Soil factors have important implications for the establishment of species that dwell on and underneath them. Dung beetles’ biological cycle is highly dependent on soil, and their nesting and feeding activities positively affect soil chemical and physical properties. We assessed the effect of the type of soil texture (loamy sand and sandy loam) and its physical properties (the amount of sand, silt, and clay), compaction, and soil moisture on dung beetle species assemblages (species richness, abundance, functional groups, and species composition) in a pastureland region of Brazilian Cerrado. We developed one dataset containing the physical soil properties (clay, silt, and sand content, soil texture classification, soil compaction, and soil moisture) and assemblage metrics of dung beetles (total richness and abundance, and abundance and richness of paracoprids, telecoprids, and endocoprids) sampled in 15 transects in introduced pastures in Aquidauana, Mato Grosso do Sul, Brazil. We report that species richness was not affected by the type of soil texture and its properties; however, there was a consistent shift in species composition between loamy sand and sandy loam soils. We also found a reduction of the total and paracoprid beetle abundances in loamy sand soil. Furthermore, soil compaction was the main driver of dung beetle species composition in pasturelands, where the increase of soil compaction negatively affected the entire dung beetle assemblage and paracoprids’ abundances.

We carried out this study in four pastureland sites that were separated by > 1 km among each one. In each pastureland site, four linear transects of 100 m were regularly distributed 100 m apart from one each other (except on one site, which had 3 transects, because the traps were removed by animals).

In each transect (i.e., replicate), we collected soil samples (~ 0.5 kg; 0–20 cm) at the three sampling points (e.g., sub-samples). The sub-samples were thoroughly mixed to obtain a representative sample to measure texture (sand, clay, and silt content), and soil moisture, and analysis of this sample provided average values for each transect. When merging the three subsets of the transects in one sample, we presented a replicate that joins the characteristics and variations observed in this space, which are independent from the nearest one. Soil compaction data were obtained using the impact penetrometer.

The texture of the soils was classified in loamy sand and sandy loam soils according to traditional particle size criteria using the USDA texture triangle, a widespread and one of the best-known soil classification systems. According to this approach, the transect soils were classified into two types of soil texture; loamy sand (Sand: 70-86%; Silt: 0-30%; Clay: 0-15%), and sandy loam (Sand: 50-70%; Silt: 0-50%; Clay: 0-20%). From the total number of transects used in this study, eight of them comprised loamy sand soil and seven of them comprised sandy loam soil.

We conducted dung beetle sampling during the rainy season (October 2010) since it exhibits the highest dung beetle activity and species richness in introduced Brazilian pastures. In each transect, we designated three sampling points, spaced 50 m apart from each other. At each sampling point, we set up four pitfall traps, spaced 10 m apart from each other in a square baited with approximately 40 g of carrion, cattle dung, human feces, and fresh grain pig dung. Traps consisting of plastic containers (15 cm in diameter, 9 cm in depth) installed at ground level. We partially filled each container with 250 ml of saline solution + neutral detergent (1.5%). We placed a plastic lid (15 cm in diameter) supported by three wooden sticks (25 cm) above each trap, suspending them 20 cm above the ground to reduce bait desiccation and prevent potential rain damage. We positioned the baits in 50 ml plastic containers at the center of each trap using a wire as a bait holder. Traps remained active for 48 hours at each transect, a time period that is widely used in ecological studies of dung beetle assemblages in tropical landscapes of America. Throughout the study, we deployed a total of 180 traps (12 traps per transect × 15 transects).