1) Phenotypic variation provides the framework for natural selection to work upon, enabling adaptive evolution. One of the most discernible manifestations of phenotypic variability is colour variation. When this variation is discrete, genetically-based colour pattern morphs occur simultaneously within a population. 2) Why and how colour polymorphisms are maintained is an evolutionary puzzle. Several evolutionary drivers have been hypothesized as influencing clinal patterns of morph frequency, with spatial variation in climate and predation being considered especially important. Despite this, no study has examined both of their roles simultaneously. The aims of this study were to: 1) examine the covariation of physiology, environmental variables, and colouration at a local scale; and 2) determine if these factors and their interplay explain broad clinal variation in morph frequency. 3) We used the lizard Liopholis whitii as a model system, as this species displays a discrete, heritable polymorphism for colour pattern (plain-backed, patterned morphs) whose morph frequency varies latitudinally. We measured reflectance, field activity temperatures, and microhabitat structure to test for differences in crypsis, thermal biology, and microhabitat selection of patterned and plain-backed morphs within a single population where colour morphs occur sympatrically. We then used data from the literature to perform a broad-scale analysis to identify whether these factors also explained the latitudinal variation of morph frequency in this species. 4) At the local scale, plain-backed morphs were found to be less cryptic than patterned morphs while no other differences were detected in terms of thermal biology, dorsal reflectance, and microhabitat use. At a broader scale, predation was the most influential factor mediating morph frequency across latitudes. However, the observed pattern of morph frequency is opposite to what the modelling results suggest in that the incidence of the least cryptic morph is highest where predation pressure is most severe. Clinal variation in the level of background matching between morphs or the potential reproductive advantage by the plain-backed morph may, instead, be driving the observed morph frequency. 5) Together, these results provide key insights into the evolution of local adaptation as well as the ecological forces involved in driving the dynamics of colour polymorphism.