2IO Edmund B. ffllson. 



that it may also enter the prototroch.^ The 28 (32?) trocho- 

 blasts are at first arranged in two roughly alternating rows en- 

 circling the embryo slightly above the equator; and the ciliary 

 plates of contiguous cells are still not united to form a continu- 

 ous ciliary girdle (Figs. 10-12). Later, extensive shif tings of 

 the cells occur in such wise that a principal circle of trochoblasts 

 is formed in a single circle completely surrounding the embryo, 

 bearing a perfectly continuous series of powerful cilia (Figs. 

 17-19). The cells in this row vary in number from 19 to 21 — 

 a fact of which no doubt is left by the study especially of acetic- 

 glycerine preparations, in which the cells may be seen with sche- 

 matic clearness. Posterior to this row lies a second row of smaller 

 elongated trochoblasts, which in the dorsal region become as 

 large as those of the principal row (Fig. 17). At this point, 

 therefore, where in so many trochophores a gap exists in the 

 prototroch, the ciliated belt is not only closed, but broader than 

 at any other point. At this point the prototroch is often three 

 cells wide; elsewhere I have not been able to distinguish three 

 rows of cilia as figured by Patten, though three such rows are 

 certainly present in Dentalhim. 



The trochophore of 24-30 hours (Fig. 17) is in the main sim- 

 ilar to that of Dentalium, as described in my former paper, but 

 the post-trochal region is relatively larger, the pre-trochal region 

 less pointed, the apical tuft shorter and broader, and the apical 

 plate less clearly marked off from the surrounding ectoblast as 

 may be very clearly seen in sagittal section. In this respect my 



^This derivation of the prototroch in Patella agrees closely with that of 

 Isclmochitou (Heath, 99), where two cells in each quadrant are likewise con- 

 tributed from i^-, and two from the second quartet except in the D-quad- 

 rant, wbere a non-ciliated dorsal gap exists from the first. I have determined 

 beyond doubt, I think, that at least two secondary trochoblasts are formed 

 in the mid-dorsal line, as shown in Figs. 10-12, where there are three such 

 trochoblasts in three of the quadrants and two in the fourth. There is further 

 no doubt whatever that the completed prototroch is closed in the mid-dorsal 

 line (Figs. 17-19). Robert describes the prototroch of Troclius as agreeing ex- 

 actly with that of Am[>hitrite and Arenicola, no cells being derived from the 

 first quartet except the primary trochoblasts. It appears to me, however, that 

 his observations do not fully establish this. {Cf. the useful comparative table 

 given by Robert at p. 420). 



