THE EXOSONE^E. 237 



The earliest stages of development closely resemble corresponding stages 

 in the Polychasta in general. Complete segmentation results in division into 

 four cells from which the four primary micromcres are derived by further 

 division. 



But I have not attempted to follow out the details of the cell- 

 lineage — a task which, in view of the minuteness of the ova and the small 

 number of stages available, would be one of peculiar difficulty. I can only 

 record the increase in the number of the micromeres until, while the four 

 macromeres still remain unaltered, they come to form a cap of small cells on 

 one side of the embryo. 



I have not succeeded in observing a stage when the micromeres have 

 completely enclosed the mass of macromeres. Or perhaps it would be more 

 correct to say, that when such an investment has taken place, some differ- 

 entiation has already been effected. Since no stage intermediate between 

 those represented respectively by figs. 52 and 53 of PL 18 has been observed, 

 it would appear probable that a process of differentiation goes on pan passu 

 with the extension of the investing layer of micromeres. This differentiation 

 results in the formation of two centres or areas of development — the dorsally 

 situated dorsal plate (PI. 18. figs. 54-57, d.pl.) and the ventral stomodceal 

 tract (PL 18. figs. 54-57 and figs. 58 & 59, st.). The former is early 

 distinguishable into anterior and posterior lobes, and consists at first of only 

 a single layer of cells. The latter at the outset appears as a small rounded 

 group of cells irregular in shape and arrangement in the middle of the 

 surface of the embryo destined to be ventral, and about midway between the 

 two poles. 



The dorsal plate sends offposteriorly a pair of outgrowths which eventually 

 meet and unite at the posterior end of the bod}-. These outgrowths, the 

 (terminal bands (g.b.), are separated from one another both on the dorsal and 

 on the ventral surface by areas composed of ectoderm (enclosing the mass 

 of yolk). Sections of embryos at this stage (PL 18. fig. 59) show that each 

 germinal band is composed of two layers, a superficial with smaller nuclei 

 — the ectoderm, — and a deeper with larger nuclei— the mesoderm. 



The stomodajal area has in the meantime been undergoing differentiation. 

 A depression has appeared on the surface which deepens and penetrates into 

 the mass of cells as a sharply-defined cylindrical pit. Latterly the stomodreal 

 rudiment sends out a string of cells which extend to the angles of the dorsal 

 plate, or more exactlv, of the posterior lobe of the latter. The destiny of 

 those clearly defined bands passing out right and left of the stomoda?um and 

 joining the apical plate dorsally, has not been followed. They very soon 

 lose their distinctness. 



Each germinal band now undergoes a process of transverse segmentation 

 (PL 18. figs. 62-65). In the region at the posterior end in which the two 

 have become united they give rise to the rudiment of the pygidial segment. 



LINN. JOURN. ZOOLOGY, VOL. XXXIV. 18 



