Mammals might be accidentally conveyed to islands on such rafts as have been described by Sir Charles Lyell, and there are instances on record of their having crossed short distances of sea by swimming. Elephants and also deer and pigs are good swimmers, the former having been known to swim for six hours at a stretch. "But," remarks Mr. Lydekker (p. 13), "it may be assumed that about twenty miles is the utmost limit which mammals are likely to cross by swimming, even when favoured by currents. Such passages as these must, however, be of very rare occurrence, for a terrestrial mammal is not likely to take it into its head to swim straight out to sea in an unknown direction. Moreover, supposing a mammal, near to a particular island, to have arrived there by swimming, unless it happen to be a pregnant female, or unless another individual of the same species but of the opposite sex should arrive soon after (a most unlikely event), it would in due course die without being able to propagate its kind." All zoologists, indeed, are quite in accord with Dr. Wallace's view as expressed in _Island Life_ (p. 74). "Whenever we find that a considerable number of the mammals of two countries exhibit distinct marks of relationship, we may be sure that an actual land-connection, or at all events an approach to within a very few miles of each other, has at one time existed." As all the European islands come under this category, their mammals exhibiting distinct relationship with those on the European continent, they all have been connected with it formerly.