THE CAUSES OF VARIATION. 239 



urchin larvae there was a tendency for the larger individuals 

 to be preserved by the action of Natural Selection. When the 

 individuals vary on an average by only 6 per cent, they are 

 all so similar in size that the advantage the larger individuals 

 have over the smaller in the struggle for existence is only 

 very slight, and hence Natural Selection can only act very 

 slowly and haltingly. Suppose, on the other hand, that by the 

 action of varying conditions of environment the larvae vary 

 on an average by 20 per cent., then the larger individuals 

 will have a very much greater advantage over the smaller 

 ones than before, and will consequently be preserved in 

 much greater proportionate numbers. Now these larvae, 

 though caused by the action of a variable environment to 

 assume a considerably larger size than they would have had 

 under a more constant one, would nevertheless under this 

 more constant condition have still been of larger size than 

 the other individuals, and hence will transmit to the next 

 generation that characteristic of increased size, though to a 

 smaller degree. Thus though the selected larvae might be 

 20 per cent, larger than the average, they only contain a 

 similar germ-plasm to larvae 6 per cent, larger than the 

 average when grown under a more constant environment. 

 Hence their offspring will only be some 6 per cent, larger 

 than those of the previous generation. 



We now see the importance of the fact established in 

 the earlier part of this paper, vi2., that the ova are more 

 sensitive to changes of environment, such as temperaiture, at 

 the time of impregnation tham subsequently. Thus, whilst 

 the temperature of the water in which these larvae develop 

 under natural conditions may show but slight variations on 

 an average in periods extending over several weeks, yet the 

 extreme variations within such a period may be considerable. 

 As, therefore, ova would probably happen to undergo im- 

 pregnation at all times in the breeding season, whatever the 

 temperature of the water, the larvae developed therefrom 

 would have a greater variability than if the ova had been 

 no more sensitive to temperature at one time than at 

 another. Though it has not been conclusively proved that 

 the ovum is specially sensitive at the time of impregnation 



