This study investigated Lewy's Phase Shift Hypothesis (PSH) for winter Seasonal Affective Disorder, which asserts that the phase angle difference (PAD) between circadian and sleep rhythms is critical in the mechanism of light's therapeutic action. Specifically, we sought to test whether a euthymic 'sweet spot' could be identified at a PAD (between temperature minimum and wake time) of circa 3 h. After a baseline week, symptomatic SAD patients (N = 43) received 8 weeks of morning light treatment. Analyses were based on SIGH-SAD ratings made at baseline and posttreatment. Also estimated pre- and posttreatment were Tmin (calculated from an algorithm based on Morningness-Eveningness self-report scores), and the phase of the sleep-wake rhythm (as assessed by daily sleep logs). It was predicted that a quadratic relationship would exist between PAD and depression ratings at baseline and posttreatment, with lowest levels around PAD = 3 h. It was further predicted that shift towards PAD = 3 h with treatment would be associated with decreases in depression with treatment. Although trends were in the expected direction, none of the three predictions were supported. Findings are discussed in terms of the study's limitations and the experimental challenge of parsing independent and interacting contributions of sleep and circadian phase.