Sunday 20 January 2013

Before the Dawn Horse

Horses are surely among the most familiar large herbivorous mammals. Yet, compared with the deer and cattle families, the horse family is remarkably small. It contains only seven living species: the horse itself, three kinds of ass, and three kinds of zebra. Yet there used to be many more, and, if we include all the extinct species, the family starts to look quite large.

Indeed, the story of horse evolution, with its steady progress towards the modern animals, with only one toe on each foot, is one that's particularly familiar. With the possible exception of the line of ever-more-upright apes leading towards modern humans, the steady line of horse evolution may be one of the most well-known, and oft-repeated, images in mammalian evolution. Of course, it's not really a line, more of a bush, with multiple different kinds of horses living alongside one another, but the general pattern is accurate.

The horses either evolved from, or are very closely related to, another family of animals called the palaeotheres. These were more or less dog-sized animals, and probably had three or four toes on each foot, just as the earlies horses did. Their exact classification, and quite how they relate to horses, isn't entirely clear. That's partly because, for most of them, we don't have very complete skeletons, and there are different opinions among scientists as to which features of those skeletons are important, and which aren't.

The paleotheres didn't die out when horses first appeared, but that's largely because they were on different continents at the time. Horses were in North America, while palaeotheres were found only in Europe and Asia. By the time horses reached Asia, in fact, the palaeotheres had long been extinct, having disappeared for other reasons, so they never got the chance to directly compete. At the other end of their evolutionary history however, it may have been a different matter.

In fact, some of the very earliest horses were European - Pliolophus, for example, lived in Britain. These early horses were so remarkably similar to the palaeotheres of the day that there has been considerable debate as to where the dividing line actually is. For example, Pliolophus, along with its more famous American cousin, Eohippus, was long considered to be an example of Hyracotherium, a palaeothere. Indeed, lot of books use the name Hyracotherium when they mean Eohippus, for exactly this reason - until fairly recently they were thought to be the same thing.

So there's a reasonable chance that the earliest palaeotheres were themselves ancestors of the first horses. But how far back can we trace the origin of the palaeotheres themselves? Part of the answer may lie in a genus of fossils named Pachynolophus. There's no doubt that Pachynolophus is a primitive member of the order of animals that includes, among other things, the horses and the palaeotheres. But quite how it is related to those animals is unclear. Some studies have concluded that it's a very primitive horse, others that it's a palaeothere, some that it's somewhere in between, and one that it's actually closer to tapirs than anything else.

A recently reported excavation in the south of France may shed some light on this. Laure Danilo and co-workers uncovered a site, next to a vineyard roughly between Carcassone and Narbonne with around seventy Pachynolophus fossils. Indeed, there wasn't much else, which may suggest that there was something special about the site that caused an unusual number of this particular species to die there - albeit probably over a long period of time. They date the site to around 53 million years ago, which is a considerable way back through the Age of Mammals, at the beginning of the Eocene, and crucially, about the time that the very first palaeotheres appeared.

The sea level was higher in those days, so that the site would have been a pleasant coastal area, even warmer and more tropical than the south of France is now. Relevant to that is the fact that Europe was, back then, much more clearly distinct from Asia than it is now; perhaps an island continent, or perhaps even a large chain of islands, like present day Indonesia. That means that whatever was going on there at the time was isolated from much of the evolution elsewhere. So it is probably specifically Europe, rather than Eurasia in general, that is the true point of origin of the horse-like animals - first the palaeotheres, and later, the horses themselves.

The skulls of these animals were about 13 cm long (five inches) so we can infer that the animal as a whole was about the size of a dog. That skull looked somewhat horse-like, but with a much narrower snout, which, taken together with the shape of the teeth, may mean that it was more suited to browsing on soft herbs than to grazing on grass. The researchers consider them to be a new species, which they have named Pachynolophus eulaliensis, for Sainte-Eulalie, where the fossils were found.

An interesting point is that, although the researchers believe that all the fossils are similar enough to be confident that they belong to the same species, they also identified two distinct forms among them. One was relatively large and solidly built, and the other much more slender. If they're right about them being the same species, a plausible explanation for that is that they represent different sexes. It's not necessarily the male that's the larger one, because there are many mammals for which the reverse is true, but when we're talking about something horse-like, it probably is.

A detailed analysis of the shape of the animal's skulls, compared with similar features in a range of other early horse-like fossils, confirms that they aren't actual horses (or at all tapir-like), as we'd probably expect, given their age. It's rather less certain what they actually are, since the fossil turns out to be quite different from other Pachynolophus species. But then, being the oldest one, that's rather what we'd expect - it's a primitive form, much closer to the ultimate origin of the group, one that had yet to develop the distinctive features that marked out the later families.

The study concludes that this animal, or at least a very close relative, was an ancestor of the known palaeotheres. Indeed, it goes as far as to say that they actually are palaeotheres, among the first of their kind, perhaps closely related to whatever Eohippus itself evolved from. The overall picture is still far from clear, but this study argues that it's all to do with France.

The south of France, they argue, was the home of the very first horse-like animals, ones that could, for the first time, be distinguished from the ancestors of animals like tapirs and rhinos. Some of them headed north, towards places like Britain, where they may have developed into true horses. Those that stayed behind in southern France evolved instead into the palaeotheres. Early northern European palaeotheres, such as Hyracotherium, are either descended from the earliest French ones, or perhaps represent an earlier migration, before the proto-horses. Either way, all the northern European forms - horses and palaeotheres alike - died out, with some of their descendants heading across Asia to America, and others being replaced by the proper French sort.

Can we really be that precise? It's probably still ambiguous, since there seem to have been quite a few new species popping up at the time, and we're at the mercy of which ones happen to have fossilised first. But if horses really owe their ultimate origin to the French, it could be because France was once an island.

[Painting by Heinrich Harder, copyright expired]


  1. As you wrote, it's a bit arbitrary to tell paleotheres apart from "true" equids, due the lack of pieces in this fascinating puzzle, but I think it's more reasonable to define equoids/equids as the branch that evolved in North America, while Paleotheriids was the endemic branch that evolved in the Eocene European Archipelago. Paleotheres became extinct after the Grand Coupure, while horses evolved in North America and reached Asia in the periods when Beringian isthmus was exposed.

    1. Yeah, the arbitrary and seemingly controversial nature of the division caused me a lot of headaches while writing this one. In the end, I just went with the definition used in the paper, and mentioned (I hope clearly enough) that it's all arguable either way. For instance, /Pliolphus/. If we define equoids as the Equidae+Palaeotheriidae clade, then they're clearly in there somewhere, but one could argue (as people have) about which side of that line they're on. It doesn't help that there's been so much synonymisation going on here, and many of the genera seem to be para- or even polyphyletic.

  2. Very well written and interesting article. I linked to it on Google+ group 'Life in the Cenozoic Era'.