CASTLE. — INBREEDING, CROSS-BREEDING, AND SELECTION. 749 



Mr. Mast's subsequent experiments, however, show that the unproduc- 

 tive A race did show its influence in the second cross-bred generation 

 (generation F 2 ), so that the cross did not eliminate wholly the unproduc- 

 tiveness inherent in the A race. For sterility and low productiveness 

 reappeared after skipping a generation, in certain of the grandchildren, 

 as will be shown by an examination of Tables X-XIV. This occurred 

 in two out of five families, following a cross between a normal female and 

 an A male. This observation shows that an entirely fertile male may 

 transmit partial or complete sterility of the female sex as a racial character 

 to his grand-daughters, though not apparently to his daughters. 



The reciprocal cross, A 9 X normal $ (Tables XV-XVIII), resulted 

 in one case in the production of females half of which were sterile (Table 

 XV), the other half being of low productiveness. This case shows that 

 a female of a race inclined to sterility may transmit that character 

 directly to her cross-bred offspring (generation F x ). The males of race 

 A in Mr. Mast's experiments transmitted sterility to the grand-daughters 

 only, never to the daughters. This difference in heredity through the two 

 sexes would seem to indicate that sterility of the female is dependent 

 upon egg structure rather than sperm structure, the eggs produced by 

 mothers of a fertile race always yielding fertile daughters. But the eggs 

 of cross-bred females, whose father was of an infertile race, produce some 

 of them fertile, some infertile females. 



Experiments of W. M. Barrows. 



Though the M and N races, after nine or fewer generations of inbreed- 

 ing, were not rendered more productive by a cross with race A, such 

 was not the case after an additional year of inbreeding. In March, 1905, 

 Mr. Barrows made crosses between the M and N races (then inbred for 

 twenty-four generations), and between each of these and the A race 

 (then inbred for fifty-four generations). See Tables III-V. The cross- 

 mated mothers in each case gave a normal result (i^i), without any in- 

 creased productiveness on account of the unrelatedness of their mates. 

 But their cross-bred offspring, with equal regularity, showed increased 

 productiveness (F 2 ), which surpassed that of either parent race under the 

 same conditions. 



At the same time Mr. Barrows made crosses between two sub-families 

 of the A race (A and A'), which had arisen from a common ancestry 

 some eleven generations previously. See Figure 1, A and A', and Table 

 VI. The two cross-mated mothers had broods considerably smaller than 

 their sub-family (A') was at the time producing, but their cross-bml 



