474 
PACIFIC SCIENCE, Vol. IX, October, 1955 
In view of what has been said above, a 
hypothetical explanation can be given for the 
relationship between pectoral counts and sea 
temperatures. The basic assumption is that 
pectoral counts in this species complex in- 
crease with decreasing temperatures. In partial 
isolation the Hawaiian island populations 
would then have developed distinctly higher 
pectoral counts. These would be higher at 
Midway at the northern end of the chain 
than at Hawaii at the southern. However, 
recent introgression from K. marginata at 
Midway could have upset this trend within 
the Hawaiian chain, giving rise to the reversed 
picture for pectoral counts within the Ha- 
waiian Islands shown in Table 5. 
The fact remains that there is more differ- 
ence between the Kuhlia populations in John- 
ston and Hawaii, which are almost similar 
in latitude, than between those of Johnston 
and Midway, which are very different. The 
conclusion seems inescapable that if members 
of the Johnston populations have entered the 
Hawaiian Islands at all, they have come in via 
the low northern islands. Why the Central 
Pacific form of Kuhlia rather than the endemic 
Hawaiian form should be present at Johnston 
remains a mystery. It does, however, bear 
out the point, previously established, that 
some elements of the Johnston biota have 
entered from the south. 
Cirrhitus alternatus-pinnulatus 
The two forms in this complex have re- 
cently been differentiated by Schultz (1950: 
548), but entirely on the basis of coloration, 
the Hawaiian C. alternatus lacking the brown 
spotting of the Central Pacific species. A 
check of the usual meristic characters in spec- 
imens from Hawaii, Johnston, and Christmas 
(in the Line Islands) shows no significant 
differentiation. As Schultz has already pointed 
out {loc. cit .), the Johnston specimens agree 
completely with the Hawaiian form. 
Chaetodon multicinctus-punctato - fasciatus 
This species pair has been separated by 
Woods (in Schultz, et al. , 1953: 571, 575, 
595) on the basis of coloration and certain 
counts. The color differences lie chiefly in the 
nature of the vertical dark bars on the nape 
and caudal peduncle. The fin ray differences 
are shown in Table 7. Woods {loc. cit.) has 
also used scale counts, but I have not been 
able to make sufficiently accurate scale counts 
in this species to be worth recording. 
The Johnston specimens agree with the 
Hawaiian form in both color and counts. 
Acanthurus sandvicensis-triostegus 
The Acanthurus triostegus complex lends it- 
self admirably to geographic analysis for two 
reasons. First, its forms are abundant and 
ubiquitous throughout much of the tropical 
Indo-Pacific, and, second, they differ in char- 
acteristics that are easily seen and calibrated. 
A preliminary analysis of geographic varia- 
tion in this complex has recently appeared 
(Schultz and Woods, 1948: 248-251). Ac- 
cording to these authors two species are rep- 
resented: Acanthurus sandvicensis in the Ha- 
waiian Islands and at Johnston, and A. 
triostegus throughout the rest of the area. The 
differences between these two lie primarily 
in the shape and extent of the mark below 
and at base of the pectoral, secondarily in 
the higher average fin counts of the Hawaiian 
species. 
The Johnston Island population, judging 
from 21 specimens taken in three Johnston 
localities, differs in no way that I can deter- 
mine from the Hawaiian form. If there is any 
admixture of A. triostegus genes in these John- 
ston specimens, it is not apparent. If, how- 
ever, populations of the A. triostegus complex 
from the next island groups to the south of 
Johnston are examined an occasional spec- 
imen turns up with more or less strong traces 
of the Hawaiian pectoral base marking. For 
a study of possible intergradation between 
the Hawaiian A. sandvicensis and the Indo- 
Pacific A. triostegus it seems advisable there- 
fore to focus attention not on Johnston but 
on the Line and Phoenix Islands to the south 
of Johnston. 
