Use of Flowering Plants to Enhance Parasitism and Predation Rates on Two Squash Bug Species Anasa tristis and Anasa armigera (Hemiptera: Coreidae)

A two-year study evaluated the effect of a flowering border of buckwheat Fagopyrum esculentum Moench on rates of egg parasitism, egg predation and adult parasitism on two squash bug species, Anasa tristis (DeGeer) and Anasa armigera Say, by comparing rates in squash fields with and without a flowering border. Furthermore, we evaluated whether there was an edge effect by comparing parasitism and predation rates in plots located in the corner of a squash field with plots located in the center of a squash field for fields with and without a flowering border. The egg parasitism rates were not affected by either treatment (flowering border or control) or plot location (edge or center). Anasa armigera egg masses only accounted for 4.3% of the total egg masses collected. The egg parasitism rates increased gradually throughout the season, peaking in the last week of August in 2017 at 45% for A. tristis egg masses. The most common egg parasitoid recovered was Gryon pennsylvanicum (Ashmead) followed by Ooencyrtus anasae (Ashmead). Adult parasitism was not affected by treatment, but A. tristis adult parasitism rates were higher in plots located on the edge of squash fields compared with plots located in the center of squash fields in 2016. Since adult parasitoid, Trichopoda pennipes (Fabricius) flies were observed visiting buckwheat flowers, future studies could explore the possibility that the flowering buckwheat may have a more impact on adult parasitism if there was a greater distance between fields with and without a flowering border.


Introduction
The loss of habitat and conversion of land to urbanization and intensive agriculture are major factors causing the loss of insect biodiversity [1]. Urban farms and gardens can increase biodiversity, ecosystem services, and climate resilience [2]. As urbanization increases, there is an increasing interest in the use of land for urban and peri-urban agriculture as a means of maintaining biodiversity and providing products to local communities [2]. The use of floral resources in urban farms can increase ecosystem services such as pollination and biological control [3,4]. Although floral resources have the potential to benefit farmers in both urban and rural areas, urban farmers are often growing organic produce for sale to local communities through farmers' markets [2]. Therefore, urban farmers have a vested interest in finding methods of biological control to manage pest populations without a reliance on pesticides.
Agricultural Research Center along with A. tristis [38]. In a recent laboratory study on the oviposition behavior of squash bugs, A. armigera was reared successfully on both yellow squash, Cucurbita pepo L. (Cucurbitales: Cucurbitaceae) and cucumber, Cucumis sativus L. (Cucurbitales: Cucurbitaceae). In paired choice tests, A. armigera was equally likely to oviposit on C. pepo or C. sativus, regardless of the species it was reared on [39].
There are two commonly occurring parasitoids of the common squash bug, A. tristis, the solitary egg parasitoid, Gryon pennsylvanicum (Ashmead) (Hymenoptera: Scelionidae) [40,41] and the adult parasitoid, Trichopoda pennipes (Fabricius) (Diptera: Tachinidae) [42,43]. In a two-year study in Maryland, G. pennsylvanicum accounted for over 99% of egg parasitism and the average rate of parasitoid emergence peaked on wild egg masses on the fifth week of July at 72.8% [44]. Reported rates of adult parasitism by T. pennipes were 12% in Virginia [34], 20% in Connecticut [43], and 20-30% in Kentucky [45]. Although there are no published field studies of parasitism on A. armigera, there was no difference in host acceptance by G. pennsylvanicum of egg masses of A. tristis and A. armigera in laboratory choice tests [38]. Predators of squash bug eggs include Araneae, Formicidae and Gryllidae [16]. Molecular analysis determined that Geocoridae, Coccinellidae, Nabidae, and web building and hunting spiders consumed squash bugs. However, it is not known which life stages they were preying upon [46]. In field tests in Kentucky, egg predation of A. tristis was low, ranging from 2.2% to 7.2% [45]. In Maryland, average predation of naturally occurring A. tristis eggs was 5.2% in a two-year study [44].
This study evaluated the use of floral resources in squash fields planted at an agricultural research center located in an urban area. This study compared rates of egg parasitism, egg predation, and adult parasitism of A. tristis and A. armigera in squash fields with a flowering buckwheat border to rates in control squash fields without a flowering border. We also evaluated whether the rates of egg parasitism, egg predation and adult parasitism varied between plots located at the edge of the field and plots located in the center of the field to determine if natural enemies attracted to flowering buckwheat were more abundant near the borders than in the center of the squash field.

Field Experiment
In 2016 and 2017, field tests were conducted in an urban location at the Beltsville Agricultural Research Center (BARC), Beltsville, MD. Squash fields were planted using Cucurbita pepo cv. Slick Pik YS26 where buckwheat, F. esculentum, was planted at a rate of approximately 45 kg/ha in a three-meter-wide border around the treated squash fields compared with control squash fields where a three-meter-wide border was mowed ( Figure 1). Each year, four squash fields were planted from seed at a rate of 3.36 kg/ha in 0.76 m rows 12-14 June 2016 and 28-30 June 2017. Planting was delayed in 2017 due to heavy rains. There were two blocks, at least 700 m apart, with a treated and a control field in each block. Squash fields were planted in different locations within the same two blocks each year. In both years, the treated and control fields were at least 50 m apart. Each field had two plots, a center plot and an edge plot. Each plot was 10 × 10 m with at least a 15-m distance between plots. There were approximately 400 squash plants in each plot. The center plot was located in the center of the field and the edge plot was located in a corner of the field so that two sides of the plot were adjacent to the border of the field. Squash fields were located adjacent to soybean, corn, and pepper fields.
Once the buckwheat plants started blooming, every squash plant in each experimental plot was checked once a week for the presence of naturally occurring squash bug egg masses and adult squash bugs of both species. In 2016, plots were checked weekly from 27 June-19 August. In 2017, plots were checked weekly from 17 July-26 August. The buckwheat plants continued to bloom throughout the experiment. All egg masses and adults found in the experimental plots were collected and brought into the laboratory. The number of eggs in each egg mass was counted. Egg masses were checked for signs of predation that leave visible evidence of damage on eggs. Egg masses were kept in Petri dishes in an incubator ( emergence were dissected and unemerged nymphs or adult parasitoids were counted. Parasitism rates were calculated by combining the number of emerged and unemerged parasitoids. The rates of nymph eclosion, parasitism, predation, or unknown egg mortality were calculated by dividing the number of eggs in each category by the total number of eggs in each egg mass. Once the buckwheat plants started blooming, every squash plant in each experimental plot was checked once a week for the presence of naturally occurring squash bug egg masses and adult squash bugs of both species. In 2016, plots were checked weekly from 27 June-19 August. In 2017, plots were checked weekly from 17 July-26 August. The buckwheat plants continued to bloom throughout the  . Each container had a hole (approximately 14 cm in diameter) in the lid covered with a fine mesh screen, a filter paper on the bottom, and two cotton plugged glass shell vials filled with water attached with a rubber band to prevent them from rolling. Adults were maintained on fresh cuttings of cucurbit plants grown in a greenhouse. After the death of an adult squash bug, the container was checked for the presence of a tachinid fly pupa. The pupa was placed in a Petri dish and kept in the incubator until the emergence of the adult tachinid fly. Tachinid parasitoids were kept in a vial of alcohol as voucher specimens.
Ethanol-preserved specimens were dehydrated through increasing concentrations of ethanol and transferred to hexamethyldisilazane (HMDS) before point-mounting. MWG identified Ooencyrtus anasae using a Leica M205C stereomicroscope with 10 × oculars and a Leica LED ring light source was used for point-mounted specimen observation [47]. All specimens were determined to genus by sight identification or using Gibson et al. [48]. Specimens were identified to species using relevant keys and primary literature [49]. All species identifications were corroborated through comparison with authoritatively identified specimens in the Smithsonian National Museum of Natural History.

Statistical Analysis
Comparisons of egg masses of the two squash bug species were conducted based only on data when both species were observed in 2017 from July week 3-August week 4. Treatment (flowering border or control), plot location (edge plots or center plots) and year (2016 or 2017) effects were examined based only on A. tristis data observed in both years from July week 3-August week 2. Nymph eclosion, eggs parasitized, and egg mortality rates were each modeled using a generalized linear mixed effects negative binomial ANOVA with offset log eggs per mass. Species comparisons were conducted within each treatment by plot location by week. Treatment by plot location comparisons was conducted within each year by week. Adult parasitism rates were evaluated for treatment, plot location, and treatment by plot location interaction effects within each squash bug species using a generalized linear mixed effects negative binomial ANOVA with offset log squash bug adults. Statistical analyses were accomplished using SAS v9.4 PROC GLIMMIX (SAS Institute Inc.: Cary, NC, USA) [50] and pdmix800.sas (SAS Institute: Cary, NC, USA) macro to obtain means comparisons letter groups [51].
Egg predation was too low to evaluate statistically. A total of 28 eggs (0.14%) were damaged by predators out of 19,640 total eggs collected. However, 21 of those 28 damaged eggs were collected in squash fields with a flowering border.
In a comparison of egg masses of the two squash bug species, there were no significant differences detected in the nymph eclosion rate, egg parasitism rate, or the rate of egg mortality from unknown causes between fields with and without a flowering border. The overall egg parasitism rate was a mean ± SE of 11.9 ± 0.8% for A. tristis eggs and 12.6 ± 3.7 for A. armigera egg masses.
Since the overall number of A. armigera egg masses collected was very low (4.3%), the analysis of treatment, plot location, week and year were performed separately for A. tristis egg masses. There were no significant differences in A. tristis egg parasitism for treatment, plot location, or year ( Figure 4).
Egg predation was too low to evaluate statistically. A total of 28 eggs (0.14%) were damaged by predators out of 19,640 total eggs collected. However, 21 of those 28 damaged eggs were collected in squash fields with a flowering border.
In a comparison of egg masses of the two squash bug species, there were no significant differences detected in the nymph eclosion rate, egg parasitism rate, or the rate of egg mortality from unknown causes between fields with and without a flowering border. The overall egg parasitism rate was a mean ± SE of 11.9 ± 0.8% for A. tristis eggs and 12.6 ± 3.7 for A. armigera egg masses.
Since the overall number of A. armigera egg masses collected was very low (4.3%), the analysis of treatment, plot location, week and year were performed separately for A. tristis egg masses. There were no significant differences in A. tristis egg parasitism for treatment, plot location, or year ( Figure  4). However, there were significant weekly differences in egg parasitism rates and there were significant interactions effects for egg parasitism (Table 1). There were also significant four-way interaction effects for treatment by plot location by year and by week for egg parasitism. In 2016 and 2017, egg parasitism rates were significantly different between edge plots in control and treatment fields in the third week of July (Figure 5a,b). In 2017, egg parasitism rates were significantly different between center and edge plots in control fields in the first week of August ( Figure 5B). In 2017, there were two incidences where only a single A. tristis egg mass was collected in a specific plot location during weekly collections. In July week 4, a single egg mass was collected in the edge plots in control fields and G. pennsylvanicum emerged from 100% of those eggs. In August week 1, a single egg mass However, there were significant weekly differences in egg parasitism rates and there were significant interactions effects for egg parasitism (Table 1). There were also significant four-way interaction effects for treatment by plot location by year and by week for egg parasitism. In 2016 and 2017, egg parasitism rates were significantly different between edge plots in control and treatment fields in the third week of July (Figure 5a,b). In 2017, egg parasitism rates were significantly different between center and edge plots in control fields in the first week of August ( Figure 5B). In 2017, there were two incidences where only a single A. tristis egg mass was collected in a specific plot location during weekly collections. In July week 4, a single egg mass was collected in the edge plots in control fields and G. pennsylvanicum emerged from 100% of those eggs. In August week 1, a single egg mass was collected from center plots in control fields and parasitoids emerged from 18.2% of those eggs ( Figure 5B). Nymph eclosion rates were significantly higher in squash fields with a flowering border with a mean ± SE of 79.8 ± 1.2% compared with 75.2 ± 1.6% in control fields and significantly higher in center plots (79.6 ± 1.5%) than in edge plots (77.2 ± 1.3%). There were significant weekly differences in nymph eclosion rates (Table 1). There were also interaction effects that significantly influenced nymph eclosion rates (Table 1). In a comparison of the interaction effects on nymph eclosion rates of treatment by plot location by year and by week, there were no significant differences in 2016 ( Figure 6A). Nymph eclosion rates were only significantly different in the combination (control field by edge plot by July week 4 by 2017) when a single egg mass was collected and there was 100% parasitism from that egg mass with 0% nymph eclosion ( Figure 6B).
Egg parasitism rates increased over the season in both 2016 and 2017. In 2016, egg parasitism rates peaked in the first week of August with a mean ±SE of 24.7 ± 3.2% ( Figure 7A). In 2017, egg parasitism rates peaked in the last week of August at 45.0 ± 7.2% ( Figure 7B). In 2016, weekly nymph eclosion rates were the lowest in the first week of August when egg parasitism rates reached their peak ( Figure 7A). In 2017, the mean (±SE) nymph eclosion rate was 49.5 ± 7.0% in the final week of August when egg parasitism rates reached 45.0% ( Figure 7B).  from that egg mass with 0% nymph eclosion ( Figure 6B). Egg parasitism rates increased over the season in both 2016 and 2017. In 2016, egg parasitism rates peaked in the first week of August with a mean ±SE of 24.7 ± 3.2% ( Figure 7A). In 2017, egg parasitism rates peaked in the last week of August at 45.0 ± 7.2% ( Figure 7B). In 2016, weekly nymph eclosion rates were the lowest in the first week of August when egg parasitism rates reached their peak ( Figure 7A). In 2017, the mean (±SE) nymph eclosion rate was 49.5 ± 7.0% in the final week of August when egg parasitism rates reached 45.0% ( Figure 7B).

Adult Parasitism
Trichopoda pennipes was the only adult parasitoid of squash bugs. There was no significant difference in the adult parasitism rate between fields with a flowering border and control fields (F = 1.1; df = 1, 7; p = 0.33). However, there was a significant difference between plot locations (F = 21.7; df = 1, 7; p = 0.002). There was a significant year by plot location effect (F = 9.8; df = 1, 7; p = 0.02). When the effects for plot location were evaluated separately for year by species effects, adult parasitism was only significantly higher in edge plots than in center plots for A. tristis adults in 2016 (F = 17.1; df = 1, 7; p = 0.004). Overall, adult parasitism was 24.6 ± 3.8%.

Adult Parasitism
Trichopoda pennipes was the only adult parasitoid of squash bugs. There was no significant difference in the adult parasitism rate between fields with a flowering border and control fields (F = 1.1; df = 1, 7; p = 0.33). However, there was a significant difference between plot locations (F = 21.7; df = 1, 7; p = 0.002). There was a significant year by plot location effect (F = 9.8; df = 1, 7; p = 0.02). When the effects for plot location were evaluated separately for year by species effects, adult parasitism was only significantly higher in edge plots than in center plots for A. tristis adults in 2016 (F = 17.1; df = 1, 7; p = 0.004). Overall, adult parasitism was 24.6 ± 3.8%.

Discussion
The use of a flowering border did not influence egg parasitism rates. These results are consistent with a study that found no difference in the number of adult G. pennsylvanicum in squash fields with floral resources compared to squash fields without floral resources [17]. Adult G. pennsylvanicum feed on squash leaf trichome exudates, which provide an energy source for foraging adults [52]. Since parasitoids have a source of nutrition from squash leaves, it may not be necessary for them to forage for nectar from flowering plants.
The egg parasitoid, G. pennsylvanicum, was the predominant parasitoid, accounting for 93.9% of eggs parasitized. In previous studies, G. pennsylvanicum accounted for >99% of egg parasitism but egg parasitism rates by G. pennsylvanicum were substantially higher in those studies [41,44]. Ooencyrtus anasae was identified as the second most common parasitoid recovered from squash bug eggs.
The identification of Ooencyrtus species is complicated by the lack of revisionary studies encompassing Nearctic taxa. Three species are known to attack A. tristis in the Nearctic: O. anasae, O. californicus Girault, and O. papilionis Ashmead. The latter two are quite different from O. anasae. Furthermore, O. johnsoni (Howard), another widespread species morphologically similar to O. anasae, was necessarily evaluated. However, it has not been recorded from A. tristis and has more metallic green reflections on the mesoscutum and metasoma, whereas O. anasae tends toward black coloration. A large integrative taxonomic study of the genus Ooencyrtus in Asia and North America is needed to conclusively differentiate between O. anasae and O. johnsoni [53].
In a previous study, egg parasitism rates by G. pennsylvanicum at BARC were much higher when squash was first planted in mid-May. In that case, the parasitism rate reached 29.6% by the second week of July, 50.5% by the third week, and 72.3% by the end of July [44]. In the current study, squash was planted in mid-late June and the parasitism rate was <1% in the second week in July and increased gradually, peaking in the end of August at 37.7%. After planting, squash bug egg masses increased rapidly. There were only eight squash bug egg masses collected during the first week of July compared with 83 egg masses in the second week and 263 egg masses in the third week. The results of the two studies at BARC suggest that G. pennsylvanicum populations primarily build up within the squash field through successive generations emerging from host egg masses and that it takes at least eight weeks after planting the squash for parasitoids to reach a population level high enough to parasitize 30% or more of squash bug eggs.
Egg predation levels were very low (<1%). However, this study only measured eggs that were damaged by predators, not eggs that were removed by predators. In a previous study, plots in squash fields were monitored daily and freshly laid squash bug egg masses were marked and collected 48 h later. The number of intact/damaged eggs per mass was counted at the time the egg masses were marked and recounted at the time of collection. Within the 48-h period, 5.9% of marked egg masses had at least one egg removed by predators and 1.0% of marked egg masses were removed entirely by predators. Predators damaged/removed 5.2% of the total number of eggs collected from marked egg masses [44]. Therefore, egg predation was underestimated because egg/egg mass removal was not accounted for. Since 75% of the damaged eggs were found in squash fields with a flowering border, it is possible that the flowering border enhanced egg predation. However, further studies measuring both egg damage and egg removal by predators are necessary to evaluate the effect of a flowering border on egg predation.
The use of a flowering border did not result in lower squash bug nymph eclosion rates. In fact, nymph survival was slightly higher in squash fields with a flowering border. Overall, squash bug nymph survival was high. Average nymph eclosion rates only dropped <60% in the last week of August in 2017 when egg parasitism rates reached their peak.
Although T. pennipes was observed visiting buckwheat flowers (Cornelius, personal observation), adult parasitism rates did not differ between squash fields with and without a flowering border. Since T. pennipes is highly mobile, distances between fields with and without a border probably need to be greater to influence adult parasitism rates. However, adult A. tristis squash bugs collected in edge plots were significantly more likely to be parasitized than those collected in center plots in 2016, suggesting that T. pennipes is more likely to search for hosts on the edge of squash fields. Squash fields were located adjacent to corn, soybean, and pepper fields. Trichopda pennipes also parasitizes other leaf-footed and stink bug species [11,54]. Therefore, parasitic flies were most likely foraging widely for hosts within the agricultural landscape. Further studies of the host searching behavior of T. pennipes are needed to develop more effective methods of enhancing adult parasitism of squash bugs.
Anasa armigera accounted for only 4.3% of egg masses, but 56.1% of the adults collected overall. Numbers of A. armigera egg masses collected from 2014-2018 have been consistently low, ranging from 2-4% of the total squash bug egg masses collected (Cornelius, unpublished). Also, nymphs of A. armigera are much less commonly seen in squash plots than A. tristis nymphs (Cornelius, personal observation). Although the reasons for the disparity in collections of adults and egg masses of A. armigera are unknown, it seems likely that differences in adult behavior could play a role. We collected any adult squash bugs that were found on squash plants when plants were searched for egg masses. We did not search for adult squash bugs hiding on the ground in the soil or under plant debris. If A. armigera adults are more likely to spend time on the plant compared with A. tristis adults, A. armigera adults were probably oversampled in relation to A. tristis adults. Most of the A. armigera adults were collected from 7-25 August 2017. In some incidences, multiple A. armigera adults were collected within one meter of each other, possibly due to aggregation behavior of the adults. Alternatively, adult A. armigera could be moving into squash fields late in the season. Currently, there are no field studies comparing the behavior of the two squash bug species. Further studies on the foraging and aggregation behavior of adult A. armigera squash bugs may shed light on the discrepancy between field collections of A. armigera adults and egg masses.

Conclusions
The proportion of people living in urban areas is projected to increase to 70% by 2050 [55]. Therefore, it is important to find ways to increase biodiversity and ecosystem services within urban habitats. As the rate of urbanization increases, urban agriculture will become increasingly more important for food security and for maintaining biodiversity [2,55]. The use of floral resources on urban farms is an important component of conserving pollinators and other beneficial insects. However, there is conflicting evidence concerning the effectiveness of floral resources for improving the biological control of pest populations.
Flowering buckwheat did not increase parasitism rates of squash bugs in squash fields. The egg parasitoid, G. pennsylvanicum, was equally distributed in center and edge plots in fields with and without a flowering border. The egg parasitism rates increased gradually over the season, as successive generations of parasitoids emerged from host eggs within the squash fields. Overall, adult squash bugs collected in edge plots were more likely to be parasitized than those collected in center plots in fields with and without a flowering border. This difference was due to increased parasitism rates of A. tristis adults in edge plots in 2016.
Pollinators and other beneficial insects were frequently observed visiting the flowering buckwheat (Cornelius, personal observation). Therefore, the use of flowering buckwheat as a border around squash fields may have a positive effect on ecosystem services despite not augmenting parasitism rates against squash bugs. However, flowering plants could also attract herbivorous pests. Further research could examine the effects of flowering buckwheat on populations of other beneficial insects and herbivorous pests. It is also possible that other species of flowering plants or that planting strips of flowering plants between rows of squash could have a greater effect on biological control of squash bugs than using flowering plants as a border. Although floral resources play an important role in increasing populations of pollinators and other beneficial insects, their role in enhancing parasitoid populations in squash fields requires further investigation.
Author Contributions: M.L.C. was responsible for the conception and design of the project and writing the first draft of the manuscript. M.W.G. was responsible for the identification of parasitic wasps and for writing revisions to the final draft of the manuscript. B.T.V. was responsible for the statistical analysis of the data. All authors have read and approved the manuscript.
Funding: This research received no external funding.