(2) We may pass now to the evidence afforded by direct observation and experiment. There is little enough of it. The best results are, perhaps, those which have been incidentally reached in the poultry-yard and on the farm in the breeding of domesticated animals. We have seen that, under these circumstances, certain parts or organs have very markedly diminished in size and efficiency; others have as markedly increased. Of the former, or decrease in size and efficiency, the imbecile ducks with greatly diminished brains have been already mentioned. Mr. Herbert Spencer draws attention[DL] to the diminished efficiency in ear-muscles, giving rise to the drooping ears of many domesticated animals. "Cats in China, horses in parts of Russia, sheep in Italy and elsewhere, the guinea-pig formerly in Germany, goats and cattle in India, rabbits, pigs, and dogs in all long-civilized countries, have dependent ears."[DM] Since many of these animals are habitually well fed, the principle of economy of growth seems excluded. Indeed, the ears are often unusually large; it is only their motor muscles that have dwindled either relatively or absolutely. If what has been urged above be valid, panmixia cannot have been operative; since panmixia _per se_ only brings about regression to mediocrity. If the effects in these two cases, ducks' brains and dogs' ears, be not due to disuse, we know not at present to what they are due. In the correlative case of increase by use, we find it exceedingly difficult to exclude the disturbing effects of artificial selection. The large and distended udders of cows, the enhanced egg-laying powers of hens, the fleetness or strength of different breeds of horses,--all of these have been subjects of long-continued, assiduous, and careful selection. One cannot be sure whether use has co-operated or not. Sufficient has now, I think, been said to show the difficulty of deciding this question, the need of further observation and discussion, and the necessity for a receptive rather than a dogmatic attitude; and sufficient, also, to indicate my reasons for leaning to the view that use and disuse, long-continued and persistent, may be a factor in organic evolution. _The Nature of Variations._ The diversity of the variations which are possible, and which actually occur in animal life, is so great that it is not easy to sum up in a short space the nature of variations. Without attempting anything like an exhaustive classification, we may divide variations into three classes. 1. _Superficial variations_ in colour, form, etc., not necessarily in any way correlated with 2. _Organic variations_ in the size, complexity, and efficiency of the organs of the body; 3. _Reproductive and developmental variations._ Any of these variations, if sufficient in amount and value to determine the question of elimination or not-elimination, selection or not-selection, may be seized upon by natural selection.