M. M. METCALF ON THE EYES AND SUBNEURAL GLAND OF SALPA. 347 



inward into the large canal which is the duct of the gland. The duct 

 and the small canals that run through the gland are lined by cubical 

 epithelium. At the level of the anterior end of the ganglion this passes 

 abruptly into the columnar ciliated epithelium of the funnel. At the 

 line of demarcation between the two kinds of epithelium the dorsal 

 wall of the duct is wanting, and the surface of the ganglion here bounds 

 the lumen of the duct. 



In the adult amaroecium the conditions are somewhat different. The 

 duct is absent, the gland and ganglion coming into close contact ; but the 

 funnel is present and its posterior portion pushes between the front por- 

 tions of the gland and ganglion. It opens freely to the former by a 

 narrow opening, but there are no definite canals running through the 

 gland and opening to the funnel. This is an indication of degeneration. 

 The gland itself is degenerate ; its walls having something the appear- 

 ance of the vacuolated notochord cells in the chick embryo. The chief 

 points of difference from the gland of clavelina shown by the gland of 

 amaroecium are : (1) the degenerate condition of the gland ; (2) the 

 absence of any duct opening into the funnel ; (3) the consequent absence 

 of any canals opening from the gland into such a duct ; (4) the fact that 

 the surface of the ganglion does not at any point lie next to the lumen of 

 the funnel, the single posterior opening from the funnel being toward 

 the gland. 



In the larval amarcecium the duct of the gland is represented by a 

 posterior prolongation of the funnel that completely separates the gland 

 and ganglion, reaching beyond their posterior limits. At one spot on the 

 ventral wall of the duct and at an opposite point on its dorsal wall the 

 epithelial lining is absent ; the surface of the gland and ganglion at these 

 points reaching the lumen of the duct. There are no canals running 

 through the gland. The presence in the larval amaroecium of a duct of 

 the subneural gland and of a communication between the ganglion and 

 the duct show that the absence of these features in the adult amaroecium 

 is due to degeneration. 



In Phallusia mammillata Julin discovered certain other structures 

 which throw great light upon certain organs in salpa I shall soon 

 describe. The condition of the ganglion and duct is practically the same 

 as in clavelina, but in addition to the small canals which open from the 

 gland into the duct there are many lateral canals connecting the duct with 

 tlie peribranchial chamber. Herdman (4) found an important variation 

 in two specimens of Phallusia mammillata which he examined. In 



