Bud development of boreal trees in spring, once initiated, is driven by ambient air temperature, but the mechanism triggering bud development remains unclear. We determined if some aspect of the diurnal or seasonal light regime influences initiation of bud burst once the chilling requirement is met. We grew 3-year-old birch plantlets cloned from a mature tree of boreal origin in light conditions realistically simulating the lengthening days of spring at 60 degrees N. To emulate the reduction in red to far-red light (R:FR) ratio between daylight and twilight, one group of plantlets was subjected to reduced R:FR ratio in the morning and evening in addition to progressively lengthening days, whereas the other group was subjected to the same R:FR ratio throughout the day. The reduced R:FR ratio of twilight advanced bud burst by 4 days compared with the reference group (P = 0.04). To assess the interplay between the fulfillment of the chilling requirement and the subsequent response to warming, we fitted a thermal time model to the data with separate parameterizations for the starting dates of heat sum accumulation in each treatment. Least-squares fitting suggested that bud development started in light regimes corresponding to late March, almost two months after the chilling requirement for dormancy release was satisfied. Therefore, shortening night length or increasing day length, or both, appears to be the cue enabling bud development in spring, with twilight quality having an effect on the photoperiodic response. If twilight alone were the cue, the difference in bud burst dates between the experimental groups would have been greater than 4 days. The result gives experimental support for the use of thermal-time models in phenological modeling.