ON THE DEVELOPMENT OF SAGITTA. 379 



are often alternately large and small; the large ones run up 

 into cells with distinct cell boundaries on each side, though 

 continuous with the muscles below, while the smaller ones 

 end abruptly. The conclusion from these facts is that in 

 Sagitta the muscle-cells retain their protoplasmic character, 

 with the nucleus at the inner surface next to the coelom, and 

 that the muscular part of the cell has a greater length than 

 the protoplasmic portion, so that in transverse sections not 

 all the muscles appear to be continuous with the protoplasmic 

 part. Since the coelom is bounded by muscle-cells there is no 

 separate peritoneal epithelium, and this conclusion is sup- 

 ported by the way in which, in adult Sagitta, the whole 

 mesoderm passes into an epithelium one cell thick at the 

 lateral lines, and by the fact that in stained larvfe nuclei are 

 seen in rows along the muscles, but not between them. 



Returning to the structure of the third-day larva, it is 

 found that the rest of the mesodermal strands is much loosei*, 

 and a coelom ic cavity is being formed, but is still crossed 

 frequently by strands of protoplasm. In the ganglion no 

 important alteration has occurred, but the nuclei seem to have 

 increased in number, and the fibrous portion is becoming 

 more conspicuous between the two groups of nuclei. 



It is during the third day that the genital cells cross to the 

 body-wall and that the transverse septum is formed, but 

 although many series of sections were cut in the hope of 

 throwing light on this question, yet none proved very success- 

 ful. Fig. 32 represents a section through the male genital 

 cells at the beginning of their movement, and shows how they 

 are connected by protoplasmic strands with the mesoderm of 

 the body-wall. In the lower half of the figure a nucleus 

 {nuc. g. e.) is seen close to the genital cell, but outside it; and 

 this is almost certainly the nucleus of one of the investing 

 cells which probably give rise to the septum. 



In fig. 33 it is seen that the alimentary canal is now quite 

 thick, and the structure of the ventral ganglion is well shown • 

 the nuclear aggregations are sharply distinguished from the 

 epidermis, and connected vontrally by a bridge of '' Punkt- 



