CASTLE. — INBREEDING, CROSS-BREEDING, AND SELECTION. 753 



same two seasons. Hence it seems not improbable that inbreeding was 

 causing a decline of fertility in series M and N, while series A uuder 

 continuous inbreeding was gradually recovering its lost fertility. 



But it is not certain that the change in productiveness of M and of N 

 was due to inbreeding ; it may have been due wholly to changed condi- 

 tions. The cultures of the two years were made by different persons, 

 and their methods may have 

 been different enough to account 

 for the difference in results. 



There can be no doubt, how- 

 ever, that a change had occurred 

 in series J, which was now twice 

 as productive as it had been the 

 previous year, equalling the M 

 and even surpassing the N se- 

 ries. This result was so surpris- 

 ing that as soon as it became 

 fully evident, a second line (A') 

 of the A series was started, 

 drawn from a different culture 

 jar. This showed a productive- 

 ness similar to that of the A 

 series the previous year. (See 

 Figure 1, A and A'.) Accord- 

 ingly there would seem to have 

 occurred a marked increase of 

 fertility in one jar of the A 

 series, but not in another. 



When the cultures were 

 transferred to a warm char ber 

 in January, 1905, M and A Oct. mm 

 responded promptly by a sharp FlGURE 3 Svnchronous cultures of 190 4-05. 

 rise in productiveness. Their 



courses through the next five months were closely parallel (see Figure 3). 

 Series N and A' responded in a similar way, but without attaining so 

 high an average productiveness, series N because of a selection of broods 

 of low productiveness (see page 756 ), series A' because of inherited low 

 productiveness. 



In the year 1904-5 two independent lines were maintained by Mr. 

 Barrows in each of the series A, 31, and N. The record of each of these 



VOL XLI. — 48 



