The general absence of structure in the interior of the large fossil trees of the Coal implies the very durable nature of their bark, as compared with their woody portion. The same difference of durability of bark and wood exists in modern trees, and was first pointed out to me by Dr. Dawson, in the forests of Nova Scotia, where the Canoe Birch (Betula papyracea) has such tough bark that it may sometimes be seen in the swamps looking externally sound and fresh, although consisting simply of a hollow cylinder with all the wood decayed and gone. When portions of such trunks have become submerged in the swamps they are sometimes found filled with mud. One of the erect fossil trees of the South Joggins fifteen feet in height, occurring at a higher level than the main coal, has been shown by Dr. Dawson to have a coniferous structure, so that some Coniferæ of the Coal period grew in the same swamps as Sigillariæ, just as now the deciduous Cypress (Taxodium distichum) abounds in the marshes of Louisiana even to the edge of the sea.

When the carboniferous forests sank below high-water mark, a species of Spirorbis or Serpula ([Fig. 431]), attached itself to the outside of the stumps and stems of the erect trees, adhering occasionally even to the interior of the bark—another proof that the process of envelopment was very gradual. These hollow upright trees, covered with innumerable marine annelids, reminded me of a “cane-brake,” as it is commonly called, consisting of tall reeds, Arundinaria macrosperma, which I saw in 1846, at the Balize, or extremity of the delta of the Mississippi. Although these reeds are fresh-water plants, they were covered with barnacles, having been killed by an incursion of salt-water over an extent of many acres, where the sea had for a season usurped a space previously gained from it by the river. Yet the dead reeds, in spite of this change, remained standing in the soft mud, enabling us to conceive how easily the larger Sigillariæ, hollow as they were but supported by strong roots, may have resisted an incursion of the sea.

The high tides of the Bay of Fundy, rising more than 60 feet, are so destructive as to undermine and sweep away continually the whole face of the cliffs, and thus a new crop of erect fossil trees is brought into view every three or four years. They are known to extend over a space between two and three miles from north to south, and more than twice that distance from east to west, being seen in the banks of streams intersecting the coal-field.

Structure of Coal.—The bituminous coal of Nova Scotia is similar in composition and structure to that of Great Britain, being chiefly derived from sigillarioid trees mixed with leaves of ferns and of a Lycopodiaceous tree called Cordaites (Noeggerathia, etc., for genus, see [Fig. 428]), supposed by Dawson to have been deciduous, and which had broad parallel veined leaves without a mid-rib. On the surface of the seams of coal are large quantities of mineral charcoal, which doubtless consist, as Dr. Dawson suggests, of fragments of wood which decayed in the open air, as would naturally be expected in swamps where so many erect trees were preserved. Beds of cannel-coal display, says Dr. Dawson, such a microscopical structure and chemical composition as shows them to have been of the nature of fine vegetable mud such as accumulates in the shallow ponds of modern swamps. The underclays are loamy soils, which must have been sufficiently above water to admit of drainage, and the absence of sulphurets, and the occurrence of carbonate of iron in them, prove that when they existed as soils, rain-water, and not sea-water, percolated them. With the exception, perhaps, of Asterophyllites (see [Fig. 461]), there is a remarkable absence from the coal-measures of any form of vegetation properly aquatic, the true coal being a sub-aërial accumulation in soil that was wet and swampy but not permanently submerged.

Air-breathers of the Coal.—If we have rightly interpreted the evidence of the former existence at more than eighty different levels of forests of trees, some of them of vast extent, and which lasted for ages, giving rise to a great accumulation of vegetable matter, it is natural to ask whether there were not many air-breathing inhabitants of these same regions. As yet no remains of mammalia or birds have been found, a negative character common at present to all the Palæozoic formations; but in 1852 the osseous remains of a reptile, the first ever met with in the carboniferous strata of the American continent, were found by Dr. Dawson and myself. We detected them in the interior of one of the erect Sigillariæ before alluded to as of such frequent occurrence in Nova Scotia. The tree was about two feet in diameter, and consisted of an external cylinder of bark, converted into coal, and an internal stony axis of black sandstone, or rather mud and sand stained black by carbonaceous matter, and cemented together with fragments of wood into a rock. These fragments were in the state of charcoal, and seem to have fallen to the bottom of the hollow tree while it was rotting away. The skull, jaws, and vertebræ of a reptile, probably about 2½ feet in length (Dendrerpeton Acadianum, Owen), were scattered through this stony matrix. The shell, also, of a Pupa (see [Fig. 442]), the first land-shell ever met with in the coal or in beds older than the tertiary, was observed in the same stony mass. Dr. Wyman of Boston pronounced the reptile to be allied in structure to Menobranchus and Menopoma, species of batrachians, now inhabiting the North American rivers. The same view was afterwards confirmed by Professor Owen, who also pointed out the resemblance of the cranial plates to those seen in the skull of Archegosaurus and Labyrinthodon.[[6]] Whether the creature had crept into the hollow tree while its top was still open to the air, or whether it was washed in with mud during a flood, or in whatever other manner it entered, must be matter of conjecture.

Footprints of two reptiles of different sizes had previously been observed by Dr. Harding and Dr. Gesner on ripple-marked flags of the lower coal-measures in Nova Scotia (No. 2, [Fig. 447]), evidently made by quadrupeds walking on the ancient beach, or out of the water, just as the recent Menopoma is sometimes observed to do. The remains of a second and smaller species of Dendrerpeton, D. Oweni, were also found accompanying the larger one, and still retaining some of its dermal appendages; and in the same tree were the bones of a third small lizard-like reptile, Hylonomus Lyelli, seven inches long, with stout hind limbs, and fore limbs comparatively slender, supposed by Dr. Dawson to be capable of walking and running on land.[[7]]

In a second specimen of an erect stump of a hollow tree 15 inches in diameter, the ribbed bark of which showed that it was a Sigillaria, and which belonged to the same forest as the specimen examined by us in 1852, Dr. Dawson obtained not only fifty specimens of Pupa vetusta (Fig. 442), and nine skeletons of reptiles belonging to four species, but also several examples of an articulated animal resembling the recent centipede or gally-worm, a creature which feeds on decayed vegetable matter (see Fig. 441). Under the microscope, the head, with the eyes, mandible, and labrum, are well seen. It is interesting, as being the earliest known representative of the myriapods, none of which had previously been met with in rocks older than the oolite or lithographic slate of Germany.

Some years after the discovery of the first Pupa, Dr. Dawson, carefully examining the same great section containing so many buried forests in the cliffs of Nova Scotia, discovered another bed, separated from the tree containing Dendrerpeton by a mass of strata more than 1200 feet thick. As there were 21 seams of coal in this intervening mass, the length of time comprised in the interval is not to be measured by the mere thickness of the sandstones and shales. This lower bed is an underclay seven feet thick, with stigmarian rootlets, and the small land-shells occurring in it are in all stages of growth. They are chiefly confined to a layer about two inches thick, and are unmixed with any aquatic shells. They were all originally entire when imbedded, but are most of them now crushed, flattened, and distorted by pressure; they must have been accumulated, says Dr. Dawson, in mud deposited in a pond or creek.