88 



blastomeres are connected as stated and with the Umiting laj^ers of 

 c^^toplasm already referred to. These two layers are continued 

 marginally into the investing protoplasm of the yolk — the so-called 

 periblast — which latter like the ventral sheet of the blastodisc is 

 connected to the protoplasmic network of the yolk. 



This picture of the conversion of the polar mass of cytoplasm 

 into the nuclear area with imperfectly separated cell masses of 

 cytoplasm around each nucleus recalls the considerations raised by 

 the late Prof. Sedgwick * from his studj' of the development of 

 Peripatus. Haidenhain | has also drawn attention again to the 

 intimate association of the cells of tissues by protoplasmic connec- 

 tions. I find the same method of division occurs in the lamprey, 

 in Lepidosteus and in Amia, and also in the fowl and in the gull. 



Gastrulation. — So far the account I have given will be 

 found to agree with the descriptions of the processes of segmentation 

 of most of my predecessors. The only important difference is that 

 the sub -blastodisc sheet of protoplasm — the so-called sub-germinal 

 periblast — is not separated from the blastomeres as described by 

 Agassiz and Whitman.;]^ I now come to the consideration of the 

 changes which follow segmentation, which are to some extent dove- 

 tailed into the process of segmentation, and which have given rise 

 to a host of theories. It will be found that the facts are even more 

 strange than the theories. 



First, with regard to the cells at the margin. It has long been 

 known that during the latter part of the segmentation phase nuclei 

 appear in the marginal protoplasm of the blastodisc where the latter 

 merges in the peripheral protoplasm (fig. II). These nuclei arise 

 in the ventral part of the series of single marginal cells, and only 

 differ from the other nuclei of the blastodisc in not leading to such 

 distinct vacuolation of the surrounding cytoplasm. Vacuoles are 

 developed, however, m this marginal syncytium, serving particu- 

 larly to separate the outer non-nucleated protoplasm of the margin 

 from the lower nucleated portion. Pari passu with the fragmenta- 

 tion internal to them the nuclei of the margin divide. The nucleus 

 elongates, and two nuclei arise by constriction. In some cases at 

 least an appearance is met with which points to chromatin segre- 

 gation, but a distinct spindle is not formed. 



* 1886, Q.J.M.S., v., 26. t Plasma und:Zelle. 



t'ioc. cit. 



