Comments on the Classification of the Percoid Fishes 1 
William A. Gosline 2 
In terms of diversification, the Perciformes 
is by far the most successful of fish orders, 
maximally represented in the inshore waters of 
tropical seas. Thus, of the 130 families of native 
Hawaiian fishes, some 41 belong to the Perci- 
formes, and 18 belong to the single superfamily 
Percoidae, which forms the subject of the pres- 
ent paper. 
From the point of view of classification, the 
order Perciformes suffers from an overabun- 
dance of representation. In fish groups that have 
undergone extensive extinction the modern 
members may be unrepresentative, but at least 
the gaps between them are usually spacious, pro- 
viding abundant material for the construction of 
discrete taxonomic pigeonholes. In the Perci- 
formes, however, the results of repeated adap- 
tive radiations seem to be living today. Some of 
these have led to a relatively high and complex 
structural reorganization (the "mesoevolution” 
of Dobzhansky, 1954). Where this has hap- 
pened, it is comparatively easy to sort lineages. 
But after the major lines of evolution have been 
extracted (as separate orders, suborders, or 
superfamilies), the remaining basal percoids 
seem to represent a central theme with numer- 
ous variations. 
Contributions toward our present knowledge 
of the percoid fishes have been made from a 
number of viewpoints, all of value. Jordan’s 
classification (1923) is based primarily on ex- 
ternal characters, whereas that of Regan (1913) 
relies heavily on the superb series of skeletons 
in the British Museum. In recent years more 
intensive investigations have been undertaken 
from two viewpoints. One is to trace a par- 
ticular structure or structural complex through 
a broad sampling of percoid fishes. This has 
been done for otoliths by Frost (1927, 1928), 
for the predorsal bones and subocular shelf by 
Smith and Bailey (1961 and 1962, respec- 
1 Contribution No. 247, Hawaii Institute of Marine 
Biology, University of Hawaii. Manuscript received 
July 9, 1965. 
2 Department of Zoology, University of Hawaii. 
tively), for the caudal skeleton by Gosline 
(1961), and for the recurrent facial nerve by 
Freihofer (1963). The other approach is to 
study as many characters as possible in a parti- 
cular percoid group, e. g., the Japanese ser- 
ranids (Katayama, 1959) and carangids 
(Suzuki, 1962). 
The present investigation began as an effort 
to ascertain the systematic position and relation- 
ships of a few percoid genera. It soon became 
apparent that, except within the narrowest 
limits, this was impossible withut some over- 
view of percoid classification as a whole. This 
paper is an effort to bring data on the supra- 
occipital crest and the jaw mechanism to bear 
on a general percoid classification and to rescue 
some main outlines of such a classification from 
the amorphous condition into which they have 
deteriorated as a result of the widely varied 
opinions recently expressed on the subject. 
MATERIAL AND ACKNOWLEDGMENTS 
The material on which this paper is based 
consists of skeletons and stained and unstained 
specimens in the University of Hawaii collec- 
tions. Much of the material is of local or 
Central Pacific origin, but for specimens un- 
available from these areas I am indebted to 
various individuals and institutions. For material 
of Bathyclupea I wish to thank Mr. Harvey 
Bullis and the U. S. Fish and Wildlife Service; 
for Pomatomus and Scombrops, Mrs. Myvanwy 
Dick and the Museum of Comparative Zoology; 
for Centropomus, Dr. Reeve Bailey and the 
Museum of Zoology; for Archoplites and 
Roccus, Dr. Warren Freihofer and Stanford 
University; for Ambassis, Mr. Pedro Gonzales 
and the National Museum of the Philippines ; 
and for Lactarius, Mr. Inocencio Ronquillo and 
the Philippine Bureau of Fisheries. 
I also wish to thank Dr. Kiyoshi Yamazato 
and Mr. Jiro Isa for translating for me the 
percoid sections of Matsubara’s books (1955, 
1963). 
409 
