Agrotis fatidica ( Hübner , 1824 ) species-group revisited , with description of two new species from the Alps and the Pyrenees ( Lepidoptera , Noctuidae )

An integrative taxonomic analysis of the European species of the Agrotis fatidica species-group is presented with special reference to the European sister taxa of A. fatidica (Hübner, 1824); in addition, a general overview of the entire species-group is given. The remarkable differences found in the barcodes of the Central and Western European populations of A. fatidica (sensu lato) led us to recognise isolated species of the A. fatidica complex. Two new species, A. mayrorum sp. n. (Northern Italy and the French Alps) and A. mazeli sp. n. (French Pyrenees) are described. The neotype of A. fatidica is designated. Agrotis luehri von Mentzer & Moberg, 1987 is treated as a subspecies of A. fatidica (stat. n.).


Introduction
Species identification and delimitation is one of the major issues in taxonomic work.In megadiverse groups such as Lepidoptera the taxonomic treatment was almost exclusively based on morphology for nearly 250 years, with integrative approaches being rare.Recently developed molecular methods, particularly DNA-barcoding, have changed this tradition dramatically (Hebert et al. 2003(Hebert et al. , 2009)).Identification of species nowadays increasingly considers molecular data, which have proved an excellent method for delimitation of large numbers of Lepidoptera on different continents (e.g.Hausmann et al. 2011aHausmann et al. , 2011b;;Hebert et al. 2013;Huemer et al. 2014b;Zahiri et al. 2014).The formation of an increasingly complete DNA barcode-library is providing a basic tool to detect cryptic diversity (Huemer 2011;Mutanen et al. 2012Mutanen et al. , 2013;;Huemer and Hebert 2016) and will be a pre-requisite for future taxonomic work.
In the year 2011 the Tiroler Landesmuseum in cooperation with The International Barcode of Life (iBOL) started a campaign to sequence the Lepidoptera fauna of the entire alpine arc.To date, more than 20,000 DNA barcodes of >3500 species have been made available.Over recent years, several overlooked cryptic species have been detected (e.g.Huemer and Hausmann 2009;Huemer and Hebert 2011;Huemer et al. 2013;Huemer et al. 2014aHuemer et al. , 2014c;;Huemer and Timossi 2014;Buchner 2015;Huemer and Mutanen 2015;Kirichenko et al. 2015;Baldizzone and Landry 2016;Gilligan et al. 2016;Kozlov et al. 2016).In this paper we reveal a striking cryptic diversity in the taxonomically difficult genus Agrotis Ochsenheimer, 1816.

Morphology and material
Extensive generic descriptions and diagnoses of Holarctic species of the Agrotis fatidica (Hübner, 1824) species-group have been published in several reviews (von Mentzer and Moberg 1987;Fibiger 1990Fibiger , 1997;;Nupponen et al. 2001;Fibiger et al. 2010), and are not repeated here.
Our study is based on about 1,500 specimens of the Agrotis fatidica species-group.Most of the material was traditionally set and dried, although a few specimens are pinned but remained unset.Genitalia preparations followed standard techniques for Noctuoidea, including everting the vesica.

DNA sequencing and analysis
Full-length lepidopteran DNA barcode sequences are a 658 base-pair (bp) long segment of the 5' terminus of the mitochondrial COI gene (cytochrome c oxidase 1).DNA extracts were prepared from a dried leg of each of 22 specimens of the Agrotis fatidica species-group from Europe (12 A. fatidica, 2 A. proverai Zilli, Fibiger, Ronkay & Yela, 2010, 5 A. mayrorum sp. n., 3 A. mazeli sp. n., Nupponen, Ahola & Kullberg, 2001 Agrotis ruta (Eversmann, 1851) xylographa Boursin, 1947 ?Agrotis patula (Walker, 1857) (status unclear) septentrionalis Möschler, 1862 kurodakeana (Matsumura, 1927) subinformis (Bryk, 1941) Taxa of the Agrotis fatidica (Hübner, 1824) species-group and the history of their interpretation The taxonomy of the Agrotis fatidica (Hübner, 1824) species-group has always been difficult to resolve.There is much similarity in habitus between taxa, slightly polymorphic facies shown by allopatric populations of a putatively identical "species" and remarkable uniformity in genitalia characters between populations.Some of the taxonomic problems concerning the ranking of some populations of the A. fatidica species-complex have been partly or satisfactorily solved over the last twenty years (e.g. Fibiger 1997;von Mentzer and Moberg 1987;Nupponen et al. 2001;Kononenko 2005;Fibiger et al. 2010;Lafontaine and Schmidt 2010).The main issues, which remain open and are reviewed here, will be solved probably only after detailed genetic investigations.The taxa of the A. fatidica species-group mentioned in the older and recent literature sources are introduced briefly below, indicating the short history of their interpretation and the basic taxonomic and biogeographical data that led to their relegation.
The type material is considered lost (Horn andKahle 1935-1937) and we here designate a neotype from the eastern Alps in order to maintain stability of the nomenclature and ground the traditional interpretation of "fatidica".Such a designation follows the qualifying conditions of section 75.3 of the ICZN Code, 4 th edition.We select an Austrian specimen, which is barcoded and which has a sequence that fits well with the DNA barcodes of the other specimens we have investigated.The sequence is in fact identical to those of four specimens.Gen. figs 1-3.1-2 Agrotis fatidica, male genitalia, 3 Agrotis proverai, male genitalia, paratype.
Apart from the Alps, published records of A. fatidica-like taxa have come from the Pyrenees, the Balkan Peninsula (Albania, Macedonia and Bulgaria) and the Caucasus region.Furthermore, two old specimens from the Frivaldszky collection (HNHM), labelled as "Corsica", are practically identical with those of the specimens from the French and Ligurian Alps.The report of "A.fatidica" from Corsica is probably based on these, almost certainly mislabelled, specimens, and the occurrence of a fatidica-group species in the island of Corsica has never been confirmed (cf.Brusseaux and Nel 2004).
Von Mentzer and Moberg (1987) excluded the presence of A. fatidica from the Ural region; however, the A. fatidica specimen illustrated in Noctuidae Europaeae 1: pl. 7, fig.2, (Fibiger 1991) is from the Urals.Moreover, an allied species, A. iremeli, is present in that region.
Agrotis fatidica var.monedula is just an individual, dark form from the Ortler Region (South Tyrol) and has no taxonomic relevance.Contrary to the statement of Poole (1989), the type-locality is mentioned in the description.
Agrotis incurva, a junior synonym of A. fatidica, was described from three highly variably coloured and shaped females and an unstated number of males collected on the high peaks of the Tyrolean Alps by Stenz (Herrich-Schäffer 1852).In the original description, brachyptery in the female sex was considered as probably the result of malformation.
It remains unclear as to whether some or all the eastern European and Asiatic populations currently referred to A. fatidica, either named (e.g.bombycia Eversmann, 1851; possibly also sajana Corti, 1932) or unnamed (e.g. the Balkanic, Caucasian and Tibetan populations), are actually conspecific with the Alpine ones or not.Material from the Altai Mts which corresponds with A. fatidica in external morphology falls within the range of intraspecific DNA barcode variation and is therefore considered as conspecific.However, the barcodes of certain Agrotis specimens, identified by their external features as A. trifurcula (Staudinger, 1892) andA. ruta (Eversmann, 1851), show no divergence from the barcodes of the European populations of A. fatidica (see also under A. trifurcula and A. patula).
Clearly, subspecific division of A. fatidica requires further study, especially with respect to the south-eastern and eastern European populations which are poorly represented in collections.

Agrotis fatidica luehri von Mentzer & Moberg, 1987 Figs 9-16
Originally, A. luehri was considered to be the Arctic sister-species of A. fatidica in which the male has shorter and broader-based lamellae of the antenna, and the female is longer-winged and with thinner antennomeres.It was described from Norway and compared with a long series of "A.fatidica" specimens from both Norway and the Alps (von Mentzer and Moberg 1987).Subsequently, however, all Norwegian populations of "A.fatidica" were said to represent A. luehri (Fibiger 1997, in Noctuidae Europaeae 3: 81).This situation is questionable as it implies that the latter has been diagnosed also in comparison to itself.The distinction of A. luehri from the alpine A. fatidica was accepted by Fibiger et al. (2010) on the basis of the characters noted above, but the logical incongruence relating to the Norwegian populations can only be explained by admitting in A. luehri a broader range of variation in morphological characters than that fixed in the original description.
The DNA barcode of 3 specimens from Norway show, however, no difference compared with those of 12 specimens from the eastern Alps (see Table 1).As it seems unlikely that two species occur together in Norway, we consider all the DNA barcodes to be sequences of A. luehri.There is some variation in DNA barcodes of A. fatidica (within some 0.4%), but the three known DNA barcodes of A. luehri nest within this range, matching exactly those of certain specimens of A. fatidica from Austria and Switzerland.
Thus, irrespective of the disputed diagnostic characters in morphology, the COI sequences support the conspecificity of A. luehri with A. fatidica.The detailed investigation of the Scandinavian specimens show, however, differences in the colouration and the wing shape between the nordic and the alpine populations (the ssp.luehri is darker and more unicolorous than the ssp.fatidica, especially the hindwings are more dark grey suffused, the melanistic specimens appear more frequently; the forewings of the ssp.luehri are narrower and more elongated).These differences and the long and complete isolation of A. fatidica and A. luehri support their distinction at subspecies level, A. fatidica luehri stat.rev.(see also Aarvik et al. 2017).

Agrotis proverai Zilli, Fibiger, Ronkay & Yela, 2010
Figs 17-20; Gen. fig. 3 The taxon is endemic to Central Italy (Abruzzo region; on the massif of Gran Sasso d'Italia), and is the only population of the A. fatidica species-group outside of the Alps ever found on the Italian Peninsula.Only males have been collected, and it seems likely that the females of this isolated species are brachypterous, as is the case in several other members of the species-group.The male genitalia and various other structural characters correspond with typical A. fatidica and its close relatives.Furthermore, DNA barcodes show a high degree of similarity to typical A. fatidica from the Eastern Alps, with only ca.0.5% divergence (n=2), and they belong to the same BIN.The genetic distance from A. mayrorum and A. mazeli is much greater, minimally 1.55% and 1.71%, respectively.
The habitus of the moths from the Abruzzi region (especially the narrower forewings and the stronger and more prominent noctuid pattern), some allometric differences, and some incongruence in the variation of colours between areas of forewing compared with A. fatidica and A. mayrorum from the Alps, indicate that the Central Italian population has differentiated sufficiently to qualify as a distinct species.17 Certain features of the SW Alpine (and the two "Corsican") specimens, distinguish them from the typical eastern Alpine A. fatidica, as well as from the A. characteristica species-complex and A. trifurcula.The main diagnostic features are the proportionally longer and slenderer branches of the male antenna, the pale hindwing of the individuals, the enhanced pale lining of the cubital vein of forewing, the generally obsolete postmedial line, and the somewhat shiny flat, not fluffy scaling.These features are partly shared with examples from Bulgaria and specimens of A. mazeli from the  NE Pyrenees, while the fatidica-like moths occurring in the Ural region have generally dark hindwings, and their forewing pattern also resemble more closely the typical eastern Alpine A. fatidica.

Agrotis mayrorum sp. n.
In the male genitalia, the valvae of A. mayrorum are comparatively the longest and narrowest within the European species of the fatidica complex, having a less dilated distal section, only slightly convex distal dorsal margin and reduced membranous ventral edge along the sacculus.The clasper (harpe) is short and broad, being distinctively shorter than in A. mazeli and on average shorter and apically more rounded than in A. fatidica; the juxta is also somewhat larger and broader, more pentagonal than in A. fatidica and A. proverai; and the dentated plate of the phallus is stronger and more densely spiny in A. mayrorum than in the related species.
The female genitalia of A. mayrorum has, in comparison with A. fatidica, proportionally a longer ovipositor, longer apophyses anteriores and a longer appendix bursae; while the ovipositor, the ductus bursae, and the appendix bursae are proportionally longer than in A. mazeli.
The species status is strongly supported by genetic distances from other taxa of the A. fatidica species-group (see below).

Agrotis sp. near A. mayrorum sp. n.
The isolated populations occurring in the high mountains of the Balkans in Bulgaria, Macedonia and Albania appear from their external morphological features to be more similar to the south-western Alpine A. mayrorum than to the geographically closer eastern Alpine A. fatidica.
The few specimens known from the Caucasus (Daghestan) show transitional features, with rather long branches of the male antennae and greyish suffused but still paler hindwing than in the typical A. fatidica populations.
The detailed taxonomic analysis of these populations and their biogeographical constraints requires further research and we cannot identify them with confidence as A. mayrorum.
Agrotis mazeli sp.n. 18 The Pyrenean populations represent a common lineage with the SW Alpine A. mayrorum, sharing the features of the male antennae and the paler (generally white or whitish) hindwing which distinguish the two new species from A. fatidica.The specific differences between A. mayrorum and A. mazeli can be found in the wing shape, the different form of the postmedial line, the shape and size of the dark dots following the terminal line and the colouration of the forewing underside.These are the same characters in which A. mayrorum and A. fatidica differ (besides the group features mentioned above).
The male genitalia of A. mazeli differ from those of A. fatidica and A. mayrorum by the short and basally broad valvae with the very typical, wide membranous ventral marginal section bordering the saccular part.The cucullus is broad with a finely pointed apex, the clasper (harpe) is rather long and slender, and the dentate subbasal plate of the vesica is weaker than in the closely related European taxa.
The female genitalia of A. mazeli have longer apophyses anteriores and a somewhat shorter appendix bursae than those of A. fatidica; the ovipositor and the ductus bursae are proportionally shorter, and the appendix bursae is shorter in A. mazeli than in A. mayrorum.
The species status is strongly supported by genetic distances from other taxa of the A. fatidica species-group (see below).
Our investigations on the DNA barcodes of certain Agrotis species occurring in the Altai Mts showed, surprisingly, no genetic distance between A. trifurcula and A. fatidica, indicating possible introgression (Huemer et al. 2017).

Agrotis innocens (Boursin, 1967)
Figs 47-50; Gen. fig. 12 This taxon was described by Boursin from three male specimens collected in the central region of Mongolia by Dr. Z. Kaszab.The specimens of this short series differ conspicuously from the sympatrically occurring and widespread A. trifurcula by their pinkish-shaded pale ochreous-grey ground colour of both wings and the reduced pattern.
The genitalia differences between the holotype of A. innocens (see Boursin 1967) and the studied specimens of A. trifurcula were subtle in the original study since the everted vesica of the two taxa were not examined.The lack of the female and the absence of A. innocens in the large samples of A. trifurcula from various places of Mongolia and the surrounding areas implied the possible conspecificity of the two taxa, especially as there are no remarkable differences in the male antennal characters.The study of the vesica of the two species revealed notable differences in the configuration of the vesica.The vesica of A. innocens is clearly longer than that of A. trifurcula, medially slightly dilated and distally evenly tapering, and is without a thickening at the nearby ductus ejaculatorius, which is typical of A. trifurcula (see the Gen. figs 10-12).Thus, A. innocens is considered here as a species distinct from A. trifurcula.After over a century of confusion regarding the identity of A. characteristica and A. robusta, and of these with other taxa of the A. fatidica species-group (e.g.trifurcula Staudinger, 1892, sajana Corti, 1932) (Corti 1932in Corti and Draudt 1931-1933;Kozhanchikov 1937;von Mentzer and Moberg 1987), these two names were recently synonymised, thus making robustana Poole an unnecessary replacement name for robusta Eversmann, preoccupied by Blanchard (see Kononenko et al. 1998;Kononenko 2005).Nevertheless, there is evidence that two concepts are still mixed under A. characteristica: that adopted by Fibiger (1990: pl. 1) in Noctuidae Europaeae Vol.1., referring to a form with comparatively short lamellae in the male antenna and non-brachypterous females which occurs in the Urals, and an Asiatic insect, referred to by Kononenko et al. (1998), which has longer lamellae, and fully-winged females.In passing, it may be noted that Corti (1931Corti ( -1933: pl. 5: pl. 5, row l) illustrated a brachypterous female of "A.robusta" while stating in the text that the female is fully developed and larger than the male.Unfortunately, the type specimen of characteristica is a female, which prevents an undisputed morphological assessment of its relationship to robusta.On the other hand, it seems likely that these two names may actually refer to the same Asiatic species, as their type localities suggest ([Inner Mongolia]: Ordos and Southern Altai, respectively).If so, the populations of the "Ural form" still remained unnamed.

Agrotis iremeli Nupponen, Ahola & Kullberg, 2001
As a close relative of Agrotis ruta (Eversmann, 1851), this species is also a member of the A. fatidica species-group which, on grounds of larval morphology, reveals a relationship with the A. segetum ([Denis & Schiffermüller], 1775) species-group, notably A. clavis (Hufnagel, 1766) (Nupponen et al. 2001).The large size and the dark reddish-brown ground colour of A. iremeli make it unlikely to be confused with its congeners.It is most like A. ruta, the ground colour of which is however always grey and less monotonous, more mottled with darker and paler areas, and has darker, more sharply defined stigmata, and a more wedged subterminal line.Some forms of A. cinerea ([Denis & Schiffermüller], 1775) might be mistaken for A. iremeli, particularly some females, but as a rule they are smaller, with more weakly crenulated antemedial line of forewing, absolutely tiny orbicular stigma, if any, smaller reniform and no claviform.
A. iremeli has been collected solely in restricted biotopes with xeric, rocky montane tundra, either in clearings within taiga forest or on the peaks of the mountains of the South Urals, at an altitude of 1300-1400 m.It appears to be endemic to the Southern Urals, in essentially similar habitats of the Northern Urals where only A. ruta was found (Fibiger et al 2010, citing K. Nupponen).

Agrotis ruta (Eversmann, 1851) Figs 51-56; Gen. figs 13-14
The only Holarctic member of the species-group.Its Nearctic populations have long been distinguished under the name A. patula.The type-locality of the species is poorly defined ("eastern Siberia").The species appears to be distributed from the Ural Mountains through Central and eastern Asia to the Pacific area and in vast parts of North America.It is highly variable in external features, with several, often remarkably dissimilar, forms and local populations.Recent molecular studies (Zahiri et al. 2014) show, however, a high consistency within the Canadian populations whereas material from the Altai Mts is distant and clusters nearest to A. proverai.Further in depth studies are necessary for the Palaearctic populations.

Agrotis patula (Walker, 1857)
The species was described from the Rocky Mountains (USA), but several large European collections house specimens under this name from Labrador and from the eastern side of the mountains of North America (named by Möschler as A. septentrionalis in 1862) and from the northern part of Pacific Asia (Japan: "Euxoa" kurodakeana Matsumura, 1927; Russia, Kurili Islands: Feltia subinformis Bryk, 1941).All these taxa have been recently considered to represent a single species, A. ruta (Kononenko 2005; Lafontaine and Schmidt 2010), mainly due to their very similar DNA barcodes.The fact that the DNA barcodes of certain closely related taxa of the A. fatidica group (e.g. A. trifurcula and A. ruta) are identical or very similar suggests that the ruta-patula complex needs to be reviewed.

Agrotis xylographa Boursin, 1948
The species was described from a single male specimen from Transbaikalia when our knowledge of the fatidica-group species was limited.At that time, separation of A. xylographa from A. ruta seemed reasonable.Subsequent study based on much more material of ruta shows the great external variation of the Asiatic populations of A. ruta (s.str.), as well as its wide range from the Ural region throughout Siberia and Mongolia to the Pacific area.The holotype of A. xylographa falls within the morphological and geographical range of A. ruta.Accepting the statement by Kononenko (1990;2005), we treat A. xylographa as a junior synonym of A. ruta.Agrotis xylographa does, however, differ from the Asian Pacific specimens of A. "patula".

Molecular results
25 specimens belonging to 4 species were successfully sequenced (Table 1) and group in 4 different clades (Fig. 61).Maximum intraspecific distances are ranging between 0% and 0.51% whereas minimum distances to the nearest neighbour are between 0.46% and 1.12% (Table 2).Agrotis mazeli sp.n. and A. mayrorum sp.n. have unique BINs (Barcode Index Number) in BOLD (Ratnasingham and Hebert 2013), whereas A. fatidica and A. proverai share their BIN.

Description of the new species
Agrotis mayrorum sp.n.Etymology.Named in honour of Marlies and Toni Mayr (Feldkirch, Austria) who collected the new species in companionship with Peter Huemer and recognized its differing morphology.
Diagnosis.The new species can be distinguished from the pale forms of A. fatidica by its lighter appearance, which is the consequence of the generally paler, milky whitish hindwings.The hindwings of the new species have only sparse, light greyish-brown irroration along the costal margin and narrow and discontinuous submarginal stripe while the narrow marginal ribbon along the variably dark, often obsolescent or deleted terminal line is whitish, as are the hindwing cilia.The discal lunule is just a fine shadow on the upperside, and the veins have a fine brownish covering.The underside of the hindwing lacks the transverse stripe (only a diffuse patch at costal margin may be present), the costal area is paler, suffused with brown, and the discal spot is rather weak, usually with a whitish inner liner on the crossvein of the cell.
By contrast, in A. fatidica, the typical specimens have dark greyish-brown suffused hindwings, with variably strongly diluted inner area, and a more prominently marked discal spot.Specimens with paler hindwings are relatively rare, but even they have broader and darker marginal field, extending in almost every specimen to the terminal line, and the discal lunule is stronger and darker, more contrasting than in A. mayrorum.The hindwing underside of A. fatidica is generally darker, usually with a broad and diffuse transverse stripe and a darker discal spot; these elements of the underside pattern are more strongly marked even in the palest A. fatidica specimens than in the new species.Another typical feature of A. mayrorum is the diffuse or obsolete postmedial line, which is less curved at the costa than in A. fatidica and A. mazeli.The dark dots tracking the fine terminal line are less prominent on the upperside than in the other two close relatives, and are almost invisible on the underside, only the thin terminal line is clearly marked.The detailed comparison of A. mayrorum and A. mazeli is given under the diagnosis of the latter species.
Agrotis mayrorum can easily be distinguished from A. proverai by the external features of the moths, being on average remarkably larger, with a more robust body and distinctly broader forewings, paler and more greyish ground colour and more prominent tessellate wing pattern, having Gen. figs 4-6.Agrotis mayrorum sp.n., paratype, male genitalia.
Gen. figs 7-9.7 Agrotis mayrorum sp.n., male genitalia, "Corsica", 8-9 Agrotis mazeli sp.n., male genitalia, 8 holotype, 9 paratype.characteristically prominent pale lining of the cubital vein and the much paler, never strongly dark filled reniform stigma.The other elements of the wing pattern (e.g. the shape and distinctness of the postmedial line, its distance from the reniform stigma, the shape and size of the arrowhead-spots, etc.) sometimes overlap with the condition in A. proverai so are unreliable, although they are usually different in the majority of the specimens.
The frontal prominence of A. mayrorum, although somewhat variable in shape and size, is less thick and more rounded or drop-shaped than the taller, more protruding frontal prominence of A. proverai.
The diagnostic features of the male genitalia of A. mayrorum are the long and slender valva (comparatively the longest and narrowest within the European species of the fatidica complex), with a less dilated distal section and only a slightly convex distal dorsal margin and reduced membranous ventral edge along the sacculus; the most often elongated-triangular cucullus with pointed apex; the short and broad clasper, which is distinctly shorter than in A. mazeli and on average shorter and apically more rounded than in A. fatidica; the juxta is also somewhat larger and broader, more pentagonal than in A. fatidica and A. proverai.The basal dilated section of the vesica is large with a well-developed, rounded conical dorsal diverticulum; the ventral carinal bar and the subbasal serrate-dentate plate of vesica are connected by an angled membranous part.The carinal bar and the dentated plate are usually directly connected in the other three close relatives and the dentated plate is regularly stronger and more densely spiny in A. mayrorum than in the related species.
The female genitalia of the three closely related species are very similar (the female of A. proverai is still unknown), although some differences can be found in the length and proportion of the ovipositor, the apophyses anteriores and the appendix bursae.Agrotis mayrorum has, in comparison with A. fatidica, proportionally a longer and more elongated ovipositor, longer apophyses anteriores and a longer appendix bursae; while the ovipositor, the ductus bursae, and the appendix bursae are proportionally longer than in A. mazeli.
Description.Wingspan: males 40-45 mm, female 31 mm.The external appearance of A. mayrorum is similar to that of A. fatidica, but has narrower, more elongated and somewhat paler coloured forewings and much paler, ochreous-shaded whitish hindwings.Forewing ground colour pale creamy beige, mottled with pale brown and grey; costal area and all veins somewhat darker grey or grey-brown and defined broadly with paler ochreous-grey stripes.Antemedial line indistinct, double, strongest at dark claviform stigma, its lower tooth relatively long.Median fascia absent; median area with a few intracellular dark streaks running from postmedial line to lower cellvein; postmedial line obsolete, appearing as a thin pale shadow defined by a few darker dots between veins.Subterminal line faint, zigzagged, pale ochreous-grey, defined by conspicuous though not very sharply marked chevron-spots; terminal line fine, black(ish), either continuous or divided into a row of fine black lunules and/or streaks; fringe pale creamy beige, occasionally with somewhat darker brown midline.Stigmata well-defined, regularly outlined by black or dark blackish-brown; claviform stigma long, apically faintly rounded, with fine blackish-brown streak connecting it to postmedial line, its filling darker brown; orbicular of variable size and shape, most often flattened-elliptical, sometimes with pointed lateral edges and darker brown centre; reniform stigma large, elliptical, darkened, particularly in its distal half.Intracellular patch between stigmata dark brown or blackish; supracellular marking between crossvein and postmedial line also well-defined, blackish.Hindwing pale whitish or milky-white with fine ochreous sheen, costal area, termen and diffuse marginal stripe suffused with ochreous-brown; veins also brownish.
Discal spot weak, small, often obsolete or missing; terminal line more or less continuous, pale ochreous-brown; cilia whitish or ivory-coloured.Forewing underside ochreous-whitish, variably suffused with pale fuscous, with faint discal dot and postmedial line; hindwing underside whitish with fine ochreous sheen, veins darkened, transversal line absent or very pale, shadow-like; discal spot also indistinct, pale brown with whitish centre.
Molecular data.BIN URI: BOLD: ACE4355.The intraspecific average and maximum distances of the barcode region are 0% (p-dist) (n=5), however intraspecific divergence may increase with additional samples from the Alps.The minimum distance to the nearest neighbour Agrotis mazeli is 1.12%, whereas the minimum distance to Agrotis fatidica is 1.71%.
Bionomics.Host-plant and early stages are unknown.The adults fly relatively late in the season within a period lasting from late July to mid-August, probably even later.Most of the specimens were obtained by light trapping using both UV tubes and mercury vapour lamp.Adult records are restricted to alpine meadows and scree at high altitudes from ca. 1900 to 2450 m a.s.l. .
Distribution.The new species is known from the south-western Alps in Italy (Alpi Cozie) and in France from the Alpes Maritimes and the Alpes-de-Haute-Provence (formerly Basses Alpes, Hautes Alpes).
Etymology.Named in honour of Robert Mazel who collected most of the type-series and generously supported our work with this important material.
Diagnosis.The new species is the allopatric sister species of A. mayrorum, sharing the group features in the male antenna (longer and finer branches than in the fatidica-line), and the white (or whitish) hindwing.The two species can be distinguished by a series of external morphological features which are recognisable in almost all specimens.These differences are ca as large as those between A. mayrorum and A. fatidica.
Agrotis mazeli differs from A. mayrorum by its somewhat narrower and more pointed forewings, which are also narrower and apically more pointed than those of A. fatidica but broader than those of A. proverai, a more sharply defined postmedial line with a clearly visible deep curve inwards at costal margin, on average smaller and more darkened orbicular and reniform stigmata, smaller and finer dark dots at terminal line, which are practically missing from the underside, and the smoothly, rather uniformly darkened forewing underside.Gen. figs 10-12.10-12 Agrotis trifurcula, male genitalia, Mongolia, 12 Agrotis innocens, paratype male genitalia, Mongolia.
In the genitalia, the males of A. mazeli can be distinguished from those of the closely related taxa by the short and basally broad valvae with the very typical, wide membranous ventral marginal section bordering the saccular part; the broad cucullus with finely pointed apex; the rather long and slender claspers; the comparatively short vinculum; the fine eversible bar of the ventral carina and the weak, dentated subbasal plate of the vesica; the rather short and obtusely conical dorso-lateral diverticulum of the inflated basal section of vesica; and the entire length of the vesica is somewhat shorter (in comparison with the total length of phallus) than in A. fatidica and A. mayrorum.
In the female genitalia, the apophyses anteriores are longer and the appendix bursae is somewhat shorter than in A. fatidica; the ovipositor and the ductus bursae are proportionally shorter, and the appendix bursae is shorter than in A. mayrorum.
Description.Wingspan: males 40-41 mm, female 32 mm.The main external features are typical of the A. fatidica species-complex.Forewing ground colour pale greyish-brown or slate-grey, with darker grey and brown irroration; costa and veins darker than ground colour and defined broadly with paler ochreous-grey intervenal stripes.Antemedial line variably but strongly marked, partly obsolete; median fascia absent; median area usually less variegated than in the other related species.Postmedial line recognisable though diffuse, pale ochreous-grey, and defined by a row of dark dots; its upper part curved inwards rather strongly around reniform stigma.Subterminal line obsolescent, ochreous-grey, defined by a row of conspicuous intervenal chevron-spots; terminal line fine, black(ish), with weak or indistinct dark dots; cilia ochreous-grey, with variably strong brown midline.Stigmata well-defined, medium-sized or relatively small, outlined by black or dark blackish-brown; claviform stigma long, dark, blackish-brown with only a slightly paler filling; orbicular stigma rather small, usually flattened-elliptical (drop-shaped) with pointed posterior edge and darker brown centre.Reniform stigma medium-sized, elliptical or lunulate, with more or less diffuse dark grey-brown outline and filling (which is stronger than in A. mayrorum); intracellular dark patch between stigmata blackish-brown or blackish.
Hindwing milky-white or light whitish-grey with fine ochreous shade, costal area, termen and diffuse marginal stripe with ochreous-brown or greyish suffusion of variable strength; veins darkened.Discal spot rounded or somewhat lunulate, often shadow-like; terminal line more or less continuous, brownish; cilia whitish with some pale brownish shading.Forewing underside rather unicolorous grey, suffused with fuscous, discal dot and postmedial line obsolescent or missing.Hindwing underside whitish with darker veins, reduced transversal line and relatively well-defined discal spot; terminal line fine, brown, without or with only very small darker dots.
Molecular data.BIN URI: BOLD: ACR9899.The intraspecific average and maximum distances of the barcode region are 0% (p-dist) (n=3), however intraspecific divergence may increase with additional samples from the Pyrenees.The minimum distance to the nearest neighbour Agrotis mayrorum is 1.12%, whereas the minimum distance to Agrotis fatidica is 2.18%.
Bionomics.Host-plant and early stages are unknown.The adults fly relatively late in the season, the specimens of the type series were collected in early August.Most of the specimens were obtained by light trapping using UV tubes and mercury vapour lamp.Adult records are restricted to alpine meadows and scree at high altitudes above the timberline (2200-2300 m a.s.l.).
Distribution.The new species is known from the eastern Pyrenees in France and Spain.The type-series originates from the French side of the Pyrenees; the occurrence of the species on the Spanish side is documented by Redondo et al. (2010, under     The distribution of A. fatidica in the Alps is discontinuous and patchy.It is widely distributed in the central eastern Alps from western Styria (south-eastern Austria) to Graubünden (eastern Switzerland), but almost completely absent from the northern and southern mountain ridges with e.g.only a single population known from the Italian Dolomites.In the western Alps just a few Swiss (Wallis) localities are reported (Wyman et al. 2015) and none southwards (Parenzan andPorcelli 2005-2006), indicating a considerable distribution gap from A. mayrorum (see Fig. 60).It seems likely that separation of these two species from a common ancestor may date back several glacial periods (see also Huemer and Hausmann 2009).
Surprisingly the nearest species, in terms of the DNA barcode, to typical A. fatidica is not A. mayrorum from the SW Alps but A. proverai from the Apennines.One might have expected that gene flow between populations from the SW Alps and the Apennines would be much more likely than from the eastern Alps to central Italy due to almost continuous mountain systems connecting the Apennines with the SW Alps.However, the barcode distance of A. proverai and A. mayrorum with 1.55% is considerably greater than that between A. fatidica and A. proverai (ca.0.5%), thus indicating a much longer isolation of these taxa.A similar pattern of distribution is documented for only few alpine species, e.g.Aspilapteryx spectabilis Huemer, 1994 (Gracillariidae), which is only known from the eastern Alps and the Central Italian Apennines but absent from the SW Alps (Huemer 2011).However, in the absence of molecular evidence these widely separated populations of this latter species are considered conspecific.

Figure 60 .
Figure 60.Distribution pattern of successfully sequenced records of the A. fatidica-complex in Europe.
see Table 1.) and processed at the Canadian Centre for DNA Barcoding (CCDB, Biodiversity