Post-Pliocene, showing that the genus existed there up to a comparatively late period; and when re-introduced it has thriven and run wild in the more temperate regions. What cause could have led to its extinction in Post-Glacial times is as yet a mystery. This genealogy of the horse, independently of its evolutionist application, is very interesting. It shows that some Eocene types were suited to continuance, and even adapted for extension, while others were destined to become altogether extinct at an early date. It shows farther that the power of continuance resided not so much in the gigantic and prominent species as in smaller forms. It is to be observed, however, that Gaudry and other orthodox evolutionists in Europe deduce the horse, not from _Eohippus_, but from _Palæotherium_, and that it is equally impossible to verify either phylogeny, since the mere sequence of more or less closely allied species in time does not prove continuous derivation. Nor indeed are we certain that one-toed horses like those now living did not exist on the dry plains in Eocene times, since the inhabitants of these plains are probably unknown to us. An amusing illustration of the probable reason of the disappearance of the missing links has recently been given by a writer not very favourable to the new philosophy. The several consecutive species may be represented by coins. We may suppose, for example, sixpences to have been coined first, then sevenpenny and eightpenny pieces, and so on up to a shilling, then pieces representing thirteen, fourteen and fifteen pence, and so on up to a half-crown or crown; but all the intervening denominations between the sixpence and the shilling, and between the shilling and the half-crown, were found practically of little use. Hence few were coined, and they soon became obsolete. Thus the antiquary would find only a few denominations, and those connecting them would be seldom or never found. It is plain that if we could suppose that nations constructed their coinage after this unthinking and empirical fashion, and that if we were justified in ascribing a similar procedure to the Creator, it might help to account for the facts as we find them, otherwise we should rather suppose that in both cases something like plan and calculation determined the selection of the species produced, whether of coins or animals. But Chance is a blind goddess, and if we instal her as creator, we must expect the work to proceed by a series of abortive experiments.