PISCES. 
165 
under repeated protest. He objects to the order of Acantho- { 
pterygians, which,, according to his views, onglit to be much ! 
more limited ; fishes in which the spines are flexible ought to be 
excluded therefrom, &c. He objects to the several families as 
being much too wide, as, he says, is shown by the frequent oc- 
currence of t)ie word generally in their diagnoses, and by the 
fact that Dr. Gunther is frequently obliged to dissolve his 
families into groups or subfamilies. He designates such groups 
as unnatural, if they contain forms which (as is so often the 
case) show some deviation from the family-type, although other- 
wise evidently most nearly related to it. He criticises all the 
instances where different values have been attached to one and the 
same character in different parts of the system. 
From these and similar remarks of Prof. Kner it appears to 
us evident that we shall hardly ever agree as regards a natural 
system of Fishes. The principle which we follow is this : we \ 
compare the individual forms, weighing their points of affinity 
against those of diversity, and until the latter are found to pre- 
dominate we are always averse to drawing a line of separation. 
Nobody can deny that the anterior tentacles of Antennarius, even 
of Mali he, the cephalic disk of Echeneis, are homologous to the 
spinous dorsal fin of a Perch ; and as these fishes have also the 
other chief characteristics of Acanthoptemjgians, we leave them 
united in the same group. We know also from experience that a 
principal character of a group may remain undeveloped in some of 
its members. Thus, although it appears a contradiction that a 
fish like Gobiesow, without spinous dorsal, should be referred to an 
order one of the chief characters of which is a spinous dorsal, 
yet, on further comparison, a greater affinity will be discovered 
with Acanthopterygians than with any of the other orders ; but we 
are obliged to characterize Acanthopterygians as fishes provided 
generally or normally with a spinous dorsal. So also we 
find the amount of identical characters in Lot'icaria and other 
Siluroids far exceeding that of the differences ; at all events, no 
one will deny that they are more nearly allied to Siluroids than 
to any other family of fish ; and to raise them into a distinct 
family, equivalent to Silw'idce, Cyprinida, &c., is a proceeding 
acceptable, perhaps, on the ground of the general appearance of 
those fishes, but certainly to be rejected in a natural system. 
Prof. Kner^s assertion, that the subdivision of the families of the 
^ Catalogue^ into groups is a proof of their want of definition, needs 
scarcely a reply. We might just as well talk of the too great 
latitude of his families, because they are subdivided into genera. 
Practically speaking, a system of Acanthopterygians (whether 
they be taken according to the definition of J. Muller or in the 
restricted sense hinted at by Prof. Kner) will be a better system 
if built up of a certain number of divisions, each division being 
subdivided into families, each family composed of several groups. 
