It is scarcely necessary to remark that in the above summary I have had reference principally to my own observations in the coal formation of Nova Scotia; but similar facts have been detailed by many other observers in other districts.[116]
[116] Especially Brongniart, Goeppert, Hawkshaw, Lyell, Logan, De la Boche, Beaumont, Binney, Rogers, Lesquereux, Williamson, Grand' Eury.
A curious point in connection with the origin of coal is the question how could vegetable matter be accumulated in such a pure condition? There is less difficulty in regard to this if we consider the coal as a swamp accumulation in situ. It is in this way that the purest vegetable accumulations take place at present, whereas in lakes and at the mouths of rivers vegetable matter is always mixed up with mud. Coal swamps, however, must have been liable to submergences or to temporary inundations, and it is no doubt to these that we have to attribute the partings of argillaceous matter often found in coal beds, as well as the occasional gulches cut into the coal and filled with sand and lenticular masses of earthy matter. To a similar cause we must also attribute the association of cannel with ordinary coal. The cannel is really a pulpy, macerate mass of vegetable matter accumulated in still water, surrounded and perhaps filled with growing aquatic herbage. Hence it is in such beds that we find the greatest accumulations of macrospores, derived, probably, in great part from aquatic plants. Buckland long ago compared the matter of cannel to the semi-fluid discharge of a bursting bog, and Alex. Agassiz has more recently shown that in times of flood the vegetable muck of the Everglades of Florida flows out in thick inky streams, and may form large beds of vegetable matter having the character of the materials of cannel. It is evident that in swamps of so great extent as those of the coal formation, there must have been shallow lakes and ponds, and wide sluggish streams, forming areas for the accumulation of vegetable débris and this readily accounts for the association of ordinary beds of coal with those of cannel, and with bituminous shales or earthy bitumen, as well as for the occurrence of scales of fish and other aquatic animals in such beds. Lyell's interesting observation of the submerged areas at New Madrid, keeping free of Mississippi mud, because fringed with a filter of cane-brake, shows that the areas of coal accumulation might often be inundated without earthy deposit, if, as seems probable, they were fringed with dense brakes of calamites, sheltering them from the influx of muddy water. It seems also certain that the water of the coal areas would be brown and laden with imperfect vegetable acids, like that of modern bogs, and such water has usually little tendency to deposit any mineral matter, even in the pores of vegetable fragments. The only exception to this is one which also occurs in modern swamps, namely, the tendency to deposit iron, either as carbonate (Clay Ironstone), or sulphide (Iron Pyrite), both of which are products of modern bogs, and equally characteristic of the coal swamps.
Where great accumulations of sediment are going on, as at the mouths of modern rivers, there is a tendency to subsidence of the area of the deposit, owing to its weight. This applies, perhaps, to a greater extent to coal areas. Thus the area of a coal swamp would ultimately sink so low as to be overflowed, and a roof shale would be deposited to bury up the bed of coal, and transmit it to future ages, chemically, and mechanically changed by pressure and by that slow decomposition which gradually converts vegetable matter into carbon and hydrocarbons. The long continuance and great extent of these alternations of growth and subsidence is perhaps the most extraordinary fact of all. At the South Joggins, if we include the surfaces having erect trees with those having beds of coal, the process of growth of a forest or bog, and its burial by subsidence and deposition must have been repeated about a hundred times before the final burial of the whole under the thick sandstones of the Upper Carboniferous and Permian.
Mention has been made of Sigillaria and other trees of the coal formation period. These trees and others allied to them, of which there were many kinds, may be likened to gigantic club mosses, which they resembled in fruit and foliage, though vastly more complex in structure of stem and branch. Some of them, perhaps, were of much higher rank than any of the modern plants most nearly allied to them. One of their most remarkable features was that of their roots—those Stigmariæ, to which so frequent reference has been made. They differed from modern roots, not only in some points of structure, but in their regular bifurcation, and in having huge root fibres articulated to the roots, and arranged in a regular spiral manner, like leaves. They radiate regularly from a single stem, and do not seem to have sent up buds or secondary stems. They thus differed from the botanical definition of a root, and also from that of a rhizoma, or root stock; being, in short, a primitive and generalized contrivance, suited to trees themselves primitive and generalized, and to special and peculiar circumstances of growth. Some botanists have imagined that they were aquatic plants, growing at the bottom of lakes, but their mode of occurrence negatives this. I have elsewhere stated this as follows:—[117]
[117] Natural Science, May, 1892.
"It is quite certain that Stigmariæ are not 'rhizomes which floated in water, or spread themselves out on the surface of mud.' Whether rhizomes or not, they grew in the soil, or in the upper layers of peaty deposits since changed into coal. The late Richard Brown and the writer have shown that they grew in the underclays or fossil soils, and that their rootlets radiated in these soils in all directions.[118] In one of my papers I have figured a Stigmarian root penetrating through an erect Sigillaria, and Logan, in his Report of 1845, had already figured a similar example. The penetration of decaying stems by the rootlets of Stigmaria is a fact well known to all who have studied slices of Carboniferous plants,[119] while Stigmariæ are often found creeping inside the bark of erect and prostrate trunks. Besides this, as I have shown in 'Acadian Geology,' in the section of 5,000 feet of coal measures at the South Joggins (including eighty-one distinct coal groups, and a larger number of soils with Stigmaria, or erect trees), Sigillaria and Stigmaria occur together, and the latter nearly always either in argillaceous soils, or sands hardened into 'Gannister,' which are often filled with roots or rootlets, or on the surfaces of coal beds. On the other hand, the numerous bituminous limestones, and calcareous and other shales holding remains of fishes, crustaceans, and bivalve shells do not contain Stigmaria in situ—the only exceptions being two beds of bituminous limestone, the upper parts of which have been converted into underclays. This section, and that of North Sydney—two of the most complete and instructive in the world—have afforded conclusive proof of this mode of growth of Sigillaria and Stigmaria.
[118] Quart. Journ. Geol. Soc., vol. ii. p. 394 (1846); Ibid., vol. iv. p. 47 (1847); Ibid., vol. v. p. 355 (1849); Ibid., vol. v. pp. 23, 30.
[119] Williamson has noticed this in his excellent Memoirs in the Phil. Trans.