Ants are an incredibly diverse and ubiquitous group of invertebrates in most terrestrial ecosystems. Although extensively sampled, the majority of ant inventories do not evaluate the effect of different sampling techniques in capturing non-traditional metrics of diversity. We aimed to quantify TD (taxonomic) and FD (functional) diversities for a local ant assemblage by integrating metrics and evaluating complementarity of pitfall traps and Winkler extractors for the leaf-litter vs. epigaeic ant faunas and to determine the effect of sampling techniques on functional composition (community-weighted means of 11 morphological traits, see Table S1 in the Excel file and Trait_base_data.csv in this dataset) and functional diversity (multi-trait morphospace measured with three different metrics, see Table S2 in Excel file). We sampled the local community (Table S3 and Table S4 in the Excel file) in an Atlantic Forest fragment using week-long pitfall traps and 1m² leaf litter samples submitted to Winkler extractors and quantified the contribution on the technique (refer to Community_data.csv in this dataset) to uniquely capture the ant morphospace by applying a new index (PWindex). Although ant TD overlapped, FD was significantly affected by the sampling technique. By controlling for TD effects, the community collected by each technique was differentially structured. Higher TD did not translate into wider morphospace for Winklers. Pitfalls recovered more functionally overdispersed assemblages. Pitfalls and Winklers overlapped in the sampling of the overall community, but each sampling method contributed with a unique spectrum to the ant morphospace. Our results suggest the importance of incorporating FD metrics in local ant inventories and the importance of sampling techniques when measuring the magnitude of FD and community structure. Our PWindex further illuminates sampling effects for ant assemblages.

File: Tables_S1-S4.xlsx

• Table_S1. Ant morphological traits used for this study, tentatively associated with functional roles and details of measurements.
• Table_S2. Multi-trait metrics used for this study, calculated over a Euclidean distance-matrix based on all traits (see Table_S1 and Table_S4 in Excel file) to calculate multi-trait metrics. Standardized effect sizes (SES) were used to deal with correlations at transect scale (N=22) between FD_PG, MNTD, and observed species richness (Sobs).
• Table_S3. List of ant species collected in the PEXJ, collected by pitfall traps (P) and with the Winkler extractor (W) during summer and winter seasons.
• Table_S4. Morphological measurements (average values in mm, recorded to the nearest 0.01mm using a Leica MZ95 stereomicroscope) for all the ant species and morphospecies collected in the PEXJ. Consult Table_S1 in Excel file for explanation of the acronyms. Only minor workers were measured to avoid allometric influence (minimum of six conspecific individuals whenever possible).

CSV Files

• File: Community_data.csv. Dataset for ant community collected in the PEXJ. Columns are as follow: area refers to edge border transects (B-1 and B-2), season (summer or winter), and method (pitfall trap of Winkler extractor).
• File: Trait_base_data.csv. Dataset for morphological measurements (average values in mm, recorded to the nearest 0.01mm using a Leica MZ95 stereomicroscope) for all the ant species and morphospecies collected in the PEXJ. Consult Excel file for additional information (Table_S1 for explanation of the acronyms and Table_S4 for information about replicates).