Light influences diverse aspects of human physiology and behaviour including neuroendocrine function, the circadian system and sleep. A role for melanopsin‐expressing intrinsically photosensitive retinal ganglion cells (ipRGCs) in driving such effects is well‐established. However, rod and/or cone signals routed through ipRGCs could also influence ‘non‐visual’ spectral sensitivity. In humans, this has been most extensively studied for acute, light‐dependent, suppression of nocturnal melatonin production. Of the published action spectra for melatonin suppression, one demonstrates a spectral sensitivity consistent with that expected for melanopsin while our own (using briefer 30 min light exposures) displays very high sensitivity to short wavelength light, suggesting a contribution of S‐cones. To clarify that possibility, six healthy young male participants were each exposed to 30 min of five irradiances of 415 nm monochromatic light (1 – 40 µW/cm2) across different nights. These data were then combined with the original action spectrum. The aggregated data are incompatible with the involvement of any single opsin and multi‐opsin models based on the original action spectrum (including Circadian Stimulus) fail to predict the responses to 415 nm stimuli. Instead, the extended action spectrum can be most simply approximated by an ~2:1 combination of melanopsin and S‐cone signals. Such a model also better describes the magnitude of melatonin suppression observed in other studies using an equivalent 30 min mono‐ or polychromatic light paradigm but not those using longer (90 min) light exposures. In sum, these data provide evidence for an initial S‐cone contribution to melatonin suppression that rapidly decays under extended light exposure.