Per-7/body Ant-l/body 
Gammarus Variability 
141 
Table 5. Pairwise comparisons in females. Captions as in table 4, 
Eye/body 
EM 
SH 
GR 
CU 
MK 
CC 
P-1 
P-II 
JC 
EM 
* 
** 
** 
SH 
- 
** 
- 
- 
- 
GR 
** 
** 
- 
** 
CU 
** 
- 
** 
MK 
- 
- 
- 
- 
- 
CC 
- 
♦ 
* 
- 
- 
♦ 
- 
P-I 
** 
- 
- 
- 
- 
P-II 
** 
- 
- 
- 
- 
- 
- 
JC 
* 
EM 
SH 
GR 
Uro-3/body 
CU MK 
CC 
P-I 
P-II 
JC 
EM 
- 
- 
* 
- 
- 
* 
- 
- 
SH 
- 
- 
- 
- 
- 
- 
GR 
** 
- 
- 
* 
* 
* 
CU 
- 
- 
- 
♦ 
♦ ♦ 
- 
MK 
* 
- 
- 
- 
CC 
* 
♦ 
- 
* 
- 
- 
- 
- 
P-I 
** 
- 
** 
- 
- 
- 
** 
P-II 
- 
♦ 
- 
* 
- 
- 
♦ 
** 
JC 
- 
- 
- 
* 
- 
- 
ters in pairwise locality comparisons using the Mann-Whitney V test. 
The results for males are shown in Table 4 and for females in Table 5. In- 
terdemic variation in length ratios is extensive. Of 288 pairwise tests, 144 
per sex, 166 or 58% show significant (p <C.05) or very significant (/?< 
.01) differences between localities. Variation is greater in males (65% of 
tests) than females (50%), which may be related to size rather than sex. 
Among characters with sexes pooled, interdemic differences are greatest 
in eye length ratio (51 of 72 comparisons, 71%) and are at roughly the 
same level for appendage ratios (antenna 1, 56%; pereopod 7, 57%; 
uropod 3, 47%). Some of this variation is due to comparison of samples 
from strongly contrasting habitats at opposite ends of the ranking. Some 
