At this point we may conceive that (2) conscious selection of the more desirable individuals appears. Through its agency, Hampshire, Shropshire, Cheviot, and Southdown sheep alike, and the others in their degree, tend toward larger size, more wool, plumper bodies, earlier maturity, greater docility, greater fertility, or whatever virtues the average shepherd may prize in a sheep. While in race traits, the breeds (uncrossed) tend to diverge from one another, in these adaptive qualities, their tendency is to run parallel-or even to converge toward greater resemblance.

With conscious selection (3), there is first a tendency to emphasize the qualities of desirable breeds. If, for example,

FIG. 50.-Typical Southdown ewe. (After Shaw.)

the Hampshire is a favorite breed, the individuals showing most distinctly black ears, legs, and face will be preferred by breeders to those having these parts pale. Again, new points of special excellence will appear in the breed and these will be deliberately emphasized, and perhaps by continuous selection a new breed will be formed having one or more of these as a distinctive trait. According to Somerville, one may chalk out on a wall any form or type of sheep he may like, and then in time reproduce it through selective breeding.

In Nova Scotia, Mr. A. Graham Bell has developed a new breed of sheep by selection, its distinctive character being in the increased milk flow, with an increased number of teats.

At Chillenham, in England, is still preserved a herd of the original wild white English cattle, from which most or all of the British breeds are said to be descended. It is stated that Lord Cawdor has offered to reproduce this herd, by selection alone, in three or four generations, using the relatively primitive Welsh cattle as his base of operations.

In general, those characters which are usually affected by selection, whether natural or artificial, are characters of degree. They are matters of more or less, a greater or less degree of strength, swiftness, size, endurance, fertility, capacity to lay

on fat, docility, intelligence, or of whatever it may be. Under ordinary conditions these characters selected are not traits of quality. They do not represent a new thing, a new acquisition, but a different degree of development of an old one, or, at most, a change in their relative arrangement, an alteration of biological perspective.

The characters which distinguish true breeds as well as true species are not of this order. They are in their essence qualitative and not quantitative. They are not, as a rule, adaptive. One set of species or race traits is as good as another, if the good qualities or adaptive qualities are represented in an equally

FIG. 52.-Heads of various British breeds of domestic cattle, showing variations in

FIG. 53. Various races of pigeons, all probably descended from the European rock dove, Columba livia.

high degree. The Southdown sheep are valued-not for their Southdown traits, but for the excellence of their mutton, a trait with which middle length of wool, tawny legs, naked faces, drooping ears, and absence of horns have nothing necessarily to do. We value these race traits only for the other qualities

FIG. 54. Skulls (in longitudinal section) of two breeds of domestic fowl, showing the large modification in the cranium: upper figure, Polish cock; lower figure, Cochin cock. (After Darwin.)

which have been in a high degree associated with them in the heredity of the race.

Under crossing and selection, much bolder attempts are possible. When parents widely divergent are crossed, many very different results are attained. In general the progeny, at least after the first generation, diverge very widely from one another. Some will have the good traits of both parent stocks; some will have the undesirable ones; some will show a mosaic of parental characters; some a more or less perfect blend of characters, this blend being definable as a finer type of mosaic. Some will diverge widely from either stock, often showing traits either remotely ancestral or wholly new. From desirable variations of this sort new races may be developed, each succeeding generation tending to give greater fixity.

In general, wide crosses or hybrids are more successful with plants than with animals, because the mutual adjustment