Classification of Percoid Fishes — Gosline 
to be the Sillaginidae (which I have not seen) 
and the Branchiostegidae (Table 1). No ex- 
planation for these apparent anomalies will be 
attempted here. 
The remaining question to be discussed is 
how Regan’s arrangement (1913) of percoid 
families agrees with other proposed classifica- 
tions. Since Patterson’s views (1964) are much 
the most radical (see above), they will be con- 
sidered first. Patterson, as previously noted, has 
suggested five independent origins for different 
percoid families. Only one point with regard to 
these percoid derivations will be added here. 
Patterson (1964:470, and elsewhere) stressed 
the primitiveness among percoids of the Scorpi- 
dae, Monodactylidae, and Kyphosidae as being 
"the only perciform families which retain both 
a toothed endopterygoid and a separate foramen 
for the hyomandibular trunk of the facial nerve 
in the lateral wall of the pars jugularis.” Al- 
though the "primitiveness” of these particular 
features is not in dispute, it would seem that 
the dentition and jaw structure of at least 
Kyphosus in the Kyphosidae and of Microcan- 
thus in the Scorpidae (Scorpididae) are highly 
specialized and have evolved a long way from 
that of Eery cop sis, from which Patterson would 
derive them. If, in fact, one were to use jaw 
structure as a basis for postulating lineages, 
Eery cop sis would seem to provide a much more 
appropriate ancestor for either Epinephelus or 
Lates than for Kyphosus or Microcanthus. 
Freihofer (1963) surveyed the various path- 
ways followed by the ramus lateralis accesso- 
rius of the facial nerve in teleostean fishes with 
particular reference to the percoids. These dif- 
ferent pathways were then grouped into pat- 
terns. Freihofer’ s Patterns 8, 9, 10, and 13 
occur among the percoids, though 13 is re- 
stricted to the family Mullidae; it should also 
be added that a rather large number of species 
distributed through various percoid families 
have the ramus lateralis accessorius absent or 
reduced. An arrangement of percoid families 
based solely on the patterns worked out by 
Freihofer would cut across that of Regan 
(1913), for Patterns 8, 9, and 10 all occur 
among families with an axillary process and 
again among families without an axillary pro- 
cess. Thus, the two methods of arrangement 
would divide percoid families in quite different 
417 
ways, although neither one is for that reason 
necessarily an incorrect indicator of relation- 
ships. 
The last arrangement of percoid families 
that will be discussed is that of Matsubara 
(1955 d,b, 1963). Matsubara raises the divi- 
sions (superfamilies) of Regan (1913) and of 
Norman (1957) to subordinal status. But more 
pertinent to the present discussion, he places 
the families accepted here as members of the 
Percoidae under four different suborders. The 
family Coryphaenidae he places in the Scom- 
brina (equals Scombroidei) . The Carangidae, 
Formiidae, Leiognathidae, Lactariidae, Menidae, 
and Rachycentridae comprise Matsubara’ s 
(1955^:ix) Carangina. The Monodactylidae, 
Toxotidae, Drepanidae, Ephippidae, Platacidae, 
Scorpididae, Antigoniidae, Chaetodontidae, 
Scatophagidae, Zanclidae, and Acanthuridae 
make up his (Matsubara, 1955^4) Chaetodon- 
tina. In the Percina Matsubara (1955^:ix-x) 
includes the remainder of the families consid- 
ered here as percoids, and adds the Cirrhitidae. 
His serial arrangement of families, though not 
his groupings, seem to follow rather closely 
Jordan’s (1923) "Classification of Fishes.” 
(Perhaps Matsubara has stated the historical 
background for his classifications of 1955 and 
1963, but both these works are in Japanese and 
I have been able to get only parts of them 
translated.) Actually, the relationships of the 
groups excluded from his Percina have been 
moot points for a long time (Gregory, 1933). 
The difficulty, if the carangoids and chaetodon- 
toids are set up as separate groupings from the 
percoids, is to know what to include in them 
and on what basis. No two authors are agreed 
on these matters. If and when carangoid and 
chaetodontoid groups can be satisfactorily de- 
fined, they should undoubtedly be recognized; 
meanwhile, however, more confusion than clari- 
fication would seem to result from such recog- 
nition. 
SUMMARY AND CONCLUSIONS 
The various percoid fishes (Superfamily Per- 
coidae) are considered to represent a single 
structural theme with numerous variations. 
There seems no need to consider the group to 
be of polyphyletic origin, and data on the 
