How the horse began to run

Hyracotherium, a close relative of Arenahippus

One of the advantages of studying fossil mammals, compared with dinosaurs, is that there are many close parallels still alive today. There isn't anything remotely like a Tyrannosaurus stalking the plains of present-day Africa, but comparing sabre-tooth cats to animals such as tigers and leopards can tell us quite a lot, with rather less need for guesswork. Also, while complete mammal fossils are still quite rare, they are, nonetheless, more common than those of animals from the more distant past.

Still, when it comes to early mammals, complete skeletons are rare enough that finding one can provide a significant opportunity to learn more about them. Species are often described on the basis of their skulls alone, since skulls tend to be the most distinctive parts of the skeleton, and you can tell a lot just from that, but having the rest of the skeleton stll attached to the skull is obviously pretty useful.

A recent report in the Journal of Mammalian Evolution described a remarkably complete skeleton of the early horse Arenahippus, with the tail being almost the only part missing.

As a side note, exactly what this animal should be called is a matter of some controversy. When specimens of the species were first found, they were thought to belong to the genus Hyracotherium, which may (or may not) be an alternative name for Eohippus, the "dawn horse" that appears at the beginning of so many charts of the evolutionary history of horses you see in museums and the like. But its probably neither, so we'll stick with the name it was given in 2002, even though there are counter-arguments to that one, too.

At any rate, whatever its called, Arenahippus is one of the most primitive members of the horse family known. It lived in the early Eocene epoch, just ten million years after the extinction of the dinosaurs, when many of the modern groups of mammals were just getting started. We don't know that later horses evolved from it, because there were lots of species of early horse living alongside each other, and while one of them must have evolved into the later ones, there's no way to tell which it was - if its even we've found yet. Those step-wise evolutionary charts you see of horse evolution don't really show exactly what evolved from what, just general pictures of what horses at a particular point in time looked like.

In reality, like all evolutionary stories, that of horses is a branching tree, although its interesting to note that Arenahippus appears to branch off even before the more famous Eohippus did, putting it even closer to the origin of the horse family:

True Equines    Mesohippus     Eohippus
     ^              |             |
     |              |             |      Arenahippus
     |              |             |           |
     ----------------             |           |
            |                     |           |
         (3 toes)                 |           |
            |                     |           |
            -----------------------           |     Palaeotheres
                       |                      |          ^
                       |                      |          |
                       ------------------------          |
                                   |                     |
                            (First horses)               |
                                   |                     |
                                   -----------------------
                                              |
                                              |

In fact, even the tree above is greatly simplified - there are a great many other fossils branching off in between the steps shown above. Nonetheless, we can see that Arenahippus diverged at a point when horses still had four toes on their front feet (although, like Eohippus, they only had three on the hind feet). In terms of its size, and to some extent, its shape, it looked more like a dog than like a modern horse.

So what can this new skeleton tell us about the life of these earliest horses? Perhaps the most obvious place to look is the legs, since the one-toed foot of modern horses is one of their most distinctive features. The tops of the thigh and upper foreleg bones are clearly rounded, with flexible hip and shoulder joints. This is quite different from modern horses, where the shape of the joints means that the limbs can only move forward and back, with very little flexibility in any other direction. The authors suggest that this would have helped in an environment more cluttered with bushes and other obstacles, rather than the open grassland that favours the gallop of modern horses. Since other evidence suggests that the area of Wyoming where the fossil was found was woodland with dense undergrowth, this makes sense.

Furthermore, the shape of the bones of the hind limb show the presence of powerful muscles, especially the calf muscle. Taking into account the shape of the knee and ankle joints, this indicates that the hind limbs would have been bent as the animal pushed itself forward and began to run - something you see in dogs, but not in horses, whose hind limbs are fairly stiff.

However, its not just the shape of the limbs that show us how the animal would have moved, but also the backbone. Reconstructions of early horses tend to show a straight backbone, as can be seen in the photograph at the top. This is how the backbone of modern horses look, and the way that the individual vertebrae lock together makes the whole structure quite rigid, a pattern also seen in other fast-running hooved animals, such as antelopes. But there haven't been many good fossils with intact backbones before, and, looking at this one, it seems the pattern isn't quite so simple.

Back flexed, legs pushing towards the midline

The vertebrae at the far end of the back, just before it joins the pelvis, were, indeed, rigid, with processes that would have locked them tightly together. But just before this was a more flexible region where the bones would have prevented the back from twisting, or from bending upwards, but would not have prevented it from bending downwards. Thus, unlike modern horses, Arenahippus could have arched its back, and most likely did so just as it began to run.

All in all, Arenahippus seems to have been a more flexible animal than a modern horse, or even than its more horse-like later relatives, such as Mesohippus. That may be partly because the later animals were bigger, and a more stable body would have made them more energy efficient while running. Arenahippus's movable knees, strong calf muscles, and flexible hips would have enabled it to push off the ground with some force, while the arching back ensured its centre of mass stayed in line. Still, it does seem to have been more rigid than, say, a modern dog and was, perhaps, just beginning on a path that would lead its later relatives (if not, necessarily, its literal descendants) to their fast-running lifestyle.

That leaves aside the question of why later horses became larger at all, requiring the change to the more familiar shape and posture we see today. That may be due to the changing climate of the time, and the spread of grasslands. Arenahippus, like other very early horses, mainly ate herbs, and perhaps fruit, browsing on low-lying vegetation, while later horses grazed on grasses. Grass is harder to digest than herbs, so that a longer digestive tract is needed if you're going to eat it. One way to increase the length of the digestive system is to increase the size of the animal its inside, and its at least possible that this was a major reason for the change. The more open environment of grasslands may also have meant that the longer stride that the size and body shape of later horses promotes would have been more useful for them than for something living among dense undergrowth.

[Pictures from Wikimedia Commons, cladogram adapted from Mikko's Phylogeny Archive]