The sediment, water, and three species of fish from 24 of Massachusetts' (r
elatively) least-impacted water bodies were sampled to determine the patter
ns of variation in edible tissue mercury concentrations and the relationshi
ps of these patterns to characteristics of the water, sediment, and water b
odies (lake, wetland, and watershed areas). Sampling was apportioned among
three different ecological subregions and among lakes of differing trophic
status. We sought to partition the variance to discover if these broadly de
fined concepts are suitable predictors of mercury levels in fish. Average m
uscle mercury concentrations were 0.15 mg/kg wet weight in the bottom-feedi
ng brown bullheads (Ameriurus nebulosus) (range = 0.01-0.79 mg/kg); 0.31 mg
/kg in the omnivorous yellow perch (Perca flavescens) (range = 0.01-0.75 mg
/kg); and 0.39 mg/kg in the predaceous largemouth bass (Micropterus salmoid
es) (range = 0.05-1.1 mg/kg). Statistically significant differences in fish
mercury concentrations between ecological subregions in Massachusetts, USA
, existed only in yellow perch. The productivity level of the lakes (as ded
uced from Carlson's Trophic Status Index) was not a strong predictor of tis
sue mercury concentrations in any species. pH was a highly (inversely) corr
elated environmental variable with yellow perch and brown bullhead tissue m
ercury. Largemouth bass tissue mercury concentrations were most highly corr
elated with the weight of the fish (+), lake size (+), and source area size
s (+). Properties of individual lakes appear more important for determining
fish tissue mercury concentrations than do small scale ecoregional differe
nces. Species that show major mercury variation with size or trophic level
may not be good choices for use in evaluating the importance of environment
al variables.