During the 2018 brown marmorated stink bug monitoring campaign in fruit orchards in the Region of Friuli-Venezia-Giulia (northeastern Italy), personnel of ERSA on August 7, and jointly with personnel of CREA on August 8, noted the presence of darkly colored H. halys egg masses, which is an indicator of parasitism (see Results for site descriptions). During the surveys, egg masses of other stink bug species were also collected when found. A portion of the egg masses collected in the field were reared in climatic chambers (26 °C, 65%RH, 16:8 L:D) until adult parasitoids emerged. For further study, adults were kept alive in glass tubes and provided with pure honey droplets as food. The remaining field collected egg masses were reared in a laboratory room in Petri dishes until parasitoids emerged and specimens were stored in ethanol for further studies. All emerged specimens were counted, identified to species and sexed.

During routine research activities of CREA personnel on July 27 and August 3, 2018, several H. halys egg masses with dark coloration were observed on Acer campestris L. trees in a parking lot near Lodi, a site close to the town of Milan (see Results for site description). All collected egg masses were reared in climatic chambers as described previously, and adult parasitoids were kept alive in rearing tubes with pure honey for further studies. Emerged specimens were counted, identified to species and sexed.

Reared specimens of Trissolcus were identified using the key to Palearctic Trissolcus provided in Talamas et al. (2017). Image Z-stacks were taken with a Canon EOS 80D camera attached to an Olympus BX51 compound scope and a Macropod imaging system from Macroscopic Solutions and were rendered using Helicon Focus (Helisoft). Voucher specimens are deposited in the Florida State Collection of Arthropods (Gainesville, FL), the USDA-ARS European Biological Control Laboratory (EBCL) (Montpellier, France), and CREA-DC (Florence, Italy). Collection data associated with a subset of these specimens are deposited in the Hymenoptera Online Database (hol.osu.edu). Morphological terminology follows Mikó et al (2007).

ats postacetabular sulcus (Fig. 7)

cs clypeal setae (Fig. 9)

eps episternal foveae (Fig. 7)

hoc hyperoccipital carina (Figs 3, 6, 8)

lo lateral ocellus (Figs 3, 6)

lpt1 lateral setal patch on mediotergite of T1 (Fig. 3)

mp mesopleural pit (Fig. 7)

ms malar sulcus (Figs 4, 7)

mtps metapleural sulcus (Fig. 2)

not notauli (Fig. 8)

of orbital furrow (Figs 4, 7)

pof preocellar furrow (Fig. 9)

slt1 setal patch on laterotergite of T1 (Fig. 3)

Following their morphological examination, all specimens were preserved in 95% ethanol and shipped to EBCL. Tables 3, 4 list the specimens and voucher information included in the analysis. Prior to DNA extraction, individual specimens were bathed three times at room temperature in molecular grade water for five minutes. Genomic DNA was nondestructively isolated from the entire specimen using the Qiagen DNeasy kit (Hilden, Germany) as published in Taekul et al. (2014) with minor modification. In Step 7, the elution buffer, warmed to 55 °C, was allowed to sit on the membrane during 15 min before centrifugation. The collected flow-through was reloaded onto the spin column to increase the DNA yield. A negative control (no insect tissue) was included in each extraction to detect potential contamination. The barcode region of the mitochondrial Cytochrome Oxidase Subunit I (COI) was amplified using the universal barcoding primers LCO1490 (5’-GGTCAACAAATCATAAAGATATTGG-3’) and HCO2198 (5’-TAAACTTCAGGGTGACCAAAAAATCA-3’) (Folmer et al. 1994). Amplification and barcode editing and analysis were done as described in Ganjisaffar et al. (2018). All sequences generated from this study and those from our custom barcode database are deposited in Genbank (Tables 3 and 4) and all residual DNAs are archived at EBCL (Tables 3 and 4). The sequences obtained were compared with sequences present in Genbank by similarity search using the Basic Local Alignment Search Tool (http://www.ncbi.nlm.nih.gov/BLASTn).

Although tracing the source of the Italian populations was not the scope of the present study, we used this barcode approach to provide better insights into the mitochondrial diversity of these two Trissolcus species, identify and exclude Asian populations that were highly divergent from Italian populations, and tentatively find similar populations. Of note, T. mitsukurii barcodes are poorly represented in Genbank database and BOLD as there is only one record published by Mita et al. (2014). For this study we had access to an unpublished EBCL database of 12 barcodes of T. mitsukurii collected in China, Korea and Japan and obtained within the framework of an ongoing USDA biocontrol program. Two sequences generated from the present study were aligned using Clustal W with the sequence from Mita et al. (2014) and the 12 sequences of our custom database.

For T. japonicus, the five sequences generated from this study were aligned with Clustal W with all barcode sequences retrieved from Genbank and BOLD (which contained some sequences not present in Genbank). Only sequences from Asian and Swiss samples were included in the dataset. For both taxa, the phylogenetic relationships among haplotypes were depicted using statistical parsimony in TCS as implemented in PopART (Leigh and Bryant 2015). This approach enabled us to display the geographical distribution of all haplotypes.

Site descriptions for recovery sites are given in Table 1. A total of 31 H. halys egg masses were collected at three survey sites in northeastern Italy on August 7–8, 2018 (Fig. 1, Table 2). Of these, 4 egg masses were not parasitized, and nymphs hatched at a rate of 98.23%; 25 egg masses were parasitized by T. mitsukurii; and 2 egg masses were parasitized by Anastatus bifasciatus (Geoffroy) (Eupelmidae) and a species of Pteromalidae. From the 25 egg masses parasitized by T. mitsukurii, a total of 424 adults emerged of which 91.75% were females. Depending on locality, the parasitism rate of eggs within single egg masses ranged from 32.14% to 100% and number of parasitized egg masses out of the total number of egg masses detected ranged among sites from 50.00% to 84.21%.

During surveys at the Codroipo site, an egg mass of a predatory stink bug belonging to the subfamily of Asopinae was collected which was also parasitized by T. mitsukurii. From this egg mass, 83.87% of the eggs (31 eggs in total) were parasitized (emergence rate 96.15%; 80.00% females); only one egg of the cluster produced a stink bug nymph, and four apparently unparasitized eggs did not hatch.

The recovery site of T. japonicus in northwestern Italy was located at 45.3031N, 9.4794E (Fig. 1). Parasitized egg masses were found on Acer campestris L. trees in a parking lot surrounded by a multi-host patchy landscape (crop fields, uncultivated fields, hedgerows), and industrial and urban areas with ornamental plants.

Field collections were made on July 27 and August 3, 2018, and a total of 45 H. halys egg masses were collected (with a mean of 25.77 eggs/egg mass). On July 27, only one egg mass was detected and was found to be successfully parasitized by T. japonicus, with the emergence of 8 specimens, all of which were males. Among the 44 egg masses collected on August 3, 21 egg masses were not parasitized and eggs hatched into nymphs with a mean rate of 82.44%, and 22 egg masses were parasitized by A. bifasciatus. Only one egg mass collected on August 3 was parasitized by T. japonicus. All eggs that were apparently parasitized (35.71% of eggs in the mass) produced adult parasitoids (9 females and 2 males), 46.43% of eggs were unhatched and 17.86% of the eggs hatched into stink bug nymphs. Ninety percent of the emerged T. japonicus were females.

Trissolcus mitsukurii is a straightforward species to identify and is separated early in the key to Palearctic Trissolcus (Talamas et al. 2017). We thus do not consider it necessary to produce an updated key to European Trissolcus that includes this species, and instead we provide the following diagnosis, with references to illustrations, that will be helpful for distinguishing T. mitsukurii from the European fauna: the clava of the female antenna is 5-merous, and the clavomeres are distinctly larger and darker than preceding antennomeres (Fig. 2); the orbital furrow is expanded where it intersects the malar sulcus (Fig. 4); the hyperoccipital carina is often present directly posterior to the lateral ocellus, but is absent between the ocelli (Figs 3, 6); the metapleuron is without setae below the metapleural sulcus (Fig. 3); a setal patch is present on the first laterotergite of the metasoma (slt1, Fig. 3). Care should be taken not to confuse the setal patch of the laterotergite with the setal patch on the mediotergite (lpt1, Fig. 3).

The key to European Trissolcus in Talamas et al. (2017) included T. japonicus to provide identification of this species given the possibility that it might become established in Europe. The specimens from northern Italy are fully congruent with the concept of this species presented by Talamas et al. (2017). Trissolcus japonicus can be separated from other species of European Trissolcus by the following diagnosis: four clypeal setae are present below the antennal insertions (Fig. 9); microsculpture is present throughout the frons (Fig. 9); the orbital furrow is expanded at its intersection with the malar sulcus (Fig. 7); the hyperoccipital carina is complete (Fig. 8); episternal foveae extend from the postacetabular sulcus to the mesopleural pit (Fig. 7); and the mesoscutum is without oblique rugae between the notauli (Fig. 8). The preocellar furrow, which extends ventrally from the median ocellus (Fig. 9), is a useful character for confirming the identity of T. japonicus, but it is not always present, and exhibits the greatest variability in males.

Specimens of A. bifasciatus were identified by GSP using the keys of Kalina (1981) and Askew and Nieves-Aldrey (2014) and the identification was confirmed by Dr. Lucian Fusu (University of Iasi, Romania) who compared them with authoritatively identified specimens.

The two voucher specimens recovered from the field in Cordenons (site 1) in the region of Friuli-Venezia-Giula yielded a similar barcode sequence of 666-bp in length. A BLAST search showed the best similarity score (99%) of this barcode sequence with T. mitsukurii (Accession No. AB971831). From the final alignment of 617-bp of 15 T. mitsukurii barcodes, a total of five haplotypes (denoted H1-H5) were recovered (Table 3). The haplotype H5 found in Italy is a new haplotype, as it did not match any haplotype found so far in Asia (Fig. 11, Table 3). From the network analysis, H5 differed by seven substitutions from the two closest haplotypes H1 and H2 from Japan and China respectively (Fig. 11).

The five voucher specimens recovered from the field in Lodi yielded a unique barcode sequence of 666-bp in length. A BLAST search showed the best similarity score (100%) of this barcode sequence with T. japonicus (Accession No. AB971832). From the final alignment of 373-bp of 60 T. japonicus barcodes, a total of seven haplotypes (denoted H1 to H7) were recovered (Table 4). The haplotype H1 found in Italy was also found in Tsukuba and Kanagawa in Japan, and in Switzerland (Fig. 12, Table 4). H1 is also the predominant haplotype found in the specimens collected in Japan (45%). Samples from China were the most diverse, displaying 4 haplotypes (H4, H5, H6, H7), although they represent only 22% of the total sampling. From the network analysis, H7 from China and South Korea differed only by one substitution from the haplotype H1 (Fig. 12). Although our haplotype analysis evidenced a best match of the Italian populations to Japanese and Swiss populations so far, we cannot entirely exclude that the haplotype H1 may also be present in unsampled populations of other parts of Asia where T. japonicus is present. Tracing the source of an introduction depends on the availability of information about population structure and may require analysis of more than one locus. To this end, a more comprehensive phylogeography study is underway which includes the COI barcode and microsatellite loci recently developed de novo in T. japonicus.