2.1 | Between-nest interactions
To quantify sharing between nests in the field, we treated selected nests in each site with a stable isotope tracer and quantified its movement into neighboring nests (Fig 1). Stable isotope tracers employ naturally occurring, non-radioactive forms of biologically relevant elements, such as nitrogen. The heavier isotope of nitrogen (15N) occurs rarely in nature, so by artificially ‘spiking’ a food with an appropriate concentration of this heavy isotope, we can trace the movement of this isotope through consumers and identify the flow of nutrients through an ecosystem (Fry, 2006). We ensured that only the treated nests had access to the isotope tracer, so if a neighboring untreated nest showed unnaturally high levels of15N, this would indicate an exchange of either workers or resources between the treated and untreated nests (i.e., no boundaries between nests). In a preliminary laboratory experiment, we detected highly elevated levels of 15N as quickly as 24 hours after feeding fire ant workers an isotope tracer, and these isotope values decreased steadily over a 72-hour period (Kjeldgaard 2020). These preliminary results indicate that the tracer must have been relatively recently consumed to be detected in workers, which would allow for an effective measure of the movement of marked food.
Sampling was conducted between August and October 2019 in six field sites in Texas, USA (Appendix S1). Habitats ranged from restored grasslands (sites O, A, and B) to mowed fields (sites C, T, and T2). Mounds were used as a proxy for individual nests. At this stage of the experiment, we could not determine if distinct mounds/ nests belonged to the same polydomous colony. After searching each field site and identifying all active fire ant nests, we selected three clusters of four to five nests at each site. One nest within each cluster was selected as the treatment nest. Clusters were separated by at least 50 m within each site to avoid potential sharing between clusters (Fig 1). Mounds were selected to represent varying distances within each cluster (between 0.4m-29.07m from the treated nest, with an average distance of 7.65m ± 0.72m) to determine any effect of distance on sharing between nests.
Similar to other studies (Goodisman et al., 2007), several fire ant nests disappeared or moved over the course of the sampling period. As a consequence, we were unable to find three nests (one nest from site T, one from O, and one from T2) after the treatment period. Each of these were untreated nests within different clusters, so their removal did not affect the number of clusters analyzed in each location. In total, we sampled from 73 fire ant nests across six sites, with 12 nests in Site A, 13 in Site B, 12 in site C, 11 in Site O, 13 in Site T, and 12 in Site T2.