Abstract Recent studies have both shown and predicted that global climate change will have a substantial influence on biodiversity. This is true especially of a global biodiversity hotspot, the Cape Floristic Region. Although the effects of predicted changes have been widely assessed for plants, little is known about how insect diversity in the region might be affected. In particular, patterns in and the correlates of diversity in the region are poorly understood, and therefore the likely affects of a changing abiotic environment on this significant group of organisms are not clear. Therefore, we investigate patterns in, and correlates of, epigaeic beetle (Tenebrionidae and Carabidae) diversity in one of the most climate change-sensitive areas in the Cape Floristic Region, the Cederberg. In particular, we determine whether epigaeic beetle assemblage structure differs between the main vegetation types in the Cederberg (Strandveld, Mountain Fynbos and Succulent Karoo), how restricted these beetles are to specific vegetation types, and which environmental variables might be associated with site-related differences in beetle richness and abundance. Sampling was undertaken during October 2002 and 2003 across an altitudinal gradient ranging from sea level (Lambert's Bay) to approximately 2000 m above sea level (Sneeukop, Cederberg) and down again to 500 m above sea level (Wupperthal) using pitfall traps. The environmental correlates of abundance and species density in the epigaeic beetles were similar to those identified previously for ants across the transect, with both taxa being positively related to several temperature variables. Several species showed habitat specificity and fidelity, and clear distinctions existed between the vegetation types across the transect. A larger proportion of the variance in tenebrionid species density was explained by environmental variables and spatial factors than for carabids. The most likely explanation for this difference is that the correlates might well reflect collinear historical processes, rather than a causal relationship between contemporary environmental variables and species density. If this is the case, it suggests that caution should be exercised when interpreting environmental correlates of species density, and making climate change predictions based on these correlates.