The 'colour-shading effect' describes the phenomenon whereby chromatic variations affect the magnitude of perceived shape-from-shading in luminance patterns. A previous study showed that in mixed colour-plus-luminance sine-wave plaids, impressions of depth in the luminance component were enhanced by non-aligned chromatic components, and suppressed by aligned chromatic components [Nature Neuroscience 6 (2003) 641-644]. Here we examine the chromatic determinants of these effects. Colour contrast was defined along the cardinal axes of colour space in order to isolate the L-M and S-(L+M) post-receptoral chromatic mechanisms. We found no difference in the potency of L-M-only and S-(L+M)-only gratings, either for enhancing or suppressing perceived depth. Moreover, the magnitude of depth-suppression was no different for any combination of depth-enhancing and depth-suppressing cardinal directions. Finally we tested whether the visual system carried the assumption that natural shading is tinged with blue, by measuring perceived depth in a colour-plus-luminance grating that was made to appear either bright-yellow/dark-blue or bright-blue/dark-yellow. However there was no difference in the magnitude of depth-suppression between conditions, suggesting that the visual system does not make any assumption about the colour of natural shading. Taken together, the results suggest that while the colour-shading effect is highly sensitive to colour contrast, it is agnostic with respect to colour direction.