European starlings (Sturnus vulgaris) exhibit seasonal changes in singing and in the volumes of the neural substrate. Increases in song nuclei volume are mediated at least in part by increases in day length, which is also associated with increases in plasma testosterone (T), reproductive activity, and singing behavior in males. The correlations between photoperiod (i.e. daylength), T, reproductive state and singing hamper our ability to disentangle causal relationships. We investigated how photoperiodic-induced variation in reproductive state modulates the effects of T on singing behavior and song nuclei volumes in adult female starlings. Female starlings do not naturally produce measureable levels of circulating T but nevertheless respond to exogenous T, which induces male-like singing. We manipulated photoperiod by placing birds in a photosensitive or photorefractory state and then treated them with T-filled or empty silastic implants. We recorded morning singing behavior for 3 weeks, after which we assessed reproductive condition and measured song nuclei volumes. We found that T-treated photosensitive birds sang significantly more than all other groups including T-treated photorefractory birds. All T-treated birds had larger song nuclei volumes than with blank-treated birds (despite photorefractory T-treated birds not increasing song-rate). There was no effect of photoperiod on the song nuclei volumes of T-treated birds. These data show that the behavioral effects of exogenous T can be modulated by reproductive state in adult female songbirds. Furthermore, these data are consistent with other observations that increases in singing rate in response to T are not necessarily due to the direct effects of T on song nuclei volume.