394 WILSON: [Vol. VI. 



period and necessitates the recognition of an intermediate stage. 

 The formation of the prototroch is, however, so intimately con- 

 nected with the first bilateral cleavages of the upper hemisphere 

 that they must be described together. 



I. {a) Bilateral Division of c^ a7id d^ ; {b) TJiird division of the 

 First Somatoblast ; ic) Fission of the Second Sojnatoblast 

 (Figs. 30-32). 



These three events occur nearly at the same time (about six 

 hours after fertilization), though there is a certain amount of 

 variation, {a) The first bilateral division always takes place in 

 the two posterior primary micromeres (c^.and d^), which divide 

 transversely into unequal parts, the anterior being the smaller. 

 The spindles are exactly radial in position {cf the spindles in 

 the same cells in the last spiral cleavage, Fig. 27), and the ante- 

 rior cells (^1-^ d^-^, Figs. 31, 32, 35) are perfectly symmetrical 

 with respect to the second cleavage-plane, which now forms the 

 median plane of the body. Since the posterior cells {c^, d^) long 

 retain their superiority in size, we may continue to call them 

 the posterior primary micromeres. The four cells thus pro- 

 duced form, as it were, an arch (Diagram II, p. 396), the ends of 

 which rest upon the girdle-cells, while the keystone is formed 

 by the posterior rosette-cell {d'^-'^). 



{b) Meanwhile a curious event happens, the exact nature of 

 which I only made out after repeated examination, both of liv- 

 ing specimens and of preparations. The first somatoblast {X), 

 namely, buds forth a small cell (;t'^) from its posterior or dorsal 

 border in the median line. This cell pushes before it the posterior 

 intermediate girdle-cell {d^-'^), which is thus carried tip into the 

 arcJi formed by d^, d^-^, d^-^, c^-'^, c^, and loses its connection zvith 

 the girdle (Fig. 31). The girdle is thus reduced from twelve to 

 eleven cells, and is interrupted in the median posterior (dorsal) 

 line. We shall see hereafter that this interesting occurrence is 

 the cause of the mediaji interruption of the prototi'och that appears 

 to be a common feature of the trochophore larva. As a rule, 

 this division of X takes place simultaneously with the bilateral 

 division of c^, d^, and the sudden transformation of the postero- 

 dorsal region is, in the living embryo, an interesting spectacle. 

 Sometimes, however, the division of X follows the others, as in 

 Fig. 31- 



