260 



ventral surface of the ethmoid was here exposed, in his specimen, an 

 error of observation, I think; the thin rostral process of the chon- 

 drocranium, that underlies the ethmoid, having probably either been 

 cut away, or shrunk away, in preparation. The mesial edges of the 

 palatine processes of opposite sides, posterior to this lozenge-shaped 

 space, diverged somewhat in all my specimens, differing in this from 

 Traquair's specimen; but in the one specimen that I have as yet 

 had carefully dissected, there was a small and irregular plate of bone 

 lying between the palatine processes, as shown in the figure, and 

 giving, when added to them, much the appearance shown by Traquair. 

 In my other, earlier preparations, where maceration, or slight and 

 successive boiling, was employed, this little plate was not found, 

 being, perhaps, overlooked and lost. The palatine processes them- 

 selves did not, apparently, in any of my preparations excepting the 

 one shown in the figure, reach the anterior end of the parasphenoid. 

 In the particular specimen used for the figures the parasphenoid over- 

 lapped slightly, ventrally, the palatine process of the premaxillary, on 

 one side, and on the other a connection between the two bones was 

 established by the little unpaired bone, the hind edge of which projected 

 backward, slightly, dorsal to the parasphenoid. Whether this little bone 

 belongs to the parasphenoid, or to the palatine process of the pre- 

 maxillary, or perhaps represents the teleostean vomer bone of the fish, 

 I can not decide, though its position, dorsal to the parasphenoid, is 

 agains tthe supposition of its being a vomer. As Pollard's subrostral 

 bone (No. 32) is otherwise represented, as will be later shown, the little 

 unpaired plate can have no relation to it unless it represents a se- 

 parate, underlying plate, of different origin, with which the tooth- 

 bearing part of the subrostral may later fuse. 



The premaxillary is traversed, through the anterior two thirds, 

 approximately, of its total length, by the main infraorbital lateral 

 canal. This canal, which I have already described in full in an 

 earlier work (No. 4), enters the bone on the mesial surface of its 

 anterior end, and runs at first laterally, downward, and forward to 

 the point where tube no. 3 of the line is given off. It then turns 

 laterally and backward to tube no. 4, and then backward, or back- 

 ward and slightly upward, to tube no. 5, which latter tube lies 

 between the premaxillary and lachrymal bones. Tube no. 3 opened 

 on the external surface of the bone, on both sides of my 30-cm 

 specimen, by an opening that lay directly postero-dorsal to the root 

 of the fourth tooth of the bone, counting from the median line of the 

 head ; tube no. 4 opening by a pore that lay directly posterior to» 



