Small, isolated populations at the edge of a species’ range (peripheral isolates) are potential hotbeds of adaptive diversity under climate change, and an interesting experimental systems in which to answer fundamental questions in ecology and evolution. We aim at exploring the genetic basis of potentially adaptive traits under a climate change scenario using as model the common sunskink Lampropholis coggeri as this animal is a well-developed laboratory model, has a well-studied phylogeographic history and ectotherms are considered to be particularly vulnerable to climate change due to their narrow tolerance ranges to environmental temperatures. Using DArT technology, we generated nearly 18000 SNPs randomly distributed in the genome in a captive breeding population previously genotyped for morphological and physiological traits. We used this data to assign paternity to individuals. We also calculated heritability of measured traits using mixed models. Our analyses suggest that a simple model without maternal is a better fit to the data. Many important physiological traits related to thermal tolerance and performance have significant differences between populations that are mantained after acclimation and are also observed on F1. Many of these traits have low heritability; these results matches Fisher's theorem that traits under strong selection have low narrow-sense heritability because of low variance in phenotypes. We conclude that the two populations studied show important phenotypic differences that are not only heritable but are under strong selection.