One assumed function of herbivore-induced plant volatiles (HIPVs) is to attract natural enemies of the inducing herbivores. Field evidence for this is scarce. Also, the assumption that elicitors in oral secretions that trigger the volatile emissions are essential for the attraction of natural enemies has not yet been demonstrated under field conditions. After observing predatory social wasps removing caterpillars from maize plants, we hypothesized that these wasps use HIPVs to locate their prey. To test this, we conducted an experiment that simultaneously explored the importance of caterpillar oral secretions in the interaction. Spodoptera caterpillars pinned onto mechanically damaged plants treated with oral secretion were more likely to be attacked by wasps compared to caterpillars on plants that were only mechanically wounded. Both of the latter treatments were considerably more attractive than plants only treated with oral secretion or left untreated. Subsequent analyses of headspace volatiles confirmed differences in emitted volatiles that likely account for the differential predation across treatments. These findings highlight the importance of HIPVs in prey localization by social wasps, hitherto underappreciated potential biocontrol agents, and provide evidence for the role that elicitors play in inducing attractive odor blends.

Field Trials: Pots containing two maize plants were randomly assigned to one of four treatments: unmanipulated control (CN), oral secretion only (OS), mechanical damage only (MD), and mechanical damage plus oral secretion (MD+OS). A pilot study on 1 January 2022 (n = 6 per treatment) did not contain the OS treatment but was included in the analysis. Two trials were conducted the following winter on 13 (n = 11) and 21 (n = 12) December 2022.  In total, 220 plants in 110 pots were used (treatment, n: CN, 29; OS, 23; MD, 29; MD+OS, 29).

To maximize emissions of HIPVs, we damaged the plants twice, once the evening before (17:00 h) and a second time the morning of the trial (7:00 h). We did this because, in response to specific elicitors in caterpillar oral secretion, plants typically initiate the de novo synthesis and release of various volatiles (Alborn et al. 1998; Schmelz et al., 2006). In maize, this takes about 8 hours and can be reinforced with repeated treatments (Turlings et al., 1993). The MD treatment consisted of using sandpaper to abrade about 2 cm² of the adaxial side of the two youngest developed leaves, avoiding the midvein. The OS treatment consisted of applying FAW regurgitant containing gut and salivary gland contents on the two youngest leaves (also on about 2 cm²) by gently squeezing the head of one caterpillar until it regurgitated. One caterpillar was used as a source of oral secretion for each pot. For the MD+OS treatment, the plants were mechanically damaged and immediately after, the oral secretion was applied directly to the wounded tissue. After imposing the second damage treatment, approximately 14 hours after the initial treatment, one randomly selected live, 3^rd FAW or VAW caterpillar was pinned to the center of each leaf. This timing coincides with when the wasps most actively forage (Prezoto et al., 2019). Live caterpillars (as opposed to freeze-killed) were used to ensure that wasps were acting as predators and not scavengers.

At 9:00 h, pots containing induced plants with pinned caterpillars were arranged in spatial blocks containing one treatment set (CN, OS, MS, MS+OS, where 1 treatment set = 1 n) in maize plots within the experimental field. Each pot was approximately 30 cm apart, forming a square with a random arrangement of treatments. Blocks were visited continuously over 75 minutes and the order in which sentinel caterpillars from the different treatments were removed in each block was recorded.

Volatile Collection: We collected volatiles as described in Turlings et al (1998). Volatiles were trapped in filters (Porapak Q adsorbent, 25mg, 80-100, Merck, Darmstadt, DE) connected to the port of each glass bottle containing the odour source for 2 hours at a flowrate of 0.4 L/min. Each filter was then eluted with 150 μL of dichloromethane (Suprasolv, GC-grade; Merck, Darmstadt, DE) into glass vials to which 10 μL of internal standards ([20ng/μL] of n-octane and n-nonyl acetate) were added. Samples were then kept at -80°C until further analyses.

Volatile analyses were carried out using a gas chromatograph (Agilent 7890B) coupled to a mass spectrometer (Agilent 5977B) with Flame Ionization Detector technology (GC-MS / FID). A 2 μL aliquot of each sample was injected in pulsed splitless mode onto a non-polar capillary column (Agilent HP-1; 30m length x 0.25mm ø x 0.25 μm thickness). After the injection, the temperature program maintained the temperature at 40°C for 3 minutes before increasing it progressively at a rate of 8°C min⁻¹ until reaching 100°C; it thence increased it at 5°C min⁻¹ until 200°C and subsequently rose it at 250°C for a 3 minutes post-run. Volatiles were identified using the NIST mass spectral library. Volatile emissions (ng hr^-1) were calculated based on the peak areas of internal standards.