Primates and hominins

Foundations

Two corrections before anything else

We did not evolve from chimpanzees. Chimpanzees are not an earlier stage of us. They are a separate species that has been evolving for exactly as long as we have, from a common ancestor that was neither. That ancestor lived around seven million years ago, and it would not have looked much like either of its descendants.

The confusion matters because it smuggles in a ladder. There is no ladder. There is a branching bush, most of whose branches are dead, and we are one surviving twig.

And a big brain did not come first. The intuition is that intelligence arrived, and intelligence made everything else possible, including walking upright to free the hands. The fossils say the reverse. Hominins were walking on two legs for roughly four million years while their brains stayed about the size of a chimpanzee’s.

Standing up was not the consequence of being clever. It came first, and whatever it was for, it was not thinking.

The evidence for the split

Until the 1960s, palaeontologists generally put the human-ape divergence at fifteen to thirty million years ago. Then biochemists began comparing proteins between species and found humans and chimpanzees far more similar than that timescale allowed. The molecular estimate came out around five million years, and it was received badly.

The molecular people were substantially right. The current estimate, from whole genomes calibrated against fossils, is around six to eight million years, and the fossil record has moved to meet it.

Genome sequencing has since added detail that a fossil cannot. Chimpanzees and bonobos are each other’s closest relatives, and both are closer to us than either is to gorillas. And, awkwardly for our self-image, a single group of chimpanzees in central Africa carries more genetic diversity than the whole human species.

The candidates

Fossils near the split are scarce, because forest floors do not preserve bone well and central Africa is under-surveyed.

Sahelanthropus tchadensis, found in Chad in 2001 and dated to around seven million years ago, is the oldest serious candidate. Its interest is the position of the foramen magnum, the hole where the spinal cord leaves the skull. In a quadruped it sits at the back; in a biped it sits underneath, so the head balances on top of the spine. In Sahelanthropus it is intermediate and argued about, and a femur attributed to the same species has produced further argument rather than resolution.

Ardipithecus ramidus, at about 4.4 million years, was described in 2009 after seventeen years of work on a crushed skeleton. It is the most informative fossil in this range and the most disruptive. It walked upright, and it had an opposable big toe for grasping branches. It lived in woodland, not grassland.

That combination broke two assumptions at once. It undercut the savannah hypothesis, which held that hominins stood up when forests thinned. And it undercut the habit of treating the chimpanzee as a living stand-in for our ancestor, because Ardipithecus is not chimpanzee-like. Chimpanzees have been evolving too, and knuckle-walking may be their specialisation rather than our inheritance.

Lucy, and the point of the lesson

Australopithecus afarensis is where the picture becomes clear rather than contested. The species spans roughly 3.9 to 2.9 million years ago, and the best-known individual, found in Ethiopia in 1974 and nicknamed Lucy, is about forty percent complete.

Lucy stood about 1.1 metres tall. Her pelvis is short and bowl-shaped, like a human’s and unlike an ape’s. Her femurs angle inward from hip to knee, which is what puts the feet under the body’s centre of mass in a walking animal. Her foot has no grasping toe.

Her brain was around 400 cubic centimetres. A chimpanzee’s is about 350 to 400. A modern human’s is about 1,350.

Fully upright, and no smarter than an ape. That single fact is the lesson.

Footprints

The strongest evidence for early bipedalism is not a skeleton at all.

In 1976, at Laetoli in Tanzania, Mary Leakey’s team found trails of footprints in a layer of volcanic ash that had been rained on, set like cement, and buried. They are about 3.6 million years old. Two or three individuals walked across, and one seems to have stepped in another’s prints.

The prints show a rounded heel strike, weight transferred along the outside of the foot and pushed off through a big toe aligned with the others. That is a human gait. Bones let you argue about how an animal probably moved. Footprints record how one actually did.

So what was it for?

Nobody knows, and the honest list of proposals is worth having.

Efficiency. Walking upright over distance costs a chimpanzee-sized animal substantially less energy than knuckle-walking. In a landscape where food is patchy, that matters.

Carrying. Two free hands can transport food, infants and eventually tools. This is attractive because it connects to everything later, and it is hard to test.

Heat. An upright body intercepts less midday sun and catches more breeze. This works well for open country and less well now that early hominins keep turning up in woodland.

Reaching. Standing to pick fruit from low branches is something chimpanzees already do, so the behaviour need not be invented, only made habitual.

The current position is that these are not competing so much as overlapping, and that no single one has been demonstrated.

What it cost

Bipedalism is not an upgrade. It is a compromise, and you are living with the bill.

The spine was a horizontal beam and is now a vertical column carrying compressive load, which is why back pain is close to universal. The knee takes forces it was not shaped for. The birth canal has to pass through a pelvis reshaped for walking, at the same time as brains were getting larger, which makes human birth unusually difficult and dangerous compared with other primates.

Evolution does not optimise. It modifies whatever was there, as lesson 5.3 described, and the modifications leave marks.

A bush, not a line

The picture of a single lineage marching from ape to human is wrong in one more way: for most of this story, several hominin species were alive at once.

At various points there were robust australopithecines with enormous chewing muscles alongside more gracile forms, and later several species of our own genus overlapping in time. The current situation, one hominin species on Earth, is unusual. For most of the last seven million years it was not the case.

Deeper

Reading locomotion from bone

The skeletal signature of habitual bipedalism is a suite rather than a single feature: an anteriorly placed foramen magnum, a short and laterally flared ilium, a valgus knee with an angled femoral shaft, a robust calcaneus, and a non-divergent hallux with a longitudinal arch. Species can and do show partial suites, which is what makes mosaic evolution more than a slogan. Ardipithecus has a divergent hallux with a pelvis showing bipedal adaptations, and Australopithecus sediba combines an australopith upper body with a derived pelvis, so any single-character diagnosis is unreliable.

The Laetoli debate

Analyses of the Laetoli trails have argued both for a fully modern striding gait and for a bent-hip, bent-knee gait, based on depth profiles of the prints and inferred pressure distributions. Experimental work with humans walking through comparable substrates has narrowed but not closed this. A separate trail excavated at the same site since 2015 includes a notably larger individual, which has been used to argue for pronounced body-size differences between sexes in A. afarensis.

Why the molecular clock won

The original protein-immunology comparisons of the 1960s were crude but produced a robust result: the human-chimpanzee difference was far smaller than the human-orangutan difference in a way inconsistent with a 15-million-year split. The subsequent conflict was resolved partly by recognising that Ramapithecus, then interpreted as an early hominin at 14 million years, is an orangutan relative. This episode is a useful case of a quantitative method overturning a qualitative consensus, and of a fossil being reinterpreted rather than a clock being wrong.

Incomplete lineage sorting

Genome comparisons show that for a measurable fraction of the genome, humans are more closely related to gorillas than to chimpanzees, and chimpanzees to gorillas than to humans. This is not a contradiction: it reflects ancestral polymorphism sorting differently in different lineages when speciation events are close together in time. It also means “the” divergence date is a distribution rather than a point, which is one reason estimates span a range.

Frontier

What the last common ancestor looked like

There is no fossil confidently assigned to the human-chimpanzee last common ancestor, and reconstructions differ substantially on whether it was a suspensory ape, a knuckle-walker or a more generalised arboreal quadruped. Since chimpanzees are derived rather than primitive, the common practice of using them as a model is increasingly questioned.

The status of the earliest candidates

Whether Sahelanthropus, Orrorin and Ardipithecus are hominins, apes on other branches, or a mixture is not settled, and the specimens are few. New material from the relevant time range in central and eastern Africa would do more to resolve hominin origins than any analytical advance.

Transcript and shot list

Target runtime about 8 minutes. Built as an order reversal: establish the sequence people assume, show the fossils that invert it, then ask the honest question about what upright walking was for. Shot types: ANIM is rendered from code with Manim, AI is a generated shot from Higgsfield with the shared style reference, REAL is agency imagery with credit, CAM is the narrator.

[00:00] S01 · ANIM · The familiar march-of-progress illustration, ape to human in a line; it is struck through and replaced by a branching bush with most branches ending.

Two corrections before anything else, because both are in almost every version of this story you have seen.

[00:15] S02 · ANIM · A tree: a common ancestor at 7 Ma, splitting into two lines of equal length, one ending at a chimpanzee and one at a human.

First, we did not evolve from chimpanzees. Chimpanzees are not an earlier version of us. They are a separate species that has been evolving for exactly as long as we have, from a common ancestor that was neither of us and would not have looked much like either. That matters because the ladder picture smuggles in a direction. There is no ladder. There is a branching bush, most of whose branches are dead, and we are one surviving twig.

[00:50] S03 · ANIM · The assumed sequence: brain grows, then hands are freed, then walking. Then the sequence is reversed: walking at 4 Ma with a small brain, brain expansion beginning at 2 Ma.

Second, and this is the lesson: the big brain did not come first. The intuition is that intelligence arrived and made everything else possible, including standing up to free the hands. The fossils say the opposite. Hominins walked on two legs for about four million years while their brains stayed roughly chimpanzee-sized. Standing up was not a consequence of being clever.

[01:25] S04 · ANIM · 1960s: palaeontologists placing the split at 15 to 30 Ma; then protein comparisons producing 5 Ma; then modern genomes settling at 6 to 8 Ma. A chimpanzee population shown holding more genetic diversity than all humans.

How do we know when the split was? Until the 1960s, palaeontologists put it at fifteen to thirty million years ago. Then biochemists compared proteins between species and found humans and chimps far too similar for that. Their estimate came out around five million, and it was received badly. They were substantially right. Modern genomes say six to eight. And genome sequencing added something a fossil cannot: one group of chimpanzees in central Africa carries more genetic diversity than the entire human species.

[02:05] S05 · REAL · The Sahelanthropus cranium, Toumaï (museum image; confirm credit line). S06 · ANIM · A quadruped skull with the foramen magnum at the back; a biped skull with it underneath; Sahelanthropus intermediate, with a question mark.

Fossils near the split are scarce, because forest floors do not preserve bone and central Africa is under-surveyed. The oldest serious candidate is this: found in Chad in 2001, dated to around seven million years. What makes it interesting is the hole where the spinal cord leaves the skull. In a four-legged animal it sits at the back. In a two-legged one it sits underneath, so the head balances on the spine. In this skull it is somewhere in between, and people are still arguing about it.

[02:45] S07 · ANIM · Ardipithecus: an upright-walking skeleton with an opposable grasping big toe, in woodland rather than grassland; the savannah hypothesis crossed out; a chimpanzee labelled DERIVED, NOT PRIMITIVE.

Then this one, at four point four million years, described in 2009 after seventeen years of work on a crushed skeleton. It is the most informative fossil in this range and the most disruptive. It walked upright. It also had an opposable big toe for gripping branches. And it lived in woodland, not grassland. That broke two assumptions at once. It undercut the idea that hominins stood up because forests thinned into savannah. And it undercut the habit of using a chimpanzee as a stand-in for our ancestor, because chimps have been evolving too, and knuckle-walking may be their invention rather than our inheritance.

[03:35] S08 · REAL · The Lucy skeleton laid out (Cleveland Museum of Natural History or National Museum of Ethiopia; confirm credit line). S09 · ANIM · Her pelvis compared with an ape’s and a human’s; femurs angling inward; a foot with no grasping toe; then brain volumes: chimpanzee 350-400, Lucy ~400, modern human ~1350.

And then it gets clear. Australopithecus afarensis, three point nine to two point nine million years ago, and this individual, found in Ethiopia in 1974 and nicknamed Lucy, is about forty percent complete. She stood about one point one metres. Her pelvis is short and bowl-shaped, like ours and unlike an ape’s. Her thigh bones angle inward from hip to knee, which is what puts the feet under the body when you walk. Her foot has no grasping toe. And her brain was about four hundred cubic centimetres. A chimpanzee’s is three hundred and fifty to four hundred. Ours is about thirteen hundred and fifty. Fully upright. No bigger-brained than an ape. That is the whole lesson in one specimen.

[04:30] S10 · REAL · The Laetoli trackway in situ (research or museum image; confirm credit line). S11 · ANIM · Ash falls, rain sets it, footprints are impressed, ash is buried; then a gait analysis: heel strike, weight along the outer foot, push-off through an aligned big toe.

But the strongest evidence is not a skeleton at all. In 1976, at Laetoli in Tanzania, Mary Leakey’s team found trails of footprints in volcanic ash that had been rained on, set like cement, and buried. Three point six million years old. Two or three individuals walked across, and one appears to have stepped in another’s prints. And the prints show a rounded heel strike, weight rolling along the outside of the foot, and a push-off through a big toe lined up with the others. That is a human gait. Bones let you argue about how an animal probably moved. Footprints show you how one actually did.

[05:20] S12 · ANIM · Four proposals as cards: energetic efficiency, carrying, thermoregulation, reaching; each with supporting and weakening evidence; none ticked.

So what was it for? Nobody knows, and here is the honest list. Efficiency: walking upright over distance costs a chimp-sized animal much less energy than knuckle-walking, which matters when food is patchy. Carrying: two free hands can move food, infants and eventually tools. Attractive, because it connects to everything later, and very hard to test. Heat: an upright body intercepts less midday sun and catches more breeze. Works well for open country, less well now that early hominins keep turning up in woodland. And reaching: standing to pick low fruit is something chimps already do, so it need not be invented, only made habitual. The current position is that these overlap rather than compete, and none has been demonstrated.

[06:20] S13 · ANIM · The costs: a horizontal spine becoming a vertical load-bearing column; the knee taking new forces; a birth canal constrained by a walking pelvis while brains enlarge.

And it was not an upgrade. It was a compromise, and you are paying for it. The spine was a horizontal beam and is now a vertical column under compression, which is why back pain is close to universal. The knee takes forces it was not shaped for. And the birth canal has to pass through a pelvis reshaped for walking, at the same time as brains were getting bigger, which makes human birth unusually difficult compared with other primates. Evolution does not optimise. It modifies what is already there, and the modifications leave marks.

[07:10] S14 · ANIM · A timeline showing multiple hominin species overlapping at several points: robust and gracile australopithecines, then several species of Homo; the present marked as a single surviving species.

One last correction. For most of this story, several hominin species were alive at the same time. Robust forms with enormous chewing muscles alongside lighter ones. Later, several species of our own genus overlapping. Having exactly one hominin species on Earth, which is the situation now, is the unusual state of affairs. It has not been true for most of the last seven million years.

[07:45] S15 · AI into ANIM · A hominin footprint in ash, held; the calendar returns, showing 22:30 on 31 December.

Seven million years ago on the cosmic calendar is about half past ten on the evening of the thirty-first of December. Everything in this chapter happens in the last ninety minutes of the year. Next lesson: what happened when the brain finally did start growing, and how many kinds of human there turned out to be.

Shot list

One AI shot, eleven code-rendered shots, three museum or research images. Estimated Higgsfield use: 2 to 4 generations including retries.