540 MAYNARD M. METO ALF, 



thought it does not disprove, any innervation of the chambers by 

 these ganglia. 



The cells of the large-celled ventro-lateral ganglia show para- 

 nuclear bodies which resemble the so-called centrosomes of the ganglion 

 cells of other Tunicata. 



These ventro-lateral ganglia in Sal^pa recall the masses of large 

 ganglion cells, which, in the Ascidians, are found between the brain 

 and the neural gland, and I believe they are probably homologous 

 structures. The function of these cell masses is doubtful, but that 

 they are in some way associated with the neural gland is rendered 

 probable by their constant position in proximity to the gland. 



In the solitary form of Cydosalpa pinnata we have the same 

 condition of the neural gland and the ventro-lateral ganglia. 



The neural gland in Salpa, in both solitary and chain forms, 

 develops from the pharyngeal (or cloacal?) epithelium (cf. Figs. 72 — 74). 

 In a young bud or embryo the pharyngeal (or cloacal ?) epithelium on 

 each side of the dorsal lamina lies close pressed to the ventral sur- 

 face of the ganglion (Fig. 72, a cross section). In an individual a 

 little older, one finds that most of this epithelium has separated itself 

 from the surface of the ganglion, but one small circular area on each 

 side remains in contact with the ganglion (Fig. 73, a parasagittal section). 

 These become the ventro-lateral chambers of the gland. As the wall 

 of the branchial chamber separates more and more from the surface 

 of the ganglion, those portions of this wall which are adjacent to the 

 two areas of adhesion are gradually drawn out into two tubes con- 

 necting the ventro-lateral chambers with the branchial chamber 

 (Fig. 74). These are the ducts of the gland. The adult condition is 

 reached by the considerable growth of these tubes, which become 

 greatly coiled, and by the enlargement of the chambers. The figures 

 referred to in this paragraph are drawn from developing buds, but 

 they represent equally well the origin of the gland in the embryo. 



No rapheal duct is present in the Cyclosalpas. 



A very small rapheal nerve wholly devoid of ganglion cells can 

 be traced into the raphe, running down behind the great rapheal 

 blood sinus. 



There is no rapheal muscle in Cydosalpa. 



The relation of these organs to the intersiphonal organs of As- 

 cidians, Pyrosoma and Doliolum will be discussed later, after a review 

 of the conditions in other species of Salpa. 



