57] THE SKULL OF AM lURUS— KINDRED 57 



the ethmoid region is comparable, to a great degree, with the condition of 

 Amiunis. The intemasal septum, except in some of the very lowest genera of 

 the family, has the same relations as in Amiurus. The vomer is usually more 

 massive and extends ventrally in between the ethmoid (Sagemehl) and the 

 cartilage. According to Sagemehl, the ethmoid has a double origin as in 

 Amiurus, the amount of ossification varying from the Amioid condition to 

 complete ossification of the intemasal cartilage. In speaking of the ventral 

 extent of the ethmoid in this group, he hoped that further investigations would 

 justify his conclusions that such a condition did exist in other teleosts. 



In the Cyprinoids (Sagemehl, 1891), there is m,ore or less invasion of the 

 cartilage of the intemasal septum by bone, but none of the species have pro- 

 ceeded as far as Amiums in this respect. In this family the ossification around 

 the ethmoid cartilage has extended on to the ventral surface and is here in 

 contact with the vomer. Concerning this region Sagemehl says: "In Folge 

 dieses Verbal tens konnen wir bei Cyprinoiden an jedem Ethmoid zwei Theile 

 unterschieden: eine diinne Knochenplatte, die annahemd dem urspriingUchen 

 Deckknochen entspricht und die lateral die Nasengrube uberdacht, und eine 

 von dieser Platte nach unten absteigende mehr oder weniger breite, aus 

 spongioser Knochensubstanz bestehende, vertikal gestellte knocheme Wand, 

 welche die beiden Nasengmben von einander scheidet, und die durch Knor- 

 pelsubstitution entstanden ist. " 



This condition is also true of Amiurus as sho\Mi in an earlier part of the 

 discussion. Hence the degree of ossification in the ethmoid region of the 

 Characinidae, Cj'prinidae and Siluridae affords another factor for grouping 

 them together. 



Of the Loricati (AUis, 1910), Trigla has a part of the intemasal septum 

 invaded by bone, but in none does the dorsal ossification, either dermal or 

 perichondrial, extend over the anterior end of the cranium. In Salmo and 

 most of the other teleosts, the ossification is entirely dermal and has no peri- 

 chondrial element connected with it, repeating the condition found in Amia. 

 From the above comparisons I think that, of all the lower teleosts, Amiurus 

 has the largest amount of ossification in this region and the steps in the forma- 

 tion of it show that it is both dermal and perichondrial in origin. The former 

 is the first of the two ossifications to appear. 



The edethmoids. Each of these is developed around the ectethmoid process 

 of either side. In the discussion of this region of the younger forms, the changes 

 which have taken place have been described and the beginning of ossification 

 in the 32 mm. larva has been noted. Perichondrial ossification had just begun 

 at that stage, while the large membranous sheet attached to the lateral edge 

 of the cartilaginous ectethmoid process had already ossified and extended 

 posteriorly above the orbit. In the 60 mm. larva, perichondrial ossifications 

 have appeared on both the dorsal and ventral surfaces of the cartilage which 



