Edaphic Collembola of Lodgepole pine Pinus contorta plantations in Cumbria, UK P.J.A. Shaw1 & M.B. Usher2 Dept. of Biology, University of York, Heslington, York YO1 5DD, UK. 1: Present address: Dept. Environmental and Geographical Studies, Roehampton Institute, Southlands College, Wimbledon Parkside, London SW19 5NN, UK. Tel. (UK) 0181 392 3457 Fax. (UK) 0181 392 3461 EMAIL: peters@max.roeham.ac.uk 2: Present address: Scottish Natural Heritage, 2 Anderson Place, Edinburgh EH6 5NP, UK. Running title: Collembola under lodgepole pine  Edaphic Collembola of Lodgepole pine Pinus contorta plantations in Cumbria, UK P.J.A. Shaw1 & M.B. Usher2 Dept. of Biology, University of York, Heslington, York YO1 5DD, UK. 1: Present address: Dept. Environmental and Geographical Studies, Roehampton Institute, Southlands College, Wimbledon Parkside, London SW19 5NN, UK. 2: Present address: Scottish Natural Heritage, 2 Anderson Place, Edinburgh EH6 5NP, UK. Running title: Collembola under lodgepole pine  Abstract Collembola were extracted from soil under stands of lodgepole pine Pinus contorta, in northern England, ranging in age from two to twenty five years. Twenty nine species were recorded, but the community was dominated by only four: Folsomia brevicauda, Friesea mirabilis, Onychiurus armatus and Pseudisotoma sensibilis. There was little indication that community composition changed in relation to the age of the stand or distance from tree base, although in both cases the tests were significant for 2 species. By contrast, more species were significantly correlated with soil water, local vegetation and fungal fruitbodies than would be expected in random data. O. armatus correlated both with abiotic factors (soil water, tree age and distance from tree base) and with the saprophytic fungus Marasmius androsaceus, which is known to be one of its preferred foods. Keywords: Collembola, lodgepole pine, forest plantations, succession. Résumé Les collemboles ont été extraits du sol situé en-dessous de peuplements de pins de Murray, Pinus contorta, âgés de 2 ans à 25 ans, dans le nord de l'Angleterre. Vingt-neuf espèces ont été relevées, dont seules quatre dominent la communauté: Folsomia brevicauda, Friesea mirabilis, Onychiurus armatus et Pseudisotoma sensibilis. Il n'y a aucun rapport déterminant entre la composition de la communauté et l'âge du peuplement, mise à part la fréquence de O. armatus qui augmente dans les plantations plus anciennes. Rien n'indique que la composition de la communauté diffère en fonction à la distance à la base de l'arbre, mais des corrélations plus importantes que celles auxquelles on s'attendrait venant de données prélevées au hasard ont été établies avec l'eau du sol, la végétation locale et les fructifications fongiques. O. armatus est en corrélation importante avec la distribution du saprophyte Marasmius androsaceus, considéré comme un de ses aliments préférés. Mots-clefs: collemboles, pin de Murray, plantations forestières, succession. INTRODUCTION Although large areas of the UK have been forested with non-native conifers in the last 40 years, their soil micro-arthropod communities have been very little studied. An indication of the degree to which the subject is under-researched comes from the work of Ford, Malcolm & Atterson (1979), who produced a book about the ecology of forest plantations without including any work on their soil fauna. More recently, an entire issue of Forest Ecology and Management (1995; volume 79 issue 1-2) was devoted to the ecology of Kielder forest (close to the location of the fieldwork described below) again without any reference to soil micro-arthropods. Collembola are the second commonest group of soil arthropods after the mites (Hale, 1967), but there appears to be only one previous study of Collembola under non-native coniferous forests in the UK, that of Poole (1961), who studied the Collembola under Douglas fir Pseudotsuga menziesii in Wales. The only other comparable work is by Usher (1969) who described Collembola under native Scots pine Pinus sylvestris in Scotland. Both authors found communities of approximately 20 species that exhibited strong aggregations in response to unknown factors, though both studies showed correlations between some species and at least some of the environmental variables that they measured. Data are presented below from a study of Collembola under a stand of lodgepole pine Pinus contorta Doug. ex Loud in northern England as part of a larger programme of research to investigate the impact of grazing by Collembola on the ectomycorrhizal fungi of lodgepole pine (Shaw, 1985). Lodgepole pine is native to the west coast of North America but has been found to grow well in the wetter areas of the U.K., and was the softwood of choice for afforesting poorly drained uplands. It is the third most important commercial timber species in the UK after Sitka spruce Picea sitchensis (Bong.) Carr and Scots pine Pinus sylvestris L., and for plantations established since 1960 is second after Sitka spruce (Locke, 1987). In addition to defining the principal members of the Collembolan community in the forests, it was decided to examine whether a range of environmental factors could be linked with their distributions. These included age of the plantation, position with respect to tree base, water content of the soil, local flora and fungi. METHODS Site Description Work was based in Spadeadam Forest, a Forestry Commission plantation in the east of Cumbria (UK), at National Grid reference NY620740. The site is at an elevation of approximately 300m, and was originally Calluna vulgaris (L.) Hull moorland. Starting in 1953 the site was converted to commercial conifer plantations, with Sitka spruce Picea sitchensis on the majority of the site and lodgepole pine on the poorest 10% of the land. Mean annual rainfall is 1150 mm, leading to the formation of ombrotrophic bogs in poorly drained areas. Soils are peaty, overlying Carboniferous sandstone, to a depth in excess of 1m. Mean pH of the top 9cm was 4.0. The site was prepared for planting by double mouldboard ploughing, creating a furrow c. 40cm deep and planting ridges c. 1.5m apart. Trees were planted at c. 1.5m separation in the middle of each ridge, and staggered in adjacent rows to minimise overlap. Mycorrhizal fungi at the site have previously been studied by Dighton, Poskitt & Howard (1986), who recorded successional changes in the fruitbody community. Extraction and Identification of Collembola Soil cores were removed from the middle of planting ridges with a steel corer of 6cm internal diameter and 9cm depth. The corer held a removable perspex cylinder which encased the soil core and allowed it to be removed intact. Immediately a core was taken it was sealed top and bottom with a polystyrene cap (to prevent escape of mobile species), and returned to ITE Merlewood upright in plastic boxes. Cores were divided into 3cm thick slices, weighed, and placed upside down in the high gradient arthropod extractor described by Usher & Booth (1984). After extraction dried cores were reweighed to estimate water content. A subsample were dried to constant mass at 105oC. Collembola were identified using Fjellberg (1982) and Gisin (1960) after clearing in lactic acid and Nesbitt's solution (Evans, Sheals & McFarlane, 1961). The extractor was calibrated by adding a known number of Collembola to sterile, moistened cores. Extraction efficiency was found to be 87% for adult specimens of the relatively large species Onychiurus armatus, but 6.7% for the small species Tullbergia callipygos. The data gave no indication that extraction efficiency of O. armatus varied with depth in soil profile (F2,6 = 0.4, p>0.05). Sampling regime Two surveys were undertaken. In 1982 samples were taken from 12 stands ranging in age from 2 to 25 years. Soil cores were taken at 20, 40 and 80cm from the base of randomly chosen trees. Three replicate trees were sampled at each site, making a total of 9 soil cores per site. In 1983 the survey concentrated on 10 stands comprising 2 replicates of 5 ages. Five of these had been previously sampled in 1982, so that a total of 17 stands were sampled during the whole programme. Cores were taken midway between randomly chosen pairs of trees. For each core taken the following information was recorded: the percentage cover of ground flora in a 50cm quadrat centred on the core, the number of sporophores of saprophytic fungi within the quadrat, and the number of mycorrhizal sporophores within 1.5m of the core. During the 1983 survey each stand was visited 3 times and 6 replicate cores were taken on each visit. A limited survey was undertaken of Collembola gut contents. Twenty adult specimens per stand of the two commonest large Collembola, Onychiurus armatus and Pseudisotoma sensibilis were cleared in lactic acid, squashed and their gut contents stained with cotton blue prior to microscopic analysis. Statistical methods Due to the highly clumped distributions typical of many Collembola, all inferential tests used were non-parametric; Kruskal-Wallis ANOVA and Spearman's rank correlation coefficient. The structure of Collembolan populations was examined by ordinating the data using principal components analysis. To avoid pseudoreplication (Hurlbert 1984), tests involving factors operating at the level of the stand (tree age and height) were run using mean population values for each stand, with a concomitant reduction in degrees of freedom. RESULTS Twenty nine species of Collembola were recorded from Spadeadam. Table 1 lists the mean density of each species over both years and all sites studied, along with summary results for non-parametric tests of differences between years, distance from tree base and depth in soil. The community was dominated by four common species: Folsomia brevicauda, Friesea mirabilis, Onychiurus armatus and Pseudisotoma sensibilis. Seven species were found to differ significantly between years. Since significance is taken as p = 0.05, tests on 29 random datasets would generate an expectance of 1.45 significant results. Applying a Poisson distribution with E = 1.45, the observation of seven significant results would occur less than one time in a hundred in random data, so it is unlikely to be an artefact caused by multiple tests. However, there was no consistent pattern to the differences between years. Fourteen species had a significantly non-random vertical distribution (in all cases due to higher densities in the surface layer). By contrast, only two species (O. armatus and P. sensibilis) were non-randomly distributed with respect to distance from tree base, which is not significantly more than would be expected from random data (p=0.17 by a poisson distribution with E = 1.45). Indications of successional trends in the data were sought by univariate and multivariate approaches. Mean densities of all species in each stand (over both years where the stand was sampled in both years) were correlated with tree age and with tree height. This test would be expected to detect early successional and late successional species, but not those associated with intermediate seral stages. Both tests found the same three species to be non-randomly distributed with respect to stand maturity; Onychiurus absoloni was negatively correlated with both age and height (rs = -0.62 and -0.58 respectively, both p<0.05 with 15 df) as was Friesea mirabilis (rs = -0.66 and -0.58 respectively, both p<0.05 with 15 df), while O. armatus was positively correlated (rs = +0.60 and +0.62 respectively, both p<0.05 with 15 df). It should however be noted that three significant results from 29 tests are not significantly more than would be found in random data (p=0.06 by a poisson distribution with E = 1.45). Results for these three species and the co-dominants Folsomia brevicauda and Pseudisotoma sensibilis are shown in Figure 1. The observation that O. armatus was significantly correlated both with stand age and with distance from tree base is worthy of further comment, since the influence of a tree on soil conditions will decline with increasing distance from its trunk. A species that increases with increasing tree age may be expected to increase in population density closer to stem base where the influence of each tree is strongest. In fact the opposite trend was found for O. armatus, which was positively correlated both with stand age and with distance from tree base (Table 1). The interpretation of this result is unclear. The densities of species occurring at over 50 animals m-2 were ordinated by principal components analysis (Gauch, 1982). The first three eigenvectors accounted for only 51% of variance, and none of the resulting axes correlated with stand age or height. These analyses give little indication that the Collembolan community changed as the pine stands matured. Collembola are known to exhibit a clumped distribution in soil (Usher, 1969), and the data were examined for evidence of multi-species aggregation. The simple option of correlating the density of each species with total density of all Collembola would be incorrect as the two parameters would clearly not be independent. Consequently the density of each species of Collembola was correlated with total density of all other species (ie the density of the nth species was correlated with total for all species except species n). The results of these tests are summarised in Table 1, showing that ten species of Collembola were positively correlated with the density of other species. This is significantly more than the 1-2 significant observations that would be expected in random data (p<0.01 by a Poisson distribution with E = 1.45). None of the correlations were negative. Non-significant correlations for minor species may simply reflect a shortage of data. One dominant species, Onychiurus armatus, had a non-significant correlation and may confidently be assumed to be distributed independently of the other species. The gut content analyses found that approximately 64% of Collembola examined had empty guts (the figure was similar for both O. armatus and P. sensibilis). Gut contents included plant cells, brown amorphous matter, fungal hyphae and spores. Only one type of fungal material was distinctive enough to be given a possible identification. This was the sporophore of a dematiaceous hyphomycete, probably Troposporella monospora (= Slimacomyces monospora), which is common on decaying Pinus needles (Ellis, 1976, p. 47) and whose spores appear to depend on grazing arthropods for dispersal (Minter, 1986). These spores were particularly common in the guts of O. armatus. The 1983 survey included observations of soil water content, local flora and fungal fruiting bodies (both saprophytes and mycorrhizal species). The results for these variables are summarised in Table 2. The plant community was sparse due to acidity and shading, with a dominance of mosses. An unexpected observation was the persistence of small amounts of bog rosemary Andromeda polifolia under closed canopies. Floral community structure changed as stands aged, with moorland plants declining and certain mosses increasing. Age-related changes in plant and fungal communities are summarised in Table 2 by correlations between each species and two parameters: stand age, and the water content of their associated soil core. These parameters were not independent: water content was significantly negatively correlated with stand age (rs = -0.20, df = 178, p < 0.01). C. vulgaris, Molinia caerulea and Polytrichum mosses were negatively correlated with stand age, while Cladonia lichens, A. polifolia, Deschampsia flexuosa, Lophozia ventricosa and Sphagnum mosses were positively correlated with water content. These were all plants of open moorland, declining as the canopy closed and the soil dried out. By contrast three bryophytes (Campylopus paradoxa, Hypnum cupressiforme and Lophocoela bidentata) were positively correlated with stand age. This line of analysis showed two species of fungi (the mycorrhizal symbiont Lactarius rufus and the saprophyte Marasmius androsaceus) to be negatively correlated with tree age. Although L. rufus is known to be associated with young conifers (Dighton, Poskitt and Howard, 1986, Shaw & Lankey, 1994) there was a paradox in the results for M. androsaceus since this fungus was also negatively correlated with water (so should be commoner in older, drier stands). Newell (1984a) found M. androsaceus to be associated with the drier horizons of coniferous soils. The data for mycorrhizal fungi show an identical pattern of successional change to that described by Dighton, Poskitt and Howard (1986) for the same forest, with L. rufus under young stands being replaced by Russula emetica as the stands aged, although the dataset presented above is smaller and less clear than that previously published. Relationships between the variables are summarised in Table 3, giving significant correlation coefficients between the 13 commoner species of Collembola (defined as those occurring at more than 50 animals m-2) and the environmental factors. Out of a total of 144 tests, 28 were found to be significant (p<0.001 by a poisson distribution with E = 7.2), suggesting that some of the distributions encountered were indeed non-random. Pseudisotoma sensibilis had two colour morphs, with 64% being dark blackish-violet (the dark form), and 36% being greenish-brown (the pale form). These two morphs were significantly co-correlated (rs=0.63, df = 178, p<0.01) but responded differently to soil water. The pale form was negatively correlated with soil water, while the dark form was positively correlated (Table 3). DISCUSSION There are few studies on the Collembolan communities of conifer plantations with which the results above can be compared. Poole (1961) studied Collembola under even aged stands of Douglas fir, and found 24 species of which 12 were recorded at Spadeadam forest (Anurida granaria, Entomobrya spp., Friesea mirabilis, Folsomia quadrioculata, Isotoma notabilis, I. viridis, Isotomurus palustris, Neanura muscorum, Neelus minimus, Pseudisotoma sensibilis, Tomocerus minor, and Tullbergia callipygos). There were conspicuous differences in the balance of species, with F. quadrioculata, I. palustris and I. notabilis being common in Poole's data but scarce at Spadeadam, while F. brevicauda and P. sensibilis showed the opposite pattern. Curiously, Poole found no specimens of the genera Onychiurus or Tetracanthella which were among the most conspicuous Collembola at Spadeadam. The frequent occurrence of the hydrophilic species Isotomurus palustris, coupled with the absence of relatively xerophilic species such as Onychiurus and Tetracanthella (Murphy, 1955) suggest the soil to have been substantially wetter in Poole's study area, although no data are given. Miles (1993) described Tetracanthella wahlgreni as an ice age relict species. Blackith (1974) described T. wahlgreni as abundant in Cladonia lichens and Sphagnum moss on Irish blanket bogs. The preponderance of Folsomia brevicauda over F. quadrioculata matches the findings of Blackith (1974), though Usher (1967, 1970) found F. quadrioculata in a P. sylvestris forest but F. brevicauda in the surrounding moorland. There were few clear signs of successional change in the Collembola data. Onychiurus absoloni was only found once in a newly planted site, and O. armatus tended to occur in older stands. No other species showed age-related trends, and 2 significant observations out of 29 species examined is not more than would be expected by chance. However, other workers have found successional changes in Collembolan communities. Usher & Parr (1977) noted changes in the Collembola of a chalk quarry as vegetation developed. In particular, Onychiurus spp. became more abundant as organic matter built up in the quarry floor. Greenslade & Greenslade (1993) found that Collembola species diversity on bauxite waste was directly linked to the diversity of plant species. These two observations refer to primary successions on ground that was initially bare and largely sterile. More relevant to the work described above were the findings of Murphy (1955) who studied Collembola at Moor House (northern England) and found that hydrophilic species such as Isotomurus sp. and Sminthurides malmgreni were replaced by Tetracanthella wahlgreni as plant debris accumulated and the site dried out. The system studied at Spadeadam started as Calluna moorland and developed into a closed canopy conifer monoculture. This caused numerous linked changes in edaphic conditions, with decreasing water content and ground flora, increasing needle litter and shading. It is worth speculating why these changes did not cause a clear alteration in the Collembolan community, at least in the first 25 years of life of a plantation. Murphy (1955), Usher & Parr (1977), and Greenslade & Greenslade (1993) all found that the most significant factor driving changes in Collembolan communities was a build-up of organic matter in the soil. The upper horizons of the pre-planting moorland soil at Spadeadam would have been almost entirely organic matter, so that afforestation would not substantially change the proportion of organic matter in the soil. This may explain the apparent lack of changes in the Collembolan assemblage. The correlations between species density and total density of all other species showed that 10 species of Collembola were significantly co-correlated (Table 1), implying the formation of multi-species clusters. Such aggregations have been suggested by previous work (Poole, 1961; Usher, 1969) and aggregation pheremones have been identified in Collembola (Joose, Verhoef & Nagelkerke, 1977), but the data above are unusual in showing that one dominant species Onychiurus armatus was distributed independently of these multi-species aggregations. This may suggest that this species uses a different aggregation pheremone to other Collembola. The 1982 survey investigated whether distance from tree base affected Collembola community composition. KruskalWallis tests found significant relationships for only 2 species, which is not more than would be expected from random data. This appears to be the first time that a Collembolan community has been analysed as a function of distance from tree base, though Zinke (1962) studied the influence of distance from tree base on soil chemistry and found a zone of elevated nitrogen and lower pH around the base of mature Pinus contorta trees. Skeffington (1983) also found depression of soil pH immediately under three species of tree in a podzolic soil. The relationships between Collembola and local environmental factors summarised in Table 3 contain more significant correlations than would be expected in random data, but no clearly defined patterns. It is noteworthy that O. armatus was negatively correlated with water (being the one species tending to occur in older, drier stands) and was also positively correlated with the saprophytic fungus Marasmius androsaceus. It is known from other work that M. androsaceus is one of the preferred fungal foods of O. armatus (Newell, 1984a;b; Shaw, 1988). A curious observation is that the three Collembola positively associated with Calluna (N. muscorum, T. wahlgreni and the dark form of P. sensibilis) were all darkly pigmented. This may allow them to warm up more rapidly in sunny heathy areas. The two colour morphs of P. sensibilis behaved quite differently with respect to soil water; the dark morph was positively associated with soil moisture while the pale morph was negatively associated. The dark morph was also positively associated with heather Calluna vulgaris. These observations are consistent with Hale (1966), who recorded these two colour morphs and noted that the pale forms were associated with mineral soils while the darker forms preferred peaty soils. Usher (1969, 1970) found that the two colour morphs had differences in spatial distribution patterns, the proportion of dark individuals varied from 25 to 59% of the population, and that the populations of the two colour morphs peaked at different times. Although apparently identical morphologically, it is tempting to speculate whether these morphs are in fact distinct species, though more work would be required to establish this. The number of species of Collembola found at Spadeadam forest (29) was comparable with results from native UK ecosystems, despite being collected under a monoculture of a non-native conifer. This shows that the widespread image of upland conifer plantations as biological deserts is inaccurate, at least for edaphic Collembola. Similar results have been found by Ozanne et al. (in press), where studies of epiphytic micro-arthropods in monocultures of alien conifers contained higher overall population densities than tropical forests, and larger species richness than native UK woodlands. These observations can be used to make a cautionary point: biodiversity may be unexpectedly high in superficially monotonous habitats and should never be pre-judged without adequate empirical data (Usher, 1994). Acknowledgements This work formed part of a PhD which was co-supervised by Dr John Dighton and funded by NERC. Fieldwork was undertaken while based at ITE Merlewood. Thanks are due to Mr TC Mitchell of the Forestry Commission for help in locating compartments at Spadeadam Forest, and for permission to sample there. Dr Juliette Frankland helped locate information on Troposporella. PJAS wishes to express gratitude to the military police at Spadeadam for their help, notably on the day that the survey coincided with a CND blockade of the site. The abstract was translated by Mrs Christine Turner.This paper could not have been produced without the Roehampton Institute's generous provision of spare time in the form of meetings, when the dataset was re-typed on a personal organiser for downloading and re-analysis. REFERENCES Blackith, R. E. (1974). The ecology of Collembola in Irish Blanket Bog. Proceedings Royal Irish Academy, 74, 203-226. Dighton J., Poskitt J. M. and Howard D. M. (1986). Changes in occurrence of basidiomycete fruit bodies during forest stand development with specific reference to mycorrhizal species. Transactions of the British Mycological Society, 87, 163-171. Ellis M. B. (1976). More Dematiaceous Hyphomycetes. CAB, London, UK. Evans G. O., Sheals J. G. and MacFarlane D. (1961). Terrestrial Acari of the British Isles. British Museum of Natural History, London, UK. Fjellberg A. (1982). Identification Keys to Norwegian Collembola. Norsk Entomologisk Forening, As NLH, Norway. Ford, E.D., Malcolm, D.C. & Atterson, J. (1979). The Ecology of Even-aged Forest Plantations. Institute of Terrestrial Ecology, UK. Gauch H. G. (1982) Multivariate Analysis in Community Ecology Cambridge University Press, Cambridge, UK. Gisin H. (1960). Collembolenfauna Europas. Museum d'Histoire Naturelle, Geneva, Switzerland. Greenslade P. and Majer J. D. (1993). Recolonisation by Collembola of rehabilitated bauxite mines in Western Australia. Australian Journal of Ecology, 18, 385-394. Hale W.G. (1966). The Collembola of Moor House National Nature Reserve, Westmoorland, a moorland habitat. Revue d'Ecologie et Biology du Sol, 3, 97-122. Hale W. G. (1967). Collembola. In Soil Biology (Eds A. Burgess and F. Raw), Academic Press, London, UK, pp. 397-412. Hurlbert S.H. (1984) Pseudoreplication and the design of ecological field experiments. Ecological Monographs, 54, 187211. Joose E. N. G. Verhoef H. A. and Nagelkerke C. J. (1977). Aggregation pheremones in Collembola (Apterygota): a biotic cause of aggregations. Revue d'Ecologie et Biology du Sol, 14, 21-25. Locke G. M. L. (1987). Census of Woodlands and Trees. Forestry Commission Bulletin 63. HMSO, London. Miles P. M. (1993). Two glacial relicts and associated species of Collembola from north Cardiganshire, Wales. Entomologist's Monthly Record, 105 23-25. Minter D. (1986). Upland foray, Sutherland. Bulletin of the British Mycological Society, 20, 17-24. Newell K. (1984a). Interactions between two decomposer basidiomycetes and a Collembolan under Sitka spruce: Distribution, abundance and selective grazing. Soil Biology and Biochemistry, 16, 227-233. Newell K. (1984b). Interactions between two decomposer basidiomycetes and a Collembolan under Sitka spruce: grazing and its potential effect on fungal distribution and litter decomposition. Soil Biology and Biochemistry, 16, 235-239. Ozanne, C. M. P., Foggo, A., Hambler, C. & Speight, M. R. (in press). The significance of edge effects in the management of forests for invertebrate biodiversity. In Canopy arthropods (Eds. N. J. Stork, J. Adis & R. Didham). Chapman and Hall, London, UK. Poole T. B. (1961). An ecological study of the Collembola on a coniferous forest soil. Pedobiologia, 1, 113-117. Shaw P. J. A. (1985). Interactions between Collembola and fungi associated with the roots of lodgepole pine, Pinus contorta. Unpublished D. Phil. thesis, University of York. Shaw P. J. A. (1988). A consistent hierarchy in the fungal feeding preferences of the Collembola Onychiurus armatus. Pedobiologia, 31, 179-187. Shaw, P. J. A. and Lankey, K. (1994) Studies on the Scots pine mycorrhizal fruitbody succession. The Mycologist, 8, 172-175. Skeffington, R. A. (1983). Soil properties under three species of tree in southern England in relation to acid deposition in throughfall. In Effects of Accumulation of Air Pollution in Ecosystems (Eds B. Uhlrich & J. Pankrath), pp. 219-231. D. Reidel, Dordrecht, Holland. Usher M. B. (1967). Studies on Population Ecology, with particular reference to soil Collembola. Unpublished PhD thesis, University of Edinburgh. Usher M. B. (1969). Some properties of the aggregation of soil arthropods: Collembola. Journal of Animal Ecology, 38, 607-622. Usher, M. B. (1970). Seasonal and vertical distribution of a population of soil arthropods: Collembola. Pedobiologia, 10, 224-236. Usher M. B. and Parr T. W. (1977). Are there successional changes in arthropod decomposer communities? Journal of Environmental Management, 5, 151-160. Usher M. B. and Booth R. G. (1984). A portable extractor for separating arthropods from soil. Pedobiologia, 26, 17-23. Usher, M. B. (1994). Biodiversity: which communities are hiding it? In Individuals, Populations and Patterns in Ecology (Eds S. R. Leather, A.D. Watt, N. J. Mills and K. F. A. Walters), pp. 265-273. Intercept Press, Andover, UK. Zinke P. J. (1962). The pattern of influence of individual forest trees on soil properties. Ecology, 43, 130-133. Table 1. Mean densities (animals m-2) and standard error (se) of all species of Collembola recorded from Spadeadam forest during 1982 and 1983, along with a summary of results of non-parametric tests examining effects due to differences between years (year), distance from tree base (dist), and depth in soil (depth). The column headed aggr lists the results of a Spearman's correlation between the density of a species and the density of all other species, which was used as a test of the tendency to be involved in multi-species aggregations. Nomenclature follows Gisin (1960). Pseudisotoma sensibilis occurs in 2 colour morphs, which were pooled for the 1982 survey but recorded separately for the 1983 work. The analyses below list results for both colour morphs pooled, using data from both surveys, but also give densities for each morph separately for the 1983 data (without statistical analyses). Species Mean se year dist depth aggr Anurida granaria 11 4 . . . . Anurida pygmaea 7 4 . . . . Brachystomella parvula 39 18 . . -** . Dicyrtoma atra 71 14 ** . -** . Entomobrya albocincta 4 4 . . . . Entomobrya nivalis 21 7 . . . . Folsomia brevicauda 2447 594 ** . -** +* Folsomia quadrioculata 4 4 . . . . Friesea mirabilis 1524 134 ** . -** +** Hypogastrura scotica 21 7 . . -* . Hypogastrura ununguiculata 4 1 . . . . Isotomurus palustris 25 18 . . . . Isotoma antennalis 4 1 . . . +* Isotoma notabilis 14 14 . . . . Isotoma viridis 42 7 . . -** . Lepidocyrtus lignorum 248 35 . . -** . Neanura muscorum 92 18 . . -** +** Neelus minimus 67 18 * . -** +* Onychiurus absoloni 226 137 . . . +* Onychiurus armatus 2341 318 ** +* -** . Orchesella cincta 4 0 . . . . Pseudisotoma sensibilis pooled 2436 251 ** -* -** +** dark morph 1239 144 pale morph 683 77 Sminthurides parvulus 39 35 ** . -** . Tetracanthella wahlgreni 336 74 . . -** . Tomocerus flavescens 4 4 . . . . Tomocerus minor 53 11 . . -** . Tomocerus vulgaris 14 11 . . . +* Tullbergia sylvatica 78 35 . . . +* Willemia anopthalma 92 53 . . . +** dist? - effect of distance from tree base on species density (data from the 1982 survey only); Year? - effect of differences between the two years of survey (confined to data from forest compartments that were sampled in both years); Abbreviations: .; p>0.05, *; differences significant at p<0.05, +*; a positive association significant at p<0.05, -*; a negative association significant at p<0.05, **; differences significant at p<0.01, +**; a positive association significant at p<0.01, -**; a negative association significant at p<0.01. Table 2. Summary results for the environmental factors measured during the 1983 survey of Collembola in Spadeadam forest. The column headed "comment" gives the nature of each named species. Columns headed "mean" and "standard error" refer to percentage cover values, except for water content (as percentage dry mass) and fungal species (number of fruitbodies recorded in stand). Columns headed "tree age" and "water" give significant values of Spearman's correlation coefficient for each species with stand age and soil water content. Parameter: comment Mean s.e. treeage water Water content 74.0 0.3 -0.20 1.0 Plants: Andromeda polifolia PP <1 . +0.20 Aulocomnium palustre Bryoph. <1 . . Calluna vulgaris PP 12 2 -0.33 +0.20 Campylopus paradoxus Bryoph. 5 1 +0.22 . Cladonia spp lichen <1 . +0.16 Deschampsia flexuosa PP 14 2 . +0.21 Dicranum scoparium Bryoph. 2 1 . . Dryopteris dilitata Fern <1 . . Hypnum cupressiforme Bryoph. 7 2 +0.22 . Lophocoela bidentata Bryoph. 1 1 +0.23 . Lophozia ventricosa Bryoph. 1 0 . +0.16 Mnium hornum Bryoph. 1 0 . . Molinia caerulea PP 1 1 -0.23 . Plagiothecium undulatum Bryoph. 5 1 . . Pleurozium schreberi Bryoph. 2 1 . . Polytrichum spp. Bryoph. 9 1 -0.47 . Rhytiadelphus loreus Bryoph. 1 0 . . Sphagnum spp. Bryoph. 2 1 . +0.16 Vaccinium myrtillus PP <1 . . Fungi: Clitocybe sp. sap <1 . . Cortinarius sp. myc <1 . +0.14 Cystoderma amianthinum sap <1 . . Galerina hypnorum sap 10 2 +0.21 . Marasmius androsaceus sap 50 8 -0.26 -0.21 Mycena galopus sap 3 1 . . Inocybe longicystis myc 9 2 . . Laccaria laccata myc <1 . . Lactarius rufus myc 21 4 -0.48 . Russula emetica myc 2 1 . . Abbreviations: . - Non significant results; Bryoph. - bryophyte; myc - mycorrhizal fungus; Sap - Saprophytic fungus; PP - phanerogamic plant. Table 3. The significant (p<0.05) Spearman rank correlation coefficients between members of the 13 commonest taxa of Collembola and local environmental factors. Data come from the 1983 survey of Spadeadam forest. soil Saprophytic Mycorrhizal water Plants Fungi Fungi content ‎‮─────────────────────────‭ ‮──────────────‭ ‮────────────── Cp Cv Df Hc Ps Gh Ma Mg Il Lr Re Dicyrtoma atra -0.20 . . . . . . . . . . . Folsomia brevicauda . . . . +0.20 . . . . . . . Friesea mirabilis +0.25 . . +0.28 . . . . . . . +0.19 Isotoma viridis . . . . . +0.18 . -0.15 . +0.18 . . Neanura muscorum . . +0.22 +0.15 . +0.18 . . . . . . Neelus minimus . . . . . 0.16 +0.27 . . +0.26 . +0.17 Onychiurus armatus -0.32 . 0.36 . . . . +0.25 . . . . Pseudisotoma sensibilis dark morph +0.18 +0.16 +0.24 . . . . . +0.16 . . . pale morph -0.20 .  . . . . . . . . . pooled . . +0.18 . . . +0.15 . . . . . Sminthurides parvulus . . . . . . . +0.25 . . +0.23 . Tetracanthella wahlgreni . . +0.20 . . . . . . . . . Abbreviations: . : nonsignificant; Cp: Campylopus paradoxa; Cv: Calluna vulgaris; Df: Deschampsia flexuosa; Gh: Galerina hypnorum; Hc: Hypnum cupressiforme; Il: Inocybe lacera; Lr: Lactarius rufus; Ma: Marasmius androsaceus; Mg: Mycena galopus; Ps: Polytrichum spp.; Re: Russula emetica; water - soil water content. Captions to Figures Figure 1. Plots of the densities of the five dominant species of Collembola (animals m-2) and the age of the lodgepole pine plantations in Spadeadam Forest. Vertical axes are logarithmic, and standard errors of the mean density are shown as bars only when they are wider than the points indicating the mean densities. Horizontal axes show the plantation age in years (note that for clarity the ages of 18 and 24 have not been numbered). a) Onychiurus spp (circles indicate O. armatus and squares O. absoloni, the latter being absent from all plantations over 3 years). b) Friesea mirabilis, c) Folsomia brevicauda, and d) Pseudisotoma sensibilis.