on such a point, that when he crossed distinct species of native plants that had not been cultivated, he never once saw in the offspring any new character ; but that, from the odd manner in which the characters derived from the parents were combined, they sometimes appeared as if new. It appears, therefore, that the point at which the line of distinction is drawn between reversion and variation depends in many cases on the state of our knowledge of the subject. In some other points, also, the relations between inheritance and variation are extremely intricate and difficult to unravel. These two principles are often spoken of as opposed to one another. The following case shows that any definition of variability implying that it is necessarily equivalent to a breach in the law of heredity is incorrect. Some kinds of sheep and cattle dogs are congenitally almost destitute of a tail ; and this Stone- henge attributes to the fact that under the old excise laws only those dogs whose tails had been removed were exempt from taxation ; so that this mutilation was universally practised until the deficiency became hereditary. The production of a tailless breed of dogs must certainly be considered a case of variation, yet in this case it is not a breach of the law of heredity, but a remarkable instance of obedience to that law, that is to say, of the transmission of the effects of mutilation. In other cases external causes produce some constitutional or otherwise imperceptible changes in the parent, and these in being transmitted to the offspring become correlated with some external or perceptible alteration, and in this way new characters may appear. This is undoubtedly a true case of variation ; nevertheless, strictly speaking it is due to the inherited effects of a cause acting on the parent. And it seems illogical to separate it in a radical manner from cases such as that of the tailless breeds of dogs. Considering, there fore, the great difficulty in which the subject is enveloped, it will be well to abandon theoretical considerations, and merely to state that characters at least practically new do undoubtedly appear in the offspring. In every group of organisms a degree of variability, sufficient to give material for the breeder to work on, probably exists. The Laplander knows and gives a name to each of his reindeer, though, as Linnaeus remarks, " to distinguish one from another among such multitudes was beyond my comprehension, for they were like ants on an ant-hill." A still more striking case is that of the old Dutch florist Voorhelm, who kept above 1200 varieties of the hyacinth, and was hardly ever deceived in knowing each kind by the bulb alone. These cases are important as showing that, even in natural objects which appear identical to the unpractised eye, perceptible differences do exist. Man can effect nothing until some of his stock begin to vary in the desired direction. But horticulturists have found by experience that when any particular character is desired, the first step is to get the plant to vary in any manner, and to go on selecting the most variable individuals even though they vary in the wrong direction, for the fixed character of the species being once broken through the desired variation will sooner or later appear. The great number of races of many domestic animals and plants for instance, of pigeons, sheep, wheat. &c, demonstrates clearly their variability in many diverse characters. In other domestic animals, however, very few distinct races exist ; yet we must not conclude that these animals have not varied. There are several causes be sides that of an inherent want of plasticity which may have been at work. It will here suffice to allude to a few of
them.
1 . I f any particular group lias not been especially subj ected to selec tion, the absence of distinct races in such a group is no proof of want of variability. This applies to asses (in England only).
2. If the breeder has not a large number of individuals to select from, the chance of the required variations occurring is very small. Hence animals kept in small lots do not form races (e.g., sheep on small holdings).
3. If intercrossing cannot be prevented, it is obvious that any variety which may appear will have no chance of being perpetuated, but will be diluted down to the normal type. This applies to cats, which, from their wandering and nocturnal habits, cannot be paired.
It may appear a truism to say that every variation must have a distinct cause, but it is a truism very often over looked. The case of twins, each born with a peculiar crook in the little finger, is instructive, for here the con clusion is irresistible that the same definite, though unknown, cause produced the mal-formation in the two children. This case may also serve to illustrate the extreme obscurity in which the causes of any given varia tion are hidden, and the great difficulty of investigating them. Some general causes which induce variability may, however, be set down.
There appears to be no doubt that organisms subjected to the unnatural and changeable conditions implied by domestication are more variable than those living in a state of nature. Thus monstrosities are comparatively frequent among domestic animals and plants.[1] Domestication causes a number of changes in the condition of life; it is therefore of interest to determine which of these are the most im portant. Contrary to what might have been expected, change of climate is not an important cause of variation. This is repeatedly shown by A. de Candolle in his Geographie Botanique ; and a change to a more genial climate is certainly not necessary, for the dwarf kidney bean, which is often injured by our spring frosts, and the peach, which requires the protection of a wall, have varied much in England. (See Acclimatisation.)
by a change in the food of the caterpillar, but there seems to be no evidence that this cause has been active in inducing variability in our domestic races. On the other hand, excess of food is probably an important cause of variability. This view was held by Andrew Knight, and the same idea is expressed in the following remark of a " great raiser of seeds :" " It is a rule invariably with us, when we desire to keep a true stock of any one kind of seed, to grow it on poor land without dung ; but when we grow for quantity we act contrary, and sometimes have dearly to repent of it."[2] Nevertheless it appears that many of the best varieties of fruit have not been produced under cultivation. Thus it is asserted that some of the finest French pears were originally found growing wild, and this was the case with an English variety of apple. The most interesting fact connected with changes in the conditions of life is that the results of such changes are capable of accumulation. It is this peculiarity that accounts for the fact that when new flowers are first introduced into our gardens they do not vary. Thus the Swan River daisy did not break from its original colour until it had been subjected to seven years of high culture. Many facts might be given showing by what slight changes of habitat the health and general development of animals and plants may be affected[3] but with these cases we are not especially concerned. (See Acclimatisation.) The causes, how ever, which induce an unstable condition of general variability are of great importance to the breeder. Of the causes not already touched on the most important is intercrossing. In considering variations under this aspect no attempt will be made to distinguish from true cases of variation the cases in which new characters are simu lated by combinations of old ones. In the first place, it is
probable that organisms propagated by sexual reproduction