173 



faces on the one side, and from their internal surfaces on the oppo- 

 site side of the head-folds. The primary optic vesicles (op) are larger 

 than the accessory ones (A. op 1 , A. op 2 ) but in other respects, they 

 are similar. The former, like the latter, are rounded in outline and 

 concave when viewed from within and, externally, they all give rise 

 to knob-like elevations where they come in contact with the outer 

 layer of epiblast. The embryo is about 3'/ 2 mm long and shows, 

 externally, 19 pairs of somites. 



Any one who is to confirm these observations, be it in Squalus, 

 must have abundance of material ranging from 3 to 3 x / 2 mm in length 

 and showing from 3 to 20 somites. 



The accessory vesicles just described are so similar in various 

 particulars with the primary optic vesicle that I feel justified in saying 

 they are serially homologous with the latter: they are formed in the 

 same way, as evaginations of the elementary brain wall, and, in 

 sections, they exhibit the same histological structure. Their sub- 

 sequent history, to be recorded later, favors the view that the ac- 

 cessory vesicles are eye-like in character. If this view be true 

 we have here a multiple-eyed embryonic form and it takes us one 

 step towards the ancestral condition. Whe should naturally expect 

 to find similar structures in other animals. It might, of course, turn 

 out to be a case of isolated survival of a primitive characteristic in 

 Squalus, without being generally preserved in Vertebrates, just as 

 nephridia are preserved in Peripatus, and are lacking in other 

 Tracheata, but it is more likely that the structures exist and from 

 their transitory nature have not hitherto been observed. I have 

 noticed similar appearing structures in Amblystoma and other Amphibia 

 but I am not ready to say, without following their complete history, 

 that they are identical with the structures I have described in Squalus. 



The presence of accessory eye-cups on the cephalic plate is 

 very remarkable but their subsequent history is even more worthy 

 of note. For sometime after discovering them I supposed that they 

 were transitory, and gave way completely to other cephalic structures 

 but, returning to the study with a larger collection of material, I have 

 been able to follow the anterior pair, step by step, through a graded 

 series of embryos, without once having lost trace of them, and to see 

 that] they enter the thalamencephalon and give rise to the pineal 

 outgrowth. The posterior pair, which are smaller, are not to be 

 followed in this definite way, they become fainter and, I believe, they 

 fade away. 



When the medullary groove is first closed, the anterior pair of 



