2O6 METCALF AND JOHNSON. 



The protuberance from the ganglion, which lies at the base of 

 the large eye, is very similar to that seen in C. pinnata (cf. Figs. 

 5 and 1 8). In the eye of C. dolichoscnna-i'irgula there is a group 

 of cells which strongly suggests comparison with the accessory 

 mass of optic cells in the curve of the horseshoe in the eye of the 

 chain C. pinnata and with the similar, though less distinct mass of 

 cells in C. chamissonis. This mass of peculiar spindle-shaped cells 

 does not show in surface view, but is shown as it appears in section 

 in PI. VIII., Figs. 5 and 10, q. They are not developed as rod-cells, 

 yet their shape, their position, their innervation, and their relation 

 to the pigment cells seem to indicate that they are to be compared 

 to the accessory portion of the eyes of the other Cvclosalpce. Yet, 

 in spite of the presence of this peculiar group of cells, the eye of 

 C. dolichosoma-virgula resembles those of the true Salpce more 

 than it does the eyes of the other Cvclosalpce. If the Cyclosalpa? 

 are more primitive than the Salp&, as many structural features 

 seem to indicate, the study of the eye of C. dolichosoma-virgula 

 would suggest a transitional stage from one group to the other. 



Another resemblance between Cyclosalpa dolichosoma-virgula 

 and the true Salpcs is seen in the fact that in the former species 

 and in Salpa runcinata-fusiformis there is found above the gang- 

 lion a peculiar ectodermal invagination which opens dorsally by 

 a relatively narrow mouth, and which probably serves as a pro- 

 tection for the ganglion and eye, no large optic chamber being 

 present in either of these species. 



The position of the large eye in the chain salpas is worth 

 noting. In most species the eye projects upwards (Salpa rnnci- 

 nata-fusifonnis, Fig. 20, PI. IX.) or forward from the dorsal face 

 of the ganglion (compare Fig. 5, PI. VIII., which shows a median 

 section of the ganglion and eye of Cyclosalpa dolicliosoina-virgjild}. 

 In Salpa pnnctata the eye projects slightly downward from the 

 antero-dorsal surface of the ganglion (Fig. 12, PI. IX.). In 

 TJiaha dciiiocratica-nnicronata the eye projects downward and 

 backward from the antero-ventral surface of the ganglion (Fig. 

 17, PI. IX.). These conditions, and the manner in which in 

 Thalia the ectoderm is drawn down in front of the ganglion, indi- 

 cate that the whole ganglion with the eye has rotated forward 

 and downward. Compare Figs. 5, 12 and 17, and note that in 



