498 South African Journal of Science 97, November/December 2001 Serpentine Ecology

tine flora is a response to the toxicity of particular metals, in particular Ni1,5 and Mg.2,11 In this study the occurrence of a large number of endemic species at upland xeric sites with high surface soil Ni (>0.3%) supports the Ni and drought hypothesis (Table 1, Fig. 2). Kruckeberg7 states that high Ni concentration in soils reduces species richness through biotype depletion. The reduction in species richness with the increased Ni in Central Queensland soils support this rationale. For example, Soil Group 3 with the highest Ni concentrations recorded on average 11 fewer species than the lower Ni (<0.3%) soils of groups 1, 2, 4, 5 and 6. The predominance of serpentine endemic shrubs (13 of 19 taxa) in this study agrees with Borhidi’s2 hypothesis that such life forms are more frequent as initial colonizers of primary Fig. 4. Comparison of species richness of open-forest strata between upland (in in situ serpentine soils. situ soils) and lowland (depositional soils) serpentine landforms of Central The total number of endemics found in the uplands (23 spp.) is Queensland. substantially greater than that found in the lowlands (15 spp.). This suggests that the predominant evolutionary processes lead- 30% of the total serpentine flora by covering <0.1% of the study ing to speciation are most likely to occur in upland, older in situ area. However, no serpentine endemic have yet been re- soils. The features of upland landforms that potentially promote corded from this community.11 the evolution of serpentine endemics include low fertility, high Ni, high Mg, a predominance of shallow soils magnifying physi- Discussion ological aridity, and a long period of landform evolution. Some Serpentine landforms in Central Queensland are character- endemic taxa that presumably originated on the uplands may ized by endemic open-forest communities in both the uplands have migrated with the movement of younger depositional soils and lowlands (Appendix 1). The distinctiveness of these land- down to the lowlands (e.g. Bursaria reevesii, Capparis thozetiana, forms has been attributed to abiotic influences such as high soil Pimelea leptospermoides, Stackhousia tryonii). Mg and Ni as well as biotic responses including physiological There is greater structural diversity, particularly in canopy drought and the evolution of an unique flora and vegeta- height, within upland plant communities than lowland plant tion.11,18–23 For example, Stackhousia tryonii and Pimelea lepto- communities (Appendix 1). Figure 5 summarizes upland envi- spermoides have evolved a specific physiological condition of ronmental conditions that are due to landform factors such as accumulating Ni.18–20 topography, aspect and elevation. Variability in soil chemistry, Brooks and others1,2,5,26 have suggested that serpentine floras water availability and fire regimes within the uplands may are subjected to two dominant selective pressures: (i) selection contribute to the greater structural diversity of these plant for Ni tolerance and (ii) selection for drought tolerance. Proctor4 communities. The abundance of heathy, sclerophyllous and others1,5,8 argue that a major force in the evolution of serpen- understorey in the uplands appears to be associated with high

Fig. 5. Illustration of four plant communities along a north–south gradient on serpentine soils at Mt Fairview, Central Queensland. Serpentine Ecology South African Journal of Science 97, November/December 2001 499

soil Ni concentrations (Figs 2, 4, 5). However,open forests in both alien and endemic herbs on serpentine. Ecology 80, 70–80. the uplands and lowlands have a localized abundance of heathy 7. Kruckeberg A.R. (1954). The ecology of serpentine soils. III. Plant species in relation to serpentine soils. Ecology 35, 267–274. understorey of species such as Leucopogon spp., Macrozamia sp. 8. Rune O. (1953). Plant life on serpentine and related rocks in the north of nov., Melaleuca bracteata, Pimelea leptospermoides and Xanthorrhoea Sweden. Acta Phytogeogr. Suec. 36, 1–139. spp. (Appendix 1). The abundance of shrubby species may re- 9. Whittaker R.H. (1954). The vegetation response to serpentine soils. Ecology 35, flect a more general nature of serpentine landform.11,19–22 275–288. The total species richness of lowland open forests is greater 10. Wild H. (1965). The flora of the Great Dyke of Southern Rhodesia with special than that of upland open forests, mainly because of a higher references to the serpentine soils. Kirkia 5, 49–86. 11. Batianoff G N., Neldner V.J. and Singh S. (2000). census and number of herbaceous species (Table 1, Fig. 4). The reduction in floristic analysis of serpentine landscapes in Central Queensland. Proc. R.. Soc. total serpentine herb diversity in uplands relative to lowlands Qld. 109, 1–30. corresponds to previous reports on the effect of elevation on 12. Cole M.M. (1973). Geobotanical and biogeochemical investigations in the serpentine flora.6 In Central Queensland11,19–22 and other areas of sclerophyllous woodland and shrub associations of the eastern goldfields area of Western , with particular reference to the role of Hybanthus 12,14–16,19 Australia, the abundance of grasses (particularly Triodia floribundus (Lindl.) F.Muell. as a nickel indicator and accumulator plant. J. appl. spp.) and other species of herbs is possibly due to their tolerance Ecol. 10, 269–320. of high Mg and/or lower Ca concentrations at serpentine 13. Severne B.C. and Brooks R.R. (1972). A nickel-accumulating plant from West- sites.6,10,26 Harrison6 found that soils low in Ca (and high in Mg) ern Australia. Planta 103, 91–94. 14. Lyons M.T., Brooks R.R. and Craig D.C. (1974). The influence of soil composi- harbour a larger number of native herbs. In addition to herba- tion on the vegetation of the Coolac Serpentine Belt in New South Wales. J. R. ceous plants, two serpentine woody endemic species, Neoroepera Soc. N.S. W. 107, 67–75. buxifolia and Callistemon sp. nov., are also reported to tolerate 15. Davie H. and Benson J.S. (1997). The serpentine vegetation of the soils with high Mg concentrations and/or a high Mg:Ca ratio.11 Woko-Glenrock region, New South Wales,Australia. In The Ecology of Ultramafic and Metalliferous Areas, eds T. Jaffré, R.D. Reeves and T. Becquer, pp. 155–162. Studies of plant–soil associations suggest that soil Ni and Mg Centre ORSTOM. Nouméa, New Caledonia. concentrations influence species composition and vegetation 16. Specht A., Forth F.and Steenbeeke G. (2001). The effect of serpentine on vegeta- structure on serpentine soils. This study has demonstrated that tion structure, composition and endemism in northern NSW,Australia. S. Afr. J. the differences between upland and lowland endemism and Sci. 97, 521–529. 17. Gibson N., Brown M.J., Williams K. and Brown A. V.(1992). Flora and vegeta- species richness are directly associated with soil chemistry.How- tion of ultramafic areas in Tasmania. Aust. J. Ecol. 17, 297–303. ever, serpentine species composition may also be influenced by 18. Batianoff G.N., Reeves R.D. and Specht R.L. (1990). Stackhousia tryonii Bailey. A other abiotic conditions. Collaboration between soil scientists, nickel-accumulating serpentinite endemic species of central Queensland. Aust. plant physiologists and ecologists is needed to examine further J. Bot. 38, 121–130. the links between endemism, species richness and soil-landform 19. Batianoff G.N. and Specht R.L. (1992). Queensland (Australia) serpentine vege- tation. In The Vegetation of Ultramafic (Serpentine) Soils, eds A.J.M. Baker, J. Proc- characteristics. tor and R.D. Reeves, pp. 109–128. Intercept, Andover, Hants. Queensland Herbarium staff, particularly John Neldner, are acknowledged for 20. Batianoff G.N., Reeves R.D. and Specht R.L. (1997). The effects of serpentine supporting this project. Taxonomic guidance was provided by Anthony Bean and on vegetation structure, species diversity and endemism in Central Queensland. In The Ecology of Ultramafic and Metalliferous Areas, eds T. Jaffré, Paul Forster and soil descriptions by Bruce A. Forster. Will Smith drafted Figs 1 R.D. Reeves and T. Becquer, pp. 147–154. Centre ORSTOM, Nouméa, New and 5. G.N.B. is grateful to Ray Specht for valuable scientific advice in early stages of Caledonia. writing the paper. The constructive critique provided by Ailsa Holland, Andrew 21. Batianoff G.N., Specht R.L. and Reeves R.D. (1991). The serpentinite flora of the Franks and anonymous referees considerably improved this document. humid subtropics of eastern Australia. Proc. R. Soc. Qld. 101, 137–157. 22. Forster B.A. and Baker D.E. (1997). Characterisation of the serpentine soils of 1. Arianoutsou M., Rundel P.W. and Berry W.L. (1993). Serpentine endemics as Central Queensland, Australia. In The Ecology of Ultramafic and Metalliferous biological indicators of soil elemental concentrations. In Plants as Biomonitors: Areas, eds T. Jaffré, R.D. Reeves and T. Becquer, pp. 27–37. Centre ORSTOM, Indicators for Heavy Metals in the Terrestrial Environment, ed. B. Markert, Nouméa, New Caledonia. pp. 179–189. VCH, Weinheim. 23. Plowman K. (2000). Capricornia (Central Queensland) Serpentinite Landscape 2. Borhidi A. (1991). Phytogeography and Vegetation Ecology of Cuba. Akademiai Nomination. Southern Queensland Natural Environment Evaluation Panel Kiado, Budapest. Conservation Nomination for Register of National Estate, Australian Heritage 3. Proctor J. (1971). The plant ecology of serpentine. II. Plant response to serpen- Commission, Canberra. (Unpubl. rep.) tine soils. J. Ecol. 59, 397–410. 24. Specht R.L. (1970). Vegetation. In The Australian Environment, ed. G.W. 4. Proctor J. (1971). The plant ecology of serpentine III. The influence of a high Leeper, pp. 43–67. CSIRO Australia and Melbourne University Press, magnesium/calcium ratio and high nickel and chromium levels in some British Melbourne. and Swedish serpentine soils. J. Ecol. 59, 827–842. 25. Webb L.J. and Tracey J.G. (1981). Australian rainforests: patterns and change. In 5. Brooks R.R. (1987). Serpentine and its Vegetation: A Multidisciplinary Approach. Ecological Biogeography in Australia, ed. A.J. Keast, pp. 605–694. W. Junk, The Ecology, Phytogeography & Physiology Series, vol. 1, ed. T.R. Dudley. Hague. Dioscorides Press, Portland, Oregon. 26. Wild H. and Bradshaw A.D. (1977). The evolutionary effects of metalliferous 6. Harrison S. (1999). Local and regional diversity in a patchy landscape: native, and other anomalous soils in south-central Africa. Evolution 31, 282–293.

Appendix 1. Vegetation and soil description of Central Queensland serpentine landforms.

Plant communities Soils ( B.A. Forster, pers. comm., 1999).

Tallopen-forest: Overstorey (30±5m;FPC20–35%): Corymbia citriodora, Euca- Group 1: Haplustox in situ soils are formed from a saprolite parent material in lyptus portuensis, E. fibrosa subsp. nov.;Mid-understorey (6 ± 4 m; FPC 25–40%): the upper mountain/high hill slopes, which is derived from highly weathered spp., Psychotria daphnoides, rainforest elements (Diospyros geminata, ultramafic rocks. Red clay loams and clays with ferrous and manganiferous Drypetes deplanchei, Neolitsea brassii, Polyscias elegans); Understorey (2 ± 1.9 m; nodules also occur throughout the soil profile. Estimated 6% of the total FPC 20–40%): Brunoniella spp., Gahnia aspera, Grewia latifolia, Lissanthe sp. nov. serpentine soil in Central Queensland, primarily occurring in the upland areas Macrozamia sp. nov., Pimelea leptospermoides, Xanthorrhoea johnsonii. of Marlborough land system with smaller amounts in Tungamull land system. Diagnostic chemical properties at 0–30 cm depth: Mg:Ca = 0.4–1.0; range Open-forest: Overstorey (18±5m;FPC30–40%): Corymbia spp., (mean). Ni concentration = 0.06–0.27% (0.16%). fibrosa subsp. nov.; Mid-understorey (6±4m;FPC10–20%): Acacia spp., Petalostigma pubescens, Psychotria daphnoides, rainforest elements; Understorey (2 ± 1.5 m; FPC 20–40%): Aristida spp., Gahnia aspera, Lomandra spp., Macrozamia sp. nov., Panicum effusum, Pimelea leptospermoides, Triodia mitchellii. 500 South African Journal of Science 97, November/December 2001 Serpentine Ecology

Tall open-forest/Open-forest: Overstorey (20±5m;FPC30–40%): Corymbia Group 2a: Ultisol/Alfisol/Inceptisol in situ soils occur on moderately weathered spp (mostly C. xanthope), Eucalyptus crebra, E. fibrosa subsp. nov; serpentine material on the mid to upper hillslopes. The stony red to brown clay Mid-understorey (6 ± 4 m; FPC 10–20%): Acacia spp., Alphitonia excelsa, loam or clay contains fewer ferrous and manganiferous nodules than Group 1 Brachychiton bidwillii, Corymbia spp., Psychotria daphnoides, rainforest elements; soils but have a higher proportion of moderately to strongly weathered ser- Understorey (2 ± 1.5 m; FPC 20–50%): Cymbopogon spp., Gahnia aspera, pentine gravels and stone, often with a redeposited silica lattice. Estimated Lomandra spp., Leucopogon spp. Macrozamia sp. nov., Panicum effusum, Pimelea 14% of the total serpentine soil in Central Queensland, occurring on uplands of leptospermoides, Themeda triandra, Tragia novae-hollandiae, Triodia mitchellii, the Marlborough/Tungamull land systems. Diagnostic chemical properties at Xanthorrhoea johnsonii. 0–30 cm depth: Mg:Ca = 5–8; range (mean). Ni concentration = 0.22–0.83% (0.27%).

Semi-deciduous vine thicket (closed-forest): Emergent (18±4m;FPC Group 2b: Alfisol/Inceptisol in situ soils occur moderately weathered serpen- 5–10%): Brachychiton rupestris, Flindersia spp., Gyrocarpus americanus; tine material on the high and lowe slopes and footslopes (and a low rise on Overstorey (10±2m;FPC30–40%): Acacia fasciculifera, Austromyrtus bidwillii, ‘Balmoral’). The stony red to brown clay loam or clay lacks ferrous and Backhousia kingii, Owenia venosa; Mid-understorey (4 ±2 m; FPC 30–35%): manganiferous nodules. Mesic sites due to increased moisture (Fig. 5). Canthium spp., Croton insularis, Diospyros humilis; Vines: Cissus spp., Melodorum Estimated 0.1% of the total serpentine soil in Central Queensland, occurring on leichhardtii; Understorey (1.5 ± 1 m; FPC 5–10%): ruscifolia, Carissa ovata, the uplands of the Marlborough/Tungamull land systems. Diagnostic chemical Turraea pubescens. properties at 0–30 cm depth: Mg:Ca = 5–8; range (mean). Ni concentration = 0.22–0.83% (0.27%) Gully vine thicket (low closed-forest): Overstorey (10 ± 2 m; FPC 55–70%): Alectryon spp., Austromyrtus bidwillii, Cryptocarya triplinervis, Neoroepera buxifolia, Niemeyera antiloga, Notelaea longifolia; Vines: Cissus oblonga, Melodorum leichhardtii; Mid-understorey (4 ±2 m; FPC 15–35%: Acronychia laevis, Canthium coprosmoides, Diospyros geminata; Understorey (2 ± 1.6 m; FPC 1–5%): Alyxia ruscifolia, Cupaniopsis wadsworthii, Gahnia aspera. Simple microphyll vine thicket (closed scrub): Emergent (12 m; FPC 5–10%): Eucalyptus fibrosa subsp. nov.; Overstorey (6±1m;FPC65%): Drypetes deplanchei, Mallotus philippensis, Niemeyera antiloga, Notelaea longifolia, Psychotria daphnoides; Vines: Pandorea pandorana. Understorey (0.1 ± 0.4 m; FPC 5%): Adiantum aethiopicum, Alyxia ruscifolia, Commelina spp., Gahnia aspera.

Open-forest/Low woodland (heathy and/or grassy understorey): Overstorey Group 3: Haplustoll in situ soils overlie partially weathered serpentine in the (14±6m;FPC15–30%): Corymbia xanthope, Eucalyptus fibrosa subsp. nov.; low hills and footslopes (Fig. 5). These soils are shallow black or dark brown Mid-understorey (6±2m;FPC20–40%): Acacia sp. nov. (Canoona), Xanthor- stony clays. These younger, upland xeric sites occupy approximately 40% of rhoea johnsonii; Understorey (2 ± 1.5 m; FPC 10–20%): Bursaria reevesii, Capparis the total serpentine soil in Central Queensland, occurring in the thozetiana, Chamaesyce ophiolitica, Cymbopogon spp. Macrozamia sp. nov., Tungamull/Marlborough land systems. Diagnostic chemical properties at Leucopogon cuspidatus, Olearia sp. nov, Panicum effusum, Pimelea leptospermoides, 0–30 cm depth: Mg:Ca = 4–11.5; range (mean) Ni concentration = 0.30–1.57% Pultenaea setulosa, Stackhousia tryonii, Triodia mitchellii. (0.6%).

Heathland/Grassland: Emergent (10 ± 2 m; FPC 5%): Corymbia xanthope; Group 3 (South Percy Island): Haplustoll in situ soils in the low hills and Overstorey (1.5±1m;FPC25–75%): Allocasuarina littoralis, Leucopogon spp., footslopes. These soils are shallow black or dark brown stony clays. There are Xanthorrhoea latifolia; Understorey (0.3 ± 0.2 m; FPC 5–15%): Lomandra spp., small laterized areas, possibly belonging to Group 2. The upland island soils are Panicum effusum, Stackhousia tryonii; Grassland (0.7 ± 0.3 m; FPC 35–75%): strongly influenced by maritime conditions, such as constant deposition of Aristida spp., Heteropogon contortus, Panicum spp., Themeda triandra. NaCl. Ni concentration unknown.

Open-forest (colluvial lowlands): Overstorey (20±5m;FPC20–3%): Group 4: Haplustoll depositional shallow to moderately deep gravelly black Corymbia xanthope, Eucalyptus tereticornis, E. fibrosa subsp. nov.; clays are formed in the footslopes and upper reaches of streams. Estimated 7% Mid-understorey (10±3m;FPC20–40%): Acacia spp., Melaleuca spp., Xanthor- of the total serpentine soil in Central Queensland, lowlands primarily occur- rhoea johnsonii; Understorey (2 ± 1.5 m; FPC 10–20%): Bursaria reevesii, Macro- ring in the Marlborough/Tungamull land systems with smaller areas in zamia sp. nov., Pimelea leptospermoides, Stackhousia tryonii, Themeda triandra. Kunwarara land system. Heterogeneous colluvial and alluvial soils, frequently intermediate between depositional and in situ soils. Diagnostic chemical Riverine open-forest: Overstorey (20 ± 5 m; FPC 20–35%): Callistemon sp. nov., properties at 0–30 cm depth: Mg:Ca = 2.5–5; range (mean). Ni concentration = Casuarina cunninghamiana, Melaleuca spp., Eucalyptus tereticornis; Understorey 0.05–0.50% (0.22%). (10±4m;FPC20–70%): Livistona decipiens, Melaleuca fluviatilis, Neoroepera buxifolia, Niemeyera antiloga; rainforest elements (Cryptocarya triplinervis, Neolitsea brassii, Terminalia porphyrocarpa).

Open-forest (alluvial plains): Overstorey (17±6m;FPC25–35%): Corymbia Group 5a: Haplustert deep depositional black and dark brown cracking clays spp., Eucalyptus fibrosa subsp. nov, E. tereticornis; Mid-understorey (5±2m; form in the alluvial plains and narrow flats adjacent to serpentine mountains FPC 15–20%): Acacia spp., Melaleuca spp., Petalostigma pubescens, Xanthorrhoea and hills. Seasonal swamp conditions may occur. Estimated 12% of the total johnsonii; Understorey (2 ± 1.5 m; FPC 10–20%): Arundinella nepalensis, serpentine soil in Central Queensland, lowlands primarily occurring in the Bothriochloa spp., Bursaria reevesii, Capparis thozetiana, Gahnia aspera, Goodenia sp. Kunwarara land system. Diagnostic chemical properties at 0–30 cm depth: nov., Grewia latifolia, Lomandra spp., Pimelea leptospermoides, Stackhousia tryonii. Mg:Ca = 5–8; range (mean). Ni concentration = 0.08–0.29% (0.19%).

Open-forest (plains): Overstorey (18±5m;FPC30–35%): Corymbia spp., Euca- Group 5b: Duplex soils composed of fine depositional grey sediment overlay- lyptus crebra, , E. fibrosa subsp. nov, E. platyphylla, E. tereticornis, Lophostemon ing brick-red clays (lateritic colluvium) occur in the northern parts of suaveolens; Mid-understorey (8±2m;FPC15–20%): Acacia spp., Allocasuarina Coorumburra (Develin Creek Flat) on broad alluvial plains. These lowlands luehmannii, Melaleuca spp.; Understorey (2±1m;FPC15–25%): Arundinella belong to the Hedlow land system. Diagnostic chemical properties at 0–10 cm nepalensis, Lomandra spp., Panicum effusum, Pimelea leptospermoides, Themeda depth: Mg:Ca = 0.5–3.2. Ni concentration unknown. triandra, Xanthorrhoea spp.

Low open-forest/Woodland (Melaleuca bracteata understorey): Overstorey Group 6: Ustic Endoaquert black cracking clays overlie a brown subsoil and (16±4m;FPC10–25%): Casuarina cunninghamiana, C. cristata, Eucalyptus deeper magnesite deposits in the swamps and floodways. Prolonged flooding tereticornis, Mid-understorey (6±2m;FPC40–70%): Melaleuca bracteata; may occur. Estimated 16% of the total serpentine soil in Central Queensland, understorey (1.5±1m;FPC15–25%): Arundinella nepalensis, Bothriochloa occurring only in the Kunwarara land system. Diagnostic chemical properties decipiens, Cyperus spp. Dichanthium spp., Gahnia aspera, Grewia latifolia, at 0–30 cm depth: Mg:Ca = 10–20; range (mean). Ni concentration = Paspalidium spp., Sorghum spp., Stackhousia tryonii. 0.14–0.31% (0.26%).