282 W, C. ALLEE 



occurs throughout the summer but is very rare during August and 

 September. As the weather becomes colder, a new period sets in, 

 but this is of much shorter duration than the spring periods. The 

 shortening cannot be wholly due to the increasing coldness because 

 animals brought into the laboratory and kept under normal con- 

 ditions do not long continue breeding. In these animals in the 

 laboratory, however, a new period of breeding begins about the 

 first of December. Staiting with a few individuals it slowly 

 increases in importance until by the middle of January, it is the 

 dominant activity of the animals. Curiously enough this is 

 much more pronounced in animals kept at temperatures about 

 5°C. than in those at 20°. In these laboratory isopods the season 

 stops about the time it is reaching its height in the field. 



One sign of the approach of the breeding season is the increased 

 tendency to collect in bunches. Bunching is apt to occur at any 

 time during the year if conditions become unfavorable, as when 

 there is a sudden drop in temperature, but the bunching tendency 

 of the breeding season is even stronger. Often these close irregu- 

 lar groupings occur, containing six or eight individuals. This 

 is especially apt to happen when the animals are stirred in a 

 current so that they are thrown against each other. The copula- 

 tion occurs much as Holmes has described for amphipods (Holmes 

 '03, p. 288) . The females may become quite helpless as the brood 

 pouch develops and unless they are clinging to some support, they 

 are often brought to the surface and float around ventral side up, 

 entirely unable to right themselves or to regain the bottom unless 

 they chance upon some solid object. 



The effect of the breeding season on the rheotactic response of 

 stream Aselli is shown in table 3. The first part of the table gives 

 the cycle of reactions due to the breeding season as shown in the 

 field experiments from April till October. The second part 

 traces the progress of these influences upon laboratory stock dur- 

 ing the winter months. One of the most noticeable changes in 

 the rheotactic response is the marked decrease in the percentage 

 of positive responses. Another almost as striking is the extreme 

 variability in animals selected at random from the breeding stock. 

 This variability is not so apparent when the same animals are 



