Atta Fabricius is an ecologically dominant leaf-cutting ant genus, the major herbivore of the Neotropics, and an agricultural pest of great economic importance. Phylogenetic relationships within Atta have until now remained uncertain, and the delimitation and identification of a subset of Atta species are problematic. To address these phylogenetic uncertainties, we reconstruct the most comprehensive phylogenetic estimate to date of Atta by employing ultraconserved elements (UCEs). We recovered 2340 UCE loci from 224 Atta specimens, which include 14 out of the 15 identifiable species from across their geographic distributions, and 49 outgroup specimens. Our results strongly support the monophyly of Atta and of the four clades that coincide with the previously recognized subgenera Archeatta Gonçalves, Atta s.s. Emery, Epiatta Borgmeier, and Neoatta Gonçalves. The Archeatta clade contains three species occurring in North and Central America and the Caribbean and is the sister group of the remainder of all other Atta species. The Atta s.s. clade is composed of two species occupying North, Central, and South America. The Epiatta clade contains seven entirely South American species and the two species of the Neoatta clade occur in Central and South America. Divergence-dating analyses identify a series of major events in the Miocene, such as the divergence of Acromyrmex Mayr and Atta 16.7 million years ago (Ma) and the crown-group origin of Atta around 8.5 Ma. Extant Atta species evolved very recently, originating in the early Pleistocene, approximately 1.8 to 0.3 Ma (crown-group ages). We provide the first evidence that Atta goiana Gonçalves belongs to the Epiatta clade and that Atta robusta Borgmeier is the species with the youngest crown-group age of 0.3 Ma. The very young ages of Atta and its component species indicate a recent, rapid radiation. Biogeographic analyses suggest that the range of the most recent common ancestor of Atta consisted of the combined North/Central America and NW South America bioregions and that one daughter lineage subsequently dispersed into South America, rapidly diversifying in the newly formed Cerrado biome and Chaco, and further dispersing into the Atlantic Forest, Caatinga, and Pampas bioregions.

We extracted DNA non-destructively from specimens using Qiagen DNeasy Blood and Tissue kit (Qiagen Inc., Valencia, CA). We quantified each DNA extraction with a Qubit 3.0 Fluorometer employing the High Sensitivity Kit (Thermo Fisher Scientific, Inc.). Sheared the DNA target size to an average fragment of 300-600 bp using a Qsonica Q800R3 Sonicator (Qsonica LLC, Newton, CT, U.S.A.). The fragmented DNA was used as input for libraries preparation using the Kapa Hyper Prep Library Kit (Kapa Biosystems, Inc., Wilmington, MA, U.S.A.) as described in Faircloth et al. (2015) with iTru Adapters protocol. We implemented all magnetic bead clean-up steps (Fisher et al., 2011) using a SPRI substitute (Rohland & Reich, 2012) as described in Faircloth et al. (2015). We used dual-indexing TruSeq adapters (Faircloth & Glenn, 2012; Glenn et al., 2019) for ligation. We amplified 15 µL of the library by PCR using the KAPA HiFi ReadyMix (Kapa Biosystems, Inc., Wilmington, MA, U.S.A.) with 25 µL of HiFi mix, 2.5 µL of each of Illumina TruSeq i5 and i7 primers, and 5 µL nuclease-free ddH₂O. We employed the thermal cycler program: 98 °C for 45 s; 14 cycles of 98 °C for 15 s, 60 °C for 30 s, 72 °C for 60 s; and final extension at 72 °C for 5 m. We purified DNA products using 1.2x speedbeads and rehydrated the purified product in 22 µL of Elution Buffer (EB, pH= 8). We quantified individual libraries using a Qubit 3.0 Fluorometer using the Broad Range Kit. We pooled eight to 12 post-PCR libraries at equimolar concentrations into 22 pools. We adjusted pool concentrations to 147 ng/µL by drying samples in a vacuum centrifuge for 45-60 m at 60 °C, and then re-suspending them in nuclease-free water. We quantified each pool concentration in a Qubit 3.0 Fluorometer with the Broad Range Kit. We enriched each pool using 9446 baits (myBaits®; Arbor BioSciences) targeting 2524 conserved loci in Hymenoptera (Branstetter et al., 2017b) for 24 h at an incubation temperature of 65 °C. Enrichment, bead-cleaning, and PCR reaction procedures followed Borowiec (2019). The resulting reactions were purified using 1.0X speedbeads and the enriched pools were then rehydrated in 22 µL EB. We quantified each enriched pool using the Qubit 3.0 Fluorometer and the Broad Range Kit. We verified the enriched pools DNA concentration with qPCR on an Applied Biosystems 7500 Real-Time PCR System (Applied Biosystems™, Thermo Fisher Scientific, Inc.) using the KAPA Library Quantification Kit (Kapa Biosystems, Inc.). We then used the resulting concentration estimates to pool the libraries at equimolar concentrations in a single final pool. We sent two final enriched pools to the University of Utah High Throughput Genomics Core Facility for quality control and sequencing of two full lanes of a HiSeq 2500 (125 Cycle Paired-End Sequencing v4) run.