Amplification and sequencing of mtDNA
In order to augment our microsatellite data, we included mtDNA previously sequenced from Seal et al. (2015; n = 61) and sequenced new samples directly for this study (n = 26). A total of 55 samples had both microsatellite and mtDNA sequenced (Table S1). Again, DNA was extracted from whole individual ants using a QIAamp DNA Micro Kit (QIAGEN), and a 779-bp sequence was obtained from the COI-tRNA Leucine-COII region of mtDNA. We used the following primers: C1-J2195 (alias CO1-RLR; 5′-TTGATTTTTTGGTCATCCAGAAGT-3′); and C2-N-3661 (alias Barbara; 5′- CCACAAATTTCTGAACATTGACCA-3′(Simon et al., 1994)). PCR mixtures and cycling profiles were identical to Seal et al. (2015). PCR products were then purified and sequenced at the University of Texas at Austin’s DNA Sequencing Facility on an Applied Biosystems 3730 DNA Analyzer. Chromatograms were visually checked and resolved in Geneious v10.2.3 (Kearse et al., 2012), and sequences were aligned using MEGA v6.06 (Tamura, Stecher, Peterson, Filipski, & Kumar, 2013) using the ClustalW algorithm (Thompson, Higgins, & Gibson, 1994). New sequences are deposited into GenBank under accession numbers MN088095-MN088120 (Table S1). mtDNA is known to have several properties that make evolutionary conclusions problematic, such as pseudogenes or nuclear insertions (numts) (Beckenbach, 2009; Cristiano, Cardoso, & Fernandes-Salomão, 2014; Martins et al., 2007; Toews & Brelsford, 2012). Therefore, we examined for stop codons in our sequences (other than at the COI-tRNA Leucine transitions) (Seal et al., 2015). Sequences were also long and frequently readable at >800bp. Two primary advantages of mtDNA sequences are that 1) sequences can be readily obtained and 2) the problems associated with mtDNA are understood and can be easily examined, whereas the problems associated with nuclear markers are more uncertain (Bowen et al., 2014; Moreau, 2009). Moreover, a preliminary examination of the microsatellite markers of T. septentrionalisused in this study supported one of the main findings of mtDNA-based studies in this species: pronounced genetic differentiation across the Mississippi River Valley (Matthews et al., 2020), thus making it unlikely that the COI sequences analyzed were numts, which usually lack variation because of purifying selection in nuclear genomes (Martins et al., 2007). Therefore, mtDNA sequences in this species are likely robust genetic tools (Matthews et al., 2020; Mikheyev et al., 2008; Seal et al., 2015).