Sensitivity analysis
A strong ecotone effect could be found for assemblages from ecoregions having many small-ranged species, as these species are more likely to have their distribution centered in the ecoregion core, as well high abundance and occurrence probability at ecoregions’ core (Andrewartha and Birch 1954; Brown 1984). Sigmodontine rodents have, in general, small range sizes (min=0.02 square degrees, 1st quartile= 4.16, median=18.30, mean=55.46, 3rd Quartile= 51.40, max= 797.37 square degrees, measured across the 350 species included in our dataset), and many species have their range totally included within the area of a unique ecoregion (Fig. S5). Furthermore, the number of species having a range smaller than ecoregion area varies geographically. While most small-ranged species occur in the Andes, we observed a considerable number of such species in Atlantic Rainforest, Cerrado, Chaco, and southwest Amazonian regions (Figs. S5-S7).
We evaluated whether results would change when analyzing aTR, aST, and aLT of assemblages of small-ranged species. To avoid area effects when classifying small-ranged species as those having their range smaller than ecoregion area (Fig. S5), we considered as small-ranged species those having a range size smaller than 4.16 square degrees, the 1st quartile of range-size values presented above. These models included values of aTR, aST, and aLT across 88 small-ranged sigmodontine species, distributed in 58 ecotone points of 14 different ecoregions, and in 81 core points of 31 different ecoregions (Fig. S6).