184 S.-Afr. Tydskr. Dierk. 1988,23(3) species richness of and forest ecosystems in the southern Cape

J.H. Koen* and W. Breytenbach South African Forestry Research Institute, Saasveld Forestry Research Centre, Private Bag X6515. George. 6530 Republic of South . Africa

Received 26 June 1987; accepted 16 March 1988

The ant fauna in fynbos and forest habitats in the southern Cape are compared. There is no significant difference in ant species richness between the two undisturbed habitat types. and the only two species common to both are Acantholepis capensis and Camponotus maculatus. The degree of Hakea sericea infestation in the fynbos negatively affected ant species richness. Ant species richness in the Afromontane forest decreased from the drier to the wetter sites. The aggressive alien ant species. lridomynnex humilis, has not invaded natural areas to any great extent.

Die mierfauna in fynbos- en inheemse woud-habitatte in die Suid-Kaap is vergelyk. Daar is geen beduidende verskille in mierspesieverskeidenheid tussen die twee onversteurde habitat-tipes nie. Slags twee spesies, Acantholepis capensis en Camponotus maculatus is in albei habitatte aangetref. Die mate van Hakea sericea besmetting in die fynbos verminder mierspesieverskeidenheid. Mierspesieverskeidenheid in die Afrikaanse berQwoud het 'n afname getoon van die droeer na die natter persele. Daar is nag geen beduidende indringing van die aggresiewe uitheemse mierspesie. lridomynnex humilis, in natuurlike habitatte nie.

·To whom correspondence should be addressed

The ant fauna of the western Cape, South Africa, is fairly 2 = 35% and Site 3 = 100% H. sericea cover infestation) well known (Slingsby & Bond 1981; de Kock 1984; at 570 and 577 m a.s.1. respectively. Rainfall in the area is Donnelly & Giliomee 1985a, b). No information is, non-seasonal with a mean of 800 mm per annum. Climatic however, available on the species composition of the extremes are experienced in the area with periods of hot, of natural and disturbed ecosystems of the southern Cape. dry berg winds or gale force winds, rain and very low Various research projects are in progress in the fynbos and temperatures with snow on the higher peaks. The soils in

.

) forest habitats of the southern Cape which aim to the Waboomskraal Valley are derived from Table Moun­ 0

1 determine and evaluate management practices, and to tain sandstone and are generally shallow to moderately 0

2 determine the relationship between faunal communities deep (G. Schafer, pers. comm.). .

d (vertebrate and invertebrate) and environmental factors. The forest study area is in evergreen high forest (Afro­ e t

a Infestation by alien vegetation (mainly Hakea sericea, an montane), 30 km to the north of Knysna, generally known d (

Australian Proteaceae) in the fynbos of the Outeniqua as the Knysna Forest (Acocks 1953). In general, the r e Mountain range is causing serious concern and the impact canopy is 20-25 m high with the following common h s i of H. sericea on the species composition (fauna and flora) species: Podocarpus falcatus, P. lalifolius, Pterocelastrus l b of the natural communities is under investigation. The tricuspidatus, Ocotea bullata, Olea capensis macrocarpa, u

P presence of alien plants (e.g. Pinus radiata) have been Rapanea melanophloeos, Apodytes dimidiata, Gonioma e

h reported to affect species diversity of ants in the western kamossi and Nuxia floribunda. Rainfall is alSo non­ t Cape (Donnelly & Giliomee 1985b). y seasonal with a mean of 1100 mm. Climatic conditions b Data on the ant communities have been gathered similar to Waboomskraal prevail in the forest but without d e during general surveys and are presented here t cold periods associated with snow. The soils in the area n to supply information on the distribution of ants in the a are variable and the underlying geology is Table Mountain r

g coastal mountain range and in the evergreen high forests

sandstone (G. Schafer, pers. comm.). e c on the plateau to the south of this range. All identified ant Three study sites were selected along a soil moisture n e specimens are lodged in the Saasveld Forestry Museum.

c gradient to represent the spectrum of forest types in the i l area. Although not directly measured, soil moisture r e Study area

d content affects the species composition of the associated n The fynbos study area IS to the Waboomskraal Valley vegetation (Koen & Crowe 1987). Sites 1 and 2 were in the u

y (33°51'S I 22°lO'E) on the north-facing slopes of the Lily Vlei Nature Reserve of the Gouna State Forest a

w Outeniqua Mountains, 30 km north-west of George. (33°56'S/23002'E), and Site 3 was on the DiepwaHe e t Fynbos communities were dominated by Proteaceae, State Forest (33°56'S I 23009' E). Site 1 (dry forest) a G

Ericaceae, Restionaceae and Cyperaceae. The most has fairly dense vegetation with an understorey of tree t e common species include Protea repens, P. lorifolia, regeneration and thorny shrubs. Site 2 (moist forest) has a n i

b Leucadendron salignum, L. eucalyplifolium and dense understorey of 2-5 m high Trichocladus crinitus and a

S Hypodiscus striatus. Three sites were chosen on a H. trees with rounded crowns. Site 3 (wet forest) consists of y sericea infestation gradient. The uninfested, mature (18- straight-stemmed and slender-crowned trees with an b

d year-old) fynbos (Site 1) is at a slightly higher elevation understorey of ferns, especially Cyathea capensis (Koen & e c (645 m a.s.l.) than the two H. sericea infested sites (Site Crowe 1987). u d o r p e R S. Afr. J. Zoo!. 1988, 23(3) 185

Methods Table 2 Formicidae collected in two areas in the Data on arthropod communities were gathered during two southern Cape with presence (+) and absence (-) indi­ separate studies: one on the effects of H. sericea on the cated community structure of plants and , and the other Fynbos on the changes in the community structure of birds and Forest insects between different forest types. Although different 51 52 53 collection methods were employed, the data is sufficient 0%· 35%· 100%· 51 52 53 to make gross comparisons between the ant communities in the two habitats. Dotichoderinae were sampled nine times during the period lridomyrmex humi/is + May 1983 to March 1984 (fynbos) and monthly from April Technomyrmex a/bipes + 1983 to March 1984 (forest) using 150 mm wide pitfall Fonnldnae traps with vehicle anti-freeze as preservative. Although Anop/o/epis custodiens + + + Marsh (1984) found pitfall trapping to be an unreliable Acantho/epis capensis + + + + way of determining relative abundances of ant species in Camponotus macu/atus + + + + + the Namib Desert, Samways (1983), de Kock (1984), Camponotus niveosetosus + Andersen (1986a, b) and Marsh (1986) found it a reliable Camponotus rufog/aucus + + method to determine species richness in various habitats. Camponotus (sp. nova) + We used 12 traps in each of the fynbos sites, set 10 m apart Camponotus sp. 2 + Plagio/epis decora + + on a contour running through the length of the study area. Myrmldnae Crematogaster Jiengmei + + 22r------, Crematogaster peringueyi + + + Me1'8nopJus peringueyi + + + N o Monomorium de/agoensis + + + o 17 Monomorium austra/e + F Monomorium S P springvaJense + E C 12 Monomorium sp. 1 + I Monomorium sp. 2 + . E ) S 0 Monomorium sp. 3 + 1

0 7 Ocymyrmex barbiger + 2 Pheidole capensis + d e

t Pheidole fore/i + a 2L-----~----~----~-----L----~----~--~ Pheido/e sp. 2

d + + ( o 2 4 6 B 10 12 So/enopsis punctaticeps + r OBSERVATION PERIOD e Strumigenys havi/andi h --a-- rOREST -+ IT/mos + + s i l SyJ/aphopsis arno/di +

b Figure I Cumulative species diverSity curves for fynbos and SyJ/aphopsis sp. 1 u + forest habitats in the southern Cape. P

Tetramorium ackermani + e

h Tetramorium de/agoensis + t

y Tetramorium grassii + + + b Table 1 Number of ant species caught per bait type for Tetramorium sp. 1 + d e a one-year period in the fynbos study site 2 t Tetramorium sp. + n Tetramorium sp. 3 + a Banana! r

g Tetramorium sp. 4 + + rum Mean! e Gen. nov. (near Pheido/e) + c Date Meat Faeces mixture Unbaited trap n e Ponerinae c i l 05/1983 5 6 4 3,75 o Discothyrea poweri -' + + r e 0611983 5 7 7 1 5 Hypopoilera nata/ensis + d n 0811983 3 10 9 8 7,5 Hypoponera u 09/1983 3 10 4 3 5 punctatissima y + a 1011983 12 12 9 3 9 Leptogenys attenuata + w e t 11/1983 9 6 10 8 8,25 Leptogenys nitida + a 01/1984 8 7 10 13 9,5 Trachymesopus G t wroughtoni + + + e 0211984 13 8 10 7 9,5 n i 0311984 8 4 11 13 9 Pseudomyrmldnae b a Tetraponera emeryi + S Mean 7,3 7,7 7,7 6,6 7,3 y

b Total 21 13 6 13 11 4

d 5td dev. of mean no. species 1 bait type = O,5J. e c 5td dev. of mean no. species 1 month = 2,2. ·Percentage Hakes present. u d o r p e R 186 S.-Afr. Tydskr. Dierk. 1988,23(3)

The traps remained in situ for a 3O-day period/month. Table 4 Ant species richness in various 'natural' and Twenty-five traps were used in each of the forest sites on a disturbed habitats. - = 'natural' sites, * = disturbed sites 25 x 25 m grid and the trapping continued for a 4-7-day period per month. Although more days trapping was done Habitat and Locality Richness Source in the fynbos, we feel confident that most species were Heathland·· 22 Andersen 1986a sampled in the forest as no additional species were added S.E. Australia (Heath land) to the cumulative total after 8 months sampling (Figure 1). Various -baits (banana/rum mixture, rotten pork) were Woodland (Mallee)·· 47 Andersen 1986b used but were discontinued in the forest after 6 months S.E. Australia because they made no difference in species captured. Mature fynbos·· 21 This paper Traps baited with similar baits (human faeces instead of S. Cape rotten pork) were used throughout the trapping period in South Africa the fynbos but did not seem to affect the number of Mature fynbos·· 14 Donnelly & Giliomee species sampled (Table 1). Detailed analysis of the W. Cape 1985 changes in species composition with bait type in the fynbos South Africa will be reported on in a subsequent paper. Namib Desert·· 27 Marsh 1986 Results South West Africa A total of 27 species was collected in the fynbos (H. sericea Dry Afromontane·· 13 This paper infested sites included) and 18 in the forest, of which only forest. S. Cape South Africa two species were shared (Table 2). Accordi~g to de Kock (1984) and Prins (pers. comm.) most of the ants collected Moist Afromontane·· 11 This paper in the fynbos are 'elaiosome hunters'. Their seed­ forest. S. Cape harvesting activities are directed towards seeds that have South Africa elaiosomes (fatty tissues). These ants do not destroy the Riparian forest·· 14 van Schagen 1986 seed & itself, but effectively disperse the propagule (Bond S.W. Australia Slingsby 1983). Four of the species collected (Crematogas­ ter liengmei, C. peringueyi, Acantholepis capensis and Upland forest·· 13 van Schagen 1986 Leptogenys nitida) do not collect seeds with elaiosomes. S.W. Australia

.

) C. peringueyi and possibly C. liengmei are omnivorous Wet Afromontane·· 4 This paper 0

1 and, amongst other things, remove the elaiosomes and forest. S. Cape 0 2

discard the seed on the soil surface (Bond & Breytenbach South Africa d e 1985). A. capensis is aphidicolous and L. nitida is a t Mature fynbos· 6 This paper a predator. d S. Cape (100% II. seriCea (

r Nine of the species collected in the forest could be South Africa infestation ) e

h considered 'elaiosome hunters' or general seed gatherers s i Mature fynbos· 13 This paper l (Tetramorium spp., Monomorium sp., Crematogaster sp., b S. Cape (35% II. seriCea u Plagiolepis sp., Syl/aphopsis spp. and Camponotus sp.). P

South Africa infestation)

e The other species are mostly carnivorous (A.J. Prins pers. h t comm.). Pine plantation· 9 Donnelly & Giliomee y Western Cape (Pinus plantation) 1985 b

d South Africa e Table 3 Monthly ant species richness in three forest t n sites Rehabilitated· 10 van Schagen 1986 a r

g coal mines

e Site 1 Site 2 Site 3 Total

c S.W. Australia n Month Dry forest Moist forest Wet forest species e Sand Mine· 6 Majer et al. 1982 in c i l 0 Eneabba (Heathland) Majer 1984 r January 6 0 6 e W. Australia d February 3 0 2 -4 n

u March 3 2 2 3

Bauxite Mine· 10 Majer et aJ. 1982 in y April 3 6 2 6 a Jarrahdale (Heathland) Majer 1984 w May 7 2 2 3 e Del Park, W. Austr. t

a June 0 0 0 0 Sand Mines· 11 Fox & Fox 1984 G

July 3 4 3 7 t (Forest with heath- e Myall Lakes in Majer 1984 August 1 4 0 4 n i N.S.W. land understorey) b September 5 2 2 6 a

S Citrus orchards· 49 Samways 1983

October 1 0 y E. Transvaal (Grassland) b November 1 0 0 1

d South Africa

e December 0 0 0 0 c u d o r p e R S. Afr. l. Zoo!. 1988,23(3) 187

As found by Majer (1984) in a similar climatic zone in canopy cover increases from the dry to the wet sites. As Australia, there was also a seasonal trend in the fynbos this study was not structured to collect tree-living ants, with lower ant species richness from May to September the possible increase in community importance of the (winter) (Table 1). Although the general arthropod survey latter in the wetter sites could not be demonstrated. revealed a distinct se3{iOnal pattern in the forest (Koen & As in the H. sericea infested sites where there is an Crowe 1987), no seasonality was apparent in the ant fauna inverse relationship between the degree of infestation (Table 3). and the plant species richness in the lower layers, there is Species richness in the fynbos and forest ant faunas a similar decrease in plant species richness in the lower were compared to that of other habitat types in the world layer from the dry to the wet forest types (unpublished (Table 4). As indicated, habitat disturbance decreases the data). In both areas, the ant fauna in the site with the number of ant species. Such disturbances are usually lowest plant species richness is characterized by tree- associated with physical change in the environment or foraging/aphidicolous species. . with the introduction of alien plant species into an The habitats associated with citrus orchards (Sam ways ecosystem, especially when such species tend to QCCur in 1983) and Mallee woodland (Andersen 1986b) have mono-specific stands. The species richness found by similarly high species diversities. This could be a Samways (1983) is higher than in the other disturbed sites consequence of a more diverse, but superficially similar (Table 4). This could be the result of sampling in 11 vegetation structure offering more foraging opportuni­ different habitats associated with citrus orchards. ties. Iridomyrmex humilis, an ~ggressive alien, has already Discussion been collected in various natural or disturbed sites in the Plant species richness in the two undisturbed vegetation western Cape (Bond & Slingsby 1984; de Kock 1984; types are fairly similar (fynbos = 47; forest = 63), and Donnelly & Giliomee 1985a, b). Concern has been there was no apparent difference in ant species richness expressed at the possible ecological effect that this between the two habitats (fynbos 21; forest 18). There species may have on the reproductive capacity of the was, however, a major difference in ant species vegetation (Bond & Slingsby 1984). It effectively composition between the areas with only two species in excludes seed-dispersing species from an area, and by co.mmon (Acantholepis capensis and Camponotus not burying the seeds, the seeds are subjected to a high maculatus). Both species are widespread in southern level of rodent predation (Bond & Breytenbach 1985). This species has fortunately not been found in any of the Africa and do not appear to be habitat specific (Skaife

fynbos sites during this study and only in one of the . 1961; Samways 1983; Donnelly & Giliomee 1985a, b). )

0 forest sites (one specimen only). As this species is

1 Species in the subfamily Ponerinae were restricted

0 usually unknowingly dispersed by man it is probable that

2 mainly to the forest sites with only one species in the it was introduced to the forest during nonnal d fynbos. This subfamily contains the most primitive e t management operations. Although collection work was a southern African ants (Scholtz & Holm 1985). d restricted to two areas it does not appear as though I. (

Ant species richness in both Australian heathland and r humilis has invaded natural areas in the southern Cape, e mature southern Cape fynbos are very similar. However, h to any great extent. s i mature fynbos in the western Cape has a much lower l This study supplies some infonnation on the species b species richness (Table 4). This could be a consequence u diversity of the ant faunas of the southern Cape, but P

of the effects of the hot, dry summers in the latter on ant e surveys conducted in more sites are needed to indicate h behaviour. The ant species richness of the H. sericea t

the geographical distribution and habitat requirements y infested sites (35% infestation = 13; 100% infesta­ b of the different species. The extension in the range of the tion = 6) has an inverse relationship to the degree of d

e Argentine ant should also be closely monitored. t infestation. This illustrates the important negative n a influence of this aggressive alien plant, not only on the r Acknowledgements g

vegetation, but also on the smaller invertebrates. Four of e c the species sampled in the 100% H. sericea infested site This study was carried out as part of the conservation n e research programme of the Forestry Branch,

c (Pheidole foreli, Monomorium delagoensis, Meranoplus i l peringueyi and Anoplolepis custodiens) forage in trees or Department of Environment Affairs. A.J. Prins kindly r e are aphidicolous (Samways 1983; A.J. Prins pers. identified the ants and G.J. Breytenbach, J. Midgley and d n comm.). The lack of elaiosome-bearing seeds in the W.J. Bond commented on an earlier draft. J. Sass and u A. y infested areas probably precludes the presence of seed­ Hendricks assisted with field work. a

w gathering ants. e t The ant species richness in the three forest sites has an References a

G So~th inverse relationship to the soil moisture content of the ACOCKS, l.P.H. 1953. Veld types of Africa. Mem. t e sites, with a three-fold difference between the dry and Bot. Surv. S.A. 28. Government Printer, Pretoria. n i

b the wet sites (dry = 13; wet = 4). Although this could ANDERSEN, A.N. 1986a. Patterns of ant community a

S possibly indicate a substrate specificity in relation to soil organization in mesic south-eastern Australia. Austr. J. y moisture content or the associated vegetation, other Ecol. 11: 87-97. b

d microclimatic factors such as temperature may play an ANDERSEN, A.N. 1986b. Diversity, seasonality and e c important role. The ant species richness decreases as community organization of ants of adjacent heath and u d o r p e R 188 S.-Afr. Tydskr. Dierk. 1988,23(3)

woodland sites in south-eastern Australia. Awtr. 1. Zool. invertebrates. Oecologia (Berlin) 72: 414-422. 34: 53--64. MAJER, J.D. 1984. Ant return in rehabilitated mines - an BOND, W. & SLINGSBY, P. 1983. Seed dispersal by ants indicator of ecosystem resilience. In: Proceedings of the in the shrublands of the Cape Province and its 4th International Conference on Mediterranean evolutionary implications. S. Afr.l. Sci. 79(6): 231-233. Ecosystems, (ed.) Bell, B., University of Western BOND, W. & SLINGSBY, P. 1984. Collapse of an ant­ Australia. plant mutualism: The Argentine ant (lridomyrmex. MARSH, A.c. 1984. The efficacy of pitfall traps for humil~) and myrmecochorous Proteaceae. Ecology 65: determining the structure of a desert ant community. 1. 1O~1037. Ent. Soc. sth. Afr. 47(1): 115-120. BOND, W. & BREYTENBACH, G.J. 1985. Ants, rodents MARSH, A.C. 1986. Ant species richness along a climatic and seed predation in Proteaceae. S. Afr. 1. Zool. 20(3): gradient in the Namib Desert. 1. Arid. Environ. 11: 150-154. 235-241. DE KOCK, A.E. 1984. Verslag oor Fynbos-Bioom-Projek: SAMWA YS, M.J. 1983. Community structure of ants Argentynse mier-opname. Unpublished report. (: Formicidae) in a series of habitats DONNELLY, D. & GILiOMEE, J.H. 1985a. Community associated with citrus. 1. Appl. Ecol.20: 833-847. structure of epigaeic ants (Hymenoptera: Formicidae) in SCHOLTZ, C.H. & HOLM, E. 1985. Insects of Southern fynbos vegetation in the Jonkershoek Valley. 1. Ent. Soc. Africa. Butterworth Publishers, London. sth. Afr. 48(2): 247-257. SKAIFE, S.H. 1961. The study of ants. Longmans, London. DONNELLY, D. & GILiOMEE, J.H. 1985b. Community SLlNGSBY, P. & BOND, W.J. 1981. Ants - friends of the structure of epigaeic ants in a pine plantation and in fynbos. Veld & Flora 67: 39-45. newly burnt fynbos. 1. Ent. Soc. sth. Afr. 48(2): 259-265. VAN SCHAGEN, J. 1986. Recolonisation by ants and other KOEN, J.H. & CROWE, T.M. 1987. -habitat invertebrates in rehabilitated coal mine sites near Collie, relationships in the Knysna Forest, South Africa: Western Australia. Western Australian Institute of Discrimination between forest types by birds and Technology, School of Biology. Bull. 13.

. ) 0 1 0 2 d e t a d ( r e h s i l b u P e h t y b d e t n a r g e c n e c i l r e d n u y a w e t a G t e n i b a S y b d e c u d o r p e R