98 THE PRINCIPLES OF HEREDITY 



animals, sheep for instance, have been improved by crossing. 1 

 Two breeds, each having in excelsis the qualities desired by 

 the breeder, are chosen and mated. The desired qualities 

 are not diminished, while the general vigour of the stock is 

 increased. 2 Or a breed having the desired qualities is crossed 

 once or twice with a breed which has them not, and whatever 

 reversion occurs in the desired qualities is eliminated in 

 subsequent generations by careful selection. 3 



163. Professor J. Cosser Ewart, also, while declaring that 

 ' ' intercrossing . . . never, so far as I have seen, results in 

 the appearance of characters absolutely new to the species," 

 insists that "though intercrossing usually results in retro- 

 gressive variation, it is indirectly an extremely potent cause 

 of progressive variations," 4 owing to the increased vigour of 

 the stock. Now, of course, it is conceivable that vigour is a 

 variation which is the cause of other variations. Probably, 

 however, it is nothing other than a manifestation of good 

 health due to the presence of certain characters and qualities 

 and the absence of certain others. Professor Ewart maintains 

 in effect indeed he does so formally that the environments 

 of the germ-cells is a cause of variation. We have already 

 seen that there are weighty reasons for believing that he is 

 mistaken. The fact that it is easy to evolve new varieties 

 from the cross-bred descendants of choice cultivated plants 

 and domesticated animals is no evidence that crossing is in 

 itself a cause of progressive variations. For not only is it 

 more easy to breed from vigorous stock, but there is no 

 reason why cross-bred stock should not produce just as many 

 progressive variations as ordinary stock in all characters 

 except those in which the parent varieties differed. Thus 

 supposing, from a plant having a flower of a certain colour, 

 two varieties were evolved by careful selection which differed 

 from the parent type and from each other only in the colours 

 of the flowers. If these varieties were crossed we should 

 certainly get no progression in colour. On the contrary, we 

 should get reversion. But there is no reason why we should 

 not get progression in shape and size. Moreover, breeders of 

 choice stock are very careful to eliminate all variations except 

 the very few they desire. In other words they endeavour to 

 suppress or eliminate variability. In this they are probably 

 successful, for, in Professor Ewart's own words, " In in-bred 



1 Animals and Plants, vol. ii., p. 73. 2 Ibid., chap, ii., p. 73. 



3 C 



4 I 

 1901. 



3 Op. cit., p. 157. 



4 Presidential Address to the Zoological Section, British Association, 



