Numerous stink bug species (Hemiptera: Pentatomidae), including Chinavia hilaris (Say), Euschistus servus (Say), E. tristigmus (Say), and Nezara viridula (L.), cause economic injury to many agronomic crops (McPherson & McPherson 2000). Stink bugs feed on developing cotton seeds and lint, which can cause shedding of young bolls, yellowing of lint, yield reduction, and transmission of the bacterial pathogen Pantoea agglomerans (Enterobactacteriaceae) that can damage seed and lint (Barbour et al. 1990; Medrano et al. 2009).

The plant-feeding habits of stink bug pests compels them to move within and between closely associated non-crop and crop habitats across farmscapes in response to changing food resources. Mimosa, Albizia julibrissin Durazz (Fabaceae), was introduced in the mid-1700s as an ornamental in the US. This exotic tree commonly grows in thickets along roadsides and in woodlands adjacent to agricultural crops across the southeastern USA. The fruit (i.e., pod) of this legume is 8 to 20 cm long and, in general, yields 5 to 16 seeds, each about 6 to 12 mm in length (Meyer 2010). Chinavia hilaris previously has been reported on mimosa in South Carolina, USA (Jones & Sullivan 1982). In southwest Georgia, the capture of C. hilaris adults and nymphs in stink bug pheromone-baited traps near mimosa in field borders adjacent to crops, and the spatial distribution of C. hilaris in these crops, indicated that this tree was a source of this stink bug dispersing into crops, primarily to cotton (Cottrell & Tillman 2015). This study was conducted to investigate the seasonal abundance and feeding habits of stink bug species in mimosa, and to provide estimates of parasitism of naturally occurring stink bug eggs on this tree.

Materials and Methods

This field study was conducted from 2013 through 2018. Over the 6-yr study, uncultivated mimosa trees in woodlands adjacent to commercial cotton fields were sampled for stink bugs at 4 sites: Grainbin (31.5656°N, 83.2979°W), Redbarn (31.5543°N, 83.3132°W), Palm (31.5369°N, 83.3321°W), and Starr (31.5719°N, 83.2969°W), in Irwin County, Georgia, USA. The number of sites and trees per site varied among yr because some of the trees were accidentally killed or hurt by growers, or did not produce fruit during a season due to drought conditions. Also, additional sites were included as mimosa trees were detected at other sites. With only 1 exception, mimosa trees were sampled weekly. In 2013, 4 trees were sampled for 5 wk at ‘Grainbin’ from 25 Jul to 22 Aug. In 2014, 3 trees were examined for 8 wk at ‘Grainbin’ from 10 Jul to 28 Aug. In 2015, 2 trees at ‘Starr’, 5 at ‘Grainbin’, and 7 at ‘Palm’ were sampled for 5 wk from 15 Jul to 12 Aug. In 2016, 9 trees at ‘Palm’ were sampled for 2 wk, on 28 Jul and 4 Aug. In 2017, 2 trees at ‘Starr’, 3 at ‘Grainbin’, 4 at ‘Redbarn’, and 5 at ‘Palm’ were examined for 7 wk on 29 Jun and from 20 Jul to 24 Aug. In 2018, 2 trees at ‘Starr’, 4 at ‘Redbarn’, and 6 at ‘Palm’ were sampled for 13 wk from 7 Jun to 30 Aug. For each mimosa sample, a whole plant, including foliage and fruiting structures, was examined for the presence of stink bugs. When necessary, a 1.5-m long cattle show stick with a small hook at the end was used to gently pull down a branch to look for insects. Species and developmental stages of stink bugs were recorded in the field using a HP iPAQ pocket personal computer (Hewlett-Packard Co., Palo Alto, California, USA). At the Grainbin site, feeding (i.e., proboscis inserted into the food source) by stink bug nymphs and adults was recorded by the plant part (i.e., leaf, stem, or fruit) each individual fed on for each tree for each sampling date in 2013 and 2014.

Naturally occurring stink bug egg masses detected in the field were brought into the laboratory and held in a walk-in environmental chamber (27 °C, 60% relative humidity, 14:10 h [L:D] photoperiod) for emergence of adult parasitoids. Trissolcus species (Hymenoptera: Scelionidae) were identified using a key to Nearctic species of Trissolcus (Talamas et al. 2015). Females of Anastatus (Hymenoptera: Eupelmidae) were identified using a key to North American species of this genus (Burks 1967). There is no key to Anastatus males. Anastatus reduvii (Howard) and A. mirabilis (Walsh & Riley) are the only 2 Eupelmidae parasitoids known to emerge from stink bug eggs in southwest Georgia, and only 1 species emerges from a single egg mass (Tillman 2016). Therefore, Anastatus males emerging with Anastatus females were assumed to be the same species. Predation of eggs was assessed as described in Tillman (2011). Voucher specimens of all insects were deposited in the USDA, ARS, Crop Protection & Management Research Laboratory in Tifton, Georgia.

All data were analyzed using SAS statistical software (SAS Institute 2010). Chi-square analyses were used to compare frequencies of parasitoid species parasitizing C. hilaris egg masses on mimosa, frequencies of plant parts fed on by nymphs of this stink bug, and frequencies on location of egg masses on mimosa (PROC FREQ). Seasonal means for the number of C. hilaris egg masses, nymphs, and adults per tree were calculated and graphed for 2013, 2014, 2015, 2017, and 2018 (PROC MEANS). Only data on stink bug species, and parasitism and predation of C. hilaris egg masses, are presented for the 2016 data set.

Results

Over the 6 yr study, 7 plant-feeding stink bug species were detected on mimosa. Chinavia hilaris was the predominant stink bug species (63.5%) (c² = 1617.7; df = 5; P < 0.001) followed by N. viridula (12.9%), E. tristigmus (10.3%), E. servus (9.7%), Thyanta custator custator (F.) (3.0%), E. obscurus (Palisot) (0.4%), and Loxa flavicollis (Drury) (0.2%). Chinavia hilaris, E. servus, and E. tristigmus were consistently present on mimosa over each yr of the study. However, E. obscurus and T. c. custator were detected on mimosa only for 3 yr and N. viridula for only 2 yr. Two females of L. flavicollis were found on mimosa in 2018. Only C. hilaris developed from nymphs to adults on mimosa. For the remaining stink bug species, only adults were detected, generally feeding on fruit.

Populations of C. hilaris were present on mimosa in Jul and Aug (Fig. 1). Observations of feeding by C. hilaris on mimosa in 2013 and 2014 revealed that nymphs (second through fifth instars) fed more frequently on fruit (98.5%) than on leaves (1.5%) (c² = 121.1; df = 1; P < 0.001). Adults (n = 41) fed only on fruit, and females laid egg masses more often on fruit (77.4%) than on leaves (22.6%) (c² = 9.3; df =1; P < 0.002). No C. hilaris were observed feeding on stems of mimosa. On 29 Jun 2017, only small seeds were present in small pods, and no C. hilaris were detected on trees. In 2018, trees flowered from early to mid-Jun. Small seeds were present in small pods from mid- to late Jun. No C. hilaris were detected on mimosa trees until pods with medium-sized seeds were sampled on 5 Jul. Only early instars (i.e., first, second, and third instars) were present on mimosa on 10 Jul 2014, and 5 Jul and 12 Jul 2018. In addition, some egg masses were detected on mimosa on 15 Jul 2015 and 12 Jul 2018. Altogether, these results strongly indicate that C. hilaris occurred and developed on fruiting mimosa. In 2017 and 2018, adult numbers decreased as pods with large seeds began senescing (i.e., drying out and turning brown). Numbers of nymphs (principally fourth and fifth instars late season), though, increased around mid-Aug, decreasing when pods senesced. A similar pattern occurred for C. hilaris numbers in Aug 2014 and 2015. Numbers of C. hilaris nymphs were relatively low in Aug in 2013 compared to the other yr, perhaps due to the drought conditions.

Only 2 species of parasitoids parasitized naturally occurring egg masses of C. hilaris. Trissolcus edessae Fouts (Hymenoptera: Scelionidae) was the most prevalent parasitoid of C. hilaris eggs on mimosa (87.8%) (c² = 136.1; df = 1; P < 0.001). Anastatus reduvii (Howard) (Hymenoptera: Eupelmidae) also emerged from eggs (12.2%) of this stink bug. Overall, 46.6% of the eggs were parasitized and 10.5% were preyed upon.

Discussion

Chinavia hilaris fed and successfully developed to adults on mimosa fruit (i.e., pods), confirming mimosa as a reproductive host plant for this stink bug. However, mimosa was not a reproductive host plant for the other stink bug species detected on this tree, and served solely as a source of food for adults. Similar to the results of the current study, populations of C. hilaris were present on mimosa from mid-Jul through Aug in South Carolina (Jones & Sullivan 1982). In the current study, adults of C. hilaris decreased in number on mimosa as nymphs began to peak early to mid-Aug, indicating that new adults developing from nymphs on the tree dispersed from this woodland food source to seek a new host plant. Timing of senescence of mimosa and likely dispersal of C. hilaris adults from this host plant coincides with the presence of fruit (i.e., bolls) on cotton, suggesting that mimosa could be a source of this stink bug in this crop. Indeed, the capture of C. hilaris adults and nymphs in pheromone-baited traps near mimosa, and subsequent presence of C. hilaris in crops near mimosa in an earlier study indicated that this tree was a source of this stink bug dispersing into cotton (Cottrell & Tillman 2015). Mark-recapture studies in the same woodland habitats of the current study demonstrated that elderberry (Sambucus nigra subsp. canadensis [L.] R. Bolli), a non-crop host plant of C. hilaris, was a source of this stink bug moving into cotton (Tillman & Cottrell 2016). Similarly, damage to apple by C. hilaris was greatest near woodlands (Mundinger & Chapman 1932), and infestations of this stink bug in soybean were consistently found on border rows next to woodlands (Miner 1966), suggesting that host plants in woodlands were sources of C. hilaris in apple and soybean.

Fig. 1.

Mean number of C. hilaris egg masses, nymphs, and adults on mimosa over time in 2013, 2014, 2015, 2017, and 2018, and percentage of senesced pods over time in mimosa in 2017 and 2018. G = green pod; S = senesced pod; s = small-sized seeds in pods; M = medium-sized seeds in pods; L = large seeds in pods.

Based on the results of the current study and previous studies, overall percent parasitism of C. hilaris eggs ranges from 16 to 49% on various host plants (Yeargan 1979; Orr et al. 1986; Jones et al. 1996; Koppel et al. 2009; Tillman & Cottrell 2016). In an earlier study, diversity of egg parasitoids emerging from C. hilaris was greater, and percent parasitism was higher in woodland habitats (41%) compared with those observed in crops (21%) (Tillman 2016). Four species of parasitoids, T. edessae, A. reduvii, A. mirabilis (Walsh & Riley), and Ooencyrtus sp. parasitized egg masses of C. hilaris in wooded areas, but only T. edessae parasitized eggs of this stink bug in crops. In the current study, both T. edessae and A. reduvii parasitized eggs of C. hilaris in mimosa. Trissolcus edessae was the principal parasitoid emerging from C. hilaris egg masses on Sesbania punicea (Cav.) Benth, another non-crop host plant of this stink bug in wetland habitats (Tillman 2015). In elderberry, these 2 parasitoid species, as well as A. mirabilis, parasitized C. hilaris in woodlands (Tillman & Cottrell 2016).

Conservation of T. edessae and A. reduvii in woodland habitats has the potential to enhance biological control of C. hilaris in agricultural ecosystems. The importance of nectar provision on parasitoid fitness has been demonstrated for various hymenopteran parasitoid species (Berndt & Wratten 2005; Araj et al. 2006). Incorporating buckwheat near soybean increased parasitism of E. servus egg masses by Telenomus podisi Ashmead in adjacent cotton (Tillman et al. 2015). Nectar provision along field edges could provide a food source to parasitoids in their woodland habitats and as they disperse from woodlands into crops. Anastatus species were the most prevalent parasitoids emerging from naturally occurring eggs of Halyomorpha halys (Stål) (Hemiptera: Pentatomidae) in ornamental nurseries in Maryland, USA (Jones et al. 2014). Trissolcus edessae and A. reduvii are 2 prevalent egg parasitoids of H. halys in the southeastern US (Tillman unpublished data). Providing nectar to T. edessae and Anastatus species near woodlands is one of many management strategies available for managing C. hilaris and, potentially, H. halys.