Revision of the genus Faucaria (Ruschioideae: Aizoaceae) in South Africa

Faucaria Schwantes is a small genus o f succulents, concentrated in south-central Eastern Cape. Distinguishing features of Faucaria are the characteristically dotted and toothed leaves and the structure of the seed capsule. A taxonomic review of the genus is presented, including relationship, diagnostic features, variation, distribution and ecology. The revision reduces the number of species from 33 to six: F. bosschearui. F.felirui, F. gratiae, F. nemorosa, F. subintegra. and F. tigrina. Two new subspecitic combinations, F. felirtu subsp b ritten iae and F. felina subsp tubercu losa have been made. A cladogram is included. A key to the taxa is provided and each species is described and illustrated. Bothalia 29.1: 35-58 (1999)


INTRODUCTION
The earliest record of a Faucaria dates from before 1707, and consists of a watercolour of F. felina subsp.felina in the Moninckx Atlas (Wijnands 1983).The Moninckx Atlas is a collection of nine volumes of water colours, painted from plants in the Hortus Medicus Amstelodamensis.Material of this specimen was proba bly collected at a location near Algoa Bay, [the presentday Port Elizabeth region], during a search for fresh water by a Dutch vessel.Herbarium material from this period could not be traced.The earliest known herbarium sheet of F. felina s.I. was made by Thunberg.who visit ed the Eastern Cape on his second journey, from September 1773 to January 1774 (Juel 1918;Dyer 1937).After Thunberg.only a few herbarium specimens of this succulent genus were deposited in herbaria until ± 1925.Even Ecklon and Zeyher.who collected extensively in the Eastern Cape between 1831 and 1832.left only a few herbarium specimens.More herbarium collections were made between 1930 and 1938 by Mmes W. van Ryneveld, G.V. Britten.L.L. Britten, I.E.King.Messrs H.W. James.G.G. Smith, and other (often amateur) col lectors.Plants from these collectors were often intro duced into living collections, especially those of the then acknowledged expert on the Mesembryanthema.Louisa Bolus in Kirstenbosch and Rondebosch.Haworth (1803) subdivided Mesembryanthemum and placed the two species known at that time, M. tigrinum and M. felin u m , later considered to be species of Faucaria.in the section Denticulata.Both species recur in the new section Ringentia (Haworth 1812) together with some species now placed in the genus Stomatium.Schwantes (1926) erected Faucaria.accommodating M. bosscheanum.M. felinum.M. lupinum.M. tigrinum and M. tuberculosum.Field collections during the nineteenthirties enabled Louisa Bolus to describe 25 species between 1933 and 1938 from living material.Her species concept was very narrow, and often based on minor dif ferences.With the exception of F. subintegra these species turned out to be synonyms of already known Faucaria species.Hartmann placed Faucaria in the Stomatium group, one of the 11 groups in the Ruschioideae (Hartmann 1991).This group is charac-terised by hom ocellular xeromorphic leaves with crystals in irregular zones, flowers with five separate nectaries, capsule with or without covering membranes, without closing bodies, and with valve wings broad to reduced.

M ATERIAL AND METHODS
Live material of more than 60 field collections were studied over the period [1985][1986][1987][1988][1989][1990][1991][1992].An extensive field study in 1991 completed the taxonomic research.SEM studies of epidermis, pollen and seed of fresh material were carried out with a JEOL JSM -5200.This material was dried using a Balzers Union CPD 020 critical point dryer with carbon dioxide and subsequently sputter-coated with platinum using a Polaron Sputter Coating Unit E5100.The morphological studies were made with a Zeiss Axiophot photomicroscope.Germination experiments with seeds of 11 accessions harvested in the wild were carried out in Petri dishes, placed in a germination incubator for 12 hours at 25°C, and 12 hours at 20°C.The material for microtome sectioning was embedded in Kulzer's Technovit 71(X) and stained with 1 % Toluidine blue in 1 NHC1.A 4% solution of formalin was used to examine the idioblasts (Hartmann 1977).The presence of tannins was established by blue colouring of tannin with a \% FeCI3 solution (Hartmann 1977).Cytotaxonomic studies were done on root tips of germinating seeds, stained with Feulgen reagent.Besides the living accessions, 300 herbarium sheets were studied from B. BOL.G, GRA.HBG.K. MO.NBG.PRE, SAM and WAG.A cladistic analysis was carried out with all eight taxa of Faucaria using Orthopteruni L. Bolus as the outgroup.Thirteen characters were investigated (Table 1).The most parsimonious cladogram was determined using Hennig86 (Farris 1988).
The generic name is derived from fa u x (= jaw.throat); Faucaria = a collection of jaws, likening the toothed leaves to the open jaws of an animal.
Distribution: western part of the Eastern Cape, wide spread in scattered populations (Figure 2A).Bolus (1930).She noted that the chief difference between F. albidens and F. bosscheana seemed to be the length of the stamens.The broad-leaved form is common in cultivation and has therefore been regarded by growers as the real F. bosscheana.All listed synonyms of F. boss cheana are based on characters such as leaf form and num ber of teeth.These characters are quite variable and may vary from koppie to koppie according to Hammer (1991).They have in common the smooth epidermis, often without conspicuous white dots, the few-toothed, uniform green to brownish coloured leaves and (not) continuous white mar gins.F. haagei was a cultivated variant of F. bosscheana, characterised by dark green broad leaves, undulating or often entire leaf margins.F. paucidens, like F. albidens, is a narrow-leaved form of F. bosscheana, described from a plant in cultivation.According to Hammer this form varies considerably around Graaff-Reinet. F. peersii is a normal variant of F. bosscheana.not distinct at specific level.
Faucaria felina subsp.tuberculosa has been regarded as a distinct species up till now.However, this taxon dif fers from F. felina only in the tubercles on the leaves.As such F tuberculosa can not be classified as a species on the tubercles only and should be regarded as a subspecies of F. felina.Some succulent lovers will have problems with the rejection of F. ryneveldiae and F. longifolia as separate species.The long, narrow, few-toothed leaves seem distinct.However, this rare leaf form occurs on widely diverging localities, e.g. in Addo (Liebenberg 7697).As genetically isolated populations develop the same leaf type, it is still hard to describe them as a single taxon, as F. longifolia and F ryneveldiae differ in other characters.The ovary top of the type specimen of F longifolia is hollow, whereas that of F. ryneveldiae is acute.Intermediates with rather long leaves and more teeth also occur.Further research must indicate whether F. ryneveldiae should remain a synonym of F. felina s.I. or regain species status.With some hesitation F. lupina is regarded here as conspecific with F felina.The differ ences are small, the leaves are greener, slender, more glossy, and the teeth normally have rather long bristles.The number of teeth is variable, from 0-10.In the notes of Haworth (1824), F. lupina was described as having longer and narrower glaucous leaves, and with longer and twice as many teeth as F felina.Salm-Dyck ( 1836) gives a more detailed description and says that this plant differs from F felina by having greener, shiny leaves, without points and more teeth with bristles.The teeth on the keels described for F. militaris occur occasionally in F. felina.F felina subsp.felina grows together with F. bosscheana in a large area.However, they do not grow sympatrically, F. bosscheana prefers a more karroid area than F felina.Intermediate populations are short-leaved, and the leaves are partly white margined and have a smooth epidermis.The leaves bear 3-6, long-needled teeth.F coronata L Bolus: 446 (1934).Type: Graham stown Dist.. between Carlisle Bridge and Riebeeck East.G .V Britten 34.fl. 15 April 1934 (BOL.holo.!).
Faucaria felina subsp.britteniae is unique in the genus because of the (partially) whitish grey epidermis.This taxon deserves the rank of subspecies by the pres ence of numerous very small dots on the leaves, papillate bristles and restricted area of distribution.Additional fea tures are the large number of teeth and the size.The most characteristic specimens with many, but rather small teeth are tound at Ecca Pass, north of Grahamstown.Plants clumped or single.Leaves spreading, 4-8 per rosette, crowded, thick, ovate-rhomboid.15-20 x 15-20 mm, keeled, with 3-6 short, whitish stout teeth on each margin, almost without bristles, some teeth may be rudimentarily developed; surface with many epidermal whitish dots, sometimes confluent on leaf margins and keel; upper surface with blunt, whitish tooth-like tubercles (this may vary from a single tubercle, to, very rarely, a complete tuberculate surface).Flowers yellow, as F fe li na.Stigmas 5, filiform, free to the base.Nectaries yellowbrown, large.Fruits and seed as in F felina subsp.felina.Distribution: Bedford surroundings; field data very scarce (Figure 4).

Specimens examined
Ecology: grassland, often camouflaged by grasses in association with many herbs.Rainfall: 493 mm.Geological formation: Beaufort group, Adelaide sub group (Karoo Succession).
As Bolus (1930) already noticed, the aquaprint of the united stigmas in C urtis' Botanical Magazine is not cor rect, in Faucaria these are free to the base.
At a glance F. felina subsp.tuberculosa can be identi fied by the tubercles on the leaves which are unique in the genus.A typical plant is compact, with whitish dots and patches and resembles F. tigrina.Cultivated hybrid ised plants often show tubercles more prominent than in wild material.
Fruit stalk absent or up to 1 x 3 mm wide, 1 mm thick.
Distribution: restricted to a radius of 3 km around Riebeeck East (Figure 2C).
Ecology: Grassy Fynbos with forest groups.The ecol ogy of the plant is peculiar: without shelter in compact clumps in sandy soil in cracks and hollows of flat sand stone plates.Rainfall: about 380 mm per annum.Geological form ation: sandstone, Karoo System.Witteberg Series, Kommadagga Subgroup.
F. gratiae and F hooleae are known only from three herbarium sheets.The descriptions do not differ signifi cantly, the type specimens originate from the same pop ulation.The plants are very small with short leaves and irregular margins, rather different from those of F. felina.Van Jaarsveld (1990) visited the locality and rediscov ered these plants.I examined this rare species at the same location in 1992, and collected it.3 km further away.
Distribution: from a single locality near Alicedale (Figure 2D) in the Eastern Cape.
When verifying the Faucaria sheets from BOL. three sheets could not be determined satisfactorily.The speci mens looked similar to F. subintegra but their leaves were different.They originate from a single locality near Alicedale.and have been collected between 1937 and 1939 by Mrs E.E.A. Archibald and Grace Britten, both lrom Grahamstown.Louisa Bolus grew field accessions in her garden, and recognised these specimens as some what different.She made an additional note on the sheet: 'may be only a variety of F. subintegra'.She referred to the habitat by her note on the sheet Faucaria nemorosa.1.e.Faucaria from the forest.She did not publish this name.The live collection of the Kirstenbosch Botanic Garden includes a plant, collected by Ernst van Jaarsveld in 1985, very much like the Archibald specimen in the Bolus Herbarium.This peculiar plant bears soft dull greyish green leaves without any sign of whitish dots, and the teeth are small and indistinct.Like the herbarium specimens ol F. nemorosa the leaves are similar to those of F. subintegra.but longer and narrower.Moreover this plant has been collected in the same area as the original F. nemorosa.The scarce, conserved fruiting material was of the normal Faucaria type, but not like those of Orthopterum.Unfortunately the plant could not be traced again in 1991.even after repeated efforts.
Study of the living material confirmed some relation ship between F. nemorosa and F. subintegra.The reduc tion of the teeth in F. nemorosa is almost complete, being reduced to small tubercles.The fruits are similar, rather small, tightly attached on the plants and often broader than long.The scarce ecological data on F. subintegra indicate differences in habitat between the two taxa.F. subintegra is restricted to shallow hollows or crevices in bare sandstone rocks in full sun under very dry circum stances.F. nemorosa grows in soil in the understorey of shrubs in partial shade.SEM micrographs of F. nemorosa show cuticular folds which are also found in F. subinte gra (Figure 5D). a common feature for Faucaria, that is missing in Orthopterum (Figure 5H).Plants clump-forming, in older stages stem-forming; stems 4 -7 mm diam.(Figure ID).Leaves 4-8, crowded, very thick, spreading, ovate-rhomboid, 20-35 x 10-15 mm, bluntly keeled at top, margins with 0-8 very short conical teeth having short bristles on each side that dis appear in older leaves; often shiny, glaucous, sometimes with a violet lustre, with minute white dots, margins and keels indistinctly whitish.Flowers with 80-92 linear petals; stamens 150-180.Nectaries yellowish green.Fruit stalk absent or 2 -4 x 2^4 mm, 1.0-1.4mm thick.Fruits tightly attached on plant, 7-10 x 9-12 mm, 6-9 mm thick; top of capsule flat, or in most fruits a raised centre.Seeds dark brown, 1.1-1.4x 1.0-1.1 mm.Flowering period: March-M ay.
Distribution: the m ost south-central part of the Eastern Cape (Figure 2E).
Ecology: Subtropical Coast.Open Valley Bushveld with grassveld in between, overgrazed, in full sun on open places on top of north-facing river bank, cavities in rock plates in Oat sandstone rock under very dry circum stances.Rainfall: 500-900 mm (Valley Bushveld) or more (Subtropical Coast).Geological formation: Sandstone of the Beaufort Group.Adelaide Subgroup (Karoo Succession).
F. subintegra possesses leaves with indistinct teeth and often rounded leaf margins and keel.The epidermis is smooth with indistinct small whitish dots.F. subindurata was described from the same area in the same publication, and is very clearly conspecific.The two names therefore have the same priority but the name F. subintegra is chosen here because it better describes the almost entire leaf margins in older leaves.
Intermediate plants have been found between F felina subsp.britteniae and F. subintegra.The leaf shape was that of F subintegra. the number and shape of the teeth those of F. felina subsp.britteniae.It is unknown if these intermediate populations are the result of natural hybridi sation.
The species was discovered as early as 1790 (Bolus 1927), probably because of its accessibility.Faucaria tigrina is a compact, crowded, distinctly flecked plant having 9-10 distinct teeth with long bristles on the short leaves.The many white-coloured flecks on the often red dish leaves resemble the lichens and reddish rocks in its natural surroundings.Some populations of F. felina subsp.britteniae at H ell's Poort.35 km west of Grahamstown and near Fort Beaufort, look like F. tigri na because of the numerous large flecks on the leaves and the long bristles.However, the leaves of F. felina subsp.britteniae are larger, not pressed against each other and the number of teeth is reduced to 5 or 6.Forma splendens.published by Jacobsen & Rowley in 1955.w as recognised on the basis of its reddish leaves, typical for some populations.F. tigrina 'Superba' is a superflu ous synonym, used in seed catalogues to differentiate good plants and seed, reintroduced by Mrs Dudley Ryder, from the usually inferior hybrids named F. tigrina (Weber 1968).

CYTOTAXONOMY
The chromosome number of Faucaria is usually list ed as 2n = 18 (Moore 1973).The chromosome number of the studied material of all species was also 2n = 18.The chromosomes are however difficult to study, and this number is therefore based on a few plates only.

CHARACTER ASSESSMENT
The classification of Faucaria is based on a combina tion of variable characters (Table 1).There could be even more variation within taxa than between taxa.These diffi culties are reflected in the key; it proved quite difficult to construct a usable one.In this treatment, considerable importance is awarded to epidermis characters, being one of the few fairly usable characters.Other characters such as number of teeth or leaf shape proved to be too variable for making reliable taxonomic decisions.Cladistic analysis is also hampered by this lack of usable information and must be regarded as uninformative.Despite these shortcomings, the authors find this monograph important as the first tax onomic treatment of this horticulturally important genus.

Leaves
The leaves of Faucaria show a distinct interspecific variation, whereas the variation in separate populations of one species is usually smaller.In the pre-juvenile stage, all species lack flecks and teeth, are fresh green, often with white margins.In the juvenile stage the subsequent leaves gradually develop the characteristics of the adult leaves (Figure 6A).In the adult stage the successive leaves are similar.In conditions of poor light the plants stay longer in the pre-juvenile and juvenile stage than in full sunlight.The ontogeny of F. bosscheana is very short, compared to the other species; it reaches the flowering stage while still in possession of juvenile leaves; the neoteny o f this species could be initiated by the very dry circumstances of its habitat, which tend to induce a short ening of its development.It has been noticed that extreme conditions promote the development o f taxa with neotenic character states (Ihlenfeldt 1971a).
The shape of adult leaves may vary from ovate-rhomboid to ovate-lanceolate, depending on species, popula tion and environment.Individual populations could be recognised in some cases on the basis of leaf characters.The leaves of Faucaria very often bear teeth on the leaf margins in the upper part.Their development is simultaneous with that of the leaf, but very young leaves bear relatively large teeth.These teeth develop succes sively, the differences in size could be large in young stages; e.g. in a leaf primordium of 1.5 mm length with the largest tooth at 1.3 mm, the ratio with two other teeth is 16:3:1.In later stages the differences between the teeth become gradually smaller.The teeth are emergences of the leaf because deeper tissue layers contribute to the for mation of the projections.The teeth consist of epidermal and hypodermal cell layers that are characterised by strongly thickened collenchymatic or sclerenchymatic cell walls.Some populations have leaves with small, poorly developed teeth on the upper side, the so-called tubercles.These usually functionless tubercles may be present on the margins or elsewhere on the upper side.Plants with tubercles on the upper surface have often been cultivated under the name F. tuberculosa.In the ontogeny, teeth appear in the 3rd or 4th leaf pair.In adult leaves the teeth are most prominent in the young stage, normally decreasing in size with age.The number and size of teeth vary with each taxon, growing circum stances, and even the population.
The teeth often bear bristles, depending on the taxon.Certain taxa like F. tigrina always possess bristles, whereas others, like F. bosscheana, usually lack them.A strong development of bristles seems to be correlated with well-developed oxalate dots in the epidermis.The presence of papillae on the teeth is also variable, they are only numerous on the teeth of F. tigrina.In F. felina s.I. the teeth are smooth to moderately papillate, some acces sions are very papillate, whereas other populations are hardly or not papillate.F. bosscheana, F. gratiae, F. nemorosa and F. subintegra show no or only a few papil lae.

Idioblasts
Idioblasts occur on the stem, leaves and receptacle of Faucaria (Figure 7A, B).They are roundish, spherical or elongated.The elongated idioblasts in the lower part of the leaves and the central tissue are orientated in the growth direction of the leaves.The subhypodermal idioblasts are visible as longitudinal or roundish dots in the epidermis.They are very often obscured by calciumoxalate dots that develop above the idioblasts.They may or may not contain bundles of needle-like calcium oxalate raphides.The idioblasts vary considerably in size: small roundish raphide idioblasts are about 80-185 [xm where as only liquid-containing idioblasts may vary between 100 and 650 pm.The most and largest idioblasts may be found at the strongly cutinized edges.Subepidermal idioblasts contain mostly tannin-like substances, idioblasts in the central tissue are often tannin-free.Dark brown-coloured tannin idioblasts occur especially in old leaves.From the outside they can be localised as small dark brown or blackish dots.It could not be established whether the percentage of tannin-like substances in decayed leaves is higher than in younger leaves.

Flowers
Flower shape and colour within Faucaria is not very variable, and only some differences in petal and stamen number can be observed between species.Variation in colour in the nectaries has some taxonomic value as was suggested by Schwantes (1957).But even within popula tions the colour can vary between green and brown.The single flower (Figure 6D  The receptacle is 2-keeled and sometimes slightly constricted at the apex, and may form a short tube up to 1 mm long.The top of the inferior ovary is flattened to conical, with 5 grooves, sometimes with another 5 alter nating less prominent grooves.Bolus (1934Bolus ( , 1937) ) emphasised especially the number of these grooves in Faucaria and the slope of the tissue between them to the centre.The middle of the ovary itself is usually elevated for 1-2 mm.A considerable variation exists in the con struction of the top of the ovary, and may even vary between flowers of the same plant.

Fruits
The top of the ovary, and immature and mature Faucaria fruits vary in shape between the taxa (Bolus 1937).The variation in shape of the top of the ovary and the immature fruits is so large that this character is not used in this study.The mature fruits show less variation among the taxa and is included only in the species description.
In the developing fruit, parenchyma tissue develops from the valves downwards, and splits the lamellae of the five septa.The top of the ripening fruit is often coni cally swollen by this developing parenchyma tissue, the grooves are well developed at this stage.In ripe capsules the parenchyma has disappeared and only five empty hollows remain of the pulp masses, so that, together with the real locules, the fruit seems to be 10-celled (Figure 9C).At first sight the capsule seems empty, because only the hollows are visible, but in fact the locules are hidden by the bent lamellae.The ripe capsule is 5-( 6 layer consists of rodlets of about 6 x 1 pm.On and between these rodlets other small, often bent rodlets are present that measure x 0.4 pm.They also found small differences in the shape of these rodlets between F. tigrina and F. felin a subsp.tuberculosa.From SEM preparations it could only be concluded that slight differ ences between the species exist, but these do not seem to be of any taxonomic use (Figure 11A-F).

Anatomy of epidermis
The wax cover on the epidermis may occur as small particles of different size (Figure 5A, B), but is obscured by the cuticular folds of the epidermis.Form, structure and occurrence of the cuticular folds of 21-36 pm wide depends on the position on the leaf and on the taxon.The folds may form cobweb-like structures in combination with small epidermal elevations of the whitish dots (Figure 5E).A more ribbed structure can be found in species without pronounced elevated dots such as F. subin tegra or F. bosscheana.The most developed cuticular fold structures can be found in species with strongly oxalateencrusted epidermis, like F. tigrina.The cuticular folds are generally less developed on the upper surface than the lower surface of the leaves.A flake-like wax cover is strongly developed in the related Orthopterum waltoniae, where the cuticular folds are absent (Figure 5H).
The epidermis of Faucaria is of the xeromorphic type (Ihlenfeldt & Hartmann 1982;Ihlenfeldt 1983) and is encrusted with calcium-oxalate crystals in the outer walls.Strongly encrusted parts of the epidermis are visi ble as whitish dots and flecks of 0.15-5.0mm or more diam., very often developed above subhypodermal tan nin idioblasts.The outer walls of the epidermal cells in these dots are dome-like and 16-26 pm thick with a thick layer of oxalate crystals, compared to the 8-13 pm thick outer epidermis of the surrounding epidermal cells.These dome-like epidermal structures may be further developed in papillae of 20-150 x 20-50 pm that are located on margins, keels and teeth of the leaves (Figure 5G).Their presence is however variable (see description of leaf).Besides these large papillae, small club-shaped papillae occur on the epidermis, which are about 7.1-9.1 x 2.8-3.9 pm in size.The outer wall of the strongly encrusted epidermis cells reaches to about half-way down the tangential walls, the thickness may vary from 2-4 pm.The inner walls of the epidermis of these cells are 3-6 pm thick, in less encrusted cells 2 pm.
The development of the epidermal calcium-oxalate crystal dots, flecks or zones differs from species to species and may be restricted to the leaf margins, as in F. bosscheana, or even absent in some populations of F. felina subsp.felina.The crystal concentrations are slight ly elevated, giving the leaves a pattern of depressions and elevations.These elevations are also partially due to the presence of subepidermal idioblasts in the mesophyll (Figure 3).
The stomata on upper and lower leaf surfaces gener ally occur in relatively less oxalate-encrusted parts between the strongly oxalate-encrusted elevated dots.The stomata occur regularly in these areas, whereas their density is very variable, ranging from 6-128 per mm2.The anisocytic and paracytic stomata are level with the surrounding epidermal cells.These superficial stomata (Figure 6E) occur also in other genera with a xeromorphic epidermis e.g.Nananthus and Rhombophyllum (Ihlenfeldt & Hartmann 1982).The guard cells measure 22-37 x 6-12 pm.The stomata may be surrounded par tially or completely by cuticular folds of the neighbour ing cells.These projections partially form a secondary outer surface.The epidermal protection of the stomata of Faucaria may be absent or almost sunken in the cuticu lar folds and/or protected by the elevated leaf surface (Figure 5A -F).This type of differentiation was described earlier from the genera O dontophorus, O ctopom a and some species of C heiridopsis (Ihlenfeldt & Hartmann 1982).Instead of surface sculpturing in the form of papillate epidermis cells like these taxa.Faucaria has cuticular folds.

Anatomy of leaves
The leaves possess a large central vascular bundle, occasionally accompanied by some parallel lateral vas cular bundles.A ring of lateral vascular bundles is also present.The distinction between assimilation parenchy ma and chloroplast-free, central, water-storing, spongy parenchyma is not very sharp.Assimilation parenchyma is absent in the dorsally lower part.A faint ring of later al vessels could be observed in the transition zone between the two tissues.

Germination and development
Fresh seed of Faucaria will germinate within a few days after sowing.The germination curve is bimodal like that of Trichodiadema spp.(Ihlenfeldt 1971b).The first seedlings appeared three days after sowing and the first germination peak appeared after six days, the second one after 12 days.New seedlings appeared up to 33 days after sowing.The percentage germination varied from 34 to 94%, with an average of 67%.In the first stage the ger mination is by means of a true operculum, which breaks off before the embryo emerges.Bregman & Bouman ( 1983), who studied seeds of numerous Cactaceae and a few Portulacaceae, found that all operculate species stud ied were succulents, whereas the inoperculate species were not.The cracks may be dorsal or lateral near the micropyle.The germination does not depend on disinte gration of the tough testa.The empty testa may still be attached to the cotyledons.It was noticed in the present study that with germination, the primary root stayed enclosed in the inner integument in some seeds.The pri mary root in these seeds may break finally through the inner integument after days or even several weeks.This delay in germination is advantageous if only a single short rain shower has fallen.
The succulent green cotyledons of the seedlings are slightly broader than long (5:4), with a clear slit between them.With their fused base they form a flat semicircular or elliptic top of the seedling.After germination they grow to about 5 mm in length and width.Dupont ( 1968) has observed seedlings of 79 genera in the Aizoaceae s.I. and found that the cotyledons of Faucaria are about average in their degree of succulence.Differences between species of this genus in size, form and colour are hardly noticeable in the seedling stages.Within some months after sowing, the seedlings develop their first leaf pair.These leaves are uniformly shiny green and possess a white, cartilaginous margin without teeth.At this stage differences between species are not yet noticeable.After a vegetative development of ± seven months to three years the first flowers develop.The flowers appear in autumn, mostly around April/May in the southern hemi sphere and in September/November in the northern hemi sphere.

Pollination biology
Faucaria sets seed easily after cross-pollination.The main flowering period is in autumn, from March till June, but occasionally flowering occurs outside this peri od.Differences in the flowering time between species are usually small; F. bosscheana flowers one month earlier than F felina s.I. in culture in the southern hemisphere.Rain may be important to induce flow'er development, in several cases the plants start flowering after the onset of the rains.Nectar does not seem to play an important role in pollination.The nectaries are not so easy to reach, insects have to pass the stamen brush.This unspecialised floral type is a representative of the cantharophilous syn drome, where pollination takes place by small malachiid beetles.Gess & Gess (1989) found small wasps (Hymenoptera.Masaridae) as one of the main pollinators of the Aizoaceae.Liede et al. (1991) found in a similar ly structured flower of Conophytum subfenestratum Schwantes (= C. p illa n sii) that both insects play a role in the pollination, but that the beetles outnumbered the wasps.Wind pollination is unlikely, as the stigmas are not exposed and dry pollen is not present (Chan 1992).The flowers of Faucaria open around 15:00-17:00.and close at sunset.It has been observed in the field that mixed populations of F a u caria.B ergeranthus and Orthopterum flower simultaneously.This could indicate a combined pollinator attraction by flower synchrony.

Fruit and seed dispersal
Faucaria has variable fruit dispersal.Some capsules become detached easily, others are persistent.The fruits can either be detached actively from the plant by the next developing leaf pair or passively when the very short pedicel w ithers.The seed is not specially adapted for dis persal and combined with the effective closing mecha nism seed dispersal of Faucaria depends on capsule dis persal.There is a close correlation between hygroscopic opening, ombrochory and germination (Hartmann 1988).In Faucaria.ejection of seeds is prevented by a combi nation of deep locules and overgrown cell lamellae.In old capsules most seeds are usually still present and only after complete deterioration, which could be after years, these seeds come free from the fruits.This is an inge nious method for these plants to overcome drought con ditions and re-establish populations after long periods of low rainfall.Observations and small experiments with seed dispersal also indicated that seeds of Faucaria are difficult to wash out from the capsules.This means that ombrochory in Faucaria is not very important.Fruit and seeds of F aucaria are not wind-dispersed (anemochorous) like nearly half of the species of the central lower Karoo community (Cowling & Roux 1987).The relatively large and heavy capsules usually stay within a few centimetres of the plants.The re-establishment on locations after unpredictable long periods of drought in the south-central Eastern Cape is probably more impor tant than dispersal.The establishment of new popula tions is difficult, because of the absence of vegetative dispersal and the poor seed dispersal.

Cultivation
Like many other species of the Eastern Cape, Faucaria is rather easy to grow in cultivation.In nature, leaf succulents in the Eastern Cape grow throughout the year, but a summer peak is apparent (Hoffman 1989).In cultivation it is attempted to restrict growth to spring and summer.The best results in the Netherlands were obtained by growing plants in a sunny place in a heated winter frame, protected from rain in the growing season by glass, watered generously in summer and early autumn.Species of Faucaria overwinter without water ing at a minimum of 5°C.Under these conditions plants grow well and remain compact.If grown in a warm greenhouse, plants at low light intensity produce long etiolated leaves.Soil is not very critical; in nature Faucaria grows in several soils.An appropriate potting mixture is 1 part compost to 2 parts sharp sand and a small amount of clay.Pots of 70-120 mm diam.are suit able.The pH can probably be below 7 according to field data (see species descriptions).Propagation is easy by seed or cuttings.Faucaria species hybridise easily in cultivation, causing frequent confusion (Hammer 1991).
To obtain unhybridised seed it is necessary to isolate species when in flower.Plants sometimes suffer from (root) mealy bugs or they may be attacked by the vine weevil, Otiorhynchus sulcatus (F.).Watering during cold or moist periods may result in rotting of the leaves, ulti mately killing off the plant.

Ecology
Faucaria is found very often under the shelter of open bushes.This 'nurse plant' phenom enon whereby seedlings readily establish them selves beneath the canopy of certain plants is common in most arid and semi-arid areas, including the Karoo (Cowling & Roux 1987).Faucaria usually grows on rocky hill slopes, on top of koppies, on open spots in the bush or more rarely in cracks of rocks.The plant has a preference for sun-facing north slopes, but is not restricted to that position.Field data recorded plants growing in gravel, sand, tillite.sandstone, shale, loam, and clay.The plants are probably not sensitive to low pH.
In this genus a very compact growth is not an adaption against aridity; F. subintegra grows in very dry circum stances and develops a rather long stem.Correlation between location factors and species seems small.The genus occurs in different rainfall regimes, the southern most populations of F. felina subsp.felina thrive exclu sively in a maximum winter rainfall area.F. bosscheana occurs in regions with mainly autumn and summer rain fall.The remaining taxa grow in areas with mainly spring and autumn rainfall.Faucaria does not grow at very high altitudes, and is mostly recorded between 600-850 m, rarely up to 1 500 m.A comparison of the distribution area with the vegetation maps of Lubke et al. (1986) indicates that Faucaria is for the greater part an element of the Tongaland-Pondoland and Karoo-Namib phytoehorological regions, which contain many succulent species from karroid origin, distributed in the thicket (Lubke et al. 1986).The distribution area of F. tigrina is within the boundaries of the southern phytogeographical centre of the Cape Floral Kingdom as defined by Olivier et al. (1983); Bond & Goldblatt (1984); Ellis (1988).The dry thickets of these river valleys were defined by Cowling (1983) as Kaffrarian Succulent Thicket, and as Valley Bushveld Thicket by Lubke et al.Like Noorsveld (Acocks No. 24) and Spekboomveld (Acocks No. 25) these thick ets are classified by Lubke et al. (1986) as Subtropical Thicket.Faucaria has been reported in all three.

Distribution and habitat
The genus is found south of 31° S and between 22° and 28° E. The main distribution area is within the southcentral region of the Eastern Cape, but it extends slight ly into the Western Cape.The genus occurs from the southeastern coast at about sea level to I 500 m in the interior of the Namib-Karoo region.Few locations are reported from the dry northwestern districts of the area, but these have been botanically poorly explored (Gibbs Russell et al. 1984).Faucaria shows a patchy distribu tion; the plants are scattered in populations of different size.Temperature and rainfall change markedly from the coast towards the interior.The more extreme tempera tures occur in the north and west of the area, with maxi mum air temperatures of more than 43°C in summer, and winter minimum falling below -3°C at Graaff-Reinet (Palmer 1989).The unpredictable rainfall diminishes gradually from south to north, the river valleys receive 500-900 mm per annum and the area at Graaff-Reinet has recorded extreme annual precipitation between 144 and 664 mm (Acocks 1988;Palmer 1989).Part of the distribution area of Faucaria is situated in the 'stem suc culent zone' as defined by Jurgens (1986).This vegeta tion type, characterised by tall long-lived succulents, only exists in regions with a moderate variability in rain fall.The diagram of Ellenberg (1981) indicates that Faucaria can grow in regions with a rainfall of 400-700 mm and a rainfall variability index of 2.5-3.5.
The present distribution of Faucaria partially fits the hypothesis of Cowling (1983).He postulated that the semi-arid river valleys of the south-central Eastern Cape, where Faucaria is concentrated, comprise an endemic centre and a relatively ancient centre of karroid taxa.He suggested that karroid species in the south-central Eastern Cape river valleys were more widespread during the last glacial era.Thicket taxa of tropical origin e.g.Euphorbia were established in the Holocene (White et al. 1941).The present-day succulent thicket com m uni ties consist of these immigrants and newly evolved species.The semi-arid climate promotes the division of large or medium-sized populations into smaller ones.
These small populations become isolated from each other, resulting in small, hut often distinct differences between the populations of Faucaria.

Conservation status
Only Hall et al. (1980) listed a conservation status for two Faucaria taxa, F. Candida and F. longidens syn onyms of F. felin a subsp.felina.The conservation classi fication was listed as indeterminate resp.uncertain, which means that insufficient data are available.As for other plant taxa it seems to be more important to protect the habitat than that of Faucaria itself.Faucaria felina s.I. prefers Eastern Cape transitional thicket, comprised of Valley Bushveld.Noorsveld and Spekboomveld.This subtropical thicket is extrem ely poorly conserved (Noorsveld = 0%, Spekboom veld = 1.8%,Valley Bushveld = 1.2%) (Hoffman & Everard 1987).More than 30% has been converted to wheal cultivation or degenerated to wasteland by overgrazing.The southern form of F. felin a subsp.felina (published as F. lupina), is threatened by human activities in its habitat around Algoa Bay.
It is preferable to assign the IUCN code 'Vulnerable' to F. tigrina, as this species is restricted to the surround ings of Grahamstown, where it is threatened by human activity such as housing development.There are about 1 700 plants in seven colonies in the wild, according to a field study of Chan (1992).According to Hilton-Taylor (1996), F. tigrina is also listed as Vulnerable.Two other species with a very limited distribution.F nemorosa and F. gratiae, are not threatened at the moment, but must be regarded as Vulnerable.About 800 plants of £ gratiae exist in two populations according to Chan and the size of the population of F. nemorosa is small but unknown.

DISCUSSION
Classification of species of Faucaria and the genus itself is problematic.On species level, taxa with striking but taxonomically difficult characters such as F ryneveldiae, remain a subject for discussion.Hartmann (1988) described eight fruit types in the Ruschioideae.Faucaria had been regarded in this scheme as being of uncertain position.In Hartmann (1993) Faucaria fruits had been classified as similar to the Leipoldtia type but lacking tubercles.However, Orthopteruni, the close relative of Faucaria.with its small tubercles, does not fit properly in the L eipoldtia type, as will be discussed later.

Relationship with other genera
Faucaria is included in the subfamily Ruschioideae of the Mesembryanthema, a monophyletie branch without tax onomic rank within the Aizoaceae (Caryophyllales).The Ruschioideae comprise 107 genera (Bittrich & Hartmann 1988) and are distributed mainly in southern Africa.
The Ruschioideae are characterised by: ovary almost always with parietal to basal placentation; nectaries crest-shaped, either as separate nectaries or in a ring, rarely flat and inconspicuous, very rarely absent; fruits with expanding keels of mainly valvular origin, never reaching to the centre of the fruit, if not hygrochastic, xerochastic, or breaking into mericarps.
The Ruschioideae are divided into 11 groups, mainly on epidermal, flower and fruit characters by Hartmann (1991).The epidermal characters employed are: epider mis mesomorphic or xeromorphic and homocellular or heterocellular.Flower features used are form and struc ture of the nectaries and the presence/absence of fila mentous staminodes.Fruit characters are very important for this classification.Hartmann (1991) placed Faucaria in the Stomatium Group, mainly on the basis of epider mis features.The problems in classifying Faucaria fruits will be discussed later (see phylogeny Frithia.M ossia and N eohenricia have morphological ly little in common with Faucaria (Table 2).Rabiea with its different floral structure and 7-10-locular Titanopsistype capsule is also not close to Faucaria.Faucaria and Stomatium differ mainly in flower, fruit and seed fea tures; morphologically the latter often looks like a small Faucaria with compact growth and toothed leaves.Flower types are.however, different, the recess flower of Stomatium with a shallow hvpanthium contrasts with the large carpet flower of Faucaria (Table 2).The scented flowers of Stomatium with small petals 0.25-0.5(-1.0)mm wide can readily be distinguished from the scentless flowers of Faucaria where the petals are 1.0-1.25 mm wide.The often club-shaped stigmas of Stomatium.1-3 mm long, differ markedly from the 8-11 mm long fili form stigmas of Faucaria.The light brown, non-woody D elosperm a-type fruit of Stomatium is not clearly dis tinct from the dark-coloured woody Faucaria capsule (Figure 10A.B).The parallel expanding keels and seed pockets are in contrast to the diverging keels and the seed-retaining syndrome of Faucaria by means of bent, split septa.The seeds of the two genera are also different, the light brown Stomatium seeds have a smooth or papil late brown testa, those of Faucaria are dark brown and tuberculate.
The sim ilarities between Chasm atophxllum and Rhinephyllum and Faucaria are restricted to compact growth and similar flower shape.Teeth on the often rounded leaves of the former two are.however, general ly indistinct or even missing, whereas in Faucaria the teeth are usually well developed and often the leaves have white margins.In Rhinephyllum.furthermore, the nectaries are often hardly separated, but in Chasmatophxllum and Faucaria they are clearly distant.The fruit of Chasm atophyllum and Rhinephyllum can not be misidentified as Faucaria fruits, mainly because of the absence of the unique stiff, upright valve wings of the latter.
Vegetatively Orthopterum is closer to Faucaria than to any other genus.Non-fruiting plants of Orthopterum with their smooth, more or less shiny appearance resem ble forms of Faucaria fe lin a , described as F. lupina, even though the leaves of Orthopterum usually have fewer teeth and white dots.
The flow ers are alm ost identical and those of Orthopterum also open in the late afternoon.The real dif ference between the two genera lies in the fruits.Nevertheless the fruits of Orthopterum are the only ones within the Ruschioideae that have close similarities to those of Faucaria fruits (see Phylogeny).
Hybridisation between Faucaria and other members of the Stomatium group is unknown.Up to now only one natural intergeneric hybrid has been reported by Hammer & Liede (1990) between Faucaria and Rhomhophyllum.

Phylogeny
The Stomatium Group is polyphyletic, a cladogram could not be constructed.The artificial character of the Stomatium Group is reflected by the absence of any diag nostic feature common to all taxa.The diversity of the group is also shown by the presence of three of the eight described Ruschioideae fruit types (Hartmann 1991).O rthopterum , Stomatium.Rhinephyllum and Chasm ato phyllum resemble Faucaria as mentioned earlier (Table 2).Only the fruits are considered suitable for a meaning ful examination of relationships within the Stomatium Group.Five out of the nine genera in the group possess D elosperm a-type fruits (Hartmann 1988).The fruits of this type have parallel expanding keels, probably an apomorphic character (Hartmann 1988) Close examination reveals that the fruits of Faucaria might be derived from those o f Orthopterum by reduction and lignification.They share split septa, which bend over the locules to resemble covering membranes.The fruits of both genera lack real covering membranes.The upper part of the split septa of Orthopterum is translucent, flex ible and raised, the lower part is partially lignified.This expanded part has disappeared in Faucaria and the lamel lae are completely lignified.Both fruits have divergent, non-broadened expanding keels.The expanding keel of Orthopterum is long, that of Faucaria short.The long awn of the expanding keel of Orthopterum is reduced to a short subulate point in Faucaria.The large, broad, reflexed, flexible valve wings of Orthopterum have been reduced to small upright wings in Faucaria, standing rigidly over the locules.The valve wings of Orthopterum are reflexed, away from the opening fissure of the locules.In spite of differences in lamellae, expanding keels and valve wings, the basic fruit structure of the genera is the same.Hie common features indicate a common ancestor, confirming the hypothesis of Schwantes (1957) that they are related.The presence of closing bodies in O rtho pterum has no influence on this relationship, as the taxon omic importance of these structures cannot be determined (Hartmann 1988).The ontogeny of the fruits also sug gests that Orthopterum and Faucaria are close relatives.Mature lruits of Orthopterum are conical and acute, pro viding space tor the raised septa.The ripening fruits of Faucaria also develop this conical top (Figure 6C), but this plesiomorphic feature disappears in later develop mental stages, as the septa are lower, and the ripe fruits are flat-or low-topped.Hartmann (1993) assumed that Faucaria fruits are similar to the L eipoldtia type, the striking feature of the genus being firm, stout valve wings fused to the valves in their basal part only and standing rigidly above the locules after the capsule has opened.Faucaria shares with the L eipoldtia type the distinct, non-broadened radi al awn and expanding keels.Other fruit types (Drosanthem um , Titanopsis) show broadened expanding keels.However, the two apomorphic characters of the Leipoldtia type (Hartmann 1983): large closing bodies and the elaborate covering membranes, are missing.The large valve wings of Ortliopterum and the small valve wings of Faucaria both occur in the Leipoldtia type.The large valve wings are common, the reduced valve wings of Faucaria are only shared with Antitnima.Orthopterum fits even better into this fruit type, as it has small closing bodies similar to those in Jordaaniella and Fenestraria.The presence of these small closing bodies does not prevent classification of Orthopterum in the L eipoldtia fruit type.The pentam erous fruits of Ortliopterum and Faucaria are also represented in this fruit type, e.g. in Antimima fruits.The absence of cover ing membranes in Faucaria and Orthopterum is the sin gle, but essential feature, preventing them from being readily classified in the Leipoldtia type.If the presence of this character prevents classification in this type, then Faucaria and O rthopterum fit into none of the fruit types described by Hartmann (1988).It must be concluded that in the present classification Faucaria cannot be satisfac torily placed.Only Orthopterum is closely related to Faucaria and together they appear to form an isolated group within the Ruschioideae.Therefore it is proposed to re-introduce the Faucaria fruit type of Schwantes (1952).This will fit into the fruit scheme of Hartmann (1988) with the following characters: Faucaria type of fruit (Schwantes 1952) Expanding keels stout, divergent with distinct radial awn.not broadened, connate in their basal part only, extending rigidly above the locules.expanding sheet pre sent.Covering membranes absent.Closing body absent or small.Genera: Faucaria Schwantes, Orthopterum Schwantes.For a cladistic analysis of the genus Faucaria itself only a few characters are usable, such as crystal dots and teeth.Four Faucaria species show good visible crystal dots dispersed over the leaves in the outer epidermal layer.The three remaining species, F. bosscheana.F. nemorosa and F. subintegra.share with the outgroup Orthopterum the presence of, mostly only a few, concen trations of crystals around the teeth and margins on the leaves.The size and number of crystal dots and teeth are dictated by the taxa and are influenced in a minor way by the habitat.In culture all species turn more greenish than in nature.
The development of hypodermal crystal concentra tions, teeth or bristles reflects the phylogeny in the ontogeny, and indicates the polarity of these characters.The ontogenetic transformations are an invaluable tool in the recognition of characters (sets of homologous states).The absence of oxalic crystal dot concentration in the leaves in the juvenile stage of Faucaria indicates that this feature is an advanced character.Because of the neotenic character of F. bosscheana.this species is con sidered basal in the genus by the phylogenetic analysis (Figure 12).The differences with the other species are.however, not so large as the diagram suggests.The dia gram reflects a close relation between F. nemorosa and F subintegra.Both show' a reduction in tooth development.The fruits are similar in form and attachment on the plant, and the two species have the same, somewhat shrubby grow th.The ecology of the two species, howev er.differs markedly: F. subintegra grows in the open, on very dry river banks.F. nemorosa is found under shrubs.A close relationship exists between F. felina s.I. and F. gratiae.F. tigrina is also related to the F. felin a Group, but is distinct in morphological and epidermal features.

CONCLUSIONS
Faucaria together w ith Orthopterum forms an isolat ed group within the Ruschioideae.The genus consists of six species.Tw o species.F. bosscheana and F. felin a s.I. occur over a large area, four are much more localised.The most common and widely distributed species, F. f e li na s.I., covers most of the diversity in Faucaria.more than three quarters of the names in the genus were based on forms of this species.The variation is continuous, only three forms could be recognised as separate taxa.giving rise to subsp.britteniae.subsp.felina and subsp.tuberculosa.Macro-morphological information was suf ficient for a proper taxonomic analysis in the genus.Epidermal characters proved important in this genus.Neither seed surface, nor pollen morphology provide taxonomically important data; similarly the structure of flowers and fruits is of limited taxonomic value.
The cladistic analysis is based on only 13 characters (Table 1) and must be regarded as preliminary.The cladogram reflects the impression that Faucaria has arisen from a toothless or few-toothed ancestor.F. boss cheana, with its neotenic tendencies, is probably the less advanced species with a relatively simple epidermis and leaf structure, with little ecological specialisation in karroid circumstances.F. tigrina, on the other hand, shows a complicated epidermis structure, is ecologically spe cialised, even showing mimicry with reddish lichens.
FIGURE 5.-Scanning electron micrographs of leaf epidermis.A, B, F. bosscheana, Bayer 4965.A, epidermis above showing rather even surface with small dispersed wax particles; B, under surface with a ribbed structure o f the cuticle and few wax particles C, F. fe li na, Van Jaarsveld 6897, under surface with cuticular folds.D, F subintegra, Van Jaarsveld 6950, under surface with coarse ribbed structure.E, F, F. tigrina, Mullins s.n.: E, under surface with raised dots and stomata between them; F, upper surface with well-devel oped cuticular folds.G, F. bosscheana, Bayer 2011b, margin with semi-globose epidermis cells.H, Orthopterum waltoniae.Hammer s.n., under surface without cuticular folds and with large flakes o f wax.Scale bars: A, D, E, 100 pm; B, C, F, G, 50 pm; H, 10 ym.

Stems o f
Faucaria taxa are short, erect or creeping, or may be virtually absent.The stems increase in length, to a maximum of about 150 mm long and up to 10 mm thick.The woody stems show a limited secondary anom alous growth and cork development.The internal struc ture consists of successive rings meristem in the phloem or pericycle.Cork develops in the inner part of the cortex or inside the endodermis.The stems are branched, so that small clumps develop.The leaves at the top o f the branches cover the stem completely, whereas older parts of the stems are mostly leafless.
FIGURE 7 -Subepidermal idioblasts in leaf section.A, B F. subinte g ra , Van Jaarsveld 6950.The dark coloured tannin contents have disappeared in some cases during preparation Scale bar: 160 pm.
FIGURE 8.-Variation in numbers of petals o f Faucaria species.
, quite different from the diverging expanding keels of F aucaria and Orthopterum.The fruits o f Stomatium and Rhinephyllum belong to the D elosperm a type, those of Chasm ato phyllum to the Drosanthemum type.This last-mentioned type has translucent covering membranes which are absent in Faucaria.Out of the genera involved, only the fruits of Orthopterum show similarities with those of Faucaria (Figure 10B, C).