Fossil plants, Vol. 2

The term Botryopterideae, first used by Renault, has been applied to a group of Palaeozoic ferns ranging from the Lower Carboniferous to the Permian and containing several genera, the distinguishing features of which are supplied by the anatomical structure of the stems or, in many cases, by that of the petiolar vascular strand. Scott[1109] subdivides the Botryopterideae into the Botryopteris and the Zygopteris sections. In an admirable monograph recently published by Paul Bertrand[1110] considerable changes are proposed in current nomenclature; he substitutes the name Inversicatenales for Botryopterideae, a designation, which as Scott remarks, is “probably too technical to command general acceptance.” A more serious criticism is that the name Inversicatenales has reference to a character (the inverse curvature of the leaf-trace in relation to the axis of the stem) which is by no means universal in the group[1111].

In the following account, necessarily incomplete, the generic terminology of Bertrand is adopted, but this decision does not carry with it any obligation to accept the name Inversicatenales. We may speak of the types of Palaeozoic ferns dealt with in the following pages as members of a group differing in many respects from any existing genera of the Filicales, and exhibiting the characteristics associated with generalised plants. Williamson, as early as 1883, spoke of Renault’s Botryopterideae as comprising “altogether extinct and generalised” types[1112]. For these generalised Palaeozoic ferns I propose to use the name Coenopterideae[1113]. This term may be adopted in a wider sense than Renault’s name Botryopterideae. The name Primofilices proposed by Arber[1114] might be employed, but the implication which it carries is an argument against its adoption. We have not yet reached a stage in the investigation of extinct types at which we are able to recognise what are actually primary or primitive ferns. The search for origins will continue; as new discoveries are made our point of view shifts and the primitive type of to-day may to-morrow have to take a higher place. The epithet primitive or primary is in reality provisional: to adopt such a name as Primofilices suggests a finality which has not been, or is likely to be, achieved. The true ancestral type—the Urform—which we strive to discover eludes the pursuer like a will-o’-the-wisp.

Seeing that the number of true ferns of Palaeozoic age has been recently considerably reduced and is likely to suffer further reduction, the consideration of such undoubted Carboniferous and Permian examples of the Filicales as are left acquires a special importance. In the first place it is natural to ask whether the Palaeozoic ferns include any types which, if not themselves ancestral forms, may serve to indicate the probable lines of evolution of existing families. It is probable that in the near future our knowledge of the Coenopterideae will be considerably extended; as yet we possess meagre information in regard to those characters on which most stress has generally been laid in the classification of recent ferns, namely the structure of the spore-bearing organs. The sporangia of Diplolabis and Stauropteris (figs. 309, A; 322) are exannulate; in the former genus they occur in sori or synangia consisting of a small number of sporangia, while in the latter they are borne singly at the tips of ultimate ramifications of a highly compound leaf. The resemblance of the synangium of Diplolabis to that of Kaulfussia (fig. 245, C) is not shared in an equal degree by the sporangia of Stauropteris, which are in some respects comparable with those of the Ophioglossaceae. In the Zygoptereae, or at least in the case of such fertile fronds as are known, and in Botryopteris (fig. 319), the sporangia occur in groups, and the pedicel of each sporangium is supplied with vascular tissue as in Helminthostachys. Another characteristic of the sporangia of the extinct types is the possession of an annulus several cells in breadth, a peculiarity which supplies a point of contact with the Osmundaceae. In the sporangia of Kidstonia we have a similar though not an identical type (fig. 256, E, p. 340). So far, then, as the evidence afforded by sporangial characters is concerned, it points to comparison with the Ophioglossaceae, the Osmundaceae, and the Marattiaceae. When we compare the steles of the stems we find a wide range of structure. All the genera agree in being monostelic; in Tubicaulis and Grammatopteris the protoxylem is exarch, in Botryopteris it is internal, while the foliar strand of Stauropteris and the stele of Ankyropteris corrugata are mesarch. The axillary branching of species of Ankyropteris suggests comparison with the Hymenophyllaceae.

The investigation of the vascular system of the petioles has afforded results which in the hands of P. Bertrand have led to conclusions in regard to inter-relationships. We must, however, not overlook the danger of attributing can excessive importance to this single criterion and of neglecting the facts of stem anatomy.

I. Botryoptereae.

Grammatopteris.

Renault instituted this genus for petrified stems from the Permo-Carboniferous beds of Autun. Grammatopteris Rigolloti[1115], the type-species, is represented by a fragment, 12–15 mm. in diameter, surrounded by crowded petioles characterised by a vascular strand in the form of a short and comparatively broad plate with the smallest tracheae at each end. The solid xylem of the stem stele (protostele) has peripheral groups of protoxylem. Nothing is known as to the form of the leaves, but sporangia similar to those of Etapteris (Zygopteris) were found in association with the stem. It is possible, as P. Bertrand suggests, that Renault’s species may be the stem of a Tubicaulis.

Tubicaulis.

Tubicaulis solenites (Sprengel)[1116]. Fig. 304.

This species from the Lower Permian of Saxony has been fully described by Stenzel[1117]. It is characterised by a very slender erect stem bearing numerous spirally disposed leaves associated with adventitious roots; the single stele (protostele) consists exclusively of tracheae, described as intermediate between the scalariform and reticulate type, surrounded by phloem. Leaf-traces are given off from the periphery of the stele where groups of smaller elements occur; these have the form of a wide-open U-shaped strand with the base of the U facing the axis of the stem. As the trace passes out towards the leaves, the ends of the U become more or less incurved. The stem is said to reach a metre in length and to bear compound fronds a metre long. The orientation of the leaf-trace with its concavity turned outwards is in striking contrast to the relation between leaf-trace and stem in recent ferns.

Tubicaulis Sutcliffii, Stopes[1118].

In this species the vascular axis, 2 mm. in diameter, is almost cylindrical and of the protostelic type with the protoxylem “near to or at the edge”: the tracheae are scalariform or reticulate. The leaf-traces, when first separated from the edge of the stele, are oval and gradually assume the curved form seen in T. solenites (fig. 304) with the convex side towards the axis of the stem. The transition from the scalariform to the reticulate type of pitting on the tracheal walls referred to by Miss Stopes has also been noticed in some recent ferns (e.g. Helminthostachys) and in Sigillaria (fig. 200, C, p. 212). The fact that the scalariform type of pitting is practically universal in the xylem of recent ferns, would seem to show that this character has been acquired in the course of evolution and retained in preference to the reticulate form characteristic of several Palaeozoic species. The distinction between the two methods of pitting is one of little phylogenetic importance.

Botryopteris.

This genus, founded by Renault on a specimen from Autun, is represented in the Lower Coal-Measures of England by Botryopteris hirsuta (= Rachiopteris hirsuta Will.), B. ramosa (= R. ramosa Will.[1119]) (fig. 306) and B. cylindrica (fig. 305), also by B. antiqua (fig. 307) from the Culm of Pettycur, Scotland.

An important characteristic of the genus is the solid stele of the stem which agrees with that of Tubicaulis and Grammatopteris, except in the central or peripheral position of the smallest tracheae.

Botryopteris forensis Renault[1120]. Figs. 309, B; 319, D–G.

The stem of this species from the Upper Carboniferous of St Étienne is 1·7 cm. x 7·5 mm. in diameter. The solid stele consists of reticulate tracheae with the smallest elements on the outer edge. The comparatively broad cortex of the type-specimen is traversed by a leaf-trace in an almost vertical course and by vascular strands passing horizontally to roots. The petioles are circular in section and their vascular strand has the form of an ω in transverse section (fig. 319, G), the three projecting arms pointing to the axis of the stem. Both stem and leaves bore large multicellular hairs, spoken of by Renault as equisetiform because of the finely toothed sheaths of which they are composed. The compound fronds had fleshy lobed pinnules with dichotomously branched veins (fig. 309, B); stomata are said to be confined to the upper surface, an observation which leads Renault to describe the plant as aquatic on evidence which is hardly convincing.

The pyriform and pedicellate sporangia are borne in groups of two to six on the ultimate divisions of the frond; the wall is composed of two layers of cells and on one side of the sporangium is an annulus several cells in breadth (fig. 319, D, F). An interesting type of sporangium described by Oliver[1121] from Grand’Croix in France may, as he suggests, belong to Botryopteris forensis; the differences between Renault’s and Oliver’s specimens being the result of the more perfect preservation of the tissues in the latter. The sporangium described by the English author is circular in section and measures 0·65 × 0·53 mm.; the wall is in part composed of a single layer of cells and in part of two to three layers, a character recalling the “annulate” sporangia of Botryopteris. Between the spore-mass and the wall is an interrupted ring of short tracheal elements similar to the xylem-mantle which occurs at the periphery of the nucellus of certain Palaeozoic gymnospermous seeds. In the absence of proof of a connexion between this sporangium and Botryopteris it is convenient to use the generic name Tracheotheca subsequently proposed by Oliver[1122]. In the recent ferns Helminthostachys and Botrychium, and, as Oliver notices, in the microsporangia of the Australian Cycad Bowenia spectabilis, vascular strands extend almost to the sporogenous tissue, but the fossil sporangium is unique in having a tracheal layer in immediate contact with the spores. These xylem elements may, as Oliver suggests, have served the purpose of conveying water to the ripening spores.

Botryopteris hirsuta (Will.)[1123].

This English species has a slender axis bearing numerous leaves with petioles equal in diameter to the stem. The surface of the vegetative organs bears large multicellular hairs. The leaf-traces resemble those of B. forensis, but the projecting teeth which terminate in protoxylem elements are less prominent than in the French species; the petioles were named by Felix Rachiopteris tridentata[1124]. As a leaf-trace passes into the stele of the stem the three protoxylem strands unite and take up an internal position in the solid stele. The stele may, therefore, be described as endarch. The small tracheae at the edge of the stele supply the xylem strands of adventitious roots.

Sporangia similar to those of B. forensis have been found in association with the English species.

Botryopteris cylindrica (Will.). Fig. 305.

A plant originally described by Williamson[1125] from the Lower Coal-Measures of England as Rachiopteris cylindrica (fig. 305) and afterwards more fully dealt with by Hick[1126], has a slender stem with a cylindrical stele characterised by well-defined central protoxylem elements in one or two groups. The leaf-traces are semi-lunar in section with the protoxylem on the flatter side. The stele of Botryopteris cylindrica (fig. 305, A) is more cylindrical in section than that of B. ramosa (fig. 306) and shows more clearly the differentiation into smaller central and larger peripheral tracheae. In the section reproduced in fig. 305, B the stele is giving off a branch almost identical in structure with the main vascular axis. Scott[1127], in referring to the inclusion of this type in the genus Botryopteris, expresses the opinion that its habit must have been very different from that of other species, and he suggests the institution of a new genus.

Botryopteris ramosa (Williamson). Fig. 306.

This species, which bears a close resemblance to Botryopteris hirsuta, was originally described by Williamson from the Lower Coal-Measures of England as Rachiopteris ramosa[1128], the specific name being chosen on account of the numerous and crowded branches given off from the main axis. The section shown in fig. 306, A, illustrates Williamson’s description of the stem as being “always surrounded [when seen in transverse sections] by a swarm of similar sections of the large and small branches, though of varying shapes and sizes.” The stele is composed of a solid and more or less cylindrical rod of xylem tracheae of the reticulate type surrounded by phloem (figs. A and D): one or more internal groups of smaller protoxylem elements occur in an approximately central position (fig. A, px). The stele is in fact endarch like those of Selaginella spinosa and Trichomanes reniforme, a feature which, as Tansley[1129] believes, probably entitles the vascular axis to be considered a primitive form of protostele. In the specimens represented in fig. 306 the phloem and inner cortical tissues were almost completely destroyed before petrifaction. The thick-walled outer cortex bears at its periphery numerous multicellular hairs. Some of the xylem strands given off from the stele no doubt supplied adventitious roots, but in most cases the outgoing branches are leaf-traces and the numerous sections of axes of different sizes seen in fig. A point to a repeated subdivision of the crowded fronds. The structure of a petiole is shown in figs. C and D. As seen in fig. C, the oval vascular strand has three protoxylem groups, px, on its flatter side; a well-defined epidermal layer is shown at e in fig. C.

Fig. B shows at a a section of a leaf-axis in the act of branching and the row of branchlets at b represents a further stage in subdivision. At sp in fig. A the section has cut through a single sporangium characterised by a group of larger (“annulus”) cells on one side of the wall.

This slender fern with its numerous repeatedly branched leaves may perhaps have lived epiphytically on more robust plants.

Botryopteris antiqua, Kidst. Fig. 307.

This species, recently described by Kidston[1130] from the Culm of Pettycur near Burntisland, is represented by sections of a small stem with a cylindrical stele 0·40 mm. in diameter composed entirely of scalariform tracheae without any recognisable protoxylem. The petioles are larger than the stem; the meristele (fig. 307) is oval with protoxylem elements on the slightly more rounded adaxial face. As Kidston suggests, this stem may belong to a scrambling plant which required support to bear its relatively large leaves. An interesting feature is the absence of projecting teeth in the leaf-trace, a character in marked contrast to the ω form assumed by the petioles of Botryopteris forensis (fig. 319, G) and B. hirsuta. This leads Kidston to suggest that the vascular strand of the petiole tends “to become more simple ... as traced back in geological time.” The greater similarity in this species between the stele of the stem and that of the petiole is probably another mark of a more primitive type.

In these three types, Grammatopteris, Tubicaulis, and Botryopteris, we have monostelic plants, for the most part of very small size, with leaf-traces varying in shape from the oblong band-form in Grammatopteris, and the oval form of Botryopteris antiqua, to the ω type represented in its most pronounced form by B. forensis. In several species the stem stele is endarch. Our knowledge of the leaves is very meagre: in B. forensis they were repeatedly branched and apparently bore small fleshy pinnules; the sporangia, though differing from those of recent ferns, may be compared with the spore-capsules of Osmundaceae as regards the structure of the annulus. The abundance of hairs on the stems and leaves of some species, the tracheal sheath in the sporangium described by Oliver[1131] as Tracheotheca (= Botryopteris?), and the apparent absence of a large well-developed lamina, may perhaps be regarded as evidence of xerophilous conditions.

II. Zygoptereae.

Corda[1132] proposed the generic name Zygopteris for petrified petioles from the Permian of Saxony, included by Cotta in his genus Tubicaulis, which he named T. primarius. Corda’s genus has been generally used for petioles of Palaeozoic ferns characterised by a vascular strand having the form of an H in transverse section (fig. 308, D). Since the generic name was instituted, information has been obtained in regard to the nature of the stems which bore some of the petioles of the Zygopteris type; and for other species of Zygopteris, the stems of which are still unknown, new generic names have been proposed. P. Bertrand[1133] retains Zygopteris for one species only, Z. primaria. Fig. 308, D, shows the character of the petiolar vascular strand; the chief points are the comparatively long cross-pieces (antennae of P. Bertrand) inclined at an angle of 45° to the plane of symmetry of the petiole axis, and the groups of protoxylem elements shown by the white patches in fig. D. In this as in other members of the Zygoptereae the main rachis of the leaf gives off four sets of branches in pairs alternately from the right and left side of the primary vascular axis. This method of branching of the stele in the primary rachis of several members of the Coenopterideae shows that the fronds bore pinnae laterally disposed, in some cases in one row and in others in two rows on each side of the rachis. In a typical fern frond, as represented by recent and most fossil species, branching of the rachis occurs in the plane of the frond, that is in the plane represented by the horizontal arm of xylem in Zygopteris primaria connecting the two antennae or cross-pieces. In the Zygoptereae the branches from the petiole vascular axis lie in a plane at right angles to that of the frond; they lie in the transverse and not in the horizontal plane. The two strands shown in fig. 308, B, 4, have been formed by the division of a single strand, 3, in the transverse plane (i.e. in the plane of the paper). As Tansley[1134] points out, a type of branching superficially similar to, though not identical with this, is seen in some recent species of Gleichenia and Lygodium. In this connexion it is worthy of note that a fern figured by Unger from Thuringia as Sphenopteris petiolata Goepp[1135] bears pinnae in two rows on the rachis which are characterised by repeated branching and by a very narrow lamina or by slender naked axes; the occurrence of this form of frond in rocks containing Clepsydropsis antiqua (fig. 308, A) suggests a possible connexion between the petrified rachis and the impressions of the leaves.

The vascular strand of the rachis of Zygopteris primaria (fig. 308, D) is simpler than that of most of the Zygoptereae and exhibits a close resemblance to the type of strand described by Renault as Diplolabis (fig. 308, C).

Diplolabis.

Renault[1136] instituted this genus for two species from the Culm beds and Coal-Measures of France based on the structure of the petioles. The stems are unknown. The main rachis has a stele similar to that of Zygopteris primaria, but distinguished by its greater similarity, in transverse section, to an X rather than to the letter H: the long transverse bar in Zygopteris is here much reduced in size. The petiole of Diplolabis forensis[1137] Ren. (fig. 308, C) has a diameter of 1·5–2 cm. From the antennae a pair of small bundles is given off alternately from the right and left side, as in Zygopteris; the members of each pair coalesce after leaving the antennae and then separate to pass into the lateral branches of the frond. The position of the protoxylem and the formation of the lateral xylem strands previous to their separation are shown in fig. 308, C. On the side of the vascular strand shown in fig. C, 2, the two lateral extensions of the antennae are converging towards one another previous to their separation and subsequent union. The ovoid sporangia occur in groups of three to six and are coalescent below with a central receptacle; they have no annulus, but the cells on the side next the receptacle are smaller than those on the external wall (fig. 309, A). The synangial form of the sorus suggests comparison with Marattiaceae.

The species described by Renault from the Culm of Esnost is regarded by P. Bertrand as identical with that described by Solms, from the Culm of Falkenberg, as Zygopteris Roemeri[1138]. Diplolabis is compared by P. Bertrand with Metaclepsydropsis, the generic name given to the Lower Carboniferous petiole described by Williamson as Rachiopteris duplex[1139].

Mr Gordon has recently described in a preliminary note a new type of stem stele under the name Zygopteris pettycurensis from the Lower Carboniferous plant bed of Pettycur[1140]: he regards the petioles attached to the stem as identical with Zygopteris Roemeri Solms-Laubach[1141]. This species, founded by Solms-Laubach on petioles only, is placed by Bertrand[1142] in the genus Diplolabis and regarded as identical with D. esnostensis Ren. The stele found by Mr Gordon may therefore be assigned to the genus Diplolabis: it includes two regions composed exclusively of tracheae and is cylindrical in transverse section. The inner xylem zone consists of short, square-ended, reticulately pitted elements and the outer zone is composed of long and pointed conducting tracheae. The scalariform protoxylem elements are situated between the two metaxylem zones. As Mr Gordon says: this type of stem occupies a position “in the Zygopteroid alliance” corresponding to that which Thamnopteris Schlechtendalii (p. 329) occupies in the Osmundaceous series. The discovery of this stem supplies another link between the two fern groups, Osmundaceae and Coenopterideae. Pelourde[1143] has described an imperfectly preserved vascular strand from a locality near Autun as the type of a new genus Flicheia esnostensis. Mr Gordon has pointed out to me that this is a partially rotted petiole of Diplolabis esnostensis (= Zygopteris Roemeri).

In their recent account of fossil Osmundaceous genera, Kidston and Gwynne-Vaughan[1144] speak of the central parenchyma of the existing medullated stele as being derived from tracheal tissue. They add that if the Zygopteroid line of descent is at all close to the Osmundaceous, we must be prepared for the existence of a Zygopteris with a solid xylem like that of Thamnopteris: “such a discovery, in fact, we hopefully anticipate[1145].” The new Pettycur stem amply justifies this prophecy. It is noteworthy that Mr Gordon’s stem affords an instance of the occurrence of a type of stele, similar in its cylindrical form and in the absence of parenchyma to that of Botryopteris, in a plant bearing leaves characterised by the Zygopteris type of vascular strand.

Metaclepsydropsis duplex (Will.) fig. 310, A[1146]. [= Rachiopteris duplex, Williamson 1874. Asterochlaena (Clepsydropsis) duplex, Stenzel 1889. Clepsydropsis, Renault 1896.]

The vascular axis of the main axis of the frond is characterised by the hour-glass shape of the xylem which consists entirely of tracheae, most of which are reticulately pitted. In a transverse section (fig. 310, A) the two ends of the stele are dissimilar; at one end of the long axis is a small bay of thin-walled tissue (phloem) enclosed by a narrow band of xylem, and at the other the bay is open and has two protoxylem groups. The latter represents the earliest stage in the production of secondary bundles: at a later stage the bay is closed by the elongation of the edges, the enclosed group of phloem is vertically extended, and the protoxylem strands are more widely separated. The curved band of xylem becomes detached as a curved arc and divides into two (fig. 310, A). In a single section of this species one often sees several strands of xylem enclosed in a common cortex with the main vascular axis; these are the xylem bundles of lateral pinnae. Metaclepsydropsis duplex shows the method of branching of the petiole vascular axis which has already been noticed in Diplolabis and Zygopteris. In reference to this feature, Williamson wrote in 1872—“I know of no recent fern in which the secondary branches of the petiole are thus given off in pairs, which pairs are distichously arranged on the primary axis, and each of which secondary petioles sustains ternary ones arranged distichously.” By slightly altering the primary stele of this type of frond, by narrowing of the constricted portion of the hour-glass and extending the lateral groups of xylem obliquely upwards, the form of stele shown in fig. 310, A, would be converted into the Diplolabis type (fig. 308, C).

Clepsydropsis.

Unger[1147] instituted this genus as a subdivision of Corda’s family Rhaciopterideae[1148], the name having reference to the hour-glass form of the vascular axis[1149]. The type-species C. antiqua (fig. 308, A) is spoken of as the commonest fossil plant in the Devonian rocks of Thuringia. In some sections the xylem has the form seen in fig. 308, A, in which an invagination of thin-walled tissue occurs at each end; in other sections (fig. 308, A′) the bays become islands in the xylem. Solms-Laubach speaks of Unger’s species as Rachiopteris (Clepsydropsis) antiqua. P. Bertrand[1150], who has recently described Unger’s plant, while recognising that C. antiqua and Metaclepsydropsis duplex closely resemble one another, draws attention to certain differences in the structure of the xylem which he regards as sufficient to justify a generic separation. The leaf-traces of Clepsydropsis are described by Bertrand as almost circular closed rings of xylem instead of an arc as in Metaclepsydropsis.

Ankyropteris.

Stenzel adopted this name for a subdivision of Corda’s genus Zygopteris, applying it to a species described by Renault as Z. Brongniarti, to a Permian species described by himself as Z. (Ankyropteris) scandens, and to Z. Lacattii Ren.; Rachiopteris Grayi Will. and Rachiopteris corrugata Will. are also included in this genus. The characters emphasised by Stenzel[1151] are (i) the double anchor-like form of the H-shaped petiole strand in which the lateral arms (antennae) are curved like the flukes of an anchor, and (ii) the emission of four rows of branches instead of two. The latter distinguishing feature no longer holds good, as Z. primaria also gives off four rows of bundles and not two as Stenzel supposed. P. Bertrand has adopted Stenzel’s genus in a narrower sense[1152].

Ankyropteris scandens Stenzel[1153]. Fig. 310, D.

This Lower Permian species is very similar to or perhaps identical with Ankyropteris Grayi (Williamson). The stem of

A. scandens was found in association with the roots of a Psaronius stem evidently petrified in situ as it burrowed, like Tmesipteris, tropical aroids, and other recent plants, among the living roots of the tree-fern. The stem, 10–11 mm. in diameter, bore fronds with an H-shaped vascular strand, small scale-leaves, and adventitious roots. The stele consists of a five-angled cylinder of scalariform tracheae surrounding an axial strand of parenchyma containing scattered tracheae of smaller diameter. This axial tissue extends as a narrow strip into each of the short and obtusely truncated arms (cf. fig. 311). A striking feature is the production of a shoot in the axil of the foliage-leaves (fig. 310, D), a manner of branching characteristic of Trichomanes (see page 365).

Ankyropteris Grayi (Will.). Fig. 311.

In describing this species, Williamson wrote—“That no classification of these fossil ferns based solely upon transverse sections of the petiolar bundles is or can be of much value, is clearly shown when tested amongst those living ferns the classification of which is chiefly based upon the sporangial reproductive organs[1154].” This is a view entirely opposed to that which inspires P. Bertrand’s recent monograph. Whether the value attached to the vascular structure of petioles as a basis of classification is upheld or not, it is noteworthy that since Williamson expressed his opinion, our knowledge of the anatomy of ferns and of the value of anatomical evidence has enormously increased. The slender stem[1155] of this Lower Coal-Measures species agrees closely with that of A. scandens; it bore spirally disposed fronds, scale-leaves, and roots. The stele has the form of an irregular five-rayed star (fig. 311) in which the relative length of the arms varies in different sections owing to the separation of the distal ends to form leaf-traces. The axial region is composed of parenchyma and associated narrow tracheae, as in A. scandens. The xylem, with protoxylem elements at the ends and especially at the angles of the arms, is completely surrounded by phloem. The cortex consists internally of parenchyma which becomes thicker-walled towards the periphery and bears multicellular epidermal hairs. A leaf-trace is detached in the form of a triangular strand and is formed by the tangential extension of the distal end of an arm of the stele. The trace, on its way through the cortex, divides into two; the outer branch gradually changes from a slightly curved band to an H-shaped meristele; the inner branch, which supplied an axillary shoot, is similar to the stele of the stem, but smaller. Scott[1156] has recently recorded the occurrence of scale-leaves (aphlebiae) in this species like those described by Stenzel in A. scandens.