Observations on the Isturgia limbaria (Fabricius, 1775) / roraria (Fabricius, 1776) complex (Lepidoptera, Geometridae, Ennominae)

The Isturgia limbaria (Fabricius, 1775)/roraria (Fabricius, 1776) complex is analyzed, taking into consideration the external morphology of the adults (coloration and pattern of upperside and underside of wings), the morphology of the female genitalia (signum and lamella postvaginalis), of the tympanal organs and of the preimaginal stages, as well as molecular data. Based on the molecular data and morphology of signum and tympanic organs, the populations of this complex can be divided into two groups, one more western (with the taxa limbaria s. str. and delimbaria), distributed east to north-western Italy and part of Germany, and one more eastern (with the taxa roraria s. str., rablensis and anzascaria), distributed west to northern and eastern Piedmont (Italy) and north-eastern and south-eastern Germany. However, there are no consistent differences between the two groups in the diagnostic characters used until now to identify the two taxa (pattern of the wing upperside and underside). Although there is a considerable genetic distance between these two groups, the correlation between molecular differences and morphological characters (size of the signum and presence ̶ or absence ̶ of a roundish lobe in the bullae tympani) is not completely constant and the two groups of populations are not completely separated from each other: some populations of the northern Apennines (taxon delimbaria) cannot be clearly attributed to one or the other group. In the absence of constant morphological characters associated with the molecular differences and in the presence of Italian populations with intermediate characters, we suggest that the different taxa of this complex be considered as subspecies of the same species, as already proposed by Povolný and Moucha (1957, 1959). However, the data available do not allow definitive clarification of the taxonomic problem and further research is necessary.


Introduction
Fabricius described Phalaena limbaria and P. roraria respectively in 1775 and 1776. Zeller (1868) described Fidonia limbaria var. rablensis (from the Julian Alps), as intermediate between roraria and limbaria, but attributed it to the latter based on the characteristics of the wing underside. Prout (1915) and Wehrli (1940) treated this taxon as Isturgia limbaria var. rablensis, distributed from northern Italy to Albania and to the southern Carpathians. Moucha (1957, 1959) downgraded roraria to a subspecies of limbaria: in their opinion the former is distributed in central and south-eastern Europe, the latter mainly in western Europe, but the populations from the Balkans and the southern Alps (taxon rablensis) have mixed characters and cannot be attributed objectively to one or the other; in addition, both limbaria and roraria are very variable and there are no differences in the genitalia, habitat and foodplant. According to these authors in the Riss-Würm interglacial period there would still have been a single species, which during the Würm glaciation would have been divided into two groups of populations, confined to different glacial refuges (south-western and south-eastern European). Later, after long isolation, the two groups of populations would have come into contact again. In Central Europe the differentiation due to the isolation period would have been sufficient to prevent hybridization between the two forms, while to the south of the Alps the isolation would have been shorter (or would not have been complete) and therefore hybridization would have been possible. According to the same authors a similar example would be represented by Polyommatus eros eros (Ochsenheimer, 1808) and Polyommatus eros eroides (Frivaldsky, 1835) (Lycaenidae).
Also Patočka (2004) considered roraria as a subspecies of limbaria. In contrast, Lehmann (1990) transferred rablensis from a subspecies of limbaria to a subspecies of roraria, based on Gelbrecht's research, who examined mainly the characteristics of the larvae and pupae and the size of the signum in female genitalia (Gelbrecht, pers. comm.); however, Gelbrecht did not examine specimens from the Julian Alps (locus typicus of rablensis), but only from Bulgaria (Vitosha). His point of view was accepted by Viidalepp (1996) and Beshkov (2001), while Huemer and Tarmann (1993), Flamigni et al. (2007) and Leraut (2009) continued to treat this taxon as Isturgia limbaria rablensis. Hausmann (2001) mentioned and figured southern Alpine populations as a model case for hybridization in the evolutionary history of species complexes.
Based on genetic data, Skou and Sihvonen (2015) attributed the populations from the Balkans and the Julian Alps to roraria, but not those from South Tyrol (Monte Baldo), which grouped in a different genetic cluster but are practically identical to those from the Julian Alps in the wing pattern. Without transferring rablensis from the status of subspecies of limbaria to a subspecies of roraria, they suggested that the limbaria/roraria species complex needs a thorough detailed analysis.
The present study aims to provide a contribution to the clarification of the taxonomic problems still unresolved, through an analysis of the different populations that make up this complex.

Material and methods
The study is based on material deposited in the collections listed in Table 1 (see section Taxonomy for details on specimens examined). These observations were also supplemented by the examination of images in the literature and on websites. Unpublished information on the morphology of larvae and pupae have been provided by Jörg Gelbrecht (Königs Wusterhausen, Germany).
Taxon delimitation was based on the combined study of external morphology of the adults (coloration and pattern of the upperside and underside wings), morphology of the female genitalia (signum and lamella postvaginalis) and tympanal organs, as well as of molecular data.
For the extraction of the female genitalia and tympanal organs the abdomen was macerated cold in a 20% KOH solution for 12 hours. Then the longitudinally sectioned abdomen was stained using chlorazol black-alcohol mixture for two minutes. Genitalia were cleaned in water and mounted on a slide in dimethyl hydantoin formaldehyde (DMHF). Tympanal organs were removed from the abdomen, cleaned and mounted on a slide.
Because the signum is stellate with a more or less elliptical shape, in each specimen the smaller diameter and the larger diameter of this ellipse were measured (both up to the ends of the tips), averaging the two diameters.
For the DNA analyses, one leg was removed from dried specimens. DNA extraction, amplification, and se quencing of the barcode region of the mitochondrial cytochrome c oxidase I (COI) gene (658 base pairs) were carried out in the Canadian Centre for DNA Barcoding, Ontario, Canada, using stand ard protocols (Ivanova et al. 2006;deWaard et al. 2008). Unfortunately, the barcoding of only one of our specimens (BC ZSM Lep 97046, from Lombardy) was successful. For this research all the data in the BOLD database (Ratnasingham and Hebert 2007) were also used: data on utilized sequences are summarized in Table 2.
The Barcode Index Number (BIN) is an online framework of BOLD database that clusters barcode sequences algorithmically (Ratnasingham and Hebert 2007).
Information on molecular data was provided by A. Hausmann (SNSB-Bavarian State Collection of Zoology, Munich, Germany), who also examined the specimens from the ZSM collection. Sequence divergence was calculated by Hausmann using the Kimura 2-parameter model (Kimura 1980) and the neighbour-joining algorithm (Saitou and Nei 1987), as implemented in BOLD.

Molecular data
The results from the molecular analysis are shown in Figure A1 (see Appendix I).
The limbaria/roraria complex genetically splits up into three genetic clusters. BOLD:AAF3598 corresponds to the taxa limbaria s. str. and delimbaria; BOLD:ACX8520 to the westernmost populations of the taxon rablensis (Lombardy and Monte Baldo in Italy, Carinthia in Austria); BOLD:AAF3597 to the remaining populations of rablensis (from the Julians Alps to Bulgaria and Romania) and to the taxon roraria s. str.. The minimum distance of AAF3598 from the other two BINs is 3.95%, while that between AAF3597 and ACX8520 is 2.0%. Further information is given in the treatment of single taxa. Figure 1 shows the distribution of the barcode clusters of the complex limbaria/roraria in northern Italy and southern Austria; Figure 2 the distribution of the taxa of the same complex in northern Italy.

Morphological characters
The diagnostic characters of I. limbaria and I. roraria are illustrated by Skou and Sihvonen (2015: 98, text-figs 37-40): wings, upperside irrorated with brown scales, terminal area dark brown in limbaria; wings, upperside striated with brown scales, terminal area light brown in roraria; hindwings, underside with radiating whitish lines in limbaria, without whitish lines in roraria. On the basis of these characters, the taxa limbaria s. str. (Figs 3,8,13,18), delimbaria (Figs 4,9,14,19), anzascaria (Figs 5, 10, 15, 20) and roraria s. str. (Figs 7, 12, 17, 22) can be identified without prob- lems (the first three as limbaria s. l.). In contrast, the identification of the taxon rablensis is problematic: on the basis of the wing underside (Figs 16,21) it should belong to limbaria (hindwing with radiating whitish lines), but the upperside (Figs 6, 11) of some specimens matches roraria, while that of others matches limbaria (above all in the most western part of its range). According   to Fajčik and Slamka (1996) the width of the dark terminal fascia is also a diagnostic character; wider in roraria (when it is present), narrower in limbaria. According to Skou and Sihvonen (2015) there are no differences in the male and female genitalia of limbaria and roraria, but in the female of roraria the lamella postvaginalis is often more sclerotized, with the posterior margin round, but they state that these differences are not clear-cut. According to the same authors the bullae tympani have a roundish lobe in limbaria, but not in roraria; however, they suggest that the reliability of this character needs to be tested on more extensive material from the whole distribution range of the complex.    J. Gelbrecht (pers. comm.) found differences in the size of the signum in the female genitalia. The morphological characters observed in the material examined by us are summarized in Table  3 ( Our observations confirm that the posterior margin of the lamella postvaginalis is often more curved in roraria and rablensis than in limbaria, but this character is variable, especially in the Italian populations, and it does not allow to identify the various taxa with certainty.

Preimaginal stages
J. Gelbrecht (pers. comm.) has bred material of the taxa limbaria s. str. (from Lüneburg Heath in Lower Saxony, Germany), delimbaria (from southern France), rablensis (from Vitosha, Bulgaria) and roraria s. str. (from Dessau in Saxony-Anhalt, Germany). Based on his observation, the larvae of limbaria and delimbaria are identical, darker green than roraria and rablensis, with brownish markings, the head capsule is more marked with black than in roraria and rablensis; the larvae of roraria and rablensis are mostly completely green, rarely with a brownish ground colour (in

rablensis
(♂♀) Forewing often with only a few dark scales (except the costa and the terminal fascia), a little more dense in the hindwing; sometimes both wings scattered with dense groups (or short stripes) of dark scales. Width of the dark terminal fascia generally just over 1 mm.
(♂♀) Fore-and hindwing scattered with dark vertical stripes; ground colour of the wings almost identical or more orange in the fore wing. Whitish radiating streaks of the hindwing present.
(altogether in the taxon rablensis) BOLD:ACX8520 Signum: (♂♀) Fore-and hindwing with identical pattern: ground colour orange scattered with dark vertical stripes (a little more dense on the hind wings).
Hindwing without a trace of whitish radiating streaks.
BOLD:AAF3597 Germany (Bavaria) rablensis), the markings of the head capsule are weak. The shape of the pupal cremaster of delimbaria is similar to limbaria (thinner and a little longer), that of rablensis is similar to roraria (shorter and thicker) (Figs 35-38). Buckler (1897) shows green and brown larvae, both from Great Britain (taxon limbaria s. str.). The morphology of the larval and pupal stages of roraria s. str. is described and illustrated in detail by Blaik and Malkiewicz (2003) (material from south-western Poland): their images of the cremaster match rather well those of Gelbrecht, but not those of Patočka (1986) (from Slovakia, as I. limbaria roraria). Images of the preimaginal stages of the complex limbaria/roraria can also be found on the web: larvae from Netherlands (UKmoths 2019 Gelbrecht also did breeding experiments (pers. comm.). The taxa limbaria and delimbaria crossbred easily; he managed to rear some adults also from the crossing between rablensis (from Bulgaria) and roraria. In laboratory conditions he also obtained a pairing between delimbaria and rablensis: the few eggs hatched but the larvae died immediately.
We have not found any images of the preimaginal stages of the Italian populations and we do not believe that breeding experiments have ever been carried out between them.

Discussion and conclusions
Populations delineated on the basis of molecular data, the size of the signum and the presence (or absence) of a roundish lobe in the bullae tympani, can be divided into two groups: one including  the taxa limbaria s. str. and delimbaria, and the other the taxa roraria s. str. and rablensis. The taxon anzascaria (not yet barcoded) can also be attributed to the latter group as it shares with rablensis and roraria a larger signum (see Table 3) and the absence of a roundish lobe in the bullae tympani; most of the populations of rablensis (from the Julian Alps to Romania) share the same BIN with roraria, while the most western ones (northern Italy, excluding the Julian Alps, and southern Austria) share their own barcode cluster which is much closer to roraria than to delimbaria and limbaria s. str.. This population is, however, not morphologically distinguishable from that from the Julian Alps. However, the diagnostic wing pattern characters until now used to identify the two species limbaria and roraria do not correlate with these two groups: the taxa rablensis and anzascaria are more similar to limbaria in the hindwing underside (whitish radiating streaks present); the pattern of the upperside wing is variable and often intermediate between limbaria and roraria in rablensis (especially in the Italian populations), while in anzascaria it is rather similar to delimbaria (forewing upperside almost free of dark scales, except the costa and the terminal fascia).
Furthermore, the two groups of populations are not completely separated from each other. Some populations of the northern Apennines (taxon delimbaria) cannot be clearly attributed to one or the other group: a barcoded specimen from the "Apuan Alps" in Tuscany (northern Apennines) shares the same BIN with delimbaria and limbaria s. str. (diverging by 1.6% from both the cluster of limbaria s. str. and the specimen of delimbaria from the Ligurian Alps). However, in the Apennines near Bologna there are females with a large signum (as in roraria s. str. and rablensis) and females with a small signum; in the latter area a female lacks the roundish lobes in the bullae tympani (like in roraria s. str. and rablensis), in the single male examined they are barely visible.
The genetic distance between the two groups of populations of the I. limbaria/roraria complex is considerable. However, the correlation between molecular differences and morphological characters is not completely constant, as shown by the example of the Apennine populations just mentioned. In the absence of constant morphological characters associated with the molecular differences and in the presence of Italian populations with intermediate characters, we believe it better, in the current state of knowledge, to consider the different taxa of this complex as subspecies of the same species, as already proposed by Moucha (1957, 1959). However, the available data do not allow a definitive clarification of the taxonomic problem and further research is necessary to consider the conspecificity of the clusters of the different taxa. Additional genetic data are necessary (in particular from northern Piedmont and the Apennines), as well as an in-depth study of the preimaginal stages of the Italian populations, with breeding experiments between them to check their interfertility.
Gelbrecht's breeding experiments suggest the conspecificity between limbaria s. str. and delimbaria on the one hand and between rablensis (from Bulgaria) and roraria s. str. on the other. However, they do not exclude the possible interfertility between Italian populations belonging to different taxa: other experiments would be necessary for this purpose. Also the immutability of differences in the larva and in the pupal cremaster should be confirmed through the study of the Italian populations.
These results emphasize the need to further investigate the patterns of mitochondrial DNA in "hybrid" populations. In our research on Ennominae (cf. Flamigni et al. 2016) we have met other cases of mismatch between DNA barcodes and morphological characters: specimens from southern Piedmont (north-western Italy) are closer to the genetic cluster of Yezognophos serotinaria serotinaria (Denis & Schiffermüller, 1775) even though their genitalia are similar to those of Y. serotinaria serotinoides (Wehrli, 1922); in addition, in the northern Apennines the male genitalia and sometimes also the female ones, although clearly attributable to the "western" taxon serotinoides, show some elements of transition towards the "eastern" nominate serotinaria, showing a parallelism with the above comments on the taxa delimbaria and rablensis. The taxon Psodos chalybaeus Zerny, 1916 (valid at subspecies rank according to Flamigni et al. (2016), downgraded to a synonym of Psodos alticolaria alticolaria Mann, 1853 by Müller et al. (2019), but with the need for further investigations) occurs in a well defined area of the Rhaetian Alps, with a different BIN and a characteristic wing pattern; however, specimens collected in the surroundings of Obergurgl (Austria: North Tyrol) in two very close localities are very similar to each other in wing pattern (intermediate between alticolaria alticolaria and chalybaeus), but they have two different BINs (corresponding to those of these two taxa). The specimens of Hylaea Hübner, 1822, from Molise (central Italy) are often more similar in wing pattern to H. fasciaria (Linnaeus, 1758) (with minimal differences in variable female genitalia), but they share the same BIN as H. mediterranea Sihvonen, Skou, Flamigni, Fiumi & Hausmann, 2014. Also the status of subspecies could be a possible alternative for the taxon mediterranea.

Diagnosis
In the female genitalia signum small (0.7-1 mm). Bullae tympani with a roundish lobe (visible only in one bulla in a female from Spain).
Distribution. Great Britain (Scotland and eastern England, now extinct), part of Germany (cf. Skou and Sihvonen 2015: map on page 97), Benelux, France (excluding the southern Alps), northern Spain (eastern Aragonese, Andorran and Catalan Pyrenees). Our specimens from Valais confirm the occurrence of the species in Switzerland, where only old records existed (cf. Reser 2010).
Remarks. According to Skinner (1984) the specimens from Ross-shire (Scotland) are larger and paler than those from southern England; the difference is also evident in the images in Ford (1972). Synonymy according to Flamigni et al. (2007Flamigni et al. ( , 2016. Diagnosis (Figs 4,9,14,19): Forewing almost free of dark scales (except the costa and the terminal fascia). Hindwing with more dark scales than forewing, but lighter than in most specimens of the nominate subspecies. The dark terminal fascia is generally narrow in the forewing (wide in the form pedemontaria), absent or very reduced in the hindwing (also in the form pedemontaria). The underside is variable, but less contrasted and less sprinkled with black than in most specimens of the nominate subspecies. The taxon pedemontaria (Figs 4b, 14b) is an individual form, sometimes prevalent in some places: it shows a wide, dark terminal fascia on the forewing and a weakly marked underside on the hindwing, which is orange-yellow, scattered with dark vertical stripes, with barely perceptible lighter radiating streaks.

Isturgia limbaria delimbaria (Staudinger, 1892)
In the female genitalia signum small (0.8-0.9 mm, exceptionally 1.2 mm in a female from the Apennines near Bologna: see Table 3). Bullae tympani with a roundish lobe (absent in a female from the same locality: see Table 3). The specimens from northern Piedmont (Graian and Pennine Alps) were not dissected.
Distribution. France SE, Italy (Piedmont, Liguria, Tuscany, Emilia-Romagna: see Fig. 2). According to Leraut (2009) its distribution in France is limited to the southern Alps; we have confirmation for the following departments: Drôme: Montagne de Chamouse (Skou and Sihvonen 2015: pl. 2, fig. 18c, e), Vaucluse, Alpes-de-Haute-Provence, Alpes-Maritimes, Var. In Piedmont it occurs mainly in the south, from the Maritime Alps to the Apennines; according to Flamigni et al. (2007) it occurs also further north, up to the Graian Alps and the surroundings of Biella (Pennine Alps); the specimens from the Biella area were examined (without dissection of the genitalia) by the first author during the preparatory work on the first volume on the Italian Ennominae. The geographic position would suggest that specimens from Corsica also belong to this subspecies: this will need to be verified through the examination of the collected material.
Molecular data. The two barcoded specimens (both from Italy) diverge from those of the taxon limbaria s. str. (from Germany and Upper Normandy) respectively by 1.2% (Ligurian Alps) and 1.6% (northern Apennines); however, they diverge from one another by 1.6% and cannot be regarded as a "common separate cluster".
Remarks. The distribution of the taxa delimbaria and rablensis in the northern Piedmont is similar to that of Crocota pseudotinctaria Leraut, 1999 andC. tinctaria (Hübner, 1799): delimbaria and pseudotinctaria occur in the surroundings of Biella, while rablensis and both pseudotinctaria and tinctaria occur further to the north-east, in Campello Monti (tinctaria also occurs in Lombardy, like rablensis). The analysis of the DNA of the northern Piedmontese populations could help to better understand the relationship between the three taxa present in this area (delimbaria, anzascaria, rablensis); the barcoding of a specimen from Campello Monti was unfortunately not successful.

Isturgia limbaria anzascaria (Staudinger, 1892)
Fidonia limbaria var. anzascaria Staudinger, 1892, Dt. ent. Z. Iris 5: 198, [Italy, northern Piedmont] Diagnosis (Figs 5,10,15,20). Large, light yellow. Forewing upperside almost free of dark scales (except the costa and the terminal fascia), hindwing upperside with a few dark scales. Forewing dark terminal fascia narrow in the male and reduced to a few blackish dots in the female according to the original description; this fascia is present in both sexes in our specimens, but it is relatively narrow in relation to their size. Hindwing dark terminal fascia rather narrow or reduced to a few scales, especially in the female. Hindwing underside yellow, scattered with a few dark vertical stripes; whitish radiating streaks present but not very evident. Although Prout (1915) and Wehrli (1940) reported also the ssp. rablensis from Macugnaga, our specimens from this locality correspond well with the description of Staudinger, differing from rablensis by the larger size (wingspan 27-30 mm in anzascaria, 23-26 mm in the Italian specimens of rablensis) and by the purest yellow, almost without dark scales; it will be possible to definitively confirm the validity of the subspecies examining a greater number of specimens.
In the only female examined signum large (1.4 mm). Bullae tympani without roundish lobe. Distribution. Valle Anzasca (Italy: northern Piedmont), on the southern side of the Monte Rosa massif (Fig. 2).
Remarks. Not yet barcoded. The size of the signum and the absence of a roundish lobe in the bullae tympani indicate a greater similarity with rablensis and roraria, but the wing upperside is closer to delimbaria, but the hindwing has a dark terminal fascia. Diagnosis (Figs 6,11,16,21). Upperside more or less scattered with groups of dark scales (or dark lines short or longer), scattered or more dense; in some specimens (mainly in the westernmost part of the distribution area) there are only a few dark scales in the forewing (except on the costa and in the terminal fascia), a little more dense in the hindwing; in other specimens (also among the most western ones) both wings are scattered with dense groups of dark scales; dark terminal fascia generally narrow, wider in Bulgarian specimens, in the hindwing it can be present or absent. Underside with fore-and hindwings more or less different in pattern and colour (identical in roraria s. str.): both wings scattered with dark vertical lines, but colour generally more orange-yellow in the forewing, more whitish yellow in the hindwing, whitish radiating streak(s) of the hindwing present (sometimes barely perceptible).
Distribution. Northern Italy (see Fig. 2), Austria (Huemer, 2013: Carinthia, Styria), Slovenia, Serbia, Macedonia, Kosovo, Albania, Greece (Western Macedonia), Bulgaria, Romania. The taxon rablensis was recorded from Albania (Djalica e Lumës) by Rebel and Zerny (1934) and by Povolný and Moucha (1957: fig. 57); the first two authors also mention Žljeb in Kosovo; its occurrence in Romania is confirmed from the Bucegi Mountains (see also Popescu-Gorj (1995)) and the Nemira Mountains; the records of I. roraria s.l. from other areas of Romania (in particular from the Cluj-Napoca area: cf. Rákosy et al. (2016)) must be verified by examining the underside of the specimens. The species also occurs in Croatia and Bosnia and Herzegovina (Skou and Sihvonen 2015), but we have not examined material from these countries. Viidalepp (1996) attributes the populations from Transcaucasus (Georgia) to this subspecies (as I. roraria rablensis). Molecular data. The westernmost populations (Lombardy and Monte Baldo in Italy, Carinthia in Austria) correspond to a separate BIN, at a distance of 2.0%, while all the others (from the Julians Alps to Bulgaria and Romania) share the same BIN of the following subspecies; however, no constant morphological character corresponds to these genetic differences and it is not possible to distinguish the specimens of the more western regions from those of the Julian Alps (both populations are very variable). The exact border between these two BINs is not known, since no specimens from the Carnic Prealps have been barcoded. Three specimens from Romania diverge into a separate cluster, but they are morphologically very similar to those from Bulgaria. (Fabricius, 1776)  Diagnosis (Figs 7,12,17,22). Ground colour of the upperside deep yellow or orange, scattered with dark vertical lines; the dark terminal fascia (when present) is generally wide; however, the upperside is often identical to that of the previous subspecies and the differentiation from it is possible only by examining the underside. On the underside fore-and hindwing with identical colour and pattern: ground colour orange-yellow or orange scattered with dark vertical stripes (a little more dense on the hindwings); hindwing without a trace of whitish radiating streaks (sometimes a streak of a slightly lighter yellow, for example in specimens from Poland and Slovakia).

Isturgia limbaria roraria
On the upperside, the dark terminal fascia is variable: according to Povolný and Moucha (1957) in Slovakia most of the individuals from the Ore Mountains have this fascia less developed than those from Vihorlat Mountains and in the first locality it is absent in a high percentage of females (but also males). In Poland (Blaik 2003) this fascia is wide and always present in the males (more or less distinct), almost always indistinct or absent in the females; specimens were collected in a locality with a large area of the forewing upperside without dark markings.
In the female genitalia signum large (1.2-1.4 mm). Bullae tympani without roundish lobe. Distribution. North-eastern and south-eastern Germany (cf. Skou and Sihvonen 2015: map on page 99), Austria (Huemer 2013: Lower Austria), southern Poland, Czech Republic, Slovakia, north-eastern Hungary. For Romania, see the previous subspecies. Viidalepp (1996) attributes to roraria s. str. the populations from Ukraine, south-eastern Russia and southern Urals. Govi (Forlì, Italy) and Melania Massaro and Rossana Pisoni (Museo Civico di Scienze naturali "E. Caffi", Bergamo, Italy) for loan of material and the Canadian Centre for DNA Barcoding (Ontario, Canada) that carried out the DNA analyses, as well as all museum curators for providing material for our study.