4o8 WILSON. [Vol. VI. 



we might have expected) ; but sometimes the reverse is the case 

 (Fig. 53). These are followed by x^, .x^ (Fig- 56). At this 

 period, therefore, the progeny of the somatoblast are approxi- 

 mately seventeen in number, as shown in the accompanying dia- 

 gram, I say approximately, because in some cases, at any rate, 



x^—x - X - X - X—x^ 

 / \ / \^ 



^2.1 j;i,l ^4,1 x^,i 



the progeny of x^ and x^ begin to divide before the equal division 

 of the two Jf's (see Fig. 55), and it now becomes extremely dif- 

 ficult to distinguish certainly between them and the cells at the 

 side of the embryo derived from the second and third sets of 

 micromeres {i.e. <?', d^, a^-'^, c^-^). A comparison of the figures 

 (43, 46, 54, 55, etc.) will show the anterior ;t:-cells passing for- 

 wards and growing over the primary mesoblasts. This is shown 

 with perfect clearness from the side in optical section (Figs. 71 

 to 73, PI. XIX). Actual sections through these stages confirm 

 the optical sections in every respect. 



During these stages the primary mesoblasts have budded 

 several times, giving rise to the group of small secondary meso- 

 blast-cells (Figs. 55 to 57) that form the posterior lip of the 

 blastopore. They may be distinguished from the .r-cells by 

 their nuclei, which are smaller and appear coarsely granular 

 on account of the character of the chromatic network. (In the 

 figures, these nuclei are made granular, while the others are left 

 clear.) 



The ninth division is shown in Fig. 59. Each of the telo- 

 blasts divides transversely into unequal parts, of which the 

 smaller is anterior. The inequality in size is much less marked 

 than in the preceding unequal division, but it is always dis- 

 tinctly recognizable. This stage is usually attained from the 

 tenth to the eleventh hour, but the specimen figured is thirteen 

 hours old. Soon afterwards (Fig. 60) the cilia are put forth in 

 the manner already described, and the embryo begins to rotate. 



Up to this point I have repeatedly observed every division of 

 the Jf-cells in the living embryos (vv^ith the exception of the 



