Science: Why the four-toed horse lost the race: If evolution is random, why has the family Equidae grown steadily larger and faster? The answer lies in the gut, says Colin Tudge

On 9 April two extraordinary species will thunder the four miles and four furlongs of the Aintree course in the Grand National. In the eyes of most 19th-century and many early 20th-century biologists, these two species demonstrate that evolution has a sense of direction and has laboured without deviation to provide the epitome of running power in the horse, and the apotheosis of intelligent life in the human being.

This conceit - that evolution really is so single-minded, and really does work towards predestined ends - fills modern biologists with unbridled rage. Yet the fact remains. Horses in particular are odd. There is nothing like a horse, and horses are like nothing else. And they do seem to have got steadily bigger and horsier in the 50 million years or so since their ancestors first appeared in the Eocene epoch, just as the 19th-century romantics said they did. So what is the truth of the matter, and why do modern biologists apparently seek to deny the obvious?

To understand, we must go back to Aristotle, who, as Bruce McFadden points out in Fossil Horses (Cambridge University Press, 1992), argued that nature and the Universe at large had an 'underlying design, order, and perfection' and that 'nature always strives towards these goals'. Many biologists in the 18th and early 19th centuries - well before Charles Darwin wrote On the Origin of Species in 1859 - embraced a primitive notion of evolution which was basically Aristotelian. They too imagined that evolution worked towards preconceived goals. In addition, most of Darwin's scientific contemporaries and predecessors were deeply religious (many were clerics, as Darwin once intended to be) and the idea that animals evolved towards a destiny seemed to reconcile the notion of evolution with the idea of God the Creator.

Such a view of evolution also smoothed the feathers of those scientists and clerics who objected to Darwin's notion that human beings in general, and they in particular, had descended from apes. It did not seem so bad if God had had us in mind all along. Accordingly, most evolutionary portraits until the last few years have shown a straight-line 'ascent' from ape to upright and cerebral human (though Darwin himself wrote of the 'descent of man').

But in the 19th and early 20th centuries very few human fossils had been unearthed. The fossils that were known in superabundance were those of horses, the creature selected by man as his companion and accomplice. These fossils seemed to show exactly the pattern that palaeontologists would like to have seen in the human record. That is, the most ancient horses of the Eocene (colloquially known as 'eohippus' but more properly named Hyracotherium) were the size of a fox and had four toes on their forefeet and three on their hind. They also had 'low-crowned' teeth; in sharp contrast to the pillar-like or 'hypsodont' molars of the modern jade. Furthermore, it seemed, the equid lineage had grown steadily larger in the intervening 50 million years, shedding toes as they went to increase their efficiency as runners, and lengthening their teeth, the better to cope with the razor-leaved grass that increasingly invaded the cooling world. O C Marsh presented this view in his drawing of 1879 (beginning with the Eocene Orohippus) and it is still being reproduced in textbooks. Such a straight-line progression is called 'orthogenesis'.

Twentieth-century biologists took exception to this idealised view for two reasons. First, they said, it is just not so. The lineage of the horse, like that of all other animals, evolved in many different directions at many different times in history. One-toed horses did evolve from three- and four-toed horses, but three-toed horses ('hipparions') none the less thrived alongside the one-toed horses until only a few million years ago (and left their footprints with those of the human ancestors Australopithecus afarensis in Africa). About six million years ago eight different genera (groups) of horses lived side by side in North America alone, three-toed and one-toed. By contrast, all modern horses (including asses and zebras) must be placed within a single genus, Equus. Furthermore, there has been no inexorable increase in size. Some of the fairly recent horse genera of North America were small, including the sheep-sized Nannipus. In short, orthogenesis is a misreading.

Second, and more profoundly, said the moderns, there is no destiny in evolution. More or less anything might happen at any one time. No lineage sets out single-mindedly to become something else; indeed there is no 'mind' in the process at all. Horses were not predestined to be horses. Neither, they add, were human beings predestined to be human beings. That is just the way things turned out.

Yet that modern no-nonsense view of horses does leave some loose ends. To be sure, the tree of horse evolution has not been a straight line. It was bushy - as was the human line. But most branches of the horse tree, both three-toed and one-toed, did tend to become more horse-like. Most in general became bigger and more hypsodont. And all the ones that were not one-toed, huge and very big-toothed have died out.

Indeed, the horse family Equidae stands in conspicuous contrast to those other great families of hoofed animals the Cervidae (deer) and the Bovidae (cattle, antelopes, sheep and goats). The modern horses, zebras, and asses are successful species but they are all much of a muchness; they are all big and barrel-bodied, and they all live on grass (that is, are grazers). But the modern cervids range from tiny 15in pudu to mighty moose, while the bovids extend from minuscule duikers to bison, buffalo and gaur. Furthermore, the deer and the bovids include both grazers and forest-living browsers (which eat the leaves of bushes). In short, the evolutionary trees of deer and bovids are far bushier than that of the equids; and in each case many different twigs of the bush have come through. The equid line may not literally have been single-minded, but it looks a great deal more single-minded than its two great rivals. Thus it seems that orthogenesis is not so easily dismissed. Yet predestiny in evolution remains an anathema. So what is really going on?

The answer, I suggest, lies in feeding efficiency. Deer and bovids are ruminants. After they swallow their food they cough it up again and chew it more thoroughly, then pass it to a vast fermenting stomach, the rumen, where obliging bacteria extract the nutrient that the animal can then absorb. Horses attempt to pull a similar trick but they lodge their fermenting bacteria in a diversion of the hind- gut, the caecum. There is no opportunity to give the food a second chewing, and there is less opportunity to absorb the nutrients that the bacteria provide because the fermenting chamber is so far down the gut. So horses have to eat more than ruminants do to get the same amount of energy. Horses, in short, are less efficient feeders than deer or bovids.

But another factor also affects feeding efficiency - body size. Big animals lose heat less easily than small ones. So, weight for weight, big animals do not need as much energy as small ones. Big animals are, in fact, much more feed-efficient than small animals. In small herbivores - say the size of a roe deer - the difference in efficiency between a ruminant and a hind-gut fermenter would be critical. But big animals are so efficient to start with, by virtue of their great size, that this difference in digestive efficiency no longer makes much difference. Hence a big modern horse can compete with a big modern bovid, as zebra demonstrate beautifully in Serengeti as they run alongside the gnu. But a small modern horse could not compete with a roe deer or a gazelle. No doubt, then, horses did strive to produce a range of grazers and browsers of all shapes and sizes just as the deer and bovids have done. But those that were smaller than an ass were simply cut off in their prime. At the small end of the market, the ruminants outdid them. Nannipus slipped briefly through the net, but no more.

We do not look after the animals that are left to us, and many antelope, cattle and deer are in danger of extinction. It is a pity, none the less, that there is no comparable range of equids to put alongside them; pudu- sized, roe deer-sized and fallow deer-sized horses. The modern Shetland is just an artefact, a domestic breed. In general, horses just don't have what it takes to be small.

But perhaps we should be grateful that gazelles and roe deer and their like have forced the horses to be so large. Otherwise the Grand National just wouldn't be the same.

(Photograph omitted)