Interactions between land use, taxonomic group and aspects and levels of diversity in a Brazilian savanna: implications for the use of bioindicators

The study was carried out in the Triângulo Mineiro region of Minas Gerais state, covering the municipalities of Uberlândia, Monte Alegre, and Nova Ponte, in south-eastern Brazil. We conducted the study in five habitat types, comprising two natural habitats (savanna and semideciduous forest), and three anthropogenic land-uses: cattle pastures (planted with introduced Urochloa grasses), soy fields (where sampling took place when plants were at the vegetative phase) and  plantations of Eucalyptus trees (≥ 6 yrs old). Ants and beetles were sampled at the same 40 sites (8 replicates per land use), and birds at 30 sites (6 replicates per land use), only some of which were the same as for ants and beetles. 

Ants that forage on ground and dung beetles were sampled using pitfall traps. Sampling took place in November and December (early wet season) 2017. In each site, eight traps were installed with traps located at the corners of a 100×100 m square, and at the mid-points of the sides of the square, keeping a minimum distance of 50 m between any two traps. All traps were at least 75 m distant from the edge of the respective land use. Traps were plastic containers (19 cm diam, 11 cm height) filled with 150 ml of a saline solution and detergent. Each trap had a wire hoop suspended over it to accommodate a small (4 cm diam, 4 cm height) plastic container for holding a dung bait. We used a 20 cm diameter plastic cover supported by three sticks to protect traps from rain. Traps were baited with ~40 g of a mixture of pig dung and human faeces (4:1 proportion) and left in the field for 48-hrs.

Birds were surveyed using 20-min point counts in the rainy season (November 2017 to March 2018). At each site, five sampling points were established, 200 m distant from each other. All surveys started at sunrise (about 6 a.m.), and all species seen or heard from each point were recorded. Each sampling site was re-surveyed in the following dry season (April to October to 2018); however, for logistic reasons we were unable to re-survey the plantation sites. 

Ant and dung beetle species were identified to species or morphospecies by comparison with named species in the Zoological Collection at the Federal University of Uberlândia (UFU) or with specialist assistance from Fernando Vaz de Mello, respectively. Vouchers of all species have been deposited at UFU´s Zoological Collection. Birds were identified directly in the field and species names follow the checklist produced by the Brazilian Ornithological Records Committee.

We classified species functionally based on primary diet, foraging location and/or behaviour, and body size, as these traits are known to be sensitive to habitat modifications and of importance for the ecosystem services provided by ants, birds, and dung beetles.

Ant species were classified according to their diet as predators, fungivores, nectarivores or omnivores, and according to their main foraging location as arboreal, epigeal (aboveground) or hypogeal (in soil and litter), based on information provided by Brown (2000) and Silvestre et al. (2003). Species were further classified into four body size categories based on our measurements of body length (Weber´s length; Brown, 1953) of 1-5 ant workers per species: 1 (< 0.75 mm), 2 (0.75-1.74 mm), 3 (1.75-3 mm), and 4 (> 3 mm).

Dung beetles were classified as coprophagous, necrophagous, frugivore, generalist or predator, according to the type of food resource each species is most often attracted to. This classification was based on over 30 years of field experience throughout Brazil by one of the authors of this study (FVM), who used multiple types of baits (e.g., carcasses, fruits, faeces) to attract and collect dung beetles, and/or on literature information. Although information about the “attractiveness” of different types of baits to dung beetles (used here as a proxy for primary diet) was not obtained directly in the sites of the present study, it is importat to note that we are not aware of any evidence of geographic or habitat variation in bait preference among tropical species of dung beetles. Dung beetles were also classified according to their foraging behaviour as: telecoprid (species that make a dung ball and roll it away for burial), paracoprid (species that store dung in tunnels dug immediately below the dung source), or endocoprid (species living within or immediately below the dung, without moving it). For this, we used the database of the Zoological Collection of the Federal University of Mato Grosso (UFMT). Whenever sample sizes allowed, 30 individuals from each species were weighed for determination of body mass (following Almeida et al., 2011), and species were classified according to the following ordinal scale: 1 (< 10 mg); 2 (10-99 mg); 3 (100-300 mg); and 4 (>300 mg).

Each bird species was classified according to its primary diet as frugivores granivore, insectivore, nectarivore, carnivore, detritivore, or omnivore, and according to the main foraging location as ground, understory/shrubby vegetation, or tree canopy, based on the Wilman et al. (2014) database and our own field experience. Using these same sources, we obtained information on mean body weights of each species and assigned them to one of five size categories: 1- (<15 g); 2 (15-39 g); 3 (40-199 g); 4 (200-599 g); and 5 (> 600 g).