250 
ERIK A : SON STENSIO 
resemble the Catopterids, to which they appear to be transitional forms, while others, 
on the contrary, such as Cheirolepis, Rhadinichthj'S were probably more primitive. 
Woodward (1891 b, p. 425), and Fritsch (1895, p. 126) have supposed that the 
family Palaeonicdidae, when better known, ought to be subdivided, and Gregory (1907, 
p. 463) considered later on that he ought also to emphasize its heterogeneous nature. 
«It is noteworthy*, he says, «that the early Heterocerci, commonly grouped together in 
the family Palaeoniscidae, include forms (e. g. Cheirolepis, Holurus, Coccolepis ) which are 
so different in several important respects that they might almost be reguarded as types 
of different families*. 
Against this it must, however, be pointed out (cf. Traquair, 1911 b, p. 8) that, 
according to what we now know, the Palaeoniscids show on the whole a rather 
uniform type of organization. In spite of the fact that they comprise certain forms that 
appear to be specialized in some respects, yet a subdivision of them into two or more 
separate families would mean the breaking of strong phylogenetic bonds or at any rate 
cause difficulties in denoting the close relationship that undoubtedly exists between 
them. On the other hand, however, we can scarcely be justified in going so far as 
Abel (1919, pp. 186—190), who includes the Catopterids among the Palaeoniscids. We 
should then be justified with about the same right in adding to the family Palaeoniscidae 
in the definition given to it by Abel, the Platysomids as well, since these, as we know 
(Traquair, 1879, pp. 379 — 38 g', must be considered as originating from the Palaeoniscids, 
to which in any case they seem to be very closely related. 
The relation between the Paleoniscids on the one hand and the Saurichthyids 
(Belonorhychids), Chondrosteids, Acipenserids and Polyodontids on the other cannot 
be discussed until later on in connection with my description of the Catopterids 
and the Saurichthyids (Part II). Here, however, I may point out that, with regard 
to the development of the membrane bones on the dorsal side of the ethmoidal 
region and of the supraorbital sensory canal, the Palaeoniscids show special conditions 
which are without parallel among the Actinopteiygii except for the Platysomids and 
Catopterids. These facts are of special interest as they seem to controvert the view 
upheld especially by Woodward (1889 c, pp. 42 — 43; 1895 b, pp. V—VIII; 1898 a, 
pp. 82—95; 1915 a, p. LXXII — LXXIV) and after him by Goodrich (1909, pp. 307 
— 32 i) that Chondrosteids and recent sturgeon fishes are directly descended from the 
Palaeoniscids. 
Finally for a large number of other conditions with regard to the Palaeoniscids 
I may refer to my account given below. 
Family Platysomidae. 
General remarks. 
Right up to quite a recent date the Platysomids were known exclusively from the 
Carboniferous andPermian formations (Traquair 1879; Zittel 1887—1889; Woodward 1891b: 
1898a; Krotov 1904). In igi 3 , however, Broom (igi 3 a) was able to show the existence 
of a species in the Lower Triassic of South Africa and the following year Lambe (1914) 
