M. M. METCALF ON THE EYES AND SUBNEUEAL GLAND OF SALPA. 361 



between the condition in Phallusia mammillata and that of other 

 ascidians, since it has no lateral ducts. These comparisons are all 

 hypothetical, with little direct evidence in their favor, yet they are well 

 worth considering. 



What is the significance of the lacunae, in the dorsal part of the ganglion 

 of salpaf They have too definite a structure to be regarded as mere 

 chance spaces between the proliferating cells. They seem to me to indi- 

 cate that the former function of the hypophysis was to aerate the brain 

 or to carry off the waste products of its metabolism, or perhaps both. 

 In many of the ascidians we find that at one point in the dorsal wall 

 of the duct of the gland the epithelial lining of the duct is wanting, and 

 the cells of the ganglion reach the lumen of the duct. This condition 

 found in .so many simple ascidians must have some meaning. It seems 

 to bear a relation to the lacunae in salpa's ganglion and to confirm my 

 interpretation of the latter. We have seen that the lacunae connect with 

 the lumen of the neural tube, the homologue of the duct of the ascidian 

 gland. 



The great variation in the subneural gland as to its presence or 

 absence in the solitary or chain forms of different species, and especially 

 the variation in the ventro-lateral outgrowths from the ganglion, would 

 indicate that these structures are now of little functional value. 



The fact that in the salpas we have present well developed eyes 

 which yet cannot be homologized with the larval ascidian eye is 

 suggestive in regard to a question which Professor Brooks discusses 

 in the main portion of this memoir, namely, the question of salpa's 

 descent from an attached ascidian-like ancestor, or from a free swim- 

 ming form like the ascidian larva. If, accepting the latter hypothesis, 

 there had been no break in the free swimming manner of life from the 

 ascidian tadpole to its descendant salpa, there would never have been a 

 time in its phylogenetic development when an eye would not have been 

 needed, and so we would naturally expect to find salpa in possession of 

 an eye directly traceable to the ascidian larval eye. This, however, is 

 not the fact. The eye of salpa is situated in a wholly different region 

 of the brain from that in which the larval ascidian eye is found. It 

 would seem, then, more natural to suppose that salpa is descended from 

 some attached form, which, having given up its free swimming life, 

 would have little use for an optic organ and would therefore have lost 

 the eye which it originally possessed. Salpa, on resuming a free swim- 

 ming habit, would need a visual organ ; which, in fact, we find. ' The 



