Feeding Ecology and Human Evolution
by Laurie Meadows
A 'web pseudopaper'[n] (draft)
The uncertainty of bones
Insights into human feeding ecology from extant great apes
Usefulness of fossil homind specimens in throwing light on on feeding ecology and human evolution
Evolution of adaptation to new feeding ecologies - plausible models
The place of the evergreen tropical forest in human evolution
    The temporally distant tropical rainforest ape: foundation of human dietary evolution
The place of the seasonal tropical forest environment in human evolution
The place of the shrubland environment in human evolution
The place of the woodland environment in human evolution
The place of the Savannah biome in human evolution
Version history, licence to copy or modify

The necessity of maternal care to offspring survival. Women invest more in offspring, whether it is
in terms of gamete size or postpartum care, relative to men. Although males contribute to the care of offspring in terms of
protection and resource allocation, varying on their level of paternal certainty, females are often ultimately responsible for
offspring survival. At the end of the chapter, Campbell sets the stage for her work on female intrasexual competition. She
proposes maternal survival is crucial for offspring survival, and thus females cannot afford to engage in violent, risky, and
dangerous competition, such as that undertaken by males.

2.9 km/hr is normal knuckle-walking speed of chimps, 4.5 km/hr is normal bipedal walking speed of humans (Rodman  and McHenry 1980)

Subject:  [Paleoanthro] Sponheimer and Lee-Thorp 2003
Date: Tue, 14 Oct 2003 23:02:21 +1300
From: Lorenzo <fmra1@naturalhub.com>
To:  "paleoanthro@yahoogroups.com" <paleoanthro@yahoogroups.com>
Many thanks for this Gerrit, very much appreciated.

After reading the paper, I tried to recall some of the detail of my personal lines of thinking were on the subject. Thanks to
google (faster than trying to find the page on my local PC!), I found some draft notes I threw up on the net last year.

"However,while the robust australopithecines may have been hard object specialists (Grine and Kay,1988 )..."

On  my arguments snipped from the web page and presented below, I tend to the same speculation -
"First, we can set aside those hominids with massive jaws, massive flat crowned teeth, and robust anchor points for the chewing
muscles. These animals are most likely to be primarily eaters of reeds and rhizomes, as their teeth are not reported to have leaf
disrupting ridges, unlike gorilla. I addition, they have very small canines which means they can move their jaws in a lateral plane,
effective in powerful grinding-chewing. They are reasonably interpreted as adapted to a specialised feeding ecology based on
wetland margin plants. As such, they cannot be informative in this discussion."

I would only add that gorilla loses the shearing ridges with wear, but differential wear of the dentine and exterior enamel
re-creates sharp edges on the eroded molar surfaces. AFAIK eroded afarine teeth have much less foliage disrupting occlusial
teeth. I would like to be corrected on this. I enquired of several researchers the fate of Pan molars occlusal surfaces over time,
and whether sharp edges of uneroded outer tooth wall developed with erosion of the softer exposed central crown dentine, but
have never received a reply.

"In tropical Africa, grasses and some sedges utilize the C4 photosynthetic pathway while trees,bushes,shrubs and forbs use the
C3 photosythetic pathway.".....", making carbon isotope analysis a useful technique for studying the diets of modern wildlife ...
Stable carbon isotope analysis provides a simple means of testing the hypothesis that australopithecines had a specialized diet of
C 3 fruits and leaves.If they ate these foods nearly exclusively,then their enamel bioapatite 13C /12C ratios should not be
significantly different from those of their C3 browsing coevals such as giraffe and kudu; however,analyses of South African
australopithecines from Makapansgat,Sterkfontein,and Swartkrans shows this not to be the case ....Instead, both A.africanus
and A. robustus have 13C/12C ratios that are highly different from those of both C3 and C4 specialists with which they are

I think the critiscms previously discussed on this list are still germane -

and I notice the authors address some of these critiscisms-

"Moreover, great variation was observed in the 13C /12C ratios of australopithecines. Indeed, A.africanus is more variable than
any other taxon occurring at the South African australopithecine sites (Sponheimer and Lee-Thorp,1999a;van der
Merwe et al.,in press ).Some of this variation may result from dietary differences within a population, but the variability more
likely reflects changes in diet _over time_ due to environmental perturbations."

The use of upland moist plateu grass was covered in the above post, and another copiously available source alluded to, along
with the matter of how durable microwear might be as a diagnostic, in my web page notes-

"Australopithecus afarensis was heavier than the chimpanzee (male A. afarensis weight is estimated at 70 kg [154 lbs], females  at 28.9 kg [64 lbs]; chimpanzee male is about 48 kg and female 42 kg [r]). The molars are large and flat, and would likely be useful for crushing the fibrous rhizomes of marsh plants such as Typha in order to extract the starch. Their teeth show signs of polishing by fine particles, which some researchers suggest might be due to the minute 'opal phytoliths' present in sedges of the family Cyperaceae, and in grasses [r]. Certainly, some Cyperus species have small edible bulbs. The mud they grow in might also be expected to have a polishing effect. The relatively thick stems of some Sorghum species have a somewhat sweet sap.[r] Enamel is thick, possibly to withstand abrasion from fibres in roots and rhizhomes, as well as adhering mud. The markedly large and slightly horizontally inclined fruit scooping incisors typical of a primarily frugivorous ape are reduced and slightly more vertically placed. In all, they are not too different from chimpanzee or human teeth in their functional aspect, except that the molars could be interpreted as adapted for a greater bulb, root, and rhizome component in the diet. To this extent (and
discounting enamel thickness), it could be argued they are as intermediate between a tropical forest feeding ecology, and a tool
assisted tree seed and bulb based feeding ecology as Ardipithecus. "

I singled out the widespread sedges of the genera Typha and Cyperus in the above snip. This was deliberate. These two genera
would be by far the most widespread, and probably contain the greatest amount of nutrient, both in the rhizome and in the
succulent shoot and in the stem.

The authors say-
".....Sedges are readily available in these environments and have been argued to be among the possible sources of the C4 signal
in australopithecines (Sponheimer and Lee-Thorp,1999a;Conklin-Brittain et al.,2002 ).Some sedges have underground storage
organs that have protein levels equal to those of most chimpanzee foods (9% crude protein ),but much lower fiber levels (16%
fiber )than foods consumed by chimpanzees (33%)(Conklin-Brittain et al.,2002 ).Equally important, these underground parts
are abundant yet inaccessible to most other mammals,making sedges a high-quality resource for which there is very little
competition....Moreover,we recently gathered sedges from a variety of localities in Kruger National Park,South Africa. Only
43% of the sedges we encountered utilized C4 photosynthesis (Fig.2 ).Consequently, it would appear that,in South Africa at
least, australopithecines would have had to have been extreme sedge specialists to explain the 33%C4 signature observed in
most individuals.More specifically,if we accept that approximately 40%of the sedges that hominids encountered used the C4
pathway,over 90%of their diet would have had to have been sedges to account for a 33%C4 signature."

43% of a species list of how many? I have argued that only Typha and Cyperus would likely be worth 'getting out of bed for'.
As stated on my web page, I don't know if Typha (in particular) is C4 or not. After this paper, I still don't know! The idea that
Typha could be -largely- responsible for the signal at the lifestage and environment of sampling ought to be easily falsifiable. I
wish the authors had. Or hadn't.

November 14th 2003 Paleoone
I googled a paper when trying answer the question of whether typha is a C3 or a C4 sedge -

Cloern JE, Canuel EA, Harris D. 2002. Stable carbon and nitrogen isotope composition of aquatic and terrestrial plants of the San Francisco Bay estuarine system
Limnol. Oceanogr.Vol 47(3). pp 713–729

The C3 saltmarsh plants Grindelia stricta, Salicornia virginica, Scirpus maritimus have similar C13 values as the sedges Typha latifolia and T. angustifolia (around -27). In contrast, the C4 saltmarsh species Distichilis spicata and Spartina foliosa have markedly different C13 values, about -14.

From this, and without specific evidence to the contrary, one might infer that Typha is a C3 plant.

However, the authors mention that marine diatomes can use both C3 and C4 pathways of carbon fixation (citing Reinfelder et al. 2000); in similar manner, I found an oblique reference to carbon pathway work in Australia that implied, but gave no details, that a suite of aquatic plants that included the genus Typha could switch between C3 and C4  pathways, depending on unresolved environmental conditions which might include temperature and aerobic vs anaerobic conditions.

Originally I wrote-
" I have argued that only Typha and Cyperus would likely be worth 'getting out of bed for'.
As stated on my web page, I don't know if Typha (in particular) is C4 or not. After this paper, I still don't know! The idea that Typha could be -largely- responsible for the signal at the lifestage and environment of sampling ought to be easily falsifiable. I wish the authors had. Or hadn't."

It looks like the 33% C4 signature in the lifestage and sample of Australopithecines measured in Sponheimer and Lee-Thorp 2003 cannot at this time be attributed to a diet with significant amounts of Typha.
 What might explain the C4 signature? Well, 40% of South African sedges use the C4 pathway, according to a personal communication by Stock to the authors. But all sedges are not created equal. Only 2 seem productive and widespread. We have provisionally discounted Typha. What about Cyperus?

Acoording to http://pi.cdfa.ca.gov/weedinfo/CYPERUSE2.html, Cyperus is a C4 plant. In addition, while it is normally associated with moist areas, it tolerates dry conditions. It seems a rather plastic genus, with Cyperus esculentus growing to 1,000 metres, and growing from warm temperate regions to the tropics.

According to -
Addy E O and Eteshola E (1984) Nutritive value of a mixture of tigernut tubers (Cyperus esculentus L.) and baobab seeds (Adansonia digitata L.).
Journal of the Science of Food and Agriculture, 35(4): 437-440

"the tubers contained 47.9% digestible carbohydrates, 32.8% oil and 3.8% crude protein.".

we have seen evidence for digging with horns and bone which might be attributabal to Australopithecus and sp. affin Homo erectus in Southern Africa. While the wear patterns on the artifacts identified strongly suggest termite mounds and not earth scratchings, the principle of scratching for food is directly transferable.


components listed are likely to be close to optimal for the primate body.

"These data suggest that even if sedges did comprise an important resource for early hominids, they were likely supplemented
with other C4 foods."

Inevitably, with Australopithicines tied to water, the gallery forest and seasonally flooded grasslands will provide seasonal
opportunities for capturing reedbuck and similar lambs/fawns (amongst others). There is no reason to suppose they were greatly
more successful in the seasonal harvest than, say, baboon. Because the evolution of the schistome  Schistosoma haematobium
resolves to about 2 mya (E. Loker, pers. comm) it is a plausible argument australopith 'haunting' a riverine/lakeside habitat (in
search of Typha), allowed its evolution as an almost exclusively human parasite. Equally, the three species of taeniid tapeworm
that affect only humans as final host resolve molecularly to around 2 mya. What the intermediate host may have been at that
early time isn't known; a good bet would be burrowing and hiding  riverine rats such cane rats or marsh rats.

So it is neither one thing nor the other. A mixed signal is expected for an animal living in a riverine mosaic, according to season
and geography. An example is on the web page mentioned above. The success of the Australopthicines might fairly be attributed
to dietary plasticity. The authors note that -
"...capuchin monkeys (Cebus apella) have large, bunodont dentition with thick enamel adapted for consuming fruits and hard
nuts. Nonetheless, up to 50% of capuchin diets can come from animal foods, although the average is closer to 25% (Fleagle,
1999;Rosenberger and Kinzey,1976;Rosenberger,1992 )."

Clearly, having large molars per se - useful for crushing succulent sedge stems in the Australopith case - is no barrier to eating
small animals ( and as the web page points out, Human and Pan molars are not hugely different morphologically, and thrashing a
small animal to death as well as tearing it to pieces renders it small and pliable enough that neither enamel 'knives' nor stone
knives are needed).

Dietary plasticity is the key to survival. This in turn is leveraged by manual dexterity, longer term memory, and social
cooperation (elephant fit these criteria if the trunk is 'manually' dextrous). The ability to store fat - save some of the oversupply in
seasonal good times - is also an advantage. And turning a liability - dependance on water - into an asset by harvesting nutritious
riverside plants adds strength to the hand. Taking the view that Australopiths were a kind of woodland bipedal 'chimpanzee', it is
axiomatic that they outcompeted evergreen tropical jungle living Pan. However, it is equally axiomatic that having adapted to
woodland/riverine mosaic Australopith were no longer competative in the relative abundance of the evergreen tropical forest in
times of climatic extreme.

The authors conclusions seem to me fair, or at least they accord with my biases, in that they align with my longstanding
assumption that Australopiths filled a woodland chimp niche, with accent on great dietary plasticity, but with primary food
sources of carbohydrate laden Typha, Cyperus, Phragmites, Pennistemen, waterlily, summer dormant corms,  high sugar Grewia
and similar berries, and tree seeds. Plus anything and everything else from termites to palm grubs, rats, monkeys, turtles,
grasshoppers, fawns, fruits -everything edible.

It is very pleasing to see this kind of paper come forth. Now lets see some estimates on interelationships of 'carrying capacity' of
the land, permissive feeding ecologies, population size, population trajectories, corelates with climate, and speculation on group
structure in varying ecozones...well, dreams are free.

Lets start with sorting highly productive sedges into C3 and C4 and publishing the results. All good field data is useful for creating defensible speculative scenarios as far as I am concerned.


A new functional interpretation of the Swartkrans early hominid bone

                          Lucinda Backwell1 and Francesco d'Errico2
                          1Palaeo-Anthropology Unit for Research and Exploration, Department of
                          Anatomical Sciences, University of the Witwatersrand Medical School,
                          Johannesburg 2193, South Africa.
                          2UMR 5808 du CNRS, Institut du Quaternaire, 33405 Talence, France.

                          Ever since Dart identified bones from Makapansgat as tools, scientific
                          consensus has fluctuated as to whether certain modified bones from
                          early hominid sites should be interpreted as artefacts, or simply the
                          result of non-human taphonomic processes. Brain and Shipman's
                          analysis of 68 specimens from Swartkrans and Sterkfontein, and their
                          interpretation of these as tools used for digging up tubers or working
                          skins has provided new data in support of the anthropic hypothesis. This
                          interpretation, however, was not validated by comparative analyses of
                          wear patterns produced by natural processes, and alternative functional
                          interpretations were not tested.

                          In this study, the results of a new archaeozoological, morphometric and
                          microscopic analysis (optical microscopy, SEM, image analysis) of the
                          Swartkrans and Sterkfontein material have been compared with those
                          obtained from reference collections of bones modified by known natural
                          agents, and from bone tools experimentally used in different activities,
                          including Brain's experimental tools.

                          Our results confirm that many of the objects were tools used by early
                          hominids. However, image analysis of the wear patterns contradict
                          previous functional interpretations. The width, orientation and
                          morphology of striations on experimental tools used to dig termite
                          mounds are consistent with those found on the Swartkrans and
                          Sterkfontein specimens, and different from those used by Brain and us
                          in digging tubers and other activities. Termites provide a rich source of
                          protein and fat, and are traditionally harvested after the rains as they
                          vacate the colony. The use of bone tools implies that early hominids
                          tapping this resource possessed the cognitive ability of developing a
                          technique able to transcend the ecological constraints of seasonal
                          availability. Moreover, breakage patterns reveal that in 96% of cases,
                          the hominids used weathered long bone shaft fragments, suggesting
                          that their gathering of bone blanks was not subsidiary to scavenging.
                          Previously undescribed traces of shaping or re-sharpening through
                          grinding have been identified, as well as sixteen new bone tools from

Without a time machine, we will never know specifically how we changed from an ape restricted to a single tropical forest on one continental mass, to an ape of worldwide distribution - an ape that now lives in all climates and geographies, from the arctic to the semi-desert, from seashore to mountain range. The only evidence we have for the changes that occurred are some fossilised bones and chipped stones from Africa, which, perhaps with adjacent regions, is recognised as the cradle of human evolution  The problem with using bones to tell the story of evolving ape species adapting to new habitats is that there is a homo-centric impulse to ascribe all ape-like fossils as human ancestors, and none as ancestors of the other currently living large hominoids, chimpanzee and gorilla [n]. The number of fossil hominids found is limited and biased - and attempts to divine relationships between these limited samples based on bone morphology has been found to be unreliable.[r] We simply don't know which animals were ancestral to humans.[n] We don't know the geographic range of the ancestral animals, and we know little of the various kinds of habitat in the various ecological zones, including animals and plants (in particular) that might have been present. We can, of course, guess from the sparse lines of evidence available.[n]

As we can only make informed guesses about the identity and importance of the various fossilised bones found so far in clarifying the picture of human evolution, any further lines of evidence, even if indirect, are useful.

In the past, the focus in paleoanthropology has tended to be on the bones, rather than the environment in which they are found. The palaeoenvironment is now attracting much more attention (Potts)[r] because everything about an animals life and body form is shaped and constrained by the environment it lives in. The food available to an animal is a function of the ecological zone it lives in. How it finds that food, how it accesses it, is a function of body form, physiology and behaviours, both innate and learned. In turn, the food resources in the environment, their physical distribution, productivity and seasonality influence group structure and behaviour.

This pseudopaper largely abandons the usual method of assuming a particular evolutionary tree and specific dates of speciation, and takes as a starting point the feeding ecology in Africa today - that is, the kinds of food available to an ape in the tropical forests, the woodlands and adjacent waters, and in more open bushland, their seasonality and physical distribution, and behavioral consequences. These ecologies are assumed to be broadly analogous to the environment of evolutionary adaptation. There have been large changes in the relative extent of  the major feeding ecologies from miocene times to the holocene. These changes are assumed to have accelerated the speed of evolution of the forest ape to the panecologic ape of today.

Many assumptions have to be made, not least of which is the reliability of drawing analogies between present food plants and animals and those present during human evolution. There is relatively little unambiguous data from which to construct an argument. As a result, such positions as this, while plausible, are speculative and leading.[n] Even so, if we are to develop strong arguments on human evolution, we need to 'fit' an ape into each broad feeding ecology available over fossil calibrated evolutionary time, show how it made its living, and show how one feeding ecology might preadapt the animal for either stepping into a further ecology, or predadapt it to climate caused major changes within its feeding ecology. In addition, where sufficient fossil hominid and associated fauna remains exist, it should be possible to infer a plausible feeding ecology A sweeping hypothetical ecology based position such as this is classically weak in that it is very difficult to test, each strand of conformity to recognised broad biological, ecological, and energetics principles buttresses the idea to the point where it becomes sufficiently plausible to present for discussion and feedback/critique.

Evolution of adaptation to new feeding ecologies - plausible models
The current broadly accepted theory of human evolution is that all African apes, including ourselves, evolved from a tropical forest dwelling ape. The detail of the sequence over time and geographic area is the subject of argument - sometimes heated.

Leaving aside the difficulties of discerning a species over both time and geography using scrappy fossil remains, interdisciplinary strands of evidence suggest a sequence of evolution that starts with a species or species of ape that was a tropical forest plant part/frugivore, speciation occuring at the forest margin to form a woodland mosaic omnivorous but still largely frugivorous ape(s), and final speciation(s) to seasonal bushland mosaic omnivore. The last phase was probably developed in the unique conditions of the Southern African bushland mosaic feeding ecology ( a case could also be made for west Eurasia, within the distribution of seasonal edible bulbs, e.g. species of Crocus and Tulipa).

And that is tool use and transmitted cultural memory, coupled with ability to switch between semi-migratory and sedentary lifeways. I argue that the overlay of culture carried humans into to a relatively unconstrained 'cross biome obligate terrestrial omnivore'. The last phase may have commenced even 2 million years ago; or as recently as 600,000 years ago. The last phase is not particulary important in this musing about the interplay between human feeding ecology and human evolution. Physiologically and anatomically, it is fair to suppose that we were fully adapted to terrestrial omnivory by the time of Homo erectus. Culture allowed us to access a wider range of plant and animal species as we radiated back into the woodland mosaic in the final phase, displacing the arboreal-terrestrial Australopithecines.

The timing of each phase, and the time within each phase, is unknowable due to lack of any reasonable sample of palaeontological or other line of evidence. Therefore the focus is on ecological niches and the possible food available to hominids in those niches. Any possible morphological or significant environmental barrier to omnivory by an African biped is also considered.

The major condition, which in all cases must be met, is that as diurnal, active primates with energy expensive brains and offspring draining female nutrional resources for a particularly long time, we must have access to nutritionally dense food.[r]

Therefore, there are two plausible models, with the first being simpler
Simple model
1. An African forest environment has never been absent of apes for at least 18 million years to present; and the earliest ancestors (not necessarily the last common ancestor) of all three African apes evolved in a forested environment, tropical, subtropical, and possibly warm temperate.
2. at some undeterminable point ancestral forms moved from the wet seasonless tropical forest to seasonal tropical forest.
3. at some undeterminable point - perhaps broadly contemporaneously - ancestral forms or sister groups of ancestral forms moved from the seasonal tropical forest  to closed and open woodland environments following stable riverside and lakeside vegetable food sources, a niche open for exploitation.
4. One or more subpopulation(s) evolved in a nutrient dense but increasingly unreliable environment highly seasonal environment. Perhaps in the seasonal woodlands of central and southern Africa, seasonal Ethiopian uplands, the open woodlands of the deep sandy soils of what is now the Kalahari; or similar specific and restricted habitat types which provide relatively large amounts of self-storing, non-spoiling calories in the form of carbohydrates and oils in tree seeds and plant underground storage organs.

More complex model
A very much weaker competing model would have a much earlier time frame -
1. One group evolved behaviourally and morphologically outside the evergreen tropical forest as it began to shrink and yield to gallery forest and woodland mosaic, perhaps 9 million years ago. Sub populations expand into wooded west Eurasia.
2. Another group remained as a primarily arboreal frugivore/folivore in the then more widespread African tropical and subtropical evergreen forest niche (prior to about 7 million years ago).
3. Widespread climate change about 7-6 million years ago shrank the tropical forest further, with grasslands expanding in East Africa. Semi-deciduous tropical forest and woodland mosaics expanded, as did the woodland ape population.
4. A sub population of the bipedal but still arboreally adapted woodland ape adjacent to the central African forest margin outcompetes the stem population still resident within the evergreen and semi-deciduous tropical forest, due in part to greater dietary plasticity in markedly dry seasons.
5. Once in the evergreen tropical forest, the bipedal animal is no longer under selective pressure for bipedal morphology, and vestigal knucklewalking adaptations are selected for.
6. One or more subpopulation(s) of the gallery forest/woodland mosaic ape evolved in a nutrient dense but increasingly unreliable environment highly seasonal environment requiring extensive ranging. Perhaps in the seasonal woodlands of central and southern Africa, seasonal Ethiopian uplands, the open woodlands of the deep sandy soils of what is now the Kalahari; or similar specific and restricted habitat types which provide relatively large amounts of self-storing, non-spoiling calories in the form of carbohydrates and oils in tree seeds and plant underground storage organs.

Other 'possible' models could be developed, but these two are plausible.

The sequence, and details of the persistance and overlap or otherwise of founding populations with diverging sub populations is immaterial to this opinion. Which particular species is derived from whichever other species and the time between speciations is immaterial. These matters are dependent of meaningful sampling of populations across their range and through time, and it is incontrovertible that we don't have a sample sufficient to draw confident conclusions. Opinion from modern analogues must suffice.

What is pivotal to this argument is identifying what kind of ape dietary a given environment will support throughout a year.

The key assumption, on which arguments such as this stand and fall, is that each environment type will support an ape best adapted to both the possibilities and constraints of that environment. It can say nothing about which model is 'correct', only that one or other is more probable than other scenarios which aren't considered here (a semi-aquatic phase, for example).

Weakness in attempting to model the paleo hominid dietary niche
Plant fossil record as poor as animal fossil record
The weakness of this approach is the fact there must be as much speculation about animal and - particularly - plant species present, as there is about which ape is ancestral to which. There are, however, fossils of animals which give clues to the plants and climate of ancient environments which are far more plentiful than ape fossils.
And in the same way that we allow our ancestors to share features we show today, we can show the same liberty to plants and assume they were broadly similar to present flora in morphology, although obviously much different in distribution.

We can add what little is known about climate from records of rate of dust deposition in marine silts, and from other indirect methodologies.

Broad typological categorisation of habitats obscures their micro-diverse feeding ecologies
A further major weakness is that habitats must be lumped together into broad designations - forest, woodland/grassland, and coastal/riverine - which don't take into account the variations in food supply, diversity, and seasonality caused by altitude, seasonality with latitude, soil constraints, coastal effects, rainshadow effects, regional topographic effects, and drainage patterns. Even in the cleared and cut-over Africa of today, and particularly in Southern Africa, there can be many ecosystems within a relatively short distance of each other, particularly where hills are relatively close to the coast line.

Diverse adjacent feeding ecologies in a 300km transect in Southern Africa
In central Mozambique, for example, upland miombo[n] woodland encounters a western granitic mountain (Gorongosa) which support open, sometimes boggy grasslands at the highest altitude, merging into low montane forest, then becoming deciduous tropical forest futher down the slopes, merging finally into closed canopy savannah wood land on the lower slopes and upland plateau. The woodland supports various ungulates, including the sable antelope and Liechtenstein's hartebeest. The plateau is cut by a trough, the southernmost part of the Rift Valley. The trough is a mix of sand and alluvium and supports dry forest in old river beds, and on the heavier soils, tall grassland with interspersed tall tree savvanah . The river floodplain has short grasses and islands of bushy thickets. The grasslands support the archetypic African herds of impala, wildebeest, zebra, buffalo, kudu, eland, bushbuck, elephant, duiker, and the usual predators. The plateau on the other side of the trough is also covered by miombo woodlands. Limestone ravines cutting into the western escarpment have dry forests on their side and tall trees on the ravine floors. As the plateau fall to the east, the miombo woodland is penetrated by marshy grasslands, with gallery forest along stream sides. Nearer the coast, areas of papyrus swamp or heath may occur. Drier areas between flood channels support palms (Hyphaene coriaceae and Phoenix reclinata) and patches of savvanah. Extensive mangrove swamps line the rivers as they become estuarine, with salt marshes fringing the mangroves. Between the river systems and their fringing vegetation near the coast, there is a low plain, lined on the seaward side by scrub covered low barrier dunes. One survey (Tinley 1997) found 74 different vegetation communities and 40 soil types in this transect. All these ecosytems, with their particular plants and animal food resources, their particular climatic condition, occur in a western upland to eastern Indian Ocean coast transect of only 300 kilometres.

Information on edible wild plants biased by ignorance of the urban living scientifically educated
The information on which existing wild plants and animals are edible (disregarding cultural conditioning) is relatively thin in an agricultural/industrial urbanised world. For example, the trees Burkea africana and Erythrophleum africanum within the above transect of central Mozambique are host to seasonal congregations of large numbers of hairy Saturnid caterpillars [n] that are eagerly sought out for food by local peoples. Many liberties have to be taken in formulating an opinion on the resources of environments that no longer exist.

Usefulness of using wild living modern Homo sapiens to model past feeding ecology
The tropical forest model uses the Chimpanzee as a 'proxy' for the tropical forest ape phase of human evolution. The woodland model lacks an extant woodland bipedal ape. As a proxy, 'wild living' humans will have to substitute, and inferences drawn in spite of about 2 million years or so of distance from historically self-observed Homo sapiens back to a bipedal phase.

Current distribution of gatherer hunters may not reflect the feeding ecology of evolution
Historically observed gatherer hunters may not necessarily be in the 'feeding ecology' where humans evolved. Unfortuneately, wild living humans have been pushed back from their natural range over the entire of Southern Africa to only those remnant groups in the most difficult environments, chiefly in arid areas of Namibia and Botswana, and seasonal woodlands of Tanzania. Even within these more difficult environments there are variations in food distribution and ease of obtaining it. There are regions with more animals than others, other regions where the sandy soils allow large stands of mongongo tree nuts to grow, and other areas with hard soils where the tuber digging sticks have to be weighted with a rock to aid penetration, for example. (This last fact is illustrative of the selective pressure a seasonal environment places on a hominid.)

Within the limits of our knowledge of food plants, their nutritional value (and variation), and their distribution in present (let alone past) day Africa, we can outline a feeding ecology for various broad climatic zones. We can make assumptions about what kind of ape lived in those ecologies, and postulate a path from forest ape to widespread obligate ground living bipedal animal.

But we must be constantly mindful that any reconstruction is still just informed guesswork, i.e. opinion, not fact.

Insights into human feeding ecology from extant great ape hominoids
How informative on evolution and diet are the fellow African Hominoids - Chimpanzee and Gorilla?

Obviously, the digestive system in general must closely fit the feeding ecology. Any degree of 'mismatch' must decrease survivability and fecundity. Therefore, the digestive system is under intense selection pressure, and necessary changes are likely to be relatively rapid. So the usefulness of 'matching' human digestive morphology and physiology with other apes depends, in part, on how close in time the split from our common ancestor was, and in part, the intensity of the selection pressure - an element that is largely unquantifiable.

The usefulness of comparing human and ape diet and physiology partly depends on how far humans and apes have evolved from each other, which depends on estimates of when the last common ancestor of the three extant African great apes lived. Mitochondrial DNA 'molecular clock' evidence suggests  The mitochondrial DNA clock also suggests 'protohumans' and 'protochimpanzees' split from a common ancestral animal more recently, within the last about 4.5 million years, and this date, or a little earlier, has cautious provisional acceptance at this time. Agreement is not universal - one recent re-calibration of the molecular clock suggest the divergence of humans from chimp was much earlier - perhaps 10 million years ago[n]. The range even for the often quoted 5 million year date could be bracketed within 7 to 3 million years ago. The very scrappy fossil record could also be interpreted to put the split further back in time, somewhere around or before 7 million years ago. Calculations of how much evolutionary time there is between chimps and humans is also influenced by the speciation rate of large bodied hominids - do they speciate slowly or relatively quickly? In any event, even using the most recent date for the common ancestor, there is a total of 9 million years (2 x 4.5 million years) of evolution between even the most similar ape to human (chimpanzee) and us.

We could argue that the more recent the split, the less time for evolutionary differences in diet and physiology to arise between us. While broadly true, different environments - tropical forest versus seasonal tropical forest versus woodland - have different food plants and animals available, with different patterns of supply, which in turn results in an evolutionary pressure which might favor different anatomies, physiologies, and foraging strategies.  An animal in a relatively rich and stable tropical environment may be under much less intense 'selective pressure' than one in a less climatically stable environment. For example, once a folivorous anatomy and physiology has been selected for, a frugi/folivorous lowland Gorilla may remain forever realtively unchallenged by hunger so long as it remains in the stable conditions of an evergreen wet tropical jungle; whereas the more frugivorous chimpanzees may experience more stress than gorilla due to sometimes erratic and sometimes seasonal fruiting patterns in seasonal tropical jungle. While about half the species of tree in mature tropical forests of West, Central, and East Africa are used by chimpanzee as a source of young leaves and/or fruit, there is often great variation in ripe-fruit crop load between trees. On average, chimpanzees must travel 10 kilometers between heavy fruiting trees [r].

While there seems little point in attempting to derive any but the most general lessons from the essentially non meat eating frugi-folivorous gorilla diet and physiology, any 'match' between human and common chimpanzee digestive adaptations and feeding behaviour point to many parallel behaviours and overlaps of food types in our two feeding ecologies. The primary functional and morphological difference affects possible feeding ecologies. Humans are poorly arboreal relative to chimpanzees.

A poorly arboreal omnivorous human in a forest is under the greatest selection pressure of all - we cannot live by eating large amounts of pithy material alone, we cannot reach fruit resources high in the canopy, and the dark cathedral depths of the forest floor are relatively unproductive. Therefore human physiology and anatomy cannot afford to be specialised, and feeding stategies have to be very flexible to be able to exploit almost any non-toxic plant or animal food. Specifically, humans would have been unlikely to even survive in a forest before humans evolved far enough to make bows, nets, spears, snares and fire.

We could argue that if a more intelligent protohuman was physiologically and morphology adapted to accessing food and surviving in an evergreen tropical forest biome, it would have. It would have displaced the chimpanzee and gorilla lineages. As these lineages exist today, clearly, once the line to Homo had substantially evolved as a woodland mosaic biped, there was no possibility of re-entering and displacing forest dwelling proto chimps in their folivorous/frugivorous niche, even although chimp and human might live, as Pygmy people do today, in the same forest. One is largely a frugivore, the other a tree seed, plant starchy storage organ, and small animal eating omnivore. As humans haven't displaced chimpanzee and Gorilla from their range (in spite of Pygmy occasionally killing chimp for food), it can be concluded that human survival in a tropical forest environment is simply another demonstration of the human tool aided, fully evolved, highly plastic 'global feeding ecology'.

Physiology and Gut Morphology
First, we can leave aside the mountain gorilla, as it is primarily a plant feeder in a specialised feeding ecology, and eats insignificant amounts of fruit [r].

Lowland gorillas and bonobo chimpanzee reflect the gut adapted to the foods of the evergreen tropical forest - a gut that deals with fruit, and for some periods, succulent leaves, rhizomes, and pithy stem cores. Ripe fruit, the major component of the chimpanzee diet is composed of 13.9% (dry weight basis) water soluble carbohydrates - almost entirely simple sugars such as glucose and sucrose [r]. The soluble sugars (and proteins) are quickly absorbed in the small intestine to meet immediate energy needs. Surplus is converted to liver and muscle glycogen, and further excess to fatty acids for storage. The sugars in ripe fruit are also associated with large amounts of fibre (33.6%, dry weight basis) [r].  It is hypothesized that 'fibre', the plant parts resistant to normal enzymatic breakdown (cellulose, hemicellulose, and lignan) is fermented in the hindgut to volatile short chain fatty acids. The suggestion is that these fatty acids are absorbed through the intestinal wall and used in cellular metabolism in place of glucose.[r] If this is so, the whole fruits and pithy plants may provide more energy than a simple analysis of the simple and complex carbohydrates present would suggest. Soluble fibre, in the form of pectins, gums, and mucilages, is also thought to be fermented by gut bacteria.

The seasonal evergreen tropical forest provides protein to the fruit and vegetable eating hominoids in the prime food, fruit. Fruit has 9.5% crude protein (dry matter basis), the 'fall back' food for hard times, piths, almost the same. Young leaves are protein rich (22.1%, dry matter basis), although not a major dietary componant (time spent feeding basis) [r]. According to Case (1978), primates have a slower postnatal growth rate than other mammals. This leads to the prediction that primates should have a relatively low protein requirement.

The human gut anatomy and physiology can be interpreted to reflect a primary dependance on vegetable food, similar to  chimpanzee. The human colon is distensible to accomodate the necessary fibrous bulk of intermediately nutrionally dense fruits and succulent stems. It is relatively long, a structural adaptation to extract as many nutrients as possible from plant food via hind gut fermentation. As with chimpanzee, food has a relatively long transit time, also to assist fermentation. The human colon also supports the bacterial fermentation of 'fibrous' plant material, deriving significant energy from ferment products (volatile short-chain fatty acids) metabolised in the liver. The human hind gut also ferments any starch that escapes digestion in the foregut.[r]

The transit time of food in the human gut argues against significant carnivory as an important factor in human evolution. The human gut transit time for food is similar to the frugi-folivorous chimpanzee - about 38 to 48 hours. In contrast, carnivores have a much shorter transit time, from 2.4 to 26 hours (varying with species). In addition, their stomach comprises about 60-70% of their digestive tract volume, in contrast to the omnivorous but plant food 'signatured' humans, whose  stomach is about 21-27% of total digestive tract volume.

The physiological processes involved in a plant carbohydrate based diet, such as that of chimpanzee, and human physiological processes are mutually inclusive. Fat contributes relatively few calories to the leaf and fruit eating howler monkey of the neotropics - about 18% of daily calories consumed [r]. The tropical forest dwelling chimpanzee is likely to have a similar percentage, or a little greater if some tree seed and animal food consumption is taken into account. The evolutionarily a-biologic western diet contains 30% or more of its calories as lipids and can't regarded as 'wild species authentic'. In fact, as long as the minimal amounts of essential fatty acids are ingested, the physiological requirement for lipid can be met by carbohydrate and protein.

<>Carbohydrate in excess of immediate energy need is converted via glycolysis to fatty acid, then stored as triglycerides, a very compact source of energy. As a result, a well fed 70kg human has sufficient triglyceride reserves to support some weeks without food. This may be the only point of significant difference between our physiology and that of chimpanzee. If humans had a morphology that could physically access the food, and the physiology to deal with some anti-feeding factors, it is quite possible that a chimpanzee diet would adequately meet our dietary needs. Conversely, a gatherer hunter diet (without bow technology) would likely meet chimpanzee dietary needs.

The morphological similarities between chimpanzee and human gut and similarities in physiology may be parallel evolution. More likely, it reflects a feeding ecology anchored in fruits, pithy plant parts, seeds, and underground plant storage organs.

Feeding apparatus
Our teeth, too, are broadly similar to chimpanzee, except our canines are much reduced, and our incisors are smaller. While chimpanzees cannot chew in a rotary fashion, as in chewing gum, it doesn't prevent them from rapidly eating large amounts of foliage. Equally, humans ordinarily use the same up and down 'chomping' that chimpanzees use.

Insights into human feeding ecology from some extinct hominoids
How informative on evolution and diet are some of the extinct African Hominoids?

This pseudopaper assumes 'model' hominids based on chimpanzee, gracile australopithicine and Homo, with 'model' adaptive behaviour in defined feeding ecologies over assumed evolutionary time span. If the models are substantially valid, the model hominoids should show morphological adaptations to the environment in which they were found and the food sources that might be available in that environment. Although outside the simple models used here, the transition from proto-hominid to hominid would be expected to show a generalised morphology adapted to both feeding zones; in other words that of an animal living in the 'edge' between the two (the edges between two ecologies are often rich in species - woodland mosaics are 'edge rich') .

Regardless of its present range into semideciduous tropical forest, chimpanzees feeding ecology includes evergreen tropical forest. Given the assumption that it evolved from an evergreen tropical forest feeding ancestor, it might be expected to show many of the morphological and physiological adaptations of that ancestor, and therefore be an informative model of the morphology of that ancestral hominoid in the evergreen tropical forest feeding ecology. Using Australopithecus afarensis as a model of an animal adapted to a patchy and complex woodland mosaic is much more speculative. In the abscence of any extant hominoid in this feeding ecology (excluding Homo sapiens), A. afarensis may be useful on the grounds that what paleoenvironmental evidence there is points to it having a woodland feeding ecology. (There is, of course, an element of circular argument here.)

We have to acknowledge the uncertainties inherent in the poor sampling of fossils, the unreliablity of cladistic analysis based on skeletal elements, the incompleteness of the samples there are, and the arguments over assignation of fossils to a species. These factors all combine to make it very unsafe to guess at a particular line of descent to explain human evolution. Basing all opinions on dietary habits soley on 'chosen species' may be misleading. But if you consider it is plausible that we evolved in a woodland and thornbush mosaic, the chances that some of the extinct hominds found so far are ancestral to us is high. So which species are in the right feeding ecology within a plausible timeframe, and what evidence is there that they are adapted to the feeding ecology assumed here?

First, we can set aside those hominids with massive jaws, massive flat crowned teeth, and robust anchor points for the chewing muscles. These animals are most likely to be primarily eaters of reeds and rhizomes, as their teeth are not reported to have leaf disrupting ridges, unlike gorilla. In addition, they have very small canines which means they can move their jaws in a lateral plane, effective in powerful grinding-chewing. The cusps on their molars are low, allowing side to side and back and forward grinding movement; the canines are relatively small and don't constrain grinding. They are reasonably interpreted as adapted to a specialised feeding ecology based on wetland margin plants or succulent stemmed giant grasses, although, like other hominids of the time (and modern chimpanzees), probably ate fruit, small animals and termites as well. As such, they cannot be informative in this discussion.

Bipedal locomotion is one of the most obvious markers of evolving humans. Knuckle walking is adequate for limited ranging in a forest environment, especially when boggy ground is not a frequent part of the feeding ecology. Bipedality might be a clue to relatively extensive ranging and food switching in habitats with patchy resources, and discontinuous forest;  therefore a mosaic woodland feeding ecology, rather than a closed canopy forest feeding ecology. The transition to bipedality might be expected to be accompanied by changes to the teeth reflecting the incoporation of woodland mosaic food sources, especially rhizomes and bulbs.

The only animal to show some morphological characteristics of a transition to bipedality (so far) is Ardipithecus ramidus. These include the position of the foramen magnum, and a distictive toe bone (contentious). The canines seem to lack honing, and are reduced - smaller than chimpanzees large canines, but larger than A. afarenis' reduced canines [r], hinting as diminishing importance of canines as weapons with increasing bipedality and weapon weilding dexterity. This 'small chimp'-like animal existed about 5.8 million years ago and for at least the next 1.5 million years in what has been interpreted as wet, thickly wooded tropical forest in middle Awash region of Ethiopia. Its teeth could be interpreted as those of an animal in a feeding ecology that was transitional between semi-deciduous tropical forest and woodland mosaic (semi-deciduous tropical forest is more probable, on basis of above equatorial latitude, and estimated 400 metres above sea level altitude).

Are there teeth intermediate between animals traversing two feeding ecologies that have similar (and sometimes identical) food species, but with pivotal points of difference? Using chimpanzee morphology as proxy model of the feeding adaptations of the ancestral ape in a tropical evergreen forest and A. afarensis morphology as the proxy model of the feeding adaptations of the ancestral ape in a complex woodland mosaic, there are these comparative differences-
1. The incisors of Ardipithecus are smaller than chimpanzee, but larger than A. afarenis. Smaller incisors have been hypothesized as co-related with smaller fruit needing less peeling with the incisors. Given that fruit range from larger than orange to small berry in both feeding ecologies, this is possible, but unlikely. Larger incisors are likely to have more biomechanical force applied to them if the front teeth are used as a vice, perhaps when stripping skin and muscle from the carcasses of small fossorial animals. A reduction in tooth size reduces the leverage on the incisors, especially when the buttressing effect of the canines is also reduced. The smaller incisors of Ardipithecus may simply be more typical of ancestoral ape, with the larger incisors of chimpanzee (atypical of Miocene apes, which also had relatively small incisors) a recent characteristic. However, the difference in size of incisors between chimpanzee, Ardipthecus, Australopithecus and Homo is not so great that it would (biomechanically) prevent any of these apes eating any of the foods within their feeding ecologies combined.
2. The enamel of the Ardipithecus premolar teeth is thicker than modern apes, but thinner than the teeth of A. afarensis; thus intermediate between evergreen forest chimp (as a model of an ancestral hominoid in that feeding ecology) and later large toothed arguably woodland hominids. Therefore, thick enamel suggests an adaptation to coping with heavy wear (perhaps increased consumption of rhizomes or other plant storage organs), or a long life history (or a little of both).

Competant bipedal gait, implying a woodland mosaic ape, could be inferred from about 3.9 million years ago, in Australopithecus afarensis (i.e using both the shortest reasonable time to speciation of 300,000 years and either A. anamensis or Kenyanthropus as an ancestor).

At the other extreme, at 6 million years antiquity, and using a 2.8 million year speciation interval, Orrorin tugenensis (granted publicly released details only 'suggest' substantial bipedalism) or Sahelanthropus tchadensis, as evidence of a bipedal hominind, an omnivorous woodland bipedal ape can be fairly argued to have existed about 9 million years ago.

Either case is conjecture; there is no evidence for the speciation interval of either ape, and the degree of bipedality of the oldest apes (Orrorin and Sahelanthropus) is in dispute, as is the idea they are ancestral to humans.

The most parsimonious choice of a human ancestor to place in a woodland mosaic feeding ecology (so far) is Australopithecus afarensis and/or A. africanus.

Australopithecus afarensis was heavier than the chimpanzee (male A. afarensis weight is estimated at 70 kg [154 lbs], females at 28.9 kg [64 lbs]; chimpanzee male is about 48 kg and female 42 kg [r]). The molars are large and flat, and would likely be useful for crushing the fibrous rhizomes of marsh plants such as Typha in order to extract the starch. Their teeth show signs of polishing by fine particles, which some researchers suggest might be due to the minute 'opal phytoliths' present in sedges of the family Cyperaceae, and in grasses [r]. Certainly, some Cyperus species have small edible bulbs. The mud they grow in might also be expected to have a polishing effect. The relatively thick stems of some Sorghum species have a somewhat sweet sap [r], as does the central pith of elephant grass, Pennisetum purpureum (free carbohydrates are fructose, glucose and sucrose)[r]. Enamel is thick, possibly to withstand abrasion from fibres in roots and rhizhomes, as well as adhering mud. The markedly large and slightly horizontally inclined fruit scooping incisors typical of a primarily frugivorous ape are reduced and slightly more vertically placed. In all, they are not too different from chimpanzee or human teeth in their functional aspect, except that the molars could be interpreted as adapted for a greater pith, bulb, root, and rhizome component in the diet. To this extent (and discounting enamel thickness), it could be argued they are at least as intermediate between a tropical forest feeding ecology, and a tool assisted tree seed and bulb based feeding ecology as Ardipithecus.

Based on only a few available features - its domed skull and possibly associated 'longer' leg bones - plus the need to show progress toward human form prior to Homo erectus at about 1.7 million years ago, it is arguable that fossil KNMER 1470,  Australopithecus/Homo rudolfensis, might be the first animal grading into the human side. The earliest possible date for the unequivocal fossils of this species 2.4 million years ago. How long the species had been in existance before this date isn't known, although the recently discovered Kenyanthropus platyops, living at and prior to 3.5 mya, could plausibly be an intermediate form [r]. What is significant is that the animal emerges at or soon after the marked aridification of East Africa in the period 2.8 million years ago to 2.5 million years ago. These drier conditions mean food sources are more widely dispersed. Water sources will become more unreliable and require migration to permanant water in the dry season. Fruiting trees also become more seasonal, but, paradoxically, the aridity challenges plants to use storage organs to concentrate carbohydrates in preparation for fast growth and reproduction when the rains come. This combination of patchiness of resources and decade to decade unreliablity exerts a strong selective pressure for decadal length memory, implying larger brain, and longer legs. Both are probably evidenced in Australopithecus'Homo rudolfensis.

There may be other plausible candidates in the fossil record; or a better candidate may be yet to be found, but the strands of body form, habitat and climate suggest A/H. rudolfensis would be at least representative of the body form, time and place in which such a lineage evolved - even if we can't be certain that it is directly ancestral. If not ancestral, it might fairly be used as a 'model' and 'time marker' of an ancestor on the line to Homo.

The molars are smaller, suggesting that either the animal had regionally evolved in a feeding ecology where the plant calorie sources were less fibrous (i.e. the less fibrous corms of Southern Africa), or tools were regularly used to beat out and collect some starches, and to grind some hard tree seeds, or both. There may be The most marked feature is a near doubling in cranial capacity. There might be many reasons for this, but complexity of food processing behaviour increases the available food in any feeding ecology. While speculative, there is some suggestion that several fossil femora unassociated with cranial remains might be those of A. rudolfensis. If so, the male animal was 172 cms (5' 8"), within the human range. This suggests that the animal had evolved to range quite widely; an essential prerequisite of highly seasonal arid shrubland.

Feeding ecologies in Africa

The evergreen tropical rainforest biome

Most of Africas evergreen rainforest (74%) is in central Africa - comprised in large part by Zaire's Congo basin - as far as Eastern Uganda. Above 1000 metres it grades into more montane/cloud forest in the uplands of eastern Zaire, Burundi and Rwanda. West African tropical rainforest, with its own typical plants and animals, comprises most of the rest (19%). The West African tropical rainforest is today present on the Atlantic Ocean coastal region in West Africa, 5 degrees above the equator, along the Gulf of Guinea coast (with a short 300 kilometre 'savvanah gap') to the Sanaga river, Cameroon.

Africa's rainforests today are 'drier' than most Asian and South American rainforests, receiving between 1600 and 2000 mm of rainfall per year, with rain in a given month only rarely exceeding 100mm (4 inches) (White, 1983). Coastal areas are wetter, as are the uplands of the eastern Congo-Zaire basin.

The wetter equatorial areas have a dry period shorter than two months. As a result, the mixed evergreen and semi-deciduous forest type doesn't display particular seasonality. The high water storage capacity of the soils means there is never a soil water deficit, meaning lianes, in particular, can be abundant. The upper canopy trees, in particular, sporadically shed their leaves before immediately recommencing growth, but not in a seasonal pattern. This forest usually has greater floristic species diversity and more leguminous trees (White, 1983).

Africas rain forest flora has less diversity than rain forests on other continents. This is usually explained as being a result of the progressive aridity since the Miocene, and the severe dry periods during the Quaternary. In the severe ice age of 18,000 years ago, the climate cooled and dried out to the extent that the shrinking forest only remained in a few refugia, resulting in the extinction of significant amounts of the (assumed) pre-existing diversity (Hamilton, 1982). Nevertheless, wet tropic rainforests are remarkable for their floral species diversity.

The warm wet tropical rain forests of equatorial Africa are relatively barren places for a ground living bipedal omnivore. The huge forest giants form a light trapping canopy far above (15 to 30 metres, with some giants penetrating the canopy and growing as high as 40 metres -130 feet). Their shade prevents the growth of anything in the dark and very humid interior except a few spindly saplings, and slow growing shade tolerant shrubs and broad leafed (often velvety or variagated) herbaceous plants mostly in the families Gesneriaceae, Melastomataceae, and Commelinaceae. Only where the leafy canopy is breached by the course of a river, or by a forest giant falling, is there the dense impenetrable tangle we associate with 'the jungle'.

1. The temporally distant tropical rainforest ape: foundation of human dietary evolution

It is universally assumed that evolution of primated towards apes started  with small (1 kg or so) arboreal quadripedal animals such as Catopithecus sp. that probably ate both fruit and insect. These fairly widespread animals evolved in Africa when it was an island continent, not yet in contact with Eurasia. At around 25 million years ago, the smaller 'monkey line' and the larger 'ape' line started to diverge from each other. There was a proliferation of apes in Africa prior to and after Africa connecting to Eurasia at the Arabian Peninsula at about 18 million years ago. They are generally described as either frugivores or folivores, and arboreal; with arguments as to whether they showed suspensory behaviour, or were quadrapeds.

By 18 million years ago, Equatorius, a somewhat primitive featured ape considered to live in dry open woodland, appears in the fossil record in East Africa. It would likely be in many different locations, but sampling biase must leave it in East Africa where fossils are accessible. The similar but more 'modern' Kenyapithecus appears around 14 million years ago. By 17 mya apes have spread 'out of Africa' into Europe and Asia. Sivapithecus (Ramapithecus) seems one of the earliest emigrants, and is suggested to have been a relatively large (around 40 kgs or so) quadripedal frugivore that included nuts and seeds in its diet. It was relatively successful, persisting until at least 7 million years ago, and might possibly be ancestral to the orangutan.

A contemporaneous group, the Dryopithicines, are recorded in Europe from about 16 million years ago to around 12 million years ago, and some believe they are ancestral to the human line as they have the most 'hominid-like' features of the Miocene apes. Oreopithecus bambolii, recorded in Europe about 11-7 million years ago, has limbs and a human-like pelvis that are intepreted by some to show it was bipedal on the ground, although still adapted for climbing. Some argue that this line is ancestral to the human lineage. Other apes living about the same time, such as Ouranoptithecus, are said to have eaten 'gritty' food, perhaps tubers and nuts, and have similar facial charactersitics to both modern apes and humans.

Whether one or any are ancestral to humans is not so important. What it does show is that from about 18 million years ago until the permissive range of the last common ancestor of chimps and humans - 10 to 3.8 million years ago - apes dispersed widely, and lived in a variety of quite different ecological niches - from semi-deciduous tropical forest, warm temperate forest, swamp forest to drier woodlands. Drawing specific dietary preferences from fossil teeth is always tentative, but interpretations include folivores, frugivores, mixed frugivore-seed eaters, nut and tuber eaters and so on.

Clearly, a primate ancestor, - frugivorous, folivorous, quadripedal forest floor scavenger - at some point abandonded 'comfortable folivo-frugivory' in a tropical to sub tropical evergreen and semi-deciduous forest environment, whether Asia or Africa, early or late, and moved into an unexploited, or underexploited, niche as a generalist feeder, primarily on the ground.

But ground living is unlikely in an evergreen forested environment - unless there is substantial folivory, in its broadest sense.

The seasonless evergreen tropical rainforest biome may never have been the 'primordial' habitat for the common ancestor of human and extant African apes. It may have simply been one of several forest habitats able to be exploited by various branches of the ape family. But it is so entrenched in our imagination it is a good starting point

A large bodied foli-frugivore in a tropical evergreen rainforest environment
Most wet tropical rainforest trees disperse their seeds by offering fleshy fruits for bats, fruit eating birds, monkeys, and apes to pick direct from the tree and carry away, to be deposited at some distance pre-packaged with fertiliser. Fallen overripe fruit and fruit dropped by monkeys are eaten by small forest antelope, pigs and forest elephants. As there are no clear seasons, the fruiting trees such as figs ripen sporadically throughout the year. A Ficus tree may ripen it's fruit over 7 to 8 days. When the tree finishes fruiting, fruit eating animals must search out the next tree that is coming into season. As around 90% the trees of the mature tropical rainforest use fruit to disperse their seed, and seasonality is slight, fruit is a major food resource in a tropical rainforest - for those able to reach it high up in the canopy. Suffice to say that fruit of at least 23 plant families are eaten by bonobo species of chimpanzee, Pan paniscus, at Wamba, in the equatorial Congo basin.[n]. Unsuprisingly, bonobo feeding ecology centres on ripe fruit, which comprises about 73% of their diet.[r]

The Wamba chimpanzees eat fruit in genera such as Diospyros, Garcinia, Mammea, Ficus, and Chrysophyllum, genera which include domesticated species, but most genera - and many of the fruit - are unique to this environment and don't have domesticated analogues. One of the most important fruits for Bonobo chimpanzees, at least, are the fruits of the 8 or so species of the tropical liane Landolphia (Wamba location) - although a total of 18 genera of lianes and vines have fruit recorded as eaten by bonobo chimpanzee. (Most, but not all, fruiting lianes are confined to the wet tropics, as an extended period of moisture deficit is inimical to water transport in a long stem. Lianes are therefore a handy marker of a moist forest environment without long periods of dryness).

 In Africa, fruit as a primary food source has not led to the evolution of a very large bodied essentially frugivorous ape. In contrast with Southeast Asia, where the large bodied orangutan (male Bornean orangutans may weigh as much as 115kgs, and are so heavy they have to walk from fruiting tree to fruiting tree) is primarily a frugivore/seedivore. As a result, the large bodied Bornean orangutan needs a relatively big territory to encompass enough fruiting trees to maintain its large size, although Bornean orangs are able to exploit the energy dense termites living in rotting logs on the forest floor. There is not usually enough fruit on a tree to support much more than one large bodied ape, and as a consequence, a  large bodied frugivorous ape is likely to be essentially solitary. As the orangutan is. The large size of the Bornean orang may be an artifact of isolation on an island without large predators. Sumatran orangs share their island with the Sumatran tiger, are consequently smaller, and rarely descend to the ground. It may be that the absence of a large bodied predatory cat 'allowed' a more varied feeding ecology for chimpanzees. There have been only one record of lions entering semi-deciduous forest and killing chimpanzees [r].

Even in the wet tropical forest fruit is in short supply in the short dry period and may be restricted in some 'abnormally' dry years. The less preferred fall back food is foliage; foliage in the whole plant sense. The young leaves of the forest giants are relatively palatable, have the least amount of deterrent tannins and other anti-feeding devices, and have the greatest amount of protein. But for an animal such as chimpanzee, which spends large amounts of time on the ground, the succulent leaf petioles, shoots and pithy stems of ground living herbaceous plants are more important. The herbaceous plants of the family Marantaceae are important, as are pithy leaf stemmed members of the ginger family, such as Aframomum sp and Renealmia sp.[r], often growing along stream edges and in gaps in the forest. These pithy, succulent stems are key to stepping across the short drier season.

In the relatively aseasonal environment of the wet tropical forest there is no need for underground plant storage organs beyond perhaps a spreading rhizomatous habit. The vine Dioscorea have an underground storage organ, and is present in the Congo basin (Dioscorea bulbifera, D. dumetorum, D. preussii, D. smilacifolia) as well as others in West African rainforest ( D. abyssinica, D. praehensilis, D. liebrestsiana). But like the legumes, it contains defenses against being eaten - alkaloidal toxins. In addition, yams are quite deeply buried, and in the wetter tropics, don't make as large a tuber as in more seasonal forest.

Nutrient dense tree seeds noted as eaten in the Congo rainforest at Wamba by the Bonobo species of chimpanzee are almost all members of the Leguminosae family - 3 species of Anthonota , Baphia species, Berlinia grandiflora, Brachystegia laurentii, a species of Cynometra, 2 species of Gilbertiodendron, Leonardoxa romii, Monopetanathus microphyllus, Pentaclethra macrophylla, Scorophloeus zenkeri and Tessmannia africana. These leguminous seeds are typically 15mmm or so wide, flattened, and with a fairly leathery or hard outer seed coat. The seeds would be most easily chewed if harvested before they are mature; but the literature is silent on this point. Leguminous trees usually disperse their seeds by the violent bursting open of a long bean like pod. Pentaclethra macrophylla, for example, has a pod over 45cm long. The tree is about 36 metres high (120 feet), so the rather large flattened red seeds are likely to be scattered over quite a wide area when released from this height.

Typical of many leguminous plants, the seeds of these kinds of trees often contain bitter, pungent, or acrid compounds to dissuade seed predation. Chimpanzee physiology must deal with many plant phytochemicals designed to dissuade browsing; it might be that they have some physiological adaptation that allows eating these seeds without 'intestinal upset'. Certainly, there is much to gain if the defenses can be overcome - P. macrophylla has about 33% fats, and a high protein content [r]. This nutrient composition is possibly typical for these tropical leguminous trees. Similarly, at Bai Hok, Gorilla eat the single seed of the leguminous tree of the genus Dialium with no apparent ill effect, where indigenous people must first leach the seed in water for an hour before it can safely be eaten raw [r]. Gorillas at Bai Hok also eat the parchment winged Pterocarpus soyauxii seed. This genus has seeds regarded as liable to cause "vertigo and vomiting" unless cooked. There is some suggestion the anti predation factor is in the outer seed coat.[r]

Seeds of some of the marantaceous understory are also eaten (in the genera Ataenidia, Haumania, Hypselodelphys, Marantochloa, Megaphrynium, Sarcophrynium, and Trachyphrynium). The only other tree seed recorded as being eaten is a member of the Meliaceae, Entandrophrogma angolense. The nutrient dense seed of Ricinodendron heudelotti is also present, although Bonobo chimpanzees do not use tools, and the kernel, contained in a hard nut, is unavailable to them. This is not the only nutrient rich tree seed at Wamba unavailable due to a hard shell. The inedible fruit of Panda oleosa contains a single large stony endocarp which encloses 3 nutty flavored oily seeds.

In the wet forest there is effectively never a soil water deficit, and where both trees and terrestrial herbaceous vegetation with palatable foliage are abundant, such as in some of the seasonally flooded swamp forests and undisturbed primary forests of the Congo basin, living is relatively easy. But a relatively small gutted ape such as the chimpanzee can't rely entirely on foliage, buds and piths, and must eat some energy dense fruit. This in turn imposes an upper limit on body size. Too large an ape can't reach all the fruit high up in the forest canopy, especially as many trees are both tall and with straight, unbranching boles for the lower two thirds. Lianes scrambling for a place in the sunny canopy 30 metres (100 feet) or more above are the 'rope ladders', but increasing body weight brings, to degree, more risk, limiting fruit foraging to robust trees and robust branchs. Young chimps sometimes have to use whippy young saplings as 'ladders' to climb trees with wide boles their arms can't encompass (some gatherer hunters in tropical forests use the same technique). Larger bodied apes in this feeding ecology, such as the Western gorilla and the Eastern gorilla, apparently [n] eat as much fruit as is available within the constraints of their feeding ecology.

Relatively large bodied primates exploiting both the abundant fruit high in the trees and the pithy herbs on the forest floor - specifically the chimpanzee - have immensely strong upper bodies and shorter hind limbs to make vertical climbing more efficient. This vertical orientation allows suspensory feeding in the canopy, and a more 'erect posture' in the trees so long as their weight is supported by their forelimbs. On the ground, the food resources are more or less immediately at hand, and awkward support of the forwardly inclined overbalanced body mass with the forelimbs (knucklewalking) is perfectly adequate for ground living in a forest. A seasonless forest feeding ecology could probably never permit selection of a fully erect and therefore bipedal body form; although some suggest suspensory behaviour might be a precursor to a bipedal animal.[r]

The largest bodied primate, the Gorilla has fully exploited the plant resources available at and near ground level in the stable wet forest environment. It has reduced its reliance on fruit, and become an ape exploiting a largely folivorus niche, and with a large gut to deal with large amounts of  vegetation. The downside is that the niche is restricted to areas of constant rainfall and fresh ground accessible foliage, such as slow growing herbaceous plants in the family Marantaceae ( e.g. genera Ataenidia and Haumania) and Commelinnaceae (e.g. Palisota sp). In higher altitude forest, nettles, Urera sp. and the buds and shoots of bamboo, Arundinaria alpina, are important. And it means constant feeding, and periods of laying about digesting the bulky plant material. Feeding on some fruit and seeds, even where the bulk of the diet comprises softer plant parts such as piths, soft petioles, flowers, buds, shoots and softer leaves - rather than coarse leaves or fibrous and abrasive grasses - has meant that the Gorilla gut hasn't needed to become specialised. Grazing herbivorous animals generally need a multiple chambered stomach to partly ferment coarse plant material as part of the digestion process. The only concession to extracting the maximum nutrient from foliage is a penchant for mountain gorillas to sometimes eat their own faeces (as do some other herbivores without multiple chambered stomachs, rabbits, for example).

Both the range of suitable environments within the seasonless evergreen forest and the range of food plants eaten is smaller than used by chimpanzee, thus limiting the possible natural distribution. A consequence of leaves and stems being constantly available and in patches of large mass, is that gorillas can move in social 'herds', or rather, 'harems' dominated by a single male. As food resources are neither patchy nor diverse, there is no need for the group to split up.

In contrast, chimpanzee use more plants parts, and more species. Because of its relative dietary plasticity and small size, they can live in social groups, but the size and disperal behaviour of the group, and of individuals within the group, is much more constrained by the feeding ecology. Because its feeding niche is 'variety', it moves from foliage to pith to flowerbud to seed to fruit according to the season and availability, and groups break up and recoalesce according to resource availablity.

The wet tropical forest is sufficiently diverse, stable and constant that an agile chimpanzee sized ape (in contrast with gorilla size primarily folivorous ape), partially bipedal or not, can live within the better part of the forests extent, largely on fruits of lianes and large forest trees, succulent plant parts on the forest floor, and the leaves and seeds of some forest trees - including the large leguminous trees that are more common in the evergreen wet tropical forest.

But not all rainforest is evergreen. Some forest has pronounced dry periods, and the trees display a kind of deciduous behaviour. This is not quite as constant and reliable an environment. Whether this 'seasonal' tropical forest or whether the wet evergreen tropical forest was the original environment in which the move to the last common ancestor of chimpanzee began is moot. But the challenges, even muted challenges, of seasonality means it is a selective milieu. And the evergreen wet tropical forest grades into the seasonal tropical forest, and even subtle behavioural responses to the foraging possibilities of adjacent ecologic niches can ultimately end in gradual evolution of body forms unlikely to be be reversed. Changes such as bipedal locomotion.

The place of the evergreen tropical forest in human evolution
This environment is a stable and rich 'spawning' ground for feeding behaviours that ultimately allow extension of the feeding range. Some of the genera of fruits here appear not only in semideciduous tropical forests further south and east, but also in southern African well watered coastal forests and woodlands.

The seeds of forest legumes must often be peeled of the seed coat to allow the nutrient dense endosperm to be eaten. These tropical legumes are much fewer further south, but the same technique can be used to render the nutrionally dense seeds of a woodland species safe.

Oily tree seeds protected by very hard seed coats are available to the animal with the technology and techniques to release them. Such techniques are needed to release the nutrients locked in Sclerocarya seeds in the seasonally dry woodlands. It is interesting to note that at Wamba, at least, the local chimpanzee culture has not learned to open the two local hard seeded trees. A seasonless tropical forest feeding ecology is clearly rich enough that the ability to exploit hard shelled tree seeds for their dense nutrients is unecessary. It may be that cultural transmission of such techniques ('memetic inheritance') only becomes selective in seasonally constrained environments.

The evergreen tropical environment is generally recognised as the environment of evolutionary adaptation that resulted in brachiation - using the forelimbs to swing from branch to branch. Obviously, this means of locomotion could only evolve in a continuous canopy. But brachiation brought with is a suite of morphological changes to the spine, pelvis, and forelimbs of apes that 'preconditioned' them to semi-upright knucklewalking, and ultimately, fully upright locomotion.
 feeding behaviour etc
"Female chimpanzees often forage alone and do not display obvious linear dominance hierarchies; consequently, it has been suggested that dominance is not of great importance to them. However, using data from a 35 year field study of chimpanzees, we show that high ranking females have significantly higher infant survival, faster maturing daughters, and more rapid production of young. Given the foraging behavior of chimpanzees, high rank probably influences reproductive success by helping females establish and maintain access to good foraging areas rather than by sparing them stress from aggression.

In many species of group-living mammals, especially those that feed on monopolizable foods such as spotted hyenas and many primates, females have frequent dominance interactions and are ranked in stable linear hierarchies (1-3). These hierarchies result from, and are maintained by, a pattern of alliances in which close relatives support each other against more distant relatives and high-ranking matrilines support each other against low-ranking matrilines (4). In the majority of studies, high rank is associated with higher reproductive success (2, 5), although this relationship is often weak, perhaps because of counter-balancing costs of high rank (6). There is debate over the extent to which the effects of rank on reproductive success are due to better access to food for high ranking individuals, or to protection from stress that results from aggression towards individuals at the bottom of the hierarchy (6,7).

Chimpanzees resemble these species in that they live in permanent social groups, and feed predominantly on ripe fruit that often occurs in monopolizable patches (8). However, they differ in that females in some populations spend more than half their time feeding alone, and most females disperse to other groups before breeding, with the result that they are usually not surrounded by relatives (8). Compared to female macaques and baboons, it is difficult to detect linear hierarchies among female chimpanzees. While some female chimpanzees clearly dominate others (9-12), dominance behavior in stable groups or stable pairs of females is uncommon, and is never observed between some dyads (9,12). In addition, when aggressive behavior does occur within a dyad, it is sometimes two-sided with no clear winner (13). These observations have led some to believe that female dominance is unimportant for reproductive success (12, 14). However, others have suggested that dominant females may gain advantages (9, 15). Here we present data from the 35 year study of the chimpanzees of Gombe National Park, Tanzania.

The chimpanzees of Gombe have been studied since 1960 (9). The 48.7 km2 park consists of series of steep valleys running from the eastern rift escarpment (1600m elevation) to lake Tanganyika (775m). The valley bottoms contain evergreen forest which gives way to semi-deciduous forest on the valley sides and grassland on the ridges (16). Since 1963, the chimpanzees of the central area of the park have been provisioned with bananas at an artificial feeding station in order to habituate them and to facilitate regular observation (9). The feeding station has been likened to an unusually long-term natural food source, and, since 1970, has been estimated to provide less than 2% of the chimpanzees' diet (17, 18). Daily observations are made of the presence, reproductive state and social interactions of individuals at the feeding station, and, since 1975, during daily all-day follows of individuals throughout their range (9). Since 1970, the habituated community has consisted of 4 to 13 adult males, 10 to 18 adult females, and 18 to 31 immatures, and has occupied a range of 6.75 km2 to 14.5 km2 spanning 3 to 6 main valleys in the middle of the park (9,19). Adult females spend about 65% of their time alone with only their dependent offspring, foraging in distinct but overlapping core areas of about 2 km2 (18, 20), while adult males are more social, travel over the whole community range, and jointly patrol and defend its borders (9, 18). While almost all males born in the community remain in the community as adults, most or all natal females visit other communities during adolescence, and about 50% emigrate permanently (21).

We assessed dominance relationships among females by examining the direction of all pant-grunts between females recorded from 1970 to 1992. Pant-grunts are the most reliable measure of submission in chimpanzees and correlate with the reception of aggressive behavior (12, 13). When we constructed dyadic matrices, many cells were empty, but by assessing two year blocks we were able to assign 88% of the females that were observed more than 10 days per year (22) as high, middle, or low ranking in each block (23). Dominance rank was not related to body weight (R2=0.01, n=15)(24,25), but individual dominance rank increased with age (26) as observed in the chimpanzees of the Mahale Mountains (10). However, a female's rank at age 21
strongly predicts her rank a decade later (R2=0.80, p=0.001, n=9) suggesting that early rank acquisition is important.

Dominance rank has a striking effect on several measures of reproductive success. First, offspring survival was significantly
related to mother's rank at the birth of her offspring. Infants of low ranking females showed much higher mortality than those of high ranking females over the first 7 years of life (Fig. 1). Second, the age that daughters reached sexual maturity was significantly related to their mother's dominance rank (Fig. 2). Daughters of low ranking females experienced their first full anogenital swelling during which adult males mated with them (27) as much as four years later than daughters of high ranking females. Age at first full swelling is strongly correlated with age at first birth (R2=0.67, p=0.01, N=8 regularly observed females). Third, there was a tendency for high ranking females to live longer (28). Finally, the annual production of offspring surviving to weaning age (5 years) is correlated with rank for mature females that were observed for at least 12 years (Fig. 3), indicating that low ranking females are unable to compensate for the higher mortality of their offspring by reproducing more quickly. All these factors combine to produce higher lifetime reproductive success in females of higher rank (29). In the analyses concerning reproduction, we excluded the female GG, because she was sterile. GG was an aggressive, masculine-looking female who occupied the highest rank and cycled regularly for 28 years but never became pregnant. If she is included in the analyses, the relationships of the rate of production of surviving infants and life-time reproductive success with rank cease to be significant (Fig. 4, 29).

Dominance probably exerts its effects in several ways. First, the high mortality of infants of low-ranking females in the first few
months of life was partly due to the infanticidal behavior of the high ranking female, PS, and her daughter, PM, who snatched and ate the infants of several females in the 1970s (9,30). Since then, high ranking females FF and GG were observed trying to snatch the newborn infant of middle ranking female GM, and females of unknown rank in an adjacent community were seen eating another female's infant (19). These observations suggest that female infanticide may be a significant, if sporadic, threat, rather than the pathological behavior of one female. However, infants are vulnerable to infanticide for only a few weeks, and rank-related effects on offspring mortality continue well beyond that age (Fig. 1). Second, the younger age at which daughters of high ranking females reach sexual maturity reflects their higher rates of weight gain (19), suggesting that high ranking females have better nutrition. Better nutrition might also account for better survival of high ranking females and their offspring. In species living in permanent groups, reduced reproductive success of low ranking females has been attributed to chronic stress due to frequent aggression from other females (7). Because female chimpanzees spend so much time alone, often going for a day or longer without seeing another female, this is unlikely to be important in this species.

High rank may confer better access to food, both by enabling a female to acquire and maintain a core area of high quality and by affording her priority of access to food in overlap areas (31). Because of the mosaic distribution of vegetation at Gombe, some female core areas are likely to contain higher quality food than others. In addition, because core areas overlap almost completely, a high ranking female may gain priority of access to preferred food sites in an overlap area. These modes of competition might explain why dominance behavior is less frequent in this species than some others and linear dominance hierarchies are hard to detect. If core areas are stable, competition is likely to be most intense when new or maturing females are attempting to establish their own core areas (15). This idea is
consistent with the frequent observations of aggressive interactions from resident females to newly immigrant females at Gombe (32) and the Mahale mountains (10), the fact that PS and PM killed the infants of neighboring females as PM reached maturity (30, 33), and observations of more frequent dominance interactions during the establishment of a new female group in captivity (15). In addition, clear dominance relationships may only be established between females whose core areas overlap, thus explaining the general lack of clearly defined linear dominance hierarchies. The stability of core areas, the acquisition of core areas, and the relationships of females whose core areas overlap are the focus of current research at Gombe.

More research is needed to understand how female chimpanzees achieve high rank. Some females, such as GG, acquired high rank by their own aggressive behavior. Other females, such as PM, have gained high rank through their mother's support. The fact that alliances with kin are sometimes important makes it all the more striking that young female
chimpanzees often disperse to other communities. This underscores the suggestion that they are 'forced' to do so in order to avoid inbreeding because their male relatives do not disperse, perhaps because of even stronger advantages to males from cooperation with relatives (34, 35). More information on the relative importance of alliances in the acquisition of female dominance rank and the influence of dominance on reproductive success in other populations of chimpanzees might clarify why only half the females at Gombe disperse, while almost 100% do so in other populations (8). Finally, if the considerable degree of reproductive skew observed in the Gombe chimpanzees also occurs in other populations, this has implications for the future genetic diversity of this endangered species. As populations become small and isolated, there is a greater chance for the genetic diversity of the population to be reduced by the successful reproduction of a few dominant individuals (36)."

The seasonal tropical forest biome

Among speculative scenarios, the suggestion that seasonal tropical and subtropical forest was the crucible of hominid evolution is plausible simply because the evironment is unstable. The behaviour patterns chimpanzees use in this environment today to 'step through' hard times - using tools to release previously unavailable 'nut' kernels, increased insect and other small animal eating, ranging behaviours along watercourses  -  are foundation behaviour patterns for carrying to even more resource patchy woodland environments. (The food species in this biome are considered in the next section in the context of the resident apes, so won't be listed here.)

In some parts of (chiefly) the West African tropics there is a period in winter when dry air masses dominate, generally for around three months. Rainfall is less than less than about 100mmm, the soil water reserves can't always keep up with transpiration and the trees respond by slowly  defoliating and resprouting new growth before the rainier weather pattern returns . This semi-deciduous adaptation allows greater ability to survive abnormally dry years and decades. The seasonal pattern of riverine gallery forests in the drier parts of Southern and Eastern Africa are probably similar, although species composition will likely be different (and a few gallery forest species provide a 'lead in' to drier woodlands - the woody scrambling fruiting climber Adeinia gummifera, for example, is found in both ecological zones.)  Eastern coastal and gallery forests penetrating dry areas might speculatively have been 'corridors' for seasonal forest dwelling hominids moving into Southern Africa, but carried the risk of becoming isolated 'islands' as the continent went through drying phases.

Quite marked changes in flora can happen even within relatively short periods. For example, the Lake Victoria basin, surrounded by seasonal semideciduous tropical forest, there is a climatically stable zone about 40 kilometres wide where the temperatures remain between  27º C (80º F) and 16º C (60º F), with a high rainfall that is well distributed throughout the year. But in a period of about 2,000 year after 6,500 B.P. it was wetter than today, and the forest extended. Then, over about a 1,000 year period it became increasingly dry, the semi deciduous forest shrank and finally gave way to open vegetation, characterised by plants of the family Capparidaceae (dry condition adapted shrubs and trees, e.g. Boscia sp) and grasses. The climate then ameliorated for the next 600 years or so, and some limited semi deciduous forest returned. At higher altitudes, dry montane Podocarpus/Juniperus procera forest also extended its range. Dry conditions returned, and montane and semideciduous forested areas shrank again in favor of open woodlands and savvanah area. The magnitude of change was muted near the lake margin, presumably due to its localised climate ameliorating effects.[r]

These sorts of marked shifts in plant communities in response to climate (with patchily distributed shrinking local 'refugia') probably happened with relatively high frequency over the course of human evolution. The fossil and pollen record is not usually able to distinguish such small scale, millenial changes. We therefore probably underestimate the selective pressures on apes living in a seasonal tropical forest environment - especially the populations living at the cusp of climatically driven vegetation change.

Seasonal subtropical forest
As noted above, the east African seasonal subtropical forest and adjacent woodland and duneforest ecozones - for example, that of broadly coastal Southern Mozambique - is not dissimilar to seasonal equatorial forest. Again, there are numerous fruiting trees, such as Strychnos madagascariensis, S. spinosa, Sclerocarya birrea, Hyphaene coriacea, Phoenix reclinata, Ficus sycomorus, Ximenia caffra, Boscia albitrunca, Diallium schlecteri, Trichilia emetica, Antidesma venosum, Ziziphus mucronata, Grewia  sp., Garcinia livingistonei, Dovyalis longispina, Syzygium cordatum, Mimusops spp., Manilkara discolor, Landolphia kirki, Tabernaemontana elegans, Vangueria infausta, Coffea racemosa. [r]

Tree seeds are seasonal, with the difficult to extract seeds of the fruit of Sclerocarya birrea perhaps being the primary nutrionally dense resource. Such subtropical forest may have extended almost to the Cape (where it might be more correctly designated warm temperate), and inland, particularly along valleys, in moister time intervals.

The place of the seasonal tropical forest environment in human evolution
Frugivorous and folivorous animals find it much more difficult to find food in these kinds of somewhat seasonal forests. Chimpanzees lose condition, and take recourse to forest floor herbacous plants, especially in the dry season. If conditions are bad they may travel quite long distances to find fruiting trees and edible terrestrial plants. (Some groups may even venture out into the woodland savannah to search for food.) A key adaptation to seasonality is the ability to store fat. Much of the range of the frugivorous orangutan is seasonal tropical forest. In the month of highest fruit production fruit constitutes 100% of its diet, with a daily energy intake for female orangutans of 7404 kcal/day. The excess calories are stored as body fat. Later in the year much less fruit is available, comprising only 21% of the diet, and with animals resorting to eating the inner bark of trees. At this time energy intake for females is only 1793kcal/day, and research has demonstrated that the animals draw on the reserves of fat laid down in the peak fruiting period.

It is very curious that chimpanzees in seasonal tropical forests are not reported as storing body fat. While a speculative idea, it may be that there are physiological differences in populations that sometimes resort to a savvanah woodland  resources in times of extended dry (e.g. Tanzanian chimpanzee populations at Mahale) and those that live in the evergreen tropical jungle in central Zaire.[r]

Regardless of season of ripening, fruits are still abundant. Richard Wranghams data on fruit observed being eaten by chimpanzees at Mahale, Kasakati and Belinge includes the fruit of 13 genera of liane (Uvaria, Ancylobotrys, Landolphia, Salacia, Canthium, Cissus, Ampelocissus, Toddalia, Glycine, Tinospora, Hypselodelphis, Antidesma, and Saba), 10 species of Ficus, and around 80 other fruiting plant genera!

The pattern of seasonal fruiting is such that even in the 'offseason', there is at least one species fruiting, and a frugivore can concentrate on that one resource, in conjunction with less preferred plant parts. Chimpanzees spend very approximately 75% of their feeding time on 'fruit and seed', about 15% of their time on leaves and pithy plant parts, and only about 5% of their time on other food categories such as small mammals, eggs, termites, and tree seeds in the sense of nuts (where available). [r]

Because the seasonal tropical forest has a period when only a few species of trees are fruiting, strong frugivory 'conditions' a hominoid to range 'horizontally' within a subset of the environment that contains these key fruit species, to exploit asynchrony of fruiting with temperature and moisture differentials that come with different altitudes, and to 'vertically' exploit fruit - from ground level shrubs to tall trees. The seasonal tropical forest contains fruiting genera present in woodland gallery forest and in woodlands themselves. Thus, the seasonal fruiting patterns, and the genera, form a natural 'lead in' to the feeding ecology of gallery forest in woodland environments. As importantly, seasonality can cause critical food shortage. This in turn rewards food switching, whether soft foliage, pith and rhizhomes, or tree seeds.

The seasonal tropical forest supports some trees which sequester calories in relatively large seeds, with the most energy dense seeds being oily seeds. At Mahale, Tanzania, Richard Wrangham has recorded chimpanzees eating a variety of tree seeds. Tree seeds available include Combretum binderanum, Julbernardia sp., various species of Brachystegia, Guibortiatessmanni, Bauhinia petersiana, Piliostigma thoningii, Afzelia sp. Acacia hockii, Parkia filicoidea, Pterocarpus tinctorius  and others. Not all potentially edible tree seeds have been identified. In the early 80's Nishida and Uehara recorded as being present at Mahale, but not eaten by chimpanzees, 4 species of Terminalia, a genus of widespread tropical tree many of which have edible kernels of fine flavor, but which are typically encased in a thick shell. The edibility of the Mahale species is unknown.

Some of these seeds have a thin edible aril surrounding them (Afzelia africana), or a thin dryish but nutritious pulp in the pod (Parkia filicoidea). In these cases, it is equivocal whether the seed is being crushed and eaten by the chimpanzees, or passing through undigested. In the case of Afzelia in particular, the immature seed is probably edible, but the hard mature seed may contain quite powerful phytotoxins [r].

Phytotoxins are much less likely to be present when the seed is protected from predation by a thick shell. Some very nutritious oily seeds are protected this way. Ricinodendron heudelotti ssp. africanum is a relative of the mongongo nut, Ricinodendron rautanenii. The nuts of this fairly widespread tree are eaten by some chimpanzee populations - and are also one of the major food sources of the Mbuti people, of the Ituri forest in Zaire R. rautanenii is a nutrient dense food - about 26% protein and 57% fats. It would be reasonable to assume R.  heudelotti has a similar analysis. It grows as far south as the forests of Angola, and shows some ability to withstand dry conditions, although less well adapted than its relative in the Southern Angolan drylands. The pulp covered seeds, flowers and pith of the oil palm, Elaeis guineensis, is eaten by chimpanzees in both West Africa and Tanzania. The fibrous pulp around the hard seed is quite oily in itself. The rich and oily kernels are enclosed within a hard shell, accessible only to chimpanzees that have learnt to use a hammer and anvil.

The cropping pattern of these trees with encased kernels is likely to be somewhat variable between species, and probably variable from year to year, depending on the pattern of the rainfall in that year. In the seasonal Malaysian forest, for example, Terminalia catappa defoliates twice a year, and also fruits twice a year [r]. It would be be reasonable to suppose African seasonal tropical forest nuts have a somewhat similar pattern. Just how long the seeds remain edible after falling is moot. Even in a humid environment, there is a good chance thick shelled tree seeds they will remain edible for weeks at least, assuming they don't germinate, and are not sequestered by rodents.

The key to sustainable life in a seasonal forest is food switching, especially in 'abnormally' dry years. This means knowing when a particular fruit is ripe, where patches of succulent leafed plants are, at what stage of ripeness a fruit is suitable to eat, at what stage of maturity to eat a potentially toxic tree seed, when a palm bud is emerging and so on. In some years, the ability to additionally switch to the most nutrient dense food the forest has to offer might make the difference between breeding success or poor conception due to low fat reserves in the female hominoid.

Some hard shelled tree seeds are particularly fatty, and often have a useful protein content. Ape teeth are not designed to open hard shelled tree seeds. But their potential value is high to an ape that is able to open them 'by hand', if it can develop the skill.

Little of the chimpanzee gross morphology seems designed for fine motor control. Chimpanzees are awkward bipeds, and their forelimbs -'arms' - are hugely powerful but their shoulders seem poorly adapted to accurate whole arm blows, or to accurate throwing. Even so, in one experiment the estimated terminal velocity of a stick weilded against a predator - in this case a stuffed leopard - was calculated to be sufficient to break the cats spine. In addition, their 'thumb' is low on the hand, and cannot oppose the fingers to create a 'tweezer' for precise manipulation of objects.

And yet lack of an opposable thumb does not negate precise control. With practice, they can learn to open nuts with precision. For chimpanzees, the problem of nut opening is demonstrably amenable to 'precision gripless' techniques which produce fine control of their hands and arms from a stable squatting position.

The fact that some [r] chimpanzees in seasonal tropical forests have learnt to open hard nuts with rocks and sticks is a powerful insight into our past. Chimpanzees in western Africa hammer open the 'nuts' of the oil palm Elaeis guineensis and the thick -shelled 'African walnut', Coula edulis, with a suitable sized stone, the nut having been previously placed either on a stone 'anvil' or on a suitable tree root. Adults and adolescents of both sexes open Coula nuts with a series of blows designed to open the nut but not smash the kernel once the thick shell is partially cracked. But it is virtually only the female chimpanzees who go through the complexities of opening Coula nuts while up the tree. Females carry both a wooden club and nuts gathered from the ground up the18 metre (60 foot) high trees and open the Coula nuts in the tree tops. They use the broad branches as the anvil, juggling the opened and unopened nuts from mouth to foot, while all the time being careful not to accidentally knock the club or rock hammer to the ground [r]. Coula trees in the coastal forests of equatorial west Africa mature their nuts over quite a long period - from December until April - making the payoff for ability in dextrous tool use, whether arboreal or terrestrial, very worthwhile - especially as the kernel is nutritionally dense, and, typical of many hard shelled nuts, without apparent anti-feeding chemicals. (According to Irvine, "nearly half the weight of the kernel is oil...[which] consists of 87 per cent oleic acid" [r].)

In one case, males have been observed bringing nuts to an old female for her to open for them. It is significant that for seeds of Panda oleosa, at least, it is virtually only the females (95%) have the patience and skill to break through the thick shell and hard spongy endocarp to retrieve the 3 or 4 oily seeds within. Only rocks are hard enough to open Panda nuts. If there are no suitable rocks nearby, the female may carry a suitable one several hundred metres back to the anvil. Panda nuts require a great deal of force to break the outer shell, but fine control to crack the inner shell without mashing the seeds within - a control which the chimpanzees demonstrate.

On several occaissions, when female chimpanzees used granite or quartizite rock to open Panda nuts, flakes of rock have sheared off [r].  The accidental creation of blades seems to be an inevitable by-product of the evolution of the seasonal forest - and woodland - 'tool assisted' feeding niche.

When fruiting is erratic (or fails) nutritionally dense resources pay a breeding or survival premium. The advantage then goes to -
i) the female ape (and her offspring) which has the best hand-eye coordination, given that results in a greater number of nuts per hour opened
ii) the male and/or female ape which cooperates in gathering tree seeds
iii) the ape which can use two hands in bringing nuts for opening
iv) for a forest dwelling agile climbing ape, the ape with the largest buccal cavity to carry nuts in
v) the female ape with the dexterity to access nutritionally dense tree seeds unavailable to males
iv) the male ape which can coerce or persuade the 'opener' to share the opened nuts
v) the ape which becomes skillful at marshalling and moving male and female 'collecters' and 'openers' from tree to tree - probably members of its own family.

A side effect of selection for using forelimbs to clutch and carry tree seeds (as noted, selection in chimpanzees, on the other hand, is for a large buccal cavity - once nuts on the ground are all eaten, they climb the tall trees to harvest more, and bring them back down to earth both in their mouth and in their one free hand [r]) and steadiness in  hand velocity control in hard tree seed opening might be an increasingly 'firmly rooted' upright stance. This coupled with better hand-eye coordination confers the ability of any adult, male or female, to quickly bludgeon to death, or seriously injure any member of its own group. Plausibly, group co-operation, female 'co-option' and male co-operation and self control within the brotherhood may have co-evolved. As Darwin pointed out, canines would no longer be the telling factor in male competition; a club is a 'great leveler' - in both senses.

Of course, such behavioural and morphological changes are only useful in a seasonally dry forest. Populations within the depths of the wet tropic forest had no such pressure to change to marked degree. The body changes concommitant with obligate bipedality might actually work against efficient access to fruit carried high in the forest canopy - especially as the lower two thirds of many tropical forest trees are unbranched.

The adaptation to being able to additionally exploit nutrient dense, hard shell encased tree seeds while retaining a 'pithy' plant food plus fruit base might have become the launching point for a broader omnivory, one anchored in tree seeds and tubers in an open woodland environment. In this model, the food availability gaps in the 'seasonal tropical forest' itself did not drive the evolution of an ability to control tools to exploit the few nutrient dense resources in that environment. After all, much less than 5% of chimpanzee feeding time is involved in opening tree seeds, and then only in those communities with that cultural 'meme'. Rather, a pre-existing non-obligatory but useful ability to access fatty tree seeds may have become a key part of a suite of related critical behavioural abilities conferring increased reproductive success in a woodland environment. Pre-formed fatty acids from oily tree seeds, converted to triglycerides and stored for fuel in adipose tissue, create an internal 'fall back' food for an animal in a erratically seasonal environment, and, as important, a 'fallback' nutrient store for partuition. The selective pressure is for cultural transmission of the tree seed opening technique across generations. The need to access fatty tree seeds only begins to becomes imperative rather than optional in the zone between seasonal tropical forest and woodland. Here, the annual learned behavioural cycle of tree seed opening (and its social implications) may ultimately have been ratcheted with accumulated subtle facilative genetic changes in brain organisation affecting behaviour and ability to finely control and co-ordinate hand, arm, and eye.

 But there no great need of finely controlled clubbing or throwing action for the agile seasonal forest ape when hunting, which for chimpanzees, is primarily a male occupation. The main prey of chimpanzees in dry seasonal forests are red colobus monkeys. These are almost all younger monkeys, from infant to adolescent. A treetop hunt, once started, is almost always successful. Hunting varies in intensity even between communities in close physical proximity. It varies according to time of year, and from year to year. Whether it is associated with times of year when there is a food shortage is equivocal. Capture of some young terrestrial animals, such as Duiker, may be purely opportunistic. In any case, even the highest meat intake in the peak of a prey hunting 'season' amounts to an average of no more than 500 grams per week (for an adult male), against from 35 to 49kgs of plant food (all types) eaten per week. And the 500 grams includes bone and skin, large in weight and low in nutritive value. A more typical meat intake is considered to be likely to be very much less than this amount - particularly for females, who are less likely to be involved in the hunt, or benefit from the spoils.

In very earliest times, whether 10 m.y.a. or 5m.y.a., the pattern of a forest ape hunting may have been similar to todays chimp - binges over days and weeks perhaps, but not always, co-related with the dry season, when food is in shorter supply and animals start to lose weight, or associated with a time of social bonding when dispersed groups in a territory form a temporary larger band. Hunting may help strengthen male bonding (at Gombe, Tanzania, 90% of kills are by males) [r].

Obviously, prey is limited to what is available. At Gombe National Park, 80% of the prey animals are red colobus monkeys, with bush pigs next most frequent, then small antelopes, baboons, and very small numbers of other vertebrates, including other monkeys. In all, the range of prey at Gombe traverses 25 vertebrate species.[r] Most often the prey are infant, juvenile, or adolescent red colobus. Adults are much less frequently caught.[r]

Chimpanzees are said to have a relatively unspecialised gut, not much different to modern humans. Whether they need to eat animals from a nutritional standpoint is moot. Certainly, as they consume the whole animal, they may absorb useful amounts of calcium from the splintered and smashed bones. But meat eating is somewhat sporadic - in the Mahale Mountains of Tanzania, 60% of the kills fall within a two month period - and don't make up a high percentage of the diet, at least by weight. In one area it is 0.3% or less, and the highest percentage recorded still averages only about 2% of the daily diet.[r] Meat intake is so low and erratic that chimpanzee, unlike human, have never become the 'primary' (definitive) host for Taenia tapeworms - a debilitating gut parasite whose definitive host is always a carnivore.[r] Even so, among chimpanzees, meat is highly valued, and so, we can fairly surmise, it would have been for the earliest semi-arboreal ancestor of us both.

Perhaps one of the most important clues to the meat eating element of our later bipedal Australopith or Australopith-like ancestry is the fact that chimpanzee do catch 'bush antelopes' and 'bush pigs'. These are terrestrial animals, and might be thought to have the best chance of escape from a knuckle walking hominoid. They form a small but significant portion of the species chimpanzees successfully hunt. Whether that translates to any greater success by an obligate biped in a closed canopy woodland or an open woodland or bushland is somewhat moot, but not particularly likely.

The seasonal evergreen forest seems the environment where key behaviours developed - food switching, communal hunting by related males, tool use by females, reliable cultural transmission of tool use, 'capture' of females to provide resources for kin, seasonally enforced ranging over greater distances and altitudes, defense of territory containing seasonally pivotal food resources. Some of these key factors for successful life in an evergreen seasonal forest are obligatory. Food switching is the most important example. Others, such as opening hard shelled oily tree seeds are useful, but not essential within that feeding ecology.  All 'preadapted' some apes to further expand into a new set of related feeding ecologies, but where the whole suite of behaviours taken as one became the key factor to survival and further adaptations.

In this model, the reason apes might move into a new feeding ecology was probably intensified competition for the food resources of the seasonal forest in a drying climatic period. The groups that were literally on the periphery of the forest - and particularly on the southern periphery where there was increased seasonality due to latitude (and perhaps also the eastern periphery where regular dry seasonality may have been associated with landform weather pattern interactions) - would be most likely to be under pressure from 'forest interior' communities with more stable seasonal forest resources. That is, in bad seasons, stronger and larger communties supported by the forest interior would need a larger range in which to find food. That range might then include peripheral woodland areas and the resources they contained. Peripheral groups may have been 'pushed' by such intraspecific competition into seasonal woodland, especially as the males in the smaller bands, unlike the females, would probably not have been able to integrate into the interior forest group, and therefore probably faced almost certain death.

Such peripheral populations would survive best - from the male perspective - if they could distance themselves from populations in the forest interior. An animal with a muscular-skeleto form reflecting adaption to semi erect suspensory feeding in a seasonal evergreen forest is at a disadvantage when very many miles have to be travelled through scrubby woodland that won't support brachiation. Both an obligate quadraped or an obligate biped can move much faster. Small changes in muscle and skeletal orientation and form that allowed more efficient ground travel away from more arboreally adapted competitors and toward patchy and seasonal resources may have been powerfully selective.

But  woodland model stands or falls on whether there are sufficient nutritionally dense resources available year-round in a woodland ecosystem to allow evolution of substantial omnivory.

The woodland biome

It has long been argued that primates that eat fruit that ripen sporadically and whose trees are scattered widely over a given territory need more memory capacity than folivores of a similar size. Primates that have to spend time and effort 'extracting' difficult-to-access energy dense foods - such as nut kernels, termites, and seeds within pods - from the environment have larger brains relative to their body size than those that are more folivorous, and the trend is magnified by the degree of omnivorous food switching (Gibson 1986).

The hominid in a woodland environment is constrained to do far more 'extracting' of plant foods, has fewer fruits available, more difficulty in finding water, faces more marked seasonality within a yearly cycle and also between years. As a result, a hominid is either omnivorous and needs a larger territory to find sufficient food, or retains a large folivorus adaptation and stays close to riverine fruit and rhizome resources.

In addition, the species of tree and shrub comprising the woodland change with latitude, soil, elevation, rainfall and so on. The lessons of one woodland environment don't always carry over to an adjacent woodland ecological zone. Over hominid evolutionary history the area of dry woodland relative to forest - in the tropics and subtropics particularly - would probably have waxed and waned with intervals of arid and moist climate. Woodland was probably far more extensive prior to human expansion 100,000 or so years ago until present [r].

The term 'woodland' covers a wide variety of feeding ecologies, most rather impoverished, and some very rich, but geographically/edaphically restricted and therefore unique. The term 'woodland' is very unsatisfactoraly broad, and therefore misleading; but it will have to do.

There are many types and intergrades of woodland - closed woodland with an almost continuous canopy of trees of about 13.5 metres (45 feet) and sparse grass underneath; open woodland of the same height, but wider spaced trees, discontinuous canopy and more grass; and woodland/bushland, similar to open woodland but with a partial substorey of bushy trees to about 6 metres (20 feet) plus grassy glades. Woodlands often gives way to deciduous forest along river margins, and to areas of wet grasslands at river deltas and lake edges. Woodland with lakes and river edges provide both varied ecozones and long 'transitional zones' between the ecozones. Transitions between ecozones are often rich in species diversity, and therefore feeding opportunities.

Large areas of woodland may be comprised of virtually one species; some fruiting trees appear sporadically in some places, in thickets in others, and may simply not be part of the makeup of a localised woodland ecozone. Food resources are therefore much more patchy than simply listing edible woodland plants would suggest. And present woodland species, composition and distribution can never be much more than speculatively suggestive of the nature of paleo-woodlands.

Closed woodland, with approximatly 950mm of annual rainfall, contains up to about 15% arboreal living larger mammals. The trees are perhaps 45 feet or so high, and there is generally an absence of a shrubby understorey. Closed woodland contains the highest percentage of tree dwelling and feeding mammals outside the generally equatorial forest habitat (in its broad sweep of types, from montane to swamp) where approximately 33% are commited arboreal feeders.[r]

Open woodlands have similar sized trees as closed woodland, but they are more widely spaced, and consequently with more grass beneath. Usually, the only arboreal animal is the galago. The large mammal species diversity is greatest in an open savanah woodland in Southern Africa, which also includes areas of damp soil dependant grasses and sedges. Fruit feeding species make up to 13% of the large mammal population, where very approximately half the large mammals in forest habit feed on fruit. Open woodlands are often threaded with rivers that have denser closed gallery woodlands along their banks.

At low to intermediate latitude a combination of seasonal drought and fire limit tree growth to those that can regenerate quickly and resist pass through of the fire front. Some trees shed their leaves in the dry season. Prolonged drought can result in overbrowsing, leaving some shrubby trees as persistantly sprouting and persistantly grazed ground level stumps until falling animal numbers later followed by some years of rain allow regrowth. These landscapes are often studded with termite mounds.

While the 'conditions of living' vary between types of woodland, broad features of African woodland can be described.  In general, the canopy is lower than forest, and more open. As a result, the understory vegetation is often more apparent. Because of the water stress in the long dry season (generally over 'winter' - variably from may to september) the trees and the understory usually drop all their leaves and remain bare until the rains in 'summer' (variably and intermittantly, from september to may). Fruit maturation is often synchronised to co-incide with the end of the dry. Some trees have swollen tap roots for water storage.  Lianes are few, and with the exception of Landolphia capensis, all in better watered sites, such as gallery forests and lakeside woodlands under Indian Oceanic influence.

Forest dwelling apes of wet equatorial regions live in a fairly stable environment. Food is generally abundant if you are a plant eater, there are many fruiting species, and fruit ripen throughout the year. Succulent herbs are available in specific zones as a fallback for hard times. It is true that preferred foods may have short seasons, but generally, food is not a problem. Woodlands are a much more erratic larders.

In general, fruiting tree species are fewer, but quite widespread. Vangueria infausta, Ximenia caffra and Sclerocarya birrea and Grewia sp. fruits, for example, are spread throughout the savanna woodlands of East and southern Africa (e.g., Fox and  Norwood-Young, 1982; Johns, Mhoro and Sanaya, 1996; Peters, 1988; Quin, 1959). Their fruiting season is usually over a short  period. Parinari curatellifolia, and Trichilia emetica fruit also have edible kernels,  Ficus sp analysed, to about 606kJ for the fruit of Diospyros dichrophylla. In the unlikely event of having to rely on these fruit alone to meet daily energy requirements of say 1,800 kcal, 4.5 kilograms of the Ficus would have to be consumed in the course of a day, or 1.2 kilograms of Diospyros dichrophylla. [n]. Grewia is a widespread genus, being present in arid woodlands and bushlands both above and below the equator.

In the dry times of year succulent foliage is hard to find, except for riverine grasses and sedges. But at this time of year Typha rhizomes and the roots of the semi-aquatic Prionum serratum, have good supplies of stored carbohydrate. The common marsh grass Phragmites autralis also has edible roots.

The question is, then, does the woodland environment yeild a reliable year round supply of nutrient dense foods?

In the equatorial grassy woodlands of northern Tanzania there are underground plant storage organs available - chiefly Vigna frutescens, V. macrorhyncha, Vatovaea pseudolablab, Eminia entennulifa and Ipomoea transvaalensis.

Relatively few seeds are available - the tree Parkia filicoidea, the waterlily, Nymphaea sp, and the small seeds of some grasses. The edible seeds of a few fruits - Parinari curatellifolia, Trichilia emetica, Adansonia, Sclerocarya birrea have already been mentioned.

Edible catipillars live on the Sclerocarya birrea trees.

Bird, monitor lizard and crocodile eggs are seasonally available, whereas reptiles such as python, young crocodile, and monitor lizards are available year round, as are small animals such as rodents, and to a lesser extent, frogs. Duikers and other small antelope are resident in woodland and grassland patches. Grasshoppers are important converters of grassland to ape protein, as they are captured more easily than mammalian herbivores. Indeed, when there are gluts of grasshoppers, baboons have been observed (Hamilton 1987) to live on them almost exclusively. [r]

Some lifestages of termites are very rich and nutritious. Termites are available all year round, with seasonal peaks.

Fruit can be regionally relatively abundant, with fruit hanging on into the difficult dry winter season, particularly Strychnos sp., Sclerocarya birrea, Grewia sp. and  Ziziphus sp.

Edible fungi can be harvested when the rains come, usually in spring and early summer.

These generalised notes on the 'woodland', or more accurately 'woodland mosaic' feeding ecology don't reflect the relative abundance, distribution, nutritional density, or reliability of food plants and animals in a given local climate, geography, soil type and drainage pattern.  There are many unique sub-feeding ecologies within the category 'woodlands'. As an illustration, two specific woodland subhabitats are presented, one equatorial (Tanzania), one warm temperate/Mediterranean.

The variable nature of a woodland - Coastal lowland lacustrine and riverine mosaic ecozones of Southern Africa.

Lake Sibaya in Southern Africa perhaps demonstrates a rich feeding 'node' that probably would have supported an omnivorous but significantly frugivorous bipedal ape as well, perhaps, as a bipedal largely 'bulb', pith, and stem feeding ape. The lake is quite large (about 60-70 km²), and is contained behind high dunes covered in coastal forest.

The dry season is about mid year, in winter, may to september. The adjacent Indian Ocean gives 1200 mm (47 inches) of rain to the eastern, oceanside shore, and about half that on the more inland shoreline, mainly in late spring, about november. The climate might be characterised as warm temperate, with about 11 C in the cool season and a high of about 29 C in summer.

Habitats include remmnant swamp forest, tall dune forest, shorter dune forest, dry coastal forest, open and closed woodland, waterlogged areas with damp adapted grass, and temporary shallow ponded water.These habitats support a wide variety of potential plant and animal food items, from crocodile eggs to mulberries.

The lake contains a crab, Hymenosoma orbiculare and a freshwater shrimp Caridina nilotica. Of the 18 species of fish, the most abundant are the cichilids Oreochromis mossambicus (Mozambique tilapia), Pseudocrenilabrus philander (southern mouth-brooder), and Tilapia sparrmanii (banded tilapia). A species of catfish, Clarias gariepinus (sharptooth catfish) and a goby, Glossogobius giuris are also relatively abundant. There are 22 species of frog with reed frogs (Hyperolius spp.), grass frogs (Ptychadena spp.), and Rana spp.being common. The Lake supports various water snakes, the water monitor, Varanus niloticus and is made dangerous by the resident hippopotamus (Hippopotamus amphibius) and Nile crocodile (Crocodylus niloticus) population. The surrounding land supports another sizable monitor, Varanus exanthematicus (veld monitor) and the ground nesting Stanley's Bustard (Neotis denhami) and numerous tree nesting species. 62 species of predominantly fish eating birds (coraments, herons, egrets, crakes, gallinules, plovers etc) either nest, roost, or feed at the lake, while reedbuck (Redunca arundinum) exploit the edaphic grasses. Other potential prey for a foraging ape include the African marsh rat (Dasymys incomtus) which breeds year round - one of about 30 rodent species - and several species of Duiker in the genus Cephalophus.

The most important foods are plants. The aquatic plants include waterlily (Nymphaea capensis), bulrush (Typha latifolia) and possibly Phragmites mauritianus (some species of Phragmites have edible roots). Damp areas support Cyperus spp, with common species having an edible bulb.

Fruiting tree and shrub species are numerous - Ficus trichopoda, Voacanga thouarsii, Syzygium cordatum (waterberry), in the tall dune forest there are the irregularly ripening fruit of the tall tree Chrysophyllum viridifolium, berries on the Cassine species shrub, and the fruits of Drypetes gerrardi (possibly edible, other species are). The shorter dune forest includes the edible fruiting species Mimusops caffra (a salt tolerant dominant species which can grow right to the high water mark), Balanites maughamii, Ziziphus macronata (important only as a famine food), Diospyros rotundifolia, Drypetes natalensis (possibly edible, other species are) and Euclea spp (probably edible, as many Euclea are, although not very palatable).

The understorey has stands of Acalypha glabrata (forest false-nettle), whose young shoots are edible. The patches of dry coastal forest include various Ficus species, Dialium schlechteri, Balanites maughamii, Manilkara discolor, and Mimusops caffra. Woodlands, both open and closed, are extensive. Many of the woodland trees have edible fruit, for example Strychnos madagascariensis, S. spinosa, Sclerocarya birrea, Trichilia sp., Hyphaene natalensis, Syzygium cordatum and Garcinia livingstonei, and Vangueria infausta . Climbers are abundant in the woodlands and on forest margins, and of the four most common species, two have edible fruit (Landolphia kirkii and L. petersiana). When the vegetation community is disturbed by overgrazing, aggressive shrubby fruiting species temporarily take over the open woodland and grasslands. These include the fruiting species Parinari capensis and Salacia krausii. The common woodland tree Terminalia sericea has palatable leaves.

What is notable about this snapshot is the large number of fruits available, most of which are likely seasonal, and the relative dearth of starchy or oily staples such as nuts and plant underground storage organs.

The stored plant foods are mostly seasonal wetland plants, mainly Typha and Nymphaea rhizomes.

The variable nature of a woodland - lacustrine woodland mosaic ecozone in equatorial Africa.
The following notes are primarily from Woodburn 1968 [r]. Adjacent to the eastern side of Lake Eyasi (200 kms south of Lake Victoria), in an arm of the Tanzanias rift valley, is a dryland mosaic of, rocky woodland, open savanna and wooded river beds. Rainfall is mainly over summer and autumn (December to May), and as in Southern Africa, the dry season is about mid year, in winter. The amount of rain, about 56 millimetres (22 inches), is similar to that of more inland Southern Africa.

Potential prey animals
The animals in this environment today are typical of the mixed grassland-woodland mosaic environment. "Elephant, rhinocerus, buffalo, giraffe, eleand, zebra, wildebeeste, hartebeeste, waterbuck, impala, Thompson's gazelle, warthog, baboon, lion, leopard and hyena are common, as are smaller animals such as anteater, porcupine, hare, hyrax, dik-dik, klipspringer, jackal, tortoise, and many others" (Woodburn).

The area has historically supported one of the last wild living Homo sapiens, peoples, the Hadza.'Modern' Hadza hunt with poisoned arrows, with metal arrow heads facilitating the killing of large animals. Therefore we must bear in mind that prior to this technology - at least 12,000 years ago[n] - fewer animals, and fewer large animals, may have been potential prey. Why? Because poisoned tip arrows are excellent ambush weapons. The animal has only to be penetrated by the point for death to be almost certain, if delayed. Without poison, the arrow has to hit a vital point for the animal to be killed or slowly bleed to death. Spears may only be effective on large animals if there are many spearthrowers, many hits, which implies ambush or escape routes being blocked.

Edible plants
Again, the most important foods are plants, and according to Woodburn, vegetable food is "always abundant even at the height of a dry season in a year of drought". In the dry season Adansonia digitata trees provide fruit and seeds, and Grewia berries remain available.

Edible insects
The grubs (and honey) of 7 species of bee are available, but the plentitude varies unpredictably between seasons and from year to year.

The place of the woodland environment in human evolution
Once in a woodland ecology, with more thinly spread resources, the ability to walk relatively large distances to find both food and water may have been key to survival. Even in more favorable well watered areas laced with water courses of various sizes and lakes, the greatest diversity of fruit bearing trees and shrubs are spread along riverine and lacustrine margins and adjacent scrubby woodlands. Bipedal ape sub groups might have comprised perhaps ten individual, and there might have been from three to ten related groups in a territory. Each sub group might have to disperse many kilometres each day to find sufficient food, and in the dry season, water.

In periods of extended aridity, for example, around the time about 1.79-1.74 mya when tropical Africa was in the grip of a dry cycle, springs fed by rainfall stored in aquifers in the adjacent Ngorongoro Highlands emerged from faults in the Rift Valley to feed local lakes and associated wetlands. These relatively small 'local refugia' provided essential water, perhaps carbohydrate rich aquatic plants, a concentration of other animals near the water, and both woodland and grassland termites to exploit. But these local reources limit population, and precipitate emmigration along 'drainage lines' where water is assured.

Patchy resources and bipedal stick-weilding stance promote evolution of reciprocal behaviours
When literally patchy resources are found, such as Cyperus bulbs in damp patches, the physical proximity pressure to share is probably both intense and critical. In times of shortage, a tension exists between quietly sequestering the resource or sharing with kin in the hope of reciprocal behaviour later. In the woodland, alternative foods are much more constrained. A club wielding bipedal ape can do huge amount more damage than a branch wielding quadripedal ape. Less physical damage to individuals or a group is incurred when (reluctant) reciprocal sharing happens. Bipedality might facilitate male stealing of resources, but because it also facilitates the ability of any adult to do severe damage to another, ultimately it must also be selective for reciprocal behaviours.

The leguminous and convolvulaceous tubers typically have fibrous and deeply sequestered roots (up to 3 metres deep for some species) - unlike the corms and bulbs of the Iridaceae of sub-equatorial southern Africa. While the energy yield of these equatorial leguminous tubers is lower than once thought [r], they are available year round, albeit a lot of effort is expended in finding, extracting, and carrying them back to a camp. In our evolutionary history, as now, it may be the females left with this task, while males seek small animals, possibly relatively unsuccessfuly - then, as now. It is more likely that very small bulbs such as those of Cyperus sp. were harvested, although these are limited to damp soils. (They have the additional advantage of being easily collected by children.) The difficult winter dry period does yield two further sources of carbohydrate - Typha and Nymphaea rhizhomes, whose safe collection is made easier by falling water levels.

In any scenario, the greatest physical and cultural plasticity is need to survive the unpredictable rigors of a woodland environment. We can imagine a selective pressure for fat storage, to make use of the seasonal overabundance of fruits when they are present. Some fruiting species provide very useful amounts of carbohydrate (Parinari curatellifolia, 88%, as the kernel is also edible), and protein (Trichilia emetica, 17% - [the seed is high in oil, and is toxic only if the seed coat is not removed first [r]]).[r]  The energy yeild varies with fruit species. Wild fruits of the Transkei, in South Africa, for example, can provide from about 165 kJ per 100 grams of edible part in the case of one Ficus sp analysed, to about 606kJ for the fruit of Diospyros dichrophylla. In the unlikely event of having to rely on these fruit alone to meet daily energy requirements of say 1,800 kcal (7536 kJ) for a small bodied hominid, 4.5 kilograms of the Ficus would have to be consumed in the course of a day, or 1.2 kilograms of Diospyros dichrophylla. [n]. Grewia is a widespread genus, being present in arid woodlands and bushlands both above and below the equator is a reliable dry season food.

While some fruits, such as Grewia berries, are easily accessible from the ground, it would be useful to retain short legs, relatively long and immensely strong forelimbs to easily climb some of the tall fruit bearing trees. One of the most valuable fruit trees is Sclerocarya birrea, for the oily kernel as well as the fruit. But fine motor control is needed to crack the hard seed and extract the kernels. Nutritious Adonsonia seeds must be crushed or ground between hardwood mortar and pestles, or between rocks. An opposable thumb and changes to the cerebral visual and motor cortices to control manual tasks would then be under selective pressure. These are the sorts of changes we see in Australopithicines.

Much more complex behaviour involving temporary food stores makes the ecozone more liveable. Such complexity may not have developed until the transition from Homo erectus to Homo sapiens. For example, Strychnos madagascariensis pulp can be dried and stored, but the knowledge that the seeds are toxic and must be removed has to be reliably transmitted. Snares might catch the secretive adult Duiker. Until such cultural ability developed, it may be that the woodland could only be reliably exploited by an ape that was both adapted to tree climbing and frugivory and wide ranging bipedal omnivory.

While the tall fruiting trees of the woodland made it essential for Australopithicines to retain tree climbing adaptations, apes pushed out of the woodland to shrubby arid grasslands (largely absent of diverse fruiting trees) had no need of such adaptations. So long as Australopithicines did not need advanced culturally transmitted behaviours to survive, they (with clear parallel to chimpanzee nut cracking behaviour) were not under selective pressure to acquire these 'mind evolving ratchets'.

Only once a hominid had evolved culture and tools to the stage of Homo erectus could this feeding ecology be re-exploited; although it may have had to have been anchored in the certainty of reef resources, and therefore restricted to the equatorial Indian Ocean coast and above.

The implication is that while the transition to a fully bipedal ape exploiting both terrestrial and arboreal food resources could only happen in the woodland mosaic, the transition to an obligate essentially terrestrial bipedal ape had to happen where arboreal morphology was no longer useful.

The Bushland Biome
Bushland has shrub or bushes up to around 20 feet high, no trees, usually poorer soils, and with very approximately 750mm of rain, about twice the rainfall as 'shrublands'. It often has areas of edaphic grassland. Bushland is perhaps more often a bushland/woodland habitat, such as parts of the Serengeti National Park. Again, usually, the only arboreal animal is the galago. There may be as much as 17% frugivorous species. It might be regarded as a lower grade woodland, and unlikely to be an environment of evolutionary adaptation, although possibly a later environment of exploitation by fully evolved Homo sapiens.

Beyond the woodland - arid shrubland

In the same way that hominid groups might have been pushed out of semi deciduous forest, down riverine gallery forest to mixed forest and woodland around equatorial and supra and sub equatorial Africa, so Australopithicine apes on the marginally survivable edge of the woodland would have been forced to adapt, move, or die out.

Dry phases put enormous pressure on populations. Cycles of aridity lasting about 23 thousand years ('Milankovitch precessional' periodicity) have been recorded in one of Kenyas stratigraphic record of diatome deposits in the period between 2.66 Ma and 2.56 Ma.. In these periods, food at the margins of productive habitats become very scarce. Populations fragment and travel far from the home territory. Under stress, populations shrink, and become isolated from each other. There is the risk of inbreeding depression reducing fitness. The greatest selective pressure is on culture as a means for the group to survive. Just as long lived elephants in Namibia respond to extreme drought by leading a group to permanant water that they last visited perhaps 3 years ago, so those adults that have explored, learned, and retained the most about location and preparation of the greatest diversity of food will tend to survive. More than culture, the genes and biochemical interaction that result in a more finely honed judicious brain, one that can make the 'right decisions' at the 'right' time, decisions involving a future extrapolated from 'evaluated' past memories, are more likely to accumulate as 'less savvy' individual and social units die out due to the effects of drought.

When moister conditions return, the population that expands from surviving residues may be more inquisitive, more plastic, and better fit to disperse widely.

The dispersal corridors are unknown. We can speculate drainage lines were followed into the Zimbabwe and Botswanan miombo. Moving along river galleries and from seasonal pan to seasonal pan, the Southern Angolan drainage system and Okavango would have been reached, then Namibia and the Atlantic coast of South Africa. Or, on the admittedly thin evidence of todays distribution, Ricinodendron heudelotti ssp. africanus whose seeds are nutrient dense, may have formed a key 'nutritional corridor' down which a hominid might move into the increasingly arid bushland/shrublands of west south African and central southern Africa, leading to a familiar and relatively reliable entry to the drylands - the related Ricinodendron rautaneii. A key point is that these areas are both cooler in winter, and seasonally dry. Therefore there are nutritionally dense bulbs and corms that 'self store' in the arid ground from year to year.

The corms, bulbs and tubers of the bushland perhaps provided a further nutrient dense resource to follow. This would have taken hominids to the varied and locally rich ecosystems of both the South African Cape and the mountains and coasts of the south east.

The Shrubland Biome
Shrubs are no more than about ten feet high at the most, usually with open dry soils between. As mentioned, rainfall is low (Kalahari thornveld, rainfall 450 mm per years) fruit feeding species make up only about 7% of the large mammal species - about half the number that open woodlands carry. The habitats become drier, more seasonal, and with only shrubs and no trees, there is an increase in the number of 'fossorial' animals. These are animals such as the aardvark, Orycteropus afer, which dig for food or dig underground burrows.

Where Dioscoria tubers in a forest environment are both buried deep and contain 'anti-feeding' compounds, shrubland tubers are usually (not always) much easier to access and more palatable. In contrast to the moist coastal plains in East Africa, where, according to Cunningham there are a low number of species of plants with edible tubers, roots or rhizomes [r] (due, obviously, to the very fact there is no significant dry or cold period), the dry west and parts of central Southern Africa have a much higher number of plants that survive the dry season as a dormant, nutrient filled, underground storage organ . But a sharp ended tool, usually a stick, is an essential tool for efficient gathering of  tubers.

Clearly, whether digging fossorial animals from their burrow, or digging tubers from the ground, a body plan permissive of an efficient, well controlled stabbing motion would be strongly selective for survival.

The Kalahari 'savannah' is the world centre of diversity for the evolution of 'geoxylic suffrutices', large underground woody storage structures which allow an above ground tree or shrub to survive seasonal and unpredicatable aridity. Counterintuitively, this allows fruiting trees such as Adansonia and Sclerocarya to survive, and in a year with good rains, fruit heavily. Others, such as Lannea edulis, the 'wild grape' ripen gradually over several months (April-May).[r]

Food in sandy shrublands is abundant in good years for most, but not all, the year.

However, the diversity of foods available is huge. Some fruits, such as Grewia, can be self picked by children for hours, and still have plenty to bring back to the camp. These fruits dry out in the winter aridity, making them even more nutritionally dense than the fresh berry. The foods of the arid bushlands traverse lizards, aardvaark, small antelope (duiker), ostrich eggs, melons, bulbs and corms, gums, seeds of mongongo (which while restricted in range, fall by the tonne and keep for over a year), Baobab seeds, birds eggs, and numerous fruits. It would require a book to do justice to the feeding ecology of this biome.

The place of the shrubland environment in human evolution

Humans are adapted to the heat, but we must have water. This in turn put a premium on the mind that developed techniques to temporarily store water, and the mind that could transmit these techniques down generations.

"Their groups were small because the desert gives food sparingly,  and often each family lived apart from other tribal members most of the year. The women did much of the work of food gathering - except for meat...the Indians couldn't make a good living from the desert alone. They lived around the edges; they haunted the lakes for wildfowl, eggs, tules, cattails, and yellow water lilies."
 - commentary of an observer of the Paiute tribespeople of the North American deserts, from earlier this century.

Counterintuitively, because the seasonality of the latitude and the Atlantic aridity lead plants to adapt by storing food and water underground, coupled with the patchy but rich lake and river resources, this is a refuge for an intelligent heat adapted hominid. These kinds of circumstances, in my view, are specific and felictous to the transitional evolution from Australopithecus to Homo, and from H. erectus to H. sapiens. It seems to me that this feeding ecology is the wellspring of the change from an ape of limited self awareness to an ape of full self awareness and consequent language communication. Homo sapiens.

During the summer 'wet' season when local pools of water are available, modern hunter gatherers are able to camp at the Mongongo groves and live relatively easily, without having to expend energy walking to find food. In late summer there are Grewia berries. Bauhinia esculenta is a water providing tuber at its best in late April but remains good for most of the winter dry season. It also provides morama beans, a 'much favored food' rich in protein and oil.

This resource is only available to a cultural ape that has the tool of fire. Whole pods are roasted, the bean removed and either chewed whole or ground. Critically, it can be dried and stored against scarcity. Raphionacme burkei flourishes all year round, and is a major source of water where a water source not available. it is fibrous, bitter, and melons are preferred. Scilla, like large spring onions, are abundant, and gathered in quantity, roast peeled and eat, although not favored. Green fleshed tsama melon Citrillus lanatus, 'constantly being cut open and eaten down to the skin'. SW Kalahari truffles Terfezia pfeilli brief appearance april may, no appearance in dry years.

Autumn is the time of greatest abundance. Bushman groups survive primarily from the gathering acticities of the females. A group of three or four females, even carrying one or two small children, do not have to travel far to gather enough roots, fruits, seedpods and berries for the entire band. A female food expedition at this time of year typically takes about an hour and a half; the exact location of the seasonal foods is known, gathering is directed to the known resources, and is not random.[r]

In Autumn, so long as rains earlier in the year had been adequate, 'eland cucumbers', Citrillus naudinianus, are plentiful - at least in well drained soils, such as the Kalahari.

The dry winter season in the Kalahari may last 6 months, and in bad years, much longer. Winter in the Kalahari is the time of greatest deprivation. As summer rain filled water holes dry out, hominids have to move near to permanant water. In some parts of the Kalahari shrubland desert there may be a few as 6 permanant water sources scattered over 1,400 square miles. Permanant water is distant from the mongongo nut groves. Winter nights in Southern Africa are quite cold today. Sufficient calories for warmth are essential at this time. In winter, females may travel for 8 or 9 hours a day, on 3 or 4 days in a week, returning to camp with perhaps 15 kgs (33lbs) of gathered food, as well as carrying a child for the greatest part of the way.

Clearly, survival in a seasonal shrubland environment depends on efficient walking. According to Lee [r], a !Kung Khoisan woman walks about 2,400 kilometres (1,500 miles) per year, almost all that distance in gathering food.

Meat eating is assumed to have become an important part of the diet of our ancestors. In its broadest sense, to include insect and reptiles, shellfish and birds eggs, probably. In the more usual sense of regulalrly eating larger animals such as antelope and the like, meat from large ungulates is unlikely to have ever been an important part of our diet until very recently - after Homo erectus, which is as far as we need to go. There will of course be spectacular community organised drives of animals over cliffs, or driven into bogs and speared, but while these are written in 'big letters' in the fossil record because they preserve well, such instances are unlikely to have much impact on day to day diet over the time period of later human evolution. It is much more likely that meat came regularly from small ground mammals, especially animals that can be trapped in their burrow such as Pleistocene equivalents of the jumping hare Pedetes Gaffer, or the ground squirrel, Euxerus erythropus fulvior.

Modern  bushmen are a small people, and their size, weight, and calorific requirements may be not that much different than those of an ancestral hominid adapted to seasonal shrubland and woodland living. A bushman kinship and allied group varies from about 20 to 40 people of all ages. Its territitory must have a permanant water source, and plant foods - in particular - sufficient to survive the dry season. We can speculate that in particularly arid decades hominids deep in the shrublands would have to emigrate to areas closer to reliable food and water, or die out. Such areas can only be lakesides, swamp edges, river edges, with resources of Trapa natans Typha latifolia, Prionum serratum, Nymphaea caerulea, Phragmites autralis.

According to Lee (1968), even in an abnormally dry year, and in the difficult winter part of the annual cycle, bushmen in the Dobe region of the Kalahari (considered one of the least productive parts) had an excess calorie intake over that needed for their body size and activity of 165 calories, and an excess of protein over that need of 33 grams. Mongongo nuts provided over half the daily calorie intake and meat about a third. The remainder comes from a variety of vegetable foods. The time spent in hunting animals worked out at ten hours per 1,000 calories. Plant food cost four hours per 1,000 calories. Even discounting the advantage of the poison arrow technology, this food comes from only a 'modest effort' of two or three days work a week, and without pressing youngsters into service [r]. Clearly, if necessary, more time spent foraging could compensate for less calories from meat.

Equally, each sub region of the arid bushland environment has its own staple basic plant food. On deep sands it is Mongongo. In areas of permanent water it might be aquatic and marshland rhizomes and bulbs. In mixed open woodland and grassland it might be bulbs and corms. Each sub habitat must have a central reliable basic plant food core. These sub habitats will tend to be unique confluences of geography, soil types, altitude, coastal versus continental effects, drainage lines and so on. The boundaries of these sub habitats can be expected to shift with changing climate, and in periods of long ardity, large areas might lose their basic food plants.

Under these conditions, a dispersal along permissive feeding ecology corridors, such as coastal strips and watered valleys, would be inevitable.

The Savannah Biome
Savannah is another term deeply lodged in popular conciousness due to TV programmes biase towards the remaining great concentrations of wild ungulates and their spectacular predators, usually on short grass, highly seasonal savannahs. Todays Savannahs, whether in Kenya or Tanzania, have been modified to a greater or lesser extent by fire. In driest areas gallery forest may trace the course of the river. For example, large areas north of the Congo basin are covered in gallery forest which then fade out into more open savvanah. Savannahs in extreme dry areas, such as the Kalahari, may be quite verdant when the rains come, but have areas with salt pans, low thorn brush and such aridity that plants have evolved adaptations to store water and resist dessication.

To an extent, todays savannahs are artificially maintained, and the term is not particularly useful. In the absence of human factors, it is likely that the boundary of the Congo forest would expand further outward, for example. And in moister periods they would again be more extensive. Upland plateau areas may attract more rainfall, with consequent longer grasses.
'Savannah' often has mosaics of larger or smaller patches of original forest or recovering forest. Heavily wooded areas may enclose savvanah patches. At a certain point in the continuum, savannah becomes woodland, which can be subdivided into various types contingent on the spacing of larger trees, and prescence or abcence of shrubs.

The percentage of arboreal  locomotion (moves and feeds in trees 90% + of the time) by larger mammals (500 grams +) in the well-known grassland savanna habitat type - the grassland plains of the Serengeti in Tanzania- is 0%. Unsuprising given the lack of trees. An open scrubby woodland savannah such as Amboseli National Park (600mm rain per year) has around 2% committed tree dwellers.[r] In more arid shrublands, such as the Kalahari thornveld, (rainfall 450 mm per years) fruit feeding species make up about 7% of the large mammal species, about half the number that open woodlands carry. Of interest is that as habitats become drier, more seasonal, and with only shrubs and no trees and only seasonal grass at best, there is an increase in the number of 'fossorial' animals. These are animals such as the aardvark Orycteropus afer, which dig for food or dig underground burrows.

The place of the Savannah biome in human evolution
There is good evidence from Khoisan hunter gathers, particularly the !Kung group of the south east margin of the Etosha Pan, where there is a grassy bushland/stunted woodland near a perennial water source, that the diet can be significantly made up of meat from animals absolutely dependant on coming to the water. This 'meat from large bodied animals' lifeway is not a useful scenario for considering an early bipedal lifestyle without the advanced technology of snares and bows and arrows.

Hunting large bodied animals is inherently dangerous. Even when approaching a wildebeest terminally affected by poison arrows, Khoisan hunters show prudent caution.

Although short grassland savannahs are not especially typical, and probably even less so prior to Pleistocene times, we can dismiss the idea that a band of small bodied bipedal hominoids of perhaps 40 kgs or so can advance across an open plain, outrun, catch hold of, restrain and kill a struggling large herbivore. No matter how many in the band, no matter how well armed with clubs.

This would not exclude fortuitously found unguarded 'kills'; generally there is quite a lot of competition from hyaena, jackal and vulture. A small hominoid venturing very far out onto short grass savvana is dangerously exposed; a short hominoid in long grass savannah is easily ambushed. And it is worth noting Stanfords comment that "modern chimpanzees in the wild have little interest in dead animals as food."

If we work back from recently extant gatherer-hunter groups that did not have bows and arrows, namely aboriginal peoples, we can observe that most of the food resources outside coastal regions were vegetable; where animals were killed, they were most likely to be small animals.

Leaving aside killing juvenile and new born large mammals, it is unlikely view that large mammals played any significant part in the evolution of the human diet up until very recently.

In summary, because the 'true' savannah supports large bodied herbivores within wide and exposed short grass plains, or poor visibility long grass plains, it is both a dangerous place for hominids and a difficult target to get close to, let alone kill (without being seriously injured). Large bodied herbivores are preyed on by large bodied felids, which require large prey, not small, but which are catholic in taste when conditions become difficult. Most hunts by wild living humans are unsuccessful. The most successful are ambush hunts, migration stream hunts, and hunts using poison arrows. None of these techniques are suitable on a savannah.

Further, it has been shown that hominids are unlikely to arrive at a felid 'kill' before other scavengers, and scavenging opportunities are too rare in the first place to be relied on as a regular source of food [r]. Scavenging for bone marrow as a significant strategy is implausable. Fortuitous capitilising on a found carcass can be expected (mainly leopard kills in woodlands), but not often.

The dry savannah biome can be dismissed as of relatively little significance to human evolution. Wet savannah, as part of an open woodland biome with gallery forest might provide useful meat at the rather compressed time of 'calving', especially of smaller antelope which tend to use adjacent woodlands as night camps anyway.

African Feeding ecologies during human evolution & speculation on the place of extinct hominoids in those ecologies

We know as little about the environment we evolved in as we do about the animals which may or may not have been ancestral to us. Broadly, Africa is a much drier place now than it was when we evolved. This pre-supposes we evolved to a substantially human form somewhere within the band ten million years ago to about 1.8 million years ago. To suggest some ideas about the evolution of the human diet we need to have an idea of the kind of ecosystems present in the areas in which we evolved. So we have to assume we had a physically unrestricted ability to disperse throughout Africa and adjacent landmasses. But we also have to assume that our distribution was limited by the distribution of the habitats that could provide food for us year round. In turn, we have to make assumptions about the kinds of plants and animals that might be present in the habitats that we re-construct as being present during 10 million years to about 1.8 million years ago.

Each one of these pre-suppositions about how and where we evolved, and what food was present is supported on weak and equivocal evidence, and informed guesswork; therefore what follows is necessarily not much more than informed speculation.

Habitats over the period of interest
Reed [r] has reconstructed the habitat at the Makapan Valley, in Northeast Transvaal, South Africa 3.2 to 2.7 million years ago as "bushland and medium density woodland". This site has yeilded more than 30,000 mammalian specimens, 24 being  A. africanus (one of the plausible ancestors of Homo in scenario 2 ). Fresh grass grazers and the prescence of some aquatic animals indicate a river and associated edaphic grassland areas transecting the woodland and bushland mosaic.

The more recent fossil assemblage (member 5, within about the last 2 million years ) suggests a much drier climate, perhaps typical of a shrubland, but with a body of water of some sort and with edaphic grasslands.

Reeds reconstruction of the habitat of the Sterkfontein Valley, near Johannesburg in South Africa at about 2.4 to 2.6 m.y.a is "an open woodland, with bushland and thicket areas". Fossils of A. africanus have been recovered from this habitat.

Again, in about the last 2 million years the rainfall and/or seasonality had changed to the point where only a grassland plains, or open wooded plains could survive. There were also likely areas of denser riparian woodland.

3 million years ago,  in the area 15 degrees above the equator in present day Chad, there were animals associated with mixed grasslands and woodlands - horse species, grassland grazers; and browers such as rhinoceros, giraffe, and elephants.

Today, the rainforest of the Southern Congo basin diffuses out into an extensive woodland known as 'Miombo'. '

What do fossil and modern ape skulls suggest about diet?
As a generalisation, the presence of  a 'saggital crest' on a modern or extinct fossil ape skull is related to eating pithy plants. Chimpanzees have a slight ridge on their skull, but most of their plant food could easily be eaten by a robust modern human. Mountain Gorilla have prounced raised nuchal crests. This may be related to a diet which includes the tougher stems of bamboo. Lowland gorillas generally don't have to deal with as challenging food, and have smaller nuchal crests. Their teeth are heavily wrinkled in ridges of hard enamel between soft dentine. These ridges are adaptation to mechanically disrupt plant cell walls of stems, buds, and leaves to release the digestible intracellular contents.

Chimpanzees have larger premolars than humans, and larger incisors. These and the large canines created a 'prognathic' face which allows a lot of room in the oral cavity for holding and manouvering bulky food for grinding between the molars. The surface of the molars are only slightly wrinkled, reflecting the more frugivorous diet.

Fossils of animals with large molars, robust mandibles, and anchorages for powerful jaw movement are probably the equivalent of the modern lowland (more frugivorous) Gorilla. These are likely terrestrial animals predominantley eating pithy, succulent stems, rhizomes, fruits and herbs.

What do fossil and modern ape skeletons suggest about diet?

One of the least obvious but most important aspects of the skeletons of a forest dwelling modern ape compared to a woodland and forest margin ape is that an upright stance allows safe killing of larger animals than chimpanzees or baboons can manage. Chimpanzees have not been recorded killing animals larger than themselves for food. When a group do kill a large animal - a fellow chimpanzee - the attack is prolonged, with the victim eventually dying from numerous bites, and from the relatively unco-ordinated pounding. This contrasts with the dispatch of smaller animals, which the chimps quickly flail to death against a branch, or simply rip apart. Baboons actvely hunt antelope fawns hiding in the long grass of woodlands. These animals are too small to endanger the baboons, and they merely seize the prey and eat it alive, without attempting to kill it first. Why? Carnivores have almost scissor-like jaws that can open wide to seize prey. They have little in the way of cheek muscles (used to shift and position food in the mouth for extended chewing), but massive clamping temporalis muscles. They run little risk of their jaws being dislocated when they grip firmly on struggling large prey. Baboons, apes, and humans have jaws hinged above the upper tooth array. An ape or baboon attempting to 'clamp' the throat of a large struggling animal runs a high risk of dislocating its jaw. A dog baboon, in spite of having massive canines, can't afford to use them in killing larger prey.

So a firm bipedal stance is permissive of using bone or wood clubs to dispatch larger animals than could be 'handled' with the unsteady  upright stance of  a knuckle walker such as chimpanzee or gorilla. Whether or not there was much opportunity, or whether opportunities were taken, is a different matter.

But a balanced bipedal stance does permit wading. Chimpanzees, knuckle walkers, are in some areas, afraid of even shallow water, and are relatively helpless in deeper water. Their present range into sub species and species seems to be significantly bounded by major river systems. Western Gorillas appear to enter water to feed on sedges on occasion, but not regularly.

Steady bipedal posture allows the exploitation of the fleshy basal stems of the swamp and river edge plant Typha and young stems of Cyperus papyifera. The papyrus exists in dense stands in some ecological areas. Crocodiles may be a constraint in some rivers and lakes, but in others the fish population is so abundant that shoreline predation is relatively rare.

The only other constraint is infection with the parasitic flatworm Schistosoma sp.. The water snails Bulinus globosus and Biomphalaria pfeifferi are vectors of Schistosoma haematobium, causing a debilitating disease 'schistosomiasis'. In the view of a specialist -
"you could argue that it was the regular visitations of primates/early hominids to water that permitted the evolution of a schistosome like Schistosoma haematobium which is almost  exclusively a human parasite... So I think its very likely early hominids had intimate and extensive contact with water, just as we do. They probably got infected and infected heavily with schisto...however, just as with the ancient Egyptians who were no doubt heavily infected, it might not have prevented them from thriving, because schisto is best thought of as a debilitating rather than acute, life threatening infection." (E.S.Loker, pers comm. January 2001)

It is plausible that S. haemotobium evolved primarily in association with Paranthropus, if the hypothesis that it was primarily a lake and riverside herbivore is accepted. Transmission to H. erectus might have occured as H. erectus radiated into and displaced Paranthropus from its range.

Fossil animals have been found in environments that have been reconstructed as woodlands and woodland gallery forests. The range of tooth morphology goes from large molars in Australopithecus anamensis, A. afarensis, A. garhi and A. robustus, and A. africanus, to human and chimpanzee sized molars in the as yet undescribed 'millenium hominid'.

Which of these has a tooth morphology that would prevent it from living in a tropical forest environment such as exists in the Congo basin today? None. Of these animals, chimpanzee and human have the smallest teeth. Chimpanzee have relatively much larger incisors than human, but molars are very similar in size and shape. Human do live as hunter-gatherers in the Congo basin - the peoples we call 'Pygmy'.

We must exclude an omnivorous proto human from the forest environment on the grounds that resources are patchy, and only fairly sophisticated bow and arrow communal hunting techniques allow the group to live there today.

Does any fossil or extant ape tooth morphology disallow eating animals? No. Again, chimpanzees probably have the smallest molars, and they have no difficulty eating  monkeys. Small animals are easily torn into pieces. As fossil hominoids are all noted as having powerful muscles, they too would easily tear apart small animals without the need for tools.

Humans, with relatively weak cheek muscles and fragile zgomatic arches can (atypically, not in ordinary mastication) generate forces of 500 pounds per square inch on the molars, and 150 pounds per square inch on the incisors.

Once brachiating apes had evolved in Africas evergreen tropical equatorial jungle, it is inevitable they would disperse. Apes at the margin of the evergreen jungle were forced into the semi-deciduous tropic jungle feeding ecology.

In like manner, as jungles shrank in drier phases, apes on the margin of the semideciduous forest adapted to a combined forest and terrestial feeding ecology of the woodland mosaic environments; but only because they retained adaptations for bipedal walking and competant tree climbing. (While the idea that a larger, bipedal ape might re-invade the evergreen tropical forest and outcompete its natal line of descent is intriguing, is is unecessarily complex.). It is likely, in my opinion, that this double adapted ape remained evolutionarily static.

The loss of substantial canines suggests it was a habitual weapon user. Gene based, behaviourally expressed, mechanisms to defuse tension may have been selected for to avoid males doing lethal harm to group members. While woodland resources are arguably no more patchy and seasonal than in semi-deciduous forest, the diversity of food types and the sub-habitats that contain them is very much greater. These complexities, with the need for fine motor control of the precision grip might have been selective for changing brain organisation, albeit size might be sufficient for a seasonal environment quite analagous the seasonal tropical forest.

A move inland and to south west Africa was underwritten by plants producing corms and bulbs in response to Southern Africas latitudinal and arid climatic seasonality. The aridity of the region made tree climbing adaptations redundant. The forelimbs shortened, the hind limbs lengthened. Supported by plant foods, the feeding ecology was dependant on acute observation and exploitation of the widest diversity of plants and animals, including lake and river resources. The technology of cooking and drying 'broke open' new food resources.

Finally, the brain honed by the richness and challenges of Southwest Africa allowed a dispersal out of Southwest Africa and across the world.

Ultimately, Southern Africas varied environments and peninsular-like isolation further refined the conceptualisation facility of the human mind and its cross generational cultural transmission. This proto sapient ape displaced its erectine progenitor as it dispersed through coastal woodlands out of Africa and into Eurasia. Able to live in most climates, most geographies.

Copyright (c)  2002  Laurie Meadows

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.1 or any later version published by the Free Software Foundation. with no Invariant Sections, with no Front-Cover Texts, and with no Back-Cover Texts. A copy of the license is included in the section entitled
 "GNU Free Documentation License".

Version History
This is an uncompleted first draft of a specialised article. Much of the draft was completed in 2001. For a variety of reasons, the article had little further work to date. But some of the specific information and some of the ideas are not easily available, accordingly, this first draft (and subsequent minor revisions), wordy and undisciplined, is published for others to use, modify, take what elements are useful.

First published 25/10/02.

Minor revised version #1 published: 26/10/02. Minor revised version #2 published: 30/10/02. Minor revised version #3 published 31/10/02. Edited version #4 published 26/04/17.