Fossil Horses--Evidence for Evolution

See allHide authors and affiliations

Science  18 Mar 2005:
Vol. 307, Issue 5716, pp. 1728-1730
DOI: 10.1126/science.1105458

Thomas Huxley, an early advocate of Darwinian evolution, visited the United States in 1876 on a lecture tour. Huxley had planned to talk about evidence for evolution based on a fragmentary sequence of fossil horses from Europe. One of Huxley's first stops was at Yale, where he studied the fossil horse collection assembled by the paleontologist O. C. Marsh during expeditions to the western territories. Huxley was so taken with the definitive evidence provided by Marsh's fossil horse collection that he used this evolutionary sequence as the focal point for his subsequent talk to the New York Academy of Sciences (1).

Since the late 19th century, the 55-million-year (My) phylogeny of horses (Family Equidae)—particularly from North America—has been cited as definitive evidence of long-term “quantum” evolution (2), now called macroevolution. Macroevolution is the study of higher level (species, genera, and above) evolutionary patterns that occur on time scales ranging from thousands to millions of years. The speciation, diversification, adaptations, rates of change, trends, and extinction evidenced by fossil horses exemplify macroevolution.

The sequence from the Eocene “dawn horse” eohippus to modern-day Equus has been depicted in innumerable textbooks and natural history museum exhibits. In Marsh's time, horse phylogeny was thought to be linear (orthogenetic), implying a teleological destiny for descendant species to progressively improve, culminating in modern-day Equus. Since the early 20th century, however, paleontologists have understood that the pattern of horse evolution is a more complex tree with numerous “side branches,” some leading to extinct species and others leading to species closely related to Equus. This branched family tree (see the figure) is no longer explained in terms of predestined improvements, but rather in terms of random genomic variations, natural selection, and long-term phenotypic changes (3).

Adaptive radiation of a beloved icon.

Phylogeny, geographic distribution, diet, and body sizes of the Family Equidae over the past 55 My. The vertical lines represent the actual time ranges of equid genera or clades. The first ∼35 My (Eocene to early Miocene) of horse phylogeny are characterized by browsing species of relatively small body size. The remaining ∼20 My (middle Miocene until the present day) are characterized by genera that are either primarily browsing/grazing or are mixed feeders, exhibiting a large diversification in body size. Horses became extinct in North America about 10,000 years ago, and were subsequently reintroduced by humans during the 16th century. Yet the principal diversification of this family occurred in North America. Although the phylogenetic tree of the Equidae has retained its “bushy” form since the 19th century [for example, see (2, 3)], advances in knowledge from fossils have refined the taxonomy, phylogenetic interrelationships, chronology, and interpretations of the ancient ecology of fossil horses.


The Equidae, a family within the odd-toed ungulate Order Perissodactyla (which includes rhinoceroses, tapirs, and other closely related extinct groups), consists of the single extant genus Equus. Depending upon interpretation, it also includes several subgenera, 8 to 10 species, and numerous subspecies (4). On the basis of morphological differences, Equus is separated into two or three deep clades within the genus. These include caballines (domesticated horse, E. caballus); zebras (three species recognized); and asses, donkeys, and related species. Recent studies of mitochondrial DNA indicate two deep clades within Equus, namely, the caballines and the zebras/asses (5). These deep clades split ∼3 million years ago (Ma) in North America and subsequently dispersed into the Old World. Equus became extinct in the New World ∼10,000 years ago, probably as a result of multiple factors including climate change and hunting by early humans. In the Old World, although its range contracted, Equus persisted and was then domesticated in central Asia about 6000 years ago from a stock similar to Przewalski's wild horse, E. caballus (sometimes considered its own species, E. przewalskii) (4).

The single modern genus Equus stands in marked contrast to a highly diverse adaptive radiation of the Family Equidae over the past 55 My that resulted in some three dozen extinct genera and a few hundred extinct species (3). Although the overall branched pattern of horse phylogeny (see the figure) has remained similar for almost a century, new discoveries and reinterpretation of existing museum fossil horse collections have added to the known diversity of extinct forms. Recent work reveals that Eocene “hyracothere” horses, previously known as “eohippus” or Hyracotherium, include an early diversification of a half- dozen genera that existed between 55 and 52 Ma in North America and Europe (6). New genera have recently been proposed for the complex middle Miocene radiation (7), although the validity of these genera is still debated.

Horse teeth frequently preserve as fossils and are readily identifiable taxonomically. They serve as objective evidence of the macroevolution of the Equidae. Horse teeth have undergone considerable changes over the past 55 My. The tempo of this morphological evolution has sometimes been slow and at other times rapid (2, 3). Primitive Eocene through early Miocene (between 55 and 20 My) horses had short-crowned teeth adapted for browsing on soft, leafy vegetation. During the later Miocene (between 20 and 15 Ma), horses underwent explosive adaptive diversification in tooth morphology. Shorter crowned browsers, which inhabited forests and open-country woodlands, declined in diversity during this time (8). In contrast, many other clades of horses evolved high-crowned teeth adapted for grazing on the extensive grasslands of more open-country biomes, which spread during the Miocene (25 to 15 Ma). Once high-crowned teeth evolved, some clades underwent a secondary adaptation, that is, they went from being grazers to being mixed feeders with diets consisting of grass and some leafy plants (9). Studies of carbon isotopes preserved in fossil horse teeth indicate that before ∼7 Ma, early tropical and temperate grasslands of the world consisted primarily of grasses that used the C3 photosynthetic pathway, whereas today these grasslands consist mostly of C4 grasses (10).

In many fossil groups, the trend toward larger body size in ancestral-descendent sequences has been termed “Cope's rule.” Early Eocene hyracothere horses classically have been compared in size to a small dog (∼10 to 20 kg), although house-cat-sized species have been discovered more recently. At the other end of the evolutionary spectrum, wild modern Equus attains a body size of ∼500 kg (3, 4). Although the 55-My-old fossil horse sequence has been used as a classic example of Cope's rule, this notion is now known to be incorrect. Rather than a linear progression toward larger body size, fossil horse macroevolution is characterized by two distinctly different phases. From 55 to 20 Ma, primitive horses had estimated body sizes between ∼10 and 50 kg. In contrast, from 20 Ma until the present, fossil horses were more diverse in their body sizes. Some clades became larger (like those that gave rise to Equus), others remained relatively static in body size, and others became smaller over time (3).

Fossil horses have held the limelight as evidence for evolution for several reasons. First, the familiar modern Equus is a beloved icon that provides a model for understanding its extinct relatives. Second, horses are represented by a relatively continuous and widespread 55-My evolutionary sequence. And third, important fossils continue to be discovered and new techniques developed that advance our knowledge of the Family Equidae. The fossil horse sequence is likely to remain a popular example of a phylogenetic pattern resulting from the evolutionary process.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.

Navigate This Article