M. M. METCALF ON THE EYES AND SUBNEURAL GLAND OF SALPA. 359 



ally separated by no distinguishable boundary. I could not from my 

 own study of the adult salpa and the later stages of development ascer- 

 tain whether the lateral ducts opened to the pharynx or to the lateral, 

 anterior prolongations of the peribranchial chamber. Dr. Brooks, how- 

 ever, tells me that at this time (the time when the lateral ducts appear) 

 the anterior portion of the peribranchial chamber is limited to two very 

 narrow lateral areas. The ducts, on the other hand, are not far from the 

 median line. The whole formation of the peribranchial chamber in salpa 

 is so much modified that it is but a slight additional change to suppose 

 that the openings of the lateral ducts of the subneural gland shifted 

 their position a little more toward the median line and into the region of 

 the pharynx, and that later, their development being somewhat accel- 

 erated, they came to be formed before the anterior horns of the peri- 

 branchial chamber develop. The greatest degree of concentration of the 

 lateral ducts toward the mid-dorsal line is seen in Salpa costata-Tillesii, 

 in which the whole gland is but a widely open pit on the mid-dorsal line 

 of the pharynx. It seems probable then that we should regard the con- 

 nection of this gland in salpa with the branchial instead of the peri- 

 branchial chamber as a secondary modification consequent upon the 

 obliteration of the lateral walls separating the two regions. Assuming 

 for the present the truth of this hypothesis, and therefore the homology 

 of the subneural gland in salpa and Phallusia mammillata, some inter- 

 esting comparisons may be made. 



The fact that in all but one of the eleven species of salpa studied 

 this gland is paired is suggestive, but in view of the fact that in Phal- 

 lusia mammillata ' and in one species of salpa (Salpa costata-Tillesii) it 

 is unpaired, it is doubtful whether in the common ancestor of the two 

 the gland was a paired structure. The transition from a paired to an 

 unpaired structure, or vice versa, would be very easy. This is seen 

 when we consider the mode of its development in salpa. 



Phallusia mammillata has a subneural gland of twofold origin ; a 

 part coming from the central nervous system ; a part from the wall of 

 the peribranchial chamber. Most of the ascidians have a simple gland 

 formed solely from the nervous system. Salpa has a subneural gland 

 formed almost solely from the wall of the branchial, originally peri- 

 branchial (?) chamber. It is not probable that the earliest ascidians had 

 a gland of twofold origin. The question then confronts us, which sort of 

 subneural gland is the more primitive? The natural though not 



1 In Phallusia mainmillata the ducts, though not the gland itself, are paired. 



