2.1 Source Host and Parasitoid InsectsFor all assays, 4–8-week-old R. dominica were reared on wheat, while S. oryzae were reared on wheat tempered to 13% grain moisture. To subculture, a total of 50 individuals were placed on 200 mL of grain in a mason jar (capacity: 473 mL) and given 14 d to mate and lay eggs. At the end of that period, adults were removed by sieving with a #10 sieve (2.00 mm; W.S Tyler Inc., Mentor, Ohio), and colonies were allowed to age for 3-weeks prior to using beetles as hosts for parasitoid rearing. Hosts used for experiments below were 2–3 weeks old. Theocolax elegans were maintained separately on two different hosts, either R. dominica or S. oryzae for at least three full generations. All colonies of hosts and parasitoids were held at 27.5°C, 60% RH, and 14:10 L:D, with parasitoids maintained in a separate environmental chamber than host only colonies to prevent cross-contamination.2.2 Odor TreatmentsOdor treatments included: 13 g of S. oryzae-damaged grain (SO-grain, hereafter) from the non-natal environment, 13 g of R. dominica-damaged grain from the non-natal environment (RD-grain), 13 g of damaged grain + conspecifics from the natal environment (Natal-grain), 10 S. oryzae adults alone (SO), 10 R. dominica adults alone (RD), and a clean (uninfested and undamaged) grain control (Ctrl). Treatments were always freshly sourced from colony material as described above, and adults were sieved out of the insect-damaged grain treatments. Grain was only pulled after colonies were 4-weeks-old. These treatments were used as odor sources for the still-air and four-way olfactometer assay. Odor treatments for the headspace characterization included: clean grain, R. dominica-infested grain, R. dominica-infested grain + T. elegans, S. oryzae-infested grain, S. oryzae-infested grain + T. elegans, psocid-infested grain, and a clean control. We included a psocid-infested grain treatment to rule out the influence of psocids in some of the replicates, which comprised incidental contamination.2.3 Four Arm OlfactometerIn a four-arm, still-air olfactometer, we evaluated the orientation and taxis of R. dominica- or S. oryzae-reared T. elegans to the odor treatments discussed above (Fig. 2). The custom-built olfactometer consisted of a central, circular (8.26 × 2.54 cm D:H) acrylic release chamber with 12 holes per cm (each of 1.75 mm D), with four abutting rectangular, glass chambers (6.35 × 6.35 × 2.54 cm L:W:H). The bottom of the olfactometer consisted of a single glass sheet (25.4 × 25.4 cm W:L). In each trial, one of the adjacent chambers was randomly selected to contain the odor treatment, while the other three remained empty. A single parasitoid was released in the center of the circular release chamber, and a sheet of glass (25.4 × 25.4 cm W:L) was immediately placed over the top of the olfactometer. The time to response of first decision, and the zone on which adults exited was recorded as either the treatment chamber (stimulus), or empty chamber (non-stimulus; one of the other three edges). Parasitoids were given 3 min to respond to the odors, and non-responders were excluded from analysis. A total of n = 15 replicate wasps were tested per treatment. After each replicate, the olfactometer was wiped down with methanol, then hexane, and allowed to dry. At the end of a day of testing, the whole apparatus was thoroughly washed with soap and water.2.4 Headspace CharacterizationTo characterize the relative difference in volatiles among treatments, a headspace collection system was used (after Van Winkle et al. 2022). Central air was scrubbed using an activated charcoal filter, then pushed through the remaining apparatus. The airflow was restricted to 1 L/min using a flow meter (Volatile Collection Systems, Gainsville, FL, USA) placed directly prior to the sample collection from the headspace chambers (10.2 × 12.7 cm D:H, 500 mL capacity) with an inlet for air and an outlet for a volatile collection trap (VCT). The headspace volatiles from the odor sources above were adsorbed for 24 h onto a VCT consisting of a drip tip borosilicate glass tube packed with 20 mg of Porapak-Q™ (Volatile Collection Systems, Gainsville, FL, USA) to adsorb volatiles with a stainless-steel screen (No. 316) on one side, and held in place with a borosilicate glass wool plug followed by a PTFE Teflon compression seal. The volatiles on the traps were eluted with 150 µL of dichloromethane (Millipore, Billerica, MA, USA) by pushing the solvent through with inert N2 gas (Ultra-high purity, Airgas, Sacramento, CA, USA) into a 2 mL GC vials with a 150 µL glass insert with polymer feet (Part #5183-2088, Agilent Technologies, Inc., Santa Clara, CA, USA). Vials were then sealed with screw caps containing a Teflon-lined septum, wrapped in PTFE tape, and stored at -20oC until GC-MS analysis (Agilent Technologies, Inc.). VCTs were washed three times with 700 µL of dichloromethane that was pushed through with N2 and then reused. To quantify the samples, 1 µL of tetradecane (190.5 ng, 99% purity, GC analytical grade, Millipore, Billerica, MA, USA) was added as an internal standard using a microsyringe (2 mL capacity syringe, Hamilton Co., Reno, NV, USA).2.5 GC-MS MethodologyAll headspace collection sample extracts were run on an Agilent 7890B gas chromatograph (GC) equipped with an Agilent Durabond HP-5 column (30 m length, 0.250 mm diameter and 0.25 mm film thickness) with He as the carrier gas at a constant 1.2 mL/min flow and 40 cm/s velocity, which was coupled with a single-quadrupole Agilent 5977B mass spectrometer (MS). The compounds were separated by autoinjecting 1 mL of each sample under splitless mode into the GC-MS at approximately 250°C. The initial oven temperature was 55°C, which was increased to 300°C at a rate of 10°C/min, where it was held for 4 min at the final temperature. After a solvent delay of 3 min, mass ranges between 50 and 550 atomic mass units were scanned. Compounds were preliminarily identified by comparison of spectral data with those from the NIST 14 library and by GC retention index and Kovats index (Adams 2009).2.6 Pilot Scale Elevator ReleasesIn order to evaluate the potential success of Theocolax elegans as an augmentative biological control agent, the ability of the parasitoid to target R. dominica and S. oryzae in a pilot-scale elevator at the USDA-ARS Center for Grain and Animal Health Research in Manhattan, KS was assessed. Theocolax elegans that had been reared for 2–3 generations on R. dominica or S. oryzae were used for this experiment. Sentinel commodity patches consisting of petri dishes (14 ⨯ 2 cm) that contained 50 g of R. dominica-infested grain or S. oryzae-infested grain without adults were placed 0.5, 1, 4, and 8 m away from a wooden box (21.6 × 29.2 × 5.7 cm W:L:H) in which 100 R. dominica-reared or S. oryzae-reared T. elegans were released in 50 g of clean, uninfested grain as a neutral refuge. In the negative control, the same setup was used, except the box with grain did not contain parasitoids to assess any background amount of parasitism in the pilot-scale elevator. After 72 h, dispersal of adult T. elegans at the release point and at each dish with grain were recorded. There was a total of n = 6 replicate releases from 10 May 2019 to 21 August 2019 and 25 June 2021 to 29 October 2021. Afterwards, dishes with infested grain were brought back to the laboratory and held at constant conditions (27.5°C, 60% RH, and 14:10 L:D) for >6 wks to evaluate progeny production and wasp emergence. Progeny and parasitoid production were evaluated daily for freshly emerged individuals until emergence of all adults ceased. Finally, grain damage was recorded by counting the number of insect-damaged kernels (IDK) and weight of damaged grain from each sample. Using total number of adults and parasitoids emerging from each treatment, a parasitism rate was also calculated at each distance.