Notes on the genus Ficinia ( Cyperaceae ) : morphological variation within the section

Generic limits of Ficinia are reviewed. Morphological variation within the section Bracteosae Pfeiff. of Ficinia Schrad. (Cyperaceae) is examined and discussed. Diagnostic characters previously in use in classification are reviewed together with a number of new characters. These have enabled the limits of many existing taxa to be refined, and have provided the basis for the establishment of a number of new taxa recently described. Characters examined include: underground organs, leaf sheaths and ligules, leaf blades, inflorescence structure, prophylls, styles, anthers, achenes, gynophores and inflorescence axes.


NOTES ON THE GENUS FICINIA (CYPERACEAE): MORPHOLOGICAL VARIATION WITHIN THE SECTION BRACTEOSAE
of a single species (Schonland, 1922), whereas at the other extreme five distinct species are recognized (Levyns, 1950).None of the work carried out so far has been entirely satisfactory and, if anything, has tended to compound rather than to clarify the confusion.
In the present investigation, morphological variation is examined in the hope that this will clarify some of the taxonomic, distributional and nomenclatural problems existing in the section Bracteosae with special attention being focused upon the F. indica/F.pallens complex.

GENERIC LIMITS OF FICINIA
Within the tribe Scirpeae, Ficinia is most closely allied to the genus Scirpus and more especially to the section Isolepis of this genus.Before Ficinia was established, many of the species now assigned to it were included in this section.Species of Ficinia may be distinguished frofn it, however, by their possession of a small, obconical, pedicellike struc ture or gynophore (also referred to as a discus) responsible for elevating the achene above the receptacle.
According to C.B. Clarke (1898) it is difficult to attribute very much importance to this structure, because it is difficult to draw the line between it and the expanded foot associated with the achene of many species of Scirpus, therefore Ficinia pulchella of Kunth is Scirpus pulchella Boeck.Furthermore, a small gynophore is also present in some species of Fimbristylis, and varies greatly in development in one of the species in this genus, namely, F. ferruginea (L.) Vahl.Schonland (1922) draws attention to the fact that an obpyramidal small stalk which becomes detached with the nut, besides being present in species of Scirpus and Fimbristylis is also found in Scleria and Diplacrum of the tribe Rhynchosporae.The main difference, however, between these genera and Ficinia lies in there being a small lobed somewhat fleshy outgrowth surrounding the base of the ovary in Ficinia and that this structure is normally three-lobed or has three main lobes.
The present investigation has shown these observations to be accurate, for all the taxa studied possessed an out-growth or outgrowths from the apex of the gynophore which enveloped all or part of the lower portion of the nut.The form of the outgrowth is highly variable within the genus, ranging from a cup-like structure to 2 -3 lobes generally joined near their bases.
The use of the gynophore as the solitary attribute characterizing the genus Ficinia has, for the greater part, been successful in uniting a large, relatively heterogeneous group of plants, which in the past were considered by some workers to comprise a number of distinct genera.
Ficinia includes a number of taxa with nuts that lack a gynophore.Among these are F. tenuifolia Kunth and F. rigida Levyns.Because of the 'ficinioid fades' of these plants, it has been generally accepted that they be included in Ficinia.Scirpus falsus C.B.
Cl. and Scirpus ficiniodes Kunth, on the other hand (to be dealt with in some detail in a separate paper), despite their 'ficinioid' appearance, have been placed in Scirpus, but not without recognition of the uncertainty of this relationship, reflected as it is, in their specific epithets.The two major attributes which dictated their inclusion in Scirpus were, firstly, the lack of a gynophore, and, secondly, the presence of six bristles surrounding each flower, a feature unrepresented in Ficinia, but characteristic of certain members of Scirpus.Therefore it is evident that the boundary between Scirpus and Ficinia is by no means clearcut nor well defined.
The 'ficinioid' appearance already mentioned is the cumulative effect of a number of minor features exhibited by a large number of species, particularly those of the section Bracteosae.It is a means by which plants of the genus can usually be disting uished quite readily from species of allied genera, often on superficial inspection alone.
The characters which collectively give this 'ficinioid fades' are the erect, and sometimes rigid, leaf blades that aie often almost cylindric and stem-like rather than flattened and leaf-like; the membranous or scarious leaf sheaths that in many instances are prolonged upwards into definite ligules, the often (but not always) condensed and highly compacted obovate to elliptical inflorescences enveloped almost completely, or only partially, in an involucre formed by the dilated bases of the outer bracts, the longitudinally striated appearance and, often, the hard, brittle texture of these bracts and sometimes of the smaller glumes within them.This 'ficinioid fades' is not commonly exhibited by members of the section Isolepiformes, which closely resemble Scirpus, and it is a typical of the subgenus Hemichlaena in which the glumes are distichously ranked and the spikelets compressed.This strongly suggests that present generic limits, defined almost exclusively as they are by the presence of a gynophore beneath the achene, are artificial, unnatural and probably too wide.There is, therefore, an important need to re-examine all species with a view to determining critical, and especially new, criteria that will effectively delimit natural groups within the genus.

DIAGNOSTIC CHARACTERS EXHIBITED BY THE SEC TION BRACTEOSAE
The characters selected by past workers to delineate the different taxa within Ficinia have not always been entirely successful in fulfilling this function.Many, with use, have proved to exhibit a considerable range in variation, extending beyond the limits originally intended for them, and are therefore considered to be unreliable.
Many taxa at their establishment were inade quately defined by too few characters.With time, and under the interpretation of numerous workers, these taxa have become heterogeneous, so that on more careful scrutiny and with the application of more critical diagnostic criteria they fell, during the present work, into relatively homogeneous sub-units worthy of independent recognition sometimes at specific, sometimes at infraspecific levels.
New characters and others previously in use that require special consideration are dealt with in sequence beginning with vegetative organs and proceeding to reproductive structures.

Underground organs
Growth is sympodial in all the taxa studied.In this sense all plants possess a rhizome, for even in the compactly tufted plants, a very abbreviated rhizome links successive shoots together.
Previous workers have made reference to the presence of stolons in some taxa of the section Bracteosae and their absence in others.These 'stolons' are essentially rhizomes, for they are underground stems clothed in leaf scales, but they are markedly elongated in comparison with the rhizomes developed in the more or less compactly tufted plants.Each gives rise to a new plant at its tip and thus the term 'stolon' is correctly applied to them (Lawrence, 1951, Willis, 1973).Two main rhizome forms, which differ in degree of develop ment rather than basic structure, are represented in the section Bracteosae (Fig. 1).

(i) Abbreviated rhizomes
Many of the taxa studied possessed abbreviated rhizomes but these, despite their short length, were not uniform.They fell into two subgroups.Little variation occurred within the first.Here the rhizome did not exceed 2 -3 mm in length.It is inconspicuous except when innovations are developed, being concealed by persistent leaf basis.
The rhizomes of successive shoots are very closely packed resulting in the development of plants with compactly tufted growth forms (Fig. 1A) for example F. capitella.
The second subgroup comprises plants in which the rhizome structure is more variable in length.
Here it ranges from 2 -25 mm and is either uniform in width or swollen apically, as in F. pallens var.pallens, to form a tuberous base to the aerial shoot.(Fig. IB).
The rhizomes are clothed in short, obtuse, striate scale bracts that are usually persistent, for example F. levynsiae.Plants falling within this second subgroup, because of their less abbreviated rhi zomes, develop a loosely tufted growth form, that, in its most extreme expression, is distinct from the compactly tufted growth form illustrated in Fig. 1A.Gradation occurred between the two forms, however, depending upon the length of the rhizome and the number of shoots developed in proximity to one another.

(ii) Elongated rhizomes
Here rhizome structure is relatively uniform.Some variation in length occurs, but all rhizomes are clearly well developed and elongated up to 150 mm in length.They are more or less uniform in width and clothed in linear-acuminate, striate, pale brown to yellow scale bracts that are persistent for the most part, but which in some taxa are lost as the rhizome became shrunken and wiry with age (Fig. 1C).Plants with elongated rhizomes have been described in the present study as stoloniferous, for example, F. dunensis.
Rhizome structure and growth form are useful in distinguishing some taxa (especially in the field).In the herbarium compactly tufted plants can only be differentiated from stoloniferous plants if stolons are included in the specimen.
Two species, F. dunensis and F. pallens, show variation in the development, or not, of stolons.F. dunensis, plants growing on sand dunes near the coast, regularly produce stolons that are usually represented on herbarium sheets.Plants from inland are compactly tufted wi{h no evidence of elongated underground stolons.This difference may be attributed to the different habitat conditions under which the plants grow, as F. dunensis is a widely distributed species.In F. pallens, stolons are present only in var.pallens, distributed along the Cape Peninsula.In var.lithosperma from the Caledon area, no stolons have been recorded.This lack of stolons may be linked to the dry climatic conditions in this area as opposed to the wetter conditions that prevail along the Peninsula.Plants of var.lithosper ma are compactly tufted, more rigid and have distinctly harder inflorescence bracts than plants of var.pallens.
This is suggestive of adaptation to drier condi tions.It is also possible that the presence or absence of stolons is to some extent genetically controlled, as the two varieties are, for the most part, allopatric in distribution, and could therefore represent indepen dent gene pools.

Leaf sheaths and ligules
Two distinct types of leaf sheath are represented among the taxa studied.The first of these is membranous, usually transparent or pale brown to red, glabrous and splitting longitudinally with age.Such leaf sheaths are eligulate, or carry very short inconspicuous ligules that seldom exceeded 3 mm in length.In the eligulate forms the membranous sheath-tissue is joined to the two margins of the thicker sheath tissue that, when extended, formed the leaf blade.Therefore the membranous tissue does not extend across the inner surface of the leaf blade, for example, F. radiata and F. latifolia (Fig. 2e & f).In the ligulate forms this membranous portion of the sheath extends across the inner surface of the blade to form a complete cylinder, for example F. arenicola and F. nigrescens (Fig. 2g & h).The sheath apices of the eligulate forms are oblique, whereas those of the ligulate forms are more often truncate.
The second type of leaf sheath is markedly scarious or papery, wrinkled, fragile and white or greyish, and generally flecked with brown markings, for example, F. deusta (Fig. 2i).In this sheath type the ligule is always well developed and up to 3 cm in length in some taxa.The ligule apices, although commonly oblique, are sometimes suboblique two-lobed, or occasionally truncate.This variation occurs not only among taxa, but within them, so that as a diagnostic character the shape of the sheath apex of this second type is unreliable.It is often difficult to distinguish the original form of the stem apex because the scarious tissue, being very delicate, tears and splits easily, giving rise to a ragged, conspicuous, but uninterpretable mass of tissue near the apex of the shoot.Despite this, the two clearly different types of sheath are useful in the separation of some taxa.In many species the leaf blades exhibit a considerable range of form.Nevertheless, if used judiciously, this character is useful as an easily observable means of distinguishing those taxa in which blade form is reasonably constant and distinct.This applies especially to taxa with blades that represent extremes in the range of form.

Most of the taxa in
Anatomical studies of leaf blades by Arnold (unpublished) suggest that internal structure will provide more valuable diagnostic features on which to distinguish taxa and on which to base relation ships, than the gross morphology of the leaves.

Inflorescences
Inflorescences in the majority of species within the section Bracteosae are characteristically compact and are surrounded by an involucre of bracts.The bases of these bracts are conspicuously dilated and envelope the lower third or more of an inflores cence.The apices of the lowest 2 -7 of these bracts are conspicuously attenuate into linear, photosynth etic blades that closely resemble leaf laminae.Despite this superficial uniformity in external appearance, considerable variation is found on dissection of the inflorescence.The structural differences found, appear to be the result of different degrees of reduction within the inflores cence.
An explanation of inflorescence structure within the section Bracteosae is given as an aid to understanding the process of reduction that is evident.
In its most complex form the inflorescence comprises a central axis (axis 1, Fig. 3A* In most taxa, reduction of the partial inflores cence is not uniform, being more pronounced at the apex of the different taxa than at their bases.This results in the condition (Fig. 3C) in which the upper bracts of axis 2 each subtend a single bisexual flower, while the lower bracts subtend a complete axis, axis 3.This state is found in F. elatior and F. ixioides.
In some species the partial inflorescence is reduced to a single axis, axis 2 (Fig. 3D).The bracts on this axis resemble glumes and each subtends a single bisexual flower.In such taxa it is common for bracts on the upper portion of axis 1 to subtend single bisexual florets, while bracts on the lower portion of the same axis subtend a complete axis, axis 2 (see Fig. 3E).Characteristic of this inflorescence form are the empty 2-8 lowest bracts on axis 1, which have lost their axillary partial inflorescences.
The ultimate stage in reduction is an inflorescence comprising a single axis, axis 1 (Fig. 3F).The lowest 2 -10 bracts on this axis are always empty, the remaining bracts each subtend a single bisexual flower.In this unispicate inflorescence form and the type shown in Fig. 3D & E, the bracts that subtend single bisexual flowers have maintained their bracteaceous nature, this is, however, not the case with the bracts of axis 2 and 3 of Fig. 3 A -C , which also subtend single flowers but which are glumaceous, being thinner, softer, smaller, more membranous and pallid and less obviously longitu dinally striate.
Accompanying the reduction of the branching axes is a reduction in the number of prophylls (p) at the base of each axis of the partial inflorescence.The inflorescence forms exhibited in Fig. 3 A -C  Although the single prophyll conditions suggests a more reduced condition from paired prophylls, this may not be the case in Ficinia.In this genus the progression in prophyll number is considered to have been from l -»2->0.This sequence is suggested firstly by the fact that taxa with a single prophyll include those with branching inflorescences, gener ally regarded as representing a condition of limited reduction in the Cyperaceae.Secondly those taxa exhibiting the most inflorescence reduction are all wanting in prophylls.In all these taxa, except one, F. nigrescens (Schrad.)J. Rayn., reduction of the branching system is complete, producing a unispicate inflorescence.In F. nigrescens reduction is not always maximal so that, together with plant bearing unispicate inflorescences, there are also plants in which all or few of the lateral branching systems (partial inflorescences) constitute single spikelets.In both these conditions prophylls were found to be wanting.
The three prophyll states in Ficinia are considered to be related to the degree of reduction in number and size of branching axes comprising each partial inflorescence.With compaction there is a compara ble increase in pressure on the internal structure of the inflorescence, this being greatest in the axil of each branch in the region of the lowest adaxially situated bract (Fig. 4A).The result being to alter its size and structure transforming it into what is commonly referred to as a prophyll.This situation represented by the single prophyll condition is found in the majority of taxa in Ficinia.
In the section Bracteosae, compaction of the inflorescence has been accompanied by a change in the nature of the inflorescence bracts.The base of each bract has to a lesser or greater degree become dilated, expanding around the partial inflorescences in the form of an involucre.The dilated portions have also become thickened, hard and rigid.The result has been to further increase the internal pressure within the inflorescence, this time concen trated in the region of the second lowermost bract of each axis (Fig. 4B), reducinglt to a further prophyll.

The degree of dilation, thickening and hardening of the inflorescence bracts decreases progressively towards the centre (apex) of the inflorescence.
Because of this the extent to which the two lowermost bracts/glumes on axes 3 and 4 (Fig. 3  A -D ) have been modified is generally less than in axes 1 and 2, so that they are not always distinguishable as prophylls.At the other extreme, the 1 -2 lowermost bracts of axis 1 exhibiting the greatest degree of dilation and hardening, and therefore the greatest pressure, are in many taxa sterile.
In those taxa, in which prophylls were wanting, the number of sterile bracts at the base of axis 1 ranged between 2 and 6.This increase over taxa with 1-2 sterile bracts (plants with double prophylls) is probably due to the overall increase in dilation and hardening of bracts along axis 1.In these taxa the increased pressure along axis 1 has not only caused the loss of the double prophyll condition but has also reduced each partial inflorescence to a single floret.Therefore, whereas in most Cyperaceae the absence of prophylls represents a less specialized condition of limited reduction, in the section Bracteosae of Ficinia this condition is synonymous with maximum reduction and specialization.
In F. pinguior C.B.C1., placed by Pfeiffer in the section Bracteosae although the inflorescence bracts have undergone dilation they have not become thickened or hardened.As this taxon exhibits a single prophyll condition, hardening of the bracts is seen as an essential factor responsible for the production of the second prophyll.Another taxon in which this may be observed is F. paradoxa (Schrad.)Nees, placed by Pfeiffer in the subsection Capitulae of the section Isolepiformes.F. pinguior is consi dered here to be more closely related to this taxon than any of the taxa included in the section Bracteosae.It is not known at this stage whether F.
pinguior with F. paradoxa should be included within the section Bracteosae or not.The evidence so far gathered suggests not, and F. pinguior is therefore tentatively excluded from the section pending revision of the genus.Another species warranting special consideration is F. ancepts Nees, also considered by Pfeiffer to belong to the section Bracteosae.This taxon has only a single prophyll per inflorescence axis.Although its bracts have undergone dilation and hardening this is not so marked as in other taxa in the section.The lowest bract and partial inflorescence in many plants is up to 1,5 cm distant below the partly compacted upper section of the inflorescence.It therefore probably represents an intermediate stage towards the development of section Bracteosae.F. anceps is most closely allied to F. fastigiata Nees.This latter species also exhibits a single prophyll condition and is included by Pfeiffer in subsection Capitulae of the section Isolepiformes.F. anceps has therefore also been omitted from this treatment of the section Bracteosae until its relationship with the rest of the genus is better understood.It should perhaps be noted at this point that prophylls are associated only with the axes that comprise partial inflorescences.It is probable that these appendages, despite their different appear ance, are morphologically equivalent to the bracts of ancestral less reduced inflorescences.This has been accepted by Blaser (1944) and Koyama (1961) but, perhaps, needs further confirmation from anatomic al and vascular investigation.The recording of the presence of prophylls in inflorescences within the section Bracteosae is an important step forward in appreciating the heterogeneity of Ficinia as it is at present delimited.Prophylls have not been recorded previously for the genus and they are certainly absent from those of its species with 'simpler' more reduced spikelets that are of the Scirpus type.

Also included by Pfeiffer in the subsection
The degree of reduction of the inflorescence is not completely constant within each species.Many taxa exhibit two or occasionally three (for example F. nigrescens) inflorescence forms.Despite this the degree of reduction of the inflorescence, together with the number of prophylls present, are extremely important characters by which to distinguish different taxa within this section.

Style and style branches
Three main types of style are represented in the taxa studied.The differences between these types depend mainly upon robustness, slenderness and on relative lengths of style and style branches.In the commonest type, the style and its branches are slender and relatively elongate, pale brown and flecked with darker markings, with the style branches obviously papillate.The lengths of the fused lower portion of the styles relative to those of their branches varied considerably within the section and proved to be a useful means by which to distinguish many species.In some taxa the fused section is short (1-2) mm in length with very long branches 6 -1 0 times the length of the fused section, for example F. capitella and F. arenicola var.arenicola (Fig. 2a).In other taxa the fused section and branches are approximately equal in length, for example F. ixioides (Fig. 2b).
The second type comprises styles that are notably broad and stout, red or pale brown with brown flecks and with the style branches granular rather than obviously papillate.The difference in appear ance of the style branches depends upon the reduced length of the papillae.Examples of this type are F. cedarbergensis and F. levynsiae (Fig. 2c).
The styles of F. radiata, which also belongs to the second type (Fig. 2d), are characteristic of this species alone.They are broad, stout and granular and either lacked branches entirely or possessed three minutes notches at the style apex that represented branches.On these features, together with a few other individual characters, this taxon was at one time placed in Sickmannia, a monotype genus distinct from Ficinia.(Such a treatment is not upheld by Arnold, 1979.)

Anthers
Anthers in Ficinia are crested.In the section Bracteosae the shape and degree of development of the crests provided a useful additional character by which to delineate many of the taxa.
Five basic forms are recognized (Fig. 2j-o).These fall into two broad groups.The first group includes those with crests as long as, or shorter than, their breadth at the base of the crests, group two includes taxa with crests twice, or more than twice, as long as their breadth at the base of the crest.
In the first group the crests are either flattened or slightly rounded, globular or triangular (Fig. 2j-1).In the second group the crests are either linearoblong with rounded apices (Fig. 2m) or linearaccuminate with pointed apices (Fig. 2n).These five forms are not altogether clearly defined and some intermediates, did exist within species, nevertheless the shape of anther crests does provide an ancillary character useful in taxonomic diagnosis.The presence or absence of spines on the crests is also useful taxonomically.In F. ixioides subsp.ixioides (Fig. 2o) the crests bear long conspicuous spines (stiff and possibly silicacontaining) while in subsp.glabra (Fig. 2n) these are minute (virtually absent).This is, therefore, a useful character in distinguishing between these infraspe cies.

Achenes
Achenes of the different taxa differed in size, shape, colour and surface texture.
Achene size ranged from 1,5-2 mm in length by 1 mm wide (F.pusilla Fig. 5a) to ± 6 mm in length by ± 4 mm in width (F.deusta, Fig. 5r).On individual plants and within taxa there is generally little size variation despite the fact that, especially in herbarium material, it is sometimes difficult to establish whether the achenes are fully mature and fertile or not.Achenes are usually three-angled or trigonous but often the ridges are not very sharp; mostly the faces equal each other in width (F.ixioides Fig. 5g), but in F. cedarbergensis (Fig. 5n) and F. levynsiae (Fig. 5p) they are unequal resulting in asymmetrical achenes.In F. deusta (Fig. 5r) and F. grandiflora (Fig. 5q) achenes are planoconvex, whereas in F. latifolia they are spherical (Fig. 5h).Achene apices also differ in some taxa.Most apices are obtuse or rounded; a few are acute often with a distinct but small beak for example F. capitella (Fig. 5c) and F. dunensis (Fig. 5j), whereas others are broad and flattened or retuse as in F. cedarbergensis and F. levynsiae (Fig. 5n & o).
The surface texture of the achenes is smooth and often polished, for example F. ixioides subsp.glabra (Fig. 5g), finely dotted (muricate), for example F. indica taxon indica (Fig. 5b), or transversely ridged as in F. capitella (Fig. 5c) and F. dunensis (Fig. 5j).Surface patterning of the achenes differs only negligibly within taxa therefore achene size, shape and surface markings are useful criteria in identify ing taxa.Achene colour is not so satisfactory.All immature nuts are more or less uniformly pale yellow to brown in colour.This colour changes with maturity to black, grey, brown or red.Although there is a difference in achene colour between species at maturity, the possibility always exists that there is a colour range as the achenes develop to full size on the plant after fertilization so that no one colour is specific for a taxon.There is also no guarantee that unfertilized achenes follow the same colour sequence during development nor that achene colour does not alter in the herbarium.Consequently not much reliance should be placed on this character in taxonomic diagnosis.

Gynophore
The presence of a gynophore below the achene was (and still is) generally accepted as the main character that distinguishes Ficinia from its allied genera.Despite this, there have been included within Ficinia several taxa that lack a gynophore.
As mentioned earlier Clarke (1898) felt that it was difficult to attach much importance to the gynophore in Ficinia.Although this may certainly be true in the diagnosis of generic limits it does not apply infragenerically.The present investigation has shown there to be little variation in gynophore form within species and subspecies but from one taxon to another gynophore shape and degree of specializa tion of the margins differed markedly thus being a satisfactory character on which to recognize many species and sometimes infraspecific sub-units.
In the section Bracteosae the gynophore enve lopes the lower portion of the achene.It shows a considerable range in variation from two or three lobed as in F. nigrescens (Fig. 5f), to cup-shaped with the upper margin crenate as in F. levynsiae (Fig. 5o) or developed into prominent finger-like projec tions as in F. latifolia (Fig. 5h).
The length of the gynophore relative to the total length of the achene is also a useful character.It ranges from 1/6 to 2/3 of the total achene length.

Inflorescence axis
After the bracts glumes and achenes have fallen, the inflorescence axis persists at the apex of the culm, remaining intact until the culm breaks up.Since the old culms often persist until after the following flowering season the form of the inflorescence axis is a readily available character with which to distinguish many taxa in the section Bracteosae.
The least reduced inflorescence axis shows a structure in which the branches associated with bract scars occur at intervals through the axis (Fig. 6a).More reduced types have branches restricted to the upper third of the axis with bract scars below (Fig. 6h).Inflorescences where the greatest reduction had taken place resulted in unbranched axes that resemble elongate empty strobili (Fig. 6n & p).Some breakage of these axes especially falling of the small branchlets probably takes place as they age and become weathered but, apart from these changes (which may be slightly misleading), persis tent axes provide a convenient and reliable character for identifying taxa.
The sequence of reduction exhibited in the inflorescence is reflected in the structure of the inflorescence axis.A number of possible lines of reduction are apparent.In the first (Fig. 6 a -e ) reduction is concentrated within the main axis which becomes progressively shortened until in F. capitella it is almost receptacle-like.Associated with this is an obvious loss in the number of branches originating from the main axis.A second possible line of reduction is indicated by Fig. 6f-j.Here the main axis has become swollen with a progressive loss in the number of primary branches, from the base of the axis upwards until in F. petraphylla all the branches are lost.Finally in Fig. 6k-p, the tendency has been for the main axis to become elongate with a total loss of branches in these taxa, except in F. nigrescens in which some members still retain their primary branches (Fig. 6m).In F. ixioides although the branches have been lost some individuals possess a single branch originating from the base of the main axis (Fig. 6k).This branch is often as well developed as the main axis, equalling it in length and giving the inflorescence axis a forked appearance (Fig. 61).This branch if present is however generally smaller than the main axis.
Ficinia exhibit well developed leaf blades.Only in a few taxa are elongate leaf blades absent, in the section Bracteosae this condition is exhibited only by F. gydomontana.Leaf blade length, although not a completely reliable character in most species, is useful in distinguishing F. petraphylla and F. arenicola from other species of the section.In these two taxa the blades are characteristically longer than the culms.Considerable variation exists in the presence and degree of development of the leaf margins.In some taxa the blades are flat, with well defined margins: in others the margins are scarcely or not represented since the blades are cylindrical.Between these extremes a complete range occurs.Metcalfe (1970) correlated the degree of infolding of the leaf margins with environmental conditions, regarding infolding as a response to dry conditions as it effectively reduces the leaf area over which transpiration can take place.This is exhibited by plants of F. ixioides and F. arenicola var.erecta.In F. ixioides subsp.whereas distributed in the wetter south-western Cape the leaf blades are lax and broad and flattened, whereas in subsp.glabra distributed in the drier north-western Cape, the blades are rigid, subcylindrical and stem-like.
the inflorescences of F. compasbergensis Drfcge, F. fascicularis Nees, F. monticola Kunth, F. pinguior C.B.C1.and F. anceps Nees, also included in the section Bracteosae by Pfeiffer (1920), only a single prophyll is situated adaxially at the base of each of these axes.Apart from exhibiting a single prophyll the first three taxa above also have branching inflorescences, considered to be a typical of the section Bracteosae in which the majority of taxa have, compact capitate inflorescences enve loped to a lesser or greater degree by an involucre of dilated inflorescence bracts.These three taxa are th erefo re, on the grounds o f in flo r e s cence form, tentatively excluded from the section Bracteosae pending revision of all the sections in the genus.

Fig. 4 .
Fig. 4.-Prophyll states: A, single prophyll state characteristic of all sections other than section Bracteosae; B, double prophyll state represented by the majority of taxa in the section Bracteosae.
Capitulae of the section Isolepiformes are F. argyropa Nees (=F.paradoxa (Schrad.)Nees var.argyropa (Nees) C.B.Cl.and F. laevis Nees.Not only do these two taxa have hardened, dilated bases to the inflorescence bracts, but they also exhibit the double prophyll condition.Their closest ally within the genus is considered to be F. indica (Lam.)Pfeiff.On the basis of this they are included here in the section Bracteosae.