108 PART I. GENERAL ACCOUNT 



cell is a little flattened from side to side (Fig. 75 C, D). Its hind end is 

 wedged in between the anterior ends of the hind pair of internal cells, which 

 are very large and meet in the sagittal plain farther back (Figs. 75Д E). In 

 addition to these three large cells there are also two groups of very small 

 polygonal cells inside the embryo, lying symmetrically on either side in the 

 anterior part underneath the external layer of cells on the concave side, 

 between the half-submerged cells and the pair of internal macromeres 

 (Fig. 75/)). 



Thus amongst the distinct groups of blastomeres developed by this stage 

 we may already distinguish a mosaic of organ primordia. The small cells of 

 the external layer represent ectoderm and the three large internal cells the 

 beginning of an archenteron whose formation is not yet complete at this 

 stage. The small internal cells, to judge from their position between the 

 ectoderm and the rudimentary archenteron, are mesodermal in nature and 

 probably mesenchymal elements. Their isolation would argue in favour of 

 this supposition. The fate of the anterior pair of half-submerged macro- 

 meres remains unknown. It is interesting to note that the hind part of the 

 embryo still lacks any differentiation into presumptive areas (Fig. 15F). 

 The large blastomeres, rich in yolk, which form this part of the embryo, 

 possibly become cells of the archenteron or they may enter into the formation 

 of the coelom. 



It is clear from this discussion that we may regard this stage as representing 

 in essence a gastrula, despite the lack of a blastocoele and of an epithelial 

 presumptive archenteron. Nor is any blastopore formed and nothing remains 

 to show its primitive position. 



The early stages of Siboglinum thus show no trace of a radial or spiral 

 type of cleavage, but instead cleavage very soon produces a bilateral sym- 

 metry. This form of cleavage in all cases known to us is a secondarily 

 modified process and presumably this is true of the cleavage of Pogonophora 

 also. The peculiar nature of cleavage in Siboglinum is evidently associated 

 with the elongated shape of the egg and the abundance of yolk and oil 

 droplets in the cytoplasm, and it is possible that in species with nearly 

 spherical eggs cleavage may follow a more primitive pattern. To all appear- 

 ances the first two cleavage planes in Siboglinum are considerably deflected 

 from their presumed primitive median position. The reason for this appears to 

 be that cleavage in a very elongated egg is so affected by the surfeit of food 

 reserves that it is forced out of the longitudinal plane into a diagonal position. 



The peculiar character of cleavage in Siboglinum appears also in its 

 inequality. Amongst the blastomeres definite groups are sharply distinguished 



