The Latero-Sensory Canals and Related Bones in Fishes. 443 



fish Hyrtl Ì8 said [69, p. 371] to have found but foui- latero-sensory 

 pores, two on either side, one at the end of the snout, and the other 

 at the tail. In Gymnarchus the secondary closure of the tubes is 

 quite certainly, as will be shown, in some way related or correlated 

 to a functional adaptation of the latero-sensory canals to the sense 

 of hearing. 



The supraorbital canal of Gymnarchus begins at the antero-lateral 

 end of the crook -shaped nasal bone of Erdl's descriptions [35\ runs 

 at first antero-mesially and then mesially along the anterior edge of 

 the snout, then dorsally and dorso-laterally across the anterior edge 

 of the snout, and finally dorso-posteriorly along the dorsal surface of 

 the snout; following, in this course, the curved line of the crook of 

 the nasal bone and then the straight line of the shank of that bone. 

 In this part of its course it lay, in the two specimens examined, in 

 an open groove on the outer surface of the nasal bone, and not in a 

 closed bony canal. Leaving the nasal the canal lies, for a short 

 distance, in the dermis, along the latero-dorsal edge of a bone called 

 by Erdl the "Pars nasalis ossis frontis", which bone is apparently 

 the homologue of the median, primary ethmoid of Sagem ehl's de- 

 scriptions of other fishes. In its course along the dorso-lateral edge 

 of this bone, the supraorbital canal lies dorso-mesial to the nasal sac, 

 the nasal bone, positively identified by its relations to the supra- 

 orbital canal, apparently having no direct relations whatever to the 

 nasal sac, unless it be by its hind end; the anterior part of the bone 

 lying wholly anterior to the sac. The prefrontal part of the canal, 

 as a whole, lies, however, mesial or antero-mesial to the two nasal 

 apertures, as it does in all teleosts. Having passed the hind end of 

 the nasal sac the supraorbital canal enters a groove on the dorso- 

 lateral edge of the frontal and here soon, as it crosses the dorsal 

 surface of the antorbital process, becomes wholly enclosed in the 

 frontal. Posterior to the antorbital process the canal again lies for 

 a certain distance in an open groove near the lateral edge of the 

 frontal, and then again entering that bone, and thei'e turning latero- 

 posteriorly, traverses what Erdl calls the outer process of the bone; 

 and, at the outer end of that process, anastomoses with the main 



