The study site consisted of an open field outside Davis, CA (approximately 0.5 km southwest of the campus of the University of California, Davis) that was bordered by farmland to the north and a creek to the south. Plants at the field site consisted largely of weedy grasses and large stands of shortpod mustard (Hirschfeldia incana (L.)), a weedy brassicaceous host of B. hilaris. Patches of another weedy host, perennial pepperweed (Lepidium latifolium L.), were also present. The site was selected because it was heavily infested with B. hilaris and other pentatomid species, including Murgantia histrionica (Hahn), Chlorochroa ligata (Say), Chlorochroa uhleri (Stål), and Thyanta custator (Fabricius). Shortpod mustard was the focus of sampling because it appears to be the favored host of B. hilaris in California; B. hilaris is typically present on shortpod mustard until it completely senesces in late summer or fall (B. Hogg, pers. obs.).

Bagrada hilaris were collected for the colony at the study site and maintained in ventilated plastic food containers (25 × 17 × 8 cm) enclosed by organdy mesh, at 28–30 °C, 30–40% RH and a 16L:8D photoperiod. The colony was provided with organic (without pesticide treatment) broccoli florets (Brassica oleracea var. italica Plenk), kale (Brassica oleracea var. palmifolia DC), and sweet alyssum (Lobularia maritima (L.)). To obtain B. hilaris eggs for sentinel cards, two sand-filled, uncovered petri dishes (90 mm diameter) were placed in each container under folded pieces of white cardstock (80 × 80 mm) to provide shade, and eggs were collected daily by sieving the sand using a No. 35 sieve (Humboldt Manufacturing Co., Elgin, IL; mesh size: 0.5 mm).

Parasitism of B. hilaris eggs was assessed using sentinel B. hilaris eggs attached to pieces (2 × 14 cm) of white cardstock (Rite in the Rain, Tacoma, WA) using Elmer’s Glue-All (Elmer’s Products Inc., Westerville, OH) (Fig. 1). To protect the sentinel eggs from sun and precipitation, larger pieces of cardstock were folded to create triangular tubes (14 cm long with 7 cm sides) within which egg cards were attached with paperclips (Fig. 1). Sampling took place in late summer and early fall, when B. hilaris numbers are typically highest in north-central California (B. Hogg, pers. obs.). Cards were placed in the field on 10, 17, and 29 September 2020 and on 15, 21, and 28 October 2020, and remained in the field for six to eight days (six days: 17 September, 15 October; seven days: 10 and 29 September, 21 October; eight days: 28 October). Depending on egg availability, either 15 (all September dates) or 10 (all October dates) eggs were glued to each card; eggs on cards were evenly spaced 1 cm apart (Fig. 1).

Figure 1. 

From left to right: A sentinel card with Bagrada hilaris eggs; a sentinel card in the field, inside a cardstock triangle to protect it from sun and precipitation; sentinel cards on the ground and in mustard foliage in the field.

Cards were placed in two locations on B. hilaris-infested shortpod mustard plants: on the ground, where most B. hilaris eggs are laid, and in the foliage (~1 m above the ground) (Fig. 1). Cards were placed in foliage to increase the potential pool of parasitoids that were sampled because parasitoids that attack pentatomids other than B. hilaris are likely to forage primarily in foliage, where most or all other pentatomids in California lay their eggs. On each sample date, two cards per plant (one on the ground and one in foliage) were placed in eight infested plants, except on 21 October when cards were placed in only six plants due to low egg availability. The same eight plants were sampled through September, but by mid-October four of the original plants had begun to senesce and contained few B. hilaris, and on each sample date in October four new infested plants were chosen as replacements, except on 21 October when three original and three new plants were sampled. Sampled plants were at least 9 m apart, and the entire distance from the first to the last sampled plant was 426 m. Cards were placed intact in plastic vials (25 mm diameter × 95 mm high) in the laboratory at 20–24 °C, 40–60% RH and 12L:12D photoperiod and were held for at least four weeks for the emergence of parasitoids, which were identified to species. Any unhatched eggs were dissected to determine whether uneclosed parasitoids were present.

Figure 2. 

A female Gryon aetherium parasitizing an egg (arrow points to the inserted ovipositor).

Eight shortpod mustard plants heavily infested with B. hilaris were selected for soil sampling at the site, such that each plant was located within 10 m of a plant bearing sentinel egg cards. Initially, undisturbed soil was sampled from under the infested plants, but few B. hilaris eggs were found, likely because the ground was too densely packed. To facilitate egg laying by B. hilaris on subsequent sample dates, leaf litter and grass and a ~1 cm-deep layer of soil was removed from a 40 cm diameter area under each plant. The soil was then sieved with a No. 35 sieve and placed back, and loose dried grass collected from the site was placed on top of the soil to provide shade. The soil was left for 6–7 days to allow B. hilaris to lay eggs and was then removed with a trowel and placed in resealable plastic bags. Soil samples were collected on four dates: 23 and 29 September and 6 and 15 October.

In the laboratory, the soil was sieved first with a No. 14 sieve (mesh size: 1.41 mm) to remove large debris, and then with a No. 35 sieve to remove soil particles smaller than B. hilaris eggs. The remaining soil particles were then examined microscopically for B. hilaris eggs, and all eggs that were found were placed in plastic vials (25 mm diameter × 95 mm high) and held at 20–24 °C, 40–60% RH and 12L:12D photoperiod for at least 4 weeks for parasitoid and B. hilaris emergence. Any remaining unhatched eggs were dissected.

Interbreeding tests were conducted with the population of G. aetherium currently undergoing host specificity testing in quarantine (originally collected in Pakistan; hereafter the “Pakistani” population) and a wild Californian colony established from individuals that emerged from the sentinel egg cards and soil samples collected in the study described above. Since G. aetherium is arrhenotokous (i.e., female progeny cannot be produced by unmated females), the production of female progeny is a sign that females were fertilized by males. A preliminary experiment with five unmated females confirmed the probable absence of thelytoky in the wild population (i.e., the females produced only male progeny).

Experiments were conducted in the USDA-ARS quarantine facility in Albany, CA at 21–26 °C, 40–60% RH, and 14L:10D. The developmental time of G. aetherium was about 22 days (range: 16–28) under these conditions, with males usually emerging 1 to 2 days before the females. To ensure the synchronized availability of males and females from both populations, three batches of B. hilaris eggs were parasitized by the two Gryon populations several days apart. All males and females used in the experiment came from these inoculation batches. After inoculation, the parasitized eggs were kept individually in 2 ml microtubes with honey as a food source for the emerging adults. Within 24 h of emergence, one male was added to the microtubes containing single virgin females. There were four treatments: Pakistani males were paired with Californian females and vice versa, and control treatments were set up by pairing males and females from the same population. If mating was not observed within 10 min of adding the male, 2 to 3 additional males were added to the microtube and left to mate for 24 h. Because parasitism by G. aetherium has been reported to be the highest right after emergence (about 12 eggs/female), steadily declining thereafter (Martel et al. 2019), single females were then transferred to glass vials (25 mm diameter × 95 mm high) containing 15 to 20 fresh (< 24 h old) B. hilaris eggs glued to a strip of cardstock (20 × 60 mm). After a 24 h exposure to the eggs, the females were removed and preserved in 95% ethanol, and the egg cards were incubated until the emergence of B. hilaris nymphs or parasitoids. Unhatched parasitized eggs (blackened) were subsequently dissected and any identifiable male or female G. aetherium (based on antennae morphology) were included in the calculation of progeny sex ratio. Ten to twenty replicates were completed for each treatment (Table 2).

Subsequent fertility of female progeny was tested by allowing the progeny resulting from the two crosses (Pakistani males × Californian females and Californian males × Pakistani females) to mate with siblings emerging from the same egg card, or if no males (or not enough) were available, with males emerging from other replicates in the same treatment. A subset of the females was then exposed to B. hilaris eggs as above. All females were paired with males within 24 to 48 h of emergence. Pairs could mate for 24 to 48 h and exposure to host eggs lasted 24 or 48 h. The fertility of 57 and 43 female progeny was tested from the Pakistani male × Californian female and the Californian male × Pakistani female crosses, respectively (Table 2).

Progeny sex ratio was calculated for each parental female as the proportion of female progeny produced; proportions were logit-transformed and then compared between treatments using ANOVA and the lm function in R version 4.0.2 (R Development Core Team 2020). Replicates with no parasitism from initial crosses (n = 4 total; n = 1 in the Pakistani male × Pakistani female control and n = 3 in the Californian male × Pakistani female cross) and subsequent fertility tests of female progeny (n = 16 total; n = 5 for the Pakistani male × Californian female cross and n = 11 for the Californian male × Pakistani female cross) were excluded from analyses.

The taxonomic work undertaken to identify G. aetherium, including molecular and morphological studies, is presented and discussed in a companion paper (Talamas et al. 2021).

Of the 972 sentinel eggs deployed on 92 cards in this study, 28 (2.88%) on three cards were parasitized by G. aetherium and 41 (4.22%) on five cards were parasitized by O. californicus (Table 1). Both species parasitized sentinel eggs on the ground and in the foliage, although most eggs attacked by O. californicus were on cards in foliage (four cards in foliage versus one on the ground). Egg parasitism by G. aetherium only occurred on cards collected on 23 September and 5 November and reached a high of 12.9% on cards on the ground on 5 November. Ooencyrtus californicus was present on all of the first four collection dates, and parasitism by it reached a peak of 17.9% on cards in foliage on 23 September (Table 1).

Of the 154 uneclosed B. hilaris eggs that were collected from soil, 37 (24.0%) were parasitized by G. aetherium and none were parasitized by O. californicus (Table 1). All unparasitized eggs contained B. hilaris nymphs that either emerged (n = 97) or died before emergence (n = 20). By mid-October B. hilaris had apparently stopped laying eggs; only one egg in soil was recovered on 15 October (Table 1). Parasitism by G. aetherium in soil peaked at 46.2% on 6 October (Table 1).

All treatments produced female progeny (Table 2). The proportion of female progeny differed between treatments (ANOVA, F_3,51 = 3.58, P = 0.02). Proportions of female progeny were higher in the two treatments that included Californian males (Table 2), although the only significant difference was between the control with Californian males and females and the cross between Pakistani males and Californian females (Tukey HSD test, P < 0.05; Table 2). In a 2 × 2 factorial ANOVA, the proportion of female progeny was affected by the origin of male G. aetherium (Pakistan/California) (F_1,51 = 10.49, P = 0.002), but not by the origin of females (F_1,51 = 0.21, P = 0.65) or by the male*female origin interaction (F_3,51 = 0.23, P = 0.63).

In subsequent fertility tests, female progeny of both crosses (Pakistani males × Californian females and Californian males × Pakistani females) produced their own female progeny (Table 2). Proportions of female progeny differed between treatments (t-test, t ₅₂ = 5.25, P < 0.001), and again were higher in the treatment that included Californian males. Indeed, because males develop from unfertilized eggs in the haplodiploid system, the treatment that initially included Pakistani males and Californian females would have produced Californian male progeny in the first generation (Table 2).