TRANSACTIONS OP SECTION D. G73 



almost always characterised by more or less reversion ; when, liowever, inter- 

 crossing results in the characters of one variety being engrafted on another, or to 

 the appearance of characters quite new to the species, we have progressive 

 variation. Judging from the results I have obtained, intercrossing of two distinct 

 varieties results, as a rule, in the loss of the more striking- characters of both 

 parents, i.e., in more or less marked reversion, the extent of the loss generally 

 depending on the difference between the forms crossed. For example, if an owl 

 pigeon is crossed with a pigeon known among fanciers as an archangel, nondescript 

 birds are obtained, which may at once, with a white fautail, give birds almost 

 identical with a blue-rock — the common ancestor of all our breeds of pigeons. 

 Intercrossing, on the other hand, rarely leads to the blending of the unaltered charac- 

 ters of two or more varieties, and it never, so far as I have seen, results in the 

 appearance of characters absolutely new to the species. In a word, the immediate 

 result of intercrossing distinct varieties is, as a rule, more or less marked reversion. 

 But though intercrossing usually results in retrogressive variation, it is indirectly 

 an extremely potent cause of progressive variation. This is due to the fact (better 

 realised by botanists than zoologists) that cross-bred offspring (first crosses) are 

 (unless the parents have been enfeebled by interbreeding) endowed with an unusual 

 amount of vigour, i.e., intercrossing is of supreme importance, not only because it 

 leads to the co-mingling of germ-plasms having different tendencies, but also and 

 perhaps chiefly because of its rejuvenating influence. The importance of this 

 rejuvenation is usually at once evident if intercrossing is immediately followed by 

 interbreeding. The persistent interbreeding of closely related forms generally 

 reduces the vigour, and, as Darwin points out, 'far from causing variability, tends to 

 fix the character of each breed ' ; ' but the intercrossing of first crosses (or of highly 

 vigorous individuals closely related in either the direct or the collateral line) 

 without appreciably weakening the constitution, often results in offspring display- 

 ing, to use Darwin's words, ' an almost infinite diversity of character.' - The 

 epidemics of Tariation, so often the outcome of interbreeding first or at least 

 vigorous recently produced crosses, are apparently partly due to the union of 

 individuals having a similar tendency checking reversion, and partly to the 

 vigour acquired by recent intercrossing. This much may be inferred from the 

 fact, that when interbreeding is persisted in the variability dwindles as the vigour 

 ebbs. 



Breeders agree with Darwin that first crosses are generally uniform, and that 

 the subsequent offspring usually vary immensely ; yet neither breeders nor 

 naturalists seem to have clearly realised that interbreeding at the right moment is 

 the direct cause of variation, while intercrossing is, except in very rare cases, at the 

 most an indirect cause of variation. 



It may be here said that it is impossible to over-estimate the importance of 

 vigour in studying variation. Without vigour no race or breed can maintain its 

 position ; without renewed vigour it is hardly likely to develop new characters. 

 The new vigour, as already explained, may be obtained by intercrossing ; but it 

 may also be acquired, especially in plants, by a change of surroundings accompanied 

 by a plentiful supply of suitable food. 



With rigid selection the gradual loss of vigour may escape notice, but when 

 selection is suspended, rapid deterioration (from the fancier's standpoint) is the 

 inevitable result. If, e.cj., a number of pigeons, good specimens of a distinct 

 breed, are isolated and left immolested for a few years, they rapidly degenerate, 

 i.e., they lose their show points (be they peaks, frills, rufls, or metallic tints) and 

 reassume the more fixed ancestral characters. If, however, the less characteristic 

 birds are eliminated, and high-class birds are from time to time introduced from 

 another loft, the vigour and the distinctive traits are indefinitely preserved. 



If the age and condition of the soma and the state of ripeness of the germ- 

 cells are potent factors, and especially if vigour counts for much, the difficulties of 

 breeders become intelligible, and the unlikeliness of intercrossing being a direct 

 cause of variation all the more evident. The most that can be expected from 



* Animals and Plants, vol. ii. p. 251. " Ihid., vol. ii. p. 254. 



