See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/230157212 Lack of disturbance as an explanation for the additive basal area phenomenon in a stratified Afrotemperate forest ARTICLE in AUSTRAL ECOLOGY · JULY 2006 Impact Factor: 1.84 · DOI: 10.1111/j.1442-9993.2006.01607.x CITATIONS READS 14 28 4 AUTHORS, INCLUDING: Michael J. Lawes Stéphane Boudreau 159 PUBLICATIONS 3,541 CITATIONS 32 PUBLICATIONS 508 CITATIONS Charles Darwin University SEE PROFILE Laval University SEE PROFILE Megan Elizabeth Griffiths University of KwaZulu-Natal 38 PUBLICATIONS 429 CITATIONS SEE PROFILE Available from: Megan Elizabeth Griffiths Retrieved on: 05 February 2016 Austral Ecology (2006) 31, 471–477 doi:10.1111/j.1442-9993.2006.01607.x Lack of disturbance as an explanation for the additive basal area phenomenon in a stratified Afrotemperate forest M. J. LAWES,1* J. J. MIDGLEY,2 S. BOUDREAU1 AND M. E. GRIFFITHS1 1 Forest Biodiversity Research Unit, School of Biological and Conservation Sciences, University of KwaZulu-Natal (Pietermaritzburg campus), P/Bag X01, Scottsville 3209, South Africa (Email: Lawes@ukzn.ac.za) and 2Botany Department, University of Cape Town, P/Bag Rondebosch, South Africa Abstract Explanations for the additive basal area (BA) phenomenon in forests are frequently given in the context of stratification and the avoidance of competition for space or resources (e.g. light, nutrients, moisture). Thus, large individuals avoid competition for light by emerging above the canopy and, in so doing, the BA of emergent individuals is often ‘additive’ to that of the rest of the individuals in a stand. The additive BA phenomenon was evident in a stratified Afrotemperate forest and was not confined to the emergent stratum but occurred also within the canopy stratum where gymnosperm BA appeared to be additive. However, there was no evidence that stand BA was at carrying capacity (i.e. presumably neither space nor nutrient resources were limiting) and there was no statistical evidence of competitive effects. We argue that avoiding competition is an insufficient explanation for how biomass accumulates in stratified forests and suggest that local variation in disturbance regime and tree lifehistory provide a more plausible and general explanation for current forest stand structure and dynamics. We suggest that the additive effect is a result of long-lived individuals persisting in the prolonged absence of disturbance to canopy and emergent strata. Key words: competition, disturbance, emergent, longevity, temperate forest. INTRODUCTION There is no general model for the determinants of total biomass and basal area (BA) in forests. One hypothesis for explaining high BA in some forests is the additive BA hypothesis. In some stratified old-growth forests the BA of emergent trees appears to be independent of, and thus ‘additive’ to, that of canopy species in a stand (Paijmans 1970; Enright & Ogden 1995; Lusk 2002). One explanation for why these forests with emergents have high BA is that the emergents are tall enough not to shade the canopy. Because emergents do not compete for canopy space or light, it is hypothesized they can develop large biomass independently of canopy species (i.e. additive BA), adding significantly to overall stand BA (Paijmans 1970; Enright 1982; Enright & Ogden 1995; Lusk 2002; Midgley et al. 2002; Lusk et al. 2003). Thus, the additive BA hypothesis is given in the context of stratification and avoidance of competition for light between the different strata. Midgley et al. (2002) have challenged the importance of the additive contribution of emergents per se, and the role of competition, in determining how biomass accumulates in forests. Using small plots (0.04 ha) in which nearest-neighbour effects could be *Corresponding author. Accepted for publication October 2005. © 2006 Ecological Society of Australia identified, they showed that stand BA or biomass accumulation was fundamentally determined by the presence of large diameter trees, even though there were no emergent individuals in their study. The BA of the canopy component, excluding the large individuals, was not significantly different from that of adjacent stands with no large diameter individuals and thus the contribution of the large diameter individuals was largely additive to that of the remaining stems within the canopy stratum. Based on this finding Midgley et al. (2002) suggested a minor role for competition among canopy trees in biomass accumulation and a stronger within-stratum additive effect of large diameter trees on stand BA in Afrotemperate forest. These authors further argued that large canopy trees gave rise to high stand densities because they packed limited canopy space more efficiently than smaller trees. Packing of large circles involves lower space losses. Lusk et al. (2003), in turn, questioned the Midgley et al. (2002) geometric explanation or ‘packing’ hypothesis, because it assumed isometry of crown expansion and BA increment of individual trees, when this was unlikely (Franco & Kelly 1998). They suggested that the association of large diameter canopy trees with high stand BA could equally arise because of the uncoupling of trunk increment from crown expansion. Here we consider an alternative explanation for the additive BA phenomenon in a stratified montane 472 M . J. L AW E S E T A L . were dominated by the conifers Podocarpus henkelii (Podocarpaceae) and Podocarpus falcatus, but also included scattered angiosperm individuals of Celtis africana (Ulmaceae) and Vepris lanceolata (Rutaceae). These trends in species dominance were repeated in the canopy stratum (Table 1) but not in the understorey. Although large Podocarpus individuals were logged from about 1890 to 1910, the composition and abundance of the dominant species in the emergent and canopy strata has not changed since then (Fourcade 1889). Since the 1950s, however, subsistence harvesting of mainly pole-sized stems, but also the occasional emergent or canopy tree, has damaged parts of the forest that are close to human settlements (M.J. Lawes, unpubl. data 2004). We excluded sample plots from anthropogenically disturbed sections of the forest from the analyses. Stems ≥2 cm d.b.h. were identified and measured in 30 0.1 ha modified Whittaker plots (Obiri et al. 2002) from the southern half of the forest. Plots were located in sets of three (20 m, 200 m and 400 m from the forest edge), each set approximately 200 m apart. For this study, BA was estimated in 20 of the plots that had experienced the least harvesting of the tree strata over the last 80–90 years. Plots were selected for their homogeneity with respect to slope and aspect. All plots were located on shallow slopes away (>30 m) from streambeds. There were no notable observed differences in superficial soils among sites and as all plots were located on one side of a large hill we assumed little variation from site to site in edaphic attributes. We excluded dead trees from our estimates of BA, even though the stump or snag occupied space that may be available to other individuals. To test for the additive versus competitive effect between strata or stratum components, we regressed the total BA of one component (a stratum or species or set of species within a stratum) against another. We did this for both within and between stratum compar- Afrotemperate forest with a large emergent gymnosperm (Podocarpus spp.) component. We argue that lack of disturbance to tree strata can give rise to an additive BA effect. We assess whether the additive BA effect is a general within-stratum phenomenon as opposed to an effect of emergents alone (Kelty 1989) or other over-topping strata. We evaluate the roles of competition and disturbance in determining stand BA, stratification and biomass accumulation. In addition, because the additive effect is most clearly described from temperate mixed broadleaf-conifer forests (Enright 1982; Lusk 2002; Midgley et al. 2002) with a large gymnosperm component, we examine the angiosperm and gymnosperm components for evidence of the effect. METHODS Data were collected in the remote 673 ha iGxalingenwa forest (30.01°S, 29.63°E). This Afrotemperate eastern mistbelt forest is located 25 km from the town of Creighton in southern KwaZulu-Natal province, South Africa. Situated on a south-facing slope between 1300 and 1600 m a.s.l., the forest is cool with a mean annual temperature of 15°C (mean monthly range: 10–19°C), and short periods of snowfall every 3–4 years. Mean annual rainfall is approximately 1060 mm with dry winters (median monthly winter rainfall = 3 mm) and wet summers (169 mm) (Schulze 1997). The forest is highly stratified with a dense emergent tier, and obvious canopy and subcanopy strata. For data analysis the demarcation of these forest strata was determined from a frequency distribution of the tree heights in the sample: (i) understorey ≥ 2 cm diameter at breast height (d.b.h.), <14 m high; (ii) canopy ≥ 10 cm d.b.h., 14 m−20 m high; and (iii) emergents = only those trees >20 m and whose crown clearly overtopped the local canopy. Emergents up to 38 m tall Table 1. Mean ± SD basal area (m2 ha−1) for the understorey, canopy and emergent strata and the total stand (including the saplings) for 20 plots in an Afrotemperate podocarp-broadleaved forest, KwaZulu-Natal, South Africa Mean ± SD Taxa Understorey Canopy Emergent Total % Mean Conifers Podocarpus henklii Podocarpus falcatus All conifers 0.7 ± 1.3 0.2 ± 0.5 0.9 ± 1.3 6.6 ± 9.7 0.6 ± 1.4 7.2 ± 9.4 17.4 ± 18.8 4.8 ± 8.7 22.2 ± 16.9 24.9 ± 22.1 5.5 ± 8.9 30.6 ± 18.4 44.3 9.8 54.5 Angiosperms Celtis africana Vepris lanceolata All angiosperms 0.4 ± 0.6 0.1 ± 0.1 7.0 ± 3.9 1.9 ± 2.6 1.5 ± 1.7 6.1 ± 3.4 1.7 ± 2.5 2.4 ± 3.0 6.4 ± 6.2 4.6 ± 5.1 4.2 ± 4.2 25.7 ± 9.6 8.3 7.4 45.8 Total 7.2 ± 3.3 13.2 ± 9.7 28.6 ± 16. 8 56.2 ± 14.7 100 © 2006 Ecological Society of Australia DISTURBANCE AND ADDITIVE BASAL AREA RESULTS Estimates of total stand BA ranged from 26 to 81 m2 ha−1 among the 20 plots, with a mean of 56 m2 ha−1. On average, this was more-or-less equally distributed between gymnosperms and angiosperms (Table 1). The BA of the emergent stratum was greater than that of the canopy stratum and was dominated by P. henkelii. BA of the canopy stratum was almost evenly distributed between gymnosperms and angiosperms, although there was a much greater representation of angiosperms than gymnosperms in the understorey (Table 1). The understorey comprised 12.8% of the total BA and was thus included in our analyses of competitive effects among strata. Evidence for competition may only be demonstrable in plots with large BA where resources may be most limiting. However, BA increased monotonically across the plots in our study and did not asymptote. We interpret this to indicate that most plots in the forest have not reached maximum BA or carrying capacity (Fig. 1). Although BA of the plots does not reach carrying capacity, there was nevertheless a significant negative correlation between the total BA of the angiosperm and gymnosperm components (r = −0.602, P = 0.005), suggesting competition on a broad level between these taxa. Examining the latter more closely, the BA of the gymnosperm emergents appeared to influence the development of understorey angiosperms with evidence of suppression of the angiosperm component in plots with high emergent gymnosperm BA (r = −0.638, P = 0.002). However, this is unlikely to be a competitive effect as much as a recruitment effect (due to greater propogule availability) because there was a strong positive correlation between gymnosperm BA in the canopy and the BA of gymnosperms in the understorey (r = 0.853, P < 0.0001). Apart from the finding © 2006 Ecological Society of Australia 90 80 Basal area (m2ha–1) isons for angiosperm and gymnosperm components. Failure to detect significant correlation between independent components, such as emergent and canopy BA, or between focal species and the ‘rest’ within a stratum, was taken as support for the additive effect and for the absence of competition. In contrast, a significant negative correlation was considered potential evidence of the existence of competition between components. The correlation or regression techniques we used are consistent with methods used by others to explore the additive BA effect (Enright 1982; Lusk 2002; Midgley et al. 2002; Midgley & Niklas 2004). Although multiple correlation analyses were conducted, we did not apply Bonferroni adjustments to correct for Type I error because we believe these to be excessively conservative under the circumstances (Gotelli & Ellison 2004). 473 70 60 50 40 30 20 10 0 Plots ordered by increasing basal area Fig. 1. Monotonic increase in the basal area of the study plots indicating that local stem basal area has not reached carrying capacity in the iGxalingenwa forest, South Africa. above, there was no statistical evidence of competition between strata of the angiosperm and gymnosperm components or these components combined (Table 2, Fig. 2). There was, however, evidence of putative competition within the emergent stratum where the BA of the emergent dominant P. henkelii was negatively correlated with that of other emergents combined (r = −0.494, P = 0.03). Although the gymnosperm P. henkelii was also the dominant canopy species, there was no evidence of competition with the other canopy species combined (r = −0.169, P = 0.47). Neither was there evidence of competition within the angiosperm canopy stratum, as the combined BA of the angiosperm dominants C. africana and V. lanceolata was not correlated with the combined data for other angiosperm species (r = −0.362, P = 0.11). The total plot BA correlated strongly with both BA of the largest tree in each plot (r = 0.698, P < 0.0001) and the number of trees with d.b.h. > 60 cm in a plot (r = 0.936, P < 0.0001) (Fig. 3). The largest tree contributed an average of 18%, and the trees with d.b.h. > 60 cm an average of 57%, to total plot stem BA. These data indicate that the BAs of the largest individuals mostly determine the pattern of biomass accumulation of the local stand. DISCUSSION Overall, it appears that the BA of the tree community is weakly controlled by competition at iGxalingenwa. There was a strong correlation between plot BA and the size of the largest tree in the plot as well as the number of very large trees in a plot. Large trees appear 474 Correlation coefficients (r) for the regression of basal area between stratum combinations and the angiosperm and gymnosperm components Emergent Total Gymnosperm Total angiosperm Emergent angiosperms Canopy total Canopy gymnosperms Canopy angiosperms Understorey total Understorey gymnosperms Understorey angiosperms Canopy Gymnosperm Angiosperm Total Gymnosperm Angiosperm Total Understorey Gymnosperm −0.198 −0.205 −0.149 −0.176 −0.400 −0.299 −0.638* −0.266 −0.396 0.338 0.415 −0.245 0.181 −0.305 −0.296 −0.051 −0.248 −0.392 −0.574* −0.087 0.283 0.853* −0.135 0.120 −0.065 0.149 0.316 0.803* −0.078 0.001 −0.602* Coefficient values marked with an asterisk (*)are significant at the P < 0.01 (two-tailed test). 60 50 40 30 20 10 0 (a) 70 Basal area (m2ha–1) Plots ordered by increasing conifer basal area 60 50 40 30 20 10 0 40 Plots ordered by increasing emergent basal area Basal area (m2ha–1) (b) 70 (c) 35 30 25 20 15 10 5 0 Plots ordered by increasing canopy basal area © 2006 Ecological Society of Australia Fig. 2. The additive basal area phenomenon: (a) the relationship between basal area of ( ) emergent conifers (Podocarpus spp.) and that of the ( ) angiosperm canopy, (b) the relationship between basal area of ( ) all emergent species and that of ( ) all canopy species, and (c) the relationship between basal area of ( ) all canopy species and that of ( ) all understorey species. Data are from 20 0.1 ha plots in a mixed broadleaf-conifer Afrotemperate forest. Basal area (m2ha–1) M . J. L AW E S E T A L . Table 2. DISTURBANCE AND ADDITIVE BASAL AREA Total plot basal area (m2ha–1) (a) 90 80 70 60 50 40 30 R 2 = 0.4879 20 10 0 0 2 4 6 8 10 12 14 16 18 20 22 Basal area of largest stem (m2ha–1) Total plot basal area (m2ha–1) (b) 90 80 R 2 = 0.8763 70 60 50 40 30 20 10 0 0 2 4 6 8 10 Number of stems DBH >60 cm in plot Fig. 3. (a) Relationship between basal area of the largest stem and that of the ‘total’ of the plot (20 plots, each 0.1 ha), and (b) the relationship between the number of stems with a diameter >60 cm and the ‘total’ basal area of the plot in iGxalingenwa forest, South Africa. to confer high BA on a stand in an additive manner, irrespective of whether the trees are emergent or not (Midgley et al. 2002). The apparent lack of competitive interactions at iGxalingenwa may be due to the phenomenon of emergence or the over-topping of one stratum by another (Enright 1982; Lusk & Ortega 2003). However, as total BA was relatively low at iGxalingenwa, and apparently well below carrying capacity, neither space nor resources were likely to be limiting and the role of competition in determining tree biomass accumulation, density and physiognomy is questionable. © 2006 Ecological Society of Australia 475 Stand stem BA at iGxalingenwa (mean = 56 m2 ha−1, range = 26–81 m2 ha−1) was well within the range of BAs reported for several temperate mixed broadleafconifer forests (New Zealand, old-growth forest, 88 m2 ha−1, Lusk 2002; 68 m2 ha−1, Enright et al. 1999; Mexico, Sierra Madre Oriental, range = 28– 71 m2 ha−1, Williams-Linera et al. 2003; Papua New Guinea, Bulolo Valley, 53 m2 ha−1, Enright et al. 1999; South Africa, Afromontane forest, mean ± SD = 47.9 ± 9.8 m2 ha−1, n = 12 forests, estimates based on stems >5 cm d.b.h., see Lawes et al. 2004 and references therein; Japan, 25.6 m2 ha−1, Hiura & Fujiwara 1999; New Caledonia, Mt. Do, 32.4 m2 ha−1, Enright et al. 1999). Thus, the mean BA of iGxalingenwa forest is not especially low and is probably a consequence of little disturbance to large trees for almost 100 years (Fourcade 1889). These comparative data suggest that many mixed broadleaf-conifer forests are likely to be below carrying capacity (see Enright 1982) and therefore neither competitive effects nor resource limitation necessarily control total BA or size-class transitions in these forests (Enright 1982; Midgley 2001). Recently, Midgley and Niklas (2004) have argued that local disturbance prevents total stand BA from being at equilibrium with the local environment. Like us, they found that total BA at a sample site was strongly correlated with the BAs of the largest trees at that site. Because local disturbance rates are an important determinant of maximum plant size, Niklas et al. (2003) argue that disturbance regimes and not competition have a central role in determining community basal stem area and how biomass accumulates in forests. Although there was a significant negative relationship between the total plot BAs of the angiosperm and gymnosperm components, suggesting a potential competitive relationship between these tree components of the forest, this trend was not evident between paired combinations of strata (see also Enright 1982; Lusk 2002). In the data of Enright (1982) there appeared to be no saturation of plot total BA. We interpret this to indicate his forest was also not at carrying capacity. The apparent independence of limits to biomass accumulation by different strata in this mixed broadleaf-conifer forest is consistent with the hypothesis that competition for resources is not the main mechanism controlling BA and growth of ageing stands. The additive BA phenomenon could suggest that conifers at iGxalingenwa are distributed independently of general forest composition or structure (Enright & Ogden 1995) and that the dynamics of the angiosperm and gymnosperm components may be uncoupled in forests where stand BAs are below carrying capacity. Certainly, the additive BA phenomenon has been frequently described from mixed broadleaf-conifer forests (Enright & Ogden 1995; Lusk 2002). 476 M . J. L AW E S E T A L . There are statistical and sampling issues that need to be resolved before we draw conclusions about the efficacy of disturbance as a determinant of total forest stem BA. For example, the additive basal effect could be an artefact of small plot size because large trees may overflow the edge and inflate estimates of BA. However, the likelihood of including the occasional large tree on the margin in a plot is the same for all plots sampled in this study and as our plots were large (0.1 ha; 20 m × 50 m) compared to other studies of the phenomenon (Lusk 2002; Midgley et al. 2002) and contained on average 5.5 ± 2.6 (mean ± SD, n = 20) large stems (d.b.h. > 60 cm), the effect of the occasional large stem is sufficiently diluted for this not to be a serious problem. In addition, the BA of the largest tree and the total BA of a plot were statistically autocorrelated. We addressed this problem by regressing the BA of the largest tree in a plot against the BA of the remaining stems (= total − largest) and there was still a significant relationship (r = 0.48, P = 0.03). In conclusion, neither the geometric/packing nor the compensation or resource use hypotheses (Midgley et al. 2002), which are based on competitive effects, are adequate explanations for the additive BA phenomenon. While partitioning of space and resources can predispose forests to large stand BA and complex vertical forest structure (Kohyama 1993; Silvertown 2004), it appears these competitive effects are not necessary conditions for the development of additive BA (Hiura & Fujiwara 1999). We suggest that the disturbance hypothesis is a more plausible and general explanation for the phenomenon. Under this non-equilibrium hypothesis, we argue that ecological disturbance keeps growth and biomass below the possible maximum, so that stratum biomass accumulation in Afrotemperate forest is a direct result of adding bigger and longer-lived trees in the prolonged absence of disturbance to the tree strata (see Enright et al. 1999). Midgley and Niklas (2004) suggested these differential size- and/or agebased effects of disturbance on tree mortality allow the evolution of large plant size and an increase in longevity. We agree with their view that latitudinal or other worldwide patterns of standing forest biomass may reflect ecological factors influencing disturbance regimes, far more than the effects of community interaction and competition. However, we draw attention to the observation that additive biomass accumulation is most prevalent in mixed broadleafconifer forests (Enright 1982; Lusk 2002; Midgley et al. 2002). Differences in their physiology from angiosperms (Bond 1989) potentially predispose conifers to large size and longevity (Lusk & Ogden 1992; Enright et al. 1999), which in turn leads to additive patterns of biomass accumulation among conifers in temperate forests responding to disturbance. ACKNOWLEDGEMENTS We thank Stefanie Pollock, John Robertson and Harriet Eeley for assistance in the field, and Pete and Bev Everett for providing accommodation. This study was made possible by financial support from the National Research Foundation (Focus area: Indigenous Knowledge Systems) of South Africa under Grant number 2053633. Any opinion, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Research Foundation. The Mazda Wildlife Fund provided logistic support. REFERENCES Bond W. J. (1989) The tortoise and the hare: ecology of angiosperm dominance and gymnosperm persistence. Biol. J. Linnean Soc. 36, 227–49. Enright N. J. (1982) Does Araucaria hunsteinii compete with its neighbours? Aust. J. Ecol. 7, 97–9. Enright N. J. & Ogden J. (1995) The Southern Conifers: a synthesis. In: Ecology of the Southern Conifers (eds N. J. Enright & R. S. Hill) pp. 271–87. Melbourne University Press, Melbourne. Enright N. J., Ogden J. & Rigg L. S. (1999) Dynamics of forests with Araucariaceae in the western Pacific. J. Veg. Sci. 10, 793–804. Fourcade H. G. (1889) Report on the Natal Forests. Watson, Pietermaritzburg. Franco M. & Kelly C. K. (1998) The interspecific mass-density relationship and plant geometry. Proc. Natl Acad. Sci. USA 95, 7830–5. Gotelli N. J. & Ellison A. M. (2004) A Primer of Ecological Statistics. Sinauer Associates, Sunderland. Hiura T. & Fujiwara K. (1999) Density-dependence and coexistence of confier and broad-leaved trees in a Japanese northern mixed forest. J. Veg. Sci. 10, 843–50. Kelty M. J. (1989) Productivity of New England hemlock/hardwood stands as affected by species composition and canopy structure. For. Ecol. Manage. 28, 237–57. Kohyama T. (1993) Size-structured tree populations in gapdynamic forest – the forest architecture hypothesis for the stable coexistence of species. J. Ecol. 81, 131–43. Lawes M. J., Midgley J. J. & Chapman C. A. (2004) South Africa’s forests: the ecology and sustainable use of indigenous timber resources. In: Indigenous Forests and Woodlands in South Africa: Policy, People and Practice (eds M. J. Lawes, H. A. C. Eeley, C. Shackleton & B. Geach) pp. 31– 75. University of KwaZulu-Natal Press, Pietermaritzburg. Lusk C. H. (2002) Basal area in a New Zealand podocarpbroadleaved forest: are coniferous and angiosperm components independent? N. Z. J. Bot. 40, 143–7. Lusk C. H. & Ogden J. (1992) Age structure and dynamics of A podocarp-broadleaf forest in Tongariro National Park, New Zealand. J. Ecol. 80, 379–93. Lusk C. H. & Ortega A. (2003) Vertical structure and basal area development in second-growth Nothofagus stands in Chile. J. Appl. Ecol. 40, 639–45. Lusk C. H., Jara C. & Parada T. (2003) Influence of canopy tree size on stand basal area may reflect uncoupling of © 2006 Ecological Society of Australia DISTURBANCE AND ADDITIVE BASAL AREA crown expansion and trunk diameter growth. Austral Ecol. 28, 216–18. Midgley J. J. (2001) Do mixed-species mixed-size indigenous forests also follow the self-thinning line? Trends Ecol. Evol 16, 661–2. Midgley J. J. & Niklas K. J. (2004) Does disturbance prevent total basal area and biomass in indigenous forests from being at equilibrium with the local environment? J. Trop Ecol. 20, 595–7. Midgley J. J., Parker R., Laurie H. & Seydack A. (2002) Competition among canopy trees in indigenous forests: an analysis of the ‘additive basal area’ phenomenon. Austral Ecol. 27, 269–72. Niklas K. J., Midgley J. J. & Rand R. H. (2003) Tree size frequency distributions, plant density, age, and community disturbance. Ecol. Lett. 6, 1–7. © 2006 Ecological Society of Australia 477 Obiri J., Lawes M. & Mukolwe M. (2002) The dynamics and sustainable use of high-value tree species of the coastal Pondoland forests of the Eastern Cape Province, South Africa. For. Ecol. Manage. 166, 131–48. Paijmans K. (1970) An analysis of four tropical rainforest sites in New Guinea. J. Ecol. 58, 77–101. Schulze R. E. (1997) South African Atlas of Agrohydrology and Climatology. Water Research Commission, Pretoria. Silvertown J. (2004) Plant coexistence and the niche. Trends Ecol. Evol. 19, 605–11. Williams-Linera G., Rowden A. & Newton A. C. (2003) Distribution and stand characteristics of relict populations of Mexican beech (Fagus grandifolia var. mexicana). Biol. Conserv. 109, 27–36.