gradually to a point, and others distinctly truncate and mucronate, does not strengthen the view that there is more than one species.

The different form of epidermal cells and variation in pattern due to thickening, noticed by M. Bertrand in T. forsteri and T. billardieri, seems also to be of little value in specific distinction, when the variation of these characters in one plant is observed.1

The leaf, which, like that of Lycopodiacea, is of the centric type, consists of several layers of spongy parenchyma bounded by a single layered epidermis, and traversed by a small axial bundle.

The parenchyma cells are large and irregular; appearing rounded in cross section, but in surface view oblong and with well-marked lateral outgrowths which give rise to the lacunar nature of the tissue. The areas of the contact of the cells are pitted and marked out by a distinct thickening of the wall; so that if the tissue has been broken up the thickened rings often separate altogether from the rest of the cell wall. The innermost layer of the cell wall is also thicker and more highly refracting than those between it and the middle lamella. In the protoplasmic layer lies an elongated oval nucleus, and numerous starch granules. Some of the parenchyma cells also contain tannin,2 forming oval drops round the cell nucleus.

The epidermis consists of a single layer of cells three or four times as long as broad, but rounded or square in transverse section. The side-walls are waved or sharply zigzag; the angles corresponding to the bases of thickening lamella which run across the roof of the cell, and anastomose forming a perfect reticulation, which is, however, not easily seen except by oblique light. In transverse sections the cells are seen to fit closely together; the outer wall or roof is flat or slightly convex, and the side walls approximately straight and vertical to the surface; the inner wall or floor fits irregularly against the parenchyma below, so that the distinction between the latter and the epidermis is by no means sharp. A thin cuticle covers the outer wall, and the wall itself is largely cuticularized: the cuticular portion running down the lateral walls of the cells for about two-thirds of their length. In many

1 M. Dangueard (loc. cit.) has added several species founded on similarly variable external characters, but he claims to support the value of his distinctions by histological differences.

They give the characteristic greenish-black colouring with ferrous sulphate and nitric acid.

cases the thickening bands crossing the outer wall, send down vertical lamellæ projecting into the cavity of the cell, and sometimes extending halfway across it. (Plate III., fig. 23.) In their contents the epidermal cells of the leaf resemble those of the subjacent parenchyma.

The stomata are numerous on both sides of the leaf, and of simple structure. They consist of a pair of simple guard-cells lying between ordinary epidermal cells; there are no subsidiary cells. The guardcells have the external wall much thickened in the middle, round the cleft, but comparatively thin at each end. There is a well-marked airspace below the stoma. Compared with those of the stem they are longer, and the transpiratory chamber larger.

The fibro-vascular bundle of the leaf is similar in structure to that of the axis, though considerably reduced. It consists of three or four scalariform tracheides, surrounded by a ring of elongated phloem cells, with finely-pitted walls and long nuclei. (Plate III., figs. 21, 22.) A sheath of larger rectangular cells encloses the whole. The bundle runs straight along the centre of the leaf to terminate in its mucronate apex, and gives off no lateral offshoots in its course.

Toward the base of the stem the leaves become smaller and pass into little subulate scales about a quarter of an inch or less in length. They have similar structure to the leaves, but contain no fibro-vascular bundle.

III.-The Sporangia.

Toward the end of the axis the leaves are replaced by sporangiferous shoots; though ordinary leaves not unfrequently occur singly among these. The axis, however, does not necessarily cease to grow, after developing sporangia: there is frequently subsequent growth by innovation, and a new leafy shoot is formed which later on develops its own series of reproductive organs. This process may be repeated several times. (Plate I., fig. 1.)

The sporangiferous shoot consists of a short cylindrical stalk, about of an inch in length, terminating in a pair of leaves of the normal type with or without the mucro. On the inner face of the shoot and at the base of these leaves is carried the sporangium. The latter when fully developed is a brown spindle-shaped structure pointed at the ends, and constricted in the middle where a transverse septum divides its cavity into two chambers. It is said that occasion

ally the constriction is almost absent, and that sometimes there are three chambers as in Psilotum1.

Viewed from the side the two portions are seen to be set at a sharp angle to one another in the full-grown sporangium. When ripe the sporangium opens by a longitudinal slit down the middle line on the side towards the axis. (Plate I., fig. 4a.)

The morphological value of these shoots has been a matter of much discussion. Up to a certain date almost all botanists had regarded the whole structure as a morphological leaf, in which a divided sterile portion carried an epiphyllous sporangium on its upper surface. This was the view of R. Brown2, Hugo von Mohl3, A. Brongniart, Springs, Kickx, and Mettenius."

Lüerssen compared the sporangium to the multilocular sporangium of the eusporangiate ferns such as Marattia; and Prantl' regards the sporangium as a modified leaf lobe, referring in support of this view to the occurrence of trifid leaves in Psilotum.

Celakovsky 10 has also described the sporangium-bearing organs as morphological fronds, comparing them in their division into barren and fertile portions with those of the Ophioglosseæ.

Juranyi," relying on developmental appearances, was the first to claim for them the value of special shoots with an apical fruit and a pair of leaves growing from below it.

Strasburger1 defends the same view with certain modifications; and Goebel1 has also adopted it, and stated it in his text-book as proved.

1 Thompson, New Zealand Ferns, Dunedin.

* Robert Brown, Prod. Flor. Nov. Holl.

* H. von Mohl, Morph. Betrachtungen über die Sporangium der mit Gefässen Versehenen Cryptogamen, 1837.

1849.

Hist. des Végétaux Fossiles, Paris, 1836.

'Monographie des Lycopodiacées," II., Mem. L'Acad. Roy. Belg., xxiv.,

"L'organe reproducteur du Psilotum triqetrum," Bull. Acad. Roy. Belg., ser. ii., vol. 29, 1870.

men.

"Ueber Phylloglossum," Bot. Zeitung, 1867.

"Die Farne der Samoa Inseln."

Prantl, "Bemerk. über die Verwandshaft verhältnisse des Gefäss-CryptogaVerhand. d. Physik. Medicin. Gesellsch. Wurtzburg, vol. ix., 1876.

10 Celakovsky, Abh. des Böhmischen Gesellschaft des Wissenschaften, vi. Folge, Bd. ii., 1882.

"Juranyi, Bot. Zeitung, 1871.

12 Bot. Zeitung, 1873.

13 Grundzüge des Syst. u. Spec. Pflanzenmorphologie, Leipzig, 1882.

Eichler and Bertrand also look upon the sporangium-bearing organ as a special shoot.

H. Graf v. Solms-Laubach2, after investigating the development of Psilotum, more especially in Ps. flaccidum, comes to the conclusion that there is nothing to justify these later views, and returns to the older opinion that the whole structure is a phyllome.

So far as our observations on the apex of Tmesipteris have been carried, we are inclined to adopt his view of the case. It is unfortunate that the conclusions drawn by various botanists as to the morphological value of the "sporangiferous shoots" in this group of plants, should have been founded almost entirely on observation of Psilotum rather than of Tmesipteris: and it seems to us that if the latter had been equally studied, recent opinion would have gone more in favour of the views of Solms-Laubach.

The phenomena seen in the development of the lateral outgrowths round the apex, as described by Solms-Laubach for Psilotum, and as sketched below for Tmesipteris, do not offer any grounds of objection to the phyllome theory.

It has to be remembered that while the habit of branching is characteristic of Psilotum, it is at a minimum in Tmesipteris. Moreover, at the apex of Tmesipteris ordinary sterile leaves occur among the fertile "shoots," rising from the growing point in a manner indistinguishable from that of the early stages of the latter. The decurrent ridge running back along the axis from the insertion of the foliage leaf is present in exactly the same form below the "sporangiferous shoot"; indicating that the true base of the leaf is far below the level of the sporangium.

Lastly, if the tissue of the "stalk" be examined it is found that its outer layers consist of that well-marked lacunar parenchyma seen in the leaf, and continuous with that of the leaves above. It is only just round the fibro-vascular bundle that the cells have the straightwalled close-fitting character of the stem ground tissue. The lacunar parenchyma is continued down the decurrent ridge just mentioned; but there is no trace of it in the tissue of the main axis, even where this shows both stomata and chlorophyll.

On histological grounds therefore, it would seem that if a separate sporangium-bearing shoot is present at all, it is surrounded by and coalescent with the base of the pair of associated leaves. There does not, however, seem to be sufficient reason for supposing the

1 Bertrand, loc. cit. 2 Solms-Laubach, Ann. du Jardin Bot. Buitenzorg, 1882.

existence of any morphological branch in the structure: it seems to be simply a leaf modified at the base in a direction which would ultimately bring about some such distinction between rachis and pinna as is seen in the ferns.

The preservation of distinctive leaf tissue in the "stalk" indicates that the latter is in a transition stage. The phylogeny of the Psilotaceæ is a question for future study to elucidate; but it seems possible that the structure of the sporangiferous shoots might result from a tendency, on the part of the leaf in a simple Lycopodium type, to close round the base of the axillary sporangium, and so gradually separate it from the main axis.

On examining the growing apex of a fertile stem, the early stages of its lateral outgrowths can be seen with great distinctness. Before the development of the sporangia, there seems to be a period of cessation or retardation of apical growth, so that, on looking at the end of a leafy stem, the termination often appears wide, instead of tapering, and leaves are well developed right to the top. (Plate IV., fig. 28.) The uppermost group of leaves curve in over the apex and protect it. By removing those uppermost leaves the actual growing point can be seen at their base, surrounded by whorls of rounded knobs, the young sporangiferous shoots. Later on the growth in length becomes rapid, and the apex assumes a tapering form; the developing shoots being separated by wider intervals. (Plate IV., fig. 27.) On examining more closely the growing region in either case it will be observed that the extreme apex is wide and dome-like, and that rounded projections grow out almost immediately behind it. They are spirally arranged, but so closely approximated as to appear like whorls. The elements of the first two, or sometimes three, cycles retain the simple form, but those succeeding show a well-marked differentiation into three lobes. One lobe is turned toward the axis, and afterward develops into the sporangium; the other two are almost from the first narrower, and soon grow beyond the third and develop into leaves. At first, however, the sporangium lobe is distinctly larger, and seems to form the end of the shoot. The sporangium lobe subsequently elongates in a direction at right angles to the main axis. of the plant, and becomes divided by a wall in a plane nearly parallel to that axis. The two chambers are at first anterior and posterior, that is, rather than upper and lower. Subsequent growth of the shoot behind it turns the sporangium forward, so that its long axis becomes vertical, and its septum horizontal.