PHYLf'lM ANNT^LATA 



485 



cephala (Fig. 379) they develop in a cavity in the operculum ; in 

 others, in the interior of the tube, between the bcdy of the worm 

 and the inner surface of the latter, or on its outer surface. In 

 some, again, though the ova do not remain in any way attached to 

 the parent wurm, they may be deposited in clumps or packets 

 enclosed in gelatinous matter. Usually they have no other 

 covering but the egg-membrane. 



The segmentation of the ovum in the Polychseta is unequal. 

 In the great majority the inequality between the megameres and 

 micromeres is very marked. In some Serpulids, however, the differ- 

 ence is very slight, and the two sets of cells are at first scarcely 

 distinguishable. In such cases the cells arrange themselves in 



Frn, 3.S0. — A, B, C, three stages in the development nf the trnchophore of Eupomatus, from 

 the side. av. anns ; J%. blastociele ; m. polar cells of tlie mesoderm ; ,mJ. niid-gnt ; )(. larval 

 head-nephrldiura ; ot. otolith ; .^7). neural iilate ; sf. stomodfeum ; -irk, preoral ciliated ring ; 

 ^rd'i post-oral ciliated rins?. (From Lang's Conf/inralive Airahhuy, after Hatschek.) 



such a way as to form the wall of a hollow sphere, the blnstida, 

 with an internal closed cavitj^, the segmentation-caxity. The 

 megameres, which may or may not have been distinct from 

 the first, lie on one side of the blastula; and soon this side 

 becomes invaginated (Fig. 380 A), the result being the forma- 

 tion of an embolic gastrula. In the great majority of forms 

 however, an e2nbolic gaslrula is formed after the manner already 

 described in the case of Nereis ; but forms of the process 

 of gastrulation intermediate between these two extremes have 

 been observed. The blastopore of the gastrula, however formed, 

 does not give rise directly either to the mouth or to the anus. 

 It becomes elongated into a slit which becomes closed up, and 

 the anus and proctodreum are formed by a fresh invagination 



