Phase analyses of the annual cycle of monthly temperature time series that date back to the eighteenth century show trending behavior that has been difficult to interpret. Negative trends in the estimated phase have been identified with precession of Earth’s axis of rotation, but the implied later onset of seasons is at odds with recent satellite measurements and with the phenological record. Positive trends in the phase and the occurrence of trends of both signs in temperature time series from geographically nearby locations have remained mysterious. This paper shows that there is a mathematical equivalence between trends in phases and seasonally differing warming trends, in particular more intense warming in winters than in summers. Using temperature time series from 16 Northern Hemispheric locations reaching back to the eighteenth century and a statistical model that can estimate the seasonal warming trends, the authors reject the hypothesis that the timing of the seasons in these locations is jointly driven by precession.