Why I’m skeptical about Lucy in the Skyfall

7 minute read

I have no trouble believing that Lucy might have fallen to her death. Why not? The Lucy skeleton has several features compatible with a lifetime of climbing, and other fossils attributed to her species, A. afarensis, have even more. So I am not primed to be skeptical of the conclusion of the new paper by John Kappelman and colleagues in Nature.

But I’m struggling to figure out why the authors decided to publish without addressing basic questions that any paleontologist would ask.

Lucy skeleton
Lucy skeleton. Credit: John Hawks CC-BY-NC-ND

The Lucy skeleton was recently on tour in the United States, beginning at the Houston Museum of Natural Science, and on this tour the bones were scanned in a high-resolution CT machine at the University of Texas. Studying the scans, Kappelman and his team decided that some of the fracture patterns are consistent with injury near the time of death, called perimortem trauma.

Their strongest evidence is the fracture pattern of the right humerus. The surface of the humeral head was crushed downward, with large chunks of the neighboring bone shattered outward and fossilized in place. Previous scientists have interpreted this damage as the product of fossilization and crushing under layers of sediment. Kappelman and colleagues examined the fragments and found two aspects that are indicative of a “green” fracture, made while the tissue was still fresh. Putting the fragments back together like a jigsaw puzzle, they saw that the entire proximal end of the bone first sheared off diagonally in a pattern called a spiral fracture. The fracture left some small shards of bone, only millimeters in size, still adhering to the fractured fossil. Kappelman and colleagues argue that if this breakage had happened with fossilization, after the soft tissue of the shoulder joint was completely gone, then these tiny shards of bone should be missing. Their presence suggests that the bone was broken near the time of death, buried with soft tissue still present, and fossilized in place.

It is not hard to find comparable proximal humerus fractures in modern clinical cases. For example, this one is from a paper on treating humeral fractures by Shane Nho and colleagues:

Radiograph of proximal humerus fracture from Nho et al. 2007.

Here is a diagram of the fragments of the Lucy proximal humerus from Kappelman and colleagues:

Lucy humerus fractures from Kappelman and colleagues 2016
Extended data figure 1 from Kappelman and colleagues (2016), showing fragments of humerus as distinct colors.

Comparing those images, you can see how the proximal humerus of Lucy resembles a fracture injury. Kappelman and colleagues build a broader case involving many other bones. Of the rest of the skeleton, the other obvious instance of damage from compression is the left distal femur, which has condyles that were jammed upward and laterally. Kappelman and colleagues show that the contour of the damaged bone is a mirror-image to the proximal end of the right tibia, the left being missing from the skeleton.

In all, they argue that some 20 bones of Lucy’s skeleton present evidence of perimortem fractures. Nearly every break present on the bones, including the broken edges of preserved fragments like the clavicle, they interpret as perimortem.

On the surface, it might seem like a watertight case. But as my Facebook feed began to pulse with messages from other anthropologists this morning, the authors defined perimortem fractures so loosely, every break in any fossil might be described as perimortem. If Lucy really had fractures on more than 75% of her preserved bones, she didn’t fall out of a tree, she fell out of an airplane.

As an example, consider the first rib. The authors do not illustrate this rib, but they argue that the rib has a “hinge fracture” at the neck, and several fractures at midshaft that were made at the time of death. They note that the first rib is almost never fractured, even in accidents where the other ribs suffer many fractures, because of the extensive soft tissue around it–dramatically, they call it a “a hallmark of severe trauma”. Their scenario for Lucy is that the extreme stresses on the shoulder at impact caused the clavicle to break, deflecting downward into the first rib and breaking it in multiple places.

But there is another process that breaks first ribs very commonly in the fossil record: Becoming a fossil.

I am not aware of any first rib from a Plio-Pleistocene hominin that is intact. The Nariokotome Homo erectus first rib appears to have a very similar pattern of breakage to the Lucy first rib. The first ribs from Dmanisi are broken, as are first ribs from Malapa and Sterkfontein. The first hypothesis to turn to for all these broken ribs is damage in the process of natural deposition and fossilization. Fossil bones are usually broken, and Lucy is no exception.

Does it matter that the authors describe some of these breaks as “hinge fractures”? Dry bones lacking the strength of their organic matrix often fracture in a straight line, like broken Greek columns, while a hinge fracture is one that changes in direction at one edge. When people see such a fracture that deflects in direction, they often attribute the fracture to fresh bone instead of dry bone.

But in reality, a fracture that passes through a cortical bone layer may deflect in a direction more perpendicular to the bone surface even if the bone is dry, beneath sediment, or mineralized. The paper uses the term “hinge fracture” many times, but neither defines it nor gives examples of how this evidences perimortem fracturing. If they examined a large collection of faunal fossils from the same context, they would find similar patterns, not from falling out of a tree but from simple post-depositional breakage.

The lack of comparison is a fatal deficit of the paper. The authors make no attempt to show that the pattern of fractures in the Lucy skeleton is different from that found in non-hominin bones from similar contexts. How many of the faunal remains from Hadar have similar fracture patterns? What about hominin bones from other sites?

In comments to Science News, Tim White provides a clear illustration of why comparisons are necessary to support the authors’ conclusion. Here in a photo is Lucy’s right proximal humerus compared to a fossil horse humerus:

Lucy humerus compared to horse

The horse did not fall from a tree.

Now, to be fair the horse doesn’t show the same degree of compression of the humeral head as Lucy. But like most paleontologists I’ve seen plenty of fossils that are much more deformed than the Lucy humerus and femur. Understanding whether that deformation is perimortem requires a close comparison with many other fossil specimens from similar contexts. None of that comparison is in this paper. No data concerning the preservation of non-hominin remains from Hadar or any other site are in this paper. If the authors want anyone to believe their analysis of the Lucy skeleton, they need to demonstrate that the fractures on the Lucy skeleton are different from those present on other fossils.

As it stands today, those paleontologists who are most knowledgeable about the Hadar fauna have been vocal: Similar fractures are common on other animal fossils including rhinos and elephants.

The paper provides no answer to this obvious criticism. It lacks minimal basic comparisons and therefore gives little reason to believe the authors’ conclusions.

The lack of data does not mean the authors’ hypothesis is necessarily wrong. I still think it is credible that Lucy may have had perimortem fractures, and I consider it possible that the humerus and femur may have compressive fracture damage consistent with a fall. But the authors frame their hypothesis as a just-so story, and only examine evidence consistent with it instead of looking for contrary evidence.

Skilled forensic anthropologists have only a limited degree of accuracy assessing perimortem versus postmortem fractures, when they are looking at buried skeletal remains that are only a few years old. With Lucy, we are talking about fractures that have been evident since the skeleton was found, and every previous scientist who examined them has concluded that they are likely postmortem damage consistent with the pattern on other fossils from Hadar. Showing that these fractures came from a particular traumatic event is going to take much better data than this paper provides.


Kappelman, John, Richard A. Ketcham, Stephen Pearce, Lawrence Todd, Wiley Akins, Matthew W. Colbert, Mulugeta Feseha, Jessica A. Maisano and Adrienne Witzel. 2016. Perimortem fractures in Lucy suggest mortality from fall out of tall tree. Nature doi:10.1038/nature19332

Nho, S.J., Brophy, R.H., Barker, J.U., Cornell, C.N. and MacGillivray, J.D., 2007. Management of proximal humeral fractures based on current literature. The Journal of Bone & Joint Surgery, 89(suppl 3), pp.44-58. doi:10.2106/JBJS.G.00648