The species was dedicated to Lucijan von Matulić (teacher at a high school in Trebinje and founder of the first Speleological Society in Bosnia and Herzegovina in Trebinje in 1911), thus it was emended to “matulici” by Folkmanová (1946).

Ilijina Pećina (as “Elias Höhle bei Trebinje” in the original description (Verhoeff 1899)) 42°43.63'N, 18°20.17'E. (Type locality of S. jugoslavicus: Vjetrenica Cave – as “grotte sur le mont ‘Brencovac’ près de Zavala en Popovo polje” in the original description (Hoffer 1937), 42°50.752'N, 17°59.028'E).

Type material: female holotype on two slides (Slide No. 266 and 267) housed by the Museum für Naturkunde, Berlin. The slides were mounted in Canada balsam, but in an inappropriate way since they are partially dried out (Figs 2, 3). Such drying may probably happen because of the mixing of the Canada balsam with a diluting-agent, like glow-oil or xylene, at a too high of a level.

Slide No. 266: cephalic capsule, mandibles, maxillae, forcipules and forcipular tergite, half of the 1^st leg-bearing segment’s tergite (Fig. 2).

Slide No. 267: posterior part of body from 12^th segment, legs missing except right 14^th leg and the 15^th pair detached. Right ultimate leg was probably not macerated in any clearing agents before slide mounting, since the muscles are well visible inside (Fig. 3). All the other parts cleared, probably via potassium hydroxide, because their muscles were dissolved.

Figures 2, 3. 

Holotype of Lithobius matulici Verhoeff, 1899 on slides from the Museum für Naturkunde, Berlin 2 slide No. 266 3 slide No. 267.

2 ♀ (HNHM chilo-6330, HNHM chilopr-377), 1 subadult ♀ (HNHM chilopr-378): Bosnia and Herzegovina, Bravenik Cave, Grab (near Trebinje), 42°35.97'N, 18°25.29'E, 20.07–20.09.2008, leg. Roman Lohaj.

A subadult female of 12 mm from the type locality cave (Verhoeff 1900; not studied). Two males and three females from the Vjetrenica Cave (type locality of male Strandiolus jugoslavicus) (Matic and Dărăbanţu 1968; not studied).

A Lithobius Leach, 1814 species (subgenus Lithobius Leach, 1814) of a length about 14–26 mm; with long antennae of 76–110 articles, reaching the posterior end of tergites 8–9 when folded backwards; ocelli absent; Tömösváry’s organ large, with a diameter 0.08–0.1 times of the length of the cephalic plate; 2+2–3+4 obtuse and short teeth on dental margin of forcipular coxosternum, porodonts large, about 2.8–3 times longer and 1.3–2 times broader than teeth; chitin-lines on the forcipular coxosternite reaching the posterior margin of coxosternite; posterior part of 14^th tergite without setae-bearing area in both sexes; legs 1–13 with long anterior and posterior accessory spines; 14^th and 15^th pairs of legs without accessory spines, without secondary sexual characters, and with the following plectrotaxy 15: -,-,(m)p,-,-/-m,mp,m,- and 14: -,-,(m)p,-,-/-m,mp,m,-; 3,4,4,3–5,5,5,5 coxal pores arranged in a single row; female gonopods with 2+2 spurs on first article, gonopodal claw bipartite.

Where differences between specimens from different caves occur, they are highlighted at the given characters.

Body length 14–26 mm (holotype 21.5 mm according to the original description; specimens from Vjetrenica Cave 20–26 mm (26 mm in holotype of S. jugoslavicus), specimens from Bravenik Cave 14–17 mm). Coloration yellowish-white in alcohol. The whole cuticle is thin and rather soft, almost transparent, wrinkled on the cephalic plate and tergites (wrinkling not mentioned for specimens from Vjetrenica Cave). Cephalic plate, forcipules and body without punctae. Cephalic plate as broad as tergite 8, about as broad as long (1.96 mm long and 2.28 mm wide in holotype, but width obviously affected there by flattening at slide-mounting; Fig. 2). Ocelli missing. Tömösváry’s organ very large, with diameter 0.08–0.1 times of the length of the cephalic plate, placed on the ventral to anterolateral margin of cephalic pleurite. Antennae composed of 76–110 articles (in holotype right antenna with 106 articles, left antenna broken and distal part missing; 85–88 articles in holotype of S. jugoslavicus and 106–110 in other specimens from Vjetrenica Cave), long (7.8–18.5 mm, 13.5 mm in holotype, 18.5 mm in holotype of S. jugoslavicus), reaching the posterior end of tergites 8–9. Most articles short, probably from secondary segmentation, with only one whorl of setae (number of whorl of setae not documented in specimens from Vjetrenica Cave but proportion of antennal articles illustrated as the same in Hoffer 1937: fig. 1). Forcipular coxosternite broad, with 2+2–3+4 obtuse and short teeth (usually 3+3 as in the holotype (Fig. 5), in the holotype of S. jugoslavicus and four other specimens from Vjetrenica Cave, and in the specimen HNHM chilo-6330 from Bravenik Cave; 3+4 in only one specimen from Vjetrenica Cave according to Matic and Dărăbanţu (1968: fig. 4c), 2+2 in specimens HNHM chilopr-377–378 (Fig. 4) from Bravenik Cave; porodonts stout and strong, about 2.8–3 times longer and 1.3–2 times broader than teeth; dentate part of the coxosternite concave, shoulder of coxosternite broad (Figs 4, 5); chitin-lines reaching the posterior margin of coxosternite (Fig. 4). Lateral edges of trochanteroprefemur and part of coxosternite extended beyond cephalic plate. Calyx of poison gland 6.5–7 times as long as wide, about ¼ situated in distal half of forcipular tibia (Figs 4, 6) (not known for specimens from the Vjetrenica Cave). Forcipular tergite narrower than cephalic plate with a ratio of about 0.8 (in holotype of S. jugoslavicus similar ratio according to Hoffer 1937: fig. 1, but about 1.1 in his fig. 8; 0.85 for another specimen from the same Vjetrenica Cave according to Matic and Dărăbanţu (1968: fig. 4b)). Lateral sides of body rather parallel, only slightly broadened at tergites 8–10. Tergites 3, 5, 8, 10, 12 and 14 posteriorly rounded, without protuberances; posterior end of tergite 14 semicircular (less pronounced in younger specimens from Bravenik Cave (Fig. 8), almost perfect in the female holotype (Fig. 7) and in the male holotype of S. jugoslavicus illustrated by Hoffer (1937: fig. 1)). Sternites 1–10 longer than broad, sternites 11–15 shorter than broad (sternites 1–11 missing and not documented in holotype). Sternite 15 in female trapeziform, posterolaterally narrower than anterolaterally, with straight posterior border, in male longer than broad according to Hoffer (1937: fig. 10, from Vjetrenica Cave, not documented from other caves). Legs elongated, 14–15^th without modifications. Length of leg articles of holotype (in mm): leg 14: trochanter+prefemur = 1.7, femur = 2.0, tibia = 2.2, tarsus 1 = 2.0, tarsus 2 = 0.8; legs 15: trochanter+prefemur = 1.6–1.7, femur = 2.1–2.2, tibia = 2.2–2.4, tarsus 1 = 2.0–2.1, tarsus 2 = 0.8–0.9. Right ultimate leg of holotype with tarsus 2 having an ‘articulated’ appearance (Fig. 3), although only collapsed as an artefact (probably caused during the mounting). Leg plectrotaxy as in Tables 1–3 (differences between cave populations given in footnotes), spines 1–6VmF and 1VmT missing in the subadult female of ~11 mm (HNHM chilopr-378). Legs 14–15 with claws of usual proportions, without accessory spines (Figs 12, 13); legs 1–13 with elongated claws and with elongated anterior and posterior accessory spines (Figs 10, 11), relative length of accessory spines highest on legs 11–12: about 0.5 of claw’s length for the anterior and 0.3 for the posterior spine (from Vjetrenica Cave Hoffer (1937: fig. 6) illustrated for leg 13 ratios of about 0.4 in both spines, while Matic and Dărăbanţu (1968: fig. 4g) illustrated for leg 10 ratios of 0.8 and 0.2). 3,4,4,3–5,5,5,5 coxal pores arranged in one line. In the original description Verhoeff (1899) mentioned 2(+1),3,4,3 as number for coxal pores in the holotype, but in fact it is 4,4,4,3 on legs 12–15 respectively; in S. jugoslavicus only legs 14–15 were documented with 5 and 4 coxal pores respectively (Hoffer 1937: fig. 10; in the text erroneously mentioned 4 and 5 respectively, which would be an unusual pattern in Lithobiomorpha). For the specimens from the same Vjetrenica Cave Matic and Dărăbanţu (1968) mentioned 5,5,5,5 coxal pores, while in the specimens from Bravenik Cave we found 3,4,4(5),3(4).

Female first genital sternite longer than wide, with 22–40 evenly scattered setae (40 in holotype; not known in specimens from Vjetrenica Cave); posterior border almost straight (Fig. 14) (not known in specimens from Vjetrenica Cave). Female gonopods with thin setae and 2+2 elongated spurs on first article (holotype in Fig. 14; unequal spurs in younger adults as in Figs 15, 16; 1+1 in a subadult specimen in Fig. 17). Lateral side of female gonopods with 7–12 moderate to long setae on first article, 5–8 setae on second and 1 or 2 setae on third article, arranged as in Figures 14–16 (only 4 setae on first article in a specimen from Vjetrenica Cave according to Matic and Dărăbanţu (1968: fig. 4d) but their drawing is probably inaccurate in this detail); dorsal side of gonopod with about 4 weak spines on second article and 1–3 minute spines on third article (Figs 14–16), medial side of female gonopods without setae (not known in specimens from Vjetrenica Cave). Gonopodal claw bipartite (on left gonopod of holotype (Fig. 14) misinterpreted by Verhoeff (1899: fig. V) as tripartite); medial tip smaller than lateral (Fig. 16).

Figures 4–17. 

Lithobius matulici Verhoeff, 1899 (holotype 5–7, 14; HNHM chilopr-377 4, 8–13, 15–16; HNHM chilopr-378 17) 4 forcipules and trunk segments 1–2, left side of forcipules with ventral view, right side with dorsal view 5 coxosternal dentation, left side with dorsal view, right side with ventral view 6 tarsungulum and forcipular tibia of the holotype (ventral view) 7–8 tergites 13–14 9 right leg 1 (anterior view) 10 claw of right leg 1 (anterior view) 11 claw of right leg 13 (anterior view) 12 claw of right leg 14 (posteriomedial view) 13 claw of left leg 15 (posteromedial view) 14 gonopods of holotype 15 female gonopod (lateral view) 16 female gonopod (anterior view) 17 subadult female gonopod (right, lateral view).

Strandiolus jugoslavicus was described by Hoffer (1937) on a single male specimen from the Vjetrenica Cave (“grotte sur le mont ‘Brencovac’ près de Zavala en Popovo polje”, 42°50.752'N 17°59.028'E) without comparison with Lithobius matulici Verhoeff, 1899, known from another cave only about 32 km away. The depository of the type is unknown, and it was not found at the National Museum in Prague (Dolejš 2015) where that part of Hoffer’s material is housed that we know to exist. However, the original description is very detailed, supplemented with illustrations, and fits in every important character with Verhoeff’s original description, but also with the holotype of matulici, as well as the fresh material studied by us. It also fits the five topotypic specimens described by Matic and Dărăbanţu (1968). The fact that Matic and Dărăbanţu (1968) identified these topotypic specimens as L. matulici (without any notes on S. jugoslavicus) also supports our conclusion that S. jugoslavicus is a subjective junior synonym of L. matulici (syn. nov.). Because also neighbouring caves might be completely isolated from each other, high-level genetic separation of cave populations might occur even in cases where no morphological differences of the specimens are obvious. Future molecular studies might easily support our decision based on morphology.

The posteriorly semicircular form of the 14^th tergite has not been highlighted for this species by the earlier authors, although it was illustrated by Hoffer’s (1937: fig. 1) drawing on the habitus of the holotype of S. jugoslavicus and Verhoeff (1899) mentioned that the posterior corners of the tergites 3, 5, 8, 10, 12, and 14 are exceptionally strongly rounded. It is present in the holotype of matulici (Fig. 7) and in our fresh specimens as well. Matic and Dărăbanţu (1968) seem to have overlooked this character, as they only mentioned that the tergites are without posterior triangular projections. Hoffer (1937) characterised the tergites as of oval in shape, but for more details he referred to his drawing with the holotype which depicts tergite 14 with rounded posterior margin.

The shape of the 14^th tergite seems to indicate a close relation of L. matulici to the members of the subgenus Thracolithobius Matic, 1962 (Zapparoli and Edgecombe 2011), especially to its type species, Lithobius remyi, described from the Gradje Cave (Montenegro), which is only 95–150 km from the known occurrences of L. matulici, and also reported from the North Albanian Merkurth Cave (Stoev 1996). As the posteriorly semicircular form of the 14^th tergite is the key character defining Thracolithobius, we could consider L. matulici as member of this subgenus, but we refrain to do for reasons of nomenclatural stability. Including L. matulici in Thracolithobius would result in a situation in which the generic name Strandiolus Hoffer, 1937 would became a subjective senior synonym of Thracolithobius Matic, 1962 according to the principle of priority (ICZN 1999: Art. 23), because its type species, Strandiolus jugoslavicus Hoffer, 1937, is synonymised in the present paper under L. matulici (see above). Strandiolus was synonymised under Lithobius (s.s.) by Stoev (1997) (also proposed earlier informally and without explanation by Folkmanová (1946) in a key) because its differential characters are either actually common in Lithobius (s.s.) – three ‘claws’ on legs 1–13, reduced leg plectrotaxy, notched lateral edges of head, absence of tergal projections, form of maxillae II – or adaptations to the cave environment – absence of ocelli, elongation of legs and narrow anterior sternites, depigmentation, high number of antennal articles – and as such of no taxonomical importance. Meanwhile, Thracolithobius Matic, 1962 is considered as a valid subgenus (Stoev 1997; Shelley 2006; Ćurčić et al. 2008; Zapparoli and Edgecombe 2011) with three species – L. dacicus Matic, 1959, L. inexpectatus Matic, 1962, L. remyi Jawlowski, 1933 – but the monopyhly of this group might be questioned. The only common character defining this subgenus is the shape of the 14^th tergite, a character that however has already been proven to vary at the inter(sub)specific level in Lithobius (Andersson 1979) and in another lithobiomorph genus, Eupolybothrus (Stoev et al. 2013; Akkari et al. 2017). Apart from this character, the members of the subgenus seem to be rather different in several other features (e.g. presence/absence of ocelli and a wart-like structure on forcipular tarsungulum) and L. matulici differs actually from the members of Thracolithobius even in an aspect of the 14^th tergite: the rounded shape is present in matulici also in females, while it is known only from males in the other species. Although at least L. remyi and L. matulici seem to be similar also in some other features (lack of ocelli, strong porodonts, coxosternal dentation) this may be also due to convergent adaptation to a similar lifestyle in cave environments.

According to this, we can expect that molecular studies will prove Thracolithobius to be polyphyletic with its members spread among different clades of Lithobius (s.l.), which would result in its synonymisation under Lithobius (s.s.); and this would be the case again even if its name would be changed here to the older name Strandiolus. In case future molecular studies give an opposite result (i.e. monophyly of Thracolithobius including L. matulici), Strandiolus might be revalidated.

Among the Lithobius species with a posteriorly rounded tergite 14, L. matulici seems to be most similar to L. remyi, but differs from that species in size (11–13 mm in remyi, 14–26 mm in matulici), number of antennal articles (56–64 in remyi, 76–110 in matulici), and the shape of the female gonopodal claw (tripartite in remyi, bipartite in matulici). From L. dacicus, L. matulici differs in size (about 12 mm in dacicus, 14–26 mm in matulici), number of antennal articles (37–61 in dacicus, 76–110 in matulici), coxosternal dentation (2+2–3+4 small and obtuse teeth in matulici, 2+2 well developed teeth in dacicus), and completeness of coxosternal chitin-lines (not reaching the posterior margin of the coxosternite in dacicus, reaching it in matulici). Lithobius inexpectatus is distinguished from L. matulici by having 12–14 ocelli (missing in matulici), by the coxosternal dentation (2+2–3+4 small and obtuse teeth and very strong porodonts in matulici, 2+2 larger teeth and slender porodonts in inexpectatus), the number of antennal articles (42 in inexpectatus, 76–110 in matulici), the presence of accessory spines on legs 14–15 (absent in matulici), the shape of the female gonopod claw (tripartite in inexpectatus, bipartite in matulici), and plectrotaxy (1–15VaF, 1–13VaT, 1–14VpT, 8–15DaP, 1–15DpP, 1–13DaF, 3–15DpF and 3–15DpT present in inexpectatus, missing in matulici).

Although no rounded form of tergite 14 is known for it, L. sketi was stated to be very similar to L. matulici, and they also co-occur in Vjetrenica Cave (Matic and Dărăbanţu 1968). The two species are readily distinguished by the accessory spines on the 14–15^th legs (present in sketi, missing in matulici), by the number and arrangement of coxal pores (5–9 per coxa arranged in 2 partly irregular rows in sketi, 3–5 per coxa in a single row in matulici), the female gonopods (1+1 spurs and simple claw in sketi, 2+2 spurs and bipartite claw in matulici), and their plectrotaxy (1–13VpP, 1–15DaP, 1–15DpP, 1–14DaF, 1–15DpF and 2–15DpT present in sketi, missing in matulici).

One lithobiomorph specimen was found in July 2014 while one of the authors, G. Balázs, was diving in Vjetrenica Cave. The specimen was in a water-filled part of the cave (Donje Vjetrenica), freely and consciously walking on the underwater bottom at a depth of 3 metres, at a distance of about 30 metres from any terrestrial microhabitats (i.e. chambers with air). This specimen was without any signs of distress (no spasms, no enfeeblement). There was no flood in the cave at that time, the water was still (not flowing), and thus a simple flushing away of the specimen from the water’s edge might be ruled out. This individual spent another 2 hours in the water, while kept captured by the diver and escaped later during photographic documentation. In the photograph (Fig. 18), the 14^th tergite of the specimen seems clearly rounded posteriorly, and thus it can be considered as L. matulici with confidence. Similar cases of lithobiomorph specimens on the bottom of water (puddles) in caves were photo-documented in Montenegro (Dobuki Do: 42°25.739'N 18°48.716'E: August 2006, Zsolt Polacsek in litt., Figs 19, 20; July 2018, Márton Mede in litt.), and from a cave in North Spain (Tibia-Fresca Cave System, 19 July 2016, see supplementary file 1: Video 1; Zsolt Polacsek in litt.).

Figures 18–20. 

Lithobius specimens from Dinaric caves 18 living Lithobius cf. matulici specimen of ca. 25 mm length from the Vjetrenica Cave (Bosnia and Herzegovina) (photo by Gergely Balázs) 19–20 Lithobius sp. under water in the Dobuki Do Cave (Montenegro) (photos by Zsolt Polacsek).