20 K. S. LASHLEY 



number of individuals. When selection involves only a few in- 

 dividuals the chance selection of a somatic variate or a composite 

 variate of the third type above may counteract the effect of pre- 

 vious selection. Only when enough individuals are selected to 

 average out the composite variates can the results be relied upon. 

 It should be pointed out that if this criticism is carried out 

 consistently it results in practical experimental indeterminism, 

 since it may be said of any experiment that it is much less ex- 

 tensive than the course of evolution and is therefore not con- 

 clusive. Perhaps the only answer that can be given to the criti- 

 cism is a selection experiment extensive enough to make the 

 chance reversal of selection improbable and throw the burden of 

 proof upon those who advance the criticism. The selection ex- 

 periment with Hydra has been prolonged until it now seems to 

 meet this requirement. 



EXPERIMENTAL DATA 



Essentially the same technique was used as in the earlier ex- 

 periment. Each polyp was kept in a separate Stender dish; the 

 food supply, Cyclops, was distributed as uniformly as possible 

 to all polyps; the dishes were cleaned and sterilized and the cul- 

 ture fluid renewed every second day, except during the winter 

 months when the polyps were not budding. New buds were 

 removed from the cultures and recorded every second day. 



The first experiment, that previously reported, was brought to 

 an end by the freezing of the food pond and for the ensuing three 

 months very little food could be obtained for the polyps. As a 

 result only 6 of the original 50 lines survived the winter, these 

 being represented by members of the last selected generation. 

 When food was again available all descendants of these polyps 

 were kept until 24 members of each of the groups, varying in 

 the required direction, were obtained. From these, 48 lines were 

 established, half of which were selected in each direction from 

 the mean number of tentacles of the entire clone. During the 

 heat of the sunmier many lines died and were replaced from others 

 of the same group so that at the end of the experiment only 20 



