no 



W. J. CROZIER. 



wise symmetrical distribution seems omitted, or suppressed (Fig. 

 4). It seems possible, if not probable, that each gland may 

 undergo a more or less cyclic sequence of growth and shrinkage. 



45 

 40 



35 

 30 



as 



20 

 15 

 10 



FREQUENCY 



012345676 10 12 14 16 



nunftERo/ GLANDS. 



FIG. 2. Showing the frequency with which different numbers of glands oc- 

 curred in a series of 166 individuals; no obviously irregular sets of glands 

 were included (see text). 



Independently of such variations, there is a still further diver- 

 sity of numbers and sizes among the different sets. As to the 

 number of glands, some examples are given in Fig. 3. In the in- 

 stances illustrated, and indeed throughout series of several hun- 



FIG. 3. Examples of symmetrical sets of "glands" (X 2/3). 



dred animals each, examined at different times since 1915? it is 

 generally true that if the number of glands in a set is low, their 

 size is individually larger than in the case of groups composed of 

 greater numbers. Within certain limits, which do not concern 



