Environmental pollution is currently identified as one of the major drivers of rapid decline of insect populations, and this finding has revitalized interest in insect responses to pollution. We tested the hypothesis that the pollution-induced decline of insect populations can be predicted from phenotypic stress responses expressed as morphological differences between populations inhabiting polluted and unpolluted sites. We explored populations of the brassy tortrix Eulia ministrana in subarctic forests along an environmental disturbance gradient created by long-lasting severe impacts of aerial emissions of the copper-nickel smelter in Monchegorsk, northwestern Russia. We used pheromone traps to measure the population densities of this leafrolling moth and to collect specimens for assessment of three morphological stress indices: size, forewing melanization, and fluctuating asymmetry in wing venation. Wing length of E. ministrana increased by 10%, and neither forewing melanization nor fluctuating asymmetry changed from the unpolluted forest to the heavily polluted industrial barren. However, the population density of E. ministrana decreased 5 to 10 fold in the same pollution gradient. Thus, none of the studied potential morphological stress indicators signaled vulnerability of E. ministrana to environmental pollution and/or to pollution-induced environmental disturbance. We conclude that insect populations can decline without any visible signs of stress. The use of morphological proxies of insect fitness to predict the consequences of human impact on insect populations is therefore risky until causal relationships between these proxies and insect abundance are deciphered.