General linear models (GLMs) were used to determine the relative importance of interspecific, ontogenetic and spatial effects in explaining variability in stable isotope ratios of nitrogen (δ15N) and carbon (δ13C) of the co‐occurring Cape hakes Merluccius capensis and Merluccius paradoxus off South Africa. Significant GLMs were derived for both isotopes, explaining 74 and 56% of observed variance in Merluccius spp. δ15N and δ13C, respectively. Spatial effects (west or south coast) contributed most towards explaining variability in the δ15N model, with Merluccius spp. off the west coast having higher (by c. 1·4‰) δ15N levels than Merluccius spp. off the south coast. Fish size and species were also significant in explaining variability in δ15N, with both species showing significant linear increases in δ15N with size and M. capensis having higher (by c. 0·7‰) δ15N values than M. paradoxus. Species and coast explained most and similar amounts of variability in the δ13C model, with M. capensis having higher (by c. 0·8‰) δ13C values than M. paradoxus, and values being lower (by c. 0·7‰) for fishes off the west coast compared with the south coast. These results not only corroborate the knowledge of Merluccius spp. feeding ecology gained from dietary studies, in particular the ontogenetic change in trophic level corresponding to a changing diet, but also that M. capensis feeds at a slightly higher trophic level than M. paradoxus. The spatial difference in Merluccius spp. δ15N appears due to a difference in isotopic baseline, and not as a result of Merluccius spp. feeding higher in the food web off the west than the south coast, and provides new evidence that corroborates previous observations of biogeographic differences in isotopic baselines around the South African coast. This study also provides quantitative data on the relative trophic level and trophic width of Cape hakes over a large size range that can be used in ecosystem models of the southern Benguela.