The study of polymorphic organisms in which sexual selection strongly drives the maintenance of the polymorphism often ignores the environmental context. However, natural selection is also crucial, often interacting with sexual selection. While most evidence suggests that sexual selection underlies the maintenance of the colour polymorphism shown by the common wall lizard, Podarcis muralis, involving up to five sympatric morphs (white -W-, yellow -Y-, orange -O-, white-orange -WO- and yellow-orange -YO), recent results using realised niche models suggest morph distribution at a geographical scale is constrained by climate. The Y and YO morphs occupy a narrow space within the niche of the other morphs, and local frequencies of W, WO and O morphs are environmentally dependent, with O and WO showing higher local frequencies in the most humid habitats. We hypothesised that morphs may behaviourally compensate for these differences at a microhabitat scale. Here, we tested this hypothesis collecting field microhabitat data from representative natural populations. Results highlighted that O lizards are partially segregated in microhabitat relative to the other morphs, tending to occupy more humid (vegetated, close to water) sites, and suggesting a physiological constraint. In contrast, the other morphs do not differ in microhabitat use, suggesting that the ecological restriction of Y and YO morphs derives from an indirect relationship between climate and population parameters (sex ratio, density) crucial for social selection, eventually involving alternative behavioural strategies. The maintenance of different colour phenotypes (morphs) in a single population occurs in complex evolutionary scenarios where several selective forces interact. Common wall lizards may be white, orange, yellow, white-orange or yellow-orange ventrally. This polymorphism is most likely driven by sexual selection, but climate probably constraints the environmental distribution of morphs, driving local morph diversity. Here, we explore if geographical differences in morph distribution result in microhabitat segregation. Our results show that orange lizards appear more frequently in humid habitats than the other morphs, suggesting a direct relationship between ecophysiology and habitat use. Although yellow and yellow-orange morphs present a narrow distribution at a geographical scale, they do not differ from the white and white-orange morphs in microhabitat, suggesting that their restricted distribution is indirectly caused by some source of environmental dependence on social selection instead of a direct physiological constraint.