Why did we start walking?

(or was it wading?)

Over 150 years after Darwin suggested that we evolved from the same ancestors as did the apes, we still don't know why we go about on two legs whilst they tend to do so on all fours.

If you look up bipedalism in a university-level text book about human evolution you'll find that lots of theories have been suggested over the years. Some say it was to free our hands for carrying things, others so we could see further. Some propose that it was to help reach for food in the branches of trees, others that it was to discourage violence through upright posturing. Some claim that that it enabled them to travel long distances more efficiently, others that it helped them do slow-speed foraging. One even suggests that the earliest bipeds, in order to gain an advantage over their rivals in foraging time, moved from the shade of trees out onto the open grassland, like mad dogs or Englishmen, into the mid-day, equatorial sun. Once there, the theory claims that they would have moved in an upright manner, by some strange logic, in order to keep cooler!

Most texts politely list them all but sensibly refrain from backing any one of them too strongly. None of them really stand out as being the obvious choice and all of them make sense in some ways, but not others.

Faced with this enigma, we might reasonably conclude that perhaps they all played a part in the solution. This would seem to be the viewpoint around which a consensus is most likely to form. It might satisfy our sense of fair play to think that they might all have contributed equally to the solution of the riddle but we should not forget that scientific theory doesn't work like politics.

There is another possibility of course. Maybe they're all wrong. Perhaps the people who thought of them just missed something or made a basic assumption that is simply false. It is throwing a spanner in the works to suggest it, but the facts indicate that this is precisely what has happened.

The standard theories can be grouped into three categories: Those that assume that bipedalism originated in the trees, those that assume it originated on the ground and those that assume it was a bit of both.

None of those seriously considered by specialists assume that it might have happened in water. This simple blind spot, it would seem, is the cause of the impasse.

Hardy’s Wading Hypothesis

The idea of a wading-origin for bipedalism is not new. 41 years and 5 months ago the eminent marine biologist, Sir Alister Hardy FRS, writing in these pages in fact (New Scientist 1960), first publicly suggested that bipedalism may have originated in apes adapted to a water-side habitat. Water, he argued, would have provided the perfect place for bipedalism to have evolved.

Sir Alister Hardy FRS, the originator of

the wading-origins model for bipedalism

The idea certainly makes perfect sense from a theoretical point of view. Natural selection predicts that traits evolve to maximize the reproductive success of the individual. This can work either directly or indirectly via kin selection. However, adaptations which offer an immediate, life-or-death survival benefit to an individual must, logically, be ranked as more significant than more subtle ones which may give milder benefit to others later. It is difficult to conceive of a more clear-cut survival benefit for an upright posture than exists in waist deep water. If an ape wades quadrupedally, its face is below the surface and it drowns. If it wades bipedally its not and it lives. Genes that code for upright posture are thus likely to thrive in apes living in water-side niches.

None of the specialists in the field have taken Hardy's idea very seriously, however. Perhaps it was because he wasn’t a paleoanthropologist, perhaps it was because he did not provide any new evidence - just a re-interpretation of old or perhaps the timing was just wrong. In 1960 most in the field had almost the exact opposite niche in mind for early hominids, namely the African savannah.

However, since then a great deal of fossil evidence has emerged of earlier bipeds that lived when Africa was not so dry and in places that were heavily forested and predominantly wet. It is hard to justify its omission from the debate today. One has to ask: If it has never even been studied, on what basis has the idea been dismissed? Somebody’s hunch? Whatever it is, it's not science.

I first learned about Hardy's idea from Elaine Morgan's beautifully written and inspirational books about the aquatic ape hypothesis. What she argued seemed to make sense to me and yet there was not even a whisper about it in the official texts.

Elaine Morgan, author of several

inspirational aquatic ape books

I was so determined to find out why academics had dismissed the idea that I went back to university, to UCL,  to do a masters degree in human evolution. I was surprised and disappointed to find that the subject was simply not on the agenda. When I raised the question with my tutor she said it was the fault of those who had proposed the idea. They had not put it forward in a strictly scientific way. Someone should make some predictions about the hypothesis and then test them, she said. So I decided to do so myself.

Apes Wade Bipedally

Precious little work had been done studying extant ape behaviour in water as they have always been considered to be rather hydrophobic. However, recent observations of western lowland gorillas at the swamps of Mbeli Bai have started to change this image. In fact, digging around the literature reveals that all four species have been observed to go into water when the need arises and, when they do, it is usually bipedally.

Male western lowland gorilla at Mbeli Bai. Orang-utans, chimpanzees and bonobos have all been observed wading bipedally like this.

I carried out a new empirical study of captive bonobos’ bipedal time at the wildlife park at Planckendael near Brussels in different substrates. I found that whilst they were bipedal only 2% of the time on land, the figure rose to over 90% when in water.

Level of bipedality in captive bonobos  in different substrates.

Bonobos were filmed with a video camera and the amount of time (in seconds) spent in a supported (where upper arms are used for balance) or unsupported bipedalism was calculated as an estimate of the time spent in that substrate.

They very rarely stepped into the moat that surrounded their enclosure, only ever doing so if the food which children threw to them fell short. But when they did go in it was almost always on two feet, even if the water was shallow enough to go in on all fours. One of the workers there, who had tracked the bonobos every day for two months told me that my observations were no fluke. He had seen individuals go into the water every day and always bipedally. These findings are exactly what the model for a wading origin for bipedalism would predict.

Redy (male, left ) and Hermien (female with infant, right)  wading bipedally at Planckendael, near Brussels.

Only once were they seen to go in on all fours.

Of course we did not evolve from extant apes, so what evidence is there that our putative ancestors waded too? Certainly they could have. If one examines the paleohabitats of the earliest bipeds - Orrorin tugenensis (5.6-6.3 mya), Ardipithecus ramidus (5.8-4.4mya), Australopithecus afarensis (3.2-4.1 mya) and A. anamensis (3.9 mya) one finds that, generally, they were wet and wooded, not dry and open. Each of them has been associated with significant bodies of water: either lakes or rivers. There are other hominid sites, whose paleohabitats seem to have been more arid, but they are young compared with the sites of Orrorin and Ardipithecus. Again this evidence is only what the wading model would expect. It certainly can’t be argued that it refutes the hypothesis.

How did Lucy Move?

So they could have waded, but did they? If the earliest bipeds were wading apes one might expect to find some kind of evidence for it in their morphology. The most complete and famous fossil record is AL 288-1 (Lucy), and it is no surprise that she has been extensively studied. Much has been written about her curious morphology but rather than a consensus view forming, the interpretation of her structure has caused much disagreement.

Some, like Owen C. Lovejoy and Robin Crompton, have argued that she walked in a fully upright, purely terrestrial, human-like way, others such as Jack Stern and Christine Berge say that it was kind of waddling gait with a 'bent-hip-bent-knee' posture, indicating that she lived a more arboreal lifestyle.

Berge has claimed that she couldn't have walked upright because she wouldn't have been able to keep her balance as her anatomy is not geared to a fully erect posture. Crompton and his collaborators have retorted that it couldn't have been a bent-hip-bent-knee gait because she would have used up too much energy and become too hot.

The arguments have certainly generated a great deal of heat but both sides have been making the same, big, dry assumption. If Lucy was a wading ape then most of their contradictions could be resolved. Buoyancy would have made it easier for her to move with a bent-hip-bent-knee gait and the cooling effect of water would have stopped her overheating whilst doing so. On the other hand, maintaining an upright posture would have been no problem for an ape as long as she was wading in water that was deep enough.

Another piece of evidence recently cast new doubts on the way she moved. In 2000 Brian Richmond & David Strait discovered that Lucy's distal radii (the ends of the arm bones nearest to the wrist) had traits analogous to ones found in chimpanzees and gorillas associated with knuckle-walking.

Because it wouldn’t make sense that she would have had two separate ways of moving on the ground (bipedalism and knuckle-walking), a rather convenient argument has been put forward explaining that these traits were evolutionary baggage from the past - even though A. africanus, a putative descendent of A. afarensis that lived only about 0.5 my later, no longer had those traits.

The most parsimonious explanation of Lucy’s knuckle-walking traits has to be that Lucy was a knuckle-walker. And, if that was the case then, logically, the simplest explanation for it was that she had three modes of locomotion for three different substrates: She climbed in trees, knuckle-walked on the ground and waded in water.

Those hips

Lucy has a very peculiar post-cranial morphology. In particular her pelvis has a characteristic platypelloid (flat front-to-back, wide side-to-side) shape with large iliac arches. They were odd enough for Charles Oxnard to conclude that she was probably not even an ancestor of ours or the African apes, but was instead evidence of a failed experiment of bipedalism. He contemplated that her mode of locomotion may have been an odd mosaic of behaviours seen in extant primates or else it was “a totally new and unknown manner of locomotion which would be unique in its own right.”

What could those traits really have been for? If we cast aside our assumption that she was a terrestrial walker and consider the possibility, instead, that Lucy was a wading ape - what explanation might be found for those traits then?

I approached the problem with a series of experiments with human subjects in a variable-depth swimming pool (or multi-pool). I wanted to find out what factors impact on wading speed. Predictably, it slowed down as the water got deeper and, specifically, I found that speed was directly related to the area of the submerged body profile. From this one would predict that any ape specialised to wading would evolve behaviours and traits to minimise this profile.

Sideways Wading

The really fascinating observations came from comparing different wading techniques. Intriguingly, I found that sideways wading, which was predictably slower than wading forwards at the shallowest depths, was actually at least as fast when the water reached chest height. Waddling, a kind of ice-skating mode, was intermediate between the two. This, remember, is from a species that can hardly be said to be adapted to moving sideways. I had never even considered the idea until my fourteen-year-old son, one of the volunteers, suggested it!

The concept might sound a little fanciful at first but extant apes have been seen to do it too. A worker at Planckendael told me that, on several occasions, he had witnessed bonobos running bipedally and sideways to get out of the rain quickly.

Wading speeds at different depths (expressed as a fraction of hip height) using different methods. Notice that at the greater depths sideways wading was as fast as the other methods and the steeper regression line.

Comparison of drag force against speed for frontal and sideways wading in humans. This used a gross underestimate of drag (based purely on submerged projected area and assuming no streamlining) but it clearly shows that the sideways mode is more efficient. Humans cannot propel themselves powerfully enough sideways to make this the fastest way of wading at all depths but maybe our ancestors could.

Sideways wading clearly creates significantly less drag than moving frontally. The excellent "Life in Moving Fluids" by Steven Vogel provides the basic fluid mechanics behind this observation but actually you only have to put your hand outside a moving car to feel the effect for yourself. Placed front-on, it gets buffeted heavily by the wind. Place it sideways and it cuts through the air like a knife. Sideways certainly would be the optimal mode of wading at all depths if humans were able to propel themselves with as much force as they do frontally.

An ape specialised in wading might therefore be expected to adopt a sideways gait to maximise its efficiency. The notion seems far fetched at first but if you dispel any pre-conceived notions and look at the facts dispassionately, I think you'll agree that the evidence that Lucy might have moved in this way is quite staggering.

Firstly, what evidence is there that Lucy would have been able to propel herself sideways with more force than we can? Quite a lot, actually. Moving sideways in water requires a fair amount of complex muscle co-ordination. The volunteers who participated in the wading experiments found that kicking the back leg out sideways (involving a strong abduction of the femur) and alternatively clawing back the front leg (adduction) seemed the best technique for the fastest sideways wading. So, we'd expect to see evidence that Lucy was stronger in this kind of movement than we are.

The skeletal and muscular structures involved in this kind of movement are illustrated in the diagram below. Basic lever theory would predict that for A. afarensis to have a stronger sideways movement than us, it should have either:

1. A greater pulling force applied (W) from the iliac arches to the greater trochanter of the femur (for abduction) or from the femur to the ischiopubic ramus (for adduction).

2. An increase in femoral neck length (N) or

3. A decreased femur length (L).

but Lucy would appear to have had all three.

Australopithecines have always been associated with long femoral necks and short femurs, large iliac arches and extended ischiopubic ramus.

Berge noted that “it appears that the australopithecine adductor musculature must have been more powerful than that of humans” and Lovejoy and others have suggested the same for her abduction.

Basic biomechanics of abduction & adduction (After Berge 1994) Hip musculature is shown in humans (left) and Lucy. In the centre Berge has assumed a human-like musculature, on the right an ape-like one.

Basic level theory indicates that Lucy would have been significantly stronger in abduction and adduction than humans.

What is the evidence that she was streamlined for sideways wading? That evidence is even stronger. The A. afarensis pelvis is often contrasted with Homo as being very much wider than ours. However this is the result of scaling up her size to human proportions and can therefore be a little misleading. If one ignores scaling and measures the pelvis itself one finds that it is not so much that her pelvis was wide laterally but that it was shallow (less than 65% of the human thickness) dorso-ventrally.

Human wading experiments show that submerged body profile is significantly reduced when wading sideways (about one third less). It follows that any ape specialising in such a mode of locomotion would evolve a body shape that reduced this profile to a minimum. Lucy’s sideways profile appears to have been less than half of her full frontal area.

Comparison of Lucy’s pelvis with a human female’s (after Abitbol 1994)

It’s not so much wide side-to-side, as flat front-to-back.

Human frontal and sideways profile at varying depths. The red line indicates the wade-swim threshold (when wading becomes faster than swimming) for this individual. Notice that the sideways profile is 30-45% less in humans. It was probably around 50% in Australopithecus afarensis.

The morphology of A. afarensis would appear to have been remarkably well suited to sideways wading although it is, of course, possible that she actually waddled through the water as predicted by Berge. No other explanation, hitherto suggested, fits a putative mode of locomotion so perfectly.

So if Lucy was a wading ape, can we imagine a scenario where her ancestors might have needed this ability? As young adolescent apes, if they had been forced to wade quickly across stretches of waist deep water even just once a week before reaching sexual maturity, it would still represent hundreds of wading events, each yielding an opportunity for selective pressure to eliminate the traits that hindered orthograde posture and bipedality. In this model the actual percentage time in the water need not be high. The threat of predation, often used as an argument against the wading-origin idea, would paradoxically significantly increase the adaptive pressure and simultaneously explain why adaptations specifically for fast wading would have evolved.

There are many habitats that would provide such opportunities for these selective pressures to work. Alister Hardy originally envisioned a coastal habitat. Elaine Morgan refined this by suggesting it was the flooding of ancestral forests leaving island niches. Others, such as Marc Verhaegen have proposed coastal mangrove swamps. I personally prefer that idea that Lucy was a “river ape” inhabiting gallery forests that surrounded rivers and lakes. This putative habitat is certainly more consistent with the fossil record and could even work in arid zones, as forestation inevitably shrinks closer to water sources as the climate becomes drier.

Hypothetical optimal ranges for quadrupedalism, bipedalism and swimming.

A – denotes probable water depth where bipedal wading was optimal for a swimming hominid. B – denotes possible overlap of quadrupedalism and swimming. C – denotes possible optimal depth for non-swimming hominid.

If our ancestors couldn’t swim there would very likely have been many niches where wading would have been the only way to get about.

The view that bipedalism may have originated in apes that were adapted to a water-side habitat is still not a popular one. However if the theory has been dismissed it is not clear on what basis this has happened because all the serious research in this area has yet to be done. Charles Oxnard wrote “though numbers may prevail in democratic institutions, in academia and in science they do not. One fact that controverts an hypothesis is more powerful than a thousand that support it.” In my study I actually found nothing to rebut the idea that bipedalism may have had a wading origin and a great deal to support it.

Algis Kuliukas

(September 2001)

Further Reading:

Abitbol, Maurice (1994). Lateral view of Australopithecus afarensis: primitive aspects of bipedal positional behaviour in the earliest hominids.
Journal of Human Evolution, 28: 211-229.

Aeillo, Leslie & Christopher Dean (1999). An Introduction to Human Evolutionary Anatomy. London: Academic Press.

Berge, Christine (1994). How did the australopithecines walk?  Journal of Human Evolution. 26:259-273.

Crompton, Robin Huw, Li Yu Wang Wejlie, Michael Günther and Russel Savage (1998). The mechanical effectiveness of erect and “bent-hip, bent-knee” bipedal walking in Australopithecus afarensis. Journal of Human Evolution 35: 55-74.

Hardy, Alister (1960). Was Man More Aquatic in the Past? New Scientist 642-645.

Morgan, Elaine (1990). The Scars of Evolution. London: Souvenir Press.

Oxnard, Charles (1983). The Order of Man. Hong Kong University Press.

Richmond, Brian. G. & Strait David. S. (2000). Evidence that humans evolved from a knuckle-walking ancestor. Nature 2000 404: 382-5.

Vogel, Steven. (1994). Life in moving fluids : the physical biology of flow Princeton University Press, Princeton.

Wheeler, P.E. (1988). Stand tall and stay cool. New Scientist. 12 May: 62-65.

See also my web-site for a full copy of the thesis: www.RiverApes.com