Journal of Tropical Ecology (2006) 22:53?64. Copyright ? 2006 Cambridge University Press doi:10.1017/S0266467405002944 Printed in the United Kingdom Competitive dominance in a secondary successional rain-forest community in Borneo Stuart J. Davies??1 and Hardy Semui? ? Center for Tropical Forest Science-Arnold Arboretum Asia Program, Harvard University, Cambridge, MA, USA ? Institute of Biodiversity and Environmental Conservation, Universiti Malaysia Sarawak, Kota Samarahan, Sarawak, Malaysia (Accepted 12 May 2005) Abstract: Competitive interactions among pioneer species may have a signi?cant impact on the course of secondary succession in tropical rain forests. We predicted that the outcome of competitive interactions in early succession would vary with soil nutrient availability. To test this hypothesis we grew seven pioneer tree species alone and in dense competitive mixtures, with four nutrient treatments: no nutrient addition, and N, P, and N plus P addition. Performance of plants grown alone and in mixtures was strongly nutrient limited. However, contrary to expectation, the competitive hierarchy among the seven species was almost identical among the four nutrient treatments. The dominant species, Melastoma malabathricum, accounted for > 70% of total stand biomass in all nutrient treatments. Seedlings of this species had higher rates of gas exchange and initial growth, and lower root allocation than the other species. Pro?les of light availability within the competitive stands indicated that light levels at ground level were well above levels at which pioneer species can successfully survive and grow, yet seedlings of species other than Melastoma remained stunted. Leaf N concentrations in all stands were 25?55% reduced by competitive interactions, and N addition increased relative competitive performance in only Melastoma suggesting that Melastoma was particularly effective at N acquisition, limiting nutrient uptake by the other species. Toward the end of the experiment individuals of Melastoma began to reproduce, suggesting that the competitive hierarchy would have changed in a longer-duration experiment. Key Words: competition, Dillenia suffruticosa, Duabanga moluccana, Ficus grossularioides, Melastoma malabathricum, Nauclea maingayi, nutrient limitation, pioneer trees, succession, tree seedlings, Trema cannabina, tropical forest, Vitex pubescens INTRODUCTION The dynamics of secondary succession vary across tropical rain-forest landscapes. Rates of species and biomass accumulation, as well as the composition of successional communities vary. This variation results from a range of interacting factors including the availabilityofplantpropagules, local climaticandedaphic conditions, and biotic interactions, all of which are affected by the nature of the disturbances that initiate successional dynamics (Brown & Lugo 1990, Guariguata & Ostertag 2001). Understanding factors constraining successional dynamics is an important objective of tropical forest ecology, and is of practical signi?cance for 1Address for correspondence: Center for Tropical Forest Science, Smithsonian Tropical Research Institute, Unit 0948, APO AA 34002- 0948, USA. Email: sdavies@oeb.harvard.edu managing the extensive areas of cleared tropical forests (Holl & Kappelle 1999). The composition and dynamics of successional communities are strongly affected by edaphic conditions (Ewel et al. 1984, Mesquita et al. 2001, Sim et al. 1992, Uhl et al. 1988). Within Malaysian forests it has long been recognized that the ?oristic composition of secondary succession varies signi?cantly across edaphic gradients (Holttum 1954). Wyatt-Smith (1963) recognized communities regenerating on land deforested and cultivated for less than 2 y to be dominated by species of Trema, Mallotus and Macaranga (Kochummen 1966), whereas communities regenerating on deforested and severely degraded land were initially dominated by species of Melastoma, Chromolaena and Dicranopteris (Kochummen & Ng 1977). This latter community, in extremelydegradedconditions, develops intoavegetation type locally referred to as Adinandra-belukar which is 54 STUART J. DAVIES AND HARDY SEMUI dominated by a distinctive suite of pioneer species, including Adinandra dumosa, Ploiarium alternifolium and Rhodamnia cinerea (Holttum 1954, Sim et al. 1992). A similar pattern of differentiation in successional dynamics across edaphic gradients is seen following natural disturbances within primary forest. For example, landslides create heterogeneous edaphic conditions within primary forest (Walker et al. 1996). Species composition and dynamics differ between the upper erosional zone and the lower nutrient-rich deposition areas at the base of landslides. In Sarawak, Malaysia, erosional zones of landslides are often colonized by Dicranopteris linearis, Melastoma malabathricum and species characteristic of degraded secondary successional sites, whereas depositional zones are usually crowded with species of Macaranga (Davies, pers. obs.). Soil nutrient availability may be the key edaphic resource differentiating these types of successional communities. Logging followed by several cycles of intensive agriculture may deplete soil nutrient reserves and lower soil pH (Grubb et al. 1994, Nye & Greenland 1960, Nykvist 1998). Comparisons of lightly and heavily degraded forests usually show lower soil nutrient levels on more degraded sites. For example, in Malaysia soil nutrient levels in Adinandra-belukar are lower than in secondary forest dominated by Macaranga species (Grubb et al. 1994, Sim et al. 1992). In addition, soil nutrient levels are lower in the upper erosional areas on landslides, and experimental nutrient addition has shown that the growth of pioneer seedlings in these sites is nutrient limited (Dalling & Tanner 1995, Fetcher et al. 1996). Furthermore, Turner (1991) found that Trema tomentosa, a species of more fertile successional communities in Malaysia, was more responsive to the addition of fertilizer than Melastoma malabathricum, a species of succession on degraded sites. Despite the evidence for the potential importance of edaphic effects on the course of early secondary succession, numerous recent studies have found that the availability of propagules is perhaps the most important constraint on secondary succession (Holl & Kappelle 1999). The most heavily degraded soils are often most distant from the seed sources of intact forest, and have a group of well-dispersed colonizing species. However, in the absence of limitations to seed dispersal, edaphic conditions and biotic interactions, including competition, may determine the composition of early successional communities. The occurrence of different successional communities on different soils in close proximity (e.g., on different zones within landslides), suggests that the outcome of competitive interactions among species may vary with edaphic conditions. In this study, we investigate the effects of soil nutrient availability on the dynamics of early rain forest succession in Sarawak, Malaysia. Seven high- light-demanding pioneer species that are characteristic of early successional communities in Sarawak were grown in competitive mixtures and without neighbours across a soil fertility gradient. Soil nutrient availability was manipulated by the addition of nitrogen and/or phosphorus. The seven species chosen for the experiment span a broad range of successional communities in the region. Melastoma malabathricum, Dillenia suffruticosa and Vitex pubescens are common to successions on degraded sites, often occurring in Adinandra-belukar. Trema cannabina and Duabanga moluccana are commonly found in early successional communities on richer or less-degraded soils, and Ficus grossularioides and Nauclea maingayi have intermediate responses. We hypothesized that the outcome of competitive interactions among the seven species would vary with soil nutrient availability. METHODS Study species The seven very high-light-demanding pioneer species selected for this experiment: Dillenia suffruticosa (Griff.) Martelli (Dilleniaceae), Duabanga moluccana Blume (Lythraceae), Ficus grossularioides Burm. f. (Moraceae), Melastoma malabathricum L. (Melastomataceae), Nauclea maingayi Hook. (Rubiaceae), Trema cannabina Lour. (Cannabaceae) and Vitex pubescens Vahl (Lamiaceae), hereafter referred to by their generic names, commonly regenerate in areas following logging and shifting cultivation, and in abandoned open areas. The species are often sympatric at the landscape scale in Sarawak (Mohizah & Davies, unpubl. data). However, our ?eld observations suggest the species differ in their speci?c habitat preferences, for example, Dillenia is very common in swampy ground (Corner 1988), whereas Duabanga is often found on fertile clay-rich soils, and on river banks (Davies, pers. obs.). The species also differ in several other life-history characteristics, notably maximum tree size and age; the smallest species, Melastoma and Trema, are not typically more than 1?4 m tall at reproductive maturity, whereas the largest species, Duabanga, grows to c. 30 m tall in rich sites. Experimental design and implementation Plants of all seven pioneer species were grown as single plants and in mixed-species competitive stands. Four nutrient addition treatments were applied to single plants and to competitive stands: control with no nutrient addition (Control), nitrogen addition (N), phosphorus addition (P), and nitrogen and phosphorus addition (NP). Competitive stands were planted in wooden tubs built with 12-mm-thick waterproof plywood. Each Competition in secondary succession 55 tub was 90.9 cm long ? 52.5 cm wide ? 30 cm deep (143.2 litres). Plants grown without neighbours were planted in wooden tubs divided into six smaller boxes of 29.7 ? 25.6 ? 30 cm (22.8 litres). Tubs were ?lled with sandy secondary forest topsoil over 5 cm of coarse gravel. The soil was collected from ?ve areas of degraded mixed dipterocarp forest, sieved through 1-cm2 wire screen and then mixed 3:1 with river sand to ensure good drainage and to enable root sampling at the end of the experiment. Tubshadholes drilled through the base andwere elevated above the ground to facilitate drainage and to ensure the roots did not escape. All tubs were in full sunlight in the grounds of the Universiti Malaysia Sarawak campus in Kuching, Malaysia (1?28?N, 110?27?E). Mean annual rainfall in Kuching exceeds 4000 mm. Tubs were kept well-watered throughout the experiment. Seeds of all species were collected from several mother trees in the Kuching area and then sown in nursery beds in February 1999. To ensure similar-sized seedlings at transplanting, planting in the seed beds was staggered by previously observed germination rate. Seed beds, with the same soil as in the experiment, were initially covered with shade cloth, but this was removed when the seedlings were c. 2 cm tall. Once the seedlings were c. 3 cm tall, they were trans- planted into the experimental tubs. Each competition stand included 98 seedlings consisting of 14 seedlings of each species. Within each competition stand seedlings were arranged in hexagonal arrays such that each seedling was surrounded by six equidistant neighbours, 7 cm apart, and all of which were of a different species. No individual had its own species as a nearest neighbour. In total there were 16 competition stands (4 nutrient treatments ? 4 replicates per treatment) with 98 plants per tub, for a total of 1568 seedlings. Of the 98 plants in each competitive stand, 38 plants (?ve for four species and six for three species) were on the edges of the stands and had fewer than six nearest neighbours. To reduce the number of times any one species had fewer than six edge plants, each replicate competitive stand had a slightly different seedling arrangement.Analyseswere conducted to assess whether edge plants with 2?5 nearest neighbours had signi?cantly different growth than plants with the full compliment of six nearest neighbours. Mean biomass of the edge plants was not signi?cantly different than plants with sixnearestneighbours, soall seedlingswere included in the statistical analyses. Five plants of each species were grown without competition in the four nutrient treatments: control, N, P and NP treatments. Single plants were planted in divided plant tubs, with six seedlings planted in each tub. Nutrients were added every 4 wk after transplanting. Nutrient addition concentrations were based on Turner (1991).Pwasadministeredas tri-superphosphate (P2O5). N fertilizer was applied as NH4NO3. For the ?rst fertilizer treatment 0.47 g m?2 of N and 1.14 g m?2 P were added. Subsequent additions were: 0.94 g m?2 of N and 2.34 g m?2 of P. Single plants and competitive stands received the same nutrient additions per unit soil surface area. Soil nitrogen and phosphorus concentrations were assessed from 5?15 cm depth in all experimental tubs before and after the experiment. Soils were sieved and air-dried prior to analysis. Total P was determined following wet digestion on a Perkin Elmer inductively coupled plasma spectrophotometer. Available P was determined followingextractionwithammonium?uoride and hydrochloric acid (Bray-2 method) followed by continuous ?ow analysis of the extracted solution with a ChemLab 4 autoanalyser (ChemLab, UK). Soil N was determined by Kjeldahl digestion. Seedling assessment Seedlings were transplanted to the experimental tubs in May 1999. The biomass of 10 randomly selected individuals per species was assessed prior to planting into the experimental tubs. Seedling survival, stem height, stem diameter (at 1 cm height), number of leaves and crown diameter (mean of longest and shortest axes) were measured every 4 wk for all seedlings. In November, above- and below-ground biomass for all seedlings in the experiment was measured. Leaf area was measured with a leaf area meter (T Delta Device, Burwell, Cambridge, England) connected to a video camera (Baxall model CD6252, UK). All plant tissue was dried to constant weight at 60 ?C. Leaf nitrogen and phosphorus concentrationsweremeasuredonup to three leaves per species per treatment for single and competitive plants. Total leaf N was measured by Kjeldahl digestion, and P content was determined following digestion with sulphuric acid. Maximum photosynthetic capacity was measured on plants grown without competition just before the end of the experiment. Measurements were conducted between 9h00 and 11h00 on clear sunny days on recently mature leaves using a portable closed gas exchange system (LI-COR Inc. Model 6200, Lincoln, Nebraska, USA). The CO2 calibration was checked daily. Across all photosynthesis measurements, mean CO2 concentration was 353 ? 2 ppm, mean PAR was 1757 ? 44 ?mol m?2 s?1, mean ambient temperature was 37 ? 0.3 ?C, mean leaf temperature was 37.6 ? 0.3 ?C, and relative humidity was 58.7 ? 0.8%. Photosynthetically active radiation (PAR) was mon- itored monthly throughout the experiment using four LI-COR quantum sensors attached to a LI-COR 1000 data logger (LI-COR, Inc., USA). At the end of the experiment, PAR pro?les within the canopy of the competitive stands 56 STUART J. DAVIES AND HARDY SEMUI Table 1. Mean (? 1 SE) soil nutrient concentrations at the end of the experiment for plants grown alone (n = 5?6) and in mixtures (n = 3?4 samples per treatment). Plants were grown without nutrient addition (Control) or with added nitrogen (N), phosphorus (P) or nitrogen and phosphorus (NP). Signi?cantdifferences amongmeans for ?mixture? and ?single? plants are indicated with different superscripts. Available P Total P % N (mg kg?1) (mg kg?1) Single plants Control 0.16 ? 0.00a 13.2 ? 0.5b 132.7 ? 1.9b N 0.17 ? 0.00a 14.2 ? 0.7b 132.8 ? 6.2b P 0.17 ? 0.01a 39.8 ? 9.9a 185.2 ? 17.1a NP 0.17 ? 0.01a 24.3 ? 5.7ab 152.5 ? 10.9ab Mixtures Control 0.17 ? 0.01a 12.3 ? 0.9bc 144.3 ? 1.8a N 0.17 ? 0.01a 11.3 ? 1.0c 138.5 ? 3.6a P 0.18 ? 0.01a 21.0 ? 2.1a 160.0 ? 8.3a NP 0.18 ? 0.01a 18.8 ? 2.7ab 146.0 ? 6.5a were measured with sensors at 5 and 20 cm above the soil, and in full sun above the canopy. Analysis Relative competitive performance (RCP) was calculated as: RCP = (Balone ? Bmixture)/Balone ? 100,whereB is plant biomass for plants grown alone and in mixtures (Keddy et al. 2002, Weigelt & Jolliffe 2003). Differences in performance amongnutrient treatmentswere testedwith ANOVAs followed by Tukey HSD post-hoc comparisons. Comparisons of rank performance among nutrient treatments for species grown alone and in mixtures were made with Kendall?s coef?cient of concordance. Logistic regression was used to test for differences in mortality rates among species and treatments. RESULTS Soil nutrient treatments Available soil P concentrations were 53?202% greater in P treatments for both competitive mixtures and single plants (Table 1). Total soil P concentrations were 15? 40% greater in P treatments for single plants, but not signi?cantly elevated in mixtures. Available P was signi?cantly lower in competitivemixtures than for single plants, but total P was not. Soil N concentrations at the end of the experiment did not differ signi?cantly among soil treatments or between single plants and competitive mixtures (Table 1). Survival Mortality of Ficus and Trema was signi?cantly greater in competitive stands than in plants grown alone (P < 0.05, 0 20 40 60 80 100 Ficus Trema 5 species Month 0 20 40 60 80 100 J J A S O N J J A S O N Ficus Trema Dillenia Duabanga Ficus Melastoma Nauclea Trema Vitex Single plants % s ur vi va l Competitive stands % s ur vi va l Figure 1. Seedling survival in competitive mixtures for seven pioneer species. Nutrient treatments were pooled for this ?gure since they had no signi?cant effect on survival (logistic regression, P > 0.5). Figure 1). More than 90% of Ficus seedlings died in the competitive stands compared with only 30% in plants grown without competition. Competitive interactions did not signi?cantly affect survival in the other ?ve species. For plants grown alone, Ficus and Trema had signi?cantly greater mortality than the other species (P < 0.05). Soil nutrient addition had no signi?cant effect on mortality rates for any species with or without neighbours (P > 0.5, data not shown). Competition in secondary succession 57 Table 2. Performance of seven species of pioneer trees in the competition experiment. Mean (? 1 SE) ?nal biomass (g) of seedlings grown without neighbours (Ficus and Trema n = 3?5, other species n = 5) and in competitive mixtures (Ficus n = 2?5, Trema n = 32?38, all other species n = 53?56 per treatment). Relative competitive performance is the per cent reduction in total biomass due to competitive interactions. Plants were grown without nutrient addition (Control) or with added nitrogen (N), phosphorus (P) or nitrogen and phosphorus (NP). For ?nal biomass, signi?cant differences among species within each nutrient treatment indicated after each mean based on log10 transformed biomass. Ficus was excluded from the ?mixtures? analysis due to limited sample sizes. For relative competitive performance, signi?cant differences among nutrient treatments indicated with different letters for Melastoma; there were no signi?cant differences in the other species. Species arranged by competitive performance ranking. Species Control N P NP Final biomass: single plants Melastoma 96.4 ? 5.2a 99.9 ? 19.3a 112.0 ? 13.7a 143.3 ? 18.3a Vitex 48.2 ? 4.7bc 62.8 ? 3.6a 60.2 ? 16.4bc 75.9 ? 7.5ab Dillenia 51.2 ? 6.5bc 80.8 ? 7.8a 107.1 ? 15.5ab 95.7 ? 8.5a Nauclea 32.2 ? 6.7c 33.4 ? 4.9ab 42.0 ? 6.2c 52.8 ? 7.1b Duabanga 72.2 ? 8.2ab 40.3 ? 2.6ab 88.4 ? 12.9a 155.0 ? 18.5ab Trema 15.2 ? 3.0d 25.7 ? 9.5b 19.8 ? 5.3d 25.1 ? 1.9c Ficus 32.9 ? 2.1c 63.2 ? 27.1ab 40.1 ? 4.1bc 63.2 ? 15.8bc Final biomass: competitive mixtures Melastoma 18.29 ? 1.73a 28.17 ? 3.02a 21.82 ? 2.01a 38.36 ? 3.00a Vitex 2.98 ? 0.30b 3.55 ? 0.45b 2.96 ? 0.48bc 5.17 ? 0.54b Dillenia 1.40 ? 0.13bc 3.25 ? 0.31b 2.47 ? 0.19b 2.37 ? 0.34c Nauclea 1.15 ? 0.10c 1.03 ? 0.14c 0.96 ? 0.08d 1.33 ? 0.11cd Duabanga 0.87 ? 0.12c 0.75 ? 0.12c 1.47 ? 0.19cd 1.21 ? 0.17d Trema 0.17 ? 0.04d 0.07 ? 0.03d 0.14 ? 0.04e 0.54 ? 0.23e Ficus 0.15 ? 0.11? 0.09 ? 0.00? 0.06 ? 0.02? 0.43 ? 0.20? Relative competitive performance Melastoma 81.0 ? 2.70a 71.8 ? 1.60b 80.5 ? 1.73a 73.2 ? 1.11ab Vitex 93.8 ? 0.97 94.4 ? 1.86 95.1 ? 0.74 93.2 ? 1.36 Dillenia 97.3 ? 0.87 96.0 ? 0.91 97.7 ? 0.24 97.5 ? 0.41 Nauclea 96.4 ? 0.63 96.9 ? 1.05 97.7 ? 0.35 97.5 ? 0.52 Duabanga 98.8 ? 0.24 98.1 ? 0.39 98.3 ? 0.21 99.2 ? 0.21 Trema 99.0 ? 0.67 99.7 ? 0.12 99.3 ? 0.26 97.7 ? 1.32 Ficus 99.3 ? 0.54 99.9 ? 0.01 99.9 ? 0.05 99.3 ? 0.32 Single plant growth: nutrient effects and rank performance Nutrient addition had a signi?cant positive effect on ?nal plant biomass in only Duabanga and Dillenia (Figure 2). Mean biomass of Dillenia seedlings was greater in P and NP addition treatments than in control plants (Table 2). NP addition led to signi?cantly greater ?nal biomass in Duabanga. For the other species there appeared to be species-speci?c differences in response to nutrient addition, however signi?cant effects may not have been detected due to limited sample sizes. Melastoma, Nauclea and Vitex, as with Duabanga, had highest mean biomass in NP addition treatments. Dillenia performed best with added P. Ficus and Trema performed best in N and NP addition treatments (Figure 2). Mean ?nal biomass differed signi?cantly among species in all nutrient treatments (Table 2). Melastoma had the highest mean biomass in control, N and P treatments and Duabanga had the highest mean biomass in the NP treatment. Nauclea and Trema had the lowest mean biomass in all four treatments. Rank performance of the species did not differ signi?cantly among nutrient treatments (Kendall?s coef?cient of concordance, W = 0.857, P < 0.05). Dynamics of competitive mixtures Melastoma dominated the competitive mixtures, account- ing for 73?79% of total stand biomass in all nutrient treatments (Figure 3a). The next most important species were Vitex with 10?12% and Dillenia with 5?9% of stand biomass. Nauclea and Duabanga had 2?5% of stand biomass, and Trema and Ficus accounted for less than 1% of total stand biomass. Nutrient addition signi?cantly affected total stand biomass, with N-treated stands having 48% greater biomass than control and P-treated stands, and NP- treated stands having 33?97% greater biomass than other treatments (Figure 3b). Despite signi?cant effects on stand biomass, nutrient addition did not affect stand composition. Rank performance was almost identical among nutrient treatments (W = 0.987; P < 0.05), with species in control, N and NP treatments having the same ranks, and just one rank difference in P-treated stands. Individual performance in competitive mixtures Compared with plants grown without neighbours, all species suffered > 70% reduction in mean individual 58 STUART J. DAVIES AND HARDY SEMUI - 5 0 0 5 0 C N P N P - 5 0 0 5 0 C N P N P - 5 0 0 5 0 C N P N P - 5 0 0 5 0 C N P N P - 5 0 0 5 0 C N P N P - 5 0 0 5 0 C N P N P - 5 0 0 5 0 C N P N P Dillenia Duabanga Ficus Melastoma Nauclea Trema Vitex a b a ab a b b c nsns ns ns ns % g ro wt h de via tio n % g ro wt h de via tio n % g ro wt h de via tio n % g ro wt h de via tio n Figure 2.Mean growth rates as per cent deviation from the mean ?nal biomass for each species. Deviations are for single plants grown in four nutrient treatments. Different letters indicate signi?cant differences among means for Dillenia and Duabanga; ns indicates no signi?cant differences among means. Competition in secondary succession 59 Figure 3. Relative biomass contribution (a) and mean stand biomass (b) for seven species of pioneer trees growing in competitive stands in four nutrient treatments: C, control; N, nitrogen addition; P, phosphorus addition; and NP, nitrogen and phosphorus addition. Different letters indicatesigni?cantdifferences inmeanstandbiomass(n = 4pernutrient treatment). Relative contributions of species to stand biomass did not differ among nutrient treatments (G-test, P > 0.5). biomass in the competitive mixtures (Table 2). Melastoma suffered least with 72?81% lower mean individual biomass in competitive stands relative to plants without neighbours. Vitex and Dillenia suffered 93?98% biomass reductions, and the other species suffered up to 99.9% reductions in growth relative to plants grown without neighbours. Nutrient addition signi?cantly affected the relative competitive performance in only one of the seven species (Table 2). In Melastoma, N and NP addition resulted in signi?cantly smaller reductions in mean individual biomass than control and P treatments. 0 25 50 75 100 0 50 100 J A S O N Dillenia Duabanga Ficus Nauclea Trema Vitex Melastoma Pe r c en t g ro wt h re du ct io n (a) (b) Pe r c en t g ro wt h re du ct io n Height growth Leaf production Month Figure 4.Timecourseof relativeperformance in (a)heightgrowthand (b) leaf production for seven pioneer species grown in competitive stands. Relative performance is the relative difference in growth for plants in competitive mixtures versus plants grown alone. Prior to strong competitive interactions, some species growth was better in competitive stands than in single plants, although this effect is magni?ed for small plants. Performance reductions due to competitive interac- tions were assessed monthly for height growth and leaf production (Figure 4). By August, the third month of the experiment, leaf production was reduced by more than 75% in Trema and Ficus, the two most heavily affected species. Height growth responded more slowly in all species. Leaf nutrient concentrations N addition did not affect leaf N in any species for single plants or competitive stands (data not shown). 60 STUART J. DAVIES AND HARDY SEMUI 0 0.5 1 1.5 2 C N P NP 0 0.5 1 1.5 2 C N P NP 0 0.05 0.1 0.15 0.2 C N P NP 0 0.05 0.1 0.15 0.2 C N P NP Competition Single SingleCompetition a a a a b b b b a a c b bcbc bc bc Nutrient treatment Nutrient treatment (b) (a) Leaf phosphorus (%) Leaf nitrogen (%) Figure 5. Mean ( ? 1 SE) leaf nitrogen (a) and phosphorus (b) concentrations for seedlings in four nutrient treatments based on n = 15?20 leaves with all species pooled. Signi?cant differences among means are indicated by different letters. However, leaf tissue N concentrations were signi?cantly lower in competitive mixtures than in single plants for all treatments (Figure 5a) and for all species with suf?cient leafmaterial for analysis. Inaddition, therewere signi?cant differences among species in leaf N content in plants with and without neighbours (Table 3). For single plants,FicusandTremahadsigni?cantlygreatermeanleaf N than all other species except Melastoma, and Dillenia, Duabanga and Nauclea had signi?cantly lower leaf N than the other species. In competitive mixtures, with Ficus and Trema having insuf?cient leaf material for analysis, Melastoma had the highest leaf N concentrations. P addition resulted in higher leaf P concentrations in plants grown with and without neighbours in most species (data not shown). Leaf P concentrations were lower in competitive stands for most species for P and NP treatments (Figure 5b). As with leaf N, species also differed signi?cantly in leaf P concentrations (Table 3). Single plants of Trema and Ficus had the greatest leaf P concentrations. Competition in secondary succession 61 Table 3. Leaf nitrogen and phosphorus concentrations for seven pioneer species grown alone and in competitive mixtures at the end of the experiment. For leaf N means based on n = 12 for all species, except Ficus (single: n = 12, competition: n = 0) and Trema (single: n = 5, competition: n = 1). Nutrient addition treatments were pooled for this analysis as they did not signi?cantly affect leaf N concentrations in any species. For leaf P, means based on n = 6 for leaves in control and N treatments pooled, except Ficus (competition: n = 0) and Trema (single: n = 2, competition: n = 0). Different letters indicate signi?cant differences among species, with Ficus and Trema excluded from the competitive mixture comparison. Species arranged as in Table 2. Species Single plants Competitive mixtures Leaf nitrogen (%) Melastoma 1.75 ? 0.09bc 1.07 ? 0.03a Vitex 1.59 ? 0.09c 0.96 ? 0.06ab Dillenia 1.19 ? 0.04d 0.89 ? 0.03bc Nauclea 1.07 ? 0.06d 0.68 ? 0.03d Duabanga 1.10 ? 0.05d 0.74 ? 0.03cd Trema 2.06 ? 0.25ab 0.93 Ficus 2.21 ? 0.09a ? Leaf phosphorus (%) Melastoma 0.085 ? 0.00ab 0.062 ? 0.01ab Vitex 0.100 ? 0.01a 0.081 ? 0.00a Dillenia 0.058 ? 0.00bc 0.051 ? 0.01b Nauclea 0.042 ? 0.00c 0.046 ? 0.01b Duabanga 0.072 ? 0.01b 0.061 ? 0.01ab Trema 0.112 ? 0.01a ? Ficus 0.111 ? 0.00a ? Light environment and photosynthetic capacity PAR was measured above and within the canopy of competitive stands before the end of the experiment (Figure 6). Plants were grown in full sun which averaged 31.4molm?2 d?1 for 5dof lightmeasurements.At20 cm above the soil within the stands, close to the mean height of Vitex and Duabanga seedlings, mean PAR was 15.4 mol m?2 d?1 (49% of full sun) and instantaneous PAR exceeded 1000 ?mol m?2 s?1 for a signi?cant period of the day. At 5 cm above the soil, close to the mean height of remaining Ficus and Trema seedlings, mean PAR was 5.1 mol m?2 d?1 (16.2%). Rates of gas exchange measured on single plants on full-sun days differed signi?cantly among the seven species (one-way ANOVA, P < 0.001). Melastoma had signi?cantly higher rates of photosynthesis and stomatal conductance than most other species in the experiment. Mean rates of maximum photosynthesis (Amax in ?mol m?2 s?1) and stomatal conductance (gs in mol m?2 s?1) for the seven species were, from highest to lowest Amax: Melastoma (Amax = 21.9, gs = 1.25), Ficus (17.5, 0.86), Vitex (15.3, 0.80), Trema (14.8, 0.83), Nauclea (14.5, 0.96), Duabanga (13.9, 0.70) and Dillenia (11.6, 0.70). Competitive dominance Plant size at the start of the experiment did not have a signi?cant in?uence on the competitive outcome (Table4).Melastoma seedlingsgrownwithoutneighbours had signi?cantly higher maximum rates of photosyn- thesis and stomatal conductance, and higher initial growth rates than the other species. These three traits were strongly positively correlated with performance in competitive stands (Table 6). Relative biomass allocation to roots in seedlings without neighbours was negatively correlated with performance in competitive mixtures 0 5 0 0 1 0 0 0 1 5 0 0 2 0 0 0 2 5 0 0 6 0 0 8 0 0 1 0 0 0 1 2 0 0 1 4 0 0 1 6 0 0 1 8 0 0 Full sun 20 cm 5 cm Time (h) PA R (? m ol m - 2 s- 1 ) Figure 6. Time course of photosynthetically active radiation (PAR) above and within (5 and 20 cm above the ground) competitive stands just prior to the end of the experiment. Data are for one representative sunny day. 62 STUART J. DAVIES AND HARDY SEMUI Table 4. Correlations (r) between mean ?nal biomass in competitive mixtures (log10 transformed) and species? traits derived from plants grown without neighbours. Amax, leaf area-based maximum photosynthesis, gs, stomatal conductance, leaf N and P concentrations, RWR, root weight ratio, LWR, leaf weight ratio, height growth in the ?rst census interval (t2 ? t3), and ?nal and initial biomass. Correlations included values for each nutrient treatment separately, so n = 28 for each analysis. Trait r P Amax 0.39 0.040* gs 0.52 0.005** Leaf N (%) ? 0.35 0.064 Leaf P (%) 0.22 0.262 RWR ? 0.45 0.015* LWR 0.27 0.166 Height growth T2?3 0.55 0.005** Final biomass 0.68 < 0.001*** Initial biomass 0.07 0.727 (Table 4), asMelastomahadamong the lowest rootweight ratios andTremaandFicushadhigh root allocation.Rapid early growth rates and high allocation to shoots enabled Melastoma to suppress the growth of the other six pioneer species in this experiment. DISCUSSION Melastoma malabathricum dominated competitive mix- tures in all four nutrient treatments in this experiment, accounting for > 70% of stand biomass in all treatments. This result was contrary to expectation for two reasons. First, ?eld surveysof secondary successional communities in South-East Asia (Ewel et al. 1984, Wyatt-Smith 1963), and elsewhere in the tropics (Mesquita et al. 2001, Uhl et al. 1988), have demonstrated differences in successional communities across edaphic gradients. Second, ?eld and nursery experimental studies, particularly in the temperate zone, have revealed differences in the outcome of competitive interactions when levels of resource availability are altered (Gehring et al. 1999, Navas et al. 1999, Reekie & Bazzaz 1989, Tilman 1987). Light pro?les within the canopy of the competitive stands at the end of the experiment suggest that the competitive outcome was not determined primarily by competition for light. Melastoma has small leaves on relatively narrow upright shoots, and does not cast a deep shade. At the end of the experiment, light levels close to the soil surface were approximately 16% of full sun with instantaneous rates above 1000 ?mol m?2 s?1 for part of the day. Most pioneer species that have been examined can grow rapidly in these light levels (Barker et al. 1997, Davies 1998, Kitajima 1994). Light levels in the forest understorey are typically around 1?3% of full sun, and are conditions in which pioneers rarely persist. Seedling growth in competitive stands was nutrient limited. Mean plant biomass for ?ve of the seven species in competitive stands was greater in at least one of the nutrient addition treatments than in control treatments. The two species without an effect, Trema and Ficus, had too fewsurviving seedlings for comparison.Despite strong nutrient limitation, N and/or P addition had no effect on the relative competitive performance of six of the seven species. If the competitive outcome in this experimentwas determined by more effective competition for nutrients by Melastoma, then it might be expected that nutrient addition would reduce competitive suppression of the other species. However, this was not the case in the experiment. Melastoma effectively limited nutrient uptake by the other species in all nutrient treatments in this experiment. One possible explanation for the result is that the addition of N or P may not have raised soil nutrient concentrations to surplus levels for Melastoma, at which other species could effectively takeup theaddednutrients. Soilsused for theexperimentwererelativelypoor,but they were not unusually poor, having similar nitrogen and phosphorus concentrations to the lowland ultisols that are widespread in Borneo (see, for example, Palmiotto et al. 2004). At least for P, nutrient addition raised soil P concentrations to levels equivalent to the more fertile ultisols on clay soils (Palmiotto et al. 2004).Alternatively, competition may have been for another below-ground resource not measured in this study (e.g. Mg; Burslem et al. 1996), Melastoma is known to accumulate Al in leaf tissues (Watanabe 1997, Watanabe & Osaki 2001). Melastoma mayhavedominated the competitive stands bymodifying soil resource conditions through an allelopathic effect. It releases high concentrations of organic acids into the rhizosphere thathelp solubilizeAl compounds, increasing Al concentrations in the rhizosphere (Watanabe & Osaki 2002). Al is toxic for many plant species and may have led to the stunted growth of the other surviving species in this study. Al concentrations were not measured in this study. The only species for which nutrient addition changed relative competitive performance was Melastoma. Nitrogen addition (N and NP treatments) decreased the effect of stand competition on Melastoma. Melastoma also had higher mean plant biomass in N and NP treatments than in control and P treatments, and higher per cent contributions to total stand biomass in N and NP treatments. Since competitive interactions greatly reduced leaf N concentrations in all stands, Melastoma appears to have beenparticularly effective atN acquisition. Thismay have beena mechanismby which it dominated the competitive stands. Further experimental work with a broader range of N availabilities would provide a test of this hypothesis. Competitive interactions signi?cantly increased mor- tality in Ficus and Trema, and growth of these two species was the most severely reduced by competitive Competition in secondary succession 63 interactions. Although Trema grew more slowly than the other species without neighbours, single plants of Ficus had among the highest gas exchange and growth rates. Ficus and Trema had the highest leaf N and P concentrations when grown without neighbours. These two species are typically found in successional communities on higher-nutrient soils (Wyatt-Smith 1963), and appear to have relatively high nutrient requirements. In full sun with low leaf nitrogen concentrations, seedlings may suffer photodamage (Bungard et al. 2000). This effect may be more severe in species with high nitrogen requirements, and may exclude pioneers of typically high-nutrient soils from poorer sites. Single plants of both Trema and Ficus had relatively better performance in N and NP treatments. Further ecophysiological study of these species in high- light, low-nitrogen environments is required. Melastoma is one of the characteristic species of Adinandra-belukar communities which grow in the very low-nutrient soils in South-East Asia. In natural successional communities in Borneo, Melastoma does not dominate early successional sites on more fertile soils. In fertile soils the chances of having individuals of other species outshadingMelastoma individualsmaybe greater, and Melastoma is highly intolerant of shade (Turner 1991). The experiment reported here lasted 6 mo. Most of the species in the experiment have life spans in excess of 10 y, and Duabanga, for example, can be a 30-m tree at maturity (Corner 1988). The outcome of the competitive interactions observed in this experiment is unlikely to have persisted over a much longer period. Although Ficus and Trema mostly died, almost all seedlings of Vitex, Dillenia, Duabanga and Nauclea survived. Melastoma usually reaches only 3?6 m in height at maturity (Corner 1988), and several individuals had started to ?ower as the experiment was ending. We predict that once Melastoma begins to change resource allocation with the onset of reproduction, and subsequently senesce, the other surviving pioneers in the experiment will increase growth rates and overtop the remaining Melastoma. In degraded old ?elds in Peninsular Malaysia, stands dominated in the early stages by Melastoma became dominated by other species after 4?8 y (Kochummen & Ng 1977). It was not known whether this change was due to shading by neighbouring successional species or to the natural senescence of the Melastoma individuals. The results presented in this study suggest that the competitive dominance of Melastoma was determined by below-ground competition. Once individuals of the other species shadeMelastoma individuals,we expect there to be a rapid shift in the competitive hierarchy, perhaps driven by a shift from below-ground to above-ground (light) competition. Further experimental studies with a longer duration are required to investigate these possibilities. ACKNOWLEDGEMENTS The research was supported by IRPA grant #05-02-09- 0508 to Unimas and SJD. We thank Professor Ghazally Ismail and the Unimas staff for facilitating the project. We thank the Forest Department of Sarawak for permission to work with the tree species. P.S. Ashton, D. King and S. Russo provided constructive comments on the manuscript. LITERATURE CITED BARKER, M. G., PRESS, M. C. & BROWN, N. D. 1997. Photosynthetic characteristics of dipterocarp seedlings in three tropical rain forest light environments: a basis for niche partitioning? Oecologia 112:453?463. BROWN, S. & LUGO, A. E. 1990. Tropical secondary forests. Journal of Tropical Ecology 6:1?32. BUNGARD, R. A., PRESS, M. C. & SCHOLES, J. D. 2000. The in?uence of nitrogen on rain forest dipterocarp seedlings exposed to a large increase in irradiance. Plant, Cell and Environment 23:1183?1194. BURSLEM, D. F. R. P., GRUBB, P. J. & TURNER, I. M. 1996. Responses to simulated drought and elevated nutrient supply among shade- tolerant tree seedlings of lowland tropical forest in Singapore. Biotropica 28:636?648. CORNER, E. J. H. 1988. Wayside trees of Malaya. (Third edition). The Malayan Nature Society, Kuala Lumpur, Malaysia. DALLING, J. W. & TANNER, E. V. J. 1995. An experimental study of regeneration on landslides in montane rain forest in Jamaica. Journal of Ecology 83:55?64. DAVIES, S. J. 1998. Photosynthesis of nine pioneer Macaranga species from Borneo in relation to life-history. Ecology 79:2292?2308. EWEL, J. J., CHAI, P. & LIM, M. T. 1984. Biomass and ?oristics of three young second-growth forests in Sarawak. Malayan Forester 46:347? 364. FETCHER, N., HAINES, B. L., CORDERO, R. A., LODGE, D. J., WALKER, L. R., FERNANDEZ, D. S. & LAWRENCE, W. T. 1996. Responses of tropical plants to nutrients and light on a landslide in Puerto Rico. Journal of Ecology 84:331?341. GEHRING, C., DENICH, M., KANASHIRO, M. & VLEK, P. L. G. 1999. Response of secondary vegetation in Eastern Amazonia to relaxed nutrient availability constraints. Biogeochemistry 45:223?241. GUARIGUATA, M. R. & OSTERTAG, R. 2001. Neotropical secondary forest succession: changes in structural and functional characteristics. Forest Ecology and Management 148:185?206. GRUBB, P. J., TURNER, I. M. & BURSLEM, D. F. R. P. 1994. Mineral nutrient status of coastal hill dipterocarp forest and adinandra belukar in Singapore: analysis of soil, leaves and litter. Journal of Tropical Ecology 10:559?577. HOLL, K. D. & KAPPELLE, M. 1999. Tropical forest recovery and restoration. Trends in Ecology and Evolution 14:378?379. HOLTTUM, R. E. 1954. Adinandra belukar. Malayan Journal of Tropical Geography 3:27?32. 64 STUART J. DAVIES AND HARDY SEMUI KEDDY, P. A., NIELSEN, K., WEIHER, E. & LAWSON, R. 2002. Relative competitive performance of 63 species of terrestrial herbaceous plants. Journal of Vegetation Science 13:5?16. KITAJIMA, K. 1994. Relative importance of photosynthetic traits and allocation patterns as correlates of seedling shade tolerance of 13 tropical trees. Oecologia 98:419?428. KOCHUMMEN, K. M. 1966. Natural plant succession after farming in Sg. Kroh. Malayan Forester 29:170?181. KOCHUMMEN, K. M. & NG, F. S. P. 1977. Natural plant succession after farming in Kepong. Malaysian Forester 40:61?78. MESQUITA, R. C. G., ICKES, K., GANADE, G. & WILLIAMSON, G. B. 2001. Alternative successional pathways in the Amazonian Basin. Journal of Ecology 89:528?537. NAVAS, M., GARNIER, E., AUSTIN, M. P. & GIFFORD, R. M. 1999. Effect of competition on the responses of grasses and legumes to elevated atmospheric CO2 along a nitrogen gradient: differences between isolated plants, monocultures and multi-species mixtures. New Phytologist 143:323?331. NYE, P. H. & GREENLAND, D. J. 1960. The soil under shifting cultivation. Technical communication No. 51, Commonwealth Bureau of Soils, Harpenden, England. 156 pp. NYKVIST, N. 1998. Logging can cause a serious lack of calcium in tropical rainforest ecosystems: an example from Sabah, Malaysia. Pp. 88?91 in Schulte, A. & Ruhiyat, D. (eds.). Soils of tropical forest ecosystems. Characteristics, ecology and management. Springer-Verlag, Berlin. PALMIOTTO, P. A., DAVIES, S. J., VOGT, K. A., ASHTON, P. M. S., VOGT, D. J. & ASHTON, P. S. 2004. Soil-related habitat specialization in dipterocarp rain forest tree species in Borneo. Journal of Ecology 92:609?623. REEKIE, E. G. & BAZZAZ, F. A. 1989. Competition and patterns of resource use among seedlings of ?ve tropical trees grown at ambient and elevated CO2. Oecologia 79:212?222. SIM, J. W. S., TAN, H. T. W. & TURNER, I. M. 1992. Adinandra belukar: an anthropogenic heath forest in Singapore. Vegetatio 102:125? 137. TILMAN, D. 1987. Secondary succession and the pattern of plant dominance along experimental nitrogen gradients. Ecological Monographs 57:189?214. TURNER, I. M. 1991. Effects of shade and fertilizer addition on the seedlings of two tropical woody pioneer species. Tropical Ecology 32:24?29. UHL, C., BUSCHBACHER, R. & SERRA?O, E. A. 1988. Abandoned pastures in eastern Amazonia. I. Patterns of plant succession. Journal of Ecology 76:663?681. WALKER, L. R., ZARIN, D. J., FETCHER, N., MYSTER, R. W. & JOHNSON, A. H. 1996. Ecosystem development and plant succession on landslides in the Caribbean. Biotropica 28:566?576. WATANABE, T. 1997. Aluminum-induced growth stimulation in relation to calcium, magnesium, and silicate nutrition in Melastoma malabathricum L. Soil Science and Plant Nutrition 43:827?837. WATANABE, T. & OSAKI, M. 2001. In?uence of aluminum and phosphorus on growth and xylem sap composition in Melastoma malabathricum L. Plant and Soil 237:63?70. WATANABE, T. & OSAKI, M. 2002. Role of organic acids in aluminum accumulation and plant growth in Melastoma malabathricum. Tree Physiology 22:785?792. WEIGELT, A. & JOLLIFFE, P. 2003. Indices of plant competition. Journal of Ecology 91:707?720. WYATT-SMITH, J. 1963. Manual of Malayan silviculture for inland forests. ForestResearch InstituteofMalaysia,Kepong,KualaLumpur, Malaysia.