sections showing the end walls recall at once the appearance of the sieve-plates of higher plants. The whole tissue may, in fact, be called a sieve parenchyma (Plate II., figs. 7, 8, and 9).

Protoplasm is abundant, especially in the middle region of the cortex; and when contracted into the centre of the cell remains connected with the wall by distinct threads. These are frequently seen to be in a line with corresponding threads of neighbouring cells, and are continuous with them. Treatment with sulphuric acid and Hoffman's blue brings out this structure very clearly, and also shows that the mass of contracted protoplasm is largely composed of convoluted filaments.

This peculiar appearance of the cell-contents is due to the presence of a fungus like that described by Treub' in Lycopodium cernuum, and by Bruchmann' in L. annotinum. As in these cases the protoplasm co-exists with the fungus, and the cell-nucleus can be clearly distinguished, so that the connexion seems to be a sort of symbiosis. The middle zone of the ground tissue is that most attacked, as was observed by Treub; but this is probably only because that zone contains more protoplasm than the peripheral layers. We have not been able to find any fruit on the fungal hyphae, neither can we trace them in from the rhizome hairs, though a few may be seen running along these hairs. Considering the physiological importance which has been attributed by some botanists3 to intracellular fungi in plant roots, it is interesting to find them here in a plant whose whole habit is so peculiar."

In the stem the cortical cells are much longer, more rectangular, and thicker walled. The side walls show numbers of minute slit-like pits in oblique parallel rows, similar to those of the sclerenchymatous fibres of Pteris aquilina. The cells contain very little protoplasm in comparison with those of the rhizome. No hypodermal specialization of the ground tissue is present. On the contrary, the epidermal cells resemble those of the ordinary ground tissue in their form and contents. Beyond the external thickening of the epidermal cells,

1 Treub, Annales du Jardin Botanique de Buitenzorg, vol. v., 1886. 2 Bruchmann.

Marshall Ward, Phil. Trans., 1887, and literature on "Mycorhiza" generally. Since this Paper was written, M. Dangueard in Le Botaniste (April, 1891), has described the Mycorhiza in Tmesipteris. We have not had time to study his account in detail, but it seems he has noted three kinds of hyphae and spores in the cells, and that he believes some of the hyphae may be referred to Nectria.

there is, therefore, little difference between the limitary and fundamental tissues.

Round the fibro-vascular system a well-marked differentiation occurs, due rather to the contents than to the structure of the cells. Immediately outside the phloem in the rhizome lies a zone of cells, one to three layers thick, containing a large quantity of starch. M. Bertrand1 has described this as the "protective sheath" ("gaine protectrice mal cáracterisée"). That term is, however, generally used in England as equivalent to endodermis, a structure which is not typically developed in Tmesipteris. There is no single layer of modified cells round the bundle, and no folding of the radial walls characteristic of a typical endodermis. It would, perhaps, be better, therefore, to use the term "phloem sheath" rather than "protective sheath." The layer is present in the stem, and even in the leaf, though in both cases less evident than in the rhizome, from the absence of the abundance of starch which characterizes it in that region; in younger parts it is represented by one cell-layer. (Plate II., figs. 6, 8, 13.)

Outside this phloem sheath lies a zone of cells conspicuous by their bright brown contents. (Plate II., figs. 5, 6, 8, &c.) In transverse section they appear as a ring of irregular patches surrounding the bundle, and looked at in a tangential section are seen to form a network with elongated meshes. The longitudinal brown masses are connected by transverse anastomoses, so that the tissue resembles in appearance some of the simpler types of laticiferous tissue, such as that of Lactuca. Closer examination shows, however, that the brown substance does not fill up special cavities, but is developed in ordinary cells of the deeper layers of the ground tissue, and the transverse anastomoses generally correspond to the cross walls between them. It may be mentioned that the brown layer is perfectly distinct in fresh specimens, and therefore not a result of preservation in spirit. No reaction is obtained by treatment of the brown substance with ferrous sulphate and nitric acid, and by this and other tests the presence of tannin cannot be demonstrated. The colour can be removed by prolonged maceration in strong Eau de Javelle, or caustic alkalis, but resists the action of acids longer than the surrounding tissues. A black colouring results from the addition of caustic potash previous to solution; a reaction that would suggest the presence of tannin were it

1 Bertrand, loc. cit. (p. 232).

not for the absence of any effect in the case of ferrous sulphate test. It differs from resins in being insoluble in absolute alcohol, but gives dense precipitate of ferric hydrate (when the sheath is dissected out and dissolved in HCl) with ammonia, showing that it contains a large percentage of iron.

The brown zone is absent in the younger parts of the stem, and consists of one layer of cells in the older part (Plate II., figs. 10, 13): in the rhizome two or three layers are often concerned in its formation. M. Bertrand1 describes this tissue as a zone "dont les parois cellulaires sont en partie gelifiées," He regards the brown matter as due to a local or general "gélification" and "humification" of the cell wall; and compares the change with that seen near the base of the stem in Psilotum. In the development of these brown contents there seems however, little in common with a mucilaginous change in the cell wall as usually understood. There is no stratification in the brown substance; it may be seen (in material preserved in spirits) in any position as regards the cell wall, and sometimes contracted into the cell cavity. The wall of the containing cell is of normal thickness throughout, even when large masses of the brown substance lie against some particular region of it. We have examined sections treated with glycerine and potash, but we can detect no thickening and swelling of the walls similar to that figured by M. Bertrand, either in spirit material or dried specimens.

There seems to be some doubt as to M. Bertrand's' use of the term "gélifiée et humifiée" as he applies them also to the external wall of the epidermis which, though brown in colour, differs greatly from the brown sheath, both in appearance and behaviour under reagents.

The fibro-vascular system forms an axial strand, very distinctly marked even to the naked eye, by the dark surrounding ring. Leaf trace bundles run obliquely across the ground tissue in a line with the midrib, so that they join the main system a little below the base of the

leaf.

The xylem forms a central, oval, or transversely elongated bipolar mass in the rhizome portion; but in the stem region there are five or six separate xylem strands. By examining transverse sections at different levels M. Bertrand1 has shown how these strands in the stem coalesce to form the single one in the rhizome. The xylem elements are large scalariform tracheides with pointed ends, similar in all

1 Bertrand, loc. cit

respects to those of ferns, and narrower forms in which the unthickened areas vary from simple circular pits to oval bordered pits. These less specialized tracheides are characteristic rather of the younger parts of the stem. The tracheides do not always occupy the centre of the bundle, where there may be two or three rows of long rectangular parenchyma cells.2 The latter, though they may be in parts absent in the rhizome, occupy the centre of the axis in the younger parts of the "stem."

Examination of the region behind the growing point shows that the tracheides develop round a central core of elongated but unlignified cells. Also that in the earliest formed tracheides, the thickening, though somewhat more irregular than in older tissue, is still of the scalariform type. We have not been able to detect any spiral protoxylem, and in the absence of this element the bundle agrees with that of Psilotum.3

The phloem, which completely surrounds the xylem, is of simple structure. It consists of elongated narrow cells, with transverse or oblique end walls, and of long tubes pointed at the ends. The walls of the cells are delicately pitted, and the tubes generally show a row of sieve plates, nearly as wide as the vessel, covered with the brightly refracting granules common in the "sieve-tubes" of the Lycopodiaceæ. We have not been able to detect pores in the pointed ends of the "sievetubes," but the transverse walls of the phloem cells are pitted and perhaps porous. The latter most resemble the elements described by Bertrand as sieve-tubes in Psilotum. The figures published by Janczewski, for Lycopodium do not show the end walls. Moreover, the phloem in Lycopodium is coloured blue by iodine, and that of most plants by iodine and sulphuric acid; but we have not been able to obtain reaction by either method in Tmesipteris.

Russow believes that the phloem walls in Tmesipteris are lignified, and compares them with the phloem of many monocotyledonous roots. We have found, however, that with all differentiating stains used the walls of the phloem elements never give the reactions of wood.

1 The Primitive Fibres of Bertrand. They are not lignified.

2 Russow, Vergleichende Untersuchungen, Mem. de l'Acad. Imperiale des Sc. de

St. Petersbourg, vii. series, tom. xix., 1873.

3 Janczewski, Mem. de la Soc. Sc. Nat. Cherbourg v.,

xxiii.

4 Janczewski, loc. cit. Recorded in L. clavatum, L. annotinum, and L. complanatum. We find it also to be the case in L. selago and L. alpinum.

5 Russow, loc. cit.

Long nuclei are seen in the phloem cells, and in sections treated with sulphuric acid and Hoffman's blue, a delicate protoplasmic network is shown lining the walls. The phloem therefore consists mainly of elements termed "cambiform tissue " by De Bary.

The growing point of the stem we have not been able to work out in detail. The primitive meristem differentiates rapidly into a dermatogen layer externally, and a procambial cylinder round the axis. An apical cell is present, but of small size, and differing little from surrounding cells. In surface view the apical cell is triangular, and segments are cut off on the three sides which rapidly subdivide by radial and tangential walls. (Plate III., fig. 18.) In the instance figured, the only one which the amount of material at our disposal has enabled us to examine, the earlier divisions do not seem to follow any very definite plan. Their arrangement is, however, closely similar to those in Psilotum, as figured by Laubach.1

The short lateral shoots developed from the rhizome near the surface of the fern-trunk are of exogenous origin, and have essentially the same structure as the main axis. The fibro-vascular bundle is well developed, and large tracheids are found extending right up to the apex. In the ground tissue a large quantity of starch is present.

II.—The Leaf.

The leaves in Tmesipteris are sessile, ovate or oblong in general outline, but slightly asymmetric, the upper border being more convex near the base, and the lower border more concave, so that the margin is slightly falcate in outline. In T. forsteri, the apex is slightly truncate, and the midrib produced into a prominent mucro. The extent of the truncation and development of the mucro is, however, extremely variable, and raises a doubt as to the value of the species which depend on the leaf characters. Thus, while T. forsteri (T. tannensis, Bernh.) is said to have acuminate fronds, T. billardieri (T. tannensis, Labill, T. truncata, Desh.) is said to be characterized by their truncation. M. Bertrand adds two forms, one with leaves both truncate and acuminate, and one with some truncate and some acuminate. The fact that in the same plant of T. fosteri there may occur leaves tapering

1 Solms-Laubach Annales du Jardin Botanique de Buitenzorg, 1884, vol. iv.