by Brian Switek
Yet, as paleontologists have learned more about horse evolution over the last half century this simple story has become more complicated. The traditional view—that the spread of grasslands sent horses moving through a particular evolutionary pathway in which all their defining adaptations were linked together—does not hold up. Part of this can be understood by looking at some strange mammals only distantly related to horses.
When the non-avian dinosaurs went extinct sixty-five million years ago, South America was an isolated island continent, and a unique array of mammals evolved there. Among the varied assemblage was a group of hooved mammals called litopterns, of which the false llama with a trunk—Macrauchenia—discovered by Charles Darwin is the most well known. But Macrauchenia was only one of the last of the litopterns, a survivor of the Great American Interchange in which so many native animals were wiped out. Much earlier in the history of the group there were two litopterns that convergently evolved to stand on one toe.
One such litoptern, Diadiaphorus, exhibited a mix of characteristics seen in early and late horses. Its skull was like that of Eohippus, with a short face and a full set of low teeth indicative of an animal that ate soft plants, but its feet were superficially similar to those of later, larger horses such as Merychippus. Like them Diadiaphorus stood only on one toe and had two side toes that did not touch the ground. This toe reduction was taken even further by its close relative Thoatherium. It, too, had a skull that was more similar to that of Eohippus than Equus, but Thoatherium had only one central toe. The only signs that it had ever had side toes were two tiny nubs along its lower leg, and this astonished paleontologists as even living horses have more prominent vestiges of their side toes (the splint bones). Since both Diadiaphorus and Thoatherium lived during the Miocene, it was clear that they had evolved more “horselike” legs even before horses themselves did.
FIGURE 83 - The left hind limbs and skulls of Diadiaphorus (left) and Thoatherium (right).
It might be tempting to look at such a case of convergent evolution and conclude that the evolution of something horselike is somehow inevitable, but the anatomy of Diadiaphorus and Thoatherium actually suggest a different conclusion. The two litopterns were clearly not going through some kind of directed sequence from less to more “horsiness” over time. They possessed a mosaic of traits that confirm that the reduction of toes, acquisition of large body size, evolution of high grinding teeth, and an elongation of the face did not have to evolve as a single suite of characteristics. To better understand how such traits evolved in horses, it is profitable to go back to the beginning.
The first step involves identifying the start of horse evolution. This is not quite as easy as it sounds. For over a century, paleontologists have been trying to figure out the relationships between the Eohippus fossils of North America and Hyracotherium of Europe. The reason for this debate is that the genera are very similar to each other and very close to the base of not only the early evolution of horses, but the radiation of other types of perissodactyls, as well. Early rhinos, tapirs, brontotheres, palaeotheres, and chalicotheres probably all evolved from something akin to what has been classically called Hyracotherium.
For many years Hyracotherium was used as a taxonomic wastebasket to which hard-to-identify fossils were often assigned. The disarray began to be rectified in 2002 when paleontologist David Froehlich reexamined many of the Hyracotherium and Eohippus fossils to see whether or not they really did fall under just one or two genera. What he found was that there were more different kinds of early horse relatives than had previously been supposed. While Hyracotherium from Europe grouped most closely with creatures like Palaeotherium, there were several different early forms at the base of the horse evolutionary tree. The early genera Sifrihippus, Minippus, Arenahippus, and Xenicohippus were all close relatives of Eohippus from North America, placed near the base of early horse evolution.
Eohippus and its close relatives flourished during the time of a global hothouse. Warm climates reached so far north that there were crocodiles living in what is now the Canadian Arctic, and much of North America was covered in subtropical forests. These lush forests provided plenty of food for the small, multi-toed horses of the day, but such succulent plants were not only available in the forest. In 1981, paleontologist Philip Gingerich described a population of two dozen Hyracotherium from a site in southern Colorado that appeared to have been a more open, woodland environment. In such a habitat the early horses could have picked leaves off shrubs, chewed on ferns, and eaten other soft plant foods that grew low to the ground. That they probably did so has been backed up by studies of microscopic wear patterns on their teeth.
FIGURE 84 - A recent phylogeny of fossil horses showing their adaptive radiations through time and their expansion to different continents.
The horses which followed the early Eohippus type were not very much different. Horses such as Orohippus and Epihippus were still relatively small forms that lived in the Eocene hothouse climate in which their ancestors evolved. As the Eocene drew to a close, though, the climate dramatically changed. By about thirty-four million years ago, the average global temperature had dropped approximately twenty-five degrees Celsius, causing an uptick in extinction that claimed the last of the archaeocetes such as Basilosaurus. Horses were not so adversely affected, and as temperatures fell in the late Eocene horses began to change. From forms akin to Epihippus evolved Mesohippus, a “middle horse” that was subtly different from its ancestors. It was a little larger, its face was a little longer, its teeth had more folds useful for chewing tougher foods, and it supported its weight on three-toed feet. In fact, Mesohippus was nearly caught in the act of a transition in that early members of the genus had a still-functional vestige of a fourth toe on their front legs while later representatives did not.
At first this might seem to fit in with the idea that there was a “main line” of horse evolution in which genera gradually transformed from one to another, but a study by paleontologists Neil Shubin and Donald Prothero published in 1989 suggested that this interpretation was wrong. Traditionally Mesohippus was treated as a kind of missing link between earlier horses and later horses, a single stock from which the horse Miohippus evolved before giving rise to a diversity of later types. Since both genera were well represented from many fossils dug out of the thirty- to thirty-three-million-year-old rock of South Dakota, however, their fossils allowed paleontologists to reappraise this classic view.
Shubin and Prothero’s findings grated against the classic story. Rather than just a single lineage of Mesohippus gently grading into Miohippus, both genera were represented by several species that overlapped in time. There was a radiation of Mesohippus species, a population of which was modified into what we call Miohippus, and species of this newly derived genus flourished alongside species of Mesohippus, which itself did not become extinct until four million years later. Just like the mammoths, populations of species split off, evolved into new species, and lived alongside their ancestral species without much further change (at least until another population speciated, in any case).
Overall, though, Mesohippus and Miohippus were not all that different from earlier horses. Miohippus embodied a shift, with its body weight on a single, central toe (even as the small side toes remained), but it was still a relatively small animal that primarily fed on shoots and leaves. As the Oligocene gave way to the Miocene twenty-three million years ago, however, major environmental changes were occurring. There was an uptick in global temperature, but the subtropical forests gave way to savanna rather than bouncing back. Grasses had been around since the Cretaceous, but it was only now that they began to occupy greater expanses of land, and many mammal groups, elephants and horses among them, responded by becoming adapted to eating grass.
It was a time when horses exhibited the highest amount of diversity. New horse genera radiated across North America, with a maximum of thirteen genera living at the same time about fourteen million years ago. Some of these line
ages, such as those containing Anchitherium and Hipparion , even crossed over into Asia and spread through the Old World.
Chewing grass can be hell on teeth, especially if the grass has little spicules of minerals in it that abrade enamel as the plant food is chewed, like silica. This means that an animal that shifted to grazing would wear down its teeth much faster than a browser would. In contrast to the conveyor-belt dental system of elephants, horses that moved into the grasslands were adapted to have a complete set of longer, high-crowned teeth that would take a lifetime to wear down.
If we were to take an x-ray of the head of a horse, we would see that the crowns of its teeth extend well into the upper and lower jaws. At only five years old the horse would still have much of its complete adult dentition intact. If we were to take another x-ray a few years later, though, we would find that the teeth did not reach as far as they once did. The horse’s near-constant chewing wears down the surfaces inside the mouth, so the alredy-formed teeth are continually pushed up through the gums to make sure there is a fresh grinding surface. Now let us assume that our horse was fortunate enough to live a long life, attaining the age of thirty-five. If we took one last x-ray we would see that the teeth would be almost entirely worn down, with just the roots extending a little way into the skull.
The shift toward these high-crowned teeth occurred about four million years after the spread of grasslands, and as selection favored higher-crowned teeth some changes in the skull were required. Horse faces, or the part of the skull in front of the eye, had generally become longer since the time of the earliest horses, but the lengthening of the teeth required that the faces of grazing horses become even longer. In their ancestral state the last teeth in the horse’s upper jaw were right below the eye socket, and as selection favored higher tooth crowns the teeth ran the risk of overcrowding the eye. The selection for higher teeth thus favored the evolution of a longer face by shifting the tooth row forward, a change that otherwise might not have occurred.
The group of horses with these high-crowned teeth belonged to the larger group Equinae, which not only included Hipparion and its relatives but the closest relatives of living horses as well. Even though these horses placed their weight primarily upon their central toe many still retained side toes for long periods of time. The side toes did not gradually shrink at a constant rate but remained so prevalent it is almost surprising that there are no three-toed horses today.
Three-toed horses were a casualty of the decline in horse diversity starting about ten million years ago. Among the first to go were the forest-dwelling anchitheres, followed by some of the earliest hightoothed equines such as Protohippus and Pliohippus. The smaller-sized horses, especially, disappeared, and those already adapted to grasslands either became adapted to the colder, drier steppe habitats that were spreading or went extinct. By about five million years ago there were only about six genera left scattered across the globe, including early members of the genus Equus. These were horses of modern aspect, but by one million years ago they were the only horses left. All the browsing horses and three-toed horses were gone, and during the Pleistocene mass extinction that wiped out the mammoths horses were also extirpated from the continent on which they had originated. It would only be a few thousand years later, when Europeans would bring domesticated horses with them to the New World, that horses would return.
A look at the evolution of body size among horses helps to underscore the pattern of their history. Even after ideas about “internal driving trends” were booted out of evolution, it was still believed that horses were a good example of a trend called “Cope’s Rule.” This is not so much a mechanism as a pattern in which the body size of a group of organisms increases over time.
Certain examples of fossil horses from successive strata would appear to illustrate this trend. The evolutionary lineages created by Marsh, Matthew, and others certainly show it, but the answer lies in an analysis of the entire swath of horse evolution. This task was undertaken by fossil horse expert Bruce MacFadden during the late 1980s, and he found a different pattern. For the first twenty-five million years of horse evolution, prehistoric horses stayed small, estimated at about 50 kilograms or less. It was only after the time of Mesohippus that significantly larger horses began to evolve, but there were still relatively small horses through the beginning of the Miocene. About ten million years ago, however, almost all of these small forms were wiped out, leaving horses that had, for the most part, independently evolved to a larger size estimated at 200 kilograms or more. At this time the small, forest-dwelling horses that had existed for about forty-five million years were decimated.The only exceptions were lineages such as the Pliocene horse Nannippus that actually became dwarfed in its evolution from a larger ancestor.
The reduction of toes, another favorite textbook trend, is similarly complex. Early horses, such as Eohippus, had four front toes and three back toes, with the underside of the toes supported by a fleshy pad. This pad was still present, though reduced, in Mesohippus, which also differed in that its fore and hind feet had three toes each.
During the Miocene radiation of horses, however, there were different modifications to this three-toed arrangement. Horses lost the fleshy pads that supported their feet and stood upon a solid hoof, and the side toes were not necessarily in constant contact with the ground. Many horse lineages retained three toes for millions of years, although the reason for this is still not completely understood. (They might have been useful when horses were walking in mud, or even more in preventing overextension of the lower leg while running.)
Even among horses that were thought to have entirely lost their side toes, recent discoveries have shown that there was a messy period in which some individuals of a genus had three toes while others had one. This was the case with a population of a close relative of Equus called Dinohippus from the approximately eleven-million-year-old Ashfall Fossil Beds in Nebraska. An understanding of development solves the puzzle of how a population alternately presents archaic (side toes) and derived (no side toes) traits.
Horses are tetrapods, descendants of the first four-legged vertebrates that clambered about the Devonian mudbanks over 365 million years ago. When horses first evolved, they inherited the developmental pattern by which fingers are produced. In this developmental pattern the “pinky” side of the foot or hand develops first, sprouting off fingers in an arc toward the other side of the hand. The reduction of horse fingers and toes, then, was probably constrained by changes in development. Some rare living horses hint that this was the case.
For hundreds of years people have noted the birth of horses with extra toes. O. C. Marsh, in two papers published in 1879 and 1892, recorded several cases of horses that had their splint bones develop into lateral toes, complete with a “hooflet.” Often there was just a single extra toe, on the inside of the leg. Through his work on fossil horses Marsh recognized this as an atavistic trait: the extra toes of the living horses were in the same place as the vestiges of the side toes in “normal” horses.
FIGURE 85 - An “eight-footed Cuban horse,” one of the rare polydactyl horses described by O. C. Marsh.
Yet at the time of Marsh’s writing no one knew anything about DNA or the mechanism of inheritance. While the role of development in the evolution of the feet of horses has not yet been extensively studied, it is apparent that, even in living horses, the genetic triggers that cause the growth of lateral toes are still present, even if they are not often expressed. This would explain why some Dinohippus had side toes while others did not. At the time of the Ashfall Dinohippus specimens, the mutation that caused a change in the regulatory genes that controlled toe production had not yet spread to the whole population. Its spread might have been more of a case of genetic drift (a chance shift in the genetic makeup of a population) than an adaptive reduction of the toes. In other words, the horses with the mutation that constrained the development of toes might have been, by chance, more reproductively successful. Dinohippus with lateral t
oes were not inferior to the single-toed forms, so the disappearance of their type cannot be considered as any kind of “improvement.”
In the opening of his 1891 summary called The Horse, the British anatomist William Henry Flower offered naturalists “insight into some of the fundamental principles of biology.” The horse, familiar to all, could be used as a starting point from which to launch into almost any topic of inquiry about biology, especially evolution:The anatomy and history of the horse are, moreover, often taken as affording a test case of the value of the theory of evolution, or, at all events, of the doctrine that animal forms have been transmuted or modified one from another with the advance of time, whether, as extreme evolutionists hold, by a spontaneous or inherent evolving or unrolling process, or, as many others are disposed to think, by some mysterious and supernatural guidance along certain definite lines of change.
But while Flower was right about the utility of the horse in understanding evolutionary concepts, our growing understanding of horse evolution has failed to tease out any great unidirectional trend, supernaturally guided or not. The overall history of horse evolution has a stop-and-go pattern shaped by drastically changing environments and evolutionary constraints. Despite the utility of the animal to the rise of modern civilizations living horses are just the last remaining twig of a previously much richer family.