Differentiation of Blenniids — Strasburg 
mean surface water temperature. These data 
have been arranged by latitude so that spec- 
imens from the northernmost islands (West- 
ern Hawaii) appear at the top, and those from 
the southernmost (Gambier Islands) at the 
bottom of the table. In counting, each fin 
ray with a separate and distinct base was 
listed as a single ray, and those rays split to 
a single base were also counted as one ray. 
Inspection of Table 1 reveals a definite, but 
not absolutely regular, decrease in mean soft 
ray count as one goes from northern latitudes 
toward the Equator. Going south from the 
Equator the mean counts gradually increase, 
ultimately approaching or exceeding those 
found in the most northerly islands consid- 
ered. Analysis of mean water temperatures 
shows that they follow the same general 
pattern. 
MEAN WATER TEMPERATURE IN 
DEGREES CENTIGRADE 
Fig. 2. Relationship between mean fin ray count and 
mean water temperature for Istiblennius edentulus. Hollow 
symbols represent data from main band of islands 
crossing Central Pacific (see text for details), solid 
symbols denote data from outlying islands. Regression 
lines calculated for main island band data only. 
Figure 2 depicts the relationship between 
mean ray counts and mean water temperature, 
all figures here being rounded to one decimal 
place. Regression equations have been cal- 
culated for data from the main island band 
(excluding Western and Eastern Hawaii, the 
Philippines, the Line Islands, the Solomons, 
and the East Indies), and are as follows: 
301 
Y = 29.523 — 0.3482 X for the dorsal fin, 
and Y = 29.447 — 0.2781 X for the anal fin. 
Since these two lines are approximately paral- 
lel it may be inferred that temperature has 
about the same effect on each fin. 
It is noteworthy that certain of the points 
(solid symbols on Fig. 2) representing data 
from areas not on the main island band occur 
at considerable distances from the regression 
lines. Fiducial limits have been set for ex- 
pected ray counts at various temperatures 
along the regression lines, using the formula 
Sy = s yx ^| 1/n + x 2 /Sx 2 (Snedecor, 1946: 
120) and values of t at the p .01 level. These 
limits together with the actual ray counts are 
presented in Table 2. 
From Table 2 it is apparent that fin ray 
counts of populations from various remote 
areas differ significantly from the values ex- 
pected on the assumption of a linear rela- 
tionship between ray counts and temperature. 
It is equally apparent that specimens from 
other remote areas have counts which are well 
in line with the expected figures. In the for- 
mer category is Eastern Hawaii (dorsal rays), 
with Western Hawaii (anal rays) and the East 
Indies (dorsal and anal rays) being borderline 
cases. The Line Island data are based on only 
two specimens, and while their ray counts 
are significantly different from the expected 
values, the small sample size precludes em- 
phasizing them. 
DISCUSSION 
It has been shown in Figure 2 that a close 
relationship exists between water temperature 
and fin ray count for I. edentulus . While the 
specimens upon which this conclusion is 
based came from a long chain of closely 
spaced islands it was also apparent (Table 2) 
that specimens from certain remote islands 
fitted quite well into this general picture. The 
Eastern Hawaiian Islands apparently represent 
a geographical region where other factors, 
presumably mutation, have offset the mean 
ray count from the expected values. Western 
