Incurvaria pirinella sp. nov., a new species of the vetulella species-group (Lepidoptera, Incurvariidae) from Bulgaria, with release of DNA barcodes for European species of Incurvaria

Incurvaria pirinella Junnilainen, Kaitila & Mutanen, sp. nov. is described from Bulgaria based on specimens collected by netting and artificial light from several low-elevation localities in Bulgaria. The species is morphologically and genetically most similar to I. triglavensis Hauder, 1912. Differences between these two species are present in external appearance and genitalia of both sexes. Additionally, I. pirinella shows a distance of 4.74% to its nearest neighbour I. triglavensis in the standard DNA barcoding marker (COI-5P). We provide preliminary observations of phylogenetic affinities of European Incurvaria and briefly discuss habitat preferences of some species. All species have distinct barcodes with minimum K2P divergences between species averaging 7.05% (range 1.2–12.8%). A world checklist of Incurvaria Haworth, 1828 is provided and DNA barcodes for all European species are here released. Finally, we document morphological variation in male genitalia within I. triglavensis Hauder, 1912.


Introduction
During several entomological expeditions to the southern Pirin mountain range in Bulgaria, the authors JJ and JPK collected specimens of Incurvaria Haworth, 1828, which originally were believed to belong to Incurvaria triglavensis Hauder, 1912, described from the Triglav mountain in Slovenia. Later, however, we noticed constant differences in the external habitus between our specimens and those of I. triglavensis as illustrated by Huemer (1993). We also observed that the Bulgarian taxon occurs in low-elevation habitats with mainly Mediterranean vegetation, whereas other species of the vetulella-group prefer alpine habitats in Central and Southern Europe. Therefore, a tissue sample of the Bulgarian taxon was sequenced for the DNA barcode region of the mitochondrial COI gene, which turned out to differ clearly from all other European species of Incurvaria. Examination of the genitalia of Bulgarian specimens showed differences from I. triglavensis in both male and female genitalia. Therefore, we describe the taxon here as a new species. We also provide DNA barcodes for all European species of Incurvaria, as well as for three other European Incurvariidae.
representing the subfamily Crinopteryginae (all other European species belong to the nominal subfamily). Therefore, only two European species of the family, Paraclemensia cyanella (Zeller, 1850) and Vespina slovaciella (Zagulajev & Tokár, 1990) are not included. Kimura 2-parameter (K2P) DNA barcode divergences were examined using the BOLD v. 4beta barcode gap analysis tool (Ratnasingham and Hebert 2007). A Neighbor-Joining (NJ) tree was constructed with Mega 7 (Kumar et al. 2016) under the Kimura 2-parameter nucleotide substitution model. A Maximum Likelihood (ML) analysis was conducted with Mega 7 under the GTR+G model for nucleotide substitutions. The node confidences were examined with 500 bootstrap replicates. Both trees were rooted on Crinopteryx familiella.
Terminology. The morphological terminology used here mainly follows Huemer (1993) and Davis (1998). In our comparisons of the newly described species with previously described species of the vetulella-group, we largely rely on the thorough revision by Huemer (1993).
To the best of our knowledge, only four other species of Incurvaria are known from the Holarctic region: I. evocata (Meyrick, 1924) described from India, (Nielsen 1982) I. takeuchii Issiki, 1957and I. alniella (Issiki, 1957 described from Japan (Okamoto and Hirowatari 2004) and I. similella Schmitz, 1969 described from the Caucasus (Schmitz 1969). These species do not belong to the I. vetulella species-group. In North America, only I. vetulella and I. circulella are known (Pohl et.al. 2016(Pohl et.al. , 2018. Incurvaria muchei Soffner, 1969 from the Caucasus has been transferred to the genus Alloclemensia Nielsen, 1981 andconsidered as synonym of A. devotella (Rebel, 1893) (Nielsen 1981). The poorly known species I. triangulifera Tengström, 1869 has recently been considered to be a junior synonym of Lampronia provectella Heyden, 1865 (van Nieukerken andKozlov 2017) and thus belongs to the family Prodoxidae. Incurvaria kivatshella Kutenkova, 1987, described from a single male specimen, was considered not to belong to the vetulella-group due to the absence of the distal projection of the phallus (Huemer 1993), but Kozlov (1996) synonymised it with I. vetulella. Incurvaria brigantinella Amsel, 1961 was described from Germany, Baden-Württemberg, Bodensee, based on a single type specimen, which is probably lost; it is therefore considered to a be nomen dubium and was not included in Fauna Europaea (van Nieukerken and Kozlov 2017). The original description including a black and white drawing of the holotype clearly shows that I. brigantinella does not belong to the vetulella species-group. It resembles I. praelatella in having rather similar forewing markings and distinctly white forewing fringe.
(original combination Adela circulella Zetterstedt, 1839) I. evocata (Meyrick, 1924  Etymology. The name of the new species is derived from the Pirin mountain range, where the new species is widely distributed. Diagnosis. Considering similar forewing ground colour and markings, Incurvaria pirinella (Figs 3-5) is externally closest to I. circulella (Fig. 14), but is easy to separate from it by the dark grey-brown fringe and distinct white spot in the forewing fringe present in I. circulella. Incurvaria pirinella is also widely allopatric as I. circulella occurs only in northern latitudes in Europe, e.g. the northern part of Sweden and Finland.
Although I. triglavensis is rather variable externally, it is always easy to separate from I. pirinella based on its relatively narrow forewing shape, paler ground colour and differences in forewing markings (Figs 6-11); see details in Suppl. material 2.
Incurvaria pirinella is easy to separate externally from Scandinavian I. vetulella (Figs 12-13) by a white narrow dorsal spot situated 2/5 from base and extending more than halfway across the forewing in I. pirinella. North European I. vetulella also typically have a distinctly broader dorsal spot, sparser forewing scaling and usually a pale medial area in the fringe, which is always absent in I. pirinella.
Male genitalia of I. pirinella are most similar to those of I. triglavensis, but small constant differences are present. I. pirinella has a stouter valva and vinculum. The valva of I. pirinella is broader in middle and the margin of the sacculus is more bulged (Figs 21-22). The vinculum of I. pirinella is relatively short and broad with clearly concave lateral margins. In I. triglavensis, the apex of the valva is more elongated and the vinculum is clearly longer and more slender with straight or slightly concave lateral margins 17e,19b,20e). The lateral arms of the transtilla are more robust (Fig. 20a), and submedial anterior projections shorter in I. pirinella (Figs 1c, 20d). The length of the transtilla's medial knob with anterior submedial projection is 0.26-0.27 mm in I. pirinella and 0.29-0.33 mm in I. triglavensis. The juxta is longer (0.95-0.98 mm) in I. pirinella than in I. triglavensis (0.8-0.85 mm) (Fig. 16, Suppl. material 1). Also, the ratio between the length of the juxta versus the transtilla's medial knob with its anterior projections is different between the species, being 3.56-3.77 in I. pirinella and 2.76-2.83 in I. triglavensis (see Suppl. material 1). The shape and angle of the distal thorn/hook-shaped projection of the phallus show variation, that is at least partly due to the position of the phallus in genitalia preparations (Figs 15d1, 18a-b); thus, this does not provide a good characteristic. However, the ratio between the maximum length of phallus versus the length of the distal projection seems to be different between the species as well (Figs 18a-b). In the examined samples, it was 3.34-3.78 for I. pirinella and 3.82-4.07 for I. triglavensis. Shape, size and ratio between length/width of abdominal segments T8 and S8 are different: See shape in Figs 23, 24, 25 and size and ratio measurement values in Suppl. material 1. Although I. triglavensis is a rather variable species, additional slightly different but sometimes   overlapping features are found in other parts: details shown in Suppl. material 1. Genitalic structures of both sexes are overall more strongly sclerotised in I. pirinella.
In the female genitalia, differences between I. triglavensis and I. pirinella are found in size of the oviscapt, especially distance from tip to bottom of basal excavation. In the examined material, it was 0.12 mm in I. pirinella and 0.09-0.11 mm in I. triglavensis (Fig. 2). The length of the apophyses posteriores is 2.70 mm in I. pirinella and 2.48-2.55 mm in I. triglavensis. Segment VIII also shows differences, but this feature shows overlap and therefore is not useful for determination of a single specimen. Differences between membranous parts of the bursa are not included as diagnostic characteristics, because they may show variation due to the age of specimens and may be altered after copulation. For more details, see Suppl. material 1 for morphometrics of genitalia of I. triglavensis and I. pirinella and photos of female genitalia (Figs 26-28, 30) of both species.
Description. Male (Figs 3-4). Wingspan 16.5-17.0 mm. Labial palp covered with brown and ochre-yellow scales, most pronounced on inner surface, 2 nd segment equal in length to diameter    of eye, pointed forward, 3 rd segment half as long, turned upward. Antenna faintly ringed with dark brown and paler greyish-brown scales, slightly serrate, reaching three-quarters length of forewing. Head with yellow tufted scales. Thorax and tegula brown. Forewing relatively broad and roundish, ground colour grey-brown; two subtriangular yellowish-white dorsal spots, first narrower, 2/5 from base, extending over half width of forewing, second smaller, close to termen; a small yellowish-white triangular costal spot rather close to apex, sometimes a second usually smaller quadrangular yellowish-white spot close to the first one; fringe as ground colour. Hindwing and fringe fuscous. Abdomen and legs pale greyish-brown. Ventral side of forewing and fringe fuscous throughout except for distinct yellowish-white costal and dorsal markings.  Female (Fig. 5). Wingspan 16.0-16.5 mm. Differs from male in darker grey-brown ground colour of forewings; two subapical costal spots always present; female dorsal spots larger and extending even more than 2/3 across forewing towards costa, the second white dorsal spot at tornus extending across the wing and sometimes reaching inner costal spot forming a complete fascia. Antennae not serrate.
Variation. The external habitus of both sexes is rather constant. Females are darker overall than males. One male has a second yellowish white costal spot near forewing apex. Sometimes the tornal spot and inner subapical costal spot form a complete fascia in the female forewing. Genitalic structures of both sexes are only slightly variable, whereas those of I. triglavensis show significant variation both externally and in genitalic structures of both sexes.
Distribution. Known from four different localities in the south-west corner of Bulgaria around the Struma river valley and its adjacent regions, which all belong to the Blagoevgrad district and the Pirin mountain range. The elevational range is wide: 200-1200 m, at least.
Biology. Many specimens were captured with insect nets during daytime, and females especially were disturbed in the evening from shrubs such as Rosa L. Males were observed swarming early in the morning, but they were also caught with artificial light late at night, which is an unusual collecting method for the species of the vetulella-group in the Alps (P. Huemer pers. comm.) and northern Europe (own observations, and although summer nights are light in the North, I. circulella comes readily to light). Incurvaria pirinella is an early species, flying at lower elevations in early May and at higher elevations in late May and June. The biology of the early stages  remains unknown. The food-plant of I. vetulella is reported to be Vaccinium L. (Ericaceae), especially V. myrtillus L. (Klimesch 1961, Wojtusiak 1976), but the life history details for the other members of the vetulella-group remain largely unknown. It is suspected that the other European members of the group are also associated with Vaccinium (Huemer 1993). However, no species of Vaccinium or Erica L. are present, at least not in most localities where we have found I. pirinella, suggesting that they are not its food-plants. Specimens of I. pirinella have been collected from completely different habitats from the other species of the vetulella-group, which in the Alps are always found in alpine areas usually above timberline and especially from north-facing slopes. In Scandinavia, I. vetulella occurs in alpine and subalpine habitats, but also boreal coniferous forest and boggy habitats. Incurvaria pirinella seems to prefer open xerothermic south-facing slopes with rich deciduous forest edges of Quercus L., and Fagus L. (Fig. 34), and a variety of shrubs, including Crataegus L., Corylus L., and Rosa, sometimes also Cornus L., Malus Mill., Pyrus L., and Chamaecytisus Link. Along with I. pirinella, I. masculella, well known as an early flying species, was collected in numbers.

Discussion
To our knowledge, detailed illustrations of female genitalia of Incurvaria have not been published before. Only the oviscapt plate was illustrated in earlier publications on the Incurvaria vetulella group

36
©Societas Europaea Lepidopterologica; download unter http://www.soceurlep.eu/ und www.zobodat.at (Laasonen et. al. 1981, Huemer 1993. The size and the shape of the oviscapt appear useful new characteristics, but the number and shape of lateral teeth of oviscapt seems to vary quite a lot even within species. We found also significant differences in the length of apophyses posteriores and minor differences in the shape and length of segment VIII and the shape and size of abdominal plates T8 and S8. DNA barcodes of I. vetulella fall into two clearly separate clusters within Europe, with I. ploessli being nested within these clusters, rendering I. vetulella paraphyletic. The first group contains samples from North European countries and the second samples from the Alps. We suspect that the I. vetulella population in the Alps actually represents an undescribed taxon, as I. vetulella shows an overall 3.46% intraspecific divergence in DNA barcodes in Europe. Genomic-based studies on whether these clusters represent a different species are underway (Huemer and Mutanen in prep.). In Central Europe, the vetulella-group contains also I. triglavensis and I. ploessli. We consider I. pirinella as a new member of this species group. Each of the Central European species appears to have a rather restricted distribution. Although I. triglavensis has also been mentioned from Balkan countries such as Bosnia and Herzegovina (Huemer 1993), further material is needed to clarify the status of the Balkan specimens, because the single barcoded specimen from Montenegro deviates significantly from other barcoded samples of I. triglavensis. Moreover, a new DNA barcode cluster was recently found from the Saualpen in Austria by Huemer (personal communication). These new findings indicate that the I. vetulella-group should be revised again.
Originally, genetically deviant specimens of I. triglavensis from the Alps and Montenegro were included in the morphometric analyses. While some morphological differences between them and other I. triglavensis as well as I. pirinella were observed, we excluded them from the final analyses (see Figs 35-36) as this was beyond the scope of our study. The illustrated I. triglavensis samples were verified not to include specimens of this putatively undescribed species.
In northern Europe, the superficially similar I. circulella occurs sympatrically with I. vetulella. Despite superficial resemblance of these species, Huemer (1993) considered that I. circulella does not belong to the vetulella-group. This is supported by the maximum Likelihood analysis of DNA barcodes, as I. circulella appears also genetically relatively distant to the species of the vetulella-group. It forms a sister lineage to the vetulella-group + I. oehlmanniella. Based on the DNA barcodes, I. oehlmanniella is nested within the vetulella-group (Fig. 33). However, a phylogenetic analysis based on DNA barcodes alone may be misleading due to a single locus and due to the small number of characters, and further studies, preferably including nuclear markers, are needed. Interestingly, Bengtsson et al. (2008) consider I. circulella as likely to feed on Betula spp., which we find unlikely as this species is often abundant in spruce "taiga" forests with plenty of Vaccinium myrtillus but no or little Betula L. (Betulaceae). Misleading habitat and food plant association may be the reason why Bengtsson et al. (2008) consider I. circulella a rare species. In Finnish Lapland it is common, especially in spruce forest habitats ("taiga" forests), but does not occur or is rare in the subalpine and alpine habitats where I. vetulella is often abundant. The males of both I. circulella and I. vetulella swarm most actively late at night and in the early morning.
In our Maximum likelihood analysis (Fig. 33), Incurvaria appears to be monophyletic, although with poor support as the distances between the species are generally large. Also, several of its basal branches are short and groupings are overall poorly supported. The vetulella-group together with I. oehlmanniella is well supported.