150 
BUL^I^ETIN OP THE BUREAU OF FISHERIES. 
CHEILODIPTERIDiE. Bluefishes. 
The bluefish, Cheilodipterus saltatrix (Linnseus), from Onset, Mass., has scales slightly over 2 mm. 
long and about 3 broad; apex broadly rounded, base flattened, slightly and obtusely angled in the 
middle; laterobasal comers rounded but evident; no radii; circuli transverse apically and basally, more 
or less angulate laterally; nucleus below the middle. The lateral circuli are often angulate only on one 
side. The apical margin is thin, and when not worn off is irregularly dentate; the coarse circuli are not 
deflected or angled below these teeth, as they are in Scomber. This scale is entirely of the scombrid 
type. 
STROMATEIDAE. Butterfishes. 
Gobiomorus gronovii (Gmelin), the Portuguese man-of-war fish, has feeble cycloid scales with 
widely spaced circuli which are more or less wavy. There are no radii. My material is perhaps imma- 
ture. In the butterfish, Poronotus triacanthus (Peck), from Woods Hole, the scales are as in Cheilodip- 
terus except that there are no signs of apical teeth, while the base often has one or two broad radial folds. 
The lateral circuli are angulate in most scales. 
The pilot-fish, Palinurichthys perciformis (Mitchill), has quadrate scales about mm. long and 
broad; laterobasal corners distinct; nucleus below the middle; circuli very distinct, the basal denser 
than the apical, the lateral vertical, parallel with the margin; four or five indistinct basal circuli. This 
scale is like that of the lake herrings, Leucichthys, and does not at all suggest that of the Scombridae and 
allies. Jordan & Evermann separated the Centrolophidse (Centrolophus and Palinurichthys) from 
Stromateidse, but Jordan has more recently made Stromateidae include both Centrolophidse and Nomeids 
{Gobiomorus). The scales would strongly suggest that Centrolophidse, at least, are a valid family. 
There is a striking general resemblance between the scale of Palinurichthys and that of Chirostoma crystal- 
linum Jordan & Snyder (Atherinidse), but the apical region of the Chirostoma scale is quite different. 
PEMPHERID.®. Deep-water Catalufas. 
Pempheris otaitensis (species perhaps wrongly determined) from the Red Sea (M. C. Z.), has scales 
which differ greatly in size on the same fish. The larger ones are about 4 mm. long and 6 broad; strictly 
cycloid, with circuli all around, transverse in the apical region; no apical radii; nucleus basad of 
middle; laterobasal comers rounded; about 15 strong basal radii, arranged fan-like, the basal margin 
strongly crenate. This scale, while very different from that of Palinurichthys, shows enough resemblance 
to suggest affinity. 
Superfamily PERCOIDEA. 
CENTRARCHIDAJ. Sunfishes and basses, 
Mr. B. A. Bean very kindly sent me from the United States National Museum a very full set of 
centrarchid scales, with numbers only, so that I might classify them without any bias derived from a 
knowledge of their generic and specific identity. This I did, and afterwards the names and localities 
were supplied. The result was that I found the group to be on the whole a very compact one, the 
principal differences being foxmd in the development of the apical teeth (the ctenoid feature), the 
density of the apical circuli, and the number of basal radii. The scales are quadrate or subquadrate in 
form, sometimes (Micropterus dolomieu) longer than broad, others (as Lepomis gibbosus) broader than 
long. No doubt the forms with well developed apical teeth are the most primitive (at least as regards 
their scales), the centrarchid type being originally ctenoid. It is interesting to note that Archoplites 
interruptus from San Francisco, the only fresh-water percoid west of the Rocky Mormtains, has very 
strongly ctenoid scales, and the basal radii with an extremely minute beading. It is remarkable that in 
other forms the ctenoid margin is lost as the scale matures, thus; 
(i) Ambloplites rupestris, rock bass; young from Sodus Bay, N. Y., show a ctenoid V-shaped apical 
area, the apical margin with many sharp teeth ; adults from Manchester, Va. , show a perfectly 
entire apical margin, though (except in latinucleate scales) the V-shaped ctenoid patch is still 
visible as a sort of latticework pattern, failing below the margin. This is very different from 
the broad ctenoid area of Archoplites, the lower margin of which, in well-developed scales, 
is nearly straight. 
