360 
PACIFIC SCIENCE, Vol. IX, July, 1955 
scales not elevated and ctenii not long; 
least depth of caudal peduncle 2.1 to 3.2 
in length of head; caudal peduncle spine 
in a sharply-defined groove .......... 5 
4b. Dorsal spines IV or V; length of longest 
dorsal ray contained 2.2 to 3.8 times in 
standard length; scales elevated and with 
long ctenii; least depth of caudal pedun- 
cle 3 to 3.5 in length of head; caudal 
peduncle spine in a shallow depression. 
(Indo- West-Pacific) Zebrasoma 
5a. Teeth fixed, not attenuate with expanded 
incurved tips, denticulate on both lateral 
and medial margins, and not over 26 in 
upper jaw; dorsal spines IX (except one 
species with VI or VII and two with 
VIII). (Circumtropical) . . . Acanthurus 
5b. Teeth movable, attenuate with expanded 
incurved tips which bear only lateral 
denticulations, and from 30 to 60 in 
upper jaw (of specimens over 75 mm. 
in standard length); dorsal spines VIII. 
(Indo-Pacific) . Ctenochaetus 
The relationship of the existing genera of 
surgeon fishes is not easily fitted into a con- 
ventional family tree pattern. Aoyagi (1943: 
196) has constructed such a tree for the 
Acanthuridae on the basis of dentition alone. 
For this one character his conclusions are well 
drawn. Naso , with its conical teeth, is listed 
as most primitive. N. lituratus (Bloch and 
Schneider) and N. unicornis (Forskal) exem- 
plify those species of this genus which have 
teeth lacking denticulations. Others, like N. 
hexacanthus (Bleeker) , have tiny denticulations 
and are higher in the evolutionary sequence. 
Prionurus , Par acanthurus, Zebrasoma, and Acan- 
thurus are progressively more specialized, 
though these four genera are basically similar. 
The teeth have become close-set, flattened, 
and strongly denticulate. Ctenochaetus , which 
has comb-like teeth, is portrayed as being 
derived from Acanthurus . 
This picture is strengthened by a considera- 
tion of food habits of the genera. The acan- 
thurids, in general, are herbivorous. Naso 
tends to feed on leafy algae such as Sargassum; 
its teeth are not efficient for feeding on slick, 
filamentous algae. The flattened, denticulate 
teeth of the next four genera might represent 
a specialization for feeding on fine algae. 
Ctenochaetus with its numerous, long, pro- 
truding teeth can feed effectively on loose 
algal filaments and other detrital material on 
the bottom. 
When, however, characters other than den- 
tition are considered, the interrelationships 
of the genera are not so simple. Naso and 
Prionurus , both with fixed caudal spines, have 
different numbers of pelvic rays. Par acan- 
thurus, which has a movable caudal spine and 
is separated from Naso by Prionurus in the 
phyletic line of Aoyagi, has a pelvic formula 
of I, 3 like Naso. It appears then that the 
reduction in pelvic fin rays from I, 5 to I, 
3 must have occurred independently in Naso 
and Par acanthurus or that the I, 5 condition 
was secondarily regained in Prionurus . The 
dorsal spines in Naso are IV to VII. In Prio- 
nurus and Paracanthurus they are VIII or IX. 
In Zebrasoma they drop to IV or V, and in 
Acanthurus and Ctenochaetus they increase 
again to VIII or IX. The scales present an 
even more perplexing problem. The sup- 
posedly advanced genera, Acanthurus and 
Ctenochaetus , have ctenoid scales which are 
less specialized than the unusual raised and 
often spinulous scales of the other genera. It 
is difficult to place Acanthurus in a more 
primitive position than Naso , however, for it. 
does not seem that a folding caudal spine and 
denticulate teeth could precede a fixed spine 
and smooth teeth. Perhaps the linear pattern 
of evolution postulated by Aoyagi would be 
less likely than one which supposes that all 
the Recent genera (except Ctenochaetus , which 
does appear to be derived from Acanthurus) 
arose from common stock at essentially the 
same period of geologic time. 
Eastman (1904*) has commented on the 
sudden appearance in the Eocene of a host 
of modern teleost types, many of which were 
