**3. Results**

Altogether, we recorded 57 species and 78,131 individuals across the three locations (Cserépfalu: 48 species, 32,490 individuals, Hollók˝o: 46 species, 23,611 individuals, and Balatonakali: 44 species, 22,030 individuals). Three species of the family Geotrupidae, 33 species of the subfamily Aphodiinae, and 21 species of the subfamily Scarabaeinae were recorded (Appendix A Table A1).

The most numerous was the Scarabaeinae subfamily (43,209 individuals, 55.3%), then the Aphodiinae subfamily (27,314 individuals, 35.0%), and finally the Geotrupidae family (7608 individuals, 9.7%). The most dominant species was *Nimbus obliteratus*, with 16,540 individuals collected, which was 21% of all individuals. In total, 18 species accounted for nearly 95% of the total abundance, while 39 species accounted for less than 1% (Appendix A Table A1).

Our models did not reveal a significant effect of locality on species richness (*p* = 0.116), abundance (*p* = 0.383), or diversity (*p* = 0.145), while habitat had an effect on diversity (*p* = 0.001), but no effect on species number (*p* = 0.070) nor abundance (*p* = 0.302). Both dung type (*p*(sp) = 0.024, *p*(abu) = 0.001) and season (*p*(sp) < 0.001, *p*(abu) = 0.022, *p*(div) < 0.001) had a significant effect. However, we did not find an effect of dung type on diversity (*p* = 0.473).

Considering habitat types, 32,917 individuals of 41 species were recorded from forests, 20,724 individuals of 47 species from wood pastures, and 24,490 individuals of 53 species from grasslands. Species richness and Shannon diversity were lower in forests than in pastures. In the case of wood pastures, these values were between the two distinct habitats. There was no difference in abundances between the three habitats (Figure 1).

Pairwise comparisons of species richness, abundances, and Shannon diversities between habitat types revealed several significant differences between forests and pastures, while wood pastures exhibited an intermediate position (e.g., species richness for sheep dung during spring and summer), except for one case (spring abundance on cattle dung) where wood pasture dung beetle number was lower than in both habitats (Figure 2, Appendix A Table A2).

**Figure 1.** Summarised dung beetle species richness, abundance, and Shannon diversity per habitat types. Box plots show medians (thick line), lower, and upper quartiles (boxes), whiskers include the range of distribution without outliers. Different letters above the boxes show the significant differences between habitats.

**Figure 2.** Habitat effect on dung beetle species richness, abundance, and diversity separately for dung type (columns) and seasons (rows). Box plots show medians (thick line), lower, and upper quartiles (boxes), whiskers include the range of distribution without outliers. Different letters above the boxes show the significant differences between habitats within a given season and dung type, applying generalised linear mixed models (GLMMs), linear mixed models (LMMs), and their pairwise comparisons. F.: forest, Wp.: wood pasture, and P.: pasture.

For composition of dung beetle assemblages, the seasonal distinction was the most noticeable e ffect according to the NMDS analysis (Figure 3A, PERMANOVA test: *R*<sup>2</sup> = 0.237, *p* = 0.001). There was no overlap between forest and grassland dung beetle samples, but wood pastures had an intermediate position between the two other habitat types (Figure 3B, PERMANOVA test: *R*<sup>2</sup> = 0.109, *p* = 0.001). Furthermore, assemblage composition showed significant di fferences among the three localities in Hungary (Figure 3C, PERMANOVA test: *R*<sup>2</sup> = 0.081, *p* = 0.001), but the e ffect of locality was smaller compared to the e ffect of season and habitat. Dung type had no significant e ffect on the composition of dung beetle assemblages (Figure 3D, PERMANOVA test: *R*<sup>2</sup> = 0.017, *p* = 0.322).

The analysis of indicator species regarding habitat, dung type, and season showed that 12 species were linked to pastures, 6 to forests, and only one to wood pastures (Table 1). Only cattle feces had three indicator species, while the other two dung types had none. Spring had 20 indicator species, while summer and autumn had less than half of that (Table 1).


**Table 1.** Dung beetle species with significant indicator values for habitat, dung type, and season.


**Table 1.** *Cont.*

**Figure 3.** Assembly composition of dung beetles according to (**A**) season, (**B**) habitat, (**C**) locality, and (**D**) dung type, applying nonmetric multidimensional scaling (NMDS).

## **4. Discussion**

Our main question concerned whether wood pastures have distinct dung beetle assemblages compared to forests and pastures. Using various types of analyses to process data on around 80,000 individuals, we found that assemblages in wood pastures showed an intermediate or transient position between the assemblages of forests and pastures, which were distinct. In Central Europe, the majority of dung beetle species inhabit open or semiopen grasslands, and there are fewer woodland species [50] (Scarabaeinae: mostly grassland species, Geotrupidae: mostly forest species, Aphodiinae: both forest and grassland species [15]). Our study showed similar patterns. Dung beetle assemblages on wood pastures showed an intermediate position between forest and pasture assemblages in composition, species richness, and Shannon diversity. In general, this position was present in all but one comparison when broken down according to seasons and dung types. This transient position is also supported by the distribution of the indicator species, wherein we found only one dung beetle (*Onthophagus joannae*) species specific for the wood pastures from the total of 57 dung beetle species and total of 19 habitat indicator species. This pattern

in assemblage characteristics probably had an e ffect on the decomposition function of dung beetles, as Buse and Entling [28] showed.

The role of habitat in structuring dung beetle assemblages is well-known. Several studies found di fferences in various parameters between forest and pasture assemblages. Ríaz-Díaz et al. [51] found significant di fferences in species composition between forest and grassland habitats. Jugovic et al. [52] found significant di fferences in species composition, species richness, and abundance among grazed and overgrown sites. Numa et al. [53] found significant di fferences for species composition, species dominance, and abundance between sites with wild versus livestock herbivory. All of these studies showed a strong effect of habitat on the structure of dung beetle assemblages. In our case, we also found significant di fferences between forests and pastures in species richness and Shannon diversity, but not for wood pastures and forests or wood pastures and pastures. This pattern is remarkable, as based on the literature. We may expect three di fferent assemblages in the three di fferent habitats. Similar studies, where forest, pasture, and a transitional habitat (usually shrubland) were involved, found significant di fferences between habitats, indicating that neither was in a transitional position [29,54]. It seems that this is the general pattern, as this was found for other taxa as well. For example, spider [37] and ground-dwelling beetle assemblages [55] of open pasture habitats and scattered trees (and other habitats) were statistically distinct. Lövei et al. [56] categorised carabids to the forest edge and matrix species as these habitats had distinct assemblages. The transitional position of wood pastures between forests and pastures is interesting, if compared to forest, forest edge, and grassland patterns. Such studies showed that all three habitats had di fferent assemblages for carabids [57], ants and spiders [58], and millipedes [59]. Therefore, the habitat effect in the forest, wood pasture, and pasture system seems to be rather di fferent from forest, edge, and pasture (or grassland) systems, highlighting the uniqueness of the former system.

Our second question was on the potential e ffects of locality, dung type, and season on assemblages. We found that dung beetle assemblages are similar across three distant locations in Hungary for all studied measures, namely for species number, abundance, Shannon diversity, and composition. The similarity of the three localities is probably linked to the overwhelming role of locally relevant factors in shaping dung beetle assemblages (and decomposition in general); thus, di fferences between habitat or managemen<sup>t</sup> types are expected to be larger than di fferences between locations [60,61]. This similarity of dung beetle assemblages across locations indicates that our results are relevant to other locations (at least in the Pannonian region), which supports the wider applicability of our findings. The habitat e ffect and lack of locality e ffect on dung beetle assemblages are probably linked to vegetation characteristics. Several studies showed that vegetation structure (e.g., plant diversity [62]), fine-scale heterogeneity in grazing intensity [63], or vegetation cover [64] highly influenced dung beetle assemblages. Our results on habitat and locality can be explained on the assumption that di fferences between forests or pastures or wood pastures are smaller across locations than among the three habitat types within a given location.

The majority of Palearctic dung beetle species are considered to be generalists, feeding on various types of dung available. This suggests that a dung type e ffect on assemblage structure is likely weak. Indeed, we could not clearly identify distinct dung beetle assemblages across the three dung types, which is supported also by a low number of indicator species associated with each dung type. However, we found a significant e ffect of dung type on abundance and species number, indicating that even generalist species can show di fferent preferences if they have a choice of several dung types [65].

Season seems to be a profound factor in shaping dung beetle assemblages. We already know from several studies that dung beetle species appear in a determined phenological order during the year [52,66,67]. We found a clear distinction in assemblage composition in spring, summer, and autumn for all assemblage measures. Indicator species are also in line with former results [68]. In autumn, the wood pasture samples were more distinct from forests and pastures, while both the spring and the summer

samples showed a greater similarity among dung beetle assemblages in the three habitats, suggesting that the seasonality influenced the effect of habitats on dung beetles.

Dung beetle assemblage composition was different in this study for habitats, seasons, and dung types. It seems that composition is a sensitive measure of differences in dung beetle assemblages across a range of conditions [51,69].

The third aim of this study was to provide information for nature conservation management. Dung beetle assemblages of Hungary and the Pannonian biogeographic region are still understudied. During this one-year sampling in three locations, nearly half of the species of Hungarian dung beetle fauna were recorded; this seems to be outstanding for the results of the mosaic-like structural complexity of the studied habitats. The conservation of dung beetles should focus on maintaining compositional diversity, as different species have different roles, and the loss of a species cannot be compensated by the presence of others [70]. Species-specific thermal tolerance is very important [71], as is poor flight capacity limiting dispersal that potentially leads to local extinctions if habitats are further degraded (e.g., Storck-Tonon et al. [72]). The provisioning of fresh dung throughout the year—irrespective of what livestock species produces it—is of major importance for all habitats. The manure of wild ungulates is not sufficient to ensure the conservation of the regional dung beetle species pool [73]. From a practical point of view, this means that if we want to maintain species-rich communities and the associated ecosystem services, we can do so by increasing the availability of the right amount and quality of manure throughout the year. Extensive animal husbandry (i.e., grazing livestock from spring to late autumn) is one way to enhance the habitat for dung beetles. It is imperative for the conservation of diverse assemblages that grazing take place in the long term. Grazing continuity has a positive effect on species numbers both for generalists and specialists [74]. The use of veterinary medical products needs to be reconsidered to avoid harm to dung beetles [14]. Long-term conservation needs to consider climate change effects and connectivity of suitable dung beetle habitats to provide dispersal possibilities. Active conservation planning is needed, as only sufficiently stable and abundant populations as well as diverse assemblages of dung beetles are able to perform their ecological functions that benefit people [15,75].
