Climate models and paleoclimatic evidence suggest that the earth's polar regions are particularly responsive to global scale climatic forcing. Empirical climatological evidence of the last century has thus far failed to confirm conclusively this regional sensitivity. We conduct an empirical study of polar sensitivity to climate forcing by comparing a high-quality, 17-year satellite-derived dataset of daily temperatures for 2.5¿ latitudinal bands to a known external forcing mechanism, the lunar phase cycle. The earth's polar regions display a temperature range of greater than 0.55 ¿C over the course of a synodic (29.53 day) month. This lunar-influenced range in temperature is 25 times larger than a similarly computed range in aggregated global temperatures over a synodic month. Temperature variations between the polar and non-polar regions also produce a pronounced temporal shift in sensible heat transfer. Strong poleward transfer of heat dominates near the full moon but the transfer substantially weakens near the new moon. It is unlikely that this sensitivity can be explained by the type of polar forcing previously identified in GCM simulations and paleoclimatic reconstructions, because of the short duration of the lunar cycle. However, it does demonstrate a new and potentially important external influence on the polar regions' climates.¿ 1997 American Geophysical Union