are usually the most liable to variation. For here the offspring has a double chance of being influenced by circum stances affecting the parents ; and by the concentration of a double set of tendencies into one individual, a better chance is given for the origin of variations produced by combinations of ancestral characters. In the same way, to a certain extent, a cross with a distinct variety produces a disturbance or loss of equilibrium in the reproductive system from which a tendency to the production of varia tions results. Thus Gartner asserts that seedlings from J)ianthiis barbatus, when crossed by the hybrid D. chinensi- barbatus, were more variable than those raised from this latter hybrid fertilized by the pure D. barbatus. Max Wichura insists strongly on an analogous result in the case of willows ; and Kolreuter says that to obtain an endless number of varieties from hybrids they should be crossed
and recrossed.Some peculiarities in our domestic races are to be attri buted to the inherited effects of habit and of disuse. Splints and ring-bones on the legs of horses appear to be certainly hereditary ; and veterinary surgeons agree in pronouncing these growths to be the result of travelling on hard roads, and of the horses being shod. The effects of disuse are clearly shown in the skeletal characters of our domestic races. These effects are well marked in tame birds, which are necessarily prevented from exercising their wings in flight. Thus in the domestic duck the crest of the sternum is less prominent, the furculum, coracoids, and scapulae are all reduced in weight relatively to that of the whole body ; the bones of the wing are shorter and lighter, and the bones of the leg longer and heavier in comparison with the same bones in the wild duck. Closely connected with this class of facts is the subject of rudimentary organs. In organisms living in a state of nature the constant pres sure of the struggle for existence tends to keep useless structures in a rudimentary condition. But domestication, in removing this pressure, does away at the same time with the principle of economy of growth ; and accordingly, we find that organs rudimentary in a state of nature become developed under domestication. Thus cultivation has made true branches out of the thorns or rudimentary branches of the wild pear. Again, the rudimentary fifth toe on the dog s hind foot becomes in some cases considerably deve loped, and forms the " dew-claw " of a few large breeds.
Correlation has probably played an important part in modifying domestic races ; for in selecting a given character man has frequently perpetuated many other peculiarities correlated with the first.
Finally may be mentioned the curious phenomena of " analogous variation." This term is applied to those cases in which varieties of one species resemble distinct but allied species. Where this occurs it is probably due to the two fonrs having originated in a common progenitor, so that modifying causes evolve similar varieties in the two cases because of the similarity of the material which these forces have to act on. Analogous variation is therefore, properly speaking, a branch of the subject of reversion, and once more points out the close connection existing between the latter phenomenon and variability.
In purely bred fowls of many races, birds may occasion ally be found closely resembling the Callus bankiva. Here the case is one of simple reversion, and has already been alluded to. The production of spangled sub-breeds of Hamburgh, Polish, Malay, and Bantam fowls is not an obvious case of reversion to a known ancestor. It may be due, however, to descent from the parent form of the (Jallinacese, considering the frequency of spangled markings throughout the order.
Selection.—Selection may be defined as the process by which the procreators of each fresh generation are chosen out of the preceding one. But with reference to the formation of our improved breeds something more than this is meant by the term. The modern development of the art, which has been distinguished by Mr Darwin as methodical selection, always implies that the breeder has before his mind an ideal form, a model on which he attempts to mould his strain. To be successful in this respect a man must not only possess in the highest degree the powers of discrimination, enabling him to determine which individuals are tending in the right direction, that is, which most nearly approach his ideal, but he must be able to decide, in the most judicious manner, as to which of his selected individuals ought to be paired together. The subject, therefore, falls naturally into two divisions (1) the discrimination of individual differences, (2) the arrangement of the pairing.
1. Discrimination of Individual Differences.—The extremely fine powers of perception and the great patience required by the breeder might be illustrated by numerous instances. Ij ir John Sebright is said to have sometimes spent several days in weighing the rival merits of five or six birds. In Germany the merino sheep farmers do not even trust their own judgment,but employ professional "sheep elassifiers" to select the best of the ilock for breeding purposes. Not merely outward form, but internal and constitutional peculiarities have been carefully attended to by breeders. Thus, Bakewell (tho first true methodical breeder of whose proceedings we have any knowledge) bred almost entirely for the early maturity and fatten ing qualities of his strain ; and again, in the improved short-horn breeds, the masses of internal fat or tallow have been increased to an almost incredible extent.[1] Professor Low states,[2] as an instructive instance of the limits that should be put to the exaggeration of any one point, that even the great Bakewell appears to have made this mistake in causing a useless accumulation of fat where it was not needed. It seems that the fat mingled less with the lean than even in the old breeds, and that it spread in layers under the skin, forming cushions of fat. He quotes a wiiter who ob serves that, having with great difficulty formed a race of cattle that would "make fat," Bakewell left his successors under the necessity of producing a breed that would "make lean." The same kind of error was at one time committed with the improved short-horns, vhose progenitors were pre-eminently good milkers; here the breeders, by attending exclusively to other qualities, have in some cases injured the milking powers of the race. Youatt is convinced that this loss is not necessarily correlated with rapid feeding qualities.[3] He asserts that by careful selection a strain may be obtained (and this has, in fact, been effected) in which the cows are first-class milkers, and when dried fatten quickly and well.
In spite of these warnings against exclusiveness and exaggeration, it should be remembered that the difficulties of selection are greatly increased by attention to several points at once. An illustration of this may be taken from the less important art of fancy breeding. An eminent fancier in speaking of the almond tumbler pigeon (a breed having five points, viz., plumage, carriage, head, beak, and eye), remarks that " there are some young fanciers who are over covetous, who go for all the above five properties at once ; they have their reward by getting nothing." Mr Darwin observes, "Vemaysmileat the solemnity of this precept, but he who laughs will win no prizes."
It should be remarked, however, that " fancying " is not governed by rules identical with those which regulate breeding for economic purposes. The fancier often has to strive after extreme abnormal development, amounting to monstrosity ; it has often been remarked that he will not tolerate anything short of this extreme divergence. On the other hand the economic breeder is prevented by solid mone tary considerations from being misled, to any great extent, by fashion. Hence, instead of the wide differences observable in "fancied" animals, we find a remarkable uniformity in certain characters among many of those bred for use. Thus all the improved races of the pig closely resemble each other in their shortened legs and muzzles, their large hairless* bodies and small tusks. Well-bred cattle of several distinct races exhibit a similar convergence of character.
2. Arrangement of the Pairing (including crossing).—We have already alluded to the remarkable case in which it seems an advan tage to the offspring that one parent only should possess the desired quality in an especial degree. This may be considered an extreme case, yet it bears some relation to the principles on which breeders usually regulate the arrangement of the pairings. Generally speak ing, individuals having certain points peculiarly well developed are matched with those excelling in other directions. It is probable that a physiological law which would formulate the exceptional cases above mentioned would also favour the more common practice
