Not the Marrying Kind

One fifth of Neanderthal DNA is found in the human genome … how much is in you? Flickr/suchosch

The idea that our ancestors interbred with other hominins like the Neanderthals or Denisovans is a truly captivating one.

The ancient DNA evidence would seem to have just emerged out of the blue, changing our thinking on how our evolution unfolded, and in a rather profound way. Is it really that amazing? Revolutionary even?

Like so many areas of science, few ideas are truly original. So much of what we write and think about today in anthropology was actually thought up during the 19th Century, well before a proper fossil record had even begun to emerge.

I guess its fair to say that there are only so many ways we can interpret our evolution, and lots of binary opposites exist and freely enter our imagination. So, much of it was dreamt up more than a hundred years ago, especially by luminaries such as Thomas Henry Huxley.

In the case of the recent evidence for interbreeding though, it’s the methods used and results found that are so remarkable. Who would have thought just a decade ago that we would have the draft genomes of several extinct human relatives sequenced? And so fast! The speed of technological change has been breathtaking.

We now have the genetic blue print for Neanderthals and the Denisovans (who ever they are).

But, we must be careful with this new information, for organisms are clearly much more than simply the sum of their genes.

Possessing a genome sequence goes only a small part of the way to understanding Neanderthal (or Denisovan) anatomy, physiology, ecology, behavior and evolution.

For example, the same gene mutation sequence might appear in the genome of Neanderthals and modern humans, but it may express different proteins in one species compared to the other, or may even be non-functional in either or both. So, its presence may not mean very much at all.

As tempting as it is to mine into the genome of these ancient hominins for simple answers to their – even our – evolution, we must be careful to resist the seductive simplicity of reductionism.

The idea of interbreeding has itself been around for some time – more than 100 years in fact – in varying guises, and is sometimes couched in terms like “degeneration,” “reticulation” or “assimilation”.

Yet, these terms have often come to mean interbreeding between long isolated (divergent) subspecies in human evolution, as was the view of my late postdoctoral supervisor Phillip Tobias.

He felt that Homo erectus, modern Homo sapiens, and everything in between, were simply subspecies (i.e. temporal or geographic variants of a very long lived single species, Homo sapiens).

Interbreeding on a large scale lies at the core of the so-called “Assimilation” hypothesis set out in the 1980s by Fred Smith and others to explain the origin of modern humans. It’s an idea that budded off from the “Multiregional” scenario developed by Wolpoff, Thorne and Wu in the early 1980s. The key difference is the strongly dominant role Africa played in our evolution, recognised in the former idea, not the latter.

Assimilation is one of four scenarios proposed during the 1970s-1980s to explain the geographic location, timing and process by which our kind (Homo sapiens or sometimes modern Homo sapiens) evolved.

Its proponents believe that modern humans exiting Africa perhaps 60,000 years ago and widely interbred with (assimilated or absorbed) the other (archaic) hominins they encountered along the way. Again, it was a process within a single species.

Supporters of Assimilation – like Fred Smith and Eric Trinkaus – have seen evidence for interbreeding in the form of anatomically mixed fossils for decades, in Africa, Europe and Asia. My late and dear friend Alan Thorne also thought the same for some of the European fossil records, as did the late Jan Jelinek.

Claims for hybrids such as the Skhul (Israel), La Ferrasie (France), Abrigo do Lagar Velho (Portugal), Mladec (Czech Republic) and Pestera cu Oase (Romania) remains, among others, have abounded in the literature. Yet, most anthropologists have been, and remain, deeply skeptical.

I have myself – in the past – toyed with this hypothesis. But, alas, I no longer find much value in its capacity to explain the patterns we see in the fossil record. There are, to my mind, far simpler explanations for fossils of mixed form, that make more sense within the context of modern evolutionary biology.

The fossil evidence for interbreeding has been widely regarded to be too ambiguous and difficult, if not impossible, to test. It also represents an example of an ad hoc explanation.

Ian Tattersall and Jeffrey Schwartz published a paper in Proceedings of the National Academy of Sciences USA (PNAS) in 1999 (vol. 96, pp. 7117-7119) addressing this very question, taking issue with claims of hybridization for the Lagar Velho child. Their article has become somewhat of a classic, concluding:

“…the Lagar Velho child’s skeleton is a brave and imaginative interpretation, of which it is unlikely that a majority of paleoanthropologists will consider proven. The archaeological context of Lagar Velho is that of a typical Gravettian burial, with no sign of Mousterian cultural influence, and the specimen itself lacks not only derived Neanderthal characters but any suggestion of Neanderthal morphology. The probability must thus remain that this is simply a chunky Gravettian child, a descendant of the modern invaders who had evicted the Neanderthals from Iberia several millennia earlier.”

Even if the genomic evidence is correct that the ancestors of Eurasians interbred with Neanderthals, and some of them also with the Denisovans, this doesn’t prove, prima facie, any fossil to be a hybrid.

Studies of hybridisation in living primates show that the offspring fail to show a predictable pattern of physical features: some look like parent species 1, some like parent species 2, and others possess an unpredictable mix of features from both.

Moreover, the signs of interbreeding disappear within a generation or two of crossing back with one of the original parent species.

So, the chances of finding a hybrid in the fossil record seem to me to be extremely remote.

I’ve spent many years looking for hominin fossils and they are near impossible to find. Really rich localities like the Cradle Caves in South Africa, Lake Turkana region in Kenya, Afar region of Ethiopia, and Atapuerca in Spain, are extremely rare.

It’s hard enough to find fossil sites, let alone ones that produce hominins, let alone ones that might contain hybrids! The odds are simply too large to consider likely.

Interpretations of the latest Neanderthal genome sequence published in Science in January of this year by Sriram Sankararaman and co-workers confirms this. Their work shows pretty firmly that interbreeding was likely to have been between distinct species: that is, between Homo neanderthalensis and Homo sapiens. Not simply populations within a species.

Supporters of all four models proposed to explain our evolution –Multiregionalism, Assimilation, Out-of-Africa with some interbreeding and Out-of-Africa replacement – have all claimed their models to be supported by the new evidence for interbreeding.

Only one thing to say here: balderdash! It simply can’t be the case that all four, mutually exclusive, models can be supported by the same dataset. If they are, then the data for interbreeding are completely meaningless!

In science, strong hypothesis make very specific and testable predictions. This also makes them, in principle, easier to refute. The new Neanderthal evidence is very specific, and it simply can’t be interpreted as supporting all four ideas.

First published in February 2014.