244 THE STRUCTURE AND AFFINITIES OF HERDMANIA CLAVIFORMIS. 



bottom upward on the surface of the gland, as it does in Fragaroides and other forms 

 (Maurice, '88, PI. 16a, Fig. 29). 



A section showing the extent of the ramification of the duct in the gland of Herd- 

 mania is shown in Figure 14. Not more than three or four branches in any one gland 

 have been seen. These never penetrate beyond the centre of the gland, and usually 

 not even as deep as that. 



The lumina of these branches are obscure, but the branches themselves are dis- 

 tinct enough from the contrast of their compact, well-stained cells with the vesicular 

 cells of the gland proper in which they are embedded. The branching of the duct 

 within the gland, interesting as it is, rarely if ever occurs in compound ascidians, 

 though this condition is common among the simple ones. The gland-cells proper 

 are of the vesicular type commonly found in this organ. The cytoplasm is reduced 

 to a narrow zone at the surface of the cell, and in this, at one side, the deeply stained, 

 crescent-shaped nucleus is a conspicuous object. The interior of the cell is clear in 

 a majority of cases and seems to be empty; but a slightly granular mass, very little 

 affected by hsematoxylin, more or less completely fills the cavity of many of the other 

 cells. The substance in the cells is probably, as is usually held, the secreted material 

 characteristic of the gland. It is, however, an interesting fact that cells in the blood 

 are common in many parts of the body and closely resemble those here described. The 

 crescent-shaped nuclei on the surfaces of these cells have their counterparts in the 

 gland-cells. 



The bodies within the gland-cells described by Metcalf for Polycyclus and Fraga- 

 roides and called by him paranuclei I have not seen in Herdmania. 



7. The Central Nervous System. The ganglion requires no special descrip- 

 tion.. It is nearly spherical, is of large size, and is situated on the dorsal side of the 

 neural gland and slightly in advance of it. 



If a dorsal or ganglionated nerve exists in the adult of this species, its connec- 

 tion with the rapheal duct is so intimate as to render it impossible to distinguish the 

 two. 



There can hardly be a doubt, I suppose, that what I have described as the rapheal 

 duct, Figures 15 and 16, is the same structure as that described and figured by Maurice 

 ('88) in Fragaroides, and by Seeliger ('93-: 02) in Ciona and Clavelina, though both these 

 authors regard the structure as nervous. But from the observations of several zoolo- 

 gists, particularly of Metcalf, the association of the two is so intimate as to make it 

 actually impossible to distinguish them. In Herdmania my observations make it 

 seem to me that the strand should be regarded as belonging to the gland rather than 

 to the ganglion. In my description of the ventral processes on the rapheal duct 



