vol. 2, pt. 2.1 A TAXONOMIC STUDY OF THE SALPIDAE METCALF. 49 



In the ganglion, accessory smaller eyes are present (fig. 19, p. 48). 

 These groups of imperfect rod-cells have no pigment associated with 

 them. The groups vary in number and position, but have about the 

 same structure, which is shown in figure 21 . They are to be compared 

 with the better developed accessory eyes of the other Cyclosalpas. 



Just behind the optic nerve, at its origin, the ganglion is swollen to 

 form a considerable protuberance whose cells, however, are not 

 developed into rod-cells as they are in the corresponding part of the 

 ganglia of some other species, as Apsteinia punctata, Apsteinia 

 asymmetrica, Salpa fusiformis, S. cylindrica, and, as I interpret the 

 relations, in the Thalias. 



The musculature of Cyclosalpa virgula has been very carefully 

 studied by Streiff (1908). For the solitary form he shows slightly 

 different arrangement of the oral retractor muscles, the dorsal re- 

 tractor being double and the ventral retractor uniting with the ventral 

 division of the dorsal retractor. The ventral horizontal 

 bands, connecting body muscles I and V, he shows as 

 continuous. We find them interrupted. The dorsal 

 longitudinal muscle band back of the upper lip he shows 

 in contact with the last sphincter muscle of the dorsal FlQ . 2 \.— cyclo- 

 lip. In our specimens these muscles are not in contact, salpa virgula, 



T-T 1 1 11 1 I'm • 1 AGGREGATED 



Doubtless all these differences in our descriptions are form.asection 

 due to differences in the specimens studied and indicate ™ B °J1 G ? ^f 



A OF THE ACCESS* 



individual divergence within the species. ory eyes in 



In the aggregated form, Streiff does not show the fromhetcalf 

 posterior branches of the last body muscle (our muscle AND Joh nson 

 IV, b) to be united above the post-abdomen (doubtless 

 again divergence between his specimens and ours), nor does he 

 describe any visceral muscle arising from the last body muscles and 

 running into the base of the post-abdomen. We suspect that here 

 Streiff failed to see a structure which was present, as he also fails to 

 note the corresponding muscle in the aggregated Cyclosalpa pinnata. 



THE CYCLOSALPAS IN GENERAL. 



In our introductory remarks, and in the course of the descriptions 

 of the several species, we have made numerous comparisons between 

 the species of Cyclosalpa. The division of the subgenus into Cyclo- 

 salpae symmetricales and Cyclosalpae asymmetricales, while truly 

 descriptive of the existing conditions, does not represent accurately 

 degrees of relationship, for Cyclosalpa bakeri, an asymmetrical form, 

 finds its nearest relative in C. floridana, an aberrant member of the 

 group symmetricales. (Compare the chart on p. 158.) Wo have 

 already noted that in the family Salpidae only the aggregated zooids 



