observations on fish scales. 
167 
These may be separated by the following key: 
Inner apical circuli usually strengthened, forming wavy longitudinal ridges; apical circuli longi¬ 
tudinal; apical margin broadly thin, with only shadowy indications of radii; basal radii about 
8, and laterobasal (ending above the basal comers) one to four on each side; scales about 
ZPi mm. long and 2 broad. Pimelometopon pulcher (subfamily Harpinae) 
Inner apical circuli not thus modified. 1 
1. Scale is very large and broad, yellowish; about 13-14 mm. long and 11-12 broad; basal radii 
numerous, usually over 20. 2 
Smaller, 7 mm. or less broad. 3 
2. Apical radii very well developed, extending from nuclear region to apex. 
Iridio bivattatus a (subfamily Julidinae) 
Apical radii very poorly developed, skin of apical region very thick. 
Tautogolabrus adspersus a (subfamily Labrinae) 
3. Basal radii more than 20, laterobasal (ending above basal comers) 7 to 10; apical radii numerous 
and close together, though imperfect; lateral circuli very dense; scale very long, about 7 
mm. long and 3% wide. Coris venusta (subfamily Julidinae) 
Basal radii less than 20 (perhaps sometimes over 20, counting rudiments), laterobasal 1 to 7; 
scale not twice as long as wide b . 4 
4. Apical radii distinct; scale about 9 mm. long and 6 broad. 
Emmeekia venusta (subfamily Julidinae) 
Apical radii indistinct or imperfect; scale about 10 mm. long and 6 broad. 
Tautoga onitis (subfamily Labrinae) 
ODACID^. 
The scaie of Odax lineatus, as figured by Gunther, is entirely of the labrid type, and is very like 
that of Coris. 
SCARIDAi. Parrot-fishes. (PI. xxxix, fig. 44.) 
Scarus croicensis (Bloch) has scales differing from those of the Labridae in being short and broad, 
shield-shaped, about 6 mm. long and a little less in width. Lateral and basal circuli excessively fine; 
basal margin broadly truncate, lobed in middle, but not in the least crenulate; basal and apical radii 
very well marked, but only the inner ones reaching the nucleus; there are about 16 apical and the same 
number basal and laterobasal; apical region papillose, but the apical margin very broadly thin and 
subhyaline. A very cyprinid-looking scale. 
Suborder Squamipinnes. Scaly-fins. 
EPHIPPIDAS. Spadefishes. 
Gunther has figured the scale of Ephippus or Scatophagus muItifasciatus. It is represented as strongly 
ctenoid, with several rows of teeth; the basal radii are feebly indicated by a few broad depressions. 
CHAJTODONTIDAS. Butterfly-fishes. 
I have Chcetodon ulietensis Cuvier & Valenciennes from Apia, Samoa; C. bricei Smith, from Katama 
Bay; and Pomacanthus arcuatus (Linnaeus). The characteristic ctenoid sculpture of these genera is 
shown in figures 38 and 39. The marginal teeth in Chcetodon are very sharp, and both the teeth and the 
truncate submarginal elements have a specialized triangular area, shaped like a thorn, superimposed, 
as it were, upon their surface. All this is exactly the same in the Atlantic C. bricei as in the Pacific 
C. ulietensis, notwithstanding the great geographical separation. 
In C. ulietensis the ctenoid area is very large and well-formed in all its details; the basal radii are 
about 17; laterobasal anglesobtuse; lateral circuli very slightly oblique. (PI. xxxix, fig. 45.) 
In the Pomacanthus the ctenoid region is much modified, the broad though sharp apical teeth each 
having a very strong rib which extends more than a millimeter basad, the successive ctenoid elements 
a Superficially these scales look exactly alike. 
b In young Tautoga scales, 5 mm. long, the length is twice the width. 
