It has been urged against the criterion of differential weathering that superglacial material is or may be thoroughly oxidized before its deposition, and that a layer of oxidized drift between layers of till may be no more than superglacial debris deposited during a minor recession of the ice.[16] We believe that this attempt to eliminate the value of this criterion rests partly on an exaggerated idea concerning the amount of superglacial material, but more especially on a failure to apprehend the real meaning of the argument for the validity of the criterion, and upon a failure to note the limitations imposed upon it by its advocates. It is not affirmed that a layer of oxidized drift between beds of unoxidized drift is _per se_ proof of two glacial epochs; but it is affirmed that if such layer of weathered drift can be shown to extend far below any possible superglacial till, into the subglacial till below, in such wise as to indicate that it is the result of subaërial exposure in a warm climate subsequent to its deposition and prior to the deposition of the overlying till, it constitutes the best possible evidence of an interglacial epoch, especially when accompanied by the corroborative testimony of other criteria. It is further affirmed that if the second sheet of drift failed to reach the limit of the first, and if the drift which was deposited by the first and never covered by the second ice-sheet, is more thoroughly and more deeply weathered than that deposited by the second, and especially if the two types of drift surface meet along a definite and readily traceable line, the argument becomes, in our judgment, irrefragable. In its application, this criterion would be infallible only in the hands of one who could distinguish between superglacial and superglacially oxidized material on the one hand, and material subaërially weathered after its deposition, on the other. [16] This point was urged at the reading of the paper at Ottawa, by Prof. C. H. Hitchcock, Mr. Upham, and others. In circumstances and relations where the weathering of the drift is not in itself conclusive, it might still have corroborative value in association with other lines of evidence.