136 
James L. Gooch and Jeffrey S. Wiseman 
completely open and relatively remote from hypogean environments, so 
the Form III morphology should be well developed there. These are the 
maximally contrasting sites and they define the extremes of the Form III 
habitat spectrum. Other localities differ less strikingly and are ranked in 
only an approximate way. 
RESULTS 
The ordinal scale scores of eye packing density, broad lobe orienta- 
tion, and lobe angle of population samples are given by sex in Table 1. 
Localities are in rank order with the most epigean at bottom. Packing 
density is relatively uniform, with most samples distributed in about a 
40:60 ratio between high- and intermediate-density scores. Only a few 
scattered individuals, constituting less than 1% of the total, have the low- 
density irregular borders indicative of the typical Form II ecotype. There 
is a shift from intermediate- to high-density eyes in increasingly open 
populations. James Creek has the highest proportion of regular-eyed in- 
dividuals, although it does not differ significantly from Emma Spring, 
using the R x C Chi-square test, with pooling of low and intermediate 
scores. However, the second to fourth ranked sites have low high-density 
scores and a 2 x 2 contingency table of the four highest ranking popula- 
tions with the five lowest, pooled by site and sex, yields a significant dif- 
ference (X- = 6.53, 1 df, p C.05). Much of the difference is contributed 
by Smoke Hole Spring, whose entire sample consists of intermediate den- 
sity eyes. When sexes pooled over populations are compared by con- 
tingency table there is a significantly greater proportion of high-density 
eyes in females (2C = 9.83, 1 df, p <^.01). We have no hypothesis to ac- 
count for this difference, but since females are smaller it may be related to 
size rather than sex. 
Table 1 shows little interdemic variation in broad lobe orientation or 
lobe angle. All pairwise contingency table tests were performed on the 
sample distributions of both characters, with broad lobe up pooled with 
equal lobe scores. Of 36 tests none indicated significant differences be- 
tween localities for lobe orientation. The overall distribution, to which all 
population samples conform fairly closely, is in the ratio 7:33:60, broad 
lobe up, equal, and down, respectively. This confirms Cole’s (1970) ob- 
servation that the lower lobe is usually broader in G. minus. Although 
lobe orientation is quite variable within populations, interdemic varia- 
tion is too low for this character to be useful in geographic studies. Lobe 
angle is only slightly more variable, with 4 of 36 pairwise tests yielding 
significant differences between sites. All involve the Petersburg I sample, 
which has a high proportion (0.49) of individuals with more obtuse 
angles. The overall ratio is 16:51:33, more acute, intermediate, and more 
obtuse angle, respectively. This character also appears to have little value 
in studies of geographical variation. 
The characters eye, antenna 1, pereopod 7, and uropod 3 lengths will 
