first steps: jerry desilva on the evolution of bipedalism

This morning I got up (at the rather early and unaccustomed hour of 3.30am) to listen to a webinar by paleoanthropologist Dr Jeremy DeSilva¹. Titled “First Steps”, his presentation was about the origins of bipedalism in the human lineage. It was a fascinating session & I thought I’d turn my notes into this post, to share with students, teachers, and anyone else interested in the topic.

Things kicked off with a reminder that the familiar iconography relating to evolution – the gradual linear progression from some ape-like ancestor to a fully upright, modern human – is, while often amusing, just plain wrong. The path of human evolution is highly branched, not linear. Nor have changes in our biology all happened in lockstep with each other; they’ve occurred at different times and different rates, to result in a mosaic of features in any particular species on our family tree.

In fact, it’s worth students remembering that even starting these sequences with something that looks like a chimp (see below) is also inaccurate. The lineages leading to modern chimps & humans diverged 6-7 million years ago (mya), and that last common ancestor probably looked quite different to both lots of descendants. (Richard Dawkins’ book The Ancestor’s Tale reinforces this concept.)

Coming back to the point about different rates & times of change: we were shown a graph of changes in brain size over time. Early hominins eg Sahelanthropus had a cranial capacity about the same as that of a modern chimp. This increased by about 20% with the evolution of the australopithicines (Lucy, A.afarensisis an example), and then increased more markedly about 2mya with Homo erectus. However, that increase wasn’t a universal thing: Homo naledi (shown in the featured image at the top of this post) retained a rather small brain despite having other relatively modern skeletal features.

But, did these changes in brain size happen at the same time as changes in locomotion? They definitely did not. Both Sahelanthropus and Orrorin, dated to 6-7 mya, had some features that supported bipedal locomotion – and this suggests that bipedalism in some form may well predate the evolution of hominins. (At this point, I’ll note that while gorillas and chimpanzees are both knuckle-walkers, they have different adaptations for this, which implies that this form of locomotion is not an ancestral one.)

Now, a chimpanzee can’t walk fully upright due to features of their spine & pelvis. They have fewer lumbar vertebrae than we do (Jerry described this as a short lower back, compared to our longer one), and their pelvis is distinctively different from ours.

comparison drawings of chimpanzee, Australopithecus africanus, and Homo sapiens pelvises and feet

Image from

You can see from this image that the chimp pelvis – specifically, the ilial bones – is elongated, with a big vertical gap between the tip of the sacrum/coccyx and the socket of the hip joint. What’s more, seen side-one those ilial bones are oriented almost parallel with the spine, which has consequences in terms of muscle attachment & operation². On the right, the modern human pelvis is shorter & more bowl-shaped, and the flaring ilial bones are oriented more to the side – in terms of how the attached muscles work, it means that we can stand & balance on one leg, whereas chimps can’t do this, and that has an impact on how we walk & run. Interestingly, gibbons, siamangs, and monkeys also have longer lower backs (more lumbar vertebrae).

Jerry went on to say that while most or all non-human primates can walk on 2 legs for brief periods of time, albeit rather clumsily, it’s hard to know how common this is. So he and a grad student took advantage of the covid-19 shutdown. They contacted North American zoos and asked them to report on the frequency of bipedalism in their primates during lockdown, when the animals would not be responding to the presence of zoo visitors & so hopefully would be showing more natural behaviours. The researchers hoped that the results of this request would allow them to test an hypothesis about the evolution of bipedalism: that if it was an ancestral feature found in primates predating the chimp-human split, and was related to the presence of a longer lower back, then modern primates with that feature would be more likely to walk bipedally on occasion.

Collectively, the zoos reported that 88% of their gibbons, and 76% of siamangs, walked bipedally at least some of the time, but the behaviour was seen far less often in other apes.

What about fossil evidence? Unfortunately the fossil record for chimps and gorillas is sparse, to say the least. However, if we go back around 12 mya, in the Miocene, there are quite a lot of primate fossils from around the northern shores of what’s now the Mediterranean and the Himalayan region, and in Africa. Jerry commented that these show long lower backs with what he called an orthograde posture, suggesting that they moved through trees and on the ground in a more upright posture rather than on all fours (pronograde). Again, this suggests an early origin for bipedalism, a suggestion supported by the separate evolution of knuckle-walking in chimps & gorillas, as I mentioned earlier. And also supported by the observation that Orrorin‘s 5 myo femur is more similar to that of those Miocene apes than it is to modern apes’ thighbones. (This is also a feature of dinosaur evolution. The iconic predator Tyrannosaurus rex was a biped – but so were the earliest-known dinosaurs; quadrupedalism evolved multiple times in the dino lineage.)

So, why did bipedal locomotion persist in the hominin lineage³? After all, said Jerry, it has multiple disadvantages. Not only are we subject to bad backs & fallen arches in our feet, but we are slow. Running at top speed, Usain Bolt can manage almost 45kph – which is phenomenal, but nowhere near as fast as a charging lion or cheetah or a fleeing antelope. So those early hominins would have been extremely vulnerable to predation, and would have struggled to run down dinner. They may well have spent the nights up in trees, coming down to forage during daylight hours, instead of at dusk & dawn. And social behaviour – someone always being on watch for predators (think meerkats) would have been important. Of course, the development of tools and use of fire would also have helped.

We’re also prone to falling, which can often be fatal, and are subject to foot & leg injury – something that’s also evident in other living primates and in our fossil record. He told us that there are foot & leg fractures seen in hominin bones from 2-3 mya – and some of those had healed prior to the individual’s death, which tells us something really interesting about social behaviour in the australopiths and early members of our own genus: there must have been some level of care (eg feeding, defence) offered to injured members of their group.

Nonetheless, despite what looks like grounds for strong selection against it, bipedalism has continued. This suggests strong adaptive significance for the behaviour & related physical adaptations, and students will be familiar with many suggestions here: the ability to carry & use tools, to carry food, to dissipate heat or spot predators, for example. Thus by 3mya, skeletal material & fossil footprints tell us that Lucy (Australopithecus afarensis) was standing and walking a lot like us. But, the story is more complicated than this makes it sound. At the same time that Lucy was alive, the structure of the “Burtele foot” shows us that another hominin species was walking in a different way from her.

Come forward in time to 2mya & the earliest Homo erectus. From the neck down their skeletons are modern – they walked as we do. But Australopithecus sediba – alive at the same time – was a hyper-pronator, with a really marked roll in each step from the outside of the heel to the big toe; quite a distinctive difference. This species also had characteristics suggesting it spent quite a bit of time in the trees, and so its musculo-skeletal system was probably the result of a trade-off (in evolutionary terms) between the two modes of locomotion.

And again, 0.25mya: Homo sapiens, H.naledi, & H.neanderthalensis all seem to have had somewhat different walks. H.naledi, with its small leg joints & flat feet, was definitely not a distance walker, but sapiens & the earlier erectus are/were capable of endurance running.

Thus (Jerry concluded), bipedalism is a mosaic feature, honed and shaped at different times in different species. Tl;dr: evolution is complicated.


¹ Dr DeSilva is a paleoanthropologist & an Associate Professor of Anthropology at Dartmouth College, USA. His studies of early human fossil foot bones have contributed to our understanding of the origins & evolution of upright walking in the human lineage.

² Jerry was asked about the impact of pelvis shape on childbirth. He responded that the changes in the hominin pelvis, which lowered our centre of mass, also altered the shape of the birth canal. While chimps give birth with the infant’s head facing forwards (which means it’s easy for the mother to “assist” at the birth), human babies’ heads & bodies rotate as they pass down the birth canal, facing first sideways & subsequently towards the back. The result is that mothers usually need someone to assist with the birth as they can’t easily guide the infant out themselves. Now, have another look at that image of pelvis structures, & you’ll see that Lucy’s was very similar to ours – this suggests that perhaps female A.afarensis mothers may have had others present at the birth of their babies: an intriguing insight into possible social behaviour in this species.

³ You could equally ask (& one of the audience did) why gorillas, chimps, & bonobos became knuckle-walkers, rather than retaining bipedalism. Jerry’s response was that big apes in trees would have major problems with falling; there’d be strong selection in favour of individuals with long arms, curved fingers, & a short lower back (all advantageous in an arboreal lifestyle). But if they subsequently started foraging on the ground (remember that adaptations for bipedalism are seen in fossils of animals that appear to have been largely arboreal), then those back, arm & hand features could really only accommodate knuckle-walking. And, as I’ve commented earlier, knuckle-walking has evolved at least twice, once in gorillas & once in chimps.

2 thoughts on “first steps: jerry desilva on the evolution of bipedalism”

  • Evolution is (or can be) really complex. So much so that even experts aren’t always able to “comprehend” it. And sometimes, there is no real “good” answer to the question “what is the story behind X?”. This can be true in regard to all biological features…

    • Absolutely – sometimes the ‘story’ may simply be that a particular trait was carried along for the ride, as other traits were selected for or against. However, asking the questions about function and evolution is important and informative.

Leave a Reply

Your email address will not be published. Required fields are marked *