Differences in leaf and root litter decomposition in tropical montane rainforests are mediated by soil microorganisms not by decomposer microarthropods

Study area

The study area is in the northern fringes of the Podocarpus National Park in the eastern slopes of the Andean Cordillera, Southeast Ecuador. Three study sites, situated at 1,000, 2,000 and 3,000 m a.s.l. were selected to represent an altitudinal gradient with moderately steep slopes of 26–31° (Moser, Hertel & Leuschner, 2007). The lower site at 1,000 m a.s.l. (S04 °06′54″, W78 °58′02″) is in the Río Bombuscaro valley and classified as evergreen submontane rainforest dominated by tree species of Arecaceae, Combretaceae, Moraceae, Monimiaceae, Rubiaceae and Sapotaceae (Homeier et al., 2008). The intermediate site at 2,000 m a.s.l. (S3 °58′18″, W79 °4′45″) is in the Reserva Biológica San Francisco on the north-facing flank of the Río San Francisco valley and consists of an evergreen lower montane rainforest dominated by trees of Arecaceae, Clusiaceae, Ericaceae, Lauraceae, Melastomataceae and Rubiaceae (Homeier et al., 2008). The highest site at 3,000 m a.s.l. (S04 °06′711″, W79 °10′58″) is near the upper Cajanuma mountain at the northwest gate of Podocarpus National Park. The forest has been classified as an evergreen elfin forest dominated by trees/shrubs of Aquifoliaceae, Bromeliaceae, Chloranthaceae, Clusiaceae, Ericaceae and Melastomataceae (Homeier et al., 2008). The climate is semi-humid with an average annual temperature of 14.9, 12.3 and 8.9 °C and annual precipitation of approximately 2,200, 3,500 and 4,500 mm at 1,000, 2,000 and 3,000 m a.s.l., respectively (Bendix et al., 2006; Homeier et al., 2010). Soil types of the study sites are alumic Acrisol (1,000 m), Gley Cambisol (2,000 m) and Podzol (3,000 m) (Soethe, Lehmann & Engels, 2006; Moser, Hertel & Leuschner, 2007). The thickness of the organic soil layers increases with altitude from 4.8 cm at 1,000 m to 30.5 cm at 2,000 m to 43.5 cm at 3,000 m (Leuschner et al., 2007; Graefe, Hertel & Leuschner, 2008). In parallel, fine root biomass increases from 2.7 to 6.2 to 10.8 t ha⁻¹ at the respective sites (Soethe, Lehmann & Engels, 2006).

Experimental design

Nylon litterbags (17 × 17 cm) with mesh sizes of 45 µm, 1 mm and 4 mm were filled with 10 g of leaf or root litter. The mesh sizes we used are typically used to evaluate the role of meso- and macrofauna in decomposition processes. The 45 µm mesh prevents access by even the smallest mites and springtails, the 1 mm mesh allows access by virtually all mesofauna species and the 4 mm mesh allows access by virtually all macrofauna species (Bradford et al., 2002; Kampichler & Bruckner, 2009; Bokhorst & Wardle, 2013). Leaf litter consisted of a mixture of freshly fallen leaves of three locally abundant tree species of each study site: Pouteria sp., Cecropia andina and Mollinedia sp. at 1,000 m, Graffenrieda emarginata, Clusia sp. and Cavendishia zamorensis at 2,000 m, Clusia sp. Graffenrieda emarginata and Hedyosmum sp. at 3,000 m. Root litter were collected by hand from the upper 20–30 cm of the soil/organic layer of respective sites and consisted of a mixture of three size classes: Small (<1 mm diameter), medium (1–2 mm diameter) and large (>3 mm diameter). The amount of individual leaf species and root size classes placed in the litterbags was chosen to resemble their amount in the litter layer and soil, respectively (see SI 1). This was done to mimic close to natural conditions in the litterbags for the colonization by the local soil animal community. The collected leaves and roots were gently rinsed with tap water to clear them from adhering soil and dried at 60 °C.

Litterbags were placed in the field in October 2008 (end of the rainy season). Bags containing leaf litter were randomly placed on top of the litter layer and fixed with nails, while those containing root litter were placed approximately 5 cm below the litter layer. Three blocks were established at each of the three altitudes with a minimum distance between blocks of 20 m. Two replicates of each treatment were placed in each block, with one replicate retrieved after 6 months and the other after 12 months.

After retrieval, the litter material in each litterbag was divided into two parts of equal mass. The first half was analyzed for dry mass, microbial biomass, basal respiration, and C and N concentrations. From the second half, Oribatida and Collembola were extracted using a modified high-gradient heat extractor (Macfadyen, 1961; Kempson, Lloyd & Ghelardi, 1963) and counted. Adult Oribatida were identified to species level or sorted into morphospecies (Balogh & Balogh, 1990; Balogh & Balogh, 2002), following the nomenclature of Subías (2018). All identified species are recorded in the Ecotaxonomy database (Potapov, Sandmann & Scheu, 2019).

Analytical procedures

Mass loss (M_loss) for both leaves and roots were calculated as M_loss = ((m₀ − m₁/m₀)) × 100, where m₀ is the dry weight of the initial litter placed in the litterbags and m₁ the dry weight of litter at harvest. To measure carbon (C) and nitrogen (N) concentrations, dried (60 °C, 72 h) leaves and roots were milled to powder (<1 mm) and analyzed using a CN elemental analyzer (Vario EL III, Elementar, Hanau, Germany).

Microbial basal respiration (BR) and microbial biomass (C_mic) were measured using a computer-controlled O₂ micro-compensation apparatus (Scheu, 1992). BR (µl O₂ g⁻¹ dry weight h⁻¹) was determined as mean O₂ consumption rates 10 to 20 h after attachment of the samples to the respirometer. C_mic was calculated from the maximum initial respiratory response (MIRR; µl O₂ g⁻¹ h⁻¹) measured after glucose saturation following the substrate-induced respiration method (SIR) of Anderson & Domsch (1978). MIRR was calculated as the average of the lowest three readings within the first 10 h and C_mic was calculated as C_mic = 38 × MIRR (mg g⁻¹ dry weight) (Beck et al., 1997; Joergensen & Scheu, 1999).

Statistical analyses

Analyses were performed using R version 3.6.0 (R Core Team, 2019). Linear mixed-effects models were used to analyze changes in M_loss, C_mic , BR, the abundance of Oribatida and Collembola, and Oribatida species richness data separately for leaf and root litter (hypotheses 1 and 2). Sampling date (6 and 12 months), mesh size (45 µm, 1 mm and 4 mm), altitude (1,000, 2,000 and 3,000 m a.s.l.) and all possible interactions were fitted as fixed factors, and block was fitted as a random factor with random intercept (R package “nlme”; Pinheiro et al., 2020). Data were log-transformed if necessary. Model residuals were checked for normality and homoscedasticity using Shapiro–Wilk test and Bartlett’s test (R package “stats” R Core Team, 2019). Differences between means were inspected using Tukey’s honestly significant difference test (R package “emmeans”; Lenth, 2020). Pearson correlation coefficients (package “stats”) were calculated to investigate relationships between M_loss, C_mic , BR, C-to-N ratio, and the abundance of Collembola and Oribatida, and Oribatida diversity separately for leaf and root litter (hypothesis 3).

Canonical correspondence analysis (CCA) performed in CANOCO 5.02 (Ter Braak & Smilauer, 2012) was used to explore the relationship between Oribatida community composition and litter characteristics (M_loss, C-to-N ratio) as well as microbial indicators (BR, C_mic ) (hypothesis 4). Only species with more than three individuals in the samples were included. Monte Carlo randomization tests using 999 simulations were used to determine the significance of the axes. Sampling date (6 and 12 months), mesh sizes (45 µm, 1 mm and 4 mm) and altitude (1,000, 2,000 and 3,000 m a.s.l.) were coded as supplementary variables not affecting the ordination. Since the global test with all litter and microbial indicators was significant, we used forward selection to identify the most important variables structuring Oribatida communities. This was done to reduce the number of explanatory variables entering the analysis while keeping the variation explained caused by them at a maximum. The forward selection procedure was stopped if a variable reached a level of significance > 0.05 and R² > 0.90 (Legendre & Legendre, 1998; Blanchet, Legendre & Borcard, 2008).

Decomposition of leaves and roots

Generally, M_loss significantly increased during the time of exposure in both leaf and root litter. Leaf litter M_loss was not significantly affected by any interaction between the three factors studied, but it was higher at 1,000 and 2,000 m compared to 3,000 m at both sampling dates (Table 1, SI 2). By contrast, in root litter the interactions between altitude and date, as well as between altitude and mesh size significantly affected M_loss. After 12 months it decreased in a linear way with increasing altitude (Fig. 1, Table 1). Further, at 3,000 m root litter M_loss in litterbags of 4 mm mesh size was higher than in litterbags of 45 µm and 1 mm mesh size (Table 1, SI 2). M_loss positively correlated with C_mic , BR and negatively with litter C-to-N ratios in both leaf and root litter. In root litter, M_loss also positively correlated with Collembola and Oribatida abundance and Oribatida richness (Table 2).

The C-to-N ratio of both leaf and root litter significantly increased with altitude from 1,000 m to similar values at 2,000 m and 3,000 m (Table 1, SI 2). In leaf litter the C-to-N ratio decreased with time, whereas the C-to-N ratio in root litter did not change significantly with time. In leaf litter mesh size did not affect the C-to-N ratio, whereas the C-to-N ratio in root litter was lower in litterbags with 45 µm mesh size. C-to-N ratios negatively correlated with M_loss and the abundance of Collembola and Oribatida, and Oribatida richness in leaf litter. In root litter, C-to-N ratios negatively correlated with M_loss, C_mic , BR and the abundance of Collembola and Oribatida, and Oribatida richness (Table 2).

Microorganisms

Generally, C_mic and BR significantly increased with time of exposure and varied with altitude in both litter types (Table 1, SI 3). In leaf litter, C_mic was higher at 1,000 and 2,000 m compared to 3,000 m, whereas in root litter, both C_mic and BR were higher at 1,000 m compared to 2,000 and 3,000 m.

In leaf litter, variations in C_mic and BR with altitude depended on time, but the effect of altitude also varied with mesh size (significant three factor interaction; Fig. 2, Table 1, SI 3). Similar to leaf litter, in root litter variations in C_mic with altitude depended on time and mesh size while variations in BR in root litter with altitude depended on mesh size (Fig. 2, Table 1, SI 3). In both leaf and root litter, C_mic and BR positively correlated with M_loss (Table 2). In root litter, C_mic and BR also positively correlated with Collembola and Oribatida abundances, and Oribatida richness, but negatively with litter C-to-N ratios.

Abundance of Collembola and Oribatida

Contrasting C_mic and BR, time of exposure as main effect neither affected the abundance of Collembola nor that of Oribatida (Table 3). Rather, the abundance of Collembola and Oribatida in both litter types varied strongly with altitude and mesh size.

Generally, in both litter types, the abundance of Collembola decreased strongly with increasing altitude (Table 3, SI 5). In leaf litter the decrease in Collembola abundance with altitude also varied with sampling date. In root litter, the variation in the abundance of Collembola with altitude varied significantly with mesh size (Table 3, SI 5 and SI 6).

Generally, contrasting the pattern in Collembola, the abundance of Oribatida in leaf litter was similar at 1,000 and 2,000 m and significantly lower at 3,000 m, whereas in root litter the abundance at 1,000 m strongly exceeded that at 2,000 and 3,000 m (Table 3, SI 5). In contrast to Collembola, in both litter types, the abundance of Oribatida generally varied with mesh size. However, in root litter the effect varied with altitude (Table 3, SI 5 and SI 6).

In both leaf and root litter the abundance of Collembola and Oribatida correlated negatively with litter C-to-N ratios and positively with the Oribatid richness (Table 2). However, in root litter the abundance of Collembola and Oribatida also correlated positively with M_loss, C_mic and BR.

Oribatida species richness

In both leaf and root litter the average Oribatida species richness per litterbag was significantly affected by mesh size and altitude (Table 3, SI 5). In leaf litter, Oribatida species richness was generally affected by the interaction between time, altitude and mesh size; whereas in root litter Oribatida species richness only varied with mesh size and altitude (Table 3, SI 5). In both leaf and root litter Oribatida species richness correlated negatively with litter C-to-N ratios and positively with the abundance of Collembola and Oribatida (Table 2). However, in root litter Oribatida richness correlated positively with M_loss, C_mic and BR.

Community structure of Oribatida

In total, 176 species of Oribatida were identified (see SI 7 for full list of species). CCA of Oribatida species in leaf litter with litter C-to-N ratio, M_loss, BR and C_mic included as environmental variables explained 12.7% of the variation of Oribatida community composition (Fig. 3A); litter C-to-N ratio accounted for 4.4% (pseudo-F = 1.9, P = 0.002), M_loss for 4.3% (pseudo-F = 1.8, P = 0.002) and BR for 3.9% (pseudo-F = 1.6, P = 0.007). The community composition of Oribatida at 2,000 and 3,000 m correlated positively with increasing litter C-to-N ratio and BR; M_loss correlated positively with the first sampling date and was associated with lower species abundance.

CCA of Oribatida species in root litter with litter C-to-N ratio, BR, M_loss and C_mic included as environmental variables explained 10.2% of the variation of Oribatida community composition (Fig. 3B); litter C-to-N ratio accounted for 7.0% of the variation (pseudo-F = 3.7, P < 0.001) and BR for 3.2% (pseudo-F = 1.6, P = 0.004). As in leaf litter, Oribatida species at 2,000 and 3,000 m correlated positively with increasing litter C-to-N ratio, but BR correlated positively with the sampling date after 12 months. In both the CCA of leaf and root litter, the 1,000 m site was separated from the 2,000 and 3,000 m sites along the first axis and the centroids of mesh size were close to the center of the ordination. As note of caution, however, most Oribatida species were rare and only occurred at certain altitudes and at specific sampling dates; for the abundance of species at each of the study sites see SI 7.

Variations in leaf and root litter decomposition with altitude

In contrast to our first hypothesis, M_loss of both leaf and root litter showed different patterns of decomposition along the altitudinal gradient. For leaf litter, M_loss did not follow the expected linear decrease with altitude, rather, decomposition rates at 1,000 and 2,000 m were similar after the 12 months of exposure. This contrasts previous studies at our study sites (Illig et al., 2008; Marian et al., 2017; Marian et al., 2019) and indicates that leaf litter decomposition cannot be explained only by the linear decrease in temperature along the altitudinal gradient studied. Potentially, the decline in leaf litter decomposition with temperature was compensated by higher precipitation at 2,000 m compared to 1,000 m (see Methods). High rainfall facilitates decomposition especially at early stages by increasing leaching of soluble compounds (Cusack et al., 2009). However, although climate is considered the primary driver of litter decomposition at large scales (Coûteaux, Bottner & Berg, 1995; Aerts, 1997), the role of climatic factors might be overridden by the variability of litter traits at local scales (Scowcroft, Turner & Vitousek, 2000; Richardson, Richardson & Soto-Adames, 2005; Fujii et al., 2017). In our study, litter characteristics such as C-to-N ratio differed strongly between the leaf litter materials exposed at the three altitudes. However, as indicated by the C-to-N ratio, leaf litter materials from the 1,000 m site were of considerably higher quality than those from the 2,000 m site (as well as the 3,000 m site), suggesting that the high decomposition rates of leaf litter at 2,000 m also cannot be explained by litter quality. Nonetheless, caution is recommended in the interpretation of this result, as other litter chemical compounds were not measured and might have influenced litter decomposition rates. Further, leaf litter M_loss was potentially modified by physicochemical interactions among the three leaf litter species placed in the litterbags and biotic factors such as microbial community composition. The fact that C_mic and BR were higher in leaf litter at 2,000 m than at 1,000 m after 6 months of exposure supports this conclusion and suggests that the leaf litter mixtures favoured the activity and abundance of microbial communities early after exposure. The generally high values of C_mic and BR in leaf litter at 2,000 and 3,000 m, despite the very high litter C-to-N ratio, also suggest that microbial communities at these sites are well adapted to decompose litter of low quality (Gholz et al., 2000; Strickland et al., 2009; Milcu & Manning, 2011; Marian et al., 2017).

In contrast to leaf litter, root litter showed the expected linear decrease in litter decomposition with increasing altitude after 12 months. Less favorable abiotic conditions, such as those at higher altitude, might affect root litter decomposition by reducing the quality of the litter material and thereby nutrient availability as reported for tropical rainforests (Vitousek et al., 1994; Tanner, Vltousek & Cuevas, 1998; Kitayama et al., 2004). This is supported by the high C-to-N ratios and low C_mic and BR, as well as the abundance of decomposer microarthropods at 2,000 and 3,000 m (compared to 1,000 m). The contrasting results between leaf and root litter decomposition rates suggest that in the studied tropical montane rainforests differences in litter chemical composition and their association with nutrient limitations among the altitudinal sites are more critical factors for the decomposition of root litter than for leaf litter. Nonetheless, a complete litter chemical composition profile is needed to support this conclusion. Moreover, leaf litter might be more susceptible than roots to effects of climatic variations as leaf litter is located on top of the soil and thereby exposed to more variable microclimatic conditions than root litter in soil (Ostertag & Hobbie, 1999; Silver & Miya, 2001). However, as root litter generally decomposed slower than leaf litter at 2,000 and 3,000 m, more buffered conditions in soil do not implicate an override of the primacy of nutrient limitations as driving factor of litter decomposition. Nonetheless, at 1,000 m the more buffered climatic condition together with the proximity of the mineral soil layer might have favoured the faster decomposition rates of roots than leaf litter.

Faunal contribution to leaf and root litter decomposition

The abundance of both Collembola and Oribatida was higher in 1 and 4 mm mesh bags irrespective of the plant litter type indicating that, as intended, 45 µm mesh size restricted the access to the litterbags by mesofauna. Thus, the different mesh sizes are a useful tool to evaluate the effects of decomposer animals on decomposition processes. However, restricting the access of mesofauna in 45 µm litterbags was more effective in Oribatida than in Collembola, indicating that the mesh size approach is limited for evaluating the role of mesofauna for decomposition processes and suggesting that it likely underestimates their effects on litter decomposition as discussed earlier (Bradford et al., 2002; Kampichler & Bruckner, 2009). Further, contrasting our second hypothesis, access by microarthropods to litterbags containing leaf litter did not vary the general decomposition rates, nor at any of the three altitudes. However, soil microarthropod access to root litter increased its mass loss at 3,000 m (4 mm mesh bags). Despite the widely assumed beneficial effects of soil microarthropods on litter decomposition, experimental evidence supporting this assumption is controversial; some studies indeed found positive effects on litter mass loss (Bradford et al., 2002; Carrillo et al., 2011; Bokhorst & Wardle, 2013), whereas others suggest their contribution to be minor or absent (Schinner, 1982; Joo, Yim & Nakane, 2003; Kampichler & Bruckner, 2009; Marian et al., 2019). Overall, our results for leaf litter support the latter and previous findings at our study sites also indicate that the decomposition of leaf litter is predominantly due to microorganisms, with the contribution of microarthropods being minor, in particular at early stages of litter decomposition (Illig et al., 2008; Marian et al., 2017; Marian et al., 2019). This finding is supported by lack of correlation of Collembola and Oribatida abundances with M_loss, C_mic and BR in leaf litter. Both decomposer groups may play a more important role at more advanced stages of decomposition when microorganisms have colonized leaf litter making it more palatable for arthropod consumers (Bardgett, 2005; Coulis et al., 2009; Das & Joy, 2009). However, the increase in root litter M_loss at 3,000 m in 4 mm mesh bags suggests that under unfavourable environmental conditions, the decomposition of low-quality litter is stimulated by soil arthropods. Several processes may have accelerated root litter decomposition including stimulation of microbial growth via nutrient mobilization, litter fragmentation and dispersal of microbial propagules (Verhoef & Brussaard, 1990; Ruess & Lussenhop, 2005; Scheu, Ruess & Bonkowski, 2005). The fact that at 3,000 m C_mic and BR in roots increased in 1 and 4 mm mesh bags supports these findings and reinforces that the contribution of microarthropods to decomposition of recalcitrant substrates is more pronounced than in readily decomposable materials (Joo, Yim & Nakane, 2006; Milcu & Manning, 2011; Gergócs & Hufnagel, 2016).

Notably, C_mic and BR varied with mesh size in both leaf and root litter with the effect in root but not in leaf litter varying with altitude. Generally, C_mic was higher in 1 and 4 mm mesh bags, while BR was higher in 45 µm mesh bags. Overall, this supports our expectation that decomposer microarthropods stimulate microorganisms, thereby enhancing their biomass. Potentially, grazing on microorganisms results in increased nutrient mobilization favouring microbial growth (Seastedt, 1984; Hättenschwiler, Tiunov & Scheu, 2005). Further, grazing may foster changes in microbial community composition resulting in more effective use of resources by microorganisms and this may explain the reduced BR in litterbags with coarse mesh size (Dilly & Munch, 1996; Chapman et al., 2013). Indeed, it has been stressed earlier that the structure of microbial communities is an important determinant of litter decomposition rates particularly in forest ecosystems (Strickland et al., 2009).

Additionally, despite the general lack of effect of mesh size on root litter mass loss, the abundances of both Collembola and Oribatida, and Oribatida species richness positively correlated with M_loss, C_mic and BR. These results support our third hypothesis and suggest that litter accumulation in nutrient-poor sites can provide more habitat space for soil decomposer microarthropods. Notably, the increase in animal abundance and diversity may enhance the faunal contribution to litter decomposition (Perez et al., 2013). Moreover, the close vicinity of roots to the mineral soil layer might have favored nutrient availability and thereby improved food resources for decomposer microarthropods (Marian et al., 2019). Further, at our study sites, the role of root exudates and mycorrhizal fungi are increasingly recognized as drivers of litter decomposition, mineralization processes and determinants of soil food webs (Marian et al., 2019; Sánchez-Galindo et al., 2019). Therefore, at 3,000 m, where the concentration of root biomass is at a maximum (Röderstein, Hertel & Leuschner, 2005; Soethe, Lehmann & Engels, 2007), soil arthropods may benefit more from root-derived resources than at 1,000 and 2,000 m either by grazing on microorganisms or directly by feeding on roots.

Interestingly, the abundance of both decomposer microarthropods in leaf and root litter did not vary significantly with sampling date. Potentially, the nutritional value of the litter material for these detritivore microarthropods changed little during the exposure time. Substrate quality may also be related to the generally low effects of soil decomposer microarthropods on leaf and root litter decomposition. Litter of different quality may attract different faunal communities and this likely contributes to the varying effects of decomposer animals on decomposition processes (Fujii & Takeda, 2012; Fujii et al., 2017).

Oribatida species richness and community structure in leaf and root litter

Similar to Oribatida abundance, the higher Oribatida species richness in litterbags of 1 and 4 mm mesh size in both litter types might be attributed to restricted access of microarthropods to the 45 µm mesh litterbags. Oribatida species richness mostly varied with altitude in both leaf and root litter. The significant decrease in Oribatida species richness with increasing altitude in leaf litter supports results of previous studies at our study sites in that species richness of Oribatida in leaf litter is driven predominantly by factors linked to altitude such as temperature, precipitation, moisture and soil pH (Illig et al., 2008; Marian et al., 2018). By contrast, in root litter, the high number of Oribatida species at 1,000 m and the similarly low numbers at 2,000 and 3,000 m suggest that apart from abiotic conditions changing with altitude other factors modify Oribatida species richness. The fact that C-to-N ratios of root litter were similar at 2,000 and 3,000 m support this conclusion and suggests that root litter quality may be a regulator of Oribatida richness; the strong correlation of Oribatida species richness with litter C-to-N ratio supports this conclusion and suggests that increasing amounts of low-quality root litter material with altitude detrimentally affect Oribatida richness at 2,000 and 3,000 m (Röderstein, Hertel & Leuschner, 2005; Maraun et al., 2013). Indeed, Illig et al. (2010) also concluded that Oribatida species diversity at the studied montane rainforests is related to litter quality as important driving factor.

Similar to Oribatida species richness, Oribatida community assemblages varied mostly with altitude in both leaf and root litter. Most of the 176 species identified were registered on the 1,000 m site, only few species were only present at 2,000 and 3,000 m, presumably reflecting less favorable climatic conditions and poor resource quality at the highest altitudes (Marian et al., 2018). Interestingly, in leaf litter certain Oribatida species including Sternoppia mirbilis, Lanceoppia sp.1 and Rostrozetes sp.6 preferentially colonized litter after 12 months of exposure with the latter two species exclusively recorded at 2,000 m. This supports our fourth hypothesis indicating that certain Oribatida species rely on food resources associated with changes in leaf litter physicochemistry during decomposition particularly at higher elevations, suggesting that Oribatida species diversity at least in part is due to resource partitioning (Marian et al., 2018). This is supported by our finding that litter C-to-N ratio works as important factor structuring Oribatida communities in both leaf and root litter. Contrasting leaf litter, increase in root litter decomposition did not affect Oribatida community structure, despite changes in C_mic and BR with time were more pronounced in root than in leaf litter. This suggests that in contrast to our expectations, Oribatida community structure in root litter is not principally linked with decomposition or the gross characteristics of microbial communities. However, potential effects of root litter decomposition on the general quality of the organic litter material may function as key factor structuring Oribatida communities as indicated by their relationship with C-to-N ratios. Additionally, as mentioned before, root exudates may also modify the general characteristics of the organic matter and become an important carbon source fuelling Oribatida communities (Pollierer et al., 2007; Dennis, Miller & Hirsch, 2010; Zieger et al., 2017). Overall, our results indicate that the environmental factors considered explained only little of the variation in Oribatida community composition in both leaf and root litter. Still, both leaf and root litter characteristics, such as C-to-N ratio, as well as C_mic , significantly correlated with Oribatida community composition and this is consistent with earlier findings (Fujii & Takeda, 2017; Marian et al., 2018).