[37] _Variation of Animals and Plants under Domestication_, i. 130, 135; ii. 288. [38] _Encyclopaedia Britannica_, article "Zoology." [39] _Variation of Animals and Plants under Domestication_, ii. 367. [40] _Variation of Animals and Plants under Domestication_, ii. 367. Why then does the cheetah inherit ancestral habits so inadequately that it is useless for the chase unless it has first learned to hunt for itself before being captured? (ii. 133). [41] _Descent of Man_, p. 33. [42] _Origin of Species_, pp. 210, 211. [43] E. S. Delamer on _Pigeons and Rabbits_, pp. 132, 103. For other points referred to, see pages 133, 102, 100, 95, 131. [44] _Origin of Species_, pp. 188, 110; _Descent of Man_, pp. 32-35; _Variation of Animals and Plants under Domestication_, ii. 289, 293. Use or disuse during lifetime of course co-operates, and in some cases, as in that of the canoe Indians, may be the principal or even perhaps the _sole_ cause of the change. [45] For the importance of panmixia as invalidating Darwin's strongest evidence for use-inheritance--namely, that drawn from the effects of disuse in highly-fed domestic animals where there is supposed to be no economy of growth--see Professor Romanes on Panmixia, _Nature_, April 3, 1890. [46] _Descent of Man_, p. 33. [47] _Descent of Man_, p. 33. [48] _Variation of Animals and Plants under Domestication_, i., 453. [49] _Descent of Man_, p. 33. [50] _Descent of Man_, p. 33. [51] Wallace shows that the changes in our domestic animals, if spread over the thousands of years since the animals were first tamed, must be extremely insignificant in each generation, and he concludes that such infinitesimal effects of use and disuse would be swallowed up by the far greater effects of variation and selection (_Darwinism_, p. 436). Professor Romanes has replied to him in the _Contemporary Review_ (August 1889), showing that this is no disproof of the existence of the minor factor, inasmuch as slight changes in each generation need not necessarily be matters of life and death to the individual, although their cumulative development by use-inheritance might eventually become of much service. But selection would favour spontaneous variations of a similarly serviceable character. The slightest tendency to eliminate the extreme variations in either direction would proportionally modify the average in a breed. Use-inheritance appears to be so relatively weak a factor that probably neither proof nor disproof of its existence can ever be given, owing to the practical impossibility of disentangling its effects (if any) from the effects of admittedly far more powerful factors which often act in unsuspected ways. Thus wild ducklings, which can easily be reared by themselves, invariably "die off" if reared with tame ones (_Variation_, &c., i. 292, ii. 219). They cannot get their fair share in the competition for food, and are completely eliminated.