MASS SELECTION. 169 



Our critics with a leaning toward the "pure-line" idea have insisted 

 that nothing but brother-sister matings should have been employed in 

 our experiments. We have several times endeavored to carry forward 

 certain high-grade families on this basis, but have been unable to secure 

 large enough numbers of offspring to make this possible; but we have 

 in several cases produced families of considerable size, descended exclu- 

 sively from a single pair of ancestors notably in the case of our pure 

 "mutant" race and in a race descended from one hooded and one wild 

 rat, which race was continued through 8 filial generations. (See p. 21, 

 Castle and Phillips.) It would have been impossible, in these and other 

 races, to make as rapid progress as we secured through selection in our 

 two principal races, for when only brother-sister matings are permitted, 

 it often happens that a mate of proper grade can not be secured for an 

 individual among its own brothers and sisters, though such a mate 

 may be found among its cousins or more remote relatives. It being our 

 first object to test the effectiveness of selection, we have made selec- 

 tion of any individual within the group (series or family with which we 

 were dealing) regardless of relationship, making the selection as rigid 

 as the maintenance of a stock of considerable size would permit. More 

 than once we have crossed the danger-line in advancing the standard 

 of selection to such an extent that only small numbers of parents came 

 up to it ; more than once we have had to relax our standard temporarily 

 in order to keep the race alive.. 



That the long-continued inbreeding of our selected races has affected 

 their vigor and fecundity is unquestionable. It is shown by the fact 

 that the plus and minus races, which had a common origin many 

 generations ago and have ever since been bred in the same room and 

 under identical conditions, if crossed with each other, produce offspring 

 of much greater vigor and fecundity than either parent strain. In this 

 our observations on the effects of inbreeding are entirely in harmony 

 with those of Darwin, Bos, Weismann, and of breeders of farm animals 

 quite generally. Miss King is credited with the view that inbreeding 

 of rats may increase their size, vigor, and fecundity, but this is cer- 

 tainly contrary to common experience with these and other animals. 

 It is probably true that under inbreeding it is possible, in exceptional 

 cases, to isolate a strain relatively immune to ill effects from inbreeding 

 (like Darwin's "Hero" morning-glory) or so inherently vigorous that 

 it succeeds in spite of inbreeding. But it is very doubtful whether 

 inbreeding of itself affects vigor other than disadvantageously. It is a 

 sufficient test to cross-breed an inbred strain, in order to ascertain 

 whether the inbreeding has increased or impaired its vigor. 



