, Per. Ciencia e investigacin
agraria 18:1 &2. p. 23-46.
Anderson, J.T., E.T. Hardy, J.T. Roach, and
R.E. Witmer. 1976. A land use and land
cover classification system for use with
remote sensor data: U .S Geological

Arce, B., C. Aguilar, R. Caas, and R.


Quiroz, 1994. A simulation model of an
alpaca system in the dry Puna of the
Andes. Agricultura! Systems 46:205-

225.
Arguelles, L. and R.O. Estrada (eds.). 1990.
Perspectivas de la Investigacin
Pecuaria para el Altiplano. Centro
Internacional de Investigaciones para el
Desarrollo / Proyecto de lnvestigacion
de Sistemas Agropecuarios Andinos.
Lima, Per. 504 p.
Baigorria, G.A., W.T. Bowen, and J.J.
Stoorvogel. 2000a. Climate/Weather
interpolation: A process-based spatial
interpolation model. Proceedings of
Annual meetings of American Society
of Agronomy - Crop Science Society of
America - Soi 1 Science Society of
America (SSA) held 5-9 Nov. 2000. SSA,
Minneapolis, Minnesota. p. 421.
(Abstract.)
Baigorria, G.A., J.J. Stoorvogel, and W.T.
Bowen. 2000b. Spatial-interpolation
rainfall model based on topography and
wind circulation. Proceedings of Annual
meetings of American Society of
Agronomy - Crop Science Society of
America - Soi 1 Science Society of
America (SSA) held 5-9 Nov. 2000. SSA,
Minneapolis, Minnesota. p. 421.
(Abstract.)
Bowen, W., H. Cabrera, V. Barrera, and G.
Baigorria. 1999. Simulating the response
of potato to applied nitrogen. In: lmpact
on a changing world. Program report
1997-98. lnternational Potato Center,
Lima, Peru. 458 p.
FAO (Food and Agricultura! Organization).
1997. Zonificacin agro-ecolgica,
Gua general. Servicio de Recursos,
Manejo y Conservacin de suelos.
Direccin de Fomento de Tierras y
Aguas, FAO. FAO, Rome, ltaly. 82 p.
Genin, D., H.-J. Picht, R. Lizarazu, and T.
Rodrguez. 1995. Waira Pampa: Un
sistema pastoril camlido-ovinos del
Altiplano rido Boliviano. IRD (lnstitut

Survey Professional Paper 964.

de recherche pour le dveloppement

Washington, D.C., USA.

(ex-ORSTOM)), La Paz, Bolivia. 299 p.

CIP Program Report 1999 - 2000

369

Grace, B. 1985. El clima del altiplano,


departamento de Puno, Per. Convenio
Per-Canad-A.C.D.I. INIPA - INIA
(Instituto Nacional de Investigacin
Agraria), Puno, Peru. 183 p.
INRENA (Instituto Nacional De Recursos
Naturales). 1996. Inventario y
Evaluacin de los Recursos Altoandinos
de la Microregin llave-Puno. Ed.
Oficina Nacional de Evaluacion de
Recursos Naturales (ONERN), Lima,
Peru.
Jongschaap, R. 2001. LINPAS model:
Dynamic simulation of rangeland
production in the Altiplano of Peru and
Bolivia. In: Proceedings - Third
lnternational Symposium on Systems
Approaches for Agricultura!
Development. CD-ROM computer file.
International Potato Center, Lima Peru.
Murillo, E. 2000. Estudio de la crianza de
llamas en el Altiplano Boliviano. Un
modelo de simulacin. Pontificia
Universidad Catlica de Chile. 70 p.
ONERN (Oficina Nacional de Evaluacion
de Recursos Naturales) 1976. El mapa
Ecolgico del Per. ONERN, Lima,
Peru.
ONERN (Oficina Nacional de Evaluacion
de Recursos Naturales) and CORPU NO
(Corporacion de Fomento y Promocin
Social y Econmica de Puno). 1984.
Inventario, Evaluacin e Integracin de
los Recursos Naturales de la
Microregin Puno. ONERN, Lima. 237 p.
Perez Mercado, R.M. 1994.
Comportamiento hdrico fisiolgico y

370

Natural Resource Management

fases fenolgicas de la Thola

(Parastrephia lepidophylla y Baccharis


incarum). Ingeniero agrnomo Thesis.
Universidad Tcnica de Oruro, Oruro,
Bolivia.
Pulgar Vida!, J. 1996. Geografa del Per,
Las ocho regiones naturales. 1Oth
Edition. Promocion Editorial Inca S.A.
(PEISA), San Isidro, Lima, Peru.
Quiroz, R., P. Zorogasta, R. Jongschaap,
C. Ibarra, C. Len-Velarde, and M.
Cruz. 2000. Grazing in the Andes with
Remote Sensing. Geoinfo Systems.
http://www.geospati a 1-on 1i ne .com/0800/
apps/apps_ 19.html
Rawlings, J.0. 1988. Applied regression
analysis: A research tool. Wadsworth &
Brook/Cole Statistical/Probability series.
Pacific Grove, CA, USA. 553 p.
Richards, J.A. 1993. Remate sensing
digital image analysis: An introduction.
Second, Revised and Enlarged Edition.
Springer-Verlag, Berlin, Germany. 340 p.
Sabins, F.F. 1997. Remate sensing:
Principies and interpretation, Third
Edition. W. H. Freeman and Company,
NY, USA. 495 p.
Statistic Analysis System lnstitute lnc.
1988. SAS/STAT; User's guide, release
6.03 Edition. Gary, NC, USA. 1028 p.
Tapia, M. 1995. Ecodesarrol lo en los
Andes Altos. Fundacin Friedrich Ebert,
Lima, Peru. 195 p.
USGS-EDAAC. 2001. USGS-EROS Data
center distributed active archive center
(NASA). http://edcdaac.usgs.gov/1 KM/
comp1 Od.html

Estimating the Spatial Variability of Weather in


Mountain Environments
G. Baigorria 1 , W. Bowen 2 , and

J. StoorvogeP

Models of crop and soil systems are useful tools for understanding the complexity of the soil-plant-atmosphere continuum. Their application, however, is
limited by the availability of weather data that drives the processes described
in such models. In this study, we describe a process-based interpolation model
being developed for estimating maximum and minimum temperatures, precipitation, and solar radiation in mountain environments. The model is
parameterized with data obtained from three weather stations set along an
altitudinal gradient of 3020 to 3590 m above sea level in the La Encaada
watershed near Cajamarca, Peru. Using an independent data set from a
fourth station within the same watershed, we show that model estimates for
daily maximum and minimum temperatures agreed well with observed data.
The accuracy of model estimates for precipitation and solar radiation is still
being evaluated.
Plant growth and soil-related processes are
strongly influenced by weather. To simulate these processes accurately, models of
crop and soil systems require weather
data, which typically include daily values
for maximum and minimum temperatures,
rainfall, and solar radiation. Each of
these variables affects, to sorne degree,
processes such as photosynthesis, evapotranspiration, and the rate of plant growth
and development, as wel 1 as soi 1-related
processes that determine water and
nutrient availability.
One of the principal limitations to these
models is the lack of weather data. This is
especially true in tropical mountain
environments where few stations are set up
to record weather data. To overcome this
limitation, and to move towards the useful
application of models in the analysis of
1

CIP, lima, Peru.


IFDC/CIP, Quito, Ecuador.
3
Wageningen University, Wageningen, Netherlands.
2

production systems and natural resource


management in mountain environments,
we have begun to look at methods for
estimating weather data inputs and their
spatial variability. Estimates of the spatial
variability of weather, along with its
temporal variability, are needed to project
spatial variation in model outputs such as
yield and soil loss by erosion. Projections
of spatial variation can be done in simulation studies through the use of a
geographical information system (GIS) that
combines spatial information on the
weather with spatial information on other
factors, such as crop, soil type, slope,
aspect, and elevation.
Methods for estimating weather data at
sites where observations are not available
include extrapolation from one or two
nearby points with observations, or interpolation between points with observations. A
model for extrapolating weather data from
a base station to adjacent mountainous

CIP Program Report 1999 - 2000

371

sites has been developed and evaluated in


the Rocky Mountains of the western USA
(Glassy and Running, 1994). This model,
however, is based on vertical (elevationrelated) corrections to the base station
variables, which may not give correct
results for conditions in tropical mountain
areas like the Andes. lnterpolation
schemes have also been examined in
mountainous areas of the USA and Europe
(Thornton et al., 1997; 2000), but these
have been done with a larger number of
stations placed across a denser network
than one finds in the Andes or other
tropical mountain areas.

values since 1983, when it was set up by


ADEFOR, a local nongovernmental
organization. The Manzanas and La Toma
stations were set up to record dai ly
temperatures and precipitation at the
beginning of 1995. At the end of 1998, we
replaced the temperature-recording charts
and rainfal 1 gauges at al 1 three sites with
automatic recording stations equipped
with a data logger and sensors (Davis
lnstruments, Davis, CA, USA) for measuring daily maximum and minimum
temperatures, rainfall, global solar radiation, relative humidity, wind speed, and
wind direction.

The paucity of weather data in the Andes


led us in 1995 to begin a study to obtain
weather data along an altitudinal gradient
that would be useful for evaluating
different weather-estimation schemes. The
study area is within a benchmark site
establ ished by the Consorcio para el
Desarrollo Sostenible de la Ecorregin
Andina (CONDESAN), located in the La
Encaada watershed near Cajamarca in
northern Peru. Sufficient data have now
been collected to begin the evaluation of
different estimation schemes, including
the development of a model for spatial
interpolation of extreme temperatures and
solar radiation (Baigorria and Bowen,
2001 ). In this paper, we provide a description of weather variability recorded in
the La Encaada study, as wel 1 as a
preliminary analysis of procedures for
estimating daily maximum and mnimum
temperatures.

By the end of 2000, we had 18 years of


daily records for extreme temperatures and
precipitation at the Usnio station, including two years (1999-2000) of daily solar
radiation data. For the other two stations,
Manzanas and La Toma, there were six
years of recorded weather data, including
two years of daily solar radiation data.

Materials and Methods


Since 1995, three weather stations have
been recording daily values for maximum
and minimum temperatures and precipitation along an altitudinal gradient in the La
Encaada watershed: (1) Manzanas (3020
m above sea level), (2) Usnio (3260 m),
and (3) La Toma (3590 m). The greatest
distance between stations is about 7 km,
which is the distance between the
Manzanas and La Toma stations. The
Usnio station has been recording daily

372

Natural Resource Management

The weather data collected in La


Encaada are being used to develop an
interpolation model for estimating maximum and minimum temperatures and solar
radiation in complex terrain. Although
only the general concepts of the model are
described here, more details on a prototype of the model are available in
Baigorria and Bowen (2001 ). Basical ly, the
model, as it has been constructed for La
Encaada, is built upon a GIS that integrates information from the three weather
stations with a digital elevation model
(DEM) containing information on latitude,
elevation, slope, and slope aspect. The
outputs from the model are presented in
raster format with a cell size controlled by
the DEM (30 by 30 m), provided by De la
Cruz et al. (1999).
The model framework for interpolating
temperature extremes and solar radiation
across the landscape is based, not on
statistical interpolation of measured
values, but rather on fundamental relationships between net flux of radiation at the
earth's surface, temperature, land-surface

characteristics, and topography. Temperature extremes are estimated as a function


of net radiation and solar radiation as a
function of temperature and atmospheric
conditions. lnterpolation is used to parameterize the model using measured data
from the three stations. After parameterization, the model then calculates
temperature extremes and solar radiation
for each grid cell, taking into account the
elevation, slope, and aspect, as well as
the effective horizon. The effective
horizon defines how a landform opposite
the given slope affects the number of hours
that the sun can be seen from the slope.

To evaluate the accuracy of the weatherinterpolation model with independent


data, we installed portable automatic
weather stations at five additional sites
within the watershed. These stations,
which were installed during December
1999, record daily temperature extremes
and precipitation but not solar radiation.
They are distributed at elevations varying
from 3250 to 3500 m above sea leve!.
Because all sites provided similar results,
model comparisons for temperature

estimations are shown for only one of the


independent data sites, Calvario (3250 m;
7.0855, 78.343W).

Results
A predominant characteristic of climate in
the La Encaada watershed is illustrated
by the monthly means shown in Table 1.
That is, temperature falls and rainfall
increases with increasing elevation. For
maximum temperatures, the annual
average decreases from 1 6C at Manzanas
(3020 m) to 14C at Usnio (3260 m) to
11 C at La Toma (3590 m). Absolute trends
are smaller for minimum temperatures: the
annual average decreases from 6C at both
Manzanas and Usnio to 3C at La Toma. In
Figure 1, a comparison of daily temperature extremes recorded during 1999 also
shows more pronounced differences for
maximum temperature among sites.
Although the annual distribution of rainfall
varies somewhat, in most months, rainfall
tends to increase with elevation (Table 1).
The relationship with elevation is better
reflected in the average annual rainfall,
which shows Manzanas receiving a total

Table 1. Monthly means far daily maximum and minimum temperatures (TMAX, TMIN), rainfall (RAIN), and
global solar radiation (SRAD) recorded at stations in La Encaada placed along an altitudinal gradient
of 3020 m (Manzanas), 3260 m (Usnio) and 3590 m (La Toma) above sea level.1
Aug
Sep
Mar
Apr
May
Jun
Jul
Oct
Nov
Dec
Variable
Jan
Feb
TMAX (C)
11
12
12
12
11
12
11
10
11
La Toma
10
10
10
14
13
14
15
14
14
15
15
14
15
14
14
Usnio
17
17
15
15
16
16
18
18
16
16
Manzanas
17
16
TMIN (C)
2
4
3
3
3
3
4
3
3
2
3
La Toma
3
5
7
6
5
6
6
7
6
6
6
6
Usnio
6
4
5
7
6
8
6
Manzanas
7
7
7
7
5
4
RAIN (mm)
13
20
57
52
73 133
74
46
56
92 139
99
La Toma
11
75
21
9
40
65
72
70
43
Usnio
92 107 11 o
20
41
50
69
43
17
3
Manzanas
73
36
55 140 105
SRAD (MJ m2/d)
20
19
22
20
20
19
18
16
La Toma
19
15
18
18
22
22
18
15
18
19
19
16
Usnio
20
16
18
17

Manzanas
1 Manzanas

21
17
19
17
21
16
17
19
16
16
(7.118S, 78.310W); Usnio (7.089S, 78.316W ); La Toma (7.062S, 78.282W).

22

CIP Program Report 1999 - 2000

20

373

22

Manzanas

3020m

Usnio

3260m

La Toma

3590m

~
(])

:s

(ti

18

(;
c..

E
:::::
E

14

x:
ctS
~

10

50

100

150

200

250

300

350

Day of year

18
Manzanas

14

3020m

Usnio

3260m

La Toma

3590m

~
(])

:s

10

~
(])

c..

E
:::::
E

x:

ctS

:.?!

50

100

150

200

250

300

350

Day of year

Figure 1. Daily maximum and minimum temperatures recorded during 1999 at three stations located along an
altitudinal gradient in La Encaada, Peru.

37 4

Natural Resource Management

of 652 mm, Usnio 715 mm, and La Toma


854 mm. For the watershed in general,
rainfall is minimal during June, July and
August. Crops grown during these months
need supplemental irrigation or they will
suffer from extreme water deficits.
The monthly means in Table 1 also show
how at each site temperature varies 1ittle
over the year, with the mean for al 1
months being within 3C of each other.
That is, there is little temperature difference between the warmest and the coldest
months. This fairly constant annual cycle
is in contrast to the much larger diurnal
range of temperature (the difference
between daily maximum and minimum
temperatures) shown in Figure 1 for
temperatures recorded during 1999. In this
case, the diurna! range appears to be two
to three times greater than the annual
range of 3C. This is a common characteristic of el imate in equatorial mountain
regions (Sarmiento, 1986).
When considering methods for estimating
weather data in mountain environments,
we thought it appropriate to first look at
the accuracy of simple procedures, such as
the use of a temperature lapse rate, which
is the decrease in air temperature with
increase in elevation (de Scally, 1997). To
obtain the La Encaada lapse rates for
maximum and minirl'Jum temperature, we
regressed annual means on elevation, as
shown in Figure 2. The slope in the given
equations represents the decrease in
temperature for each meter increase in
elevation above a given point. Thus, for
each 100-m increase, the maximum
temperature would be expected to decrease 0.93C, whereas the mnimum
temperature would decrease by 0.53C.
These lapse rates are within the range of
values presented for a diverse set of
studies done worldwide by de Scally
(1997).
Using these lapse rates and temperatures
recorded at the Manzanas station during
the year 2000, we then calculated daily
values for maximum (TMAX) and mini-

18

~---------------.

- ..

TMAX=0.0093x+44.41

R2 = 1

........

....

.....

-- ..

- - .. -

TMIN = -0.0053x + 22.501


R2 =0.8013
3000

3100

3200

3300

3400

3500

..
3600

Elevation (m)

Figure 2. Linear regressions relating changes in


maximum (TMAX) and minimum (TMIN)
temperatures to changes in elevation (slope equals
temperature lapse rate).
mum (TMIN) temperatures for a point at an
elevation of 3250 m (230 m above the
Manzanas station). The calculations were
done simply by subtracting the difference
in temperatures based on the estimated
lapse rates and elevation (TMIN = 1.22C;
TMAX = 2.14C) from daily values recorded at the Manzanas station. The point
being estimated was equivalent in elevation to the Calvario station where weather
data were collected on site during 2000;
hence, these data served as a check on
the accuracy of the estimates. A comparison of the mnimum temperatures
calculated from the lapse rate versus those
measured in Calvario is shown in the top
graph of Figure 3. The estimates of daily
mnimum temperature calculated based
solely on lapse rate tended to be much
less than the actual measured values.
Although there was a significant linear
relationship between calculated and
measured values, the regression line was
far removed from the one-to-one 1i ne.
When the interpolation (simulation) model
described earlier was used to calculate
minimum temperatures, there was much
better agreement between simulated and
measured values (bottom graph, Figure 3).
By taking into account other information
provided by the DEM, specifically, slope,
aspect, and effective horizon for the
Calvario site, the interpolation model

CIP Program Report 1999 - 2000

375

TMIN observed (C)

10
9
y= 1.1256x - 0.8894
R2 =0.7419

8
7

~
"C

:5
E
;

1-

10

TMIN observed (C)

Figure 3. Relationship between minimum temperature (TMIN) observed at the Calvario station (3250 m) and that
estimated using only the temperature lapse rate (top figure) ora simulation model that incorporates the effect of
topographic parameters obtained from a digital elevation model (bottom figure). The broken line is the regression
line and the salid line is the 1:1 line.

376

Natural Resource Management

proved to be more accurate for estimating


daily minimum temperatures.
For estimating daily maximum temperatures, differences between the lapse rate
and interpolation model were less dramatic. In fact, the lapse-rate method
provided slightly better results, although
both methods provided good approxi mations to the one-to-one line, with similar
scatter. The closeness of fit by the two
methods is not shown here, but it can be
illustrated by linear regression: whereas
the equation was y = 0.8954x + 2.94
(R2 = 0.8619) when only elevation and the
temperature lapse rate were used, it was
y= 0.7693x + 3.25 (R2 = 0.8261) using the
interpolation model linked to a DEM.

Discussion
A preliminary analysis of the interpolation
model being developed for La Encaada
indicates that it can be successfully
applied to estimating the spatial variability of daily maximum and minimum
temperatures within the watershed. Further
analysis should focus on evaluating daily
estimates for rainfall and solar radiation.
The interpolation model described here
has recently been expanded to include
spatial estimates of rainfall based on
topography and wind circulation (Baigorria
et al., 2000).
A critica! question for future work is how
accurate such an interpolation model
might be for other watersheds where only
one or two weather stations exist, or where
a few weather stations are spread across a
larger region. In an attempt to address
these questions, we have formed a partnership with the Peruvian lnstitute for
Agrometeorology and Hydrology
(SENAMHI) that includes an evaluation of
the interpolation model using weatherstation data from throughout Cajamarca
Province.
Because few watersheds in the Andes have
even one weather station, another approach we are investigating is the use of
remote-sensing techniques to estimate

weather data. For example, Diak et al.


(2000) have validated a fairly simple
process for obtaining real-time estimates of
downwelling longwave radiation and
incident solar radiation using data from
geostationary operational environmental
satellites (GOES). Dubayah and Loechel
(1997) also used GOES data to estimate
solar radiation and demonstrated that such
estimates could also be made for complex
terrain when GOES data are linked to a
DEM. In the long run, satellite-based
estimates may provide the most accurate
and cost-effective approximations of
weather data for mountain environments.

References
Baigorria, G.A. and W.T. Bowen. 2001. A
process-based model for spatial interpolation of extreme temperatures and solar
radiation. In: Proceedings of the Third
lnternational Symposium on Systems
Approaches for Agricultura! Development [CD-ROM computer file]. CIP,
Lima, Peru.
Baigorria, G.A., J.J. Stoorvogel, and W.T.
Bowen. 2000. Spatial-interpolation
rainfall model based on topography and
wind circulation. In: 2000 agronomy
abstracts. ASA/CSSA/SSSA, Madison,
WI, USA. p. 421.
De la Cruz, J., P. Zorogasta, and R.J.
Hijmans. 1999. A digital atlas of natural
resources in Cajamarca. Production
Systems and Natural Resources Management Department Working Paper
No. 2. CIP, Lima, Peru. 49 p.
de Scally, F.A. 1997. Deriving lapse rates
of slope air temperature for meltwater
runoff modeling in subtropical mountains: An example from the Punjab
Himalaya, Pakistan. Mountain Research
and Development 17:353-362.
Diak, G.R., W.L. Bland, J.R. Mecikalski,
and M.C. Anderson. 2000. Satellitebased estimates of longwave radiation
for agricultura! applications. Agricultura! and Forest Meteorology
103:349-355.
Glassy, J.M. and S.W. Running. 1994.
Validating diurnal climatology logic of
CIP Program Report 1999 - 2000

377

the MT-CLIM model across a climatic


gradient in Oregon. Ecological Applications 4:248-257.
Sarmiento, G. 1996. Ecologically crucial
features of climate in high tropical
mountains. In: Vuilleumier, F. and M.
Monas (eds.). High altitude tropical
biogeography. Oxford University Press,
Oxford, UK. p. 111-145.
Thornton, P.E., S.W. Running, and M.A.
White. 1997. Generating surfaces of

378

Natural Resource Management

daily meteorological variables over


large regions of complex terrain. Journal
of Hydrology 190:214-251.
Thornton, P.E., H. Hasenauer, and M.A.
White. 2000. Simultaneous estimation
of daily solar radiation and humidity
from observed temperature and precipitation: An application over complex
terrain in Austria. Agricultura! and
Forest Meteorology 104:255-271.

Andean Roots and Tubers and


other Crops
Scientist and Farmer
Partners in Research far the 21 st Century

Effect of Viruses UMV, UVC, PapMV-U, and PLRV on


Ulluco Production and Their Control
C. Lizrraga, 1 M. Santa Cruz, 1 G. Lpez, 2 and S. Fuentes1

Eight viruses infect ulluco (Vllucus tuberosus Caldus), an Andean tuber crop.
Three of them, potato leafroll virus (PLRV), Andean potato latent virus, and
potato virus T, also infect potato (Solanum tuberosum L.). Arracacha virus A
and papaya mosaic virus, ulluco isolate (PapMV-U) infect other crops. PLRV,
one of the most damaging potato viruses, was detected in healthy potato
plants growing in the field next to PLRV-infected ulluco plants, indicating that
viruses can be disseminated among different Andean tuber crops under
natural conditions. The viruses with the highest incidence were PapMV-U and
three others, ullucus virus C, ullucus mild mottle virus, and ullucus mosaic
virus. In two field experiments in Junn, Peru, ulluco plants of accession MH290 infected with ullucus virus C or ullucus mosaic virus, and ulluco plants of
native variety Jaspeado infected with PLRV or PapMV-U yielded about 30%
less than the healthy control plants. Also, during field exposures, only a low
percentage of healthy plants were reinfected with viruses. These results
indicate that the production and use of virus-free ulluco seed tubers is justified in this environment and, therefore, the use of better seed by farmers has
been promoted, particularly in the La Libertad community in Junn, Huancayo
Department.

Ulluco follows potato as the second most


cultivated tuber in the Andean region, and
as a popular food for more than 1000 years
(Hodge, 1951 ). lt is planted as a monocrop
or in association with other Andean tubers
such as potato, oca (Oxals tuberosa), and
mashua (Tropaeolum tuberosum) (Tapia,
1992). Both potato and ulluco are economically important for farmers. The main
ulluco-producing area in Peru is Junn
Department (3500 to 3800 m), located in
the central Peruvian highlands (OIA,
2001 ). Ulluco is the crop with the highest
commercial value in the La Libertad
community in Junn.
1
2

CIP, Lima, Peru.


Universidad Nacional del Centro del Peru, Huancayo, Peru.

Four viruses, ullucus virus C (UVC),


ullucus mild mottle virus (UMMV), ullucus
mosaic virus (UMV), and papaya mosaic
virus, ulluco isolate (PapMV-U), were
detected before 1993 in ulluco plants from
Peru and Bolivia (Brunt et al., 1982).
Toledo et al. (1994) also reported the same
viruses infecting in vitro ulluco accessions
of Peru and Ecuador from the germplasm
of the Universidad Nacional Mayor de San
Marcos (UNMSM), Lima, Peru. Viruses are
the main cause of degeneration (yield
reduction) of vegetatively propagated
crops such as potato (Salazar, 1996). Virus
control is based on two principal measures: genetic resistance and the use of
healthy (virus-free) seed (Salazar, 1996).
Farmers produce their own ulluco seed
CIP Program Report 1999 - 2000

381

tubers, generation by generation, thus


permitting the accumulation of viruses and
degeneration of the crop. Because no work
on genetic resistance has been done in
ulluco, the production of better seed is the
measure most 1ikely to control viruses and
increase productivity at the farm level.
Understanding the viruses, e.g., their
detection and importance, is the first step
in producing better seed.This paper presents an overview of the
relevant research done in ulluco since
1993. ldentifying the viruses in ulluco,
developing reliable detection methods,
and eliminating viruses using thermotherapy and meristem-tip culture were
accomplished first. This research was
followed by determining which viruses are
important in Peru based on their distribution, incidence, and effect on yield. After
this was accomplished, the use of better
seed by farmers was promoted, with
emphasis in the La Libertad community.
lmplications of these results on the epidemiology of potato viruses infecting Andean
tuber crops are also discussed. Parts of this
work have been published elsewhere
(Lizrraga et al., 1997; Lizrraga et al.,
1999; Villavicencio, 1 999).

Materials and Methods


Virus identification
Accessions from CIP's in vitro collection
held in trust, and plants from farmers'
fields in Peru (see Table 1) were evaluated
for virus infection by double antibody
sandwich, enzyme-linked immunosorbent
assay (DAS-ELISA). Antibodies to detect
PVT were supplied by the Scottish Agricultura! Science Agency. For PLRV detection,
a commercial conjugated enzyme (lgGAP) (BIOREBA, Longmont, CO, USA) was
used. Rothamstead Experimental Station
initially supplied antiserum to AVA. Antibodies for other viruses were prepared at
CIP.
The viruses used as positive controls were
isolated from ulluco accessions and

382

Andean Roots and Tubers and other Crops

maintained in plant species as indicated in


Table 2.

Virus eradication and production of


healthy planting materials
Ulluco line Jaspeado was cleaned up
through thermotherapy (40C and 25C, for
6 h each, two cycles/d for 25 days) and
meristem-tip culture (on 4.6 g MS basic
salt medium (Murashige and Skoog, 1962)
containing 2% sucrose, 2 ppm calcium
pantothenate, 0.5 ppm gibberellic acid,
and 0.7% agar, pH 5.6). Virus elimination
~as confirmed by both DAS-ELISA and sap
inoculation to indicator plants. For
micropropagation, plantlets were maintained between 18 and 20C, 2000 lx
intensity and illuminated for 16 h/day. Six
months later, cleaned up plantlets (classified as virus-free) were established in a
steril ized substrate in an insect-proof
screenhouse before transplanting in the
field. Certified seed (third generation) for
promotion and distribution to farmers was
~roduced from basic seed (second generat1on); both were field-grown.

Reinfection of virus-free ulluco plants


Jaspeado virus-free seed tubers were given
to three farmers in La Libertad in 1996
and their fields were tested for virus '
reinfection for three growing seasons. Leaf
samples from 80 plants/field selected at
random were collected and evaluated by
~AS-ELISA. Plants and tubers belonging to
d1fferent seed categories from seedproducing fields in La Libertad during the
1997/1998 and 1999/2000 growing seasons
were also evaluated.

To evaluate transmission of PLRV under


natural conditions during the 1999/2000
growing season, tubers of healthy potato
varieties Canchn, Revolucin, and
Perricholi (20 plants/cultivar) were planted
next to ulluco Jaspeado infected with
PLRV in La Libertad. Foliage and tubers
from potato and ulluco plants were tested
for virus infection by DAS-ELISA.

Table 1. lncidence of ulluco viruses in accessions from different countries and from departments of Peru
maintained at CIP's in vitro germplasm collection and in plants from farmers, fields (Junn and
Huancavelica, Peru).

Samples
from:
In vitro
collection

Field 2

lnfected {%}1
Accessions
uve UMV PapMV-U UMMV PLRV APLV
{No.}
Argentina
39
76
45
90
66
21
59
Bolivia
77
66
46
62
22
52
22
4
Colombia
90
60
90
90
30
90
Ecuador
5
39
26
50
90
o 26
Peru
258
74
61
65
39
29
35
Apurimac
4
30
o 60
o 30
90
o
Amazonas
3
o
o
o
35
35
Ancash
25
90
66
39
31
27
66
Ayacucho
8
45
o
o 21
60
38
68
33
16
30
Cajamarca
63
76
61
o o
o
o 45
o
Cerro de Paseo
2
82
93
65
73
36
41
43
Cusca
Huanuco
2
o 90
o
o
o
90
o o
o
Junn
1
o
o
o
o
La Libertad
6
65
45
o 35
45
o 22 32
7
68
57
32
Lima
6
90
o 35
90
90
Piura
90
74
Puno
38
77
74
61
25
33
Total
383
72
65
40
31
{%}
57
37
Country
Department

Peru
Ju nin
Chicche
S.Juan deJarpa
Huaracayo
Huancavelica
Pazos

Total
{%}

PVT
9
16

AVA

15

14

o
o
o o
o o

o o
o o
o 12
o
21
o o
o o
o
13
o o
o o
o o
o o
o o
o
33

42
43
41

68
62
56

65
62

51
49
51

o
o

21
24
17

nt 3
nt
nt

o
o
o

72

57

43

58

24

nt

67

52

42

61

22

180
180
180

72

180
720

Note: See Table 2 for full names of viruses.


1 The arcsin ../ percentage transformation (Steel and Torrie, 1980).
2 Data from Villavicencio, 1999.
3 nt = not tested.

Effect of viruses on ulluco production


Cropping season 1995/96. Tubers of
accession MH-290 infected by viruses
UMV, UVC, PapMV-U, and
UMV+UVC+PapMV-U were planted in a
randomized block design with five
treatments (including a healthy control)
and four replications of 132 plants/plot in
Huancayo, Junn (3280 m). Carlos Arbizu
(CIP) provided the virus-free MH-290
accession (from Bolivia).

Cropping season 1998/99. Tubers of


variety Jaspeado infected by viruses UMV,
UVC, PapMV, and PLRV were planted in
La Libertad (3500 m), following the same
experimental design as before, but with
40 plants/plot.
In both cases, fol iage growth (height) of
plants was measured 5 months after
planting. Only the inner rows were
harvested and the yield recorded. Autoinfection (percentage of virus-infected

CIP Program Report 1999- 2000

383

Table 2. List of viruses infecting ulluco.


Virus isolated from
Genus
Virus
accession (origin)/
Virus maintained in 2
1
lnstitution
Comovirus UVC (Ulluco Virus C) UH 009 (UNMSM)/
Chenopodium quinoa
Huancayo - Junn
Willd .
Potyvirus
UMV (Ulluco Mosaic
UH 009 (UNMSM)/
Nicotiana benthamiana
Huancayo - Junn
Virus)
Domin .
Potexvirus
PapMV-U (Papaya
UH-009 (UNMSM)/
C. mura/e L.
Mosaic Virus, ulluco
Huancayo - Junn
isolated)
Tobamovirus UMMV3 (Ullucus Mild U-016-83 (CIP)/Cerro N. clevelandii Gray x N.
Mottle Virus)
bigelovii (tarr) S. Wats.
de Paseo
Luteovirus
PLRV (Potato Leafroll MH-290 (CIP)/
Ul/ucus tuberosus
Virus)
Huancayo - Junn 4
(MH-290)
Tymovirus
APLV3 (Andean Potato UP-271 (UNMSM)/
N. clevelandii x N.
Latent Virus)
Puno
bigelovii
Trichovirus pVf3 (Potato Virus T) MH-463 (CIP)/Cusco
C. quinoa
Nepovirus
AVA 3 (Arracacha
UP-254 (UNMSM)/
C. quinoa
Virus A)
Puno

lnfecting other crops

Oca, mashua

Potato
Po tato
Potato, oca, mashua
Arracacha

Germplasm collection from UNMSM (Universidad Nacional Mayor de San Marcos, Lima, Peru) and CIP (lnternational
Patato Center, Lima, Peru).
2 Ali virus isolates, except PLR\/, were mechanically transmitted.
3 Virus transmitted by true seed repo rted in other hosts, but not yet confirmed in ulluco.
4 PLRV was isolated from the naturally infected accession and maintained in the same plant.

progeny tubers from a virus-infected plant)


was determined by DAS-ELISA testing of
sp routs on 1O tubers/p lant from se lected
plants of each treatment. Data were
analyzed using MSTAT or SAS softwa re
(SAS 1989) .

To compare th e behavior of Jaspeado basic


seed with infected seed, tubers we re
planted in a randomi zed block des ign with
four treatments (basic, infected, from
positive selection , and from farm er) and
with four replications of 40 plant eac h in
La Libertad during the 1997/98 growin g
sea son. This experiment was repeated at
the sa me site during th e 1998/99 growi ng
seaso n, except that the treatments were
virus-free tubers, ce rtifi ed seed, and
infected tubers (one tuber of 25 g or fi ve,
5-g tubers planted together).

were found infecting the crop during these


studi es.
Over 90% of the plants eva luated, both
from th e in vitro co ll ection and farm ers '
fields, were infected with viruses .
Complex infections (two or more viruses)
were common. Viruses with the hi ghest
in cidence and distribution were uve,
UMV, PapMV-U , and UMMV (Tab le 1).

Results
Virus identification and viral incidence
Ei ght viruses infect ullu co (Table 2). Of
th ese, viruses PLRV, APLV, PVT, and AVA

384

Andean Roots and Tubers and other Crops

Figure 1. Ulluco plant infected with complex of


viruses, showing mosaic and growth reduction
(right).

Only plants severely affected by


compl exes of viruses showed evident
symptoms in the field (Figure 1).

Virus eradication and reinfection of virusfree ulluco plants


Th ermotherapy, alternating between 40(
and 25( in two cycles daily, eliminated
viruses from meristems in Jaspeado plants
infected with up to three viruses. UMMV
and PapMV-U were the most difficult to
eliminate. The varieties Canario, Tarmea,
and Picado de Pul ga were cleaned of
viruses following the same procedure.
Th ermotherapy with a continuous
temperature of 38-40( is less efficient
because ulluco plants become extremely
stressed, causing elevated plant mortality.

field exposure, when PLRV infection was


over 50% (Tab le 3). Ulluco plants from
basic seed (two field exposures) and
certified seed (three field exposures),
analyzed during th e 1997/98 cropping
season , were approximately 40% infected
with PLRV. But PLRV was not detected in
certified seed tubers tested in 1999 (Table
4). Most were infected with APLV, UMV,
UVC, and UMMV (Table 4).
PLRV was detected by DAS-ELISA in sorne
of the healthy potato plants growing next
to PLRV-infected ulluco (two plants eac h
of potato cu ltivars Canchn and
Revolu ci n), as well as in healthy ulluco
plants (5 of 20 control plants) . However,
sprouts of the harvested potato tubers
tested negative for PLRV.

The reinfection rate of virus-free ullu co


was low in farmers' fields until the third

Table 3. Reinfection (%) 1 of virus-free ulluco Jaspeado plants in farmers' fields in La Libertad, Junn, Peru
(3500 m).
Field

Planting season

1996-1997
Farmer 1
Farmer 2

Farmer 3

PLRV
APLV
UMV
uve

1.3
1.3
1.3
1.3

2.5
2.5
2.5
2.5

PLRV

10.0 6.6

PLRV

2.5 3.5

1997-1998

1998-19992

Not planted

Not planted

UMV
uve
PapMV-U
UMMV
PLRV
APLV
Not

5
2.5
1.3
2.5
11
2.5

4.8
3.5
2.5
3.5
7.0
3.5

UMV
uve
PapMV-U
UMMV
PLRV
APLV

~lanted

27.5 9.8
12.5 7.2
5.0 4.8
13.8 7.6
51 .3 11 .0
13.8 7.6
Not

~lanted

Note: See Table 2 far full names of viruses.


1 confidence limits (p = 0.05).
2 Plants were evaluated far all ulluco viruses, but only those detected are indicated in the table.

Table 4. Reinfection (%) 1 of Jaspeado ulluco plants and seed tubers from virus-free ulluco produced in field
in La Libertad, Junn , Peru (3500 m).
Samples 2 Categor'f Evaluated
UVC
UMV PapMV-U
UMMV
PLRV
APLV PVT AVA
Plants
Virus-free
48
O
O
O
O
o
o nt 4 O
Basic
105
43.7
43.7
O 43.7 41 9.5 11 6.0 nt
O
eertified
58
128.5 128.5
34.5
97.5 4813.1 2811 .8 nt
O
Tubers
eertified
49
3313.4 4514.2
O 2712.7
o 5714.1 o o
Note: See Table 2 far full names of viruses.
1 confidence limits (p = 0.05).
2

Plants and tubers from cropping season 1997-1998 and 1999-2000, respectively.
Viru s-free = 1st generation in screenhouse; Basic = 2nd generation in field; Certified
4 nt = not tested .

= 3rd generation in field.

CIP Program Report 1999- 2000

385

Effect of viruses on ulluco production


The onl y symptoms associated with viru s
infection observed during both cropp in g
seaso ns was a temporal mosaic in ullu co
pl ants infected with UMV and PapMV-U.
Viruses UMV or uve significantly red uced
y ield of MH-290; as d id PLRV and PapMVU on Jaspeado plants, (Tab le 4 and Figure
2). Growth (height) was also signifi ca ntl y
di fferent in MH-290, but not in Jaspeado
(data not show n). Autoinfection was clase
to 100% for UMV, uve, and PapMV-U in
accessio n MH-290, but was lowe r (88% for
UM V, 22% for uve, and 20% for PLRV) in
Jaspeado (data not shown).
Pl ants from infected seed- eith er naturally
or artifi ciall y-from field expe rim ents,
farmers, or pos iti ve se lection from
apparentl y hea lth y plants yie lded fewer
com merc ial-qu ality tubers than those from
virus-free, basic, or certifi ed seed (Tab les 5
and 6), alth ough th e differences were not
always stat isti ca ll y signi fica nt. Th e
prod uctivity of plants from seed tubers of
different sizes (five tubers of 5 g vs 1 tuber
of 25 g) was simil ar (Tab le 6).
A total of 1140 kg of hi gh quality Jaspeado
seed tubers were di stributed among 12
farm ers from Pazos, e hi cche, La
Espera nza, and La Libertad during 19961999 . Over 80% of the fa rmers who
rece ived promotional seed showed interest
in culti vatin g and propagatin g it to
in crease their inco me. Sorne farmers even
pl anted co mmercia l fields (ave rag in g

13.9 kg

around 1.0 ha). Data is being co ll ected to


eva lu ate th e benefit.

Discussion
O ur obse rv ati ons and those from To ledo et
al. (199 4) indi ca te th at v iruses u ve,
PapMV-U, UMV, and UMMV are
w idespread in ulluco (Tab le 1). Hi gh viral
in cidence has also been repo rted in
Bo li v ia and Ecuado r (D uqu e and H ermann ,
1994; Badani et al., 1997). Ei ght viru ses
have bee n fo und in fec tin g ullu co . Viral
incidence data of PLRV, APLV, PVT, and
AVA were not ava il ab le befare this stud y.
Th e info rm atio n co mpil ed here on th e
identificat ion and distributi on of viru ses in
ullu co permits recog nizin g dissemin atin g
viru ses and those restri cted to ce rtain
geographi ca l areas (Tab le 1). lt is of
ep id emio log ica l interest that PLRV, A PLV,
PVT, and AVA naturall y infect other
Andean crops, espec iall y patato (Tab le 2)
(Janes and Kenten, 1978; Li z rraga et al. ,
1997; Li z rraga et al. , 2000). Farm ers'
trad itio nal croppin g systems (ulluco and
potato in mi xed crop ping, ullu co fields
beside patato fields, or ulluco pl anted in
fields w here patato was prev iou sly grown )
favor vi rus dissemination betwee n crops.
Th at could be occurring with PLRV, one of
the most damag ing patato viru ses, and thi s
study suggests th at, in Peru, potato seedproducing fi eld s should be far from both
patato and ullu co f ield s. Under expe rim ental co nditi ons, PLRV was transmitted by
th e gree n peach ap hid (Myzus persicae)

19.8 kg

14.1 kg

Figure 2. Effect of viruses on yield of ulluco Jaspeado in secondary infection with PLRV and PapMV-U. Tubers
harvested from 20 plants per each treatment.

386

Andean Roots and Tubers and other Crops

Table 5. Average yield of tubers (kg) from ulluco plants (MH-290 and Jaspeado) with secondary infection
(seed tubers were virus infected), planted in two different cropping seasons and places in Junn, Peru.
Cropping season,
Treatment
Yield 1
Yield
place and variety
Total weight
Commercial quality
reduction (%) 2
1995-1996
Healthy control
26.9 a
16.0 a
(Huancayo)
UMV
19.1 be
10.5 be
29
27
MH-290
uve
19.7 be
10.7 be
PapMV-U
24.2 a
14.5 ab
10
UMV +uve+ PapMV-U
16.7 be
10.5 be
38
Healthy control
19.8 a
nt 3
1998-1999
UMV
16.4 a
nt
17
(La Libertad)
uve
19.5 a
nt
2
Jaspeado
PapMV-U
14.1 b
nt
29
PLRV
13.9 b
nt
30
Note: See Table 2 for full names of viruses.
1 Means within columns followed by the same letter do not differ significantly (p = 0.05). lnner rows of 66 and 20
plants for MH-290 and Jaspeado, respectively.
2 Compared with total yield of healthy control.
3 nt = not tested.

Table 6. Average yield of tubers (kg) from ulluco Jaspeado, planted in La Libertad, Junn, (cropping
seas ons 1997-98 and 1998-99).
Yield 1
Cropping
Treatment
Yield
reduction (%)2
sea son
Total weight
Commercial quality
1997-98
Basic
18.90 a
13.19 a
lnfected 3
14.11 ab
10.27 ab
22
From farmer
13.35 b
9.39 b
29
Positive selection
9.86 b
7.86 b
40
1998-99
Virus-free (5 g)
37.27 a
28.96 a
Virus-free (25 g)
36.79 a
25.09 ab
13
eertified (5 g)
35.66 a
19.62 e
32
eertified (25 g)
34.09 a
20.95 be
28
lnfected (5 g)
36.80 a
19.52 e
33
lnfected (25 g)
31.98 a
18.74 e
35
1 Means within columns followed by the same letter do not differ significantly (p = 0.05). Plot of 40 plants.
2 Compared with commercial quality yield of healthy control.
3 lnfected with at least one of following viruses: UMV, UVC, PapMV-U, and UMMV.

from ulluco to potato and vice versa


(Lizrraga et al, 1997). A similar situation
might also occur with APLV in potato, and
with PapMV-U, AVA, and PVT in other
Andean crops. Until virus dissemination
under these conditions is better understood, precautions shou Id be taken when
the germplasm col lection of several
Andean crops are planted in the same
fiel d.
The field experiment carried out at two
sites with two genotypes showed that
viruses can cause significant yield reduc-

tion in ulluco when infected seed tubers


are used (secondary viral infection) (Table
5). In both experiments, no yield reduction
occurred when plants were artificially
infected (primary infection) during their
growing period (data not shown). In other
ex peri ments with one of the genotypes,
plants from virus-free tubers also yielded
more than those from infected seed, and
plants from basic seed (one field multiplication) yielded more than plants from seed
tubers from farmers or positive selection
(Table 6). During the field experiments,
ulluco plants infected with only one virus

CIP Program Report 1999 - 2000

387

did not show symptoms and were more


vigorous than those from farmers' fields
nearby that were heavily infected by virus
complexes. Th is demonstrates the risk of
using seed from positive selection.

environment interactions. For example,


MH-290 is from Bolivia, Jaspeado is
adapted to Junn, the virus isolates used in
the field experiments were from Junn, and
the experiments took place in Junn.

Virus control in ulluco can be achieved by


eradication, and by using thermotherapy
and meristem-tip culture to regenerate
healthy plants. Stone (1982) used a
combination of chemotherapy and meristem-tip culture to eradicate four viruses
(UMV, UVe, PapMV-U, and UMMV). The
eradication of viruses from ulluco using
thermotherapy is more difficult than from
potato, because ulluco does not tolerate
high temperatures (40-42e) for long
periods. In general, potyviruses are more
readily eliminated from their hosts by
meristem-tip culture than potexviruses and
tobamoviruses in sorne hosts (Stone, 1982).
That could explain why PapMV-U and
UMMV were the most difficult to eliminate from infected plants.

Conclusions

UMV and potato viruses PLRV and APLV


were most frequently found reinfecting
virus-free ulluco plants in La Libertad
(Tables 3 and 4). PLRV and UMV are both
aphid transmitted. This result is contrary to
that obtained in studies of reinfection of
virus-free potato plants in a nearby location, where infection by aphid-transmitted
viruses (PLRV and PVY) was low
(Bertschinger, 1992). Although PLRV was
detected in virus-free potato plants growing next to PLRV-infected ulluco, the virus
was not detected in the tubers. lt is
possible that PLRV infects potato plants
late in the growing season and the virus
moves slowly from plant to tubers. Or
PLRV from ulluco may be a different strain
requiring a longer time to translocate to
tubers as it is adapting to the host.
Lizrraga et al. (1996) observed that
ulluco becomes infected by PLRV four
months after graft-inoculation and that the
virus has an irregular distribution in the
ulluco plant. The differences in the
autoinfection of accession MH-290 and
Jaspeado cou Id be due to host-pathogen-

388

Andean Roots and Tubers and other Crops

Knowledge of the viruses infecting ulluco,


reliable detection methods, and the
eradication of viruses using thermotherapy
and meristem-tip culture resulted in the
production of healthy ulluco plants of
Jaspeado, the variety preferred by Junn
farmers. The main importance of this work
is demonstrated by greater productivity of
better quality seed provided to farmers in
La Libertad, and their increasing acceptance and demand for it. This work also
provides the groundwork for organizations
engaged in cleaning up ulluco to produce
better seed and to facilitate the potential
interchange of valuable virus-free
germplasm.
lt has been confirmed that ulluco has a
high incidence of viral infection, with
frequent mixed viral infections. And it has
been demonstrated that UMV, uve,
PapMV-U, and PLRV can lower yield in
the ulluco crop in Peru. Finally, this
research supports the premise that viruses
(e.g., PLRV) infecting one plant species
can eventually adapt in their ability to
infect other species growing in close
association for extended periods.

Acknowledgments
This work is part of the collaborative
program Biodiversity of Andean Root and
Tuber erops funded by the Swiss Agency
for Technical eooperation. Thanks to
R. Burns of the Scottish Agricultura!
Science Agency, Edinburgh, Scotland, UK,
for supplying PVT antibodies, and R.
Woods from Rothamsted Ex peri mental
Station, Harpenden, U K, for AVA antiseru m. Sorne of the tables, with additional
information, have been translated from
Spanish from Lizrraga et al. (1999) with
the kind permission of Fitopatologa.

References
Badani, A.C., S. Conzales, C. Plata, and
E.N. Fernandez-Northcote. 1997.
Incidencia de virus en papalisa (Ul/ucus
tuberosus Loz) en Cochabamba, Bolivia.
In: IX Congreso Internacional de
Cultivos Andinos, Libro de resmenes.
Universidad Nacional de San Antonio
de Abad del Cusco (UNSAAC), Centro
de Investigaciones en Cultivos Andinos
(CICA), and Asociacin ARARIWA,
Cusco, Per. p. 15.
Bertschinger, L. 1992. Modeling of potato
virus pathosystems by means of quantitative epidemiology: An exemplary
case based on virus degeneration
studies in Peru. Ph.D. thesis. Swiss
Federal lnstitute of Technology, Zurich,
Switzerland. 111 p.
Brunt A.A., S. Phillips, R.A.C. Jones, and
R.H. Kenten. 1982. Virus detected in
Ullucus tuberosus (Basel laceae) from
Peru and Bolivia. Annals of Applied
Biology 101 :65-71.
Duque, L.M. and M. Hermann. 1994.
Erradicacin de virus en melloco
(Ul/ucus tuberosus Caldas). Carta de
SEFIT - Sociedad Ecuatoriana de
Fitopatologa 3(6):1-3.
Hodge, W.H. 1951. Three native tuber
plants of the high Andes. Economic
Botany 5:185-201.
Jones, R.A.C. and R.H. Kenten. 1978.
Arracacha virus A, a newly recognized
virus infecting arracacha (Arracacia
xanthorrhiza: Umbelliferae) in the
Peruvian Andes. Annals of Applied
Biology 90:85-91.
Lizrraga, C., M. Querci, M. Santa Cruz,
l. Bartolina, and L.F. Salazar. 2000.
Other natural hosts of potato virus T.
Plant Disease 84:736-738.
Lizrraga, C., M. Santa Cruz, J.L. Marca,
and L.F. Salazar. 1999. La importancia
de los virus que infectan a Ullucus
tuberosus Caldas en el Per.
Fitopatologa 34:22-28.
Lizrraga, C., M. Santa Cruz, and
L.F. Salazar. 1997. Progress in identify-

ing viruses infecting Andean root and


tuber crops. Program report 1995-1996.
lnternational Potato Center, Lima, Peru.
p. 156-58.
Lizrraga, C., M. Santa Cruz, and
L.F. Salazar. 1996. First report of potato
leafroll virus in ulluco (Ul/ucus
tuberosus Caldas). Plant Disease 80:344
(Abstract)
Murashige, T. and F. Skoog. 1962. A
revised medium for rapid growth and
bioassays with tobacco tissue culture.
Physiologia Plantarum 15:473-497.
OIA (Oficina de Informacin Agraria).
2001. Estadstica Agraria - Junn.
Ministry of Agriculture, Peru. URL http:/
/www.minag.gob.pe
Salazar, L.F. 1996. Potato viruses and their
control. lnternational Potato Center,
Lima, Peru. 214 p.
Steel, R.C. and J.H. Torrie. 1980. Principies and procedures of statistics, A
biometrical approach. McCraw-Hill
Book Co, NY, USA. 623 p.
SAS. 1989. SAS lnstitute INC., SAS/STAT
User's Cuide, Version 6, 4th Edition,
Volume 2. Cary, NC, USA. 943 p.
Stone, O.M. 1982. The elimination of four
viruses from Ullucus tuberosus by
meristem-tip culture and chemotherapy.
Annals of Applied Biology 101 :79-83.
Tapia, M. 1992. Cultivos marginados d~ la
regin andina. In: Hernandez Bermejo,
J.E. and J. Len, (eds), Cultivos
marginados: Otra perspectiva de 1492,.
FAO, Rome, ltaly and Jardn Botnico
de Crdoba, Spain. p. 123-129.
Toledo, J., U. Jayasinghe, J. Anguerre,
R. Estrada, and M. Hermann. 1994.
Avances en los estudios de
enfermedades virales en Ullucus
tuberosus. Agro-Sur 22:14. (Abstract)
Villavicencio, A. 1999. Distribucin y
reinfeccin de virus del olluco (Ullucus
tuberosus Caldas) en la Sierra Central.
Ingeniero Agrnomo Thesis. Universidad
Nacional del Centro del Per.
Huancayo, Per. 58 p.

CIP Program Report 1999 - 2000

389

Compositional Changes of Oca Tubers Following


Postharvest Exposure to Sunlight
M. Hermann 1 and C. Erazo 2

Oca is an edible starchy tuber grown in the Andes. To improve its culinary
quality, harvested oca is typically exposed to direct sunlight (sunned) for
several days prior to consumption. Five native cultivars were examined to
determine the changes in nutritional composition of oca as a result of sunning.
This practice increased the proportions of dry matter, soluble solids, and
sugars (mainly sucrose) and decreased total acids, due mainly to a large
reduction in malate and glutarate. High levels of oxalate, an anti-nutritional
factor, were found, ranging from 306 to 539 and 251 to 451 mg/100 gin
edible matter of freshly harvested and sunned tubers, respectively. Sunning
reduced oxalate levels in dry matter by an average of 26%. Further study will
be necessary to fully assess the food safety of oca and to provide recommendations for maximum daily intake.

Oca (Oxalis tuberosa Molina) is an edible


starchy tuber grown in the high Andes.
Native Amerindian communities of Bolivia
and southern Peru regard oca as a valuable
source of food, second in importance to
the potato among root and tuber crops.
Oca tubers are usually baked or cooked in
stews, although rural children sometimes
eat them raw. Cooked oca has a mushy or
floury texture depending on the starch
content, and a sweet, slightly acid taste.
Raw oca is tart and pleasantly crunchy.
After harvesting, oca tubers destined for
direct consumption are traditionally
exposed to direct sunlight for a week or
two. This treatment is referred to as
sunning, or soleado, and is known to
enhance sweetness and improve culinary
quality.
The chemical composition of oca has not
been studied in detail, although we do
1

CIP, Lima, Peru.


2
Universidad Catlica, Quito, Ecuador.

know that starch accounts for the majority


of tuber dry matter and represents between
5% and 21 % of fresh matter (Cortes,
1978). According to Gross et al. (1989),
the most important sugars occurring in oca
are sucrose and glucose; there are also
traces of raffinose and stachyose. Protein
accounts for about 1% of fresh matter
(Gross et al., 1989; Kays et al., 1979), but
the quality is low (Gross et al., 1989).
Oca provides a good source of vitamin C
(38 mg/100 g (Collazos, 1974)), potassium,
and iron (Kays et al., 1979). Recent work
in New Zealand revealed high levels of
soluble oxalate in oca, ranging from 92 to
221 mg/100 g edible matter (Ross et al.,
1999). However, more comprehensive data
on the composition of acids and sugars,
which are responsible for the peculiar taste
and presumably postharvest changes in
sweetness and quality, are not currently
available in oca.

The objective of the present study was to


investigate the compositional changes

GIP Program Report 1999 - 2000

391

occurring in oca during sunning. Analyses


concentrated on starch, sugars, and acids
as these are the most important chemical
constituents in defining taste, texture, and
nutritional value.

Materials and Methods


Five native cultivars of oca were selected
(see Table 1), which are consumed directly, rather than processed into storable
products. Single plants were grown from
tubers planted in six-liter pots in an opensided, insect-proof, quarantine greenhouse
situated near Quito, Ecuador. Temperature
and light conditions were close to those of
the surrounding equatorial highland
environment (altitude 3000 m). The pot
substrate consisted of a m ix of peat, sand,
and volcanic soil. The plants were fertilized with a compound fertilizer, and
watered as required. A randomized block
design with three replications was used,
with pots spaced at intervals of 40 by 70
cm. Tubers were harvested eight months
after planting, when all accessions had
formed mature tubers and the aerial parts
of the plants had died. Half the harvested
tubers of each plant were reserved for
immediate chemical analysis and the
other half for sunning prior to analysis.
The sunning treatment started on the day
following harvest. Tubers were placed on a
reflective surface (aluminum foil) in the
greenhouse and left for 1 O days. They were
thus exposed to full sunlight and protected
from rain. Temperature sensors connected
to electronic ONSET "stow-away" data
loggers were placed in the center of one
oca tuber and in the surrounding air.

Temperatures were recorded at half-hourly


intervals.
Because of limited laboratory capacity,
only one pooled sample from the three
replications of an accession was analyzed
for chemical constituents. Sugars and
acids were extracted from freeze-dried
samples by immersing in water at 70C for
30 minutes. A high performance anion
exchange chromatograph (HPAEC) was
used to determine the levels of sugars and
acids. Starch was physically extracted
from fresh tuber tissue by macerating in a
kitchen blender, fol lowed by three cycles
of washing and starch sedimentation. Both
starch and HPAEC measurements were
taken in duplicate. Soluble solids were
measured with a portable refractometer
and levels of dry matter were determined
by freeze-drying.
Starch yields throughout this paper are
expressed as the weight of air-dried starch
obtained per tuber fresh weight. The water
content of air-dried starch was found to
range between 1 0% and 12%.

Results
Although their edaphic space was restricted by being pot-grown, the oca plants
developed normally, showing a typical
ontogenetic sequence (slow initial development fol lowed by vigorous shoot and
foliage growth, flowering, tuber bulking,
and senescence). Plant and final tuber size
and plant habit were comparable to those
usually observed for oca under field
conditions.
During sunning, maximum tuber flesh
temperatures reached 35C in the early

Table 1. Oca cultivars used in this study and their attributes.


Country and locality
Altitud e
Common
Collector
(m)
number
na me
2130
ECU-1018 Not available Ecuador, Los Dos Puentes,
Laja
2660
ECU-1025 Oca amarilla Ecuador, Baber, Laja
Peru, Chamis, Cajamarca
CA-5054 Kile
3100
MU-028 Ch'iyara apilla Peru, Collini - Uta Uyu, Puno 3900
HN-1146 Oca amarilla Argentina, Colanzul, Salta
3700
392

Andean Roots and Tubers and other Crops

Latitude Longitude

Comments

0359'S 7913'W Varieties typical of


the country of origin
0336'S 7911'W
0709'S 7830'W
1619'S 6918'W Good culinary quality
2253'S 6512'W Especially sweet

afternoon, and were about 1OC higher


than maximum air temperatures. Minimum
tissue temperatures occurred at dawn and
varied between 7 and 1 OC.

'C

Q)C")OQf'-

~~~~;;~~~~~

......

-.::j"'

C\J

"I";'"

~ .e

-.::j"'

C\J C\J

~ ~Lcir--.:~::~~::g~~~R~

Composition of freshly harvested oca


tubers.
Dry matter and physical ly extractable
starch accounted for 10.4-14.3% and 5.29.2% of freshly harvested edible matter,
respectively (Table 2). Total sugars comprised 2.1-3.6% and total organic acids
amounted to 2.6-3.4% of fresh matter. Of
the total sugars in freshly harvested tubers,
about half was sucrose and about a quarter
each of fructose and glucose. Among the
organic acids, glutarate and malate were
most abundant and tartarate least. There
were notable differences between accessions, especial ly regarding the content of
starch total sugars, ~nd single acids.
Accessions ECU-1025 and CA-5054, which
had the highest dry matter and starch
content, were also high in total sugars, but
had average acidity. HN-1146 had almost
twice the oxalate content of CA-5054, but
its total acid content was only 11 % higher.
This suggests the existence of considerable
variation in the acid profiles of different
oca cu ltivars. However, the statistical
significance of these results cannot be
determined due to the fact that only one
pooled sample was studied.

C.Nr--.:t-c.oC\JOOOC.OC\J~O~

~~~

LL

C.0

et')

C\J~

C.O C.0

C") et')

L{)

C")~~

C\J

a>c a.M. t -.
e: OMO>OC.OMC\JM-.::l"'OO>-.::l"'C.0
e: C\J
t - L{) C\J O> O> C\J e.o 00 e.o L{)
::::11
OOO>OOOC.OC\JC.OMM
(1)

L{) et') ~

"'O

cu

Cl
o
o

Light-microscopic inspection of tuber


tissue taken from al 1 five accessions, from
both freshly harvested and sunned tubers,
yielded no evidence of the presence of
insoluble oxalate crystals.

Compositional changes dueto sunning


Sunning resulted in several substantial
changes in tuber composition, which were
broadly consistent across accessions in
magnitude and direction. Table 2 shows
compositional changes for fresh, edible
matter. In comparison with freshly harvested tubers, sunning increased dry
matter content through water evaporation,
soluble sol ids, and total sugars, but sharply

e::
o
E

en

C2.

E e:
o(.) ::J

.l!:S Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl
ClCl~EEEEEEEEEE

(.)

.E
cu

..e::

(.)

e-.: cu

:e
-~
:e
ca ca
Q)

t-

>

CIP Program Report 1999 - 2000

393

decreased levels of starch and organic


acids. For example, sugar levels almost
doubled in CA-5054 and HN-1146, while
total acids in ECU-1025 were reduced to
less than a third of the level in freshly
harvested material. The decrease in acids
was due to large reductions in glutarate
and malate, typically to less than 30% of
the values of freshly harvested oca,
whereas oxalate, succinate, and tartarate
showed a slight overall increase. This
increase was probably due to a concentration effect resulting from dehydration of
the sunned tubers.
In contrast to fresh matter, dry matter
analysis (Table 3) showed a reduction in
levels of oxalate and succinate in all five
accessions, whereas tartarate was reduced
in three accessions. Si mi lar tendencies
were seen for sugars. Fructose and glucose
concentrations were generally higher in
the edible matter of sunned tubers compared with freshly harvested material
(Table 2), whereas analysis of dry matter
(Table 3) showed that both sugars tended
to diminish in sunned tubers. Sucrose
accounted for most of the increase in total
sugars after sunning, in both edible matter
and dry matter.
Figure 1 shows oca starch grains of freshly
harvested and sunned tubers, illustrating a
reduction in grain size, which suggests
starch degradation in the sunned material.

Figure 1. Oca starch granules of freshly harvested


(left) and sunned (right) oca tubers, cultivar ECU1025, at the same magnification.

Discussion
This study is the first to provide extensive
data on the carbohydrate and organic acid
composition of freshly harvested oca
tubers. lt is also the first to address the
compositional changes associated with
sunning, a traditional postharvest technique to render the tubers more palatable.
The oca cultivars used in this study are
morphologically diverse and represent
different agro-ecologies found in Argentina, Ecuador, and Peru. These data are
therefore not constrained by using geographically limited material and allow
inference to be made concerning oca
throughout the species range.
In agreement with earlier studies (Cortes,
1978; Cross et al., 1989; Kays et al.,
1979), oca was shown to consist mostly of

Table 3. Effect of sunning on oca tuber composition relative to dry matter.


Relative change due to sunning (%)
Range (% dry matter)
Variable
(freshly harvested = 100%)
Freshly
Sunned
ECU-1018 ECU-1025 CA-5054
MU-028
harvested
Starch
43-68
14-36
61
29
53
32
16-28
18-30
113
105
Total sugars
136
107
9-14
12-20
129
Sucrose
143
158
129
4-8
Fructose
3-7
95
85
98
83
3-6
Glucose
3-8
86
70
119
82
Organic acids
20-32
5-12
35
23
40
46
8-13
0.7-3.2
Glutarate
11
6
10
27
Mal ate
5-10
0.8-3.3
28
11
59
31
2.1-5.2
1.4-2.8
Oxalate
85
47
91
95
2.1-2.6
1.8-2.2
Succinate
76
76
87
90
0.3-1.2
Tartarate
0.4-1.3
160
147
70
84
394

Andean Roots and Tubers and other Crops

HN-1146
34
106
116
88
102
37
24
27
54
85
89

starch (about 50% of total dry matter) and


sugars. Sucrose was the most important
sugar fol lowed by fructose and glucose.
Acidity in oca is not only attributable to
oxalate, but also to other organic acids,
particularly glutarate and malate. However, oxalate is the most important acid
from a dietary point of view, being an antinutritional factor. Dietary oxalate binds to
calcium, preventing it from being used for
essential functions in human metabolism.
In addition, oxalate forms insoluble salts
with calcium, which can lead to the
development of kidney stones (Holmes et
al., 1995).
Levels of oxalate in freshly harvested
(306-539 mg/100 g edible matter) and in
sunned oca (251-451 mg/100 g) (Table 2),
were considerably higher than those found
in most other starchy staples (<100 mg/100
g). Only aroids such as taro (Colocasia
esculenta) have similarly elevated oxalate
levels; however, most of the oxalate in
these species occurs as insoluble calcium
salts (Holloway et al., 1989). In contrast,
Ross et al. (1999) found that oxalate in oca
appears in its soluble form. The present
study supports these findings, as oxalate
crystals were not observed in oca tissue
when viewed under the microscope.
Oxalate ranges in this study were 2-2.5
times greater than those reported in fieldgrown material in New Zealand (Ross et
al., 1999). Oxalate levels are known to be
influenced by environmental factors,
particularly nitrogen fertilization, and this
raises the question of whether cultivation
in pots led to the high levels observed in
this study.
Although high, the levels of oxalate found
in oca in this study are still only half those
reported for wild and cultivated species in
the Amaranthaceae, Chenopodiaceae, and
Polygonaceae (Libert and Franceschi
1987; Guil et al., 1997). These leaf '
vegetables are considered safe for human
consumption in moderation. Furthermore,
oca is mostly baked or boiled before being
consumed, procedures that are known to

significantly reduce oxalate levels (Ross et


al., 1999).
Further studies to identify the bio-availability of oxalate in oca are needed before
final nutritional recommendations can be
made. The findings of this study suggest
that, at the current state of knowledge, it
would be premature to promote oca for
regular and high consumption, as the
oxalate intake would be considerable and
could pose health risks especially for
infants.
This study suggests that sunning oca tubers
results in considerable compositional
changes. These improve the eating quality
of oca through an increased sugar content
and reduced acidity. Unfortunately, the
anti-nutritional factor oxalate is one of the
compounds least affected by these
changes. Nevertheless, oxalate intake per
consumed oca dry matter would, in the
case of the five accessions studied, have
been reduced through sunning to 47-95%
of the initial values (mean: 74%).
Much oca is processed into khaya, a dry,
storable product with a similar density to
cork. This has a bland, floury taste when
reconstituted with water and cooked.
Processing oca involves a number of steps,
including freezing, leaching, squeezing
and drying. Processing aims to remove
most of the solutes from the tuber, and is
probably very efficient in removing
oxalates. The compositional changes that
occur during processing of oca would be
worthy of further study.

References
Cortes H. 1978. Avances de la
investigacin en oca. In: Tapia, M.E.
and M. Villarroel (eds.). Proceedings of
the First lnternational Congress on
Andean Cultivars held 25-28 October
1977 in Ayacucho, Peru. lnter-American
lnstitute for Cooperation on Agriculture,
La Paz, Bolivia. p. 227-243. (in
Spanish)
Cross, R., F. Koch, l. Malaga, A.F. de
Miranda, H. Schoeneberger, and
CIP Program Report 1999 - 2000

395

L.C. Trugo. 1989. Chemical composition


and protein quality of sorne local
Andean food sources. Food Chemistry
34:25-34.
Guil, J.L., l. Rodrguez-Garca, and
E. Torija. 1997. Nutritional and toxic
factors in selected wild edible plants.
Plant Foods and Human Nutrition
51:99-107.
Holmes, R.P., H.O. Goodman, and
D.G. Assimos. 1995. Dietary oxalate
and its intestinal absorption. Scanning
Microscopy 9(4):1109-1120.
Holloway, W.D., M.E. Argall,
W.T. Jealous, J.A. Lee, and J.H.
Bradbury. 1989. Organic acids and
calcium oxalate in tropical root crops.

396

Andean Roots and Tubers and other Crops

Journal of Agricultura! and Food


Chemistry 37:337-341.
Kays, S.J., T.P. Gaines, and W.R. Kays.
1979. Changes in the composition of the
tuber crop Oxalis tuberosa Molina
during storage. Scientia Horticulturae
11 :45-50.
Libert, B. and V.R. Franceschi. 1987.
Oxalate in crop plants. Journal of
Agricultura! and Food Chemistry
35:926-938.
Ross, A.B., G.P. Savaga, R.J. Martin, and
L. Vanhanen. 1999. Oxalates in oca
(New Zealand Yam) (Oxalis tuberosa
Mol.). Journal of Agricultura! and Food
Chemistry 47:5019-5022.

lmpact of Downy Mildew on the Yield of Quinoa


S. Danielsen1 ' 2 , S.E. Jacobsen 1 , J. Echegaray1 , and T. Ames 1

Downy mildew (Peronospora farinosa (Fr.) Fr) is the most important disease of
quinoa (Chenopodium quinoa Willd.) in the Andes. To quantify the impact of
the disease on grain yield under natural conditions, a field experiment with
eight quinoa cultivars with and without application of fungicides, was carried
out in Huancayo, Junn, Peru (3200 m), where the crop is grown traditionally.
Area under the disease progress curve (AUDPC) values were calculated,
based on evaluations of disease severity (percentage leaf area affected). The
cultivar Utusaya, originating from the Bolivian salt desert (200 mm annual
rainfall), was strongly affected by downy mildew, which caused complete
defoliation, premature maturation, and a yield loss of 99%. Even in the most
resistant cultivar, yield was reduced 33%, indicating the destructiveness of
this disease.

Quinoa has been cultivated in the Andean


highlands for severa! thousand years. lt is
one of the most important crops in the
region and has great potential because it is
highly nutritious and resistant to various
adverse abiotic factors such as drought,
frost, and saline soils (Galwey, 1989;
Jacobsen et al., 1998; Jensen et al., 2000).
Downy mildew is endemic in Bolivia,
Chile, Colombia, Ecuador, and Peru,
where it is the most important disease of
quinoa (Alandia et al., 1979; Aragn and
Gutirrez, 1992). Downy mildew has also
been detected in Canada (Tewari and
Boyetchko, 1990) and Europe (Hockenhu 11
and Jacobsen, KVL, Copenhagen, Denmark, 1998, pers. comm.).
Downy mildew reduces the photosynthetic
area of the plant due to the development
of chlorotic and necrotic spots on the
leaves and causes leaf loss. The few
reports on downy mildew of quinoa deal
1

with pathogen detection, symptom description, host specificity identification,


and screening for resistance in the field
(Alandia et al., 1979; Aragn and
Gutirrez, 1992; Bonifacio and Saravia,
1999; Ochoa et al., 1999; Otaz et al.,
1976). Despite its importance and wide
dissemination, very little is known about
the effect of downy mildew on quinoa
grain yield.
This work is part of the Quinoa Project
CIP-Danida, whose research on downy
mildew of quinoa began in 1999. Research
included in the project is characterization
of the pathogen using molecular and
phenotypic markers, development of
disease evaluation methods, and assessment of the impact of downy mildew on
grain yield and quality. The objective of
the study reported here was to determine
the yield loss under natural, high disease
conditions caused by downy mildew in
eight quinoa cultivars with different
agroecological adaptations.

CIP, Lima, Peru.


The Royal Veterinary and Agricultura! University, Copenhagen,
Denmark.

CIP Program Report 1999 - 2000

397

Materials and Methods

Experimental design

The tria! was carried out in the experimental field of CIP, Huancayo, Junn, Peru,
(3200 m) during the 1999/2000 growing
season under conditions of natural downy
mildew infection.

The experiment was a split-plot design


with four replicates. The principal plots
were those that received the fungicide
treatments and the subplots were the
different cultivars.

Soil and climate


The soil is characterized as clay to clay
loam with a pH of 7.6-7.9. The diurna!
temperature ranged from 4-9C min and
15-26C max, and relative humidity
between 20 and 95%. Total precipitation
during the growing period was 400 mm.
Two additional irrigations were made.

Soil disease control


To prevent the attack of soilborne diseases
caused by Rhizoctonia and Fusarium, the
seeds were treated with PCNB and
benomyl at sowing, and the plants treated
later in the season as needed.

Cultivars
Eight quinoa cultivars with different
geographical origins were included in the
study: Utusaya (salt desert); LP-48, La
Molina 89 (coast); Blanca de Juli,
Kancolla, and Jujuy (highlands); and
Amarilla de Marangan and lngapirca
(val ley). Utusaya and LP-48 are the
earliest maturing cultivars; La Molina 89,
Amarilla de Marangan, and lngapirca the
la test.

Each plot consisted of six rows 3.5 m long,


with a distance between rows of 60 cm.
Two rows of maize were planted between
each principal plot to limit fungicide drift
between treated and untreated plots. A
path of 1 m was established between
subplots.

Evaluations of downy mildew


Ten plants of the four center rows of each
plot were selected at random for the
evaluations. The first evaluation was
performed 69 days after sowing (DAS) and
once a month thereafter, three times in
total. Disease severity was evaluated on
three leaves/plant selected at random. A
leaf was taken from the lower, middle, and
upper part of each plant sampled, and
rated according to the scale developed by
Danielsen and Ames (2000). The average
of the three leaves represents the disease
severity value of the plant. Based on these
measurements, AUDPC (area under the
disease progress curve) was calculated
according to the equation of Campbell and
Madden (1 990):
n-1

AUDPC

= ~, ( Y;+

Y;+, )/2 X ( t;+1-t)

Control of downy mildew


Two weeks after sowing, treated plots were
sprayed with a commercial dose of
metalaxyl (systemic fungicide). Three
weeks later mancozeb (contact fungicide)
applications were begun. The mancozeb
treatment consisted of three applications
spaced one week apart. This cycle was
repeated throughout the growth period to
keep infection at a mnimum. The last
application was given at the end of the
rainy season, one month before the
beginning of harvest.

398

Andean Roots and Tubers and other Crops

where n is the number of evaluations, y is


disease severity, and t is the number of
days after sowing when the evaluation is
done; (t/ y) = (O, O) is included as the first
evaluation. The advantage of using
AUDPC values, rather than single severity
measurements, is that AU DPC reflects
disease progress throughout the whole
growing season. Single severity readings
do not capture changes caused by environmental fluctuations.

Harvest
Seed yield was estimated from harvest of
the four center rows. The number of plants
in each plot was counted. Harvest was
done when the plants were mature,
starting 131 DAS with Utusaya and LP-4B,
and ending 165 DAS with La Malina 89,
Amarilla de Marangan, and lngapirca.

Amarilla de Marangan
100 - . - - - - - - - - - - - - - - - - - - - .

?f.

- - Untreated
__._ T reated

80

Cll

!!!
Cll
1ii

60

"O

40

:g

20

.!l1

69

Statistical analysis
The effect of cultivar and fungicide on
yield and AU DPC, and the effect of
fungicide treatment on the number of
plants per plot were analyzed by ANOVA.
The correlation between yield and AU DPC
was analyzed with correlation analysis
(SAS System version 6.12, SAS lnstitute
lnc., Cary, NC).

98

131

Days after sowing


Utusa ya
100

?f. 80
Cll
!!!
Cll
60
1ii
.!l1
"O
40

20

Results

98

69

131

Days after sowing

The plants were harvested between 131


and 165 DAS. In general, the plots without
fungicide application matured between
9 and 1 3 days earl ier than the plots with
fungicide application. The number of
plants/plot was reduced significantly
(P = 0.017 4) due to the effect of downy
mildew. For this reason, a covariance
analysis was not made to correct the yield
according to the number of plants/plot.

Figure 1. Disease progress curves in twa quinaa


cultivars with twa levels af dawny mildew (with and
withaut fungicide). The difference in AUDPC far the
entire growth seasan between treated and untreated is
significant in bath cultivars, P = 0.0432 and 0.0001
far Amarilla de Marangan and Utusaya, respectively.
5000

DTreated
Untreated

400)

The severity of downy mildew measured


as AUDPC was significantly higher in the
untreated than in the treated plots (Table
1, Figures 1 and 2). According to AUDPC
values, Utusaya was the most susceptible
cultivar and La Molina 89, Amarilla de
Marangan, and lngapirca the most
resistant (Figures 1 and 2). In ali cultivars
the level of downy mi ldew was kept at a
minimum in the fungicide treated plots
with no significant difference in AUDPC
between cu ltivars.
Downy mildew caused a significant yield
reduction in ali cultivars (Table 1). Yield
loss ranged between 33% (Amarilla de
Marangan) and 99% (Utusaya). The
correlation analysis showed a strongly
significant negative correlation between
AUDPC and yield (r = -0.73, P = 0.0001).

oa.. :mo
o

::>

2CXXl

<(

1CXXJ

<tl

en

::::>

=>

::::>

--,
Q)

"O

ro

>-

""'!"

::-

o....
_J

::::>

--,

oo
e

<tl

::,,::::

-~o.
<tl
Ol

<tl

in

:2
Q)

"O

;::

C1l
00

<tl

<tl

:2

_J

<(

<tl

Cultivar

Figure 2. AUDPC values in quinaa cultivars with twa


levels af dawny mildew (with and withaut fungicide).
Calumns marked with different letters are significantly
different (P < 0.05). Calumns far treated and
untreated are campared separately.

Discussion and Condusions


This confirms the observations of Bonifacio
and Saravia (1999) who pointed out that

CIP Program Report 1999 - 2000

399

Table 1. Seed yield and AUDPC values in eight quinoa cultivars with two levels of downy mildew (with and
without fungicide treatment).
AUDPC 2
Yield (kg/ha) 1
Cultivar
3
3
Untreated Yield loss (%)
Treated
Untreated
Difference
Treated
Salt desert cultivar
579
2106
23
4636
4057 ***
Utusaya
99 ***
Coastal cultivars
4092
1021
384
2790
LP-48
75 ***
2406 ***
7686
3263
256
852
La Malina 89
58 ***
596 *
Highland cultivars
1465
Blanca de Juli
5891
510
2976
75 ***
2466 ***
3083
408
1829
Kancolla
4748
35 *
1421 ***
1846
544
2478
J~uy
5424
66 ***
1934 ***
High altitude valley cultivars
Amarilla de M.
7559
5073
33 ***
303
882
579 *
144
lngapirca
6570
3562
46 ***
1207
1063 ***
Leve! of significance: * P = 0.05, *** P = 0.001 (ANOVA)
1 The harvested area of each plot was 8.4 m2. The yield was converted into kg/ha by multiplying g/parcel with a factor

1.19.
2 Area under the disease progress curve.
3 The plots were sprayed with metalaxyl (systemic) and mancozeb (contact) to control downy mildew.

late cultivars in general are more resistant


to downy mildew than early cultivars. That
was true for the most resistant cultivars in
this study: lngapirca, La Molina 89, and
Amarilla de Marangan. Despite low
AUDPC values, these three cultivars
registered yield reductions ranging from 33
to 58%. This indicates that even a mild
attack of downy mildew can cause a
substantial reduction in seed yield. These
findings may mean that the AUDPC fails
to sufficiently describe the disease.
The severity measurement only considers
the foliage left on the plant and not the
leaves that have been shed as an effect of
the disease. Furthermore, severity is a
relative parameter that does not take into
account total size of the plant or foliage
density. However, a strongly significant
negative correlation between AUDPC and
yield was found, which means that under
these conditions AUDPC explains most of
the yield variation. Another promising
method, the measurement of canopy
reflectance, is presently being developed
at CIP to detect late blight in patato and
downy mildew in quinoa.
Because of their high susceptibility to
downy mildew, cultivars originating from
400

Andean Roots and Tubers and other Crops

the Solivian salt deserts (e.g., Utusaya) are


restricted to areas with low precipitation
where downy mildew is nota problem.
Cultivars originating from the coast,
valleys, or highlands can give satisfactory
yields in the presence of downy mildew.
But the 33-58% yield reductions experienced by representative cu ltivars from
those areas u sed in th is study demonstrate
the severe effect downy mildew has on
quinoa yield.

Acknowledgments
We are grateful to Carlos Hualhuas and
Lorenzo Safra for handling the seeds.
Thanks to Felipe Mendiburu for helping
with the statistics. This study was supported by the Danish government (RUF
grant no. 90929).

References
Alandia, S., V. Otaz, and B. Salas. 1979.
Enfermedades. In: Tapia, M.,
H. Gandarillas, S. Alandia, A. Cardozo,
A. Mujica, R. Ortiz, V. Otaz, J. Rea,
B. Salas, and E. Sanabria, (eds.).
Quinua y Kaiwa. Editorial llCA,
Bogot, Colombia, p. 137-148.

Aragn, L. and W. Gutirrez. 1992. El


mildiu en cuatro especies de Chenopodium. Fitopatologa 27:104-109.
Bonifacio, A. and R. Saravia. 1999.
Evaluacin de la resistencia al mildiu
en quinua. In: Proceedings of the Tercer
Taller de Preduza en Resistencia
Duradera en Cultivos Altos en la Zona
Andina, 27-29 de Septiembre de 1999.
Cochabamba, Bolivia, p. 49-59.
Campbell, C.L. and L.V. Madden. 1990.
lntroduction to plant disease epidemiology. John Wiley & Sons, NY, USA. 532 p.
Danielsen, S. and T. Ames. 2000. El mildiu
(Peronospora farinosa) de la quinua
(Chenopodium quin()a) en la zona
andina. Manual prctico para el estudio
de la enfermedad y del patgeno.
lnternational Patato Center, Lima, Per,
32 p.
Galwey, N.W. 1989. Quinoa. Biologist
36:267-274.
Jacobsen, S.-E., A. Mujica, and O. Stolen.
1998. Salt tolerance of quinoa during
germination. Agronoma Tropical
(Maracay) 48:359-366.

Jensen, C.R., 5.-E. Jacobsen,


M.N. Andersen, N. Nunez,
S.D. Andersen, L. Rasmussen, and
V.O. Mogensen. 2000. Leaf gas exchange and water relation characteristics of field quinoa (Chenopodium
quinoa Willd.) during soil drying.
European Journal of Agronomy
13:11-25.
Ochoa, J., H.D. Frinking, and Th. Jacobs.
1999. Postulation of virulence groups
and resistance factors in the quinoa/
downy mildew pathosystem using
material from Ecuador. Plant Pathology
48:425-430.
Otaz, V., P.C. Agu lar, and A. Can ah u a.
1976. Resistencia en quinua (Chenopodium quinoa) al mildi (Peronospora
effusa). Fitopatologa 11 :47-49.
Tewari, J.P. and S.M. Boyetchko. 1990.
Occurrence of Peronospora farinosa
f.sp. chenopodii on quinoa in Canada.
Canadian Plant Disease Survey
70:127-128.

CIP Program Report 1999 - 2000

401

Quinoa: An Alternative Crop for Saline Soils


in the Andes
S.-E. Jacobsen1 , H. Quispe 1 , and A. Mujica2

Salt tolerance mechanisms were studied in two varieties of quinoa


(Chenopodium quinoa Willd.), and in one variety of amaranth (Amaranthus
caudatus L.). The experiment was conducted in pots in a greenhouse at CIP,
Lima, Peru. The amaranth demonstrated very little ability for regulation of
leaf water potential and stomatal conductivity, and the plants died at high
salinity levels. Quinoa behaved as a facultative halophyte, accumulating salt
ions in the tissue. This mechanism adjusted leaf water potential, enabling the
plants to maintain cell turgor and limit transpiration under saline conditions.
The accumulation of salt indicates that quinoa may be used to clean saltcontaminated soils. The characters most sensitive to salinity were stomatal
conductance, leaf area, and plant height. Therefore, screening for only minor
reduction in, for instance, plant height may be a means of identifying varieties
for saline soils.

During the past three decades, severa!


developing countries in the dry regions of
the world have greatly expanded their
areas of irrigated lands to produce the food
needed by their rapidly growing populations (Brady and Weil, 1999). lnitially,
large increases in food crop production
were stimulated by expanded irrigation. In
many areas, however, the need for good
soil drainage was overlooked, and,
therefore, the process of salinization has
accelerated. As a consequence, salts have
accumulated to levels that are adversely
affecting crop production.
Soil salinization is one of the major
problems in agriculture, particularly
because saline soils are found primarily in
arid regions where drought, extreme
temperatures, and nutrient deficiency go
hand in hand, and where scarce precipitation and high evaporation hinder a
1
2

CIP, Lima, Peru.


Proyecto Quinua CIP-DANIDA, Puno, Peru.

leaching out of the salts that accumulate


in the upper soil layers. lt is estimated that
between 340 andas muchas 950 billion
square ~i lometers, equ ivalent to about
20% of the arid and semiarid soils of the
world, or 6% of the world land area, are
saline (Flowers et al., 1986). In addition,
there is an increase in salinization due to
irrigation, which is estimated to affect
50% of irrigated land.
Most crops are sensitive to salt. Mechan isms of tolerance for salt in halophytes,
species that prefer saline conditions, have
been studied in the search for alternative
crops, and to enhance understanding of
how to improve tolerance in existing
crops. Plants may survive salinity by
synthesizing organic salutes, which is an
important physiological characteristic of
the halophytes. They have the ability to
accumulate salutes to enable continued
growth and prevent a water deficit or an
excess of ions such as sodium (Na),

CIP Program Report 1999 - 2000

403

potassium (K), and chloride (CI), which


can be toxic (Flowers and Yeo, 1986;
Flowers et al., 1986; Jacoby, 1999; Prado,
1999). Salt stress affects plant growth and
modifies plant physiology. High salt
concentrations in the environment induce
water deficits in plants. A physiological
response to salt stress is diminished
stomatal conductance, which reduces
transpiration for maintaining cell turgor,
reduces CO 2 intake, and inhibits photosynthesis. This causes reduced biomass, leaf
area, and plant height (Katerji et al., 1994;
Reddy and lyengar, 1999).
The Andean highlands are affected by salt
problems, particularly in the salt deserts of
the southern part of Bolivia and the central
Altiplano of Peru and Bolivia around Lake
Titicaca, in raised beds of the altiplano
(Canahua and Cutipa, 2001 ), and in large
parts of the cold agroecological zones
called suni and pampa (Tapia, 2001 ). Only
a few crops can be grown in these marginal areas. One of them, quinoa, can be
cultivated under extreme conditions of
salinity, drought, and cold (Jacobsen and
Mujica, 2001 ). The Andean crop gene
bank of the Universidad National del
Altiplano (UNA), Puno, has 1800 quinoa
accessions from the Andean regan with
wide genetic variability for many characteristics, including salt tolerance (Tapia,
1980).
Quinoa seeds of the 103-accession Peruvian core collection held at UNA (Ortiz et
al., 1999) were selected based on a
geographically stratified, non-overlap_ring
sampling procedure. They were germ1nated in salt concentration levels of 0.6 M
to identify accessions with salt tolerance
(Jacobsen et al., 2001 ). Quinoa accessions
showed different responses to salt concentrations, so that 15 accessions of the core
collection were initially selected as
tolerant. However, cotyledonous leaves
were observed in only four of these
accessions, the others died. Thus, when
screening for salt tolerance in quinoa, a
seed germination study is not sufficient.

404

Andean Roots and Tubers and other Crops

Further assessment of seedling and plant


development under salt stress is required.
The purpose of this study was to evaluate
the leve! of salt tolerance in quinoa and
amaranth when applying saline water
throughout the entire growth period. The
effect of the salt on various characters was
estimated.

Materials and Methods


A greenhouse trial was conducted at CIP,
Lima, Peru, from March to July 1999.
Plants were grown in individual pots, each
containing 5 kg of substrate. The substrate
consisted of 2 parts silty soil, 1 part sand,
and 1 part compost. Daily average temperature ranged from 16.5C in July to
23.9C in March. Average relative humidity was between 76% in March and 88%
in July. The study included two quinoa
bred varieties (Utusaya, regarded as the
most salt-tolerant material, and 03-26-0036
from the UNA gene bank) and one amaranth variety (Osear Blanco, from
Huancayo, Peru). Utusaya comes from the
salt desert of the southern Bol ivian Altiplano. The amaranth variety was included in
the study for comparison with quinoa, as it
is considered a saline-susceptible crop.
Plants were treated with nine levels of
salinity: electrical conductivity (EC) at
25C of < 1, 5, 1 O, 15, 20, 25, 30, 35, and
40 mS/cm (miliSiemens per centimeter).
The salinity gradient was established
according to the EC of the irrigation water.
The salinity of the drainage water and
saturated soil extract was monitored to
determine the salinity of the substrate,
which was adjusted to maintain salinity at
the predetermined levels. The salt sourc~
was seawater with the following compos1tion:
Bicarbonates (mg/I)
Boron (mg/I)
Carbonates (mg/1)
Calcium (mg/1)
Chlorides (mg/I)
Sulfates (mg/I)

156

o
397
13,582
2134

Potassium (mg/I)
Magnesium (mg/I)
Sodium (mg/I)
Nitrates(mg/1)
pH
EC (mS/cm)

469
1120
7018
37
7
43

Salinization was induced at the beginning


of floral bud formation by irrigating with
diluted seawater to obtain the specific
salinity levels. For the <1 mS/cm treatment, tap water with EC of 300 S/cm was
used. The pot substrate was irrigated at
75% of field capacity (FC) to obtain full
FC, corresponding to 24% soil moisture
content. The experi ment was done in a
randomized complete block design, with
27 treatments and 5 replications. Thus, the
data were processed under a factorial
arrangement (3 x 9). The analyzed variables were leaf water potential, stomatal
conductance, leaf area, plant height, root
weight, stem weight, stem diameter,
panicle weight, panicle length, seed yield,
and harvest index. Leaf water potential
was measured with a Scholander 80325
pump (Labconco, Kansas City, MO, USA),
stomatal conductance was measured with
an AP-4 porometer (Delta-T Devices,
Cambridge, U.K.), and leaf area was
measured with a leaf area meter Li-Cor LI
3000 (Li-Cor, Lincoln, NB, USA).
For analysis of salt tolerance, saline
treatment vs control treatment (TS/TC) was
used. TS was the average of the three most
saline treatments (30, 35, and 40 mS/cm),
and TC the average of the three least
saline treatments (<1, 5, and 10 mS/cm)
(Royo and Arages, 1995). Values close to
1 demonstrated tolerance, and values
el ose to zero demonstrated h igh susceptibi l ity to salinity. Analysis of variances
were run (SAS, 1988), and Tukey's means
separation test was used to compare
averages.

Results and Discussion


Highly significant differences between
varieties and between the analyzed
variables were found for TS/TC. The

amaranth variety was most susceptible to


salinity, with an average TS/TC of 0.15
over ali measured variables, followed by
the quinoa varieties Utusaya (TS/TC =
0.59) and 03-26-0036 (TS/TC = 0.70). For
amaranth, the characteristics least affected by salinity stress were leaf water
potential and stomatal conductance,
indicating that this species does not adjust
its physiological activity to stress, resulting
in low biomass and seed production
(Figure 1).
For quinoa, stem diameter and harvest
index were the variables least affected by
salinity, with values of TS/TC close to 1.
Utusaya had higher harvest index under
salinity (higher TS/TC ratio) than 03-260036, but 03-26-0036 had higher seed
yield (Figure 1).
Stomatal conductance was very sensitive
to salinity, as were leaf area, plant height,
root and stem weight, leaf water potential,
and seed yield. The same sensitivity to
salinity has also been shown for barley
(Hordeum vulgare L.) (Royo and Aragez,
1995).
Differences in seed yield were highly
significant between varieties tested.
Utusaya had a significantly higher average
seed yield over treatments (2.4 g 0.95)
than 03-26-0036 (2.0 0.60 g) and amaranth (0.5 0.40 g).
The highest seed yield for quinoa was
obtained at an EC of 15 mS/cm (3.8 0.78
g for Utusaya and 2.5 0.27 g for 03-260036). At ECs higher than 15 mS/cm, yield
began to decrease in both varieties. In
amaranth, the highest yield was obtained
at an EC of 5 mS/cm (1 .1 0.22 g).
With respect to the behavior of quinoa in a
saline medium, it was previously reported
that salt concentrations between 8.1 and
16 mS/cm are lethal for growth (Cari,
1978), which is much lower than we found
here. Bosque (1998) demonstrated that the
stomatal resistance of the quinoa variety

Real, an ecotype from the salt desert of


Bolivia, ranged from 4.8 to 16.1 s/cm

CIP Program Report 1999 - 2000

405

Stem diameter (mm)

(A)

Harvest index (%)

(A)

~
1

~
~

Panicle weight (dm (g))

(BC)

Panicle length (cm)

(BCD)

Seed yield (g/plant)

(BCD)

Stem wei'ght (dm (g))

(CDE)

Leaf water potential (MPa)*

(CDE)

~
1
'1

Plant height (cm)

,,

,,,,,,,,,

:-...-...................................

.............

........................................................................

(DE)

~
1

Root weight (dm (g))


2

Amaranth
Osear Blanco

(E)

SI

Leaf a rea (cm ) *

(E)

Stomatal conductance (cm/sec)*

(F)

Quinua
03-26-0036

1
~

'1

~'''''''"
1
1

0.1

0.2 0.3

0.4

0.5 0.6

0.7

0.8

Quinua
Utusaya
1

0.9

1.1

Figure 1. Salinity tolerance according to TS/TC values in quinoa varieties (03-26-0036 and Utusaya) and in the
amaranth variety Osear Blanco.* = Measured at anthesis. A-F indicate significant differences at the 95% level as
an average between the two quinoa varieties and the one of amaranth. (dm = dry matter; MPa = MegaPascal)
(seconds per centimeter) under salinity of
7.77 mS/cm and from 12.2 to 28.2 s/cm at
11.88 mS/cm; increasing salinity tended to
increase this resistance.

Conclusion
Amaranth demonstrated very 1ittle abi 1ity
to regulate leaf water potential and
stomatal conductivity, and the plants died
at high salinity levels. According to the
TS/TC factor, amaranth was four times
more susceptible to salinity than quinoa in
relation to seed yield. Quinoa demonstrated the ability to accumulate salt ions
in its tissue to control and adjust leaf water
potential. That enabled the plants to
maintain cell turgor and limit transpiration
under saline conditions, thus avoiding

406

Andean Roots and Tubers and other Crops

physiological damage from drought or


even potential death.
Sorne of the characteristics that were
measured, such as leaf area, biomass
production, seed yield, and harvest index,
showed better responses under moderate
saline conditions (10-20 mS/cm) than
under lower EC, indicating that quinoa is a
facultative halophyte. The study also
indicated that one of the characters most
sensitive to salinity was plant height.
Therefore, screening for only a minor
reduction in this character may be a
means of increasing seed yield of quinoa
in saline soils. The study also demonstrated that the measure of salinity
tolerance defined above, TC/TS was
suitable for comparisons of speies and
varieties.

References
Bosque, H.D. 1998. Ecophysiological
analysis of drought and salinity stress of
quinoa (Chenopodium quinoa Willd.).
MSc thesis in Soi 1 Science and
Eremology. Faculty of Science, Faculty
of Agricultura! and Applied Biological
Sciences, University of Gent.
lnternational Center for Eremology,
Gent, Belgium. 122 p.
Brady, N.C. and R.R. Weil. 1999. The
nature and properties of soils. 121h
edition. Prentice-Hal 1, lnc, Upper
Saddle River, NJ, USA.
Canahua, A. and Z. Cutipa. 2001.
Produccin de la quinua en waru-waru:
Perspectivas y limitaciones. In:
Jacobsen, S.-E. and Z. Portillo (eds.).
Memorias, Primer Taller Internacional
sobre Quinua - Recursos Geneticos y
Sistemas de Produccin, 10-14 May
1999, Universidad Nacional Agraria La
Molina, Lima, Peru. CD-Rom available
from CIP, Lima, Peru.
Cari, A. 1978. Efectos de la salinidad y
fertilizacin potsica en dos variedades
de quinua (Chenopodium quinoa Willd).
MSc thesis. Universidad Nacional
Tcnica del Altiplano, Puno, Peru. 66 p.
Flowers, T.J., M.A. Hajibagheri, and
N.J.W. Clipson. 1986. Halophytes. The
Quarterly Review of Biology 61 (3):
313-337.
Flowers, T.J. and A.R. Veo. 1986. Ion
relations of plants under drought and
salinity. Australian Journal of Scientific
Research 13:75-91.
Jacobsen, S.-E. and A. Mujica. 2001.
Avances en el conocimiento de
resistencia a factores abiticos adversos
en la quinua (Chenopodium quinoa
Willd.). In: Jacobsen, S.-E. and
Portillo (eds.). Memorias, Primer
Taller Internacional sobre Quinua Recursos Geneticos y Sistemas de
Produccin, 10-14 May 1999,
Universidad Nacional Agraria La
Molina, Lima, Peru. CD-Rom available
from CIP, Lima, Peru.

z.

Jacobsen, S.-E., E. Ruiz, A. Mujica,


J.L. Christiansen, and R. Ortiz. 2001.
Evaluacin de accessiones de quinua
para la tolerancia a salinidad. In:
Jacobsen, S.-E. and Z. Portillo (eds.).
Memorias, Primer Taller Internacional
sobre Quinua - Recursos Geneticos y
Sistemas de Produccin, 10-14 May
1999, Universidad Nacional Agraria La
Molina, Lima, Peru. CD-Rom available
from CIP, Lima, Peru.
Jacoby, B. 1999. Mechanism involved in
salt tolerance of plants. In: Pessarakli
M. (ed.). Handbook of plant and crop
stress. Second edition. Marcel Dekker,
lnc., NY, USA. p. 97-123.
Katerji, N., J.W. van Hoorn, A. Hamdy,
F. Karam, and M. Mastrorilli. 1994.
Effect of salinity on emergence and on
water stress and early seedling growth
of sunflower and maize. In: Agricultura!
water management. Elsevier Science.
Netherlands. p. 81-91.
Ortiz, R., S. Madsen, E.N. Ruiz-Tapia,
S.-E. Jacobsen, A. Mujica-Sanchez,
J.L. Christiansen, and O. Stolen. 1999.
Validating a core collection of Peruvian
quinoa germplasm. Genetic Resources
and Crop Evolution 46:285-290.
Prado, F. 1999. Fisiologa y bioqumica del
estrs: respuestas de las plantas al
ambiente. In: Jacobsen, S.-E. and
A. Mujica (eds.). Fisiologa de la
resistencia a sequa en quinua
(Chenopodium quinua Willd.). First
lnternational Course. Quinoa Project.
CIP-Danida, Lima, Peru. p. 21-29.
Reddy, M. and E. lyengar. 1999. Crop
responses to salt stress: Seawater
application and prospects. In:
Pessarakl i, M. (ed.). Handbook of plant
and crop stress. Second edition. Marce!
Dekker, lnc., NY, USA. p. 1041-1068.
Royo, A. and R. Arages. 1995. Efecto de
la salinidad sobre diversos caracteres
morfo-fisiolgicos y sobre el
rendimiento en grano de la cebada. In:
Investigacin agraria, produccin y
proteccin vegetal. Instituto Nacional

de Investigacin y Tecnologa Agraria y


Alimentacin. Madrid, Spain. p. 71-84.

CIP Program Report 1999 - 2000

407

Tapia, M. 1980. Origen, distribucin


geogrfica y sistemas de produccin de
la quinua. In: 1 Reunin sobre gentica
y fitomejoramiento de la quinua.
Proyecto PISCA/UNTA/IBTA/llCA/CllD,
Puno, Peru.
Tapia, M. 2001. Zonificacin
agroecolgica del cultivo de la quinua

408

Andean Roots and Tubers and other Crops

(Chenopodium quinoa Willd.). In:


Jacobsen, 5.-E. and Z. Portillo (eds.).
Memorias, Primer Taller Internacional
sobre Quinua - Recursos Geneticos y
Sistemas de Produccin, 1 0-14 May
1999, Universidad Nacional Agraria La
Molina, Lima, Peru. CD-Rom available
from CIP, Lima, Peru.

Enriching the Portfolio: CIP's Global


and Regional Partnerships
Scientist and Farmer
Partners in Research for the 21 st Century

CONDESAN: Watershed Use Planning


Throughout most of the Andes, farmers must make decisions at the community
level, as well as on their individual farms. Resolving issues such as increasing
irrigation efficiency, landslide and soil erosion prevention, improving pest control,
increasing range productivity, and combating the scarcity of fuel wood require a
landscape or watershed vision. Therefore, an organizing principie for the Consortium for Sustainable Development of the Andean Ecoregion (CONDESAN) is land
management at the watershed scale. As a partnership project with the lnternational Patato Center, the Consortium works to incorporate the output from CIP
scientists and others into successful development projects.

Working at a watershed rather than farm level, however, is accompanied by three


challenges: (1) collecting and evaluating the quality of the data at this higher
scale of aggregation; (2) promoting awareness, interest, and participation at both
the individual community and watershed level; and (3) establishing logical links
between improved watershed management and practica! livelihood issues. The
Consortium is addressing these challenges in a number of ways. Work is being
undertaken to develop robust geographic information systems (GIS) methodologies
that permit local NGOs to build and modify natural resource maps with municipal
authorities as described in the following papers. Work is being done to reinforce
existing village committees that manage local irrigation systems, common
grazing resources, or bioreserves with watershed-level information. The Consortium is working at the policy level to encourage the development of appropriate
government incentive programs and attract private entrepreneurs to invest in
projects that improve both local incomes and natural resource management.

CIP Program Report 1999 - 2000

411

Putting Natural Resource Management on the Map:


Using GIS as a Tool for Soil Conservation Planning in
Two Small Andean Watersheds
J. Posner1, C. Bussink2, R. Hijmans2, J. de la Cruz2, H. Willet3, P.A. Sanchez4

The paper describes a practical approach to developing a georeferenced


natural resource database that is locally constructed and managed. With the
new GIS (geographic information system) software and hardware available
today, various types of natural resource information can be collected and
combined to allow local users to visualize new relationships in GIS output
maps. The maps can be prepared and modified according to specific criteria.
This capacity helps operationalize the concept of the watershed for use in
development planning. The data collected for this study were used to create
maps for two watersheds that show zones for alternative soil conservation
interventions. This work was a collaborative effort by the Cajamarca office of
the Peruvian Proyecto Nacional de Manejo de Cuencas Hidrogrficas y
Conservacin de Suelos (PRONAMACHCS), the non-governmental organization Asociacin para el Desarrollo Rural de Cajamarca (ASPADERUC),
municipal authorities of La Encaada and Asuncin, and the lnternational
Potato Center (CIP). CONDESAN (Consorcio para el Desarrollo Sostenible de
la Ecorregin Andina) sponsored the project.

Steep topography, abrupt changes in


ecology due to elevation, the importance
of irrigation, and a new emphasis on
environmental management are all factors
that make watersheds an often-used unit
for rural development and natural resource
management projects in the Andes
(Rhoades, 1998). In addition, the desire to
develop more participatory approaches to
rural development has given impetus to
focusing on micro-watersheds (500020,000 ha), where the inhabitants are
neighbors who are members of the same
irrigation committees and often serve
together on municipal councils.
1

CONDESAN/CIP, Lima, Peru; present address: University of


Wisconsin, Madison, WS, USA.
CIP, Lima, Peru.
3
SNV/PRONAMACHCS, Cajamarca, Peru.
4
ASPADERUC, Cajamarca, Peru.
2

One challenge to this approach is


operationalizing the concept of watershed.
For many municipal and project officials it
is difficult to move beyond site-specific
interventions (such as promoting a new
crop or constructing terraces) to a watershed perspective, or to use a watershed
vision to set priorities for site-specific
activities (Farrington and Lobo, 1997).
In order to address this issue, a team of
CIP/CONDESAN scientists, agronomists
from the non-governmental organization
(NGO) ASPADERUC (Asociacin para el
Desarrollo Rural de Cajamarca), and
municipal officials, worked together to
develop a methodology to incorporate
different types of local information (topography, hydrology, political limits, etc.) into
a Geographic lnformation System (GIS)
CIP Program Report 1999 - 2000

413

matrix. Then, in collaboraton with the


Peruvian national soil and water conservation agency PRONAMACHCS (Proyecto
Nacional de Manejo de Cuencas
Hidrogrficas y Conservacion de Suelos),
algorithms were defined that permitted a
simplification of the information and the
identification of zones for soil conservation interventions.
The overall goal of this activity is to
develop a cost-effective and robust
methodology that will enrich the local
debate on natural resource management
with information that is collected and
processed locally. We initiated the work in
northern Peru in the Cajamarca Department, first with a watershed within the
municipality of La Encaada (15,700 ha),
and then in Asuncin (8100 ha) (Figure 1).

Materials and Methods


Data management
Data managed in a GIS can be used for a
variety of purposes. The goal of this
activity would be to develop a solid
database for different types of community
planning. The first output were maps and a
zonification system for soil conservation
interventions that could be used by
PRONAMACHCS, the national soil and
water conservation agency. lt was assumed that slope, ground cover, soil type,
and rainfall intensity would be the most
important factors in the zonification.
Figure 2 summarizes the flow of information in developing the maps. First, contour
lines were digitized and interpolated,
resulting in elevation models. These were
used to subdivide the watershed into
catchments and to make a slope map.
Ground cover and soils maps (approximately 1 :25,000) were then drawn using
aerial photos and existing soil maps,
'walking the landscape', and digging soil
pits. Based on this work, slope classes,
ground cover classes, and soil depth units
were defined. Political boundaries between communities, roads and paths, and
local reference points (clinics, schools,

414

Enriching the Portfolio: CIP's Global and Regional Partnerships

and churches) were also georeferenced


and included in the GIS database. The GIS
work for La Encaada was completed at
the lnternational Potato Center using
existing databases developed for other
projects (De la Cruz et al., 1999). For
Asuncin, ASPADERUC scanned the maps
for importing into Autocad, a mapping
program, where the features of the map
were digitized on screen. Additional
landmarks were georeferenced with a
global positioning system (GPS). Data
were imported into IDRISI (IDRISI software
(http://www.clarklabs.org/)) for processing.

Developing interventions
After the data collection phase, algorithms
were devised following the flow diagram
in Figure 2 to divide the watershed into
soil conservation zones. In collaboration
with the PRONAMACHCS team and
published tables (PRONAMACHCS, 1998),
a set of rules was established to identify
potential soil conservation nterventions
(no intervention, sloping terraces, infiltration ditches, and reforestation) on cropped
land (Table 1).
There was a consensus that soi 1 conservation measures were not of great
importance for slopes of less than 5% and
that no interventions other than vegetation
restoration and forest plantations were
possible on slopes of more than 40%. The
main interventions would focus on the
intermediate slope types, which were
categorized into two groups: 5-15%
(terracing possible) and 16-40% (infiltration ditches, terracing only possible if the
soils are deep).
The six ground cover classes originally
used in the data collection phase were
simplified to three, based on 'permanency'
of ground cover: zones of permanent
vegetation (grasslands, forest), zones of
annual disturbance (annual crops plus
associated short-term fal low), and degraded zones where little vegetative cover
remained. With respect to crop type,
cereal (maize, wheat, barley, oats), tuber
(potato, oca, ollucu), and legume (peas,

76'

68'

72'

o
Colombia
Ecuador

4'

4'

CAJAMARCA
Department

Brasil

Altitude [m]

CJ 1-1000
CJ 1001 - 2000
c---12001 - 3000
CJ 3001 - 4000
-

4001-5000

CJ >5000

12'

12

.~

.;:::

oco

16'

16'

so

72'

76'

68'

Figure 1. Location of the study watersheds near Cajamarca, Peru.


lentils, lupins) crops were all collapsed
into an 'annual cropping' category. The
reason is that in this region of Peru there is
only one rainy season (November- March,
500-800 mm/year) resulting in similar
growth patterns for each of these different
food crops. When funds are limited, soil

conservation in areas with permanent


cover (pastures, forests) is not a priority nor
is the rehabi 1itation of seriously degraded
areas. Therefore, the focus of the
interventions would be on areas in annual
crops and their associated fallows.

CIP Program Report 1999 - 2000

415

Elevation
model

Slope

Slope
classes

Ground
cover

Ground
cover
classes

Soil

1-------11---

Eros ion
risk

1----...__~soil depth1-------1~~----'

classes

Basemaps

C=>

Decision table

Results

Process
Proposal at
commnunity
level

Figure 2. lnformation flow in developing a soil conservation zonification.


Areas in annual crops were crossed with
four slope classes to create an erosion risk
table for annual cropping: low risk on
slopes of less than 5%; medium risk on
slopes of 6-15%; high risk on slopes from
16-40%; very high risk on slopes of more
than 40%.
According to the soil classification used,
there were 209 different units in La
Encaada (Jimnez, 1996) and 114 units in
Asuncin (Jimnez, 1998). But because
decisions on soil conservation interven-

tions (terraces, ditches, and reforestation)


with these non-volcanic soils would be
based primarily on soil depth rather than
on such characteristics as soil texture or
interna! drainage, it was considered
appropriate to simplify the soil categorization into just two categories: less than 60
cm deep, and more than 60 cm deep.
Few small watersheds in the Peruvian
Andes have rainfall stations that measure
both total rainfall and maximum 30minute rainfal 1 intensities. Because both of

Table 1. Decision rules far soil conservation interventions in annual cropping areas.
Soil depth
Erosion potential
Proposed intervention
Shallow<60 cm
Low
No intervention
Medium
Sloping terraces
High
lnfiltration ditches
Very high
Re-establish native vegetation
Deep>60 cm
Low
No intervention
Medium
Sloping terraces
High
Sloping terraces/infiltration ditches
Very high
Commercial forests
416

Enriching the Portfolio: CIP's Global and Regional Partnerships

the project micro-watersheds have one


station, neither has two, it is difficult to
identify differences in rainfall within the
watershed. As a result, although daily
rainfall data were collected and used in
subsequent crop (DSSAT) and watershed
models (SWAT), it was not possible to use
variations in rainfal 1 amount and intensity
within the watershed as a factor in this
typology.
Once the rules were agreed upon, the
maps were drawn, and the classification
was ground checked. At that point, the
relative areas for each type of intervention
were calculated for the watershed, by
catchment, and by community. Maps of
several communities were enlarged to
1 :2000 and used to initiate participatory
discussions about potential conservation
interventions.

Results and Discussion


In spite of the general ly steep topography
in the Andes and the relatively smal 1 size
of the watersheds (8000-15,000 ha), we
found that working at a scale of 1:25,000
with contour lines at intervals of 25 m was
acceptable for our objectives. At this scale
we were able to identify important
differences on the landscape and in
sufficient detail for conservation planning.
The two watersheds selected are fairly
typical of agricultura! zones in the Andes.
High elevation native pastures, primarily
of Calamagrostis, Festuca, and Bromus
grasses (Snchez, 1995; 1999),
predominate in La Encaada (45%), but
constitute only 15% of the lower Asuncin
watershed. Nevertheless, both watersheds
are dominated by steep slopes (> 15%),
and cropped land plus its associated fal low
make up 33-45% of the watershed area.
On less steeply sloping land, planted
pastures are common, mostly consisting of
ryegrass (Lolium perenne) and clover
(Trifolium repens).
La Encaada has greater expanses of flat
land (slope 0-5%) used for annual cropping than does Asuncin (23% vs. 4%). In

Asuncin, nearly half of the annual


cropping is conducted on slopes of more
than 40%. In contrast, in La Encaada
only 12% of the cropped land is on slopes
that steep. This suggests that the
sustai nabi 1ity of production systems in
Asuncin is in greater jeopardy than it is in
La Encaada, and rather than soil conservation interventions, alternative
production systems (such as improved
pastures, more fruit trees or forestry) must
be a high priority for municipal authorities.
Figure 3A and 3 B are maps of soi 1 conservation intervention zones by watershed. At
each site, the no-intervention zone covers
the largest area (63% in La Encaada,
54% in Asuncin); these are the areas that
are either in permanent cover or in annual
cropping but on mild slopes. Zones
suitable for sloping terraces, infiltration
ditches, and reforestation are the next
largest areas (30% in La Encaada, 44% in
Asuncin). Terracing and creating
infiltration ditches on cropped lands will
res u lt in better soi 1 and water management
and should improve crop yields. On the
other hand, reforestation will mean taking
land out of annual cropping so the tradeoff must be examined. The third major
category is rehabilitation, which implies
rebuilding soil and vegetative cover in
zones that have been degraded, primarily
by man's activities (La Encaada 7%,
Asuncin 2%).
When the interventions map is superimposed on the community boundaries map,
critica! areas, those with high priority for
soil conservation, can be identified. For
example, both La Torre community in La
Encaada (highlighted with parallel lines
in Figure 3A) and Shirac in Asuncin
(highlighted with parallel lines in Figure
3B) are high priority areas. In both, nearly
one hundred percent of the area is in
cropland and 81 % in the former and 96%
in the latter would need terracing, infiltration ditches, or should be put into
permanent cover according to these
algorithms.

CIP Program Report 1999 - 2000

417

795000

805000

800000

810000

Land use
Annual

aopp;ng
Degraded

1D

Zonification
kea (ha)
No intervention
444
lnftration ditches
1655
Sloping terraces
2422
186
Commercial forest
Nativeforest

446

Rehabilitation

1085
9503

(\J Corrwnunity boundaries


:" \ .: Undefined community bounclaries

9215000

9215000

9210000

9220000

9210000

i!!!!!!'!"'!1iiiii;;;;ij2~!!!!'!"

aooooo

795000

805000

810000

Figure 3A. La Encaada soil conservation zonification with community boundries.


772000

776000

780000

Area ha)

124
1241

483
70
1662
9192000

9192000

75
4467
(\) Community boundaries

9188000

9188000

9184000

9184000

kilomeler.;
1
Praeclil:xlllTMi:one 17s

772000

776000

Figure 38. Asuncion soi l conservation zonification with community boundries.


418

Enriching the Portfolio: CIP's Global and Regional Partnerships

780000

An additional step in this process was to


take the results of the zonification exercise to individual communities to see if
they would be a useful aid in local soil
and water conservation planning. A blowup (1 :2000) of a portian of the watershed
map covering the single community of La
Torre in La Encaada was taken into the
field to do a participatory zonification
exercise. An evaluation of this exercise
with the PRONAMACHCS officials led to
two main conclusions. First, the map
showing the zonification of interventions
was helpful in enriching the conversation,
b~cause the map did not coincide exactly
w1th the local vision of natural resource
management. And second, having a
georeferenced base map was useful to
agency technicians because they could
accurately locate and inventory existing
structures (terraces, waterways) and plan
new i nterventions.
Not surprisingly, the change in scale from
a 15,700 ha watershed to a 220 ha community resulted in sorne locally
'unacceptable' general ities. Based on
further community-level discussions, the
team added three components to the
program.
With the help of farmers, project
agronomists are making community
land-use maps based on local criteria
(Olivares et al., 2000).
Recent aerial photographs (approximately 1 :15,000) are being blown up to
1 :3000 and used as the basis of natural
resource discussions in the community.
The team is redoubling its efforts to
focus on irrigation and potable water
issues (Delgado, 1998; Soto, 1996).
With these additions to the flow diagram
shown in Figure 2, the process and database are now serving the needs of district
(watershed authorities) and community
leaders. For example, the database and
mapping capabilities are being used to
draw land-use maps based on local
definitions, to identify irrigation canal
catchment basins, to map ali the springs

potentially useful for drinking water, and


to identify zones for orchard expansion.

Conclusions
Digitizing existing maps and
georeferencing sorne additional field
measurements can result in useful databases for natural resource management.
What is equally important is that more and
more local organizations (government
agencies, NGOs, university laboratories)
are acquiring the software and experience
necessary to apply and improve the
methodology. At the level of a watershed
like La Encaada or Asuncin, the methodology allows municipal and agency
personal to identify soil conservation
intervention zones or zones suitable for
new production systems. At the local
community level, the system is being
adapted to help envision and salve local
problems, especial ly around water issues.

References
De la Cruz, J., P. Zorogasta, and
R. J. Hijmans. 1999. Atlas digital de los
recursos naturales de Cajamarca.
Department of Natural Resource
Management working document No. 2.
lnternational Patato Center, Lima, Peru.

49 p.
Delgado Loayza, R. 1998. Evaluacin del
recurso hdrico de la microcuenca de
Ro Asuncin-Cuenca del Jequetepeque.
ASPADERUC/CONDESAN. Cajamarca,
Peru. 43 p.
Farrington, J. and C. Lobo. 1997. Scalingup participatory watershed development
in India: Lessons from the lndo-German
Watershed Development Programme.
Natural Resource Perspectives 1 7:1-6.
Jimnez Medina, M. 1996. Estudio
de suelos de la Microcuenca "La
Encaada" (semi-detallado).
ASPADERUC/CONDESAN, Cajamarca,
Peru. 73 p.
Jimnez Medina, M. 1998. Estudio de
suelos-Distrito de la Asuncin (semi-

detal lado), ASPADERUC/CONDESAN.


Cajamarca, Peru.

CIP Program Report 1999 - 2000

419

Olivares, M., G. Muoz, E. Torres,


and P.A. Snchez. 2000. Estudio
sensible de suelos de La Encaada
PRONAMACHCS-MIMA/
ASPADERUC, Cajamarca, Peru. 1O p.
PRONAMACHCS. 1998. Manual tcnico
de conservacin de suelos y aguas.
Agencia Cajamarca, Peru. 1O p.
Rhoades, R.E. 1998. Participatory
watershed research and management.
Gatekeeper Series No. 81. lnternational
lnstitute for Environment and
Development, London, U K. 20 p.

420

Enriching the Portfolio: CIP's Global and Regional Partnerships

Snchez Vega, l. 1995. Estudio botnico


de la microcuenca de La Encaada.
ASPADERUC/CONDESAN. Cajamarca,
Peru. 39 p.
Snchez Vega, l. 1999. Estudio botnico
de la microcuenca Asuncin.
ASPADERUC/CONDESAN. Cajamarca,
Peru.
Soto Hoyos, J.F. 1996. Diagnstico del uso
actual y potencial hdrico en las
Microcuencas de La Encaada y
Tambomayo. ASPADERUC/
CONDESAN. Cajamarca, Peru. 90 p.

Land-Use Change in the Cajamarca Catchment, Peru,

1975-1996
C.B. Bussink and R.J. Hijmans

We used land-use maps for different years and areas in the Cajamarca catchment in the northern Peruvian Andes to study changes in land use between
1975 and 1996. Despite population growth, agriculture has not been intensified nor extensified. Neither did we find evidence for crop encroachment into
the higher parts of the catchment. The agricultural area has decreased and
the proportion of fallow land has increased. There has been a significant
increase of shrubs and bare soil at the expense of natural pastures, perhaps
indicating overgrazing. The area planted to specific crops has fluctuated, but
there has been a clear decrease in the area planted to barley and an increase
in wheat. The area with exotic tree species has also increased.

The global increase in population has led


to important changes in agriculture,
through both intensification and
extensification of production.
Extensification refers to the use of more
land for agriculture, while intensification
refers to using agricultura! land more
intensively, e.g., with shorter fallow
periods. Both processes can have a
significant negative impact on the environment. Agricultura! extensification leads
to the loss of natural habitats, and intensification can lead to the loss of biodiversity
as well as to environmental pollution, as it
often goes hand in hand with increased
pesticide and ferti 1izer use. However,
intensification may also lead to increased
economic land value and incentives to
conserve land resources through such
measures as soil conservation (Tiffen and
Mortimore, 1994).
lt is important to note, however, that
patterns of land use change can be rather
different between regions and countries,
within countries, and even within village
territories. Particularly in areas of heterogeneous terrain, such as the Andes,

different patterns of land-use change are


likely to occur simultaneously; there are
reports of intensification as well as abandonment from different parts of the Andes
(e.g., Mayer, 1979; Herv and Ayangma,
2000; Wiegers et al., 1999). Understanding
the patterns and causes of land-use change
in the Andes could lead to the development of policies that would help avoid
sorne of the more negative consequences
of these changes. lnsight into the current
processes of change could also help in the
design of development interventions.
In the present study, we describe land-use
change in the Cajamarca catchment
between 1975-1996. We used a geographic information system (GIS) to
compare a number of large-scale maps
(1 :25,000) from different years and zones.
The Cajamarca catchment is in the
northern Peruvian Andes (Figure 1 ). lt is
regarded as a region with a distinct
ecology, characterized by the jalea or
grassland vegetation, in the highest areas,
being a transition zone between the more
humid paramo Andes in the north and the
drier puna Andes in the south (Troll, 1968).

CIP Program Report 1999 - 2000

421

7'00'

Cajamarca
catchment

kilometers
5 10

o
!'8"1.5'

rJO'

78"{):}'

Figure 1. Location of the Cajamarca catchment in Peru and the study areas: 1975/78 (the whole catchment),
1991 (A), 1992/1993 (B), 1996 (C), and the zone with three studies (D).
Near Cajamarca town, at 2650 m, average
daily temperature is about 13C and yearly
precipitation is 720 mm. The temperature
is rather constant over the year, while
diurna! differences are about 17C. There
is a clear rainy season between October
and April (De la Cruz et al., 1999). The
altitude in the catchment is between 2000
m to 4200 m, with about 79% between
2600 m and 3800 m. The distribution of
crops by altitude has been used to distinguish main agroecological zones: the
cu ltivated pasture zone (irrigated val ley
bottoms), the maize zone (lower slopes),
the tuber zone (higher slopes), and the
natural pasture zone (jalea) (Kohler, 1986;
Kohler and Tillman, 1988; Seifert, 1990;
422

Enriching the Portfolio: CIP's Global and Regional Partnerships

Tapia, 1996). Land-use systems in this area


are different from those found in the
central and southern Andes of Peru (see,
e.g., Mayer, 1979; Morlon, 1992). For
example, there is not much communally
managed land in Cajamarca, and household access to different production zones
is limited.
Between 1940 and 1993, the Cajamarca
Department had an annual population
growth rate of about 1 .5%, and the smaller
Cajamarca Province (which roughly
coincides with the Cajamarca catchment)
had a growth rate of 1.7%. Although this is
below the national average (2.2% for the
same period), Cajamarca has a high

growth rate for areas in the Peruvian Andes


(INEI and UNFPA, 1996; INEI, 1995;
Seifert, 1990; Wiegers et al., 1999).

April). The maps were digitized' using


Arclnfo (ERSI, Redlands, CA, USA) and
analyzed with ArcView (ESRI) and
Microsoft Excel.

Materials and methods


Land-use change
Land use

Our point of reference is a series of three


land-use maps for 1975, 1977, and 1978.
Together, these maps cover the whole
Cajamarca catchment (211,854 ha),
except for a small part in the southwest
(25,932 ha). They also cover sorne areas
outside the catchment, which we excluded from our study. We compared these
maps with six land-use maps of smal ler
zones within the Cajamarca catchment,
for different years between 1991 and 1996.
To faci 1itate the analysis and because of
complementarity in location and similarity
of years, we merged the maps for 1975,
1977, and 1978 (Gozalo et al., 1977;
Landa et al., 1978; Landa and Johansen,
1978) to form what we will refer to as the
"1975/78 map". We merged three maps for
1991 (zone A, 19,748 ha (UNC, 1991 a,
1991 b, 1991 c)) and two maps for 1992 and
1993 (zone B, 6,916 ha (Chiln, 1993;
Saldaa, 1994)). Zone C is based on one
map for 1996 (18,732 ha (M. Jimenez,
UNC, unpublished)) (Figure 1). We also
made a single map for the "1990s,"
combining zones A, B, and C, using the
most recent data when there was an
overlap between zones.
All the maps were made using the same
methodology. With air photos (most at a
scale of 1 :20,000) and fieldwork, map
units were delineated and the percentage
of different land-use classes within each
unit was estimated. Land-use classes
include broad categories of vegetation as
well as different crops. For example, one
map unit could contain 30% maize, 20%
wheat, 40 % shrub land, and 10% natural
pasture. We simplified the maps in sorne
cases by merging classes: rye and oats;
potatoes and other tubers; and pulses
(beans, lupines, lentils, and green peas).
Fieldwork was mostly carried out during
the main growing season (November to

To study changes in land use, the 1975/78


data were compared with 1991, 1992/93,
and 1996. All comparisons were made on
the areas in common to the two maps. For
a small area of 4,369 ha (zone O in Figure
1 ), we had an overlap of three studies
(1975/78, 1991, and 1996). For all zones,
we calculated the fallow ratio as the area
of land in fallow divided by the total
cultivated area (crops and fallow land).
Tabulating areas for one zone for two
different dates al lowed overal 1 change to
be assessed. As one land unit on a map
can contain many land-use classes, we
applied the following two rules to estimate
which land-use classes changed into
which other classes. For all land units, the
following applies:
lf a land-use class is present in both
time periods, we assume that the
minimum of the two areas occupied the
same location within that land unit in
both periods (i.e., no change).
The differences in area of one land-use
class are proportionally subtracted from
(in the case of an increase) or added to
(if there was a decrease) the a reas of al 1
new land-use classes. The proportion is
based on the change in area.
These rules are likely to result in conservative estimates, meaning that sorne
changes may be underestimated, because
they do not take into account changes in
location of land-use classes within map
units.

Results
Land use in 1975/78

In 1975/78, the largest land-use class in


the Cajamarca catchment was natural
pastures (40%); 38% of the land was
cultivated (30% annual crops, 5% fallow,
CIP Program Report 1999 - 2000

423

and 3% cultivated pastures) (Table 1). The


most important annual crops were barley
and wheat. Annual crops and fallow land
were relatively over-represented in the
three study zones in 1975/78, in comparison with land use in the catchment as a
whole (Table 1 ).
Land-use change

In all study areas, there was an increase in


the area with shrubs (+ 10% for the 1990s),
forest (+3%), wheat (+3%), and fallow
(+0.6%), whereas there was a decrease in
the areas with barley (-7%), pulses (-1 %),
and natural pasture (-14%) (Figure 2).

Changes in the other land-use classes were


general ly smaller and inconsistent: in
sorne zones there was an increase and in
other zones there was a decrease.
Most of the individual land-use classes in
1975/78 had, for the most part, the same
land use as in 1990 (Table 2). This was
especially true for land with shrubs,
forests, cultivated pasture, and crops
(around 70%). On average, over al 1 zones,
new forests have been planted on former
natural pasture (38% of the new forest),
cropland (21 %), and on bare soil and shrub
land (19%). In the 1990s, 15% of the
former area of bare soil was cultivated

Table 1. Land use in Cajamarca by different areas and years (%).


Cultivated
Natural
Shrubs
Forest
Ye ar
Are a
pasture
pasture
1975(78
o
40
13
3
Catchment
1975(78
o
27
8
1
Zone A
1991
26
3
1
5
1975(78
15
1
1
Zone B
30
23
1
1992/93
20
3
1975(78
7
Zone C
32
6
1
7
6
1996
23
10
1975(78
4
8
1
Zone
30
1990s
16
19
5
5
A+B+C

Crops

Fallow

30
45
37
35
34
37
33
39
35

5
9
12
9
9
7
8
9
9

Bare soil
& rocks
9
10
15
9
11
10
12
9
12

Land use

Forest
Cultivated pasture
Maize
Wheat

13 Zone A (1975178-91)

Barley

mZone C (1975178-96)

C Zone B (1975178-92/93)
C Zone A+B+C

Rye& Oats
Pulses

Tubers
Fallow
Bare soil I rocks
-25

-20

-15

-10

-5

10

15

20

Area change {%)

Figure 2. Change in land use far three zones with different time series (1975/78-1991, 1975/78-1992/93, and
1975/78-1996).
424

Enriching the Portfolio: CIP's Global and Regional Partnerships

Table 2. Change in land use from one class to another between 1975/1978 and the 1990s (expressed in
percent of area of each land-use class in 1975/78).
1990s
Cultivated
Natural
Bare soil /
Shrubs
Forest
Crops
Fallow
Total
1915na
pasture
pasture
rocks
Natural
15
11
39
20
6
3
5
100
pasture

12

100

100

73

11

66

100

10

15

19

43

100

12

15

57

100

Shrubs

69

Forest

69

10

Cultivated
pasture
Crops
Fallow
Bare soil J
rocks

with crops again, 5% was fallow, and 1 %


was covered with cultivated pastures. New
areas with bare soils or with shrubs were
mostly in areas that had natural pastures or
crops in 1978.
There was an average decrease of 4% of
the total arable land area (Table 1), which
includes crops and fallow land, roughly
irrespective of altitude (Figure 3). The
relative proportion of fallow land over crop
land increased. In 1978, there was one ha
of fallow for every 5.6 ha of cultivated
land (including fallow), while in the
nineties there was one ha of fallow for
every 4.8 ha of cultivated land.
Land-use change in zone D, for which we
had data for three years (1975/78, 1991,
Relative area (%)
70

60
50

40

01975(78

30

1990s

20
10

Figure 3. Agricultura! area (crops and fallow) by


altitude class far 1975/78 and the 1990s.

and 1996) confirms the major trends: there


is an increase in the areas of bare soil,
shrubs, forests, fallow, and wheat, and an
important decrease in the areas with
barley and natural pastures. Although for
most classes the changes between 1975/78
and 1991 and 1975/78 and 1996 point in
the same direction, the size of change can
be quite different (Figure 4).

Conclusions and Discussion


We have not found any evidence for
intensification of land use in the
Cajamarca catchment between 1975 and
1996. The total area under cultivation had
decreased si ightly, and the fal low ratio
had increased. Neither did we find evidence for an extension of the agricultura!
frontier into higher areas; this is often
suggested to be happening.
Ou r resu lts are in 1i ne with the observations of Frias (1995) that, despite
population growth, the importance of
agriculture for employment is decreasing
in both absolute and relative terms.
However, the results seem to contradict
Seifert (1990), who found that fallowing
had disappeared on farms of less than
5 ha. This could be because these small
farms, although they are the most common

CIP Program Report 1999 - 2000

425

Relative area

01975/78
1991
[!11996

Land use

Figure 4. Area of land use classes far three years (1975/78, 1991 and 1996) in one zone (Zone D; Figure 1).
farm size, only cover 10% of the cultivated area (Seifert, 1990). The decrease of
area planted to tubers and the increase of
area under cultivated pasture coincides
with Frias's (1995) study for the whole
Cajamarca Department. However, instead
of the decrease in wheat described by
Frias (1995), we found an increase.
The largest change is an increase in shrub
land and bare soil, mainly at the expense
of natural pasture. Overgrazing may have
led to soi 1 eros ion and loss of soi 1 ferti 1ity,
which in turn, may have led to an increase
in unpalatable shrubs. Alternatively, it
might be a sign of vegetation recuperating
because of a decrease in grazing pressure.
The i ncrease in forested areas match es the
pattern of agricultura! dis-intensification
and retraction. Also, over the past few
decades, there has been considerable
public investment in forestation in
Cajamarca, particu larly through a local
non-governmental organization.
One complication with our data is that we
compared not only across different years
but also across different zones. The only
zone for wh ich we had data for three
periods was zone D. However, it appears
that for sorne land-use classes, the effect

426

Enriching the Portfolio: CIP's Global and Regional Partnerships

of classification error overrides the effect


of years. In zone A, the areas with shrubs,
fallow land, and bare soil were probably
overestimated at the expense of natural
pasture, because it is quite unlikely that
much of the soil that was bare in 1975/78
would be cultivated again 20 years later,
as our data suggest. Rigorous definitions of
these land-use classes are necessary to
improve this type of study. Crops are easier
to classify, and the results for zone D
indicate that there is considerable variation between years in areas with specific
crops.
For estimates of crop area, the timing of
the survey is very important. In the offseason, there would be much more fallow
and the natural vegetation would be more
likely to be classified as bare soil. The
spatial effect of rotation on land-use
patterns should not be of much importance
for this study because it should average
out within and between land units.
This study is descriptive in nature and does
not offer insights into the factors that drive
the changes we observed nor in other
important aspects of land use, such as
productivity. Further research, including
detailed fieldwork, would be needed for

this. At the other end of the spectrum lies


the question about the representativeness
of our results for a larger area such as the
northern Peruvian Andes, which, for
example, could be studied with satellite
images. Representativeness is an important
issue because the Cajamarca catchment
may not be typical. Because the catchment is located near an important
departmental capital, land use may have
been relatively intensive for a long time;
then, over the past few decades, there
may have been more intensification in
other areas. Whereas in our study area, the
area of exotic forest (eucalyptus, pine)
increased, natural forests in nearby areas
are believed to have decreased dramatically over the last 30 years (H. Willet,
personal communication; ONERN, 1975).
The time period we studied is obviously
arbitrary and completely driven by the
presence of land-use maps. The results
could be put into a wider context using
census and household data in which an
important milestone would be land reform
of the early 1970s.

Acknowledgements
We thank Jorge de la Cruz, Felipe de
Mendiburu, Edwin Rojas, and Percy
Zorogasta for assistance; Jiefar Diaz,
Wilfredo Poma (UNC), and Pablo Sanchez
(ASPADERUC) for providing access to old
land-use maps; CONDESAN for partly
funding this research; and Joshua Posner
and Dominique Herv for reviewing of this
paper.

References
Chiln C., M.I. 1993. Estudio del uso
actual de la tierra de la intercuenca del
rio Sambar-Distrito Baos del Inca.
Master's thesis. Universidad Nacional
de Cajamarca, Peru.
De la Cruz, J., P. Zorogasta, and R.J.
Hijmans. 1999. Atlas digital de los
recursos naturales de Cajamarca.
Natural Resource Management Working
Document No. 2. lnternational Potato
Center, Lima, Peru. 49 p.

Frias, C. 1995. Pobreza campesina. Slo


un problema rural? Cajamarca:
economa, espacio y tecnologa.
lntermediate Technology Development
Group (ITDG), Lima, Peru. 153 p.
Gozalo, V.J., C. Landa, and M. Garcia.
1977. Uso Actual de la tierra de los
PIAR's Cajamarca y San Marcos. Vol. l.
Distritos de Jesus, Namora, Matara y
parte de los de Cajamarca Llacanora y
San Marcos. Ministry of Agriculture,
Food, and Technical Cooperation,
Belga, Cajamarca, Peru.
Herv, D. and S. Ayangma. 2000.
Dynamique du l'occupation du sol dans
une communaut agropastorale de
!'altiplano central bolivien. Revue de
Gographic Alpine 2:69-84.
INEI (Instituto Nacional de Estadstica e
Informtica). 1995. Per: Poblacin
total por rea urbana y rural, segn
departamentos, provincias y distritos.
Instituto Nacional de Estadstica e
Informtica, Lima, Peru.
INEI (Instituto Nacional de Estadstica e
Informtica) and UNFPA (United
Nations Population Fund). 1996. Per:
Proyecciones departamentales de la
poblacin 1995-2015. Instituto
Nacional de Estadstica e Informtica
and United Nations Population Fund,
Lima, Peru.
Kohler, A. 1986. Sistemas de uso de suelos
en laderas de Cajamarca - Per. Boletin
de Lima 47:31-40.
Kohler, A. and H.J. Tillman. 1988.
Campesinos y medio ambiente en
Cajamarca. Mosca Azul, Lima, Peru.
Landa, E.C. and A. Johanson. 1978. Uso
actual de la tierra de los PIAR's
Cajamarca y San Marcos. Vol. 111.
Estudio de parte de los distritos de
lchocan, San Marcos, Encaada
Cajamarca y Chetilla. Ministry of
Agriculture, Food, and Technical
Cooperation, Belga, Cajamarca, Peru.
Landa, E.C., A. Johanson, and M. Garcia.
1978. Uso actual de la tierra de los
PIAR's Cajamarca y San Marcos. Vol. 11.
Zona Norte, Distrito de Baos del Inca y
parte de la Encaada y Llacanora. Zona

GIP Program Report 1999 - 2000

427

Sur, Parte de los distritos de San Marcos


e !chocan. Ministry of Agriculture,
Food, and Technical Cooperation,
Belga, Cajamarca, Peru.
Mayer, E. 1979. Land use in the Andes:
Ecology and agriculture in the Mantaro
Valley of Peru with special references to
potatoes. lnternational Potato Center,
Lima, Peru. 115 p.
Morlon, P. 1992. Comprendre l'agriculture
paysanne dans les andes centrales.
Prou-Bolivie. National lnstitute of
Agricultura! Research, Paris, France.
HONREN (Oficina Nacional de
Evaluacin de Recursos Naturales).
1975. Inventario, evaluacin y uso
racional de los recursos naturales de la
zona sur del departamento de
Cajamarca. Volumen l. National Office
far the Evaluation of Natural Resources,
Lima, Peru.
Saldaa A., M.O. 1994. Estudio del uso
actual de la tierra de los caserios de
Chim-Chim, Carahuanga y la Collpa en
la cuenca del rio Quinuario - Distrito
Baos del Inca. Thesis, Universidad
Nacional de Cajamarca, Peru.
Seifert, R. 1990. Cajamarca: Va
campesina y cuenca lechera. Privately
published, Cajamarca, Peru.
Tapia, M.E. 1996. Ecodesarrollo en los
Andes altos. Friedrich Ebert Foundation,
Lima, Peru.
Tiffen, M. and M. Mortimore. 1994.
Malthus controverted: The role of

428

Enriching the Portfolio: CIP's Global and Regional Partnerships

capital and technology in growth and


environment recovery in Kenya. World
Development 22(7):997-101 O.
Trol!, C. 1968. The cordilleras of the
tropical Americas. In: Trol!, C. (ed.).
Geoecology of the mountainous regions
of the tropical Americas. Collequium
Geographicum 9, lnstitut de Universitat
Bonn, Drummlers Verlag, Bonn,
Germany. p. 1 5-56.
UNC (Universidad Nacional de
Cajamarca). 1991 a. Uso actual de la
tierra. Map at 1 :25.000 scale. Convenio
CUSO - UNC. Proyecto de desarrollo
rural integral de Jesus. Universidad
Nacional de Cajamarca, Cajamarca,
Peru. (Unpublished.)
UNC (Universidad Nacional de
Cajamarca). 1 991 b. Uso actual de la
tierra. Map at 1 :25.000 scale. Servicio
Silvo Agropecuaria (S.E.S.A).
Universidad Nacional de Cajamarca,
Cajamarca, Peru. (Unpublished.)
UNC (Universidad Nacional de
Cajamarca). 1991 c. Uso actual de la
tierra. Map at 1 :25.000 scale. Grupo
Polivalente de Proyectos Sociales Chim Shaullo. Universidad Nacional de
Cajamarca, Cajamarca, Peru. (Unpublished.)
Wiegers, E.S., R.J. Hijmans, D. Herv, and
LO. Fresco. 1999. Land use intensification and distintensification in the upper
Caete Valley, Peru. Human Ecology 27

(2): 319-339.

GILB: An Update
The Global lnitiative on Late Blight (GILB) brings people together with the aim of
increasing and enhancing global effarts devoted to solving the problem of late
blight in patato, the world's most costly biotic constraint to patato production.
GILB was initiated in early 1996 when farty-one participants from developing and
industrialized countries met at a Project Design Meeting at CIP headquarters in
Lima, Peru, to plan a three-phase, ten-year program with specific priorities far
each phase. lnterest in GILB has grown rapidly, to the point that, in the year
2000, the GILB newsletter reached more than 633 late blight workers in over 77
countries.
GILB priorities were further modified and refined during a 1999 GILB-organized
global conference with 165 participants from 40 countries. Participants reviewed
the list of priorities established in 1996 and analyzed achievements under those
priorities. Because impressive progress had been made toward meeting the initial
priorities, in part through rapid advances in molecular technologies, new priority
areas were established to guide GILB activities far the next three years. The
highest-priority areas far GILB are breeding far host resistance, studying the
pathogen and host-pathogen relationship, promoting integrated pest management
(IPM), and providing training and infarmation.
Members at the GILB'99 Conference agreed that infarmation is a critica! need.
Therefare, GILB has concentrated its efforts on developing a Global Late Blight
lnfarmation System available online at the GILB web address (www.cipotato.org/
GILB). This online system includes
databases on Phytophthora infestans populations and their characterizations,
a directory of people interested and working in particular areas (indexed by
geographic area and theme),
a P. infestans bibliography,
links to other potato late-blight-related websites,
laboratory techniques,
avai lable research materials, and
links to relevant research databases.
A global catalog of late blight resistant patato varieties is being compiled and
will be available both online and as a CD-ROM. The catalog will include data
such as synonyms, parentage, breeder, releasing institution and date, whether it is
protected or not, source (where it can be obtained), and area of resistance (countries).
GILB promotes high-priority research by improving communications between and
among researchers and institutions and encouraging the transfer of technology.
Through GILB, research is facilitated by improving access to infarmation to late
blight working groups, especially in developing countries. The Global Late Blight

CIP Program Report 1999 - 2000

429

lnformation System and the GILB newsletter, along with sponsored symposia,
meetings, and workshops supported by GILB, all contribute to attaining these
goal s.

GILB members will next meet at the GILB Global Conference 'Late Blight:
Managing the Global Threat' to be held 11-13 July 2002, in conjunction with the
1 Sth Triennial Conference of the European Association of Potato Research (EAPR)
which will take place 14-19 July 2002 in Hamburg, Germany.

This report, particularly in the first seven papers in the potato section, show the
preadth and depth of CIP's research on Late Blight.

430

Enriching the Portfolio: CIP's Global and Regional Partnerships

SIUPA: A New lnitiative


The huge growth of urban populations in the developing world over the past thirty
years and the increasing dependence of city dwellers on agriculture have been
majar factors in the establishment of the CGIAR Strategic lnitiative on Urban and
Peri-urban Agriculture (SIUPA). CIP's role as SIUPA convener is based on more
than a decade of work with urban homegardens in the Philippines and severa!
years of collaborative work with rootcrop processing and pig-raising enterprises in
Hanoi, Vietnam. In Lima, Peru's capital city, CIP has also been involved in
nutrition-related studies. Finding common cause with many international and
national research and development efforts that have emerged in recent years, CIP
and other CGIAR Centers recognize the urgent need for increased research to
improve both the food security and poverty alleviation effects of urban agriculture
while ameliorating negative environmental impacts.
The following three papers describe the first outputs of this new lnitiative. Understanding and improving the environmental effects of waste water from rootcrop
processing and pig-raising enterprises in Hanoi require a range of disciplines and
institutional involvement. CIP and the lnternational Water Management lnstitute
(IWMI, Sri Lanka) have teamed up with a local environmental research institute
and local authorities to implement the research activities. The Philippines paper
emphasizes the diversity of institutions involved in urban agriculture and describes the integration of SIUPA into the broad coalition of national agencies
being assembled. Research on the dairy sector in peri-urban Lima demonstrates
the strong influence urban centers have on certain agricultura! sectors and explores possible changing scenarios in the sector given shifts in production
practices and in urban markets.

CIP Program Report 1999- 2000

431

Multi-Sectoral lnitiatives for Urban Agriculture in


Metro Manila, Philippines
D. Campilan 1, R. Boncodin 1, and C. de Guzman 2

Recent multi-sectoral initiatives to address needs and opportunities for urban


agriculture have led to the expansion of city farming activities in Metro
Manila, Philippines. The CIP-convened Strategic lnitiative for Urban and Periurban Agriculture (SIUPA) has contributed to improved coordination and
enhanced capacity for urban agriculture in the city. Local systems of urban
agriculture in Metro Manila range from backyard and continer gardens for
meeting subsistence food needs, to commercial vegetable farms for increased
household income. Through various policy, research, and extension interventions, public- and private-sector groups serve as key actors to support
household food security and livelihood of urban farming households. A key
element in this collaborative effort is the partnership between SIUPA and the
Philippines' lnter-Regional lntegrated Research, Development and Extension
Program on Urban Agriculture (llRDEP) in addressing key problem areas in
urban agriculture-food safety, production systems, and environmental
protection.
Urban agriculture has a long history in
Metro Manila, Philippines, but over the
past two years the number of agricultura!
activities in the metropolis has increased
dramatically. Prompted by the worsening
economic and environmental situation,
urban households have pursued urban
agriculture as a means for improving their
food security and livelihoods and the
environment. Extension services by
government agencies, research and
extension programs of academic institutions, poi icy development by national
legislative bodies, and coordinating
mechanisms of the national agricultura!
research system (NARS) all support urban
agriculture in Metro Manila. Opportunities
exist for collaboration between the CGIAR
Strategic lnitiative on Urban and Peri1
2

Uruan Agriculture (SIUPA) and the


Philippines' lnter-Regional lntegrated
Research, Development and Extension
Program on Urban Agriculture (llRDEP).
This report presents a comparative analysis
of priorities common to the two programs
and identifies potential collaborative
activities.

Overview of Metro Manila


The city of Manila evolved from a 16th
century Spanish settlement along the Pasig
River, a majar waterway flowing into the
South China Sea on the western side of
Luzon lsland. Early city dwel lers u sed the
river as a means of transport, but also
fished and grew food crops along its banks.
These pioneering activities were the
earliest forms of urban agriculture.

CIP-UPWARD, Los Banos.


University of the Philippines Los Banos, Philippines.

CIP Program Report 1999 - 2000

433

Since that time Manila has grown into a


bustling metropolis and has become the
country's seat of government, as well as its
business and financia! center.
In the early 1970s, Manila and 16 of its
adjoining cities and municipalities were
joined into a geopolitical unit known as
Metropolitan (Metro) Manila (Figure 1).
Subsequently, the Metropolitan Manila
Commission (MMC) was established
through Presidential Decree Number 824
issued in 1975. The MMC acted as a
central government in the Metro Manila
area, establishing and administering
programs and providing urban services,
and performing general administration,
executive, and policy-making functions,
including raising taxes. The MMC was
converted into the Metropolitan Manila
Authority (MMA) in 1990, and finally into
the Metro Manila Development Authority
(MMDA) in 1995.

Figure 1. Metro Manila and the Philippines.

434

Enriching the Portfolio: CIP's Global and Regional Partnerships

Like other urban centers in the developing


world, Metro Manila is beset by problems
of high population density, poverty, and
environmental pollution. As the national
capital region (NCR), Metro Manila
occupies a total land area of 636 sq. km.
Among the 1 7 component cities and
municipalities, Quezon City ranks first in
terms of land area and population (Table
1 ). lts popu lation is estimated at 1 0.5
million, and is growing at an average of
2.3% annually. Government statistics
indicate a population density of 14,865
persons/sq km, an estimated 1 O mi Ilion
people.
According to the 1995 census of the
National Statistics Office, Metro Manila
had almost 2 million households, 40% of
which were below the poverty line (set in
1997 at US$100.83 (P4840) (the 2000
exchange rate was P48 = US$1) per month
far a family of five) (Mangahas, 1999).
Thirty-five percent of the urban population
live in slum dwellings or squat on private

Table 1. Metro Manila's cities and municipalities,


area and population (Philippine Statistical
Yearbook, 2000).
City/municipality
Land area Population
(sq km)
('000)
National Capital Region
636.0
10,492
Kalookan
55.8
1,233
Las Pias
41.5
499
Makati
29.9
524
Malabon
23.4
356
Mandaluyong
26.0
304
Manila
38.3
1,673
Marikina
437
38.9
Muntinlupa
46.7
393
National Capital Region
636.0
10,492
Navotas
254
2.6
489
Paraaque
38.3
Pasay
13.9
363
13.0
Pasig
582
10.4
Pateros
64
Quezon City
166.2
2,160
San Juan
10.4
129
Taguig
33.7
510
Valenzuela
47.0
521
or public land. Half of the population of
Quezon City, the largest and most populous of Metro Manila's 17 cities and
municipalities, 160,000 families, are
squatters (Banal, 2001 ).
In 1997, the annual per capita poverty
threshold (the amount of money needed to
satisfy basic nutritional requirements (2000
calories per day) and other basic needs) for
Metro Manilans was about US$298
(P14,299); this is higher than in any other
region in the country. More than 800,000
people had incomes below this threshold.
The unemployment rate in Metro Manila is
the highest in the country at 16% (more
than 700,000 persons) (Philippines Statistical Yearbook, 2000).
Compared to other regions in the Philippines, food expenditure per family is
higher in Metro Manila (US$1162
(PSS,786)) than in the rest of the country
(US$576 (P27,645)) (Department of Agriculture, 1997). About 10% of this amount
is spent on vegetables, fruits, and roots and
tubers; 28% on meat and dairy products;
and 12% on fish products. Non-food items

such as education and housing account for


the other half of the household's average
expenditures.
With access to food in adequate quantity
and quality seriously limited for urban
households, it is not surprising that nearly
one-third (32%) of children in Metro
Manila are underweight and one-fifth
(22%) have stunted growth (Laa et al.,
1998). lron, vitamin A, and iodine deficiencies, the most prevalent forms of
undernutrition in the Philippines (Bayani et
al., 1999), have been attributed to a
decreasing consumption of vegetables,
particularly in urban areas where the
average daily per capita vegetable
consumption is reported at 133 g/day
compared with 149 g/day in rural areas.
Villavieja (1989) reported that the major
factors contributing to the declining trend
in vegetable intake were changing
lifestyles and dietary habits as a result of
increasing urbanization, lack of supply,
high price of vegetables, and diminishing
purchasing power.
As a result of poor urban planning and
management, Metro Manila faces serious
problems of traffic congestion, pollution,
and poor waste disposal. A 1999 survey by
Asiaweek ranked Metro Manila 25th out of
40 of the most livable cities in Asia, lower
even than the Philippine cities of Davao
(ranked 17th) and Cebu (22nd). Metro
Manila suffers from severe air pollution 90 mg/m 3 suspended particles compared
with the best city in Asia, Fukuoka, Japan,
at 26 mg/m 3 (Asiaweek, 1999). Metro
Manila produces over six tons of solid
garbage everyday. Nearly half (45%) of
this is kitchen wastes that can be
composted and most of the rest (41 %) is
glass, metals, plastics and paper that could
be recycled (MMDA, 2001 ).

Local lnitiatives for Urban Agriculture


A survey carried out more than a decade
ago revealed that only about 14% of the
households surveyed in Metro Manila
engaged in food production through

GIP Program Report 1999 - 2000

435

backyard gardening. About 83% of the


respondents gave lack of available space
as their reason for the absence of backyard
gardens (Villavieja, 1989). In the past few
years, however, there has been a rapid
expansion of urban agriculture activities in
the metropolis, stimulated by the worsening economic and environmental situation.
Urban agricultura! in Metro Manila is
either individual or communal, and for
subsistence or commercial purposes.
Urban garden systems range from growing
crops in recyclable containers to commercial-scale crop production.
Container farming is particularly popular
in densely populated areas of Metro
Manila such as Quezon and Makati; in the
slum areas, Mandaluyong, Paranaque, and
Las Pias; and in flood-prone areas,
Malabon and Valenzuela (Duldulao,
2001 ). Vegetable crops are grown in
recycled tin or plastic containers placed in
the yard, on windowsills, and on rooftops.
Portability is an important advantage of
container gardening, because the plants
can be easily moved in times of flooding
or for safekeeping or aesthetic purposes.
The MMDA, other government agencies,
and NGOs have organized households to
establish communal gardens. Successful
examples are found in Muntinlupa,
Quezon, and Mandaluyong where groups
of householders are assisted in obtaining
access to vacant areas in their communities and provided with training and access
to inputs.
An emerging phenomenon in Metro
Manila is the proliferation of commercialscale urban gardens. These are found in
large housing subdivisions, such as in
Quezon, Las Pinas, and Paranaque, where
there are large areas of unused, prvate
land. In North Fairview, Quezon City, for
example, more than 70% of the land in an
undeveloped 50-ha subdivision has been
cultivated by poor householders in the
neighborhood. Residents of nearby slum
areas use these prvate lots to grow

436

Enriching the Portfolio: CIP's Global and Regional Partnerships

vegetables that are sold in Metro Manila


markets. These urban farmers enter into
informal, temporary land use arrangements
with prvate landowners and the
homeowners' association, often with local
government agencies acting as mediator
and guarantor.
The most common crops grown in these
commercial urban farms are quick-growing
leafy vegetables such as pechay, mustard,
lettuce, broccoli, green onion,
sweetpotato, upland kangkong and
alugbati (Malabar spinach). Coriander,
chives, and celery are popularly cultivated
herbs. In most cases, the choice of crop is
determioned by the financia! resources of
the farmer. Fresh produce is immediately
sold or delivered to nearby wet markets or
to major supermarkets within Metro
Manila that cater to high-income consumer groups. lt is also not uncommon for
traders and middlemen to bid for vegetables in garden plots befare harvest.
What is not certain is what percent of the
produce from these gardens is directly
consumed by urban farming households.
Governmental extension services
In 1998 the government launched a formal
national Urban Agriculture Program (UAP),
to be coordinated by the Office of the
Presidential Assistant on Food Security
(OPAFS) in partnership with eight government agencies and non-governmental
organizations.
By virtue of a special arder from the
Office of the Department of Agriculture
Secretary, OPAFS spearheads and coordinates implementation of various urban
agriculture programs in the Philippines,
including the llRDEP. In Metro Manila, the
program is jointly implemented by the
Department of Agriculture Regional Field
Unit, various local government units, and
the Metro Manila Development Authority
(MMDA).
The UAP program targets urban agricultura! activities at three levels:

household, such as backyard and


container gardens,
community, such as garde,ns and farms
jointly managed by local informal
groups, and
institutional, such as farms managed by
schools, civic, and religious groups and
cooperatives.
UAP supports farmers by:
negotiating the use of public and
prvate lands for use by urban farming
groups,
providing training in crop management,
providing planting materials and access
to financia! support for needed inputs,
mobilizing resources to set up support
infrastructure, and
monitoring and evaluating programs.
By December 2000 the program has
supported 1596 beneficiaries-householders, government employees, and
students-in cultivating a land area of
129,409 m 2 (Table 2) (Duldulao, 2001 ).
Private sector involvement

Several NGOs and prvate-sector groups


promote and support urban agriculture in
Metro Manila, but often as part of broader
programs on health and nutrition, environmental sanitation, and cleanliness and
beautification campaigns. Among these
are two programs under prvate foundations set up by the Office of the First Lady
in two recent adm in istrations: the Piso
Para sa Pasig (One Peso for Pasig River)
which raises funds for waste management
for the Pasig River, and the Gulayan at
Bulaklakan (Vegetable and Flower Gardening Group) which conducts information
campaigns and provides technical/
resource support to com01unity vegetable
and ornamental gardens in the metropolis.
The organic farming movement in the
Philippines has spurred the establishment
of commercial farms within and around
Metro Manila, producing organically
grown food and ornamental crops, live-

Table 2. Coverage of the UAP in Metro Manila as of


December 2000 (Duldulao, 2001 ).
City/
Area Total beneficiaries
municipality (sq m)
Kalookan
20,000 50 households
Las Pinas
1,300 200 households
Makati
320 1O households
Malabon
20,000 62 households
Mandaluyong 10,000 30 households
Manila
28,000 130 students and
employees
Muntinlupa
15,000 30 households
Paraaque
10,450 50 households
Pasay
450 50 government employees
Pasig
300 100 students and
government employees
Pateros
2,000 50 households
15,789 700 households
Ouezon City
San Juan
500 30 government employees
Taguig
800 54 households
Yalenzuela
4,500 50 government employees
129,409 1,596
Total
stock and dairy products, and planting
materials and fertilizers. Weekend markets, organized in two locations in Metro
Manila, serve as outlets for organic
farmers.
The Philippines' organic farming industry
is still in its infancy, contributing less than
1% to total agricultura! output, but it is
growing by between 10% and 20%
annual ly. Domestic production is about
Pl 50 mi Ilion (US$1 = P48) and imports of
processed organic food products amount to
about P150 mi Ilion (Canono, 2000).
The Organic Producers and Traders Association (OPTA) has about 350 members
throughout the country, and is affiliated
with the lnternational Organic Farming
Movement (IOFM). lt is estimated that
more than half of OPTA's output is sold
and consumed in Metro Manila.
OPTA conducts training and seminars,
publishes a quarterly magazine, provides
technical consultancy services, and
lobbies for policies supporting organic
farming, such as incentives for producers
and traders, certification and accreditation, and food safety regulations. Recently,

CIP Program Report 1999 - 2000

437

it has initiated the piloting of urban


organic farming models called

Hanapbuhay Organiko Pang Entreprenor


(Organic Livelihood and Entrepreneurship)
or HOPE.
Research and extension support
academic institutions

by

Aside from school gardens establ ished to


support food production in urban communities, academic institutions within or close
to Metro Manila have also undertaken
research and extension activities to
develop and field-test urban farming
technologies.
Research at the Central Luzon State
University (CLSU), north of Metro Manila,
has led to the development of receptacle
farming. Using a 300-m 2 farm, researchers
have generated technologies that address
the limitations of land, light, and water
that confront urban agriculture. Since
1999, CLSU has evaluated various recyclable containers (such as rubber tires, tin
cans, plastic containers, sacks, and junk
metal items) for growing 60 varieties of
vegetables, spices, cereals, and edible
ornamentals (Nitural, 2001 ).
In partnership with the Asan Vegetable
Research and Development Center
(AVRDC), CLSU also conducts a peri-urban
vegetable research and development
project, whose main goal is to provide
Metro Manila with a regular supply of
fresh vegetables and other perishable farm
products. Funded by the German government, the project fol lows a systems
approach and emphasizes farmer participatory research on crop management, pest
and disease control, soil and crop nutrition, and socioeconomics.
In 1995, the University of Philippines at
Los Baos (UPLB) initiated an Urban
Agriculture Planning Workshop to discuss
various aspects of local urban agriculture,
identify gaps in present research and
development, and formulate proposals for
bridging those gaps (Urban Agriculture
Research and Development Committee,

438

Enriching the Portfolio: CIP's Global and Regional Partnerships

1995). One offshoot of the workshop was


creation of the Urban Agriculture Project
(UAP) under UPLB's Agro-Industrial
Program of the Col lege of Agricu ltu re.
UAP served as an extension project aimed
at strengthening the capabilities of local
government units, state col leges and
universities, and NGOs in the transfer and
promotion of urban agricultura! technologies (De Guzman and Banatlao, 1999).
The project was formally launched and
implemented in Los Baos, Calamba, and
Sta. Rosa (which are municipalities in the
process of urbanizing), and later in
Quezon City, Metro Manila.
Aside from this effort, UPLB has established a 350-m 2 demonstration garden
displaying protected (covered) cultivation
of vegetables and various models of
growing containers. To further promote
urban agriculture among its students, the
university has also recently instituted an
undergraduate course in urban horticulture.
Similarly, Cavite State University (CvSU)
has been promoting various urban agricultura! technologies at its Sanayan sa

Kakayahang Agrikultura (Agricultura/


Ski/Is Training) (SAKA) project site.
Examples on show include protected
(greenhouse) and semi-protected (net
houses and net-covered growing beds)
cultivation and soil-less culture, including
hydroponics of vegetables and cut flowers.
The university has extended technical
support services to various cooperatives
involved in cut-flower production in
nearby Tagaytay City and. other peri-urban
areas in the province.
Over the past three years, an urban school
farm has been piloted by the Operation
Brotherhood Montessori Center at its
campus in Las Pinas City, in southern
Metro Manila. The 6000-m 2 farm is the
offshoot of a partnership with Kasetsart
University in Thailand to promote exchange of expertise and experiences in
school-based farming. Managed by
teachers and students, the school farm
produces commercial horticultura! crops,

particularly premium spices, as well as


compost and planting materials. Part of the
farm produce is suppl ied to food establ ishments in Metro Manila, and the rest is sold
at the school market. Another 5-ha periurban farm is being establ ished in nearby
Cavite Province to provide an area for
expanding the pilot integrated farm model.

Development of a supportive policy


framework
Traditionally, priority in agricultura! policy
development in the Philippines has been
given to the rural sector. More recently,
however, key policy initiatives have been
undertaken to pass legislation and formulate policy guidelines to support the
development of agriculture in the urban
sector.
In 1999, the Committee on Agriculture of
the Philippines House of Representatives
initiated deliberation on urban agriculture
policies, with Quezon City Congressman
Dante Liban leading this movement. A
series of consultations was conducted with
state colleges and universities to determine institutional capacities for carrying
out urban agriculture programs. The
underlying goal is to achieve legislation
authorizing allocation of part of the
Department of Agriculture's budget for
urban agriculture efforts.
Meanwhile, under the new Agriculture and
Fisheries Modernization Program (Republic
Act 8435 of 1998), the Bureau of Agricultura! Research (BAR) has been tasked to
coordinate, plan, monitor, evaluate, and
integrate all relevant research and development activities in the country. This BAR
mandate covers both rural and urban
agriculture, and has led to the formulation
of an urban agriculture research, development, and extension agenda (discussed in
the following section).
Although general policy support to urban
agriculture is in its early stages, there have
already been policy interventions specific
to urban forestry. As early as 1988, urban
forestry gained formal recognition through

Memorandum Orders 198 and 199 of the


Department of Environment and Natural
Resources (DENR). These served as the
basis for the creation of a DENR Urban
Forestry Division in Metro Manila. Subsequently, six cities in Metro Manila
established their respective Greening
Offices. The Luntiang Kamaynilaan (Green
Metro Manila) Program planted 100,000
tree seedlings over two years (although the
survival rate was only 65%).
Subsequently, the 1990 Master Plan for
Forestry Development provided for peopleoriented forestry designed to raise the
quality of the environment in urban
centers in the Philippines (DENR, 1990).
Under this policy, a tree-to-person ratio of
1 :4 and a forest park-to-person ratio of
1:100,000 were envisioned. In Metro
Manila, the target was to establish at least
60 forest parks. Efforts to achieve these
targets are still underway, mainly through
the various city and municipal governments in Metro Manila and through the
coordination of the MMDA.
Meanwhile, recent legislation passed in
the Philippines Congress provided for the
formulation and implementation of
appropriate measures for solid waste
management and control of air pollution.
The Ecological Solid Waste Management
Act (Republic Act Number 9003 of 2000)
established the legal basis for an integrated ecological approach to solid waste
management. lt requires city and municipal governments to reduce their waste
disposal by 25% within five years through
segregation, recycling, composting, and
re-use. Local government units are to be
assisted by a newly created National
Commission on Solid Waste Management,
which has been given the task of carrying
out research and supervising agricultura!,
hospital, household, and industrial waste
disposal nationwide. The law also requires
the mandatory separation of garbage at its
source and the integration of waste
management in school curricula
(Bainbridge, 2001; Donato, 2001 ).

GIP Program Report 1999 - 2000

439

Republic Act Number 8749, also known as


the Philippine Clean Air Act of 1999,
provides for the institution of a holistic
national program of air pollution management within the framework of sustainable
development. This law emphasizes the
need for pollution prevention rather than
control, and the setting up of a guarantee
mechanism for clean-up and environmental rehabi 1itation.
Research, development and extension
coordination

For the first time in the history of the


Philippines national agricultura! research
and development system, an integrated
program on urban agriculture was initiated
by BAR in 2000. The lnter-Regional
lntegrated Research, Development, and
Extension Program (llRDEP) on Urban
Agriculture is a five-year program (BAR,
2000) established through a network of
research, training and extension, and
academic institutions working on urban
agriculture issues.
An inter-agency Core Planning Team (CPT)
was formed and given the task of formulating a program plan that spells out priorities
for research, development and extension.
The CPT identified three key problem
areas for urban agriculture in the Philippines: low production and productivity; a
weak marketing system; and environmental degradation. Consequently, a research,
development, and extension plan, due to
be launched in 2001, was developed. lt
focuses on food safety, an input management system, production systems, saving
biodiversity, socio-economics and marketing, environmental protection, manpower
capability building, and policy advocacy
(Table 3).
Opportunities for Collaboration with

SIUPA
The CGIAR Strategic lnitiative on Urban
and Peri-urban Agriculture (SIUPA) has
selected Metro Manila as one of its global
network of field research sites. SIUPA has
worked closely with the CPT of the

440

Enriching the Portfolio: CIP's Global and Regional Partnerships

Philippines' llRDEP on Urban Agriculture,


particularly in identifying possible areas
for col laboration. A comparison of the
action plans drafted by the two programs
shows three key aspects of the llRDEP with
potential links to SIUPA's program agenda:
food safety, production systems, and
environmental protection (marked by bold
lines in Figure 2). SIUPA's program
framework for urban/peri-urban agriculture
(UPA) consists of five thematic areas for
R&D (i.e., supply of UPA products, UPA
and livelihoods, environment and health,
agricultura! and non-agricultura! uses, and
cross-cutting themes) which are subcategorized into three levels of R&D
interventions (i.e., household/community,
institutional, and policy).
Using SIUPA's program framework, the
CPT analyzed llRDEP's list of priority
projects in terms of prioritization, coverage, and distribution (Table 4). llRDEP
concentrates on two themes in urban
agriculture that are of global significance:
supply of urban/peri-urban agriculture
products and environment and health
(Table 4). Moreover, it tends to work
mainly at the level of households and
communities, whereas there is very
limited effort to address institutional and
poi icy issues.
SIUPA-llRDEP collaboration has been
formalized through a project called
Capacity Development to Urban Agriculture Research Development and Extension
in the Philippines.
The collaborative project seeks to:
help the llRDEP develop link with
programs and organizations supporting
global efforts to promote urban agriculture,
identify and respond to training-related
needs of individuals and institutions
involved in the llRDEP, and
develop mechanisms for information
dissemination and exchange relevant to
the program agenda of the llRDEP.

Table 3. llRDEP objectives and proposed strategies (BAR, 2000).


Objectives
Strategies
1. To establish the status of urban
Gather information through surveys and documentation of
agriculture in the country.
existing urban agriculture projects.
Review literature on urban agriculture.
Conduct national conferences and workshops.
2. To adapVdevelop technologies for
ldentify, collect, verify, package, and pilot test technologies on
crops, animals, and fisheries to
urban agriculture.
Select and develop crops, livestock, and fish for the urban
urban environment.
environment.
Develop production technologies for limited space and adverse
conditions.
3. To enhance the capabilities of local
Disseminate information through technical and popular
government units, cooperatives
publications.
Conduct multi-level training among stakeholders.
and individuals in managing
agricultura! and fisheries production Establish technology demonstration projects on urban agriculture.
Hold consultative briefing and orientation.
and processing in urban areas.
lnitiate community organizing.
4.To develop positive behavioral
lnstitute social mobilization.
values among urban agriculture
Hold agricultura! exhibits and contests.
practitioners.
Conduct nutritional deficiency survey.
5. To improve the nutritional status of
Introduce, promote, and produce nutritious vegetables.
urban households.
Monitor and evaluate nutritional status of the participants.
Convert urban waste into compost.
6. To promote a sustainable
Treat and recycle household waste water and city sewage water.
environment through urban
Use recycled containers as production input.
agriculture.
Use botanical pesticides and biological predators as methods of
insect and disease control.
Analyze pesticide residue and toxic metals in urban farms.
Develop test kit for detecting chemical residues in vegetables.
Assess and develop the production and marketing scheme.
7. To determine and develop the
Establish market information center and linkages.
production and marketing scheme
on urban agriculture.
8. To advocate policies far sustainable Prepare and provide policy papers far decision makers.
urban agriculture.

llRDEP
Socio-economics and marketing
Food safety
Production systems
Environmental protection
Input management system

SIUPA
Supply of perishable/processed products
Urban agriculture and livelihoods
Environment and health
Agriculture and non-ag uses
Cross-cutting themes

Saving biodiversity

Figure 2. Links between action plans of llRDEP and SIUPA.

CIP Program Report 1999 - 2000

441

Table 4. Mapping IRRDEP projects using the SIUPA program framework (Campilan, 2001).
Theme
Household/community
lnstitutional
Policy
Supply of UPA
Development of year-round vegetable
products
production systems.
Development of fertilizer, growing media.
Edible landscaping.
Receptacle/container gardening.
lmprovement of varieties of vegetables.
UPA and
Development of marketing strategies
Development of marketing
livelihoods
and assessment of product flow.
strategies and assessment
of product flow.
Environment and Analysis far toxic heavy metals.
health
Assurance of toad safety in edible
crops.
Detection of pesticide residues.
Collection and selection of crop
varieties.
Recycling of river and sewage water.
Development of odorless compost
making.
Agricultural/nonagricultural uses
Cross-cutting

Four sets of future activities were proposed.


Training on participatory methods for
assessing urban agricultura! systems.
Participation in international conferences and trainings on urban
agriculture.
Literature support to urban agriculture
research, development and extension.
Monitoring and evaluation of capacity
development.
Capacity development is critica! to
llRDEP's success, for three main reasons.
1. Urban agriculture is a new program area
for research, development and extension in the Philippines. As a pioneering
effort in the country, one of the primary
challenges of the llRDEP is to facilitate
the creation of a critica! mass of urban
agricultura! professionals and institutions.
2. Urban agriculture draws on knowledge
and expertise from a variety of disciplines (agriculture, environmental

442

Enriching the Portfolio: CIP's Global and Regional Partnerships

Generation of data on UPA Generation of


lmpact of GO, NGOs Ol UPA. data Ol UPA.
studies, health and nutrition, social
sciences, management and poi icy
planning, etc.) Those who now work on
urban agriculture received training in
these specialized areas. There is a need
to develop their capacities toward a
more holistic and integrated approach
to urban agriculture research, development, and extension.
3. To effectively carry out its program
agenda, llRDEP-Urban Agriculture
requires adequate information support.
However, this is constrained by limited
efforts and resources in the Philippines
to create databases, provide literature
access, and disseminate relevant
information.

References
Asiaweek. 1999. Asiaweek best cities
1999. http://www.asiaweek.com/
asiaweek/features/asiacities/acl 999/
data/manila.html
Banal, Conrado R. 2001. Breaktime.
Philippine Daily lnquirer, 20 February
2001 issue. p. 82.

Bainbridge, A. 2001. Garbage act is first


law signed by GMA. Philippine Daily
lnquirer, 27 January 2001. p. 1.
BAR (Bureau of Agricultura! Research).
2000. lntegrated inter-regional research,
development and extension agenda for
urban agriculture in the Philippines.
Proposed 2000-2002 workplan. BAR,
Quezon City, Philippines.
Bayani, E.M., M.B.T. Flores, M.R. Enteria,
and M.L.A. Vega. 1999. NNC's role in
promoting vegetable production and
utilization. Paper presented at the
Symposium-Workshop-Exhibit on
Vegetables by the Society for the
Advancement of the Vegetable lndustry,
21-22 January 1999 in Los Banos,
Laguna, Philippines. (unpublished)
Campilan, D.M. 2001. Global trends in
urban agriculture: lnsights, issues and
initiatives. Paper presented at the
National Conference on Urban
Agricultura! Systems in the Philippines,
15-1 7 January 2001, Quezon City.
(unpubl ished)
Canono, J.F. 2000. Philippines: Organic
products-organic markets, brief 2000.
USDA Foreign Agricultura! Service
GAi N Report #RP001 5.
De Guzman, C.C. and P.P. Banatlao. 1999.
The Urban Agriculture Project of the
Agro-Industrial Development Program of
UPLB-CA: Promoting agriculture in
urban communities. Paper presented at
the AIDP-PITAS Lecture Series, 8 Jan.
1999, Operations Room, Samonte Hall,
UPLB. (unpublished)
Delfinado, RD. 1998. food remains top
item in Filipino family's expense list.
The Philippine Star, 9 Dec. 1998.
DENR (Department of Environment and
Natural Resources). 1990. The 1990
Master Plan for Forestry Development.
DENR, Quezon City, Philippines.
Donato, A.E. 2001. Abalos wants
campaign on waste management.
Philippine Daily lnquirer, 1 Feb. 2001.
p. 1.
Duldulao, V.A. 2001. Gulayan at
Bulaklakan: A component of the

government's urban agriculture


program. Paper presented at the
National Conference on Urban
Agricultura! Systems in the Phil ippines,
15-17 January 2001, Quezon City.
(unpublished)
Department of Agriculture. 1997. The food
and agribusiness yearbook, Dept. of
Agriculture, Quezon City, Philippines.
Laa, R.O., G.M. Villavieja, and
C.M. Cerdea. 1998. Comparison of the
nutritional status of 0-59 month-old
Filipino children using Philippines and
lnternational Standards. DOST-FNRI,
Taguig, Metro Manila, Philippines.
Mangahas, M. 1999. Monitoring Philippine
poverty by operational social indicators.
Paper presented to the World Bank's
Poverty Reduction and Economic
Management (PREM) Network, PREM
Week 99, 13-14 July 1999, University
of Maryland, Maryland, USA.
(unpublished)
MMDA. 2001. Official website. http:
wwww.mmda.org.
Nitural, P.S. 2001. Receptacle farming jn
urban agriculture: The CLSU model.
Paper presented at the National
Conference on Urban Agricultura!
Systems in the Philippines, 15-17
January 2001, Quezon City, Philippines.
(unpublished)
NSCB. 2000. Philippine Statistical
Yearbook. NSCB, Quezon City,
Philippines.
Asiaweek. 2000. The Asiaweek Quality
of Life lndex. 15 December 2000.
p. 46-47.
Urban Agriculture Research and
Development Committee. 1995. Urban
Agriculture Research and Development
Committee report. UPLB College of
Agriculture, Los Banos, Philippines.
Villavieja, G.M. 1989. Vegetable
consumption and nutrition. In:
Proceedings of the National Vegetable
lndustry Review and Planning
Workshop, 15-17 November 1989,
PCARRD, Tagaytay City, Los Banos,
Laguna, Philippines. p. 235-248.

CIP Program Report 1999 - 2000

443

Peri-Urban Milk Production in Peru: Farm Strategies


and Policy Options
T. Bernet1 , C. Gomez 2, J. Julca 2, G. Prain1, J. Senz2

This study assesses past, present, and future development trends of the dairy
sector in the surroundings of Lima, Peru. Results show that peri-urban milk
producers currently enjoy favorable market conditions, thanks to increased
competition among local milk buyers, who seek to exploit spare processing
capacity. Yet, this situation is likely to change in the near future as milk
production rapidly expands. The resulting drop in price will be likely to put
many small and medium farms out of business, whereas large farms will be in
a better position to confront lower prices, as a result of considerably lower
average production costs (economies of scale) and higher productivity,
mainly due to better feed quality. The only viable strategy for small and
medium milk producers is herd growth and the creation of producer associations to strengthen their market position versus suppliers of inputs and
purchasers of milk. Nonetheless, improvements in feeding and reproduction
practices will be essential to achieve faster herd growth. For large farms, the
key to improved profitability will be to efficiently manage the different
contracted specialists involved in production.
Peru, like other developing countries, has
experienced strong urban growth during
the last few decades. Today, 72% of its 25
million inhabitants live in cities. Metropolitan Lima alone makes up 30% of
Peru's total population (INEI, 2000). As
migration from mountain regions to coastal
urban centers continues, urban consumer
preferences increasingly determine the
production priorities of the domestic
agricultura! sector, e.g., where production
and processing takes place, what farm
strategies are implemented, and what
technologies are applied. The geographic
proximity between producers and consumers is particularly important for fresh milk
and its derivatives, because of their high
perishability and bulkiness. This paper
1
2

CIP, Lima, Peru.


Universidad Nacional Agraria La Molina, Lima, Peru.

assesses how the dairy sector around Lima


is likely to evolve and what farm strategies and policy options would most
meaningfully confront future challenges.

Methods
Secondary literature, farm surveys (Julca,
2000; Senz, 2000) and census data (INEI,
1995), and milk collection statistics were
used to characterize Lima's dairy sector.
Based on that, three typical dairy farms
were defined (in terms of herd size): small
(1 O cows), medium (35 cows), and large
(220 cows). A farm-household optimization
model (Bernet, 2000) was used to assess
current profitability levels and the impact
on profits and production driven by
potential production context changes (e.g.,
feed prices and qual ity, herd management,
milk prices, and production levels). The

CIP Program Report 1999 - 2000

445

three farm types were modeled separately,


to take into account the particular production setting in each case (Table 1 ), e.g.,
production levels, feed quality, and
product and factor prices. Smal 1 and
medium farms are pure stables; the large
farm includes an additional 1O hectares of
cropland for fodder production.

0.40
0.35
Vi
(/)

2.
Sf
(j
c.
CD

0.30
0.25

()

0.20

Results
0.15

Characteristics of Lima's dairy sector


The city of Lima is by far the most important market for the national dairy industry.
Lima's inhabitants account for more than
80% of national consumption of industrially-processed dairy products (McBride
1997). The capital city's market dominance for fresh dairy products is likely to
continue, given the continued growth rate
of the city of more than 6% annually.
Consequently, as a production region,
Lima has the advantage of having its
market next door. This geographic advantage has led to the recent establishment of
additional milk processing plants by
different companies in peri-urban Lima
(USDA 1999). Hence, farm-gate milk
prices are considerably higher in Lima
compared to the other milksheds (Figure
1). These prices reflect two things: first,
'geographical protection', because milk
from other milksheds faces higher transportation and refrigeration costs when shipped
to Lima; and, second, increased competi-

1992 1993 1994 1995 1996 1997 1998 '1999

Figure 1. Development of real producer prices far milk


(1992-1999) in the most important Peruvian
milksheds (base = Dec. 1999). Source: Ministry of
Agriculture (1992 to 1999).
tion between local milk buyers trying to
reduce their current processing overcapacity with higher milk purchases. Also,
municipalities have stimulated the stronger competition for milk purchases as they
have started to use locally-produced milk
for government-subsidized food programs
in the surroundings of Lima, e.g., Vaso de
Leche, which provides milk to underprivileged children in shantytowns.
The wide variation in herd size among the
dairy farms near Lima is explained by the
vast supply of feed sources in the region.
There is an extensive market for fresh corn
and ingredients for feed concentrate (e.g.,
cotton seed, fish meal, and leftovers from
asparagus and flower processing).

Table 1. Characterization of small, medium, and large farms in coastal Peru near Lima.
Farm type
Small
Medium
Size of herd
< 20
20 to 100
Milking technology
by hand
by hand/machine
Artificial insemination
no/(yes)
yes/(no)
Labor force
family
contracted
On-farm specialists
no
no
High-quality feed mix
no
no/{yes)
Own silage production
no
no
Milk production (l/lactation)
3650
6800
Body weight of cow (kg)
450
650
First insemination date (mo.)
20
17
Calving interval (mo.)
> 15
14
Mortality rates
high
low
Sources: Avni, 1996; Julca, 2000; Senz, 2000.

446

Enriching the Portfolio: CIP's Global and Regional Partnerships

Large
> 100
machine
yes
contracted
yes
yes
yes/(no)
7800
650
17
14
low

Because stall-feeding dominates, the


majority of farmers keep more animals
than they can feed with their own land. In
fact, most dairy farms are no more than
cattle stables. Located on low-cost desert
land, these farms purchase fodder and feed
concentrate daily or weekly from specialized producers or intermediaries. They sell
manure to regional crop growers every
couple of months. The difference between
smaller and larger farms is marked by
differing milk production and technology
levels, and varying feed quality and
management standards (Avni, 1996). In
general, al 1 these aspects improve as farm
sizes become larger (in terms of herd size)
(Table 1 ).

Description of the farm types

livelihood security. Milk production is a


particularly important income source for
elderly people, who have limited options
for other work (Julca, 2000).

Medium and large farms have far better


feed and reproduction practices. They
commonly use corn with cobs and feed
concentrate m ixes that they prepare onfarm. Since the dairy cattle on these farms
are grouped by age and lactation stage,
the quality and quantity of feed are
adjusted to the animals' specific nutrient
requirements. Family labor is commonly
replaced with contracted labor in these
stables, including different specialists such
as veterinarians, nutritionists, and accounting personnel. The main differences
between medium and large farms are
lower average production costs and higher
milk yield per cow. The latter is driven,
first, by the higher quality of feed farmers
are able purchase thanks to their stronger
position to negotiate, and, second, by
better herd management due to access to
better qualified specialists.

Small farms are run by poor people within


urban settlements (Figure 2). Milk production is liquidity-constrained and deficient
in both feeding and reproduction practices.
This is expressed in low body weights,
delayed first-insemination dates, and high
mortality rates of the cattle. Small farms
are pure stall-feeding systems with miniOverall, medium farms are more equitymum stable infrastructure, located next to
constrained in comparison to larger farms,
farmers' houses. Feed qual ity is low,
which are stronger financially. Therefore
because smal 1 farmers strongly depend on
the infrastructure and machinery used for
local intermediaries and the feed quality
milk production is much more advanced
they offer. Because farmers often lack
_ on large farms. Cows are always milked
access to high-quality feed and equity,
by machine, in contrast to small and
they mainly use corn without cobs and
medium farms, where hand milking is
wheat bran to feed their dairy cattle.
common (Senz, 2000). Because larger
Often, milk production is a complemenstables general ly possess crop land, they
tary economic strategy to improve
grow part of their feed themselves.
Current profitability levels of dairy farms

Figure 2. An example of a small farm in Lurin, a costal


town 1Omiles south of metropolitan Lima.

Model results show remarkable effects of


economies of size among different farm
sizes; profits dramatically increase with
larger herd sizes (Figure 3). The increased
profit per cow in larger farms is primarily
related to better feed and herd management, as expressed by i ncreased body
weight, milk production, and herd fertility
(e.g., lower mortal ity rates, shorter calving
intervals, earlier first inseminations). This
is largely due to the hiring of nutrition and

CIP Program Report 1999- 2000

447

us - $

....----------.
2CXX)

Meat

O Milk
1500

lnfrastructure

Livestock
fZ.I sanitation

500

lnc~

(JIB Capital

l
1CXXJ

Labor
ITT Fodder (corn
t:;J with cobs)
EEI Feed
~ concentrate

Costs

Small
Medium Large
(10 cows) (35 cows) (220 cows)

Figure 3. lncome and cost structure per cow of small,


medium, and large farms.
reproduction specialists. A second factor is
lower input prices, which is a result of
improved negotiation power, particularly
for purchasing feed. lmproved economies
of scale also significantly lower average
costs for labor, infrastructure, and capital.
lmpact of feed and herd management
Feed plays a predominant role in all farms.
Feed not only accounts for more than 50%
of direct production costs, but also represents considerable indirect costs due to the
labor required to organize, mix, and
distribute it.
The quality of corn is crucial. On small
and medium farms, profits decrease by
around 30% when corn is used without
cobs, despite its lower price. Profits are
also hit when very old corn with poor
nutrient quality is used. When they lack
their own crop land, small and medium
farms are forced to purchase higher levels
of relatively expensive feed concentrate to
maintain a balanced nutrition for their
animals. On large farms, profit decreases
only by approximately 10% when cob-less
corn is purchased. Here, own high-quality
corn compensates to sorne extent for the
deficient nutrient content of purchased
low-quality corn.
Different herd management practices, as
described earl ier, also ha ve a substantial
effect on profits. With major improvements

448

Enriching the Portfolio: CIP's Global and Regional Partnerships

in herd management, profits can be


increased by as much as 50% on all farms .
Nonetheless, milk prices and milk production levels have the strongest impact on
farm profits. The lower the current profits,
the greater is the effect. A milk price
increase/decrease of 15% affects small
farmers' profits by plus or minus 60%. A
milk price collapse of this magnitude
therefore threatens the viability of milk
production in small and medium farms.
Large farms on the other hand are in a
better position to withstand such milk
price reductions. Although a decrease in
milk prices of 30% could induce profit
losses of almost 60%, large farms with an
average of 220 cows would still gain
approximately US$370 per cow per year,
or US$80,000 in total.
However, it is important to emphasize that
milk prices are more profit-relevant than
milk production levels. Whereas an
increment in milk price results in a direct
increment in profit, an increment in milk
production also induces higher feed costs
to reach the higher production level.
Therefore, an increase in milk production
is only half as beneficia! as an equal
increase in milk price.

Conclusions
Promising farm strategies
Milk prices in Lima will tend to fall when
the Lima market reaches a certain level of
saturation for bulky and perishable dairy
products. Therefore, the most important
strategy for farmers is to increase herd size
to benefit from better economies of size.
Small farm owners can simulate increased
herd size by forming local producer
associations linked to a milk collection
center to strengthen their negotiation
position vis-a-vis milk buyers and feed
providers. Such associations would provide
better access to production-relevant
information and technologies and could
faci litate other relevant activities for its
members (e.g., collective purchase and

preparation of feed concentrate, contracts


with fodder-corn producers, implementation of artificial insemination). Such
associations also help improve farmers'
record-keeping, a precondition for successfully implementing improved farm and
herd management, which ultimately
increases both herd size and profits (Julca,
2000).
For medium farms, which are especially
vulnerable to milk price decreases, herd
and feed management improvements are
crucial to achieving maximum herd
growth. When herd sizes cannot be
sufficiently increased in the short-term,
farmers should consider merging their
herds to achieve the necessary farm size.
Merging among medium farms is also
1ikely to provide the necessary equ ity to
acquire cropland for achieving higher feed
security. Own corn silage production
would reduce the risk of seasonal feed
scarcity.
Large farms are in the best situation to
withstand decreases in milk price. Large
farm owners might consider starting to
process their own milk if the number of
industrial milk processors declines given
the higher consumer prices for dairy
products in situations marked by reduced
competition. In any case, on the production side, large farms must concentrate on
maintaining optimal farm management
practices by contracting highly specialized professionals in the fields of animal
nutrition, breeding, and reproduction.

Prospective policy interventions

For small farmers, policy interventions


should target the creation of milk producer
associations. To generate secure income
for these farmers, the public sector should
purchase the milk for its food programs
directly from these associations.

fact, in the long-term, it may be more


constructive for the national economy if
policy interventions emphasize the
promotion of milk production in the
highlands rather than on the coast (Bernet,
2000).

References
Avni, S. 1996. La problemtica lechera en
el Peru. Estado de Israel - Fongal Lima
- Agencia Norteamericana para el
Desarrollo Internacional, Lima, Peru.
Bernet, T. 2000. The Peruvian dairy sector:
Farmer perspectives, development
strategies and policy options. PhD
Thesis. ETH No 13830, Zrich,
Switzerland.
INEI (Instituto Nacional de Estadstica e
Informtica). 1995. Censo agropecuario
1993. INEI, Lima, Peru.
INEI. 2000. Demographic indicators of
Peru in 1998 and 1999. Instituto
Nacional de Estadstica e Informtica,
Internet information (http://
www.inei.gob.pe/).
Julca, J. 2000. Caracterizacin productiva
de pequeos ganaderos lecheros del
Valle de Lurn. Thesis, Facultad de
Zootecnia, Universidad Agraria de La
Molina (UNALM), Lima, Peru.
McBride, E. 1997. Poltica alimentaria e
importacin de alimentos de origen
animal en el Per. Thesis, Facultad de
Zootecnia, Universidad Agraria de La
Molina (UNALM), Lima, Peru.
Senz, J. 2000. Uso de un modelo de
simulacin para la optimizacin de la
ganadera lechera en la zona de Lima.
Thesis, Facultad de Zootecnia,
Universidad Agraria de La Molina
(UNALM), Lima, Peru:
USDA. 1999. Peru: Dairy situation 1999.
USDA, Foreign Agricultura! Service,
Gain Report #PE9017, U.S. Embassy,
Lima, Peru.

For medium and large farmers, explicit


policy interventions are less relevant. In

CIP Program Report 1999 - 2000

449

Agro-processing Waste Assessment in Peri-urban


Hanoi
D. Peters1, Do Duc Ngai2, Dang Thi An 2

Agro-processing provides a way to add value to crops and generate income


for processors, but the wastes generated by processing may present an
environmental health hazard. Seasonal medium-sized processing activities
tend to generate more wastes than can be utilized or managed. Cassava and
canna starch processing activities in three villages near Hanoi, Vietnam,
generated almost 1.45 million cubic meters of waste water during the 19992000 processing season. The organic matter contained in the waste water,
whether at processing points or in downstream ponds, exceeded critical
values set by the government. During the same season, 51,750 t of solid
wastes were generated. Cassava solid wastes were generally used or sold as
feed but canna wastes were dumped along roadsides or in ponds. Local
residents perceived the waste water and solid wastes as environmental health
hazards that need to be addressed.
lntegrated crop cu ltivation and livestockraising-especial ly pig-fattening-is a
majar source of income for thousands of
families living in the peri-urban areas
around Hanoi. Families who can include
agro-processing within their crop-livestock
systems are able to earn considerably
higher incomes and make a bigger contribution to the economic development of
the area. However, there are environmental and social costs to these enterprises
that could outweigh the added value they
provide. The salid and liquid wastes from
processing are a source of environmental
pollution and a health hazard to local and
neighboring communities. Large-scale
industrial processors often possess the
capital to invest in waste processing that
will add further value to the residue orto
manage waste materials that cannot be
further processed into useable products.
1

CIP-Hanoi, Vietnam.
lnstitute of Ecology and Biological Resources (IEBR), Hanoi,
Vietnam.

Also, these enterprises are usually required


by government to undertake waste management. At the other end of the scale,
very small-scale farmer-processers tend to
generate only limited amounts of waste
which they can absorb into their crop1ivestock systems-waste water in the
fields and salid residue as livestock feed.
Medium-sized, household-based processors, however, generate too much waste to
be absorbed on-farm, yet do not have the
capital to invest in residue processing or
waste management. When these processors are concentrated in a small area, the
waste they generate may overwhelm the
absorptive capacity of the community.
Such is the case of three starch-processing
communities in Hoai Duc District, Ha Tay
Province, on the outskirts of Hanoi.
Processing activities are concentrated in
three villages, Cat Que, Duong Lieu, and
Minh Khai, where 6000 processors support
a population of 30,000 residents with
starch processing and associated products,

GIP Program Report 1999 - 2000

451

such as noodles, maltose, candy, and pigs


(which are fed on cassava starch-processing residues). In Duong Lieu, the biggest
processing commune with 2400 households, almost 29% of households are
engaged in cassava starch processing, and
more than 6% in canna starch processing,
and many others in starch-related activities (Table 1). The seasonal nature of
processing activities-from November to
March to accommodate the harvest
season-further contributes to the concentration of waste in a small area during a
short period of time.
In order to safeguard the economic
potential of the agro-processing component of crop-livestock systems in
peri-urban Hanoi, a better understanding is

Table 1. Processing and ather types af ecanamic


activities, and hausehalds (%) engaged in these
activities in Duang Lieu Cammune, Vietnam.
Household economic activities
(%)
Cassava starch processing
28.73
Canna starch processing
6.43
Re-filtering cassava starch
6.43
Cassava cleaning and grating services
2.69
Drying canna salid waste far sale
9.53
Re-filtering canna salid waste
1.46
Maltase processing
6.66
Canna noadle processing
2.96
Sugar processing
0.09
Rice noodle processing
3.92
Candy processing
1.46
Tofu processing
0.36
Mung bean
0.68
Cassava salid waste trading
0.41
Canna salid waste trading
0.09
Cassava and canna roats trading
1.50
Rice seed ing for malta se processing
0.14
Agricultura! labor
9.17
Industrial labor
4.15
Mushroom growing with canna salid
0.05
waste
Mushroom growing with sawdust
0.50
Pig raising
64.25
Alcohol making
0.96
Ceramic tile cutting for starch settling
0.09
tanks
Drying cassava so lid waste for fuel
19. 70
Other (i.e., small traders, repairer,
38.62
handicraft, engineering)

452

Enriching the Portfolio: GIP's Global and Regional Partnerships

needed of the environmental and social


impacts, and measures need to be taken to
reduce negative effects. A research project
was launched with the following objectives:
to quantify the waste water and solid
waste generated by cassava and canna
starch processing enterprises and
determine the potential impact of such
wastes on the environment and on
human health through assessing the
volume and characteristics of the waste
water at various points, the volume and
characteristics of the solid wastes at
various times, and local perceptions of
the environmental and health problem?
associated with the wastes,
to devise action research interventions
to reduce the volume of wastes at their
sources through improved processing
technology, and
to devise action research interventions
to reduce waste by identifying alternative uses.
This paper reports on the first of these
objectives.

Materials and Methods


Volumes of waste water and solid waste
were assessed by interviewing 280 respondents (processors and non-processors) from
three processing villages and one nonprocessing village (Table 2). Samples of
waste water were collected at various
points along the canals of both processing
and non-processing villages and from
dispersa! ponds and lakes. Sample collection and analyses were conducted in
accordance with environmental rapid
appraisal guidelines for each environmental indicator of waste management (UNEP,
1994).
lt was not possible to begin project
activities until March 2000, near the end
of the processing season. So, the data
presented here do not reflect the situation
at the height of the processing season.

Table 2. Households interviewed at processing and non-processing villages.


Villages
Cassava processing Canna processing Starch filtering
Non-processing
households
households
households
households
Duong Lieu
45
15
3
38
Cat Que
37
1
1
35
Minh Khai
24
10
6
30
Son Dong
o
Total (no.)
106
26
10
138

Benefits and costs of starch processing


All households involved in agro-processing
of starch and related products are also
farmers. The inclusion of starch processing
in the household's activities more than
doubles household income (Figure 1); the
increased benefits from the added activity,
and the system interactions that enhance
earnings from pig-raising through using
byproducts as feed, more than offset the
reduced earnings from crop production and
other on- and off-farm labor. However, the
waste water and solid waste generated by
the processing may have environmental
and social costs, including water contamination and effects on health, reported
negative impact of waste water on crops,
reported negative impact on fish stocks,
visual pollution and smell, and possible
effects of rotten solid wastes on human
health.
'000 VNDl11ouseholdlyear
(US$1 = 14,500 VND)

30000
DProcessor
Non-processor

20000
15000
10000

Crops

Starch

Pig

101

74
70

35
280

Local people's perceptions concerning the


wastes

Results and Discussion

25000

Total

Other

Total

Figure 1. Average household income of processing


villages versus non-processing villages.

Local residents, whether processors or nonprocessors, whether in processing villages


or non-processing villages, are well aware
of the negative effects of processing. All
respondents agreed that the drinking water
during the off-season was much cleaner
and that waste water was mixed with
irrigation water.
Nearly all of the respondents in the
processing villages said that the solid
waste smelled and looked bad (Table 3).
Almost all the non-processors thought that
the decomposing waste was harmful to
their health and even 84% of the processors admitted so. Most of the people in the
processing villages have had solid waste
dumped in front of their houses, and this
has been a source of conflict among the
residents.
In the processing villages, the nonprocessors expressed themselves freely and
assertively concerning the problem with
pollution caused by processing and the
negative effects on human heath. The
processors, on the other hand, were
hesitant to admit the severity of the
environmental problems caused by
processing (Table 3). The most notable
impact in the non-processing village was
waste water in canals flowing by the nonprocessing village. The residents in the
non-processing village commonly complained about the poHution and bad smell
in their village associated with activities
in the processing villages, and pleaded for
solutions.

CIP Program Report 1999 - 2000

453

Waste water assessment


Waste water is generated by cassava and
canna starch processing, starch refining,
and to a lesser extent, maltose processing
and canna starch noodle processing.
Canna starch processing generates more
waste water than does cassava processing
beca use:
cassava is peeled and requires less
water for cleaning whereas canna roots
are difficult to peel and so remain
unpeeled and
canna goes through a second and third
manual extraction (by stepping on it)
which generates more waste water.

On average, processing 1 t of cassava


roots generates 10.7 m3 of waste water and
processing the same amount of canna
generates 12.9 m3 (Table 4). Data from the
district and village leaders showed that
these three villages purchased and processed 75,000 t of cassava roots and
50,000 t of canna roots during the 19992000 processing season, thus generating
almost 1.45 million cubic meters of waste
water during this season.
Results of laboratory analysis of waste
water samples are shown in Table 5. At the
point of processing, concentrations of
pollutants were generally higher in canna
processing waste water than in cassava

Table 3. Perceptions of processors and non-processors toward the wastewater and salid waste generated
by starch processing (% of households responded "yes").
Processing villages
Non-processing
Questions
Processor Non-processor
villa ge
(N=142)
(N=103)
(N=35)
22
7
Ponds and lakes can still be used as clean water far
washing and cleaning.
77
78
87
Ponds and lakes are dirtier in processing season.
01
Ponds and lakes in processing villages are dirtier than non100
99
processing village.
51
48
o
Noticed plants nearby waste water ditches and canals have
died.
21
28
Saw children unintentionally swimming in wastewater.
58
23
28
20
Saw children get diseases or illness from falling into
wastewater.
o
Salid waste looks bad and dirty.
89
100
o
100
93
Noticed bad smell from salid waste.
84
98
63
Decomposed salid waste has negative effects on human's
health.
o
Noticed processors dumping salid waste in front of other
92
98
peoples' houses.
o
85
Noticed such dumping behavior cause conflict among
53
peo ple.

1 Oin this column indicates question was not relevant to non-processing villagers.

Table 4. Volumes of wastewater generated from various steps in processing.


Can na
Cassava
Processing procedure
Water supply
Water waste
Water supply
Water waste
(m 3/ton root)
(m 3/ton root)
(m 3/ton root)
(m 3/ton root)
2.0
1.9
4.5
4.1
Root cleaning
1st starch separation
4.5
4.4
5.0
4.3
2nd starch separation
4.5
4.0
4.5
4.4
o
o
0.7
0.5
Re-separate salid waste
11.0
10.7
14.7
12.9
Total
454

Enriching the Portfolio: CIP's Global and Regional Partnerships

Table 5. Characteristics of starch-processing waste water collected at the point of processing, general
industrial waste water, and critica! values far regulatory compliance (mg/I).
Parameters
Parameter concentrations from various waste waters
Critica! value 1
Waste water at
Waste water in
General
Allowed Allowed
processing (mg/I)
canals (mg/I)
industrial
Value 2 Value 3
Cassava
Canna
Processing Non-pro. waste water
villages villages
(mg/I)
pH
4.6
5.8
6.9
7.55
6.85
5-9
5.5 - 9
BOD5
551.5
486.8
710
25
710
<50
<100
COD
6214.8
7378.8
1162
45.7
1162
<400
<100
SS
1466
3012.6
501
35
501
<100
<200
os
5735.7
7385.8
1.34
0.51
1340
Total P
34.82
72.37
24.21
25
24.21
<6
<8
Total N
168.1
199.3
102.5
12.87
102.5
<60
<60
Am-N
47.52
35.19
98.84
<1
<10
Mn
0.44
2.21
1.36
0.5<
1.36
<1
<5
so 4216
69.25
30
CN0.035
0.001
0.01
<0.1
<0.2
Note: 8005: biological oxygen demand; COD: chemical oxygen demand; SS: suspended solids; OS: dissolved solids;
Am-N: ammonium nitrogen; Mn: manganese; sol-: suttate ion; CN-: cyanide ion.
1 Based on Vietnam Wastewater Discharge Standards (MOSTE, 1995), Vietnam's standard No 5945 - 1995.
2 Allowed to be discharged only into water bodies used for navigation, irrigation, or aquatic breeding and cultivation.
(TCVN 5945-1995).
3 Allowed to be discharged only into the specific water bodies permitted by authoritative agencies.

waste water, but all measured values


exceeded the critica! values set by the the
Ministry of Science, Technology, and
Environment (MOSTE) in its Vietnam
Wastewater Discharge Standards (MOSTE,
1995). These standards stipulate that waste
water with such levels of contamination
may be discharged only into water bodies
specified by authoritative agencies.
The concentrations of pollutants were
much lower (but still above critica! levels)
in the canals, even after waste water had
become diluted and had had a chance to
settle. By the time the water reached the
non-processing village, the pollution
concentrations were below the critica!
levels set by the government. lt is important to emphasize, however, that these
samples were col lected at the end of the
processing season when minimal amounts
of pollutants were being introduced into
the canals. A follow-up study is currently
underway to col lect waste water samples
during the peak of the processing season.

Solid waste assessment


Solid waste is generated during root
cleaning, peeling (cassava), and separation and re-separation (canna). (Cassava
produces more solid waste than canna
because cassava does not go through
second or third separation/extraction.)
On average, cassava generated 47%
sol id waste of the roots processed and
canna generated 33%. In the 1999-2000
season, 51,750 tons of solid wastes were
generated.
Solid waste from cassava starch processing
comprises root skin, fibrous residue, and
black starch. Root skin waste vares from
5% of root weight if the roots are peeled
by machi ne to 10% from hand peel i ng.
Two to three days after processing the
fibrous residue begins to ferment, changes
color from beige to brown, and acquires a
fou 1 smel I; after one week it starts to
decompose. Black starch (only about 2.5%
of root weight) is a byproduct consisting of
starch and other non-soluble organic

substances with little value.

CIP Program Report 1999 - 2000

455

Canna roots, with a sinuous and nonuniform shape, create much solid waste
during processing-comprising root skin
and fibrous residue. Cleaning canna roots
creates flaky solids and root skin accounting for about 5% of root weight. Solid
waste created during the first starch
separation stage consists of fiber, starch,
nutrients, and about 80-90% water.
Because canna starch is difficult to
extract, this residue still has a high starch
content, so it undergoes another manual
separation. After the second separation,
another household may take the solid
waste and put it through yet one more
cycle of separation to extract the very
small amount of starch remaining. Altogether about 5-7 kg of starch can be
extracted from one t of waste. The processors are willing to take the pain to extract
canna starch three times because canna
starch and noodles command much higher
prices and provide higher profits than does
cassava starch.
Laboratory analysis showed that cassava
solid waste contains 87.5% moisture and
12.5% dry matter. The dry matter is 69%
starch, 0.4% NPK, and 30% fiber, and
hence makes good pig and fish feed. Of
the 35,000 t of cassava solid waste
generated in 1999-2000, processors fed
11 % to their own pigs, and sold 71 % as
raw material for feed; only 18% was
wasted. Can na sol id waste, on the other
hand, contains less moisture (about 80%)
and more dry matter (20%), but the dry
matter contains only 4% starch and 0.6%
NPK, the rest being fiber that pigs cannot
easily digest. Canna solid waste is therefore hardly ever used as pig feed, and 90%
of it is abandoned to decompose or
dumped into lakes and ponds. In 19992000 about 15,000 t of sol id canna waste
was disposed of in this way. However, in
2001 a new model canna-starch separator
was introduced into the processing villages. lt has greater extracting efficiency
and creates extremely fine residues that
simply float away with the waste water, so
the amount of canna waste was greatly
reduced and little was abandoned in the
456

Enriching the Portfolio: CIP's Global and Regional Partnerships

villages. This solves the problem of canna


sol id waste, but leads to more suspended
solids in the waste water.
From waste to worth
Sorne of the waste water is already being
used productively. There is no widespread
use of waste water on crops, but the
drainage system allows the possibility of
diverting waste water onto growing crops
in situations where there is insufficient
irrigation water. Anecdotal information
suggests that this can produce problems in
sorne crops. Sorne local residents said that
wastewater had been used to irrigate crops
before, but the effort fai led because the
experiment "lacked scientific rigor."
There have been no communal-level
efforts to treat the waste water, but in 1 996
the Vietnamese Government allocated 6
billion VND (US$1 = 14,500 VND) to
construct a waste water processing factory.
Construction is sti 11 not completed, and the
government is now negotiating to sell the
unfinished factory to a private company to
process organic fertilizer or animal feed
from the cassava sol id waste. lt appears
that the system of drainage ponds and the
wide drainage channel at the final outflow
point from the largest processing village
acts as sedimentation ponds, partially
cleaning the water on its journey out of
these villages and towards neighboring
villages. Sediments from this canal are
used as fertilizer. The potential of waste
water as a nutrient-rich irrigation source
for certain crops is currently being evaluated, along with a detailed assessment of
the functioning of the existing 'sedimentation' ponds and channels as treatment
processes for the waste water.
Already more than 80% of the cassava
solid waste is being used productively,
primarily as pig and fish feed but also for
other innovative purposes. For example, in
Cat Que village it is being used as a
medium for growing mushrooms. Moist
waste is tightly-packed inside shells made
of dried waste. The waste holds moisture
and its nutrients are available for the

growing mushrooms. Canna waste cannot


be u sed as feed, but a smal 1 amount is u sed
as fertilizer and, when formed into small
bricks and sun-dried, as a poor quality fuel.
These local recycling strategies are
important, but the prices for existing
recycled products are so low that there is
little incentive and other innovative uses
or management techniques are needed.
Pham et al. (1992) and Demo-os et al.
(2000) were able to increase the protein
level by 10-20 fold when manufacturing
livestock feed with protein-enriched
cassava or sweetpotato starch sol id waste.
This could be a potential venue for
processing and adding value to the
cassava waste in Duong Lieu. However,
Pham's and Demo-os's research was
conducted in laboratories with chemical
substances and sophisticated equipment,
and, therefore, the process may need
majar modifications before it can be
applied on-farm.

Conclusions
In the past, villagers caught fish and
mussels in the waterways near their
vil lages, but toda y there are no fish left.
As a result of processing activities, the
major canal that flows through the three
processing villages and past sorne nonprocessing ones has been filled with
mud and waste for several years. Sorne
sections of the canal are now only 0.5 m
deep. During the heavy rains between July
and August, roads in the processing
villages are now frequently flooded up to
0.2-0.3 m deep.
The decomposing or decomposed abandoned solid wastes, combined with
intensive pig-raising as an associated
enterprise, create foul odors that pollute
the processing and non-processing villages
alike. Sorne villages have set aside areas
for solid waste dumping, but the quantities
of wastes generated by processing have so
far overwhelmed this limited infrastructure.
Now that the extent, characteristics, and
impacts of the waste water and solid waste

have been assessed, sorne solutions are


needed to manage these wastes. Proposed
solutions for managing the waste water
include building community stabilization
ponds to settle the organic matter in the
water and re-using the waste water on
various crops that may resist, or even
benefit from, the organic matter in the
water.
A follow-up project is underway to test the
possibility of processing and re-using the
waste water on various crop fields. The
waste water is being tested for irrigating
water taro, which seems to thrive on the
organic matter and nutrients in the waste
water. While the water is re-used to
irrigate water taro, the field simultaneously serves as a stabilization pond until
the water is released into an adjacent field
a week later. This new trial will examine
the effects of waste water on the yield of
various crops and on soil structure.
Proposals to be investigated for managing
the solid waste include production of
commercial feeds and efficient fuels, and
other innovations, such as using canna
solid waste to raise worms for pig feed.

References
Demo-os, R.A., T.S.J. Valdez, and M.C.
Mapili, Jr. 2000. Feeding value of
protein-enriched sweet patato pulp for
broilers. Philippines Journal of Veteran
Animal Science 26:41-50.
MOSTE (Ministry of Science, Technology,
and Envioronment). 1995. Vietnamese
government standard of environment.
Hanoi Publishing House. Hanoi,
Vietnam.
Pham, C.B., M.R.L.Y. Lat, T.J. Ramirez,
M.J. Quinlat, and L.J. Pham. 1992.
Enriching cassava protein using solid
state fermentation. BIOTECH,
University of the Philippines at Los
Banas, Philippines.
UNEP (United Nations Environmental
Program). 1994. Overview of
environmental indicators: State of the
art and perspectives. IVM, Netherlands.

CIP Program Report 1999 - 2000

457

Acronyms and Abbreviations


ADB
ADFA
AESA
AEZ
AFLP
AFRENA
AMMI
APLV
APMV
ARARIWA
ARTC
ASAR
AUDPC
AVA
AVRDC
BADC
BMZ
Bt
BU
BW
CAE
CAPS
CEC
CENA
CIAT
CllD
CIMMYT
CIP
cm
CMD
CMF
CONDESAN
CPRI
CPRO
CTCRl/ICAR

d
DAS
DAT
DADO
DAN IDA
DEM
DFID
DM
DNA

Asian Development Bank


average of diseased foliage area
agro-ecosystem analysis
agroecological zoning
amplified fragment length polymorphism
African Resource Network in Agro-Forestry (Uganda)
additive main effect and multiplicative interaction
Andean potato latent virus
Andean potato leafroll virus
Association for Andean Technical-Cultural Promotion (Peru)
Andean root and tuber crops
Asociacin de Servicios Artesanales y Rurales (Bolivia)
area under the disease progress curve
Arracacha vi rus A
Asian Vegetable Research and Development Center (Taiwan)
Bangladesh Agricultura! Development Corporation
Federal Ministry of Economic Cooperation and Development
(Germany)

Bacillus thuringiensis
blight unit
bacteria! wilt
cellu lose acetate electrophoresis
cleaved amplified polymorphic sequence
Cation exchange capacity
Civil Engineers Network Africa, South Africa
Centro lnternacinal de Agricultura Tropical (Colombia)
Centro lnternacinal de Investigaciones para el Desarrollo (IRDC)
lnternational Maize and Wheat lmprovement Center (Mexico)
lnternational Potato Centro (Centro Internacional de la Papa)
centimeter
cassava mosaic virus disease
cubic feet per minute
Consortium for the Sustainable Development of the Andean
Ecoregion
Central Potato Research lnstitute (India)
Centre for Plant Breeding and Reproduction Research (Netherlands)
Central Tuber Crops Research lnstitute, lndia/lndian Council of
Agricultura! Research
day
days after sowi ng
days after transplanting
District Agriculture Development Office (Nepal)
Royal Danish Ministry of Foreign Affairs
digital elevation model
Department for lnternational Development
dry matter
deoxyribonucleic acid

CIP Program Report 1999 - 2000

459

EARO
EARRNET
EC
ECABREN
ELISA
EPA
ESARC
FAO
FAPESP
FFS
FMV
FORTIPAPA
FPR
FSP
FU
GCA
GILB
GIS
GOES
Gpi
GPS
GTZ
GUS
GxE

h2
HDP
HPAEC
HYV
IAF
IAl-ISP
IARC
ICM
ICRAF
ICRISAT
IDM
IDRC
IFAD
IFAS
IFDC
IFPRI
llRDEP
llTA
ILRI
INERA

460

Acronyms and Abbreviations

Eastern Agricultura! Research Organization (Ethiopia)


Eastern Africa Rootcrops Research Network (Uganda)
electrical conductivity
Eastern and Central Africa Bean Research Network (Uganda)
enzyme-linked immunosorbent assay
Environmental Protection Agency (USA)
East and Southern Africa Regional Centre (Kampala, Uganda)
Food and Agriculture Organization of the United Nations
Funda<;ao de Amparo a Pesquisa de Estado de Sao Paulo
(Brazil)
farmer field school
feathery mottle virus
Fortalecimiento de la Investigacin y Produccin de Semilla de
Papa en el Ecuador
farmer participatory research
Forages for Smallholders Project, (CIAT, Colombia)
fungicide unit
general combining ability
Global lnitiative on Late Blight
geographic information systems
Geostationary Operational Environmental Satellite
glucose-6-phosphate isomerase
global positioning system
Gesellschaft fur Technische Zusammenarbeit (Germany)
b-glucuronidase
genotype by environment
hour
heritabi 1ity
high demand and production growth
high performance anion exchange chromatograph
high yielding variety
lnter-American Foundation (USA)
lnter-American lnstitute for Global Change Re~earch, lnitial
Science Program (Brazil)
international agricultura! research center
lntegrated crop management
lnternational Centre for Research in Agroforestry
lnternational Crops Research lnstitute for the Semi-Arid Tropics
integrated disease management
lnternational Development Research Center (Canada)
lnternational Fund for Agricultura! Development (USA)
immunofluorescence antibody staining
lnternational Fertilizer Deveopment Center (USA)
lnternational Food Policy Research lnstitute (USA)
Philippine's lnter-Regional lntegrated Research, Development
and Extension Program on Urban Agriculture
lnternational lnstitute of Tropical Agriculture (Nigeria)
lnternational Livestock Research lnsitute (Kenya)
National potato and sweetpotato programs of the national
agricultura! research institute of D.R. of the Congo

INFOANDINA
INIA
INIAP
INRA
INTA
IOFM
IPGRI
IPM
IRR
IRRI
ISO

J
K
KARI
kg
1

LB
LEACHM
LMF
m
mm
masl
MCL
mo
MRR
n
N/m2
NAFTA
NAR
NARO
NARS
NASH
NCM
NDVI
NGO
NPV
NUE
OFE
OM
OPEC

p
p
PAGE
PapMV-U
PARC
ppb
PBS
PCA

PCR
pep

CONDESAN information network


Instituto Nacional de Investigacin Agraria (Peru)
Instituto Nacional Autnomo de Investigaciones Agropecuarias
(Ecuador)
lnstitut National de la Recherche Agronomique (France)
lnstitute Nacional de Tecnologia Agropecuaria (Argentina)
lnternational Organic Farming Movement
lnternational Plant Genetic Resources lnstitute (ltaly)
integrated pest management
interna! rate of return
lnternational Rice Research lnstitute (Philippines)
lnternational Organization far Standardization
joule (unit of light energy)
potassium
Kenya Agricu ltu ral Research 1nstitute
kilogram
liter
late blight
Leaching Estimation and Chemistry Model
leafminer fly
meter
millimeters
meters above sea level
maximum contaminant level
month
marginal rates of return
nitrogen
Newtons/m2 (1 kg m per second squared)
North American Free Trade Association
net assimilation rate
National Agricultura! Research Organization (Uganda)
national agricultura! research systems
nucleic acid spot hybridization
nitrocellulose membrane
normalized difference vegetation index
nongovernmental organization
Net present value
n itrogen use efficiency
overland flow element
organic matter
Organization of Petroleum Exporting Countries
probability
phorphorus
polyacrylamide gel electrophoresis
papaya mosaic virus, ulluco isolate
Pakistan Agricultura! Research Center
parts per billion
pre-basic seed
principie component analysis
polymerase chain reaction
peptidase

CIP Program Report 1999- 2000

461

PGI
pH
PI
PIC
PLRV
PNS-PRODISE
PO
PoGV
PRAPACE

PRECODEPA
PRGA
PROINPA
PRONAMACHCS
PSTVd
PTL
PTM
PVT
PVX
PVY
QTL
RAPO
rAUDPC
RCBD
rDNA
RFLP
RGR
RH
RKN
RRA
RRD
RS
SAAS
SAPPRAD
SARDl-UMCOR
SARRNET
SCA
SCRI
SDC
SDRF

sos
SEINPA
SE NASA
SG

462

Acronyms and Abbreviations

potato genetic identification


acidity
proteinase inhibitor
polymorphic indes content
potato leafroll virus
Programa Nacional de Semillas del Proyecto de Desarrollo
Integral de Semillas (Peru)
parent offspri ng
Phthorimaea operculella Zeller granulovirus
Programme Rgional de l'Amlioration de la Culture de la
Pomme de Terre et de la Patate Douce en Afrique Central et
de l'Est (CIP network)
Programa Regional Cooperativo de Papa (CIP network in
Central America and the Caribbean)
Participatory Research and Gender Analysis (CGIAR lnitiative)
Proyecto de Investigacin de la Papa, National Patato Research
Program (Bolivia)
Proyecto Nacional de Manejo de Cuencas Hidrogrficas y
Conservacin de Suelos (Peru)
potato spindle tuber viroid
pathogen tested list
potato tu ber moth
potato vi rus T
potato X potexvirus
potato vi rus Y
quantitative trait loci
randomly amplified polymorphic DNA
relative area under the disease progress curve
randomized complete block design
ribosomal DNA
restriction fragment length polymorphism
relative growth rate
relative humidity
root-knot nematode
rapid rural appraisal
Red River Delta, Vietnam
reducing-sugar
Shandong Academy of Agricultura! Science (China)
Southeast Asian Program for Potato Research and Development
Sustainable Agricultura! and Rural Development lnitiativeUnited Methodist Committee on Relief, (D.R. Congo)
Southern Africa Root Crop Research Network
specific combining ability
Scottish Crop Research lnstitute
Swiss Development Cooperation
single-dose restriction fragment
sodium dodecyl sulfate
Semilla e lnvestigacion en Papa (Peru)
Servicio Nacional de Sanidad Agraria (Peru)
specific gravity

SIUPA
SLA
SM-CRSP
SPCFV
SPCSV
SPFMV
SPLV
SPVD
SQR
SSR
t

TAC
TCRC
TPS
UAP
UMMV
UMV
UNEP
UNHEVAL
UNICEF
UPGMA
UPWARD
USAID

uv
uve
w
WEPP
WUE

CGIAR Systemwide lnitiative on Urban and Peri-Urban


Agriculture
specific leaf area
Soil Management Collaborative Research Support (USA)
sweetpotato chlorotic fleck virus
sweetpotato chlorotic stunt virus
sweetpotato feathery mottle virus
sweetpotato latent virus
sweetpotato virus disease
standards for qualitative resistance
simple sequence repeats
metric tonne (1000 kg)/ton (2200 lbs)
Technical Advisory Committee of the CGIAR
Tuber Crops Research Center (Bangladesh)
true potato seed
Urban Agriculture Program (Philippines)
ullucus mild mottle virus
u l lucus mosaic virus
United Nations Environmental Program
Universidad Nacional Hermilio Valdizan (Peru)
United Nations lnternational Children's Emergency Fund
unweighted pair group method with arithmetic mean algorithm
Users' Perspective with Agricultura! Research and Development
(CIP network)
United States Agency for lnternational Development
u ltra-violet (1 ight range)
ullucus virus e
watt (unit of light)
Water Erosion Prediction Project
water use efficiency

CIP Program Report 1999 - 2000

463

Author lndex
Alarcn, Lidia. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru. p. 95.
Aley, Pedro. Crop Protection, lnternational
Potato Center, A.P. 1558, Lima 12, Peru.

p. 105.
Amaro, J.R. Universidad Nacional del
Centro del Peru, Huancayo, Peru. p. 69.
Ames, Teresa. Crop Protection,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 397.
Amoros, Walter. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru.
p. 185.
Asmunati, Rini. Mitra Tani, Yogyakarta,
Indonesia. p. 331.
Baigorria, Guillermo. Production Systems
and Natural Resource Management,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 361, 371.
Barrea, O. Fundacin para la Promocin e
Investigacin de Productos Andinos
(PROINPA), 191 Calle Honduras, Barrio
Petrolero, Sucre, Bolivia. p. 143.
Barreda, Carolina. Production Systems and
Natural Resource Management, A.P.
1558, lnternational Potato Center, Lima
12, Peru. p. 361.
Bejarano, Carlos. PROINPA Foundation,
Cochabamba, Bolivia. p. 251.
Bello, Vernica. Crop lmprovement and
Genetic Resources, lnternational Patato
Center, A.P. 1558, Lima 12, Peru.
p. 267.
Beltrn, Gregario. Crop lmprovement and
Genetic Resources. lnternational Potato
Center, A.P. 1558, Lima 12, Peru. p. 63.
Bernet, Thomas. Social Science,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 445.
Bi, Y.P. Shandong Academy for
Agricultura! Sciences, Jinan, Shandong,
China. p. 219.
Boncodin, Raul. CIP-UPWARD, PCARRD
Complex, Los Banos, Laguna,
Philippines. p. 433.

Bonierbale, Merideth. Crop lmprovement


and Genetic Resources, lnternational
Potato Center, A.P. 1558, Lima 12, Peru.
p. 49, 95, 167, 185.
Bowen, Walter. IFDC/lnternational Patato
Center, LAC, A.P. 17-21-1977, Quito,
Ecuador. p. 355, 371.
Brck, Holger. lnstitute of Plant Nutrition
and Soil Science, Kiel University,
Germany. p. 273.
Burg, Kornel. Austrian Research Center
Seibersdorf, Environment and Life
Science, A-2444 Seibersdorf, Austria.
p. 303.
Bussink, Coen B. Production Systems and
Natural Resource Management,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 413, 421.
Cabello, Rolando. Crop lmprovement and
Genetic Resources, lnternational Patato
Center, A.P. 1558, Lima 12, Peru.
p. 197, 207.
Campilan, Dindo M. UPWARD
Coordinator, CIP-UPWARD, PCARRD
Complex, Los Banos, Laguna,
Philippines. p. 239, 433.
Caedo, Vernica. Crop Protection,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 117.
Carbajulca, Doris. Crop lmprovement and
Genetic Resources, lnternational Patato
Center, A.P. 1558, Lima, Peru. p. 295.
Casso, Ricardo. PROINPA Foundation,
Cochabamba, Bolivia. p. 251.
Cervantes, Jim C. Crop lmprovement and
Genetic Resources, lnternational Patato
Center, A.P. 1558, Lima 12, Peru.
p. 303.
Chacn, Maria G. lnternational Potato
Center, Apartado 17-21-1977, Quito,
Ecuador. p. 87.
Chien Dao Huy. Root Crop Research
Center, Vietnam Agricultura! Science
lnstitute, Hanoi, Vietnam. p. 211 .
Cipriani, Giselle. Crop lmprovement and
Genetic Resources, lnternational Patato

CIP Program Report 1999 - 2000

465

Center, A.P. 1558, Lima 12, Peru.


p. 267.
Cisneros, Fausto. Crop Protection,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 155, 161.
Cotrina, Lucio. CARE, General Santa Cruz
659, Jess Maria, Lima, Peru. p. 225.
Cruz, Mariana. Production Systems and
Natural Resource Management, A.P.
1558, lnternational Potato Center, Lima
12, Peru. p. 361.
Dang Thi An. lnstitute of Ecoiogy and
Biological Resources (IEBR), Hanoi,
Vietnam. p. 451 .
Danielsen, Solvig. CIP/current address:
The Royal Veterinary and Agricultura!
University, Copenhagen, Denmark.
p. 397.
de Guzman, Constando. University of the
Philippines, Los Banos, Philippines.
p. 433.
de la Cruz, Jorge. Production Systems and
Natural Resource Management,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 413.
de Mendiburu, Felipe. Statistics,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 105, 143.
Devaux, Andre. Andean Potato Project,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 251.
Diaz, Luis. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru. p. 63.
Do Duc Ngai. lnstitute of Ecology and
Biological Resources (IEBR), Hanoi,
Vietnam. p. 451 .
Douches, Daniel S. Dept. of Crop and Soil
Sciences, Michigan State University,
East Lansing, MI, USA. p. 117.
Echeandia, Edda. Training and
Communication, lnternational Potato
Center, A.P. 1558, Lima 12, Peru. p. 16.
Echegaray, Judith. Crop Protection,
lnternational Patato Center, A.P. 1558,
Lima 12, Peru. p. 397.
El-Bedewy, Ramzy. lnternational Potato
Center, SSA, P.O. Box 25171, Nairobi,
Kenya. p. 77.
Erazo, Cristina. Universidad Catolica,
Quito, Ecuador. p. 391.

466

Author lndex

Espinoza, Jorge. Crop lmprovement and


Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru.
p. 185.
Ewell, Peter. lnternational Patato Center,
SSA, P.O. Box 25171, Nairobi, Kenya.
p. 77.
Forbes, Gregory A. lnternational Patato
Center, LAC, Apartado 17-21-1977,
Quito, Ecuador. p. 39, 87.
Fuentes, Segundo. Crop Protection,
lnternational Patato Center, A.P. 1558,
Lima 12, Peru. p. 267, 381.
Fuglie, Keith O. lnternational Potato
Center, Kebun Percobaan Muara, Jalan
Raya Ciapus, Bogor 16610, Indonesia.

p. 211.
Garrett, Karen A. Kansas State University,
Manhattan, KS, USA. p. 225.
Garry, Guillemette. CIP/Present address:
c/o 7 rue de Concise, 53940
St. Berthevin, France. p. 39.
Gastelo, Manuel. Crop lmprovement and
Genetic Resources, lnternational Patato
Center, A.P. 1558, Lima 12, Peru. p. 63.
Gaur, P.C. Central Patato Research
lnstitute (CPRl/ICAR), Shimla, H.P.,
India. p. 219.
Ghislain, Marc. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru. p. 27,
167, 175, 267, 295.
Golmirzaie, Ali M. Department of
Horticulture, University of Arkansas,
Fayetteville, AR 72701, USA. p. 175.
Gomez, Carlos. Universidad Nacional
Agraria, La Molina, Lima 12, Peru.
p. 445.
Gomez, Rene. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru.
p. 175.
Gonzles, Eliana. Crop Protection,
lnternational Patato Center, A.P. 1558,
Lima 12, Peru. p. 39.
Gonzales, Silvia. PROINPA Foundation,
Cochabamba, Bolivia. p. 251.
Grande, Enrique. Crop lmprovement and
Genetic Resources, lnternational Patato
Center, A.P. 1558, Lima 12, Peru.
p. 219.

Gutarra, Liliam. Crop Protection,


lnternational Patato Center, A.P. 1558,
Lima 12, Peru. p. 105, 143.
Gysin, Ren. CIP/Home address:
Oberlandstr. 20, 3700 Spiez,
Switzerland. p. 27.
Hakiza, J.J. Patato Research Program,
National Agricultura! Research
Organization, Kabale, Uganda. p. 77.
Heath, Jeremy J. Kansas State University,
Manhattan, KS, U.S.A. p. 225.
Hermann, Michael. Crop lmprovement
and Genetic Resources, lnternational
Ptato Center, A.P. 1558, Lima 12, Peru.
p. 273, 281, 391.
Herrera, Carmen. Crop lmprovement and
Genetic Resources, lnternational Patato
Center, A.P. 1558, Lima 12, Peru.
p. 295.
Herrera, Maria del Rosario. Crop
Protection, lnternational Patato Center,
A.P. 1558, Lima 12, Peru. p 27.
Hidalgo, Osear A. Project CIP/SDC, P.O.
Box 2122, lslamabad, Pakistan. p. 239,
245.
Hijmans, Robert J. Natural Resource
Management, A.P. 1558, lnternational
Patato Center, Lima 12, Peru. p. 69, 87,
323, 413, 421.
Hoa, T.D. Hue University of Agriculture
and Forestry, Hue, Vietnam. p. 343.
Hossain, Abu Enamder. Tuber Crops
Research Center (TCRC), Joydebpur,
Gajipur, Bangladesh. p'. 259.
Hossain, Mohammad. Tuber Crops
Research Center (TCRC), Joydebpur,
Gajipur, Bangladesh. p. 259.
Huaccho, Luisa. CIP/Present address: St.
Hilda's College, Oxford University,
Oxford, UK. p. 323.
Huamn, Zosimo. Pro Biodiversity of the
Andes (ProBioAndes), Lima 12, Peru.
p. 49, 175.
Huber, Jurg. Federal Biological Research
Centre for Agriculture and Forestry,
lnstitute for Biological Control,
Heinrichstr. 243, D-64287 Darmstadt,
Germany. p. 123.
Hung, LQ. College of Agriculture and
Forestry, Ho Chi Minh City, Vietnam.
p. 343.

llangantileke, Sarath G. lnternational


Patato Center, SWA, e/o IARI Campus,
Pusa, New Delhi 110012, New Delhi,
India. p. 259.
Jacobsen, Sven-Eric. DANIDA/
lnternational Patato Center, A.P. 1558,
Lima 12, Peru. p. 397, 403.
Jaramillo, Raul. lnternational Patato
Center, LAC, A.P. 17-21-1977, Quito,
Ecuador. p. 355.
Jayasinghe, Upali. lnternational Patato
Center, ESEAP, Kebun Percobaan Muara,
Jalan Raya Caipus, Bogar 1661 O,
Indonesia. p. 259.
Johnson, Nancy. CGIAR Program on
Participatory Research and Gender
Analysis, lnternational Center for
Tropical Agriculture, Cali, Colombia,
p. 331.
Jurez, Henry S. lnternational Patato
Center, A.P. 1558, Lima 12, Peru and
Universidad Nacional Agraria La
Malina, La Malina, Lima 12, Peru.
p. 69.
Julca, Jessica. Universidad Nacional
Agraria, La Malina, Lima 12, Peru.
p. 445.
Kadian, Mohinder S. lnternational Patato
Center, SWA, C/o IARI Campus, Pusa,
New Delhi 110012, India. p. 259.
Kelm, Michael. lnstitute of Plant Nutrition
and Soil Science, Kiel University,
Germany. p. 273.
Khoang, Le H. Post Harvest Technology
lnstitute, Ho Chi Minh City, Vietnam.
p. 343.
Kriegner, A. Austrian Research Center
Seibersdorf, Environment and Life
Science, A-2444 Seibersdorf, Austria.
p. 303, 315.
Kroschel, Jurgen. University of Hohenheim
(380), D-70593 Stuttgart, Germany.
p. 123.
Lagnaoui, Aziz. Crop Protection,
lnternational Patato Center, A.P. 1558,
Lima 12, Peru. p. 117, 123.
Lama, Tara L. Patato Development
Section, Ministry of Agriculture,
Kathmandu, Nepal. p. 239, 245.
Landeo, Juan A. Crop lmprovement and
Genetic Resources, lnternational Potato

GIP Program Report 1999- 2000

467

Center, A.P. 1558, Lima 12, Peru. p. 27,


63, 225.
Lemaga, Berga. PRAPACE, P. O. Box
22274, Kampala, Uganda. p. 129.
Leon-Velarde, Carlos U. lnternational
Livestock Research lnstitute and
Production Systems and Natural
Resource Management, lnternational
Potato Center, A.P. 1358, Lima 12, Peru.
p. 289.
Li, J.H. Gansu Academy for Agrkultural
Sciences, Lanzhou, Gansu, China.
p. 219.
Lizrraga, Charlotte. GILB, lnternational
Potato Center, A.P. 1558, Lima 12, Peru.
p. 381.
Lpez, Glcerio. Universidad Nacional del
Centro del Peru (U NCP), Huancayo,
Peru. p. 381.
Mahmood, A.A. Tuber Crops Research
Center (TCRC), Joydebpur, Gajipur,
Bangladesh. p. 259.
Malagamba, Patricio. Training and
Communications, lnternational Potato
Center, A.P. 1558, Lima 12, Peru. p. 16.
Manrique, Kurt. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru.
p. 281.
Martinez, J. Fundacin para la Promocin
e Investigacin de Productos Andinos
(PROINPA), 191 Calle Honduras, Barrio
Petrolero, Sucre, Bolivia. p. 143.
Mendoza, Alejandro. Universidad
Nacional Hermilio Valdizan, Hunuco,
Peru. p. 225.
Mihovilovich, Elisa. Crop lmprovement
and Genetic Resources, lnternational
Potato Center, A.P. 1558, Lima 12, Peru.
p. 95.
Milla, Susana. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru.
p. 295.
Monosalva, Patricia. Dept. of Plant
Pathology, 4024 Throckmorton Plant
Sciences Center, Kansas State
University, Manhattan, KS, USA. p. 27.
Mujica, Angel. Projecto Quinua CIPDANIDA, Universidad Nacional del

468

Author lndex

Altiplano, Av. del Ejercito 329, Puno,


Peru. p. 403.
Mujica, Norma. Crop Protection,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 161.
Mwanga, R.O. National Agricultura!
Research Organization, Namulonge
Agricultura! and Animal Production
Research lnstitute, P.O. Box 7083,
Kampala, Uganda. p. 303.
Nelson, Rebecca J. CIP/Mac~night
Foundation, Cornell University, lthaca,
NY, USA. p. 27, 39, 49, 225.
Nga Do Thi Bich. Root Crop Research
Center, Vietnam Agricultura! Science
lnstitute, Hanoi, Vietnam. p. 211
Nguyen Thi Hoa. Root Crop Research
Center, Vietnam Agricultura! Science
lnstitute, Hanoi, Vietnam. p. 211 .
Nguyen, T. Lan. Hong Duc University,
Thanh Hoa City, Vietnam. p. 343.
Nio-Liu, David O. CIP/present address:
lntercollege Graduate Program in
Genetics, Penn State University, 103
Tyson Building, University Park, PA
16802, USA. p. 27.
Nez, Jorge. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru.
p. 167.
Oanh, N.T.K. Hanoi Agriculture
University, Hanoi, Vietnam. p. 343.
Ojeda, Luis. Crop lmprovement and
Genetic Resources, A.P. 1 558,
lnternational Potato Center, Lima 12,
Peru. p. 295.
Ojha, D.N. Project CIP/SDC, P.O. Box
2122, lslamabad, Pakistan. p. 245.
Ojiambo, P.S. lnternational Potato Center,
SSA, Nairobi, Kenya. p. 77.
Olanya, O. Modesto. lnternational Potato
Center, SSA, Nairobi, Kenya. p. 77.
Orrego, Ricardo. Crop Protection,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 225.
Ortiz, Osear. Social Science, lnternational
Potato Center, A.P. 1558, Lima 12, Peru.
p. 225.
Pacheco, Miguel A. Instituto Nacional de
Investigacin Agraria, Cusco, Peru.
p. 225.

Panta, Ana luz. Crop 1mprovement and


Genetic Resources, lnternational Potato
Center, A.P. 1558, lima 12, Peru.
p: 175.
Parraga, A. Universidad Nacional Daniel
Alcides Carrion, Oxapampa, Cerro de
Paseo, Paseo, Peru. p. 225.
Prraga, Adelmo. Universadiad Nacional
Andrs Avelino Cceres, Oxampampa,
Peru. p. 69.
Perazzo, Giovana. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, lima 12, Peru.
p. 175.
Prez, Ana luz. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru. p. 95.
Prez, Wilmer. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru. p. 27,
39, 49.
Peters, Dai. lnternational Potato CenterHanoi, Sol 4 B25, Nam Thanh Cong,
lang Ha, Dong Da, Hanoi, Vietnam.
p. 451.
Pinedo, Hapita M. Crop Protection,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 39.
Posner, Joshua. University of Wiseonsin,
Madison, WS, USA. p. 413.
Prain, Gordon. Social Science,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 445.
Pravatiner, Miguel. Universidad Nacional
Agraria, la Molina, Lima 12, Peru.
p. 161.
Priou, Sylvie. Crop Protection, A.P. 1558,
lnternational Potato Center, Lima 12,
Peru. p. 105, 143.
Quiroz, Roberto. Natural Resources
Management, A.P. 1558, lnternational
Potato Center, Lima 12, Peru. p. 361.
Quispe, Hipolito. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru.
p. 403.
Raymundi, Rub. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru. p. 49.

Reinoso, Jorge. Centro de Investigaciones


en Recursos Naturales y Medio
Ambiente, Puno, Peru. p. 361.
Rivera, Maria. Crop Protection,
lnternational Potato Center, A.P. 1558,
lima 12, Peru. p. 39.
Rivera, Mitchel D. Universadiad Nacional
Andrs Avelino Cceres, Oxampampa,
Peru. p. 69.
Roca, William H. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru.
p. 175.
Rodrguez, Flor. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, lima 12, Peru.
p. 167, 175.
Roncal, Erminia. Instituto Nacional de
Investigacin Agraria, Cajamarca, Peru.
p. 225.
Rossel, Genoveva. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru.
p. 295, 315.
Saenz, Jorge. Universiadad Agraria
Nacional la Molina, lima 12, Peru.
p. 445.
Salas, Alberto. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru. p. 39,
49.
Salas, Clara. Crop Protection, lnternational
Patato Center, A.P. 1558, Lima 12, Peru.

p. 105.
Salazar, luis F. Crop Protection,
lnternational Potato Center, A.P. 1558,
lima 12, Peru. p. 267.
Sanchez, Pablo A. Asociacin para el
Desarrollo Rural de Cajamarca
(ASPADERUC), Cajamarca, Peru.

p. 413.
Santa Cruz, Magnolia. Crop lmprovement,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 39, 381.
Sattelmacher, Burkhard. lnstitute of Plant
Nutrition and Soil Science, Kiel
University, Germany. p. 273.
Simon, Reinhart. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, lima 12, Peru. p. 27.

CIP Program Report 1999- 2000

469

Son, Nguyen T. Hanoi Agriculture


University, Hanoi, Vietnam. p. 343.
Sporleder, Marc. CIP/University of
Hohenheim (380), D-70593 Stuttgart,
Germany. p. 123.
Stoorvogel, Jerse. Wageningen University,
Wageningen, The Netherlands. p. 355,
371.
Suarez, Victor. Social Science,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 251.
Taipe, Marco V. lnternational Potato
Center, Apartado 17-21-1977, Quito,
Ecuador. p. 87.
Tenorio, Jose. CARE, General Santa Cruz
659, Jess Maria, Lima, Peru. p. 225.
Thanh, P.D. Thang Binh District
Agriculture and Rural Development
Bureau, Quang Nam Province, Vietnam.
p. 343.
Toledo, Judith. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru.
p. 175.
Torres, Roger. Servicio Nacional de
Sanidad Agropecuaria (SENASA), Baos
del Inca, Cajamarca, Peru. p. 143.
Torres, Shei ly. Crop Protection,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 27.
Trognitz, Bodo. CIP/Present address: Dept.
Biotechnology Austrian Research
Center, A-2444 Seibersdorf, Austria.
p. 27.
Trognitz, Friederike. CIP/Present address:
Dept. Biotechnology Austrian Research
Center, A-2444 Seibersdorf, Austria.
p. 27.
Upadhya, Mahesh D. Crop lmprovement
and Genetic Resources, lnternational
Potato Center, A.P. 1558, Lima 12, Peru.
p. 197, 207, 225.
Valdivia, Roberto. Centro de
Investigaciones en Recursos Naturales y
Medio Ambiente, Puno, Peru. p. 361.

470

Author lndex

Valencia, Cesar. CARE, General Santa


Cruz 659, Jess Maria, Lima, Peru.
p. 225.
van de Fliert, Elske. CIP-ESEAP, Kebun
Percobaan Muara, Jalan Raya Caipus,
Bogor 1661 O, Indonesia. p. 331, 343.
Vera, Alcira. Crop Protection, lnternational
Potato Center, A.P. 1558, Lima 12, Peru.
p. 155.
Villamn, Francisco. CIP/Present address:
University of Wisconsin, Madison, WI
53706, USA. p. 167.
Villar, A. Servicio Nacional de Sanidad
Agropecuaria (SENASA), Baos del
Inca, Cajamarca, Peru. p. 143.
Walker, Thomas S. Social Science,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 219, 251.
Waugh, Robbie. Scottish Crop Research
lnstitute, lnvergowrie, Dundee 002
5DA, UK. p. 167.
Willet, Has. SNV/Proyecto Nacional de
Manejo de Cuencas Hidrogrficas y
Conservacin de Suelos
(PRONAMACHCS), Cajamarca, Peru.
p. 413.
Wiyanto. Mitra Tani, Yogyakarta,
Indonesia. p. 331.
Zegarra, Octavio. Crop Protection,
lnternational Potato Center, A.P. 1558,
Lima 12, Peru. p. 123.
Zhang, Dapeng. Crop lmprovement and
Genetic Resources, lnternational Potato
Center, A.P. 1558, Lima 12, Peru.
p.267, 295, 303, 315, 323.
Zorogasta, Percy. Natural Resource
Management, lnternational Potato
Center, A.P. 1558, Lima 12, Peru.
p. 361.
Zuiga, Noemi. Instituto Nacional de
Investigacin Agraria, Lima 12, Peru.
p. 225.

Selected Publications by CIP Staff from 1999-2000


Arce, P., M. Moreno, M. Gutierrez,
M. Gebl-uer, P. Dell'Orto, H. Torres,
l. Acua, P. Oliger, A. Venegas,
X. Jordana, J. Kalazich, and
L. Huluigue. 1999. Enhanced resistance
to bacteria! infection by Erwinia
carotovora subsp. atroseptica in
transgenic patato plants expressing the
attacin or the cecropin SB-37 genes.
American Journal of Potato Research
76:169-177.

Bejarano, L., E. Mignolet, A. Devaux,


N. Espinola, E. Carrasco, and
Y. Larondelle. 2000. Glycoalkaloids in
patato tubers: The effect of variety and
drought stress on the a-solanine and achaconine contents of potatoes. Journal
of the Science of Food and Agriculture
80:2096-21 OO.
Bentley, J. and G. Thiele. 1999.
Bibliography: Farmer knowledge and
management of crop disease.
Agriculture and Human Values
16:75-81.
Bernet T. and C. Len-Velarde. 2000.
lncome effects of fodder and herd
management on small-scale milk
producers in the northern Peruvian
Andes. Livestock Research for Rural
Development 12(3). (Internet journal)

Bernet T., J. Julca, J. Saenz, and G. Prain.


2000. Peri-urban milk production in
Lima, Peru: Assessing farmers' decisionmaking within a changing market.
Livestock Research for Rural
Development 12(4). (Internet journal)
Bernet, T. and C. Len-Velarde. 2000.
1ncome effects of fodder and herd
management on small-scale milk
producers in the northern Peruvian
Andes. Livestock Research for Rural
Development (12)3.

Boncodin, R., D. Campilan, and G. Prain.


2000. Dynamics in tropical
homegardens. Urban Agriculture
1:10-12.

Braun, A., G. Thiele, and M. Fernndez.


1999. La escuela de campo para MIP y
el comit de investigacin agrcola
local: plataformas complementarias
para fomentar decisiones integrales en
la agricultura sostenible. Manejo
Integrado de Plagas 53:1-23.
Braun, A.R. and E. van de Fliert. 1999.
Evaluation of the impact of sweetpotato
weevil (Cylas formicarius) and of the
effectiveness of Cylas sex pheromone
traps at the farm level in Indonesia.
lnternational Journal of Pest
Management 45:101-11 O.
Campilan, D. and G. Prain. 2000. Selfassessment as an approach to
evaluating participatory research: An
Asian experience. In: Lilja N., J. Ashby,
and L. Sperling (eds.). Assessing the
impact of participatory research and
gender analysis. CGIAR SWP-PRGA
(Consultative Group on lnternational
Agricultura! Research Systemwide
Program-Participatory Research and
Gender Analysis), Cali, Colombia.
p. 172-182.

Campilan, D., G. Prain, and


C.L. Bagalanon. 1999. Evaluation from
the inside: Participatory evaluation of
agricultura! research in the Philippines.
Knowledge, Technology, and Policy
11 : 11 4-1 3 1.
Caedo, V. and A. Lagnaoui. 2000. Estado
actual del uso de feromonas sexuales en
el Centro Internacional de la Papa
(CIP). In: Arning l. and A. Lizrraga
(eds.). Control etolgico: Uso de
feromonas, trampas de colores y de luz
para el control de plagas en la
agricultura sostenible. RAAA, Lima,
Peru. p. 73-80.
Castillo, G. and D. Campilan. 1999.
Nurturing a network: Lessons in the
evolution of the UPWARD program. In:
Niehof, A. and P. Terpstra (eds.).
Households in an interdisciplinary

CIP Program Report 1999 - 2000

471

perspective. Wageningen University,


Netherlands. p. 3 7-46.

Chavarriaga, P., M.M. Maya, J. Tohme,


M.C. Duque, C. Iglesias,
M. Bonierbale, S. Kresovich, and
G. Kochert. 1999. Using microsatellites,
isozymes, and FLPs to evaluate genetic
diversity and redundancy in the cassava
core collection and to assess the
usefulness of DNA-based markers to
maintain germplasm collections.
Molecular Breeding 5:263-273.

Chen, D., M. de la Vina, T. lnukai,


D.J. Mackill, P.C. Ronald, and
R.J. Nelson. 1999. Molecular mapping
of the blast resistance genes, Pi44(t),
derived from a durably resistant rice
cultivar. Theoretical and Applied
Genetics 98:1046-1053.
Cisneros, F. and N. Mujica. 1999. Impacto
del cambio climtico en la Agricultura:
Efecto del fenmeno del Nio en los
cultivos de la Costa Central. In: Per:
Vulnerabilidad frente al cambio
climtico. CONAM, Lima, Peru.
p. 115-135.

Collins, W.W., E.E. Carey, l.G. Mok,


P. Thompson, and O.P. Zhang. 1999.
Utilization of sweetpotato genetic
resources to develop insect resistance.
In: Clement, S.L. and S.S. Quisenberry
(eds.). Global genetic resources for
insect-resistant crops. CRC Press, Boca
Raton, FL, USA. p. 193-205.

Crissman, C.C., J.M. Antle, and


J. J. Stoorvogel. 2000. Tradeoffs in
agriculture, the environment and human
health: Decision support for policy and
technology managers. In: Lee D.R. and
C.B. Barret (eds.). Tradeoffs or
synergies? Agricultura! intensification,
economic development and the
environment. CABI Publishing,
Wallingford, UK. p. 135-150.

Danielsen, S., S.-E. Jacobsen,


J. Echegaray, and T. Ames. 2000.
Correlacin entre mtodos de
evaluacin y prdida de rendimiento en
el patosistema quinua-mildi.
Fitopatologa 35:242-248.

472

Selected Publications by CIP Staff

Day-Rubenstein, K. and K. Fuglie. 1999.


Resource allocation in joint publicprivate agricultura! research. Journal of
Agribusiness 17:123-134.
Day-Rubenstein, K. and K.O. Fuglie. 2000.
The CRADA model for public-private
collaboration in agricultura! research.
In: Fuglie K.O. and O.E.
Schimmelpfennig (eds.). Publ ic-private
collaboration in agricultura! research.
lowa State University Press, Ames, IA,
USA. p.155-174.

Di Feo, L., S.F. Nome, E. Biderbost,


S. Fuentes, and L.F. Salazar. 1999.
Etiology of sweet patato chlorotic dwarf
disease in Argentina. Plant Disease
84:35-39.

Di Feo, L., S.F. Nome, E. Biderbost,


S. Fuentes, and L.F. Salazar. 2000.
Etiology of sweetpotato chlorotic dwarf
disease in Argentina. Plant Disease
84:35-39.

Erselius L.J., M.E. Vega-Snchez, and


G.A. Forbes. 2000. Stability in the
popu lation of Phytophthora infestans
attacking tomato in Ecuador is
demonstrated by cellulose acetate
assessment of glucose-6-phosphate
isomerase. Plant Disease 84:325-327.
Estrada, R.O. and R. Quiroz. 2000.
Technological and institutional changes
affecting mixed crop-livestock
production systems in the Andes. In:
Tulachan P.M., M.A.M. Saleem,
J. Maki-Hokkonen, and T. Partap (eds.).
Contribution of livestock to mountain
livelihoods: Research and development
issues. ICIMOD-SLP-FAO-CIP
(lnternational lnstitute for lntegrated
Mountain Development-Systemwide
Livestock Program-Food and
Agriculture Organization of the United
Nations-lnternational Patato Center),
Kathmandu, Nepal. p. 83-93.

Fawzia, A., E.G. Karuri, and


V. Hagenimana. 1999. Sweet patato
ketchup: Feasibility, acceptability, and
production costs in Nairobi, Kenya.
African Crop Science Journal 7:81-89.
Finckh, M.R. and R.J. Nelson. 1999.
Phylogenetic and pathotypic analysis of

bacteria! blight race 3. European Journal


of Plant Pathology 105:743-751.
Fuentes, S. and L.F. Salazar. 2000. La
mosca blanca como transmisor de virus.
In: Valencia L., La mosca blanca en la
agricultura Peruana. Lima, Peru.

p. 55-72.
Fuglie, K.O., C. Narrod, and
C. Neumeyer. 2000. Public and private
investments in animal research. In:
Fuglie K.O. and O.E. Schimmelpfennig
(eds.). Public-private collaboration in
agricultura! research. lowa State
University Press, Ames, IA, USA.

p. 117-151.
Fuglie, K.O. 1999. Conservation tillage
and pesticide use in the Cornbelt.
Journal of Agricultura! and Applied
Economics 31:133-147.
Fuglie, K.O. 2000. Trends in agricultura!
research expenditures in the United
States. In: Fuglie, K.O. and O.E.
Schimmelpfennig (eds.). Public-private
collaboration in agricultura! research.
lowa State University Press, Ames, IA,
USA. p. 9-23.

Fuglie, K.O. and O.E. Schimmelpfennig


(eds.). 2000. Public-private
collaboration in agricultura! research.
lowa State University Press, Ames, IA,
USA. 354 p.

Fuglie, K.O., V.S. Khatana,


S. llangantileke, J.P. Singh~ O. Kumar,
and G. Scott. 2000. Economics of potato
storage in northern India. Quarterly
Journal of lnternational Agriculture

39:131-148.
Fuglie, K. 1999. lnvesting in agricultura!
productivity in Indonesia. Forum
Penelitian Agro Ekonomi 1 7: 1-16.

Ghislain, M., O. Zhang, O. Fajardo,


Z. Huamn, and R. Hijmans. 1 999.
Marker-assisted sampling of the
cu ltivated Andean potato Solanum
phureja collection using RAPO markers.
Genetic Resources and Crop Evolution

46:547-555.
Ghislain, M., M. Bonierbale, and
R. Nelson. 1999. Gene technology for
potato in developing countries. In:
Hohn, T. and K.M. Leisinger (eds.).

.Biotechnology of food crops in


developing countries. Springer Wien,
NY, USA. p. 135-140.

Golmirzaie, A.M., A. Panta, and J. Toledo.


1999. Advances in the conservation of
root and tuber crops. In: Benson, E.
(ed.). Plant conservation biotechnology.
Taylor and Francis, London. p. 165-178.
Graves, C. (ed.). 2000. La papa, tesoro de
los Andes. lnternational Potato Center,
Lima, Peru. 210 p.
Hagenimana, V. 1999. Micro-scale
enterprise approach to sweetpotato and
potato improvement systems. In:
Kwarteng, J. (ed.). Enhancing
postharvest technology generation and
dissemination in Africa and Mexico
City, Mexico. Sasakawa Africa
Association, CASIN, Geneva,
Switzerland. p. 20-26.

Hagenimana, V., M.A. Oyunga, J. Low,


S.M. Njoroge, S. Gichuki, and
J. Kabira. 1999. The effects of women
farmers' adoption of orange-fleshed
sweetpotatoes: Raising vitamin A intake
in Kenya. ICRW, Washington, OC, USA.

24 p.
He, W., Z.M. Zhang, and Y. Wang. 1999.
GILB meeting in Ecuador and late blight
research progress. Chinese Potato
Journal 13:182-183.

Hijmans, R.J., C.A. Forbes, and


T.S. Walker. 2000. Estimating the global
severity of potato late blight with a GISlinked disease forecast model. Plant
Pathology 49:697-705.

Hijmans, R.J., K.A. Garrett, Z. Huamn,


O.P. Zhang, M. Schreuder, and
M. Bonierbale. 2000. Assessing the
geographic representativeness of
genebank col lections: The case of
Bolivian wild potatoes. Conservation
Biology 14:1755-1765.

Hijmans, R.J., M. Schreuder, J. de la Cruz,


and L. Guarino. 1999. Using GIS to
check coordinates of germplasm
accessions. Genetic Resources and Crop
Evolution 46:291-296.
Huaccho, L. and R.J. Hijmans. 1999.
Creacin de una base de datos

CIP Program Report 1999- 2000

473

georeferenciada de la distribucin
global de la papa. Sistmica 1 :19-24.

Huamn Z., C. Aguilar, and R. Ortiz.


1999. Selecting a Peruvian sweetpotato
core col lection on the basis of
morphological, eco-geographical, and
disease and pest reaction data.
Theoretical and Appl ied Genetics
98:840-845.

Huamn, Z., R. Hoekstra, and


J.B. Bamberg. 2000. The inter-genebank patato database and the
dimensions of available wild potato
germplasm. American Journal of Potato
Research 77:353-362.

Huamn, Z., R. Ortiz, O.P. Zhang, and


F. Rodriguez. 2000. lsozyme analysis of
entire and core collections of Solanum
tuberosum subsp. andigena patato
cultivars. Crop Science 40:273-276.

Jacobsen, S.;.E., B. Joernsgaard,


J.L. Christiansen, and O. Stolen. 1999.
Effect of harvest time, drying technique,
temperature and 1ight on the
germination of quinoa (Chenopodium
quinoa). Seed Science and Technology
27:937-944.

Jensen, C.R., S.-E. Jacobsen,


M.N. Andersen, N. Nuez,
S.D. Andersen, L. Rasmussen, and
V.O. Mogensen. 2000. Leaf gas
exchange and water relations of field
quinoa (Chenopodium quinoa Willd.)
during soil drying. European Journal of
Agronomy 13:11-25.

Jorge, V., M. Fregene, M.C. Duque,


M. Bonierbale, J. Tohme, and
V. Verdier. 2000. Genetic mapping of
resistance to bacteria! blight disease in
cassava (Manihot esculenta Crantz).
Theoretical and Applied Genetics
101 :865-872.

Khatana, V.S., S. Arya, and


S.G. llangantileke. 1999. Decline in
sweet patato cultivation in India with
special reference to the state of Bihar.
Asian Agri-History 2:93-110.

Kumar, J., R.J. Nelson, and R.S. Zeigler.


1999. Population structure and
dynamics of Magnaporthe grisea in the

474

Selected Publications by CIP Staff

India Himalayas. Genetics 152:971984.


Lemaga, B. 2000. Joint effort averts food
shortage. AgriForum (Newsletter) 12:11.

Len-Velarde, C., R. Quiroz,


P. Zorogasta, and M. Tapia. 2000.
Sustainability concerns of livestockbased livelihoods in the Andes. In:
Tulachan P.M., M.A.M. Saleem,
J. Maki-Hokkonen, and T. Partap (eds.).
Contribution of livestock to mountain
livelihoods: Research and development
issues. ICIMOD-SLP-FAO-CIP
(lnternational lnstitute for lntegrated
Mountain Development-Systemwide
Livestock Program-Food and
Agriculture Organization of the United
Nations-lnternational Patato Center),
Kathmandu, Nepal. p. 183-202.

Lizrraga, C., M. Querci, M. Santa Cruz,


l. Bartolini, and L.F. Salazar. 2000.
Other natural hosts of potato virus T.
Plant Disease 84:736-738.

Mihovilovich, E., H. Mendoza, and


L.F. Salazar. 2000. Combining ability
for resistance to sweetpotato feathery
mottle virus. HortScience 35:13191320.
Mujica, E. and J.C. Alurralde. 2000. La
gestin integral del agua en
Cochabamba. Sntesis de un foro
electrnico. Comisin para la Gestin
Integral del Agua en Cochabamba y
Consorcio para el Desarrollo Sostenible
de la Ecorregin Andina, Cochabamba,
Bolivia. 254 p.
Mujica, A. and S.-E. Jacobsen. 1999.
Agrobiodiversidad de las "aynokas" de
quinua (Chenopodium quinoa Willd.) y
la seguridad alimentaria. Revista
lngeneria Quimica y Desarrollo
Regional 6:26-28.
Mujica, A. and S.-E. Jacobsen. 2000.
Potencial y perspectivas futuras de la
quinua (Chenopodium quinoa Willd.) en
el Per. Mosaica Cientfico (1999)
2:56-60.
Narrod, C. and K. Fuglie. 2000. Private
investment in livestock breeding with
implications for public research policy.

Agribusiness: An lnternational Journal


16:457-470.

Ordez, M.E., H.R. Hohl, A. Velasco,


M.P. Ramon, P.J. Oyarzun., C.D. Smart,
W.E. Fry, G.A. Forbes, and L.J. Erselius.
2000. A novel A2 population of
Phytophthora, similar to P. infestans,
attacks wild Solanum species in
Ecuador. Phytopathology 90:197-200.
Ortiz, O. 1999. Understanding interactions
between indigenous knowledge and
scientific information. lndigenous
Knowledge and Development Monitor
7:7-1 o.
Ortiz, O. 2000. Los agroqumicos no son
la nica opcin para luchar contra las
plagas. In: El Medio ambiente en el
Per: Ao 2000. Instituto CuantoUSAID (United States Agency for
lnternational Development). p. 83-94.

Ortiz, R., S. Madsen, E.N. Ruiz-Tapia,


S.-E. Jacobsen, A. Mujica-Sanchez,
J.L. Christiansen, and O. Stolen. 1999.
Validating a core collection of Peruvian
quinoa germplasm. Genetic Resources
and Crop Evolution 46:285-290.
Prain, G. and M. Hermann. 2000. Guide
to lndonesian sweetpotato genetic
resources. CD-ROM. lnternational
Potato Center, Lima, Peru.
Prain, G. 1999. Households and social
networks in agricultura! research. In:
N iehof, A. and P. Terpstra (eds.).
Households in an interdisciplinary
perspective. Wageningen University,
Netherlands. p. 47-65.
Prain, G. 1999. Overview: Local
maintenance of crop biodiversity in
the Philippines. In: Conservation and
change: Farmer maintenance of agrobiological diversity in the Philippines.
UPWARD, Los Baos, Philippines.

p. 1-13.

J. Hagman. 2000. Synthesis:


Farmers' management of diversity in
local systems. In: Almekinders, C. and
W. de Boef (eds.). Encouraging
diversity: The conservation and

Prain, G. and

development of plant genetic resources.

lntermediate Technology Publications,


London, U K. p. 94-1 OO.

Prain, G. and M. Piniero. 1999. Farmer


management of rootcrop genetic
diversity in southern Philippines. In:
Prain, G., S. Fujisaka, and M. Warren
(eds.). Biological and cultural diversity:
The role of indigenous agricultura!
experimentation in development. ITP,
London, UK. p. 92-112.

Prain, G.,

J. Schneider, and C. Widyastuti.

2000. Farmers' maintenance of


sweetpotato diversity in Iran Jaya. In:
Almekinders, C. and W. de Boef (eds.).
Encouraging diversity: The conservation
and development of plant genetic
resources. lntermediate Technology
Publications, London, UK. p. 54-59.
Pray, C.E. and K.O. Fuglie. 2000. The
prvate sector and international transfer
of agricultura! technology. In: Fuglie,
K.O. and O.E. Schimmelpfennig (eds.).
Public-private collaboration in
agricultura! research. lowa State
University Press, Ames, IA, USA.
p. 269-299.
Priou, S. and M. El Mahjoub. 1999.
Bacteria! and fungal diseases in the
major potato-growing areas of Tunisia.
European Plant Protection Organization
(EPPO) Bulletin 29:167-171.
Priou, S., L. Gutarra, and P. Aley. 1999.
Highly sensitive detection of Ralstonia
solanacearum in latently infected
potato tubers by post-enrichment ELISA
on nitrocellulose membrane. European
Plant Protection Organization (EPPO)
Bulletin 29:117-125.

Priou, S., L., Gutarra, H. Fernndez, and


P. Aley. 1999. Sensitive detection of
Ralstonia solanacearum (race 3) in soil
by post-enrichment DAS-ELISA. ACIAR
Bacteria! Wilt Newsletter 16:10-13.

Quiroz, R., C. Len-Velarde, and


W. Bowen. 2000. FSR from a modelling
perspective: Experiences in Latn
America. In: Collinson, M. (ed.). A
history of farming systems research.
FAO (Food and Agricu lture
Organization of the United Nations),
Rome, ltaly, and CABI Publishing,
Wallingford, UK. p. 342-354.

CIP Program Report 1999 - 2000

475

Quiroz, R., W.T. Bowen, and A. Gutarra.


1999. lntegrating remote sensing and
dynamic models to assess pasture and
livestock production at the ecoregional
level: Developments in the Altiplano.
In: Thornton, P.K. and A.N. Odero (eds.).
Proceedings of the workshop on
ecoregional research at ILRI, Addis
Ababa, 5-8 October 1998. ILRI,
Nairobi, Kenya. p. 97-103.

Raymundo, A.K., A.M. Briones,


E. Y. Ardales, M. T. Perez,
L.C. Fernandez, J.E. Leach, T.W. Mew,
M.A. Ynalvez, G.C. Melaren, and
R.J. Nelson. 1999. Analysis of DNA
polymorphism and virulence in
Philippine strains of Xanthomonas
oryzae pv. oryzicola. Plant Disease
83:434-440.

Reeves, A.F., O.M. Olanya, J.H. Hunter,


and J.M. Wells. 1999. Evaluation of
potato varieties and selections for
resistance to bacteria! soft rot.
American Journal of Potato Research
76:183-189.
Rivera, B., J. Posner, and E. Mujica. 2000.

Educacin de posgrado y manejo de


recursos naturales en la ecorregin
andina. Consorcio para el Desarrollo
Sostenible de la Ecorregin Andina y
Universidad de Caldas, Manizales,
Colombia. 192 p.

Roa, A.C., P. Chavarriaga-Aguirre,


M.C. Duque, M.M. Maya, M.W.
Bonierbale, C. Iglesias, and J. Tohme.
2000. Cross-species ampl ification of
cassava (Manihot esculenta)
(Euphorbiaceae) microsatellites: Allelic
polymorphism and degree of
relationship. American Journal of
Botany 87:1647-1655.
Roncal, J., L. Gutarra, and S. Priou. 1999.
Rapid differentiation of strains of
Ralstonia solanacearum by restriction
analysis of PCR-amplified fragments.
ACIAR Bacteria! Wilt Newsletter
16:7-10.

Salazar, L.F., G. Mller, M. Querci,


J.L. Zapata, and R.A. Owens. 2000.
Patato yellow vein virus: lts host range,
distribution in South America and

476

Selected Publications by CIP Staff

identification as a crinivirus transmitted


by Trialeurodes vaporariorum. Annals of
Applied Biology 137:7-19.

Salazar, L.F., l. Bartolini, and V. Flores.


2000. Evidence for the existence of
pyyNTN in the Andes and a hypothesis
towards its origin. Fitopatologa
35:87-90.
Scheidegger, U. and G. Prain. 2000.
Support to diversity in patato seed
supply. In: Almekinders, C. and W. de
Boefs (eds.). Encouraging diversity: The
conservation and development of plant
genetic resources. lntermediate
Technology Publications, London, UK.
p. 232-236.

Scott, G., M. Rosegrant, and C. Ringler.


2000. Global projections for root and
tuber crops to the year 2020. Food
Policy 25:561-597.

Scott, G.J., R. Best, M. Rosegrant, and


M. Bokanga. 2000. Roots and tubers in
the global food system: A vision
statement to the year 2020. CIP-CIATIFPRl-llTA-IPGRI (lnternational Potato
Center-Centro Internacional de
Agricultura Tropical-lnternational Food
Poi icy Research lnstitute-lnternational
lnstitute of Topical Agriculturelnternational Plant Genetic Resources
lnstitute). Printed at CIP, Lima, Peru.
111 p.

Sherwood, S.G., R. Nelson, G. Thiele, and


O. Ortiz. 2000. Farmer field schools in
potato: A new platform for participatory
training and research in the Andes.
LEISA 16(4):24-25.
Sherwood, S. and J. Chenier. 2000.
ANAFAE and COLABORA: Lessons from
experiences with collaborative networks
for sustainable agriculture and natural
resource management. In: NGOresearch partnerships (RESPAR).
lnternational lnstitute for Rural
Reconstruction, Cavite, Philippines.
33 p.
Sherwood, S.G. and N.T. Uphoff. 2000.
Soil health: Research, practice and
policy for a more regenerative
agriculture. Applied Soil Ecology
15:85-97.

Spooner, D.M., A. Salas, Z. Huamn, and


R.J. Hijmans. 1999. Potato germplasm
collecting expedition in southern Peru
(Departments of Apurmac, Arequipa,
Cusco, Moquegua, Puno, and Tacna) in
1998: Taxonomy and new genetic
resources. American Journal of Potato
Research 76:103-119.

Trognitz, B.R., S. Carrin, and


M. Hermann. 2000. Expression of stylar
incompatibility in the Andean clonal
tuber crop oca (Oxalis tuberosa Mol.,
Oxalidaceae). Sexual Plant
Reproduction 13:105-111.
Van de Fliert, E. 1999. Women in IPM
training and implementation in
Indonesia. In: Van de Fliert, E. and
J. Proost (eds.). Gender and IPM: Crop
protection practices and strategies.
Royal Tropical lnstitute, Amsterdam.
p. 71-78.

Vega-Snchez, M.E., L.J. Erselius,


A.M. Rodriguez, O. Bastidas,
H.R. Hohl, P.S. Ojiambo, J. Mukalazi,
T. Vermeulen, W.E. Fry, and G.A.
Forbes. 2000. Host adaptation to potato
and tomato within the US-1 clonal
lineage of Phytophthora infestans in
Uganda and Kenya. Plant Pathology
49:531-539.

Vivanco, J., M. Querci, and L.F. Salazar.


1999. Antiviral and antiviroid activity of
MAP-containing extracts from Mirabilis
jalapa roots. Plant Disease 83:11161121.

Walker, T., P. Schmiediche, and


R. Hijmans. 1999. World trends and
patterns in the potato crop: An
economic and geographic survey.
Potato Research (42):241-264.

Walker, T., S. Swinton, R. Hijmans,


R. Quiroz, R. Valdivia, M. Holle,
C. Len-Velarde, and J. Posner. 2000.
Technologies for the tropical Andes.
Policy Brief No. 3. In: Pender, J. and
P. Hazel 1 (eds.). Focus 4 series.
Promoting sustainable development in
less-favored areas. IFPRI, Washington,
OC, USA.
Walker, T. 2000. Reasonable expectations
on the prospects for documenting the
impact of agricultura! research on
poverty in ex-post case studies. Food
Policy 25:515-530.

Wiegers, E.S., R.J. Hijmans, D. Herv,


and LO. Fresco. 1999. Land use
intensification and desintensification in
the Upper Caete valley, Peru. Human
Ecology 27:319-339.

Zeddam, J.L., J.L. Rodriguez, M. Ravallec,


and A. Lagnaoui. 1999. A noda-like
virus isolated from Spodoptera eridania
(Cramer) (Lep.: Noctuidae). Journal of
lnvertebrate Pathology 74:267-274.
Zhang, O.P., G. Cipriani, l. Rety,

A. Golmirzaie, N. Smit, and D.


Michaud. 2000. Expression of protease
inhibitors in sweetpotato. In: Michaud,
D. (ed.). Recombinant protease
inhibitors in plants. Landes Bioscience,
Georgetown, TX, USA. p. 167-178.

Zhang, O.P., J. Cervantes, Z. Huamn,


E. Carey, and M. Ghislain. 2000.
Assessing genetic diversity of sweet
patato (lpomoea batatas (L) Lam.)
cultivars from tropical America using
AFLP. Genetic Resources and Crop
Evolution 47:659-665.

CIP Program Report 1999 - 2000

477

Global Contact Points


CI P Headquarters
lnternational Potato Center (CIP)
Avenida La Universidad 795, La Molina
Apartado 1558
Lima 12, Peru
Tel: +51 1 349 6017

Fax: +51 1 317 5326


E-mail: cip@cgiar.org
Website: www.cipotato.org

Networks
CONDESAN (Consortium for the Sustainable Development of the Andean
Ecoregion)
(same address, telephone, and fax as CIP
headquarters)
E-mail: condesan@cgiar.org
Website: www.condesan.org
Contact: Elias Mujica, Acting Coordinator
GILB (Global lnitiative on Late Blight)
(same address, telephone, and fax as CIP
headquarters)
E-mail: gilb@cgiar.org
Website: www.cipotato.org/gilb
Contact: Wanda Collins, GILB Coordinator

GMP (Global Mountain Program)


(same address, telephone, and fax as CIP
headquarters)
E-mail: r.quiroz@cgiar.org
Contact: Roberto Quiroz, Program Coordinator
SIUPA (CGIAR Strategic lnitiative on
Urban and Peri-Urban Agriculture)
(same address, telephone, fax as CIP
headqu arters)
E-mail: g.prain@cgiar.org
Website: www.cipotato.org/siupa
Contact: Gordon Prain, SIUPA Coordinator

Latin America and the Caribbean


(LAC)
Regional Office Peru
(same address, telephone and fax as CIP
headquarters)
Direct tel: +51 1 317 5315
E-mail: cip-lac-office@cgiar.org
Website: www.cipotato.org/regions/
lac.htm

CI P Regional and Liaison Offices

LATIN AMERICA ANO


(LAC)
EASTANOSOUTHEAST
ASIAANOTHE
PACIFIC (ESEAP) ,.,..., .. .. .. . '

SUB-SAHARANAFRICA(ssA)

478

Global Contact Points

SOUTH ANO WEST ASIA (swA)

Contact: Fernando Ezeta, LAC Regional


Representative
Andean Potato Project (Papa Andina)
Bolivia, Ecuador, Peru
(same address, telephone and fax as CIP
headquarters)
E-mail: a.devaux@cgiar.org or
thiele@cip.org.ec
Website:www.cipotato.org/papandina
Contacts: Andr Devaux, Project Coordinator, and Graham Thiele, Participatory
Research and Training Specialist
Liaison Office Ecuador
lnternational Potato Center
Estacin Experimental Santa Catalina
Km 17.5 Panamericana Sur
Sector Cutuglagua Canton Meja
Apartado 17-21-1977
Quito, Ecuador
Tel: +593 2 690 362/690 363/694 923
Fax: +593 2 692 604
E-mail: cip-quito@cgiar.org
Website: www.quito.cipotato.org
Contact: Gregory Forbes, Liaison Scientist

Sub-Saharan Africa (SSA)


Regional Office Kenya
lnternational Potato Center
PO Box 25171
Nairobi, Kenya
Tel: +254 2 632 054
Fax: +254 2 630 005 or 631 499
Telex: 22040
Cable: CIPAPA, Nairobi
E-mail: cip-nbo@cgiar.org
Contact: Peter Ewell, SSA Regional
Representative
Liaison Office Uganda
lnternational Potato Center
e/o PRAPACE (see address below)
Contact: Berga Lemaga, PRAPACE Coordinator

Networks
PRAPACE (Regional Potato and
Sweetpotato lmprovement Program for East
and Central Africa)

Plot 106, Katalima Road, Naguru


PO Box 22274
Kampala, Uganda
Tel: +256 41 286 209
Fax: +256 41 286 947
E-mail: prapace@infocom.co.ug or
be.rga@imul.com
Contact: Berga Lemaga, PRAPACE Coordinator

South and West Asia (SWA)


Regional Office India
lnternational Potato Ce~ter
e/o IARI Campus, Pusa
New Delhi 110012, India
Tel: +91 11 585 0201
Fax: +91 11 573 1481
E-mail: cip-delhi@cgiar.org
Contact: Sarath llangantileke, SWA
Regional Representative
Project CIP-SDC (Potato Development
Project for Bhutan, Nepal, and Pakistan)
PO Box 2122
lslamabad, Pakistan
Tel: +92 51 925 5067 and 925 5040, ext.
3121
Fax: +92 51 925 5034
E-mail: ohidalgo@isb.comsats.net.pk or
o.hidalgo@cgiar.org
Contact: Osear A Hidalgo, Project Leader

East and Southeast Asia and the


Pacific (ESEAP)
Regional Office Indonesia
lnternational Potato Center
Kebun Percobaan Muara,
Jalan Raya Ciapus
Bogor 1661 O, Indonesia
Tel: +62 251 317 951
Fax: +62 251 316 264
E-mail: cip-bogor@cgiar.org
Website: www.eseap.cipotato.org
Contact: Keith Fugl ie, ESEAP Regional
Representative
Liaison Office China
lnternational Potato Center
e/o The Chinese Academy of Agricultura!
Sciences

CIP Program Report 1999 - 2000

479

12 Zhong Guan Cun South Street


West Suburbs of Beijing
Beijing, People's Republic of China
Tel: +86 1 O 6897 5504
Fax: +86 1 O 6897 5503
Telex: 716 22233 or 222720 CAASCN
Cable: AGRIACA
E-mail: cip-china@cgiar.org
Website: www.eseap.cipotato.org/cipch ina
Contact: Yi Wang, Liaison Scientist
Liaison Office Vietnam
lnternational Potato Center

Sol 4 825
Nam Thanh Cong
Lang Ha, Dong Da
Hanoi, Vietnam
Tel and fax: +84 4 835 5494
E-mail: cip-hanoi@fpt.vn or
d.peters@cgiar.org or dpeters@fpt.vn
Contact: Dai Peters, Postharvest Specialist

Networks
ANSWER (Asian Network for Sweetpotato
Genetic Resources)
c/o CIP-ESEAP Regional Office
JI. Raya Ciapus
Bogor, Indonesia

480

Global Contact Points

Telephone: 62-251-317951
Fax: 62-251-316264
E-mail: CIP-Bogor@CGIAR.Org or
S.Mahalaya@CGIAR.Org
Website: www.eseap.cipotato.org/answer/
Contacts: Algerico M. Mariscal, ANSWER
coordinator

UPWARD (Users' Perspectives with


Agricultura! Research and Development)
PCARRD Complex
Los Baos, Laguna, 4030 Philippines
c/o IRRI
PO Box 3127
Makati City, 1271 Philippines
Tel: +63 49 536 0235
Fax: +63 49 536 1662
E-mail: cip-manila@cgiar.org
Website: www.eseap.cipotato.org/u pward
Contact: Dindo Campilan, UPWARD
Coordinator

* Central and Eastern Europe, Transcaucasia, and Central Asia (ECA): closed
in 2000

You might also like