What Killed the Hominins of AL 333?

Over 36 years since its discovery in Ethiopia’s Afar Depression, the 3.2 million year old skeleton of Lucy is still the most famous in all of paleoanthropology. Older fossil humans have been found, as have more complete remains, but none have generated the same swell of interest that has virtually turned these Australopithecus afarensis bones into the mascot of an entire scientific subdiscipline. This is a boon to scientists – to have an almost instantly-recognizable specimen to use as a hook in science communication – but, despite her small stature, Lucy has cast a long shadow that has obscured a spectacular collection of her relatives who died under mysterious circumstances.

In 1975, just a year after the discovery of Lucy, scientists found the remains of multiple individuals of the same species within the same slice of geologic time. The site was given the designation AL 333, and it yielded over 200 bones representing multiple individuals of different ages. At a time when Australopithecus afarensis was the oldest human known, this assemblage was given the popular name “The First Family.”

Most of the AL 333 fossils had eroded out of their parent layer and were found scattered over the hillside, but a few were found in the rock itself. Though no single skeleton was as complete as Lucy’s, the collection of fossils provided anthropologists with a rich store of information about growth and variation in the newly-discovered species.* Yet bones do not only tell us about the lives of the animals they once supported. They also can tell us something about the way those animals died and what happened to their bodies after death. The presence of so many human bones at one small site raises the question of how this group met their demise.

One of the first steps in reconstructing the last moments of the AL 333 australopithecines is figuring out how many there were in the first place. This is not easy. Estimates have fluctuated from five to twenty two, and this is attributable to the scrappy nature of the bones.

Most of the skeletal elements found at AL 333 were bits and pieces. There were no well-preserved skeletons. (The only articulated parts were a partial hand and foot.) In such circumstances paleoanthropologists must survey the entire collection for duplicate parts that allow them to calculate a minimum number of individuals. In a 2005 study of the AL 333 fossils, anthropologist J. Michael Plavcan and co-authors noted that the minimum number of individuals in the dental sample was nine, whereas the minimum number of individuals based upon parts of the body (excluding the head) was only three or four. There is no way to know whether the individuals which left teeth behind are also represented by these postcranial elements. A greater number of individuals may have contributed to the postcranial sample – Plavcan and colleagues estimated between five and eight – but, since the collection is so fragmentary, it is impossible to detect overlap with the dental sample. The number of individuals at AL 333 is impossible to know with absolute certainty.

For the sake of argument, though, let’s tentatively accept the count forwarded by Donald Johanson (one of the discoverers of the AL 333 fossils) as 17 individuals, including nine adults, three adolescents, and five children. That is quite a large number of A. afarensis to find in one place, and the general assumption has been that these individuals all died at once.

One of the earliest hypotheses to account for the AL 333 site was that the hominin family had been swept up by a local flood. This was supported by a 1992 study conducted by Stefan Radosevich, Gregory Retallack, and Maurice Taieb. Brown siltstone at the locality was interpreted as evidence of flooding, and remnants of the ancient soil were taken as a sign that the hominins lived in a dry woodland habitat near a streambed. While Radosevich, Retallack, and Taieb could not confirm what killed the australopithecine group, it appeared that they died during or just after a flood. Their bodies rotted and their bones began to be scattered (explaining why there were no articulated skeletons), and within a matter of months another flood buried the fragmented remains.

This is a common burial scenario for many fossil accumulations. The well-known, 130 million year old fossils of the England’s Wessex Formation were buried when heavy rains created slurries of plant debris and soil that picked up the carcasses of dinosaurs and other animals as the sloppy mass swept down the prehistoric hills. In a different fashion, the Cretaceous deposits along Alaska’s Colville River were created when the seasonal thaw created floods which quickly overtook and buried the Arctic dinosaurs which lived in the lowlands. Obviously these sites are much, much older than AL 333, but they underscore the role floods and debris flows have played in creating the fossil record. Local, catastrophic floods both kill and bury, and they can even preserve the remains of animals which died sometime previously.

The death-by-flood hypothesis seemed simple enough, and the lack of traces left by carnivores at AL 333 seemed to rule out the idea that the hominins had been prey or that their bones had accumulated at the site over a longer period of time. (Although, as Radosevich and co-authors noted, it is possible that “such evidence could have been obscured during preparation [of the fossils.]”)  This hypothesis, of a group of 17 hominins caught up in a flood, led to some rather dramatic reenactments, such as the imaginary scene of a sleeping family of A. afarensis finding themselves suddenly inundated in Bones, Stones, and Molecules by David Cameron and Colin Groves (p. 59-60). Even so, there was no conclusive piece of evidence which showed that the A. afarensis individuals had drowned. Their bodies had almost certainly been buried by flooding, but the idea that a flood had swept up an entire family was mostly a matter of inference.

Stories gleaned from the fossil record are always subject to revision, of course, and in 2008 paleontologist Anna Behrensmeyer published her reassessment of the AL 333 site. A pioneer in the study of how animals perished and became preserved in the fossil record (called taphonomy), she did not find any evidence of the catastrophic flooding event. Instead the clues in the rock presented a more perplexing scenario.

The bulk of Behrensmeyer’s study considered the geology of the hominin-rich site within the context of the surrounding area. At the time the AL 333 deposit was created about 3.2 million years ago, the local landscape had been transformed from a lakeshore into a plain crossed by meandering rivers and streams. Behrensmeyer wrote “The overall landscape that the A.L. 333 hominins inhabited, or perhaps occasionally traversed, would have been a relatively featureless, seasonally dry, grass-dominated plain (on a scale of kilometers) where the slight depression created by the dying channel may have hosted a different type of vegetation, including bushes or trees that were able to grow under conditions of more stable soil moisture.” Such areas may have acted as naturally-created paths for animals moving between the grasslands and more closed-in woodlands.

Within this ancient landscape, AL 333 site itself was a narrowed part of a branched, dried out channel system. The fossils found within it – from the hominins to scraps of crocodile, bovid, elephant, and fish – were deposited in a shallow dip within that system when water flowed through it again for a brief period. Once filled, this swale dried up again and the ancient, clay-rich soils became established on top of the fossil-bearing layers (coming significantly after the hominins, contrary to the findings of Radosevich and co-authors).

Behrensmeyer confirmed that the hominin bones were deposited when a dry channel was refilled due to seasonal flooding (perhaps in more than one event), but her reconstruction of the ecological setting makes it unlikely that the hominins were killed in a catastrophic flood. The site did not represent a ravine in which the group would have been trapped, as dramatized by Cameron and Groves, but was actually a small depression only about a foot and a half deep. Nor was the flow of water especially powerful. The flow left sediment which filled in the swale, “gently covering rather than scouring preexisting sediment and organic remains.” Additionally, the hominin bones and other fossils at the site were fragmentary and not in the well-preserved condition expected for animals suddenly overcome and rapidly buried. The bones found at AL 333 may have been scattered over a wider area on the plain and condensed into the dip, and, regardless of whether they died at the specific site or not, the A. afarensis individuals were already piles of bones by time they were buried.

Drowning in a flash flood can now be ruled out as a cause of death, even if a more gentle flood created the fossil deposit. What remains unclear is what killed the australopithecines in the first place. This is difficult to study because the AL 333 site is a mix of bones which could have been transported from the surrounding area. We may not be looking at the original crime scene. Frustratingly, Behrensmeyer does not go into detail about alternative scenarios in her paper, but comments she has made elsewhere have hinted that she and other researchers might pin carnivores as the culprits.

As cited in the geological paper, the dry channel environment the australopithecines perished in was likely used as a path by animals. If this is so, carnivores may have been prehistoric highwaymen on these natural roads, and Behrensmeyer has proposed that the AL 333 bones might be the leftovers of a massacre. A 2003 abstract about AL 333 presented at the Paleoanthropology Society meeting concluded “The evidence argues against death caused by an unusual flood or miring and leaves open the possibility of predation or another cause of sudden mass mortality.” This hypothesis popped up again in 2008, both in a Science News article and comments made during promotion for Lucy’s North American museum tour. In the latter article, especially, Behrensmeyer is quoted as saying that the australopithecines might have been the unfortunate victims of “surplus killing” by a saber-toothed cat or other large carnivore. In these circumstances, initially described by carnivore expert Hans Kruuk in 1972, predators kill far more prey than they can eat at one time. This has been observed among wolves, brown bears, and spotted hyenas, among other animals.

Research into the fate of the AL 333 australopithecines is ongoing, and, unfortunately, no fully-detailed analysis of the predation hypothesis has yet been published. I can only imagine that it is in the works. Nevertheless, we can still survey the available information to see if the AL 333 site is consistent with the activity of large predators.

The first hurdle is in determining whether there was any First Family to speak of at all. Many of the bones were found on the surface of the site after they eroded out of the rock, making it impossible to figure out how how they were laid down. There are are few bones which were found in situ, though, so those might provide some indications of whether there was one deposition event or multiple. If the bones were washed into the swale from their resting places in multiple events, then the individuals probably did not belong to a single group but instead died and were disarticulated in the same area.

It may also be worthwhile to reexamine the associated bones for signs as weathering. In one of her classic studies, Behrensmeyer showed that bones go through a characteristic sequence of breakdown when left exposed on the surface. If bones from AL 333 show different degrees of weathering it would indicate that they spent different amounts of time exposed and therefore do not represent a single catastrophic event. Radosevich and colleagues considered this in their study and did not report any signs of weathering, but it is worth reexamining the bones and scrutinizing new fragments which have been found since the initial excavations. It is entirely possible that the AL 333 australopithecines did not die all at once.

But, as we currently lack this level of resolution, let’s consider the idea that the australopithecines actually died as a group. If they were all victims to a predator, then the AL 333 assemblage would be a remarkable instance of surplus killing. In order for such events to occur, however, the predator must catch and kill the prey animals, and it is difficult to imagine a group of 17 or more australopithecines being killed at once by a single predator. If such an event actually transpired, it may have involved a group of predators, as has been seen among wolves in North America and the spotted hyenas of Africa that Kruuk observed.

The potential culprit is also unknown, although here the AL 333 bones may be a little more helpful. Previously Johanson has argued against the predation hypothesis by stating that the hypothetical predator did not leave characteristic toothmarks on the hominin bones, but not all carnivores consume prey in the same way. Spotted hyenas, for example, are bone crushers and probably would not have left very much at all behind. (Their prehistoric cousins at China’s Dragon Bone Hill, which did consume humans, left little more than skullcaps to the fossil record.) Cats, on the other hand, are the most hypercarnivorous of all carnivoran mammals, meaning that they specialize on the soft parts of bodies and largely eschew bones. (And, even in cases where cats consume bones, sometimes those parts can pass through the digestive system intact. In forensic experiments with leopards, Travis Pickering and Kristian Carlson found that parts of the baboon finger  sometimes passed through the cat’s system with soft tissue still intact. Delicate finger and toe bones may still make it into the fossil record even after they have been eaten!) Big cat kills often leave parts of the limbs, shoulders, hips, and skull behind while obliterating the spine and much of the hands and feet. The bones represented at AL 333 appear to be consistent with this, although it should be remembered that the grouping of bones may be biased by other factors – such as ease of transport – if the assemblage was washed into place.

If the AL 333 accumulation was created by a predator, then the pattern of damage appears to indicate that a big cat was responsible. As Behrensmeyer noted, the list of potential culprits includes saber-toothed cats. The wide-ranging Homotherium and Megantereon, as well as the false-sabercat Dinofelis, were all potential predators of australopithecines. Unfortunately, however, the way these cats caught, killed, and consumed their prey has been debated for over a century, and there are relatively few sabercat table scraps that allow paleontologists to investigate what these predators would have done to a carcass. One of the best collections of available clues comes from an accumulation of bones created about 20,000 years ago by Homotherium serum in Friesenhahn Cave, not far from modern-day San Antonio, Texas.

Described in the middle of the 20th century and reanalyzed in 1995 by anthropologists Curtis Marean and Celeste Ehrhardt, the cave site was a den of Homotherium serum. It contained abundant remains of both the sabercat – including Homotherium kittensincluding Homotherium kittensincluding Homotherium kittens – and juvenile mammoths. Skull elements of the mammoths were the most common, and many of their bones showed toothmarks left by the feeding cats. In particular, the damaged bones confirmed that Homotherium used its heavy-duty incisor teeth to nibble, tear, and scrape at carcasses that had already had most of the fleshy parts removed. (The frequency of toothmarked bones was similar to what was found in leopard dens.) While the Friesenhahn Cave and AL 333 sites are different in regards to their age, location, composition, and mode of accumulation, the remains of the AL 333 australopithecines are consistent with what Homotherium was capable of. Furthermore, if multiple australopithecines were slaughtered in a surplus-killing event then the cat would have had ample soft-tissue resources to feed on before having to scrape away at the bones (thereby explaining the apparent lack of direct damage to the bones which would have occurred late in carcass utilization).

Of course, a surplus killing scenario brings up difficult questions of its own. How could a group of between nine and seventeen australopithecines been massacred within moments of one another? Were some of the adults slain in the process of protecting the younger individuals? Was there a solitary hunter, or was the group attacked by a multiple assailants? (Lions live in prides, but the social behavior of sabertoothed cats continues to be debated.) Such a large kill would have undoubtedly attracted the attention of other predators and scavengers, as well, which leaves the question of how the bodies were processed by carnivores after the initial event.

What caused the demise of the AL 333 australopithecines remains unknown. The initial hypothesis – that they were a family drowned and buried by a flash flood – was simple and neat, but Behrensmeyer’s reanalysis of AL 333 forces us to throw out this narrative. Nor does the predation or surplus killing scenario presently have much direct evidence to support it. We may never know the exact circumstances of what killed these humans. If the AL 333 deposit represents an accumulation of bones washed off the nearby plain and mixed together – a good possibility, given the scrappy nature of other fossils found in association – then the circumstances by which each individual australopithecine died are almost entirely obscured. Marks on the bones left by a predator, scavengers, and weathering might give us some idea of what happened, but we can’t investigate the original crime scenes for crucial clues. For now, AL 333 remains one of the most important fossil human sites ever found, as well as one of the most mysterious.

*(This spurred a long-running debate over how different male and female A. afarensis actually were. Whereas Owen Lovejoy and co-authors have proposed that male and female A. afarensis only differed about as much as the sexes in our species as a way to support their particular hypothesis about the social behavior of early humans, J. Michael Plavcan and colleagues have called Lovejoy’s methods into question and reconstructed A. afarensis as exhibiting a greater degree of difference between males and females. We may have to wait for more complete A. afarensis skeletons before we can confirm the degree of sexual dimorphism in this species.)

Top image: The Australopithecus afarensis skeleton known as “Lucy.” Image from Wikipedia.


Antón, M., Salesa, M., Turner, A., Galobart, A., & Pastor, J. (2009). Soft tissue reconstruction of Homotherium latidens (Mammalia, Carnivora, Felidae). Implications for the possibility of representations in Palaeolithic art☆ Geobios, 42 (5), 541-551 DOI: 10.1016/j.geobios.2009.02.003

Anna K. Behrensmeyer (1978). Taphonomic and Ecologic Information from Bone Weathering Paleobiology, 4 (2), 150-162

Anna K Behrensmeyer (2008). Paleoenvironmental context of the Pliocene A.L. 333 “First Family” hominin locality, Hadar Formation, Ethiopia GSA Special Papers, 446, 203-214 : 10.1130/2008.2446(09)

Kruuk, H. (2009). Surplus killing by carnivores Journal of Zoology, 166 (2), 233-244 DOI: 10.1111/j.1469-7998.1972.tb04087.x

Marean, C. (1995). Paleoanthropological and paleoecological implications of the taphonomy of a sabertooth’s den Journal of Human Evolution, 29 (6), 515-547 DOI: 10.1006/jhev.1995.1074

PLAVCAN, J., LOCKWOOD, C., KIMBEL, W., LAGUE, M., & HARMON, E. (2005). Sexual dimorphism in revisited: How strong is the case for a human-like pattern of dimorphism? Journal of Human Evolution, 48 (3), 313-320 DOI: 10.1016/j.jhevol.2004.09.006

Radosevich, S., Retallack, G., & Taieb, M. (1992). Reassessment of the paleoenvironment and preservation of hominid fossils from Hadar, Ethiopia American Journal of Physical Anthropology, 87 (1), 15-27 DOI: 10.1002/ajpa.1330870103

Reno, P., McCollum, M., Meindl, R., & Lovejoy, C. (2010). An enlarged postcranial sample confirms Australopithecus afarensis dimorphism was similar to modern humans Philosophical Transactions of the Royal Society B: Biological Sciences, 365 (1556), 3355-3363 DOI: 10.1098/rstb.2010.0086

Go Further

Subscriber Exclusive Content

Why are people so dang obsessed with Mars?

How viruses shape our world

The era of greyhound racing in the U.S. is coming to an end

See how people have imagined life on Mars through history

See how NASA’s new Mars rover will explore the red planet

Why are people so dang obsessed with Mars?

How viruses shape our world

The era of greyhound racing in the U.S. is coming to an end

See how people have imagined life on Mars through history

See how NASA’s new Mars rover will explore the red planet

Why are people so dang obsessed with Mars?

How viruses shape our world

The era of greyhound racing in the U.S. is coming to an end

See how people have imagined life on Mars through history

See how NASA’s new Mars rover will explore the red planet