observations on eish scales. 
153 
nucleus, very coarse concentric circuli, and about 15 or more very long and strong sharp spines along the 
apical margin. Thus the scales agree in all essential features with those of Aphredoderus, and in their 
marginal teeth closely resemble Xenochares and Distichodus in the African characinid fauna. 
The Aphredoderidae seem quite out of place next to the Centrarchidae. Jordan says of them: 
“ Probably the most primitive of all living Percoid fishes, showing affinities with the Salmopercae.” 
Aphredoderus “agrees with the Berycoid fishes in scales and structure of the fins, and Boulenger places 
it with the Berycidae. ” However, the resemblance to the berycoid scales is remote. In the Aphre¬ 
doderidae the apical teeth are articulated, as it were, on the margin, and in both the living and fossil 
fishes are movable. In the berycids they are in¬ 
stead rigid projections on the marginal area, into 
which they are basally prolonged. 
Regan has recently referred the Aphredoderidae 
to the Salmopercae. 
KUHLIID^. 
Kuhlia rupestris (Cuvier & Valenciennes) from 
Mauritius (M. C. Z. 5710)has rather large subquad¬ 
rate scales, about 10 mm. long and broad, the latero- 
basal comers rectangular, and the lower margin 
nearly straight. The nucleus is about 6 mm. from 
the base; apical region typically ctenoid, of the usual Percoid type, the apical teeth numerous and 
sharp, the submarginal elements truncate (style of Lagodon in Sparidae); basal radii delicate, about 
eleven. This agrees very well in general type with Lucioperca. 
PERCIDiE. Perches. 
The typical percid scale is subquadrate, ctenoid, with nucleus apicad of the middle, and strong 
basal radii. The American subfamilies are separated on the scales as follows: 
Basal radii many or few (e. g., 20 in some Hypohomus, 6 in some Boleosoma ); basal margin not very 
deeply crenate; submarginal apical elements short and broad.Etheostominae 
Basal radii few; submarginal apical elements quadrate; scales larger.. 1 
1. Basal margin very deeply crenate..Percinae 
Basal margin not very deeply crenate.Luciopercinae 
In the Old World we find the following subfamilies: 
Basal margin very deeply crenate or lobed; dorsal fins distinct; no canine teeth {Perea). . .Percinae 
Basal margin moderately crenate or lobed.• - ...• 1 
1. Submarginal apical elements short and broad; marginal teeth long ( Zmgel ).Etheostominae 
Submarginal apical elements longer, more quadrate; marginal teeth usually shorter. ... 2 
2. Dorsal fins distinct (and other characters)... Luciopercinae 
Dorsal fins united (etc.).Acermmae (Acenna) 
I have not examined the genus Percarina. 
In spite of the great development of Percidae in North America, the group must surely have origi¬ 
nated in the Old World, because (1) it is more varied in type there, and (2) it is more difficult to separate 
the European subfamilies than the American on their scales. However, the palaeontological evidence 
indicates considerable antiquity for the percids both in Europe and America, and even if it is true that 
the living American subfamilies are of Old World origin, this does not prove anything regarding the 
history of the much older forms known as fossils. 
The Lucioperca-Acerina type of scale is doubtless the primitive one in the family. It is difficult to 
find much difference between the scales of these genera, so different in other respects; but in Acerina 
the sub marginal apical elements are more conspicuously broadened basally, approaching the condition 
Fig. 13.— Aphredoderus sayanus (Aphredoderidae). Marginal 
teeth. Bureau of Fisheries. 
/ 
