236 SCIENCE PROGRESS. 



the environmental condition which in previous researches 

 has generally been examined most completely, and which 

 has been found of such powerful influence. The reason 

 of this is twofold. In the first place, nothing is known as 

 to the natural food of these larvae ; and in the second place, 

 they seem to be able to develop equally well whether food 

 is present or absent. In all probability they feed on the 

 other minute organisms with which the sea water teems. 

 If this is so, and if in the early stages of development sup- 

 plies of food are of importance to growth, it follows that 

 when a small number of larvse are allowed to develop in a 

 given volume of water they ought to grow faster than when 

 there is a large number, for then there is only a much 

 smaller proportionate amount of food material available. 

 Experiment proved that within certain limits the size of the 

 larvse is not influenced by the number developing together. 

 In every experiment the number of larvse present was esti- 

 mated by counting under the microscope those present in a 

 known volume of water, and it was found, as a mean of 

 thirty experiments in which there were on an average 4000 

 larvse per litre present, the average size was almost exactly 

 the same in thirty other experiments in which the number 

 of larvse was about 10,000 per litre. It was in fact found 

 that the larvse were undiminished in size if they were 

 kept under about 30,000 per litre. It must be concluded 

 therefore that in the first few days of their development 

 these larvae have enough food material stored in their bodies 

 for their immediate wants. 



In all the observations thus far described reference has 

 been made only to the body lengths of the larvse. As has 

 been mentioned, the arm lengths were measured in addition. 

 Hence it is possible to institute a comparison as to the 

 effects produced in both cases. These effects, contrary to 

 what one would suppose, run on by no means parallel lines. 

 Thus, as we have just seen, the body lengths are practically 

 unaffected by the number of larvse developing in a given 

 volume of water. The arm lengths react most markedly, 

 so that with 4000 larvse per litre the aboral and oral arm 

 lengths are respectively 13 '4 and 15*9 per cent, shorter than 



