M. M. METCALF ON THE EYES AND SUBNEURAL GLAND OF SALPA. 329 



indicated structures which I regard as probably homologous, one with 

 the dorsal pair, the other with the posterior pair of smaller eyes of 

 Cyclosalpa pinnata, chain form. Fig. 10, Plate LIV, shows a vertical 

 section through the ganglion and eyes in the direction of the long axis of 

 the eye. In the large eye is seen an irregular mass of rod cells which 

 show no indication of the typical structure except in the thickening of 

 their walls. They are irregularly polyhedral in form and are not notice- 

 ably elongated. There is but slight indication of two regions of the eye 

 corresponding to the two regions in the larger eye of the chain form of 

 Salpa runcinata-fusiformis. This indication is given by the arrange- 

 ment of the pigment cells. There is a single continuous layer of these 

 between the mass of rod cells and the ectoderm of the optic sheath. 

 This layer is, however, thicker and more dense in two regions, one on the 

 ventral face of the eye near the ganglion, p', the other on the dorsal face 

 near the apex, p". The positions of these more developed portions of the 

 pigment layer correspond to the positions occupied by the two masses of 

 pigment cells in the two regions of the larger eye in the chain forms of 

 Salpa hexagona, Salpa runcinata-fusiformis and other species. In this 

 way we have a slight, but decisive, indication of the division of this eye 

 into two regions which, when typically developed, show the structure 

 found in the larger eye of the chain form of Salpa runcinata-fusiformis. 

 On the posterior face of the ganglion there is another large mass of 

 irregularly polyhedral cells with thickened cell walls, closely resembling 

 the rod cells of the dorsal eye (Fig. 10, Plate LIV). What may be the 

 meaning of these cells I am unable to say with certainty. The cells 

 much resemble the rod cells of the large "dorsal eye of the same species ; 

 the posterior pair of eyes in the chain Cyclosalpa pinnata, though much 

 smaller, occupy about the same position ; the pigment spots in the gan- 

 glion of pyrosoma are in a position corresponding to the ventral portion 

 of this mass of cells. I think it very probable that this structure is 

 homologically an eye, having the same relation to the posterior pair of 

 small eyes found in Cyclosalpa pinnata that the large dorsal eye of Salpa 

 costata-Tillesii bears to the unpaired eye of the former species. On the 

 opposite side of the ganglion, beneath the dorsal eye and in front of the 

 optic nerve, there is still another mass of similar, imperfect rod cells, ex, 

 and, in close connection with them, a patch of deeply pigmented cells. 

 Neither the rod cells nor the pigment cells show so well in the section 

 figured as in the adjacent sections. The structure of this mass of cells 

 and its position, comparable to the position of the dorsal pair of small 



