You are on page 1of 7

Shrinkage, Porosity and Bulk Density of Foodstuffs

at Changing Moisture Contents


J. E. LOZANO, E. ROTSTEIN and M. J. URBICAIN

ABSTRACT interphase velocity w becomes important. The transport


Dependabledata on bulk density,volumetric shrinkagedue to water theorem is:
loss and porosity are needed to model processessuch as drying,
packagingand storing.Experimentaldataarepresentedfor all three .Iv~(~)ddV = .fvb$$V + m&(t) !w.n dA (3)
properties.It is possibleto model the water-loss-based
bulk shrink- dt
agecoefficientto obtain a predictiveequationbasedon composition for the case J/ = 1 it follows:
of the foodstuff. From this, a generalizedcorrelation is obtained
which predicts bulk shrinkagecoefficient knowing only the initial dSb 1
-=- &b(t) w-n dA (4)
moisture content of the food. Porositiesfor the foodstuffs consid-
ered can be predictedthrough suitablecorrelations,but there is no 4 w
generalizedequationspanningall foods. where sb is the bulk shrinkage coefficient (Lozano et al.,
1980):
INTRODUCTION VW)
sb = ___ (5)
MANY FOOD PROCESSING OPERATIONS can be suc- vb,o
cessfully designed or studied on the basis of sound engi-
neering modelling. There is a family of operations which Thus, in this approach it is necessary to know or to be
involve the interaction between a water containing food- able to predict the change of sb as a function of moisture
stuff and water in the surrounding atmosphere at the pre- content. This type of information is also needed when
vailing temperature and pressure, such as drying, packaging, simplified effective diffusivity models are used (Charm,
storing and other similar operations. Modeling in these cases 1978).
is based on properly setting up the heat and mass transfer There are very few data on porosity, bulk density and
equations, coupling them by means of a suitable hypothe- bulk shrinkage coefficient as a function of moisture content.
sis (usually, local equilibrium) and solving the resulting Kilpatrick et al. (1955) studied volume shrinkage of pota-
system of differential equations. This approach requires a toes and other vegetables as drying proceeds. Charm (1978)
knowledge of transport properties in general and porosity reported on volumetric contraction of meat and potatoes.
in particular. Chirife (1969) provided data for apples and potatoes.
Representative heat and mass transfer equations and the Gijrling (1958) studied shrinkage during the drying of
role porosity play in them, have been recently outlined by macaroni. Suzuki et al. (1976) investigated the shrinkage
Lozano et al. (1980). When these equations are used, there in dehydration of root vegetables. Shrinkage and porosity
is need to account for the change in porosity and the over- of apple tissue at different moisture contents were reported
ail shrinkage of the samples as they lose moisture. One pos- by Lozano et al. (1980).
sible approach is the use of a fictitious length z (Rom$n It would be useful to use this information as a basis to
et al., 1982), so that every change of real length Ax can be obtain general predictive correlations. Gcirling (1958),
transformed into the change of fictitious length through the Chirife (1969) and Charm (1978) have used the following
expression: analogy of the thermal expansion equation;
A Vb
AZ = LAX ( > --@AX (6)
(1 f xl P vbo
where &, is the bulk density of the sample and p is defined In this case, there is need to obtain the specific linear
below: shrinkage coefficient value for each foodstuff and Q!is not
valid over the entire range of X.
P( p) Kilpatrick et al. (1955) suggested an equation valid for
l--E x=l) the early stages of drying. It requires prior knowledge of
the initial moisture content and the density of the dried
The above model requires data on porosity and bulk material.
density as a function of moisture content. Suzuki et al. (1976) developed three equations which
An interesting potential application of data on volume apply to three different drying models: uniform drying,
change as a function of moisture content results from the core drying, and semicore drying. The first model results
drying theory recently put forward by Whitaker (1980). in two alternate equations: one needs data for equilibrium
Cellular tissue making up the solid foodstuff may be re- moisture contents and bulk density, while the other requires
garded as a multiphase system and, by making use of the the initial moisture content and bulk density of the mater-
transport theorem and the averaging theorem, the macro- ial. The second and third model need the initial and equilib-
scopic transport equations are developed. In this case the rium values for moisture and bulk density. Suzukis and Kil-
Patricks equations are summarized in Table 1 to facilitate
comparison with the one developed in the present work.
Authors Lozano, Rotstein and Urbicain are affiliated with Planta The objective of this paper is to present data on bulk
Pilot0 de lngenieria Quinica, (UNS-CONICETI, 8000 Bahia Mama, density, porosity, and bulk shrinkage coefficients for sev-
Argentina. eral foods, as they change in moisture contents. The data
are correlated to simplify further use. Both in the case of
Volume 48 (1983)-JOURNAL OF FOOD SCIENCE-1497
SHRINKAGE, POROSITY AND BULK DENSITY. ..

the bulk shrinkage coefficient and porosity an attempt was The procedures for measurement of bulk density and particle den-
made to obtain predictive correlations based as composi- sity were the same as reported in a previous contribution (Lozano
tion. The attempt was particularly successful in the case of et al., 1980).
sb> which could be predicted over the entire moisture
range with only a knowledge of the initial moisture content Calculations
of the food. Porosity, E, was calculated on the basis of the density measure-
ments:
MATERIALS & METHODS pp - Pb
e=- (12)
THE FOODS SELECTED were pears, carrots, potatoes, sweet PP
potatoes, and garlic. Typical compositions are shown in Table 2
(Watt and Merrill, (1963). Except for garlic the samples were cut where po, the particle density, is based on the particle volume ex-
from the fresh product in the shape of cylinders, 1 cm in diameter, cluding all internal pores which are connected to the surrounding
4 cm long. Two sets of garlic samples were prepared: whole pieces air space and pb is the bulk particle density which is based on the
or slices obtained cutting one piece in half along the longest axis. particle volume including these internal pores. The density pp was
In the case of carrots the cylinders were cut in an angle with respect determined by measuring the particle volume in a pneumatic psycro-
to the longest axis, so that the sample had balanced amounts of the meter in which the compressed air volume included the volume of
. inner core and the external phloem and parenchymatic tissue. air inside the open internal pores and pb was determined by using
Drying was carried out in a conventional pilot air drier, with a mesh buoyant force to measure the volume of water displaced by the
wire tray and through circulation of air. Air velocity was 1 m/set particle when immersed.
in all cases. For foodstuffs containing starch the dry bulk tempera- The bulk volume at each moisture content was calculated from
ture was 40 f lC, relative humidity 35%. For other foodstuffs, the bulk density and the corresponding sample weight W:
60 * lC and 30%. After drying, all samples were tightly wrapped
in polyethylene film and placed into a thermostatic chamber for 24 ,= (13)
hr, to homogenize water and temperature profiles across the sample. pb

Table 1 -Previous models for bulk shrinkage coefficient prediction


RESULTS & DISCUSSION
Model Equation FIG. 1 SHOWS bulk density vs moisture content for the
foodstuffs studied. It can be seen that there is no homo-
Kilpatrick et al (1955) X + 0.8
Sb = (7) geneity of behavior. Bulk density increased in carrots and
X0 t 0.8 pears with decreasing moisture content, while in the case of
Suzuki et al. (1976) potato, sweet potato and garlic it increased up to a certain
Uniform drying (a) =-
X+a
csi moisture content and then decreased with decreasing mois-
Sb Xo+a ture content. Particle density (Fig. 2) for all foods but car-
a = x,$&1)+& rots increased with decreasing moisture contents. For car-
rots it increased to X/X, = 0.075 and then decreased as X
(bl sb = bX+c (9) tended to zero.
The bulk and particle densities can be correlated with an
equation of the type:
b =%fTI .I x7
C = 1 +b-p,,,-, p=h+llc +pexp(-q:) (14
A0 A0
Core drying model Sb = ++I (10)
0

K = , .jxa + )pb,o x~
(X,+l)p, x0-xX,

semicore drying model = rX+n (11)


sb
X0

= x, (1 - E)
r
xo-x,-zz~xo-xe+~-l)

(x, + 1 b,,,o
E = (Xo+l)Pb,e

= Pb,ec - (1 - X)Pb,e
z
pb,o

~Xo-xe-Z(~Xo-Xe+~-1)
n = xo-xe-z(xo-xe-t~-l)

Table P-Typical composition of foodstuffs studied

x0 xcw
Foodstuff (kg/kg) (kg/kg) 1
01 0.2 03 04 05 0.6 W 08 09

Carrot 7.55 0.64 - 0.36 VL


Garlic 2.28 0.80 - 0.20
Pear 7.03 0.83 0.17
Fig. I-Bulk density vs moisture content (Units of pp are 103
Potato 4.50 0.15 0.80 0.05
kg/m3). (1) Carrot; (2) Pear; (3) Potato; (4) Sweet Potato; (5)
Sweet potato 2.55 0.22 0.75 0.03
Garlic (whole pieces); (6) Garlic (sliced).

1498-JOURNAL OF FOOD SCIENCE-Volume 48 (7983)


Table 3 shows the corresponding values of the constants for carrots and pears and in Fig. 4 for potatoes, sweet pota-
and the coefficient of determination r2 obtained from ap- toes and garlic. The corresponding coefficients for Eq (14)
plying at least square nonlinear parameter algorithm (Mar- are shown in Table 3. The shape of the other porosity
quardt, 1963). The one exception is the particle density of lines indicate that the structural changes in the other food-
sweet potato for which the fit of Eq (14) is poor and the stuffs are of a different nature.
following correlation is more adequate (r* = 0.97): Fig. 5 shows the change in bulk shrinkage coefficients of
the foodstuffs, as a result of drying. The bulk shrinkage
(p,), = 1.553 - 4.954; + 4.630 (;) 1.051 (15) coefficient is plotted as a function of moisture content.
0 0 With the exception of carrots and pears, the slope of sh
becomes noticeably less steep for X/X, < 0.15. This is
Note that carrots are the only ones showing the pecu-
important because it indicates that all linear predictive
liar result of decreasing pp in the last stages of drying. Since
equations for sh will fail to cover the entire range 0 <
the porosity measured is that due to open pores, the radical X/X, < 1. Not less important is the fact that the range 0 <
drop in measured porosity appears to be due to a closing X/X, < 0.15 is very significant in modelling drying opera-
off of these pores, as shown by apples by Lozano et al. tions. For instance, for potato (Crapiste and Rotstein,
(1980). The dotted line in Fig. 2 projects the change in pp
that would have occurred if the pores had remained open. 1982) it corresponds to X = 0.68, a value which is in equi-
Porosity values as a function of moisture content were librium with an atmosphere of relative humidity @ > 0.9.
calculated using Eq (12). The results are plotted in Fig. 3 In other words, this range of X at which there is a change of
slope in sb is the one where most of the modeling and
drying simulation is done. As a result, the ability to predict
values of sh within this range should be a required feature
of the models under consideration. Fig. 6 indicates that
this is accomplished by the new correlation.
PP
1.5

E
I..

------___

01 0.2 03 04 0.5 0.6 M 09


x/x0

Fig. 3-Porosity dependence on moisture con tent: experimental


(full line) and predicted (dotted line) data. (1) Pear; (2) Carrot.
03 a2 03 0.4 a5 a6 0.7 00 09
X/X

Fig. 2-Particle density as a function of moisture content (Units of


PP are 103 kg/m3). (11 Carrot; (2) Pear; (3) Potato; (4) Sweet po-
tato; (5) Gailic.

Table d-constants and coefficient of determination corresponding


to Eq (14)

Foodstuff density h !2 P q r2 20

Carrot bulk 0.984 0 0.224 1.800 0.97


particle 1.497 -0.294 -0.253 39.793 0.96
particle (true 1.497 -0.294 0.033 36.820 0.97
IO
porosity)
Pear bulk 1.251 -0.153 -0.107 1.33x10- * 0.97
particle 0.832 0.220 0.632 2.775 0.97
Potato bulk 1.202 -0.148 0.259 15.507 0.96
particle 1.234 -0.117 0.085 19.040 0.97
Sweet 0 a2 0.3 OX 0.5 06 a7 00 a9 1.0
potato bulk 1.266 -0.219 -0.319 6.700 0.96 X/x,
bulk (whole 3.260 1 .I 72 -2.325 -0.395 0.96
piece)
Fig. 4-Porosity as a function of moisture content. Experimental
Garlic bulk (sliced) 1.130 -0.567 0.187 -0.866 0.95
(full line) and predicted (dotted line) data. (1) Potato; (2) Sweet
particle 2.694 0 -1.316 -0.1638 0.95
potato; (31 Garlic.

Volume 48 (1983)-JOURNAL OF FOOD SCIENCE-1499


SHRINKAGE, POROSITY AND BULK DENSITY. .,

Volume shrinkage modeling high degrees of oversaturation and they approach asympto-
It appears as desirable to be able to predict volume tically the density of solid sugar, Eq (21) can be used over
shrinkage of a foodstuff without having to measure the the entire range of moisture content:
property itself. This can be accomplished provided compo- Xsp
sition data are available such as that shown in Table 2, i.e. pm (X) = 0.997 exp (0.413 ~ (2.1)
amount of the main constituents per unit weight of dry x+&J
matter. Since the focus is on volume, the other information The other intervening densities are regarded as constant,
required is density of the same constituents. neglecting swelling due to water sorption. The correspond-
For the foodstuffs under consideration, the main non- ing values (Kirk-Othmer, 1964) are:
water constituents are sugar, starch and the cell wall mater-
ials. Cellular membrane material can be neglected because pst = 1.64 x lo3 kg/m3
of its small contribution to the total weight. The weight pew = 1.55 x 1O3 kg/m3
fraction of the above constituents results from a material By systematic addition of volume contribution of con-
balance on a representative sample and it can be calculated stituents, it is possible to predict the bulk shrinkage coeffi-
from the following exprgssions: cient with the following expression:
. %%i
&=J (16) sb = I/-$+?(x) I[
z +Z di
J. mshj i 0
where :
%w
XIX= (17) (23)
zz +Z di
J. msgj i
dsi (xsg+ pm (W MM
xst c (18) Y= (24)
+ z di
x0
J -%I! i

where rn%j is the amount of sugar j and di the amount of


nonsugar dry matter i involved in the material balance.
M=k+Xst (25)
Note that the total amount of dry matter, md, is: Pew Pst

c 1 -Em;) Psn (X0>


md =jrnsg~ .+C i di (19) t= (26)
1 -E(X) p,(X)
and:
&g+xcw+xst=1 (20)
Density of sugar solutions has been reported by Honing

sb,l
(1953). Assuming water is mainly present as a solvent for
the sugars, considering that sugar solutions tend to reach

1.,
I
1 n7C / /- b

0 0.1
x/x0

Fig. b-prediction of bulk shrinkage coefficient at low moisture


Fig. 5-B&k shrinkage coefficient of different foodstuffs as a func- contents. la) Eq (22); /b) linear extrapolation of Eq (22) through
tion of moisture con tent. 0.2 < x/x, G 1.

1500-JOURNAL OF FOOD SCIENCE-Volume 48 (19831


The pore volume was added through use of porosity While Suzukis and Kilpatricks equation fit the sb data
values. In the case of carrots total porosity, as described reasonably well at high X/X, values, they fit the data less
above, was considered. well than the presently proposed sh vs X/X, correlation
Eq (22) requires a knowledge of porosities to predict when X/X, < 0.15. Moreover, they fail to indicate the
bulk volumes and as such it is not very useful. Fortu- curvature in sb vs X/X, which is encountered when X/X,
nately, sample calculations show that: [ + 1 in the range; < 0.15. Fig. 7 shows the case of sweet potato, to illustrate
0.01 <(X/X,><l. this point.
Thus, Eq (11) can be approximated by: Having shown that Eq (27) is a good predictive correla-
tion for foodstuffs where &, M and X0 are known, a rea-
s,=fl;+,(x) (27) sonable task is to see whether a more general expression
0 can be built on its basis so as to predict the bulk shrinkage
coefficient starting from less information about the food-
Expanding Eq (21) in the form of a power series and stuff. Fig. 8 shows that the final value of the bulk shrink-
substituting it in Eq (22), it is possible to write se as a age coefficient at X = 0 for the different foods is a
direct function of moisture content: function of X0. The representative correlation, with r2 =
0.991, is:
Sb=~+c$+(j$---)+F(~)2 (28) 0.966
Sb,f = (33)
where: X0 + 0.796
0.997 M + xsp Fig. 5 shows that the sb vs (X/X,) lines are ordered in a
B= (29) particular fashion. The ordering is such that smaller sb
Xo+xsg+MPsnp at X/X, = 0 corresponds to larger initial moisture content.
The separation between lines increases as X/X, decreases.
It follows that all sb points could be made to coincide in
C= (l+;+TM)- (30) one by defining a function of [I - (X/X,)]. A suitable
function to accomplish this end is:
X 0.018
S,,= 0.161 + 0.816- + 0.022 f?Xp ( -) + f (l-2 (34)
(31) x0 X + 0.025

where f is related to the bulk shrinkage coefficient at X =


F = 0.085 (32) X0, Eq (3 l), through the equation:
The error in sb with respect to Eq (27) is less than 1% f = 0.209 - s,,f (35)
for X = 0, which is the most unfavourable case. Fig. 9 is a plot of the experimental values adjusted as
Fig. 6 shows in more detail the behavior of Eq (22) suggested in Eq (34), sb - f [ 1 - (X/X,)], as a function of
at low moisture contents (0 < X/X, < 0.15). It can be seen XIX,. It also shows the lines corresponding to Eq (34).
that the proposed correlation does show the significant de- It can be seen that the fit of experimental values to the
crease in the rate of change of sb, while a linear extrapola- correlation is quite satisfactory, the coefficient of determi-
tion would fail to do so.

SWEET POTATO
A Experimental 0 Carroot
04 0 Pcor
0 mtoto
t A sweet Potato

2 dc a

fig. 7-Bulk shrinkage coefficient of sweet porato as a function of


moisture content. (at Eq (221; (bl Eq 17); (cl Eq 181; (d) Eq 191;
id Eq (10); tfj Eq (11). Fig. g--Bulk shrinkage coefficient of different foodstuffs at X = 0.

Volume 48 /1983)-JOURNAL OF FOOD SCIENCE- 1501


SHRINKAGE, POROSITY AND BULK DENSITY. ..

nation being r * = 0.97. Data for onions from other authors and 0.30-0.55 m/set. Thus, as long as it is conventional
(Mazza and Lemagner, 1980), plotted in the same figure, air drying and changes in drying conditions are not too
also fit the proposed correlation well. Thus, it can be stated drastic, the results are valid. As far as the authors know,
that Eq (33), (34) and (35) provide the means to predict there are no similar data available for other drying proce-
sb accurately in the entire range 0 < X < Xc for a variety dures. As to the influence of sample shape, this paper
of foods, requiring only a knowledge of the fresh food reports data corresponding to cylinders 1 cm in diameter,
moisture content. 4 cm long and, in the case of garlic, there are data in addi-
The generalization cannot be extended with equal suc- tion corresponding to slicing the original cylinder. Suzuki
cess to the modeling of porosity. It is possible to write et al, (1976) used 1 in. cubes; Mazza and Lemagner (1980)
Eq (12) in terms of the constituents as: dried onion slices. The implication is that the correlations
suggested are not sensitive to shape.
E=l--o --!I- (36)
(X0 + 1) PbP sb Pb Notation
a constant defined in Eq (8)
The initial bulk density can be approximated to an average AI area enveloping volume V, m2
Pb,o = 1.019 x lo3 kg/m3, which represents any of the b = constant defined in Eq (9)
foodstuffs tested with less than 3% error. A way to gen- B= constant defined in Eq (22)
eralize sb has been discussed above. But the difficulty ap- C constant defined in Eq (9)
pears when considering pp (X), which varies widely in c-1 constant defined in Eq (23)
value and behavior from one foodstuff to the other, as d = mass of nonsugar constituent, kg
shown in Fig. 2. An attempt to obtain a predictive equation D= constant defined in Eq (24)
on the basis of water content and composition, results in: f = adjusting parameter, defined in Eq (28)
F= constant defined in Eq (25)
CM +p+
xsg pX )Pb,o h = constant defined in eq (9)
sg S K= constant defined in Eq (10)
e=l- (37)
II = constant defined in Eq (9)
tx, + 1) sb m= mass of constituent, kg
As shown in Fig. 4, the results provide an approximation M= constant defined in Eq (20)
for garlic, sweet potato and potato; the prediction of values n = outward normal vector, m
n = constant defined in Eq (11)
for pear is poor and it fails to predict the porosity of car-
rots with changing X (Fig. 3). P = constant defined in Eq (9)
A valid question is how sensitive the data are to differ- 9 = constant defined in Eq (9)
ent drying conditions and sample shape. Data by Kilpatrick r = constant defined in Eq (11)
et al. (1955), Suzuki et al. (1976), and Mazza and Lemag- S = shrinkage coefficient, m3/m3
ner (1980) are quite close to the data reported here. All t = time, s
authors used conventional air drying. Kilpatrick et al. x = length, m
(1955) did not report sample shape or drying conditions, x= moisture content, kg/kg dry matter
although they referred to tunnel drying. Suzuki et al. v= volume, m3
(1976) used 40C dry bulb temperature and 30% relative w= interphase velocity, m/s
humidity and air at 0.6-0.7 m/set. Mazza and Lemagner W= wet sample weight, kg
(1980) used 40.5-60C, an unreported relative humidity Z = fictitious length, m

specific linear shrinkage coefficient


constant defined in Eq (18)
constant defined in Eq (19)
porosity, m3/m3
relationship defined by Eq (2 1)
density, kg/m3
relative humidity
constituent concentration, kg/kg dry matter
typical transport property
0.6 Subscripts
b = bulk
cw= cell wall material
d = dry matter
= equilibrium
0 carrot
. Pear
1 F =
0 =
final
initial
0 Potato p = particle
. sweet mtato sg = sugar
v cdic sn = sugar solution
* cnion(t.hzza,198ol sp = sweet potato
st = starch
0.1 I
OS a1 a2 03 0.4 05 0.6 Q7 QS 09 1. REFERENCES
Charm, S.E. 1978. The Fundamentals of Food Engineering,
3rd. ed. AVI Publishing Company, Inc.. Westport. CT.
Chirife, J. 1969. El Encogimiento y su Influenpia en la Interpreta-
Fig. g--Shrinkage coefficient generalized correlation. cidn de1 Mecanismo de Secado. Industria Y Quimica.
-Continued on page 1553
1502-JOURNAL OF FOOD SCIENCE-Volume 48 (1983)
change in texture scores occurred with time of holding and
the scores were higher (flesh was firmer), than those of the
corresponding samples from fish held in ice.
Cod held in MRSW readily absorbed salt (Table 1) and
l ICE at 9 days, 58% of the panelist thought the samples tasted
too salty. If the ratio of fish to brine were higher, the
n MRSW absorption of salt should not be as much a problem. Total
solids content decreased with time of holding in ice, but
/ increased for fish in MRSW due to absorption of salt (Table
1). The TMA content increased for fish held in either
media, and generally fish held in MRSW had a TMA content
slightly less than that of fish held in ice (Fig. 1). Flavor
scores and TMA values were correlated both for fish in ice
(r = -0.933, p < 0.05) and in MRSW (r = -0.969, p <
0.01). Use of TMA content for an index of spoilage for cod
held in MRSW may prove difficult because a sample (12
days) with an acceptable flavor score had a substantial TMA
content and an unpleasant odor in the raw state. Changesin
TVA content (Fig. 1) paralled those in TMA content and
were significantly correlated with flavor scores: (Ice, r =
. ICE -0.993, p < 0.001; MRSW, r = -0.946, p < 0.05).
n MRSW f REFERENCES
AOAC. 1975. Official Methods of Analysis. 12th ed. Association
of Official Analytical Chemist, Washington, DC.
Bullard. F.A. and Collins, J. 1979. An improved method to analyze
trimethylamine in fish and the interference of ammonia and
dimethylamine. Fish. Bull. U.S. 78: 465.
Collins, J., Reppond, K.D., and Bullard, F.A. 1980. Black rockfish.
Sebastes melanops: changes in physical, chemical, and sensory
properties when held in ice and in carbon dioxide modified
refrigerated seawater. Fish. Bull. U.S. 77: 865.
Friedemann, T.E. and Brook, T. 1938. The identification and
determination of volatile alcohols and acid. J. Biol. Chem. 123:
---.
161
Kramer, D.E., Nordin, D.M.A., and Gardner, L.J. 1977. A com-
parison of the quality changes of Alaska pollock and Pacific cod
IL I I I I during frozen storage at -28C. Fish. & Mar. Serv. Tech. Rpt.
0 ##753(Canada).
3 6 9 12 Lemon, D.W. and Regier, L.W. 1977. Holding of Atlantic mackeral.
Scromber Scrombus, in refrigerated seawater, J. Fish. Res. Board
TIME OF HOLDING, DAYS Can. 34: 439.
Reppond, K.D., Bullard, F.A. and Collins, J. 1979. Walleye pollock.
Fig. l-Regression of trimethylamine (TMA) and total volatile acid Theragra chalcogramma: physical. chemical and sensory changes
when held in ice and in CO2 modified refrigerated seawater. Fish.
(TVA) content on time of holding Pacific cod in the round in ice Bull. U.S. 77: 481.
and CO;, modified refrigerated seawater (MRSW). Tomlinson, N., Geiger, S.E., Gibbard. G.A., Smith, D.T., Southcott,
B.A., and Boyd, J.W. 1978. Improving landed quality of ground-
fish by modification of refrigerated seawater. Fish. and Mar. Serv.
Tech. Rpt. #783 (Canada).
M S received 4125183; accepted 5110183.
statistically (p < 0.05) different. Comments from the Referenceto trade names does not imply endorsementby the
judges indicated that low texture scores were due to soften- National Marine Fisheries Service, NOAA, to the exclusion of other
products that may be suitable.
ing of the flesh. For cod held in MRSW, no significant

SHRINKAGE, POROSITY & BULK DENSITY. . . From page 1502

Crapiste. G.H. and Rotstein, E. 1982. Prediction of sorptional equi- Marquardt. D.W. 1963. An algorithm for least squares estimation of
;l$~ data for starch containing foodstuffs. J. Food Sci. 47(5): nonlinear parameters. SIAM J. 11: 431.
Mazza, G. and Lemagner, M. 1980. Dehydration of onion: some
G&ling, P. 1958. Physical phenomena during the air drying of food- theoretical and practical considerations. J. Fd. Technol. 15: 181.
_..~~_._ -~~Fundamental
stuffs. In - ~~~~~~~~~~~~~
~~ ..~_
Asoects of the Dehvdration of Food- Roman. G.N.. Urbicain. M.J.. and Rotstein. E. 1982. kinetics of
stuffs. Soceity of Chemical Industry, London. The Macmillan the approach to sorptional equilibrium by.a foodstuff. AIChEJ.
Company, New York. (in press).
Honing, P. 1958. Principles of Sugar Technology, Vol. 1. Else- Suzuki, K.. Kubota. K., Tsutomu. H., and Hosaka, H. 1976. Shrink-
vler Pub. Co., Amsterdam. age in dehydration of root vegetables. J. Food Sci. 41: 1189.
Kilpatrlck, P.W., Lowe, E.. and Van Arsdel, W.B. 1955. Tunnel Watt, B.K. and Merrll. A.M. 1963. Composition of Foods. Agr.
dehydrators for fruits and vegetables. In Advances in Food Handbook No. 8. USDA-AR& Washington, DC.
Research, Vol. 6, p. 360. Academic Press, New York. Whitaker. S. 1980. Heat and mass transfer in granular porous media.
Kirk-Othmer Encyclopedia of Chemical Technology. 1964. In Advances in Drying, Vol. 1. Hemisphere-McGraw-Hill, New
2nd. ed. J. Wiley and Sons, Inc., New York/London/Sydney. York.
Lozano, J.E., Rotstein. E.. and Urbicain, M.J. 1980. Total porosity M S received 7129182; revised 12127182: accepted l/21/83.
and open pore porosity in the drying of fruits. J. Food Sci. 45:
1403.

Volume 48 (1983)-JOURNAL OF FOOD SCIENCE- 1553 I

You might also like