Reservoirs comprise an expanding proportion of global freshwater resou
rces. New multimetric approaches to biological monitoring, such as the
index of biological integrity, have been useful in streams; similar a
pproaches in reservoirs might aid managers concerned about the reservo
irs they manage. Electrofishing data from Tennessee Valley Authority r
eservoirs were used to evaluate the applicability of biomonitoring to
those reservoirs. Individual metrics and overall index scores exhibite
d substantial annual variation. Within-year variation in scores from a
supplemental data set suggests that annual changes in scores may not
represent real fluctuations in resource condition. Species accumulatio
n curves demonstrate that standard samples consisting of ten timed run
s were not adequate; the sample error was high. Some metrics are usefu
l, but they require more testing; other metrics are clearly not approp
riate for use in a integrative multimetric index. Finally, the reservo
irs included in this study may be similarly degraded; independent data
from separate regions and watersheds are needed to resolve this and o
ther issues. Although attaching qualitative judgements (e.g. excellent
to poor) is premature, the concept of a reservoir biomonitoring index
deserves more study, especially the development of more comprehensive
and systematic sampling programmes designed to resolve sample adequac
y and sources of variation in sample data.