Rabdophaga rosaria (H. Loew, 1850)

IDENTIFICATION OF R. ROSARIA View in CoL

In this study, a whole larval body was used for each sample to make DNA extraction more reliable. Before DNA extraction, we confirmed the pinkish-white body colour and the presence of sternal spatula with a U-shaped emargination that are the distinguishing larval morphological features of R. rosaria (e.g. Möhn, 1955; Nijveldt & Yukawa, 1982). Thus, based on the larval morphological features and the shape of rosette galls, we identified, before DNA analysis, all larval samples from Japan, Korea, China and Russia to be R. rosaria .

In addition, all specimens collected from The Netherlands by J. Bijkerk, from Poland by M. Skrzypczyńska, from the UK by C. Former and

*Collector: CF, C. Former; DC, D. Collet; GJ, G. Japoshvili; JB, J. Bijkerk; JY, J. Yukawa; KMH, K. M. Harris; KY, K. Yamagishi; MS, M. Skrzypczynska; MT, M. Tokuda; NU, N. Uechi; PWP, P. W. Price; SS, S. Sato; TI, T. Ishizaki; WGK,W.-G. Kim.

† The bold accession numbers mean that these data were obtained from GenBank or from Bowser et al. (2018); otherwise, the present data have been newly deposited in the DNA Data Bank of Japan (DDBJ), European Molecular Biology Laboratory (EMBL) and GenBank nucleotide sequence databases.

‡ Salix × multinervis is a common hybrid between S. aurita and S. atrocinerea in the UK ( Newsholme, 1992).

§ Asphondylia yushimai (Cecidomyidi: Asphondylini), ‘the soybean pod gall midge’.

K. M. Harris and from Soldotna, Alaska by D. Collet were identified as R. rosaria by the respective collectors, who are cecidomyiid taxonomists ( Table 3), except for some specimens collected by D. Collet from S. barclayi and S. sitchensis in Girdwood, AK, USA. In light of the genetic analysis by Bowser et al. (2018), we tentatively identified them to be Rabdophaga strobiloides (Osten Sacken, 1878) , which was expressed as R. cf. strobiloides following Bowser et al. (2018), a tentative identification.

MOLECULAR PHYLOGENETIC ANALYSIS FOR RABDOPHAGA GALL MIDGES

Forty-nine ethanol- or acetone-preserved specimens of the aforementioned gall midges were subjected to molecular phylogenetic analysis ( Table 3). For every individual, total DNA was extracted from the whole body with the DNeasy tissue kit (Qiagen, Japan) following the manufacturer’s instructions. A region of the cytochrome c oxidase subunit I (COI) gene of the mitochondrial DNA was amplified, purified, sequenced and electrophoresed following the methods described by Yukawa et al. (2003). The primers used for the amplification were as follows: forward, 5′-GGATCACC TGATATAGCATTCCC-3′ (COIS); and reverse, 5′-CCC GGTAAAATTAAAATATAAACTTC-3′ (COIA). This primer set has been used effectively for the analysis of intra- and interspecific variations in Cecidomyiidae (e.g. Uechi et al., 2003, 2004; Yukawa et al., 2003; Tokuda et al., 2004). The nucleotide sequence data reported in this paper have been deposited in the DNA Data Bank of Japan (DDBJ), European Molecular Biology Laboratory (EMBL) and GenBank nucleotide sequence databases with the accession numbers shown in Table 3.

For comparison, we obtained one sequence dataset of the COI region for European Rabdophaga heterobia (Loew, 1850) from GenBank and two for North American R. strobiloides from Bowser et al. (2018), as representatives of each species. Unfortunately, datasets for other species were not available or, if any, the range of the COI region used in the other species did not overlap sufficiently with what we analysed for R. rosaria , in which all positions containing gaps and missing data were eliminated, and there were a total of 421 positions in the final dataset. To include the datasets for R. heterobia and R. strobiloides , we inevitably reduced the positions from 421 to 376 by excluding discordant regions.

The sequence data were analysed with the ML method using the software package MEGA6 ( Tamura et al., 2013). Analyses were conducted using Tamura’s three-parameter model ( Tamura, 1992). Rate variations among sites were modelled with gamma distribution (+G, parameter = 0.25, shape parameter = 1). All positions containing gaps and missing data were eliminated. The resulting trees were subjected to bootstrap analysis ( Efron, 1982; Felsenstein, 1985) with 1000 replications for the ML cladogram. To estimate the time of diversification into different clades, we referred to the rates of 2.0 and 2.3% pairwise sequence divergence per million years for Hawaiian Drosophila spp. ( Diptera : Drosophilidae ) ( DeSalle et al., 1987).