to a very limited extent, and in certain groups only. It frequently happens that a character existing in one of the parents is transmitted more powerfully to the offspring of the sex to which that parent belongs than to the opposite sex. The large and important subject of secondary sexual characters hinges entirely on this phenomenon. The re semblance between prepotency and sexual limitation be comes clear when we remember that where the offspring are of one sex it may be impossible to distinguish between these forms of heredity. The most interesting point con nected with secondary sexual peculiarities in relation to the subject of breeds is, that they are sometimes found in domesticated animals whose nearest wild congeners show no such limitation of character. Thus in the sheep, the males of certain races differ greatly from the females in the shape of their horns, in the development of fat in the tail (in certain fat-tailed breeds), and in the outline of the fore head. These differences are interesting because, so far as we know, similar secondary sexual differences are not found in the nearest allied wild species of sheep. On the other hand, secondary peculiarities which originally distinguished the sexes are in some cases diminished or removed by domestication. Thus our improved breeds of pigs have to a large extent lost the formidable tusks of the wild boar. The existence of secondary sexual characters gives a striking illustration of another important law of inheritance. This law asserts that the age at which any character first shows itself in the offspring is the same as that at which it appeared in the parent. Now, secondary sexual characters those, for instance, presented by the male sex have appa rently been developed by sexual selection, and this force can only be brought to bear on variations occurring in adult animals. If, then, the male offspring do not develop the selected peculiarities until they arrive at puberty, the age at which it appeared in their male parent, it is clear that they cannot differ from the female until the age of puberty arrives. And this, is well known to be the case, for at an early age the sexes are usually undistinguish- able by any secondary characters. (See Descent of Man,
vol. i. chap, viii.)The interesting form of inheritance exemplified by the transmission through the female line of diseases necessarily confined to the male sex has been already alluded to. This latency of male characters is clearly illustrated by what frequently occurs to old hens. It is well known that a large number of female birds, when old or diseased, partly assume the secondary male characters of their species. Waterton (Essays on Nat. Hist.} gives a curious instance of a hen which had ceased laying, and had assumed the plumage, voice, spurs, and warlike disposition of the cock. The opposite case of the assumption by the male of female characters is illustrated by the fact that capons sometimes acquire the sitting instinct of the hen.
The possibility of characters existing in a latent condition is of the utmost moment to the breeder, since upon it de pends the possibility of reversion or atavism. Reversion is a matter of extreme importance to the breeder, for it is one of the serious hindrances to the progress of his art. Since the time of the famous Bakewell during last century, Lei cester sheep have been bred with the most scrupulous care, yet grey -faced, black-spotted, or wholly black lambs occasion ally appear. In this case the most careful selection has been necessary to battle against the tendency of the original colouring of the sheep to reappear. And in all cases of selection it is this tendency that has to be struggled against by the breeder. On this principle the gardener looks over his beds and weeds out the " rogues." Even from seeds gathered from the finest cultivated varieties of the heart s-ease (Viola tricolor], plants perfectly wild both in flowers and foliage are frequently produced. The proxi mate cause of any particular case of reversion is utterly obscure ; but some of the general causes may be set down. It is frequently asserted that domestic animals or culti vated plants, when allowed to run wild, always revert to the original parent form of the species. This assertion appears to rest on insufficient evidence, and to be an exaggerated statement of what is known on the subject. Nevertheless some weight must be allowed to it. Pigs have run wild in various parts of the world, and have every where acquired the general characters of the wild pig, and the young have re-acquired the longitudinal stripes. This last character is interesting, since it is not in any way a direct result of the changed conditions of life, as the thicker bristles and increased size of the tusks might be supposed to be. Another well-established cause of reversion is crossing. The case is exceedingly striking when the offspring of a cross do not resemble any near progenitor, but throw back to very remote ancestors. In illustration may be mentioned the experiments on pigeons detailed in the Variation of Animals and Plants under Domestication (vol. i. p. 200). There can be but little doubt that all our domestic races of pigeons have descended from Columba livia, the wild rock pigeon ; the common dovecot pigeons exhibit the coloration of the parent form, and the most purely-bred fancy breeds, when of a blue colour, often show these characteristic marks. One of the above-men tioned experiments consisted in pairing a " mongrel female barb fantail with a mongrel male barb spot, neither of which mongrels had the least blue about them." It appears that blue barbs are exceedingly rare, that the spot has been known as a pure breed for nearly 200 years, and that a white fantail throwing any other colour is almost an unknown occurrence ; nevertheless the offspring from the above two mongrels were of exactly the same blue tint over the whole back and wings as that of the wild rock pigeon from the Shetland Islands. Moreover, every characteristic mark of the wild pigeon was repeated in their mongrel offspring. This experiment demonstrates in the most striking way the tendency of a cross to produce reversion. The same result was also obtained by pairing black Spanish cocks with hens of various white breeds. In this case the offspring reverted to the red colouring cf Galhis bankiva, which may be safely ranked as the parent form of our domestic fowls. In these instances the offspring revert to a character originally possessed by the ancestors of both parents, and here the cross is in no way essential to the reversion ; it merely acts as a disturbing cause (although, probably, no other equally strong disturbing power could be named). In these cases reversion to a character of any degree of antiquity may occur. In the other class of cases where the character to which the offspring revert is one given by a single cross with a distinct variety, the tendency to re version becomes weaker in each generation removed from the cross, and may ultimately be obliterated. The length of time requisite to effect obliteration has formed a subject of discussion. The question can hardly be answered, but the fact that it has been asked shows at least that oblitera tion may in some cases be effected in a practically finite period. In other cases even characters gained in this way by a single cross seem incapable of extermination. Fowls have been known to exhibit a Malay character, due to a cross with that breed forty years previously.
differing from the parent, we set it down at first sight as an instance of variability. But on the discovery being made that the peculiarities characterizing the offspring are derived from a remote ancestor, it can no longer be so considered, and must be attributed to reversion. Many cases of apparent variation are due to this cause. Thus
Gartner declares, and his experience is of the highest value