340 AETHTJR WILLEY. 



the primary branchial canals in their capacity as distinct 

 structures deserving a special name do not exist ; and, in the 

 second place, it is precisely the way in which the atrial cavi-. 

 ties arise in the bud that shows conclusively that they are 

 perfectly homogeneous structures, and do not originate from 

 two ''^bauches," hypoblastic and epiblastic respectively. 



The question then arises as to what the " culs-de-sac hypo- 

 blastiques" which were seen by van Beneden and Julin might 

 be, and it would appear to be not very difficult to answer. 



A transverse section through the pharyngeal region of a 

 larva of Clavelina lepadiformis is shown in fig. 29, from 

 which it is seen that the branchial sac has a peculiar and 

 characteristic structure at this stage, brought about by the 

 occurrence of certain definite foldings in its wall. It consists 

 of a median portion and two lateral portions. The median 

 ventral wall forms a deep fold with high cylindrical cells ; 

 this is the endostyle. The lateral portions consist also of two 

 large folds with a lower cubical epithelium. Between the 

 lateral folds and the ectoderm are seen the two peribranchial 

 cavities, which have begun to spread themselves out. Stigmata 

 are present at this stage, but this particular section does not 

 pass through any. A somewhat oblique section is represented 

 in fig. 28, so that the atrial cavity is only seen on one side 

 with two stigmata in connection with it, one of which, how- 

 ever, has apparently not yet broken through. 



The section also involves the opening of the neuro-hypo- 

 physial canal into the branchial sac ; but I will not enter into 

 the question of the nervous system here, reserving it for the 

 next " study." 



The stigmata arise by simple fusions occurring at regular 

 intervals between the wall of the branchial sac and the floor 

 of the atrial cavity on each side. The " culs-de-sac hypo- 

 blastiques" of van Beneden and Julin can be nothing else than 

 the two lateral divisions of the branchial sac which I have just 

 described. But it was largely on the strength of the supposed 

 part which these "culs-de-sac hypoblastiques" took in the 

 formation of the atrial chambers that van Beneden and Julin 



