South African JOllmoi of Botany 2001 8 7 58· 64 Copynghr €) NISC Ply Lid Pnn /ed III Soutll Afllca - Ail ng/lls reselved SOUTH AfRICAN JOURNAL OF BOTANY ISSN 0254-8299

Vegetation ecology of the southern Free State: Vegetation of the drainage channels

PW Malan" , HJT Venter' and PJ du Preez'

, Department of Biology, University of North West, Private Bag X2046, Mmabatho 2735, , Department of Botany and Genetics, University of the Free State, PO Box 339, B/oemfontein 9300, South Africa * Corresponding author. e-mail: pwmsct@unw001 .uniwest.ac.za

Received 1 February 2000, accepted in revised form 26 September 2000

Little information is available on the vegetation of the dam sample plots, A TWINS PAN classification, refined southern Free State, A phytosociological analysis of the by Braun-Blanquet Procedures, resulted in 9 distinct riparian shrub communities of the southern Free State vegetation units grouped as two major communities. is presented. Releves were compiled in 38 stratified ran-

Introduction

The necessity of detailed ecological studies as a basis ed with low-lying drainage channels and other re levant bot­ for sound land-use planning, management and research is tom land habitats, This study also fits In with a comprehen­ widely recognised (Pentz 1938, Codd 1949, Bayer 1970, sive phytosociolog ical programme under the Grassland Foran et ai, 1986, Bosch el al. 1987), One of the major proj­ Biome Project (MentiS and Huntley 1982, Scheepers 1987). ects in this respect is the Grassland Biome Project. The Grassland Biome of South Africa covers approximately 27% Study area of Ihe country (Rutherford and Westfall 1994), As a result of in tensive agricu ltural practices and urbanisation, together The present study includes the southern Free State and is with industrialisation, the deterioration of the grassland led situated to the south of the 29°00'5 latitude and to the west to concern amongst decision-makers, resulting in the launch of the 27°00' E longitude, encompassing approximately of the Grassland Biome Project (Mentis and Huntley 1982), 27 bOOkm'. One of the first priorities of this project IS to determine the Two biomes can be distinguished in the study area, name­ location and extent of the major vegetation types within the ly the Grassland and Nama- Biomes (Rutherford and Biome (Mentis and Huntley 1982). Description of the plant Westfall 1994). According to Acocks (1988) the vegetation of communities of the Zastron Area (Malan el ai, 1999), those these biomes in the southern Free State is divided into six associated with the rocky outcrops of the dry south-western different veld types. Towns situated in the study area are Free State (Malan el al. 1998) and Acacia communities (from north to south) Bloemfontein, Petrusburg, Fauresmith, (Malan el ai, 2000, submitted ), all form a broad spectrum of Wepener, Zastron and Bethulie (Figure 1). According to Low the vegetation of the southern Free State. and Rebelo (1996). the grasslands of the Grassland Biome According to Du Preez (1991) the Acacielea karoo in the are dominated by a single layer of grasses. Trees are southern and eastern Free State represents the thickets absent, except in a few localised habitats (Low and Rebelo usually situated on well-developed levees along rivers, 1996). The Nama-Karoo mainly consists of grassy, dwarf streams and drainage lines, and is also present on clayey shrubland (Low and Rebelo 1996), soils on the low level terraces and flood plains adjacent to The ra infall is erratic, especially in the western part of the rivers, Vegetation of this class may also be found on deep study area from where it increases in an easterly direction alluvial or colluvial soils on gradual soils on gradual foot­ from a 300-400mm/annum rainfall interval to a 600-800mml slopes of hills and ridges, usually situated near drainage annum rainfall interval. Slow-draining streams are common lines and rivers (Bredenkamp et a/. 1989, Bredenkamp and owing to the flatness of the terrain. Bezuidenhout 1990). The Acacia karoo-Rhigozum tricholo­ The habitat is fairly unstable, due to seasonal flooding and mum Community associated with higher-lying areas is dis­ drying that, together with the frequent overgrazing of the cussed elsewhere (Malan el ai, 2000, submitted ). area, have caused an advanced state of degradation of the The aim of this study was therefore to claSS ify and vegetation. The soil is often severely eroded and the vege­ describe the shrub communities in the southern Free State. tation obviously disturbed. A detailed descnption of the study This report primarily deals with shrub communities associ at- South Arrican Journal or BOlany 2001 67- 58·64 59

Republic of South Africa

Petrus burg .. • Bloemfontein ..N 600 mm Rainfall isohyct

• w.\ ener " Fauresmrth 30 0 S \

500 km

o so 100 km

Figure 1: The location of the stu dy area in rela tion to towns area and environmental attributes is given elsewhere (M alan and Rhus lancea res pectively dominate these major com­ et al. 1998). muni ties . The communities are classified as follows:

Methods Acacia karroo-Protasparagus laricinus Major Community Releves were compiled in 38 stratified random sample plots. 1.1 Acacia karroa-Salix babylonica Commu nity Care was taken to avoid sampling of severely degraded 1.2 Ziziphus mucronata-Protasparagus laricinus vegetation and excessively disturbed areas. Community Plot size was Fixed at 1DOm 2 which is in accordance with 1.2. 1 Rhus pyroides-Lycium hirsutum Sub-community the plot size used by Scheepers (1975), Bredenkamp and 1.2.1.1 Setaria verticillafa -Chenopodium album Variant Theron (1978), Rossouw (1983), Van Wyk (1983), Muller 1.2.1.2 Crassula lanceo/ata-Acacia karroa Variant (1986) , Du Preez (1991) , Bezuidenhout (1993) and Fuls 1.2.1 .3 Rhus pyroides-Protasparagus laricinus Variant (1993). For every plant species present in the sample plot, a 1.2.2 Heteromorpha trifo/iata-Nidorel/a resedifolia cover·abundance value was estimated according to the Sub-Community Braun-Blanquet scale (Mueller-Dombois and Ellenberg 1974). 2 Rhus lancea-Rhus burchellii Major Community Two-way indicator species analysis (TWINSPAN ; Hill 2.1 Olea europaea-Rhus lancea Community 1979b) was applied to the floristic data set in order to derive 2.2 Rhus erosa-Rhus lancea Community a first approximation of the plant communities of the area . In 2.2 .1 Diospyros austro-africana-Rhus lancea order to determi ne vegetation gradients and the relationship Sub-community with environmental variables, Detrended Correspondence 2.2.2 Heteropogon contortus-Rhus /ancea Analysis (DECORANA, Hil l 1979a) was applied to the floris­ Sub-community tic data set. Additional data such as soil forms and the rock­ 2.3 crispa-Rhus lancea Community iness of the soil surface were also noted. Taxon and au thor names comply with those of Arnold and Description of the communities De Wet (1993). Acacia karroo·Protasparagu5 laricinus Results and Discussion Major Community

In the analysis of the vegetation, nine distinct vegetation This widespread shrub community is mainly associa ted with units were identified. Two major communities, grouped into the low-lying areas and undulating plains along public roads. five distinct plant communities were identified. Acacia karroa According to Carr (1976), Acacia karroo is more widely dis­ tributed than any other Acacia species and occurs in varying 60 Malan. Venter and du Preez cl imatic conditions in all provinces of the Republic of South poorly drained, with water vi si ble on the surface weeks afte r Africa, eastern Botswana, Swaziland and . Although good rains. The soil is drier than that of the Acacia karroo­ Acacia karmo is adaptable to a wide variety of cond itions Salix babylonice Community. Habitat disturbance due to and soil forms (C arr 1976) , Ihe location of this major-com­ co ntinuous overgrazing and trampling of livestock occurs. munity is restricted to areas where wi nd-blown sand and Species listed in specie s group F, with Ziziphus mucrona­ gravel eroding from higher-lyin g areas often cover the soil ta and Celtis africana the only exclusive diagnostic tree surface. species , differentiated this vegetation unit (Table 1). The vegetation is characterised by Acacia karma (species Protasparagus lancinus , P suaveolens and Lycium cineri­ group G, Table 1). According to Carr (1976), A. karroo was um (species group F) are the most abundant shrubs. The probably the first of the South Africa n acacias to have drawn reed, Cyperus longus, normally associated with wetlands the attention of early botanists. Other widely spread trees (Eckhardt et al. 1993). the forb Tagetes minuta and th e include Ziziphus mucronata (species group F), Olea climber Clematis braclliata, are the only other abundant and europaea, Diospyros Iycioides (species group H) and Rhus differentiating species (species group F, Table 1). Other con ­ lancea (species group M). Protasparagu s laricinus (species spicuo us species in clude the trees , Acacia karroa (species group F) is th e only abundant shrub (Table 1). The sedge group G), Olea europaea (species group H), Diospyros Cyperus longus (species group F) is also common. Iycioides (species group I) and Rhus lancea (species group Graminoids are scarce and are restricted to patchy occur­ Pl. Th e height of these trees vari es between 4-5metres. rences of Setaria verticillata (species group B), Eragrostis An average of 15 species per releve , was record ed . obtusa (species group F), Themeda triandra and Sporobolus Th is plant commun ity is divided into two distinct sub-com­ fimbriatus (species group M). munities (Table 1). An average of 13 species per releve was recorded for th is major comm unity. 1.2.1 Rhus pyroides-Lycium hirsutum Two distinct commun ities further characterise this major­ Sub-community community (Tabl e 1). Th is bottom land vegetation occurs on seasonally wet, well­ 1.1 Acacia karroo-Salix babylonica Community drained sandy soils of riverbeds. The habitat is fairly unsta­ ble due to seasonal flooding and drying. Overgrazing of pas­ This commun ity represents azonal wetland-like vegetation lure and trampling of liveslock frequently occurs. and is present in depressions or other bottomland situations. Specie s listed in species group D differentiate this sub­ These bottomland situations generally have soils with a commun ity. Lycium hirsutum and Rhus pyroides (species higher (>50%) cl ay content than those of upland areas group 0 ) are the only abundant and differentiating shrubs (Land Type Survey Staff, in press). No rocks or large stones and Panicum maximum the only differentiating grass (Table are visible on the soil surface. The soli is poorly drained and 1). Other conspicuous trees and shrubs include Ziziphus low-lying areas are inundated for lon g periods after rains. mucronata (species group F), Acacia karroo (species group Species of species group A (Tabl e 1) ch aracterise this G), Diospyros Iycioides (species grou p I) and Rh us lancea community. The diagnos tic woody species are the exotics (species group P) with Celtis africana (sp ecies group F) and Salix babylonica, Schinus molle and Nicotiana glauca, the Olea europaea (species group H) less abundant (Tab le 1). reed Phragmites australis, the grass Tetrachne dregei, as Three distinct variants further characterise this plant com­ we ll as the shrublets Asclepias fruticosa and Melilotus alba munity (Table 1). (Table 1). According to Stirton (1978), Nicotiana glauca is indigenous in north-western and central Argentina , 1.2.1.1 Setaria verlicillata-Chenopodium album Variant Paraguay and Bolivia and commonly occurs along public roads in dongas and on river beds. According to Bromilow The Setaria verticillata-Chenopodium album Variant occurs (1995), it is thought to have been introduced into Namibia in in full sun and is restricted to th e disturbed and overgrazed horse fodder du ring the German occupation . Phragmite s zones of dry riverbeds . Duplex soils are prominent. Large australis sometimes is heavily grazed by livestock and gen­ trees are absent and the maximum height of tree species erally does not exceed one metre in he ight. Acacia karroa does not exceed two metres. (species group G) has an extremely high cover-abundance The two diagnostic species are pioneers , which colonise and completely dominates the vegetation, with Diospyros disturbed habitats. They are the grass Setaria verticillata Iycioides (species group I) also generally encountered and the forb Chenopodium album (species group B, Table (Table 1). 1). The absence of Olea europaea (species group H) also An ave rage of eight species per releve was recorded for differentiates this vegetation un it. Acacia karroo (species this community. group G) is inconspicuous and has an ave rage height of only 0.3 metres. Conspicuous species include the sedge 1.2 Ziziphus mucronata-Protasparagus laricinus Cyperus longus (species group 0), with Rhus pyroides Community (species group 0 ) and R. lancea (species group P) small, but nu me rous . Forbs are inconsp icuou s and are restricted to This plant community ca n be found on the relalively well­ Chenopodium album (speci es group B), Tagetes minuta, drained , sandy so ils of ri ve rbanks . Also present are place s Artemisia alra and Zinnia peruviana (species group F), as with clayey and ca lcareous soil. These places are generally well as the climber Clematis brachiata (species group F). Sou th African Journal of Botany 2001 . 67 58·64 61

Table 1: Phytosociological table of the drain age channels In the An average of 18 species per releve was recorded (Ta ble study area 1 ).

Major Community I 1 2 1.2.1.2 Crassula lanceolata-Acacia karroo Variant Community 11.1 1.2 1 2.1 1 2.2 12.3 Sub·Community I 1.2.1 .21 1.1 1·2 1 Variant 1 .1 1 .2 1 .3 1 1 1 I Th is enti ty is associated with wet riverbeds at the footslopes of dolerite hills. Th e undergrowth of th is vegetation unit 111121614466 15555 15556122615544 1335314555 155166 1 occurs in the full shade of the tree canopy. The soil is sandy Releve Number 19990901990012222 122201660199891649715999 198800 I 1678096145171012314652 16731236718666 11978 146945 1 with fine dolerite stones from ups lope (often 200mm deep) visible on the surface. This habitat is wetter than that of the Number of 10000001 11 1111111 111101 11010000 1000010000100000 I species/Releve 18689891972514456 155471138165791765915897 155835 1 Setaria verlicillata-Chenopodium album Va ri ant. This variant is not well developed . Crassula lanceolata SPECIES GROUP A and Pollichia campestris (species group C) are the only dif­ Salix babylonica 1121311 1 I Asclepias fruUcosa 11111111 I ferentiating species. Acacia karroo (species group G), the Tetrachne dregei IR11RRI I most abundant tree species, genera lly exceeds three Meliiotis alba JR RRR I I metres in height. Rhus pyroides (species group D), Celtis Nicotiana glauca 1111 1 1 I Schinus molle I 121 1 12 africana, Ziziphus mucronata (species group F) and Phragmites australisl13 R 1 I Diospyros Iycioides (s pecies group I) are more abundant SPECIES GROUP B Setaria verticillata I 12341 1 1 I I here than in the Setaria velticillata-Chenopodium album Chenopodium album IR1221 I R I R Variant. The presence of Olea europaea (species group H) SPECIES GROUP C a lso differentiates this varian t from the Setaria ve riicilfata­ Grassula [ancaolata 1 IR++11 Pollichia campestris 1 IRRRIR R R Chenopodium album Variant (Table 1) . Olea europaea is SPECIES GROUP 0 shrub-like and generally does not exceed two metres in Lycium hirsutum [ 1"" 1+"'1'12' 1 1 height. Rhus -pyroides I 1221R I 221 13111 1 1 2 Cyperus longus I 3 132111212 I 331 I 1 An average of 15 species per releve was recorded (Table Panicum maximum I 12111 1+ .1 111211 I 1 ) SPECIES GROUP E Heteromorpha trifoJiata I 13421 1.2,1.3 Rhus pyroides-Protasparagus laricinus Variant NldoreHa resedifolia I 1"' 1 SPECIES GROUP F Ziziphus mucro nata I 1'2"1'2131"'R I32' 1 Th is vegetation un it also occurs in dry riverbeds with a thick Protasparagus (50-1 00mm) layer of dolerite gravel, which gives It a rocky laricinus 2 11 1R1 1122RI 1221 1 11 l' appearance, covering the soil surface. The deeper soil lay­ Tagetes min uta 1"RRI'" 1'2"1"1 I I Clematis brachiala 11R1Rj1++1 12 1 IRRI I I ers are clayish and deep « 100mm deep). Cellis africana 11111112221 111 I I I Diagnostic species are absent (Table 1). The most abun­ Protasparagus dant woody species are Rhus pyroides (species group D) suaveolens I IR 1+ 1 11 1R11 l ' I 1 I Lyclum cinereum [ IR 11 RI 1R I I I I and Diospyros Iycioides (species group I). Acacia karroo Eragrostis obtusa I IR I I R 11 I I R I 1 (species group G) and Rh us lancea (species group P) are, Artemisia afra I 11 I 11R I I I I RI Zinnia peruviana I I'R I R I R I I R I I I although inconspicuous, the most abundant trees. The SPECIES GROUP G sedge Cyperus longus (species group D) often forms dense Acacia karroo 15552411111 132331211 121 11 2 SPECIES GROUP H stands. Noteworthy is the absence of Celtis africana Olea europaea I 11111 12 IR1 1'2" 1 12 (species group F) and the scanty occurrence of Acacia kar­ SPECIES GROUP I roo (species group G) and Olea europaea (species group Diospyros Iycioides 1222121112RRj3222 132 1'2221'2' 1 Diospyros austro H). Graminoids are scarce and restricted to Panicum maxi­ africana 11 112R21411 I mum (species group D) , Themeda tdandra and Sporobolus SPECIES GROUP J fimbdatus (species group 0). Heteropogon contortus 1 122211 R An average of 15 species per relev€ was recorded (Table Rhigozum obovatuml 1121 I 1 ). SPECIES GROUP K Rhus ciliata I +1 13111 1 12121 121 SPECIES GROUP L 1.2.2 Heteromorpha trifoliata-Nidorella resedifolia Aristida diffusa I 112 21 321 12 I Sub-community SPECIES GROUP M Rhus erosa I 13223123121 SPECIES GROUP N Thi s community is restricted to bottomland situations where Euclea crispa poorly drained calcareous so ils occur. The soil is ca lcare­ subsp. ovata I 11 '1 1111221 SPECIES GROUP 0 ous-clayey and calcareous stones are visible on its su rface . Themeda triandra I 11 12 21 221 2121 2 The soil associated with th is sub-community were found to Sporobolus be generally more clayey and trampled than those associat­ fimbrjatu5 1 12 1'2 "1 12 SPECIES GROUP P ed with the Rhus pyroides-Lycium hirsutum Sub-community. Rhus lancea 1 R I221 12311111 IR21 155341213+1531115241 I The vegetation is overg razed and the sh rub Rhus burchelli l 11 IR 11 I 111 11 2411244111 11 1 Heteromorpha tdfoliata, generally not higher than two metres, and the forb Nidorella resedifolia (species group E) 62 Malan. Venter and du Preez are diagnostic (Table 1). Acacia karroo , generally not and south-west. The soil is well-dra ined and sandy with exceeding two metres in height, and Ziziphus mucronata (2- exposed rocks covering less than 20% of the soil surface. 4 metres high) (species group F) are the most abundant tree Fine gravel virtually covers the soi! surface, especially in the species. Other tree species , include Olea europaea (species lower-lying areas. group H) and Rhus lancea (species group P). with R ciliata Rhus erosa (species group M) is very conspicuous and (species group K) the only locally abundant shrub (Table 1) . the on ly differentiatin g species present (Ta ble 1). Rhus The absence of Diospyros Iycioides (species group H) , also lancea (species group P), together with R. erosa (species differentiates this sub-community from the Rhus pyroides­ group M) and R. burc/lelli, (species group P) often fo rm Lycium hirsutum Sub-community (Table 1). dense stands , especially in the lower-lying areas (Table 1). Two sub-communities further differentiate this plant com­ 2 Rhus lancea-Rhus burchellii Major Community rnunity (Tabl e 1).

The Rhus lancea-Rhus burchellii Major Community is chiefly 2.2.1 Diospyros austro-africana-Rhus Jancea associated with dra inage channels at the foots[opes of low Sub-community hills. The habitat is more rocky than that of the Acacia kar­ roo -Diospyros Iycioides Major Comm unity and shows strong This sub-community is encountered on well-drained sandy affinities to the vegetation associated with the rocky out­ soils of south-west facing slopes where deep dongas are a crops in the southern Free State (Malan et al. 1998). Big common phenomenon. Signs of overgrazing are obvious. dolerite rocks (>2m in diameter) from upslope often cover Th e soil varies from shallow « 100mm) to deep (>200mm). more than 50% of the soi l surface. Th e soil su rface is often The soil is less rocky than in the Heteropogon contortus­ severely eroded and dongas (donga erosion) are conspicu­ Rhus lancea SUb-community. ously observable. This sub-community lacks exclusive diagnostic species Rhus lancea and R. burchellii (species group P) are the (Table 1). Diospyros Iycioides (species group I) and Rhus only widely dispersed woody species. Rhus burchellii lancea (species group P) are the most conspicuous tree (species group P) is diagnostic for th is vegetation group species present. Rhus laneea , here less prominent than in (Table 1). Rhus erosa (s pecies group M) and R. ciliata the Heteropogon contortus-Rhus lancea Sub-community, (s pecies group K) have a more restricted distribution (Tabl e generally does not exceed two metres in height. Rhus erosa 1). Rhus erosa , especially, is common to dominant on more (species group M) and Diospyros austro-africana (species arid mountain slopes and hil ls (Venter and Joubert 1985 ). group I) are the most prominent shrubs. The absence of Sign s of overgrazing are obvious in some cases, with the Heleropogon contortus and Rhigozum obovatum (species pioneer species Aristida diffusa prominently present group J), as well as the presence of D'ospyros Iycioides and (species group L, Table 1). Themeda triandra and D. austro-africana (s pecies group I) further differentiate th is Sporobolus fimbriatus (species group 0) are the on ly other sub-community from the Heteropogon contortus-Rhus relatively common graminoids (Table 1). lancea Sub-community (Table 1). An average of seven species per rei eve WQS recorded (Ta ble 1) 2.2 .2 Heteropogon contortus-Rhus lancea Three communities are distinguishable (Table 1). Sub-community

2.1 Olea europaea·Rhus lancea Community This sub-co mmunity is associated with well-drained sandy­ rocky soi ls. The aspect is south. Rock slabs cover 60% of Th is vegetation un it is restricted to the drainage channels at the so il surface. Fine gravel and dolerite stones are visible the base of rocky hills next to public roads. The aspect is on the su rface, leaving the rock slabs virtually unexposed. south and south-east. Soil of the Mispah Form is typical of Heteropogon contortus (species group J), absent in the this habitat with surface-rock percentages exceeding 20%. Oiospyros ausfro-africana-Iancea Sub-community, is the Big dolerite rocks are cornman, covering 20-50% of the soil only diagnostic grass present with the shrub Rhigozum obo­ surface. The soil is shallow «100mm deep) and rocky. valum (species group J) being the on ly diagnostic shrub Diagnostic species are absent, but the vegetation is rather (Table 1). This sub-community lacks species from species characterised by the exceptional dominance of Rhus lancea group H, which are evident in the Oiospyros austro-africana­ (species group p, Table 1). Olea europaea (s pecies group Rhus lancea Sub-community (Table 1). A further character­ H). Diospyros Iycioides and D. austro-africana (species istic of this vegetation unit is the partial dominance of Rhus group I) are the only other conspicuous woody species pres­ lancea and R. burchellii (species group P) res pectively. ent. Aristida diffusa (s pecies group L) and Rhus burchellii Rhus ciliata (species group K) is widespread and is the only (species group P) are the only prominent grass and shrub other prominent shrub. The grasses Themeda Iriandra species respective ly (Table 1). (species group 0) and Aristida diffusa (species group L) are An average of seven species per releve was recorded the only noteworthy graminoids present (Table 1). (Table 1). 2.3 Euclea crispa-Rhus /ancea Community 2.2 Rhus erosa-Rhus fancea Community The Rhus erosa-Rhus lancea Community is also found The Euclea crispa~Rhus lancea Community is restricted to along the footslopes of rocky hills. The aspects are south the rocky low-lying drainage channels at the foo t of dry South African Journal of Botany 2001. 67 58-64 63

north-facing hill slopes. Big dongas are common with a thick Major Community) are to the left of the diagram. Along Ihe layer of gravel (100-200mm) cove ri ng the soil surface. The second axis a moisture gradient exists. Plant communities in soil is shallow «100mm), rocky (>60%) and of the Mispah wet areas are situated at the top . The moistu re decreases Form (Land Type Survey Staff, In Press). towards the bottom with the communi ti es associated with The shrub Euclea crispa (species group N) differentiates dry habitat situated at the bottom of the diagram (Figure 2). th is plant community (Table 1). Rhus lancea (species group The distribution patte rn of the vegetation of drainage P) is to tally dominant and is the only conspicuous tree pres­ channels apparently depends primarily on the va riations in ent (Table 1). Graminoids are poorly represenled and only irradiation (Fu ls 1993) and rockiness of the soil surface. The Themeda triandra and Sporobolus fimbriatus (species group hierarchical classification of the vegetation stresses the cor­ 0) are worth mentioning (Tab le 1). relation between habitat and communities in the study area, as well as the relations hips between communities . Ordination Discussion In the scatter diagram the distribution of syn taxa along the first and second axes of the DECORANA ordination is given The vegetation of the drainage channels of Ihe southern (Figure 2). No clear discontinuity exists between the diffe r­ Free State is in a state of degradation. All Ihe plant commu­ ent communities. It was thus decided to restrict the ordina­ nities identified in this study could be related to specific envi­ tion to the two major ·communities. Clear discontinuity can ronmental conditions in the field. It is important to note that, be observed and the major plant communities are restricted although some of these communities show strong affin ities to specific areas in the diagram. The first axis illustrates a with the vegetation associated with rocky outcrops in the gradient that may be related to altitude and rockiness of the study area (Malan et al. 1998), all these communities were soil surface. Plant commun ities associated with the rocky encountered in bottomland situations and drainage chan­ drainage channels and the footslopes of hills (Rhus lancea­ nels. Rhus burchellii Major Community) are to the right of the dia­ Plant communities are dynamic and may display changes gram while commu nities of the [owlands and undulating ter­ in species composition according to the habitat. According to ra in with Jess rocky soil {Acacia karroo-Diospyros Iycioides Eckhardt (1993) , field in a good condition is usually charac­ terised by a high basal cover of high ecological status species. The [ow cover of ecological important grasses is nralnage Olannels Lo.. l"!ld~ further ind icative of veld in a poor cond ition (Van Oudtshoorn Undul a tlng Pl .. ,ns PQOts l opes ot 11111s 1991). The vegetation associated with drai nage cha nnels can therefore be described as field in a poor condition. The

Les s Rocky ' So~l Roclcy most important woody species are Acacia karma, D;aspyros Broded ~oils Iycio;des, Olea europaea, Protasparagus laricinus, Rhus (M;llnly Dong" burchellii, R. ciliata , Rhus lancea and Ziziphus mucronata. Bro::ionl Acacia karma, Diospyros Iycioides and Rhus lancea are the

I most important differentiating species. Themeda triandra 'H L ~:::.':~~~::,C"'.' .•. ,,,.,,,,., ", and Sporobolus fimbrialus are the most abundant grasses. b.""" ...... ·..... , ...... t ...... , ...... , : Signs of overgrazing are obvious wit h the pioneer species nO ~ i 0 Aristida diffusa prom inently present. Other species com­ .t ~ ~ monly occurring in overgrazed areas are Artemisia arra, I.'" 1 Asclepias fruticosa, Chenopodium album, Clematis brachia­ ,.~ :-> ta , Lycium cinerium, Nidore/la resedifolia, Protasparagus • \ suaveolens, and Tagetes min uta (Malan 1997). According to I ~ Van Oudtshoorn (1991) , A. diffusa is an unpalatable grass N .. i ~ and can be seen as indicative of overgrazed pasture. ~ I 0 " Aristida diffusa also frequently occurs on rocky ou tcrops in the southern Free State (Malan et al. 1998). 1 In contrast to the mountainous habitats , the comm unities \ 1 t Q, associaled with drainage channels display generally a lower ;; species diversity. The habitat is fa irly unstable, due to sea­ 1 .~ sonal flooding and drying which , together with the frequent ,j " overgrazing of the area, play an important role in the degra­ " dation of the vegetation . The strong presence of certain woody species within the drainage channels is ascribed to the favourable moisture regimes in these areas. Habitat dis­ AxIS 1 turbance (especially in areas along public roads) freq uently occurs . Figure 2: A DCA ordination of the major riparian shrub communities Ross (1948) stated that conlinuous degradation of natural in the southern Free State; (a) Acacia Karoo-Diospyros /ycioides pasture in many parts of the world is visible . Dense strands Major Community: (b) Rhus lancea~Rhus burchellii Major of fire resistant mature tree and shrub species supress the Community 64 Malan, Venter and du Preez

growth of grasses and fire thus its detrimental effect HIli MO (1979a) DECORANA· a FORTRAN program for detrended (Trollope 1974). Observations made in the Eastern Cape correspondence analysis and reciprocal averaglng_ Department show that Acacia karroa was able to survive eight years in of Ecology and Systematics. Cornell University, lthaca, New York succession by coping after each burn (Trollope 1974). HIli MO (197gb) TWINSPAN· a FORTRAN program arranging mul+ Ecologically sound conservation management pro­ livariate data In an ordered two·way table by classification of Indi· viduals and attributes Cornell University. Ithaca, New York grammes should take this present delineation of the riparian Land Type Survey Staff (in press) Land Types of the maps 2924 - shrub communities as the basis for future management Koffiefontein and 3024 • Colesberg . Mem . agric nat Resources. planning. The restricted distribution of these communities S. Afr No 14 emphasises th e fact that these areas should be given high­ Low AB . Re belo AG (eds) (1 996) Vegetallon of South Africa. er conservation priority. Lesotho and SwaZi land . Department of Environmental Affairs and Tou rism . Pretona. References Malan PW (1997) Vegetation ecology of the southern Free State Unpublished PhD dissertation. University of the Free State. Acocks JPH (1988) Veld types of South Afrrca 3rd edn . Mem bot BloemFontein Surv South Africa 57 : 1-146 Ma lan PW, Venter HJT, Ou Preez PJ (1998) Vegetation ecology of Arnold TH, De Wet Be (1993) of southern Africa: names and the southern Free State: Shrubland communities of the rocky out· distribution. Mem. bot. Surv S Afr. 62: 1-825 crops. South African Journal of Botany 64: 1· 17 Bayer AW (1970) Plant ecology in the service of man In southern Malan PW, Venter HJT, Du Preez PJ (1999) Vegetation ecology of Africa South African Journal of Science 66. 71 -77 the sou thern Free Slate: 1 Plan t communities of the Zastron Bezuidenhout H (1993) Syntaxonomy and synecology of the west· Area South African Journal of Botany 65 · 260·269 ern Transvaal grasslands. Unpublished PhD dissertation Malan PW, Venter HJT, Du Preez PJ (2000) Vegetation ecology of University of Pretoria . Pretoria the southern Free State: The Acacia communities. South AfTican Bosch OJH, Janse van Rensburg FP, Truter S du T (1987) Journal of Botany (submitted) Identification and selection of benchmark sites on Lilhofic soils of Mentis MT, Huntley BJ (1982) A description of the Grassland Blome the western Grassland Biome of South Africa. Journal of the ProJect. South African Natura l Science Programme 62: 1·29 Grasslands Society of South Africa 4: 143·147 MUeller·Dombois 0 , Ellenberg H (1974) Aims and methods of veg A Bredenkamp GJ , Bezuidenhout H (1990) The phytosociology of the etatlon ecology. Wiley, New York Faan Memtjes Nature Reserve in the western Transvaal grass· MOiler DB (1986) Plantekologle van die Willem Pretorius Wildtuin land, South Africa . South African Journal of Botany 56: 54·64 Ph. D. dissertation, Un iversity of the Orange Free State . Bredenkamp GJ, Joubert AF. Bezuidenhout H (1989) A reconnais· Bloemfontein sance survey of the vegetation of the plains in the Potchefstroom· Pentz JA (1938) The value of botanical survey and the mapping of Fochville·Parys area. South African Journal of Botany 55 : 199· vegetation as applied to farming systems In South Africa. Mem 206 bot. Surv. S Afr 19: 1·15 Bredenkamp GJ, Theron GK (1978) A synecological account of the Ross JC (1948) Land utilization and soli conservation in the Union Suikerbosrand Nature Reserve. 1 The phytosociology of the of South Africa. BuBetin Div. Soil Can. and Ext. , Dept. of Witwatersrand geological system. Bothalia 12: 513·529 Agriculture, Pretoria Bromilow C (1995) Problem pla nts of South Africa. Briza Rossouw LF (1983) 'n Ekologiese studle van die boomgemeen· Publications, Arcadia , Pretoria skappe van die Bloemfontein-omgewing , OranjeNrystaat. MSc Carr JO (1976) The South African Acacias. Conservation press th esis, University of the Ora nge Free State, Bloemfontein (PTY) LTD Johannesburg Rutherford MC, Westfall RH (1994) Biomes of southern Africa· an Codd LEW (1949) The application of ecology to agricultural prob· objective calegonzalion. 2nd edn Mem. bot. Surv. S. Afr 63 1-94 lems in South Africa. In: Statement and communication of the Scheepers JC (1975) The plant ecology of the Kroonstad and African Regional Scientific Conference, 17·28 October, Bethlehem areas of the Hlghveld Agricultural Region. DSc thesis. Johannesburg, pp 115·119 University of Pretoria, Pretoria Du Preez PJ (1991 ) A syntaxonomical and synecological study of Scheepers JC (1987) Grassland Biome ProJect: Proceedings of the the vegetation of the south· eastern Orange Free State and rei at· workshop on the classification and mapping. S Afr. Ecosys. Prog. ed areas with special reference 10 Korannaberg. PhD disserta· Occ. Rep. Series 16.1·31 lion. University of the Orange Free State, Bloemfontein Stirton CH (1978) Indringerplante. Mooi, maar gevaarllk. Eckhardt HC (1993) A synecological study of the vegetation of the Department of Nature Conservation , Cape Provincial north-eastern Orange Free Stale. Unpublished MSc thesis, Administration , Cape Town Un iversity of Pretoria, Pretoria Trollope WSW (1974) Fire as a method of controlling macchia (fyn· Eckhardt HC, Van Rooyen N, Bredenkamp GJ (1 993) Wetland plan t bas) vegetation on the Amatole mountains of the eastern Cape communities of the Vrede·Memel Warden area, north·eastern Proceedings of the Grasslands SOCiety of South Africa 8 35·41 Orange Free State. Navors. nas. Mus. Bloemfontein 9: 246·262 Van Oudtshoorn FP (1991) Gids tot grasse van Suid·Afrika. Briza Foran 80, Bastin G, Shaw KA (1986) Range assessment and mon ~ Publications, Arcadia, Pretoria itoring in arid lands: the use of classification and ordination in Va n Wyk S (1 983) 'n PJanlekologiese studie van die Abe Bailey· range survey. Journal of Environmental Management 22 : 67·84 natuurreservaat MSc thesis, Potchefstroom University for Fuls ER (1993) Vegetation ecology of the northern Orange Free Christian Higher Education, Polchefstroom State. Unpublished PhD dissertation. University of Pretoria, Venter HJT, Joubert AM (1 985 ) KlimpJante. bome en struike van die Pretoria Oranje·Vrystaat. 2nd ed . PJ de Villiers, Bloemfontein

Edited by GJ Bredenkamp