338 JOHNS HOPKINS UNIVEKSITY MORPHOLOGICAL MONOGRAPHS. 



What relationships between species does this evidence from the eyes favor ? 

 First, the fundamental uniformity of plan in the eyes of the different 

 species shows what is universally recognized, namely, a common ances- 

 try; second, we see certain instances of the closest agreement in the 

 structure between the eyes of two or more species, indicating a natural 

 group. Cyclosalpa pinnata and Cyclosalpa Chamissonis form one such 

 quite distinct group. The eye of the chain Cyclosalpa pinnata passes 

 through a Chamissonis stage; indicating that the latter species is the 

 more primitive of the two. Again, the almost perfect uniformity of 

 structure in the eyes of Salpa runcinata-fusiformis and Salpa Africana- 

 maxima indicates that these two form a natural group. But the fact 

 that the manner of innervation of the large dorsal eye of the chain form 

 of these species points to a reversal of position during the development 

 similar to that described for Cyclosalpa pinnata, indicates that these two 

 salpas fall into a larger natural group with the cyclosalpas. This group 

 includes also the other salpas in which a similar reversal of the position 

 of the eye has taken place : i. e., Salpa cylindrica, Salpa hexagona, Salpa 

 costata-Tillesii, Salpa cordiformis-zonaria and Salpa democratica-mucro- 

 nata. Quite distinct from this large group is another natural group 

 including Salpa scutigera-confederata and Salpa bicaudata, in which no 

 reversal of the eye has taken place. This is shown by the manner of 

 innervation. The large eye of the chain Cyclosalpa pinnata passes 

 through a stage when in shape, position and innervation it resembles 

 the adult eye of the chain Salpa scutigera-confederata ; so it is safe to 

 say, so far as the evidence from the study of the eyes goes, that the group 

 including the latter species is more primitive than the large group in 

 which we find Cyclosalpa pinnata and Salpa runcinata-fusiformis. The 

 imperfect condition of the rod cells of Salpa scutigera-confederata and 

 Salpa bicaudata indicates perhaps that these forms have somewhat 

 degenerated from the ancestral type, to which, however, they are more 

 closely related than any other species so far as regards the shape, 

 position and innervation of the eye. 



Between the members of the larger group we find varying degrees 

 of resemblance in the eye ; Salpa cylindrica stands about midway 

 in the group. Salpa democratica-mucronata shows the greatest diverg- 

 ence from this species. Salpa runcinata-fusiformis and Salpa Africana- 

 maxima agree quite closely with it. Salpa cordiformis-zonaria stands 

 quite far removed. The Cyclosalpas form a distinct group with a 

 more highly specialized eye than Salpa cylindrica, yet bearing a close 



