THE "TRITON" TXJNICATA. 107 



and 11 on Plate XVIII., but may be further back as represented by Keferstein 

 and Ehlers in their pi. ix. fig. 1. It may advance forward, so as to touch 

 the 3rd muscle band (see PI. XIX. fig. 1), but is never found outside the 3rd 

 intermuscular space. 



The ganglion is very opaque, and it is difficult to make out its constitution. 

 Fig. 8 on Plate XTX. shows its anterior end with four nerves, two large and 

 two small, arising from it. Grobben * has apparently not noticed the smaller 

 pair (PI. XIX. fig. 8, n'), but he describes a median anterior nerve which I could 

 not find in any of my specimens, unless it be the nerve shown at n in Plate 



XVIII. fig. 12, which is drawn from an individual having apparently only 

 three anterior nerves. As in other Tunicates, where the matter has been 

 investigated, the nerve cells are all in the outer layers of the ganglion, and the 

 centre is formed of a mass of delicate interlacing fibres and granular matter. 

 Fig. 12, Plate XIX., shows this arrangement well. The nerve cells are 

 ovate, unipolar, bipolar, or multipolar, rarely the latter. They are finely 

 granular, and have distinct nuclei and nucleoli (see PI. XIX. fig. 13). 



On the ventral surface of the ganglion there lies a dark mass which must be 

 the neural gland, but of which I was unable to make out the structure definitely. 

 It gives rise anteriorly to a very delicate duct which runs directly forwards to 

 open at the prebranchial zone into the funnel-shaped depression mentioned 

 above (see PI. XVIII. figs. 8 and 12). This duct is wide where it emerges from 

 below the ganglion, and its wall is formed of distinct polygonal cells (see PL 



XIX. fig. 8 n.d). It rapidly narrows, however, as it runs forwards, and the 

 cell elements lose their distinctness, so that in the part immediately in front 

 of the 3rd muscle band (PL XVIII. fig. 8, n.d) it is very difficult to make 

 out any structure in the wall. In front of this point it again becomes more 

 distinct, and the cells vary from fusiform to squamous in their character (PL 

 XIX. fig. 9, n.d) up to the point where the duct joins the funnel-shaped de- 

 pression. 



The length of this neural duct varies with the positions of the ganglion and 

 of the aperture in the prebranchial zone. The normal arrangement is shown in 

 figs. 8 and 11, Plate XVIII., while in fig. 1, Plate XIX., it is abnormally 

 short, on account of the unusual position of the ganglion. The aperture 

 in the prebranchial zone is always placed in the median dorsal line upon the 

 most anterior point of the spirals formed by the peripharyngeal band, and 

 therefore in the 2nd intermuscular space. Grobben and also Keferstein and 

 Ehlers figure it in the 1st intermuscular space, an arrangement which I have 

 never seen. Although the peripharyngeal bands encroach upon the 1st inter- 

 muscular space at the two sides (see PL XVIII. fig. 11), they always, in the 

 specimens which I have examined, dip posteriorly at the ventral and dorsal 



* Loe. tit., p. 9. 



