MUSTELUS L^VIS. 97 



bend. It first turns laterally and slightly forward, and 

 connects with the distal end of the supra-orbital canal, 

 this bend being short and being apparent in sections, but 

 not in dissections. It then turns sharply backward, in the 

 line of, and apparently as a direct continuation of, the 

 supra-orbital canal, and continues a short distance almost 

 directly backward, parallel to the lateral edge of this part 

 of the head. It then turns laterally and forward in a short 

 rounded angle, the hyomandibular canal arising at this bend 

 almost as a direct continuation backward of the infra-orbital 

 canal anterior to the bend. Running laterally and forward 

 a short distance the infra-orbital canal reaches the lateral 

 edge of the head, where it turns upward and forward on 

 to the lateral surface of the head, and, continuing in this 

 same direction, soon reaches a point approximately ventral 

 to the anterior edge of the eye. There it curves gradually 

 backward in a short bend, and then runs backward below 

 the eye and upward behind it, between it and the spiracular 

 opening, thus encircling about one half the orbit. Dorsal 

 to the spiracular opening the canal turns upward and 

 forward, and, approximately dorsal to the hind edge of the 

 eye, anastomoses with the hind end of the supra-orbital canal. 

 It then turns sharply backward, upward, and mcsially, and 

 so continues to the point where it joins the lateral end of 

 the supratemporal commissure. There the canal turns 

 almost directly backward, and continues backward as the 

 lateral line of the body. 



Along this infra-orbital section of the main lateral canal, 

 that is, from its anterior end to the point where the supra- 

 temporal cross-commissui"e is given oif, there were, in my 

 embryo, 180 tubules of varying size and length, all leading 

 directly and independently from the canal, and opening on 

 the external surface by one or more surface pores. Along 

 this same length of canal there were 110 sense organs, all 

 innervated by branches of the buccalis and oticus facialis. 

 The tubules in certain parts of the canal lay regularly one 

 between each two successive sense organs, this being markedly 



