Reply to: Charlier et al. 2018. Mudslide and/or animal attack are more plausible causes and circumstances of death for AL 288 (‘Lucy’): a forensic anthropology analysis. Medico-Legal Journal 86(3) 139–142,2018

Introduction

Lucy (A.L. 288-1, Australopithecus afarensis) (Figure 1) ranks as perhaps the most famous as well as one of the most complete fossil hominin skeletons ever discovered. It has been the focus of detailed study for 45 years and is at the centre of a long running debate about whether early hominins combined bipedalism and climbing behaviors. ¹ A detailed description of the skeleton by Johanson et al. ² concluded that, with only a few exceptions, the breaks in the skeleton were likely created postmortem by the sorts of processes usually associated with fossilisation. In contrast, we argued that a subset of the breaks are instead consistent with perimortem fractures, breaks that occurred at or near the time of death. ³ We focused primarily on compressive breaks at several of the long bone joints, and presented criteria for assessing if a break occurred perimortem – there is no sign of healing, the edges of the broken bones are sharp and clean, bone fragments and slivers of bone are preserved at the site of the fractures, and the fractured articular end of the bone was driven into itself by an impact with the articular surface of its paired bone – all of which suggest that bone-into-bone compressive fracturing occurred by impact while the bone was fresh and the periosteum and joint capsule were intact. ³ OPEN IN VIEWER

Perimortem or postmortem?

The first question that should be evaluated is the nature of the breaks themselves: are they perimortem, as we suggested, or postmortem, as long maintained? At first, Charlier et al. ⁴ appear to agree with us by stating that:

According to the pattern of the fractures, this mechanism would have occurred on Lucy’s corpse at the moment of death or shortly after (peri-mortem period), at least on ‘fresh’ bones, i.e. wet and without loss of collagen substrate. (p 2)

Mudslides

Once these data suggested that a subset of the breaks in Lucy’s skeleton were perimortem in nature, we evaluated a large number of hypothetical mechanisms that could potentially produce this pattern of fractures. ³ After detailed consideration, we concluded that an impact following a fall from considerable height was the most likely mechanism because of the close match between the nature of the fractures in Lucy’s skeleton and those documented in numerous clinical cases, the absence of any similarity to what is the typical preservational state witnessed in vertebrate fossils from the discovery site and the Hadar Formation, and other data from the discovery bed. Charlier et al. ⁴ do not directly dispute our preferred hypothesis or the clinical data that we cite. Instead, they argue that two from among the many other possible mechanisms that we investigated (and concluded to be less likely than a fall from height) are more likely to have created the fractures even though they contradict themselves as to whether the breaks are perimortem or postmortem in nature.

The first of these mechanisms, a mudslide (that the authors variously include under “flooding/mudslide/landslide” [p 2]), is one that we considered in the Supplementary Information published with our article. ³ Charlier et al. ⁴ maintain that we state “that trauma lesions cannot be those [sic.] due to flooding” (p 2) but this statement is not accurate; at the close of our detailed assessment of this mechanism, we do not state that it cannot be a cause but instead rank it as less likely than a fall from considerable height. The bone-into-bone compressive fractures oriented along the long axes of the limbs that we identified at Lucy’s shoulders, knees, and ankle joint, along with spiral fractures in the humeral and tibial shafts, together suggest that the arms and legs were outstretched at the time of injury as is known to occur when a conscious victim braces their limbs during a fall in anticipation of impact. This configuration of abducted and extended limbs in turn drives the bones into one another at their articulation. Our later work ⁵ examined the cross-sectional properties of Lucy’s long bones and showed that her limbs have especially thick cortical bone. This observation in turn suggests that high forces would have been required to produce, for example, the four-part proximal humerus fracture in the right shoulder, itself a unique signature of the proposed mechanism of injury. ⁶

Although the overall fracture pattern documented in Lucy’s skeleton could potentially be produced in a conscious victim in flood waters by blunt force trauma at impact with floating or stationary objects, it is very unlikely to occur in an unconscious or already drowned flood victim (conditions which follow rapidly once a victim is caught in a violent flood) because the limbs would be limp, not actively outstretched. The authors fail to describe how the action of rolling a body could produce this particular subset of compressive fractures. Furthermore, in floods with energies high enough to produce severe blunt force trauma, the limbs and head are frequently dismembered and, once separated, often widely dispersed by the high velocity flow.^7,8 Charlier et al. ⁴ curiously dismiss an article ⁹ that we cited about flash flood mortality statistics because it is “old” (p 2), but the date of an article is hardly a reason to dismiss its conclusions. The fact that skeletal elements from all four of Lucy’s limbs, her torso, and skull were recovered from a single sedimentary unit (see below) within a limited area strongly suggests that dismemberment was very unlikely to have occurred but, if it had, that the flow was not powerful enough to widely disperse the body parts. These observations draw into serious question the possibility that a mudslide or flood produced these fractures.

Sedimentological data from fossil locality A.L. 288 where Lucy was recovered that were overlooked by Charlier et al. ⁴ also fail to provide any evidence for a mudslide or flood. We discussed these data in detail in our article. ³ The unit that contained Lucy’s skeleton is reported as a 0.5 m thick crevasse splay sandstone, not a mudstone, deposited in a shallow distributary channel in a region of low vertical relief on the distal floodplain in proximity to a lake. Johanson et al. ¹⁰ reconstruct this setting as “a slow-moving stream” (p 380), not a flood, mudslide, or high energy deposit. It is interesting that Coppens, a co-author on this article ¹⁰ and now a co-author with Charlier et al., ⁴ apparently disagrees with this earlier conclusion. Charlier et al. ⁴ cite Johanson et al. ¹¹ (their reference 2) as the source for their assertion that Lucy was dispersed over “a slight distance of almost 50 to 100 m on a sandy layer of 40 to 50 cm thick” (p 2) but a close reading reveals that there is no mention of this point in that article, and we are not sure why it was cited. Charlier et al. ⁴ further suggest that impacts with “boulders and fallen trees which are carried along by fast-flowing waters” (p 2) of the flood or mudslide or landslide could have created the fractures; if so, where are the boulders? Nothing coarser than medium-grained sand (particle size 0.25–0.5) is preserved in the discovery bed, and some of this sand still adheres to Lucy’s right proximal humerus.

Another inconsistency with the hypothesis of a flood or mudslide of sufficiently high energy to produce the compressive fractures preserved in Lucy’s joints is found in the state of preservation of the other fossils collected from the same sandstone at fossil locality A.L. 288 where Lucy was discovered. Johanson et al. ¹² state that the “sandstone horizon [is] characterized by superb fossil preservation yielding such delicate items as crocodile and turtle eggs, rodent skulls and even crab claws” (p 562). If such “delicate” items had been transported by a high energy flood or mudslide along with “boulders and fallen trees”, they would have been destroyed rather than preserved in such a superb state. The fossilised remains of monkeys, bovids, squirrels, equids, and elephants were also recovered from locality A.L. 288. We have examined these original fossils (not replica casts) and observed that they do not preserve any evidence of high-energy breakage; they are identical to the preservational state witnessed in the other fossils from the Hadar Formation and differ dramatically from the compressive fractures seen in Lucy. ¹³

If instead of flood or mudslide damage, the compressive fractures in Lucy’s skeleton were produced by taphonomic processes associated with, for example, burial in the sediments, we would expect to witness similar types of compressive breakage in the other fossils recovered from the discovery unit. This is not the case. The contrast between all of the other A.L. 288 fossils and Lucy is marked, and this dramatic difference makes the compressive bone-into-bone fractures preserved in her skeleton all the more unusual. In summary, there is no supporting evidence from the sedimentological context of the discovery bed or from the other fossils recovered from locality A.L. 288 for a flood or mudslide, making it highly unlikely that this mechanism was responsible for producing the fractures preserved in Lucy’s skeleton.

Crocodile attack

Charlier et al. ⁴ next evaluate two marks on Lucy’s skeleton that they contend provide evidence of a crocodile attack. Apparently using, but not citing, the widely known assertions by Johanson et al. ² that “[a]n oval puncture defect is present in the inferior aspect of the [ischial] tuberosity and is possibly the result of carnivore activity” (p 424), and that “[t]he superior surface of superior pubic ramus bears a second carnivore puncture mark” (p 424), the latter point also noted by Johanson et al. (p 380), ¹⁰ Charlier et al. ⁴ propose that these are crocodile tooth puncture marks and state that:

[t]heir morphology is fully consistent with carnivore tooth marks, including crocodile bites (circular depressions with a conic sectional shape, well-defined margins with punched-in appearance). (p 2)

In order to illustrate and measure these two features, Charlier et al. ⁴ provide in their Figure 2 images of 3D scans taken of a replica resin cast of Lucy’s innominate but how the scans were obtained is not described. Unfortunately, the authors appear unaware of the fact that the renderings are mirror images of the fossil; their Figure 2(a) to (c) shows a right innominate even though the actual innominate is a left (compare their Figures 1 and 2), and the images are also incorrectly described in the figure caption. OPEN IN VIEWER

Close examination shows that the 3D scan of the ischial feature provided by Charlier et al. ⁴ in their Figure 2(b) is of insufficient resolution for an accurate diagnosis. We have studied the original fossil, and in our Figure 2(a) illustrate the actual left innominate (A.L. 288-1ao) with close-up stereo photographs of her ischium. We also provide comparative stereo photographs of the ischia from a human (Homo sapiens, Figure (2b)) and chimpanzee (Pan troglodytes, Figure 2(c)), both housed at the University of Texas at Austin. Hominin and hominoid ischial tuberosities are often marked by sometimes numerous single and paired foramina, and Figure 2 shows that ischial foramina of Lucy (Figure 2(b): single foramen in centre in stereo; second foramen noted by arrow), the human (single foramen), and the chimpanzee (paired foramina) are similar in size and shape. There is no evidence of any surface damage around the perimeter or within Lucy’s foramina or those of the other two specimens that might suggest that these are puncture features. These comparative data demonstrate that Lucy’s ischial foramina are normal and within the range of variation for hominins and hominoids. It is clear that these features are not tooth puncture marks.

In contrast, we agree with Johanson et al. ¹⁰ that the feature on the superior surface of Lucy’s pubis is a puncture mark. We discussed this feature in detail, ³ and it is surprising that Charlier et al. ⁴ do not acknowledge this fact. Once again, the 3D scan provided in their Figure 2(a) is of insufficient resolution to accurately diagnose this feature. Our Figure 3 is a stereo photograph of Lucy’s actual superior pubic ramus and shows that, contra Charlier et al., ⁴ this is not a “circular depression[s] with a conic sectional shape” (p 2). It is instead a tiny ∼2 mm long shallow puncture feature framed by linear cracks in the surface of the bone that caused it to fold and flex into itself by about 0.4 mm in depth, demonstrating that it most likely occurred when the bone was fresh, one of our criteria for a perimortem fracture. OPEN IN VIEWER

After their detailed discussion of these features and their proposal that a crocodile attack is a possible cause of Lucy’s death, Charlier et al. ⁴ once again reverse themselves and in the conclusion state that:

There is no evidence to confirm that these marks were made peri-mortem – they could just as likely have been inflicted by some natural means post-mortem (falling rocks, skeleton being tossed around in water/mud and impacting hard objects, etc.). (p 3)

Although it is possible that the puncture mark on the pubis was created by a tooth, our detailed examination of Lucy’s complete skeleton ¹³ revealed that this tiny feature is the only puncture mark on the entire skeleton. The absence of any other puncture marks, including opposing tooth marks, contra Charlier et al., ⁴ draws into question their suggestion that this feature is a crocodile tooth puncture mark because crocodile attacks are usually associated with numerous puncture marks produced by multiple crushing bites along with a tendency for the body to be ripped apart. ¹⁴ If the diminutive Lucy had been in the mouth of even a small to medium-sized crocodile, given their large number of teeth, we would expect to find multiple puncture marks along with evidence of crushing. The absence of any other puncture features and crushing on Lucy’s skeleton dramatically lowers the likelihood that this tiny isolated puncture fold feature represents evidence of a crocodile attack. We would also be surprised for Lucy’s skeleton to be as complete as it is if she had been attacked or scavenged by a crocodile. We hypothesised that this impact feature, like those found in the other parts of her skeleton, was created when the body hit the ground. ³

Conclusion

While we welcome the contribution by Charlier et al. ⁴ to the debate about the possible cause of death of perhaps the most famous hominin fossil known to science, we respectfully suggest that a more thorough critique of our work on Lucy should be based on an examination of the actual fossil rather than replica casts. A meaningful critique must also consider all data, including a detailed analysis of the nature of the breaks (perimortem or postmortem), the geological context where Lucy was discovered, the preservational state of Lucy and the other fossils recovered from the A.L. 288 fossil locality, and the comparative anatomy of closely related species. When considering all of these data together, we maintain that a fall from considerable height and the resulting impact produced by a vertical deceleration event is the most likely explanation for what we identified as perimortem bone-into-bone compressive fractures in Lucy’s skeleton.