OBSERVATIONS ON FISH SCALES. 
145 
Submarginal elements of apical area not appearing concave below 23 
23. Submarginal elements of apical area hat-shaped. 
Cottogaster shumardi {Giitaxd). Wabash River, Ind. (B. W. Evermaim). (Percidae) 
Submarginal elements of apical area quadrate. 
Meniicirrhus saxatilis (Bloch & Schneider). Woods Hole, Mass. (Scisenidse) 
Suborder Synentognath. 
BELONID/E. Marine gars. 
The hoimdfish, Tylosurus acus (Lacepede) has irregular elongate (broad) scales with conspicuous 
concentric circuli, but no radii whatever. They thus recall the scales of Salmonidse. 
EXOCtETIDiE. Flying-fishes. 
Cypsilurus heterurus (Rafinesque) has large broad thin scales, about 10 mm. broad and 8 long;, 
nucleus a short distance below middle; circuli strong, concentric, in some scales much more widely 
spaced laterally, and then often angled; margin wholly cycloid; about 4 or 5 strong basal radii (more in 
latinucleate scales), none apical. Except for the presence of radii, this is very similar to Tylosurus, 
It seems also to lead toward the scombrid type. 
HEMIRHAMPHID^. Halfbeaks. 
Hyporhamphus unifasciatus (Ranzani), from Woods Hole, Mass., has very broad scales, about 5 mm. 
long and 8, or slightly over, broad; apical margin broadly rounded; laterobasal comers rounded but 
evident; middle of base with a broad median lobe, which may be bilobed; basal radii two or three,, 
widely diverging; nucleus near middle, but indistinct; apical field entirely covered with very fine 
transverse circuli, except a slender sculptureless submarginal band; basal field with transverse more 
widely spaced circuli, bending upwards and still more widely spaced in the lateral fields, and ulti- 
mately, when not reaching the margin, forming acute angles with the adjacent circuli of the apical 
field. The transverse dense apical circuli remind one of Xystcema (Gerridae), but in Xystmma there 
is no sign of lateral angulation, and the apical marginal region is entirely different. 
The scale of Hyporhamphus is evidently a further development, much more extreme, of the type 
of Cypsilurus . It also seems to point clearly in the direction of the scombrids. 
SCOMBRESOClDyE. Sauries. 
The billfish, Scomhresox saurus (Walbaum) represents a still more extreme development along the 
lines of Hyporhamphus. The thin scale is about 5 mm. long and 8 broad, without angles or radii. The 
circuli are everywhere completely transverse, denser on the apical than the basal half, the two halves 
separated in some cases by more or less of an interval about the middle. At each extreme side is a 
thin zone, sculptureless except for a few irregular lines. The scales are of the same general type as 
those of the Scombridse and Cheilodipteridae. 
So far as seen, the synentognathous families may be separated thus: 
Apical and basal circuli about the same i 
Apical circuli transverse, denser than basal 2 
1. Radii absent Belonidae 
Basal radii strong Exocoetidae 
2. Without radii Scombresocidae 
With radii Hemirhamphidae 
If doubt should arise owing to the apical circuli of Scombresocidae not being much denser than the: 
basal, the family will still be easily recognized by the completely transverse character of the circuli, 
so that the effect of lateral angulation is lost. Dr. Jordan states that Regan (1911) recognizes an order 
Synentognathi, with one suborder for the belonids and scombresocids, and another for the hemirham- 
phids and exoccetids. The scale characters would not have suggested this arrangement. 
