The desire to stop the current dramatic loss of biodiversity has been a major stimulus for many vegetation ecologists to unravel the mechanisms responsible for the coexistence of species. After the Rio Janeiro Convention many ecologists were convinced that nature conservation would gain strong societal support if they could prove that the loss of species would have important negative effects on the ecosystem functions that are relevant to society. I conclude that in order to understand such possible effects, it is necessary to analyse the effects of individual species on those ecosystem processes that we consider to be relevant in the context of specific questions. The great challenge for the near future is to scale the effects of plant species on their local environment up to the level of the whole planet, so that we learn about possible feedbacks that might regulate or destabilize those characteristics of the globe that are essential to our society.

Question: Is there evidence for limiting similarity in the timing of fruit production by a bird-dispersed plant community? Is the rate of fruit removal in each plant species inversely related to fruit availability in other species? Can simple measurements of fruit phenologies (i.e. temporal changes in fruit availability) obscure important fruit attributes that influence their removal by birds?

Location: Vancouver Island, British Columbia, Canada.

Methods: Periods of fruit availability were measured in ten woody angiosperm species for two years. In the second year, the fate of individual fruits was quantified to disentangle dates of fruit maturation, removal and mortality from measurements of availability.

Results: Null model analyses of fruit availability distributions showed no evidence for limiting similarity. However, fruit removal rates of most plant species were correlated with their relative abundance in the community, indicating fruits were removed more rapidly when other fruits were less abundant. Species with similar periods of fruit availability often had different dates of fruit maturation, rates of fruit removal and fruit persistence times, indicating fruit availability measurements can obscure important bird-fruit interactions.

Conclusions: Competition for dispersers appears to occur. However, it has not resulted in limiting similarity in fruit availability distributions. A likely explanation for this discrepancy is that fruit availability distributions often confound several important fruit attributes that can independently influence fruit removal by birds.

Nomenclature: Hitchcock & Cronquist (1994).

Questions: Which demographic and life-history differences are found among 95 sympatric tree species? Are there correlations among demographic parameters within this assemblage?

Location: Central Amazonian rain forest.

Methods: Using long-term data from 24 1-ha permanent plots, eight characteristics were estimated for each species: wood density, annual mortality rate, annual recruitment rate, mean stem diameter, maximum stem diameter, mean stem-growth rate, maximum stem-growth rate, population density.

Results: An ordination analysis revealed that tree characteristics varied along two major axes of variation, the major gradient expressing light requirements and successional status, and the second gradient related to tree size. Along these gradients, four relatively discrete tree guilds could be distinguished: fast-growing pioneer species, shade-tolerant subcanopy species, canopy trees, and emergent species. Pioneers were uncommon and most trees were canopy or emergent species, which frequently had low mortality and recruitment. Wood density was negatively associated with tree mortality, recruitment, and growth rates when all species were considered. Growth rates varied markedly among and within species, with pioneers exhibiting far faster and less variable growth rates than did the other species. Slow growth in subcanopy species relative to canopy and emergent trees was not a simple consequence of mean tree size, but apparently resulted from physiological constraints imposed by low-light and other conditions in the forest understorey.

Conclusions: Trees of Amazonian rain forests could be classified with some success into four relatively distinctive guilds. However, several demographic and life-history traits, such as those that distinguish early and late successional species, probably vary along a continuum, rather than being naturally grouped into relatively discrete categories.

Nomenclature: Ribeiro et al. (1999).

Abbreviation: BDFFP = Biological Dynamics of Forest Fragments Project.

Question: The evolution of general flowering and mast fruiting of dipterocarps in tropical rain forest has been explained by different mechanisms. We studied whether the abundance of flowering conspecifics influences the recruitment success of tropical forest trees in Borneo.

Location: Sepilok Forest Reserve, Sabah, Malaysia. Method: We examined the recruitment success of 17 species of Dipterocarpaceae in a 640-ha sample area over two flowering events in 2001–2002 using seed traps and quadrats. Seed predation experiments were used to investigate if post-dispersal seed survival was density-dependent.

Results: We found a negative relationship between the percentage of flowers abscised per individual and the number of flowering conspecifics and a positive relationship between the percent of flowering trees that produced viable seeds and the number of flowering conspecifics. However, we found no evidence of decreasing pre- and post-dispersal predation with increasing numbers of flowering conspecifics. High levels of flower abscission, and pre- and post-dispersal seed mortality, resulted in near-complete recruitment failure of most species, with only three species successfully recruiting in only one year. One of these, Parashorea tomentella, satiated seed predators over a large area, while Hopea beccariana and Shorea multiflora only recruited in small isolated clumps. Seed predation experiments suggest that post-dispersal seed survival was positively density-dependent in the short-term.

Conclusions: Increased density of flowering conspecifics may contribute to increased likelihood of successful cross-pollination during low intensity flowering events and, in some cases, to enhanced probability of short-term seed and seedling survival. Both processes may contribute to the evolution of mast fruiting and general flowering in Southeast Asian lowland dipterocarp forests.

Questions: How do biological invasion patterns of Larix kaempferi seedlings change with different microhabitats along an elevational gradient on a volcano? How are seedling attributes such as establishment, competitive ability and morphological plasticity, advantageous to the invasion of stressful disturbed areas?

Location: Mount Koma, Hokkaido, Japan.

Methods: Seed sowing experiments and natural seedling censuses were conducted with L. kaempferi and the dominant native tree Betula ermanii. Seed germination, seedling survival and allocation were investigated on three microhabitats (bare ground, Salix reinii patch and Larix understorey) in three elevational zones for three years.

Results: Seed germination was higher in Larix understorey than in bare ground and Salix patches, but did not differ between elevations. Survival rates were not different between elevations and microhabitats. Larix had a higher survival rate than Betula. Larix showed the highest natural seedling density in Salix patches, independent of elevational differences, while Betula density was nearly zero. Larix seedlings changed allocations between microhabitats, while the ratio of leaf to total biomass was constant. In bare ground Larix was more stunted and branched and increased its allocation to the roots. This form is adaptive to windy, nutrient-poor environments. Larix seedlings were taller and more slender in Salix patches, indicating that the priority was light acquisition in shaded habitats. Little change in Betula allocation was detected.

Conclusion: Invasive species establish themselves more efficiently than native species in every microhabitat investigated, especially at higher elevations, by having higher survival and growth rates derived from superior seedling performance. Plant communities on and above the tree line are modified by the biological invasion.

Abbreviations: BA = Bare ground; H = high elevational zones; L = Low elevational zone; LU = Larix understorey; M = Middle elevational zone; PFD = Photon flux density; RPFD = Relative photon flux density; SP = Salix reinii patch.

Nomenclature: Ohwi (1975).

Question: How does fire influence species richness and diversity in subtropical grassland in southern Brazil?

Location: Recurrently burned grassland in Porto Alegre, Brazil (30°03′ S, 51°07′ W; max. altitude 311 m a.s.l.) and abandoned grassland near São Francisco de Paula, Brazil (29°47′ S, 50°22′ W; ca. 900 m a.s.l.).

Methods: In the burned grassland, between-year changes in community composition and single-plot diversity, species number and vegetation structure were analysed in two consecutive years for plots with different time since last fire. Responses to fire of individual species were analysed. At the abandoned site, diversity, species number and vegetation structure were examined.

Results: Species number and small-scale species turnover were highest ca. one year after the burn, and decreased as caespitose grasses increased in cover with time since fire until reaching a stable, but less diverse state three to four years after a fire. The abandoned grassland showed higher dominance of caespitose grasses and lower richness and diversity.

Conclusions: Fire clearly leads to a short-term increase in species richness and diversity at the plot scale, as competitive interactions are being reduced and recruitment possibilities are high in early post-fire vegetation development. Overall community composition does not change after a fire. While small herbs seem to be slightly favoured in the early post-fire environment, no clear group of fire following species (absent in vegetation unburned for longer) was observed. The results indicate that the community is adapted to the current fire regime and is being maintained under the influence of fire.

Question and Location: How does soil burn severity and early post-fire tree composition affect long-term understorey vegetation dynamics in the coniferous forests of eastern Canada?

Method: Vegetation dynamics were assessed using paleoecological methods and a chronosequence analysis of extant stands. The relation between environmental factors and succession was evaluated using ordination techniques on the chronosequence data. Understorey succession was studied by regression analysis on the chronosequence data and through within-site Markovian transition probabilities between successive 1-cm layers of plant macroremains from soil organic matter profiles.

Results: Initial tree composition (Picea mariana and Pinus banksiana) had little effect on understorey composition. Soil burn severity (measured as the thickness of the residual forest floor humus) significantly affected temporal changes in understorey species. Following fires of high severity, stands underwent a gradual paludification with a net increase in Sphagnum and ericaceous shrubs (Ledum groenlandicum), and a decrease in feathermosses. Paludification was accelerated after low severity fires, which led to the dominance of Sphagnum less than 200 years after fire, and of L. groenlandicum shortly after fire. In situ paleoecological work confirmed results obtained with the chrono-sequence analysis.

Conclusions: One vegetation gradient related to time after disturbance is insufficient to account for the full complexity of long-term changes in understorey composition following fire. Current forestry practices that protect the forest floor humus may induce a premature paludification.

Abbreviations: AMS = Accelerated mass spectrometry; GCC = Global climate change; HS = High severity; LS = Low severity; TSF = Time since last fire.

Nomenclature: Marie-Victorin (1995) and Montgomery (1977) for vascular plants; Anderson et al. (1990) for bryophytes and Lévesque et al. (1988) for macrofossils.

Questions: Do past disturbance, soil nutrients, or species diversity predict the invasion success of the alien tree Pittosporum undulatum in an island montane rain forest? What are the consequences of its invasion for forest composition and species diversity?

Location: Blue Mountains, Jamaica.

Methods: Censuses of trees ≥ 3 cm DBH in permanent plots in four sites within ca. 7 ha; 1974–2004 (intensive sites) and in 16 plots within 250 ha; 1990–2004 (extensive plots).

Results: Pittosporum was unrecorded in the intensive sites before a severe hurricane in 1988: by 2004 all four sites were invaded. Pittosporum had invaded 25% of the extensive plots in 1990 and 69% in 2004, where its basal area increased from 0.5 ± 0.4 (SEM) m².ha⁻¹ in 1990 to 2.8 ± 1.3 m².ha⁻¹ in 2004. It had zero stem mortality and diameter growth rate exceeded that of native species fourfold. Pittosporum's basal area in the extensive plots in 2004 was positively related to the stand basal area damaged in the 1988 hurricane and negatively related to soil N concentrations. Pittosporum invasion was unrelated to stand-level tree species diversity in the extensive plots but as its basal area increased over time the basal area of native species and stand-level diversity declined.

Conclusions: There are no obvious functional attributes of Pittosporum unrepresented in the native tree flora although it has high photosynthetic efficiency compared with native trees. More widespread invasion of these forests by Pittosporum seems inevitable since hurricanes, which accelerated the invasion, affect these forests frequently.

Nomenclature: Adams (1972) except where other authorities given.

Question: How does competition by grasses, grazing by indigenous large herbivores, and their interaction affect the establishment, growth and survival of transplanted native woody and herbaceous seedlings on an abandoned agricultural field?

Location: West Coast Renosterveld, Cape Floral Region, South Africa.

Methods: Indigenous shrub seedlings were planted in different treatments, where either grass competition or herbivory or both were manipulated. Survival, growth and canopy cover of the seedlings were measured on a monthly basis over 14 months, and compared between treatments.

Results: Experimental transplanting of indigenous shrubs into an old field showed that most of the plants investigated competed for resources with grasses on the field, and competition negatively affected the seedlings throughout the experiment. Mortality was higher, and growth was reduced for seedlings exposed to grass competition. Herbivory alone had no significant impact on the target species with the exception of Olea europaea ssp. africana. There was no significant interaction between competition and herbivory.

Conclusion: Reduction of herbaceous competition significantly accelerates shrubland recovery on abandoned agricultural fields in renosterveld.

Abbreviations: NPNW = Not protected from herbivory and not weeded; NPW = Not protected from herbivory and weeded; PNW = Protected from herbivory and not weeded; PW = Protected from herbivory and weeded.

Nomenclature: Germishuizen and Meyer (2003).

Question: What are the shape, abundance, size, and structural characteristics of sapling patches in an old-growth Pinus jeffreyi-mixed conifer forest with a relatively intact disturbance regime?

Location: Sierra San Pedro Martir, Baja California, Mexico.

Methods: Regeneration was quantified by sampling sapling patches on seven 1200 m permanently marked line transects. For all patches intersected, tree size, species, age, and patch canopy cover were recorded. Patch structural characteristics were statistically compared to data from unbiased forest inventory plots. The minimum regeneration patch was defined as three saplings in a 49-m² area.

Results: We sampled 54 sapling regeneration patches. Patch size varied from 6.6 to 674.8 m². A small portion of the forest was in patches (patch fraction = 3.8%, patch abundance = 8.5 per ha). The majority of the patches were small; 64.8% were less than the mean patch area of 100.1 m². Patches were shaped as a rhombus. For all forest characteristics, mean values inside patches were significantly different than those from the unbiased forest inventory plots.

Conclusions: These forests have a fine-grained pattern of regeneration patches. Our largest patch size of 674.8 m² is the smallest reported in Pinus jeffreyi, P. ponderosa, and mixed conifer forests; other studies have probably had more difficulty delineating regeneration patches because of forest ingrowth from fire exclusion. Frequent fire, irregular seed crops, and seed dispersal by small mammals and birds could create this regeneration patch regime. High variation in nutrient availability after fire could also contribute to increased stand patchiness.

Nomenclature: Hickman (1993) for vascular plants; Kays & Wilson (2002) for mammals.

Abbreviation: Pj-mc = Pinus jeffreyi-mixed conifer; SSPM = Sierra San Pedro Martir.

Question: In non-linear physical, or chemical, systems dynamic instability limits predictability and external fluctuations cause interesting, and sometimes counter intuitive, effects. We ask how these generic properties of complex systems are seen in vegetation change.

Methods: An interacting particle system is used to simulate possible developments of a plant community that was observed for 30 years on a test area in the Lüneburger Heide (Germany). We investigate simulated trajectories for five plant types over several decades. The internal updating rule, simulating the interactions between the five plant types, is completely deterministic, the only sources of stochasticity being the initial conditions and the external climatic variations.

Results: Though the results of our simulation model share many aspects with those of stochastic models with probabilistic transition matrices, the solution manifold of our non-linear model goes beyond the possibilities of stochastic models. Among many stable developments, chaotic trajectories were also found. Climatic fluctuations increased the frequency of certain plant types. These properties of vegetation dynamics match the long-term field observations. They reflect generic features of complex systems.

Conclusions: The presented non-linear model can provide insights into the dynamics of vegetation change, even though the mechanisms observed in the real world are modelled only in a rather general way. The rich behavioural repertoire, resulting from deterministic internal dynamics, can simulate observed properties that are beyond the possibilities of models with probabilistic internal dynamics.

Nomenclature: Corley et al. (1981); Corley & Crundwell (1991); Tutin et al. (1964–1983); Grolle (1983); Grolle & Long (2000).

Abbreviations: CA = Cellular automaton; IPS = Interacting particle system; M = Moss.

Question: Do water gradients produce patterns of responses to stress and competition similar to those induced by nutrient gradients?

Location: French Alps.

Methods: We established a split-plot design in a calcareous grassland, with watering and fertilization as main plot treatments and competition as subplot treatment. We followed individual and competitive responses of transplants of the three potential dominant grass species: Bromus erectus, Brachypodium rupestre and Arrhenatherum elatius, in all plots during two growing seasons. Changes in natural relative abundances of the three grass species were also monitored.

Results: The growth and the relative abundance of A. elatius were primarily stimulated by nutrient addition and those of B. rupestre by water addition, whereas B. erectus decreased in abundance and had a very low flexibility with enhanced resource supply. Competition intensity increased for all species with both watering and fertilization and the ranking in competitive responses did not change with treatments: A. elatius > B. rupestre > B. erectus.

Conclusions: Patterns of dominance were efficiently explained by stress tolerance abilities and competitive responses for dry and poor sites, and wet and rich sites for B. erectus and A. elatius respectively, whereas competitive responses were poor predictors of dominance for B. rupestre in wet and nutrient-poor sites. Further studies are needed to assess the potential role of other processes, such as increasing competitive effect on light with increasing age as well as interference, to explain the dominance of this conservative competitor type of species in wet and nutrient-poor sites.

Abbreviations: C = Control plot; F = Fertilized plot; PAR = Photosynthetically active radiation; RNE = Relative Neighbour Effect index; W = Watered plot; WF = Watered and Fertilized plot.

Nomenclature: Tutin et al. (1964–1980).

The invasion of a target community by a non-indigenous plant species includes the stages of arrival, establishment and spread, which tend to depend on different characteristics of the invasive species and its context. While the mechanisms behind the invasion of highly disturbed ecosystems are well known, our understanding of the invasion process in undisturbed or weakly disturbed ecosystems is much more limited. Here we propose that, once a non-indigenous species has arrived to a new ecosystem and become established, the likelihood that it spreads, and thus becomes invasive, may depend on just one or very few characteristics, called ‘triggering attributes’ (TA). We propose that a TA is a vegetative or regenerative attribute discontinuously distributed in comparison to the resident community. This attribute allows the species to benefit from a resource that is permanently or temporarily unused by the resident community. We present an original study case and examples from the literature to illustrate our approach, and we also propose some ways to test it in different ecosystems.

Abbreviation: RGR = Relative growth rate; SLA = Specific leaf area; TA = Triggering attribute.