Systematics and biology of some species of Micrurapteryx Spuler (Lepidoptera, Gracillariidae) from the Holarctic Region, with re-description of M. caraganella (Hering) from Siberia

Abstract During a DNA barcoding campaign of leaf-mining insects from Siberia, a genetically divergent lineage of a gracillariid belonging to the genus Micrurapteryx was discovered, whose larvae developed on Caragana Fabr. and Medicago L. (Fabaceae). Specimens from Siberia showed similar external morphology to the Palearctic Micrurapteryx gradatella and the Nearctic Parectopa occulta but differed in male genitalia, DNA barcodes, and nuclear genes histone H3 and 28S. Members of this lineage are re-described here as Micrurapteryx caraganella (Hering, 1957), comb. n., an available name published with only a brief description of its larva and leaf mine. Micrurapteryx caraganella is widely distributed throughout Siberia, from Tyumen oblast in the West to Transbaikalia in the East. Occasionally it may severely affect its main host, Caragana arborescens Lam. This species has been confused in the past with Micrurapteryx gradatella in Siberia, but field observations confirm that Micrurapteryx gradatella exists in Siberia and is sympatric with Micrurapteryx caraganella, at least in the Krasnoyarsk region, where it feeds on different host plants (Vicia amoena Fisch. and Vicia sp.). In addition, based on both morphological and molecular evidence as well as examination of type specimens, the North American Parectopa occulta Braun, 1922 and Parectopa albicostella Braun, 1925 are transferred to Micrurapteryx as Micrurapteryx occulta (Braun, 1922), comb. n. with albicostella as its junior synonym (syn. n.). Characters used to distinguish Micrurapteryx from Parectopa are presented and illustrated. These findings provide another example of the potential of DNA barcoding to reveal overlooked species and illuminate nomenclatural problems.


Introduction
With more than 2000 described species, the family Gracillariidae represents one of the most diverse groups of small moths (De Prins and De Prins 2015). Many species of gracillariids remain to be discovered, especially in the tropical regions (Lees et al. 2013;Brito et al. 2013) but also in the Palearctic (Laštůvka et al. 2013;Kobayashi et al. 2013;Kirichenko et al. 2015) and Nearctic regions (Davis and Deschka 2001).
During a DNA barcoding campaign of leaf-mining insects from Siberia carried out in 2011, we discovered a genetically divergent lineage of Micrurapteryx feeding on the Siberian peashrub Caragana arborescens (Fabaceae). Preliminary barcoding data showed pronounced divergence in the COI barcoding fragment from European specimens of M. gradatella. Examination of the genitalia revealed that it was clearly different from European M. gradatella.
In their taxonomic review of the Palearctic Micrurapteryx, Kuznetzov and Tristan (1985) called M. gradatella the species found in Siberia mining "yellow acacia" (= Caragana arborescens). They also stated that despite the confusion in the Russian lit-erature about various names applied to specimens mining Caragana in Siberia, in their estimation there was only one species present, which they deemed to be M. gradatella. Subsequent works (Noreika 1997;Kuznetzov and Baryshnikova 1998;Kuznetzov 1999) followed Kuznetzov and Tristan (1985).
Contrary to these authors, our findings indicated unequivocally that at least two species were present. This raised the question of whether the Caragana-feeding lineage from Siberia represented an undescribed Micrurapteryx species. Two unavailable names have been used in the literature to refer to a species feeding on C. arborescens in Siberia: Parectopa caraganella Danilevsky and P. caraginella Danilevsky (Hering 1957;Dovnar-Zapol'skiy 1969;Kuznetzov and Tristan 1985;De Prins and De Prins 2015). The lingering confusion about the identity of Caragana-feeding Micrurapteryx in Siberia is partly due to the lack of a detailed description of M. gradatella in Europe and an over-reliance on wing pattern characters without examination of genitalia. Only recently both female and male genitalia of M. gradatella have been illustrated (Bengtsson and Johansson 2011), but that description was very brief.
Based on differences in morphology and DNA sequence data (mitochondrial and nuclear), we assess that there are two species of Micrurapteryx in Siberia, M. caraganella and M. gradatella. We present elaborated morphological re-descriptions of the adults of both species. In addition, we compare the morphology and DNA barcodes with other European and North American Micrurapteryx, as well as some related species of Parectopa developing on Fabaceae whose barcodes clustered near Micrurapteryx.
The availability of the binomen M. caraganella with authorship attributed to Hering (1957) is discussed. We show that the Nearctic Parectopa occulta Braun, 1922 in fact belongs to Micrurapteryx (comb. n.) and is closely related to the Palearctic M. gradatella, and is re-described. In addition, based on examination of type specimens the North American Parectopa albicostella Braun, 1925 is shown to be a junior synonym (syn. n.) of M. occulta. Finally an assessment of morphological characters are presented that distinguish Micrurapteryx from Parectopa. Thus, in all cases sampling was done in urban ecosystems, on planted bushes of Caragana spp., on C. arborescens in all localities, additionally on C. frutex (L.) K. Koch and Medicago sativa L. in Omsk and C. boisii C. K. Schneid. in Novosibirsk. In all localities, except two, both the damaged leaves (carrying mines) and live insects (larvae in mines or pupae in cocoons on leaves) were collected; in Ulan-Ude and Chita, only empty mines were found which were preserved as herbarium vouchers. For comparative purposes, in early July 2015 we also collected mines with live larvae of M. gradatella on Vicia amoena in suburb of Krasnoyarsk (Yenisei river bank, near Karaulnaya biostation) and Parectopa ononidis (Zeller, 1839) on Trifolium pratense L. in suburb of Krasnoyarsk (Yenisei river bank, Skala Berkut).

Sampling
Mined leaflets as well as larvae feeding in mines and pupating on leaves were photographed in nature and in the laboratory with a digital camera Sony Nex3 (in laboratory, the photographs were taken through a Zeiss STEMI DV4 binocular microscope).
Adults of M. caraganella examined in this study were obtained by rearing larvae and pupae collected on C. arborescens in July-August 2013-2015 and on C. frutex in July 2015. Six larvae and seven pupae were preserved in 96% ethanol, including a specimen on Caragana boisii, for genetic and morphological analyses. In addition, 70 larvae were left to complete their development in glass jars (200 ml) lined with filter paper on the bottom, in laboratory conditions (22 °C, 55% RH, LD 18:6 h photoperiod). As leaflets of the host plant dry quickly, mined leaflets were collected with a short section of twig; the latter was tightly wrapped in paper tissue and moisturized every second day, following guidelines of Ohshima (2005). Twelve pupated larvae, collected in nature as well as those that pupated in the laboratory, were transferred to Petri dishes (90 mm in diameter) on filter paper and kept until the adults emerged. In the Petri dishes, the humidity was regulated by adding few drops of water to a small cotton ball attached inside the lid. In total 32 larvae out of 70 larvae pupated and 30 adults emerged. Larvae of M. gradatella (5 specimens) and P. ononidis (4), collected near Krasnoyarsk, were grown in the same conditions as above and 2 adults of each species emerged.
Samples of Micrurapteryx salicifoliella, M. occulta, Parectopa lespedezaefoliella Clemens, 1860 and P. robiniella Clemens, 1863 from North America, as well as M. gradatella and M. kollariella from Europe were also examined. All specimens used in this study for both genetic and morphological analyses are listed in Tables 1 and Suppl. material 1: Table S2.

DNA sequence analysis
Sequence data for the barcode fragment (Hebert et al. 2003) were collected to estimate the barcode gap between M. caraganella and the related species. In addition, we sequenced two nuclear genes: histone H3 and 28S rDNA (28S) for M. caraganella and M. gradatella as an independent source of data to confirm the large divergence observed in the barcode fragment between these two species.
DNA from 22 specimens of M. caraganella, seven specimens of M. gradatella and one Parectopa ononidis was extracted, PCR amplified and sequenced at INRA (Orléans, France). DNA was extracted using NucleoSpin® tissue XS kit, Macherey-Nagel, Germany, according to the manufacturer's protocol. The COI barcoding fragment, 658 bp, was amplified via PCR at the standard conditions for the reaction. PCR products were purified using the NucleoSpin® Gel and PCR Clean-up kit Macherey-Nagel, Germany and sequenced by the Sanger method with Abi Prism® Big Dye®Terminator 3.1cycle sequencing kit (25 cycles of 10 s at 96 °C, 5 s at 50 °C, 4 min at 60 °C). Sequencing was carried out using a 3500 ABI genetic analyzer. All sequences were aligned using CodonCode Aligner 3.7.1. (CodonCode Corporation).
DNA for the remaining samples was extracted and barcoded at the Canadian Centre for DNA Barcoding (CCDB, Biodiversity Institute of Ontario, University of Guelph) using the standard high-throughput protocol described in deWaard et al. (2008). In addition, 109 samples of North American species Micrurapteryx salicifoliella, M. occulta, and Parectopa robiniella, earlier barcoded by other colleagues, were also included in the analysis (Table 1).
The resultant sequences, along with the voucher data, images, and trace files, are deposited in the Barcode of Life Data Systems (BOLD) (Ratnasingham and Hebert 2007;www.barcodinglife.org) and the sequences were deposited in GenBank. All data are available through the following dataset (http://dx.doi.org/10.5883/DS-MICRURA) Intra-and interspecific genetic distances were estimated using the Kimura 2-parameter model implemented within the analytical tools available in BOLD. We also used BOLD to obtain Barcode Index Numbers (BINs) (Ratnasingham and Hebert 2013). A neighbor-joining (NJ) tree was constructed using MEGA 5.05 (Tamura et al. 2011).

Morphology
The external morphology of M. caraganella and the related species of Micrurapteryx and Parectopa was studied (  Table S2). Genitalia dissections and slide mounts prepared by PT (TRB slide numbers) and NK (NK slide numbers) followed Robinson (1976); those prepared by JFL (MIC, JFL, and USNM slide numbers) followed Landry (2007).
Genitalia imaged by PT were photographed with a Leica DFC 450 digital camera through Leitz Diaplan GMBH microscope. Those imaged by JFL were photographed with a Nikon DS-Fi1 digital camera mounted on a Nikon Eclipse 800 microscope at magnifications of 40× or 100× and Nikon's NIS 2.3 Elements was used to assemble multiple images from successive focal planes into single deep-focus images. All photos and illustrations were processed, adjusted, and assembled into plates with Adobe Photoshop. Terminology of the genitalia follows Klots (1970) and Kristensen (2003); body larval chaetotaxy Kumata (1988), and that of the head Davis and Wagner (2005). Scanning electron microscope (ESEM) digital images of pupae were taken with a Hitachi TM1000.
Pinned specimens were photographed with a Canon EOS 60D with a MP-E 65 mm macro lens. They were placed on the tip of a thin plastazote wedge mounted on an insect pin, with the head facing toward the pin and the fringed parts of the wings facing outward. This ensured that there was nothing between the fringes and the background. Lighting was provided by a ring of 144 LEDs covered with a white diffuser dome (Fisher 2012 and references therein). The camera was attached to a re-purposed stereoscope fine-focusing rail. Sets of 30-65 images in thin focal planes were taken for each specimen and assembled into deep-focused images using Zerene Stacker, and edited in Adobe Photoshop.

DNA barcodes
In total, 157 DNA barcodes of specimens of the genera Micrurapteryx and Parectopa were analysed in this study:   (Table 1, Figure 1. A Neighbor-Joining tree, based on COI barcode fragment, generated under the K2P nucleotide substitution model, of the studied taxa. Each specimen is identified by its Sample ID code (see Table  1). Branch lengths represent the number of substitutions per site. BIN numbers from BOLD system are given in parentheses for all clusters. There are 56 mutations and 9.2% interspecific distance between Micrurapteryx caraganella and M. gradatella. sample ID -NK440) was not successful but their sequences with the genes 28S and histone H3 were obtained. There was a perfect correspondence between Barcode Index Numbers (BINs) membership and the known species (Fig. 1). The sequences of M. caraganella formed a distinct cluster (Fig. 1). We found 56 diagnostic substitutions in the barcode fragment between M. caraganella and M. gradatella (Suppl. material 1: Table  S3). There is a clear barcode gap in the genus Micrurapteryx with a mean intraspecific divergence of 0.24% versus a nearest-neighbour (NN) distance averaging 5.84%. The lowest interspecific distance (2.0%) was observed between M. gradatella and specimens from North American M. occulta reared from Lupinus (Table 2). With DNA-barcoding, we identified Parectopa ononidis on Trifolium pratense in Siberia (Krasnoyarsk, Yenisei, Skala Berkut, 5.VII.2015, sample ID NK463) (Fig. 1), which is a new insect record for Siberia.
Within studied species, M. gradatella showed low intraspecific variability (0.02%) with ten specimens originating from one locality in Finland and one locality in Siberia (Table 2). All specimens from Finland, collected on Lathyrus linifolius shared the same haplotype. One mutation was observed in a Siberian specimen of M. gradatella (sample ID -NK459) sampled from a second host, i.e. Vicia amoena.
North American specimens of M. occulta formed a single large cluster belonging to one BIN (BOLD:AAD5802) which was nested close to M. gradatella within Micrurapteryx. Intraspecific variability at 1.66% was higher than for other species studied here but the geographic sampling was correspondingly much greater, covering 38 localities spanning the continent from East to West.

Nuclear genes
We obtained sequences of the nuclear gene histone H3 and 28S rRNA D1-D3 for 23 specimens (17 specimens of M. caraganella and 6 specimens of M. gradatella, Table 1). Both H3 and 28S unequivocally delimit two distinct species with 3 and 2 diagnostic nucleotide substitutions respectively ( Fig. 2 Diagnosis. Superficially, this species can be confused with M. kollariella (Figs 17,20,(30)(31)47), widespread in Europe east to Kazakhstan. However, the latter can be distinguished by its forewing pattern with wider costal strigulae and white dorsal margin not denticulate. In male M. kollariella, the coremata are very long; the valvar apex is more protruded than in M. gradatella; the saccular apex has a strong, incurved bifurcate tooth; and the phallus is anteriorly widened and deeply invaginated and with fine lateral serrations (Figs 30,31); in female M. kollariella, S6 is weakly sclerotized and less developed, the antrum is widest near the ostium, and the signa are a pair of finely denticulate plates (Fig. 47); in M. gradatella the antrum is elongate, cylindrical and widest more anteriorly. For differences with M. caraganella, see under that species.
Head. Frons and vertex white, sometimes with intermixture of brown scales on vertex, around eyes and at base of antenna. Labial palpus white, rather long and slender, upturned, spotted with dark brown in medial and apical segment; maxillary palpus about half of apical segment of labial palpus, outer side fuscous. Antenna fuscous, scape and pedicel white ventrally, remaining articles ringed with paler colour; pecten absent.
Thorax. Dorsum and venter white, tegulae dark brown. Legs white, tibiae and tarsi annulated with dark brown; fore coxa and femur grey outwardly. Forewing dark brown in ground colour with white markings; costal margin with 5 white strigulae; first three almost parallel, oblique and bent outwards; first costal strigula with basal half parallel to costa, then oblique and fragmented; second often obsolescent; fourth and fifth semicircular, often both touching opposite margin; dorsal margin white in basal two-thirds, with two or three white projections, the more distal one almost touching the first costal strigula; apical spot black, not quite touching 5 th strigula; cilia white around apex to tornus, with dark brown tips forming a line which projects a little at apex; hindwing grey ochreous, cilia pale grey.
Abdomen. Brown dorsally and white latero-ventrally. Segment 7 in the male with pair of coremata of thin scales about half width of sternum (Fig. 13). In the female sternum 6 more strongly sclerotized with a slight convexity on the proximal margin ( Fig. 18).
Pupa. Maximum length 5.5 mm; width 1.3 mm; vertex just shorter than frons. Frontal process (cocoon cutter) a transverse ridge strongly and irregularly dentate; frontal setae not visible, clypeal setae paired, very reduced and nearly contiguous. Antenna extended to abdominal segments A9; forewing to A5 or A6; hind leg to A10 or slightly longer than abdomen. Setae D1, L1 and SD1 present on abdominal segment A1-A7. Patočka and Turčáni (2005) report seta D1 on segment 7 but this was not found in the specimens examined. Cremaster consisting of a ring of five pairs of small spines, dorsal pair slightly enlarged and more closely set, two ventral pairs very small.
Larva. Very similar to M. kollariella and M. caraganella. Last larval instars of this species were studied in detail by Grandi (1933) and no structural differences were discovered. For description, see M. caraganella below.

Micrurapteryx caraganella
Hering's distinction in a key constitutes, however unintentionally, a valid description and thus makes the name Parectopa caraganella nomenclaturally available with Hering as the author. Despite being woefully inadequate, the "description" provided in Hering's key minimally meets the criteria expressed in Article 13.1.1 of the Code, namely that a name published after 1930 (but before 1960) "be accompanied by a description of definition that states in words characters that are purported to differentiate the taxon" (International Commission on Zoological Nomenclature 1999). It is worth noting that the description of the mine in association with the host plant provides a more useful diagnosis in the present case. Because the mine constitutes the work of an animal it could be construed as a condition for availability (Code article 12.2.8). However, such evidence is not admissible to assess the availability of names published after 1930.
Given its year of publication, a type specimen is not even required. Did Hering have voucher material of that species from Siberia when he wrote his 1957 work? He only mentioned the name "Buhr" at the end of the key couplet, who is presumably the person who communicated the information to him. He did not indicate how he obtained the name he attributed to Danilevsky. Even if so, the existence of voucher specimens would not affect the attribution of the name to Hering. In a catalogue of leaf-mining insects, Dovnar-Zapol'skiy (1969) cited "Parectopa caraginella Dan." (this seems to be a misspelling of caraganella) as a species feeding on Caragana described by Danilevsky from Western Siberia without any further reference or indication. As such, that citation has no nomenclatural value. Kuznetzov and Tristan (1985) correctly discounted the names Parectopa caraganella Danilevsky and Parectopa caraginella Danilevsky as nomenclaturally unavailable. Indeed, despite being cited by several authors, no original publication by Danilevsky where either spelling of the name is mentioned seems to exist. It is intriguing that no authors who cited or attributed the names to Danilevsky gave any indication or reference where those names were seen in the first place.
Diagnosis. The forewing pattern of M. caraganella is very similar to that of M. gradatella and the two species are separable with certainty only by examination of the genitalia. In male genitalia, M. caraganella differs mainly by the presence of a sharp, prominent tooth on the middle of the ventral margin of the valva. This character allows distinguishing easily this species from all other congeners. In female genitalia, the antrum is ampulla-shaped with lateral broadenings, whereas it is almost cylindrical in M. gradatella. The cremaster differs in pupae of the two species: there are three pairs of little spines in M. gradatella (Patočka and Turčáni 2005)  Head. Frons and vertex white, sometimes sprinkled with brownish grey. Palpi white; labial palpus rather long and slender, upturned, with apically forked dark brown band on median segment and sometimes apical one ringed with grey; maxillary palpus slightly more than half length of apical segment of labial palpus, spotted with fuscous outside. Antenna as in M. gradatella.
Thorax. Legs and thorax as in M. gradatella. Forewing dark brown in ground colour with white markings; costal margin with 5 white strigulae, the first four curving outwards, the fifth inwards, the first long and strongly oblique, the fourth often indistinct; dorsal margin with basal ⅔ white, this fascia denticulate inwards, often linked irregularly with costal strigulae; apical spot black with some mixture of paler scales, surrounded by circular white line including 5 th costal strigula; cilia and hindwing as in M. gradatella. Abdomen. Brownish grey dorsally and white ventrally, apical segment with lateral dark grey spot in the female. Segment 7 of male similar to M. gradatella. Sternum 6 of female as in gradatella but posterior margin more rounded.
Male genitalia (Figs 26,27). Tegumen short, triangular at apex, with no setae; tuba analis membraneous, without subscaphium, produced beyond tegumen, very similar to M. gradatella. Valva longitudinally cleft, costal region with sinuous margin, cucullus lobe rounded; sacculus with large, sharp tooth in middle of ventral margin and apex ventrally produced into strongly sclerotized toothed process with two pointed ends. Phallus tubular, about 0.9x length of valva, slightly bent in apical third, with small broadenings at base, a few small teeth on medio-ventral and dorsoapical walls and 2-3 larger denticles before apex; vesica with rather large patch of microspines and a thin, long cornutus apically pointed. Segment 7 with a pair of coremata of thin scales almost as long as width of sternum.
Pupa (Figs 49-54). Maximum length 4.2 mm; width 0.9 mm. Head setae as in M. gradatella. Frontal process (cocoon cutter) a transverse ridge strongly and irregularly dentate. Antenna extended to abdominal segment A7, A8 or A10; forewing to A5, A6 or A7; hindleg from posterior margin of A7 to just beyond apex of abdomen. Setae D1, SD1 and L1 present on abdominal segment A1-7. Cremaster consisting of ring of five pairs of small recurved spines, two dorsal pairs slightly enlarged and more closely set, ventral pair very small.
Head. Frons elongate, extended to epicranial notch, dorsal apodemes well developed, margins of epicranial notch with slight enlargement, on each side of caudal half while in M. gradatella these margins are regular; chaetotaxy with all three MD setae present, P2 very reduced; six stemmata on each side, arranged in 2 groups: first with 1 ventrad to A3, 2 between S2 and A3, 3, ventrad, near S2; second group in oblique line close to antenna. Mandible with 4 dorsal teeth and two ventral; both lateral setae present.
Body. Cuticle densely covered with very minute hairs, except on pronotal plate and small, symmetrical areas; chaetotaxy rather similar to that of Acrocercops-group (Kumata 1988): L setae bisetose on all segments except A9, SV bisetose on T1 and unisetose on T2-3, proprioceptor MD1 and MV3 present on T2-3 and A1-9; prolegs on A3-5 and A10. Most setae are inconspicuous, particularly the D and SV groups.  Biology (Figs 65-76). The species usually mines the leaves of Caragana arborescens (Figs 65-69) but some individuals (i.e. larvae in mines) were also found on C. frutex (Figs 70-71), C. boisii (Fabaceae) (Figs 73, 74) and on the herbaceous Medicago sativa (Fabaceae) (Fig. 72). The mine is a roundish or slightly branched blotch (branches are short, 2-5 mm long) above the midrib (Figs 68-69). Often a long, narrow tunnel is visible on the lower surface of the leaf (Fig. 71). The mine quickly develops into an upper-surface flat blotch with digitate channels, occupying half or an entire leaflet (Figs 68, 69, 72, 73), similar to M. gradatella (Figs 59-62). Fresh mines are white (Figs 66-72) with larvae visible when examining the mines with backlighting (Figs 73). The larva consumes all layers of palisade parenchyma and partly damages the layers of spongy parenchyma. Since not all spongy parenchyma is eaten, the colour of the mine can be slightly greenish yellow. Larvae eject frass out of the mine by protruding the rear end of their body through a slit (up to 7 mm long) on the underside of leaves. Larvae can leave their mines (Fig. 74) and begin a new one, either on the same or a neighboring leaflet.
Pupation (Figs 75-76). Pupation takes place outside the mine, usually on the lower surface of a leaflet where the larva spins a transparent, glossy cocoon, locating it usually perpendicular to the midrib, as in case of M. gradatella (Figs 63-64). Silk deposition by the prepupa induces a slight buckle in the leaf so that presence of the cocoon can be detected from above by the curved appearance of the leaflet. Occasionally pupation may also occur on the upper side of a leaflet, at the base along the midrib (Fig. 75).
Phenology. In Siberia, M. caraganella has two generations. The overwintering stage is not known (but is likely to be as a pupa or adult); neonate larvae of the first generation usually occur in early June. Adults fly in early July. The second generation develops from mid-July until the end of August.
Ecology and host plant range. Leaf mines of the new species were most commonly found in Siberia on Siberian peashrub, Caragana arborescens (Fabaceae), a plant widely used for different purposes: as an ornamental, for erosion control, as a source of nectar for bees, and for nitrogen fixation (Shortt and Vamosi 2012  These findings suggest that M. caraganella is an oligophagous insect with a preference for C. arborescens. In the Central Siberian garden SB RAS (Novosibirsk) in July 2012, NK also found a few mines of M. caraganella on C. boisii, an allied plant originating from China. In July 2015 in Omsk (Victory park), NK recorded mines of M. caraganella on C. frutex (native in Siberia). In the same location and at the same time bushes of C. arborescens were observed to be heavily attacked by M. caraganella (Figs 66,67), whereas bushes of C. frutex growing in vicinity (20 m from the damaged C. arborescens) were hardly colonized by the insect. In Omsk, on the same plot, NK also found the occasional mines of M. caraganella on the herbaceous legume Medicago sativa growing near heavily infested Siberia peashrub C. arborescens.

Micrurapteryx occulta
Other specimens examined. See Tables 1, Suppl. material 1: Table S2. Diagnosis. Superficially, M. occulta is virtually indistinguishable from the other species treated here, especially when the substantial amount of individual variation in coloration is taken into account. Most specimens have the head, thorax, costal and dorsal margins and strigulae of the forewing white, contrasting sharply with the dark brown disk and ground color. However, in several specimens, the white areas are obscured by a suffusion of dark-tipped scales which gives them an overall dark, peppery appearance. The genitalia of both sexes are amply different from M. salicifoliella, the only other North American species (Figs 16,22,28,29,48). When compared to Palearctic Micrurapteryx, its genitalia are most similar to those of M. gradatella, from which it differs in having a single elongate cornutus and the latero-medial tooth projecting, whereas M. gradatella has a second cornutus consisting in a small, separate spine and its latero-medial tooth is elongate and flat. In the female genitalia of M. occulta, the posterior sclerotized papillate section of the ductus bursae is slightly shorter relative to the anterior membranous section, or less than half the length from the antrum to the corpus bursae; in M. gradatella, the papillate section extends to about two-thirds of the ductus length. The two species are closely related morphologically, genetically, and biologically. (Figs 6-10). Wingspan 8.7-11.7 mm (average 10.1 mm; 44 specimens).

Description of adult
Head. Frons and vertex white in most specimens, or dark from admixture of dark brown scales in dark specimens. Labial palpus shape as in M. gradatella, outer surface of article 2 dark brown, inner surface from all white to nearly all dark brown; article 3 variously ringed with dark brown in distal half in many. Antenna dorsally fuscous throughout, ventrally with scale, pedicel, and in many ¼ to ⅓ of flagellum white; pecten absent.
Thorax. Dorsum white in pale (most) specimens, predominantly dark brown peppered with white in dark specimens. Tegulae dark brown. Legs as in M. gradatella.
Forewing. Pattern very similar to that of M. gradatella, but rather variable: in several specimens, dark portion of disk with pale-based, dark-tipped scales giving the appearance of pale suffusion; white dorsal margin in some specimens obscured by suffusion of dark-tipped scales; terminal portion between strigulae 4 and 5 and around apical spot rufous in specimens with white costa and margin. Forewing of darker specimens with overall peppery appearance.
Male genitalia (Figs 32-39, Suppl. material 2-4: Figs S01-S34). 32 preparations examined. Very similar to M. gradatella. Tegumen about 0.2× length of valva, with long and thin peduncular arms, apex subtriangular or subconical, with jagged edge, sometimes slightly indented. A pair of elongate lamellae about as long as tegumenpeduncular arms bracing the sides of anal tube, their distal portion with oblique wrinkles. Anal tube with 1 or 2 setae in few specimens, without seta in most. Latero-medial spine of phallus simple in most specimens, bitoothed in some specimens (including the holotype of P. albicostella), tri-toothed observed in one specimen (Table 3), the spine projecting dorso-laterally from the phallus surface. Notes about synonymy and variation. The synonymy of Parectopa albicostella with Micrurapteryx occulta is here established based on examination of the type specimens of both nominal species. Braun described each species on the basis of a single specimen, which she reared. The holotype of M. occulta is a female reared from Vicia caroliniana, and that of P. albicostella a male reared from an unspecified "vetch" (Fabaceae). We were not able to barcode the types. However, barcoded specimens of both sexes with genitalia corresponding to each of these nominal species cluster within a single, cohesive BIN (BOLD:AAD5802) comprised of specimens spanning a transcontinental geographic range. This cluster also includes specimens reared from different Fabaceae hosts that match the respective types in genital morphology and external appearance. Despite some morphological and genetic variation among examined specimens, we cannot find any consistent character to keep these two nominal taxa separate. Braun (1925) indicated that M. albicostella was closely allied to M. salicifoliella Chambers (Fig. 11), P. thermopsella Chambers, and M. occulta Braun, "but separated from all of them by the dark head and thorax and the white costal edge." We observed that these colour characteristics vary individually among all specimens examined, including among M. salicifoliella. For example, a pair of M. occulta with identical barcodes reared from leafmines on the same lupine plant from British Columbia (specimens CN-CLEP00121158 and CNCLEP00121159) shows the male with a dark head and thorax as well as a darkened dorsal edge as exhibited by the male holotype of P. albicostella, whereas the female has a white head, thorax, and costal edge as in the female holotype of M. occulta. In fact, the holotype of P. albicostella has the thorax predominantly dark peppered with white scales (Fig. 8, not really "streaked" as Braun described). Although this might suggest sexual dimorphism in colouration, both colour patterns (and others) were observed in each sex among the other specimens that we examined.
The genitalia of both Braun holotypes are not distinguishable from those of other barcoded specimens in BIN BOLD:AAD5802, as well as from several additional nonbarcoded specimens examined. Although minor variations in several features were observed, these do not exhibit a clear geographic pattern (Table 3).
In male genitalia (32 preparations examined, Figs 32-39, Suppl. material 2-4: Figs S01-S34), for example, the lateromedial tooth of the phallus is simple in most specimens (Fig. 33) but double in a few western specimens (Figs 35, 39, including the M. albicostella holotype from Utah), with one from British Columbia showing a suggestion of blunt doubling, and even one eastern specimen from New Brunswick with a triple tooth (Fig. 37); the apical lobe of the sacculus is variously pointed or somewhat rounded (rounded in M. albicostella holotype from Kentucky); the curvature of the apex of cucullus varies from well rounded to nearly straight; and a single or a pair of fine setae are present on the membranous part of the anal tube in some specimens (Fig. 32). The anal seta character is uncommon in Gracillariinae -it may have been overlooked -and seems inconstant at the specific level. One seta is present in one male M. kollariella examined (Fig. 30).
In female genitalia (17 preparations examined), the number of signa varies from 2 to 8 (average 5), and the relative length and thickness of the antrum, sclerotized portion of the ductus bursae, and ostium notch vary slightly in proportions with no significant gap (Figs 44-46, Suppl. material 5-6: Figs S37-S52).
On "Parectopa" thermopsella (Chambers, 1875). Braun (1922Braun ( , 1925 also alluded to the relatedness of P. thermopsella to M. albicostella, M. occulta, and M. salicifoliella, highlighting slight differences in forewing streaks, and this suggests superficially a similar external appearance and forewing pattern. It is not known whether Braun had seen authentic Chambers specimens of P. thermopsella. Chambers (1875) mentioned his P. thermopsella as "closely allied" to P. lespedezaefoliella (type species of Parectopa), P. robiniella, and M. salicifoliella, but his description of the larval mine immediately after that statement makes it unclear whether he was referring to the larval habits, the external appearance of the adult, or both. Both P. lespedezaefoliella and P. robiniella (Fig. 12) have forewing patterns unlike Micrurapteryx species but the larval mines are similar in appearance. The identity of Gracilaria [sic] thermopsella Chambers, 1875 remains unknown. The type locality is Spanish Bar, Colorado, and the host plant is a species of Thermopsis (Fabaceae). It has been included in Parectopa by subsequent authors (Braun 1925, McDunnough 1939, Davis 1983 but no type or other Chambers specimens seem to exist (Don Davis, pers. comm. to JFL, 2015).
Note on transferring occulta from Parectopa to Micrurapteryx. Despite the long-standing combination of occulta/albicostella with Parectopa, DNA, the forewing pattern, and genitalia clearly indicate greater relatedness to members of Micrurapteryx.
Biology Distribution. Micrurapteryx occulta is here recorded from across North America in the northern half of the continent, in Canada from the Maritime Provinces (Newfoundland, New Brunswick, Nova Scotia) to British Columbia, north to northernmost Yukon; in the United States it has been found in Connecticut (D.L. Wagner, pers. comm.), Kentucky, Illinois (T. Harrison, pers. comm.), Colorado (E. van Nieukerken, pers. comm.), Utah, Nevada, and California.

DNA barcoding and the status of Micrurapteryx species.
Siberia has a rich fauna of Lepidoptera which is still very poorly documented (Sinev 2013). So far, about 50 species of Gracillariidae are known to occur in Siberia on woody plants (Tomilova 1973;Dovnar-Zapol'skiy and Tomilova 1978;Kuznetzov and Baryshnikova 1998;Sinev 2008) but most of the region remains unexplored. Here, we confirm the existence of a distinct species of Micrurapteryx, namely M. caraganella feeding on plants from the genus Caragana (mainly on Siberian peashrub C. arborescens) and occasionally on Medicago sativa (Fabaceae) in Siberia based initially on differences in DNA barcodes. The status of M. caraganella is also supported by nuclear data, male and female genital morphology and biology.
In a review of Palearctic Micrurapteryx by Kuznetzov and Tristan (1985) considered that the Caragana-feeding Micrurapteryx present in Siberia were all referable to M. gradatella. However, it is clear from their description and illustrations of that species that it is markedly different in male and female genitalia from what is regarded as M. gradatella in Europe. Instead, their M. gradatella corresponds to our concept of M. caraganella.
In North America DNA barcodes revealed that a single species with a wide continental distribution is present, but that a significant amount of morphological variation was found among numerous specimens, supporting the synonymy of two long-standing nominal species, Parectopa albicostella and P. occulta. Barcodes and morphology also supported the transfer of Parectopa occulta to Micrurapteryx.
The average interspecific divergence for the DNA barcode fragment found within Micrurapteryx (11.5%) is similar to other Gracillariidae such as Cameraria and Phyllonorycter (Langmaid et al. 2011). The relatively high level of DNA barcode divergence found between M. caraganella and M. gradatella contrasts with the limited differentiation in the two nuclear genes sequenced (i.e. H3 and 28S) ( Table 2). The striking difference in the level of divergence between mitochondrial and nuclear genes could be caused by maternally inherited symbionts such as Wolbachia (Kodandaramaiah et al. 2013). A study on Wolbachia infection of both species is needed to confirm the role of this endosymbiont on the levels of mitochondrial and nucleotide diversity observed.
Host range in Micrurapteryx. The genus Micrurapteryx comprises species feeding on more than twenty different genera of legumes, and a host shift from Fabaceae to Salicaceae (Suppl. material 1: Table S1). In North America M. occulta has been recorded on several different genera of Fabaceae hosts (Caragana, Lathyrus, Lupinus, Melilotus, Vicia) (see specimens examined in DS-MICRURA dataset and Suppl. material 1: Tables S1, S2). Historic records of P. thermopsella (reared from Thermopsis in Colorado) may also be referable to this species (although no authentic specimens of this nominal species are known). However, most individual Micrurapteryx species are specialized on one or two host plant genera. Our findings add more evidence to the prevalence of relatively high levels of host plant specialization. Micrurapteryx sophorivora Kuznetzov & Tristan, 1985 is restricted to Sophora sp. M. gradatella is known to feed only on Lathyrus and Vicia and is found in North Europe exclusively on Lathyrus linifolius. In Siberia, Micrurapteryx caraganella can occasionally colonize other Caragana species, besides C. arborescens, for example C. frutex and C. boisii. The species is also able to develop on the herbaceous legume Medicago sativa. Such a host shift from a woody shrub to a herbaceous plant is uncommon in Gracillariidae, which typically have strict diets and where occasional host shift usually do not take place between structurally different plant species. We recorded a new host genus (Medicago) only in one location in Siberia (Omsk) where M. caraganella was highly abundant and was severely defoliating C. arborescens, and thus could disperse to a nearby herb. It is possible that Medicago does not represent a normal food plant for M. caraganella, but that its occurrence on that host resulted from a local mass-occurrence and a consequent "spill-over effect". Such a phenomenon is reported in other leaf miner species, including the horse-chestnut leafminer Cameraria ohridella Deschka & Dimić, 1986, a recent invasive pest of horse chestnuts (Aesculus hippocastanum L.) in Europe (Šefrova and Laštuvka 2001). Along with outbreaks and co-presence of the related maples (Acer spp.), mines of the horsechestnut leafminer can be found on maples, although in lower abundance (Gregor et al. 1998;Péré et al. 2010). Similarly, Ectoedemia occultella (Linnaeus, 1767) (Nepticulidae), an abundant leaf miner of birch trees (Betula spp.), has been once reported to feed on an unrelated willow Salix pentandra L. (Johansson et al. 1990). In other gracillarid species, we have occasionally observed atypical host shifts, and these events often are associated with elevated population numbers, e.g. in Phyllonorycter hilarella (Zetterstedt, 1839) (from Salix spp. to Populus tremula L., observations by MM, see also Bengtsson and Johansson 2011); P. sorbi (Frey, 1855) (observations verified by barcoding by MM from Sorbus spp. to Prunus padus L., P. domestica L., P. avium L., Malus spp. and Crataegus spp.; also a record verified by barcoding on Chaenomeles sp. (C. Doorenweerd, in litt.)); and Phyllocnistis labyrinthella (Bjerkander, 1790) (from Populus spp. to Salix pentandra, observations by MM). We observed significantly lower abundance of Micrurapteryx caraganella mines on Medicago, which we consider supporting the spill-over hypothesis, but on the other hand, we have not monitored the presence of mines on Medicago over several years, and therefore cannot exclude the possibility that it is part of the normal diet of M. caraganella.
Differential diagnoses of Micrurapteryx and Parectopa. The original descriptions of these two genera (for Parectopa : Clemens 1860: 209;for Micrurapteryx: Spuler 1910: 409) focused exclusively on external features of the head, antennae, palpi, wing shape, and venation, as was customary at that time. Spuler (1910: 409) defined Micrurapteryx on account of the apex of the forewing being tail-like (hence the name): this appearance results from a thin "pencil" or line of dark fringe scales at the apex of the forewing which stand out from the surrounding white fringe scales, and thus make the wing appear "tailed". This appearance is further accentuated by a rim of white scales between the apical dark spot and the base of the "tail". In Parectopa, there is also a thin line of dark fringe scales at the apex of forewing but the dark outer edge of the fringe surrounds it so that it does not look "tailed".
In describing Parectopa, Clemens (1860: 209) presented the description of the forewing venation first, emphasizing (italics in his text) the lack of "costal nervure" (Sc?) and the "three-branched" median vein (instead of four, as when CuA1 and CuA2 are both present, meaning these two veins are coincident or fused).
Vári (1961) cited verbatim the original descriptions of both genera and added genitalia characters as well as venational and leg details, but his re-descriptions do not provide clear distinctions between the genera other than for venation. In his treatment of Micrurapteryx he stated "Probably allied to Parectopa, but differing from it by [forewing] veins 2 (CuA2) and 3 (CuA1) being stalked and the male genitalia" (Vári 1961: 55), as opposed to being coincident in Parectopa. This venational feature is indicated by Spuler (1910: 409, legend of fig. 160) to be variable in Micrurapteryx. The value of these minor venational differences has not been assessed. The genitalia characters as provided by Vári are not easily comparable between the two genera (see Suppl. material 1: Table S5). Despite indicating that he "greatly restricted" Parectopa and reinstated Micrurapteryx as valid, Vári did not list which species he examined for both genera, although it can be assumed from his discussion that these included at least the type species.
In addition to DNA barcodes that cluster species into different sets of BINs and segregated Micrurapteryx from Parectopa, we noted several morphological characters not formulated by previous authors that distinguish the two genera from each other (Figs 77-86). These character states are likely mixtures of apomorphies and plesiomorphies. Without a phylogenetic framework for the genera of Gracillariinae and a more comprehensive mapping of characters across genera, it remains premature to assign character polarities and apomorphies that would support either the monophyly of each genus, or whether Micrurapteryx and Parectopa form a single monophyletic clade and should be combined. However, the differences are compelling enough to support the proposed new combinations. Provisional diagnoses for each genus follow. The characters presented are not meant to be exhaustive. We focused on abdominal and genital characters, and did not examine wing venation nor other skeletal features. We did not conduct a comprehensive survey of all the species currently attributed to each genus. However, the character states given here were present in all those examined (listed in Table 1 and Suppl. material 1: Table S2).
Contrastingly, character states shared by the examined species of Parectopa: Forewing with pattern of short costal and dorsal streaks, dorsal margin concolorous with disk, apical spot absent (Fig. 12). Male abdomen with S1-2 venulae sinuate and anteriorly extended into free apodemes projected beyond anterior margin of sternum (Figs 23, 78). T7 well sclerotized, transverse, anteriorly margined with antero-lateral corners prolonged into tapered strut which abuts similar structure of S7. S7 sclerotized with thickened anterior margin. Intersegmental membrane 6-7 with pair of very long coremata (longer than width of abdominal segment). T8 elongate-conical with posterior margin lined with dense row of flatly broadened scales. S8 completely membranous, reduced, indistinct. Pleura 8 with pair of elongate coremata with scales in transverse, fan-like arrangement (Fig. 80). Female abdomen with S1-2 similar to male but venulae straight. S6 weakly sclerotized, not markedly distinct from other sterna. Male genitalia (Fig. 82)  In conclusion, our study documents another example of how DNA barcoding can help to reveal overlooked species and clarify taxonomic issues (Jin et al. 2013;Landry et al. 2013;Lees et al. 2013;Mutanen et al. 2013;Huemer et al. 2014). Moreover, our analysis highlights the need for a careful revision of Parectopa and Micrurapteryx in the Nearctic and Palearctic Regions, particularly in the context of a broader phylogenetic analysis of the Gracillariidae.