330 
BULLETIN OP THE UNITED STATES FISH COMMISSION. 
Pseudopriacanthus) should certaiuly be omitted from the Serranidce^ as should Apogon 
and its allies, among which are apparently the two aberrant Cuban genera, Liopropoma 
and Chorististium. We are not quite sure that Bypticus, Percichthys or even Morone 
should be retained among the Serranidce, but we prefer for the present to regard the 
three groups typified by these genera as sub-families of rather than to 
attempt to define them as distinct families. 
The Percidw (Perea, Stizostedion, Aspro, JEtheostoma) stand near Morone and Per- 
cichthys, but they are fairly well defined by the greatly increased number of vertebrae, 
a character associated with their fresh-water life and northern distribution. From 
all Serranidee they further ditter in having two anal spines. Near also to the Serra- 
nido; is the singular group of North A.merican sun fishes or (Jentrarchidee. In these, 
however, the pseudo-branchiae are obsolete and the general form of the body is notably 
different. 
The American genera of Serranidee fall naturally into S’x groups, which we may 
call sub-families, although the relation of two of these groups (Serraninw and Epi- 
nephelince) is almost too close for such designation. These six groups, Orammistina;, 
Epinephelince, Serranince, Anthimm, Latinm, and Perciehthyince, are typified respectively 
by the genera Rypticus, Epinephelus, Serramis, Anthias, Morone, and Percichthys. 
They are easily recognized, and in most cases readily defined, notwithstanding the 
fact that most European naturalists have confounded Epinephelus with Serranus, 
while some have placed in the genus Anthias not only genuine Serranince, but also 
species of the Sparoid group of Lutjanince. 
The American Grammistmw all belong to a single genus Rypticus, Grammistes 
having no American representatives. 
The recognized genera of Epinephelince are fairly well defined, and although most 
of them are closely allied to each other, there is no evidence of any intergradation of 
such a character as to efface our lines of generic division. 
The Serraninm are equally varied among themselves, but the lines of division are 
less sharply marked, and certain species form connecting links between one genus and 
another. Authors have divided the group or sub family into numerous genera, but 
these genera are not separated by strongly marked characters, and most of them might 
well be placed as sub-genera under Serranus. 
The species of Anthiinm are' not numerous, but the generic differences in this group 
are more marked than those sejiarating the genera of the two groups just named. The 
representation of this group in our fauna is not extensive enough to give us a clear 
idea of the mutual relations of the forms included in it. 
We unite under the sub family name of Latince (the more usual name Labraeince not 
being available, as the name Labrax properly belongs to the genus Eexagrammus, of 
which name it is a later synonym) certain Serranidee, which show more marked re- 
semblances to the genus Perea. These are characterized in general by the divided 
dorsal, the villiform teeth, and by the absence of a supplementary bone to the max- 
illary. A majority of the species are Asiatic, and but few of these have been studied 
by us. Some of them (Onidon) are said to lack the pseudo-branchim, but this is prob- 
ably not true. Some of the Latince (Lateolabrax) approach the Percidee in the increased 
number (35) of the vertebrae, which are about 25 in most of the others, but they differ 
from the Percidee in this, at least, that none of the latter (Perea, Stizostedion, Etheostoma, 
Acerina, Aspro, etc.), have more than two spines in the anal fin, while all the Latince 
