116 SELECTION IN CLADOCERA ON THE BASIS OF 



month) period of the experiment the mean for the minus strain for 1 

 two-month period (December 1916-January 1917) fell to within 48 

 seconds of that for the plus strain, yet, from consideration of the 

 data for this line for the 21 months at the end of the experiment, no 

 one would be inchned to question the validity of a real and large 

 difference in reactiveness between the two strains. Further, periods 

 of local shifts in reactiveness — changes which are beUeved to be 

 purely environmental and non-genetic — are seen in other lines, e. g., 

 Lines 695 and 740 (see pages 44 and 86), in which there is not evi- 

 dence of an effect of selection. The approximation of the curves 

 for the plus and minus strains during 1914 and again in 1915 

 seems to have been due to differential local environmental influences 

 (see also pages 140-142). 



There is an additional suggestive feature of the general course 

 of the curve for the Line 757 selection data. It is in the marked 

 general lowering of the mean reaction-time for both the plus and 

 minus strains of Line 757. The curve starts at 743 seconds for the 

 plus strain^ and at the close of the experiment ends at 124 seconds.^ 



The average reaction-time for the plus strain for the first 9 

 months of selection was 788 seconds and for the last 9 months 313 

 seconds, while the intervening 3 year-periods of data show a pro- 

 gressive and consistent lowering of this reaction-time as follows: 

 684, 605, 458. The form of the curve suggests that this lowering 

 had not ceased, but was continuing when the experiment closed. 



The curve shows a shght lowering of the reaction-time for the 

 757 minus strain, but the amount is small and the mean reaction- 

 times for the longer periods of the experiment do not indicate as 

 large and consistent lowering in this as with the plus strain. The 

 means for the minus strain for the longer periods are 839, 783, 707, 

 770, and 626 seconds. 



Extent of the Change in Reactiveness of the Line 757 Strains. 



The lowering of the mean reaction-time for the 757 plus strain 

 may be examined further. When it is recalled that a mean reaction- 

 time of 788 seconds (that for the plus strain for the first longer period) 

 means that by far the greater part of the individuals failed to react^ 

 to the light stimulation during the test (actually 77.2 per cent of the 

 individuals of the plus strain during the first nine-month period 

 failed to reach an end of the tank) and that a mean reaction-time 

 of 313 seconds indicates that a great preponderance of the individuals 

 responded to the light stimulation (only 15.1 per cent, less than one- 

 tenth as many as during the earlier period, failed to respond during 

 the final nine-month period) the decline in reaction-time from 788 

 seconds to 313 seconds takes a still greater significance. Not a 



^ Average for 2J^ months' data, 88 individual records. 



^ Average for 1 month's data, 42 individual records. 



' That is, reach an end of the tank during the 15-minute period of the test. 



