50 J. N. Am. Benthol. Soc., 2000, 19(1):50–67  2000 by The North American Benthological Society Use of periphyton assemblage data as an index of biotic integrity B. H. HILL 1,5 , A. T. HERLIHY 2 , P. R. KAUFMANN 2,6 , R. J. STEVENSON 3,7 , F. H. MCCORMICK 1 , AND C. BURCH JOHNSON 4 1 National Exposure Research Laboratory, US Environmental Protection Agency, 26 W. Martin Luther King Drive, Cincinnati, Ohio 45268 USA 2 Department of Fisheries and Wildlife, Oregon State University,  US Environmental Protection Agency, 200 SW 35th Street, Corvallis, Oregon 97333 USA 3 Department of Biology, University of Louisville, Louisville, Kentucky 40292 USA 4 OAO Corporation,  US Environmental Protection Agency, 200 SW 35th Street, Corvallis, Oregon 97333 USA Abstract. Periphyton assemblage data collected from 233 stream site-visits (49 in 1993, 56 in 1994, and 128 in 1995) throughout the Mid-Appalachian region were used to develop a periphyton index of biotic integrity (PIBI) based on 1) algal genera richness; 2) the relative abundances of diatoms, Cyanobacteria, dominant diatom genus, acidophilic diatoms, eutraphentic diatoms, and motile dia- toms; 3) chlorophyll and biomass (ash-free dry mass) standing crops; and 4) alkaline phosphatase activity. Thirty-seven diatom genera and 38 non-diatom genera were collected. The relative richness and relative abundance (RA) of these genera were used to calculate the RA metrics of the PIBI. PIBI scores ranged from 48.0 to 85.1 among the 233 site-visits with an overall regional mean (1 SE) of 66.1  0.5. The 10 metrics and the PIBI were correlated with 27 chemical, 12 physical habitat, and 3 landscape variables. Overall, PIBI was inversely correlated with stream depth, stream water color, and Fe. Component metrics were significantly correlated with several chemical (Al, acid neutralizing capacity, Cl, Fe, Mn, N, Na, P, pH, Si, SO 4 , total suspended solids), physical habitat (channel embed- dedness, riparian disturbances, stream depth, stream width, substrate composition), and landscape (% of the watershed in forest, agriculture, and urban land uses) variables. Canonical correlation analysis revealed significant correlations between the 10 PIBI metrics and 4 significant environmental gradients related to general human disturbances (stream acidity, stream substrate composition, and stream and riparian habitat). Analysis of variance revealed significant differences in PIBI scores for lowland vs highland streams, and among stream orders. Annual differences were explained by dif- ferences in the proportions of sampling sites in lowland streams in each year. The univariate distri- bution of PIBI scores was used to set threshold PIBI values for the assessment of ecological condition in Mid-Appalachian streams. Key words: periphyton, IBI, EMAP, regional scale, stream condition, multimetrics. The use of biological communities as indica- tors of water quality is evolving as our under- standing of the interactions between water qual- ity and the integrity of biological communities improves (Karr et al. 1986, Hughes et al. 1991, Dixit et al. 1992). Stream periphyton assemblag- es, although not widely used, are ideally suited for water-quality assessment (Patrick 1973, Ste- venson and Lowe 1986, Rott 1991, Round 1991, van Dam et al. 1994, Stevenson and Pan 1999). 5 E-mail address: hill.brian@epa.gov 6 Present address: National Health and Environmen- tal Effects Laboratory, US Environmental Protection Agency, 200 SW 35 th Street, Corvallis, Oregon 97333 USA. 7 Present address: Department of Zoology, Michi- gan State University, East Lansing, Michigan 48824 USA. Analysis of periphyton assemblages may focus on taxonomic or non-taxonomic features. Taxo- nomic descriptors (e.g., diversity indices, taxa richness, indicator species) are commonly used, and are described by several researchers (e.g., Patrick 1973, Palmer 1977, Rodgers et al. 1979, Weitzel 1979). Non-taxonomic measures (e.g., biomass and chlorophyll per unit area) can also be useful for detecting effects not indicated by taxonomic analysis. For example, toxic pollut- ants may cause sublethal (i.e., reproductive) ef- fects that would not immediately be detected by taxonomic descriptors such as taxa richness, but might be indicated by changed chlorophyll con- tent, or enzyme activity. A summary of non- taxonomic measurements is presented in Weitz- el (1979). There have been 3 basic approaches to using