Subjective evaluations of brightness contrast were obtained for Munsell chips of varying hues, lightness values and sizes against a white background illuminated in turn by four white illuminants at four background luminances in the mesopic region. Chip lightness was the strongest variable, resulting in a family of monotonic functions relating brightness contrast ratings to chip lightness. At the lowest light levels a nearly linear relationship was found between chip lightness and subjective ratings. As light levels increased, the relationship became curvilinear such that at the lowest contrast, chromatic brightness became important for subjective ratings. Within the level of precision for this experiment, chip hue and size as well as the spectral power distribution of the illuminants did not significantly impact the overall brightness contrast ratings. The subjective ratings of brightness contrast were analysed as a function of the luminance contrast of the stimuli as determined by conventional photometry and by three models of mesopic vision. The response functions relating subjective rating to the stimulus characterized in terms of these models were similar to those obtained with conventional photometry.