Thorpe S.K.S., Holder R.L., and Crompton R.H.

This work is reported in the journal Science 1st June 2007

The Science web site abstract can be found here:

The full text of the article can be found here:

Media coverage links:

Based on computer simulations of the mechanics of motion in fossil human ancestors such as the famous 'Lucy' skeleton, our research group has long argued that early human ancestors would have walked upright, rather than semi-crouched, as the old 'up from the apes' view

has suggested (See e.g. : How Lucy Walked, JHE1998, J Anat 2004, Lucy musculoskeletal model and J R Soc Interface 2005). But we have not been able to say where such upright walking originated. Now, research on the orang utan, suggests that upright walking may have been a basic element of the lifestyle of the earliest ancestors of modern apes, including humans, which would have been tree-dwelling specialists on ripe fruit, living among the fine branches of tropical forest trees.

It is generally thought that the apes separated from the monkeys some 24 million years ago, filling a specialized niche as ripe-fruit eaters, moving and foraging among the fine branches at the periphery of tropical forest trees.

All living apes, and all their close relatives, which are known from fossils as much as 17-21 million years old, show skeletal adaptations to upright posture of the trunk, a posture today associated with climbing on high-angled branches and tree-trunks, suspension by the forelimbs from near-horizontal branches and vines, and of course our own two-legged upright walking, or bipedalism.

However the traditional ‘up from the apes’ view (see image above) of the origins of human walking holds that bipedalism evolved in a ground-dwelling, quadrupedal knuckle-walking ape as it moved out of forest, into open country. A knuckle-walking origin is still favoured by quite a few workers, who look to chimpanzees for a model of the appearance and habits of the common chimpanzee-human ancestor.

As this image indicates, the ‘knuckle-walking’ model has been taken by some to imply a period in which early human ancestors would have had a less-than- fully upright posture when walking, as a consequence of compromise between arboreal and terrestrial life. Others, including ourselves, have argued with increasing success that a period of such ‘bent-hip, bent-knee’ gait is unsupported by the fossil evidence, and biomechanically unlikely.

In the absence of any evidence that our earliest ancestors differed either in brainsize or dietary adaptation from other great apes, the presence of adaptations for habitual bipedalism has for many years been the criterion used to distinguish between fossils of early human relatives and those of other apes.

But recently discovered fossils such as Orrorin, from East Africa, appear to show evidence of bipedality at as much as 5-8 million years ago, a range similar to DNA-based dates for the genetic separation of humans and chimpanzees. On which side of the division do they lie, or could they be ancestors to both?

From 4.5 to 3.5 MYA fossils from undoubted human relatives, such as Australopithecus anamensis, and of course Australopithecus afarensis, (‘the species to which ‘Lucy’ belonged) have hindlimbs adapted for terrestrial bipedalism, but forelimbs arms characteristic of tree dwellers. Of course, not all of these may be direct human ancestors.

Importantly, all these fossils are found in predominantly woodland environments. It isn’t until 2 MY that human ancestors are found associated with predominantly open environments, and then they are members of our own genus, Homo, with limb proportions rather similar to our own.

Humans are not the only living ape to be able to walk bipedally: all living great apes (the gorilla, pygmy and common chimpanzee and humans) can and do walk bipedally on occasion. , Our own work has shown that curiously the ape which has the most human-like (that is upright) bipedalism is not our closest genetic relative, the rather terrestrial chimpanzee, but the more distantly related and almost entirely tree-dwelling orangutan.

As the only great ape which remains in the ancestral ape niche, orangutans are a vital model for an understanding of the evolution of limb adaptation in the apes in relation to its environment. In the ancestral, fine-branch, niche, both access to fruits, and crossing gaps within the trees require an ability to navigate very thin terminal tree branches which are liable to bend under body mass.

The logical conclusion from the fossil, environmental and experimental evidence that is described here is that upright, straight-legged walking originally evolved as an adaptation to tree-dwelling. But how can this be? Apes and monkeys nearly all increase the flexion of their fore- and hindlimb-joints when moving upon small, flexible branches. This acts to lower the centre of gravity of the body and hence reduce the chance of toppling, and also reduces branch vibration. Some living monkeys and apes, of course avoid the risk of toppling altogether by suspending themselves underneath the branches.

Our hypothesis was that bipedal locomotion in the trees might actually be advantageous for arboreal apes on the thin terminal branches because their long prehensile toes can grip multiple small branches and thus maximize stability, while freeing one or both hands for balance, feeding or weight transfer while crossing gaps in what primatologists call the ‘terminal branch niche’.

To test this hypothesis we studied Sumatran orangutans for 1 year. Following them from when they woke up in the morning to when they made their night nest and recording every minute a range of variables, including the type of locomotion they used, how many branches/lianas they used to support their body weight, and the diameter of the supports to indicate how flexible the supports were.

We then statistically compared the associations between bipedalism and quadrupedalism and suspension (which as explained earlier are the locomotor behaviours that are expected to be associated with flexible supports) with the number of supports used and the diameter of the supports.

We found that when orangutans are moving on the single largest, most stable supports, they used quadrupedal gaits; when they move on slightly smaller supports they use suspension, but when they move on multiple really small supports (<4cm in diameter) they use bipedalism.

There are 2 important aspects to note about their bipedalism:

Thus we have shown how hand-assisted bipedalism could have evolved in the original/basic ape habitat, the terminal branch niche, to navigate the very smallest branches where the tasty fruits are and the smallest gaps between tree crowns.

Our results:

We can only speculate on how and why chimpanzees and gorillas became knuckle-walkers. But following Jonathan Kingdon, we think that in the late middle Miocene (about 10-12 mya), the tropical forests of East and Central Africa experienced a series of cyclical climatic changes, which resulted in the repeated opening and closing of the forest as the climate became drier or wetter. During times when the forests were receding, apes began to encounter more and more gaps between trees that they couldnt cross in the forest canopy. So, we suggest that:


Play Sequence-S1 Play Sequence-S2 Play Sequence-S3

From a conservation perspective:

Sumatran orangutans are predicted to be extinct in the next decade if habitat destruction continues at the rate it is today. The fact that orangutans are so important in understanding our own evolution further highlights the importance of conserving the rainforests, the ecosystems and the orangutans within them.


With this in mind perhaps any further research in this area should be part of a broader project that studies associations between orangutan locomotor behaviour and habitat use, thus identifying key habitat requirements that have implications for their survival and/or reintroduction in logged/degraded habitats.


All pages © PREMOG 2007 | | Last updated 24/5/2007