Discovery-defense strategy as a mechanism of social foraging of ants in tropical rainforest canopies

We defined six 30-m radius circular plots separated by 250 – 800 m. At each sampling plot, we chose the four tallest canopy trees, making a total of 24 trees. Between May and June 2018, we climbed each of these 24 trees using the single rope technique. We carried out observations at the highest possible point of the tree crown where the climber was safe (21.5 ± 5 m), usually between 10:00 and 17:00 h. Three baits were placed as far as possible from the trunk on different branches of each focal tree (n = 72 baits). Placing baits on different branches implies exposing them to distinct foraging trails, guaranteeing the independence of our observations. One tea spoon of bait (a fifty-fifty mixture of tuna with honey) was wrapped in a paper napkin and placed on the branches. We performed observations on each bait every fifteen minutes, from 15 min (first observation) to 195 minutes (thirteenth observation) after placement of the bait, in order to evaluate species’ change over time. During these observations we counted the number of individuals of each ant species at the baits (only if they were in contact with the bait). When it was impossible to count the exact number of ant workers because they were very tiny (e.g., Pheidole and Azteca species) and/or numerous individuals were present on the bait (e.g., more than 20 ant workers), we estimated the number visually, giving a specific number for our estimation, for example: “50 workers”. We avoided removing ant workers from the baits as much as possible, especially the discoverer species (i.e., the first ant species to discover each bait). To investigate the pool of ant species from each focal tree, using forceps, we collected voucher specimens outside the baits, i.e., ants that were foraging on trees and not feeding on the baits, especially at the end of the observation period.