The subgenus Monotarsobius in the Iberian Peninsula with a description of a new pseudo-cryptic species from Northern Spain revealed by an integrative revision of Lithobius crassipes L. Koch, 1862 (Chilopoda, Lithobiomorpha, Lithobiidae)

Abstract The widespread European centipede species Lithobius (Monotarsobius) crassipes L. Koch, 1862 was revised using an integrative approach incorporating sequence data and morphology. The partial mitochondrial cytochrome c oxidase subunit I (COI) barcoding gene was amplified and sequenced for 21 individuals from northern Spain, France and Germany as well as for individuals of three other species of the subgenus Monotarsobius Verhoeff, 1905. The dataset was used for molecular phylogenetic analysis and genetic distance determination. In addition, Monotarsobius specimens from more than 100 localities in northern Spain, France, and Germany were morphologically investigated. Both morphological and molecular data indicate that specimens from the Navarre and Gipuzkoa provinces, northern Spain, represent a distinct pseudo-cryptic species, only differing in some minor characters from L. crassipes. The new species L. (Monotarsobius) crassipesoides sp. n. is described and compared to L. crassipes in detail using morphology and morphometric statistics for body, head, and antennae length, number of ocelli and coxal pores, as well as the starting leg for legpair spines Vmt and DaP. The Iberian and European records of L. crassipes are discussed. The subspecies L. crassipes morenoi Garcia Ruiz, 2014 from Southern Spain is elevated to species as L. morenoi stat. n. A checklist, distribution map and key to all five species of Monotarsobius of the Iberian Peninsula are presented.

(GRETIA) collection in Nort-sur-Erdre, France. The French specimens used for DNA extraction came from the forest of the Gȃvre (collected by E. Iorio), in the municipality of Le Gâvre in Loire-Atlantique department and are kept in the SMNG collection. All material is preserved in 70 % or 96 % ethanol (DNA vouchers) respectively. During the German Barcoding of Life project Myriapoda , Wesener et al. 2015, 2016 22 specimens were collected for sequencing from various localities in Germany, but also from eastern France, Austria and Wales. The vouchers are deposited in the Zoologische Staatssammlung München (ZSM).
Sixty-four specimens of L. crassipesoides sp. n. from Spain were morphologically studied. For L. crassipes 55 specimens from 36 localities in Germany and 131 specimens from 60 localities in France were studied (see supplementary Table 1 for localities and collection numbers). Only adult specimens were used for morphological characters. Characters mentioned in the description of L. crassipesoides sp. n. are very similar or like in L. crassipes, unless stated otherwise. All specimens for molecular study were investigated for morphology prior to DNA extraction.
Illustrations. For scanning electron micrographs (SEM), samples were dehydrated through an ethanol series, dried in a desiccator overnight, and mounted on aluminum stubs before being sputter coated with gold-palladium. SEM images were taken using a JEOL JSM-6510LV microscope, and samples were removed from stubs and returned to alcohol upon examination.
Preserved specimens were imaged with a Leica M165 C or Leica DM5500B stereo microscope and DFC295 camera. Focus-stacked images were assembled from 10-25 source images using the software package Leica V4.5. All images of the z-stacks are available online at VIRMISCO (www.virmisco.org, Christian et al. 2017).
All figures were later edited using Adobe Photoshop CS4. Maps were created with ArcGIS 10.
DNA extraction and molecular analysis. At the SMNG, DNA was extracted from 2-4 legs each of six specimens of L. crassipesoides sp. n., three specimens of L. crassipes and one specimen of L. forficatus (Linnaeus, 1758) (Table 1). Total genomic DNA was extracted using the Qiagen DNAeasy Blood & Tissue kit following the standard protocol except that tissue was incubated for 48 h. All specimens were later deposited in the collections of the SMNG.
Polymerase chain reaction (PCR) was used for amplifying the COI bar code fragment using the primer pair LCO1490 and HCO2198 (Folmer et al. 1994). The following thermocycling profile was used to amplify fragments of COI: predenaturation at 95° C for 1 min, 35 cycles of 40 s at 94° C, 40 s at 51° C, and 1 min at 72° C, final extension step for 5 min at 72° C. All PCR mixes had a total volume of 10 µL and contained 1 µL template, 0.1 mM of each primer, 4 × 0.15mM dNTPs [Peqlab], 1 × PCR Buffer containing 2 mM MgCl 2 [Peqlab], and 0.2u Polymerase [Peqlab]. All fragments were sequenced in both directions by Biodiversity and Climate Laboratory Centre, Frankfurt, Germany.
As part of the GBOL project at ZSM, 11 specimens of L. crassipes, five specimens of L. curtipes C. L. Koch, 1847, and six L. austriacus (Verhoeff, 1937) were extracted Table 1. Species, localities, GenBank accession numbers, and repository accession numbers for all specimens analyzed. and sequenced by the Canadian Centre for DNA Barcoding (CCDB, Guelph, Canada) using standardized, high-throughput DNA extraction, PCR amplification and bidirectional Sanger sequencing (http://www.ccdb.ca/resources.php). For PCR and sequencing, a primer cocktail (Hebert et al. 2004) was used. See also Wesener et al. (2015) and Wesener et al. (2016).
All 32 new sequences were deposited in GenBank (see Table 1 for accession numbers). Alignment and phylogenetic analysis. All obtained sequences were checked via Blast searches (Altschul et al. 1997); no contaminations were discovered. The sequences were aligned by hand in ClustalX ver. 1.83 (Chenna et al. 2003). One sequence for L. crassipes (Eitzinger et al. 2013) was downloaded from GenBank. As outgroups sequences Eupolybothrus cavernicolus Komerički & Stoev, 2013(Stoev et al. 2013, Stenotaenia linearis (C.L. Koch, 1835) (Wesener et al. 2015) and Cryptops parisi Brölemann, 1920(Wesener et al. 2016 were downloaded from GenBank (see Table 1 for accession numbers). The final dataset for the phylogenetic analysis included 26 sequences and the alignments consisted of 657 bp (COI mtDNA). To find the best substitution model, Modeltest as implemented in MEGA 7 (Kumar et al. 2016) was utilised. The lowest Bayesian Information Criterion score (BIC) was obtained for the Tamura-Nei model plus gamma distribution with indifferent sites (Tamura and Nei 1993) (BIC 11667.16). Maximum likelihood analyses were conducted in MEGA7 (Kumar et al. 2016). The tree with the highest log-likelihood (-5445.7387) is shown. Initial tree(s) for the heuristic search were obtained by applying the neighbour-joining method to a matrix of pairwise distances estimated using the maximum composite likelihood approach. A discrete gamma distribution was used to model evolutionary rate differences among sites (five categories (+G, parameter = 1.0671)). The tree ( Fig. 1) is drawn to scale, with branch lengths measured in the number of substitutions per site. The bootstrap consensus tree inferred from 1000 replicates (Felsenstein 1985) is here used as the best estimate of the phylogeny of the analysed taxa ( Fig. 1). The final trees were edited using Adobe Illustrator CS4. Mean uncorrected pairwise distances between terminals (transformed into percentages) were determined using MEGA7 (Kumar et al. 2016).
Statistical analysis. The software package PAST version 3.14 (Hammer et al. 2001) was used for analysis.
Box and jitter plots: For each sample, the 25-75 percent quartiles are drawn using a box. The median is shown with a horizontal line inside the box. The minimal and maximal values are shown with short horizontal lines ("whiskers"). Each value is plotted as a dot. To show overlapping points more clearly, they have been displaced using a random "jitter" value controlled by a slider.
Usually nonparametric statistics have been used, while parametric statistics have been restricted to interval data. Prior to the parametric t-test normality was tested using three statistical tests for normal distribution: The Shapiro-Wilk test (Shapiro and Wilk 1965), the Jarque-Bera-Test (Jarque and Bera 1987) and the Anderson-Darling-Test with significance estimated according to Stephens (1986). For the Jarque-Bera-Test and Anderson-Darling-Test also a significance test has been included based on Monte Carlo simulation, with 10,000 random values taken from a normal distribution has been included. If one of these normality tests rejected the null hypothesis of equal distribution only non-parametric statistics were used for the following tests.
In the xy graph of the canonical variate analysis (Figs 9-11) 95 % concentration ellipses have been plotted. Their calculation assumed normal distribution for the values of the discriminant factors. They estimate a region where 95 % of population points are expected to fall, i.e. they are not confidence regions for the mean.
The following characters have only been investigated in L. crassipesoides sp. n. from Spain and L. crassipes from Germany: the head length, the number of coxal pores on legpair 12-15 and the start of the ventral median spine on the trochanter.
Terminology. The terminology of external morphology follows Bonato et al. (2010). The coloration as seen in alcohol material under a stereo microscope follows the colour terminology of Köhler (2012). The

Molecular analysis
The monophyly of both Lithobius crassipesoides sp. n. and L. crassipes is well supported with bootstrap values of 99 and 98 respectively ( Fig. 1 Statistics of body length. A comparison of jitter box plots shows that the specimens of L. crassipesoides sp. n. from Spain are usually smaller than those of L. crassipes (Fig. 2). The basic descriptive statistic parameters are given in Table 2.
The performed tests for normality (Table 3) showed significant differences from a normal distribution within the specimens from France, while there had been no significant deviation between the two sexes in the specimens from Spain and Germany. Based on the significance of the French samples it seemed to be wise to use nonparametric tests solely.   Using the Mann-Whitney-U-Test for mean, no significant differences in body length have been found between males and females of L. crassipesoides sp. n. from Spain (p=0.108) and between L. crassipes from France (0.433) and Germany (p=0.067). Therefore, it was justified to pool the samples. These pooled samples showed, that specimens of L. crassipesoides sp. n. are significantly shorter than specimens of L. crassipes from France (p<0.001) and Germany (p<0.001), while the two samples of L. crassipes do not significantly differ (p=0.081).

Statistics of head length
A comparison of jitter box plots shows, that the specimens of L. crassipesoides sp. n. from Spain usually have shorter heads than those of L. crassipes from Germany (Fig. 3). The basic descriptive statistic parameters are given in Table 4.
The performed tests for normality (Table 5) showed significant differences from a normal distribution within the specimens from Germany in the Shapiro-Wilk W and Anderson-Darling A tests, while there had been no significant deviation between the two sexes in the specimens from Spain. Based on the significance of the German samples it seemed to be wise to use non-parametric tests solely.
Using the Mann-Whitney-U-Test, no significant differences in head length have been found between males and females of L. crassipesoides sp. n. (p=0.105) and L. crassipes from Germany (0.655). The pooled samples show that head length was significantly shorter in specimens of L. crassipesoides sp. n. than in specimens of L. crassipes from Germany (p<0.001).  Statistics of antennae length. A comparison of jitter box plots shows, that the antennae of the specimens of L. crassipesoides sp. n. from Spain are usually smaller than those of L. crassipes (Fig. 4). The basic descriptive statistic parameters are given in Table 6.
The performed tests for normality (Table 7) showed low significant differences (p between 0.05 and 0.01) from a normal distribution only within pooled specimens from France and Spain, while there had been no significant deviation between the two sexes in the specimens from all countries and the pooled specimens from Germany. Based on the low significance of the French and Spanish samples it seemed to be wise to use non-parametric tests solely.
Using the Mann-Whitney-U-Test, only low significant differences in antennal length have been found between males and females of L. crassipesoides sp. n. from Spain (p=0.046) but not between L. crassipes from France (0.655) and Germany (p=0.555). The very low significance in specimens of L. crassipesoides sp. n. solely is interpreted  as a stochastic effect and as there are no sex-specific differences in both of the L. crassipes samples they are pooled for comparison of the two species. These pooled samples showed that the antennae of specimens of L. crassipesoides sp. n. are significantly shorter than in specimens of L. crassipes from France (p<0.001) and Germany (p<0.001, while the two samples of L. crassipes do not significantly differ (p=0.632).

Statistics of number of ocelli.
The number of ocelli was counted on each side of the head separately. For statistical evaluation, the average of the two sides has been used, as the separate measures are not independent samples (Lamprecht 1992). This approach also allows using ocelli numbers from a single side if the other was not countable.
A comparison of jitter box plots shows that the specimens of L. crassipesoides sp. n. from Spain usually have a lower number of ocelli than specimens of L. crassipes from France and Germany (Fig. 5). The basic descriptive statistic parameters are given in Table 8.
Using the Mann-Whitney-U-Test, no significant differences in the number of ocelli have been found between males and females of L. crassipesoides sp. n. from Spain (p=0.063) and L. crassipes from France (0.438) and Germany (p=0.074). The pooled samples show that the number of ocelli in specimens of L. crassipesoides sp. n. is significantly lower than in specimens of L. crassipes from France (p<0.001) and Germany (p<0.001, while the two samples of L. crassipes do not significantly differ (p=0.293).
Statistics of number of coxal pores. A comparison of jitter box plots shows that specimens of L. crassipesoides sp. n. usually have a lower number of coxal pores on leg-   pair 12 to 15 and that females usually have a higher number of coxal pores on them (Fig. 6). The basic descriptive statistic parameters are given in Table 9.
In Table 10 the differences in coxal pore numbers between males and females as well as between the different species have been tested using the Mann-Whitney-U-Test. This test showed that females usually have a higher number of coxal pores than males. This had been more obvious in specimens of L. crassipes from Germany (significant p values in all legs) and also more obvious on the coxopleura of legpair 14 and 15 (significant p values in both species). Therefore, comparisons have to be made sex specific. These sex specific tests showed significant differences in the number of coxal pores on legpair 15 solely in males, while in females it was exactly legpair 15 that showed no significant p value. Pooled samples showed significant p values for coxal pore numbers in all legs, but this comparison is of doubtful value as it represents an average of the sex specific differences. Statistics of legpair DaP spine. The start of the anterior dorsal spine (DaP) at the prefemur has sometimes been checked on both sides, sometimes on one side solely, depending on the availability of intact specimens with at least one leg being complete on one side at a given legpair position. Again, the average value has been used for the statistical calculation if both values had been available. Due to the high number of specimens with missing legs the number of samples was much lower than in the previously processed characters, see Table 11 for basic descriptive statistic parameters.  A comparison of jitter box plots shows that the spine DaP usually starts later in specimens of L. crassipesoides sp. n. from Spain than in specimens of L. crassipes (Fig. 7), with an overlap at legpair 12 and 13 in females solely.
Using the Mann-Whitney-U-Test, no significant differences in the starting position of spine DaP have been found between males and females of L. crassipesoides sp. n. from Spain (p=0.209) and L. crassipes from France (p=0.340), while there was a low significance between males and females from Germany (p=0.024). The pooled samples show that spine DaP starts significantly later in specimens of L. crassipesoides sp. n. than in specimens of L. crassipes from France (p<0.001) and Germany (p<0.001). However, the two samples of L. crassipes differ significantly (p=0.001).

Statistics of median ventral spine at the trochanter
The start of the median ventral spine at the trochanter (Vmt) has sometimes been checked on one side solely. The basic descriptive statistic parameters are given in Table 12. A comparison of jitter box plots shows that the spine Vmt usually starts earlier in specimens of L. crassipesoides sp. n. from Spain than in specimens of L. crassipes (Fig. 8), with an overlap at legpair 13.
Using the Mann-Whitney-U-Test, no significant differences in the starting position of spine Vmt have been found between males and females of L. crassipesoides sp. n. (p=0.777) and L. crassipes from Germany (p=0.771). The pooled samples show that spine Vmt appeared significantly earlier in specimens of L. crassipesoides sp. n. than in specimens of L. crassipes from Germany (p<0.001).

Canonical variate analysis
A canonical variate analysis (CVA, linear discriminant analysis with several groups) has been performed to improve the separation of the two species. As complete datasets are necessary for such a comparison only samples from Germany could be used to represent L. crassipes. All morphometric characters evaluated above have been used and sex specific groups have been defined. Additionally, the ratios of body length and antenna length to head length have been included. The CVA (Fig. 9) allowed a complete separation of all specimens, as shown by the convex hulls. Also, the 95 % concentration ellipses show a complete separation, although there was one female of L. crassipes (SMNG VNR 15170-2 b) that fell outside the concentration ellipse.
Although differences between the sexes are clearly visible, even CVA does not allow separating them completely. Three discriminant factors have been extracted, explaining 85.8 %, 12.1 % and 2.1 % of the observed variance.
The factor loadings of discriminant function one show that the start of spine DaP and the body length have the highest separating power for the two species (Table 13). Ocelli number, start of spine Vmt and antennal length are also of value. Discriminant factor one could be interpreted as a "size factor". It is remarkable that head length and both ratios are of negligible influence. They are just redundant expressions of the size factor, which is best represented by body length.
Discriminant factor two could be interpreted as a "sex factor", with the highest factor loading on the number of coxal pores on legpair 15. Ocelli number and start of DaP are also of importance, while the numbers of coxal pores on the other legs are redundant characters.
Factor three is of no practical value and explains only a negligible part of the variance (Fig. 10).
Based on these interpretations a reduced model could be defined, which has a lower separating power but requires less measuring. Fig. 11 shows the first two discriminant factors. The parameters of the reduced model are given in Table 14.

Molecular and morphological comparison of L. crassipesoides sp. n. and L. crassipes
Both molecular analysis (Fig. 1) and morphology support the hypothesis that the northern Spanish L. crassipes-like specimens represent a species genetically and morphologically distinct from L. crassipes. The two species are morphologically similar and probably closely related, but can be distinguished as both males and females. While L. crassipes is widely distributed in Europe, L. crassipesoides sp. n. is currently only known from northern Spain (Fig. 17).

Statistical analysis
The main weakness of the morphometric statistical evaluation is the fact that the measured specimens are not identical with the sequenced specimens. This means that the possibility of having other cryptic species among the material cannot be excluded. The rejection of the normal distribution in the French samples in body length, in the German samples in head length and in samples from Spain and France in antennae length might be interpreted as a hint for more cryptic lineages.

Canonical variate analysis
The moderate dropping of the eigenvalue of axis1 from 10.6 to 9.84 (7 %) and of axis2 from 1.49 to 1.23 (1.5 %) shows that no crucial variables have been removed when generating the reduced model. This is also visible by comparison of Fig. 9 versus Fig. 11, as the overlapping of the convex hulls changes only at a small extent. To check if the measures antennae length and head length represent discriminating power beside a size factor, their quotients had also been included in the complete model of the CVA. As visible in Table 13 head length shows only a low, negligible factor loading of -0.014, while antennae length shows a distinctly higher loading of -0.125 and the quotients AL/HL and BL/ AL show lower loadings, but these are still higher than the loading of HL solely. Also a model without AL and HL was calculated, but the factor loadings of AL/HL and BL/ AL did not change their values. This means that in contrast to HL AL has a discriminant value beside a size factor, and that using AL solely in a model would be the best choice. As seen in Fig. 10 axis3 shows no clear separation for any group. This is also reflected by the low eigenvalue, explaining only 2.1 % of the overall variability in the complete and 1.1 % in the reduced model. Furthermore axis1 and axis3 show the highest factor loadings on the same variables, marked in bold letters in Table 14. For that reason it is justified to omit axis3 in further considerations.

Lithobius crassipes on the Iberian Peninsula and elsewhere in Europe
Currently, all records of L. crassipes on the Iberian Peninsula as well as in other parts of Europe need to be carefully checked against L. crassipesoides sp. n., which could also be dispersed by human transport , Stoev et al. 2010). In the southern part of Spain more pseudo-cryptic species of the Lithobius crassipes-group may also exist. L. crassipes is formally known from the Pyrénées-Atlantiques department (French Basque country) (Iorio 2014(Iorio , 2016, which is adjacent to the Spanish Navarre province. However, as mentioned above, L. crassipesoides sp. n. occurs only a few hundred meters or few kilometers from the French border and can be assumed to occur at least in the most western region of the French Pyrenees. Interestingly, some authors have previously given details on morphology of Spanish L. crassipes. Serra (1980) described specimens of L. crassipes from the province of Almería as follows: 6-11 mm body length, 6-11 ocelli in 2 or 3 rows, a DaP spine on legpair 10 (or even sometimes on legpair 7) to legpair 15. Perhaps his description included specimens of both L. crassipes and L. crassipesoides sp. n. Finally, Serra (1982) mentioned the same details as above for some L. crassipes from the provinces Asturias and Almería but only gave the plectrotaxy of legpair 15, not the preceding legpairs.
Here we give a brief review of historic descriptions of L. crassipes for characters which have been found to be useful. Descriptions of L. crassipes elsewhere in Western Europe quoted a body length of 9-12 mm (Iorio 2008, 2010, Iorio and Labroche 2015 or 9-13 mm (Meinert 1872), up to 13.5 mm (Barber 2009); Koch (1862) gave the body length in "lines" (= lignes in French, an old measurement unit), thus his 4-4.5 lines may correspond to 9.02-10.15 mm. Manfredi (1957), in her description of L. crassipes stictonotus Manfredi, 1957 from southern Italy, a synonym of typical L. crassipes according to Bonato et al. (2016), has recorded 11.5-14 mm. Only Brölemann (1930) mentioned a small size, with 6.5-10 mm. Also, the number of ocelli ranged from 8 to 11 in two rows (Brölemann 1930), 10 to 11 in three rows (Koch 1862), 8 to 12 in two or three rows (Meinert 1872, Iorio 2008, 2010 or 9 to 13 in two or three rows (Barber 2009). Manfredi (1957) gave 12 or 13 ocelli for L. crassipes stictonotus. The DaP spine starts at legpair 10 (or even on legpair 7 to 9) to legpair 15 according to Brölemann (1930) and Eason (1964). Brölemann (1930) noted that the DaP spine is only sometimes present on legpair 15. Despite several hundred of specimens of L. crassipes having been examined from France and Germany (this study and unpublished data), no specimen with an absence of a DaP on legpair 15 was observed. Brölemann (1930) may have had specimens of L. crassipesoides among his material.
In Sweden, western specimens of L. crassipes seems to differ from the south-eastern Swedish L. crassipes (including also those from Denmark) as underlined by Andersson (1981). In the west, a DaP spine is only present on the last one, two or three legpairs, or a DaP may even been absent. In the south-east, L. crassipes have a DaP spine at least from legpair 12 to 15, but usually on more legs, 88 % from legpair 10 to 15, and it can even start on legpair 5 (Andersson 1981). The western Swedish specimens have a body length of 8-10 mm, and 8-10 ocelli; the south-eastern specimens are 9-11 mm long with 9-12 ocelli (Andersson 1981). Andersson et al. (2005) also record both forms, but without any supplementary information. We note that the south-eastern Swedish and the Danish specimens correspond well with our L. crassipes from France and Germany.

Diagnosis. Small member (body length 6.4-11 mm) of the subgenus Monotarsobius.
Antennae with 20 articles, short, 2.6 times longer than head, 1/4 of body length. 5-11 ocelli, mostly 8 or 9, in two or three rows with one larger posterior ocellus. Legs with species-specific plectrotaxy; legpair 14 and 15 thickened in both sexes, much more so in males; legpair 15 without accessory apical claw, in males with a depression in the posterior half of tibia. Female gonopod claw tridentate. L. crassipesoides sp. n. differs generally from other Iberian members of Monotarsobius in the presence of a depression in the posterior half of the legpair 15 tibia. It differs from L. osellai and L. morenoi in having more than one row of ocelli; from L. blascoi in having only 20 antennal articles; and from L. crassipes in smaller body length, shorter antennae, lower number of ocelli, the DaP spine starting posteriorly from legpair (12) 13, and the different location and size of the male depression on legpair 15 tibia.
Etymology. Derived from the morphological similarity to Lithobius crassipes.
Tergites. Surface slightly rough, glossy. Posterior border of T1 feebly concave or straight, T3 to T5 feebly concave, T8 to T15 distinctly concave, T16 feebly to distinctly concave. Posterior angles of T9, T11 and T13 mostly obtuse or rounded with no trace of lobes or triangular projections.
Legs. Tarsus and metatarsus fused on legpair 1 to 11. On legpair 12 and 13 the tarsal-metatarsal articulation is indistinct. Penultimate and ultimate legpairs (14,15) are densely covered with pores. Last two legpairs are thickened in both sexes, much more so in males. Without accessory apical claw on legpair 15.
Plectrotaxy. The plectrotaxy of legs of L. crassipesoides sp. n. is given in Table 16. It differs from L. crassipes in the absence of a DaP spine up to legpair 11, DaP very rarely present on legpair 12, rarely present on legpair 13 to 14, while almost always present in L. crassipes on legpair 10 to 15, frequently on legpair 9, rarely also on legpair 7 and 8 or only on legpair 11 to 15 (Table 17).
Distribution. So far only known from Navarre and Gipuzkoa provinces, northern Spain (Fig. 17). Some records are only a few kilometers from the French border. L. crassipesoides sp. n. is therefore expected also in the Western Pyrenees in France.  Habitat. The species was mostly found in the leaf litter and under the bark of dead wood in mountain deciduous forests from 550 to 1300 m a.s.l.
Remarks. L. crassipes was recorded from the Navarre region from Irati (Salinas 1990) and several locations near Quinto Real Herrera 1980, 1982). Barace and Herrera (1980) also gave a brief description of 12 males and eight females. Most of the material studied by Herrera (1980, 1982) and Salinas (1990) was available for re-examination at MZNA (38 specimens) and was confirmed to be L. crassipesoides sp. n. Therefore, all records of Herrera (1980, 1982) and Salinas (1990) of L. crassipes are hereby assigned to L. crassipesoides sp. n. Finally, as mentioned above, the description of Serra (1980) for L. crassipes is possibly composite and perhaps partially concerns L. crassipesoides sp. n.
Colour. General body colouration varies from pale horn colour to cream colour and to yellow ochre (Fig. 12B) or rarely with very light reddish tint.
All current records of L. crassipes on the Iberian Peninsula are doubtful and need further verification (see discussion above). Some records from the literature have been assigned here to L. crassipesoides sp. n. or L. morenoi. Distribution. Known from two caves (Sima LQ-14, Abuchite, Luque; Sima de la Sierrezuela/Sima Fuente del Francés, Carcabuey) in Andalusia and from Jerez de la Frontera (Cádiz), Spain (Fig. 17).

Lithobius (Monotarsobius) morenoi
Remarks. This subspecies was described by Garcia Ruiz and Baena (2014) based on specimens from two caves in the Cordoba province in Andalusia. It clearly differs from L. crassipes crassipes in the low number of ocelli (only three), the enlarged Tömösváry's organ, and the absence of some spines, particularly of Vmt, VaF, and DaH on legpair 15. Garcia Ruiz and Baena (2014) also discuss the record of L. crassipes by Serra (1982) coming from Cv. de las Motillas, Jerez de la Frontera (Cádiz), which was also considered by them as probably representing L. crassipes morenoi. The specimens of Serra (1982) have the same morphological peculiarities regarding number of ocelli (three), Tömösváry's organ and modification of male legpair 15 tibia as the subspecies L. c. morenoi. All records from L. c. morenoi are from the same geographical area in Andalusia, localized near the cities of Seville and Cordoba. Therefore, we are elevating L. c. morenoi species status and assigning the Jerez de la Frontera records of L. crassipes to L. morenoi stat. n. Distribution. In Spain (Fig. 17) known from two localities in the Sierra de Gredos, in central Spain (Matic 1968, Serra 1982, and one uncertain record from the Sierra de Grazalema, Andalusia (Voigtländer and Reip 2013). There is also one doubtful record from Malta (Kime 2003).

Key to the species of Lithobius (Monotarsobius) of the Iberian Peninsula
Key is valid for adult specimens only.
-Body 8.5-12 mm long (extreme values 7.9 and 13.2). Usually, DaP starts from LP7 to LP11 until LP15, very rarely from LP12 or LP13. Male dorsal depression of LP15 tibia longer, starts in the first third of LP15 tibia (Fig. 15B)