44 
MEMOmS OF TEE QUEENSLAND MUSEUM, 
The species examined are — 
Chmtodon epMppnon C. & V. Queensland. 
„ octofasciatus Bloch. „ 
„ flavirost7ds Gunther. „ 
„ ii'ifasciai us Park. „ 
,, xilietoisis C. & V. Samoa. 
„ hricei Smith. Atlantic. 
Ileniochus cJirysostonvus C. & Y. Queensland. 
Chelmon rostrahis C. & V. Queensland. 
SCALES OP PO]MACANTHIi\LE. 
(1) The scales of Holacanthus and Chcetodontoplus are entirely of the same 
general type as those of FomacantJms, (Cf. Bull. U. S. Bur. Pish., xxxii, 
p. 167, f.'39.) 
(2) These scales are very different in their ctenoid area from those of the 
Clicetodoniinm, and taking this fact along with others, it may perhaps he 
a question whether the PomacantJiinm should not stand as a distinct family 
Fomacanthidm, 
(3) Having in view the (diaracters of related families, there can be no doubt, I 
think, that the scnle-eharacters of the Fomacanthince are strongly modified 
from the more primitive type of the group, which is much more nearly 
approached by the Chwlodontin-ce. The special modification is however 
entirely in the ctenoid (apical) area. 
(4) Among the Ilolacantliince, Holacanthus hicolor shows the nearest approach 
to Iho cluetodontine scales, having the lower (mesad) part of the ctenoid 
area with small separate elements, instead of the usual continuous long 
rods or ridges. There is some indication of this sort of thing in 
Fomacantkus arcuai us. 
(o) HolacanUius bicolor, with its well defined converging basal radii, appears 
to approach a relatively primitive type ; but its ctenoid area, with long 
parallel rods, seems extremely modified. There is a singular resemblance, 
doubtless quite superficial, to the ctenoid area of Aphrododerus (Bull. 
U. S. Bur. Pish., xxxii, p. 153, f. 13). 
(6) The Chadodontoplus scale with its comparatively narrow form and rounded 
base is peculiar, but the ctenoid area agrees with that of II. bicolor. The 
differences are much less significant than the resemblances. 
