E. T. NEWTON ON FISHES* TAILS. 93 



and Calamoichthys of N. African rivers, the tail of the former being 

 apparently diphycercal, but internally slightly upturned, much as in 

 the cod-fish. 



4. Acipenseroidei. This group, which has been established by 

 Dr. Traquair (Palseontographical Society, 1877), to include the 

 recent Chondrosteidce and certain fossil forms, is represented in the 

 Devonian Rocks by Cheirolepis ; and in the Coal Measures, Permian, 

 Trias and Lias, by such genera as Palaoniscus, Amhlypterus and 

 Pygopterns, all of which have tails resembling stage 4 or 5. The 

 Chondrosteidce are represented in the Lias by Chondrosteus, which 

 has a persistent notochord and a strongly heterocercal tail, like its 

 recent representatives the sturgeon, Spatularia, and Scapirhynchus. 

 Stage 3. The sturgeon is known from the London Clay. 



5. Cephalaspidce. These fishes, characterized by having their 

 heads covered by a continuous shield, are the oldest representatives 

 of the class Pisces, The Cephalaspis, from the Upper Ludlow, has 

 a heterocercal tail representing stage 3. The Pteraspis, from the 

 Lower Ludlow, is the earliest fish yet found, and so far as known 

 the tail was heterocercal. 



6. Placodermi. This is another remarkable group of fishes, only 

 known in the Devonian and Carboniferous strata. One of these, 

 Coccosteus, had a truly diphycercal tail, and persistent notochord, 

 and possibly the others were the same. Stage 1. 



7. Acanthodidce. This, like the last group, is confined to the 

 Devonian and Carboniferous rocks. These fish have the body 

 covered with shagreen, and have long spines to the dorsal, pectoral 

 and ventral fins. Their tails are strongly heterocercal and represent 

 stage 4. 



Taking the Ganoids as a whole, we see that their earliest re^u'e- 

 sentatives were heterocercal, Cephalaspis, &c, but in the oldest beds 

 where they occur in any numbers, we find every form of tail, from 

 the diphycercal, as in Coccosteus and Glyptolajmus, to the extremely 

 heterocercal, as in Tristichopterus, while, as M. Alex. Agassiz 

 truly states, we have no such form of tail as that of Lepidosteus, 

 which makes a very close approach to the externally homocercal 

 form, until we get to the Secondary strata, where Lepidotus and 

 Pachycormus are perhajDs the most highly developed forms, and 

 judging from M. Agassiz's restorations, had homocercal tails. 



Teleostei. The Teleostei are not certainly met with in the fossil 

 state until we get up as far as the Cretaceous Rocks, but there we 



