Reviews — Fossil Fishes in British Museum. 125 



With the dismissal of the Chondrostei the special interest of the 

 present volume commences, as in the treatment of the typical Meso- 

 zoic Ganoids — be it remembered that " Ganoid " is now only a con- 

 ventional term — a considerable amount of " redistribution of seats " 

 takes place with regard to forms with which the general palaeonto- 

 logist is already familiar. As co-ordinate with Chondrostei, the 

 author here deals with two further suborders of the Actinopterygii, 

 viz. the Protospondyli and the iEtheospondyli, and part of a third, 

 the Isospondyli. 



Protospondyli. 



The Protospondyli, abbreviate heterocercal fishes, with complex 

 mandible, no interclavicular plates, and having the vertebral axis 

 represented by a persistent notochord or by vertebrae which, at least 

 in the caudal region, have the pleurocentra and hypocentra distinct 

 and alternating, are, in fact, nearly the equivalent of what since 

 Huxley's time have been called Lepidosteidae or Lepidosteoidea, but 

 with Lepidosteus left out and Amia included. For the disappearance 

 of the " Amioidea " as a suborder I have long been prepared by the 

 obvious similarity in cranial characters between Amia and such 

 genera as Lepidotus and Eugnatlms. And yet these have osseous, 

 rhombic, brilliantly ganoid scales, while those of Amia are rounded, 

 thin, and flexible — are, in fact, " cycloid " in the most Agassiziau 

 sense of the word. The condition of the squamation as an ordinal or 

 subordinal, or even family characteristic, here receives its crowning 

 blow, after having been utterly shaken by the occurrence of Palaeon- 

 iscidae with round, imbricating scales {Trissolepis, Fritsch, Crijphio- 

 lepis, Traquair) ; and a still further interesting case is afforded by 

 Mr. Woodward's Semionotid genus JEthiolepis, in which the body- 

 scales are rhombic, while those towards the tail are " cycloidal." 



The families included by Mr. Woodward under Protospondyli 

 are the Semionotidae, Macrosemiidae, Pycnodontidae, Eugnathidae, 

 Amiadae, and Pachycormidae. In the first of these families we have 

 deep-bodied " Stylodonts " like Dapedius included along with fusi- 

 form " Sphaerodouts " like Lepidotus, for the mere shape of teeth is 

 as of little classificatory importance as that of scales. Then as 

 regards the Pycnodontidae, Mr. Woodward has made a special and 

 elaborate study of this remarkable family, bringing out many new 

 and important structural facts, such as the fusion of the pterygo- 

 palatine arcade with the base of the skull, etc. As regards tiie 

 systematic position to which he has assigned these fishes, 1 have 

 every reason personally to be gratified with his results. It is nuw 

 many years ago since I imagined that I had successfully combated 

 the idea of Egerton that the Platysomidae were structurally related 

 to the Pycnodontidae, maintaining as I did that while the former were 

 of Acipenseroid affinity, and indeed close to the Palaeoniscidae, the 

 structure of the internal skeleton in the latter (Pycnodontidae) 

 pointed rather to their place being in "the great Lepidosteoid series 

 of Ganoids " than in any other. Nevertheless, in his most recent 

 systematic work Prof, von Zittel still maintains that the Pycno- 

 dontidae are probably the descendants of the Platysomidae, and asserts 



