190 CERTAIN LAWS OF VARIATION. 



the earlier stages of development than in the later, but 

 what is the numerical expression of this difference? 

 Such an expression can be obtained in two ways: di- 

 rectly, as the result of experiment ; and indirectly, from 

 certain considerations as to the rate of growth and per- 

 sistence of variations. 



To determine the effect of environment on growth, 

 almost any organism can be made use of, but it would 

 obviously be exceedingly troublesome and laborious to 

 work with the higher organisms, such as mammals. In 

 them the earliest stages of embryonic development 

 would be especially difficult to reach. With many of 

 the lower organisms, however, all such difficulties are 

 avoided, and an inexhaustible supply of material can 

 readily be obtained. For these reasons the author at- 

 tempted to investigate the question at issue by observa- 

 tions on the larvae of sea-urchins. The method of ex- 

 periment has already been indicated in Chapter IV., 

 so that it is unnecessary to refer to it again. 



The object in view was to determine the permanent 

 effect of some abnormal environmental condition, act- 

 ing at various periods of development, on the size of the 

 larvae. It was found that their growth practically 

 ceased after 6 or 8 days, and hence any effect then 

 found to be present was fixed and ineradicable, so far 

 as the larval stage of the organism was concerned. The 

 most convenient environmental condition to work with 

 proved to be temperatiire, for it is easily controlled, 

 and the effect produced may be considerable. Thus 

 larvse kept at 10 C. during growth are some 25 per 

 cent, smaller than those kept at 20 0.* 



*Phil. Trans. 1898, B. p. 481. 



