Fish Fauna of Johnston Is. — Gosline 
nichthys cookei-laticaudata , Gymnothorax euros- 
tus-buroensis , Kuhlia sandvicensis-marginata , 
Cirrhitus alternatus-pinnulatus , Chaetodon multi- 
cinctus-punctato-fasciatus , and Acanthurus sand- 
vicensis-triostegus. A uniform treatment for all 
six species would be desirable in order to 
enable comparison of all six area by area. 
Unfortunately the availability of specimens 
makes this impossible. 
Before these species are dealt with, it seems 
advisable to provide certain background in- 
formation. The Johnston collections were 
originally made to check whether the endemic 
Hawaiian fishes were really species or merely 
subspecies. It was felt that intergradation be- 
tween the Hawaiian endemics and their Cen- 
tral Pacific counterparts would occur at John- 
ston if anywhere. It does not occur there 
(or elsewhere) among any of the fishes here 
investigated, and on the basis of absence of 
intergradation (the term is here used in con- 
trast with introgression) the Hawaiian ende- 
mics must be considered full species. But 
would the Hawaiian endemics interbreed with 
their Central Pacific counterparts if both were 
present? There is no way of determining this 
at Johnston, because the two never occur 
together there (or elsewhere) . When a Central 
Pacific fish is represented by a variant in Ha- 
waii, either the Central Pacific form (e.g., 
Kuhlia marginatd) or the Hawaiian form (e.g., 
Muraenichthys cookei , Gymnothorax eurostus , 
Cirrhitus alter nat us, Chaetodon multicinctus , and 
Acanthurus sandvicensis) , or neither, but not 
both, occurs at Johnston. That this proves 
nothing regarding the interbreeding poten- 
tialities of the two geographic forms has been 
discussed in an earlier paper (Gosline, In 
press). The failure to be able to determine 
whether interbreeding between the Hawaiian 
endemics and their Central Pacific counter- 
parts would or does occur makes it impossible 
to prove whether the Hawaiain forms are full 
species or merely subspecies. This matter has 
also been discussed elsewhere (Gosline, op. 
cit.). The point here is that the failure to settle 
the matter has led to considerable zoogeo- 
469 
graphic misunderstanding concerning en- 
demism in the Hawaiian inshore fish fauna. 
For example, Jordan and Evermann (1905: 
32) conclude that about 50 per cent of the 
species of Hawaiian shallow water fishes are 
endemic; Fowler’s (1928) treatment of the 
same fishes would give a far lower percentage 
of endemism, perhaps 15 per cent. This ap- 
pears to be a disagreement concerning the 
number of endemic Hawaiian fishes, but in 
reality it is a difference in viewpoint regarding 
how many Hawaiian endemics should be 
treated as full species. That one viewpoint is 
correct and the other incorrect will probably 
never be proved. About all that can be said 
is that in general the Hawaiian offshoots of 
Indo-Pacific species are more distinctive than 
those that occur anywhere else. Since I believe 
that Jordan and Evermann’s interpretation of 
Hawaiian endemism in fishes brings out this 
point more clearly than Fowler’s and since 
no real intergradation can be demonstrated 
between Hawaiian and Central Pacific forms, 
it seems preferable to side with Jordan and 
Evermann. 
Muraenichthys cookei-laticaudata 
The Hawaiian form, Muraenichthys cookei , 
was described by Fowler (1928: 41, fig. 9)- 
In 1943 Schultz (p. 53) synonymized Fowler’s 
species with Muraenichthys laticaudata (Ogil- 
by) described from Fiji. In 1949 Schultz and 
Woods (p. 172) recognized both species, 
differentiating them on the basis of the more 
anterior position of the dorsal origin in rela- 
tion to the anus in M. cookei. The same basis 
of differentiation is used by Schultz (in Schultz, 
et al 1953: 72-73). No other differences 
between the two species are known. The rela- 
tionship between the dorsal origin and the 
anus in specimens of the M. cookei-laticaudata 
complex from several localities is shown in 
Table 3. (In the table total lengths have not 
been given since there is no evidence of a 
change in the dorsal-anus relationship with 
growth.) Several points can only be suggested 
by this table since the within-sample variabil- 
