Millipede and centipede assemblages on the northern and southern slopes of the lowland Altais, southwestern Siberia, Russia (Diplopoda, Chilopoda)

Abstract The total species richness in the myriapod assemblages of the lowland Altais near Charyshskoe Village, Altai Province, southwestern Siberia, Russia is estimated to be at least 19 species from ten genera, eight families, five orders, and two classes. The following species are new to SW Siberia: Lithobius (Ezembius) ostiacorum Stuxberg, 1876, L. vagabundus Stuxberg, 1876, and L. (Monotarsobius) nordenskioeldii Stuxberg, 1876, while L. (E.) proximus Sseliwanoff, 1880 and L. (M.) insolens Dányi & Tuf, 2012 are recorded for the first time from the Altai Province of Russia. A species of Strigamia which is morphologically similar to Strigamia cf. transsilvanica (Verhoeff, 1928) has been found in the study area but its true specific identity is yet to be determined. The seasonal dynamics of myriapod assemblages in terms of the species diversity, density, sex-age structure, and vertical distribution along the soil profile have been studied with regard to the different slope exposures.


Introduction
Despite the recent increased interest in the myriapod fauna of southwestern Siberia (Mikhaljova et al. 2007, 2014, 2015, Mikhaljova 2009, 2017, 2014a, b, c, 2016a, b, c, 2017a, b, Nefedieva et al. 2014, 2015, Nefediev 2016, the biodiversity and ecological characteristics of myriapods in the study area of the lowland Altais, a transition zone from the plains of the southwestern Siberia to the mountains of southern Siberia have not been studied to date.
The material was collected using the standard soil fauna sampling techniques practiced in Russia (Ghilarov 1987) by taking 5 soil samples per study site, hand-sorting each 10 cm layer down to 30 cm until fauna penetration, with the sample area totaling ¼ m 2 . Soil samples were taken three times during summer 2016, starting at the beginning of summer (31 May-2 June), through mid-summer (12-13 July) to late summer (22-23 August). Also we collected additional faunistic material in nearby localities by hand sampling in the summers of 2015-2017. The total number of studied millipedes and centipedes was 684 and 666 specimens, respectively.
The distribution of recorded species in soil samples was analyzed using CANOCO for Windows 4.5 (ter Braak and Šmilauer 1998). Following lengths of gradient in species data we selected Redundancy analyses (RDA) using environmental variables, i.e. exposure (south/north), month, depth of soil sample and sample ID. The significance of models was evaluated using Monte Carlo tests with 499 permutations. For the  evaluation of significance and effect of tested environmental variables forward selection was applied. The effect of selected significant environmental variables (month, depth) for predicting the distribution of individual species was tested using Generalized linear models (GLM) with evaluation of AIC.
The material treated here was collected by A.M. Alenov (A.A.), E.V. Andreeva Remarks. The julid L. tigirek has been collected outside its terra typica for the first time. The above records on the northern slope show the species preference for more humid habitats.

Remarks.
The above records of the julid S. latisupremus are the southwesternmost for the species. In the study localities, the species demonstrates no preference between investigated habitats as regards different slope exposures.
Remarks. The results of this study suggest that S. clavatipes prefers the southern slope, in spite of its highly ecological valence.  Distribution. Southern Siberian boreal range with isolated Yenisei population: this species has previously been recorded in the Yenisei River area, Krasnoyarsk Province and Irkutsk Area (central and eastern Siberia, respectively) (Zalesskaja 1978), also recently found in northern Mongolia (Poloczek et al. 2016), Altai Province (Nefediev et al. 2017c) and Republic of Altai (Nefediev et al. 2017d).

Class
Remarks. The above record of L. ostiacorum, recently announced at the 17th International Congress of Myriapodology (Nefediev et al. 2017c), can be considered as the first formal find of the species in SW Siberia. In the study localities, the species was found more frequently on N facing habitats.

Lithobius (Ezembius) proximus Sseliwanoff, 1880
Lithobius proximus -Zalesskaja 1978: 125-126 Remarks. The above find of the species, recently announced at the 17th International Congress of Myriapodology (Nefediev et al. 2017c), can be considered as the first formal record of it in the Altai Province, SW Siberia. In the investigated area, L. proximus is very rare and shows no significant differences in its distribution between slopes.
Remarks. In the study localities, L. sibiricus shows a higher abundance on the northern slope.  Distribution. Trans-Palaearctic: the species displays extremely wide distribution in Europe, Asian Russia, the Near East and the Arabian Peninsula, also in northern Mongolia. In Siberia L. curtipes has been reported from the Novosibirsk, Omsk, Tyumen and Tomsk areas, the Altai and Krasnoyarsk provinces and the Republic of Altai (Nefediev et al. 2016d(Nefediev et al. , 2017b.

Lithobius
Remarks. Despite a wide geographical range, and its high ecological valence, in the study area, the species inhabits mainly the northern slope.
Remarks. The above record of L. insolens, recently announced at the 17th International Congress of Myriapodology (Nefediev et al. 2017c), can be considered as the first formal record of the species in the Altai Province, SW Siberia. In the study area, the species has significant preference for the southern slope. A single ♂ with aberrant numbers of antennomeres (22+24 vs. 20+20 in original description) was found.
Material examined. 1 juv. (ASU), Russia, southwestern Siberia, Altai Province, Charysh District, ca. 4.5 air-km SE of Charyshskoye Village, site 1 on N slope, soil sample 3 (0-10 cm deep), 13.07.2016, leg. Kh.N., S.N., V.S. Distribution and remarks. Until recently this species was been known only from its terra typica in the Krasnoyarsk Province, central Siberia, Russia. New records of L. nordenskioeldii in the Altai Province, as announced at the 17th International Congress of Myriapodology (Nefediev et al. 2017c), and in the Republic of Altai (Nefediev et al. 2017d) seems to indicate the low level of species abundance in the Altai region.
Remarks. The species identity of this new record is delayed pending an examination of additional material of specimens with similar diagnostic characters from the Republic of Altai.
Distribution. Originally described from the Yenisei River basin, Krasnoyarsk Province, central Siberia (Zalesskaja 1978), the species has been found recently in the Altai Province and Republic of Altai (Nefediev et al. 2017c, d), both SW Siberia, Russia.
Remarks. The above finding of L. vagabundus, recently announced at the 17th International Congress Myriapodology (Nefediev et al. 2017c), can be considered as the first formal record of the species in southwestern Siberia. In the study region, the species was very rare in all biotopes. Remarks. The identification of the above recorded specimens to the species level is impossible due to their early instars or lack of legs.
Remarks. Apparently a very euryoecious species, A. macrocephalus has currently been recorded mainly from habitats on the southern slope. Distribution. Central-Palearctic temperate range: widespread from Central Europe and the Caucasus, S. pusilla is found in the Urals, SW and central Siberia and N Mongolia (Bonato et al. 2012;Poloczek et al. 2016;Nefediev et al. 2017c, d).
Remarks. In the study area, the species was found rarely and on the northern slope only.
Remarks. Although both specimens resemble S. transsilvanica, the study area is far from the known distribution of the species. Aside from the possibility of human introduction of this species in the Charysh District, the presence of a possible undescribed species similar in morphology to S. transsilvanica could be tested by molecular methods in the future.
Remarks. In the study region, E. koreanus appears to be found mainly on the northern slope.
Remarks. In the study area, E. retusidens inhabits both slopes, and is one of the most dominant species.

Results and discussion
The myriapod fauna of the study area comprises at least 19 species from 10 genera, 8 families, 5 orders and two classes (Diplopoda and Chilopoda).
The species richness in the millipede assemblages was found to be very low and similar on both slopes (I J = 0.83). Thus, 5 diplopod species are known to occur on both slopes (Megaphyllum sjaelandicum, Sibiriulus latisupremus, Orinisobates sibiricus, Schizoturanius clavatipes and Altajosoma sp.), whereas Leptoiulus tigirek inhabits the northern slope only (Table 1).
The total species richness in the centipede assemblages is twice as high compared to the millipede one, with 10 and 12 species recorded on the southern and northern slopes, respectively. Most Chilopoda species are common to both slopes, namely, Lithobius  Table 1).
The julid L. tigirek, which has recently been included in the Red Data Book of the Altai Province (Nefediev 2016), has been collected outside its terra typica for the first time, thus also expanding the eastern range limit of the species (Figure 6). The julid S. latisupremus has previously been known from the Smolenskoe and Altaiskoe districts in the Altai Province and from the Shebalino District in the Republic of Altai (Mikhaljova et al. 2014). The current record of the species is the westernmost known to date ( Figure  7). The species identity of Altajosoma sp. is delayed pending a revision of the variation in Altajosoma bakurovi bakurovi (Shear, 1990), which the currently recorded diplomaragnid is close to in the shape of colpocoxites of posterior gonopods and in their distal parts, but differs significantly in the large posterior angiocoxal processes.  The species diversity of Diplopoda is very low on both slopes. The julid M. sjaelandicum predominates on the dry southern slope, ranging from 44 to 60 % of the total millipede abundance, whereas S. latisupremus tends to dominate on the more humid northern slope, ranging from 44 to 73 % of the total diplopod abundance (Figure 8). The latter species may also be considered as a codominant species on the southern slope (23-36 % of the total millipede abundance), while the rest of the millipede species are rare or very rare on the southern slope. Codominants of the northern slope appear to be M. sjaelandicum and O. sibiricus with 22 % of the diplopod abundance. The RDA model also reveals the pattern of millipede distribution (Figure 9) explaining 20.3 % of the variability in species data. Of the tested environmental variables, slope exposure (south/north) and time of sampling (month) are significant (F = 9.88, p = 0.002 and F = 3.42, p = 0.018, respectively). Of the recorded species, M. sjaelandicum and S. clavatipes predominate on the southern slope.
Species diversity of Chilopoda is low on the southern slope: two species predominate, in particular, L. (M.) insolens, ranging from 34 to 72 % of the total chilopod Figure 9. RDA ordination biplot of the distribution patterns of millipedes in soil samples on the study slopes. Environmental variables significantly contributing to the prediction are in bold. The whole model is statistically significant (F = 4.73, p = 0.002) and explains 20.3 % of variability of species data, the X-axis explains 16.5 %. 48 ± 10.8 ind./m² on the southern slope, and from 9 ± 1.2 to 22 ± 13.6 ind./m² on the northern one, gradually declining from June to August in both habitat types ( Figure 12). Of the recorded species, abundance of the only julid, S. latisupremus, are significantly affected by the time of sampling as the population decreases from June to August (GLM: F = 6.92, p = 0.010). The numbers of centipedes on the northern slope are twice as high compared to the southern one. The seasonal dynamics of Chilopoda density ranges from 20 ± 6.8 to 27 ± 19.6 ind./m² on the southern slope, and from 31 ± 0.0 to 47 ± 11.6 ind./m² on the northern one, the highest being in June and August and the lowest in July in both habitat types ( Figure 13).
The age structure will be considered here, using the dominant species as an example. Thus, in the age structure of the julid M. sjaelandicum population on the southern slope, juveniles predominated during the summer, and their abundance varied from 100 % of the population in June to 70 % in July and August. In contrast, in the julid S. latisupremus, overwintering adults predominated at the beginning of summer (with 75 % of the population), producing juveniles, which started to prevail in the middle of summer (with 76 % of the population).
The age structure in the population of the lithobiid L. (M.) insolens is as follows: adults predominate at the beginning of summer on both slopes, ranging from 70 to 100 % of the population, while young individuals emerge in the middle of summer in amounts equal to the total numbers of males and females, and this ratio is maintained until the late summer. The sex ratio is close to 50:50 during summer on both slopes, but on the southern slope only females exceed males twice over by the end of summer. In the -age structure of E. retusidens on the southern slope, the abundance of juveniles Figure 11. RDA ordination biplot of the distribution patterns of centipedes in soil samples on the study slopes. Environmental variables significantly contributing to the prediction are in bold. The whole model is statistically significant (F = 4.12, p = 0.002) and explains 15.2 % of variability of species data, the X-axis explains 10.3 %. is 3 times higher than in adults. On the northern slope, the ratio of adults and juveniles is equal at the beginning of summer, while in the middle and late summer adults start to prevail to become twice as abundant. For adults, the females steady prevailed, outnumbering males from 2 to 5 times throughout the season in both habitats.
Regarding the vertical distribution in the soil profile, more than 80 % of millipedes prefer the upper soil layer to a depth of 10 cm on both slopes. Diplopods are very rare in the litter, especially on the dry southern slope (where they numbered less than 1 %),  but the numbers are about 15 % more on the humid northern slope, with maximum penetration in depth to no more than 20 cm ( Figure 14). With regard to the vertical distribution in the soil profile in centipedes, we observe the preference of chilopods to the upper soil layer. Thus, approximately 80 % of centipedes of the total chilopod abundance has been reported from the top 10 cm layer on both study slopes, with the maximum penetration in depth to no more than 30 cm. Centipedes are very rare in the litter, accounting for about 1 % on the dry southern slope and about 13 % on the more humid northern one ( Figure 14). As the depth of the sample is a significant variable for RDA model, we tested its power to predict the distribution of individual species. Abundances of the geophilomorph E. retusidens and the lithobiomorph L. (M.) curtipes are the only species significantly affected by depth of sample. The geophilomorph prefers deeper soil layers and the lithobiomorph prioritizes the surface and upper soil layers (GLM: F = 6.41, p = 0.013 and F = 4.01, p = 0.048, respectively). This is not surprising, as the preference for the upper layers of soil by L. (M.) curtipes is well known (Tuf 2002(Tuf , 2015. The ability of geophilomorphs to penetrate to deeper soil layers is documented and also recorded, using subterranean pitfall traps, too (Tuf et al. 2017).

Conclusions
1. The species richness of millipedes is found to be very low in both habitat types studied, on the northern and southern slopes, whereas the centipede species richness is assessed as twice as high. The total richness comprises at least 19 species, belonging to ten genera, eight families, five orders, and two classes.