The exact position of the tree in relation to the underlying coal is shown in the above section. Fig. 4. Immediately over the coal there is a bed of hard shale, six inches in depth, in which no fossils are found; this is overlaid by a softer shale abounding in coal-plants; all the upright stems were rooted in the six-inch shale. Upon clearing the base of this tree, a complete set of conical tap-roots was discovered, arranged as in the annexed sketch. Fig. 5. There are four large tap-roots in each quarter of the stump, and five inches below these another set of smaller tap-roots; the total number amounting to eighteen. The horizontal roots are seen to branch off in a regular manner, the base being divided into four nearly equal parts by deep channels, extending from the centre to the points _i, k, l, m_. [Illustration: _Fig. 5._ _The under surface of Fig. 3, showing the conical tap-roots of the tree, 1/12 natural size._] Mr. Brown remarks, that these short thick tap-roots were evidently adapted only to a soft wet soil, such as we may conceive was the nature of the first layer of mud deposited upon a bed of peat which had settled down slightly below the level of the water. He supposes, from the presence of a layer of shale without fossil plants immediately over the coal, that the prostrate stems and leaves which occur in such large quantities in the next superincumbent bed, fell from trees growing on the spot, and were entombed in layers of mud held in suspension in the water, which at short intervals inundated the low marshy ground on which they grew; for had the plants been drifted from a distance, he conceives they would also occur in the first layer of shale, as well as in those higher up. [Illustration: _Fig. 6._ STEM OF A SIGILLARIA BROKEN OFF CLOSE TO THE ROOTS, 1/12 _natural size_.] Having thus shown that the _Sigillaria alternans_ was provided with roots adapted for a soft muddy soil, Mr. Brown next describes the specimen represented in Fig. 6, which is the stem of the same species of tree broken off near the roots; the hollow cylinder of bark (_a_) having been bent down and doubled over by the pressure of the surrounding mud, so as effectually to close up the aperture, leaving only a few irregular cicatrices, of three or four inches in length, converging at the apex; the structure, arrangement, and number of the tap-roots, as well as the horizontal ramifications, are similar to those in Fig. 5. This fossil clearly explains the nature of the "_dome-shaped_" plant figured in the "Fossil Flora of Great Britain."[148] [Footnote 148: The figures 3, 4, 5, 6, and the descriptions, are from the paper of Richard Brown, Esq., published in the Journal of the Geological Society of London, for March, 1849, entitled, "Description of erect Sigillariæ, with conical tap-roots, found in the roof of the Sydney Main Coal, in the Island of Cape Breton."]