EXPEKIMENTS ON LTMN^A. 161 



of water ; the other by placing two different quantities of 

 animals, from the same mass of eggs, in two aquaria of equal 

 size. All the conditions of exi^itence, and above all the supply 

 of food, were kept at the optimum. Consequently all the 

 animals were under equally favourable conditions, irrespective 

 only of the volume of water which fell to each animal's share ; 

 this varied at most between 100 and 2,000 cubic centimetres. 

 In both experiments the results were similar (fig. 43) : the 

 smaller the volume of water which fell to the share of each 

 animal, the shorter its shell remained ; and, moreover, it made 

 no differenee, with regard to the length the shell attained in the 

 different groups of animals, whether each isolated individual 

 had from the first a definite quantity of water allowed to it, 

 as in the first series of experiments, or whether several indi- 



FlG. 43. — Four equally old shells of Lr/mn<ea stagnaliB, hatched from the same mass of ova 

 but reared in different volumes of water ; a, in U)0 oiibic centimetres ; 6, in 250 cuBic 

 centimifcres ; c, in 600 cubic centimetres ; and d, in 2,000 cubic centimetres. 



viduals living together had a larger volume of water to share 

 among them in the same proportion. Thus I succeeded, under 

 conditions of existence otherwise identical, in establishing a 

 curve of growth for the Lymnsea corresponding to the volume 

 of water. This curve (fig. 44) shows that the favourable effect 

 of an increase of volume of water is highest between 1 00 and 500 

 cubic centimetres for each individual, and that it then gradually 

 decreases, till, at 5,000 cubic centimetres, it would seem to cease 

 entirely ; i.e. an increase of volume above this maximum has, 

 as it appears, no further effect whatever upon the rapidity of 

 growth. Thus the optimum of the volume of water which allows 

 the greatest possible length of shell to be attained by a Lymnsea 

 within a given time lies approximately between 4,000 and 5,000 

 cubic centimetres ; to determine the point exactly was impos- 

 sible for various reasons. ' The woodcut (fig. 43) exhibits the 



