Effects of thirty-five years of afforestation with Pinus radiata on the composition of mesic mountain near Stellenbosch

D.M. Richardson and B.W. van Wilgen South African Forestry Research Institute, Jonkershoek Forestry Research Centre, Stellenbosch

The fynbos vegetation of Biesievlei, Jonkershoek, was surveyed Introduction and described in 1945. In 1948 the catchment was afforested The Jonkershoek Forest Influences Research Station, as it was with Pinus radiata. This paper presents results of a reassessment of the vegetation in 1984 using the same then known, was established in 1935, chiefly to investigate methods that were used in 1945. Afforestation has reduced the the effects of afforestation on streamflow. Biesievlei is one cover of the vegetation (excluding P. radiata) from 75% to 20%. of seven experimental catchments that was afforested for this The total number of species was reduced by 58% from 298 to 126. At least 190 species found in 1945 were not found in 1984, purpose. A study of the fynbos vegetation of Biesievlei was and at least 18 species were added to the list. The mean undertaken in October - November 1945 (Rycroft 1950). This density was reduced from 260 to 78 m - 2. Only stream study was the first to analyze sclerophyllous fynbos vegetation bank vegetation, comprising mainly large-leaved sprouting using quantitative methods. Its main purpose was to develop shrubs, persisted in a relatively unmodified state. Away from the stream , annuals, geophytes and hemicryptophytes were methods for sampling vegetation in the fynbos but the results dominant. Dominant spe.cies in the pre-afforestation flora were provide a unique opportunity to assess the effects of afforesta­ not resilient to afforestation. Serotinous Proteaceae, woody tion on the composition of the indigenous vegetation. Invasion small-leaved sprouting and myrmecochorous shrubs and large­ by alien trees and shrubs is a major threat to the conservation leaved sprouting shrubs have been virtuall y eliminated. Certain groups may re-establish after clearfelling, but others, notably of fynbos. Seagrief (1950), Cowling et al. (1976) and Milton various groups of shrubs may have been permanently (1976) have described fynbos plant communities under pine eliminated. The implications for weed control in conservation plantations or infestations near Grahamstown, on Table areas are discussed. Mountain and at Jonkershoek respectively, but no quantitative S. Afr. J. Bot. 1986, 52: 309-315 information is available on the effects of afforestation on the Die fynbos-plantegroei van Biesievlei , Jonkershoek is in 1945 floristics and structure of indigenous vegetation. This study ondersoek en beskryf. Die opvanggebied is in 1948 met Pinus was aimed at quantifying the effects of 35 years of suppression radiata bebos. In hierdie artikel word die resultate aangebied van 'n 1984-heropname van die plantegroei waarin dieselfde by a dosed-canopy stand of Pinus radiata. This may be used metodes gebruik is as die in 1945. Bebossing het die to represent the effects of severe invasion by alien trees and, plantegroei-bedekking (P. radiata uitgesluit) vanaf 75% tot 20% by assessing the effects of such 'invasions', certain hypotheses verminder. Die totale aantal spesies is met 58% verminder, van on the effects of weed-clearing operations can be proposed. 298 tot 126. Ten minste 190 van die spesies wat in 1945 gevind is, is nie weer in 1984 opgemerk nie en ten minste 18 spesies is bygevoeg tot die lys. Die gemiddelde plantdigtheid is The study area verminder vanaf 260 na 78 plante m - 2. Slegs oewerplantegroei, The Jonkershoek Valley (33°57'S, 18°55'E) is situated to the hoofsaaklik herspruitende struike met groot blare, het bly south-east of Stellenbosch in the south- Province, voortbestaan in 'n relatief onveranderde toestand. Weg van die stroom, was jaarplante, geofiete en hemikriptofiete dominant. (Figure 1). It is surrounded on three sides by Dominante spesies in die voor-bebossingsflora was nie bestand mountains formed predominantly from sandstones of the teen bebossing nie. Serotiene Proteaceae, houtagtige, smalblaar, Group. The Biesievlei Catchment, 27 ha in spruitende en mirmekochore struike en grootblaar, spruitende struike is so te se geelimineer. Sekere groepe mag hervestig na extent, lies on the south-facing slopes of the valley. The area kaalkap, maar andere, veral verskeie groepe struike is dalk is underlain by Cape Granite with a small band of Malmes­ permanent verdring. Die gevolge vir onkruidbeheer in bury Shale in the upper catchment. The soil of the upper bewaringsgebiede word bespreek. slopes is a grey-brown loam on partly decomposed shale while S.·Afr. Tydskr. Plantk. 1986, 52: 309-315 at the lower elevations, grey heavy loam soils on yellow clay Keywords: Afforestation, alien plants, ecosystem stability, predominate (Rycroft 1950). Elevation ranges from 290m at fynbos, Pinus radiata the stream-recording weir to 580 m at the highest point. The mean aspect and slope of the catchment (each measured at 200 random points) are 226° and 19° respectively (Rycroft 1950). The upper portion of the catchment has steep rounded D.M. Richardson* and B.W. van Wilgen ridges separated by steep-sided gullies; the average slope of South African Forestry Research Institute, this portion is 26°. The lower portion flattens out below 355 m J onkershoek Forestry Research Centre, Private Bag X 50 I I, to a gentle slope of about 9° (Rycroft 1950). Stellenbosch, 7600 Republic .of South Africa The climate of the area is mediterranean [Koppen's (1931) *To whom correspondence should be addressed humid-mesothermal type Csb] with a dry summer and the average temperature of the warmest month below 22°C. The Accepted 4 March 1986 mean annual rainfall is 1427 mm, of which about 630Jo falls 310 S.-Afr. Tydskr. Plantk., 1986, 52(4)

Figure I The Jonkershoek Valley near Stellenbosch showing the position of Biesievlei.

between May and August (unpublished records, Forestry aethiopicum and Leucadendron sa!ignum was estimated on Branch). each plot. Each species was assigned to a life form (after Rycroft (1950) described the natural vegetation of the area Raunkiaer 1934) and the cover of each life form on each plot and distinguished 13 communities. Dominant species included was determined. The life forms were stem succulents, epi­ Protea burchellii, P. neriijolia [listed as P. !epidocarpodendron phytes, megaphanerophytes and mesophanerophytes, micro­ by Rycroft (1950)], Leucadendron sa!ignum (Proteaceae), phanerophytes, nanophanerophytes, chamaephytes, hemi­ Anthospermum aethiopicum (Rubiaceae), Rhus angustijolia cryptophytes, geophytes, helophytes and hydrophytes, and (Anacardiaceae), Clijjortia cuneata (Rosaceae), and Widdring­ therophytes. The proportion of the flora in each life form tonia cupressoides (Cupressaceae). Nomenclature in this paper was then determined. follows Bond & Goldblatt (1984). Families with 15 or more In November 1984, 39 years after the initial study, the species (number of species in brackets) were: Asteraceae (59), vegetation was re-surveyed. Two hundred quadrats were Fabaceae (36), Iridaceae (31), Poaceae (24), Cyperaceae (17), positioned at random in the area. Seven of these fell outside Oxalidaceae (17) and Proteaceae (15). Ten species of Restiona­ the catchment boundaries and were excluded. A portable steel ceae and nine of Ericaceae were present. The vegetation of frame was used to demarcate each of the remaining 193 quad­ upper Biesievlei was last burnt approximately 19 years before rats. Data were collected in such a way as to be directly compar­ afforestation (Van Wyk 1977). Vegetation in the lower part able to Rycroft's (1950) survey. In addition, we also estimated of the catchment was burnt in a wildfire in 1942 (unpublished the cover of individual species in each quadrat. Specimens records, Forestry Branch). In 1947 the vegetation around which could not be identified in the field were collected, and individual planting sites was slashed but not burnt. Almost these were identified by matching them with herbarium speci­ the whole catchment (980Jo) was afforested with P. radiata mens at Jonkershoek. Specimens that could not be identified in 1948. Areas not afforested were steep cliffs where planting (very small individuals or seedlings with no flowers or fruit) 1 was not possible. Trees were planted at 1200 stems ha- • were given serial numbers so that an accurate assessment of the Thinnings to 740, 494, 320 and 158 stems ha- 1 were carried number of species present was made. out in 1959, 1964, 1971 and 1976 respectively. Trees were As Rycroft (1950) did not give cover values for individual pruned to 7 m in 1962. The catchment was clearfelled in 1985. species, direct comparison of dominant components of the vegetation was not possible. In order to compare the dominant Methods components of the vegetation in 1945 and 1984, lists of the In the 1945 survey, 200 quadrats of 0,1 m2 each (0,40 x most important species encountered in the two surveys were 0,25 m) were located at random in the catchment. Species compiled. Rycroft (1950) described 13 vegetation communities were listed on each quadrat. Individual plants rooted in the based on his survey. From these accounts, we listed all those quadrats were counted on each quadrat to estimate density. species described by Rycroft variously as abundant, charac­ Total plant cover and the cover of all Poaceae; Restionaceae teristic, chief, common, conspicuous, dominant, frequent, and Cyperaceae; geophytes; and the species Anthospermum principal, prominent or well represented. We will use the term S. Afr. J. Bot. , 1986, 52(4) 311

50 dominant to describe these species in further discussion. For 1945 the 1984 survey, an importance value (sensu Mueller-Dombois n = 200 & Ellenberg 1974) was calculated for each species as the sum of relative density, relative frequency and relative cover. Relative density was defined as the number of individuals of the species in all quadrats, expressed as a percentage of the total number of individuals of all species. Relative frequency o~~~~~2~~~~~g~R~~~~~ was defined as the frequency of the species (the number of I I I I I I I I I I I I I I I I I I I I I ~w~~N~M~~~~~w~~~oo~m~ ~ quadrats in which it occurred divided by the total number Number of individual plants per quadrat of quadrats), expressed as a percentage of the sum total of frequency values for all species. Relative cover was defined as the sum of all cover estimates for the species in all quadrats, expressed as a percentage of the sum total of cover values for all species. Species in both lists were assigned to reproduc­ tive and growth form guilds (see Table 5) in order to compare the proportional distribution of the dominant flora in such 90 guilds. 80

70 Results ~"' ~ 60 Plant canopy cover 0" 0 50 Cover values for the vegetation categories recorded from both E surveys are given in Table 1. The total cover of species § 40 z excluding Pinus radiata has been reduced from 74,7 to 19,70Jo 30

following afforestation. The cover of each category of plants 20 has been reduced markedly except for geophytes. Rycroft 10 (1950) did not give standard deviations for the mean values

which prevents statistical treatment of the data. ol,()~~~~?3~~!;gf6:gg~R<·······< =- . I I I I I I I I I I I I I I ..-co;:~N~c;;~;~u;[:RW$ Number of individual plants per quadrat Table 1 Percentage plant cover for categories of vegetation in two surveys in the Biesievlei Catchment Figure 2 Comparison of plant density (plants per 0, I m2 quadrat) in in 1945 and 1984 two surveys in Biesievlei in 1945 and 1984.

Mean cover (O?o) Vegetation component 1945 1984 recorded by Rycroft (1950) were not found in 1984. This figure is potentially as high as 82%, which would be the case All plants 74,7 19,7 if our 54 unidentified specimens are all species not found by 12,3 2,3 Poaceae Rycroft. Similarly, between 18 and 71 species were added to 17,9 1,6 Cyperaceae and Restionaceae the list. Afforestation thus results in the elimination of a very Geophytes 6,6 4,4 large proportion of the species. Some species survive in Anthospermum aethiopicum 2,9 0,1 Leucadendron sa/ignum 1,2 0,1 reduced numbers despite afforestation. Some species not found in pristine fynbos appear after afforestation. Many of these are small herbaceous plants or cosmopolitan weeds. Alien woody plants found in Biesievlei in 1984, although not Plant density recorded on the quadrats, were Acacia melanoxylon, Hakea Plant density as estimated from the two surveys is compared sericea, Pittosporum undulatum and Solanum mauritianum. in Figure 2, which reflects a marked decline in the number of plants per quadrat. The mean plant density has declined Species dominance by 70% from 260 plants m - 2 in 1945 to 78 plants m- 2 in Examination of Rycroft's (1950) community descriptions 1984. Only one quadrat (0,5% of the total) contained no revealed 73 dominant species. These are listed in Table 2 under plan,ts in the 1945 survey whereas 37 quadrats (19,2% of the the corresponding reproductive and growth form guilds. The total) were empty in 1984. The maximum plant densities 30 most important species in 1984 and the corresponding recorded in 1945 and 1984 were 1390 and 700 plants m- 2 frequency, density and cover values are given in Table 3. Of respectively. Rycroft's 13 original communities, only the stream bank community persists 39 years after afforestation. The remainder Species diversity and turnover of the understorey vegetation consists of a sparse cover The change in the number of species per quadrat shown in dominated by a few species (mainly Oxalis spp., Briza maxi­ Figure 3 reflects a reduction in alpha diversity due to af­ ma, Hypochoeris radicata and Halleria elliptica) (Table 3). forestation. The mean number of species per quadrat declined from 8,5 in 1945 to 2,1 in 1984. The total number of species Life form spectra encountered in the catchment was 298 in 1945 and only 126 Rycroft (1950) gave life form data for Biesievlei as the number in 1984. We were unable to identify 54 specimens from the of species in a given life form expressed as a percentage of 1984 sample due to the nature of the material. Of our sample the total number of species present. Comparative figures for of 126 species, 54 were common to Rycroft's (1950) list, 18 the two surveys are given in Table 4. Chamaephytes have been were not found in 1945 and 54 could not be identified to markedly reduced by afforestation while the proportion of species level. This means that at least 64% of the 190 species hemicryptophytes has increased. Other life forms have 312 S.-Afr. Tydskr. Plantk., 1986, 52(4)

40 1945 Table 2 Seventy-three dominant species from Ry­ n = 200 croft's (1950) description of plant communities of Biesievlei in 1945. (See text for criteria used to select "§"' 1\l 30 species) ::J r:::r Herbaceous plants 0 Q; Annuals: Sebaea aurea .0 E 20 Geophytes: Corymbium glabrum, C. vil/osum, Micranthus z::J alopecuroides, Pro/asparagus rubicundus Non-seasonal hemicrytophytes: Bobartia indica, Cannamois virgata, Carpha capite/lata, Ficinia filiformis, F. indica, 10 Hypodischus albo-aristatus, Ischyrolepis subverticellata, Restio filiform is, Tetraria bromoides, T. burmanii, Zantedeschia aethiopica. Seasonal hemicrytophytes: Berkheya armata, Dryopteris bergiana, 0 1 5 10 15 20 Ehrharta longifolia, Pennisetum macrourum, Number of species per quadrat Pentaschistis curvifolia, Pentaschistis sp., Plagiochloa uniolae, Pteridium aquilinum, Themeda triandra. Other herbs: Cassytha ciliolata, Gunnera perpensa, Senecio pinifolius

Woody small-leaved plants ( ~ leptophyllous) 50 Sprouters: Brunia nodiflora, Diosma hirsuta, Helichrysum teretifolium, Montinia caryophyllacea, Muraltia heisteria, Phylica spicata, Struthiola myrsinites, Widdringtonia cupressoides 40 :: "§"' Reseeders: "0 ::J - Ericaceae: Erica coccinea, E. imbricata "'r:::r 0 1984 Myrmecochorous: Anthospermum aethiopicum, Cullumia Q; 30 n = 193 .0 ciliaris, C. setosa, Indigojera cytisoides, E Psoralea aphylla ·z::J - Wind dispersed: Helichrysum odoratissimum, Oedera sp., 20 Senecio subcanescens, Stoebe plumosa

Woody large-leaved plants (;;;, nanophyllous) Sprouters: - Bird dispersed: Hal/eria el/iptica, flex mitis, oleoides, Olea europaea ssp. africana - Other: stellatifolium, , ,:,~=:: Freylinia lanceolata, Leonotus leonurus, 0 1 2 3 4 5 6 7 Leucadendron salignum, Metrosideros Number of species per quadrat angustifo/ia, Otholobium fruticans, 0 . obliquum,

2 Podalyria calyptrata, Protea nitida, Rhus Figure 3 Comparison of the number of species per quadrat (0, I m ) angustifolia, Rhus tomentosum, Salvia in two surveys in Biesievlei in 1945 and 1984. africana-caerulea Reseeders: - Serotinous: Leucadendron rubrum, Protea burchellii, remained at much the same proportion of the flora. The in­ P. neriijolia crease in mega- and mesohanerophytes and reduction in micro­ - Myrmecochorous: Bolusafra bituminosa (?), Cliffortia phanerophytes can be attributed to differences in interpretation cuneata, C. pterocarpa when assigning species to life forms. For example, we classified - Wind dispersed: Othonna quinquidentata the species Brabejum stellatifolium, africana, - Bird dispersed: Chrysanthemoides monilifera Maytenus oleoides and Olea europaea as mega- or meso­ Succulents phanerophytes. These species were all present in 1945 which Erepsia anceps can only mean that they were called microphanerophytes by Rycroft (1950). eliminated (Table 5). There is also a noticeable increase Reproductive and growth form guilds in the relative proportion of bird dispersed, large leaved The proportion of the dominant flora in terms of reproductive sprouters. and growth form guilds for the two surveys is compared in Table 5. Dominant species in 1984 are mainly herbaceous Discussion plants (630Jo) and woody, large-leaved shrubs (230Jo). Annuals, The validity of the basis of comparison geophytes and hemicryptophytes are dominant away from the This study aimed to repeat the work of Rycroft (1950) and stream bank whereas large-leaved sprouting shrubs form the the data were collected in exactly the same way as was done dominant guild in the riparian zone. Small-leaved sprouting in that pioneering study. The science of vegetation ecology shrubs and large-leaved shrubs (especially those that regenerate has advanced considerably since then. If our aim had been from canopy-stored seeds) have been greatly reduced or to describe the vegetation, we would have taken such advances S. Afr. J . Bot., 1986, 52(4) 313

Table 3 The thirty most important species in the species (Table 5). However, such differences in methodology Biesievlei Catchment in 1984. Importance was defined would have prevented direct comparisons. Through direct as the sum of relative density, relative frequency and comparison we have been able to quantify changes in cover, relative cover (see text). For explanation of guild codes density and in broad life form composition. We also listed see Table 5 cover by species on each plot and have used this to compile a list of the dominant flora (Table 3). We have compared "O;g .... Q) t:: ;::l Q) ;::l this to the subjectively determined list of dominants in 1945. ro oo ;> C Such a comparison must be seen as coarse and the results .~ N u ;> E ~ Q) ·- .... ;., V> l u indicative of broad trends rather than an exact account of gs u ~ E t:: t:: "0 .OS "O..c Q) V> ~ .... changes in composition of dominant species. Given the 0 ~ ;::l t:: ~ t:: 0 .... ;:!: c:r ro .:: ro 0. Q) v_!g Q) magnitude of changes that have occurred, we feel that such & e .... .§ Species c:;: 00 t:.L.. ~5 ~ a comparison is both valid and useful. Oxalis lana/a GEO 25 ,4 13 ,4 1,5 37 ,0 Briza maxima ANN 12,4 16,8 1,7 35 ,8 Effects of afforestation on vegetation Oxa(is purpurea GEO 13,0 13,4 2,0 32,6 Afforestation has caused a clear reduction in the number of Hypochoeris radicata ANN 21,8 4,9 0,9 21,1 species present, and in the cover and density of remaining Brabejum stellatifolium LLSO 4,7 0,7 2,6 15 ,2 species in Biesievlei. Afforestation has also changed the relative Halleria el/iptica LLSB 8,3 1,7 1,2 11 ,8 importance of various reproductive and growth form guilds Rubus rigidus LLSB 5,7 0,7 0,7 7,0 in the remaining vegetation. The most important families in 4,7 0,5 0,8 6,7 Berkheya armata GEO the impoverished flora are Oxalidaceae, Poaceae and Astera­ Pteridium aquilinum SHE 3,1 1,2 0,7 6,2 ceae whereas Proteaceae, Rosaceae and Cyperaceae were Stoebe cinerea SLW 1,0 0,1 0,9 4,7 Rhus angustifolia LLSO 2,6 0,3 0,6 4,5 found to be most important in terms of cover in an un­ Oxalis bifida GEO 3,1 1,8 0,1 4,4 disturbed fynbos community at Jonkershoek (Van Wilgen Oxalis incarnata GEO 0,5 2,3 0,2 4,3 1981). A tall closed shrubland with a rich assemblage of Maytenus oleoides LLSB 1,0 0,1 0,8 4,1 species has become a very sparse understorey with relatively Grass sp . SHE 2,6 1,7 0,2 4,0 few species under the closed pine canopy. Similar patterns Protasparagus rubicundus GEO 2,1 0,2 0,5 3,7 of vegetation change are evident under severe infestations of Ehrharta bulbosa SHE 2,1 1,5 0,1 3,4 pines and other species such as Hakea sericea and Acacia Myrsiphyllum scandens GEO 3,1 0,5 0,3 3,3 saligna (Van Wilgen & Richardson 1985; Macdonald & Senecio sp. ANN 2,1 1,2 0,1 3,1 Richardson in press). Populations of many species have Tetraria bromoides SHE 1,6 0,5 0,4 3,0 crashed in response to afforestation and the resultant 'stable Rhus tomentosa LLSO 2,1 0,3 0,2 2,5 state' consists of pines, relictual fynbos populations, cosmo­ Plectostachys polifolia SLW 1,6 0,3 0,2 2,2 politan weeds and other fugitive herbaceous plants. Serotinous Pellaea pteroides SHE 2,1 0,3 0,2 2,1 non-sprouters (Protea spp.), woody small-leaved sprouters and Tetraria cuspidata NHE 1,0 0,2 0,3 2,0 Tetraria involucra/a NHE 1,0 0,1 0,3 2,0 woody small-leaved myrmecochorous species have been Erica hispidula SLE 1,0 0,1 0,3 1,8 eliminated. The number of dominant sprouting shrub species Ischyrolepis gaudichaudianus NHE 2,1 0,4 0,1 1,8 has been reduced from 21 to 3 (Table 5). Herbaceous plants Olea europaea ssp. africana LLSB 1,6 0,2 0,0 1,8 have become dominant by default as other species have been Tetraria sp. NHE 0,5 0,1 0,3 1,7 eliminated. This does not necessarily imply that they have Cliffortia cuneata LLM 2,1 0,2 0,1 1,6 become more abundant. Of the species listed as dominant in 1984 (Table 3), only the small, deep-rooted forb Hypochoeris radicata is a new species. Only large-leaved shrubs of the stream bank community have persisted in significant numbers. Table 4 Plant life form distribution in Biesie­ Natural communities on permanently moist sites on Table vlei in two surveys expressed as a percentage of the flora. The figure in brackets for the 1984 Mountain showed similar resistance to modification following survey is the mean cover of the life form cal­ invasion by pines (Cowling et at. 1976). These authors found culated from 193 quadrats that the Osmitopsis asteriscoides- Berzelia lanuginosa com­ munity was 'not severely affected' by invasion. The stream Li fe form 1945 1984 bank above the 335 m contour in Biesievlei was not fringed Epiphytes 0,3 0 (0) by tall woody plants prior to afforestation (Rycroft 1950), Mega- and mesophanerophytes 0 3,2 (3,9) but seedlings of Kiggelaria ajricana and the alien weed Sola­ Microphanerophytes 6,4 3,2 (I ,0) num mauritianum have become established in this zone. The Nanophanerophytes 23,5 24,8 (4,6) pines have provided perching sites for frugivorous birds and Chamaephytes 23,2 5,6 (0,4) thus new foci for the deposition of seed. Hemicryptophytes 14,1 38,4 (4,9) Geophytes 22,1 22,4 (4,4) Disturbance and stress as ecological factors in Hydro- and helophytes 2,3 1,6 (0,2) afforested habitats Therophytes 8,1 0,8 (I, 7) Disturbances are defmed as 'mechanisms which limit the plant Total number of species 298 126 biomass by causing its partial or total destruction' (Grime 1979). Uninvaded mountain fynbos habitats are subject to disturbance in the form of fires at intervals of 10 to 40 years. into account. For example, much larger plots would have been Dominant species in habitats subject to disturbance are used and the classification of species into life forms (as done generally resilient under a particular disturbance regime. Stress by Rycroft) could have been done along more meaningful factors are defined as phenomena which 'restrict photo­ lines, such as we have attempted to do with the dominant synthetic production, such as shortages of light, water, and 314 S.-Afr. Tydskr. Plantk., 1986, 52(4)

Table 5 Reproductive and growth form guilds of the dominant species of Biesievlei in 1945 (See Table 2) and 1984 (See Table 3)

1948 1984

Reproductive Number of O?o of the Number of %of the and growth dominant dominant dominant dominant Life form form guilds species flora species flora

Herbaceous plants Annuals ANN I I 3 10 Geophytes GEO 4 5 7 23 Non-seasonal hemicryptophytes NE 10 14 4 13 Seasonal hemicryptophytes SHE 9 12 5 17 Other Herbs HER 3 4 0 0

Woody, small-ieaved ( ~ leptophyllous) Sprouters SLS 8 II 0 0 Seeders Ericaceae SLE 2 3 I 3 Myrmecochorous SLM 5 7 0 0 Wind dispersed SLW 4 6 2 7

Woody, large-leaved (::;;~· nanophyllous) Sprouters Bird dispersed LLSB 4 6 4 13 Other LLSO 13 18 3 10 Seeders Serotinous LLC 3 4 0 0 M yrmecochorous LLM 4 6 I 3 Wind dispersed LLW I 0 0 Bird dispersed LLB 0 0 Succulents sue I 0 0 Total 73 30

mineral nutrients' (Grime 1979). Dominant species in disturbed affected by the increased stress (Table 4). Woody, large-leaved habitats are able to withstand low levels of stress. An un­ serotinous re-seeders such as Protea species have no viable natural increase in the intensity of stress (such as that caused strategy to survive such stress and may be permanently by afforestation) could, however, be expected to eliminate eliminated. Seeds of these species have a low dispersal effi­ many elements of the flora (Grime 1979). Afforestation with ciency and do not persist in the soil (Bond 1980). Sprouting pines has increased stress in a number of ways. Pines are more shrubs may also be permanently eliminated as seed production effective users of available resources; they grow faster and is not usually vital to the survival of sprouters, and no large reach a much greater size than the indigenous plants. Litter seed banks can be expected to remain. Myrmecochorous fall from pines causes considerable suppression of the under­ shrubs, although severely reduced in the above-ground flora, storey vegetation. Litter fall in a 35-year-old P. radiata stand may re-establish from soil-stored seed banks. Wind dispersed at 1onkershoek was estimated at 372 g m- 2 yr. - 1 (Versfeld species may recolonize from outside the area. Other species 1981). Litterfall in coastal fynbos amounts to between 72 and survive in certain habitats. This may be due to the inherent 84 g m- 2 yr. - 1 in 9-year-old vegetation (Mitchell et a!. in tolerance of these species to stress or to ameliorating factors press), while at Jonkershoek the figure is 217 g m- 2 yr.- 1 that reduce the magnitude of stress. Woody large-leaved in 22-year-old vegetation (F.J. Kruger & A.J. Lamb un­ sprouting shrubs persist near streams where moisture is published data). Vapour from decomposing P. radiata litter abundant and where P. radiata grows least well (Grey & inhibits growth of certain species (Lill & McWha 1976). Taylor 1983). Shading will further affect photosynthesis and germination of understorey plants. The relative contribution of each of Implications for conservation in invaded habitats these stress factors to the suppression of elements of the The results of this study can be useful for assessing priorities natural vegetation is not known but the combination of factors for weed clearing in the fynbos. Afforestation has created a has undoubtedly caused the demise of several groups of new community with properties and attributes entirely dif­ species. Invaded habitats can be seen as analogous to af­ ferent from the pre-afforestation fynbos which was resilient forested sites. Invaded habitats are subject to increased stress in relation to the disturbance regime (fires) under which most which, if the invasions are severe enough, will eliminate many of its component species evolved. Following disturbance elements of the previously dominant flora. (clearing and/or fire) in heavily invaded sites, the altered The reproductive output of the community has been clearly community probably cannot return to its pre-invasion state. reduced in proportion to the reduction in density and cover This means that the vegetation has been stressed to a point of the vegetation. The following question then arises: which where it is no longer resilient to the disturbance regime under species will be able to regenerate following a disturbance such which its dominant components have evolved. This can be as clearfelling or fire? Examining the response of reproductive tested following clearfelling in Biesievlei by monitoring suc­ and growth form guilds to stress may answer this. Geophytes cession on larger, permanent plots. were the only group away from streams apparently not Two important points arise with regard to the practical S. Afr. J. Bot., 1986, 52(4) 315

implications of this study. Firstly, dense closed-canopy infesta­ DIAMOND, J. 1985. How and why eroded ecosystems should be tions of pines or other large trees and shrubs will alter the restored. Nature 313: 629-630. composition of fynbos in such a way that clearing alone will GREY, D.C. & TAYLOR, G.I. 1983. Site requirements for commercial afforestation in the Cape. S. Ajr. For. J. 127: probably not restore the original flora. Shrubby elements of 35-38. the flora will be eliminated. In the future it may be necessary GRIME, J.P. 1979. Plant strategies and vegetation processes. John to repair such impoverished ecosystems ('restoration ecology' Wiley, London. sensu Diamond 1985). Secondly, infestations which are cur­ KOPPEN, W. 1931. Grindruss der Klimakunde. De Gruyter, rently sparse should not be left to become dense (and thus Berlin. cause the natural vegetation to become stressed) before KRUGER, F.J. 1981. Conservation: South African heathlands. In: Heathlands of the world. B. Analytical studies, ed. Specht, clearing. Priority should be given to the clearing of sparse R.L. pp. 231-234, Elsevier, Amsterdam. or moderate infestations as restoration of the areas will be LILL, R.E. & McWHA, J.A. 1976. Production of ethylene by difficult should weed populations become dense. incubated litter of Pinus radiata. Soil Biology and Bio­ Given that dense, closed-canopy infestations have already chemistry 8: 61-63. radically altered the composition of the vegetation, and that MACDONALD, I.A.W. & RICHARDSON, D.M. in press. Alien clearing will not restore the natural vegetation, should such species in terrestrial ecosystems of the fynbos biome. In: The ecology and control of biological invasions in South Africa, ed. infestations be cleared? Dense infestations have a number of Macdonald, I.A.W. & Kruger, F.J. Oxford University Press, disadvantages besides causing drastic changes in community Cape Town. structure and species diversity. Surface water resources and MILTON, S.J. 1976. Understorey vegetation of a pine plantation, the scientific, aesthetic and recreational value are reduced, and Bosboukloof, Jonkershoek, S.W. Cape. B.Sc. Hons. thesis, fire hazard is increased (Kruger 1981; Van Wilgen & Richard­ Univ. of Stellenbosch. son 1985). We suggest that these areas be cleared of dense MITCHELL, D.T., COLEY, P.G.F., WEBB, S. & ALLSOPP, N. Litter fall and decomposition processes in the coastal fynbos infestations, but that this work should proceed at a lower vegetation, south-western Cape, South Africa. J. Ecol. (in priority than the clearing of sparse or moderate infestations, press). to allow for restoration. Restoration in these areas must MUELLER-DOMBOIS, D. & ELLENBERG, H. 1974. Aims and concentrate on the re-establishment of shrubby species. methods of vegetation ecology. p. 547, Wiley, New York. RAUNKAIER, C. 1934. The life forms of plants and statistical Acknowledgements plant geography. p. 632, Clarendon Press, Oxford. We thank F.J. Kruger for suggesting that this study be RYCROFT, H.B. 1950. A quantitative ecological study of the undertaken and for useful comments on the manuscript. An vegetation of Biesievlei Catchment, Jonkershoek. Ph.D. thesis, anonymous referee provided valuable criticism which led to Univ. of Cape Town. SEAGRIEF, S.C. 1950. Studies on the plant ecology of Fern the improvement of the manuscript. The study forms part Kloof near Grahamstown. M.Sc. thesis, Univ. of South Africa. of the Conservation Forestry Programme of the Department VAN WILGEN, B.W. 1981. Some effects of fire frequency on of Environment Affairs (Forestry Branch). fynbos plant community structure at Jonkershoek, Stellenbosch. S. Afr. For. J. 118: 42-55. References VAN WILGEN, B.W. & RICHARDSON, D.M. 1985. The effects BOND, P. & GOLDBLATT, P. 1984. Plants of the Cape Flora of alien shrub invasions on vegetation structure and fire - A descriptive catologue. 11 S. Afr. Bot. Suppl. Vol. 13: behaviour in South African fynbos shrublands: A simulation 1- 455. study. J. appl. Ecol. 22: 955-966. BOND, W.J. 1980. Fire and senescent Fynbos in the Swartberg, VAN WYK, D.B. 1977. Die invloed van bebossing met Pinus southern Cape. S. Afr. For. J. 114: 68-71. radiata op die totale jaarlikse afvoer van die Jonkershoek COWLING, R.M., MOLL, E.J. & CAMPBELL, B.M. 1976. The strome. M.Sc. thesis, Univ. of Stellenbosch. ecological status of the understorey communities of pine forests VERSFELD, D.B. 1981. Litter fall and decomposition in stands of on Table Mountain. S. Afr. For. J. 99: 13-23. mature Pinus radiata. S. Afr. For. J. 116: 40- 50.