Predatory and Parasitic Insects Associated with Urophora cardui L. (Diptera: Tephritidae) Galls on Canada Thistle, Cirsium arvense L. (Asterales, Asteraceae) in North Dakota

Simple Summary Plant gall-inducing insects can be attacked by a diverse and unique assemblage of predatory and parasitic insects. We surveyed the insect fauna associated with galls induced by the fly Urophora cardui L. on Canada thistle, a perennial weed. We found that the thistle gall fly and its primary parasitoid wasp, Pteromalus elevatus (Walker), were widespread throughout the study area. In addition, we recovered a checkered beetle, Isohydnocera tabida (LeConte), from a subset of galls, which is a new host record and provides vital information on the little-known immatures of this predatory species. This study adds to the taxonomic literature of gall-inhabiting insect species and I. tabida. Abstract We surveyed the insect fauna associated with Urophora cardui L. (Diptera: Tephritidae) galls on Canada thistle, Cirsium arvense L. (Asterales, Asteraceae), in parts of the northern Great Plains, U.S., by field-collecting galls and rearing or dissecting out the insects. We also examined the relationships between gall biomass and insect density and biomass. Urophora cardui were widespread, and the gall biomass was positively correlated with fly density and fly biomass. We recovered Isohydnocera tabida (LeConte) (Coleoptera: Cleridae) from galls in two counties, which represents a new host record and provides vital information on the little-known immatures of this predatory species. Pteromalus elevatus (Walker) (Hymenoptera: Pteromalidae) was the dominant parasitoid that emerged from the U. cardui galls. Individual galls typically only had one insect species, and occasionally both U. cardui and P. elevatus were present, but it was rare for other insects to be present in galls housing I. tabida. This study adds to the taxonomic literature of gall-inhabiting insect species and provides new information on the predators of U. cardui, specifically a little-known clerid beetle species.


Introduction
Urophora cardui L. (Diptera: Tephritidae) is a biological control agent of the noxious weed Canada thistle, Cirsium arvense L. (Asterales, Asteraceae) native to Europe, and this insect induces stem galls. Canada thistle is the primary host plant for U. cardui, although this insect is associated with two other thistle species, C. creticum (Lam.) and C. setosum (Willd.) Besser ex M. Bieb. [1][2][3]. A classical biocontrol program against C. arvense began in North America in the mid-1900s, and between 1963 and 1996, six insect species were deliberately released or redistributed in order to suppress C. arvense populations [4,5]. Urophora cardui was first released in Canada in 1974 [6] and in the U.S. in 1977 in California [7]. It has since been released in multiple other states and is now widely distributed within Europe and across the U.S. and southern Canada [7][8][9][10].
Little is known about the insect species that occur within U. cardui galls in North Dakota, a north-central U.S. state, although the parasitoid community from nearby states has been studied [10]. Therefore, while monitoring the field releases of Canada thistle biocontrol agents [22,23], we collected additional U. cardui galls from Canada thistle in order to quantify the identity and density of insects associated with U. cardui galls and examine the relationships between gall size and insect density and biomass. We also provide the first report of a previously unknown predaceous beetle within the thistle galls, Isohydnocera tabida (LeConte) (Coleoptera: Cleridae).

Site Description and Sampling Methods
We sampled U. cardui galls on Canada thistle in 2012 and 2013 from field sites in the eastern and western parts of North Dakota, U.S. (Table 1). These sites were part of a statewide Canada thistle biological control project, where in 2004, the North Dakota State Department of Agriculture released adult stem-mining weevils, Hadroplontus (Ceutorhynchus) litura Fabricius (Coleoptera: Curculionidae) (see Prischmann-Voldseth et al. [22,23] for details). In 2009, we established permanent transects at each site to enable the systematic collection of samples, with three 15.9 m-long transects extending out from the original weevil release site at the linear edge of the thistle patch, which was created by anthropogenic activity (e.g., a road or field edge), creating four inter-transect spaces [22]. In 2012, we collected U. cardui galls from May 2 to May 10, prior to the emergence of adult flies, in two to four of the inter-transect areas at each site (i.e., replicate) listed in Table 1. In each replicate, one person harvested galls for five minutes by clipping plant stems approximately 0.5-1.5 cm above and below the gall. Galls were placed in labeled self-sealing plastic bags and transported back to the lab in plastic coolers with ice packs. In 2013, we collected galls in the first week of May and focused on sites in Bowman and Stark counties where we had found I. tabida the previous year.
To assess insect emergence, within 48 h of sampling, we placed the galls in individual plastic cups (29.6 mL Solo ® , Lake Forest, IL, USA) with muslin fabric covering a hole in the lid. The cups were maintained at room temperature (25 • C ± 5 • C) in the lab and monitored twice per week until an insect emerged, after which, the cups were checked every one to three days. In 2012, we monitored the cups until the end of June (11 d after the last insect emerged from a gall), after which, we considered emergence complete. In 2013, we monitored the galls daily after the first insect emergence event, and if no insects had emerged from a gall by the end of the observation period (i.e., 24 June 2013), the gall was dissected using a scalpel in order to obtain information on any remaining insects.
After collecting the insect data, we dried the galls to a constant mass in an oven at 49 • C and quantified their dry mass using a digital scale (Sartorius, model 1412, Westburg, NY, USA). We quantified the gall length (2a) and diameter along two axes at the widest point of the gall (2b,2c) using electric digital calipers (model 784EC, Sona Enterprises, Santa Fe Springs, CA, USA). We calculated the gall volume using the formula for an ellipsoid (V = 4/3πabc).
We placed the insects into labeled glass vials or plastic microtubes until they were identified, with the latter specimens stored in 70% ethanol only in 2013. We used White and Korneyev [1] to identify U. cardui and an online key [24] and other references to identify parasitoids [25,26]. We used Knull [27] and Optiz [28] to identify clerid beetles to species, which was verified by Dr. John Leavengood Jr. (john.m.leavengood@usda.gov). We deposited voucher specimens in the North Dakota State Insect Reference Collection, which is housed in the North Dakota State University Entomology department.

Statistical Analyses
Data were analyzed using JMP ® Pro 15.0.0 (SAS Institute Inc., Cary, NC, USA, 2019). Prior to analysis, we transformed the following variables to normalize their data distributions: SQRT (X + 0.5) for insect densities and log (X + 1) for gall and fly dry mass. We used a correlation analysis to test the associations among gall characteristics and U. cardui density and mass. The analyses included data from all viable galls, regardless of whether insects other than U. cardui were present, although correlations between gall characteristics and fly mass did not include data from galls where no U. cardui emerged. We used an ANOVA to examine the s of sample location on dependent variables, followed by a Tukey's Honest Significant Difference post hoc test for mean separation. A non-parametric Kruskal-Wallis rank sum test, followed by a Steel-Dwass post hoc test for mean separation, i.e., a non-parametric test equivalent to Tukey's Honest Significant Difference post hoc test [29,30], were used to analyze the differences in emerged wasp density between locations, with sites lacking wasps excluded from the analyses. Non-viable galls, or galls where no insects emerged or which, when dissected, lacked any living or dead insects, were not used in the analyses as these galls were likely more than a year old and all insects had previously emerged.

Results
Urophora cardui were widespread, and both males and females often emerged from the same gall (54% of galls with at least one fly in 2012), although only females (28%) and only males (17%) emerged from a sizeable percentage of galls. Across all sites, the mean (± SE) number of U. cardui adults that emerged from viable galls ranged from 1 to 16, with an average fly density of 3.26 ± 0.14 per gall in 2012 and 2.44 ± 0.12 in 2013 when using data from galls with at least one fly. In 2012, the location impacted the mean number of total U. cardui adults (both sexes) that emerged from the galls (df = 8262 , F = 6.89, p < 0.0001), as well as male (df = 8262 , F = 6.93, p < 0.0001), and female densities (df = 8262 , F = 2.80, p = 0.005; Table 2). The dry mass of all adult flies that emerged from the galls (i.e., total fly biomass per gall, 2012 data only) ranged from 0.8 to 43.2 mg and averaged 12.1 ± 0.6 mg (mean ± SE), with the highest fly biomass being from galls collected in Bowman and Griggs counties (df = 8262 , F = 10.17, p < 0.0001; Table 2). Using data from all the galls, the dry mass of an individual U. cardui adult averaged 3.6 ± 0.1 mg. When the data were limited to the galls where only a single U. cardui emerged, the average dry mass of one fly was 3.5 ± 0.2 mg for males (n = 31) and 3.4 ± 0.2 mg for females (n = 46). 1.83 ± 0.15 n/a n/a n/a 1.71 ± 0.23 n/a n/a Bo 211 196 23 160 65 2.74 ± 0.15 n/a n/a n/a 1.92 ± 0.17 n/a n/a 1 Sites are arranged from east to west; the three sites in Stark county were considered unique when they were established, but are adjacent to one another, and so we sampled them collectively in 2013. 2 C = galls collected, V = galls with at least one insect, +I = galls with at least one I. tabida, +U = galls with at least one U. cardui, +P = galls with at least one P. elevatus. 3 Data from galls with at least one U. cardui, M = males, F = females. 4 Data from galls with at least one P. elevatus, M = males, F = females. 5 Insects only recovered from one gall, so there's no variance, within a column values sharing the same letter are not significantly different at alpha = 0.05.
During the gall dissections, we found the larvae, pupae, and adults of a clerid beetle species, I. tabida, inside U. cardui galls, as well as I. tabida pupae within U. cardui puparial cases (Figure 2), which appears to be a new host record. Isohydnocera tabida were only recovered from galls collected in the southwest corner of the state. When I. tabida were found, typically only one specimen was present per gall, although in 2013, two specimens were recovered from one gall, and in another case, three adults emerged from a single gall. In 2012, 11 specimens were recovered, and in 2013, when we focused our sampling efforts in that geographical area, we recovered 43 specimens from Stark and Bowman counties ( Table 2). Insects 2022, 13, x FOR PEER REVIEW 5 of 13 During the gall dissections, we found the larvae, pupae, and adults of a clerid beetle species, I. tabida, inside U. cardui galls, as well as I. tabida pupae within U. cardui puparial cases (Figure 2), which appears to be a new host record. Isohydnocera tabida were only recovered from galls collected in the southwest corner of the state. When I. tabida were found, typically only one specimen was present per gall, although in 2013, two specimens were recovered from one gall, and in another case, three adults emerged from a single gall. In 2012, 11 specimens were recovered, and in 2013, when we focused our sampling efforts in that geographical area, we recovered 43 specimens from Stark and Bowman counties ( Table 2). Pteromalus elevatus (Walker) (Hymenoptera: Pteromalidae) (syn. Habrocytus elevatus) was the most common parasitic wasp found (Figure 3), and male and female P. elevatus co-occurred within galls (Table 2). A few other parasitic wasp species were collected in the Ichneumonidae, Braconidae, and Chalcidoidea families, including what we believe was a Schizoprymnus sp. and a wingless wasp that was likely a Eupelmus (Macroneura) species, although the specimens could not be definitively identified and were extremely rare compared to P. elevatus. The number of individual parasitic wasps that emerged from a single gall ranged from 1 to 7, and when focusing on galls where at least one P. elevatus emerged, the density of adults per gall was similar among all sites (both sexes, Pearson's χ 2 = 7.17, df = 7, P = 0.412; males, Pearson's χ 2 = 6.93, df = 7, P = 0.436; females, Pearson's χ 2 = 0.83, df = 6, P = 0.991).

Discussion
Here, we report the first occurrence of Isohydnocera tabida reared from U. cardui galls on Canada thistle. Our observation of I. tabida immatures inside the galls and a pupa inside the puparium of U. cardui suggests that the clerids are preying on the gallflies. Clerid beetles, also known as checkered beetles, are typically brightly colored and patterned as adults and prey on beetles and other insects inhabiting wood, although some have been reported to eat pollen, grasshopper eggs, and immature Lepidoptera and Hymenoptera [28,[31][32][33], while Morrill et al. [34] have reported adult and larval Phyllobaenus dubius (Wolcott) preying on wheat stem sawfly larvae, Cephus cinctus Norton (Hymenoptera: Cephidae), within wheat stems.
Isohydnocera was previously within the genus Hydnocera, with insects in the latter genus having been reported to feed on cryptic insect larvae in multiple orders, including woodborers, stem-miners, and gall-makers [31,32,35]. Within Isohydnocera, I. curtipennis is one of the best-known species. In addition to consuming wood-boring insects [36], I. curtipennis adults were reared from Euura salicis-nodus Walsh (Hymenoptera: Tenthredinidae) galls on willow (Salicales: Salicaceae) trees in Fort Collins, Colorado [37] and Gnorimoschema gallaesolidaginis Riley (Lepidoptera: Gelechiidae) galls on goldenrod (Asterales: Asteraceae, Solidago sp.) plants [32]. In the latter case, Sabrosky [32] concluded that an I. curtipennis larva had parasitized a single G. gallaesolidaginis pupa from a plant collected near Medora, Kansas, and noted that the adult beetle emerged from the gall in May through a single round exit hole.
De Smet-Moens [38] reported occasionally collecting an unidentified clerid beetle larva in the genus Phyllobaenus or Isohydnocera feeding on Orellia ruficauda (F.) (Diptera: Tephritidae) on Canada thistle in southern Montana. Price [10]  Isohydnocera tabida has been reported in multiple locations and habitats in Canada and the central and eastern U.S., and appears to have a widespread distribution. Searching for "Isohydnocera tabida" in SCAN (Symbiota Collections of Arthropods Network, scanbugs.org) [39], which has specimen data from over 225 North American arthropod specimen and observation collections, resulted in 10 images of adults and 172 specimen records from two Canadian provinces (Québec and Ontario), and sixteen U.S. states. We found similar results when we searched the Global Biodiversity Information Facility open access database [40,41]. The NDSIRC (North Dakota State Insect Reference Collection, NDSU, Fargo ND) has 15 specimens of I. tabida adults from North Dakota, South Dakota, and South Carolina, including some islands in the latter state (Table 3). With regard to the literature records, Chapin [42] indicates that I. tabida is found along the eastern coast of the U.S., into Canada and west to Kansas and Nebraska, and Wolcott [43] reports the presence of I. tabida in Ontario, Canada and five U.S. states: Alaska, Illinois, Indiana, Kansas, and Wisconsin. Isohydnocera tabida has been collected from the salt marshes in North Carolina [44] and from three of the Boston Harbor Islands (i.e., Great Brewster, Middle Brewster, and Ragged) off the Massachusetts coast [45]. Specimen records appear to be for adults, which have been collected from April (Texas) to August (Massachusetts, Wisconsin), and are most frequently collected in June and July. The records indicate that adults were collected by sweeping or beating vegetation (e.g., ferns, grasses, weeds, and herbaceous plants) in meadows or in fragmented and disturbed landscapes (NDSIRC records, [27,39,42,46]). Mawdsley [47] suggested that I. tabida was likely found in prairie habitats, and in Iowa, multiple adults were collected from lead plants (Fabales: Fabaceae, Amorpha canescens Pursh) at Rahel's Prairie within the Hitchcock Nature Center [48]. In Wisconsin, adults were knocked off black locust branches (Fabales: Fabaceae, Robinia pseudoacacia) or hand collected from a walnut tree (Fagales: Juglandaceae, Juglans sp.) [46]. One adult was collected from milkweed (Gentianales: Apocynaceae, Asclepias syriaca L.) in Bowling Green, Ohio [49]. Adults have also been collected using Malaise and light traps [39].
Even though multiple I. tabida adults have been collected, there are virtually no records or information on the life history of the immatures. Böving and Champlain [31] and Böving and Craighead [50] present identification keys for subfamily or genus and illustrations of larval Cleridae, although information on I. tabida is lacking. Knull [27] indicated that I. tabida was found in plant stems infested with tumbling flower beetle larvae (Coleoptera: Mordellidae), which likely refers to the reports from East Falls Church, Virginia, where pupae were found in annual plant stems infested with mordellid larvae [31] and from Staten Island, New York, where Leng and Davis [51] indicated that I. tabida preyed on mordellid larvae in the stems of annual plants in June. Our discovery that I. tabida larvae and pupae occur within U. cardui galls on Canada thistle could explain both the lack of information on immatures and the widespread distribution of adults, as this weed has a cosmopolitan distribution. However, the clerid could also be preying on parasitoid wasps within the galls. Larvae of a related species (Phyllobaenus) have been observed consuming Microbracon (Hymenoptera: Braconidae) and Eurytoma pupae within the galls of the moth Gnorimoschema baccharisella Busck (Lepidoptera: Gelechiidae) on coyote brush plants (Asteraceae: Baccharis pilularis) [52].
Information on U. cardui predators is scarce, although some moth larvae, e.g., Metzneria spp. (Lepidoptera: Gelechiidae), Homoeosoma spp. (Lepidoptera: Pyralidae), and Eucosma spp. (Lepidoptera: Tortricidae) have been reported to prey on larvae of other Urophora species [17]. Insect-induced plant galls are also vulnerable to predation by vertebrates such as mice and birds [53], with higher predation rates on large galls [54], although Freese and Zwölfer [14] indicated that predation of U. cardui galls does not depend on gall size. Parasitoids are considered to be the primary cause of mortality for larval U. cardui [9,17].
In the Palearctic region, Eurytoma robusta and E. serratulae are the most common hymenopteran parasitoids of U. cardui larvae, and the two species can co-occur within a single gall [3,21]. In contrast, the ectoparasitoid P. elevatus was the dominant parasitoid species in our study. Pteromalus elevatus primarily parasitizes tephritid flies that feed on plants within the Asteraceae family [55][56][57]. This species is associated with U. cardui throughout Europe, where it is generally considered a minor parasitoid [11,58,59]. In southern Finland, P. elevatus was present at 25 out of 75 sites [58], and in northern Germany and Denmark, Johannesen and Seitz [18] found P. elevatus parasitoids at 13 of the 20 sites they sampled, and its rate of parasitizing U. cardui ranged up to 11%. However, the general conclusion is that in the Palearctic region, P. elevatus does not have a major impact on U. cardui populations [9,17,21].
In the Nearctic region, U. cardui were released in Canada to control Canada thistle starting in 1974 [6], and by 1978-1979, 3% of fly larvae were parasitized by P. elevatus. Similar to our observations, Peschken and Derby [20] found P. elevatus was the dominant parasitoid of U. cardui in New Brunswick, Canada, and did not find any E. robusta or E. serratulae. They indicated that P. elevatus was accidentally introduced to North America (Newfoundland, Canada), and suggested that the levels were too low to negatively impact U. cardui [20]. Parasitoids associated with U. cardui were surveyed in the western U.S. (Washington, Oregon, Idaho = 13 sites; Montana = 6 sites), and the Midwest (North Dakota = 1 site; South Dakota = 15 sites) [10]. Price [10] reared 11 parasitoid wasps from U. cardui galls, and P. elevatus was the dominant parasitoid associated with U. cardui at all their study sites, although they reported a low rate of parasitism (i.e., 11.4% of galls) and concluded that parasitoids did not limit the control efficacy of U. cardui. A previous survey of insects associated with Canada thistle in Montana did not list U. cardui, and the authors indicated that this species had not established after its initial release in 1978 [38,60].
We reared a small number of uncommon parasitic wasp species from U. cardui galls, including an unidentified species of Schizoprymnus. Price (2014) rarely recovered Schizoprymnus from U. cardui galls and did not find any in North Dakota [10]. Wasps in this genus typically parasitize beetle larvae but are also known from dipteran and lepidopteran hosts [26], so they may have been parasitizing either U. cardui or I. tabida, although we did not find the latter and this wasp species at the same location.
The average density of U. cardui per gall at our study sites was similar to that in other areas, including the Midwest and western U.S. (1.16 ± 0.14 SE adult U. cardui emerged per gall) [10], Denmark and northern Germany (3.6 ± 1.7 SD cells per gall, ranging from 1 to 12) [18], eastern France (3.86 ± 2.43 cells per gall, ranging from 1 to 23) [14], central Europe (3.38 ± 0.66 cells per gall with a range of means from 2.29 to 4.91) [14], and Canada (5.8 ± 3.7 SD cells per gall, ranging from 1 to 17) [16]. We found that gall size was positively related to U. cardui density, and others have also documented positive correlations between gall size and the number of larval chambers within the gall and the number of U. cardui larvae and emerged adults [6,10,16,18,61]. Freese and Zwölfer [14] found that the more U. cardui larvae within a gall, the lower the larval mortality, the higher the larval mass, and the higher the fitness of the resulting adults.
Exposure to air is believed to trigger U. cardui pupation [8,11], which typically occurs in the spring when galls become wet and the callus tissue blocking the larval tunnels degrades [13]. Because moisture facilitates pupation and adult emergence [13], and we placed detached galls in plastic cups without supplemental moisture, our insect densities are likely conservative, which is supported by our observations of low levels of live and dead larvae, pupae, and adults in dissected galls. We found little to no evidence of U. cardui death due to pathogens.
In conclusion, P. elevatus was the primary parasitoid of U. cardui on Canada thistle across North Dakota, with certain sites having higher parasitoid densities than other areas. Isohydnocera tabida appeared to be geographically restricted to the SW corner of the state, and our observations suggest that this species preys on U. cardui and that their immatures use the galls as protected overwintering sites, as has been suggested for other clerid beetle species [35]. Isohydnocera tabida may also opportunistically feed on parasitoids, as evidenced by the fact that clerid beetles rarely co-occurred with other insects within a gall.