Indigenous Salvia species from southern Africa are popular traditional medicines for the treatment of a variety of conditions. They produce fragrant volatiles that can be isolated as essential oils. Some of these volatile organic compounds may play a role in the biological activities of the extracts. Three indigenous Salvia species, Salvia africana-lutea, S. lanceolata and S. chamelaeagnea, were selected for this study as they are commonly used in traditional medicine in South Africa, and the essential oils from these species have potential for commercialisation. Although some studies have described the essential oil compositions and some biological activities, only single composite samples were used. The aim of this study was to investigate the intra- and interspecies variation of the essential oils, sampled over a wide geographical area and using a representative sample size, to encourage commercialisation of the essential oil. Essential oils were isolated from individual plants using conventional hydrodistillation of the aerial parts, harvested from several localities. Gas chromatography coupled simultaneously to mass spectrometry/flame ionisation detection (GC-MS/FID) was used to identify and quantify the volatile constituents. The essential oils of S. africana-lutea consisted mainly of terpinene-4-ol + β-caryophyllene (1.4 - 29.0%), T-cadinol (1.2 - 20.0%), α-eudesmol (trace - 23.0%) and β-eudesmol (trace - 26.0%), those of S. lanceolata comprised mainly terpinene-4-ol + β-caryophyllene (4.3 - 31.0%), α-humulene (2.3 - 15.0%), bicyclogermacrene (trace - 37.0%) and spathulenol (trace - 25.0%), while the essential oils of S. chamelaeagnea were characterised by δ-3-carene (trace - 18.0%), limonene (1.6 - 36.0%), viridiflorol (9.8 - 61.0%) and 1,8-cineole (not detected - 11.0%). The compounds identified in the essential oils of the three selected Salvia species have been identified in other Salvia essential oils. To add to the novelty of this study, the superior resolving power of two-dimensional gas chromatography was demonstrated through analysis of selected essential oils. Many additional compounds were identified, and previously co-eluting compounds were clearly separated. Chemometric modelling of the GC-MS data using SIMCA P+ 14 software allowed distinct clustering patterns to be discerned. The unsupervised principal component analysis model revealed separate clusters for the three species, confirming substantial chemical differences between their essential oils. Quantitative, rather than qualitative differences were evident when individual essential oil samples representing the same species, were compared. For each species, two chemically distinct groups were observed and unique marker compounds could be identified. This study has contributed detailed information on the major and minor volatile compounds present in the essential oils of the three Salvia species investigated.