CERTAIN LAWS OF VARIATION. 217 



That want of adaptation should, as a rule, lead to in- 

 creased variability may not seem at first sight obvious, 

 but a little reflection will show, I think, that it is only 

 what one ought on a priori grounds to expect. Thus 

 supposing a group of organisms is subjected to unfav- 

 ourable environmental conditions, so that they are all 

 decreased in size by, on an average, say 10 per cent. 

 We know by experience that all organisms are not 

 equally affected by changed conditions. Some of them 

 are more resistant than the average, and others less. 

 Consequently, supposing the group of organisms had 

 originally varied between such and such limits, then on 

 reduction of their average size by this 10 per cent., 

 some of the small individuals, and likewise some of the 

 large ones, will be diminished by perhaps 12 per cent, 

 in size, and others by only 8 per cent., and hence the 

 range of variation of the whole group of organisms will 

 be widened. 



Actual proof of this contention has been obtained by 

 the author in the case of sea-urchin larvae. Thus, in 

 experiments of one kind and another, some 412 series of 

 measurements, each on 50 individuals, have been made 

 on the body length of Strongylocentrotus larvae after 6 

 or 8 days' growth under various favourable and unfav- 

 ourable conditions. The variability of each series was 

 determined by a method it is unnecessary to mention in 

 detail here,* that of the whole of the larvae being taken 

 as 100. All the series of larvae grown under so-called 

 " normal " conditions (i. e., in jars of sea-water kept at 

 as constant a temperature as possible) and of other 

 larvae which did not vary from the normal by more than 

 * Vide Phil. Trans., 1895, B. p. 617. ^ 



