1877.) 57 [Cope. 
of priority will require that the name Ganozdei should be retained for this 
division. 
The tribe Actinoptert as left in my latest paper above quoted (May, 
1877), has slightly different boundaries from those originally prescribed for 
it, since the Chondrostei are now excluded from it. As then and now under- 
stood, it is nearly identical with the Yeleostec of Miller, a name which I 
should adopt for it, were it not that some of his Ganotdei and numerous 
extinct forms with unossified vertebral column belong to it. For such 
fishes the name of Miiller is too glaringly inapplicable to be employed. 
After excluding the extinct genera of Hyopomata which are clearly 
Crossopterygia and Chondrostet, there remains a numerous assemblage, 
whose relationships to existing types of fishes have never yet been ascer- 
tained. I refer especially to the families of the Lepidoides, Sawroides and 
Pycnodontes, of the Poissons Fossiles of Agassiz, and other forms subse- 
quently described; among others, the Dorypterus of Germar. The only bond 
which retained these forms in connection with the fossil Crossopterygian 
fishes, the rhombic and enameled scales, may be safely disregarded in 
view of the important characters of the skeleton which declare their affini- 
ties to be diverse; the more as some of the latter (Celacunthida) have 
rounded seales, and Leptolepis and other genera referred by Agassiz to the 
Sauroides, have cycloid scales. The heterocercal character of the tail of 
some of them, is of but little greater weight. I have already shown that 
fishes presenting this character (Lepidosteus, Amia) do not differ in other 
respects from other Actinopteri, while the still lower isocercal condition is 
often seen in the latter. Further, the extinct genera do not agree among 
themselves in this respect, some bring heterocercal, and some isocercal.. 
The question remains as to the proper location of the families just named, 
in the tribe Actinopter?. It has been impossible to discover all of the 
characters necessary to the fullest elucidation of this question, but the 
greater number of them have been satisfactorily ascertained. The follow- 
ing results are therefore approximations to the truth which I believe that 
future researches into the osteology will confirm. At the least they are 
much nearer to an expression of nature than any yet attained. 
As regards the general affinities represented by the terms P/ysostomi 
and Physoclysti, there is no doubt that the Lepidoides and Sauroides exhibit 
the former. This is seen in the uninterrupted conjunction of the parietal 
bones (where it has been possible to observe the parts), and in the ab- 
dominal position of the ventral fins, and extent of the maxillary bone; as 
well as in the less important features of the absence of all ctenoid charac- 
ters of scales and preoperculum, lack of spinous rays, etc. The Pycno- 
dontide present in general similar characters, and add nothing which 
should separate them widely from the Lepidoid genera of Agassiz, especial- 
ly the family of the Dapediide. Like these, however, they approach nearer 
to the Physoclysti in the anterior continuation of the interneural spines as 
far as the skull. This character is found also in some Physostomous fishes, 
z. e. the greater number of Characinide, the Elopide, Umbrida and some 
PROC. AMER. PHILOS. soc. xvii. 100. H 
