(3) _Inorganic Products formed during a time of Ice Recession._ The recession of the ice after a maximum of advance would leave a land surface more or less affected with marshes and ponds. In such situations, bog iron ore might accumulate, if conditions were favorable. Such ore beds, buried by the drift of a later ice advance, would have a significance comparable to that of forest beds, except that they would give less definite information as to climate, and would be correspondingly less trustworthy. Should such ore beds be found in such relations as to prove that the underlying and overlying bodies of drift were deposited by ice sheets which extended great distances further south, their significance would be enhanced. From the thickness of the ore beds some inference might be drawn as to the length of time concerned in their accumulation. But because of the variable rate at which bog ore may accumulate, such inference should be used with caution. Concretions of iron oxide might be formed in the marshes or in ill-drained drift areas where accumulations of greater extent were not made. A subsequent incursion of the ice might incorporate these nodules with its drift, wearing and striating them as other stones, and depositing them as constituent parts of the later drift. Such iron nodules in the later drift would mean a recession and re-advance of the ice with some considerable interval between, although not necessarily an interval sufficiently warm or long to be regarded as an interglacial epoch.[15] Calcareous concretions, like those of the loess, would possess a like significance, in like relations. While in themselves these inorganic products of a time of ice recession might fail to be conclusive of separate ice epochs, they might have much corroborative significance when associated with other phenomena. An inter-till iron ore bed, associated with a forest bed which indicated a warm climate, would be most significant. [15] This point concerning iron nodules was suggested to the writer by Mr. W. J. McGee. The absence of knowledge of ore beds between sheets of till, and the absence from an upper bed of till of concretions of iron and lime carbonate formed during a recession of the ice, would be no proof that interglacial epochs did not occur. These products were probably formed in relatively few localities. They stood good chance of destruction at the hands of the returning ice, and they may exist, where they have not been discovered, or where their significance has not been understood. Their absence is at best no more than negative evidence.