Stable isotope analyses
Workers were dried in an oven at 60˚C for 24-48 hours and then stored in airtight vials prior to processing. The abdomens of all ants were removed prior to weighing to avoid the effects of stomach contents on isotopic signatures (Tillberg et al., 2006). To achieve appropriate weights for each sample, five to ten workers per sample were pooled and chopped in glass vials to fine homogeneous powders using small scissors (Fig 1). Approximately 0.400 mg of each sample was weighed into tin capsules (Costech Analytical Technologies Inc., Valencia, CA, USA) using a microbalance (Mettler Toledo, Columbus, OH, USA). All samples were analyzed at the Texas A&M University Stable Isotopes for Biosphere Science Laboratory (https://sibs.tamu.edu/) using a Delta V Advantage Isotope Ratio Mass Spectrometer coupled with a Costech Elemental Analyzer and Thermo ConFlo IV Universal Interface (Thermo Fisher Scientific, Waltham, MA, USA). All baseline samples (collected before mounds were fed the tracer) were analyzed before any post-feeding samples to ensure that natural abundance values were not influenced by memory effects from the high levels of 15N in spiked samples. Nitrogen isotope ratios are presented in \(\delta\) notation:
δ15N (‰) = [(Rsample – Rstandard)/Rstandard] x 103
where Rsample is the15N/14N ratio of the sample and Rstandard is the15N/14N ratio of the atmospheric N standard (Mariotti, 1983; Coplen, 2011). Precision was 0.1‰.
To verify that our isotope tracer methods could detect resource sharing over 30-m distances, we conducted a proof of concept experiment using a population of tawny crazy ants (Nylanderia fulva ) near College Station, Texas. This species forms a single supercolony throughout its invaded range in North America, in which workers regularly share collected resources with each other and occupy transitory nests (Eyer et al., 2018; Wang et al., 2016). Our results from this experiment confirmed that the isotope tracer is highly successful at detecting sharing in unicolonial populations at distances that were relevant for our study (i.e., up to 28.4m from the treated area; Kjeldgaard et al.,unpublished data ).