— 205 — 



in Belonc, tlicse two tcleosts being the only ones in which it is known to occiir. In Belone the bone 

 is indistinguishably fused vvith the premaxillary, while in Elops it is found as an independent ossicle. 



4. Themaxillaryof the mail-cheeked fishes has, on its anterior end, a process thatmaybe called 

 its ascending process. This process gives articulation to the articular process of the premaxillary, 

 and itself articulates, by the intermediation of a päd of semi-cartilaginous tissue, with the dorsal 

 surface of the ascending process of the vomer. This process of the maxillary is apparently found 

 in all the Acanthopterygii and Anacanthini of Günther's Classification, but it is not evident in all 

 descriptions of the Physostomi. In those fishes in which it is found it varies greatly in its developmont, 

 and may even be found as two instead of as a single process. 



5. The vomer of fishes is primarily a bone, doubtless paired, that is developed in relation 

 to tooth-bearing plates on the dorsal surface of the mouth cavity; and it was primarily limited to 

 the roof of that cavity. In certain teleosts, however, this tooth-bearing plate has accjuired a dorsal 

 limb which may be said to consist of a head and two ascending processes, one on either side; and 

 these ascending processes are quite certainly formed by the fusion, with the tooth-bearing plate, 

 of the pre-ethmoid (septomaxillary, Amia) bone of either side. When the pre-ethmoids are found 

 as independent ossifications the vomer is without ascending processes. The ascending process of 

 either side gives articulation to the ascending process of the corresponding maxillary. 



6. The septomaxillary of the Amphibia and the higher vertebrates is probablv represented 

 in fishes by the antorbital bone of Amia, that bone being developed in protective relation to the 

 infranasal portion of the latero-sensory canals. This antorbital, latero-sensory bone is found in 

 Polypterus and Elops, as well as in Amia, and is possibly also found in certain of the Siluridae (Pollard). 

 In Polypterus it fuses with the premaxillary to form an infranasal process of that bone. 



7. In Macrodon there is a bone, called by Sagemehl the accessory palatine, that is apparently 

 developed in the maxillary breathing valve of the fish, and that is accordingly the homologue of the 

 so-called vomer of Polypterus. It has never been recognised in any other teleost. 



8. The palato-quadrate articulations with the ethmoid region of the skull differ considerably 

 in different ones of the mail-cheeked fishes examined. In Scorpaena and Sebastes there are two of 

 these articulations, one with an anterior palatine process of the ethmoid cartilage and the other with 

 the ectethmoid; the lachrymal also articulating with the latter bone. In Cottus the articulation 

 with the ectethmoid is suppressed, the palatine there being bound to the lachrymal and that bone 

 alone articulating with the ectethmoid. In Trigla, this posterior articulation, with the ectethmoid, 

 is much as in Cottus, while the anterior articulation, with the ethmoid cartilage, has been largely 

 replaced by a swinging articulation (or attachment) of the lachrymal and palatine with the antero- 

 lateral corner of the nasal bone. In Dactylopterus this latter articulation is still more pronounced. 



9. The quadrate has, in most if not in all teleosts, a posterior process which forma the posterior 

 boundary of the symplectic groove on the internal surface of the bone and has supporting relations 

 with the preopercular. This process of the quadrate is not found, as a part of that bone, in the bony 

 ganoids, but it is elsewhere represented both in Amia and Lepidosteus. In Amia it has fused with 

 the symplectic to form a process of that bone that gives a supplementary articulation to the mandible; 

 while in Lepidosteus it is the independent, so-called preoperculum of Parker's descriptions (interoper- 

 culum, Collinge), which, in that fish, has an independent articulation with the quadrate. In the 

 Muraenidae, the process and the symplectic seem to both be indistinguishably fused with the body 

 of the quadrate. 



