Stigmellamultispicata Rociene. & Stonis, an Asian leafminer on Siberian elm, now widespread in eastern North America (Lepidoptera, Nepticulidae)

Abstract Stigmellamultispicata Rocienė & Stonis, 2014, previously known from the single male holotype from Primorye, Russia, is reported as a new invasive species mining leaves of Siberian elm, Ulmuspumila L., in eastern North America. Both adults and leafmines have been reported from many sites as unidentified Nepticulidae since 2010. Crucial for the identification was a match of the DNA barcode of a single larva collected on Ulmuspumila in Beijing with adults from North America. The single larva constitutes a new record for China. Stigmellamultispicata is closely related to the European S.ulmivora (Fologne, 1860), feeding likewise on Ulmus, but differs in details of external morphology and genitalia, particularly in the female, where S.multispicata has a remarkable elongated narrow ovipositor, suitable for oviposition in underside hairy leaf vein axils, where all mines start. In North America S.multispicata is the only Ulmus-feeding nepticulid with green larvae. Currently the species is known from USA: Illinois, Indiana, Iowa, Maryland, Massachusetts, Minnesota, New York, Ohio, Tennessee, Wisconsin, and Canada: Ontario and Québec. In Sagaponack, on Long Island, New York, larvae have been reported to occur en masse on Siberian elms from at least two sites. The current distribution could be reconstructed thanks also to many online photographs from observation websites. The species is redescribed, with the first descriptions of female, larva, and leafmine, and compared with S.ulmivora, which is fully redescribed. The two native North American nepticulid Ulmus leafminers, S.apicialbella (Chambers, 1873) and Ectoedemiaulmella (Braun, 1912), are diagnosed and new provincial and state records are provided. A key to linear mines on Ulmus in North America is provided. We suspect that trade of live plants through nurseries played a role in the sudden spread of this invasive species.


Introduction
The North American insect fauna has been enriched by an influx of numerous alien species. Most invasions in temperate North America originate in Europe, far fewer in Asia (Mattson et al. 1994(Mattson et al. , 2007. Among phytophagous insects on woody plants, the number reported in 2007 as arriving from Europe is 310 species, compared with only 77 coming from Asia (Mattson et al. 2007). The increasing global travel and transport of goods results in an ever-growing number of established alien insect species, estimated to be more than 40 over a five-year period (Work et al. 2005). While the number and proportion of non-native Lepidoptera is not as large as in some other insect groups, around 250 of the 5431 named Canadian and Alaskan species (4.6%) are thought to be introduced ). Thanks to DNA barcoding, previously overlooked introductions have been recognized, particularly in taxonomically poorly studied groups (Landry et al. 2013). Here we report another case of a lepidopterous insect introduced from Asia, where DNA barcoding played an essential role in the identification.
Beginning around 2010, James Vargo found numerous specimens of an unknown Stigmella species that he had not seen before, while light collecting on his property in Indiana, USA. Material was sent to specialists Don Davis and David Wagner, neither of whom recognized the moths as belonging to a known American species. EJvN got some of these specimens when borrowing unidentified Nepticulidae from the Mississippi Entomological Museum in 2010, and also failed to name them. A near match of the DNA barcode with a larva collected in 2013 from a leafmine on Ulmus in Beijing, China, was striking, but initially regarded as coincidental.
In October 2015, DOG received a sample of green nepticulid larvae from Ms. Lee Foster in Sagaponack on Long Island, Suffolk County, New York, that were descending en masse on silk strands from Siberian elm (Ulmus pumila L.) . Since the two known North American Nepticulidae on elm (see below) have yellow larvae, he contacted EJvN for identification. It soon was clear by comparing DNA barcodes and the emerging moths that these were the same species as the one found in Indiana. The property in Indiana also has several old Siberian elm trees, but leafmines were not seen until 2018 (J. Vargo, pers. comm. to EJvN). Meanwhile, more records with the same DNA barcode turned up in the BOLD database (Ratnasingham and Hebert 2007). The match with the Chinese elm miner now made more sense, and the male genitalia and external features appeared to match the illustrations of the recently described Stigmella multispicata Rocienė & Stonis, 2014, based on a single male holotype from Primorye, Russia (Stonis and Rocienė 2014). It was therefore likely that we were dealing with a recent invasion. Also, several online photos of leafmines and adults, including on BugGuide (2018) and iNaturalist (2018), could now be matched with S. multispicata (Table 1).
The Nepticulidae (pygmy moths) of North America were revised almost 40 years ago (Wilkinson 1979, 1981, Wilkinson and Scoble 1979, Wilkinson and Newton 1981Newton and Wilkinson 1982), but these studies were based on the relatively poor material then available in some museum collections only; no new fieldwork was carried out, and the publications are hampered by the lack of good color illustrations of moths and leafmines. In recent years, extensive collecting and genetic analyses have shown that the nepticulid fauna of North America is much richer than previously thought and also that more revisionary work of old types is needed. The first results of this research were summarized in the recent catalogues of global and Canadian Nepticulidae (van Nieukerken et al. 2016, . Also, a revival of leafmine studies, partly made possible through online resources (BugGuide 2018, Moth Photographers Group 2018) resulted in an influx of new distribution records.
The Nepticulidae of East Asia are best known from Far East Russia, especially the Primorye region, by studies of J. Stonis (formerly R. Puplesis) and his students . The study on Chinese nepticulids is still in its infancy [two papers deal with species feeding on Fagaceae (van Nieuker ken and Liu 2000, ] whereas studies on the Japanese fauna are now progressing (Kemperman et al. 1985, Hirano 2013, 2014, Yagi and Hirowatari 2017. We will here redescribe Stigmella multispicata, compare it with the European S. ulmivora (Fologne, 1860) and the other North American species of Nepticulidae feeding on Ulmus, and discuss the probability of its invasion from Asia into North America.

Material
The material of S. multispicata from the United States and Canada originates from several sources; the few adult specimens from Canada were collected during the School Malaise Trap Program (Steinke et al. 2017). Leafmines in China were collected by EJvN during a collaboration between groups in The Netherlands and China in the 1980s (van Liu 2000) and again during a 2013 visit of EJvN to Beijing and Tianjin. More data were obtained from internet searches and barcoded specimens on BOLD (Ratnasingham and Hebert 2007). * The single photographed adult was said to be part of a "mini invasion" at a light on this night. ** This specimen was sent to the authors and is also listed under material examined. *** According to the observer the tree was either Ulmus parvifolia or pumila, we think it is U. pumila.
Material of the other species discussed here mostly originates from the collections of the Naturalis Biodiversity Center, Leiden, Netherlands. The Material examined section provides only the basic locality data, all details on registration numbers, genitalia slides, collectors etc. are provided in the Suppl. material 1.

Methods
For the collection of leafmines in China see van Nieukerken and Liu (2000). Since the collecting of adults and mines in North America was done by several different persons, there are no general methods to describe here. Genitalia were prepared according to our standard procedures, usually including DNA extraction, which were described earlier in detail ( , van Nieukerken et al. 2010.
Measurements of genitalia were obtained from digital images, using calibrated scaling in the Zeiss AxioVision software; we used a 20× objective for male genitalia and 10× or 20× for female genitalia. Capsule length was measured from vinculum to middle of uncus; valva length from tip of posterior process to ventral edge, excluding the sublateral process; phallus length was measured along the sclerotized tube, from tip, excluding any protruding vesica parts. Total bursa length includes all of the internal genitalia from cloaca to anterior edge of bursa; apophyses were measured from abdominal tip to anterior tip of apophyses. Genitalia measurements are rounded off to the nearest 5 μm. Forewing length was measured from tip of fringe to attachment on thorax, with a Zeiss SV11 stereo-microscope at a magnification of 20×. Antennal segment counts include scape and pedicel; they were counted on photographs or directly under the same stereo microscope.
Photographs of moths were made with an AxioCam MRc 5 digital camera attached to a motorized Zeiss SteREO Discovery V12, using the Module Extended Focus, Zeiss AxioVision software, to prepare a picture in full focus from a Z-stack of ca 10 to 40 individual photos. Leafmines were photographed by EJvN with an AxioCam HRc camera on a Zeiss Stemi SV11 stereo-microscope, without extended focus. Photos by CSE were taken either with a Canon EOS Rebel XSi SLR digital camera, and MP-E 65 mm macro lens (Figure 34), or with a Nikon D50 digital camera and AF Micro Nikkor 105mm lens ( Figure 39) and with a Macro Twin Lite MT-24EX flash unit. Genitalia were photographed with an MRc 5 camera on a manually operated Zeiss Axioskop H, without using extended focus. Photographs were edited with Adobe Photoshop ® (various versions), avoiding any change to the real object, but backgrounds are cleaned of excess debris and artifacts by using healing brush and clone tools; tone and contrast are adjusted and some sharpening is used.

Stigmella multispicata Rocienė & Stonis
Diagnosis. In North America S. multispicata is the only Stigmella species with the combination of black frontal tuft, white collar, and single fascia. Stigmella quercipulchella (Chambers, 1882) is relatively similar, but has an additional silver patch at tornus, is slightly larger and has more strongly purple reflections across the forewings. This combination of characters is also diagnostic in East Asia, but there remains a possibility that similar species will be discovered. From the closely related European S. ulmivora it differs by the white collar (dark in ulmivora) and the entirely dark antennae (those of S. ulmivora have the distal 7-8 flagellomeres white); S. ulmivora is also slightly larger and has more antennal segments. The female of S. multispicata differs from all more or less similar species by the obvious long ovipositor, visible even without dissection. Male genitalia differ from those of S. ulmivora in the shallowly indented uncus and the very short sublateral processes of the transtilla; the female genitalia are easily recognized by the long apophyses; the ductus spermathecae has no spines in contrast to S. ulmivora. Some species in the S. rhamnella group have superficially similar male genitalia, but they have no juxta, and the moths are externally very different. Stigmella multispicata leafmines are characterized by the egg placement on the leaf underside in vein axils, larval exit on leaf underside, and green to blue-green larval color.
Leafmine . Egg always deposited in vein axils on leaf underside, beneath the trichomes; leafmine a long narrow upper side gallery or corridor mine, running through leaf, usually not along veins and not crossing midrib; slightly curved, but  many mines make a U-turn near the end. Frass initially in narrow medial black line, later becoming contorted, brown and almost filling mine. Larval exit on leaf underside.
Larva (Figs 23,25,26). Bright green to blue-green, probably feeding with venter upwards (analogy with S. ulmivora, but not positively observed); head capsule translucent brown. Larvae descending by silken threads, sometimes en masse, spinning a brown cocoon on debris.
Life history. Larvae and leafmines found in China in October; in North America larvae were observed from 15 June to mid-July and from 19 October to 6 November. Moths were found on 26 May and from 8 July throughout August to 6 September (with a peak between 10 and 15 August), a few late records from 22 and 26 September and 2 October. Moths reared from October mines emerged in the laboratory between 25 March and 19 May. The species has at least two annual generations, maybe more. Adults are frequently found at light.
Distribution (Figs 28,29). Presumed to be native in Russia: Primorye and China: Beijing. Almost certainly introduced in North America: Canada: Ontario, Québec; United States: Illinois, Indiana, Iowa, Maryland, Massachusetts, Minnesota, New York, Ohio, Tennessee, Wisconsin. The species has been found in the urban environment, in farmland and in more natural habitats. Table 1 lists the online photographs that we recognized as representing S. multispicata.
Remarks. The extensive collections of Chinese microlepidoptera of Nankai University were searched in vain for this species (Li Houhun, personal communication to EvN). It is possible that specimens can be found in other collections in China, such as the Zoological Institute in Beijing.
Material examined. 19♂, 26♀, 3 sex undetermined, 5 larvae, mines. China: 1 larva (green, dried out, destructively extracted for DNA), China, Beijing, Beijng Botanical garden -Wofosi, E.J. van Figures 3, 4, 10 , 11, 13, 15, 17, 27, 30-33 Nepticula ulmivora Fologne, 1860: 199 Diagnosis. Stigmella ulmivora can be separated from S. multispicata by the slightly larger size, the dark collar, and the antennae with the terminal 7-8 flagellomeres white. In Europe and North America there are no other Stigmella species with the same combination of characters. The male genitalia are very similar to those of S. multispicata, but have a deeper indentation in the uncus, and longer and more distinct sublateral processes of the transtilla. The female differs by the blunt ovipositor and the spiny ductus spermathecae. Leafmines differ from those of S. multispicata by the egg position not being in vein axils; in Europe mines are inseparable from those of S. ulmiphaga (Preissecker, 1942). Due to the variability of mines of S. ulmivora, they sometimes are difficult to separate from those of S. lemniscella (Zeller, 1839), from which the yellow larva emerges through the leaf upper side, not the underside as in S. ulmivora.
Leafmine (Figs 30-33). Egg on leaf underside, against a vein. Mine a highly variable gallery, ranging from short and filled with dense frass in thick leaves (usually in the sun) to long and narrow, often partially following a vein, with frass either linear or becoming contorted, mines sometimes much winding. Larval exit on leaf underside.
Larva (Figure 32). Bright green, feeding with venter upwards; head capsule translucent brown. Larvae descending by silken threads, spinning a brown cocoon on debris or on tree trunks.
Life history. Bivoltine, or possibly partially univoltine in northern parts of Europe. Larvae in June to early July, again in August to November. Adults recorded from May (a single April record) to early July and again in August.
Distribution. Widespread throughout Europe, east to the Volga region in Russia (Johansson andNielsen 1990, van Nieukerken 2017). The species occurs both in natural habitats and on trees in cities, often in large numbers (EJvN, personal observations).

North American Ulmus leafminers
Previously only two Nepticulidae were known to feed on Ulmus in North America: Stigmella apicialbella (Chambers, 1873) and Ectoedemia ulmella (Braun, 1912). This is a much poorer fauna than the seven European species ( van Nieukerken 1986, Puplesis 1994, and in Asia the number is probably still higher, but for several species that are potentially Ulmus feeders the hosts are as yet unknown. Identification of the North American Nepticulidae mines and adults reared from Ulmus is straightforward. For convenience we provide a key that distinguishes these from other insects that form partially or entirely linear mines. Primary blotch mines on elm are formed by additional species of Lepidoptera (Coleophoridae: Coleophora; Gracillariidae: Cameraria, Phyllonorycter), Coleoptera (Buprestidae: Brachys; possibly also Curculionidae: Tachygonus), and Hymenoptera (Tenthredinidae: Fenusa). For a complete key, see Eiseman (2018).  Braun, 1963 has been reared from elm but no details of the mine were recorded (Braun 1963). The above description is based on mines of an undetermined species collected on Ulmus alata in NC by T. S. Feldman and on U. pumila in MA by C. S. Eiseman. -Mine longer than 2 cm and eventually more than 1 mm wide, with extended portions that do not follow veins; frass at least initially forming a narrow central line; larva feeding as a miner throughout its development .  (Braun 1917, Newton andWilkinson 1982). Ulmus alata constitutes a new host record.

Key to North American Ulmus (linear) leafmines
Leafmine (Figs 36-39). Egg on either leaf surface, may be against a vein, but never in leaf axil. Mines linear, usually rather straight, partly following veins, or more con- torted; frass variable, from narrow linear to contorted, green or brown, sometimes completely filling the mine. Larval exit on leaf upperside.
Larva ( Figure 39). Yellow, feeding with dorsum upwards; head capsule brown. Larva spinning a brown cocoon on debris.
Life history. Bivoltine, or possibly trivoltine (Braun 1917). Larvae in June to early July, again in August to October. Adults recorded from April to early July and again in August.
Distribution. Widespread in Eastern North America, positive records from: Canada: New Brunswick, Ontario (BOLD: BIOUG33718-A12), Quebec (van Nieukerken 2018), USA: Alabama*, Connecticut*, Georgia*, Illinois, Indiana, Kentucky (Chambers 1873), Massachusetts, Mississippi*, New York*, North Carolina, Ohio (Braun 1917), Tennessee, Vermont. States without reference are new records: from states with asterisk we have as yet only seen vacated mines; from the other states the occurrence is confirmed by adults or DNA barcodes of larvae.
DNA barcodes. We have three DNA barcodes, all with BIN BOLD: ACG9146 ( Figure 44).
Leafmine . Egg usually on upper leaf surface, often against a vein. Early mine a narrow linear tract with broken narrow linear frass, sometimes filled with frass, at first often winding, then straighter, often following a vein, later mine abruptly widening into an irregular blotch with scattered frass. Larval exit on leaf upper side, or cocoon spun inside mine, often in center of blotch (Braun 1917) (Figure 43).
Larva. Pale yellowish white, feeding with venter upwards, ganglia usually obvious; head capsule translucent brown. Larva spinning brown cocoon inside mine or on debris.
Life history. Larvae found from July to early October, possibly in two generations, but it is also possible that this represents one extended generation. Adults recorded from May to August. From larvae collected in August, the adults emerged the following year.
DNA barcodes. We have six DNA barcodes, all with BIN BOLD: AAJ6172 ( Figure 44).

Discussion
Morphology and DNA barcoding show clearly that the North American and Asian populations of Stigmella multispicata are conspecific, suggesting a recent invasion by this species. We think that the invasion must have been from Asia to North America and not the other way around. Although the North American fauna is not yet completely known, a species that is now so abundant and widespread would have been hard to miss by earlier collectors, but not a single older specimen has been found in collections. Also its occurrence on the introduced Siberian elm and (so far) absence from native elm species speaks in favor of an Asian alien. The close relationship between S. multispicata and the western Palearctic S. ulmivora in the Palearctic S. ulmivora species group (van Nieukerken et al. 2016), further supports its Palearctic origin, even though we have no older records than the empty leafmines collected in Beijing in 1984. The DNA barcoding results show that the North American population is slightly different from the single barcode from China: Beijing. This could mean that the source for the invasion is somewhere else in East Asia.
The date of introduction is difficult to assess. Our oldest North American records are from 2010, but since those already are numerous, the species must have been well established several years before that. Since nepticulids are able to spin their cocoons on many substrates, including tree trunks, branches or other objects, we assume that an introduction of larvae or pupae inside cocoons with plants of Siberian elm, or even another plant or object, is most likely; transport of insect invaders with live plant material is considered a major pathway for exotic herbivores (Liebhold et al. 2012, Brockerhoff andLiebhold 2017). Current global traffic has resulted in several recent examples of introductions of leafminers from North America into Europe and vice versa (van Nieukerken et al. 2012b, 2012d, Landry et al. 2013. As far as we know this is only the second instance of a leafmining moth from Asia introduced into North America, the other instance being Caloptilia triadicae Davis, 2013 on the Chinese Tallow Tree Triadica sebifera (L.) (Davis et al. 2013). We suspect that the apparent rapid spread of S. multispicata may have been aided by the nursery trade. A quick internet search found nurseries in Illinois and Tennessee that will ship Siberian elm plants almost anywhere in the US; although phytosanitary certificates are typically required by states receiving nursery stock, inspection of plants may be less common after they arrive at their destination and even careful inspection of potted or bare-root plants could easily miss the small cocoons. It was recently shown that trade amongst tree nurseries throughout Europe contributed to the spread of the citrus longhorn beetle, Anoplophora chinensis (Forster, 1771) (Eschen et al. 2015).
There is a parallel with another invasive Asian leafminer, the weevil Orchestes steppensis Korotyaev, 2016(Korotyaev 2016, also a leafminer of Siberian elm. The species was first found in North America around 2003, then misidentified as the European O. alni (Linnaeus, 1758) (Anderson et al. 2007, Looney et al. 2012, and has since spread widely over the United States, including the west (Looney et al. 2012). Also in Asia this species has probably been spreading, and is reported as a local pest species in China (Li et al. 2017). It is possible that S. multispicata is also more widespread in Asia, but overlooked due to the limited number of people studying lepidopterous leafminers, and the fact that the mines in small numbers are inconspicuous.
The Siberian elm was introduced in the United States around 1860 and is widely planted for windbreaks and lumber, and now itself an invasive species, particularly in pastures, roadsides and prairies throughout the Midwest and Great Plains regions (Swearingen and Bargeron 2016). It is uncertain whether S. multispicata also will colonize American elm species; the American species of Ulmus are in a phylogenetically different subgenus, Oreoptelea (Spach) Planchon, whereas all Palearctic species belong to subgenus Ulmus L. (Wiegrefe et al. 1994). Larvae of the Siberian Elm specialist Orchestes steppensis have not yet been confirmed on American elm species (Looney et al. 2012), but adults have been found feeding on U. americana (Anderson et al. 2007), and in Ottawa in 2018 we observed some Orchestes mines on U. americana in association with abundant O. steppensis mines on U. pumila. Another invasive Palearctic leafminer, the tenthredinid Fenusa ulmi Sundevall, 1844, has been recently noted to colonize the indigenous North American species of Ulmus (Anonymous 2015).
At the time of writing, Stigmella multispicata already is widespread in eastern North America, from eastern New York to western Iowa, and from Minnesota and Québec to Tennessee. Although the collected material gives a good insight already, online observation websites have been very helpful in providing a quick survey of the distribution, as this source alone was responsible for five of the state records. So far, S. multispicata does not appear to have risen to the status of a damaging pest. Although Siberian elm is no longer widely planted as an amenity tree, many specimens remain in urban and residential areas, where the descending larvae and aesthetic damage could be a concern. At this point the damage by S. multispicata is local, and probably not yet a great problem. In most places, the mines of O. steppensis outnumber those of S. multispicata. We advise to follow a prudent course and monitor both species simultaneously to document their spread and impact.