A re-examination of variation associated with environmentally stressed organisms.

Variation is an essential feature of biological systems. Populations adapt to dynamic environments, in part, because of this variation. In this review, we re-examine phenotypic variation, especially in organisms living in polluted environments. A recent goal of ecotoxicology is to understand the sublethal effects of exposure to pollutants, e.g. responses to endocrine-disrupting contaminants. While variation is an inherent quality of organisms, variance is a statistical measure of the variation of a trait. Increased variance has been associated with organisms living at the perimeter of a population's range, introduced into novel environments, or exposed to pollution. Some researchers have proposed increased phenotypic variance in exposed populations as an evolutionary mechanism, and others have suggested its use as a biomarker. While we agree that variance often increases in the exposed population, we also recognize that the opposite phenomenon occurs. That is, variance can decrease from exposure to pollution. Altered variance in the exposed population-leading to heteroscedasticity-could result in erroneous conclusions (Type II errors). We suggest that exposure to endocrine-disrupting contaminants could influence the health of populations in ways that are not always represented by measures of central tendency, and that variance and distribution should also be examined in environmentally stressed wildlife.