Replicated outdoor experimental streams were exposed to control water and 10 and 30 μg/L selenium (introduced as sodium selenite) to evaluate whether standard laboratory test results accurately predict the effects of selenium on fish in freshwater ecosystems and to evaluate effects found in field studies as reported in the literature. Effects on the survival, growth, and reproduction of bluegill sunfish and the early life stage of their progeny were the end points measured in a 356-d study. Adult bluegills exposed to 10 μg/L had survival reductions (p ≤ 0.05) during the final 98-d period, and those exposed to 30 μg/L experienced complete mortality. Adult growth was reduced at 10 μg/L during an intermediate period (p ≤ 0.05). Exposures of adults at both 30 and 10. μg/L for 40 weeks before spawning resulted in reduced embryo and larval survival and produced larvae with a high incidence of edema, lordosis, and internal hemorrhaging (not all were statistically significant at p ≤ 0.05). This study confirms that standard laboratory tests, which expose fish only through the water pathway, have underestimated the toxicity of selenium, and that 10 μg/L of selenium in the water of a natural ecosystem may adversely affect bluegills, as concluded by other investigators in several field studies.