44 
ERIK A : SON' STENSIO 
a description of the two genera , and in connection with this established his family «des 
Coelacanthes*. As the name itself indicates, the type genus for this family was Coelacanthns. 
Among his «Coelacanthes» Agassiz now includes a number of different forms. Besides 
Coelacanthus and Macropoma he mentions the following genera: Holopthychius Agassiz, 
Glyptosteus Agassiz, Phyllolepis Agassiz, Glyptelepis Agassiz, Psammolepis Agassiz, Hoplopygus 
Agassiz, Uronemns Agassiz, Undina Munster, Ctenolepis Agassiz, and Gyrosteus Agassiz. 
The forms mentioned were for the most part preserved in a very fragmentary manner 
and the definition of the family that Agassiz was able to give was consequently, as he 
points out himself, very incomplete. For this reason alone it is therefore easy to explain 
why the family should have a certain heterogeneous composition, apart from the principles 
on which Agassiz in other respects based his classification. It is curious, however, that 
he considered it possible to incorporate Macropoma with the «Coelacanthes» only with 
hesitation, while he places even Placoderms among them without any reservation. He 
certainly points out the considerable analogies between Macropoma and Coelacanthus, 
but at the same time he believes he can show differences between them of a profound 
nature. The close mutual relationship between Coelacanthus, Undina and Macropoma was 
thus not clear to Agassiz, a fact that appears still more in a later work, his «Mono- 
graphie des poissons fossiles du vieux gres rouge*- (pp. 59—109), published in 1844. 
In this work Agassiz considers that he has gained a somewhat deeper insight into 
the organization of his «Coelacanthes». He puts forward as their most important character 
the cycloid ,«enamel-coated* scales and the «hollow» nature of the bones. In other res¬ 
pects he finds them closely connected with his «Sauroids». From the latter he has now 
also transferred the following to the «Coelacanthes»: Platygnathus Agassiz, Dendrodus 
Owen, Lamnodus Agassiz, and Cricodus Agassiz; and from the Cephalaspids Chelonichthys 
Agassiz, which he now gives under the name of Asterolepis Eichwald. Glyptosteus is now 
called Bothriolepis Eichwald and Psammolepis is termed Psammosteus Agassiz. He does not 
restrict himself in this conqection to asserting more decidedly his previous views on 
Macropoma, but goes considerably farther. It is not only Macropoma that in his opinion 
fits in badly with the real «Coelacanthes» but also Undina as well, and he is also 
doubtful whether a number of other forms such as Bothriolepis , Asterolepis and Psammosteus, 
did not in fact represent a special type, quite different from the Coelacanthus-type. 
In an appendix, published in the work just mentioned together with the last part 
of the description of the «Coelacanthes», Agassiz, 1844, pp. 107—109, considers that he 
could with certainty add to them even the recent genus Arapaima and probably the 
similarly recent genera Heterotis, Osteoglossum and Ami a, as well, a view that is 
emphazised more strongly by Vogt in 1845 (pp. 53—68). Moreover AgasSiz has in letters 
expressed the opinion that Ceratodus Agassiz too would belong to the «Coelacanthes» 
(Gunther, 1871, p. 559). Finally Egerton in 1861, pp. 51—55 adds to the «Coelacanthes» 
as defined by Agassiz Tristichopterus Egerton and Dipterus Sedgwick and Murchison. 
In 1845 Pictet (pp. 23 i —236 ) closely follows the views put forward by Agassiz 
in his «Recherches». On the other hand Giebel (1848, pp. 219—227, 270—280), three 
years later, had apparently already perceived to some extent the heterogeneous character 
of the «Coelacanthes-» as conceived by Agassiz: The fossil forms included among them 
by the latter are thus found in Giebel’s «Fauna der Vorwelt» divided among three 
