52 

 is equal and rr.f^iilar. This arranpe^e-^t of the blastoT^pres , 



hov/ever, is of very short duration, for soon a separation 



ce 

 takes nla* betv/een the cells of the lov;er quartet and 



two of therr, roll away fro the olan^^ of separation in one 

 direction; the other two Tovinc^ out in the onoosite direc- 

 tion. In this -i^ration the blastorreres "ove throuPh an 

 angle of 45 dep:rpes or ^ore, and finally co-e to lie in 

 such a -position as to forii; a semicircular nlate as shov/n 

 in Figures 13 and 14. The ser^aration and rotation of the 

 cells of one quartet seeTis to be constant in its occurence; 

 but the final arrangement of the blastomeres is not always 

 as rep:ular and definite as that shov/n ii the figures. At 

 tiS'es they are rr.ore loosely a-^d irregularly connected, and 

 may assume rela'ive "ositions si'-ilar to that shown by 

 Lletschnilroff for Oceania armata in Figure 34, Plate 1, of 

 his "Embryologische Studinn." In the case referre' to the 

 blastomeres are ?.o srtroa.' out that the individuals, with 

 three exceptions, touch only one of their fellows, thus 



