Lone Survivors (41 page)

I have been working on trying to date the Broken Hill skull more precisely for about fifteen years, with a number of collaborating scientists and even mineralogist colleagues at the Natural History Museum. The main methods we have been using (see chapter 2) are ESR (electron spin resonance) on a tooth fragment from the skull and uranium-series dating from various bones and sediment from the site. It would normally take great care and courage to remove an enamel fragment from one of the precious teeth of the Broken Hill skull, but a previous accident worked in our favor in this case. Some unknown member of staff or a visitor had accidentally knocked the corner from one of the molars and, rather than reporting it, had simply glued the piece back on! When one of my eagle-eyed colleagues, Lorraine Cornish, spotted this, she dissolved the glue and we had the perfect fragment of enamel to date. But one of the unknowns in ESR dating is the degree of radiation received by the enamel fragment since it was buried, and this has to be reconstructed from site data, which in the case of Broken Hill was severely lacking, with the complete destruction of the original location by mining. However, chunks of sediment and bone breccia were saved from the mine, in some cases because they also contained interesting minerals, and others were collected after the discovery of the skull, so these could be measured to help reconstruct the burial environment.

One of the worst-case scenarios in ESR dating is underwater burial, since water interferes with the accumulation of the ESR signal. There was plenty of evidence from the mine records that the level where the skull had been found had to have water regularly pumped out as it actually lay below the existing water table. However, two other clues I already mentioned became critical here. First, the “skin” that the miners thought they had found was actually layered sediment impregnated with minerals, which must have been laid down relatively horizontally and could not have formed under water. Records made at the time of discovery stated that it lay near the skull and tibia, at a steep angle, suggesting that it had fallen from higher up. Second, we know that the skull was covered in, and even contained, many bones of small animals such as shrews, and the mine records clearly document layers of small mammal bones at a much higher level, far above the place where the skull was actually found. So it seems likely that quarrying at the base of the sediments led them to constantly collapse down, and the skull was almost certainly derived from the higher levels of the site, above the water level.

Now, when we factor everything together, the ESR signal in its tooth enamel suggests that the skull is actually between 200,000 and 300,000 years old. And two other age estimates from the femur fragment and the so-called skin suggest the real age could be closer to 200,000 rather than 300,000 years. It is possible that the skin accumulated above the level of the skull and femur before they all collapsed down, so it could be somewhat younger than them, but certainly there is nothing here to indicate that this assemblage is anything like 500,000 years old. Such a surprisingly young age is not contradicted by the artifacts known from the site, which have early Middle Stone Age affinities, nor from studies by the paleontologists Margaret Avery and Christiane Denys of the small mammal accumulations closely associated with the skull, which in species represented match those known from African sites like Twin Rivers, dated in the range 200,000 to 300,000 years old. If the Broken Hill skull, one of the best-preserved relics of
Homo heidelbergensis
, is actually less than 300,000 years old, what does this mean for our models of human evolution and for the origin of our species?

The result does have important implications for our reconstructions of recent human evolution because, as I explained, the Broken Hill fossil has been a cornerstone of the assumed gradual evolutionary sequence from archaic to modern humans in Africa. Dating Broken Hill to about 500,000 years placed it some 300,000 years older than the first known modern humans, allowing plenty of time for the necessary changes to happen. But the new dating makes Broken Hill only somewhat older than Omo 1 at about 195,000 years, and perhaps close in age to the more modern-looking Florisbad and Guomde fossils. This would imply either that there was a very rapid evolutionary transition to the earliest modern humans about 250,000 years ago, or that Africa contained great variation in its human populations at that time. Could that variation have even extended to the coexistence of different human species? We already discussed the puzzling variation between the two apparently contemporary Omo Kibish fossils 1 and 2, with skull 1 looking decidedly modern and skull 2 having a more primitive braincase, with an angled back, and we mentioned the rather archaic-looking rear of the Herto adult skull. Moreover, there are other primitive-looking fossils in Africa (such as those from Ngaloba and Eyasi in Tanzania) that overlap the dates we currently have for the oldest modern-looking humans, and I will discuss a particularly striking example next. All of this means that I am reconsidering many of my previous views on the origin of our species in Africa, and I now think we need to talk about origins, rather than a single point of origin.

I pointed out in the previous chapter that the nature of the manufacturers of Paleolithic tools from many parts of Africa remains completely unknown, since there are no associated fossils. This is especially true of artifacts from West Africa, where the oldest known fossil, from the Iwo Eleru rock shelter in Nigeria, is thought to be less than 15,000 years old. This poorly preserved skeleton was excavated from basal sediments at Iwo Eleru in 1965 by the archaeologist Thurstan Shaw and his team and was associated with Later Stone Age tools. That latter fact alone would indicate a relatively young age, and a radiocarbon date on a piece of charcoal suggested an age of about 13,000 years. The skeleton, and particularly the skull and jaw, was studied in 1971 by Don Brothwell, my predecessor at the Natural History Museum, and he argued that while the specimen could be related to recent populations in West Africa, it actually looked rather different from them. I studied the skull for my Ph.D., with surprising results. I also found that it did not closely resemble recent African populations, but in its long and low shape it was actually closer to early moderns such as those from Skhul, and even to more primitive specimens such as Omo 2. This was decidedly odd for such a young skeleton, and so I recently collaborated in a new study of the specimen with the archaeologist Philip Allsworth-Jones, the dating expert Rainer Grün, and the anthropologist Katerina Harvati. We first checked with Thurstan Shaw whether there were any hints that the skull could have been much older than previously suggested, and there were none. With the help of the Nigerian archaeologist Philip Oyelaran, I obtained a fragment of bone from the skeleton and passed it to Grün in order to check its age directly. His determination from a direct uranium-series age estimate is that the bone is unlikely to be older than 20,000 years, consistent with the stratigraphy and associated archaeology and radiocarbon date.

Finally, could Brothwell and I have been wrong about the unusual shape of the skull? Harvati used state-of-the-art geometric morphometric scanning techniques on an exact replica of the skull (which is now in Nigeria) and found, as we did, that it was quite distinct from recent African crania, and indeed from any modern specimen in her comparative sample. Her results placed the skull closest to late archaic African fossils such as Ngaloba, Jebel Irhoud, and Omo 2—all thought to be at least 140,000 years old. So what does this mean? Because of the poor preservation of Pleistocene bones in West Africa, we have no other data on the physical form of the inhabitants of the region during the whole of the Pleistocene, so we have to be careful in interpreting an isolated specimen such as Iwo Eleru. But it does not seem to be diseased or distorted, and does indeed seem to indicate that Africa contained archaic-looking people in some areas when, and even long after, the first modern-looking humans had appeared. Support for this view comes from the work of the anthropologist Isabelle Crevecoeur. Her restudy of the numerous Ishango fossils from the Congo showed that these Later Stone Age humans were similar to Iwo Eleru not only in age but also in the surprisingly archaic features found in their skulls, jaws, and skeletons.

African fossils Ngaloba (Laetoli H.18,
Top
) and Iwo Eleru (
bottom
). They resemble each other, despite Iwo Eleru being less than 20,000 years old, compared to Ngaloba's 140,000-year-old archaic features.

Africa today has the greatest internal genetic variation of any inhabited continent, and its skull shapes show the highest variation. This is usually attributed to its greater size, larger ancient populations, and deepest time lines for humanity. But could those time lines go back even farther than we thought? Did the early modern morphology evolve gradually and then spread outward from a region like East Africa, completely replacing archaic forms within Africa and then outside (as mtDNA data would suggest)? Or could there have been a version of assimilation or multiregional evolution within Africa, with modern genes, morphology, and behavior coalescing from partly isolated populations across the continent? Given its huge size, complex climates, and patchworks of environments, Africa could have secreted distinct human populations just as easily as the rest of the inhabited world. So was the origin of modern humans there characterized by long periods of fission and fusion between populations, rather than representing a sudden single event? And was the replacement of the preceding late archaic peoples not absolute, so that they were partly absorbed by the evolving moderns rather than completely dying out? In which case, did early
Homo sapiens
forms, and even the preceding species
Homo heidelbergensis
, survive alongside descendant modern humans?

This might explain archaic aspects in the shape of the Herto, Omo 2, and Iwo Eleru crania. In part they resemble archaic crania like Broken Hill, assigned to
Homo heidelbergensis,
so is this mosaic anatomy just a primitive retention from more ancient ancestors, or is it a sign of gene flow from contemporaneous African populations that still retained such features? My gut feeling is that some (but clearly not all) of the “ancient” DNA markers being picked up outside of Africa and used to argue for gene flow from non-African archaics will turn out to be traces of admixture that had actually happened in Africa. (A good example of this is the microcephalin gene discussed earlier.) Those traces were then carried from there in the modern human dispersal(s), followed by the operation of selection and drift on those populations, producing frequency changes in the genes when comparing the groups with each other and with their African counterparts. So while some archaic genes certainly were picked up by interbreeding outside of Africa, some were also acquired before the exodus, and yet others could even have been added in Africa, after it.

We still lack the amount of genetic data for African populations that we have for people from Europe and North America, but Africa is beginning to catch up. Charla Lambert and Sarah Tishkoff analyzed thousands of samples to reveal several deep and ancient population clusters, and, as we saw, Michael Hammer and his colleagues found evidence of archaic genes in three samples of moderns, but especially in West Africans. Now they have taken this work further by analyzing about half a million bits of genetic coding in samples of Mandenka (Senegal), Biaka pygmies (Central African Republic), and San (Namibia). They found strong evidence for split times of more than 100,000 years, predating the exodus from Africa, and they detected evidence of ancient admixture (with unknown “archaic” human groups) in both the Biaka and San. Philipp Gunz and his colleagues also recognized this from the great variability they found in late archaic/early modern African crania, as this edited extract of their conclusions shows.

Our fossil AMH [anatomically modern human] data suggest that before there was isolation by distance [=drift] from Africa, there already existed (at least temporally) isolation by distance within Africa. Seemingly ancient contributions to the modern human gene pool have been explained by admixture with archaic forms of
Homo
, e.g. Neanderthals. Although we cannot rule out such admixture, the proposed ancestral population structure of early AMH suggests another underestimated possibility: the genetic exchange between subdivided populations of early AMH as a potential source for “ancient” contributions to the modern human gene pool. Any model consistent with our data requires a more dynamic scenario and a more complex population structure than the one implied by the classic Out of Africa model. Our findings on neurocranial shape diversity are consistent with the assumption that intra-African population expansions produced temporarily subdivided and isolated groups. Separated demes (population subdivisions) might have partly merged again, whereas others left Africa at different times and maybe using different routes, and still others probably also remigrated to Africa.