Collembolan Assemblages Research Articles

Tree crowns broken off by windstorms are an unstable life raft for Collembola

There are a number of ways to clear the aftermath of a windthrow disturbance of forest stands, the most common practice being to remove all broken trees and broken-off crowns lying on the ground. This practice leads to complete exposure of the soil, which deprives soil invertebrates (including Collembola) of the protection of trees that affords them a chance of surviving. Accordingly, following a windthrow disturbance of pine stands in 2017, a three-year study of collembolan assemblages was undertaken in stands spared from salvage logging. We aimed to test the effect of three different levels of disturbance (severely, moderately and least disturbed stands with a canopy cover of 0–20%, 20–60% and 60–90%, respectively) on the survival of Collembola assemblages and to determine its association with changes in the soil environment and in the LAI index. Additionally, in the severely and moderately disturbed stands, Collembola were sampled between crowns of fallen trees and under the crowns. There were no significant differences in density, species richness and proportions of individuals of belowground “soil” and aboveground “epedaphic” species between the Collembolan assemblages that were associated with the degree of windthrow disturbance and time since disturbance. The study confirmed the presence of a significantly higher number of species and proportion of “epedaphic” species, and a lower proportion of “soil” species in the assemblages sampled under fallen tree crowns than between crowns. Analysis of principal response curves (PRC) yielded unexpected results as it indicated that these differences were significant only in the first year post-disturbance, thus suggesting a very short-lasting protective effect of tree crowns on Collembola, RDA analysis with preselected factors from environmental variables of interest (LAI of standing and fallen tree crowns, soil respiration, soil temperature and humidity, soil pH and soil nitrogen and carbon content) indicated the LAI index as significant for the Collembolan assemblages in the first yearpost-disturbance, soil moisture in the second year, and soil temperature in the third year. This sequence of significant indices over a three year period is compatible with the fallen crowns becoming more and more thinned as a result of needles falling off (from shade to full exposure to sunlight). We nevertheless postulate that at least some trees or their crowns lying on the ground should be left in place during clearance of windthrow-affected tree stands to facilitate restoration of the soil biota.

Urbanization and greenspace type as determinants of species and functional composition of collembolan communities

Urban regions are rapidly expanding worldwide resulting in biotic homogenization and loss of ecological functions in urban ecosystems due to management practices targeting at satisfying aesthetic and health demands of urban residents. These practices also modify living conditions and food recourses of soil invertebrates thereby affecting the structure and functional diversity of soil animal communities including collembolans. Here, we assessed the response of the community composition and functional diversity of collembolans as a major component of soil food webs to urbanization (suburban vs urban region) and greenspace types (including forest and four park-associated greenspaces: lawn, lawn with shrubs, lawn with trees, and lawn with shrubs and trees). Our results highlight that both urbanization and greenspace type significantly affect soil properties and community structure of collembolans. The negative effect of urbanization and park-associated greenspaces on species and functional composition of collembolan communities were likely due to both changes in soil abiotic conditions and bacterial community composition, whereas the reduction of collembolan functional traits likely resulted from changes in soil abiotic conditions and fungal community composition. In park-associated greenspaces richness and diversity of bacterial communities were highest in lawns with trees and lowest in forests. By contrast, species richness and diversity of fungal communities were highest in lawns with shrubs, but, similar to bacteria, lowest in forests. Community composition and functional traits of collembolans were more homogeneous in urban than suburban greenspaces pointing to reduced functioning of collembolan assemblages in urban areas. Overall, our results suggest that changes in soil properties and bacterial communities caused by urbanization and greenspace type are important factors contributing to taxonomic homogenization of collembolan communities, while the loss of functional traits of collembolan communities in urban greenspaces is likely caused by changes in soil properties and fungal community composition.

Spontaneous regeneration of Collembola assemblages in a raised bog after human-induced disturbance

Peatlands in anthropogenically transformed landscapes play an important role in biodiversity conservation. Most Baltic raised bogs have been completely destroyed by drainage and peat extraction. Collembolan and plant assemblages from an old peat mine (post-excavation pits and dykes overgrown by pine trees) were compared to those of intact ombrotrophic raised bog habitats (open bog and bog-pine forest) 70 years after excavation cessation. The plant cover and collembolan structure in regenerating pits were similar to those of the assemblages in natural open raised bogs. Rare tyrphobiont and tyrphofilous Collembola species occurred in both habitats. The assemblages in open bog habitats differed significantly from those in natural pine bog forest and dykes overgrown by pine trees. The assemblages of the natural open raised bog and pine bog forest were the most diverse and included numerous highly specialized Collembola species. The assemblages occurring on wooded dykes were significantly different from those at other peatland sites and were more similar to those of upland pine forests. The analysis of environmental factors (determined by moss Ellenberg numbers) showed that moisture was the most important factor affecting Collembola occurrence in natural and disturbed habitats. The occurrence of moss species also had a significant influence on the species composition of the collembolan assemblages. Our study revealed that seventy years of spontaneous ecosystem regeneration after peat excavation have led to habitat restoration for peat bog mesofauna. These results have important implications for conservation and restoration planning in disturbed peatlands.

Collembolan Assemblages Response to Wild Boars (Sus scrofa L.) Rooting in Pine Forest Soil

Collembola are an important component of soil communities in all terrestrial ecosystems. In temperate coniferous forests, they are one of the most numerous invertebrate groups, and disturbances that change their density and structure may have negative effects on soil fertility and productivity. Our goal was to determine whether intensive rooting in the forest floor by wild boars affects edaphic Collembola. Soil samples from three paired rooted and non-rooted plots in Scots pine stands were taken twice a year to study the impact of such bioturbation on forest collembolan assemblages. Substantial changes in the taxonomic and functional structure of the collembolan assemblages were identified in all disturbed plots. The abundance and number of species significantly decreased in the bioturbated forest floor. The shares of atmobiotic and hemiedaphic springtails increased at the expense of epedaphic forms. Most of the differences were evident shortly after grubbing but were not significant a few months later. The decline in moisture in disturbed soil could be an explanatory factor causing the differences in the structure and abundance of collembolan assemblages between the bioturbated and intact plots. Our study revealed that large mammals ubiquitous in forest ecosystems can be an important disturbing factor for soil microarthropods. Intensive wild boar rooting in the forest floor had a strong negative effect on the occurrence and abundance of Collembola. This kind of bioturbation also modified the functional structure of assemblages, which in turn may have important consequences for the soil food web and above- and belowground interactions.

Does the enrichment of post-arable soil with organic matter stimulate forest ecosystem restoration—A view from the perspective of three decades after the afforestation of farmland

The restoration of forest ecosystems on former agricultural land faces numerous challenges resulting mainly from differences in soil properties between post-arable and continuously forested ground. We revisited a 34-year-old but very well-documented experiment with enrichment in organic matter of the arable sandy soil aiming to accelerate humus formation and soil biota regeneration in Scots pine plantation. The goal of our study was to determine whether the addition of a mixture of pine bark with sawdust or lupine cultivation would change the selected soil properties and, as a result, would shift soil microarthropods of post-agricultural plantations towards those occurring in pine monocultures on continuously forested land. The effect of tillage depth was also studied.The soil properties of the experimental plantation differed significantly from those of the forest soil with respect to the pH, nitrogen (N) content, C/N ratio, sum of exchangeable base cations and degree of soil saturation with base cations. Moreover, the diversity and taxonomic and functional structures of Collembola assemblages differed considerably from the assemblages of pine monoculture of the same age growing on continuously forested land. The expectation of the experiment was not met; however, a certain significant ameliorative effect on some soil properties was observed with some treatments. The total organic carbon content (TOC), C/N ratio and cation exchange capacity increased as a result of the addition of bark and sawdust. In plots where deep tillage was applied, the TOC and N contents were lower than those in plots with shallow soil cultivation. A significant increase in the number of hemiedaphic forms at the expense of euedaphic and atmobiotic forms was observed in plots with lupine, while a mixture of pine bark and sawdust caused an increase in the number of euedaphic forms.Soil pH was identified as a key parameter that determined the difference between forest and post-agricultural collembolan assemblages. The variability of the structure of post-agricultural assemblages depends on N and exchangeable Mg2+ contents and hydrolytic acidity. The enrichment of post-arable soil with organic matter and tillage had long-lasting effects on soil properties but did not stimulate restoration of forest collembolan assemblages on former farmland.

Functional collembolan assemblages induce different plant responses in Lolium perenne

Interactions between functional groups of soil fauna and plants are poorly explored although they drive functional processes such as nutrient availability and therefore plant performance. Here, we investigated the separated and combined effects of two collembolan functional groups on soil properties and growth of Lolium perenne, a typical grass species from temperate grasslands. Under microcosm conditions, we established four treatments based on the presence and combination of two collembolan functional groups: (1) control without Collembola; (2) epedaphic species; (3) euedaphic species; (4) epedaphic + euedaphic species. After five months of experiment, we measured the effect of those treatments on both nutritional and morphological variables of L. perenne and on soil properties. Collembolan presence stimulated plant performance. Individuals of L. perenne growing with euedaphic species presented higher numbers of leaves and nutrient contents compared with individuals growing with epedaphic species. Further, the combination of both collembolan functional groups enhanced plant performance and soil nutrient availability, demonstrating that a functionally diversified soil fauna assemblage cause overyielding of ecological processes. Our results provide evidences of complementarity interactions between different functional groups of soil fauna causing overyielding of primary production.

Seven Decades of Spontaneous Forest Regeneration after Large-Scale Clear-Cutting in Białowieża Forest do not Ensure the Complete Recovery of Collembolan Assemblages

The long-term effects of large-scale disturbance on forest ecosystem processes and structure are poorly understood. To assess the effects of large-scale clear-cutting on the taxonomic and functional structure of collembolan assemblages, 18 plots were established in the Polish part of Białowieża Forest. All plots, situated in a mixed Tilio-Carpinetum broad-leaved forest, had eutrophic Cambisol developed on rich glacial deposits. The Collembola assemblages in the stands that had naturally regenerated on large-scale clear-cuts performed at the beginning of the 20th century were compared to those in old-growth forests (i.e., the endpoint of stand development following stand-replacing disturbance). Collembolans, one of the most numerous soil microarthropods, are successfully used to assess the consequences of forest management and ecosystem restoration. Our study tested whether seven decades of spontaneous forest development after large-scale anthropogenic disturbance ensures the complete recovery of the soil Collembola. Using complementary taxonomic and life-form approaches, we provide evidence that the collembolan assemblages associated with the tree stands that had spontaneously developed in large harvesting plots distinctly differed from those in old-growth deciduous forests in this region despite seven decades of regenerative forest succession. The species diversity of the assemblages in the naturally regenerated tree stands was significantly lower, and their life-form structure was noticeably different from those in the reference forests. Moreover, the shift in the functional group structure of the collembolan assemblages in the stands that had regenerated after clear-cutting indicates that their activity seven decades after disturbance is concentrated mainly on the decomposition of the litter in the upper layers, whereas the processes controlled by these organisms in the deeper soil layers are not fully restored.

Forest-management types similarly influence soil collembolan communities throughout regions in Germany – A data bank analysis

European forest management practices create broad-scale mosaics of tree-species composition and developmental stages, inducing gradients in environmental conditions at landscape scales. However, consistent effects of specific forest-management types on general characteristics of soil microarthropod assemblages could not always be determined. This study aimed to reveal generalizable effects of similar forest-conversion measures using collembolan data extracted from the German soil-zoological data warehouse “Edaphobase”. Data of collembolan assemblages and environmental variables from forest-conversion studies were analyzed from three forest types – deciduous, coniferous and mixed forests – in two German regions (a northern and a southern) with various sites in two areas each. We hypothesized that forest management has a strong effect on soil Collembola and the different forest types induce consistently dissimilar assemblages independent of region. Abundance, diversity and species richness of collembolan assemblages were compared between forest types. Community composition was analysed with cluster analysis and canonical ordinations to reveal similarity patterns and the driving environmental variables. Contrary to our hypothesis, forest conversion caused only small differences in quantitative community variables, most likely due to site-specific influences and high variability. More than a third of the considered species showed preferences for stands with either deciduous or coniferous trees. The collembolan assemblages of the two regions differed from each other, indicating separate species pools. Nonetheless, as opposed to the quantitative metrics, different species compositions did consistently occur between forest types across regions, revealing a gradient from deciduous over mixed to coniferous stands. Collembolan assemblages were primarily associated with dominant tree species (forest stands) and humus forms, while quantitative environmental parameters had low explanatory power. Therefore, forestry measures induce qualitative differences in the soil-habitat, leading to habitat-specific environmental filtering of the regional collembolan species pools and a mosaic of soil microarthropod communities at the landscape level.

Springtails (Collembola) in the Subpolar Landscapes of the Northern Hemisphere

An overview of previously published and new information on the collembolan fauna and assemblage structure in the polar desert zone is presented. So far, 71 springtail species from 37 genera and 11 families have been reliably recorded within the zone. Eleven species are added to the previously known fauna of Franz Josef Land, and the very north of Novaya Zemlya has been surveyed for the first time. Even the much better known fauna of Bolshevik Island, Severnaya Zemlya is enriched by 3 species. The known species richness of springtails of Ellef Ringnes Island, Canadian Arctic Archipelago, is also increased from 8 to 13 species. Most genera except Folsomia and Hypogastrura include only 1, more rarely 2 species in each study region of the polar desert zone. Species with circumpolar distribution patterns comprise more than 60% of the total list, but only 10 species are common to all the three provinces; this obviously indicates a certain regional specificity. Besides, the faunas of different provinces (and regions within a province) vary markedly in the proportion of species with more southern distribution patterns. Nevertheless, the collembolan assemblages in all the three provinces of the polar desert zone are rather similar at the structural level, this being a direct consequence of the general depletion of the complexes against the background of high total abundance and less pronounced habitat specificity of the common species.

Soil arthropod responses to controlled traffic in vegetable production

In this study, the effects of a controlled traffic (CT) management system on the soil arthropod assemblage in a Red Ferrosol (krasnozem) soil in north-west Tasmania were examined. Individual soil cores were collected at three depths over two seasons (winter and spring, 17 months apart), and soil fauna extracted using Berlese-Tullgren funnels. All arthropods were identified to the level of order and counted. Data were analysed to assess the effects of a controlled traffic system on the overall abundance and ordinal richness of the arthropod assemblage, as well as abundance and species richness in the collembolan assemblage. Multivariate analysis was used to examine the differences in the ordinal composition of the arthropod assemblage. The data were variable between seasons, with significant increases in arthropod and collembolan abundance (p < 0.01) at all depths being evident under controlled traffic in spring. Arthropod richness was significantly greater in spring (p < 0.1), but a similar effect was not measured for collembolan richness. Overall arthropod abundance was greater under CT in winter (p < 0.1), although by depth, differences between treatments were evident only in the deepest samples for arthropod abundance (p < 0.05) and arthropod and collembolan richness (p < 0.1). While it was not possible to separate the relative impacts of traffic and tillage, it is concluded that the controlled traffic system had positive effects on the abundance and diversity of soil arthropods.

Forest plant community as a driver of soil biodiversity: experimental evidence from collembolan assemblages through large‐scale and long‐term removal of oak canopy treesQuercus petraea

Plant–soil interactions are increasingly recognized to play a major role in terrestrial ecosystems functioning. However, few studies to date have focused on slow dynamic ecosystems such as forests. As they are vertically stratified by multiple vegetation strata, canopy tree removal by thinning operations could alter forest plant community through tree canopy opening. Very little is known about cascading effects on soil biodiversity.We conducted a large‐scale, multi‐site assessment of collembolan assemblage response to long‐term canopy tree removal in sessile oakQuercus petraeatemperate forests. A total of 33 experimental plots were studied covering a large gradient of canopy tree basal area, stand age and local abiotic contexts.Collembolan abundance strongly declined with canopy tree removal in early forest successional stage and this was mediated by negative effect of understory plant community composition changes, i.e. shift from moss and forb to tree seedling, fern, shrub and grass species. Negative effect of this composition shift on collembolan species richness was largely offset by positive effect of the increase in understory plant species richness. This gives support to both the plant mass‐ratio and functional diversity hypotheses. Collembolan functional groups had contrasting response patterns, which were mediated by different ecological factors. Epedaphic (r‐strategist) abundance and species richness increased with canopy tree removal in relation with the increase in understory plant species richness. In contrast, euedaphic (K‐strategist) abundance and species richness declined with canopy tree removal in early forest successional stage in relation with changes in understory plant community composition and species richness, as well as microclimatic conditions.Overall, our study provides experimental evidence that forest plant community can be a strong driver of collembolan assemblages. It also emphasizes the role of trees as foundation species of forest ecosystems that can shape soil biodiversity through their regulation of understory plant community and ecosystem abiotic conditions.

Isolated cork oak trees affect soil properties and biodiversity in a Mediterranean wooded grassland

Mediterranean wooded grasslands are multipurpose systems that support high plant and animal diversity levels and are habitats of European importance (i.e., 6310 – Dehesas with evergreen Quercus spp.). Moreover, these systems offer a number of agro-ecosystem services such as forage production, soil carbon sequestration, nutrient recycling and soil protection. The scattered trees enhance the ecological complexity of grassland influencing the soil properties, the herbaceous layer diversity and composition and the soil communities. Understanding how isolated trees influence the other components of the system is essential to comprehend their role supporting high levels of above and below ground biodiversity and ecosystem services.In the present study, we present a hypothetical framework of the effects of isolated trees on soil properties, plant and soil fauna assemblages, the latter here represented by the class Collembola. The floor litter and the associated input of organic matter to the soil was a key factor linking the components of the tree-soil-biodiversity system in a Mediterranean cork oak wooded grassland.Topsoil C increased by +50% under the tree canopy in comparison with the areas beyond the tree canopy. Plant diversity was lower under tree canopy, but contributed to enhance the total species richness of the grassland. Collembolan diversity was higher under the peculiar conditions beneath the tree canopy. Relationships between plant and collembolan species emerged.The findings of this study suggest that isolated trees have direct and indirect effects on soil properties, plant and collembolan assemblages, hence they can influence the ecological processes of wooded grasslands, with implications for food webs, nutrient cycling and productivity of the agro-ecosystem.

“Collembola of the Western Taimyr:” Forty years later

A quantitative study of Collembola was carried out in 2010 in the area of the former Tareya Field Station of the Komarov Botanical Institute of the USSR Academy of Sciences, situated on the Taimyr Peninsula within the central part of the typical tundra subzone. The present-day collembolan assemblages were compared with those revealed 40 years ago, based on the results of a study performed in the region during the International Biological Program. Despite the absence of notable changes in the flora and vegetation, collembolan assemblages of all the plant associations studied in 2010 were clearly different from the complexes that had existed in the same territory at the end of the 1960s. Nevertheless, both the local collembolan fauna and the studied collembolan assemblages did not become more similar to their southern analogues and still fit well the characteristics typical of the region’s zonal position.

Response of collembolan assemblages to plant species successional gradient

Abstract It has been acknowledged that soil organisms play a significant role in nutrient cycling and thus may affect productivity and competition within plant assemblages with potential effects on vegetation trajectories. However, few studies have considered a single conceptual framework referring to both above–belowground linkages and plant succession. Consequently, we lack knowledge on the effects of plant successional processes on the dynamics of soil biota. Given this dearth of information, our study aimed to describe the dynamics of a major group of soil fauna, namely collembola, during the secondary succession on chalky slopes edging the Seine River. We selected five different plant assemblages as representative of a chronosequence: short grassland, tall grassland, encroached grassland, shrubland and forest. Our results clearly highlighted a strong response of collembolan assemblages to vegetation gradient. The changes in collembolan diversity were due to a high turnover rate in early successional stages and to nestedness in late successional stages. In addition, each collembolan life form had a specific response to vegetal succession. Euedaphic assemblages increased progressively during the succession while epedaphic did not follow a clear pattern. Our results also indicated a link between the dynamics of collembolan assemblages and plant life forms (e.g. trees and graminoids) rather than plant diversity. A focus on life forms, for plants and soil biota, seems to provide a good framework to study linkages between above and belowground biota.

Springtails (Hexapoda, Collembola) of Tundra Landscapes of the Kola Peninsula

A quantitative study of collembolan assemblages on the Barents Sea coast of the Kola Peninsula was performed. Altogether, 104 collembolan species of 55 genera and 16 families were found. Analysis of the faunal composition revealed its high similarity to that of the faunas of the neighboring regions of Norway as well as a marked decline in species richness as compared to the Siberian analogues. A high share of littoral forms with increased vagility and the predominance of species with “southern” (= boreal) and “western” distribution patterns are the main characteristics of the fauna studied; only a few true arctic forms were found, always with low abundance and occurrence levels. Most of the studied collembolan assemblages are characterized by the dominance of a few species which inhabit a part of or the entire range of plant communities. As a result, collembolan assemblages under various types of vegetation cover in different landscape elements often show only insignificant differences.

Changes in soil faunal assemblages during conversion from pure to mixed forest stands

Replacement of native deciduous forests by coniferous stands was a common consequence of former European afforestation policies. However, these changes have proven to lead to serious ecological problems. Therefore, re-establishing mixed forests with native tree species became an increasingly popular management strategy to fulfil the demands of multi-functional forestry. We report about changes in collembolan assemblages and microbial performances during conversion of pure coniferous stands to mixed forests. The study was carried out in the Black forest area (SW of Germany), where a gradient of conversion from pure spruce stands (S1) to equally mixed stands (spruce, beech, and fir) at S4, through two intermediate stages (S2 and S3) was selected. Results clearly indicated strong modifications of the collembolan communities with an enrichment of the assemblages over the course of the conversion process. Mean species richness increased by 47% from S1 to S4 accompanied by diversity indices (Shannon and Simpson) higher at S4. Significantly different soil biota assemblages were found at each phase of the conversion process. Spatial turnover and nestedness contribute almost equally to the modification of assemblages from S1 to S2 while later on in the mixing process only spatial turnover was acting. Concomitantly, significant shifts in the functional structure of the collembolan assemblages were depicted, deep-dwelling collembolan (euedaphic) being, surprisingly, the most responsive group. In contrast, neither microbial nor coarse environmental parameters were influenced by the factor “conversion phase”. We suggest that stimulation of Collembolan communities after the mixing process was mainly due to the input of more suitable food sources and/or microhabitat increases. Our findings underline the crucial role of aboveground processes on the belowground system, with the intensity and scheme of the mixing-process as important factors to consider when aiming at soil biodiversity improvement in forest systems.

Rainforest Collembola (Hexapoda: Collembola) and the insularity of epiphyte microhabitats

Abstract. 1. This study compares species composition of collembolan (Collembola:Hexapoda) assemblages associated with rainforest soil, forest floor leaf litter and epi-phyte-associated leaf litter deposits, canopy foliage and bark surfaces in a subtropicalrainforest in southeast Queensland.2. The results of analyses of similarity show that in both winter and summer, thespecies composition of collembolan assemblages differs significantly between each ofthe microhabitats studied. These results also confirm earlier work showing verticalstratification of species composition in collembolan assemblages associated with leaflitter suspended in epiphytes.3. Many of the species occurring in leaf litter suspended in epiphytes were notfound on intervening bark surfaces. This supports a hypothesis suggesting thatepiphytes within rainforest canopies can be thought of as habitat islands forarthropods.Key words. ANOSIM , Collembola, epiphyte, island biogeography, rainforest, verti-cal stratification. IntroductionCollembola are often the most abundant hexapods collected instudies of arthropods in rainforest canopies (Kitchinget al., 1993; Palacios-Vargas & Go´mez-Anaya, 1993;Guilbert et al., 1994, 1995; Palacios-Vargas et al., 1998, 1999;Yanoviak et al., 2003, 2004), and have long been known tooccurinawiderangeofrainforestmicrohabitats,includingsoils,forest floor leaf litter andhumic soils andleaf litter suspendedinepiphytes (i.e. Delamare Deboutteville, 1951). Collembola arealso the frequent prey items of mites, ants, micro-spiders,pseudoscorpions, small lizards, frogs and insectivorous birds(Hopkin, 1997). It thus seems likely that in addition to their sig-nificantroleindecompositionprocesses,theCollembolamaybeimportant in sustaining the diversity of these predator groupswithinrainforestcanopies.Many authors have suggested that the presence of epiphytesin rainforest canopies may play a n important role in sustainingthe abundance and diversity of Collembola and other arthro-pods in rainforest canopies (Palacios-Vargas & Go´mez-Anaya,1993; Guilbert et al., 1994, 1995; Kitching et al., 1997; Palacios-Vargas et al., 1998, 1999). The humic soils associated with rain-forest epiphytes have also been likened to habitat islands withinthe matrix of the rainforest canopy (Paoletti et al., 1991).Although speculative, this suggestion is tantalising since it maylink studies of epiphyte-associ ated arthropod assemblages to awell-established field of ecological theory, i.e. MacArthur andWilson’s(1967)theoryofislandbiogeography.Rodgers and Kitching (1998) have shown that the relation-ship between collembolan assemblages in forest floor leaf litterand epiphyte-associated litter dep osits (hereafter, referred to assuspended litter) is vertically stratified and accords with somepredictions of island biogeographic theory (assuming that for-est floor leaf litter is a source of species colonising leaf litterdeposits suspended within the canopy). This relationship wassuch that the similarity of species composition was significantlygreater between forest floor and lower canopy suspended litterassemblages than between the forest floor and suspended litterin the upper canopy (Rodgers & Kitching, 1998). The remote-ness of the island (i.e. lower canopy cf upper canopy) was sug-gested to provide some explanation for these results (Rodgers& Kitching, 1998).In this article, studies of this epiphyte island hypothesisare extended. We suggest that if epiphytes are indeed insularhabitats, this will be reflected in the comparative species compo-

Springtails (Hexapoda, Collembola) in karst landscapes of the Pinega State Reserve

The collembolan fauna of the Pinega State Reserve (Arkhangelsk Province) was examined. The study area is highly specific due to its karst topography. In all, 126 Collembola species were recorded in the reserve, a significant fraction of their diversity being restricted to karst habitats. Several species that had never been recorded in plain forest biotopes were of special interest. Such forms with distribution ranges mainly limited to the tundra zone comprised about 10% of the local fauna. Regarding the longitudinal sizes of the ranges, this group was not homogeneous and included both widespread circumpolar and east-Palaearctic (Siberian) species. Several arctoalpine forms were also recorded. Most of these arctic species were not numerous over the area studied and in general, the collembolan assemblages at the coldest sites of the reserve did not differ significantly from those typical of the forest zone, with the exception of two species, Desoria tshernovi and D. inupikella, which often predominated in the karst communities. The origin and appropriateness of the term relict for the revealed “insular” populations of arctic species are discussed.

Response of collembolan communities to land‐use change and grassland succession

This study focuses on the long‐term changes of collembolan communities occurring after the conversion of arable land to managed grassland. We analysed collembolan communities at grassland sites of different age that had been gradually converted over a period of 50 yr. Abundance and biomass responded rapidly and very positively to the conversion of arable land to grassland, while species richness was not affected. Collembolan assemblages changed only little during grassland maturation. The impact of land‐use change on community structure was more obvious at the functional level because the colonization processes observed in our study mostly relied on hemiedaphic species. Vegetation and soil parameters were good predictors of collembolan community structure during development of managed grassland. The present study demonstrated that past landscape patterns and processes like land‐use conversion and subsequent succession had a considerable impact on the present day pattern of species richness and community composition of Collembola within a landscape. Our results strongly differ from those obtained for other invertebrate groups, highlighting on the one hand the very diverse reactions of invertebrates to a common factor, and on the other hand the need to survey more than one taxa in order to draw conclusions on effects of land‐use change on faunistic communities.

Collembolan communities in a peat bog versus surrounding forest analyzed by using self-organizing map

The structure of Collembolan assemblages in a peat bog surrounded by forest in the French Pyrenees was analyzed using self-organizing map (SOM). The dataset included 68 samples (48 in the peat bog, 20 in the forest) for 63 species. Peat bog communities were shown to be always well separated from forest communities, with very few overlap in terms of species composition. From the dataset, five community groups (two in the forest, three in the peat bog) were defined by SOM, indicating an unexpected heterogeneity. No special assemblage was clearly differentiated at the interface zone of forest-peat bog. These variations were strongly correlated with a series of environmental parameters, substrate water content, Sphagnum and grass cover. An apparently homogeneous peat bog may therefore host rather different communities, the ecotone being not necessarily the most species-rich zone. Manipulating wetland habitats for biodiversity conservation should therefore not be done before having carefully established the real pattern of biodiversity in the area.