Frequently the iron and manganese which were originally finely disseminated through shale, sandstone, etc., are subsequently concentrated into bodies of comparatively pure ore, and very commonly this concentration takes place by a process of re-solution of the iron and manganese and re-deposition by replacement with limestone, or, more rarely, with some other material. The limestone or other material which thus acts as a precipitant is often in the same series of strata from which the iron and manganese were removed, and thus these two substances, which were once in a finely disseminated condition, may be converted into deposits of comparatively pure ore and yet remain in the same general series of strata in which they were originally deposited. A remarkable case of this is seen in the iron deposits of the Penokee series in Michigan and Wisconsin,[16] to be mentioned again on page 370. It has also been suggested by H. D. Rogers[17] that certain siderite deposits in the Coal Measures were formed by the conversion of finely disseminated sesquioxide of iron into carbonate of iron by organic matter, and the subsequent segregation of the carbonate as now found in layers and nodules. The surface waters that carry the iron and manganese to the strata being deposited at a given time are sometimes derived from areas in which iron predominates, sometimes from areas in which iron and manganese are both abundant, and sometimes, though rarely, on account of the scarcity of such regions, from areas in which manganese largely predominates over iron. If iron and manganese were always precipitated from these waters in similar chemical forms and under the same conditions, it would be expected that the strata deriving their iron and manganese from surface waters would contain those substances in the same relative proportions as they had existed in the rocks from which they were derived, and that they would be in an intimately mixed condition. Such is doubtless often the case, or at least approximately so; but it is also often the case that iron and manganese occur in separate deposits, yet in close proximity to each other and often alternating along the same horizon. Besides this, the two substances frequently form parts of the same deposit and yet are distinctly separate from each other. In such cases the question arises as to why the iron and manganese are not intimately mixed in the form of a manganiferous iron ore, as would be expected if they had been precipitated together. Moreover, deposits sometimes occur which are composed largely of manganese ore, with little or almost no iron, and when the source of the manganese is looked for, we often find that the rocks which probably supplied it contained both manganese and iron, and that the iron was present in a much larger proportion as regards the manganese than in the new deposit. Here again the question arises as to why the iron and manganese are not in the same relative proportions in