4i6 WILSON. [Vol. VL 



at p. 404, and I refer again to Diagram III, which shows its 

 marginal cells {cf. Figs. 25, 26, 31, 33, 36, 47 to 48 : it will 

 of course be understood that in such a Mercator's projection 

 some distortion is inevitable, which appears mainly in the exces- 

 sive size of a^, Jfi, c^, d^). It will be observed that a^-"^, b'^^, and 

 ^•2 lie opposite the respective cleavage-lines between the macro- 

 meres, — a point of importance in the subsequent orientation. 

 As the closure of the blastopore proceeds, c^^, b"^-^, and c^-^, 

 which from the first lie somewhat outside the others, are, as I 

 believe, left entirely behind, so as to form no part of the blasto- 

 pore-margin ; whereas the other cells are pushed downwards, 

 and their descendants form the lip of the blastopore, as shown 

 in the Diagram IV. To substantiate this statement I shall be 

 obliged to describe the history of the cells somewhat in detail 

 by returning to an earlier period. I have shown that, in the 

 course of the fifth cleavage, the secondary micromeres {a^, b"^, c^) 

 divide somewhat unequally, in a left-handed spiral (as seen from 

 above), the larger cell {a^-^, ^H ^2.1^ in each case lying some- 

 what above the other and in contact with the girdle-cells (Figs. 

 25, 26, 33, III). In their later divisions (Figs. 36, 45) the upper 

 cells divide vertically, so as to give rise to a series of large 

 polygonal cells, lying below the prototroch ; these I shall call 

 XhQ post-trochal cells {cf. Figs. 75, 76). It is between these cells 

 and the entomeres that the head-kidney takes up its definitive 

 position and begins to elongate. They are clear, show a marked 

 tendency to vacuolation, and it is these cells that I have sug- 

 gested at p. 403 may possibly be comparable with the velar or 

 external excretory organ of the molluscan veliger. 



Meanwhile the smaller or lower cells {a^-"^, b'^-''-, c^^) also 

 divide, but in a different manner. The first division is hori- 

 zontal (more or less oblique) and also unequal, a smaller cell 

 being budded off above, next the prototroch. That this takes 

 place in all of three of the cells in the same manner, I only infer 

 from their later arrangement ; but I have several times observed 

 the actual division, in the manner described, in c^-s^ and it seems 

 fair to conclude that it holds for the others also. The first divis- 

 ions of the third group of micromeres {a^ to d^), which mean- 

 while take place, are also horizontal {cf. Figs. 49, 50, 53). 



We have now arrived at a stage, however (Figs. 52 to 55), 

 where (T^-^ and a^-'^ can with certainty be recognized as the right and 



