196 TRANSACTIONS OF THE AMERICAN PHILOSOPHICAL SOCIETY 



distally and toothed (Tate Regan, 1929, p. 316). In the typical forms (Fig. 78) the palatine 

 has become rod-like and freed from the pterygoid but retains its attachment to the maxilla 

 (cf. Starks, 1926, p. 185). The metapterygoid, according to Regan (1911/ (Ostariophysi), 

 p. 563) has moved forward over the top of the quadrate and usurped the place of the 

 pterygoid, while the pterygoid itself has disappeared, except in the Bagridae, where it 

 forms a small plate behind the palatine. No trace of these primitive conditions is evident 

 in the chondrocranium of a 10 mm. Amiurus figured by Kindred (1919, PI. I). Here the 

 hyomandibular bears a deep sinus on its front border in the place where a metapterygoid 

 would be looked for, while the functional pterygoid bar is in its usual position connected 

 with the anterior inferior border of the quadrate. The palatine has already become long 

 and rod-like, the orbits are very small and far forward; in general this foetal chondrocranium 

 rather fully foreshadows the adult skull. The circumorbital series is often reduced to a 

 thread or absent, the preopercular and opercular apparatus variously reduced. The post- 

 temporal has been practically annexed to the skull and apparently the supracleithrum has 

 disappeared. In the midst of these and other specializations the orbitosphenoid element 

 is strongly developed, perhaps because it is needed to stiffen the immense skull and nuchal 

 spine. 



Still greater specializations appear in connection with the peculiar modifications of the 

 air-bladder and of the Weberian apparatus. Perhaps in order to support the massive skull 

 the first vertebra forms a disc "rigidly united to the basioccipital and to the second, third 

 and fourth vertebrae, which are ankylosed to form a complex to which the fifth is rigidly 

 attached and with the parapophyses ankylosed to the centra" (Tate Regan, 1929, p. 316). 



The branches of the transverse processes of the complex vertebrae have different rela- 

 tions with the parts of the air-bladder and of the "supracleithrum" (= supratemporal) 

 in the different families (Tate Regan, 1929). 



In passing we may allude to the well known fact that the under side of the skull (Fig. 

 79) of many catfishes bears a certain resemblance to a crucifix (Gudger, E. W., 1925). 

 The crown is formed by the opposite tripus of the Weberian apparatus. The arms of the 

 cross are formed by the transverse processes of the complex vertebrae. The arms of the 

 figure on the cross are the ascending processes of the parasphenoid. The "inscription" 

 across the base of the cross is furnished by the vomerine tooth-patch. No better example 

 perhaps could be found of a class of fortuitous resemblances between wholly unrelated 

 objects, which the late Professor Bashford Dean called "Unnatural History Resemblances" 

 (1908). Is it any wonder then that still closer resemblances are often produced between 

 form patterns that are related? 



The resemblances of the cephalic shield of the loricariid catfishes (Fig. 80) to those of 

 the cephalaspid ostracoderms are evidently to be referred to "convergence." In these 

 heavily armored forms the siluroid skull attains its highest specialization. In Plecostomus 

 (Fig. 80) the small, almost tadpole-like mouth bears several rows of minute, curved, rod- 

 like teeth in the premaxillae and dentaries. The width across the quadrate-articular joints 

 is about twice the length of the jaw itself. This is braced posteriorly by the transversely 

 widened pterygo-entopterygoid, which articulates firmly at two contacts with the side 

 wall of the ethmoid, below the large olfactory fossa. The palatine extends forward as a 

 long process overlapping the reduced maxilla. The very large plate-like supratemporal 

 overrides the reduced ascending limb of the cleithrum and acts as a cover for a bony tunnel 



