The two-spined trap-jaw ant, Odontomachus haematodus (Linnaeus), is native to South America and was introduced to the United States via Mobile, Alabama, in the mid-twentieth century. Misidentification as a native congener for several decades has inhibited the study of the ecology of O. haematodus in its introduced range within the Gulf Coastal Plain. We conducted a multi-method survey to: 1) assess the abundance of O. haematodus across urban parks in Hattiesburg, MS, the northernmost published record in this species’ exotic range, 2) establish which sampling methodologies are most effective for conducting urban surveys of O. haematodus, and 3) study the nesting preference of O. haematodus in urban parks. We observed O. haematodus in 6 out of 10 urban parks in Hattiesburg, MS. Of the six methodologies we compared, the most effective techniques for detecting O. haematodus were opportunistic collections, pitfall traps, and surveying the bases of trees for active nests. We found nests most often along the bases of trees in urban parks, without apparent nest-choice preferences for specific types or sizes of trees. These results are consistent with the assessment that O. haematodus is well established in this region and can nest in a wide range of microhabitats.

Specimen identification

Collected ant specimens were identified to species using easily observable morphological characters in several published keys (Brown 1976; MacGown et al 2014; Schmidt and Shattuck 2014). Characters enabling identification of the reported specimens to genus Odontomachus included a large body size (> 0.8 mm; Figure 2A-B), elongated mandibles inserted medially on the anterior of a head that is longer than wide (Figure 2C), and a conical-shaped petiole that is taller than long (Figure 2D). Characters used to identify specimens to species included Weber’s length being greater than 3 mm (Figure 2A), overall dark color (Figure 2A-B), petiole with transverse striae (Figure 2D),  elongate and spiniform metasternal processes (Figure 2E), and a lightly shagreened dorsal surface on the first tergite of the gaster (Figure 2F).

Describing diel activity

Most Odontomachus are reported to be nocturnal, with peak colony activity at night and typically observable but reduced activity during the day (MacGown et al 2014). To ascertain the best time of day to perform active collection methods that assume the target is in its foraging period (opportunistic collection and baiting), we first observed round-the-clock activity of a single large O. haematodus nest. This colony was first identified in June 2021 and is nested at the base of a water oak, Quercus nigra, tree on an empty lot road verge one block away from Town Square Park in Hattiesburg, MS (31.3221, -89.2864). Once an hour, from 16:00 on 15 June 2024 to 15:00 on June 2024 (excluding the hours of 112 amto 4 am when access was not available), the number of ants entering and leaving the nest per a 60-second observation period was recorded, with notes taken of how many of those ants were carrying foraged items into the nest, and when observable, what types of foraged items were carried. The times of day with the highest incidence of foraged items being transported into the nest were chosen to conduct the multi-park survey.

Comparing sampling methods

We compared six common methods for sampling ants: 1) the use of Berlese funnels to extract ants from leaf litter samples, 2) pitfall traps, 3) Vienna sausage (high-protein) baits, 4) Pecan Sandies ® (mixture of carbohydrate, lipids, and protein) baits, 5) opportunistic collections, and 6) nest mound surveys in their ability to detect O. haematodus. We used all six methods to collect ants in each of 10 park sites distributed across Hattiesburg, Mississippi (Figure 1B) between 20 June 2024 and 7 August 2024. Following preliminary observations of peak foraging activity in the early morning hours (Figure 3), we conducted all opportunistic collections, bait sampling, leaf litter sampling, and nest searches between the hours of 5:30 am and 10 am on mornings without rain.

Following a standard baiting procedure established for other ants (Castracani et al 2020), within each site,40 scintillation vials (20 ml each) arranged in 4 transects were left uncapped for 30 minutes. Each of the four transects contained 10 vials spaced 2 m apart. Vials contained one of two bait types: 1/8 of a canned Vienna sausage (Bryan Foods, distributed by Sweet Sue®, San Diego, CA) or 1/8 of a Pecan Sandie cookie (Kebler®, distributed by Ferrero USA, Parsippany, NJ). Within each park, one transect of each bait type was placed in mowed lawn (sod) and the other in the leaf litter, pine straw, or mulch that surrounded landscaped park plants. After 30 minutes, vials were capped and then immediately frozen.

For 30 minutes at each site (following deployment of baits), a trained myrmecologist (able to visually identify ant morphospecies with reasonable reliability) walked haphazardly tthroughthe park, visually inspecting the ground and sides of human structures or plants. When ants were encountered, 1 to 3 voucher specimens of each newly encountered morphospecies per park were collected using a pair of forceps (for larger species) or a hand-held aspirator (for smaller species). Opportunistically collected ants were immediately placed into 95% ethanol during this process.

We used standard methods, approximating work by Sabu et al (2011), to extract ants from sampled leaf litter. Ten 25 x 25 cm leaf-litter quadrat samples (> 5 m between each sample, haphazardly spaced based on availability of leaf litter in each park) were taken from each site and transported to the lab. Ants were separated from leaf litter for 48 hours using Berlese funnels (Model #2831, Bioquip Products, Inc., Gardena, CA) equipped with 40W incandescent bulbs. At the base of each funnel was affixed a collection beaker containing 50 mL of 95% ethanol.

The bases of trees in each park were visually searched for nest mounds of O. haematodus. The first 50 trees starting from each park entrance were surveyed. For this survey, a tree was defined as any plant > 1.5 m in height and > 5 cm in diameter at 0.25 m height. Surface-visible nest mounds of O. haematodus are easily recognizable byloose soill, detritus, and plant debris piled against the base of logs, trees, or other structures, with large 4 mm to 11 mm diameter entrance tunnels being typical (Figure 4). An observed nest was only counted in this survey if a specimen of O. haematodus was successfully collected from the nest and verified morphologically using established characters (Figure 2).

At each park, we laid 10 in-ground covered pitfall traps per site. Each trap consisted of a 200 mL plastic cup filled with 30 mL of 5% soapy water. Location of pitfalls within each park was haphazardly chosen among low-human-traffic areas of the park (i.e., flower bed, behind fence, or off-trail in bushes), with 2 meters between each trap. Pitfalls were left in place for 24 hours before contents were collected into 95% EtOH for later identification.