CASTLE. — INBREEDING, CROSS-BREEDING, AND SELECTION. 735 



recorded as sterile, and in some cases each parent was then placed with 

 a new mate to ascertain, if possible, in which of the two parents the 

 sterility lay. In some cases one or other of the two was found to be 

 fertile by a different mate, but in no case were both found to be fertile 

 by new mates. This result indicates that the sterility observed is not 

 relative, but absolute ; for an animal which is sterile toward one mate is 

 sterile toward all mates. 



Brief History of the A Series. 



The experiment which was begun first and has been continued longest 

 constitutes what we shall call the A series (see Table I and Figure 1). 

 During the first year of this experiment (1901-2), when it was carried on 

 by F. W. Carpenter, no record was kept of the number of sterile pairs, or 

 the total number of young produced by each pair (except in generation 6, 

 Table I), attention being directed chiefly to the effects of the inbreeding 

 on variability. But beginning with the thirteenth generation, when A. H. 

 Clark took charge of the experiment, records as complete as possible were 

 kept of the number of fertile pairs and the number of young produced by 

 each pair. Whenever a female was accidentally killed or escaped from the 

 breeding-jar (as sometimes happened when pupae were being removed) 

 before she had had an opportunity to lay her full quota of eggs, her 

 brood was not used in making up the average for the generation, though 

 the pair to which she belonged was recorded as fertile. 



From generation 13 to generation 24 inclusive, Sept. 1902-June, 

 1903, the experiment was in charge of A. H. Clark and showed results 

 of a pretty uniform character. Between seventeen and eighteen per 

 cent of the pairs formed were sterile. The average number of young 

 produced by a fertile pair was about 39 ; the maximum number, 145. 



As a control on the results given by the A series during this period, 

 pairs were from time to time taken from a stock-jar established in the 

 following way. In October, 1902, flies captured at three different 

 localities several miles apart were placed in a large covered jar within 

 which was a smaller open jar containing fermenting banana. The flies 

 soon increased, forming a vigorous colony, which was given no further 

 attention beyond the occasional addition of fresh food. Three pairs 

 taken from this jar in the latter part of October, 1902, produced remark- 

 ably vigorous broods numbering respectively 529, 560, and 2G0 young, 

 an average of 450 young to a brood. Two pairs taken from the same 

 jar about five weeks later gave a similar result, the broods numbering 

 623 and 361 respectively. 



