"The roots of the preceding fossils repeatedly ramify as their distance from the stem increases, and ultimately terminate in broad flattened points. The whole of the spreading roots of these trees (the _Sigillariæ_) cover only an area of thirty square feet each; whilst those of the Lepidodendron (Fig. 1), whose stem is only two or three inches larger in diameter, covered a space of two hundred square feet. Since it is well known, from numerous examples, that the Lepidodendra were lofty trees, with spreading branches, which therefore required wide bases for support, may we not conclude that Sigillariæ of the species described were, on the contrary, trees of low stature, without heavy branches?" I cannot quit this subject without again adverting to the remarkable phenomenon mentioned In a previous note, namely, that in the bed of pulverulent earth--the _under-clay_--on which the coal invariably reposes, the roots (or Stigmariæ) of large trees are generally the only organic remains met with. The constant occurrence of these fossils in the under-clay, and their rarity in the coal and shale, was long ago pointed out by Mr. Martin, Dr. Macculloch, and other geologists; but the importance of the fact was not appreciated till Mr. Logan drew attention to it. In the Welsh coal-field, through a depth of 1,200 feet, there are sixty beds of coal, each of which lies on a stratum of clay abounding in Stigmariæ. In the Appalachian coal formation of the United States, the same phenomena occur. Thus it appears that the under-clay is the natural soil in which the roots (_Stigmariæ_) of the Sigillariæ and Lepidodendra grew; the coal above it is composed of the carbonized stems and foliage of those trees; and the roof or coal-shale is formed by the leaves and branches of a forest overwhelmed and buried beneath the transported detritus of distant rocks. These phenomena may be explained by supposing that a plain, densely clothed with a luxuriant intertropical vegetation, was either inundated by an irruption of the sea, or overwhelmed by a flood, from the sudden breaking up of the barrier of an inland lake; or by the subsidence of the country on which the forests grew. But when we find an uninterrupted series, in which triple deposits of this character are repeated through many thousand feet, the solution of the problem is beset with difficulties, which the hypothesis of repeated periodical subsidences, however ingenious, does not, in my opinion, remove.[149] [Footnote 149: See Wonders of Geology, pp. 669, 718, 731.]