366 TUNICATA. 



it has a dorso-ventral position. Only later does it shift farther back 

 and come to lie ventrally. This observation is of importance in 

 connection with the condition of this organ in Amphioxus, where 

 it undergoes a similar displacement. 



The digestive or pyloric gland arises as a caecal outgrowth at the 

 boundary between the stomach and the intestine ; it, however, soon 

 branches repeatedly, and these ramifications extend over the surface 

 of the intestine where, by anastomosing, they form a network (Fig. 

 175 A, dr, p. 379). It has been homologised by Willey (No. 54a) 

 with the hepatic caecum of Amphioxus. 



Peribranchial, atrial, or cloacal cavity. The first rudiments which 

 lead to the development of the peribranchial cavity are found, shortly 

 before the larva hatches, in the form of a pair of ectodermal invagina- 

 tions lying dorsally at the boundary between the sensoiy vesicle and 

 the trunk-ganglion, called by Metschnikoff, who was the first to 

 observe them, the cloacal vesicles (Fig. 167, cl, p. 356). Two 

 diverticula grow out from the pharynx towai"ds these invaginations, 

 one on each side, and fuse with them, thus giving rise to the first 

 gill-slits (Fig. 168, k, p. 358). According to Kowalevsky, a second 

 pair of these slits (k") soon forms in Phallusia behind the other in 

 the same way. If the interpretations of Kowalevsky and Seeliger 

 are correct, the cloacal vesicles, by enlarging, give rise to the paired 

 halves of the peribranchial cavity. In this case, the latter would be 

 lined throughout with ectoderm, and the wall of the pharynx, which 

 is perforated by the gill-slits, would on its inner side be covered with 

 entoderm and on its outer with ectodermal epithelium. We should 

 then perhaps be justified in homologising the peribranchial cavity of 

 the Ascidians with the atrium of Amphioxus; we can hardly, in any 

 case, doubt the homology of the gill-slits in these two groups. 

 Another view has, however, been adopted by van Beneden and 

 •Iulin (Xos. 9 and 10). According to these observei\s, the first gill- 

 slit arises through the fusion of a rather long entodermal diverticulum 

 with the cloacal vesicle of the same side which, according to these 

 authors, is never very large. The Ascidian larva at this stage is 

 exactly in the condition in which Appendicular i a remains through- 

 out life, the pharynx, in the latter, communicating through a 

 branchial passage on either side with the exterior. These passages 

 represent a pair of gill-slits, and this pair, in the Ascidian as in 

 Appendicularia, remains the only pair. \n the Ascidian, the branchial 

 passages are considerably enlarged secondarily, and in this way the 

 peribranchial or atrial cavity arises. Since these passages, according 



