Leiobunum rupestre species group : resolving the taxonomy of four widespread European taxa ( Opiliones : Sclerosomatidae )

Within the central European opilionid fauna the widely used species names Leiobunum rupestre Herbst, 1799 and Leiobunum tisciae Avram, 1968 pose taxonomic and distributional problems. In addition, Nelima apenninica Martens, 1969 is close to L. tisciae in terms of external and genital morphology, but is specifi cally distinct. While coxal denticulation is largely lacking in N. apenninica, the validity of the genus Nelima Roewer, 1910 is questioned again. In addition, Leiobunum subalpinum Komposch, 1998, a recently described novelty from the eastern Alps, is closely related to L. rupestre. The four species are combined as the morphologically defi ned Leiobunum rupestre species group. Except for L. subalpinum, they were found to be allopatrically distributed from the Carpathians across central and Northwest Europe to the south-western Alps. The latter species is locally sympatric and partly elevationally parapatric to L. rupestre. Leiobunum tisciae is a recently introduced name and here recognized as a junior synonym of a number of taxa described much earlier, of which L. gracile Thorell, 1876 is re-introduced as oldest available name. Detailed morphological and distributional data for all taxa are presented.


Introduction
A common problem with the Sclerosomatidae Simon, 1879 (harvestmen) is their morphological plasticity as well as their general uniformity, making it often diffi cult to draw well-recognizable species borders.Even more problematic, especially in the tropics, is the differentiation of higher taxonomic levels from species category upwards (Hedin et al. 2012).Applying this diffi culty to widespread, closely related and morphologically similar taxa can further hinder discrimination.Here the similar species Leiobunum rupestre (Herbst, 1799), L. subalpinum Komposch, 1998, L. tisciae Avram, 1968 and Nelima apenninica Martens, 1969 are investigated and summarized as the Leiobunum rupestre species group, though being generally grouped with two different genera.The four species occupy large parts of central and southeastern Europe, parts of Scandinavia (Finland, Norway, Sweden), locally Great Britain, continental Italy and south-western France, but sympatric occurrences are unknown (Fig. 2), except for L. subalpinum whose distribution is nested within the area of L. rupestre.Discrimination of L. rupestre and L. tisciae has been under debate since Martens (1978) included L. tisciae into the central European fauna, even though male genital morphology was properly described and is straightforward.A number of European countries stated the presence of one or two of these species thereafter (e.g., Great Britain: Davidson 2009;Denmark: Enghoff 1988;Czech Republic and Slovak Republic: Šilhavý 1981;Italy: Chemini 1980;Marcellino 1971;Poland: Staręga 2004;Norway: Stol 2003, 2010a, b;Estonia: Staręga 1978;Tomasson et al. 2014;Latvia: Tumšs 1963;Spuņgis 2008), but a complex morphological and taxonomic situation was evident.
The description of L. tisciae by Avram (1968) was shortly followed by that of Nelima apenninica by Martens (1969), who soon recognized that both species are fairly similar in terms of genital morphology, but he was primarily concerned to discriminate Leiobunum rupestre against L. tisciae (Martens 1978).Up to today, N. apenninica remains poorly known, being recorded from widely scattered localities.These four species raise taxonomic problems and need to be well defi ned.L. tisciae and N. apenninica having been described relatively recently, and especially because L. tisciae spans a large European area, making it likely that it has been described under an older name before.L. subalpinum, a novelty from the eastern Alps, added a further complexity to species recognition in this group of locally common long-legged harvestmen.Taking this argumentation into account and based on representative reference material, we here revalidate the species status, synonymy and distribution of the representatives of the L. rupestre species group.The taxa Leiobunum nigripalpe Simon, 1879 and "Leiobunum sp.", the latter a recent immigrant to western central Europe (Wijnhoven et al. 2006), are not part of the L. rupestre species group and are not treated here.

Material and methods
Material investigated: 54 series of L. rupestre, 40 series of L. tisciae and 20 series of N. apenninica were checked; details are listed in the Material sections (see Appendix).Also literature sources, which we feel present reliable records, are listed there.Original line drawings were produced using a camera lucida attached to a Leitz dissecting Microscope and a Leitz Laborlux Microscope.Measurements were taken by means of a micrometer disc using the same optical devices.All measurements are given in mm.

Distributional data
From a total of 2600 text-recognized pdf fi les of opilionid literature, those matching the search strings 'apenninica', 'rupestre' or 'tisciae' were selected.Following our established synonymy further, references were extracted manually and all identifi able distributional records of the L. rupestre group were georeferenced.Of these records only a limited number of the recorded specimens was available to us and geographic reconstructions relied on our representative samplings and sound publication data.We assume Šilhavý (1981) and Staręga (1976Staręga ( , 2004) ) to have correctly discriminated L. rupestre and L. 'tisciae' (L.glabrum in Šilhavý 1981) and their data already demarked a North-to-South running borderline between the two species in the Czech Republic and Slovakia (Šilhavý 1981) and in Poland (Staręga 2004).We further interpreted all records east of this line (e.Staudt 2016).All reconstructed localities are given in Fig. 2 and Appendix 1. Up to-date distributional data of Leiobunum subalpinum was available from Komposch (1998Komposch ( , 2009b) ) and we did not search for any additional records.

Remark
Of the representative material investigated, we found the four hypothesized species well separable upon a number of somatic and male genital morphological characters, which are summarized in Table 1 for convenient use.Species appear largely allopatric, except for L. subalpinum, which is fi rmly nested within the area of L. rupestre.In addition, parapatric and locally sympatric occurrences of two species under discussion in East-Central Europe cannot be excluded, especially in South Poland, the central Alps and close to the Czech/Slovak border area (Fig. 2).Komposch, 1998 is considered most closely related to L. rupestre (Komposch 1998) and, as such, is also regarded as a member of the L. rupestre species group.As it is easily discriminated from all discussed forms by its dark coxal markings and, apart from L. rupestre, by its restricted alpine range, it is not featured here (see Komposch 1998;Komposch & Gruber 2004 for Table 1.Comparison of somatic characters of Leiobunum apenninicum (Martens, 1969), L. gracile Thorell, 1876and L. rupestre Herbst, 1799. L. subalpinum Komposch, 1998 is not included, because it is very similar to L. rupestre and is easily identifi ed by its black coxal markings.For comparison with L. rupestre we refer to Komposch (1998).If not otherwise mentioned, characters apply to males only.Characters outlining a single species are given in bold, more variable characters are given in italics.The combination of several characters for determination is advised.Palpal spination is variable, apparently the appearance in ordered rows is an individual character.Body coloration varies with age and population, aged specimens are darkest.Herbst, 1799 L. gracile Thorell, 1876 L. apenninicum (Martens, 1969) without spines, sometimes one 1-3 small spines on each side, spine present rarely absent in both sexes dorsal coloration in both sexes dorsal coloration in both sexes dorsal coloration without pair of white spots on with small white spots on each with pair of white spots on each each area along the median line area along the median line area along the median line in females no lateral coloration in females lateral coloration in females lateral coloration of opisthosoma with a blurred of opisthosoma with a blurred of opisthosoma with a blurred parallel pattern of zigzag lines parallel pattern of zigzag lines, parallel pattern of zigzag lines, best seen at low magnifi cation best seen at low magnifi cation details) except for general remarks; its geographic range is indicated in Fig. 2. In addition, the poorly known Leiobunum nigripalpe Simon, 1889, described from the western Alps (France), was reinvestigated.However, the type series represents a composite collection of at least two different small Nelima Roewer species, both unaffi liated to the L. rupestre species group.This holds true, too, for the specimen depicted as Leiobunum nigripalpe in Martens (1978), representing a yet to be identifi ed Leiobunum species.

Taxonomic history
At fi rst glance rather complex, but many distributional records from East central, Northwest and North Europe in reality refer to Leiobunum 'tisciae' (see below, L. gracile).L. rupestre is a montane species and well known from low to mid-altitudes of the central and eastern Alps and mountainous areas in the Czech Republic and north of the Alps in Germany.Its genital morphology is quite characteristic and was fi rst reliably depicted by Šilhavý (1948, fi gs 4-6; 1956, fi gs 437-439), fully in accordance with drawings in Martens (1978, fi g. 778) and Komposch (1998, fi gs 9-10).Leiobunum populations in the Carpathians, Poland (Staręga 1976), southern Finland (Heinäjoki 1944), along the Baltic coast hinterland and Denmark were also largely affi liated to L. rupestre (Martens 1978, fi g. 787), though erroneously.These older determinations mostly neglected the genital morphology of these crucial northern populations, which was reconsidered after Avram (1968) published a new species, L. tisciae, clearly distinct from rupestre but similar in external morphology.Due to its wide distribution, the majority of available names for the eastern and northern populations can be assigned to the synonym of its oldest name L. gracile (see below).

Diagnosis
A medium-sized Leiobunum species with blackish upper side, except for broad, white markings on Ceph disto-laterally in male (Figs 1A, 3A), broad, blackish irregular saddle-like marking all over the length of the body in female with extended white markings laterally from Ceph to abd Area II and a cross stripe all over area V (Figs 1B, 3B).In both sexes series of para-median lines of small white spots on the abd areae I-V are absent.Contrasting yellowish underside including coxae of all appendages.Wings of truncus penis broad, broadest in lower third part of truncus (Fig. 5A-B).Coxa IV with retrolateral row of granules.PEDIPALPS.Male (Fig. 4A-B): Fe with irregular rows of marked triangular spines ventrally; dorsally nearly unarmed, ventral spines more slender than in L. gracile; Pt slightly armed with pointed denticles laterally; Ti markedly convex in basal part and concave at ventro-distal end, therefore forming a well-marked S-bend, fi eld of fi ne dark coloured granules all over the ventral side, well to be seen; Ta continuously and slightly bent to ventral.Female (Fig. 4G-H): similar to male, less spines on Fe and Pt, few spines on Ti proximally and no ventral granulation, Ta strait, slightly bent distally.
GENITAL MORPHOLOGY (Fig. 5A-C).Truncus penis stout, in ventral/dorsal view from basal opening to insertion of glans slightly and continuously tapering; distal wings covering less than half of truncus.Wings consist of two independent parts: a ventral shell-like plate, sharply cut off horizontally at upper third of wing structure, from its upper margin continuously extending to lateral side of truncus, thus partly embracing the lateral wings which extend from dorsal side of truncus via lateral side to ventral side.Their lower two thirds are covered by the ventral plate.Surface of the upper margin of the lateral wings and membranes adjacent to the truncus is invaginated forming a long double-walled internal sack.At the very distal end of truncus a small membranous oval sack with a distal opening is attached ventrally.
From lateral view truncus slightly curved (concave on ventral side), slightly tapering towards glans.Wings massively enlarged on ventral side.Glans stout, in lateral view tapering towards stylus.Distribution (Fig. 2) Main area are the eastern Alps from the eastern part of Switzerland, all over Austria, the German and Slovenian alpine parts and extending to the north-western Balkan Peninsula.Records extend southwards to Serbia (CJM6363, CIK1336, CIK1337), Croatia (Babić 1916;Novak 2004b) and Bosnia-Hercegovina (Martens 1978;Novak 2005).To the Northeast, the area extends into the Czech Republic: many localities in Bohemia and Moravia, mainly in the northern and southern mountains marking the borders to Austria, Germany and Poland (Šilhavý 1981).
For Poland, Staręga (1976) indicated a number of "rupestre" records in the southern mountains, the Karkonosze (in German "Riesengebirge") close to the Czech border.From the opposite Czech side, in the same mountain stock, the Krkonoše, Šilhavý (1981) pinpointed records for (true) rupestre, which he compared to "glabra" from his country.As these close-to-border Polish localities are situated in mountainous areas, too, they probably belong to rupestre as well.Besides, the Polish 'tisciae' (i.e., gracile, below) records are all situated in lower stretches of central and eastern parts of the country (Staręga 1976(Staręga , 2004), far from any (possible) mountainous rupestre record.The L. rupestre area as depicted in Martens (1978: fi g. 787) needs to be reduced, substracting the distribution area of.L. 'tisciae' (i.e., gracile, below), stretching from the Carpathians in a broad northerly area to the Baltic coast up to South Finland and in the West to Denmark and northern Germany (Fig. 2).However, the remaining distribution area of L. rupestre slightly changed with respect to the Balkan area, then largely unknown, and new records in central Germany were added.

Ecology
This is a forest species with a high demand of air moisture.Therefore, it is confi ned to moist places close to streams and creeks, on tree trunks, under rotten wood and on moist rock faces.It hardly occurs in open areas, in agricultural land, urban gardens or on house walls in villages and towns, if not situated in moist forests (Martens 1978), or at higher altitudes.Martens (1978) indicated height records from 260 m to 2160 m a.s.l. in the Austrian Alps, but likely included localities for the closely related L. subalpinum.Komposch & Gruber (2004) set the Austrian uppermost rupestre limit at 1680 m a.s.l. and for subalpinum at 2100 m a.s.l.
In all cases populations of rupestre are confi ned to mountainous areas of different altitudes.Only in the Saxonian Elbsandsteingebirge does rupestre live at the bottom of deeply recessed valleys which are permanently moist and where sun radiation is largely excluded from the narrow valleys by the steep rock faces.Komposch, 1998 Fig. 2 Leibunum subalpinum Komposch, 1998: 20.Leiobunum rupestre Gruber, 1966: 46 (partim).Nelima religiosa Roewer, 1910: 912 (in

Taxonomic history
Rather uncomplicated.Because of its extreme similarity to L. rupestre, including genital morphology, this species went unidentifi ed in many former high-altitude reports on L. rupestre until its formal description in 1998.Leiobunum rupestre and L. subalpinum occur in sympatry and local syntopy, but coxal markings allow easy fi eld identifi cation, so no further misindications were recorded.

Diagnosis
A medium-sized Leiobunum species very similar to L. rupestre, except for black coxal markings in both sexes and white markings on Ceph including a white area in front of the tuber oculorum in males.Coxa IV without granules.For further details refer to Komposch (1998).

Description
BODY.See Diagnosis.
DORSAL PATTERN.Extremely similar to L. rupestre.
PEDIPALPS.Very similar to L. rupestre except for a medio-basal swelling on Ta in males.
GENITAL MORPHOLOGY.Very similar to L. rupestre.The wing-free basal part of truncus shorter than in L. rupestre.
Distribution (Fig. 2) Restricted to the eastern Austrian Alps, mainly on the southern macroslope in the federal states of Salzburg, Steiermark and Kärnten (Komposch 1998;Komposch & Gruber 2004;Komposch 2009b) extending into adjacent Slovenia (Novak et al. 2006).Type locality is Gößnitztal in the National Park Hohe Tauern in Kärnten.

Ecology
This a mountainous species of the eastern Alps.Its distributional range extends from about 1500 m to 2100 m a.s.l. with rare outliers down to 430 m in moist and cool ravines and up to 2200 m.Vertical rockfaces mostly of siliceous origin are preferred (Komposch 2009a, b;Komposch & Gruber 1999).
There is a zone of sympatric and sometimes even syntopic occurrence together with L. rupestre between 500 m and 1400 m.Strongholds of L. rupestre in that area extend from 500 m to 1000 m with rare occurrences below (down to 200 m) and above (up to 1400 m) (Komposch 1998).

Taxonomic history
A rather complex situation developed because old species names with type localities in various areas in Scandinavia or in (present) eastern locations were never checked in detail.This is partly due to the fact that names were based on juvenile specimens and partly to the unavailability of former authors to access specimens of northern (Baltic) and southern (Alpine) origin for direct comparison.Consequently, northern populations, originally described as Liobunum gracile Thorell, 1876 and Liobunum laeve Thorell, 1876, were permanently affi liated to L. rupestre, starting with Tullgren (1906a, b), followed by Heinäjoki (1944), Staręga (1976) and Martens (1978).The situation became even worse when Avram (1968) described L. tisciae Avram, 1968 from the Tisza valley in Hungary.By external morphology, this species is similar to L. rupestre, which Avram did not take into account.Martens (1978) (Tullgren 1906b).Yet, both have been described from Sweden, Province of Scania, from where Tullgren (1906b) confi rmed an adult specimen.Therefore, it seems a reasonable decision to re-establish the oldest available name, Leiobunum gracile Thorell, 1876, and place L. laeve Thorell, 1876, L. norvegicum Strand, 1910and L. tisciae Avram, 1968 in its synonymy.L. gracile has page and even line priority over L. laeve.

Diagnosis
A medium-sized Leiobunum species with blackish upper side, except for broad white markings of Ceph disto-laterally in male (Fig. 3C), broad blackish irregular markings on abd area I-V of the body, in female with mottled white markings laterally from Ceph to posterior areae, white lateral marking on abd area V (Fig. 3D).In both sexes two paramedian small white spots on abd areae I-V, more conspicuous in male (most similar to L. apenninicum).Contrasting yellowish underside including coxae of all appendages.
Wings of truncus penis slender, slightly tapering proximally.
DORSAL PATTERN (Fig. 3C-D).See Diagnosis.Tu oc in both sexes smooth, without spines, in rare cases one spine present.
PEDIPALPS.Male (Fig. 4C-D): Fe with irregular rows of large triangular spines ventrally often with single hair on top; dorsally few on distal part, ventral spines broader and more massive than in L. rupestre, where they are more slender and without hair on top.Pt with scattered denticles la and do; Ti slightly concave ventro-distally, forming a fl at S-bend, denticles and hairs on ventral side of palpal tibia light-coloured, hard to be seen; Ta mostly straight, bent only in distal third.Female (Fig. 4J-K): in all members similar to male, but armament less conspicuous, Ti more slender, Ta strait.
GENITAL MORPHOLOGY (Fig. 5D-F).Penis similar to that of L. apenninicum.Truncus stout, from ventral/ dorsal view from basal opening to insertion of glans slightly and continuously tapering to the internal sac of the wings; lateral wings (ventral view) markedly enlarged, smoothly rounded thus nearly eggshaped (irrespective the lateral membranes running down the truncus).At its very distal end the wing structure close to the glans insertion opens to a small opening, extending into two long nearly parallelsided double-walled sacks to the lower end of the wings.A lateral somewhat invaginated discharged membrane in the distal half of the wings unites the dorso-lateral and the latero-ventral part of the wings (lateral view) allowing for a possible infl ation of the wing structure when the sacs are fi lled with secretion.
From lateral view truncus slightly curved (concave on ventral side), slightly tapering towards basis of stylus.Wings enlarged on ventral side, slightly rounded convexly.Glans broad at truncus insertion, in lateral view continuously tapering towards stylus.
Carpathian Arc with many localities documented in collections, but few published papers available, not very far eastward and not in the southern lowlands (Murányi & Lengyel 2006;Weiss 1996;Cirdeî 1960).
In the Southwest the area extends into Serbia (Bor Dubasnica, Mala tisnica, CJM 6363).There are scattered records for the western part of Ukraine, mainly as an extension of the Carpathian area (Bartoš 1939;Cîrdei 1960;Staręga 1978;Morin 1931Morin , 1934;;L. Koch 1870).The many records in Poland pertain to the eastern two thirds of the country, extending to the Baltic Sea (Staręga 1978(Staręga , 2004, rupestre; for true Polish rupestre see above) and including former Westpreußen.Occurrences in Russia are reported from former Ostpreußen (Le Roi 1914) and from St. Petersburg (CJM 3032), extending to Northeast Poland (Staręga 1976(Staręga , 2004)).Estonia: records along the Baltic Sea Coast (Tomasson et al. 2014).Latvia: partly detailed locality maps (Tumšs 1963;Spuņgis 2008).Finland: only in the southern part, with detailed map based on monitoring records (Heinäjoki 1944), presence confi rmed by Uddström et al. (2013).Germany: records along the Baltic Sea Coast and its hinterland (up to Hamburg), dating back to Kraepelin (1896), Le Roi (1914) and Rabeler (1929), but apparently presently very rare, last record in Schleswig-Holstein, Lübeck in 2007 (CJM 6111; see Martens 1978).Denmark: early records by Hansen (1884), more recent records from all over the country remained largely unchanged between about 1960 (Meinertz 1964) and 1987 (Enghoff 1988), with slight distributional differences, formerly missing from Funen and rare in South Jutland and Himmerland, but very common and abundant there in 1987 (Enghoff 1988), including Funen, South Jutland and Himmerland.Since then, it has vanished, has not been recorded for many years, but it was rediscovered recently and is presently extremely rare.In 2008, four individuals were present on four out of 64 sites (Enghoff et al. 2014).Sweden: except for the early records of Thorell (1876) Strand 1900, norvegicum).Great Britain: Derbyshire (Martens 1978), unrecorded until 2008 when it was found around Aberdeen and Dunbennan, Huntly (Davidson 2009, CJM 6358-6361), in Scotland apparently spreading and in Great Britain presently known from seven counties (British Arachnological Society 2016).
The present distribution of L. gracile and that in the recent past obviously represent two different ecological-defi ned origins.One belongs from lowland to montane and riverine forests, the other one comprises secondary urban habitats to which the species was (most likely) transferred by human activity.Such "anthropogenic" populations may suddenly increase, occupy a large area and sometimes collapse again, for reasons little understood.Others are spot-like with little infl uence on local faunas.At least for a period of roughly 120 years such dynamics can tentatively be reconstructed (see Discussion).

Ecology
A euryoecious species.In Slovakia, including the Carpathian Arc, recorded from sheltered places in oldgrowth forests on tree trunks and on rock faces (Šilhavý 1981).It is common in the inundation zone of rivers like the Tisza, e.g., frequently reported from tree hollows of Populus (Avram 1968, tisciae) and Salix trunks (Avram 1968;Csizmazia et al. 1966, nigripalpis;Kolosváry 1963Kolosváry , 1965Kolosváry , 1966a, b), b).This also holds true for East Poland where gracile lives in forested habitats (Neple, Bug River, AXLS 958).Weiss (1996) indicated its occurrence from Romanian Transylvania forests of the hilly and montane belt.
In all other areas, which we believe to be secondary, L. gracile is confi ned to strongly modifi ed and human-infl uenced habitats.There it competes with other species preferring house walls like Opilio saxatilis C.L. Koch, 1839, O. parietinus (Degeer, 1778) and recently with O. canestrinii Thorell, 1876.
In central Europe the latter presently is by far the dominant species in human habitats and may have partly caused the noticeable decline of O. saxatilis, O. parietinus and L. gracile (see Discussion).

Taxonomic history
Though only recently described, the status of this large species was accepted by subsequent authors (Chemini 1980;Komposch 2009).However, similarity of genital morphology with L. 'tisciae' (now L. gracile, Martens 1978) later led Martens to the assumption that the scattered distribution of apenninica might represent disjunct area parts of L. gracile.This remained unpublished due to the emerging complex taxonomic situation in L. gracile/L.rupestre.Yet, many details separate both species as is further confi rmed by molecular genetic analysis (J.Martens, unpublished data).
The generic affi liation of apenninica is diffi cult to establish.Certainly, it does not belong to the European bulk of small-bodied Nelima species like N. sylvatica Simon, 1879, N. gothica Lohmander, 1945and N. sempronii Szalay, 1951, but is a large and long-legged species similar to L. gracile, less to L. rupestre.
Due to the lack of coxal denticles in the type material and according to the present defi nition of sclerosomatid genera it was originally placed in Nelima.However, based on only a single external character, this is an artifi cial affi liation, that we fi nd especially compromised in the species-pair gracile and apenninica.Leiobunum gracile shows a strong variation of number and placement of coxal denticles down to zero (CJM6357, female) and their frequent appearance in Nelima apenninica (AXLS1739 two to seven on coxa IV, in both sexes) underlines the false generic placement of apenninica.Consequently, in accordance with similar genital morphology and external characters we prefer to place apenninica in Leiobunum.

Diagnosis
A medium-sized Leiobunum species with blackish upper side, except for broad white markings of Ceph disto-laterally in male (Figs 1C, 3E), broad blackish irregular markings on abd areae I-V of the body.In female with mottled white markings laterally from Ceph to posterior areae, white lateral marking on abd area V and a white cross stripe all over areae VI (Figs 1D,3F).In both sexes a series of para-median small white spots on the abd areae I-V, more conspicuous in male (a character similar to L. gracile).
Contrasting yellowish under side including coxae of all appendages.Wings of truncus penis robust, slightly tapering proximally.

Description
BODY (Figs 1C-D, 3E-F).In both sexes dorsal granulation consisting of fl at, pointed granules, regularly spaced, well to be seen with 250× (comparable to L. gracile, see Šilhavý 1981: fi g. 2); eye mound in both sexes with 1-3 small spines on each side, rarely absent.
PEDIPALPS.Male (Fig. 4E-F): Fe with few massive spines, normally restricted to third ventro-distal part, only single spines on dorso-distal side or absent; Pt unarmed except for few small spines dorsally; Ti slighty infl ated, rather short thus appearing pretty massive, brush-like hair fi eld on ventral side, a peculiar character of its own; no spines clearly visible, likely for being of the same light colour as the embedding hair fi eld; Ta continuously bent ventrally over its whole length.Female (Fig. 4M-L): Similar to male.Fe nearly smooth, spines even less than in male, tibia less infl ated and slightly concave ventrally, no hair fi eld; tarsus straight, distally inconspicuously bent ventrally, more slender and slightly longer than in male.
GENITAL MORPHOLOGY (Fig. 5G-J).Penis similar to L. gracile.Truncus stout and parallel-sided in the basal half from where slightly and continuously tapering to the lateral lowest parts of the wings, from there slightly enlarged to middle of the wings, then tapering to basis of glans.Lateral wings (ventral view) markedly enlarged, nearly parallel-sided over nearly their total length, pretty sharply merging to distal part of truncus.At the disto-lateral margin of the wings (close to the truncus/glans joint) wings open to a narrow slit leading to two ventro-lateral pockets.A lateral membrane in the distal half of the wings unites the dorso-lateral und the ventro-lateral part of the wings which allows for a possible infl ation of the wing structure when the pockets are fi lled with secretion.From lateral view truncus slightly curved (concave on ventral side), slightly tapering towards middle part, then enlarging again.
Wings massively enlarged on ventral side, slightly rounded convexly.Glans slender, in lateral view tapering towards stylus.
Distribution (Fig. 2) It was long considered a rare and local species (Martens 1969;Komposch & Gruber 2004;Komposch 2009;Chemini 1980) and only a few specimens were available for the original description from the Apennines (Martens 1969).Even fewer became known from the apparently isolated area at the Italian border to Austria (Chemini 1980) and in Austria itself (Martens 1978;Komposch 2009) and the Southwest Alps (Martens 1978).Presently, the area seems to be subdivided into three major parts, viz., French and Italian southwestern Alps, southern Austrian (eastern Tyrol) and Italian Alps (South Tyrol), and scattered records in the Apennines chain.Type locality is Pratovecchio in the Apennines.Yet, there was little additional locality information since its original description until the species was recently rediscovered in the Southwest French Alps in high frequency, namely in the Mercantour National Park and the Vallée des Merveilles (CJM; ALS leg., JM leg).Disregarding its specifi c habitats and general low sampling activity for long-legged species in the Northwest Alps, the species might have been neglected in large parts of its distributional area.Prior to its description, L. apenninicum has likely been confused with L. rupestre, as is obvious from Trossarelli's (1943) records for the southwest Alps.The isolated record from Calabria (Gozo 1908, L. rupestre; included in Fig. 2) requires confi rmation.

Ecology
This is a forest species.During the day males rest on tree trunks and rock faces; females, though not exclusively, prefer to hide on the ground in leaf litter and thus are not easy to locate.Consequently, only males have long been known, because they are easier to spot and collect.The species is locally pretty common wherever the daytime microhabitats of females are found.Otherwise, it is readily encountered at night, and this technique revealed large populations in the Mercantour National Park in the French Alps.

Discussion
Systematics of the four central European species Leiobunum rupestre, L. subalpinum, L. gracile and L. apenninicum highlights general problems related to European sclerosomatid taxonomy.Firstly, in this case old, but available names, have been largely neglected and were not compared to species known from the area.Subsequent species proposals, i.e., Leiobunum norvegicum or Leiobunum tisciae ignored this point and caused considerable taxonomic confusion.Old names, though available, were not considered in detail and simply put aside.On the other hand, Leiobunum subalpinum, a species very similar to L. rupestre but not rare at all in its limited alpine area, escaped taxonomic recognition for about 200 years.Furthermore, the artifi cial division of Leiobunum into Leiobunum s. str.and Nelima proposed by Roewer (1910) and maintained for a century seems to be inappropriate in the case presented here.The Roewerian separation is based solely on the presence (Leiobunum) or absence (Nelima) of coxal denticles, but does not place "Leiobunum" and "Nelima" species into reasonable morphological and much less preferably monophyletic units, which is substantiated by results of modern molecular genetics (Hedin et al. 2012).
In our example, "Nelima" apenninica normally does not generally bear such denticles, but its male genital morphology is astonishingly similar to that of L. gracile to an extent that one might even suppose conspecifi ty."Leiobunum" gracile is reported to have highly variable numbers of coxal denticles (Avram 1968) and partly none, while our material also revealed several individuals of "Nelima" apenninica to have few denticles, too.Molecular genetic analyses indeed demonstrate close relations (J.Martens, unpublished data), confi rming the specifi c differences.The taxon apenninica belongs to Nelima only by Roewer's genus defi nition, genetically and genital-morphologically it is affi liated to a group of central European Leiobunum.However, "Leiobunum", presently a holarctically distributed genus, is heterogeneous as well and refers to a number of morphologically and genetically distinct lineages, which are far from being disentangled (Hedin et al. 2012).From our material it is also apparent that an increase of the number of denticles, as well as their presence on more than one coxa is linked to an increased body size of the individual, at least in apenninica and gracile.Such a variable character is certainly inappropriate to serve for differentiation of higher level taxa within a highly heterogeneous phylogenetic group.A molecular-based re-organization, not only of the European Leiobunum/Nelima complex, is in urgent need.
The distributional analysis revealed that Leiobunum gracile and L. apenninicum, though closely related, show a very different distributional pattern.Leiobunum apenninicum is a conservative forest species, apparently confi ned to a few scattered areas and presently with little perceivable tendencies to enlarge its distribution.
The closely related L. gracile, however, massively enlarged its area in close affi nity to man-made habitats around the Baltic Sea coast.In this area of expansion its populations considerably fl uctuate in numbers and it can constrict its area down to local disappearance.Parts of the Carpathian Arc and the adjacent mountains in Southwest Poland and Slovakia and low-altitude riverine forests in Hungary may represent the original area of the species, though this is diffi cult to substantiate."Natural" populations may be indicated by hilly and montane forests, also riverine localities where high population densities are obvious (Avram 1968;Csizmazia & Kolosváry 1966;Kolosváry 1963Kolosváry , 1965Kolosváry , 1966aKolosváry , b, 1969;;Kolosváry & Homonnay 1967).
However, it remains arbitrary to solidify area limits of an original and an expansive population based on ecological comparison of historical records.It is thus only hypothesized that starting from the Carpathians the Baltic coast area and its hinterland have been colonized, possibly by help of human-mediated transport.All Baltic authors (Rabeler 1929;Kraepelin 1896;Le Roi 1914;Heinäjoki 1944;Tumšs 1963;Meinertz 1964;Enghoff 1988;Enghoff et al. 2014;Staręga 1978Staręga , 2004) ) emphasize the close proximity to human settlements in their area, such as in gardens, specifi cally on house walls and similar environments, while forest habitats played a minor role in all ecological characterisations.In the Baltic area, this presumed invasion of new areas was successful, but to the West, the North Sea coast was not reached, at least not in a continuous belt, with the westernmost localities around Hamburg (Kraepelin 1896).Isolated outposts are known from the British Isles (Martens 1978) including Scotland (British Arachnological Society 2016), where the population seems to prosper.
A reasonable proof of newly gained territories is the fact that in the newly colonized areas population densities are high in "disturbed" habitats, remain stable for a certain period and after years or even decades collapse or fi nally completely disappear.This holds true for the population in Denmark, stable at least from the mid 1960s (Meinertz 1964) to 1987, when Enghoff (1988) undertook another monitoring and found an even more extended territory inhabited.Presently, L. gracile has largely disappeared from Denmark (Enghoff et al. 2014).Likewise in northern Germany only a few isolated records were reported between 2000 and 2009 (Staudt 2016), while the species appeared common at the beginning of the last century (Kraepelin 1896;Rabeler 1929;Le Roi 1914).Other occurrences remained spot-like from the very beginning and apparently disappeared soon after the fi rst discovery in Great Britain (Martens 1978, tisciae) or gained new area spots far away in Scotland and other parts of the U.K. (British Arachnological Society 2016).
Reasons for such extreme population fl uctuations are unclear and have not been surveyed in detail.
Population decline was obvious for long prior to the invasion of Opilio canestrinii to central Europe, which started about 1970 (Martens 1978, as ravennae, fi rst German records in Stuttgart 1974 CJM1619;Mainz 1975 CJM1525;Berlin 1976 CJM1553).For the year 1987 Enghoff (1988) stated that L. gracile and O. canestrinii lived side by side in high population densities in Denmark.Only gracile disappeared, canestrinii remained to the present and is common everywhere (Toft 2004, pers. comm.).However, it must be taken into account that long-lasting effects of high population density of O. canestrinii may have caused the decline of L. gracile.The same is assumed for Opilio parietinus in central Europe.
The temporal invasion of the Baltic dates back for at least 120 years, probably considerably longer.Kraepelin (1896) mentioned gracile as not being rare for the Hamburg area whereas Heinäjoki (1944) stated that gracile invaded Finland and gained northward territory only in the 1920s.Monitoring of the Finnish arachnid fauna was pretty detailed at that time and earlier occurrences would have been certainly noticed as indicated by Heinäjoki (1944), who documented the extension to about 61.6° N up to Hämeenkyrö, until today the northernmost known record (Uddström et al. 2013).For Denmark records date back to the 19 th century, when Hansen (1884) mentioned the species.

Leiobunum rupestre group Leiobunum rupestre
accepted this novelty but at that time lacked suitable fresh material to identify populations from northern Germany and Denmark as conspecifi c.Instead, he erroneously treated one species under two names, namely North German and Scandinavian populations under L. 'rupestre' and those from the Carpathians and single individuals from the British Isles under L. tiscae, besides occurrences of L. 'rupestre' in the Carpathian Arc as well(Martens 1978).Šilhavý (1981)was the fi rst to recognize two very similar Leiobunum species in (former) Czechoslovakia: L. rupestre in the West (now Czech Republic) and L. 'glabrum' in the East (now Slovakia).
Strand (1900)6b)ded his species L. gracileThorell, 1876 and L. laeve Thorell, 1876on juvenile specimens from South Sweden, although he mentioned only a single male.His original material (kept at the Museum of Natural History Stockholm) defi nitely concerns the types and was examined.It was found, that it indeed only contains juveniles, as alreadyTullgren (1906b)assumed they represented juveniles of L. rupestre.Likewise,Tullgren (1906b)assumed L. norvegicum Strand, 1900 to be identical with L. rupestre, but refrained from making a fi nal decision, not having seen the type material.Describing L. norvegicum,Strand (1900)depicted a female of the L. rupestre complex (from Kristiania [now Oslo]; leg.Sept.1899)andlaterMartens (1978)agreed with Tullgren's view, including L. norvegicum in the synonymy of a broad species L. rupestre.Considering the allopatric distribution of L. rupestre and L. 'tisciae' we now can fi rmly state that all Leiobunum 'rupestre', based on material from Scandinavia, belong to L. 'tisciae'.But as L. 'tisciae' is a recently proposed name, older names are to be preferred.Albeit L. gracile and L. laeve are oldest, their description is based on subadults