To whom correspondence should be addressed at present address: De- partment of Biology, Laurentian University, Ramsey Lake Rd., Sudbury, ON P3E 2C6, Canada. Fax: (705) 675-4859. E-mail: gpyle@nickel. laurentian.ca. Ecotoxicology and Environmental Safety 48, 202 } 214 (2001) Environmental Research, Section B doi:10.1006/eesa.2000.2016, available online at http://www.idealibrary.com on Toxicity of Uranium Mine-Receiving Waters to Caged Fathead Minnows, Pimephales promelas G. G. Pyle,* S. M. Swanson,- and D. M. Lehmkuhl* *Department of Biology, University of Saskatchewan, 112 Science Place, Saskatoon, Saskatchewan S7N 5E2, Canada; and -Golder Associates, 10th Floor, 940-6th Avenue, S. W., Calgary, Alberta T2P 3T1, Canada Received May 15, 2000 Larval fathead minnows (Pimephales promelas) were placed at four exposure sites for 7 days in each of 5ve lakes surrounding the Key Lake uranium mine in northern Saskatchewan, Canada. Fish placed in lakes receiving Mo-contaminated mill e8uent demonstrated higher mortalities than those placed in lakes re- ceiving Ni-contaminated mine-dewatering e8uent, which was not signi5cantly di4erent from reference sites. No signi5cant di4erences were detected in 5sh growth among the study lakes because of the high (90%) mortality in Fox and Unknown lakes. Principal components analysis characterized exposure sites by total and dissolved metal concentration. Stepwise multiple re- gression of 5sh mortality on principal components generated from total metal data revealed that principal component 1 could account for 84% of the variance associated with 5sh mortality. Careful examination of the metals that correlated strongly with principal component 1 and with 5sh mortality suggested that dietary Se toxicity probably resulted in the di4erential fathead minnow mortality observed among study lakes.  2001 Academic Press Key Words: 5sh; in situ toxicity testing; larval fathead min- nows; Pimephales promelas; principal components analysis; ura- nium mine; dietary metal toxicity; nickel; molybdenum; selenium. INTRODUCTION Northern Saskatchewan, Canada, is home to some of the top-producing uranium mines in the world. In parti- cular, Cameco's Key Lake uranium operation is the largest uranium mine in the world, with average annual U  O  production exceeding 510 kg (Saskatchewan Energy and Mines, 1999). During operation, Key Lake uranium ore is extracted from an open-pit mine and is milled on site. Treated pit-dewatering and mill e%uents are released to surrounding aquatic systems, causing an elevation of metal concentrations in receiving waters above background con- centrations (Cameco Corp. et al., 1995). Mill e%uent has previously been characterized as having high concentra- tions of molybdenum (Mo), whereas dewatering e%uent has high concentrations of nickel (Ni) (Cameco Corp. et al., 1995). Consequently, both Ni and Mo have been identi"ed as &&metals of concern'' in aquatic receiving systems around Key Lake. Traditionally, environmental assessments of mine-receiv- ing waters have made use of univariate sample-characteriza- tion techniques to identify potential toxicants to indigenous biota (Maund et al., 1999). Organisms occurring in indus- trially contaminated natural systems are rarely, if ever, exposed to single contaminants. Instead, they are exposed to complex mixtures of contaminants. Attempting to relate biological e!ects to single contaminants in a complex system through univariate analysis is di$cult and often unrealistic (Maund et al., 1999). Ecotoxicologists have recently recognized this short- coming and have begun using multivariate techniques to examine several environmental variables simultaneously as a more realistic approach to relating biological e!ects to complex mixtures (Maund et al., 1999). Multivariate tech- niques allow variables from a data set containing many variables to be analyzed simultaneously, rather than indi- vidually as in univariate methods. In this way, mutual variance among variables can be investigated allowing for a more realistic characterization of the system under study. Although such analyses are limited by their inability to clearly identify causal relationships, they serve as powerful heuristic tools. In this study, larval fathead minnows (Pimephales promelas) were placed in metal-contaminated and reference lakes for 7 days, after which growth and mortality were observed. Each exposure site per lake was characterized by both total and dissolved metal concentrations, which were ordinated by principal components analysis (PCA). The "rst 202 0147-6513/01 $35.00 Copyright  2001 by Academic Press All rights of reproduction in any form reserved.