Gammarus Variability 
145 
morphological profile of the common epigean ecotype was done, ex- 
tending the pioneering work of Holsinger and Culver to additional pop- 
ulations, both sexes, and larger sample sizes. Second, the level of inter- 
demic variability was determined and variability was related to the 
epigean-hypogean habitat spectrum and to other factors. Last, the 
overall pattern of interdemic variation was evaluated, which also may be 
applicable to other freshwater species. 
All characters displayed high intrademic variability, and eye density 
and length ratios of eye and appendages also varied significantly among 
demes. Lobe orientation and lobe angle scores were distributed quite 
uniformly over populations and thus have less value in geographical 
studies in the Appalachians. Growth of appendages was approximately 
linear on body length in the adult size range, without significant dif- 
ferences in slope between sexes or large differences in intercepts. Growth 
equations were different enough among populations, however, to give 
coefficients of determination usually less than 0.70 on locality mean 
lengths. Female amphipods averaged 73% the length of males, but 
proportionate size of sexes varied widely. We have no evidence as to 
genetic or directly environmental causes of size differences. Females also 
had statistically significant smaller appendage length ratios than males. 
Interdemic character variation was both systematic, that is habitat 
scale related, and nonsystematic. The former is evidenced by the finding 
that Form III demes in shaded, secluded springs near cave or conduit dis- 
charge usually had more irregular eyes and more elongate appendages 
than more open populations. At Emma Spring and Cunninghams, at 
least, there are no known Form II populations. This suggests that 
morphology is an in situ adaptation and not the result of genetic mixing 
with cave ecotypes. These intergradational populations would be ex- 
pected to possess slightly reduced eyes as well. Although we have in- 
cluded eye length ratio in the composite character scale there is no clear 
trend toward eye reduction (Table 3). We conclude that ecotypical 
characters are intergradational among epigean demes and that the 
characters rank roughly in the same order as habitats scaled from fully 
epigean to semihypogean. 
The conformation of character and habitat rankings is very inexact. 
Aside from sampling errors and imprecision in assigning habitat rank, 
there is extensive nonsystematic and statistically significant interdemic 
variability. Some of this variability may actually be systematic adapta- 
tion to habitat variables not apparent to the observer. Minckley and 
Cole (1963) found variation, mostly of setation patterns, in G. minus in 
two Kentucky streams to be associated with lotic and lentic microen- 
vironments, aquatic vegetation, and the presence of the probable com- 
petitor Gammarus bousfieldi. Dickson (1977) noted that type and amount 
of food influenced pigmentation and body and antenna length in the 
troglobitic species Crangonyx antennatus. 
A baseline datum in the present study is the existence of extensive 
