B othalia 34,2: 141-153 (2004) V e g e t a t io n o f h i g h - a lt it u d e fen s and r e s t io m a r s h la n d s o f th e H o t t e n t o t s H o lla n d M o u n t a in s , W e s t e r n C a p e , S o u t h A f r ic a E.J.J. SIEBEN * f , C. BO U C H ER * & L. M U C IN A * Keywords: canonical correspondence analysis. C ape W etlands F ynbos. phytosociology, plant co m m unities, svntaxonom y ABSTRA CT Seepages occurring at high altitudes in the Hottentots Holland M ountains (H H M ) (W estern C ape Province. South Africa) were subject to a phytosociological survey. Relevé sam pling m ethod and classification procedures o f the floristic-sociological (Braun-B lanquet) approach as well as num erical data analyses (num erical classification and ordination) were used to reveal syntaxonom ic patterns and characterize the position o f the syntaxa along m ajor environm ental gradients. Nine plant com m unities were recognized, three o f which were classified as associations, follow ing formal syntaxonom ic and nom enclatural rules o f the floristic-sociological approach Most o f the studied m ire com m unities were dom inated by low-growing clonal restios (R estionaceae). whereas som e consisted o f other types o f gram inoids. The most im portant species determ ining the structure (and function) o f the m ire com m unities on sandstones o f the HHM include restios A n th o c h o rtu s crin a lis, C h o n d ro p e ta lu m d e u s tu m . C . m u c ro n a tu m , E le g ia in te rm ed ia . E. th yrs ife ra . R e s tio su b tilis. R . p u r p u r a s c e n s . cyperoids E p is c h o e n u s villo su s. F ic in ia a rg yro p a , grasses E h rh a rta se ta ce a subsp. se ta cea . P e n ta m e r is h ir tig lu m is as well as shrubs B er ze lia sq u a rro s a . C liffo rtia tricu s p id a ta . E ric a in te n a lla r is and G ru b b ia ro sm a rin ifo lia . P ro tea la c tic o lo r and R e s tio p e r p le x u s dom inate a rare shale band seep­ age com m unity. There are tw o m ajor groups o f com m unities—the fens (dom inated by carpets o f A n th o c h o r tu s c rin a lis and other low-grow ing species) and the restio m arshlands (m osaics o f low tussocks o f R es tio s u b tilis and tall C h o n d ro p e ta lu m m u c r o n a ­ tu m ). The degree o f soil (and water) m inerotrophy was found to be the most im portant differentiating feature betw een the m ire (fen and restio m arshland) com m unities studied. The soils in the centre o f mires were found to have high contents o f peat and show ed very little influence from the underlying sandstone. The soils along the m ire m argins had a greater adm ixture o f m in er­ al soil derived from the sandstone or shale bedrock. IN T R O D U C T IO N The Cape m ountain ranges are im portant catchm ents for high-quality drinking water. A vast quantity o f this w ater is stored in the basem ent rocks before it is released into river system s via seepages (H ew lett 1982). Seepage is a very general term for an area w here w ater percolates through the upper soil layers (m ostly lying over a layer o f im penetrable rock) and usually form s the source o f rivers. It includes the m ajority o f wet slopes as well as m any m ires o f the C ape m ountains. The quality o f w ater is largely determ ined by the soil conditions that percolat­ ing w ater encounters prior to its em ergence at the surface and on its way to a river, hence seepages are o f great im portance to river ecosystem s (B osch el a l. 1986). H ow ever, the total surface area o f seepages that influ­ ence the w ater quality o f rivers is variable and m any seepages can be out o f touch with the river system for a long tim e. T his was called the Variable Source Area C oncept by H ew lett (1961). A fter heavy rains the Source A rea o f a river expands and m uch w ater, that was stored in basem ent rocks or in fens for a long tim e, m oves into the river. During its storage in the seepage areas, w ater is stained by tannins leached from decaying plant litter. T his explains the brow n colour that is characteristic o f m any o f W estern Cape rivers, particularly after heavy rains (K ing & Day 1979; D allas & Day 1993). Some seepages located in high-altitude areas supporting fynbos vegetation have soils rich in peat —accumulated organ­ ic material (Gore 1983)—they can be classified as mires. * D epartm ent o f B otany & Z oology. U niversity o f S tellenbosch. Private Bag X 1 . 7b02 M atieland. S tellenbosch. South A frica. + author for correspondence; E-m ail: e.j.j.siebenCa buw a.nl M S. received: 2001-10-15. These are defined as wet. swampy habitats characterized by peat> soils, regardless o f the chemico-physical properties of the peat and water captured in the peaty soils. Concepts such as bogs and fens refer to specific types o f mires (Gore 1983). The term bog refers to the strictly ombrotrophic (rain-fed) and usually oligotrophic mires found in places with high pre­ cipitation. whereas fens are the mires (minerotrophic or tran­ sitional ) that are fed to a larger extent by water that has per­ colated through the mineral substrate. The Cape high-altitude mires generally qualify as fens; true bogs are rare in the south­ ern hemisphere, although they do occur on some steep southfacing slopes in Western Cape mountains. In the mountains, mires are found at the sources of the rivers and in watersheds. Riparian mires can be formed in places where the floodplain is very wide and the area remains inundated long after a flood has receded. There are four types o f seepages linked to the drainage network o f rivenne system s recorded in the Cape m oun­ tain ranges, namely; 1. w ell-drained slo p e se e p a g e s supporting soils o f the F em w ood form (Fry 1987); this type show s a high level o f variability and can be characterized by the increased presence o f B runiaceae; C am pbell (1986) in his structur­ al classification o f the Fy nbos B iom e. classified the veg­ etation o f these seepages as Wet E ricaceous Fy nbos; 2. low-altitude va lley see p a g es characterized by high soil water levels and peaty Cham pagne soils (Fry 1987); Boucher (1978) described the E ric a -O s m ito p s is Seepage Fy nbos in this habitat in the Kogelberg Biosphere Reserve. A subty pe o f the low-altitude valley seepages occurs on tem porarily wet sandy soils. It is dominated by E leg ia fila c e a (Tay lor 1978); 142 B othalia 34,2 (2004) 3, the high-altitud e f e n s , situated at the sources o f rivers: C am pbell (1986) has classified the vegetation o f this h a b ita t as S n eeu k o p A zo n al R estio id F y n b o s and O tterford W et Proteoid Fynbos; 4, in order to distinguish betw een the different degrees o f m inerotrophy in the m ires described in this study, we w ant to introduce the term re stio m a r s h la n d s for the bet­ ter-drained sites at the edges o f m ires o f the Fynbos B iom e, in contrast to the ‘fe n s’ being situated in the cen­ tre o f the m ire. T he restio m arshland supports a vegeta­ tion structural type called S neeukop A zonal R estioid Fynbos (C am pbell 1986). M ost o f the A frican sw am ps (including m ires and o ther types o f m arshlands) are dom inated by grasses and sedges (Van Z inderen B akker & W erger 1974; W eisser & H ow ard-W illiam s 1982; T hom pson & H am ilton 1983; R ogers 1995, 1997). T he fens and restio m arshlands o f the Fynbos B iom e are conspicuously different due to the dom inance o f (often endem ic) R estionaceae (C am pbell 1986). D espite their peculiarity, the w etland ecosystem s o f the C ape m ountain ranges have received little atten­ tion (B oucher 1988; R ogers 1997) and hence deserve a closer look. T his study describes vegetation types found in the rare and poorly studied high-altitude fens and restio m arshlands located in the region o f richest precipitation in W estern C a p e —the H ottentots H olland M ountains (H H M ). T he floristic com position o f these vegetation types and their relationship to m ajor ecological factors are the m ain foci o f this paper. M A TER IA L A N D M E T H O D S S tu d y a r e a : lo c a tio n , c lim a te , g e o lo g y a n d s o ils T he m ajority o f vegetation sam ples used for this paper were recorded from the H H M (betw een 33° 56' S and 34° 03' S latitude and 18° 57' E and 19° 09' E longi­ tude). T his area is situated betw een the tow ns o f Stellenbosch, F ranschhoek and G rabouw in W estern C ape, South A frica—the region with the highest rainfall in W estern C ape and possibly also in the entire South A frica. There are m any peaks reaching above 1 000 m altitude, w here num erous and extensive m ires have developed. We have added som e additional vegetation sam ples from the Du T oitskloof M ountains, w hich are located to the north o f the HHM and som e from the G roenland M ountains, southeast o f the H H M . M ost o f the fens in the HHM are found in the Palm iet R iver catchm ent, although the R iviersonderend and Eerste R ivers are fed by w ater from fairly extensive mire system s. Tw o other rivers originating in the area, the Berg R iver and the Lourens R iver, barely receive w ater from m ires as they originate on very steep m ountain headw alls. The clim ate o f the Fynbos Biom e in the southw estern C ape is classified as m editerranean, with hot, dry sum ­ m ers and m ild, wet w inters. In the K óppen (1931) system the clim ate o f the area is classified as C s b —having m esotherm al (C) clim ate with a w arm , dry sum m er and average tem peratures above 22°C and relatively wet w in­ ters (sb). M ost m ountains o f the Fynbos Biom e have a rainfall betw een 1 000 and 2 000 mm m ean annual pre­ cipitation (M A P), but in the w ettest areas (such as the HH M ) it m ight exceed 3 000 mm (Schulze 1965). M ost low -lying localities receive m uch less—up to 750 mm near the coast, and m ostly less than 400 mm M A P in the interm ontane valleys (Fuggle & A shton 1979). The m ountains in the Fynbos Biome play a m ajor role in influencing precipitation and evaporation. Extrem ely high regional variation in precipitation (Figure 1) occurs due to the w indw ard-leew ard geom orphological dichoto­ my in the m ountains and the fact that w inds can sw eep unhindered over the coastal plains (D eacon e t a l. 1992). Schulze (1965) described a strong gradient in rainfall w ith in cre asin g a ltitu d e — 50 m m o f p re cip ita tio n increase per 300 m increase in altitude. The regional rainfall pattern in the m ountains is variable, depending on aspect o f slope as well as frequency and strength o f northw esterly and southw esterly w inds. M ountains can Mean Annual Precipitation (MAP) 18*4ff 34*00"—' 19*12- Rainfall (in mm) 0 -8 0 0 801 -1200 1201 -1600 1601 -2000 2001 - 2400 2401 - 2800 2801 - 3600 No Data N 17 FIG U R E l .—G rid with M ean Annual P re c ip itatio n in H o tten to ts H olland M ountains (Source: C om puting C entre for W ater R esearch; grid data co m p u t­ ed from w eather data from nine official w eather stations in vicinity, extrapolated using m ethodology o f D ent e t a l. 1987). B othalia 3 4 2 (2004) 143 also receive m uch precipitation (not registered in the rain gauges) from m ist (K erfoot 1968; Fuggle & Ashton 1979) associated with the sum m er trade w inds, locally known as ‘so utheasters’. Sixty per cent o f the precipitation falls in the w ettest four m onths spanning June to Septem ber. This precipita­ tion is m ostly in the form o f rain, but snow regularly occurs in w inter on the m ountain peaks (Fuggle & Ashton 1979). W ind speeds are the highest in w inter on the m ountain tops (Fuggle 1981). This is in contrast with the wind patterns in the surrounding low lands, show ing the highest speed in sum m er. The geology o f the HHM is dom inated by sandstones and shales o f the Table M ountain G roup, especially at the higher altitudes. Som e o f the valleys are underlain by A chaean C ape G ranite and M alm esbury G roup shales, but no m ires were recorded on these latter substrates. The Table M ountain G roup sandstones (T M S) are m ainly from the N ardouw Form ation (De Villiers et a l. 1964). H igh-altitude shale bands o f the C ederberg Form ation accom panied by tillites o f the Pakhuis Form ation occur throughout the area. T he shale bands are situated at high­ er altitudes than m ost o f the seepage areas and virtually all the vegetation types (w ith one exception) described in this study, are reported from sandstone. Soils o f the m ires are classified as belonging to the C ham pagne Form (Soil C lassification W orking G roup 1991). The am ount o f peat is variable; tow ards the outer edges o f seepages, the soil contains a greater portion o f m inerotrophic m aterial (m ainly sand) originated from sandstone. Here the prevailing soil form is the Fem w ood Form . In this study, we have used soils as a m ajor crite­ rion for the location and delim itation o f the fens. M e th o d s o f d a ta c o lle c tio n Vegetation w as studied by means o f plot sampling. T he plots w ere laid out in selected fen and m arshland habitats in a representative way (W esthoff & Van der M aarel 1978). The size o f m ost o f the plots w as 10 x 10 m (som etim es less in the case o f sm all vegetation stands), as this w as con sid ered to be sufficient to record the species rich n ess. In each plot, each plant species w as recorded and the co v er-abundance o f each w as e s ti­ m ated using the m odified B raun-B lanquet sam pling scale (B arkm an e t a l. 1964). T hirty-three plots w ere recorded in the H H M . We also added a fu rth er tw o sam ­ ples from the Du T o itsk lo o f M ountains m erely to point out that sim ilar com m unities also o ccu r in the n eig h ­ bouring m ountain ranges (A ppendix 2). Soil sam ples w ere only co llected from the topsoil (A -horizon). T he particle size d istrib u tio n o f the soils is d eterm ined for the fractions finer than 2 m m 0 . A fter drying and g rin d ­ ing the soil to break up any co ag u latio n , the soils w ere shaken through a set o f sieves. T he sieves o f the fo l­ low ing raster size w ere used: 500 ^ m (to separate coarse san d ), 106/vm (to separate m edium sand) and 63 pirn (to separate fine sand). T he fraction that passes through all the sieves is com posed o f silt and clay. A cidity and resistance w ere m easured in w ater-satu rat­ ed soils using a pH m eter (O rion 4 20A ) and a co n d u c ­ tivity m eter (Y SI 3200). T he fraction o f organic m atter was m easured by titration according to the W alkleyB lack m ethod (W alkley 1935; N o n -A ffiliated Soil A nalysis W ork C om m ittee 1990). A list o f environm ental variables and other relevant inform ation is presented in Table 1. TABLE I —Explanatory variables used in the canonical ordination. Data type: 1: interval scale: R: ratio-scale variables. N: nom inal-scale variable Name of variable Data type Methods of measurement estimation, scales, and identity of states for the nominal variables Soil reaction (pH ) I M easured in H -0 using pH -m eter O R IO N 420A . O rganic m atter (O R G A N IC ) R Percentage calculated by titration using the W alkley-B lack m ethod (W alkley 1935). Slope (SL O P E ) R A ngle m easured by slope-m eter in degrees. G eology N D eterm ined using geological m ap (S A C S ) and by field observations. States: S andstone (S an). T illite (T il). Shale (S ha). G ranite (G ra). A lluvium (Q ua). Soil type N D eterm ined after Fry (1987) and Soil C lassification W orking G roup (1991) and on basis o f field o bser­ vations (ch aracter o f lop soil layer and soil depth). S tates (Soil Form s): M ispah. M agw a. O akleaf. F em w ood. G len ro sa. C ham pagne. Aspect N M easured by com pass and classified into quadrants. States: N .W . S & E. A ltitude (A L T IT U D E) R IX*termined from orthophotographic m aps (scale 1: 10 000); expressed in m. R esistance (R E SIST A N ) R R esistance to an electrical current (£2) m easured using a YSI m odel 3200 conductivity m eter. G ravel (G R A V EL) R °tc o f coarse m aterial (>2 m m ) in soil sam ple. C oarse sand (C SA N D ) R o f coarse sand (0 .5 -2 m m ) in soil sam ple. M edium sand (M S A N D ) R o f m edium sand (1 0 6 -5 0 0 //m ) in soil sam ple. Eine sand (ESA N D ) R o f tine sand (6 3 -1 0 6 //m ) in soil sam ple. Silt (SILT) R Soil depth (S O IL D E P T ) R D epth o f soil until bedrock; estim ated average expressed in cm . o f silt (< 63 //m ) in soil sam ple. B edrock cover (B E I)R (X 'K 1 R E stim ate o f co v er (% ) o f bedrock. Large cobbles (L G _C O B B ) R E stim ate o f co v er (% ) o f large c obbles (1 3 -2 5 cm O ). Sm all cobbles (SM .C O B B ) R E stim ate o f co v er (9 t) sm all cobbles (6 -1 3 cm 0 ) . P ebbles (PE B B ) R E stim ate o f co v er (% ) pebbles (2 -6 cm O ). 144 M e th o d s o f d a ta h a n d lin g a n d p r e s e n ta tio n The vegetation sam ples were stored in a database in the form at o f the National Vegetation D atabase (M ucina e t al. 2000) and using the database-m anagem ent softw are Turboveg (H ennekens 1996b; H ennekens & Scham inée 2001). The original cover-abundance data was trans­ form ed into percentage form at using Turboveg. The data were classified using Tw o-way Indicator Species Analysis (T W IN SPA N) (Hill 1979) follow ed by manual table-sorting using MEGATAB 2.0 (H ennekens 1996a), aim ed at im provem ent o f coincidence betw een the groups o f relevés and groups o f species. Differential species for com m unities and their groups were identified on the basis o f fidelity. The differential species, constant com panions and the dom inant species are identified in each case. A dif­ ferential species is a species that can be used to differenti­ ate betw een one com m unity or a group o f com m unities and the rest at the sam e syntaxonom ic level (W esthoff & Van der M aarel 1978; M ucina 1993). The difference in the frequency o f more than two presence classes (40% ) was taken as sufficient for a species to be considered to be dif­ ferential for one o f the com m unities under com parison. H ow ever, a species can also be ranked as differential on the basis o f distribution o f cover values (dom inant versus non-dom inant) am ong relevés o f the com m unities under com parison. A dom inant species is a species that is con­ stant and has an average cover o f m ore than 25% (W esthoff & Van der M aarel 1973). A constant com panion is a species that occurs in m ore than 60% o f all the sam ­ ples in a com m unity and is not considered differential at the sam e tim e. Less frequent species recorded in plots are listed in Appendix 1. The com m unities are described here at the level o f association or rankless vegetation type com ­ parable to the level o f association. We refrained from defining units o f the higher syntaxonom ic ranks due to limited extent o f the data and local character o f the study. We have used only tw o inform al groups o f the com m uni­ ties based on the habitat characteristics (fens vs. restio m arshlands). The relationship betw een the environm ental data and the vegetation data was determ ined using m ultivariate techniques. We follow 0 k la n d (1996), w ho m ade a case for validity o f use o f both ordination and constrained ordination as com plem entary approaches, both direct and indirect gradient analyses were perform ed The four vari­ ables, coarse sand, m edium sand, fine sand and silt, together com prise 100% in every soil sam ple so they are not independent from each other. A ccording to the rec­ om m endations o f Ter Braak & Sm ilauer (1998) this sort o f (com positional) data, has to be log-transform ed prior to analysis. C orrespondence A nalysis (C A ) was adopted as the in d irect g ra d ie n t an aly sis tec h n iq u e , w hile C anonical C orrespondence A nalysis (C C A ) was used to perform the direct gradient analysis (see Ter Braak 1986; Jongm an e t al. 1987 for details on the techniques). The C anoco 4 program suite (Ter Braak & Sm ilauer 1998) was used to perform CA and C CA . B othalia 34,2 (2004) taxom ic nom enclature (W eber et a l. 2000). Vernacular nam es, using a com bination o f im portant taxa and veg­ etation structure (E dw ards 1983), were coined for all plant com m unities as well. R ESU LTS The fens (and most o f the restio marshlands) of the region have a high cover o f Restionaceae. Only a few seep­ age types are dom inated by grasses or sedges, such as C a rp h a g lo m era ta , E p isch o en u s spp., lso lep is p ro lifer and P e n n ise tu m m a c ro u ru m . Som e R estionaceae typically occurring in the seepages are A n th o c h o r tu s c rin a lis, C h o n d ro p eta lu m m u cro n a tu m , E leg ia th yrsifera and R estio su b tilis. Two gram inoids such as E h rh a rta seta ce a subsp. seta ce a and E p isch o en u s villo su s are also common. The follow ing C om m unity G roups and C om m unities have been revealed in our data: C o m m u n ity G r o u p A : F e n s Fens form the wettest parts of the seepages—they are poorly drained and contain much peat. The low-grown restio A n th o c h o rtu s crin a lis is usually dominant and forms dense mats in between the tussocks o f cyperoid E p isch o en u s villo ­ s u s . Five fen com m unities were distinguished: the Communities A 1 and A2 occur on steep slopes and experi­ ence somewhat better drainage than the flat-habitat fen com ­ munities (Com munities A3, A4 and A5). C o m m u n ity A l : P r o te a la c tic o lo r - H ip p ia p ilo s a T all S h r u b la n d (Table 2, relevés 1, 2) The Com m unity A l is peculiar due to its link to shale bands. P ro tea la ctico lo r is the dom inant species and forms a dense shrubbery 2 -4 m tall. This is the only form o f proteoid fynbos recorded on the slope seepages in HHM . The herb layer covers over 80% and is dom inated by tussocks o f E p isch o en u s villo su s and mats o f R estio p e r p lex u s . O ther im portant species include S e n e c io u m b e lla tu s, S er ip h iu m p lu m o su m ( = S to eb e p lu m o sa ), H ip p ia p ilo sa and O x a lis tru n ca tu la . The stands o f this com m unity were recorded on the eastern slopes o f Somerset Sneeukop at a very high altitude (about 1 400 m). The habitat receives a very high annual rainfall (more than 3 300 mm ), some o f it in the form o f snow, which might persist longer on the southern than on the northern slopes o f the mountain. A dense mist blanket covers the mountain especially in sum ­ mer. It is purported to contribute a considerable additional am ount o f am bient precipitation (M arloth 1903). Cam pbell (1986) refers to this vegetation (in structural terms) as Otterford Wet Proteoid Fynbos. C o m m u n ity A 2: h le g ia th v r s ife r a - C e n te lla e r ia n th a S h o r t C lo s e d H e r b la n d N o m e n c la tu r e o f ta x a a n d p la n t c o m m u n itie s (Table 2, relevés 3 ,4 ) T he nom enclature o f plant species follow s G erm ishuizen & M eyer (2003). T hree o f the w ell-sam pled plant com m unities w ere nam ed according to the rules for syn- This com m unity is found near the sources o f the L ourens R iver on the w estern slopes o f Som erset Sneeukop (1 1(X) m) at high altitudes and receives a high B othalia 3 4 2 (2004) 145 precipitation. The upper herb layer is formed by the dom ­ inating E le g ia th y r s ife r a , whereas the low er herb layer is form ed by a multitude o f species, such as S en e c io u m b ella tu s . H ip p ia p ilo s a and E ric a cu rv iflo ra . This com m uni­ ty, like the P ro tea la c tic o lo r -H ip p ia p ilo s a Tall Shrubland, is quite atypical for the seepages o f the studied area. The differential species, C a rp a co ce sp er m a c o ce a . C en tella eria n th a , O th o n n a q u in q u ed e n ta ta and U rsinia ecklo n ia n a , are all more com m on in typical ericaceous fynbos (Sieben 2003). The E le g ia th y r s ife r a -C e n te lla er ia n th a C om m u­ nity occurs on steep slopes on sandstone. com m unity also represents a typical form o f w hat is described by C am pbell (1986) as Sneeukop A zonal R estioid Fynbos. In one relevé. C a r p h a g lo m e r a ta was recorded as the dom inant sp ec ies—a situation usually encountered in w et habitats at low er altitudes. C o m m u n ity A 3: A n th o c h o r tu s c r in a lis - E le g ia in te r ­ m e d ia T all C lo s e d K e stio la n d A n th o c h o r tu s c r in a lis . Floristically and structurally this C o m m u n ity A5: R e s tio b ifu r c u s - A n th o c h o r tu s c r in a lis S h o r t C lo s e d R e s tio la n d (Table 2. relevés 15-18) A fu rth e r seep ag e (Table 2, relevés 5 -8 ) S c ie n tific n a m e : A n th o c h o r to c r in a lis - E le g ie tu m in te r ­ m e d ia te a ss. n o v a h o c lo co Holotypus: Table 2, relevé 7 This is an extrem ely species-poor seepage com m uni­ ty, w hich is lim ited to the D w arsberg M ountains in the Berg River catchm ent. The dom inant vegetation stratum is a dense layer o f E le g ia in te rm e d ia , w hich grow s 1.2 to 1.5 m tall. Linder (1987) has not recorded this species outside the C ape P eninsula, but K ruger (1978) found it on the Dw arsberg. In this study, it was recorded in several other locaties in the H H M . The low er herb layer o f the com m unity is dom inated by A n th o c h o r tu s c r in a lis and is less dense. D w arsberg receives m ore than 3 000 mm rainfall per annum and the com m unity is found in the w ettest, extrem ely peaty habitats, surrounded by stands o f the F icin io a rg yr o p a e -E p isch o en e tu m villosi and Tetrario c a p illa c e a e -R e s tie tu m su h tilis. E p isc h o e n u s v illo su s, S e n e c io c r is p u s and the m oss C a m p y lo p u s s te n o p e lm a as well as the species m entioned above are the only species present in the vegetation. C o m m u n ity A 4: F ic in ia a r g y r o p a - E p is c h o e n u s v illo ­ s u s S h o r t C lo s e d R e stio la n d (Table 2, relevés 9 -1 4 ) S c ie n tific n a m e : F ic in io a r g y r o p a e - E p is c h o e n e tu m villo si a ss. n o v a h o c lo co H olotypus: Table 2. relevé 9 T his is one o f several seepage com m unities dom inat­ ed by the restio A n th o c h o r tu s c rin a lis. T his clonal sp e­ cies form s dense m ats and is often found intertw ined with E h rh a r ta s e ta c e a subsp. se ta c e a , C liffo r tia tric u sp id a ta and S e n e c io c r is p u s . The vegetation is m uch short­ er than in the previous com m unities. The tallest restio present is E le g ia g r a n d is . w hich grow s taller than 0.5 m together with the tussocks o f E p is c h o e n u s v illo su s. In the lim ited open spaces, low -grow n F ic in ia a rg y ro p a and A n th o x a n tu m to n g o can be found. T his com m unity as well as the C om m unity A 5. resem ble sim ilar vegetation described from the Table M ountain (G lyphis e t a l. 1978: E ric a m o llis Fynbos C om m unity; L aidler et a l. 1978: R e s tio - H y p o la e n a S ubcom m unity). Both com m unities are associated with peaty soils that are w aterlogged for most o f the tim e. F ic in ia a r g y r o p a -E p is c h o e n u s villo su s type is also d o m in ated by com m unity resem bles the previous one closely and it might also be considered as a subassociation o f the F ic in io a r g y r o p a e -E p is c h o e n e tu m villo si. The m ain d if­ ference is in the absence o f S e n e c io g r a n d iflo r u s and F ic in ia a rg y ro p a . but this com m unity also has som e d if­ ferential species o f its ow n, such as G la d io lu s c a rn e u s, R e s tio b ifu r c u s . R e s tio c o r n e o lu s , T etra ria c a p illa c e a and C h o n d r o p e ta lu m m u c r o n a tu m (the last-n am ed reaches far above the dom inant herb layer). N otable is the occurrence o f P r io n iu m se r r a tu m — a typical shrub in C ape m ountain stream s (S ieben 2003). C o m m u n ity G r o u p B: R e stio M a r s h la n d s The habitats supporting all four com m unities o f C o m ­ m unity G roup B are better drained than those o f C o m ­ m unity G roup A. The soils are largely o f m inerotrophic origin and the peat content is low. They are often found on the edges o f the m ires and the w ater drains into the fens. The restio m arshlands are richer in species than the fens. The C om m unities B 1 and B2 show transitional fea­ tures betw een restio m arshlands and fens, through o ccu r­ rence o f species such as S e n e c io c risp u s. C o m m u n ity B l: P la ty c a u lo s d e p a u p e r a tu s S h or t C lo s e d R e stio la n d (Table 2 .re le v é s 19-21) T his is the only seepage type dom inated by restio P la ty c a u lo s d e p a u p e r a tu s . w hich form s dense green m ats and is the m ost conspicuous differential species o f this com m unity. The tussock-form ing R e s tio s u b tilis is the co-dom inating elem ent. Together they form the low er herb stratum . O ne em ergent 1.5 m tall restio. C h o n d r o ­ p e ta lu m m u c ro n a tu m . form s its ow n stratum . E p is c h o e ­ n u s v illo su s. E le g ia n e e s ii and T etra ria c a p illa c e a are significantly shorter. A part from the eponym ous, the only other differential species o f this com m unity are the grass P e n ta m e r is h ir tig lu m is and the geophyte K n ip h o fia ta b u la r is , flow ering after a fire. We believe that this com m unity, although only recorded in our study in three sam ples, due to being visible after a recent fire, is quite com m on in the Palm iet R iver catchm ent. It seem s to occupy an interm ediate ecological position betw een the F ic in io a r g y r o p a e -E p is c h o e n e tu m v illo si A ssociation (from peaty soils) and the T etra rio c a p illa c e a e - R e s tie ­ tu m s u b tilis A ssociation (from m ore m in ero tro p h ic soils). The localities o f this com m unity receive less rain­ fall than o th er seepage types, with a M A P o f just over 2 (XX) mm. 146 B othalia 3 4 2 (2(X)4) f^i i/“> ( N (N - r^, (N - c^i - f^i (N - X U. E- 22 ' t a_ u. E- 22 Tf D£ X f- 22 - t <N 3C Q ® o> r~~ oc r- X s sC vC r*1 I CL Li . H 22 •' t - sC f^i sC f N X — — H 22 ^t O - <N O ' ' í i /l x CL u. 02 ^T fN O ' - ( N c*“i fN X CL _L. E- 02 ' t 04 v~, X CL u. h- ffl ci X CL u. E- 22 a. H SC e*', — t- 2 : r*'. c*“i “ <N OC DC X (N <N vC sO <N X ■'í •' t Í N »Ti - ( N r^- *t *t <N Tf ÍN *t <N 't r- *í sD fN - -■ - Í N <N ~ - * ac Q. u. H ' í’ £L — t- X CL u. E — 05 <N X L. E Cu U — X QQ r*~, CN CM - - O «o >c ■'t o. H 02 - <N O - c^, vC Tf L. t CL U 22 - CL u. 02 - - O' O ' C *"i i/"> ■*T 30 x u. t- — 3C rj ° C+', DC O' X — u. i- < I/-, —+ - r~ rj ° <N X X = — u. E- < ir. + + + — -C + “ >c (N - - X r~ X u u. H < ir. ---- + + - <r. - X 3C r~ r- 2= 0£ tt. E- < ir. - -t <N </%O' - n >/-. - r^( ÍN Tf *f LL H < •rt + + -C - - r*-t = — X t- < •*t + Í3 ~ r) <N *t ro - - rj = UJ u. f- < - - rj r**t ri - c V, x 2Í u. H < -r —< - 3 <N -• - - ® *t - 0£ (JUE O' rsj (N sC O' O' 02 ÍL f- < -t oc fN -■ 22 U. E- < r~ rj rj O' - O rj = 22 LL H < sC rsi rj X - ° - = o I 22 X í- < r^, V. <N ri r-' - u I E ' I I •*t fN ^t rj - - - x -> cr. í- < -• <N fMri 5 I 5 í — H < -■ ■y'. t- < rj r^. - o - - - - *r - aí & y; < x 0£ -y. V'. < - “ s |1 1 £ 5 -cT ^ !í 7 V -2 5 ^ ?■ A •- ? 5 ■2 S' ^ c < < £ £ £ £ <<<< < < r r x ZI I 'T .'T .y r .y r. < < < < < < < f l l Í 5f 5 1 1 1 :1 < z -2 2 ■ 2 . 2, <; t fM I— < < <<< ^ >* i ' ^ .r c ’E 3 E 3 E 3 e 5 £ E £ 2 3 M 1 ! 1 1 1 J 1 ? !1 i "3 'tl o ■U£? .i 0£ U. 3! 2- 2 c £ a. ac t ? 3- v l5 i2 l.£f^ v. v ■a ÍtjC O ll II i r -2 - il H T. I * Êf í ' | -2 -c i i 'o3 | | 147 B othalia 34,2 (2004) C3 + U- E + _C — m E + . o + — E 5 C Tf E a — c i E E + 5 + — + — + + C E r r + E ---- + ----- + _ + — E - + + —+ + c3 + - - v ~ . r *~, rr. s + E E — E — s + ■C — - — :o E - r 1- .c + + E + E^ aj ------ E — . Cu - . a + a a E + E E + E E n n — ■'* + « « E E E E E — C S »*} »*", -1- + E + -r + + E ■* — Z si s i s i si r j r s c j <N n Ï Ï Ï Ï Ï _ n n rj < < < SC Z ■=; Ï Ï cc x scccecc ccc ecc c r j 5 5 : x / 5 ? * 5 l -C •s 5 « -S 2 a » a .< 2 !fc 5 * a ;c 5 " c ■£ * i e 3 3 •§ 1 g t 5 -c 'c 5 ? £ ■fe k a: ac k i Uji l 1 4 5- * j- s f l l S .e .c E sc ac r cc r £ x 5 21 CC £ £ < .c .e ^ i H 5 . 1 1 1 ^ W^ ^ ac < < x r £ r r r v, v , v . ir. i/". </-. < << < < < x x x x x x 148 B othalia 34,2 (2004) B othalia 3 4 2 (2004) C o m m u n ity B 2: E r ic a a u tu m n a lis - R e s tio p u r p u r a s c e n s T all C lo s e d R e st io la n d 149 am ongst others, B ru n ia a lo p e c u r o id e s. B e r z e lia s q u a r ­ ro sa . E r ic a fa s tig ia ta , G r u b b ia ro sm a r in ifo lia and R e s tio b ifid u s. (Table 2, relevés 22, 23) This is one o f the tw o com m unities described from riparian m ires. From the point o f view o f hydrology and species com position these com m unities are closely relat­ ed to the Restio M arshlands, hence they are classified as such in this study. These com m unities occur along the highest reaches o f the rivers (in this study all were sam ­ pled in the Palm iet R iver catchm ent) and have a m ixture o f seepage and riparian elem ents m aking up very species-rich com m unities, both in com parison with other seepage types as well as with riparian com m unities. The E ric a a u tu m n a lis -R e s tio p u r p u r a s c e n s com m unity has a m ore prom inent tall herb layer than m ost o f the seepage com m unities. The dom inant species is R e s tio p u r p u r a s ­ c e n s , but E le g ia r a c e m o s a and E. th y r s ife r a are also abundant. In the low er stratum . A n th o c h o r tu s c rin a lis is conspicuous. Som e species such as H ip p ia p ilo s a and S e n e c io c r is p u s are shared with C om m unity G roup A. This type can best be described as a R estioland. because small and big restio species are its m ost im portant struc­ tural constituents. It is difficult to determ ine the differen­ tial species in a situation where relevés are few. but A riste a b a k e r i, C liffo r tia o v a Iis, E le g ia r a c e m o s a . H ip p ia p ilo s a and an unidentified species o f A steraceae. might serve as possible candidates. E ponym ous E r ic a a u tu m n a lis is endem ic to the H H M . The com m unity occurs in the highest reaches o f the W esselsgat River, w here riverbanks are steep and rocky. C o m m u n ity B 3: G r u b b ia r o s m a r in ifo lia - R e s tio a ff. v e rsa tilis M e d iu m C lo s e d S h r u b la n d (Table 2. relevés 2 4 -2 6 ) This is the other type o f riparian m ire found along high altitude stream s. The main difference from the pre­ vious type is the shape o f the banks, w hich are tlatter and less rocky in this vegetation. The dom inant sm all restio here is R e s tio aff. v e r s a tilis . com pared with A n th o ­ c h o r tu s c r in a lis in com m unity B2. T his com m unity has a tall herb stratum (reaching 1.0-1.5 m ), dom inated by the shrubs B e r z e lia sq u a rro sa , B r u n ia a lo p e c u r o id e s and G r u b b ia ro s m a r in ifo lia and restio id s R e s tio p u r p u r a s c e n s and C h o n d r o p e ta lu m m u c ro n a tu m . The m ost closely related riparian co m m u ­ nity is the E r ic o -T e tr a r ie tu m c ra s s a e (S ieben 2003), w hich shares m any E ric a species with the G r u b b ia ro s­ m a r in ifo lia -R e s tio aff. v e rsa tilis C losed Shrubland. The most closely related seepage com m unity is the T etra rio c a p illa c e a e -R e s tie tu m s u b tilis . A species that is shared with this com m unity is R e s tio aff. v e rsa tilis. w hich is the dom inating elem ent o f the ground layer. The G r u b b ia r o s m a r in ifo lia -R e s tio aff. v e rsa tilis C om m unity is ty p i­ cal o f situations w here river banks are not steep and there is a lot o f lateral seepage. It can be described as a shrub­ land because o f the high cover o f shrubs o f Ericaceae and B runiaceae. There are m any differential species, m ost o f w hich are shared w ith eric a ce o u s fy n b o s and the E r ic o - T e tr a r ie tu m c r a s s a e in particular. T hese are. C o m m u n ity B 4: T e tra ria c a p illa c e a - R e s tio S h o r t to T all C lo s e d R e s tio la n d s u b tilis (Table 2, relevés 2 7 -3 5 ) S c ie n tific n a m e : T etra rio c a p illa c e a e - R e s tie tu m s u b ­ tilis a ss. n o v a h o c lo co H olotypus: Table 2. relevé 27 T his is the m ost com m on type o f seepage com m unity in the area, w hich can be characterized by the absence o f S e n e c io cr isp u s. The dom inant restio is R e s tio s u b tilis , w ith A n th o c h o r tu s c r in a lis and R e s tio aff. v e rs a tilis as co-dom inants. A typical characteristic is the m osaic form ed by patches o f low vegetation o f sm all restios and sedges (R e s tio s u b tilis, R . aff. v e rsa tilis. T etra ria ca p illa c e a and E p is c h o e n u s v illo su s) and patches o f tall v eg­ etation consisting only o f C h o n d r o p e ta lu m m u c ro n a tu m . T his species does not resprout after fires, like m any other seepage species, but regenerates from seed. It tends to dom inate the com m unity, because the old plants form a thick litter layer on the soil beneath it. w hich seem s to prohibit other (aggressively spreading) clonal species from grow ing there. D ifferential species o f this co m m unity are few . because m ost species are shared with the G r u b b ia ro s ­ m a r in ifo lia - R e s tio aff. v e r s a tilis C losed S h rubland. D iagnostic features are m ostly the dom inance o f R e s tio s u b tilis and the occurrence o f C h r y s ith r ix species. As in the case o f the riparian seepage types, this com m unity contains num erous shrub species, such as the differential species G ru b b ia r o s m a rin ifo lia and B e r z e lia sq u a r r o s a . but they do not grow very tall. T his com m unity occurs on m ore m inerotrophic soils than the form er com m unities. N evertheless, the soils are very acidic and highly organic. In one case it was found in a riparian zone and it is closely related to the riparian seep ag e types d esc rib ed ab o v e. T he co m m u n ity described by B oucher (1978) as C h o n d r o p e ta lu m -R e s tio T ussock M arsh seem s to be quite sim ilar, but the d o m i­ nant sm all restio in the K ogelberg is not R e s tio s u b tilis but R a m b ig u u s and m any other species are absent in the K ogelberg com m unity. (G rad ien t a n a ly s e s The m ost im portant environm ental factors that com e out o f the C CA (F igure 2) are slope and altitude. T his is m ainly due to the outlier com m unities o f A l and A 2. w hich are located at a higher altitude and on steeper slopes than any other o f the m ire com m unities. They also have, together w ith the riparian mire com m unities B2 and B3. the highest values for rockiness. It is interesting to see that there is a sharp contrast in the fraction o f soil particle sizes: the fraction o f coarse sand is an im portant environm ental variable and the com m unities with a high fraction o f coarse sand are the riparian m ires (B 2 and B3) 150 B othalia 34.2 (2004) FIG U R E 2 ,— B iplot o f constrained o rd in atio n (C C A ) o f m ire vegetation o f H ottentots H ol­ land M o u n ta in s, fe atu rin g A xes 1 and 2 with position o f com m unity sam ples and se­ lec ted e n v iro n m en ta l v a ri­ ables. C om m . AI and C om m . A 2 ,0 ;C o m m .A 3 ,+ ; C om m . A 4, □ ; C o m m . A 5, ▼ : C om m . B2 and C om m . B3, X; C om m . B3, A; C om m . B4. o. and re stio m a rsh lan d s (B 4) w hich are the m ost m inerotrophic m ires in the m ire system . On the other hand, there are the com m unities w hich have a high frac­ tion o f fine sand and silt, w hich represent the fens in the m iddle o f the m ire system w here peat form ation occurs (especially A 3, but also A4 and A 5). The axis that is form ed by the variable o f coarse sand, fine sand and silt reflects a gradient in m inerotrophy or in w aterlogging. T he fens (A 3, A 4 and A 5) have the highest values for organic m atter contents and soil depth. T hey are also m ainly found at the higher altitudes because this is w here the highest rainfall occurs. D IS C U SS IO N O ne o f the m ost im portant questions that is raised from the results o f this study is w here the high-altitude m ires o f the Fynbos Biom e fit into the w orld-w ide typol­ ogy o f fens and m ires. A lthough the m ires regularly form the sources o f the rivers, they are clearly very different from the European spring ecosystem s (Z echm eister & M ucina 1994). Sw am ps and bogs w ith a high gram inoid cover are found extensively in the boreal zone o f the northern hem isphere (Sjors 1983) and the vegetation co v er o f sw am ps and bogs in A frica is also m ostly d o m i­ nated by gram inoids (T hom pson & H am ilton 1983). G ore (1983) distinguishes betw een om brotrophic and m inerotrophic m ires, based on the origin o f the water. In om brotrophic seepage, a thick layer o f peat has d evel­ oped and there is no m ore contact with the m ineral su b ­ strate. The w ater originates exclusively from rain, which results in very oligotrophic conditions. The w ater from m inerotrophic m ires seeps through the m ineral substrate into the m ire, so it is richer in nutrients than the o m ­ brotrophic m ire. O m brotrophic m ires can only exist in very hum id clim ates such as the blanket bogs o f the British Isles and the elevated bogs o f northern Europe; they are quite rare in the southern hem isphere. A ctually, the distinction betw een om brotrophic and m inerotrophic m ires is m ore like a gradient, an idea expressed by Sjors (1983). O m brotrophic m ires are on the one extrem e o f this gradient and all m ires that do not feed exclusively on rainw ater m ake up the rest o f this gradient. Sjors (1983) also gives a m ore detailed subdivision o f European mires: topogenous m ires (influenced by stag­ nant w ater), soligenous m ires (influenced by seepage), lim nogenous m ires (influenced by floodw aters) and om brogenous m ires (influenced by rainw ater). Solige­ nous and lim nogenous m ires are both associated with rivers. Soligenous m ires form around springs and lim ­ nogenous m ires occur in the floodplains along the low er reaches o f rivers. The m ires described in this study are all o f the so lig en o u s type. The different com m unities described in this study are situated along a gradient from dry to m oist. In the F ic in io a r g y r o p a e -E p is c h o e n e tu m v illo si A ssociation, occurring in the centre o f the m ire, w ater stagnates m ore because the drainage is slow. On the edges, the T etra rio c a p illa c e a e -R e s tie tu m s u b tilis A ssociation, w hich has a faster drainage, will prevail. Further tow ards the m argins, com m unities dom inated by C h o n d r o p e ta lu m d e u s tu m can occur, but these w ere not recorded during this study. Two com m unities can occur tow ards the centre o f very wet m ires, nam ely the A n th o c h o r to c r in a lis -E le g ie tu m in te r m e d ia e A ssociation o r the I s o le p is p r o life r - B u lb in e lla n u ta n s Tall C losed Sedgeland described from the Du T oitskloof M ountains. Both c o m m u n ities are ex trem ely p o o r in sp ecie s, because of the specific stresses that occur under w ater­ logged conditions. It is clear that this gradient, from B othalia M 2 (2004) 151 w ell-drained to poorly drained or from the edge to the centre o f the m ire, is also very prom inent in the o rdina­ tion diagram s. This gradient was also found by B ragazza & Gerdol (1999) in som e m ires in the southeastern A lps. The other distinguishing feature that show s clearly in the ordination diagram s is the im portance o f the substrate, as can be seen from the P ro te a n m n d ii-H ip p ia p ilo s a Tall Shrubland from the shale band. A conspicuous thing about the vegetation o f m ires o f the Fynbos Biom e is that they are dom inated by the clo n ­ al restios A . c r in a lis , P. d e p a u p e r a tu s and R. su b tilis (L inder 1985). Because these species tend to cover everything, the vegetation is relatively poor in species. Clonal reproduction is often coupled to environm ental plasticity, so the species can tolerate slight differences in the environm ent. The diversity o f m icrosites is much bigger than the species richness suggests (Price & M arshall 1999). The tall restio C. m u c ro n a tu m only regenerates from seed after fires and usually occurs in large, dense m onotypic stands when m ature. It does not support much vegetation underneath it. The dead m ater­ ial from previous generations can form dense accum ula­ tions o f debris and this creates a very unfavourable sub­ strate for other species. In m arshes elsew here in the w orld, clonal sedges and grasses take the place o f the clonal R estionaceae record­ ed in this study. It is generally accepted that the clonal grow th form is an adaptation to the stress o f w aterlog­ ging. This is confirm ed by the investigations by Soukupová (1994) o f three clonal gram inoids. A fter w aterlog­ ging there is an increase in clonal m odules. Specht (1981) review s m any o f the problem s that sclerophyllous plants have to overcom e in seasonally w aterlogged areas. It has becom e clear from this study that the mires in South Africa are very different from those in the northern hem isphere. Although they are vulnerable to predicted cli­ mate change (Rutherford et al. 1999). there is very little know ledge about the fens and m ires o f the southern hem i­ sphere. In order to be able to make general statem ents about mire ecosystem s, more attention should be paid to the mire ecosystem s in countries like South Africa. REFEREN CES B A R K M A N . J.J.. D O IN G . H & S E G A L . S 1964. K ritische Bem erkungen und Vorschláge /u r quantitativen V egetallonsanalysc. A c ta B o ta n ic a N e e r la n d ic a 13: 394 4 1 9 B O S C H . J.M .. A L L E T S O N , D J .. JA C O T G U IL L A R M O D . A .F.M .G .. K IN G . J M & M (X )R E . C .A . I9K6. R iver response to c atch ­ m ent conditions. In J.H . O 'K e e ffe . T h e c o n s e r v a tio n o f S o u th A fr ic a n r iv e r s. South A frican N ational Scientific P rogram m es Report No. 131. B O U C H E R ,C . 1978. C ape H angklip area II The vegetation. B o th a lia 12: 4 5 5 -4 9 7 . B O U C H E R , C . 1988. W etland vegetation o f the Fynbos B iom e. In J M K ing. H y d r o lo g y a n d h y d r o b io lo g y in th e F y n b o s B io m e 3 4 -4 0 . E cosystem s P rogram m es O ccasional R eport N o. 26. C S IR . Pretoria. B R A G A Z Z A . L. & G F R IX )L . R 1999. H ydrology, groundw ater chem istry and peat chem istry in relation to habitat c o nditions in a m ire on the southeastern A lps o f Italy. P la n t E c o lo g y 144: 24 3 -2 5 6 . C A M P B E L L . B.M 1986. A classification o f the m ountain vegetation o! the Fynbos Biom e M e m o ir s o f th e B o ta n n a l S u r v e y o f S o u th A fr ic a N o. 50: 1 - 1 2 1. D A L L A S . H .F .& D A Y .J.A . 1993. T h e e ffe c t o f w a te r q u a lity v a r ia b le s o n r iv e r in e e c o s y s te m s : a review . W ater R esearch C om m ission Report T T 61/93. Pretoria. D E A C O N . H .J.. JURY. M .R . & E L L IS. F. 1992. Selective regim e and tim e. In R .M . C ow ling. T h e e c o lo g y o f f y n b o s ; n u tr ie n ts , f ir e a n d d iv e r s ity : 6 -2 2 . O xford U niversity Press. C ape Tow n. D E N T . M .C .. LY N C H . S.D . & S C H U L Z E . R E. 1987. M a p p in g m e a n a n n u a l p r e c ip ita tio n a n d o th e r r a in fa ll s ta tis tic s o v e r s o u th e r n A fr ic a . A gricultural C atchm ents R esearch Unit R eport No. 27. W ater R esearch C om m ission R eport N o. 109/1/89. DE V IL L IE R S . J.. JA N S E N . H. & M U L D E R . M .P. 1964. D ie g e o lo g ie v a n d ie g e b ie d tu s s e n W o rc e s te r e n H e r m a n u s . G eological Survey, Pretoria. ED W A R D S. D. 1983. A broad-scale structural classification o f v eg eta­ tion for practical purposes. B o th a lia 14: 7 0 5 -7 1 2 . FRY. M. ST. L. 1987. A d e ta ile d c h a r a c te r iz a tio n o f s o ils u n d e r d iffe r ­ e n t fy n b o s - c lim a te - g e o lo g y c o m b in a tio n s in th e s o u th w e s te r n C a p e . M .Sc. thesis, U niversity o f Stellenbosch. F U G G L E , R.F. 1981. M a c ro -c lim a tic p a tte r n s w ith in th e F y n b o s B io m e F in a l R e p o r t. N ational Program m e for Environm ental Sciences, Fynbos B iom e P roject. C ape Tow n. F U G G L E . R.F. & A S H T O N . E.R . 1979. C lim ate. In J. Day, W .R. Siegfried. G .N . Louw & M .L. Jarm an. F y n b o s e c o lo g y : a p r e ­ lim in a r y s y n th e s is . South A frican National Scientific P rogram ­ m es R eport N o. 40: 7 -1 5 . G E R M IS H U IZ E N . G. & M E Y E R . N .L. (eds). 2003. Plants o f southern Africa: an annotated checklist. S tr e litz ia 14. N ational B otanical Institute. Pretoria. G L Y P H IS . J., M O L L . E J . & C A M P B E L L . B .M . 1978. Phytosociological studies on Table M ountain. South A frica. 1. The Back Table. J o u r n a l o f S o u th A fr ic a n B o ta n y 44: 2 8 1 -2 8 9 . G O R E . A J.P . (ed.). 1983. E c o s y s te m s o f th e W o r ld 4 M ire s: sw a m p , b o g , f e n a n d m o o r. A . G e n e r a l stu d ie s . E lsevier. A m sterdam . H E N N EK EN S. S.M . 1996a. M E G A T A B — a v is u a l e d ito r f o r p h y to s o c io lo g ic a l ta b les. Version 1.0. U ser's G uide. G iesen & G eurts. Ulft. H E N N E K E N S . S .M . 1996b. T V R B O V E G : s o ftw a r e p a c k a g e f o r in p u t, p r o c e s s in g , a n d p r e s e n ta tio n o f p h y to s o c io lo g ic a l d a ta . Version July 19% . U se r's guide. IB N -D LO & L ancaster U niversity, W ageningen & L ancaster. H E N N E K E N S . S.M . & SC H A M IN É E . J .H J . 2001 T U R B O V E G . a com prehensive data base m anagem ent system for vegetation data. J o u r n a l o f V e g e ta tio n S c ie n c e 12: 5 8 9 -5 9 1 . H E W LE TT . J .D. 1961. S o il m o is tu r e a s a s o u r c e o f b a s e flo w fr o m s te e p m o u n ta in w a te r s h e d s . S outheastern E xperim ental Station. U SD A Forest Service Paper 132. H E W L E T T . J.D . 1982. P r in c ip le s o f f o r e s t h y d ro lo g y. U niversity o f G eorgia P ress. A thens. G eorgia. H IL L . M .O . 1974. C orrespondence analysis: a neglected m ultivariate m ethod. A p p lie d S ta tis tic s 23: 3 4 0 -3 5 4 . H IL L . M .O. 1979. T W IN S P A S — a F O R T R A N p r o g r a m f o r a r r a n g in g m u ltiv a r ia te d a ta in a n o r d e r e d tw o w a y ta b le b y c la s s ific a tio n o f in d iv id u a ls a n d a ttr ib u te s . C ornell U niversity. Ithaca. New York. JO N G M A N . R .H .G .. TER B R A A K . C J F & VAN T O N G E R E N . O .F.R . 1987. D a ta a n a ly s is in c o m m u n ity a n d la n d s c a p e e c o lo ­ gy. P udoc. W ageningen. K F R F O O T . O . 1968. M ist p re cipitation on vegetation. F o re stry A b s tr a c ts 29: 8 -2 0 . K IN G . J.M & DA Y. J.A . 1979. H ydrologv and hydrobiolo gy. In J. Dav. W .R. S iegfried. G .N . Louw & M L . Jarm an. F y n lw s e c o l­ o gy a p r e lim in a r y sy n th e s is . South A frican N ational S cientific P rogram m es R eport No. 40: 2 7 -4 2 . K O P P E N . W. 1931. G r u n d r ifi d e r K lim a k u n d e . De G ruyter. Berlin. K R U G E R . F J 1978. South A frican heathlands. In R.L. S pecht. E c o s y s te m s o f th e W o rld 9. H e a th la n d s a n d r e la te d sh r u b la n d s B A n a ly tic a l stu d ies : 1 9 -8 0 . E lsevier, A m sterdam . L A ID L E R . D . M O LL . E J .C A M P B E L L . B.M & G L Y P H IS. J 1978. Phytosociological studies on Table M ountain. South A frica. 2. The Front T M e J o u r n a l o f S o u th A fr ic a n B o ta n v 44: 2 9 1 -2 9 5 . L IN D E R . H P 1985. C onspectus o f the A frican species o f R estionaceae B o th a lia 15: 3 8 7 -5 0 3 . L IN D E R . H P 1987. The evolutionary history o f the P oales/R estion a le s—a hypothesis. K ew B u lle tin : 42: 2 9 7 -3 1 8 . M A R L O T H . R 1903 R esults o f experim ents on Table M ountain for ascertaining the am ount o f m oisture deposited from the south­ east clouds. T r a n s a c tio n s o f th e R o y a l P h ilo s o p h ic a l S o i iety o f S o u th A f r u a 14: 4 0 3 —W>K. M U C IN A . L. 1993. N o m e n k la to risch e und sy n ta x o n o m ish e D efim tionen. K o n /e p te und M ethoden In L. M ucina. G 152 B othalia 34,2 (2004) G rabherr & T. E llm auer, D ie P fla n z e n g e s e lls c h a fte n Ó s te rr e ich s . T eil I. A n th r o p o g e n e V eg e ta tio n : 1 9 -28. G ustav Fischer, Jena. M U C IN A . L ., B R E D E N K A M P, G J . , H O A R E , D .B . & M C D O N A L D , D J . 2000. A national vegetation database for South A frica. S o u th A fr ic a n J o u r n a l o f S c ie n c e 96: 4 9 7 ,4 9 8 . N O N -A F FIL IA T E D SO IL A N A LY SIS W O R K C O M M IT T E E . 1990. H a n d b o o k o f s ta n d a r d s o il te s tin g m e th o d s f o r a d v is o r y p u r ­ p o s e s . Soil S cience Society o f South A frica. Pretoria. 0 K L A N D , R .H . 1996. A re ordination and constrained ordination alter­ native o r com plem entary strategies in general studies? J o u r n a l o f V eg e ta tio n S c ie n c e 7: 2 8 9 -2 9 2 . P R IC E . E .A .C . & M A R S H A L L , C . 1999. C lonal plants and en v iro n ­ m ental heterogeneity. P la n t E c o lo g y 141: 3 -7 . R O G E R S , K .H . 1995. R iparian w etlands. In G .I. C ow an. W e tla n d s o f S o u th A fr ic a : 4 1 -5 2 . D epartm ent o f E nvironm ental A ffairs and T ourism , Pretoria. R O G E R S , K .H . 1997. F reshw ater w etlands. In R .M . C ow ling, D M. R ichardson & S.M . P ierce. V eg e ta tio n o f s o u th e r n A fric a : 3 2 2 -3 4 7 . C am bridge U niversity P ress, C am bridge. R U T H E R F O R D , M .C .. PO W R IE . L.W . & S C H U L Z E , R E. 1999. C lim ate change in co nservation areas o f South A frica and its potential im pact on floristic com position: a first assessm ent. D iv e r s ity a n d D is tr ib u tio n 5: 2 5 3 -2 6 2 . S C H U L Z E . B .R . 1965. C lim a te o f S o u th A fr ic a . P a r t 8 . G e n e r a l s u r ­ vey. G overnm ent P rinter, Pretoria. * S IE B E N , E J J . 2003. T h e r ip a r ia n v e g e ta tio n o f th e H o tte n to ts H o lla n d M o u n ta in s , W estern C a p e , S o u th A fr ic a . Ph.D . thesis. U niversity o f S tellenbosch. S JO R S , H. 1983. M ires o f S w eden. In A J .P G ore, E c o s y s te m s o f th e W o rld 4 . M ire s: s w a m p , b o g , f e n a n d m o o r. B R e g io n a l stu d ies: 6 9 -9 4 . E lsevier. A m sterdam . SO IL C L A SS IFIC A T IO N W O R K IN G G R O U P 1991. G r o n d k la ss ifik a s ie . E e n ta k s o n o m ie s e s is te e m v ir S u id -A fr ik a . D epartm ent o f A griculture A ffairs. Pretoria. A P PE N D IX 1. S O U K U P O V Á , L. 1994. A llocation plasticity and m odular structure in clo n a l g ra m in o id s in re sp o n se to w a terlo g g in g . F o lia G e o b o ta n ic a & P h y to ta x o n o m ic a 29: 2 2 7 -2 3 6 . SPEC H T, R.L. (ed.). 1981. E c o s y s te m s o f th e W o rld 9. H e a th la n d s a n d r e la te d sh r u b la n d s . B . A n a ly tic a l s tu d ie s . Elsevier, A m sterdam . TA Y LO R , H.C . 1978. C apensis. In M J .A . W erger. B io g e o g r a p h v a n d e c o lo g y o f s o u th e r n A fr ic a : 171-229. Junk, T he H ague. TER B R A A K . C.J.F. 1986. C anonical C orrespondence A nalysis: a new eigenvector technique for m ultivariate direct gradient analysis. E c o lo g y 67: 1167-1179. TER B R A A K , C J.F . & S M ILA U E R . P. 1998. C A N O C O R e fe r e n c e M a n u a l a n d U s e r 's G u id e to C o n o c o f o r W in d o w s. S o ftw a r e f o r C a n o n ic a l C o m m u n ity O r d in a tio n (v e r s io n 4 ). M icrocom puter Pow er , Ithaca, N ew York. T H O M P S O N . K. & H A M IL TO N . A C . 1983. Peatlands and sw am ps o f the A frican continent. In A J.P . G ore, E c o s y s te m s o f th e W o rld 4 . M ir e s: s w a m p , b o g , f e n a n d m o or. B . R e g io n a l stu d ies : 3 3 1 -3 7 3 . E lsevier. A m sterdam . VAN Z IN D E R E N B A K K E R , E M & W E R G E R , M .J.A . 1974. E n­ vironm ent, vegetation and phytogeography o f the high-altitude bogs o f L esotho. V eg e ta tio 29: 37^49. W ALKLEY, A. 1935. An exam ination o f m ethods for determ ining organic carbon and nitrogen in soils. J o u r n a l o f A g r ic u ltu r a l S c ie n c e 25: 5 9 8 -6 0 9 . W E B E R , H E., TH EU R ILLA T. J.-P. & M O R A V EC . J. 2000. Inter­ national Code o f Phytosociological N om enclature, edn 3. J o u r n a l o f V e g e ta tio n S c ie n c e I I: 7 3 9 -7 6 8 . W E IS SE R , P. J. & HO W ARD -W ILL1 A M S, C. 1982. The vegetation o f the W ilderness Lakes system and the m acrophyte encroachm ent problem . B o n te b o k 2: 19-40. W ES TH O FF , V. & VAN DER M A A R E L . E. 1978. The BraunB lanquet approach. In R.H . W hittaker. C la s s ific a tio n o f p la n t c o m m u n itie s: 2 8 7 -3 9 9 . Ju n k . The H ague. Z E C H M E IS T E R . H. & M U C IN A . L. 1994. Vegetation o f European sp rin g s: h ig h -ran k sy n ta x a o f the M o n tio -C ard a m in e te a . J o u r n a l o f V e g e ta tio n S c ie n c e 5: 3 8 5 -3 9 9 . List oi less frequent species having one or tw o occurrences in the relevé table. Sequence: nam e o f species, the field code o f the relevé(s) and cover-abundance (in brackets) A g a th o s m a p e n ta c h o to m a 262 (2a) A s k id io s p e r m a c h a r ta c e u m 262 (2a) A s k id io s p e r m a e s te r h u y s e n ia e 164 (2b), 202 (+) B e r z e lia la n u g in o s a 262 (+) B le c h n u m ta b u la re 131 (+) B o b a r tia g la d ia ta 259 (+) C h iro n ia d e c u m b e n s 262 (r) C l i f f o r tia g r a m in e a 165 (1) C liffo r tia r u s c ifo lia 233 (2a) C o r y m b iu m c o n g e s tu m 131 (1) C o r y m b iu m c y m o s u m 131 (+) D ic r a n o lo m a b illa r d ie r i 123 (2a) D is a tr ip e ta lo id e s 123 (2m ), 127 (2a) E d m o n d ia p in ifo lia 131 (+) E h r h a r ta r a m o s a 259 (1) K o g e lb e rg ia v e rtic illa ta 127 (+) L o b e lia ja s io n o id e s 134 (+) L y c o p o d ie lla c a ro lin ia n a 123 (+) O s m ito p s is a fra 241 (+) O x a lis n id u la n s 202 (+) P e n ta m e r is th u a r ii 233 (2b) P e n ta s c h is tis p a llid a 135 (+) P ro te a c y n a r o id e s 128 (+), 233 (r) P so r a le a a c u le a ta 241 (2a) R a s p a lia v irg a ta 136 (2a) R e s tio b ifa r iu s 131 (+) R e stio in te r m e d in s 128 (3) R e s tio o b s c u r u s 262 (1) R e s tio p e d ic e lla tu s 259 (+), 262 (3) S c h iz a e a te n e lla 127 (1) E h r h a r ta s e ta c e a subsp. u n iflo r a 111 (2m ) E p is c h o e n u s c o m p la n a tu s 131 (1), 201 (+) E p is c h o e n u s g r a c ilis 262 (2a) E r ic a lo n g ifo lia 262 (r) E u c h a e tis g la b r a 201 (1) S e n e c io c o le o p h y llu s 134 (2a) S e n e c io p u b ig e r u s 233 (1) S e n e c io r ig id u s 131 (+) E u r y o p s a b r o ta n ifo liu s 232 (2m ), 233 (1) F e lic ia c y m b a la r ia e 111 (1) F ic in ia cf. in v o lu ta 259 (3) F ic in ia sp. 233 (1) F is s id e n s p lu m o s u s 111 (2m ) G e is s o r h iz a u m b ro sa 131 (+) H e lic h r y s u m c y m o s u m 233 (2a) H y m e n o p h y llu m p e lta tu m 131 (2m ) H y p o c h a e r is r a d ic a ta 259 (+) Is c h y r o le p is tr iflo r a 262 (1) I s o le p is d ig ita ta 135 (2m ) S o n c h u s o le r a c e u s 259 (+) S ta b e r o h a c e r n u a 234 (1) S ta b e r o h a va g in a ta 124 (+), 189 (+) S e r ip h iu m p lu m o s u m (juvenile) 188 (r) T e tra ria p illa n s ii 201 (1), 259 (2a) T e tra ria th e r m a l is 124 (r) T o d ea b a r b a ra 111 (1), 259 (2a) U rsin ia d e n ta ta 233 (1), 241 (+) U tric u la r ia b is q u a m a ta 163 ( + ) V illa rsia c a p e n s is 163 (1), 164 (+) W a h len b e rg ia p ro c u m b e n s 201 (D .2 4 1 (2b) W a tso n ia b o r b o n ic a 241 (r). APPENDIX 2 Selected data on sampled vegetation properties and geographical location of two relevés in Du Toitskloof Mtns and relevés in Hottentots Holland Mms. No., releve number; Com., community; FC, field code; As., aspect; SI., slope in degrees; Area, sampled area (m~); CE2. cover of shrub layer (%); C E 1, cover of herb layer (7r); CB). cover of moss layer (%); HE2, height of shrub layer (m); HE 1, average height of herb layer (m); SR. species richness B othalia 34,2 (2004) r- 153 rN — (N fN ITi IT) IT, (N (N (N fN (N O i a, i a, p p — a O O O O' o o p O' O' O' O' O' O' 8 OJ 8 C8\1 O' O' O' O' O' O' O' O O O' O o O — — oo rsl ——O' O' o = = 0C O' O DC O' O' O O' O — — CC<N<N<N-ÍN (N IT, (N Tf O' 0C •Ti »/“) p — (N NC ^>C rf 53 E t> E W u c UJ O ï; " E E -c y u E t; U w E E oc. u. eii « to cc +J í'J cc X . O O cc Ú Ú 3 .8 eii £ !3 rt E Cn Cr. £ C * £ £ S £ 2 £ •O -3 "3 T3 £ 2 tX j 0 U O r' r: os cL E u u cí. c. a g t I . os a: oE « * 8c . 8c . 8. 2 "O c -0 O' o^ O' O' 0s O' , it — x O' i- <r, — <C >C v~. < *-. < *-; rp, r*"i f t rr C u~. vC ir, X 5 M N - S S Í Í Í “ £ £ £ C- C- CL CL C* CJj CJj CJj ^ Í I * c e c c c c c C C C ~ "3 -a t3 ^ O C > ^ ^ ai c oo 2 g- y s- ■o j a / c 2 O' O' O' r: í /. s3 s E O' O O' O ' O ' O ' O ' O ' O ' O ' O ' O ' O ' O ' U w ts — £ £ úi / C ir , O — — ac O' O' o E C/5 C/5 Ui O' — rj C ~ T c . J= C 'j W 0> CC rsS OS rsS0£ O' H E ■s -s; U U U UJ d. *—• rr* E E O i/~, ly, ir; c o c o c o c O ' O ' O f- vd O O' Q £ ff sc sc rin »r> i/", «/“> ir; *r> ‘Z’', 04 »ri O' O' O' O' o— — c «Tï l/”, i/% »0 m o o c & O' — — — — r<-, 888 o c o c o o o o O ' O ' O ' O ' O ' ^ K. O' — — — <N l/~, — ir> C C O C O C * T i C C ^ - 25 — fN r- i Ti g g e c g c ^ c a . a . o . c . c . 13c c ” -u > a£ I * c = /. c u u u u u a j u o o1 w4> i j s j o p ; c x v. v. i yr , T — C C C C C C C C C C C C C C C C C ± <N — f ^ * r ;i^ D C '^ tO ' — O' t <N t t T t <N (N — — <*“, — t t t t j S o o •o c io «/% c x o ^ (N ^ r j T r*~. T rTr . rTs v~. r) fN IO <N © T rr, r*“, n c o 5 í o S o' 5 » 8 » 5 5 C O r . i ^ > C l Z Z. ac X' x g y s S s S c ! ^ . ^ i r. r ^ i v) ~ ' ir C © * 1 0 , i /i 0 vi C < U N UJ O C O C U ■ — <n r~ * O' W > » » — f U r r , * t O ' X — <N r r , V", >C >r . — ' t ' T r ' O C ' C i n n r l rN N n N Nn Nr , Nr , Nt n4 i i r,N « N c - Cf SC S•' r S - C - í'-. -f' , - n - N- -n < , r ^ r » x x 3C X ~ t O '- Sc X M C tu , n , 3. S c 1 *r 't rr- < < < < < < < < < < < < < < < < E® B B B f f l aB SSSaa rr, -T \ r, v C f ' - — — 1— ' •— —■ — — — '— — N N N N I S N N N N N r , O C O ' C — *Nr r , > r > r i >C r ^ 9 C O C — «N rr, " T V , > c r ^ a e o - c IT, — <N r , *t ir,