Altitudinal distribution and habitat use of the common wall lizard Podarcis muralis (Linnaeus, 1768) and the Horvath’s rock lizard Iberolacerta horvathi (Méhely, 1904) in the Kočevsko region

The study reports on the distribution and habitat use of two lizard species in the Kočevsko region: Horvath’s rock lizard and common wall lizard. Extensive sampling across an altitudinal span of 200 to 1,100 m a.s.l. in the study area revealed 62 localities with populations of both or either species. At 11 of these localities (18%) species occurred in syntopy, at 42 locations (68%) only common wall lizards were found, while at 9 locations (14%) only Horvath’s rock lizards were recorded. Both species occurred across the entire altitudinal span but exhibited an opposite pattern of relative abundances and frequencies, which increased with increasing altitude in Horvath’s rock lizard and with decreasing altitude in common wall lizard. The habitat use of common wall lizard was more general (it was found in seven habitat types) than Horvath’s rock lizard that was registered only in three habitat types with rocks.

The common wall lizard has the largest distributional range of all species of the genus Podarcis Wagler, 1830 (Gasc et al. 1997, Sillero et al. 2014. Previous studies revealed that this species originated from multiple glacial refugia (Gassert et al. 2013, Salvi et al. 2013, and multiple lineages were identified within three Mediterranean peninsulas (Iberian, Apennine and Balkan; Salvi et al. 2013). Its widespread distribution expands across most of Central Europe, the northern part of Iberian Peninsula, large parts of the Apennine and the Balkan Peninsulas and stretches to the east into North Turkey (Gasc et al. 1997, Sillero et al. 2014). The northernmost native distribution is probably still unresolved because results from a recent genetic study suggested that the population on Jersey (Channel Islands, UK) and in the Chausey archipelago may be of native origin (Michaelides et al. 2015), while in the past it has been believed that the species distribution does not extend beyond the Netherlands and that common wall lizards found in UK were introduced (Arnold 1995). In Slovenia, it is relatively common and widespread (Tome 1996, Mršić 1997, Tome 2001, Krofel et al. 2009).
The species tandem studied here is not unique, since other Iberolacerta-Podarcis species pairs with completely or partly attitudinally segregated distributional patterns have been observed also in the Iberian Peninsula, where other species of Iberolacerta occur (e.g. Moreira et al. 1999, Arribas et al. 2006, Monasterio et al. 2010. In several parts of the species range, I. horvathi populations tend to be denser at higher altitudes (e.g. De Luca 1989), while density in P. muralis follows an opposite trend (e.g. Krofel et al. 2009Krofel et al. , Žagar et al. 2013. In general, both species are found on rocky substrates with sparse vegetation (Arnold 1987;Arnold & Ovenden, 2004;Arnold et al. 2007;Cabela et al. 2007;Žagar et al. 2013), except that Horvath's rock lizards are more associated with rocks, while common wall lizards occur in a wider variety of different habitats (Arnold & Ovenden 2004).
In this study, an extensive sampling across an altitudinal span of 200 to 1,100 m a.s.l. was conducted in the Kočevsko region in order to comprehensively recognise syntopic and allotopic occurrence, altitudinal distribution and habitat use of Horvath's rock lizard and common wall lizard.

Materials and methods
The study was limited to the Kočevsko region, where we collected data on the presence and relative abundances (using transect line counts) of the study species in the period between 2006 and 2015 ( Fig. 1, Annex 1). Part of the data collected in the 2006-2008 period was obtained within the framework of a diploma thesis (Žagar 2008a) and was published in a study of habitat use of reptile community in the Kočevsko region (Žagar et al. 2013). Specifically, in that work we included information on the altitude, exposition, vegetation cover and habitat type of 10 reptile community members, from which we included for I. horvathi and P. muralis finds from 33 localities that are also included in this analysis. The data of the 2009-2015 period was collected within the framework of a PhD thesis (Žagar 2016). Species recognition was done by either coming very close to the lizard or photographing it, to inspect the position of scales on the head or colouration of the throat region. The species can readily be identified upon either of these characteristics (Tome 1999, Arnold & Ovenden 2004). We did not distinguish sex or age of individuals in this data set (Annex 1).
Locations were described as allotopic, when all visits of that location confirmed the presence of only one species, and syntopic, when both species were found at least once during the same visit. Transect line counts (Buckland et al. 1993) were conducted in one or up to three replicates (Annex 1). We summed all observations per transect and corrected for the number of times that we walked that transect (divided by number of replicates) to calculate the frequency of individuals recorded on each transect. We grouped transects into five altitudinal belts, each encompassing 200 m of altitude (Tab. 1). Thereupon, we determined the relative abundances for each altitudinal belt by summing up frequencies of individuals and dividing it by the summed length of transects inside each altitudinal belt.  We compared the observed frequencies of lizards (corrected for replicated transect visits, Tab. 1) in five altitudinal belts with expected frequencies (if species were equally distributed across the altitudinal span in the study area corrected for the total distance of surveyed transects in each altitudinal belt) using the Chi square test.
For assessing habitat use, one of the seven different habitat types were assigned to each transect occupied by study species: (i) natural rock area, (ii) urban area, (iii) agricultural land, (iv) water bank, (v) road, (vi) artificial rock area, and (vii) open forest (Annex 1). Habitat types describe the typical areas where transects were located in the study area. Natural rock areas were naturally occurring rock cliffs and screes, urban area included backyards, cemeteries and house ruins, agricultural land included grasslands, pastures and crop fields, water banks were banks of rivers, streams or lakes, roads were gravel or asphalt roads, artificial rock areas comprised of any rocky ground or walls originating from human activities, and open forest were located in forests with <85% crown coverage (Žagar et al. 2013). We calculated the relative proportion of allotopic and syntopic populations in each habitat type to present it graphically (Fig. 2).

Syntopic and allotopic occurrence
Results represent a dataset of 62 localities, at which one or both study species, Horvath's rock lizard and common wall lizard, were found in the Kočevsko region within the 2008-2015 period (Fig. 1, Annex 1). Both species were found to occur in syntopy at 10 localities (16%), common wall lizard was allotopic at 43 locations (69.5%) and Horvath's rock lizard at 9 locations (14.5%) (Fig. 1). Syntopic populations were found across the whole altitudinal span but with the majority of them located at middle altitudes (average altitude of syntopic populations (N = 10) was 620 m a.s.l., lower quartile range = 512 m a.s.l., upper quartile range = 813 m a.s.l., Annex 1). The lowest syntopic population was found at the entrance to Bilpa cave at 200 m a.s.l. and the highest at Kameni zid at 1,061 m a.s.l. (Annex 1).

Altitudinal distribution
The highest relative abundances of Horvath's rock lizard were determined for the highest altitudinal belt (900-1099 m a.s.l.) and relative abundances decreased with decreasing altitude (Tab. 1, Fig. 2). The opposite pattern was observed for the common wall lizard; relative abundance was highest at the two lowest altitudinal belts (100-299 and 300-499 m a.s.l.) and decreased with increasing altitude (Tab. 1, Fig. 2). Results of the Chi square test to compare observed frequencies of lizards (Tab. 1) in five altitudinal belts with expected frequencies (if species were equally distributed across the altitudinal span, see also Methods) showed significant differences between expected and observed frequencies for both species (for Horvath's rock lizard: χ 2 = 145.27, df = 4, P < 0.0001); for common wall lizard: χ 2 = 11.31, df = 4, P = 0.0233). Results of comparing altitudes from all finds of both species also showed a statistically significant difference between the species (Horvath's rock lizard, N = 71, median = 948 m a.s.l., and common wall lizard, N = 205, median = 430 m a.s.l., Mann-Whitney U tests: U = 1425, Z = 10.10, P < 0.0001).

Habitat use
The study species were found in seven different habitat types; Horvath's rock lizard in three and common wall lizard in seven of them (Fig. 3, Annex 1). Allotopic populations of common wall lizard were found in all seven habitat types, syntopic populations in all three habitat types where Horvath's rock lizard was found: in natural and artificial rocky habitats and in open forests (Fig. 2)  To check whether the changes in relative abundance with altitude can be explained by observed differences in habitat use between the species due to changes in habitat availability across the altitude, we decided to repeat the comparison of altitudes between the species by including findings in only three habitat types that occurred throughout the altitudinal range and were used by both species (artificial and natural rock areas and open forest). Results showed that in these habitat types, too, Horvath's rock lizard was found at significantly higher altitudes (N = 71, median = 948 m a.s.l.) than common wall lizard (N = 113, median = 507 m a.s.l..; Mann-Whitney U tests: U = 815, Z = 9.09, P < 0.0001).

Discussion
In conclusion, we have found that in the Kočevsko region, Horvath's rock lizard and the common wall lizard -two lizard species, which exhibit a high resemblance in overall body plan and many ecological characteristics -occurred across the entire altitudinal span but exhibited an opposite pattern of relative abundances and frequencies, which increased with increasing altitude in Horvath's rock lizard and with decreasing altitude in common wall lizard. The observed pattern of habitat use suggests that the common wall lizard occupies here a more diverse array of habitat types than Horvath's rock lizard.
Jointly, the opposite pattern in relative abundances across the altitudinal span and wider use of habitat types of the common wall lizard compared to Horvath's rock lizard suggest that the species segregate to some extent in their spatial distribution and spatial niches in the Kočevsko region. However, compared to other studies of distribution of these two species (see introduction), our results showed an interestingly high altitudinal overlap in the distribution of the two species and relatively high proportion of syntopic populations (16%), as well as an overlap in three habitat types. Previous studies reported that syntopic populations of studied species occurred only in a limited zone of middle altitudes, while Horvath's rock lizard was found in allotopic populations at higher altitudes and the common wall lizard in allotopic populations at lower altitudes (De Luca 1989, Lapini et al. 2004, Cabela et al. 2007, Rassati 2010. So far, this is the first observation of syntopic populations found across the entire altitudinal span of an area for these two species. This may be due to the specific topography of the Kočevsko region where altitudes do not exceed 1,100 m a.s.l. (Perko & Orožen Adamič 1998), whereas other study areas had higher altitude ranges (over 2,000 m a.s.l. in the Alpine region or up to 1,757 m a.s.l. at Velebit). Horvath's rock lizard was found there in places up to the highest peaks in Velebit ( De Luca 1989) or up to 2,000 m a.s.l. in the Alps (De Luca 1989, Lapini et al. 2004, Cabela et al. 2007, Rassati 2010.
The found between-species differences in altitudinal distribution, not only in the Kočevsko region but elsewhere, reinforce that Horvath's rock lizard is a high-altitude species that can also occur in lowlands but on rarer occasions, while the common wall lizards' populations are most abundant in lowlands and become less dense at higher altitudes. Recent research revealed that both species also exhibit differences in physiological characteristics and that Horvath's rock lizard has adaptations that are potentially advantageous in high-altitude areas that are climatically thermally more restrict (lower yearly average air temperatures and shorter activity periods for lizards) compared to lowlands. For example, the study species differ in seasonal variation of their preferred body temperatures in terms that Horvath's rock lizard exhibits a more accurate thermoregulation across the seasons than the common wall lizard (Osojnik et al. 2013). The differences between the species were also observed on the cellular level where Horvath's rock lizard had higher potential metabolic activity than the common wall lizard, which may be advantageous in thermally restrictive environment together with more precise thermoregulatory behaviour as exhibited by Horvath's rock lizard (Žagar et al. 2015