(ProQuest: ... denotes formulae omitted.)

Introduction

The dynamics of arthropod communities in forest ecosystems can be understood only when based on the trophic connections established between them (Scheu, 2002). Given the importance of many of the species in detritic food chains, and the influence of their activities on the rates of decomposing organic compounds and the dynamic of nutrients (Bird et al., 2000), terrestrial arthropods from litter and soil are a useful indicator of forest ecosystem conditions and changes within them (Hole, 1982; Kopeszki, 1992; Curry, Good, 1992; Hogervorst et al, 1993; Hoekstra et al., 1995). This presented an argument for the recommendation to use soil and litter arthropods for the assessment of the biological effects of silvicultural practices of different intensities. These practices included intensive cuts, the use of pesticides and fertilizers and also soil bedding (Bird et al., 2000).

Measuring litter diversity through direct investigations is costly and therefore the use of indicators becomes necessary (Gaston, 2000; Kerr et al., 2000; Ekschmitt et al., 2003). The fact that species considered functionally redundant (according to the theory of redundancy) can gain a new functional significance due to their interaction with other species (Andren et al., 1995), and that changes in species composition is not a random process (Andren et al., 1995; Wolters, 2001) indicate the great stability which can be exhibited by litter biota. Therefore, species diversity is recommended as an indicator of ecosystem changes (Didham et al, 1996; Gibb, Hochuli, 2002). This argument is further strengthened by the fact that litter biota functioning is assured by the multiplication of roles maintaining a high level of specific diversity in the community (Wolters, 2001).

The basic elements of trophic webs are not strictly specialized species, but they are trophic groups of species which have equivalent roles. Detritophagous organisms consume amalgamations of particles and substances or microorganisms, and cannot be delimited by the main component of their diet (Scheu, Setälä, 2001). Since soil and litter organisms have a flexible diet they can not be appropriately assigned to a certain trophic module. Dependent on the availability of resources, these organisms can feed on algae, fungi, detritus or microorganisms, or they can be phytophagous or predators (Ponge, 1991; Walter, 1987; Maraun et al, 1998).

Although soil and microclimate parameters are generally significantly correlated with diversity, despite different measurement types, they are unsatisfactory predictors explaining less than 50% of the variation (Ekschhmitt et al., 2003). Diversity is also frequently subjected to variations across sub-samples (CV = 20-60%), mainly due to the variation in microclimatic parameters (Ekschhmitt et al., 2003). These variations can be avoided by limiting the correlation with environmental factors. Therefore, some authors suggested estimating diversity at higher taxonomical levels (Gaston, 2000; Kerr et al., 2000).

We assume that litter levels micro-climatic conditions. Therefore, the dynamics of litter communities are determined by the availability of resources and by the rates of mass and energy transfer. These rates are determined by the structure and composition of the arboretum and whether it is natural or modified.

Epigeal communities remain an effective measure of changes at the ecosystem level.

This study emphasizes the beneficial use of diversity in epigeal communities as an instrument for investigating changes in ecological systems as a result of silvicultural practices. Here, we used litter invertebrate fauna which is richer than the invertebrate fauna in the fermentation layer (Evans et al., 2003). Important differences between the larval and imaginai stage with respects to their belonging to a certain trophic category determined that the focus on the diversity of epigeal fauna should be placed on the latter stage.

The overall goal of the study is to compare the diversity of epigeal fauna in two forest stands, one subject to stronger anthropic interventions over a longer period, and thus hypothesizing that the greater diversity of the fauna will correspond to a greater diversity in flora and to less human intervention. If results confirm the hypothesis, this will offer practitioners arguments for selecting near-natural techniques in silvicultural management.

Material and methods

Study sites

The study was conducted in two stands situated in a mixed lime-oak-hornbeam forest (Donila et al., 1990) located in the plain region of Southern Romania near Bucharest (40°38'7.77° N, 26°9'18.76" E). The stands have the following structure: ( 1 ) oak wood (Querceta roboris). with a dense litter and highly covered soil, less shrubbery with average-developed vegetation characteristic of the plains: Arum orientale M. B i e b. and Pulmonaria officinali! L.; (2) oak wood (Querceta roboris) with derived arboreta: Acer campestre L., and Carpinus betulus t., dense litter, with average-developed shrubbery containing mostly Crataegus nomogyna Ja cq., and Ligustrum vulgare L., and vegetation consisting of associations of Brachypodium sylvaticum (H u d s.) Beau v., Geum urbanum L., Pulmonaria officinola L., Asperula sp., and Dentaria bulbifera L. A more detailed comparison of the two stands is presented in Table 1.

Treatments

Both stands form part of a forest created as an artificial regeneration plantation 70 years ago. Vegetation in the first stand (referred as "site A") consists mainly of Quercos robur L. pedunculate oak, resulting from selective cuts and artificial regeneration, so that other species were replaced by this pedunculate oak.

The vegetation of the second stand (referred as "site B") consists of pedunculate oak and other tree and shrub species, which are present in different proportions, thus generating the characteristic mosaic of mixed oak forests.

Sampling

Pitfall traps were used to collect epigeal fauna. Traps consisting of transparent plastic containers with a diameter of 15 cm, containing a preserving liquid consisting of a small quantity of detergent added to 4% formaldehyde solution, to modify the surface tension of the liquid, were buried completely up to the container top. For each site, traps were placed in lines often at a distance of 1.5 m apart. This configuration ensured the best representation of fauna in the probes. Sampling was performed in May, June, July, and August, which corresponded to the vernal and estival periods. Traps were installed for a period of one week each month. These periods reflect the main stages in the vegetation cycle of leafing, blooming and fruiting.

Samples were transported to a laboratory and the arthropods were identified and stored in 80% isopropyl alcohol. All invertebrates were classified to the family level, except for Pseudoscorpiones, Araneae, Opiliones, Acari, Isopoda, Diplopoda and Chilopoda which are considered recognizable taxonomical units. Individuals considered to be at the same taxonomical level were assigned to functional groups (Bird et al., 2000). The insects were also determined up to the order level, except for beetles (Coleoptera), flies (Diptera), wasps (Vespoidea), and true bugs (Heteroptera). The abundance of each taxon was recorded. Due to the fact that the trophic category can differ between larval and imaginai stages, the larvae were not accounted for.

Statistical analysis

To compare the diversity of the two sites, we used three indices: species diversity and two indices of similarity, Jaccard's and Sorenseris (Magurran, 1998).

Jaccard's index: nc / (nA + nB - nc) [Equation 1]

Sorensons index: 2 nc / (nA + nB) [Equation 2]

where nA and nB are the number of species in sites A and B, and nc = number of common species.

In addition, similar to methods employed in the study by Petrisor (2000), we computed diversities using Shannons informational entropy and compared them for the two communities using a test proposed by Magurran (Magurran, 1998). In addition, the specific frequency distributions during each month individually and for the entire period of study were calculated using the χ^sup 2^ test for "goodness of fit" of the two empirical distributions (Hutcheson, 1970). Shannon's informational entropy is denned as

... [Equation 3]

where h is the informational entropy, expressed in this case as nits; pi represents the relative frequency of the species i; and s represents the total number of species in the community (Petrisor, 2000; Magurran, 1998; Hutcheson, 1970).

The comparison of two empirical entropies is performed using the following test

... [Equation 4]

where,

... [Equation 5]

Under the null hypothesis (i.e., no difference between the two distributions, tested against the alternative "any difference"), the test follows a t (Student) distribution which contains the following number of degrees of freedom (Petrisor, 2000; Magurran, 1998; Hutcheson, 1970)

... [Equation 6]

Hutcheson (1970) proposed a chi-square (2) test for comparing specific frequency distributions of two communities. The z test is computed using the formula

... [Equation 7]

where X^sub n-1^^sup 2^ is the test statistic; under the null hypothesis, its distribution is 2 with n-1 degrees of freedom; Ri is the frequency of species i, in the reference period; and Oi is the frequency of species i, in the studied period (Petrisor, 2000; Hutcheson, 1970).

All computations were performed using an Excel® spreadsheet, except for the p-values which were computed by a software application implemented via the Internet (Arsham, 2004).

Results

Taxonomic considerations

Individuals collected in all samples were found to belong to five classes of arthropods: Arachnida (Pseudoscorpiones, Araneae, Opiliones, Acari), Malacostraca (Isopoda: Oniscidea), Diplopoda, Chilopoda and Insecta. Over the entire period, 9439 individuals were collected in May (6932 in Site A and 2507 in Site B), 4638 in June (2640 in Site A and 1998 in Site B), 5125 in July (2231 in Site A and 2894 in Site B), and 265 (81 in Site A and 187 in Site B) in August (Fig. 1 and Table 4). This gave totals of 11884 individuals collected at Site A and 7583 at Site B (Fig. 2). For the distribution at the order and family level, please refer to Table 2. From this, 12 taxons were present in site A only, 12 in site B only, while 51 were found in both sites.

In both sites, the Insecta had the highest relative abundance: 84.77% in site A and 77.54% in site B. The groups of arthropods with high relative abundance were: Collembola (67.96% in site A and 51.64% in site B), Araneae (7.20% in site A and 8.25% in site B), Formicidae (4.19% in site A and 7.39% in site B), Acari ( 3.95% in site A and 4.19% in site B), Opiliones (1.63% in site A and 6.04% in site B), Carabidae (2.64% in site A and 3.44% in site B), Sciaridae (2.99% in site A and 2.29% in site B), lsopoda (1.64% in site A and 3.08% in site B), Tipulidae (1.03% in site A and 1.71% in site B) and Phoridae (0.72% in site A and 2.34% in site B). All these taxonomic groups, except for the three families of Diptera (Sciaridae, Tipulidae and Phoridae), are characteristic of the epigeal fauna. We assume that the high abundance of the families of Diptera in our captures was a result of sampling bias; we used pitfall traps without covers.

Community considerations

Table 3 displays a simple comparison of diversity in the two sites based on species richness, and similarity measured using Jaccard's and Sorenson's indices. The values indicate that the sites are more than 50% similar, regardless of the period or index used in comparison. The results of comparing informational entropy for the two communities are presented in Table 4. Tests were significant at 0.001 in May, July and overall, but not in June (p = 0.226) and August (p = 0.376), thus indicating greater diversity in site B in May, June, and overall, and in site A in the remaining periods. Table 5 presents the results of the comparison performed using the chi-square test to compare the specific distributions. All tests exhibited significance at 0.001, indicating significantly different distributions for each month individually, and also for the entire study period.

Discussion

Species richness appears to be greater in community A for each month individually and also overall. However, comparisons based on informational entropy indicate greater diversity in community B during May and June, and greater diversity in community A during July and August. Monthly differences are significant only in May and July. For the entire period of study, community B exhibits significantly greater diversity. While results appear to be different when using the two methods, it is worth mentioning that comparisons based on Shannon's entropy index lead to sound results and are sustained by statistical significance. Specific distributions significantly differ for all months, for the entire period of study and also in all scenarios.

Our findings suggest that epigeal community diversity is greater in forests with increased tree species, and sustain the hypothesis according to which these communities can indicate structural changes (Evans et al., 2003).

The biodiversity data shifts with proportional increase in taxonomic resolution (Doledec et al., 2000). The literature reviews identified numerous papers in which sites, samples or treatments are compared based on ordinal-level abundance data, similar to this study. Since many of these processes operate at the species level, although there are some exceptions (New, 1996), few of these studies make a significant contribution to our understanding of the processes at work. A good practice is to allocate material to the species or morpho-species level if possible. Similarly, a beneficial practice is to treat all samples as separate entities throughout processing unless there is a very good reason to aggregate them before processing. Keeping them separate provides greater flexibility to the analysis. This is especially appropriate to detect levels of between-sample variation or species turnover, or to estimate species richness by examining the rate of species accumulation by sample (Brose, 2002; Cam et al., 2002). An additional issue arises when material from a single species includes a mixture of developmental stages. This occurs in insect taxa which undergo incomplete metamorphosis, including bugs and grasshoppers, and also in many other invertebrate taxa, such as millipedes and molluscs. Presuming that data refers only to adults, when creating a database for such material it is important not to aggregate them if subsequent analysis treats all material as having equal status. At the macro-taxonomical level of order and family, a similar number of individuals or, otherwise a significant difference can be obtained from two compared sites (Grove, 2003; Krebs, 1989; Nitzu et al., 2009).

From an ecological viewpoint, we conclude that the greater diversity of site B, correlated with the fact that its structure is the closest to the natural one, suggests that silvicultural practices closer to the natural model provide for greater diversity, and this leads to increased stability under pressures induced by anthropogenic impact (Tomescu, Savu, 2002).

Finally, from a statistical methodology viewpoint, the results herein emphasize the utility of using the t-test in the comparison of Shannon's informational entropy index when comparing diversity (Magurran, 1998).

Conclusion

The functions of forest ecosystems change gradually from forests with a structure closer to the natural one to those subject to silvicultural interventions such as selective cuts and artificial regeneration across the grathent of anthropogenic impact; the endpoints can easily be distinguished functionally based on the structure of plant communities, and herbivorous and detritivorous organisms.

Our results indicate that diversity in epigea! communities is greater in forest sites with more tree species, suggesting that silvicultural practices closer to the natural model increase the diversity of forest ecosystems. Therefore, forest management has an effect on the richness and diversity of litter fauna, due to the interactions of trophic guilds. Consequently, the surrogacy of higher taxa can be seen as a valuable approach when it is impossible to identify the specimens at a low taxonomic level in a reasonable period of time, and in a context of limited financial resources. However, further studies are required to test whether the results herein are specific to the studied systems or if they can be generalized to "different" types of forest management.

Translated by the authors

English corrected by R. Marshall

Acknowledgements

The authors wish to thank M. Ianculescu, Member of the "Gheorghe lonescu-Cis. esti" Academy of Agricultural and Forestry Sciences and Adrian-Iovu Biri? of the University of Agricultural Sciences and Veterinarian Medicine in Bucharest for their valuable comments and suggestions which helped to improve the quality of this manuscript, and to R. Marshall for his editorial assistance with proofreading the manuscript.

References

Andren, O., Bengtsson, J., Clarholm, M., 1995: Biodiversity and species redundancy among litter decomposers. In Collins, H.P., Robertson, G.P., Klug, MJ. (eds), The Significance and Regulation of Soil Biodiversity. Kluwer Academic Publishers, Dordrecht, p. 177-188.

Arsham, H., 2004: The p-values for the popular distributions. Program available on-line at http://home.ubalt. edu/ntsbarsh/Business-stat/otherapplets/pvalues.htm.

Bird, S., Coulson, R.N., Crossley, D.A. Jr., 2000: Impacts of silvicultural practices on soil and litter arthropod diversity in a Texas pine plantation. For. Ecol. Manage., 131: 65-80.

Brose, U, 2002: Estimating species richness of pitfall catches by non-parametric estimators. Pedobiol., 46: 101-107.

Cam, E., Nichols, J.D., Sauer, J.R., Hiñes, J.E., 2002: On the estimation of species richness based on the accumulation of previously unrecorded species. Ecography, 25: 102-108.

Curry, J.P., Good, J.A., 1992: Soil faunal degradation and restoration. Adv. Soil Sci., 17: 171-215.

Didham, R.K., Ghazoul, I., Stork, N.E., Davis, A.J., 1996: Insects in fragmented forests: a functional approach. Trends Ecol. EvoL, 11: 255-260.

Doledec, S., Olivier, J.M., Statzner, B., 2000: Accurate description of the abundance of taxa and their biological traits in stream invertebrate communities: effects of taxonomic and spatial resolution. Arch. Hydrobiol, 148: 25-43.

Donila, N., Chiritä, C., Stänescu, V. (eds), 1990: Types of forest ecosystems in Romania (in Romanian). ICAS, Bucharest, 389 pp.

Ekschmitt, K., Stierhof, T., Dauber, J., Kreimes, K., Wolters, V, 2003: On the quality of soil biodiversity indicators: abiotic and biotic parameters as predictors of soil faunal richness at different spatial scales. Agr. Ecosys. Environ., 98: 273-283.

Evans, A.M., Clinton, P.W., Alien, R.B., Framptond, C.M., 2003: The influence of logs on the spatial distribution of litter-dwelling invertebrates and forest floor processes in New Zealand forests. For. Ecol. Manage., 184: 251-262.

Gaston, K.J., 2000: Biodiversity: higher taxon richness. Prog. Phys. Geog., 24: 1 17-127.

Gibb, H., Hochuli, D.F., 2002: Habitat fragmentation in an urban environment: large and small fragments support different arthropod assemblages. Biol. Cons., 106: 91-100.

Grove, S.J., 2003: Maintaining data integrity in insect biodiversity assessment projects. ). Insect Cons., 7: 33-44.

Hoekstra, J.M., Bell, R. T., Launer, A.E., Murphy, D.D., 1995: Soil arthropod abundance in coastal redwood foresteffect on selective timber harvest. Environ. Entomol., 24: 246-252.

Hogervorst, R.F., Verhoef, H.A., van Straalen, N.M., 1993: Five-year trends in soil arthropod densities in pine forests with various levels of vitality. Biol. Pert. Soils, 15: 189-195.

Hole, F.D., 1982: Effects of animals on soil. Geoderma, 25: 75-112.

Hutcheson, K., 1970: A test for comparing diversities based on the Shannon formula. J. Theor. Biol., 29: 151-154.

Kerr, J.T., Sugar, A., Packer, L, 2000: Indicator taxa, rapid biodiversity assessment, and nestedness in an endangered ecosystem. Cons. Biol., 14: 1726-1734.

Kopeszki, V.H., 1992: A first attempt using soil dwelling collembolan species Fosomia candida (W i I e m) and Heteromurus nitidus (Templete n) asan active bioindication in a beech forest ecosystem. Zool. Anzeiger, 228: 82-90.

Krebs, C.J., 1989: Ecological methodology. Harper and Row Publishers, New York, 654 pp.

Magurran, A.E., 1998: Ecological diversity and its measurement. Princeton University Press, Princeton, 179 pp.

Maraun, M., Migge, S., Schafer, M., Scheu, S., 1998: Selection of microfungal food by six oribatid mite species (Oribatida, Acari) from two different beech forests. Pedobiol., 42: 232-240.

New, T.R., 1996: Taxonomic focus and quality control in insect surveys for biodiversity conservation. Aust. J. Entomol., 35: 97- 106.

Nitzu, E., Olenici, N., Popa, I., Nae, A., Bins, I., 2009: Soil and saproxylic species (Coleóptera, Collembola, Araneae) in primeval forests from the northern part of South-Eastern Carpathians. Ann. For. Res., 52: 27-54.

Petriçor, A.I., 2000: Study about the degradation of the lentie ecosystems in Pantelimon lakes (Bucharest, Romania). In American Statistical Association (ed.), Proceedings of the Section on Statistics and the Environment. Alexandria, VA, 10: 68-70.

Ponge, J.E, 1991: Food resources and diets of soil animals in a small area of Scots pine liner. Geoderma, 49: 33-62.

Scheu, S., Setälä, H., 2001: Multitrophic interactions in decomposer food webs. In Tscharntke, T., Hawkins, B. A. (eds), Multitrophic Interactions in Terrestrial Systems. Cambridge University Press, Cambridge, p. 223-264.

Scheu, S., 2002: The soil food web: structure and perspectives. Eur. J. Soil Biol., 38: 11-20.

Tomescu, I., Savu, A.D., 2002: The relationship between diversity and stability in forest ecosystems (in Romanian). In "Constantin Brincu;i* University - Engineering Faculty (ed.), Proceedings of University's Day 8th International Conference, "Constantin Brâncusi" University, Tärgu Jiu, p. 1-4.

Walter, D.E., 1987: Trophic behavior of "mycophagous" microaithropods. Ecol, 18: 226-229.

Wolters, V, 2001: Biodiversity of soil animals and its function. Eur. J. Soil Biol., 37: 221-227.

Author affiliation:

DRAGOS A. SCAUNACU1, ALEXANDRU I. PETRISOR2, FÍNICA M. IVANOV

1 Faculty of Ecology, Ecological University of Bucharest, bd. Vasile Milea nr. IG, sector 6, cod 061341, Bucharest, Romania; e-mail: scaunasu_dragos@yahoo.com, finica_ivanov@yahoo.com

1 Faculty of Urbanism, "Ion Mincu" University of Architecture and Urbanism, str. Acaderaiei nr. 18-20, sector 1, cod 010014, Bucharest, Romania; e-mail: a.i.petrisor@gmail.com