In our original theory of ice ages (1) we proposed that Quaternary glaciation is initiated by the migration of the geographic poles to regions of thermal isolation-that is, to the Arctic Ocean and Antarctica. Glacial-interglacial alternations were explained by the variability in the moisture supply for glacial growth, a variability resulting from the alternations of ice-covered with ice-free states of the Arctic Ocean.Although we still adhere to the fundamentals of this twofold model, important modifications have been made on the basis of new material that has become available. The changes in the model which apply to the glacialinterglacial aspect are, in summary, as follows: (i) the freezing of the Arctic Ocean surface, initially postulated to have occurred at the termination of a glacial stage (about 11,000 years ago in the case of the Wisconsin, or Wuirm, stage) must have occurred considerably earlier; (ii) the moisture source for the growth of the southern half of the North American ice sheet must have been, not the Arctic Ocean, but the rich southerly source usually assumed for the growth of the entire ice sheet; (iii) the termination of a glacial stage is a consequence of a dearth of atmospheric moisture for precipitation-a dearth caused by glacial growth rather than by freezing of the Arctic Ocean surface as a result of the lowering of sea level below some critical sill depth, as originally postulated.Our purpose in this article is to present the revised theory, including a discussion of the time of onset of glaciation and a consideration of the heat Dr.