FLORA OF TROPIGAL AFRICA. FLORA OF teORICAL ARRIC A. EDITED BY SIR DAVID PRAIN, C.M.G., C.J.E., LL.D., F.R.S. DIRECTOR, ROYAL BOTANIC GARDENS, KEW. VOL. VI.—SECTION 2. ULMACEH TO CYCADACE SH. PUBLISHED UNDER THE AUTHORITY OF THE SECRETARY OF STATE FOR THE COLONIES. LONDON: LOVELL REEVE & CO., Limirsp, Publishers to the Wome, Colonial and Lndian Grobernments, DATES OF PUBLICATION OF THE SEVERAL PARTS OF THIS VOLUME. Part I. pp. 1-192 was published March, 1916. no Ales ay 198-972 e November, 1917. PREFACE. —_—~~e It was found necessary in 1904, for reasons then explained by the editor, to divide Volume IV. of the “ Flora of Tropical Africa ”’ into two sections. The same expedient has had to be adopted in the case of the present volume. The preparation of two-fifths of this final section of Volume VI. has been due to the kind collaboration of Dr. A. B. Rendle, F.R.S., Keeper of Botany, British Museum. Members of the Kew staff have supplied the rest of the text, with the exception of the account of the Gnetaceew. A melancholy interest attaches to that account ; it is one of the last contributions to botanical knowledge made by its gifted author, the late Professor H. H. W. Pearson, F.R.S., of Cape Town, who had bestowed especial attention on this group of plants. The results of recent study have created a difficulty with regard to the position the G'netacee should occupy in a conspectus of natural orders. In other floras of the series to which this work belongs, that order, which includes the living representatives of the class Gnetales, has been placed among the Gymnosperme. The Gnetales, however, exhibit characters that do not wholly conform with the conception of the Gymnosperme which is usually accepted. In 1908 Arber and Parkin expressed the opimion that the Gnetales are “ gym- nosperms nearly related to angiosperms,” derived from a common ancestral stock, the Hemiangiosperme. In 1912 Lignier and Tyson advanced arguments in favour of the treatment of the Gnetales as an early, highly specialised branch of the Angiosperme, which has retained many features inherited from pre-angiospermous ancestors. The attempt to resolve this and other cognate difficulties has been materially assisted by the valuable advice and practical help of Dr. O. Stapf, F.R.S., Keeper of the Herbarium, Kew. The ¥ vi PREFACE. circumstance that in the Gnetacee the character of “ nakedness of the ovule’’ is manifest, admits the formal retention of the order in the Gymnosperme, and allows the accordance to it, in our ordinal conspectus, of a serial position which conforms with the practice adopted in other floras of the series in which the “ Flora of Tropical Africa ” finds a place. In the definition of the Gnetales, prepared for this conspectus by Dr. Stapf, due allowance is, however, made for the peculiar and ambiguous position of the class. The terminology employed in describing the organs of reproduction is intended to reflect our belief that, in spite of the nakedness of the ovule, the affinities of the Gnetales are with the angiospermous rather than with the gymnospermous phylum. Associated with this difficulty as to the nine of the Gnetales is the more serious one created by the modern view as to the relation- ship which these two great phyla bear to each other. The phe- nomenon of gymnospermy, first pointed out by R. Brown in 1826, was turned to taxonomic account by Brongniart as early as 1828. But when Brongniart recognised in the Gymnosperme a distinct natural group, he regarded that group as an integral part of the Dicotyledones. Tt was not until 1864 that A. Braun proposed the treatment of the Gymnosperme as a division of the Phanerogame equal in status to the Dicotyledones and the Monocotyledones taken in conjunction. The evidence in favour of the view, adopted by Hooker in 1876, that the two primary phanerogamous divisions, Angiospermae and Gymnosperme, represent quite distinct lines of evolution, is so strong that this disposition is now generally accepted. The series of Colonial Floras was initiated in 1859, and the first work in this series was issued in 1861, while the view expressed by Brongniart in 1828 was still generally accepted. In that flora the Gymnosperme were accordingly regarded as a subclass of the Dicoty- ledones, and were assigned a serial position immediately after the really dicotyledonous subclasses and immediately before the mono- cotyledonous ones. The view that the Gymnosperme are a subclass of the Dicotyledones has not been maintained in every Colonial Flora completed since 1861. In one instance the Gymnosperme have been regarded as a group, within the Dicotyledones, equal in status to the whole of the angiospermous subclasses of that class. In other cases the Gymnosperme have been advanced to the rank of a distinct class, comparable in status with the Monocotyledones on the one hand, the Dicotyledones on the other. Neither of these two modifications of the original view necessitates a change in the PREFACE. vii serial position assigned to the Gymnosperme in 1861, which has persisted in those Colonial Floras published since that year. The acceptance of the modern view as to the relationships of the Gymnosperme no doubt justifies the action of those who would modify that serial position. But what may be lawful is not neces- sarily expedient. Considerations of practical convenience render it desirable to adhere in this work to the serial arrangement adopted in 1861. The maintenance of that arrangement derives considerable support from the action taken by Bentham and Hooker in the “Genera Plantarum” in 1880. These authors, while presenting the Gymnosperme in such a fashion as to indicate that this group is not to be regarded as similar in status to the Dicotyledones and the Monocotyledones, have nevertheless accorded it a serial position between the two classes of Angiosperme. In following that example here, an adequate safeguard against any possible misunderstanding is provided by the subjoined synoptic statement, prepared by Dr. . Stapf, wherein the natural position and the divisions, according to modern views, of the Gymnosperme that occur in tropical Africa, are clearly indicated :— Division PHANEROGAMZE (Siphonogame). Subdivision A. ANGIOSPERMZ. Class 1. Dicotyledones. Class IT. Monocotyledones. Subdivision B. GYMNOSPERM A. Class [. Gnetales. Order i. GNETACER. Class IT. Coniferales. Order i. PINACEZ. Order ii. TAXACEZ. Class JIT. Cycadales. Order i. CYCADACE®. The admission of the Gymnosperme to the rank of a phylum distinct from that to which the rest of the Phanerogame belong, creates another difficulty which is harder to settle than the one disposed of above. This difficulty is concerned with the appli- cability to the Gymnosperme of the conventional terminology applied to the “flower.” Attempts to maintain the old uniformity of terms have been frequent; these attempts have proved far from successful. Those who lay most stress upon function have found FL. TROP. AFR. VI, SECT. II. a2 vill PREFACE. themselves impelled to extend the incidence of the term “ flower,” and of the terms for various parts of the flower, so as to render these applicable to all the phyla that are connected by homologies within the sexual sphere of reproduction. As a logical consequence even the sporangiophore of a horsetail and the fertile frond of a fern, thus become “ flowers.” The value of this procedure as a means of correlating what might otherwise appear wholly unrelated is doubtless great. But, even where considerations of a practical nature may be neglected, we soon come to realise that such endeavours to attain uniformity are attended by a double disadvantage. They blunt the perception of the very divergent lines along which the great plant-phyla have been evolved ; they obscure the appreciation of the correspondingly different structures in which the evolution of these phyla has resulted. Those practical considerations which condition the preparation of a technical work like the present flora cannot, however, be dis- regarded here. For purposes of discrimination and of classification it is of the utmost consequence that fundamental differentiations of structure be expressed in suitable and distinctive terms. Here, therefore, the term “flower” and all that this word connotes is confined, as it has been by Arber and Parkin, to the Angiosperme, a phylum characterised by the evolution of that definite collection of organs with its normally cyclic structure, its varied and specialised envelopes, its stamens, closed carpels, styles and stigmas, and its peculiar type of fertilisation. The ambiguous position of the G'netales, and the eclectic treatment con- sequently accorded to them in this work, have already been explained. Leaving that class out of account, we recognise in the Gymnosperme a phylum within which the development of the reproductive system has not gone beyond the evolution of structures, termed here cones or strobiles, with usually scale-like leaves, abaxial (dorsal) pollen- sacs, adaxial (ventral or marginal) openly exposed ovules, and a corresponding mode of fertilisation. These features are more fully brought out in the definitions of the Gymnosperme, and of the classes which that division includes, prepared by Dr. Stapf for the conspectus of the orders contained in this section of the “Flora of Tropical Africa.” The definitions of these gymnospermous classes, of necessity somewhat fuller than the corresponding definitions of the classes of the Angiospermae, are as condensed as the special circumstances permit ; in each case, PREFACE. ix some of the more commonly employed alternative terms applied to particular organs’ have been added in brackets after those adopted in this work, so as to assist the student in correlating them correctly. The account of the Cycadacee calls for a word of explanation. It was hoped that this might have been prepared by Sir W. T. Thiselton-Dyer, F.R.S. Circumstances have unfortunately frus- trated the hope that his exact and extensive knowledge and his practised judgment in dealing with the members of this interesting and important order could be utilised in full. The work, none the less, is still largely his. The material in the herbarium at Kew, on which the account here printed is mainly based, has been arranged and annotated by him with such care and critical insight that the task of preparing that account has involved little more than the drafting of co-ordinated descriptions of the various genera and species. In addition to those whose assistance has been acknowledged in previous volumes of this work and whose collections have been made use of in the preparation of the present section, thanks are due to the following for material from the areas mentioned in each case :— I. Upper Gurnea.—M. Pobéguin, French Guinea; N. W. Thomas, Sierra Leone ; A. C. Miles, Gold Coast ; C. O. Farquharson, Southern Nigeria; E. W. Foster, P. H. Lamb, A. J. Richardson, and T. Thornton, Northern Nigeria. III. Nizz Lanp.—E. Chiovenda, Eritrea; R. E. Massey, Abyssinia and Soudan; R. E. Drake-Brockman, Somaliland ; R. A. Dimmer, T. D. Maitland, W. R. Rutter, and J. D. Snowden, Uganda; E. Brand, W. J. Dowson, and A. Linton, British East Africa. IV. Lower Gurynea.—H. H. W. Pearson, Angola, Damaraland and Great Namaqualand. VI. MozamsBiquE Distr.—A. F. Appleton, Mrs. QO. Craster, R. E. Fries, and 8. Schénland, Rhodesia. For the loan or gift of types and other material thanks are also due to Mr. A. Baudon, Professor A. Bertolini, Dr. A. Fischer von Waldheim, Professor H. Lecomte, Dr. A. B. Rendle, and Professor H. Schinz. The regions into which the area of this Flora is divided are defined in the preface to Volume VII. For the detailed topography the third edition the ‘‘ Spezial-Karte von Africa,” published at Gotha x PREFACE. by J. Perthes in 1893, has again been used. The geographical distribution of the species described in this section has been worked out by Mr. N. E. Brown, A.L.S. The manuscript has been prepared for the press and the proofs have been checked by Mr. C. H. Wright, A.L.S., Assistant in the Herbarium, Royal Botanic Gardens, Kew. jD a Kew, August 20, 1917. CONTENTS. a a Page CoNSPECTUS OF THE ORDERS : ; 2 A - ‘ * -) xiii Order CXXTIT, Ulmacee , ; : : < - ° 1 OXXIiIa. Barbeyacer . : 4 ‘ . Pe CXXIIIs. Cannabinacee . ; : : “ + 10 CXXIIIc. Moracee , : = : : . re aus CXXIiIp. Urticacexr - : . : . 240 CXXIV. Myricacee : : ; : = BUT CXR. Chetines © 0 a5 BS CXXVI. Salicines . ‘ s : : s 3 SEG CXXYVII. Ceratophyllee . . ° ° . . 326 CXXVIII. Gnetacez . ‘ : . : . - 328 CXXIX. Pinacee . . : s : : . 333 CXXIXa. Taxacee . : . ° . . - 3838 CXXX. Cycadacex: . . : ‘ ‘ . 344 Addenda . : : : : F » 355 xi CONSPECTUS OF THE ORDERS CONTAINED IN THE SIXTH VOLUME.—SECTION 2. CLASS I.—DICOTYLEDONES. SUBCLASS IIIl.—MONOCHLAMYDEZ. Series VII,—Unisexuales (continued).—Flowers unisexual. Ovary syncarpous or monocarpous ; styles as many as the carpels, often 2-partite ; ovules solitary or 2 collateral. Seed albuminous or exalbuminous. Herbs, shrubs or trees. Perianth calycine, small or none. CXXIII. Untmacem. Flowers polygamous. Perianth 4-5-merous. Stamens as many as the perianth-lobes and opposite to them, or more; filaments not inflexed. Ovule pendulous. Fruit a samara or drupe. ‘Trees or shrubs. Leaves alternate. Style 2-partite. CXXIIIA. Barszyacem. Flowers diccious. Perianth 3-4-merous, Stamens 6-9; filaments not inflexed. Style simple; stigma linear-oblong, flattened. Ovule pendulous. Fruit dry, indehiscent. Seed exalbuminous. Tree. Leaves opposite. CXXIIIz. CANNABINACER. Flowers diccious, rarely moneecious. Male perianth 5-partite; female small, cup-like. Stamens 4, opposite the perianth-segments. Stigmas 2, filiform. Ovule pendulous. Fruit an achene; albumen scanty. An aromatic herb. Leaves alternate or the lowest opposite. CXXIIIc. Moracem. Flowers monecious or diccious. Perianth usually 4-merous. Stamens usually as many as and opposite to the perianth-segments ; filaments inflexed or straight in bud. Carpels 2, the posterior rarely developed or represented by an empty chamber; styles 2 or 1. Ovule pendulous. Fruit an achene, nut or drupe. Albumen present or absent. Trees or shrubs, more rarely herbs. Leaves alternate, rarely opposite. CXXIIIp. Unticacem. Flowers moneecious or diccious (polygamous in Parietaria). Male perianth 4-5-partite, rarely monophyllous. Stamens generally as many as and opposite to the perianth-segments; filaments inflexed in bud. Female perianth 3-5-lobed or -partite. Style (when present) simple; stigma capitate or penicillate. Ovule erect or ascending, Fruit an achene. Albumen scanty or none. Herbs or shrubs, more rarely trees, sometimes with stinging hairs. Leaves alternate or opposite, generally penninerved, often with cystoliths. CXXIV. Myricace#. Flowers unisexual, in spikes, without perianth. Stamens 2-16; filaments short, free or more or less united; anthers erect. Ovary 1-celled ; xiii xiv CONSPECTUS OF THE ORDERS. style short. Ovule solitary, erect, orthotropous. Fruit a small drupe. Albumen 0. Shrubs or trees, usually aromatic. Leaves alternate, penniveined. CXXYV. Casvartnem. Flowers unisexual, in spikes or strobiles. Male flower: Perianth-segments 1 or 2, circumscissile. Stamen 1; filament inflexed in bud. Female flower: Perianth 0. Ovary 1-celled; style short. Ovules 2. Fruit a nut enclosed by the persistent woody bracts. Albumen 0. Leafless trees or shrubs, Branches articulated at the nodes where they bear 4 to many scales in a whorl. Series VIII.—Ordines anomali. Orders nearest allied to those of Series VII, Unisexuales, but not sufficiently closely so to be joined to any one of them. CXXVI. Sautcrnrs. Flowers dicecious, solitary under each bract of cylindrical catkins, more rarely in ebracteolate racemes. Perianth 0. Disc of 1-2 glands, or cup-shaped. Stamens 2 to many. Ovary 1-celled, with 2-4 parietal placentas, 2- to many-ovuled. Fruit a 2-4-valved capsule. Seeds small or minute, with silky hairs from the funicle, exalbuminous. Trees or shrubs. Leaves alternate; stipules small and deciduous, or leafy and persistent. CXXVII. Crraropnyitte®. Flowers unisexual, axillary. Perianth thinly herbaceous, multipartite. Stamens many. Ovary 1-celled, 1l-ovuled. Fruit a nutlet. Seed pendulous, exalbuminous; embryo straight. Submerged aquatic herbs. Stems elongated, leafy all over. Leaves verticillate, 2-fid or dichotomously divided into filiform or linear segments. Flowers solitary at each leaf-whorl. SUBDIVISION GYMNOSPERMA. Pollen-sacs and ovules borne on modified leaves (scales: sporophylls), which are grouped spirally or in whorls or pairs—or the ovules cauline, terminal and solitary— on separate axes of the same or of different individuals (monecious and dicecious species) ; the groups of sexual leaves forming, or the solitary terminal ovules entering into the formation of—(1) ‘‘cones’’ (strobiles: flowers) without any specialized envelope, although frequently subtended and in the earliest stages covered by squami- form cataphylls [Coniferales: Cycadales], or (2) ‘‘ florets’’ (flowers of most authors) with a perianth-like envelope in the g and an ovary-like envelope in the ? sex [G@netales]. Pollen-sacs (microsporangia: anther-cells) 2—co on the underside of the g scales (microsporophylls) or around the scale-stalk or grouped into ‘‘ anthers” borne on free or more or less fused filaments (stamens, (netales). Pollen-grains (microspores) on germination producing 1-3 vegetative cells (prothallium) and a very rudimentary antheridium which gives rise to 2 usually immotile, more rarely motile [Cycadales: Ginkgoales| male cells. Ovules (macro- or megasporangia) sessile or subsessile and borne directly or indirectly (by the intercalation of a scale-, ligule-, or aril-like appendage), on the upper side [most Conifere] of the expanded 9 scales (macro- or megasporophylls : carpels) or in direct continuation of the axis of the cone [some Conifere] or the floret [Gnetales], never enclosed in a mono- or poly- earpellary ovary with a style and stigma (hence termed ‘‘naked”’). Integument 1, sometimes with an accessory outer perfect or imperfect envelope (outer integument : aril: epimatium). Nucellus large, free from the integument only in the upper region, Endosperm, formed before fertilization (prothallium), filling the whole embryo-sac or only a portion of it [Gnetum sometimes] and apically bearing rudi- mentary archegonia whose egg-cells after fertilization undergo embryonic divisions CONSPECTUS OF THE ORDERS. XV which, however, lead normally to the development of only one perfect embryo. Fruiting ‘‘ cones’ usually much enlarged, either with the seeds more or less covered by the indurated scales (or their appendages) or enclosed among the fused and fleshy scales, the fruit finally resembling a berry or drupe; or reduced to 1 or 2 (rarely more) exposed seeds borne on the dry or fleshy axis of the cone; fruiting ‘‘ florets ”’ nut- or drupe-like, according to the coriaceous or half-fleshy, half-bony structure of the mature envelope. Seeds, if wholly covered or enclosed, with a crustaceous to woody testa—if partly or wholly exposed, with an externally fleshy and internally woody testa or enveloped in a more or less fleshy aril. Embryo axile in the copious fleshy or oily, rarely starchy endosperm, straight ; cotyledons 2-«. Woody plants, mostly trees, sometimes with short, tuberous stems. Leaves mostly coriaceous, usually simple (often linear or much reduced and squamiform), or pinnate to bipinnate ({Cycadales}. <‘ Cones” terminal or axillary, solitary or clustered (rarely spicate or paniculate), inconspicuous at the time of pollination or more or less vividly coloured ; ‘*florets”” in spikes, Pollination mostly by wind; pollen deposited directly on the micropyle or nucellus. CLASS I—GNETALES, Diclinous, dicecious or moneecious. Male floret (pseudo-bisexual in Welwitschia) : envelope formed of one pair or two decussating pairs of scales, free or united ; stamens 1-8 with the anthers sessile or subsessile on the summit of a stout central axis or [ Welwitschia] 6 with the filaments connate at the base around a central superior barren ovule. Female floret : envelope an ovary-like utricle with the ovule naked, erect, orthotropous ; ovule with a single integument produced into an elongated tubular micropyle which protrudes through the orifice of the utricle and with or without an aril. Seed albuminous, enclosed at maturity in the hardened utricle which is either differentiated into a fleshy outer and a hard inner layer or entirely coriaceous and sometimes [ Welwitschia] produced laterally into 2 wings. Embryo straight with 2, rarely 3, cotyledons. Erect or scandent, virgate or leafy shrubs, trees or woody climbers or [ Welwitschia] a woody plant of unique form consisting of a stout tuberous hypocotyl and an early arrested depressed stem-apex, with true vessels in the secondary wood, without resin canals. Foliage-leaves two or more, rarely in whorls of 3, opposite, simple. Florets (flowers) few or many, in axillary, rarely terminal, unisexual or bisexual spikes; spikes at the time of pollination greenish or yellowish, rarely vividly coloured [ Welwitschia]. Seeds with their envelopes more or less enclosed in the dry or fleshy spikes or exposed and drupe-like [Gnetum]. CXXVITI. Gnuracza. Only order. CLASS II.—CONIFERALES. Diclinous (normally), moncecious or dicecious. Male cones (male strobiles: male flowers) mostly catkin-shaped, made up of verticillate or spirally arranged scales (stamens: microsporophylls), bearing dorsally, or rarely around the scale-stalk, 2-15 pollen-sacs (microsporangia) ; pollen-saes dehiscing variously ; pollen-grains roundish, with or without vesicular appendages; generative cell producing 2 immotile male cells. Female cones (female strobiles: female flowers: female inflorescences) usually catkin-shaped, subsessile or pedancled, made up of verticillate or spirally XVI : CONSPECTUS OF THE ORDERS. arranged scales (carpels: macrosporophylls) bearing 1 to many ovules at or above the base either directly.or on an interposed usually scale-, ligule- or aril-like appendage (ovuliferous scale: placenta: aril: epimatium); or the cone reduced to a few subterminal or 1 truly terminal ovule, supported by a number of barren scales ; scales variously shaped, usually much larger than the ovules and covering them, or smaller and the ovules more or less exserted. Ovules (macrosporangia) erect or reversed, usually orthotropous, rarely anatropous ; integument 1, sometimes with an outer covering more or less surrounding the seed, or completely enclosing it except at the micropyle and even fused with it (outer integument: aril: epimatium) ; micropyle widely open or only a minute perforation ; nucellus free upwards from the integument, sometimes protruding through the micropyle ; pollen-chamber 0. Mature cones formed of the more or less enlarged and indurated scales (and their appendages) euclosing the seeds (typical cones), or of more or less fleshy scales fused into berry-like structures (galbules) ; or (if much reduced) with the seeds exserted from the unchanged or modified subtending scales. Seeds erect or recurved, ovoid or ellipsoid, often compressed and winged by the adhesion of a portion of the ovuliferous appendage or by outgrowths of the testa: testa membranous, coriaceous, crustaceous, wingless or winged or partly (inner layer) crustaceous or woody and partly (outer layer) fleshy or coriaceous and the seed then drupe-like ; nucellary membrane thin, free from the testa except at the base. Endosperm fleshy and oily, rarely starchy. Embryo 1 (rarely 2 or 3), axile, erect, cylindric-clavate, usually slender ; cotyledons 2 to many in a whorl; radicle superior or inferior. Trees or shrubs, mostly evergreen and coriaceous. Leaves opposite, verticillate or spirally arranged, scale-like or needle-shaped, or linear, rarely lanceolate to ovate, entire, rarely serrulate or lobed, very often heteromorphous. Male cones terminal on short leafy shoots, or axillary, sessile or peduncled, usually subtended by small imbricate cataphylls, solitary or gathered in heads, spikes or umbels. Female cones terminal or axillary, solitary, rarely in clusters or spikes. CXXIX. Pinacez. Female cones of very numerous or few simple or appendaged scales, all or a part of them bearing 1-8 ovules which are always shorter than their scales and covered by them, when mature representing typical “‘ cones” or berry-like galbules. Seeds enclosed, without an aril or arilloid covering. CXXIXa. Taxacem. Female cones usually much reduced, consisting of some barren and 1 to few (uppermost) fertile simple scales, the latter bearing 1 (very rarely 2) ovules, or the ovule solitary and terminal, supported by barren scales ; ovule usually more or less exceeding its scale; mature cone little altered or its axis and scales becoming more or less fleshy. Seeds usually exserted, with or without an arilloid covering. CLASS III.—CYCADALES. Dicecious. Male cones (male strobiles: male flowers) typical. cones; scales (stamens: microsporophylls) flat or more often thickened upwards and up to cuneate-peltate, coriaceous, close, bearing on their underside numerous (often over 100-700) pollen-saes (microsporangia), scattered or more often arranged in small substellate groups of 3 or 4, dehiscing longitudinally. Pollen-grains ellipsoid or subglobose ; generative cell producing 2 ciliate motile male cells (spermatozoids). Female cones (female strobiles: female flowers) similar to the male, but usually larger ; lower and uppermost scales (carpels: macrosporophylls) usually barren, the others fertile, either representing flat, toothed to pinnatifid stalked blades, bearing CONSPECTUS OF THE ORDERS. XVil the ovules (usually more than 2) in slightly hollowed-out notches on each side of the stalk up to and including the base of the blade [Cycas] or more often unguiculate- peltate with only 1 ovule from near each of the-inner angles of the base of the head. Ovules (macrosporangia) orthotropous, sessile or subsessile, obliquely erect [Cycas] or reversed and parallel to the claw, directed towards the axis of the cone; integument 1, stout, perforated at the apex by the micropyle ; nucellus free upwards from the integument, produced into a beak and passing into the micropyle, its apical tissues breaking down and forming a pollen-chamber in the mucilaginous liquid ‘of which fertilization takes place. Mature cones little changed except for the harder texture of the more or less spreading scales. Seeds large, globose to oblong, drupe- like ; testa of a fleshy, variously coloured outer coat and a hard inner layer, inside which there is a membrane containing the vascular system of the ovule and formed of the innermost strata of the integument together with portions of the nucellus. Endosperm fleshy. Embryo 1, axile, subcylindrical, borne on a filiform, spirally coiled up suspensor ; cotyledons 2, rarely 1 (by suppression) or 3-6 ; radicle superior, surrounded by a cap-like coleorrhiza, Woody plants with a stunted, tuber-shaped stem or a columnar trunk of varying height, usually simple, rarely forked and always densely covered upwards with the scars or persistent bases of the cataphylls. Leaves mostly of 2 kinds, namely scale-like cataphylls acting as bud-scales and exstipulate fronds, both produced in alternating series ; the cataphylls ovate to subulate, mostly woolly ; the fronds gathered in dense terminal crowns, pinnatisect or pinnate, rarely 2-pinnate, often very long, coriaceous, usually disarticulating at the base; rhachis generally straight, rarely involute in vernation; pinne straight or involute. Strobiles subsessile or borne on ebracteate or bracteate peduncles, solitary or in clusters below and often close to the apex of the stem or [Cycas] the female cones truly apical, their axis continuing growth after maturation and producing a series of cataphylls, followed by one of fronds, a process which may be repeated many times at varying intervals during the life of the plant, CXXX, CycapacEea. Only order. FLORA OF TROPICAL AFRICA. Orpver CXXIII. ULMACE2. (By A. B. Renp1z.) Flowers (in tropical African genera) polygamous, regular, apetalous. Leaves of the perianth free or more or less united, calycine, generally 4 or 5 in number, segments imbricate or valvate. Stamens as many as and opposite to the perianth-lobes rarely more ; filaments not inflexed; anthers erect in bud, 2-celled, dehiscence longitudinal. Ovary superior, 1-celled; ovule solitary, pendulous from or below the apex of the cell, anatropous, with two integuments ; style (in African genera) central, 2-partite, branches stigmatose on the upper face. Fruit a flattened samara or an ovoid to globose drupe. Endosperm wanting or scanty. Embryo straight or curved.—Trees or shrubs. Leaves alternate, in two rows, simple, often unequal at the base. Stipules lateral, free, generally small, caducous. Flowers small, in axillary branched cymes, the fertile sometimes solitary. Species about 130; widely distributed in both temperate and tropical regions. Sus-Orpur I. Ulmoidew.—Flowers in the axils of scars of previous year’s shoots ; ‘as a samara ; wate straight aes iy . 1. HoLoprenea. Sus-Ornper II. Celtidiotdes. pias tn the axils of foliage-leaves of the same season ; fruit a drupe ; embryo curved. Branchlets unarmed (in African species) : Leaf-margin entire or weakly or irregularly toothed (serrate in C. kraussiana) ; fruit 24-9 lin. long ; cotyledons broad . 2. CELTIS. Leaf-margin minutely toothed or seeuiines fruit about 1 lin. long ; cotyledons narrow 3. TREMA. Branchlets armed with axillary spines ... 4. CHRTACME. 1. HOLOPTELEA, Planch.; Benth. et Hook. f. Gen. Pl. iii. 352. Flowers hermaphrodite with a few male. Perianth simple, calycine, divided. nearly to the base into 3-8, often unequal lobes. Stamens as many as the lobes but generally more, up to12. Filaments erect, long, free, ultimately exserted ; anthers roundish, 2-celled. Ovary rudimentary or absent in male flowers, in the fertile stalked, compressed, passing above into a pair of short spreading style-arms, FL. TROP. AFR. VI. SECT. I. PART I. B 2 CXXIII. ULMACEZ (Rendle). [Holoptelea. which are densely stigmatose on the upper face; ovule solitary, pendulous from the apex of the cell. Fruit dry, indehiscent, form- ing a roundish flat samara with a broad membranous veined wing.— A tree. Leaves alternate, distichous, short-stalked, penniveined, 1-nerved, punctulate on the lower face with small round cystoliths. Stipules lateral, small, scarious, caducous. Species : one, in India, Ceylon, and Cochin-China, hitherto not recorded from Tropical Africa. 1. H. integrifolia, Planch. in Ann. Sci. Nat. 3me. sér. x. 266. A large spreading deciduous tree 50-60 ft. high, with ash-grey (brown- violet, according to von Doering) pustular bark; shoots glabrous. Leaves elliptic, acuminate, base rounded or cordate, margin entire _ (often serrate on seedlings and shoots), 3-5 in. long, 14-23 in. wide, glabrous, upper face shining, midrib prominent beneath, veins 5-7 on each side, ascending and uniting below the margin. Flowers in short branched inflorescences at the scars of fallen leaves, usually male and hermaphrodite, the latter above. Sepals concave, obovate, about ? lin. long, puberulous on back. Filaments glabrous; anthers puberulous. Ovary compressed, pubescent, the stalk lengthening as the fruit ripens. Samara 3-1 in. broad, notched at the top, the broad wings beautifully veined; stalk }-} in. long, jointed in the middle.—Planch. in DC. Prodr. xvii. 164; Hook. f. Fl. Brit. Ind. v. 481. Ulmus integrifolia, Roxb. Corom. Pl. t. 78. 3 od Guinea. Togo; Atakgama, in a mountain wood at 1200 ft., Doering, ! An interesting addition to the flora of tropical Africa. Based on a single specimen with flower kindly lent for determination by the Director of the Royal Botanic Gardens, Berlin. 2. CELTIS, Linn.; Benth. et Hook. f. Gen. Pl. iii. 354. Flowers polygamous, the fertile generally hermaphrodite. Male: Calyx membranous, deeply divided, sometimes to the base, into 5, rarely 4, concave imbricate segments. Stamens as many as the sepals; filaments free, not incurved; anthers bluntly and shortly ovate. Torus densely hairy. Fertile: Calyx and torus as in male; stamens smaller than in male. Ovary sessile; ovule pendulous from the apex, anatropous; style central, 2-partite, branches undivided or 2-fid, densely papillose-stigmatose on the upper face and margins. Drupe fleshy, ovoid to globose, sometimes slightly keeled ; endocarp bony, rough, sometimes keeled. Embryo curved ; cotyledons very broad, applied face to face.—Trees or shrubs, deciduous or evergreen, unarmed in tropical African species. Leaves alternate, short-stalked, generally more or less coriaceous, often with acuminate apex, and unequal at the base, margin entire or serrate, penniveined, 1- or 3-nerved. Stipules lateral, free, caducous. Inflorescences male or androgynous, in laxly panicled cymes or fascicled, axillary or at the lower leafless nodes of young shoots; Celtis. } COXXIII. ULMACEZ (Rendle). 3 fertile flowers with longer stalks, solitary or few in the higher axils.— Mertensia, H. B. et K., Nov. Gen. et Sp. ii. 30. Momisia, F. G. Dietr. Vollst. Lex. Gaertn. Nachtr. v. 122. Solenostigma, Endl. Prodr. Fl. Norfolk. 41. ; Species about 60, widely dispersed in temperate and tropical regions, especially in the northern hemisphere. Stigmas undivided : ; aves with serrate margin... ike a3 w. Ll. C. kraussiana, Leaves with entire margin or with a few weak teeth fe above ae ies ae ee