cular scar, and situated in verticils at the top of well-marked nodes of the stem.

In tree-ferns the leaf-bases are large and usually without a distinct articulating surface. The vascular bundles are numerous. Protopteris has rounded leaf-scars with a large horseshoe-shaped bundle of vessels above and small bundles below. Caulopteris has large elliptic or oval leaf-scars with vascular scars disposed concentrically.. Palæopteris,* of Geinitz, has the leaf-scars transversely oval and the vascular bundles confluent in a transverse band with an appendage or outlying bundle below. Stemmatopteris has leafscars similar to those of Caulopteris, but the vascular bundles united into a horseshoe-shaped band.

2. Subdivision of Sigillariæ in Accordance with their Markings. The following groups may be defined in this way; but, being based on one character only, they are of course in all probability far from natural:

1. Sigillaria, Brongniart. Type, Sigillaria reniformis, Brongniart, or S. Brounii, Dawson.-Stem with broad ribs, usually much broader than the usually oval or elliptical tripunctate areoles, but disappearing at base, owing to expansion of the stem. Leaves narrow, long, three-nerved.

2. Rhytidolepis, Sternberg. Type, S. scutellata, Brongniart.Ribs narrow, and often transversely striate. Areoles large, hexagonal or shield-shaped, tripunctate. Leaves as in last group. Rings of rounded scars on the stems and branches mark attachment of fruit. It is possible that some of the smaller stems of this group may be branches of trees of group first.

3. Syringodendron, Sternberg. Type, S. organum, L. and H., S. oculata, Brongniart.-Stems ribbed; areoles small and round, and apparently with a single scar, or three closely approximated. These are rare, and liable to be confounded with decorticated examples of other groups; but I have some specimens which unquestionably represent the external surface.

4. Favularia, Sternberg. Type, Sigillaria elegans of Brongniart.-Leaf-bases hexagonal, or in young branches elliptical, in vertical rows, but without distinct ribs, except in old or decorticated stems. Fruit borne in verticils on the branches bearing transverse rows of rounded scars. Leaves somewhat broad and longitudinally striate.

* This name, preoccupied by Geinitz, has been inadvertently misapplied to the Devonian ferns of the genus Archeopteris.

5. Leioderma, Goldenberg. Type, S. Sydnensis, Dawson. Ribs obsolete. Cortical and ligneous surfaces striate. Vascular scars double, elongate longitudinally, and alike on cortical and inner surfaces. Areoles in rows and distinct; stigmaria-roots striate, with small and distinct areoles.

6. Clathraria, Brongniart. Type, S. Menardi, Brongniart.— Areoles hexagonal, not in distinct rows, but having a spiral appearance. Some of the plants usually referred to this group are probably branches of Favularia. Others are evidently fragments of plants of the genus Lepidophloios.

3. Internal Structures of Sigillaria-Stems.

I long ago pointed out, on the evidence of the external markings and mode of growth, that the stems of Sigillaria must have been exogenous, and this conclusion has now been fully confirmed by the microscopic researches of Williamson, not only in the case of Sigillariæ, but of Lepidodendra and Calamodendra as well. Confining myself to my own observations, three types of Sigillariæ are known to me by their internal structures, though I cannot certainly correlate all of these with the external markings referred to above.

1. Diploxylon, in which the stem consists of a small internal axis surrounded by a very thick inner bark and a dense outer cortex, A fine example from the South Joggins is thus described:*

"The axis of the stem is about six centimetres in its greatest diameter, and consists of a central pith-cylinder and two concentric coats of scalariform tissue. The pith-cylinder is replaced by sandstone, and is about one centimetre in diameter. The inner cylinder of scalariform tissue is perfectly continuous, not radiated, and about one millimetre in thickness. Its vessels are somewhat crushed, but have been of large diameter. Its outer surface, which readily separates from that of the outer cylinder, is striated longitudinally. The outer cylinder, which constitutes by much the largest part of the whole, is also composed of scalariform tissue; but this is radially arranged, with the individual cells quadrangular in cross-section. The cross-bars are similar on all the sides and usually simple and straight, but sometimes branching or slightly reticulated. The wall intervening between the bars has extremely delicate longitudinal waving lines of ligneous lining, in the manner first described by Williamson as occurring in the scalariform tissue of certain Lepidodendra. A few small radiating spaces, partially

* "Journal of the Geological Society of London," November, 1877,

occupied with pyrites, obscurely represent the medullary rays, which must have been very feebly developed. The radiating bundles passing to the leaves run nearly horizontally; but their structure is very imperfectly preserved. The stem being old and probably long deprived of its leaves, they may have been partially disorganised before it was fossilised. The outer surface of the axis is striated longitudinally, and in some places marked with impressions of tortuous fibres, apparently those of the inner bark. In the cross-section, where weathered, it shows concentric rings; but under the microscope these appear rather as bands of compressed tissue than as proper lines of growth. They are about twenty in number. This tree has an erect, ribbed trunk, twelve feet in height and fifteen inches in diameter, swelling to about two feet at the base.

2. Favularia Type.-This has been well described by Brongniart and by Renault,* and differs from the above chiefly in the fact that the outer exogenous woody zone is composed of reticulated instead of scalariform tissue, and the inner zone is of the peculiar form which I have characterised as pseudo-scalariform.

3. Sigillaria Proper.—This I have illustrated in my paper in the "Journal of the Geological Society" for May, 1871, and it appears to represent the highest and most perfect type of the larger ribbed Sigillaria. This structure I have described as follows, basing my description on a very fine axis found in an erect stem, and on the fragments of the woody axis found in the bases of other erect stems:

a. A dense cellular outer bark, usually in the state of compact coal-but when its structure is preserved, showing a tissue of thickened parenchymatous cells.

b. A very thick inner bark, which has usually in great part perished, or been converted into coal, but which, in old trunks, contained a large quantity of prosenchymatous tissue, very tough and of great durability. This "bast-tissue" is comparable with that of the inner bark of modern conifers, and constitutes much of the mineral charcoal of the coal-seams.

c. An outer ligneous cylinder, composed of wood-cells, either with a single row of large bordered pores,† in the manner of pines

* "Botanique Fossile," Paris, 1881.

These are the same with the wood-cells elsewhere called discigerous tissue, and to which I have applied the terms uniporous and multiporous. The markings on the walls are caused by an unlined portion of the cellwall placed in a disk or depression, and this often surrounded by an

་

and cycads, or with two, three, or four rows of such pores sometimes inscribed in hexagonal areoles in the manner of Dadoxylon. This woody cylinder is traversed by medullary rays, which are short, and composed of few rows of cells superimposed. It is also traversed by oblique radiating bundles of pseudo-scalariform tissue proceeding to the leaves. In some Sigillaria this outer cylinder was itself in part composed of pseudo-scalariform tissue, as in Brongniart's specimen of S. elegans; and in others its place may have been taken by multiporous tissue, as in a case above referred to; but I have no reason to believe that either of these variations occurred in the typical ribbed species now in question. The woody fibres of the outer cylinder may be distinguished most readily from those of conifers, as already mentioned, by the thinness of their walls, and the more irregular distribution of the pores. Additional characters are furnished by the medullary rays and the radiating bundles of scalariform tissue when these can be observed.

d. An inner cylinder of pseudo-scalariform tissue. I have adopted the term pseudo-scalariform for this tissue, from the conviction that it is not homologous with the scalariform ducts of ferns and other acrogens, but that it is merely a modification of the discigerous wood-cells, with pores elongated transversely, and sometimes separated by thickened bars, corresponding to the hexagonal areolation of the ordinary wood-cells. A similar tissue exists in cycads, and is a substitute for the spiral vessels existing in ordinary exogens.

e. A large medulla, or pith, consisting of a hollow cylinder of cellular tissue, from which proceed numerous thin diaphragms towards the centre of the stem.

These structures of the highest type of Sigillaria are on the one hand scarcely advanced beyond those of Calamopitus, as described by Williamson, and on the other approach to those of Cordaites, as seen in specimens presented to me by Renault.

Finally, as to the fruit of Sigillariæ, I have no new facts to offer. The strobiles or spikes associated with these trees have been variously described as gymnospermous (Renault) or cryptogamous (Goldenberg and Williamson). I have never seen them in place. Two considerations, however, have always weighed with me in reference to this subject. One is the constant abundance of Trigonocarpa

hexagonal rim of thickened wall; but in all cases these structures are less pronounced than in Dadoxylon, and less regular in the walls of the same cell, as well as in different layers of the tissues of the axis.

M

and Cardiocarpa in the soil of the Sigillaria forests, as I have studied this at the South Joggins. The other is that the rings of fruit-scars on the branches of Sigillaria are homologous with leaf-scars, not with branches, and therefore should have borne single carpels and not cones or spikes of inflorescence. These are merely suggestions, but I have no doubt they will be vindicated by future discoveries, which will, I have no doubt, show that in the family Sigillariacea we have really two families, one possibly of gymnospermous rank, or at least approaching to this, the other allied to the Lepidodendra.

CRYPTOGAMIA.

(Acrogenes.)

Family LEPIDODENDREÆ; Genus LEPIDODENDRON, Sternberg.

These are arboreal Lycopods having linear one-nerved leaves, stems branching dichotomously, and with ovate or rhombic leaf-bases bearing rhombic leaf-scars, often very prominent. The fruit is in scaly strobiles, terminal or lateral, and there are usually, if not always, macrospores and microspores in each strobile. The young branches and stems have a central pith, a cylinder of scalariform tubes sending out ascending bundles to the leaves through a thick cellular and fibrous inner bark, and externally a dense cortex confluent with or consisting of the leaf-bases. Older stems have a second or outer layer of scalariform fibres in wedges with medullary rays, and strengthening the stem by a true exogenous growth, much as in the Diploxylon type of Sigillaria. The development of this exogenous cylinder is different in amount and rate in different species.* This different development of the exogenous axis is accompanied with appropriate external appearances in the stems, and the changes which take place in their markings. These are of three kinds. In some species the areoles, at first close together, become, in the process of the expansion of the stem, separated by intervening spaces of bark in a perfectly regular manner; so that in old stems, while widely separated, they still retain their arrangement, while in young stems they are quite close to one another. This is the case in L. corrugatum. In other species the leaf-scars or bases increase in size in the old stems, still retaining their forms and their contiguity to each other. This is the case in L. undulatum, and generally in those Lepidodendra which have large leaf-bases. In these species the

* See "Memoirs of Dr. Williamson," in "Philosophical Transactions," for ample details.