354 E. W. MACBRIDE. 



Professor Lankester's later paper he asserts that he and 

 Willey showed that there were no atrial folds at all; 

 since the atrial cavity originated as an insinking formed 

 between two free edges which were only the metapleura. 



A renewed and careful examination of my preparations 

 has led me to a different conception of the origin of the 

 atrial cavity. I hold that there are atrial ridges, and that 

 the first rudiments of these make their appearance long 

 before the stage in which they were observed by Lankester 

 and Willey. 



In transverse sections of larvEe of the latest stage to which 

 they can be artificially reared — that is to say, of larvte at the 

 period of the formation of the mouth — it can be observed that 

 the section of the body has a very different shape in the 

 pharyngeal from what it has in 6he posterior region of the 

 animal. In specimens preserved in osmic acid it can be seen 

 that in the pharyngeal region there are two latero-ventral 

 ridges, sometimes extraordinarily dilated. By carefully 

 following up the sections it can be seen that the cavities 

 contained in these ridges are extensions of the cavities of the 

 first myotome (somite) on each side (fig. 1). This pair of 

 myotomes was proved by me to have an independent origin 

 from the alimentary canal, and was compared to the collar 

 I'egion of Balanoglossus, for which reason its cavity on each 

 side is termed the collar cavity in this paper. The dilation 

 of the collar cavities and the consequent extension of the 

 latero-ventral ridges vary a good deal, and seem to depend 

 on the amount of fluid contained in them. As we follow the 

 sections back the collar cavities become more and more 

 restricted to the latero-ventral angles of the animal, and 

 between them and the pharynx is interposed on each side the 

 splanchnoccele, that is the coelomic tube formed by the 

 coalescence of the ventral portions of the posterior myotomes 

 (figs. 2 and 3). Our knowledge of the formation of the 

 splanchnoccele is most unsatisfactory. Kowalevsky's obser- 

 vations on this point do not rest on sections. In larvae of 

 the age under discussion, in the anterior region the splanch- 



