Climate, floods and river gods: environmental change and the

Meso–Neolithic transition in southeast Europe

Clive Bonsall
School of Arts, Culture & Environment, University of Edinburgh, Old High School,
Infirmary Street, Edinburgh EH1 1LT
C.Bonsall@ed.ac.uk

Mark G Macklin
Institute of Geography and Earth Sciences, University of Wales, Aberystwyth SY23 3DB
mvm@aber.ac.uk

Robert W Payton
School of Agriculture, Food and Rural Development, University of Newcastle upon Tyne,
Newcastle upon Tyne NE1 7RU
r.w.payton@ncl.ac.uk

Adina Boroneanþ
Institute of Archaeology “V Pârvan”, Str Henri Coandã 11, sect 1, Bucureºti, România

Keywords

Climate, rivers, floods, Mesolithic, Neolithic, Iron Gates, southeast Europe

Abstract

A conspicuous gap in the radiocarbon record of the Iron Gates Mesolithic suggests that many riverbank
sites were abandoned between c 8250 and 7900 cal BP.1 This period of site abandonment is linked to
increased flooding along the Danube, which can be correlated with a distinct global climatic oscillation. The
implications of these environmental changes for the interpretation of Lepenski Vir and the timing of the
Meso–Neolithic transition in the northern Balkans are examined. There is growing evidence of climatic
instability during the Holocene and its effects on river systems. We suggest that climate-related flooding
had a significant impact on human settlement and use of riverine environments in southeast Europe during
the middle Holocene, and may even have been an important stimulus of culture change.

1 An archaeological discontinuity in the Iron
Gates 8250–7900 cal BP
The Anglo-Romanian excavations at the Late
Mesolithic–Early Neolithic site of Schela Cladovei
between 1992 and 1996 (Boroneanþ et al 1999) have
generated the largest series of 14C dates for any
Stone Age site in the Iron Gates gorge section of
the Danube Valley (figure 1). The means of 44 AMS
dates on human bones and artefacts of terrestrial
mammal bone range from 8105 to 6695 BP, with a
distinct gap between 7460 and 7100 BP (8260–
7900 cal BP) (figure 2).
The site of Vlasac, which lies some 80 km
upriver from Schela Cladovei, has a large series of
radiometric and AMS 14C dates for archaeological

Before Farming 2002/3_4 (2) 1

Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ
Figure 1 Stone Age sites in the Iron Gates.
samples covering roughly the same time-span
(figure 2). The means of 18 age measurements
range from 7935 to 6790 BP and there is a gap
between 7440 and 7000 BP (8255–7840 cal BP),
which is broadly coincident in time with that
recorded at Schela Cladovei.2
Thus, there are no 14C dates from either site in
the age-range between c 8250–7900 cal BP.
Furthermore, four other sites from the Iron Gates
with multiple 14C age measurements — Padina (on
the right bank of the Danube in Serbia) and Icoana,
Ostrovul Banului and Ostrovul Corbului (on the
Romanian side) — also lack dates in that same
time-range (figure 3).3 These data suggest that
many Mesolithic riverside settlements in the Iron
Gates reach of the Danube were abandoned c 8250
cal BP, with some sites being re-occupied after c
7900 cal BP.
2 The 8200 cal BP global climatic oscillation
This period of site abandonment in the Iron Gates
corresponds in timing with a distinct global climatic
cooling phase, known as the ‘8200 BP event’ (eg,
2 Before Farming 2002/3_4 (2)
Alley et al 1997; Johnsen et al 2001). The Holocene
has been characterised by significant short-term
climatic variability. Oxygen isotope and
palaeotemperature records from Greenland ice cores
show a number of centennial-scale cooling events,
of which the ‘8200 BP event’ is arguably the most
pronounced. The δ18O profile from the GISP2 ice-
core shows this event to have had a duration of
around 330 years, from c 8290–7960 cal BP (Nesje
& Dahl 2001). The 8200 BP event appears to have
been accompanied by an abrupt change in
atmospheric and oceanic circulation patterns, which
at the regional level was reflected in changes in
weather and precipitation patterns (eg, Street-Perrott
& Perrott 1990; Stager & Mayewski 1997).
The 8200 BP event is registered in a variety of
climatic archives from Europe. For example, it has
been recognised in oxygen isotope records from the
Ammer Lake in southern Germany (von Grafenstein
et al 1998) and from a speleothem in south-west
Ireland (McDermott et al 2001). Lake sediment and
palaeobotanical records from the Alpine region
indicate cooler and wetter conditions at that time
(Haas et al 1998).
 Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ

Schela
Vlasac
Cladovei
Figure 2 Radiocarbon mean ages for Schela Cladovei and
6500
Vlasac — see figure 1 for site locations. Data are from Srejoviæ
6600 & Letica (1978), Bonsall et al (1997, 2000, and unpublished
6700 data) and Cook et al (2002). The 14C ages on human bone have
been corrected for the freshwater reservoir effect according to
6800 Method 1 of Cook et al (2002).
6900
7000

7100 7900 cal BP

Age (14C yr BP)

7200

7300

7400
8250 cal BP
7500

7600

7700

7800
7900

8000

8100

8200

Figure 3 Radiocarbon mean ages for Padina, Icoana,

Ostrovul Banului and Ostrovul Corbului — see Figure 1 for
site locations. Data are from Radovanoviæ (1996) and
Boroneanþ (2000).

Padina Icoana Ostrovul Ostrovul

Corbului Banului
6500

6600

6700

6800

6900
7000

7100 7900 cal BP

Age (14C yr BP)

7200

7300

7400
8250 cal BP
7500

7600

7700

7800
7900

8000

8100

8200

8300

Before Farming 2002/3_4 (2) 3

Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ
Importantly, in terms of human settlement of
river valleys, a number of river systems across
Europe experienced marked increases in the
frequency and magnitude of floods broadly
coincident with the 8200 BP event. There is
evidence for a major period of flooding on several
rivers in central Europe (Becker & Schirmer 1977;
Leuschner et al 2000), Britain (Macklin & Lewin, in
press) and the Danube delta where there was a
major channel avulsion (Ghenea & Mihaelescu
1991). However, the clearest record of severe
flooding at this time comes from the middle Durance
valley in the French Alps where a period of
substantial river aggradation is securely dated to
between 8250 and 7950 cal BP (Miramont et al
2000).
3 Floods and settlement relocation in the
Iron Gates
Typically, the Iron Gates sites are located close
to the bank of the Danube, and archaeological
remains survive between 5 and 13 metres above
the average summer water level (prior to dam
closures downstream in 1972 and 1995). It follows
that they would have been vulnerable to major
floods, which would have resulted in inundation of
the lower parts of the sites. Moreover, many Stone
Age settlements within the gorge (eg, Vlasac) were
situated at the foot of steep slopes, which river
erosion and undercutting could have destabilised,
leading to landslides creating an additional hazard
to human occupation. Therefore, it is likely that a
sustained period of more frequent and, probably,
larger floods between c 8250 and 7900 cal BP,
would have caused people to abandon their
existing settlements and move them to higher
locations. It is not suggested that Stone Age
people abandoned the Iron Gates reach of the
Danube, since there is clear evidence to the
contrary (Bonsall et al 1997, 2000).
A key question must be, is there any evidence
for major flooding around 8000 cal BP in the Iron
Gates, in the form of damage to archaeological
structures and/or accelerated sedimentation of the
valley floor? At present it is difficult to give a clear
4 Before Farming 2002/3_4 (2)
answer to this question, for a number of reasons.
Evidence of major flooding may have been
overlooked in earlier accounts of the Iron Gates
sites (see below), and virtually all the known sites
are currently inaccessible owing to artificial flooding
following construction of the Iron Gates dams (figure
1). At least one site, Schela Cladovei, is still
accessible. However, in floodplain sites of this age
stratigraphic evidence of flooding often would not
persist in the sedimentological record because of
incorporation of any flood deposits into surface soil
horizons by human activity and biotic processes
of pedoturbation. Only in cases of deep burial by
rapidly deposited river alluvium is the evidence of
bedded flood deposits likely to survive in the
subsoil. Where this occurs such deposits should
theoretically overlie a buried A horizon marking the
former topsoil, but this may have been eroded during
the flood event. Generally, the thickness of flood
sediments is unlikely to be very great in the more
elevated parts of the floodplain away from the river
channel, but could locally be thicker in depressions.
At Schela Cladovei direct dating of the Holocene
alluvial sediments has been attempted (Fuller et al
1994). An infra-red stimulated luminescence (IRSL)
age of 8.2±1.3 ka was obtained (using a ‘partial
bleach’ approach) for a sample collected from above
a Mesolithic skeleton. Although this is a single age
determination with a large error margin, it is at least
consistent with a major phase of floodplain
sedimentation between 8250 and 7900 cal BP.
4 Lepenski Vir — settlement or sacred place?
Another site that may show evidence of the effects of
flooding during the period in question is Lepenski Vir,
but the interpretation of this site is not straightforward.
Situated 2.5 km upriver from Vlasac on the right
bank of the Danube in Serbia (Figure 1), Lepenski
Vir has a number of features that set it apart from
other Late Mesolithic/Early Neolithic sites in the
Iron Gates. They include (i) buildings with mortared
floors and elaborate stone fittings, (ii) burials that
(apparently) were deliberately emplaced within or
beneath some of the buildings, (iii) numerous
carved sandstone boulders (‘sculptures’ and
 Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ

‘altars’), (iv) an unusually high frequency of other
decorated (‘symbolic’) items, and (v) a location
seemingly chosen to face an unusual and imposing
landscape feature — the trapezoidal-shaped
‘mountain’ of Treskavac (figure 4).
‘spiritual home’, perhaps — then transferring it to
another location would not have been an option,
and they are likely to have resorted to other
measures to combat the threat from the river.
Several previous authors have suggested that the
carefully-prepared mortared floors and substantial
stone fixtures of the buildings of ‘Lepenski Vir I’
were designed to resist flooding, although these
authors have tended to view Lepenski Vir and the
other Stone Age sites of the Iron Gates as seasonal
encampments subject to annual floods throughout
their existence (Gimbutas 1991:285; Chapman
2000:195). The present authors would accept the
first part of this scenario, but not the second.

Lepenski
Vir
6500

6600

6700

6800

6900
7000

7100 7900 cal BP

7200
age BP

7300

7400
14C

Figure 4 The distinctive trapezoidal mountain of Treskavac on the 8250 cal BP

Romanian bank of the Danube, opposite Lepenski Vir (photo: Clive Bonsall). 7500

7600
It is possible that Lepenski Vir originated as an
7700
ordinary settlement like Schela Cladovei or Vlasac,
7800
in the period before 8200 cal BP. But many authors
have concluded that during the final Mesolithic and 7900
Early Neolithic it was treated as a ‘sacred place’, 8000
used primarily as a centre for burial and ritual. The
8100
trapezoidal-plan buildings have been variously
8200
interpreted as ‘sanctuaries’, ‘shrines’ or ‘temples’
(Srejoviæ 1972; Gimbutas 1991; Chapman 2000). 8300
Interestingly, it is the only site in the Iron Gates
which has 14C dates whose mean ages fall within Figure 5 Radiocarbon mean ages for Lepenski Vir. Data are from

the period 8250–7900 cal BP (figure 5). Quitta (1972) and Bonsall et al (1997, 2000). The 14C ages on human

If the Stone Age people of the Iron Gates bone have been corrected for the freshwater reservoir effect according

regarded Lepenski Vir as a sacred site — their to Method 1 of Cook et al (2002).

Before Farming 2002/3_4 (2) 5

 II XXXVI XXXII 6560±100 II IX 6630±100 XLIV XVII XXXIII XIV XIX XXV

6
?

7210±200 27 32 6815±70 23

33 18
I.3 ? 35 57 24
20 31
phase uncertain
66
15
43 ?

59 60 44 64 49 25 26 14 5
P
6620±100 51 12 19
56

Before Farming 2002/3_4 (2)

46 26' 13 6
65 61 48 21
45 4 6970±60 47 52 34 11 16 6820±100 28
I.2 55
22

Architectural phase
42 36 7335±70 54 7040±100
10
Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ

7430±160 62 40 39 6860±100 1 6870±65 37 3 63 50 6845±100 9 29

I.1
2 38 47' 7 30

41 58 53 8 17

54 'LVI' building with pottery

P Pottery stratified between superimposed buildings

Building with one or more stone sculptures

Building with one or more stone 'altars'

Figure 6 Stratigraphic relationships of the Lepenski Vir ‘shrines’, organised according to Ivana Radovanoviæ’s (1996) architectural phases. Associated finds of Starèevo pottery and carved boulders (‘altars’ and sculptures) are

shown. Designed primarily as a test of Radovanoviæ’s relative chronology of the site, the chart highlights the fact that sculptures are found in buildings that have radiocarbon ages as old as 7335±71 BP (c 8150 cal BP).

Furthermore, three superimposed buildings with associated sculptures and ‘altars’ (21, 22, 29 — bottom right of the chart) are stratigraphically older than a human burial, which on evidence of dietary tracing using stable isotopes

probably pre-dates 7900 cal BP and the appearance of pottery at Lepenski Vir — for details, see Bonsall et al (2000:128–130). Notes: 1 Information on superpositioning of buildings is from Srejoviæ (1972) and Radovanoviæ (1996).

2 The occurrence of stone sculptures and ‘altars’ is based on the catalogue by Srejoviæ & Baboviæ (1983), which has several discrepancies with Srejoviæ’s (1972) list. 3 Only a selection of the ‘LVII’ buildings identified by Srejoviæ

have been included in the chart; stratigraphic relationships with ‘LVI’ buildings are poorly documented, and not entirely convincing. 4 14C dates in italics are weighted means of two or more age measurements on the same sample.

There has been much speculation about the
significance of the stone sculptures. It seems
generally agreed that they were not just works of
art, but intimately associated with the religious
beliefs of the community. Some authors have
argued that the figural sculptures — with carved
faces that are often described as ‘fish-like’ or ‘half-
fish, half-human’ — depict ancestors or ‘river gods’
(eg, Srejoviæ 1972, 1989; Srejoviæ & Baboviæ 1983;
Handsman 1991; Gimbutas 1991). Radovanoviæ and
Voytek (1997:28–29) interpreted their appearance
and proliferation at Lepenski Vir as part of a process
of intensive ‘ideological integration’ among a
Mesolithic forager population, which was under
increasing territorial and economic pressure from
farming societies that were becoming established
in the surrounding areas. The weakness of this
interpretation is that sculptures occur in some of
the earliest-dated houses of Lepenski Vir (figure
6), which probably pre-date the earliest known
Neolithic sites on the Hungarian or Wallachian
Plains by several centuries.
The carved boulders appear to have been
intimately associated with the architecture of
Lepenski Vir, and were often found inserted into
sockets in the mortared floors. In some houses
that lacked carved boulders, holes in the mortared
floors have been interpreted as the sockets for
sculptures that were subsequently removed
(Srejoviæ 1972:81). Significantly, neither the
sculptures nor the mortar-floored buildings can be
traced back before c 8200 cal BP, and it is possible
that both phenomena appeared around that time.
This raises the intriguing possibility that the
ideological integration referred to by Radovanoviæ
and Voytek was not a response to the threat from
neighbouring farmers, but to increased flooding
along the Danube. For the inhabitants of the Iron
Gates between 8250–7900 cal BP the Danube was
still the principal source of food (Bonsall et al 1997,
2000) and the main artery of communication and
social interaction; but it had become more prone
to extreme and unpredictable floods. The cause of
such dramatic events would likely have seemed
mysterious. The source of major flooding in the
Iron Gates reach would have been geographically
Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ
remote, generated by excessive rainfall and/or
snowmelt many hundreds of kilometres upriver in
the Alps, Carpathians and other mountainous areas
that form the headwater catchments of the Danube.
Therefore, an equally, if not more valid interpretation
of the Lepenski Vir sculptures is that they were
apotropaic, intended to appease or achieve
dominance over the unseen powers responsible for
unpredictable and potentially devastating floods,
which became much more frequent (and possibly
of greater magnitude and duration) between 8250
and 7900 cal BP.4
5 Floods and ‘Proto-Lepenski Vir’
Srejoviæ believed that the first settlement at
Lepenski Vir had been confined to a narrow strip of
land along the riverbank. This ‘Proto-Lepenski Vir
settlement’ comprised the remains of some eight
structures, represented essentially only by
remnants of stone-bordered hearths; no other traces
of the floors or walls of the original structures
survived. These were located in a zone about 12
metres wide that was apparently separated from
the Danube by a narrow floodplain some 15 metres
wide (Srejoviæ 1972:45–46).
There is no radiocarbon or (convincing)
stratigraphic evidence to support Srejoviæ’s
assertion that the Proto-Lepenski Vir structures
are older than the mortar-floored buildings of his
main Lepenski Vir I phase. Located in the lowest
part of the site, they would have been at most risk
from higher-than-normal floods, and the simplest
explanation is that they are the remains of mortar-
floored buildings, contemporaneous with those of
Lepenski Vir I higher up the slope, which were
severely damaged by extreme flood events between
8250 and 7900 cal BP.
6 Wider implications
6.1 Archaeological potential of the Iron Gates
The construction of two dams across the Danube
— Iron Gates 1 between 1964 and 1972 and Iron
Gates 2 between 1977 and 1995 — led to artificial
Before Farming 2002/3_4 (2) 7
Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ

Schela Lepenski
Vlasac Padina
Cladovei Vir

6300 7250 cal BP

6400

6500

6600 7500 cal BP

6700

6800

6900
7000

7100 7900 cal BP

Age (14C yr BP)

7200

7300

7400
8250 cal BP
7500

7600

7700

7800
7900

8000

8100

8200

8300

Figure 7 Radiocarbon mean ages for Lepenski Vir, Padina, Schela Cladovei and Vlasac plotted against the 7500–7250 cal BP flood episode.

flooding of the majority of the known Mesolithic
and Early Neolithic sites. As a result,
archaeologists have tended to assume that all of
the archaeologically interesting areas are now
submerged, and no further excavation is possible.
However, if during the period c 8250–7900 cal BP
settlements were transferred to higher positions
relative to the river, then some of those sites can
be expected to have survived. It is worth
emphasising that riverbank erosion has been the
principal mechanism leading to exposure and
subsequent discovery of Stone Age settlements in
the Iron Gates. This, and the fact that survey work
was focused on valley floor areas that were to be
flooded following dam closure, has determined the
present archaeological distribution pattern.
Hitherto, except in caves (Boroneanþ 2000), very
little archaeological prospection has been
undertaken in more elevated areas where there are
fewer natural exposures.
If the above interpretation is correct, it follows
that there is significant scope for further research
in the Iron Gates, focusing on a crucial period in
the prehistory of the region — that during or
immediately prior to which pottery and farming
made their appearance in the region.
6.2 Beyond the Iron Gates: the spread of farming
into the Danube Basin
The period from 8250 to 7900 cal BP was when
Neolithic farming spread through much of southeast
Europe. During that time the agricultural frontier
moved from northern Greece to the Hungarian Plain.
River valleys and their generally fertile, stoneless
soils developed from river alluvium appear to have
been the primary conduits for agricultural expansion,
the first Starèevo–Körös settlements being
established along the Danube and its major
tributaries in northern Serbia and southeast Hungary
c 7900 cal BP. In other words, it took 300–400 years
for the agricultural expansion to cover a straight-line
distance of c 600 km. Yet some later agricultural
expansions over considerably greater distances

8 Before Farming 2002/3_4 (2)


apparently were accomplished more rapidly. For
example, the Linearbandkeramik (LBK) expansion
from the Hungarian Plain to the Netherlands and
the spread of farming through the British Isles, both
extending over c 1000 km, occurred in less than
two centuries.
In southeast Europe, it can be shown that
farming reached the middle Morava Valley (150 km
from the Danube) by c 8100 cal BP (Whittle et al,
in press). However, it appears to have been another
150–200 years before farming villages appeared on
the Hungarian Plain. Did increased flooding between
c 8250 and 7900 cal BP delay the spread of
agriculture onto the valley floors of the Danube and
northern tributaries such as the Tisza, Maros and
Körös? Or were the first farming settlements
established in the Danube Basin before 7900 cal
BP, but subsequently buried by accelerated river
alluviation during the ‘8200 BP event’?
We incline toward the former interpretation.
Starèevo–Körös sites typically are located on
upstanding alluvial landforms such as levees,
islands and low terraces that would have been
extremely vulnerable to large floods (Gimbutas
1991; Manson 1995; Whittle 1996). Even more
vulnerable to flooding would have been the
associated fields, crops and livestock, probably
often located in lower-lying parts of the valley floors.
6.3 Later flooding episodes and Neolithic
settlement of the Hungarian Plain
The major period of flooding between c 8250 and
7900 cal BP was not the only such episode
registered in the Danube catchment during the
early–middle Holocene. A significant increase in
flood frequency and magnitude has been
documented in many European river systems
between c 7500 and 7250 cal BP (Becker &
Schirmer 1977; Leuschner et al 2000; Macklin &
Lewin, in press).
Judging by the distribution of 14C dates from
Lepenski Vir, Padina, Schela Cladovei and Vlasac,
intensive use of the principal riverside settlements
of the Iron Gates came to an end during this period
(figure 7). Only sporadic use of the sites is recorded
Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ
thereafter, until the Eneolithic and in many cases
the pre-Roman Iron Age. This raises the possibility
that a further period of major flooding forced the
long-term abandonment of these riverside locations.
Since by that time agriculture had supplanted
aquatic resources as the basis of the Stone Age
economy, there was perhaps little incentive for
people to return to their former settlements.
6.4 Starèevo, Vinèa and LBK
Not long after the Iron Gates sites were abandoned,
many Starèevo–Körös settlements in the Middle
Danube Basin ceased to be occupied. A recent
study by Whittle et al (in press), which produced
nearly seventy AMS 14C dates for early Neolithic
sites on the Hungarian Plain, shows that few, if
any, Starèevo culture occupations are later than
7400 cal BP. Coincidentally, this is also the
approximate date of the emergence of the LBK
culture on loessial soils of the higher areas of the
Hungarian Plain.
Are these cultural events linked, and are they
related to the period of increased flooding recorded
across Europe at c 7500–7250 cal BP? Did more
frequent and/or longer-lasting overbank flooding
force early farmers to move away from low-lying
alluvial areas, and explore the agricultural potential
of other environments? And was this the catalyst
for the LBK expansion across the loess belt of
central Europe?
As noted above, early agricultural expansion
through the Balkan Peninsula and onto the
Hungarian Plain tended to follow river valleys.
Explanations for this have generally stressed soil
fertility and better availability of water (Sherratt 1980;
Roberts 1998). Whilst it is true that young, often
naturally fertile, well-drained, stoneless soils occur
on river alluvium, it must be recognised that soils
of alluvial valleys form a mosaic, with substantial
areas of more sandy and more clayey, often poorly-
drained soils. This mosaic was of variable suitability
for agriculture but also provided a range of
exploitable natural resources and habitats, including
wetlands. Conversely, in much of the Balkan
Peninsula shallow calcareous soils (Rendzinas)
Before Farming 2002/3_4 (2) 9
 Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ

IRON HUNGARIAN PLAIN FLOOD HUMAN RESPONSE TO CLIMATE-RELATED CHANGES

GATES FLOODPLAINS LOESS BELT HISTORY IN FLOOD FREQUENCY
7000

Vinca
Quiescent phase. Reoccupation of floodplain areas by the Vinca and Tisza cultures.
v

Tisza
v
Consolidation of settlement in the loess zone. No significant reoccupation of former riverbank

LBK
7200 settlements in the Iron Gates.

? Active phase. Increased flooding leads to abandonment of riverbank settlements in the Iron Gates
and of valley floor farming settlements on the Hungarian Plain, prompting exploration of the agricultural
7400
potential of other areas — reflected, e.g., in the initial LBK expansion across the loess belt.
Starcevo-Körös
Neolithic

7600 Quiescent phase. Reduced flood frequency allows farming settlements to be established (belatedly)
Early

on alluvial soils of the Hungarian Plain. Agriculture and pottery also appear in the Iron Gates.
Lepenski Vir

Many former riverbank settlements are reoccupied, but importance of aquatic resources declines as
v
Age (cal. yr BP)

7800 farming comes to dominate the Iron Gates economy.

Active phase. Increased flooding leads to temporary abandonment of many riverbank settlements in
8000 the Iron Gates and (probably) establishment of new sites on higher ground. Lepenski Vir assumes
? the role of 'sacred place', used as a centre for burial and ritual. Accelerated sedimentation of floodplains
in the Middle Danube basin may have resulted in burial/destruction of the evidence of earlier human
8200 occupation. Possibility of 'false starts' to the Neolithic on the Hungarian Plain.

8400
Late Mesolithic

8600 Quiescent phase. Intensive exploitation of aquatic resources in the Iron Gates reach of the Danube
from small 'village' settlements on both banks of the river.

8800

9000
Absence or scarcity of settlement evidence Major period of flooding

Figure 8 Cultural chronology of the Middle Danube basin from 9000 to 7000 cal BP and human response to changes in flood frequency.

that dominate the limestone uplands would have favour or constrain agricultural land use. The
been much less attractive to early farmers; likewise predominantly northwestward (LBK) expansion of
the drought-prone soils of the grassland areas of agriculture in the period beginning c 7400 cal BP
the Hungarian Plain (Chernozems) would have been started along the axis of the deciduous forest–sub-
less suitable for early Neolithic agriculture. humid grassland transition within the loess belt of
The mosaic of soils on river levees and the Hungarian Plain. We would argue that the
backswamps would have been variably affected by attraction of these areas can be attributed to a
increased flooding. Low-lying parts of the floodplain favourable combination of climate, soil conditions
with clayey soils would have experienced deeper, and the presence of open areas initially precluding
more extensive and more prolonged inundation the need for woodland clearance. Under the climatic
leading to a periodic expansion of wetlands. Well- conditions and natural vegetation communities of
drained loamy or silty soils on the higher lying areas woodland and grassland that prevailed during the
closer to the river channels, which were favoured middle Holocene, Luvisols (Brown soils) and
for settlement and agriculture, would have been Chernozems form fertile soil resources with stable
subject to more frequent bank erosion and structure, good organic matter levels and good
deposition of coarse-textured alluvial sediments that water storage capacity.5 Rainfall in this zone is non-
locally would have buried cropland or grazing areas. limiting for cereals and the land resource base
The aeolian loess belt that stretches from the would have provided an attractive ‘niche’ for
steppes of Russia to the Paris Basin supports a agricultural expansion at a time when the river
variety of soils (FAO 1974–81; CEC 1985; Bronger valleys of the Hungarian Plain had become less
1991). Although these are all stoneless and silty suitable for farming because of increased flooding.
in texture, they display a range of contrasting Because of their stoneless character, texture and
chemical and hydrological properties that may fertility, making them easy to till, the soils

10 Before Farming 2002/3_4 (2)

 Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ

encountered would have had many properties in
common with some of the soil resources previously
exploited in alluvial valleys. The more acid loessial
Brown soils (Luvisols) of the deciduous woodland
zone further to the northwest also provided valuable
agricultural land, assuming adequate manuring.
Following the 7500–7250 cal BP flood phase,
the river valleys of southeast Europe would rapidly
have returned to their former state, allowing farming
settlements to be re-established on soils now often
enriched by recent additions of alluvium. This may
have been a critical factor in the appearance and
proliferation of Vinèa and Tisza culture settlements
in river valleys of the central and northern Balkans
from c 7250 cal BP (Gimbutas 1991).6
7 Conclusions
Until just over a decade ago, it was assumed by many
archaeologists that the Holocene in Europe was
characterised by relatively limited climatic variability.
Recent studies of ice cores, lake and cave sediments
have overturned this traditional view. At the same time
geomorphological studies of river systems have
shown them to be particularly sensitive to small-scale
changes in climate in both the prehistoric and historic
periods that are not registered in the palynological
record of vegetation change. This has major
implications for the understanding of the European
archaeological record, since some of the largest
concentrations of ancient human populations and
settlements occurred in river valleys.
What we have shown in this paper, with examples
drawn from a single major river basin in central and
southeastern Europe, is how climate-related flooding
impacted on human settlement and use of riverine
environments, and may even have been an important
stimulus of culture change (figure 8).
We believe that this poses new challenges, and
suggests a possible future direction, for archaeological
research in which greater emphasis is placed upon
the dynamic relationship between human behaviour
and river environments. There is an urgent need for
more research into the long-term flood histories of
European rivers to achieve better chronological and
spatial control of individual flood episodes. This should
be conducted in parallel with systematic
archaeological and pedological investigations of river
environments where cultural remains are stratified
within fluvial sediments and afford opportunities for
high resolution dating.
Acknowledgements
We thank László Bartosiewicz, Gordon Cook, Dimitri
Tsintjilonis, Barbara Voytek and Ann Wintle for their
comments on the first draft of the paper, and Alasdair
Whittle for sight of a prepublication copy of a
forthcoming article. The 1992–1996 excavations at
Schela Cladovei were funded by the British Academy,
the Carnegie Trust for the Universities of Scotland,
and the Munro Lectureship Fund and the Hayter Fund
of The University of Edinburgh. The radiocarbon dates
were provided by the Oxford Radiocarbon Accelerator
Unit and paid for by the UK Natural Environment
Research Council.

1
The chronological evidence discussed in this paper is based on a combination of archaeological and non-archaeological 14C, dendrochronological,
luminescence and ice-core ages. For the sake of consistency, we quote all ages in calibrated years BP (before present).
2
Two radiometric 14C dates on charcoal from Vlasac (Z-268 and Z-264 – Srejoviæ & Letica 1978:129) have been omitted from Figures 2 and 7, for taphonomic
reasons. These samples appear to be from grave fills. If so, the charcoal is likely to be redeposited and could comprise material of differing ages.
3
A number of radiocarbon dates from Padina on bones of humans and animals that may have ingested significant amounts of freshwater fish have
¶
been omitted from figures 3 and 7. They comprise five radiometric 14C ages on human bones (Burleigh & Zivanoviæ 1980) and three AMS 14C ages
on dog bones (Whittle et al, in press). For the humans there are no associated δ15N measurements, which would allow a correction for the freshwater
reservoir effect to be applied, and for the dogs there is insufficient information on the dietary end members to perform a reservoir correction.
4
The possibility that some of the sculptures were ‘apotropaic’ is discussed by Srejoviæ (1972:111), and the idea is implicit in the writings of some later
authors. However, none of these authors has linked the sculptures to increased flooding along the Danube, as suggested in this paper.
5
Evidence from buried soils developed in loess from various parts of Europe, models of soil evolution and established rates of soil-forming
processes indicates that the major soil categories of the loessial Hungarian Plain (Chernozems, Luvisols, etc) had sufficient time to develop by the
end of the Mesolithic and would have had broadly similar properties to soils under sub-humid grasslands and deciduous forest today. However, the
forested Luvisols would have been less acid than modern equivalents that have remained under forest, as decalcification and leaching would have

Before Farming 2002/3_4 (2) 11

Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ

been less advanced. Loess and silty alluvium are normally well supplied with easily weatherable silicate minerals and provide initially base-rich
parent materials that are stoneless and have good water-holding capacity. Hence they would have formed relatively fertile and desirable soils for
agricultural use under the climatic conditions and vegetation communities that prevailed in central Europe during the middle Holocene. Such parent
materials continue to provide land with high agricultural capability in those parts of Europe today. Under early agriculture natural soil fertility would
have declined more rapidly in the formerly forested Luvisols than in the Chernozems, once the forest nutrient cycle had been broken.
6
Careful scrutiny of radiocarbon dates from sites on the Hungarian Plain suggests there is a hiatus in the settlement record at the transition from
Early to Middle Neolithic. Reliable 14C ages for Starèevo occupations are invariably older than c 6450 BP (7400 cal BP) while 14C ages for Vinèa or
Tisza culture occupations are invariably younger than c 6300 BP (7250 cal BP).

12 Before Farming 2002/3_4 (2)

Climate, floods and river gods: environmental change and the
Meso–Neolithic transition in southeast Europe

Bonsall, Macklin, Payton, Boroneanþ

References

Alley, RB, Mayewski, PA, Sowers, T, Stuiver, M, Taylor, KC & Clark, PU 1997. Holocene climate instability:
a prominent widespread event 8200 yr ago. Geology 25:483–486.

Becker, B & Schirmer, W 1977. Palaeoecological study on the Holocene valley development of the River
Main, southern Germany. Boreas 6:303–321.

Bonsall, C, Lennon, R, McSweeney, K, Stewart, C, Harkness, D, Boroneanþ, V, Payton, R, Bartosiewicz,

L & Chapman JC 1997. Mesolithic and Early Neolithic in the Iron Gates: a palaeodietary perspective.
Journal of European Archaeology 5(1):50–92.

Bonsall, C, Cook, G, Lennon, R, Harkness, D, Scott, M, Bartosiewicz, L & McSweeney, K 2000. Stable
Isotopes, radiocarbon and the Mesolithic-Neolithic transition in the Iron Gates. Documenta Praehistorica
27:119–132.

Boroneanþ, V 2000. Paléolithique Supérieure et Épipaléolithique dans la Zone des Portes de Fer. Bucureºti:
Silex.

Boroneanþ, V, Bonsall, C, McSweeney, K, Payton, R & Macklin, M 1999. A Mesolithic burial area at Schela
Cladovei, Romania. In Thévenin, A (ed) L’Europe des Derniers Chasseurs: Épipaléolithique et Mésolithique.
(Actes du 5e colloque international UISPP, commission XII, Grenoble, 18–23 septembre 1995). Paris: Éditions
du Comité des Travaux Historiques et Scientifiques:385–390.

Bronger, A 1991. Argillic horizons in modern loess soils in a Ustic soil moisture regime: comparative studies in
forest-steppe and steppe areas from eastern Europe and the United States. Advances in Soil Science 15:41–90.

¶
Burleigh, R & Zivanoviæ, S 1980. Radiocarbon dating of a Cro-Magnon population from Padina, Yugoslavia,
with some general recommendations for dating human skeletons. Zeitschrift für Morphologie und
Anthropologie 70:269–274.

CEC 1985. Soil map of the European Communities 1:1,000,000 (9 sheets and explanatory text). Luxembourg:
Commission of the European Communities, Directorate-General for Agriculture.

Chapman, J 2000. Fragmentation in Archaeology: Peoples, Places and Broken Objects in the Prehistory of
South Eastern Europe. London: Routledge.

Before Farming 2002/3_4 (2) 13

Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ

Cook, GT, Bonsall, C, Hedges, REM, McSweeney, K, Boroneanþ, V, Bartosiewicz, L & Pettitt, PB 2002.
Problems of dating human bones from the Iron Gates. Antiquity 76:77–85.

FAO 1974–81. FAO–UNESCO Soil Map of the World 1:5,000,000. Volumes I–X. Paris: UNESCO.

Fuller, IC, Wintle, AG & Duller, GAT 1994. Test of the partial bleach methodology as applied to the infra-red
stimulated luminescence of an alluvial sediment from the Danube. Quaternary Science Reviews (Quaternary
Geochronology) 13:539–543.

Ghenea, C & Mihaelescu, N 1991. Palaeogeography of the lower Danube valley and the Danube delta
during the last 15 000 years. In Starkel, L, Gregory, KJ & Thornes, JB (eds) Temperate Palaeohydrology.
New York: Wiley:343–364.

Gimbutas, M 1991. The Civilization of the Goddess: The World of Old Europe. San Francisco: Harper.

Haas, JN, Richoz, I, Tinner, W & Wick, L 1998. Synchronous Holocene climatic oscillations recorded on
the Swiss Plateau and at timberline in the Alps. The Holocene 8(3):301–309.

Handsman, RG 1991. Whose art was found at Lepenski Vir? Gender relations and power in archaeology. In
Gero, J & Conkey, M (eds) Engendering Archaeology. Oxford: Blackwell:329–365.

Johnsen, SJ, Dahl-Jensen, D, Gundestrup, N, Steffensen, JP, Clausen, HB, Miller, H, Masson-Delmotte, V,
Sveinbjörnsdottir, AE & White, J 2001. Oxygen isotope and palaeotemperature records from six Greenland
ice-core stations: Camp century, Dye-3, GRIP, GISP 2, Renland and NorthGRIP. Journal of Quaternary
Science 16(4):299–307.

Leuschner, HH, Spurk, M, Baillie, M & Jansma, E 2000. Stand dynamics of prehistoric oak forests derived
from dendrochronologically dated subfossil trunks from bogs and riverine sediments in Europe. GeoLines
11:118–121.

Macklin, MG & Lewin, J In press. River sediments, great floods and centennial-scale Holocene climate
change. Journal of Quaternary Science.

Manson, JL 1995. Starèevo pottery and Neolithic development in the central Balkans. In Barnett, WK &
Hoopes, JW (eds) The Emergence of Pottery: Technology and Innovation in Ancient Societies. Washington:
Smithsonian Institution Press:65–77.

McDermott, F, Mattey, DP & Hawkesworth, C 2001. Centennial-scale Holocene climate variability revealed
by a high-resolution speleothem δ18O record from SW Ireland. Science 294:1328–1331.

Miramont, C, Sivan, O, Rosique, T, Edouard, JL & Jorda, M 2000. Subfossil tree deposits in the middle Durance
(southern Alps, France): environmental changes from Allerød to Atlantic. Radiocarbon 42(3):423–435.

Nesje A & Dahl SO 2001. The Greenland 8200 cal. yr BP event detected in loss-on-ignition profiles in
Norwegian lacustrine sediment sequences. Journal of Quaternary Science 16(2):155–166.

14 Before Farming 2002/3_4 (2)

 Climate, floods and river gods: Bonsall, Macklin, Payton, Boroneanþ

Quitta H 1972. The dating of radiocarbon samples. In Srejoviæ, D (ed.) Europe’s First Monumental Sculpture.
New Discoveries at Lepenski Vir. London: Thames & Hudson:205–210.

Radovanoviæ I. 1996. The Iron Gates Mesolithic. Ann Arbor (Michigan): International Monographs in Prehistory,
Archaeological Series 11.

Radovanoviæ I & Voytek BV 1997. Hunters, fishers or farmers: sedentism, subsistence and social complexity
in the Djerdap Mesolithic. Analecta Praehistorica Leidensia 29:19–31.

Roberts, N 1998. The Holocene, 2nd Edition. Oxford: Blackwell.

Sherratt, A 1980. Water, soil and seasonality in early cereal cultivation. World Archaeology 11: 313–330.

Srejoviæ, D 1972. Europe’s First Monumental Sculpture. New Discoveries at Lepenski Vir. London: Thames
& Hudson.

Srejoviæ, D 1989. The Mesolithic of Serbia and Montenegro. In Bonsall, C (ed) The Mesolithic in Europe.
Edinburgh: John Donald:481–491.

Srejoviæ, D & Baboviæ, L 1983. Umetnost Lepenskog Vira. Belgrade: Jugoslavija.

Srejoviæ, D & Letica, Z 1978. Vlasac. A Mesolithic Settlement in the Iron Gates. Beograd: Serbian Academy
of Sciences and Arts Monograph DXII.

Stager, JC & Mayewski, PA 1997. Abrupt early to mid-Holocene climatic transition registered at the equator
and poles. Science 276:1834–1836.

Street-Perrott, A & Perrott, RA 1990. Abrupt climate fluctuations in the tropics: the influence of Atlantic
Ocean circulation. Nature 343:607–612.

von Grafenstein, U, Erlenkeuser, H, Müller, J, Jouzel, J & Johnsen, S 1998. The cold event 8200 years ago
documented in oxygen isotope records of precipitation in Europe and Greenland. Climate Dynamics 14:73–81.

Whittle, A 1996. Europe in the Neolithic: The Creation of New Worlds. Cambridge: Cambridge University
Press.

Whittle, A, Bartosiewicz, L, Boriæ, D, Pettitt, P & Richards, M In press. In the beginning: new radiocarbon
dates for the Early Neolithic in northern Serbia and south-east Hungary Antaeus 25.

Before Farming 2002/3_4 (2) 15