. 17
( 43 .)


that storage is compatible with mobile societies in the form of strategically
placed caches in a seasonally revisited landscape (Stopp, 2002), a very likely
tactic of the Moderns in the cold environments of the Eurasian Plains. Storage
among hunter“gatherers buffers predictable seasonal variation and is therefore
most common in highly seasonal environments. With an increase in storage
there is a decrease in the number of residential moves (Kelly, 1983). Humans
today are chararcterised by having among the highest levels of adipose tissue
of all mammals (including Arctic species) and this may be a relic of having
evolved in highly seasonal and unpredictable environments (Pond, 1978, 1999).
Storing fat in adipose tissue permits humans to build up a considerable energy
reserve and fat soluble vitamins (Bunn & Ezzo, 1993). These authors conclude
that Plio-Pleistocene (and later) hominids who faced continuous uncertainty in
their food supply had the problem of balancing essential nutrients and that this
might only have been met by the presence of stored nutrients and energy in fat
96 Neanderthals and Modern Humans

that could make up for dietary imbalances. Individuals carrying fat do not have
to draw on muscle tissue to meet their energy needs so that individuals with
an ability to store fat would have had a ¬tness advantage. One way in which
fat reserves could have been accumulated was through gorging on meat, as
occurs in contemporary San bushmen and Hadza (Bunn et al., 1988; Hitchcock,
Humans are neither strict hunters nor scavengers. It is clear today that Ne-
anderthals and Moderns were opportunistic and hunted, foraged or scavenged
depending on circumstances (Marean & Kim, 1998). Research papers using
stable isotope (Bocherens et al., 1991, 1999; Fizet et al., 1995; Richards et al.,
2001) and buccal microwear analyses (Lalueza et al., 1996) have led the au-
thors to conclude that Neanderthals consumed mammalian herbivore meat. The
stable isotope data come from ¬ve specimens in three central European sites in
Belgium, France and Croatia, all on or at the edge of the Eurasian Plain. They
span a huge period of time. The buccal microwear data come from a few more
sites, mainly in French but also from the Levant and the Gibraltar Devil™s Tower
child. Nevertheless, though suggestive, the data are too limited to generalise
across the entire geographical range and the huge spans of time involved. The
conclusion that the Gibraltar Devil™s Tower individual was mainly carnivorous
(Lalueza Fox & P©rez-P©rez, 1993) illustrates the dif¬culty. If we assume that
this individual was not anomalous then, on the ecological evidence from the
Gibraltar sites (Finlayson & Giles Pacheco, 2000; see below), we would have to
accept that, while a large proportion of the diet would have been meat it would
not have been exclusively so.
The problems associated with high protein intake, especially in skeletally-
robust hominids, and the need to include fat and carbohydrate have been outlined
by Cachel (1997). Additionally, high protein intake may have a negative effect
on pregnant women which may explain reductions in protein intake by hunter“
gatherers at certain times of the year (Speth, 1991). In the Mediterranean lands
Neanderthals may have had dif¬culty in obtaining fats from mammalian her-
bivores that would have been leaner than their counterparts in the high latitude
plains. This dif¬culty may have been alleviated by consumption of marine mam-
mals “ in Vanguard Cave, Gibraltar, during the last interglacial Neanderthals
consumed monk seal Monachus monachus and probably dolphins. The options
presented by the Mediterranean environments in terms of insect larvae, fruits,
nuts, roots and tubers and marine molluscs in coastal sites, appear to have been
exploited by Neanderthals, thus minimising the effects of total dependence
on mammalian meat. The conclusion that Neanderthals consumed mammalian
herbivore meat is undeniable, it has been known for a long time. What we can-
not infer, however, is that that is all they consumed. Perhaps the individuals on
the plains only ate such meat but, then again, there would not have been much
Comparative behaviour and ecology 97

else that could have been eaten and herbivores would then have been abundant.
Plains dwellers would initially have had access to terrestrial mammalian fat
and freshwater ¬sh in localised areas. These Moderns were less robust than the
Neanderthals and the problem of loss of calcium resulting from a high protein
intake would have been reduced (Cachel, 1997). It is interesting to note that
in Arctic hunter“gatherers, group size and sociality is constrained by the cost
of acquiring adequate amounts of fat (Cachel, 1997). The ease of acquisition
of fat by Moderns exploiting the herbivore biomass on the Eurasian Plain may
be a contributory factor in the sociality and large group sizes of these people
(Gamble, 1999; see below).
There is a view that links changes in food types consumed by humans and
an increase in the diversity of food types taken from the middle Palaeolithic
to human population pulses (Stiner et al., 1999, 2000). According to this view
humans would have initially selected ˜slow™ prey, that is prey in which capture
time was minimised, and then moved to more mobile prey once the slower
prey had been depleted (Stiner et al., 1999, 2000). My view is that the only
generalisation that we can make about diet is that humans have for a long
time been able to eat a wide range of foods. People from at least the time of
the common ancestor of Neanderthals and Moderns have been opportunistic
omnivores capable of handling a wide range of foods, animal and almost cer-
tainly vegetal. I predict spatio-temporal differences at all scales in response
to spatio-temporal resource heterogeneity. There is no theoretical reason or
empirical evidence to propose that changes across time should be linear or uni-
directional. If there is a case to be made for the diversi¬cation of the range of
prey exploited and methods used by humans, then it is only after the Last Glacial
Maximum (LGM) and especially towards the Pleistocene“Holocene boundary
as large mammalian herbivores became regionally depleted (Holliday, 1998;
Elston & Zeanah, 2002). The subsequent evolution of food production may be
a development of this process (Chapter 8; Diamond, 2002)
Stiner et al. (1999) suggest that Palaeolithic human population growth de-
pended on variations in small game “ overexploitation depressed the popu-
lations of certain prey leading to hunting of less favourable types. Only four
Italian and two Israeli sites were used in the analysis and the temporal scale
of resolution of the faunal data did not match the ¬ner-scale variability of the
late Pleistocene climate (Chapter 6) so that it is not possible to conclude that
communities were insensitive to climate. The claim is made on the basis of
species composition comparisons and does not take abundance into account.
Their analysis of prey composition through time is ¬‚awed. Inter-site data are
lumped. The relative contributions of marine mollusc and vertebrate abundance
are compared even though the methods of estimation differ “ number of iden-
ti¬ed skeletal specimens are used, instead of minimum number of individuals,
98 Neanderthals and Modern Humans

to estimate vertebrate prey but minimum number of individuals are estimated
for marine molluscs and the two are readily compared. Claimed trends in prey
size reduction are statistically insigni¬cant with considerable overlaps. In any
case the size differences may re¬‚ect inter-site differences. For example, all their
estimates of humeral shaft diameter for the early periods (200“70 kyr) are from
Hayonim Cave in Israel and it is impossible to determine whether later changes
are due to temporal shifts or simply because other sites (which might always
have had smaller size categories) were being sampled. Even more seriously,
in the case of marine molluscs different species within the same genus (e.g.
Patella), that are known to differ in size in the wild, are lumped in size compar-
isons and we are therefore left with the uncertainty of the extent to which the
observed trends simply re¬‚ect different proportions of species in each sample.
Possible biases due to inter-site and inter-species variations may therefore be
superimposed on the claimed temporal patterns and these alternatives have been
The availability of tortoises would have been reduced during Oxygen Isotope
Stages (OIS) 4 to 2. Northern Italy is currently at the edge of the geographic
range of Testudo hermanni and well outside that of T. graeca (that in any
case may be a recent introduction); in Stiner et al.™s (1999) study, the tortoise
disappeared faster in Italy than in Israel where they were not lost altogether.
The reductions in sea level generated by cooling would have disconnected the
Italian caves from the immediate coastal environment (Kuhn, 1995) and could
have also reduced the marine mollusc contribution. Such environmental factors
could also explain their observations. Relative abundance trends in other prey
would result from tortoise and marine mollusc reduction and need not re¬‚ect
real increases. Recent work in Gibraltar (Finlayson & Giles Pacheco, 2000)
indicates that vertebrate community composition was similar throughout the late
Pleistocene but climate altered vegetation and the local availability of species.
Stiner et al.™s (1999, 2000) study cannot even be regarded as indicative because
it extrapolates from the scale of a handful of local sites, some of which may
not even be independent of each other, to a global scale. Whether or not there
was ever a broad spectrum revolution (see Chapter 8) we certainly cannot infer
it from these studies. Ultimately, climate seems to have been the key factor in
the affairs of the Palaeolithic humans of the Mediterranean and further.
Clearly, in more varied regions Neanderthals were omnivores. It is more likely
that the over-dependence on meat in marginal areas re¬‚ects the increasing stress
to which these populations were subjected. The reality is that we have increasing
evidence that Neanderthals across a huge time span and going as far back as the
last interglacial at least were consuming marine resources including molluscs,
seals and probably ¬sh and cetaceans, just as other contemporary humans were
doing in similar situations at the same time in South Africa (Deacon, 1989).
Comparative behaviour and ecology 99

Deacon (1989) has argued that African Middle Stone Age (MSA) subsistence
behaviour should be regarded as ˜modern™, there being no evident difference in
subsistence ecology. Acheulian sites in South Africa are tied to valleys and water
sources in the coastal platform. MSA/LSA (Late Stone Age) sites are found high
up in the Cape Mountains as well as on the coast and there is frequent use of
rock shelters. MSA populations ate meat and marine and molluscs (source of
minerals) but there is no evidence of ¬shing or hunting of ¬‚ying birds. We also
know that in the right conditions, for example in central Africa, the harvesting of
freshwater resources was happening in the MSA (Brooks et al., 1995). Similarly,
the Neanderthals occupying the topographically heterogeneous Mediterranean
belt from Iberia in the west to, at least, Crimea in the east exploited a wide range
of foods that included large mammals, small mammals and birds, tortoises,
marine molluscs and probably even marine mammals, ¬sh and plants (Stiner,
1994; Finlayson et al., 2000a). Such a varied diet was probably a re¬‚ection of
the micro-spatial and seasonal variability in resource availability in these areas
and these would have also reduced the risks associated with overdependence
on speci¬c prey items. In Israel Moderns and Neanderthals hunted the same
animals but Moderns differed from Neanderthals in having a more seasonally-
speci¬c hunting strategy (Liebermann & Shea, 1994).
Humans have therefore been consuming a broad spectrum of prey, when avail-
able in suitable environments, from at the very least the last Interglacial and
probably much further back. The opportunistic humans would have optimised
foraging tactics and these would have varied temporally and spatially, and at
different scales, depending on resource availability. The degree to which Mod-
erns and Neanderthals were specialised hunters is also likely to have been very
¬‚exible. Clearly, Moderns in the open plains of Eurasia probably specialised in
particular types of herding prey at particular times of the year but Neanderthals
on the edge of the plains were probably very similar (Gamble, 1986; Mellars,
1996). Moderns and Neanderthals in the heterogeneous mid-latitude belt would
have varied from specialised to generalised hunting in accordance with the na-
ture and dispersion of their prey. I would therefore predict a higher probability of
specialisation on the open plains than in the more heterogeneous landscapes. In
the latter case I would expect a mosaic of specialisation“generalisation, related
to environmental and climatic features, that is independent of hominid taxon.
Recent evidence from south-western France (Grayson & Delpech, 2002) that
shows specialisation in particular resource taxa by Moderns and Neanderthals
corroborates this view. It is hardly surprising that the authors should ¬nd no
difference in the level of specialised hunting between the Mousterian and Au-
rignacian of this region. The two populations responded to similar terrain in
a similar fashion, a situation not dissimilar to that in the Levant (see below).
In any case we must be aware that the meagre data available to us lacks the
100 Neanderthals and Modern Humans

resolution that some authors would like and it is not possible to substantiate
global theories on this basis.
Gamble (1995) compiled a database of 588 sites in his north-central (NC),
south-east (SE) and east Mediterranean (ME) regions. These regions coincide
approximately with the Eurasian Plain (NC), the heterogeneous mid-latitude
belt (ME) and an intermediate region (SE) between the two. Gamble (1995)
provided data from archaeological sites and palaeontological sites, the latter
with no human activity. Since the data recorded presence or absence of species
in each site, density biases were avoided. I have re-analysed these data (Table
5.1) “ I estimated species availability to humans from the palaeontological
data. The data in Table 5.1 provide the following information: (a) availability
“ the frequency of each species in each region; (b) selectivity “ the difference
between presence in archaeological sites and the expected presence from the
palaeontological data; and (c) relative differences in selectivity between regions
within time periods allocated to Middle Palaeolithic, early Upper Palaeolithic
and late Upper Palaeolithic. The data were too fragmentary for the late Upper
Palaeolithic to be compared with the other two periods. The following patterns
emerged from the data.

Predominantly plains species

These were mammoth, horse and reindeer. All three were actively selected by
humans on the Plains (selected equates to hunted or scavenged in all cases).
The availability of the three species in the mid-latitude belt (MLB) was low.
Mammoths were selected as encountered but horse and reindeer were actively
selected. Mammoth and reindeer were selected in the plains at a higher rate
than in the MLB in the Middle and early Upper Palaeolithic. Horse was also
selected at a higher rate in the Plains in the Middle Palaeolithic but the trend
was reversed, although not in equal intensity, in the Upper Palaeolithic. This
trend may re¬‚ect the southern range shift of the horse with the onset of the
LGM. The dataset is incomplete for the giant deer, the elk and the saiga but we
may tentatively place them as Plains species occurring at intermediate levels of
availability, the ¬rst two species being actively selected and the saiga selected
at the rate of encounter. The availability of giant deer in the MLB was low but
they were actively selected. Elk and saiga appear to have been largely absent.
There is a suggestion of a trend towards higher rate of exploitation of giant
deer in the MLB in the Upper Palaeolithic, possibly re¬‚ecting a similar range
change response to that for the horse.
Table 5.1. Analysis of mammalian herbivore consumption by humans in Europe. The first three columns record the availability
(AV) of each species for all regions and time periods: +, occur statistically significantly greater than expected; =, no statistical
difference from expectation at random; ’, occur statistically significantly below expected. pl, Plains; int, intermediate region
between plains and heterogeneous belt; het, heterogeneous belt. The next three columns record selectivity (Sel) by humans for
all regions and time periods. The remaining columns record the observations by time periods: MP, Middle Palaeolithic; EUP,
Early Upper Palaeolithic; LUP, Late Upper Palaeolithic. Statistically significant positive relationships are in dark grey;
insignificant relationships are in light grey; statistically significant negative relationships are in white. See text for interpretation


Species pl int het pl int het pl int het pl int het pl int het

Mammoth + = ’ + = = + ’ ’ + ’ ’
Rhinoceros = = = + + = + ’ ’ + = ’
Bos = = = = = + = = = = = =
Megaceros + = + = = = ’ = +
Alces + + = ’
Horse + = ’ + + + + ’ ’ = = = + = ’
Red deer ’ = + + = = = = = = = =
Reindeer + ’ ’ + + + + ’ ’ + ’ ’ = = =
Sus = ’ + = = =
Ibex ’ + + + = + ’ = + = = =
Chamois + +
Saiga =
Roe deer = = = = ’ = = = =
Selected 9 4 6 5 0 1 3 0 1 1 0 0
Encountered 4 5 4 4 7 5 4 6 4 1 2 1
Avoided 0 2 0 1 4 5 1 2 3 0 0 1

From: Gamble (1995).
102 Neanderthals and Modern Humans

Predominantly heterogeneous landscape species

These were red deer and ibex. Ibex were actively selected but red deer were
selected as encountered. The availability of the two on the plains was low and
both were actively selected. Ibex were selected at a higher rate in the MLB
than on the plains in the Middle Palaeolithic but there was no difference in the
early Upper Palaeolithic suggesting a greater specialisation in ibex hunting in
the Middle Palaeolithic in the MLB. There was no difference in the case of the
red deer, between regions or periods. The dataset is incomplete for the wild
boar and chamois but we may tentatively place them as MLB species occurring


. 17
( 43 .)