14 Mr. A. S. Woodward on some 



the powerful development of the dentition. Thus the Pycno- 

 donts may be closely related to other ganoid fishes with 

 normal skulls, and their nearest allies may exhibit a quite 

 ordinary opercular and branchiostegal apparatus if the mouth 

 be less displaced or the dentition less powerful. 



Having premised so much, it only remains to emphasize 

 again the contention of Traquair * that the Pycnodont skele- 

 ton, considered apart from the head, is completely " Lepid- 

 osteoid" in character. This emphasis is all the more necessary 

 since, even in a modern handbook, which is authoritative in 

 most respects, the exploded errors of thirty years ago are 

 allowed undue influence, and the "special" difference between 

 Pycnodonts and the Palaeozoic Platysomidas is said to consist 

 in " the dentition and arrangement of the jaws " f- As a 

 matter of fact, all the fundamental differences are in the trunk. 

 In Pycnodonts the tail is atrophied-heterocercal, the rays of 

 the dorsal and anal fins are supported by a single series of 

 endoskeletal elements of equal number, and there are no 

 infraclavicles. Even the most specialized of known Platy- 

 somidai never exhibit the faintest approach to this combination 

 of characters ; the heterocercal tail and multiple-rayed median 

 fins are as pronounced in the most specialized as in the most 

 generalized types. The summit of Platysomid evolution 

 seems to have been attained by the anomalous Permian genus 

 Dorypterus. The only known link between the grade of fish 

 to which Platysomus belongs and the higher grade in which 

 the Pycnodonts must be placed is the Triassic family of 

 Catopteridge, where the upper caudal lobe is hemiheterocercal, 

 while the rays of the dorsal and anal fins are nearly as few as 

 their supports, which are ranged in only one series. Present 

 knowledge is therefore in favour of the supposition that the 

 Pycnodonts are a secondary modification of the higher or 

 " Lepidosteoid " grade, which was first reached through 

 comparatively generalized types of fishes resembling the 

 Catopteridte. 



In fact, making allowance for the morphological changes 

 necessitated by a reduced branchial outlet and a forwardly 

 displaced mouth — further, admitting that the anomalous 

 arrangement of the cranial roof-bones is the result of degene- 

 ration — there is little to distinguish the Pycnodonts from 

 Lepidotus, Dapedius, and their allies. It is certainly note- 

 worthy that Lepidotus itself, which has the most highly 

 developed tritoral dentition, exhibits very few branchiostegal 



* R. H. Traquair, "On the Structure and Affinities of the Platy- 

 somidge," Trans. Roy. Soc. Edinh. vol. xxix. (1879), pp. 382-386. 

 t K. A. von Zittel, ' Grundziige der Palaeontologie ' (1895), p. 574. 



