Soil Animal Communities of Planted Birch Stands in Central Finland

The aim of this study was to compare soil animal communities in planted birch (Betula pendula) stands of different origin with each other and with natural forests. Six 30-year-old birch stands were investigated, three planted after clear-cutting of spruce stands, and three planted on cultivated soil. The faunal communities were markedly different in plantations established on spruce forest soil and on arable soil. “Birch after Spruce” communities were relatively similar to those of coniferous forests, though the population densities were generally lower. “Birch after Field” communities were sparse and could be characterised as “impoverished forest communities”, except in Lumbricidae and Enchytraeidae that have affinities with deciduous forests and cultural landscapes. Soil conditions are not sufficient to explain the differences between the forests. Colonisation and transport by man may determine the presence of certain species, especially earthworms. These in turn affect soil properties, and compete with or otherwise have negative effects on other soil fauna. Thus the community differences between different forests are an outcome of several factors: soil characteristics, site history, colonisation ability and interspecific interactions.


Introduction
Nordic coniferous forests have undergone many changes under the human influence.The old slash-and-burn agriculture that after clear-cutting and some years of cultivation left the ground open for natural succession favoured development of deciduous stands, mainly composed of the birch species (Betula pendula and B. pubescens), but also other species such as Alnus incana, Popu-lus tremula and Sorbus aucuparia.If left undisturbed, communities are gradually replaced by the two climax tree species, the conifers Scots pine (Pinus sylvestris) and Norway spruce (Picea abies).In the 20 th century, the slash-and-burn practice was abandoned, and the more modern forestry approach tended to grow spruce or pine in monocultures, saplings of deciduous trees being destroyed mechanically or by using herbicides.This strongly reduced the proportion of deciduous Silva Fennica 39(1) research articles or mixed stands in the total forest area.During the last two decades an opposite tendency has prevailed: a mixture of deciduous trees, silver birch (B.pendula) being the most favoured, is recommended in practical forestry.In addition, abandoned fields are most usually reforested by silver birch, as a result of which young birch stands of homogeneous age structure are now commonly seen (Finnish Forest Research Institute 1992).In the future, expected warming of the global climate will probably further contribute to the change of boreal forests in favour of broadleaved trees (Kellomäki et al. 1996).
Facing these trends, it is important to know more about biological changes in forest soils accompanying the alterations in stand composition.Will the community structure in coniferous forest soil after replanting with birch resemble that of a natural deciduous forest?To what extent does the decomposer community of a birch stand established on cultivated land change towards a forest soil community, or does it rather retain the character of arable soil?What are the mechanisms that regulate these changes?
The forest floor at boreal and temperate latitudes harbours an abundant and diverse soil faunal community, but there are also marked differences between different kinds of forests.In coniferous forest soils with low pH, microarthropods and enchytraeids predominate, while in less acid deciduous forests "macrodecomposers" such as diplopods, isopods and earthworms contribute essentially to the total biomass (Macfadyen 1963, Petersen andLuxton 1982).Related differences are even found within deciduous forests on different soil types (Schaefer and Schauermann 1990).The soil fauna of Nordic coniferous forests is relatively well known; several studies have been published, together covering all major animal groups.On the other hand, quantitative assays in deciduous forests are scarce and casual, almost restricted to a single paper (Axelsson et al. 1984).The microarthropod fauna of a beech forest in Denmark was thoroughly studied by Luxton (1981a,b,c,d,e) and Petersen (1980).
The present paper summarizes the results of a recent study, the aim of which was to compare the decomposer communities of anthropogenic birch stands of different origin with each other, and with those of natural forests and arable soils at the same latitude.Detailed results on each main group of organisms have been or will be published elsewhere (Huhta 2002, McLean and Huhta 2002, Huhta and Niemi 2003, Räty and Huhta 2003a, 2004a, b, Huhta et al. 2005).As the community structure is thoroughly described in these papers, the present report tends only to summarize the species-level data, and mainly concentrates on soil properties and vegetation, faunal biomasses and general conclusions.

Site Description
Six planted birch (Betula pendula) stands were investigated.Three of these were planted 28 to 33 years prior to the study after clear-cutting of spruce stands ("Birch after Spruce", symbol BS), and three were planted 26 to 43 years earlier on arable soil ("Birch after Field", BF).The soil type of the BS sites is ferric podsol.The fields on the BF sites were established on earlier coniferous (spruce) forest soil and cultivated for decades or even centuries before reforestation.All the sites are located in central Finland, 30 to 90 km north of the town of Jyväskylä (62°24'-63°00'N, 25°30-50'E).The stand characteristics are as follows: Three "natural" deciduous stands were also studied, but these differed so much from each other that comparison with the planted stands was hardly possible.Therefore, the results of the reference sites are presented in another paper (Huhta et al. 2005), and only briefly referred to in the present context.

Vegetation and Soil Analyses
The ground vegetation was analysed in July 1998 by estimating the coverages of vascular plant species from ten random 1 m 2 plots at each study site.
The following soil variables were determined from three samples (n = 6) taken in 1998 (May, August and October), separately for the 0-3 and 3-6 cm soil depths: -pH H2O in soil-water suspension (5:25 m/m).
-Dry matter and water contents, drying overnight at 80°C.-Organic matter content as loss on ignition (4.5 h at 550°C).-Mineral nitrogen (NH 4 + and NO 3 -) was extracted in 2-M KCl and measured according to SFS standard 3032.The results were calculated as mg N/m 2 in the whole depth of 0-6 cm.-Microbial biomass and basal respiration were determined using the substrate induced respiration (SIR) method (Anderson andDomsch 1978, Nordgren 1988).The results were calculated as mg C/g dry mass and mg C/cm 3 .
The following variables were determined only once and on different dates: -Soil texture: particle size distribution of mineral soil was determined by dry sieving down to size class 0.2 mm, and smaller fractions using SYMPATEC diffraction spectrophotometer.-Water retention was determined using a pressure plate chamber at potentials -0.3, -1, -5, -10 and -100 kPa.Retention at -5 kPa was used as water holding capacity (WHC).Matrix potential (pF) was calculated on the basis of WHC and water content measured at each sampling (topmost 5 cm of soil excluding litter).-Soil structure: content (% of weight) of water-stable aggregates was determined by Viljavuuspalvelu Oy, Mikkeli (SFS-EN-ISO 9002).

Faunal Analyses
Soil faunal samples were taken in May, August and October 1998 for macroarthropods, August and October 1998 for microarthropods, May, Aug. and Oct. 1999 for nematodes and enchytraeids, and May and Oct. 1999 for lumbricids.The sam-ples, each consisting of six sample units, were taken separately for each animal group, and randomly from a representative area of ca. 30 × 30 m independently of the total size of the similarly managed area.The macroarthropod and lumbricid samples were 25 × 25 cm blocks of topsoil, depth 0-5 cm for arthropods and 0-10 cm (two 5-cm layers separately) for earthworms.Other samples were taken with cylindrical steel corers (area 10 or 25 cm 2 ), and the layers 0-3 and 3-6 cm, and also 6-9 cm for nematodes and enchytraeids, were treated separately.Macroarthropods were extracted using large Tullgren funnels (Huhta 1972), microarthropods using a modified high gradient apparatus (Macfadyen 1961), enchytraeids by wet funnels with O' Connor's (1962) heating regime, nematodes from ca. 1.5 cm 3 subsamples with the wet funnel technique of Sohlenius (1979), and earthworms using the wet extraction technique of Huhta and Koskenniemi (1975).In addition, formaline extraction (Daniel et al. 1992) was separately applied for the earthworm Lumbricus terrestris (three 0.5 m 2 plots per sample).
All major animal groups were identified to species level whenever possible (genera level for nematodes).The data from all soil layers were summed and the results were transformed into numbers per square metre.

Estimation of Biomasses
Collembola were counted in size classes at intervals of 0.1 mm (up to 1.2 mm) or 0.2 mm, and their biomasses were estimated by species using the regression formulae of Tanaka (1970) and Petersen (1975).For species not found in these, regressions for species of similar body shape were used.The weights for adult Oribatida were mainly taken from Luxton (1975), and for missing species interpolated from species of similar size and shape.Average individual weights for Prostigmata, Astigmata and juvenile Oribatida were taken from Huhta et al. (1986a).Weights for Mesostigmata were obtained from the data of Huhta et al. (1979Huhta et al. ( , 1986a,b),b), or using the regression formula of Persson and Lohm (1977).Abrahamsen's (1973) equations of body volume and density were used to calculate the biomass of Enchytraeidae.Earthworms were weighed fresh after keeping 48 h on moist paper to empty their guts, and transformed to dry weights by assuming a dry mass content of 16%.Total biomasses of Nematoda were estimated on the basis of average individual weight taken from Huhta et al. (1986a).The biomass determinations for macroarthropods follow Huhta (2002).

Statistical Analyses
Differences between the site types (BS, BF) were tested by two-way analysis of variance (Site type, Date) on sample averages of each site and date.Because of low number of replicates (n = 3), the level of p < 0.1 was also considered to indicate significance.Differences between site replicates (BS1,2,3 and BF1,2,3) were separately tested by ANOVA (Site, Date; n = 6).For DM, OM, pH and SIR the two soil depths were tested separately.As there were few significant time effects or Site × Date interactions, and these are not important in the present context, only the main effects are presented.The biomass data were ln-transformed prior to the analyses.Biomasses were tested only at group level; the tests at species level have been conducted on numbers of individuals and presented in other papers (see Introduction).
Community ordination was carried out on microarthropods and enchytraeids using Non-Metric Multidimensional Scaling (NMS).It was done using the PC-ORD programme (version 4; MjM Software Design) in NMS Autopilot mode, with the Sørensen (Bray and Curtis) index as distance measure.For microarthropods, species present only as one specimen in the whole material were discarded.The data were ln-transformed prior to the analyses.The measured soil variables were included in the ordinations; for pH, OM and SIR only the data of 0-3 cm were used, which was considered more descriptive to the study sites.

Vegetation
The plant communities differed considerably between the birch stands established on forest soil and on arable soil (Table 1).The grasses Deschampsia flexuosa, Calamagrostis arundinacea and Agrostis capillaris were characteristic to the "Birch after Spruce" sites, followed by some common forest herbs such as Melampyrum sylvaticum and M. pratense, Rubus saxatilis, Maianthemum bifolium and Trientalis europaea.Dwarf-shrubs that dominate in coniferous forests were found only sporadically.A. capillaris was also abundant in one "Birch after Field" site, but all other predominant plants in BF were herbs of cultural origin: Silene dioica, Fragaria vesca, Geum rivale, Filipendula ulmaria, Veronica chamaedrys etc.

Soil Characteristics
Organic matter and water contents in the deeper soil layer (3-6 cm) were higher in "Birch after Field" than in "Birch after Spruce" (F = 4.51, p = 0.055, and F = 60.3, p < 0.001, respectively), indicating a more even vertical distribution of soil organic fraction in BF (Fig. 1).pH was also higher in "Birch after Field" (F = 12.6, p = 0.004 and F = 18.05, p = 0.001 for top and bottom layers, respectively).Clay content and fine fraction of mineral soil were clearly higher in the "Birch after Field" sites (F = 6.03, p = 0.070, and F = 17.33, p = 0.014, respectively), reflecting the fact that clay and silt soils have been preferably taken for cultivation, in comparison with coarse soils.WHC was higher (F = 5.47, p = 0.079) and average pF lower in BF (F = 3.35, p = 0.092), which indicates a better water availability.Contents of mineral nitrogen varied considerably between replicates, but the average NO 3 -concentration was higher in BF (F = 3.68, p = 0.08).Content of water stable aggregates, indicating earthworm activity, was also much higher in the BF sites.Total microbial biomass (SIR) per dry mass of soil (0-3 cm layer) was about 50% higher in "Birch after Spruce" (F = 23.03,p < 0.001), probably due to high fungal biomass (McLean and Huhta 2002), and basal respiration was also higher (0-3 cm; F = 7.87, p = 0.016) (Table 2).However, when calculated per soil volume, the differences were less obvious (SIR: F = 4.78, p = 0.049; BAS: F = 0.119, p = 0.976), and basal respiration in the 3-6 cm depth was even higher in BF (F = 5.61, p = 0.036).The site replicates differed significantly from each other, in BF concerning all soil variables except dry matter (p < 0.001 in most cases), and in BS concerning pH, particle size, SIR and basal respiration.

Macroarthropoda
The macroarthropod communities and biomasses were relatively similar at all study sites.When significant differences in biomasses between "Birch after Spruce" and "Birch after Field" were observed, they were higher in BS (Coleoptera and Diptera larvae, Homoptera and Chilopoda; Huhta 2002).The coleopteran community did not differ much between different kinds of birch stands, while there were more differences in Araneida.The spider communities were fairly similar to   those of coniferous forests, while in Coleoptera some of the most abundant species are rare in spruce forests.All dominant species in the study sites are common members of the forest floor fauna, and differ from those of open terrain.The populations in "Birch after Field" were generally low, and the communities could be characterised as impoverished forest communities.There were considerable differences between replicates of the same kind of forests, especially in Araneida.
(Biomass data for macroarthropods were given by Huhta (2002), and are not repeated here.)

Microarthropoda
There were significant differences between birch stands established after spruce forest and after cultivation, and between these and natural deciduous forests (Huhta and Niemi 2003, Huhta and Ojala, unpublished).The communities of "Birch after Spruce" were rather similar to those of natural spruce forests, but the population densities were lower.The populations in "Birch after Field" were low, especially in Oribatida, and in all cases when significant differences were found, they were in favour of BS.Total biomass of Collembola and Oribatida differed significantly between BS and BF.All the dominating species in the study sites are common members of the forest soil fauna.In terms of biomass, the most dominant species were Isotoma spp.among Collembola, Lysigamasus vagabundus and Veigaia nemorensis in Mesostigmata, and Nothrus silvestris, Ceratozetes gracilis and Hemileius initialis in Oribatida (Table 3).NMS ordination reveals a distinct grouping of the microarthropod communities (Monte Carlo permutation: p < 0.032) (Fig. 2).The BS sites are clearly separated in the ordination space both from each other and from the BF sites.Among BF, site BF1 is located most remote from the BS sites, while BF2 and BF3 overlap and lay between BF1 and BS.All the environmental factors except NH 4 are significantly correlated with the ordination axes, most distinctly with axis 3, with which pH, NO 3 , WHC, clay, fine fraction and aggregate content are in positive, and organic matter, SIR and pF in negative correlation (Table 4).

Enchytraeidae
Similarly to microarthropods, the enchytraeid communities also showed marked differences between the birch stands established on forest soil and on arable soil, but even more between replicates of similarly managed forests (Räty and Huhta 2004b).Total numbers and populations of most species were lower in "Birch after Field" than in "Birch after Spruce".Cognettia sphagnetorum, the "superdominant" enchytraeid in coniferous forests, was the second dominant in the BS sites.Fridericia ratzeli had the highest biomass in both site types, and strongly dominated the enchytraeid biomass in BF (Table 3).In the NMS ordination the enchytraeid communities of the BS and BF sites are located in opposite directions from the origo, though different samples are rather scattered and partly overlapping (Monte Carlo permutation: p < 0.020).As in microarthropods, site BF1 is distinctly separated from the other BF sites and BS (Fig. 2).Correlations between the axes and environmental factors are in agreement with those for microarthropods (Table 4).

Lumbricidae
Two of the "Birch after Spruce" sites harboured a community of Dendrobaena octaedra, Aporrectodea caliginosa and two or three species of Lumbricus, while D. octaedra alone was present in BS3.All these species were present in the three "Birch after Field" sites, where their total biomass, ranging from 14 to 26 g (d.m.) m -2 , was 2.6 times the average in "Birch after Spruce".(Species-level data for lumbricids, including biomass, are given by Räty and Huhta 2004a.)

Nematoda
Total numbers, diversity, and densities of most taxa and feeding groups were lower in "Birch after Field" than in "Birch after Spruce".The "BS" sites were relatively similar in community structure to natural deciduous forests, but the "BF" sites were dissimilar with both (Räty and Huhta 2003a).

Total Biomass
The average total biomass of all major soil animal groups except Lumbricidae was higher in "Birch after Spruce" than in "Birch after Field".However, the earthworm biomass showed an opposite trend, being lowest (virtually zero) in BS3 where the summed biomass of other groups was highest, and much higher in BF than in BS (Fig. 3).As the biomass of earthworms was several times higher than that of all remaining groups together, "Birch after Field" harboured, on the average, ca.2.5 times more soil animal biomass than did "Birch after Spruce" (Table 3).

Community Structure in Different Forests
It became obvious that even 30 years after reforestation the soil animal communities of planted birch stands established on spruce forest soil and on cultivated soil differ markedly from each other.Generally, the communities of "Birch after Spruce" are relatively similar to those of original spruce forests, though the population densities are usually lower.Concerning most faunal groups, the "Birch after Field" communities are poorer in densities, biomass and diversity, but they mainly share the same dominant species.They can thus be characterized as "impoverished" forest communities, having little in common with those of arable soils under cultivation.However, the animal groups also differ from each other in the degree of "community specialisation": Coleoptera and Collembola have relatively similar communities in different forests, Araneida, Mesostigmata and Nematoda deviate more, and in Oribatida, Enchytraeidae and Lumbricidae the differences between study sites are conspicuous.Earthworms are the only group to have a considerably higher biomass in "Birch after Field" than in "Birch after Spruce", and they largely determine the total faunal biomass.

Resource Quality and Soil Structure
Coniferous forest soil and litter are optimal habitats for such animals as microarthropods and enchytraeids.Most of the species are microbidetritivores, either fungal feeders or consumers of dead organic matter with fungal growth inside (Petersen 1971, Luxton 1972, Springett and Latter 1977, Latter and Howson 1978, Siepel and de Ruiter-Dijkman 1993).The amount and quality of organic matter are considered an important determinant for the fungal-based soil food web (Swift et al. 1979).The average plant community differs between the different study sites: grasses dominate in "Birch after Spruce", herbs in "Birch after Field" and dwarf-shrubs in coniferous forests.
As litter derived from different plants differs in quality, this might result in differences in the litter-feeding fauna.An experiment was carried out by Ojala and Huhta (unpublished), in which the preference of Collembola and Oribatida for different litter substrates was tested.No species showed preferences for birch litter over spruce or over grasses and herbs; thus the presence of birch did not explain their abundances in the study sites.Most species that were equally common in BS and BF, preferred either grass, herb or fern litter, or two of these.This is in accordance with Maraun and Scheu (2000), who concluded that the forest type or litter quality is of minor importance for the density and community structure of oribatids.
The study sites also differ in the physical structure of the topsoil.In "Birch after Spruce" there is a loose microstructure among litter and bases of grass tussocks, whereas in BF there is a scanty litter layer with more distantly-spaced bases of herbs.Physical structure of the soil surface is probably an important factor to explain the community structure of spiders.Small web-building species (Linyphiidae that form the majority of the spider community of forest floor (Huhta 1971) require small spaces for anchorage of their webs.This probably explains the greater numbers and diversity of spiders in "Birch after Spruce" stands than in "Birch after Field".For euedaphic animals such as enchytraeids the physical structure is less important.

Soil Acidity
pH is generally regarded as one of the most important environmental factors for soil fauna.Many species are intolerant to acid environments, but several oribatids, the enchytraeid Cognettia sphagnetorum and the earthworm Dendrobaena octaedra reach the highest densities in raw humus soils with low pH (Standen and Latter 1977, Hågvar and Abrahamsen 1980, Standen 1980, Koskenniemi and Huhta 1986, Didden 1993).The species diversity of Oribatida is highest in acid soils, while most enchytraeids are acid-sensitive; Standen (1980) and Nowak (2001) have observed a positive correlation between pH and the number of enchytraeid species.As the "Birch after Field" sites have higher soil pH than the "Birch after Spruce" sites, the species richness of enchytraeids is expected to be higher, and that of oribatids lower in BF than in BS.This is supported by our observations in oribatids, but not in enchytraeids.Increase of enchytraeid diversity and decrease of oribatid diversity in "Birch after Spruce" in relation of original spruce forest can also be expected, which agrees with the observations.pH also contributes to explain the differences in community structure of nematodes.The higher soil pH in "Birch after Field" originates from the time of cultivation, since arable soils are routinely neutralized by liming, the effect of which is very long-lasting (Aarnio et al. 2003).In "Birch after Spruce" an increase of pH in comparison with original spruce forest is expected due to deciduous litter (Mikola 1985).
However, pH alone can not explain all the variation between the study sites.Soil pH may in fact be less important than previously believed, since moder soil of pH 5.5 may harbour an abundant oribatid community with similar species structure as acid raw humus soils (Maraun and Scheu 2000).Hågvar (1990) also has shown that some "acidophilic" oribatids actually reproduce better in more neutral soil in the absence of competition by other microarthropods.

Dispersal
Long-lasting cultivation is expected to decimate the populations of forest soil arthropods.Ability to colonise vacant environments is thus a factor that may influence the possibilities of forest species to invade the BF sites after reforestation.For flying insects such as most Coleoptera this is hardly a problem.Spiders are also good colonisers due to their capability for aeronautic dispersal.Epigeic spiders, beetles and large predatory mites have good dispersal ability by land.Many nematodes and mesostigmatid and astigmatid mites have capacity for phoretic dispersal (Nicholas 1983, Karg 1989, 1993).In contrast, dispersal ability of small wingless soil animals may be poor.
Regrettably, virtually nothing is known about dispersal in Enchytraeidae, Oribatida and edaphic Collembola.Onychiurus armatus (Collembola) and Cognettia sphagnetorum (Enchytraeidae) have been observed to spread only a few centimeters per week in the laboratory (Sjögren et al. 1995, Sjögren 1997, Salminen and Sulkava 1996).In connection with the present study, a dispersal experiment was carried out by Ojala and Huhta (2001), which indicates that most microarthropod species have potential capacity for land colonisation over the distance of 30 m (from forest edge to the sampling sites) in the time of 30 years that had elapsed after reforestation.Even the species with the slowest rate of dispersal (Medioppia subpectinata, Oribatida) had viable populations in all the BF sites.
Passive transport by man is known to play a decisive role in the dispersal of earthworms (Lee 1985).Species invasions to new localities are thus accidental, and the species assemblage at a certain site is largely determined by chance.The same is probably true in Enchytraeidae, which are similar to earthworms in general biology.Contrary to BS, the "Birch after Field" sites have a long cultural history before reforestation, and most potential species could be expected to have colonized the sites.All potential earthworm species as well as diverse enchytraeid communities were in fact present in the BF sites, while the earthworm community in the most isolated BS site (BS3) solely consisted of the ubiquitous species Dendrobaena octaedra.However, the latter site harboured a diverse community of Enchytraeidae.

Interspecific Interactions
There was an overall inverse relationship between the abundance of earthworms and other animal groups.Earthworms had a high average biomass in the "Birch after Field" sites where microarthropods, enchytraeids and nematodes were less abundant, and with the exception of enchytraeids also their communities were less diverse.Among the "Birch after Spruce" sites, earthworms were virtually lacking in BS3, where all other groups showed the highest biomass.
There is both correlative and experimental evidence that earthworms can suppress populations of other soil animals.Schaefer and Schauermann (1990) found in German beech forests that the abundances and biomasses of Lumbricidae and Enchytraeidae were inversely related.Huhta and Viberg (1999) have shown that the earthworm Dendrobaena octaedra reduces the population of Cognettia sphagnetorum.A negative relationship between earthworms and nematodes has been observed by Yeates (1981).Hyvönen et al. (1994) showed that the presence of the earthworm Dendrobaena octaedra suppresses the populations of nematodes, and suggested that earthworms consume microfauna as part of their diet.In connection with the present study, Räty and Huhta (2003b) carried out an experiment where the earthworms Lumbricus terrestris and Aporrectodea caliginosa reduced substantially the enchytraeid and nematode populations.
Physical structure of the topsoil is related to the abundance of burrowing earthworms that mix the soil layers and consume the litter layer which, during its decomposition, provides nutrition for litter-feeding fauna and fungal-based soil food web.Thus earthworms probably affect negatively the populations of several soil animals, e.g.spiders and microarthropods, by disturbing their microhabitats and reducing their resources.Maraun and Scheu (2000) also concluded that disturbance and reduction of habitable space by earthworms is a major determinant of the community structure of Oribatida.

Conclusions
The community structure of soil fauna is markedly different in birch stands established on earlier spruce forest soil and on arable soil.The communities in "Birch after Spruce" are relatively similar to those of coniferous forests, though the population densities are generally lower.The communities of "Birch after Field" are generally sparse and could be characterised as "impoverished forest communities", except Lumbricidae and Enchytraeidae that have affinities with deciduous forests and cultural landscapes.
The soil characteristics in birch stands established on earlier forest and cultivated soils differ from each other to such an extent that this can be expected to influence the decomposer communities for several decades after reforestation.Some of these characteristics (clay and fine mineral matter) reflect the original soil properties, some others (pH and soil stratification) result from agricultural practices, and some are secondary properties resulting from the former (organic matter, water holding capacity, nitrification).
However, the prevailing soil conditions are not enough to explain all observed differences in soil fauna.At least for earthworms, probably also for enchytraeids and many other edaphic animals, colonisation abilities and accidental transport by man may determine the presence or absence of certain species at a site.The long cultivation his-tory of the "Birch after Field" sites "guarantees" the presence of the common burrowing earthworms, while isolation from cultural landscape may prevent colonisation in "Birch after Spruce".On the other hand, the biomass of earthworms also depends on soil acidity, and simply due to low pH may remain lower in BS than in BF.
Earthworms, further, have marked impacts on several soil properties due to their burrowing activity and feeding on litter.They also directly or indirectly compete with or otherwise have a negative effect on other soil faunal components.Thus the observed structure of the soil communities and the differences between "Birch after Spruce" and "Birch after Field" are an outcome of a complex and interacting set of factors: soil characteristics, site history, dispersal and interspecific interactions.

Fig. 3 .
Fig. 3. Soil faunal biomasses (dry matter /m 2 ) on the study sites; means of 2 or 3 samples (please note the different scales).

Table 1 .
Field-layer vegetation of the study sites in July 1998 (coverage %; species exceeding cover of 0.5% at any site are included).

Table 3 .
Biomasses (mg d.m./m 2 ) of the most dominant species in biomass, and totals of each main group (ANOVA main effects shown on the row 'total').