M. M. METCALF ON THE EYES AND SUBNEURAL GLAND OF SALPA. 341 



the anterior part of the neural canal (Diagram 2, A). In the larval 

 ascidian we find a number of cells in a corresponding position, whose 

 inner ends abut on a mass of pigment (Diagram 2, B). These cells are 

 slightly evaginated, and in the hollow of the evagination there are one 

 or more "lenses," 1 each of which has been formed by the migration 

 of one of the neural cells into the hollow of the optic evagination. In 

 the vertebrate pineal eye (Diagram 2, D) we have this evagination 

 carried much further and the lens formed in a different way, but the 

 rod cells correspond to those of the larval ascidian eye in their relation 

 to the neural canal, their pigmented ends being toward the cavity of the 

 neural tube, or rather of the outgrowth from the neural tube. Salpa 

 eye (Diagram 2, C) forms a connecting link in structure between the 

 larval ascidian eye and the vertebrate pineal eye, and though it is 

 formed in a different way from a different part of the brain and forms 

 no part of the phylogenetic series, still it indicates a probable phase in 

 the structural development of both the pineal and lateral eye of verte- 

 brates. 



DIAGRAM 2. 



B 



A optic unit of amphioxus eye (?). 

 B =. optic unit of clavelina eye. 



C = optic unit of salpa eye. 



D = optic unit of vertebrate lateral or pineal eye. 



In the optic unit of certain species of salpa we find a cell interposed 

 between the rod cell and the pigment cells. I refer to the intermediate 



1 1 have often found several such lens-like bodies in the eye of clavelina sur- 

 rounded by the mass of pigment granules. 



