Dung Beetle Assemblages Attracted to Cow and Horse Dung: The Importance of Mouthpart Traits, Body Size, and Nesting Behavior in the Community Assembly Process

Dung beetles use excrement for feeding and reproductive purposes. Although they use a range of dung types, there have been several reports of dung beetles showing a preference for certain feces. However, exactly what determines dung preference in dung beetles remains controversial. In the present study, we investigated differences in dung beetle communities attracted to horse or cow dung from a functional diversity standpoint. Specifically, by examining 18 functional traits, we sought to understand if the dung beetle assembly process is mediated by particular traits in different dung types. Species specific dung preferences were recorded for eight species, two of which prefer horse dung and six of which prefer cow dung. Significant differences were found between the functional traits of the mouthparts of the dung beetles attracted to horse dung and those that were attracted to cow dung. Specifically, zygum development and the percentage of the molar area and the conjunctive area differed between horse and cow dung colonizing beetles. We propose that the quantitative differences in the mouthpart traits of the species attracted to horse and cow dung respectively could be related to the differential capacity of the beetles to filtrate and concentrate small particles from the dung. Hence, the dung preference of dung beetles could be related to their ability to exploit a specific dung type, which varies according to their mouthpart traits. Moreover, we found that larger and nester beetles preferred cow dung, whereas smaller and non-nester beetles preferred horse dung. This finding could be related to the tradeoff between fitness and parental investments, and to the suitability of the trophic resource according to the season and species phenology.


Introduction
Dung beetles (Scarabaeidae: Scarabaeinae, Aphodiinae; Geotrupidae) are distributed in all biogeographic regions and include more than 8000 species [1]. They are mainly coprophagous, a feeding habit which arose from saprophagy in the Cretaceous, leading to a species radiation during mammal diversification in the Cenozoic [2,3]. After the extinction of megafauna, in some places (e.g., Italy, Spain), the feces of domestic livestock became the beetles' main trophic resource in human-dominated landscapes [4]. In other places (e.g., tropical and subtropical habitats), however, dung beetles became more generalist species associated with the greater diversity and abundance of other types of resource, such as omnivorous dung, vertebrate and invertebrate carrion, fungi, and fruits [5,6].

Study Area and Sampling Design to Estimate Trophic Preferences
The study area is located in the sub-mountain pastures of the central Apennine Range (Pesaro-Urbino province, Italy) with an altitudinal range between 500-900 m a.s.l. The climate of the area is temperate. Within the province, we selected five pastures with an extension of between about 20 and 60 hectares (Isola del Piano, Catria, Nerone, Calamello, Pietralata) (see Tonelli et al., [38,51] for further details on the study area and pastures), and within each pasture we selected three sampling sites, separated by at least 500 m to guarantee spatial independence [52,53]. In each sampling site, we placed square with four standardized pitfall traps (CSR model in Lobo et al., [54]) positioned at the corners, separated by about 50 m (four traps × three sampling sites × five pastures = a total of 60 traps). Two of the traps were baited with cow dung (500 cm 3 ), while the other two were baited with horse dung (500 cm 3 ). The dung was frozen to kill any dung beetles, predators or Diptera larvae that may have altered the results. The data of the two pitfall traps baited with the same type of dung at each sampling site were coupled; hence, there were a total of 30 sampling points (two dung type × three sampling sites × five pastures). Traps were exposed for 48 h every 15 days from June to November 2013 and in May and June 2014. The data of each sampling day were added together to obtain the cumulative abundance and richness of the assemblages. The dung that was used to bait the traps was ivermectin free because this veterinary medical product could alter dung attraction [55].

Functional Traits
In total we measured 18 functional traits: 11 morphological (seven associated with the body and four with the mouthpart), and seven behavioral. All traits are associated directly or indirectly with ecological functions. We present a summary table (Table 1) with the potential biological and functional importance of each trait. For further details on the measurement protocols of each trait see Supplementary Materials Files S1 and S2, and Tonelli et al. [56].   [57]. This method measures species specificity and fidelity to an ecological status (in our case, fidelity to a dung type). Species with a significant IndVal (p < 0.05) were considered to prefer that type of excrement. The analysis was performed using PC-ORD 5 [58].

Functional Diversity
Taking into consideration all measured functional traits and all species, we investigated the effect of dung type on functional diversity using single trait indices: CWM, FDvar, and FRO. CWM (community weighted mean) represents the expected functional value of one trait in a random community sample. It is the community mean of a trait weighted by the relative abundance of the taxa [59]. FDvar (functional divergence) can be defined as the variance of a trait in a community, whose squared residuals are weighted by the abundance of the species [60]. It is constrained between 0-1; high values of functional divergence are related to a high degree of niche differentiation, whereas low values indicate a low level of niche differentiation [61]. FRO (functional regularity) is the regularity or evenness of the trait values and abundance in the observed range [62]. It is constrained between 0-1, showing higher values when each species has the same distance from its neighbors and each species is present in the same abundance [62].
The functional diversity indices were calculated using FDiversity software [63]. Dummy variables were created for each category of categorical traits. We used linear mixed models (LMM) to evaluate dung type effect (cow vs. horse) on functional diversity indices. Sampling sites were added as random factors in the LMM procedure in order to avoid any potential effect related to differences in the pastures where the data were collected [64]. When the null hypothesis was rejected, we used Fisher LSD for mean comparison (p < 0.05). We evaluated the assumptions of normality and homogeneity of variances using graphical analyses of residuals. Statistical analyses were performed with InfoStat version 2020 [65].

Functional Diversity
The CWMs showed differences in 10 functional traits (p < 0.05): four related to body morphology (fresh body mass, hind tibiae length, abdomen length, and wing load), one related to mouthpart morphology (zygum development) and five to ethological traits (nesting pattern, nest type, nest depth, horizontal nest distance, and phenology). Specifically, fresh body mass, hind tibiae length, abdomen length, and wing load showed higher values in cow dung. Regarding mouthpart morphology, we found that the CWM of zygum development was higher in horse dung. As regards ethological traits, we found that the CWM of nesters (telecoprid and paracoprid with a horizontal and vertical relocation of trophic source) was higher in cow dung, whereas non-nesters that do not relocate dung showed a higher CWM in horse dung. Finally, phenology varied significantly between dung type, with a higher CWM in species that are active all year or in summer-autumn seasons in the cow dung, while the CWM of species actives in autumn, winter and spring was higher in the horse dung.
The FRO of the percentage of molar filtering area was higher in horse dung (p < 0.05). The FDvar of the percentage of the conjunctive area and the phenology of species actives throughout the year was higher in cow dung (p < 0.05) ( Table 4).

Discussion
In the present study we investigated the attractiveness of horse and cow dung to dung beetles using a comprehensive range of functional diversity indices and functional traits. Our investigation showed the presence of dung beetle species with a trophic preference for horse or cow dung. We also found quantitative differences in some mouthpart traits (zygum, conjunctive and molar filtering area) between dung beetles attracted to horse dung and those attracted to cow dung. Furthermore, our results suggest that larger and nester beetles prefer cow dung, whereas smaller and non-nester beetles prefer horse dung. Finally, it was found that phenology could be important to dung beetle trophic selection.

Trophic Preferences
Our results regarding trophic preferences are consistent with those of Barbero et al. [25] and Dormont et al. [18,29,66]. Previous studies conducted in Europe have found all the species examined in this study in a great variety of dung types [67,68]. Nevertheless, comparing our results with these studies, which explicitly evaluated trophic preference, we found that Aphodius fimetarius (Linnaeus, 1758) and Colobopterus erraticus (Linnaeus, 1758) were generally more attracted to cow dung [18,66]. For the other species (Aphodius coniugatus (Panzer, 1795), Bodilopsis rufa (Moll, 1782), Esymus pusillus (Herbst, 1789), Geotrupes spiniger Marsham, 1802, Chilothorax conspurcatus (Linnaeus, 1758) and Labarrus lividus (Olivier, 1789)), a trophic preference for cow or horse dung was recorded for the first time. Hence, if this feeding predilection is stable through the species range, it should be investigated in other regions [25]. We also found some species exclusively in horse or cow dung, but their low abundance does not allow us to draw firm conclusions ( Table 2).

Mouthparts
Significant differences in the mouthpart traits of the dung beetles attracted to cow dung and those attracted to horse dung were found ( Figure 1). Specifically, the zygum, conjunctive and molar area development seems to play an important role during the community assembly process in these two types of dung. The zygum is the proximal part of the epipharynx covered by setae. Its development varies according to the species [69,70]. The conjunctive is a structure whose surface consists of lamellae and it is inserted between the distal and molar lobe of the mandibles [71]. The molar surface consists of tightly packed transverse ridges with numerous pores that connect the molar surface with a system of deeper channels [71,72]. The specific function of these traits is not yet totally clear [71], but their variability seems to be correlated with the type of dung that the beetle feeds on. Indeed, hard and soft feeders show different mouthpart features [41][42][43][44]69,71]. For example, the zygum is more developed in hard dry pellet feeders [69], whereas the development of the conjunctive area is greatly reduced. On the other hand, hard feeders show extensive development of the anterior part of the molar area [71]. In any case, all of these studies draw a distinction between hard feeders and soft feeders based on a qualitative descriptive evaluation of the morphology of their mouthparts.
The quantitative differences between the mouthpart traits of the species attracted to horse and those attracted to cow dung could thus be related to their differential capacity to filtrate and concentrate small particles from the dung. The greater development of the zygum in dung beetles attracted to horse dung and the reduced variation found in the conjunctive and zygum of these beetles could be associated with the higher overall fiber content and larger fibers found in horse excrement. Horse dung could thus influence the dung beetle community species selection based on these mouthpart features, which may be related to the elimination of coarse fibers. We also found particular mouthpart features associated with cow dung, namely the ratio between the filtering area and the total area of the molar lobes, which showed a diminished level of functional regularity. Hence, we propose that this trait could be linked to the process of concentrating the small particles in cow dung, which has a higher moisture content.
Horse and cow dung differ in terms of their physical and chemical features [27,29]. One of the most important differences lies in their relative fiber content. Horse dung contains more overall fiber and the fibers themselves tend to be of a larger size [27]. Moreover, horse dung has a lower moisture content [24]. Holter [73] proposed that dung beetles collect dung with their maxillary palps; the large particles are subsequently brushed out by filtration setae on the mouthparts, and the remaining paste is then squeezed by the molar lobes while superfluous liquid is led away from the pharynx through the filtration channels. This process concentrates the remaining small particles, which are then ingested [74]. Hence, there is no evidence that dung beetle mouthparts are involved in the trituration of large particles, and their role seems limited to the elimination of coarse fibers and the Life 2021, 11, 873 9 of 13 concentration of small particles from the aqueous medium. To our knowledge, our study is the first attempt to study the influence of mouthpart traits from a quantitative standpoint.
Life 2021, 11, x FOR PEER REVIEW 9 of 14 lobes while superfluous liquid is led away from the pharynx through the filtration channels. This process concentrates the remaining small particles, which are then ingested [74]. Hence, there is no evidence that dung beetle mouthparts are involved in the trituration of large particles, and their role seems limited to the elimination of coarse fibers and the concentration of small particles from the aqueous medium. To our knowledge, our study is the first attempt to study the influence of mouthpart traits from a quantitative standpoint.

Body Size and Nesting Behavior
Our results suggest that larger and nester beetles prefer cow dung, whereas smaller and non-nester beetles prefer horse dung. Body size is one of the most important dung beetle traits which is very sensitive to changes in trophic availability and quality [38]. Food quality determines adult dung beetle body size and influences several reproductive outputs [34,35,75]. Larger dung beetles need more dung for nesting, but this necessity is plastic and the amount of dung that is needed may vary according to dung quality and dung type [35,76]. Our results suggest that cow dung could be more suitable for larger dung beetles and can probably maximize their fitness. This result, however, is not in agreement with those reported by other authors who found an optimal investment strategy in dung beetles colonizing horse dung. For example, Hunt & Simmons [34] found that females of the Onthophagus taurus need about 20% less dung when provisioning with horse dung, and hence received greater fitness returns per unit of investment and experienced lower provisioning costs in terms of the minimum amount of dung required to produce a surviving offspring than females provisioning with cow dung. Likewise, Moczek [75] observed that almost twice the amount of cow dung was required to yield the body size obtained with half the amount of horse dung. These investigations attribute their findings to the higher quality value of horse dung. However, dung quality can be related to the

Body Size and Nesting Behavior
Our results suggest that larger and nester beetles prefer cow dung, whereas smaller and non-nester beetles prefer horse dung. Body size is one of the most important dung beetle traits which is very sensitive to changes in trophic availability and quality [38]. Food quality determines adult dung beetle body size and influences several reproductive outputs [34,35,75]. Larger dung beetles need more dung for nesting, but this necessity is plastic and the amount of dung that is needed may vary according to dung quality and dung type [35,76]. Our results suggest that cow dung could be more suitable for larger dung beetles and can probably maximize their fitness. This result, however, is not in agreement with those reported by other authors who found an optimal investment strategy in dung beetles colonizing horse dung. For example, Hunt & Simmons [34] found that females of the Onthophagus taurus need about 20% less dung when provisioning with horse dung, and hence received greater fitness returns per unit of investment and experienced lower provisioning costs in terms of the minimum amount of dung required to produce a surviving offspring than females provisioning with cow dung. Likewise, Moczek [75] observed that almost twice the amount of cow dung was required to yield the body size obtained with half the amount of horse dung. These investigations attribute their findings to the higher quality value of horse dung. However, dung quality can be related to the diet of the animals, showing spatiotemporal variation with seasonal consequences for the trophic preferences and reproductive outputs of dung beetles [77], which could account for the difference in our results.
Another interesting result regards dung beetle nesting behavior. We found that nester species, such as paracoprids and telecoprids, which show relocation behavior [78], prefer cow dung. On the contrary, non-nester species, which lay eggs directly in the dung pat where the entire development takes place, clearly prefer horse dung [79]. This interesting finding is counterintuitive because, generally, non-nester species need to lay their eggs in dung pads characterized by a high water content to avoid desiccation, which could hinder larval development. On the other hand, paracoprid and telecoprid dung beetles bury dung under the soil, where water loss is prevented and coarser and drier dung may not affect larval development [80]. However, our results might also be explained by the phenology of the species that prefer horse or cow dung. Indeed, cow dung was clearly preferred by dung beetles that are active throughout the year or in the summer-autumn period, whereas horse dung was selected by species with a phenological peak during autumn, winter and spring. In the region under study, an average annual precipitation of 1000 mm was recorded with maximum levels in spring and autumn. Hence, non-nester species, such as the Melinopterus and Nimbus species, can exploit horse dung regardless of its lower moisture content because they are at their phenological peak during wet seasons.
Finally, as a caveat, we must recognize that the collection method can influence the results in trophic preference experiments [81]. Indeed, even though pitfall traps make it possible to conduct large-scale controlled experiments, dung beetles cannot escape from such traps, and this method could reflect the pattern of initial colonization by adult beetles (i.e., findability). On the other hand, collecting with dung pats reflects both findability and preference [18].

Conclusions
We confirmed the presence of species specific dung beetle trophic preferences for horse or cow dung. Moreover, we found differences in the mouthpart traits of dung beetles between these two excrements. The development of the zygum, as well as the percentage of molar and conjunctive areas, appear to play an important role in the community assembly process and dung choice. This could be related to the difference in the moisture and fiber contents of horse and cow excrement, which selected species based on their ability to filtrate and concentrate small particles. Although dung beetles use volatile organic compounds to discriminate between dung types, their choice could also be related to their capacity to make use of a particular excrement type, which may vary according to their mouthpart morphology. Moreover, larger and nester beetles showed a preference for cow dung, whereas smaller and non-nester beetles preferred horse dung. Larger dung beetles probably maximize their fitness by using cow dung. Generally, non-nester species need a more humid and stable resource (such as ruminant dung), but the presence of these species during the wet season could explain their choice of horse dung. Further studies using more dung types are needed to better understand the role of functional traits in dung beetle trophic preferences and consequent community assembly processes.