516 H. H. SWINNERTON. 



the etlimoid plate — the nasal septam of the salmon (Parker, 

 73, p. 128) — has appeared. At its upper end this mesethmoid 

 cartilage expands laterally and fuses with the upper ends 

 of the parethmoid cornua. 



The ethmoid region of the larval salmon (25 mm.) presents 

 one or two important differences from that of the stickleback. 

 Looked at from beneath it is broad but not rectangular, for 

 whilst both cornua (PL 31, fig. 47, e. 2^i'; e. p. c.) are recognis- 

 able, the anterior border is not straight, but has grown out into 

 a triangular rostrum (r,). At the same time on the dorsal sur- 

 face the nasal septum has extended to the anterior end of this. 



In the stickleback the epiphysial cartilage (PL 28, fig. 3, 

 ep. c.) has grown backwards, but the epiphysis is still closely 

 related to its anterior edge. Laterally it is continuous with the 

 hinder ends of the supra-orbital bands [sh. h.), and thus closes 

 in a space which may be called the anterior fontanelle {a.f.). 

 In thereof of a salmon 20 mm. loug (fig, 61) this cartilage is 

 expanded laterally, and, except for its continuity behind with 

 the post-orbital process [sh. p.),a,ndiov the fact that the prosen- 

 cephalon overlies the dorsally excavated mesethmoid cartilage, 

 the whole roof presents the same condition as that above 

 described for the stickleback. In an older salmon (25 mm.) 

 the anterior fontanelle has become covered in by an inward 

 growth of the epiphysial cartilage and supra-orbital bands, 

 except for a smaller space over the epiphysis {ep.f.) and a larger 

 one [a.f.) at the anterior extremity which corresponds to that 

 marked m. n. c. by Parker (73, pi. iv, fig. 2, and pi. vii, fig. 4) 

 and to the median fontanelle described by Winslow (98, p. 186) 

 in the trout. This fontanelle is now bounded in front by 

 the more fully developed mesethmoid cartilage, now no longer 

 overlapped by the prosencephalon. 



Through the kindness of Professor Howes I have been 

 enabled to examine the head of a larval Amia 19 mm. long. 

 The roof of this was in exactly the same condition as that 

 just described for the 25-mm. salmon, even to the presence 

 of the small foramen above the epiphysis, ft is safe, thei'efore, 

 to assume that here also the tegmen cranii is formed, not by 



