The endemic New Caledonian conifer Agathis ovata occurs as an emergent tree in fire-prone shrublands (maquis), and fire-sensitive rainforest. Growth, survivorship and recruitment over 5 yr were compared for populations from forest and maquis on ultramafic substrates in New Caledonia to investigate whether demographic behaviour varied in response to the strongly contrasting forest and shrubland environments.

Growth of seedlings and of small (30–100 cm height) and large (100 cm height; 5 cm DBH) saplings was slow, but varied significantly among stages, site types and years. The greatest difference in growth rates was among stages, seedlings growing 0.34 cm.yr⁻¹, small saplings 1.06 cm.yr⁻¹ and large saplings 2.13 cm.yr⁻¹. Tree DBH increased by only 0.05 cm.yr⁻¹ and, based on these rates, individuals with DBH of 30 cm are estimated to be more than 700 yr old. Few trees (3.5%) produced cones in any year and seedling recruitment was low, but some recruitment was recorded each year in both maquis and forest. Rates of recruitment per parent were highest in forest (1.28.yr⁻¹, cf 0.78.yr⁻¹), but the higher density of trees in maquis meant that overall recruitment was greater there (92 ha⁻¹.yr⁻¹, cf 56 ha⁻¹.yr⁻¹). Seedling mortality ranged from 0.9 to 2.9% among years with no significant difference between maquis and forest. No sapling mortality was recorded, but annual tree mortality ranged from 0 to 1.4%. Evidence from a recently burned site indicated that while trees may survive fire, seedlings and saplings do not. Post-fire seedling recruitment per ha from surviving trees was four times lower than in unburned sites, but growth rates were four times higher.

Similar demographic attributes, including high survivorship, low growth rate and low rates of recruitment over a long reproductive life, characterize Agathis ovata populations in both maquis and rainforest in New Caledonia and are indicative of a broad tolerance of light environments that is unusual among tree species. These demographic attributes help to explain the long-term persistence of the species in these strongly contrasting habitats.

Abbreviations: CEC = Cation exchange capacity; NDF = neighbourhood density function.

An intense lateral blast devastated Mount St. Helens in 1980, but forest understory species survived in some north-slope ‘refugia’. We explored the effects of refugia on colonization of barren pumice in 1997 and 1998, 18 yr after the eruption. The seed rain of 23 colonizers came mostly from populations that had previously established in refugia. Parachutists had small, vagile seeds, parasailors had winged seeds, and tumblers were blown along the ground. The latter two groups are heavier and dispersed more slowly, but are more likely to survive. The proportion of the vegetation represented by wind-dispersed species increased with distance from refugia. Parachutist's density declined with time and proximity to refugia. As vegetation adjacent to refugia developed, populations of parasailors and tumblers expanded, foreshadowing their dominance in more remote pumice. Refugia played a critical role in determining the rate and course of succession by providing fertile islands that permitted pioneers and dry meadow species to establish near barren pumice. Species that survived in refugia played a negligible role in colonization. This study showed that when refugia contrast sharply with new substrates, they accelerate recovery by facilitating the invasion of pioneer species.

Nomenclature: Taxonomy follows the Integrated Taxonomic Information System (ITIS) ( http://www.itis.usda.gov).

In French Guiana, inselbergs in the form of granite outcrops rise abruptly from the surrounding rain forest. They constitute isolated islands of a special type of vegetation restricted to this peculiar substrate. Shrub granitic vegetation, organized in thickets on open exposed rocks of inselbergs, are described using the Braun-Blanquet method combined with Correspondence Analysis. This phytosociological study revealed only one particular shrub community on each inselberg, including predominantly evergreen and sclerophyllous shrubs, especially microphanerophytes, belonging to the Clusiaceae, Myrtaceae and Bombacaceae. These outcrop communities exhibit species endemic to the Guianas region and also species rare in French Guiana. Affinities with flora of other inselbergs and vegetation types in South America are examined and discussed. Reasons for observed floristic and structural changes in each community are also discussed.

Abbreviation: CA = Correspondence Analysis.

Nomenclature: Boggan et al. (1997).

We documented the occurrence of a 1934 blowdown in a subalpine forest in northwestern Colorado, USA. Prior to the blowdown, the stand was dominated by old-growth Picea engelmannii - Abies lasiocarpa forests. Although blowdowns are believed to trigger outbreaks of Dendroctonus rufipennis (spruce beetle), we found no detectable increase in beetle caused mortality. Forest recovery was by both release of the previously suppressed regeneration and by new seedling establishment. Both recovery pathways were dominated by Abies. The blowdown thus caused a shift in species dominance from Picea to Abies; 65 yr after the blowdown, the fallen logs and tip up mounds continue to provide favourable habitat for seedling establishment of both species. The present study shows that the legacy of blowdowns can influence forest dynamics for decades following the disturbance event.

We test to what extent mean environmental conditions and environmental heterogeneity are related to species richness in a regular geographical grid system (UTM) of 10 km × 10 km in the NE Iberian Peninsula (i.e. Catalonia, ca. 31900 km²). Species richness of each UTM quadrat was estimated by compiling a large database (more than a million records) from bibliographic references and atlases. Mean environmental conditions of each quadrat were derived from climatic maps. Environmental heterogeneity was estimated from the diversity of geological substrates and climatic classes in each quadrat. The increase in effective (real) area due to topographic complexity was also considered (derived from the digital elevation model). The statistical analysis was performed by a weighted analysis of deviance assuming a negative binomial error distribution. The results suggest that species richness in the study area is a function of both within-quadrat heterogeneity (specifically, effective area, heterogeneity of geological substrates, heterogeneity of January temperature) and mean environmental conditions (mean annual temperature, Thornthwaite moisture index and aspect). All these parameters showed a positive relationship with species richness. Quadrat heterogeneity accounted for ca. 2/3 of the explained deviance, suggesting the importance of environmental heterogeneity when using a geographical grid system. The study fits well with earlier results on the importance of climatic parameters on plant species richness and provides one of the few rigorous, quantitative, coarse-scale studies testing environmental heterogeneity in plant species richness.

Statistical models of the realized niche of species are increasingly used, but systematic comparisons of alternative methods are still limited. In particular, only few studies have explored the effect of scale in model outputs. In this paper, we investigate the predictive ability of three statistical methods (generalized linear models, generalized additive models and classification tree analysis) using species distribution data at three scales: fine (Catalonia), intermediate (Portugal) and coarse (Europe). Four Mediterranean tree species were modelled for comparison. Variables selected by models were relatively consistent across scales and the predictive accuracy of models varied only slightly. However, there were slight differences in the performance of methods. Classification tree analysis had a lower accuracy than the generalized methods, especially at finer scales. The performance of generalized linear models also increased with scale. At the fine scale GLM with linear terms showed better accuracy than GLM with quadratic and polynomial terms. This is probably because distributions at finer scales represent a linear sub-sample of entire realized niches of species. In contrast to GLM, the performance of GAM was constant across scales being more data-oriented. The predictive accuracy of GAM was always at least equal to other techniques, suggesting that this modelling approach is more robust to variations of scale because it can deal with any response shape.

Abbreviations: GLM = Generalized Linear Model; GAM = Generalized Additive Model; CTA = Classification Tree Analysis; ROC curve = Receiver Operating Characteristic curve; AUC = Area Under the Curve.

The interactions between plants of different species, age or size play an important role in the dynamics of an ecosystem and can induce specific structures. These interactions can be studied by analysing the spatial structure of the corresponding bivariate patterns. The intertype L₁₂-function has recently been successfully used in many papers for that purpose. However, when interpreting the results obtained with ecological data, at least two different null hypotheses – independence or random labelling – can be appropriate, depending on the context of the study and the nature of the data. As these two hypotheses correspond to different confidence intervals, an inappropriate choice of the null hypothesis can lead to misinterpretations of biotic interactions when studying ecological data. This problem has rarely been mentioned in the literature.

In this paper we clarify the differences between these two null hypotheses, and illustrate the risk of misinterpretation when using an inappropriate null hypothesis. We review the main characteristics of these two hypotheses, and analyse the spatial structure of both real data from forest stands and simulated virtual stands of different structures. We demonstrate that the risk of misinterpretation is quite high, and that extreme misinterpretations, i.e. cases leading to opposite conclusions in terms of spatial interaction, can occur in a significant number of cases. We therefore propose some guidelines to help ecologists avoid such misinterpretations.

Nomenclature: Boggan et al. (1997); Rameau et al. (1989).

Abbreviation: CSR = Complete spatial randomness.

Variation partitioning by (partial) constrained ordination is a popular method for exploratory data analysis, but applications are mostly restricted to simple ecological questions only involving two or three sets of explanatory variables, such as climate and soil, this because of the rapid increase in complexity of calculations and results with an increasing number of explanatory variable sets. The existence is demonstrated of a unique algorithm for partitioning the variation in a set of response variables on n sets of explanatory variables; it is shown how the 2n − 1 non-overlapping components of variation can be calculated. Methods for evaluation and presentation of variation partitioning results are reviewed, and a recursive algorithm is proposed for distributing the many small components of variation over simpler components. Several issues related to the use and usefulness of variation partitioning with n sets of explanatory variables are discussed with reference to a worked example.

Abbreviations: AVE = Average variation explained; CCA = Canonical Correspondence Analysis; CO = Constrained ordination; DCA = Detrended Correspondence Analysis; IU = Inertia units; pCO = Partial constrained ordination; RDA = Redundancy Analysis; TI = Total inertia; TVE = Total variation explained; VE = Variation explained.

Soil seed bank and floristic diversity were studied in a forest of Quercus suber, a forest of Quercus canariensis and a grassland, forming a vegetation mosaic in Los Alcornocales Natural Park, southern Spain. The soil seed bank was estimated by the germination technique. In each community patch, diversity, woody species cover and herbaceous species frequency was measured. Three biodiversity components – species richness, endemism and taxonomic singularity – were considered in the vegetation and the seed bank. Forest patches had a soil seed bank of ca. 11200–14100 seed.m⁻² and their composition had low resemblance to (epigeal) vegetation. The grassland patch had a more dense seed bank (ca. 31800 seed.m⁻²) and a higher index of similarity with vegetation, compared with the forests nearby. The complete forest diversity was 71–78 species on 0.1 ha, including 12–15 species found only in the seed bank; the grassland species richness was higher (113 species on 0.1 ha). We discuss the role of soil seed banks in the vegetation dynamics and in the complete plant biodiversity of the mosaic landscape studied.

Nomenclature: Valdés et al. (1987).

We studied the relationship between plant N:P ratio, soil characteristics and species richness in wet sedge and tussock tundra in northern Alaska at seven sites. We also collected data on soil characteristics, above-ground biomass, species richness and composition. The N:P ratio of the vegetation did not show any relationship with species richness. The N:P ratio of the soil was related with species richness for both vegetation types. Species richness in the tussock tundra was most strongly correlated with soil calcium content and soil pH, with a strong correlation between these two factors. N:P ratio of the soil was also correlated with soil pH. Other factors correlated with species richness were soil moisture and Sphagnum cover. Organic matter content was the factor most strongly correlated with species richness in the wet sedge vegetation. N:P ratio of the soil was strongly correlated with organic matter content. We conclude that N:P ratio in the vegetation is not an important factor determining species richness in arctic tundra and that species richness in arctic tundra is mainly determined by pH and flooding. In tussock tundra the pH, declining with soil age, in combination with Sphagnum growth strongly decreases species richness, while in wet sedge communities flooding over long periods of time creates less favourable conditions for species richness.

Nomenclature: Hultén (1968).

Structural and compositional changes were analysed over the course of 400 yr of post-fire succession in the sub-boreal forests of west-central British Columbia. Using a chronosequence of 57 stands ranging from 11 to 438 yr in age, we examined changes in forest structure and composition with complementary PCA and DCA ordination techniques. To determine stand ages and timing of tree recruitment, approximately 1800 trees were aged. Most early successional forests were dominated by Pinus contorta, which established rapidly following fire. Abies lasiocarpa and Picea glauca ×engelmannii were also able to establish quickly, but continued to establish throughout the sere. Few Pinus contorta survived beyond 200 yr, resulting in major changes in forest structure. In some stands P. contorta never established, which led to considerable variation among stands less than 200 yr old. The oldest forests converged on dominance by Abies lasiocarpa. Vascular plant diversity decreased during succession whereas canopy structure became more complex as gap dynamics developed. Although these sub-boreal forests contain few tree species, successional changes were pronounced, with structure changing more than composition across the chronosequence.

Abbreviations: CWD = Coarse Woody Debris; DCA = Detrended Correspondence Analysis; DBH = Diameter at Breast Height; PCA = Principal Component Analysis.

Nomenclature: Douglas et al. (1989–1994).

In this field study we analysed the regional and local scale effects of disturbance and climate on altitudinal treelines dominated by Nothofagus pumilio in northern Patagonia. We compared two regions west and east of the Andes at 40° S, slopes with warm vs cool aspects and undisturbed vs locally disturbed treelines. This spatial framework allowed us to test (1) for differences among treelines affected by different types of local disturbance and (2) the traditional hypothesis that low temperature limits treeline. Contingency tables and ANOVA showed that local disturbance occurred more frequently than expected on slopes with cool aspects, steep slope angles and concave slope configuration. Disturbed treelines were locally lowered with longer ecotones and lower krummholz growth rates and vegetation cover than undisturbed treelines. Three-way ANOVA showed the significant influences of study area (regional climate) and aspect (local climate) on treeline elevation, krummholz growth rates and density, tree density and vegetation cover, while accounting for local disturbance. Treeline elevations were higher east of the Andes reflecting the more continental climate in Argentina than in Chile, plus regional impacts of volcanic eruptions. Tree density and vegetation cover were greater west of the Andes reflecting greater precipitation in Chile. Within study areas, local climate had different influences on treeline elevations and krummholz growth rates west and east of the Andes. We predict that increased tree growth and upslope advance of treeline in response to global warming is more likely in Chile than in Argentina near 40° S, unless precipitation increases east of the Andes. To test these predictions, we recommend research be stratified to account for the influences of local disturbance, which may confound climatic impacts. In northern Patagonia, suitable control (undisturbed) study sites will most likely be found at upper slope positions with low slope angles, simple microtopography and straight topographic configuration.

Nomenclature: Muñoz (1966).

Abbreviation: TRMI = Topographic relative moisture index.

The influence of canopy trees and shrubs on understorey plants is complex and context-dependent. Canopy plants can exert positive, negative or neutral effects on production, composition and diversity of understorey plant communities, depending on local environmental conditions and position in the landscape. We studied the influence of Prosopis velutina (mesquite) on soil moisture and nitrogen availability, and understorey vegetation along a topographic gradient in the Sonoran Desert. We found significant increases in both soil moisture and N along the gradient from desert to riparian zone. In addition, P. velutina canopies had positive effects, relative to open areas, on soil moisture in the desert, and soil N in both desert and intermediate terrace. Biomass of understorey vegetation was highest and species richness was lowest in the riparian zone. Canopies had a positive effect on biomass in both desert and terrace, and a negative effect on species richness in the terrace. The effect of the canopy depended on landscape position, with desert canopies more strongly influencing soil moisture and biomass and terrace canopies more strongly influencing soil N and species richness. Individual species distributions suggested interspecific variation in response to water- vs. N-availability; they strongly influence species composition at both patch and landscape position levels.

The response of fragmented vegetation to human impact has been analysed in degraded and degrading areas in Tanzania (Lake Manyara). Phytosociology was integrated with GIS and remote sensing data as follows: (1) a land cover/land use map was obtained by analysing remote sensing data and conducting field verification; (2) phytosociological relevés were randomly sampled in woodland vegetation patches using the map; (3) the pattern of land cover/land use around the relevés was described; (4) gradients of land use intensity (human impact) were obtained based on the descriptions; (5) the response of vegetation types to impacts of gradients was calculated using fuzzy set theory.

Two complementary gradients of human impact were defined based on land cover/land use analysis of the remotely sensed data: one related to cultivation intensity and the other to grazing intensity. Response functions of vegetation types (defined by numerical classification) to these gradients demonstrated that the vegetation types are strongly related to the degree of human impact and that the corresponding vegetation patches show different degrees of permeability to the species of the surrounding landscape.

Nomenclature: Anon. 1954–.

The nearest-neighbour technique is used to infer competition and facilitation between the three most abundant species in a semi-arid region of western South Africa. Relationships among the shrubs Leipoldtia schultzei and Ruschia robusta, which are leaf-succulent members of the Mesembryanthemaceae (‘mesembs’) and Hirpicium alienatum a non-succulent Asteraceae, were compared on two adjacent sites with different histories of browsing intensity. Competition was more prevalent and more important than facilitation. The only evidence for facilitation was found at the heavily-browsed site where the palatable Hirpicium was larger under the unpalatable Leipoldtia. Generally the prevalence and importance of competition was reduced at the heavily-browsed site. Strong evidence was obtained for intraspecific competition in each of the three species; also, competition was evident between the two mesembs, where Leipoldtia was competitively dominant over Ruschia, although neither species inhibited Hirpicium. Minimal competition between the mesembs and the asteraceous shrub was interpreted in terms of differentiation in rooting depth, and competition within the mesembs, in terms of overlap in rooting depth. The mesembs had the bulk of their roots in the top 5 cm of soil, while the asteraceous shrub had the bulk of its roots, and all its fine roots, at greater depths. The shallow-rooted morphology of the mesembs is well adapted to utilize small rainfall events, which occur frequently in the Succulent Karoo, and do not penetrate the soil deeply. Modifications of existing methods are applied for analysing nearest-neighbour interactions.

Nomenclature. Taxonomic nomenclature follows Arnold & De Wet (1993) except for the use of the term Mesembryanthemaceae.

The flora of the deciduous forests at the base of the north Estonian limestone escarpment is species rich, with an exceptionally high number of rare bryophyte species. Relationships between species richness of bryophyte and herb layers and biotic and environmental conditions were studied, using General Linear Mixed Models. Human disturbance (waste deposit, tree damage etc) was significantly negatively correlated with species richness of both plant layers. Soil nitrogen content was negatively and soil water retention positively correlated with bryophyte species richness, while herb richness was unrelated to soil factors. After eliminating the effects of environment, negative correlations in species richness and cover between the bryophyte and herb layers were discovered on finer scales (1 m²), referring to biotic interactions. This relationship was obscured with the simple correlation analysis. On the other hand, the positive correlation in species pools between the bryophyte and herb layers (0.1 ha) was insignificant. The species pools of both bryophyte and herb layers were significantly positively correlated with the species richness of the tree layer. In summary, bryophyte and herb layer richness responded differently to environmental conditions, but human disturbance significantly decreased the richness of both layers. Due to the uniqueness and small area of these forests we recommend protection and restoration of disturbed sites.

Nomenclature: Ingerpuu et al. (1994) for bryophytes, Leht (1999) for phanerogams.

Abbreviation: GLMM = General Linear Mixed Model.