CASTLE. — INBREEDING, CROSS-BREEDING, AND SELECTION. 739 



In generation 43 of the A series we have indications of the possible 

 return of a period of depression. Ten fertile pairs produced on the 

 average 62.8 young, the maximum number being 81. Two other pairs 

 were sterile. During the next four generations (44-47, reared June- 

 August, 1904), no record was kept of the number of young produced. 

 Twenty pairs in all were used, all of which proved fertile. The young 

 constituting the 47th generation emerged as imagoes about August 15, 

 1904, and contrary to previous usage were not separated, but were left in 

 the jar where they had emerged, to mate freely inter se. About October 

 10, animals were withdrawn from the jar to start a new series of 

 generations. In the meantime two generations had probably been 

 produced, allowing of possible mating of cousins, rather than of brothers 

 and sisters, as in all previous generations. See page 751 as to the effect 

 of mating cousins. 



On October 10, 1904, Mr. W. M. Barrows took charge of certain of 

 the A jars, removing ten pairs from jar A 47.2 to form what we shall 

 call generation 50. At least eight of the ten pairs were fertile, but they 

 showed wide variation in the size of their broods, ranging from 12 to 239 

 young. The average size of the six complete broods, that is, broods 

 reared in jars where the mother lived for at least three weeks, and in 

 which the food was properly fermented, was 100. The average was 

 higher still, viz. 187 in generation 51, but it fell to 63 in generation 52, 

 probably as a result of low temperature. It rose again sharply in gen- 

 eration 53, following a transfer to a warm chamber, and continued to rise 

 in the two following generations until an average of 308 was attained 

 in generation 55. Following the attainment of this, for the A race, un- 

 precedented maximum, there was a steady fall in the average to 135 in 

 generation 59, where the experiments for the year came to an end. 



A second branch of the A family (A', Figure 1), taken from the miscel- 

 laneously bred descendants of jar A 47.3, ran a course nearly parallel 

 with that of the branch already described, but attained a maximum aver- 

 age of only 156. The two were kept under identical conditions. 



Loio Productiveness Characteristic of the A Stock. 



Looking back over the history of the A series for fifty-nine generations of 

 almost continuous inbreeding (Figure 1), we see that the average number 

 of young produced by fertile pairs had never been as high as 200 until 

 the fifty-third generation (winter of 1905) ; usually it was under 100; 

 and about 1 in 5 of the pairs formed had been sterile (at least in the pe- 

 riod between the twelfth and forty-fourth generations). Control cultures 



