Responses to sublethal concentrations of mercury were determined and compared for several populations of the mayfly, Isonychia bicolor from sites on two mercury polluted rivers in Virginia (USA). Results from laboratory respiration experiments indicated that small nymphs were more sensitive to mercury than larger nymphs, and that sensitivity increased with temperature. In addition, mayfly nymphs from polluted sites on the less contaminated South River were found to be more resistant to sublethal doses of inorganic mercury than nymphs from a reference site as indicated by smaller changes in respiration following exposure. Results of preexposure treatments suggested a genetic basis for the observed tolerance. No evidence of tolerance was observed in nymphs from the more contaminated Holston River. Field surveys revealed no significant differences in age structure, growth, or survival between populations at reference and polluted sites on the South River. However, growth and survival of nymphs from the polluted site on the Holston River were significantly slower than at the reference site during the summer. In fact, shortly after the summer generation hatched, nymphal density at contaminated sites declined to zero. This apparent local extinction may have been due to increased toxicity of mercury at warmer temperatures. Results of reciprocal transplant experiments conducted at sites on the South River suggested that population differences in maintenance costs observed in short-term laboratory experiments, may have been reflected by population differences in fecundity in long-term field experiments. When nymphs from the reference site on South River were transplanted and allowed to develop at a contaminated site, lower fecundities were observed than for the native population. No population differences in fecundities were observed on the Holston River. The relationship between allozyme genotype to survival of I. bicolor nymphs to acute mercury exposure was also tested. The probability of survival and individual times to death (TTD) were found to be significantly different among genotypes at the Glucose Phosphate Isomerase (GPI) locus. This was true at both summer and winter temperatures and for populations from two separate, unpolluted streams. However, genotypes identified as sensitive and tolerant in these experiments showed no consistent relationship with environmental mercury levels in polluted rivers. Therefore, the use of allozyme variants as a biomarker to assess evolutionary change in populations due to mercury pollution may be impractical in freshwater systems. / Ph. D.