Classification of Percoid Fishes — Gosline 
both today and in the fossil record, little- 
known beryciform-like fishes lead out in all 
morphological directions from the central 
berycoid core. Patterson’s other suggested points 
of origin for percoids are not core berycoids, 
but rather are peripheral forms for which little 
knowledge is available : Sphenocephalus, Aipich- 
thys, the Pycnosteroididae, and the Dinoptery- 
gidae. What is known about these four groups 
gives me, at least, no feeling of assurance about 
even their beryciform affinities. Indeed, the 
various percoids Patterson suggests as separate 
derivatives from them (Serranidae, Menidae 
and Carangidae, Chaetodontidae, Centrarchi- 
dae) would seem to me to be a far more 
close-knit group than the various forms from 
which they were supposedly derived. 
That some of the lesser known, "peripheral” 
fossil berycoids like Aipichthys, Sphenocepha- 
lus , Dinopteryx, and Pycnosteroides may prove 
to be nearer the ancestral percoid type than the 
"core” berycoids seems quite probable, if only 
because the reduction in ossification which may 
well have led to the lack of knowledge of the 
"peripheral” berycoids is also a step in the 
direction of the percoids. Stated conversely, the 
"core” berycoids, except possibly the Polymixii- 
dae, seem to be too completely ossified to have 
been percoid ancestors. 
So far the objections to deriving different 
basal percoids from separate berycoid stocks 
have been of a general nature. Some discussion 
of the supraoccipital crest will, I think, provide 
evidence against certain of the specific deriva- 
tions postulated by Patterson. 
The supraoccipital crest attains extensive 
development only in the teleosts. Among lower 
teleosts the supraoccipital tends to be a small 
bone at the rear of the skull that does not 
extend forward between the parietals. In gen- 
eral, the size of the supraoccipital bone is quite 
closely associated with the size of its crest: 
where the crest is large the supraoccipital is 
large, and in groups where the crest is low or 
lacking the supraoccipital may disappear, as in 
some congrid eels. In the acanthopteran fishes 
the supraoccipital crest may become very large; 
here the deeper-bodied fishes tend to have larger 
crests. 
Among deep-bodied fishes, however, there 
are two quite different types of crest and, 
411 
though each has doubtless arisen many times, it 
seems improbable that one would give rise to 
the other. In strongly swimming, compressed 
fishes the body musculature tends to extend far 
forward over the head. This musculature has its 
attachment in large part to the supraoccipital 
crest which extends forward within it. In the 
berycoid Homonotichthys (Patterson, 1964:284, 
fig. 35) and in numerous percoids the supra- 
occipital comes forward between the frontals 
or its crest extends forward over the frontals. 
In a whole series of percoid families, including 
the Carangidae, Coryphaenidae, and Priacanthi- 
dae, an even greater anterior extension of the 
supraoccipital crest is brought about by the 
development of a median ridge on the paired 
frontals. In all these fishes the crest is buried 
in musculature and is not broadly exposed on 
the surface, its upper rim being more or less 
knife-edged. Finally, the cranial roof below 
the crest is not especially vaulted or otherwise 
distended. 
A quite different type of supraoccipital crest, 
constructed in another way and serving another 
principal function, is that of the acanthuroids, 
chaetodontids, Antigonia, etc. Here, in essen- 
tially slow-moving, spinous fishes, the supra- 
occipital extends up and back over the nape as 
a sort of protective shell. (This same type of 
development is found in some of the lower 
teleosts, e.g., catfishes.) One can envision this 
construction as arising from the condition in 
fishes like the modern berycoids Holocentrus or 
MyrtprisUs . In these the supraoccipital crest is 
small and extends directly back from the upper 
surface of the skull; its upper border has flat- 
tened out somewhat. If, in deeper-bodied forms 
such as the extinct holocentrid Caproberyx, this 
type of supraoccipital crest, along with the 
whole posterior portion of the cranium, were 
to be raised up and expanded over the nape, 
then the development of a posteriorly vaulted 
skull continued up and back as a rooflike su- 
praoccipital crest would occur. In such a fish as 
Chaetodon, which has this type of construction, 
there is very little muscle attachment to the 
high, broad, and strong supraoccipital crest. 
The vaulting at the rear of its cranium adds 
structural strength to the crest base (and also 
has the curious result of leaving the brain 
resting in the bottom of a high, empty vault). 
