by Brian Switek
It would not be long before other paleontologists picked up where Gaudry had left off. During the 1870s a Russian émigré and former student of Ernst Haeckel named Vladimir Kovalevskii rifled through cabinets of bones in Parisian museums, including Cuvier’s collections from around the city itself. Within these collections lay the the bones of Anchitherium, which had existed in just the right place and time to bridge a major gap in horse evolution, and the correlation of features from Palaeotherium on the end through Anchitherium and Hipparion to Equus seemed to suggest adaptation to changing environments. Living horses were known to be grazers, with high-crowned teeth adapted to eating grass. Palaeotherium and Anchitherium had lower-crowned teeth that were much more similar to those of browsers, like tapirs, and so Kovalevskii, who did accept Darwin’s mechanism for evolution, not only had a transitional series but also a trigger for the mechanism of natural selection to act. (Although his scientific work on the evolution of horses was significant, Kovalevskii was never welcomed into professional paleontological circles and, beset by depression and financial woes, he committed suicide by chloroform in 1883.)
Kovalevskii’s general series was very similar to the linear transition of horses that T. H. Huxley presented in his 1870 address to the Geological Society of London. Huxley’s address was primarily concerned with presenting evidence from paleontology for evolution, including representative types of the transition from reptiles to birds and from terrestrial carnivores to whales. In both cases, however, fossil remains were only representative of the probable way in which evolution proceeded. Horses, however, presented a direct line of descent, and as such were a perfect example of how evolution was affected through successive modification.
When we consider these facts, and the further circumstance that the Hipparions, the remains of which have been collected in immense numbers, were subject, as M. Gaudry and others have pointed out, to a great range of variation, it appears to me impossible to resist the conclusion that the types of the Anchitherium, of the Hipparion, and of the ancient Horses constitute the lineage of the modern Horses, the Hipparion being the intermediate stage between the other two.
This three-step path from Anchitherium to Equus would only be the start. The theory of Darwin and Wallace predicted that there would be an even older equid that would connect Anchitherium to an even more ancient ancestral type, and Huxley felt that a small creature called Plagiolophus (originally called Palaeotherium minor by Cuvier) fit the role nicely. Although cautious about placing the little perissodactyl in the direct line of descent, Huxley still felt that it was at least representative of the form of the creatures from which later horses evolved.
The evolution of the horse thus seemed almost completely understood. Even after Huxley began to shift his research interests away from paleontology, horses remained a particularly good example of evolution. The fossil horses were the stars of his 1876 lecture tour of the United States, where Huxley summarized the views he had earlier presented to the Geological Society:Seven years ago, when I happened to be looking critically into the bearing of palaeontological facts upon the doctrine of evolution, it appeared to me that the Anchitherium, the Hipparion, and the modern horses, constitute a series in which the modifications of structure coincide with the order of chronological occurrence, in the manner in which they must coincide, if the modern horses really are the result of the gradual metamorphosis, in the course of the Tertiary epoch, of a less specialized ancestral form.... That the Anchitherium type had become metamorphosed into the Hipparion type, and the latter into the Equine type, in the course of that period of time which is represented by the latter half of the Tertiary deposits, seemed to me to be the only explanation of the facts for which there was even a shadow of probability.
The procession of forms made sense in terms of anatomy and fossil sequence, but a new set of discoveries would overturn what had been Huxley’s prime example of evolution. These would be made not in the Old World, but in the New.
By the middle of the nineteenth century some European paleontologists were becoming worried about the state of their science. It seemed as if most of the available fossil strata had all been thoroughly investigated, and even these were separated by gaps in the geological record that could not be found in Europe. The German paleontologist Ferdinand Roemer even went as far as to tell his students to avoid populated areas and seek fossils in wilder, unsettled regions. This point was apparently not lost on one of his students, O. C. Marsh.
Marsh’s professional career got off to an unusual start. Through the connection to his wealthy uncle, the entrepreneur George Peabody, Marsh was able to wrangle a professorship in paleontology at Yale. It would be the first such official post in the nation. The college did not actually have the money to pay Marsh, but Peabody endowed the school with funds to build a museum for his nephew, and provided him with an allowance that kept him living comfortably. Since he was not being paid by Yale, Marsh had no teaching duties or other responsibilities placed upon him. He was free to pursue his professional career as he saw fit.
FIGURE 79 - Paleontologist O. C. Marsh.
During the time Marsh was studying anatomy and paleontology in Germany prior to his Yale appointment, the general consensus was that there was no strong evidence that ancient horses had ever lived in North America. Recalling such statements Marsh later wrote:I heard a world-renowned professor of zoology gravely inform his pupils that the horse was a gift from the Old World to the New, and was entirely unknown to America until introduced by the Spaniards. After the lecture, I asked whether no earlier remains of horses had been found on this continent and was told in reply that the reports to that effect were too unsatisfactory to be presented as facts of science. These remarks led me, on my return, to examine the subject myself.
The opportunity for Marsh to prove this professor wrong came soon enough. In 1868, while traveling to the end of America’s unfinished transcontinental railroad, he heard that some strange bones had been recovered from a well in nearby Antelope Station, Nebraska. Smelling an opportunity, Marsh tipped the conductor to hold the train while he went to see if the rumors were true, and soon found what appeared to be the bones of diminutive horses. Marsh named his “toy” horse Equus parvulus, later renamed Protohippus.
This was just the beginning of Marsh’s fossil horse discoveries. Accompanied by teams of students (and sometimes cavalry escorts) Marsh collected the bones of numerous fossil horses from almost the entire geological span of the previous fifty-five million years. These were all brought back to the Peabody Museum at Yale for study. Marsh had gathered a much more complete and detailed sequence than the lineage Huxley proposed in 1870.
Marsh was certain he had found the direct line of descent from small, many-toed ancestral horses to large, single-toed descendants. The horse pedigree ran from the four-toed Orohippus of the Eocene through the three-toed Mesohippus to the Miocene forms Miohippus and Protohippus and to the one-toed Pliocene Pliohippus before culminating in the familiar Equus of the Pleistocene on. It was a simple, beautiful sequence that left no room to doubt that horses had evolved in North America. The sequence preferred by Gaudry, Kovalevskii, and Huxley only documented migrants that had left the continent of their birth.
FIGURE 80 - A visual summary of horse evolution as proposed by O. C. Marsh. In Marsh’s diagram, there are clear, unidirectional trends in toe reduction, tooth height, and, though it cannot be seen here, size.
Huxley was faced with this wall of evidence when he met with Marsh during his lecture tour and could do little except discard his shorter European horse genealogy for Marsh’s more comprehensive one. For every question Huxley had, Marsh seemed to have a bone in response. As he wrote to his wife of the meeting, Huxley “turned upon [Marsh] and said: ‘I believe you are a magician; whatever I want, you just conjure it up.’” Huxley was most taken with the four-toed Orohippus , and predicted that a five-toed horse would eventually be discovered in even older Eocene deposits. (In jest, Huxley drew a cartoon o
f the elusive “Eohomo” riding the “Dawn Horse”.)
Huxley quickly changed his lecture to include the new information. During his American lecture tour he argued Marsh’s case for a North American origin of horses, and Huxley concluded:Thus, thanks to these important researchers, it has become evident that, so far as our present knowledge extends, the history of the horse-type is exactly and precisely that which could have been predicted from a knowledge of the principles of evolution. And the knowledge we now possess justifies us completely in the anticipation that when the still lower Eocene deposits and those which belong to the Cretaceous epoch, have yielded up their remains of ancestral equine animals, we shall find, first, a form with four complete toes and a rudiment of the innermost of first digit in front, with probably, a rudiment of the fifth digit in the hind foot; while, in still older forms, the series of the digits will be more and more complete, until we come to the five-toed animals, in which, if the doctrine of evolution is well founded, the whole series must have taken its origin.
Although neither Huxley nor Marsh was aware of it at the time, the remains of tiny Eohippus were with them as they looked over the remains of its younger relatives. As Marsh wrote to Huxley two months later, “I had him ‘corralled’ in the basement of our Museum when you were there but he was so covered with Eocene mud that I did not know him from Orohippus. I promise you his grandfather in time for your next horse lecture if you will give me proper notice.” With four front toes and the vestiges of a fifth, the tiny horse relative corresponded almost exactly to what Huxley had proposed would be found. Even more interesting was that this creature appeared to correspond closely to a creature that had been described years before under the name Hyracotherium.
In 1832, William Colchester found two fragments of a fossil jaw in Woodbridge, England. They were brought to the attention of Richard Owen, who supposed that they had belonged to some kind of monkey, but as more fragments from similar deposits were found, including a partial skull, Owen realized that the animal was not a monkey at all. He described it under the name Hyracotherium in 1839, but just what type of mammal it was remained unclear. In life it would have had “a resemblance to [the physiognomy] of the Hare, or other timid Rodentia,” though in affinity Owen thought Hyracotherium was probably a member of the “natural family of the Hog tribe.” He even described a second species of the “rabbit-like” Hyracotherium in 1843 on the basis of several teeth, but the discovery of Eohippus allowed the affinities of Hyracotherium to be understood. The small creature was no hog relative at all but was more closely related to the earliest known horses, although naturalists disagreed as to whether Hyracotherium or Eohippus should be the proper name for this early form.
FIGURE 81 - T. H. Huxley’s cartoon of an “Eohomo” riding his vision of what an “Eohippus” might look like.
Unfortunately Marsh was not able to make good on his promise of an even earlier horse ancestor with five complete toes, but by illustrating the graded evolution of the horse he had unquestionably verified what Darwin’s theory predicted. Darwin himself appreciated the evidence that Marsh had provided his theory, noting that the paleontologist’s work “afforded the best support to the theory of evolution which has appeared in the last 20 years.”
The evolutionary series seemed to be complete. There were so many horse fossils that it was easy to create displays showing how horses had gone from small, generalized animals to the powerful beasts of burden we know today. Marsh’s “direct line of descent,” however, would not hold up against the mounting wave of horse fossils that continued to come in from the western United States. There were simply too many extinct species to array in a single straight line.
In a 1907 review of fossil horses, the paleontologist G. W. Gidley surveyed the fossil horses from the previous thirty-four million years and found that many of these multi-toed fossil horses were not directly ancestral to living forms. In fact, at any point in time there seemed to be as many as four genera of horses living alongside one another. He envisioned the evolution of horses in a branching pattern, with each family being self-contained. Yet, even if the general patterns were apparent, the exact connections between types were still difficult to pin down.
Walter Granger complemented Gidley’s work in 1908 with a revision of the geologically older Eocene horses of North America that formed the base of the later radiation Gidley had cataloged. These earlier fossils were especially problematic, as they were close to the ancestry of not just horses, but also of the massive, horned brontotheres and other kinds of perissodactyls as well. Virtually the entire swath of perissodactyl diversity had been derived from very similar early types that resembled Hyracotherium, thus placing the branching pattern of horses into an even more tangled evolutionary pattern.
These more detailed problems persisted even after the efforts of Gidley and Granger. In 1924 W. D. Matthew, another scientist in H. F. Osborn’s stable at the American Museum of Natural History, published a paper entitled “A New Link in the Ancestry of the Horse.” In his introduction Matthew wrote:The series of American Tertiary ancestors of the horse is one of the classic examples of evolution provided by the fossil record and the most complete and convincing among the mammals. Nevertheless, it is well recognized by those who have made a special study of it, that, while the broader lines of descent are beyond reasonable question, there are definite gaps between some of the successive stages and many minor problems as to the details of phylogeny.
One such gap was between the late Pliocene horse Pliohippus and the Pleistocene genus to which modern horses belong, Equus. While Pliohippus was large and the toes on the lateral sides of its feet had been reduced practically to nubs, it still had some peculiar features, such as a pit on the either side of its face, which distinguished it from modern horses. There had to be an intermediate form between it and Equus if it was the true ancestor of modern horses. Matthew believed that he had found such an animal. He called it Plesippus, and while Pliohippus would later turn out to be further removed from the ancestry of modern horses than had been previously thought, Plesippus did possess transitional features that linked it between archaic horses and the earliest Equus.
Even as paleontologists recognized the branching pattern of horse evolution, however, the linear iconography favored by Marsh held fast. The tension between these ideas was embodied by a review of horse evolution Matthew published in 1926. In general, it seemed apparent that horses got bigger, lost their “extra” toes, developed elongated faces, and evolved high-crowned teeth fit for grazing all at the same time. The traits were not obtained one at a time, but were all gradually modified in one direction. This could clearly be seen in the updated evolutionary trajectory Matthew published (Eohippus-Orohippus-Epihippus-Mesohippus -Miohippus-Parahippus-Merychippus-Pliohippus-Plesippus-Equus) which was taken as representing the successive “stages” of horse evolution. From oldest to youngest, each genus represented a “step forward” for horses. In this view, even if we had only the skeletons of the beginning (Eohippus) and end (Equus) of the series it would still be possible to predict all the intermediates along the graded chain.
This progression was in stark contrast to the full diversity of fossil horses Matthew recognized. In a more technical diagram, representing the span of time each genus was found and their connections to each other, it was clear that as many as seven different genera overlapped in time. Ancestors even persisted alongside their descendants in some cases, but Matthew did not pay much attention to this. He lumped most of what were considered “side branches” under the heading “Divergent Lines of Equidae” in this paper, implying that they diverged from the “main line” of evolution from Eohippus to Equus. Hence Hipparion was presented as a dead-end offshoot, as was a South American form with an enigmatically large nasal cavity called Hippidium, and many others. For Matthew, the “main branch” horses had given rise to radiations of “divergent lines” at various times during history, and any genus that expired without pushing the next “stage” of h
orse evolution forward could be pushed to the side.
FIGURE 82 - A simplified version of horse evolution as forwarded by W. D. Matthew. Even though Matthew recognized that horse evolution was more “bushy” with different groups overlapping in time and many not being ancestral to living forms, the most popular visualization of horse evolution was this modified form of Marsh’s diagram.
Even so, Matthew was more amenable to natural selection than many of his colleagues. He was more interested in how changing climates and environments affected organisms, and to him it was clear that the evolutionary success of the horses was dependent upon the spread of grasslands. The earliest horses, such as Eohippus, browsed on leaves in warm forests during the beginning of the Eocene, but as the global climate became cooler and more arid most horses became adapted to life on the grassland. The reduction of toes, the evolution of high-crowned teeth, and other fluctuating characteristics all seemed to be tied to a shift to life on the plains.
Matthew’s emphasis on environment would later be complemented by his student, George Gaylord Simpson, in his 1951 popular summary Horses, in which evolution by natural selection was given pride of place despite occasional references to “main lines” of descent. In the classic view, the shift from browsing to grazing had driven the larger trends seen in horse evolution. The larger size, high-crowned teeth, and other adaptations had all been selected for as forests gave way to grasslands, especially in the latter part of the Cenozoic (occurring alongside the evolution of grazing elephants).
The basic story went something like this: by the time Eohippus had evolved, western North America was covered in warm forests that provided plenty of soft food for the little horses to eat. The successors of Eohippus, such as Orohippus and Mesohippus, continued this lifestyle, but by about nineteen million years ago the climate was becoming cooler and drier. Grasslands overtook forests, so horses were adapted to life on the plains. This resulted in a proliferation of horses adapted to eat grass rather than leaves, such as Merychippus and Hipparion (among many others), which exhibited many of the classic horse characteristics. The evolution of high-crowned teeth adapted to grinding grass, long legs tipped in a reduced number of toes for fast running on hard ground, long faces, and other traits all arose from this environmental shift.