108 MEMOIKS OF THE NATIONAL ACADEMY OF SCIENCES. 



The "primitive streak" of Caldwell does not seem to be present in the larva of P. architecta. 



The "nephridial pit" is of ectodermal origin. 



The mesoderm arises, for the most part, from the lips of the blastopore. Archenteric 

 diverticula are not present in the larva of /'. architecta, but there is a sac-like formation of 

 mesoderm cells in the anterior end which forms the lining of the preoral lobe and which gives 

 rise to a mesentery between the lobe and collar cavities. 



The lining of the collar cavity does not arise from a mesodermal sac. It is formed by 

 isolated mesoderm cells which arrange themselves on the somatic wall leaving the splanchnic wall 

 practically without any lining. 



In the larva of /'. architt eta the mesodermal lining of the trunk cavity is complete, covering 

 both the somatic and splanchnic walls, and it seems probable that it arises from cells forming the 

 base of the nephridial diverticula. There is a mesentery between the cavities of the trunk and 

 collar. 



A stomodseum and proctodeum are not present. The blastopore becomes the mouth, the 

 anus arises quite late in the early life of the embryo, and the rectum is formed as an outgrowth 

 of the blind end of the archenteron. 



The nephridial canals, at least, have their origin in a single median pit which soon branches 

 into two intercellular tubes. We have not found anv evidence that the excretory cells of the 

 nephridia are formed from free mesoderm cells attaching themselves to the blind end of the 

 nephridial canals. 



The " neuropore" and "subneural gland," which Masterman has described, do not exist in the 

 Actinotrocha examined, although imperfectly preserved specimens show unusual structures which 

 might lie taken for these organs. 



Masterman's "subneural sinus" is not present either, although there is a space below the 

 ganglion which is free from mesodermal strands. The "atrial grooves''' which Masterman says 

 exist are present in the larvae we have studied, but we can not consider that they have the signifi- 

 cance that he assigns to them. Occasionally grooves are found which might be comparable to 

 Masterman's "oral grooves, '" but they are due to imperfect fixation. The stomach diverticula 

 exist in one species that we have examined, but they do not impress us as being of notochordal 

 nature, as Koule and Masterman have claimed. 



There is a subepidermal layer of nervous tissue throughout the body. Extending anteri- 

 orly from the ganglion, which is situated on the median dorsal surface of the hood, are three 

 longitudinal nerves, which finally become continuous with a nervous ring running around the 

 edge of the hood. From the posterior side of the ganglion two parallel tracts of nerve fibres 

 issue and pass posteriorly along the dorsal collar wall until they reach the circle of tentacles, 

 where most of them follow the line of insertion of the collar trunk mesentery, and give rise to a 

 collar nerve ring. The nerve fibres from the edge of the preoral hood do not pass up to the 

 ganglion from the point of attachment of the hood on to the collar wall, as Masterman has 

 described, but they make a sharp turn, running posteriorly and obliquely along the lateral and 

 ventral wall of the collar, where they form two definite nerve tracts which become lost in the 

 region of the collar nerve ring. While there may be nerve fibres passing from the ganglion out 

 in all directions over the surface of the hood, we have not been able to make them out, nor do we 

 find any definite nervous tract running along the dorsal or ventral wall of the trunk segment. 



There is one pair of retractor muscles extending from the region of the ganglion to the 

 collar walls, in the region of the first and second pairs of tentacles, and besides these, in one 

 Actinotrocha that we have examined, there is another pair extending from the sides of the ganglion 

 to the. ventral w all of the hood. In this latter Actinotrocha there is an extensive layer of muscle 

 fibres in the wall of the hood and also a ring of fibres around the edge of the latter. A pair of 

 longitudinal muscle tracts extend from the region of the ganglion, along the dorsal wall of the 

 Actinotrocha, to the perianal ring, and there is a similar pair of tracts extending along the ventral 

 wall of the collar and trunk. A ring of muscle fibres run parallel with the ring nerve, between 

 the collar and trunk segments. Beside these muscle tracts there is a layer of circular fibres in 



