7o2 PROCEEDINGS OF THE AMERICAN ACADEMY. 



already low in generation 6, and kept on declining until, in January, 1903 

 (generation 18), the average had sunk to 21. From this point on a 

 gradual rise followed, which ended abruptly in generation 29 (September, 

 1903). Then the productiveness fell off until, in January, 1904, the 

 average brood numbered 29 young. Again it began to rise, as it had 

 done the previous year. The rise was more rapid than in 1903, and 

 reached a point somewhat higher. It also culminated (generation 41) 

 sooner (in May rather than in September), and then began to decline 

 again. In the fall of 1904 the productiveness of the A series was just 

 about what it had been the previous spring, but showed a tendency to 

 decline. Early in January, about the time recovery had begun in the 

 two previous years, the cultures of the A series, together with those of 

 series M and iV, were transferred to a warm chamber. A remarkable 

 increase in productiveness followed, which culminated in March (genera- 

 tion 55). The A broods then averaged 308 to a pair, but from that 

 time on they steadily diminished in size, until at the end of May they had 

 fallen to 135. 



Each year the A series showed a gradual rise of productiveness fol- 

 lowed by an abrupt decline. The annual rise follows a depression which 

 occurs under fall conditions, in which low temperature is doubtless an 

 important factor. Out of doors the rise would doubtless come in the 

 spring, but in the warm laboratory it sets in earlier and culminates 

 earlier. The culmination occurs in the three successive years in Septem- 

 ber, in May, and in March respectively. Each year the A series reached 

 a higher maximum, until one branch of it at least seems to have become 

 fully normal in productiveness, though another branch (A 1 , Figure 1) 

 had still about the same productiveness in 1904-5 as in the previous 

 year. 



The M and N series (Figure 2) during the first nine months of 

 inbreeding (generations 1-14) showed no marked change in productive- 

 ness. The M series showed on the whole a rise, the N series a fall, in 

 productiveness. Both were evidently influenced by the same set of 

 external conditions because the M and N curves generally show corre- 

 sponding elevations and depressions. The N cultures were on the whole 

 more vigorous ; they showed no sterility whatever, whereas sterility did 

 occur sporadically in the M series. Nevertheless the N series was appar- 

 ently declining in fertility (perhaps because it had been above the normal 

 in fecundity), while M was rising. In the second year of inbreeding 

 (generations 19-28) both the M and the N series were less productive 

 than in the previous year. The reverse was true of the A series in the 



