574 CRESSWELL SHEAKEK. 



gut wall above the anal vesicle. Close examination of the 

 sections shows them first as two cells in the ventral wall of 

 the stomach, and then the gut. The change that has to do 

 mostly with bringing this about is the great increase in the 

 dorsal surface of the gastrula and the consequent narrowing 

 of the blastoporal surface, changing the large ventral to a 

 small ventro-lateral surface. At a time when the anal open- 

 ing of the gut has not been established they occupy about the 

 mid-region of the archenteron. At the period when the anus 

 breaks through they have already moved into the anal end. 



The blastocool during this time is still small, and has not 

 undergone the great increase it shows shortly after this 

 period, as only a trace of it can be seen between the gut and 

 the ectoderm. This adds somewhat to the difficulty of deter- 

 mining how the various steps in the process take place. The 

 primitive trochophore about this time begins to assume its 

 typical shape ; up to this the round shape of the gastrula has 

 been retained. During early gastrulation before the division 

 of the coelomesoblast cell, as shown in Text-fig. 17, I have 

 been quite unable to distinguish it from any of the other endo- 

 derm cells. No conspicuous cell is seen forcing its way into 

 the segmentation-cavity as shown by Hatschek (17) and Soulier 

 (38), and I believe that both these investigators have been 

 mistaken in their identification of the coelomesoblast cell. 

 The cell shown in Hatschek's figs. 25-36, and in Soulier's 

 figs. 25-27 and 33 and 34, and identified by them as the 

 coelomesoblast, are really the right and left portions of the 

 ectomesoblast. At a later stage they give rise to the head- 

 kidneys. The real coelomesoblast at this period still lies in 

 the gut- wall, and not free in the blastoca^l. 



In the late gastrula stages the right and left portions of 

 the larval mesoblast appear as shown in 'Pext-fig. 18. In all 

 respects these cells answer to the mesoderm cells of Hatschek's 

 figs. 25-37, fig. 9 of this paper corresponding to Hatschek's 

 fig. 33. By a comparison of figs. 9, 10, 12, 13, 15, 16, the 

 various changes will be seen by which these cells are trans- 

 formed almost entirely into the head-kidneys. In fig. 16 the 



