A thesis submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Doctor of Philosophy, 2017 / Under current and previous global climate change, environments are changing and have changed at a rapid rate. Species with the potential to undergo adaptive radiation are likely to survive environmental change. The genus Rhabdomys is widespread in southern Africa, occurring along the east-west rainfall gradient in South Africa. Rhabdomys may have undergone adaptive radiations in the past, which may have resulted in the current suite of species in various habitats of different aridity. Some Rhabdomys species also occur in sympatry in some locations in South Africa. The aim of my study was to investigate adaptive variation in Rhabdomys by studying the behaviour of 5 populations, representing 3 Rhabdomys species, across South Africa. Using selected taxa, my approach was, firstly, to describe variation in two traits, personality and spatial cogntion, well known for showing environmentally-linked (i.e. adaptive) variation. Secondly, I manipulated the development of exploratory and anxiety behaviour to assess the limits of the adaptive variation (i.e. test the nature of the reaction norm of the characters measured). I first established the taxon-level personality of 4 taxa (2 sympatric) in 5 standard behavioural tests. Generally, the semi-desert living R. pumilio was the boldest together, surprisingly, with R. d. dilectus occurring in grasslands of central South Africa, contradicting previously published results. Comparatively, R. bechuanae from central South Africa and R. dilectus from far north-eastern South Africa, also occurring in grasslands were less bold, even though R. bechuanae is sympatric with R. dilectus in central South Africa. My data indicate adaptive variation at the extreme populations and possibly character displacement in the sympatric populations. In the next chapter, I investigated whether early rearing environment shapes exploratory behaviour and anxiety responses of R. pumilio and R. bechuanae. I predicted that using an interspecies cross-fostering protocol would reveal a gene x environment interaction on behaviour, so that fostered offspring would display an intermediate behaviour phenotype compared to their non-fostered siblings. I showed that a novel rearing environment mostly did not influence the adult behaviour of cross-fostered inidividuals. This indicates genetic constraints on exploratory behaviour and anxiety responses. Next, I tested whether physical rearing environment shapes exploratory behaviour and anxiety responses. I reared semidesert R. pumilio, sympatric R. bechuanae and R. dilectus and allopatric R. bechuanae under either no cover or high cover for 2 generations. The taxa were mostly similar and altering the phyical housing condition did not alter behaviour, but there were small differences between the taxa in exploratory behaviour. In the final experimental chapter, I established whether the environment predicts the spatial cognition in semi-desert R. pumilio, sympatric R. bechuanae and R. dilectus and an allopatric population of R. dilectus from far north-eastern South Africa. The populations showed very similar performance in a modified Barnes maze, indicating a possible phylogenetic constraint on spatial cognition. Overall, my study suggests that there is adaptive variation in personality but not spatial cognition. In contrast to previous studies in the genus, alterations to the social and physical environments failed to separate out genetic and environmental effects (i.e. reaction norm) that would potentially provide the mechanisms for adaptive variation within and between species. The similarity in spatial cognition between taxa and similar responses to environmental modification indicate phylogenetic constraints on traits that were predicted to vary geographically. / XL2018