386 RAMSAY H. TRAQUATR ON THE 
characters can be used in the definition of a group except such as are 
altogether absent from every other with which it is compared. For example, it 
is perfectly legitimate in comparing the Acipenseroid and Lepidosteoid 
Ganoids to bring forward the presence of infraclavicular plates in the former 
and their absence in the latter, although these plates are also present in 
Crossopterygii, because the latter are in their turn separated from the 
Acipenseroidei by other important characters, such as the presence of jugular 
plates and the lobation of paired fins. What we require is that the assemblage 
of characters shall be exclusive. And it certainly seems to me that the 
characters assigned by Professor Youne to his Lepidopleuride are, both taken 
individually, and in the aggregate, quite insufficient either to characterise a sub- 
order, or to differentiate the fishes therein included from the Lepidosteide of 
Professor Hux.ey ; while, on the other hand, the wide gulf which exists between 
the Platysomide and Pycnodontide, in certain very essential points of structure, 
is ignored. 
The suborder Lepidopleuridz must, therefore, in my opinion, be abandoned, 
and the affinities of the Platysomidee traced in another direction. 
Affinities of the Platysomide with the Paleoniscide. 
The Platysomide agree with the Paloniscide in the following points :— 
1. The vertebral axis is notochordal; the neural and hzemal spines are short; 
and there are two sets of interspinous bones, proximal and distal, supporting 
the rays of the dorsal and probably also of the anal fin. 
2. The fins are fulcrated. 
3. The rays of the median fins are more numerous than their supporting 
interspinous bones, which they overlap with their proximal extremities ; they 
are closely set, closely jointed, and in the anterior part of each fin the demi- 
rays are closely imbricated. 
4, The caudal fin is completely heterocercal and acipenseroid in aspect, and — 
the upper margin of the body-prolongation is set with a row of pointed imbri-— 
cating V-scales. 
5. The paired fins are similar in structure in both families, though the ventrals _ 
are small in some Platysomidze, and not observed in others. The shoulder 
girdle is composed of the same elements, which in both families have an exceed- 
ingly similar shape, and include well-marked infraclavicular plates. 
6. The osteology of the head is morphologically similar in both. The cranial — 4 
roof bones correspond, plate for plate, both in number and in their relative posi- 
tions to each other, and the nasal opening on each side is situated between the — 
median superethmoidal and the dermal anterior frontal, As in the Palonis- 
cide, the preoperculum extends forwards on the cheek, and the other oper- 
