1858.] HUXLEY — CEPHALASPIS AND PTERASPIS. 279 



diagnostic marks of a Ganoid, visible in the external skeleton, to two 

 — the presence of " fulcra " and the articulation of the scales by 

 gomphosis. The rest of the essential characters of the Ganoids are 

 entirely derived from the soft parts — the brain, the heart, the branchiae, 

 and the air-bladder. A Ganoid is in fact distinguished from any 

 other fish by the following peculiarities. 



The optic nerves form a chiasma ; the bulbus aortae is rhythmi- 

 cally contractile, and provided with several series of valves ; the 

 branchiae are free ; there is an air-bladder connected by an open duct 

 with the intestine ; the ventral fins are abdominal. These essential 

 characters are shared by only six genera of existing fishes — Lepidos- 

 teus, Polypterus, Amia, Acipenser, Scapirhynchus, and Spatularia 

 — which are no less singular in their distribution than in their ana- 

 tomy. All are essentially freshwater fishes ; all are found in the 

 northern hemisphere; three — Lepidosteus,Amia,sa\& Spatularia — are 

 exclusively North American ; Polypterus is only known in the Nile, 

 while Acipenser is common to Europe, Asia, and North America. 



Now what evidence have we that either Cephalaspis or Pteraspis are 

 in the proper sense Ganoids ? There is nothing about their dermal 

 covering peculiarly characteristic of Ganoids ; and as to the rudiment- 

 ary state of ossification of the vertebral column, there are Teleostean 

 fishes (e. g. Helmichthys) quite as imperfect in this respect as any 

 Ganoid. 



Without doubt there is a singularly close resemblance, in the struc- 

 ture of the dermal plates, between Cephalaspis and Megalichthys — 

 the last being very probably a true Ganoid ; but the point of differ- 

 ence is noteworthy : it is precisely the characteristic ganoin-layer 

 which is absent in Cephalaspis. 



On the other hand, the arrangement of the hard tissues in Pter- 

 aspis reminds one almost as strongly of Ostracion, an undoubted 

 Teleostean. 



The existing fishes to which Cephalaspis presents the nearest 

 resemblance in form, viz. Loricaria and Callichthys, are Siluroid 

 Teleosteans, and not Ganoids ; and, if we take the immediate allies 

 of Cephalaspis and Pteraspis, viz. Coccosteus and Pterichthys, their 

 analogies with Siluroids, such as Bagrus and Boras, are as strong as 

 those with Acipenser. 



A careful consideration of the facts, then, seems to me to prove 

 only the necessity of suspending one's judgment. That Cephalaspis 

 and Pteraspis are either Ganoids or Teleosteans appears certain ; but 

 to which of these orders they belong, there is no evidence to show. 



If this conclusion is valid, it is clear that the ordinary assumption, 

 that the earliest fishes belonged to low types of organization, falls to 

 the ground, whatever may be the relative estimation in which the 

 different orders of fishes are held. 



But it is said that the great development of the dermal skeleton, 

 combined with the rudimentary condition of the endo-skeleton, shows 

 that these early fishes occupied a low place within their own group. 



Mere a-priori argumentation on such questions as these would be 

 a waste of time ; but, happily, we can put the principle involved in 



