896 JOUENAL OF THE ROYAL HORTICULTURAL SOCIETY. 



" (4) By selecting those best plants which, upon trial, produce superior 

 progeny, the whole variety may be slightly or considerably improved. 



" (5) Since the plants of each succeeding generation also vary, by 

 repeatedly choosing the best the variety or race is further improved. 



(6) In many cases crossing increases the average vigour of the 

 progeny, but in other cases it decreases the average vigour, size, or other 

 desirable characteristics. 



" (7) In all cases crossing increases variation, as a rule, both toward 

 better plants and toward poorer ones, thus giving opportunity for selecting 

 from among the best plants individuals which are superior, as progenitors 

 of varieties, to any individuals which could have been secured without 

 crossing. 



" (8) New varieties can best be founded upon one to a dozen superior 

 selected or cross-bred seedling plants used as parents. 



" (9) Very large numbers of individuals must be used from w^hich to 

 select or breed, in order that mother plants may certainly be discovered 

 from which superior varieties will spring." 



The Use of Variation Illustrated. 

 Prof. Hays introduces some interesting diagrams giving graphic ex- 

 pression to Quetelet's law of variation. Two strains of pure-bred Wheats 

 are shown, the individuals of which diverge very little from the average, 

 except a few at both extremes, which are either very good or very poor, 

 and it is only from a few of the extreme good ones that future progress 

 can be made. These two strains are then crossed together, and the result 

 is very striking. 



There are still large numbers which retain the average value, but at 

 the two extremes the good and poor forms are more numerous and more 

 extreme than in the case of either of the two pure-bred strains. These 

 graphic diagrams show at a glance the powerful value of cross-breeding 

 in securing variation (both good and bad) quickly, e.g. to take a single 

 character of the above, i.e. yield in grams per plant." 



(1) The pure-bred "Fife" ^Yheat individuals varied in yield from 

 ^ to 5 grams. 



(2) The pure-bred " Blue Stem " Wheat individuals varied in yield from 

 1 to 5^ grams ; while 



(3) The ''hybrid " race, between the two, produced individuals which 

 varied in yield from 0 to 11^ grams. 



Further breeding from those individuals giving the greatest yield 

 shows that the maximum yield is by no means maintained, though the 

 average yield is gradually and surely increased. 



Curiously enough, some plants, which individually yielded well, pro- 

 duced progeny decidedly low in yielding power, thus demonstrating that 

 the " force of heredity of the family, race, and species powerfully combats 

 the new characteristics and tends to reduce the new forms to conformity 

 with the old. Thus, in trying to raise the yield above the average, we 

 must eliminate not only the poor but the average blood, retaining only the 

 blood of the few which have the greatest ability to produce progeny with 

 exceptionally large yield." 



From these experiments follow two important facts : — (1) That we 



