In environmental science, the variability of biological responses in natural or laboratory populations is a well known and documented phenomenon. However, while an extensive literature aims to explain and understand the origin of variability, few try to use it as a demonstration of the population's response facing a stress. We propose here a theoretical framework that explores various patterns of variability both within and among populations, and seeks methods useful in bioevaluation methodologies. We also introduce the concept of “ecotoxicological niche” to characterize the ability of a population to endure contamination.