469 



i^:.^ 



IS. 



Fig. 3. Three cells from the lumen of the neural gland, disintegrating to form 

 the secretion of the gland. 



Engyra pilularis, Cynthia echinata, Styrla plicata, 

 Polycarpa glomerata, Polycarpa variaiis (the rapheal pro- 

 longation in these two species of Polycarpa is short and bifurcated), 

 Ciona intestinalis, Diazona violacea, Rhopalaea neapo- 

 litana, Bostrich obranchus manhattensis (not Molgula 

 raanhattensis of Verril, which is a distinct species). In 

 Bostrichobr an chus, as in Cynthia papulosa, there is much 

 glandular tissue in the raphe. In the other forms named merely the 

 duct of the neural gland is prolonged into the raphe, there being 

 little, if any, secretory activity. 



It is interesting to note that this rapheal portion of the neural 

 gland may be present in one species and be absent in another of the 

 same genus ; e. g. Styrla plicata has this region of the gland well 

 developed in the form of an elongated tube running far down into 

 the raphe, while the american variety of Styrla aggregata (which 

 Vereil has named Cynthia partita) has no rapheal portion of the 

 gland. 



This rapheal portion of the gland lies in the place formerly 

 occupied by the trunk region of the larval nerve tube. It would be 

 interesting to know whether it has arisen in situ by the metamor- 

 phosis of certain cells of the trunk region of the larval nerve tube, 

 much as the main portion of the gland arises from a more anterior 

 region of the larval nervous system ^). I have as yet been unable 

 to obtain favorable material for the study of this point. 



1) I can fully confirom Julin's description of the origin of the 

 gland from the wall of the larval nerve tube, immediately behind the 

 sensory vesicle, and opposite to the area from which the ganglion is 

 proliferated. Ectrinascidia turbinata and Molgula manhat- 

 tensis give diagrammatic demonstrations of this fact. 



