Acta Herpetologica 15(1): 3-14, 2020 ISSN 1827-9635 (print) © Firenze University Press ISSN 1827-9643 (online) www.fupress.com/ah DOI: 10.13128/a_h-8448 Has the West been won? A field survey and a species distribution model of Iberolacerta horvathi in the Alps Giuseppe De Marchi1,*, Giovanni Bombieri1, Bruno Boz1, Fausto Leandri2, Jacopo Richard3 1 Associazione Faunisti Veneti, Venezia (VE), Italy 2 World Biodiversity Association onlus, c/o Museo Civico di Storia Naturale, Verona (VR), Italy 3 Veneto Agricoltura, Agenzia Veneta per l’Innovazione nel Settore Primario, Legnaro (PD), Italy * Corresponding author. Email: dromasardeola@gmail.com Submitted on: 2019, 2th June; revised on: 2019, 8th September; accepted on: 2019, 2nd November Editor: Dario Ottonello Abstract. The Horvath’s rock lizard (Iberolacerta horvathi) is a rupicolous mountain species endemic of the eastern Alps and northern Dinaric range. The species has its known western limit of the distribution in the Veneto region of Italy. It is not known whether the species is really rare in Veneto or whether the area has been insufficiently sur- veyed. In addition, it is not known whether the westward distribution of the species is limited by a physiographic or by a climatic barrier. During the period 2016-2018, 118 sites were surveyed in the Veneto and Trentino-Alto Adige regions. Four new occurrences of Iberolacerta horvathi were discovered in Veneto that: 1) largely fill the gap between the westernmost known site and the closest site to the east; 2) extend further west the known distribution by 9 km. In addition the species was confirmed in three already known sites. A species distribution model was developed with the software MaxEnt, using 100 occurrences from Italy, Austria and Slovenia. The best model shows that the distri- bution is explained by the asperity of their habitat, the sedimentary bedrock, the aspect, the average temperature of the coldest quarter, the rainfall seasonality and the average summer rainfall. The last variable appears as the most likely responsible for the rarefaction of the species at its western limit. In addition, the species distribution model suggest that the Horvath’s rock lizard might be present in some additional mountain groups where it has so far not been found yet. Keywords. Horvath’s rock lizard, Maxent, rainfall, temperature, roughness, aspect, Bedrock, Veneto. INTRODUCTION The present distribution of reptiles in Europe is much dependent on its climatic history, in particular to the Pleistocene oscillations. After the climate improvement following the last glacial maximum (18,000 BP) many species spread from southern refugia and have colo- nized very large areas (Joger et al., 2007). Some species were able to spread from their refugia in a more limited way due to ecological constraints or to the presence of geographical barriers like the Alps for species from the Apennines peninsula and the Pyrenees for species from the Iberian peninsula (Joger et al., 2007). The opposite happened to some cold adapted species, like the Dinolac- erta mountain lizards of the southern Dinaric chain and the species of Iberolacerta of the Iberian Peninsula: they have likely reduced their range and have become restrict- ed to high altitude isolated mountain peaks (Ortega et al., 2016). The conservation prospect of these high altitude species is likely going to worsen due to the ongoing cli- mate change (Le Galliard et al., 2012). A different case is the distribution of the Horvath’s Rock Lizard (Iberolacerta horvathi), an endemic species of the northern Dinaric chain and eastern Alps (north- 4 Giuseppe De Marchi et alii west Croatia, Slovenia and adjoining northeastern Italy and southern Austria) (Speybroeck et al., 2016), a spe- cies phylogenetically related to a group of mountain dwelling lizards of the Iberian Peninsula (Carranza et al., 2004, Crochet et al., 2004). As a matter of facts, the known presence localities in Italy and Austria are in areas that were glaciated during the last glacial maximum (Ivy- Ochs et al., 2009). So, the species should have colonized the eastern Alps from refugia in the southeastern Alps or from the Dinaric chain. Unfortunately, it is difficult to study how this colonization happened due to insuf- ficient knowledge of the distribution of this species, as suggested by the recurrent discoveries of its presence in new mountain groups (e.g., Žagar et al., 2014). This lack of data is likely the result of the difficult accessibility of the rocky cliffs inhabited by the species and by the pos- sible misidentification with the similarly looking (Žagar et al., 2012) and frequently syntopic common wall lizard, Podarcis muralis (Cabela et al., 2007). In Italy, Horvath’s rock lizard is well distributed in the Friuli region (Lapini et al., 2004; Rassati, 2010; Rassati, 2012) but its presence appears to be particularly scattered west of the Piave Riv- er in the Veneto region. The westernmost site was discov- ered in 1993 close to the village of Listolade (Lapini and Dal Farra, 1994; Lapini et al., 2004), a site that was about 70 km away from the nearest known site. In spite of sub- sequent discoveries (Rassati, 2010; Rassati, 2012; Lapini, 2016) and quite extensive surveys (Tormen et al., 1998; Bonato et al., 2007; Bonato, 2011; Cassol et al., 2017) there is still a gap of around 30 km between the site of Listolade and the closest one to the east. Therefore, it is not known whether the species is present in this large gap and even west of Listolade. Moreover, it would be interesting to discover whether the species has reached its potential distribution or whether its westward post glacial expansion is still lagging behind its potential distribution, as it still happens for some plants (Svenning and Skov, 2007; Willner et al., 2009; Dullinger et al., 2012) and ani- mals (Araújo et al 2008; Pinkert et al., 2018). In addition it is not known which factors might limit or have limited the westward spread of the species. In order to fill these gaps in the knowledge of the species, the research has the following aims: 1) to improve the known distribution by surveying mountain groups west of the Piave River in the Trentino- Alto Adige and Veneto regions of Italy; 2) to correlate the presence of the species in Ita- ly, Austria and Slovenia to climatic, geological, physi- ographic and ecological parameters in order to esti- mate, through a species distribution model, whether the western limit of the distribution has likely been fixed or whether new sightings further west should be expected. MATERIAL AND METHODS Field survey Surveys were conducted during 94 days in the period April 2016 - September 2018 in 118 sites by one to three researchers per site. 20% (n = 24) of the sites were visited more than once. Surveyed areas were concentrated west of the Piave River but a few sites were investigated also just east of it. Due to the inac- cessibility of many cliffs, investigated sites were not randomly distributed but were inevitably concentrated close to roads and mountain paths. Species identification was rarely performed by catching the specimens with a long stick with a cotton loop. More common- ly, species identification was performed by analysing pictures taken with a telephoto lens, an efficient technique also when lizards are on inaccessible rocky cliffs (De Marchi et al., 2019). Identification characters from the frequently syntopic common wall lizards are as in Corti et al. (2010). The coordinates and altitude of the sites were obtained by hand held GPS devices. In addition, the sites where Horvath’s rock lizards were found were characterized by the bedrock obtained from the geological map of Italy viewed on the Geo- portale Nazionale (http://www.pcn.minambiente.it) and by the habitat classification obtained from the “Carta della Natura” of Veneto and “Carta della Natura” of Friuli-Venezia Giulia (1:50.000) (http://www.geoviewer.isprambiente.it/). Species distribution model (SDM) In order to elaborate the potential distribution of the spe- cies in Italy and in neighbouring Austria and Slovenia, a pres- ence only modeling approach was chosen due to the lack of reliable absence points particularly outside our study area. A maximum entropy approach was developed with the software Maxent (Phillips et al., 2006; Merow et al., 2013; Phillips et al., 2017), which proved to be quite successful even when occur- rence data were few and in presence of somehow biased sam- ples (Elith et al., 2006). As occurrence data of Veneto are few and restricted to a too small area for statistical analyses, pub- lished data related to Italy (Lapini et al., 2004; Rassati, 2010; Rassati, 2012; Lapini, 2016; Lapini, 2017; Rassati, 2017), Austria (Tiedeman, 1992; Ortner, 2006; Cabela et al., 2002; Cabela et al., 2007) and Slovenia (Žagar, 2008a; Žagar, 2008b; Krofel, 2009; Cafuta, 2010; Žagar et al., 2011; Žagar et al., 2012; Osojnik et al., 2013; Žagar, 2016; Zauner, 2018) were collated. Such a wider area helps to reduce the risk that some local climatic conditions might wrongly highlight the importance of an environmen- tal variable that in fact is not a relevant one (Phillips, 2006). Unfortunately, some published occurrences lacked coordinates but have only the altitude and a general description of the col- lecting site: only when the description was considered suffi- ciently detailed to be confidently associated to a single 300 m × 300 m pixel (see later) the occurrences were used by extract- ing the coordinated from Google-Earth. As some occurrences are aggregated due to sampling bias, some occurrence localities were manually removed so that the remaining ones are at least 5Westward expansion of the Horvath’s rock lizard one km from the closest ones. Eventually, the 100 occurrences of the Horvath’s lizard were joined together with a background dataset of 10,000 points to randomly sample the analysed area (lat 45.35°N to 47.00°N, lon 11.20°E to 15.00°E). The environmental layers (all at European scale) were as follows. 19 bioclimatic variables at 30-arc seconds (about 1km) resolution derived from the monthly temperature and rainfall values for the period 1970-2000 (http://www.worldclim.org/ bioclim). Slope, Roughness (the degree of irregularity of the surface, calculated by the largest inter-cell difference of a cen- tral pixel and its surrounding cell, as suggested by Wilson et al., 2007) and Aspect (the compass direction that a slope faces with values ranging 0-360°) layers calculated with the GIS software QGIS (version 2.8.6) from a 25m digital elevation model (Layer: DEM-v1.1-E40N20) downloaded from the Copernicus Land Monitoring Service (https://land.copernicus.eu/). A 250m veg- etation layer, Corine Land Cover clc_12 (version 18_5), again downloaded from the Copernicus Land Monitoring Service. A geological layer from the 1 : 5 Million International Geological Map of Europe and Adjacent Areas (IGME5000, Asch, 2003) with bedrock types reclassified in the following five categories: crystalline (all igneous and metamorphic rocks), carbonates (limestone, dolomite or carbonates as predominant bedrocks), secondary carbonates (limestone, dolomite or carbonates locally present but not as predominant bedrocks), other sedimentary rocks (like sandstone, conglomerate et clay) and undifferentiat- ed bedrock. All environmental layers were resampled to a reso- lution of 300 m using the package “raster” in R (Hijmans and van Etten, 2016). While Maxent is partly able to balance between model fitting and model simplicity, an information approach was adopted to check whether simpler models with fewer variables performed better (Warren and Seifert, 2011; Zeng et al., 2016). First, a reduced set of environmental variables was pre-select- ed, based on ecological or biological knowledge. Out of the 19 available bioclimatic variables, BIO1 (Annual Mean Tempera- ture), BIO4 (Temperature Seasonality), BIO5 (Max Tempera- ture of Warmest Month), BIO6 (Min Temperature of Coldest Month), BIO10 (Mean Temperature of Warmest Quarter), BIO11 (Mean Temperature of Coldest Quarter) BIO12 (Annual Precipitation), BIO 15 (Precipitation seasonality), BIO18 (Pre- cipitation of Warmest Quarter) and BIO19 (Precipitation of Coldest Quarter) were selected. Then, a reiterative process of model formation and stepwise removal of the least contribut- ing variables was utilized (Zeng et al., 2016). Unfortunately, most temperature related variables (BIO1, BIO5, BIO6, BIO10, BIO11) were highly correlated (pearson coefficient >0.7). In a similar way most rainfall related variables (BIO12, BIO18, BIO19) were highly correlated. Therefore, Maxent models were run in order to select the single most predictive temperature variable and the single most predictive rainfall variable. This was done (Table 2) by running Maxent models (regularization multiplier was set at 1 and regularization values were linear- quadratic-product) alternatively containing only one of the cor- related variables for temperature (BIO1, BIO5, BIO6, BIO10, BIO11) and only one of the correlated variables for rainfall (BIO12, BIO18, BIO19). Among the alternative models the one with the lowest Akaike information criterion corrected for low sample size (AICc) calculatd with the software ENM- TOOLS version 1.4.4 (Warren and Seifert, 2011; Shcheglovito- va and Anderson, 2013) was selected. The AICc is an estima- tor of the relative quality of statistical models for a given set of data, trading-off between the simplicity of the model and the goodness of fit of the model (Warren and Seifert, 2011). Then, simpler models were constructed by stepwise removal of the variable with the lowest permutation importance (Zeng et al., 2016) and the AICc of the resulting models was calculated in order to select the final set of variables. As ‘‘feature class’’ and ‘‘regularization multiplier’’ affect the performance of the model (Morales et al., 2017) the model was fine tuned by constructing models with 7 levels of the regularization multiplier (0.05, 0.1, 0.5, 1, 2, 4, 10) and three regularization values (Linear-Quad- ratic, Linear-Quadratic-Product and Hinge). Finally, the best model (Regularization value = Linear-Quadratic; Regularization multiplier = 0.1) was run 100 times with bootstrap resampling and the average model was calculated. The quality of this final model was tested by calculating the area under the curve (AUC) of the receiver operated characteristic (ROC) plots in order to discriminate a species’ model from a random model. Finally, the percent contribution and permutation importance of the envi- ronmental variables to the final Maxent model were measured (Phillips, 2006). RESULTS The Horvath’s Lizard was found in four new sites (Valle di San Lucano, Val del Grisol, Casoni and Forra del Piave) and its presence was confirmed in three sites (Monte Tudaio, Val Zemola and Listolade) already known in literature (Fig. 1). Fig. 2 shows the habitat of the four new occurrences for Horvath’s Lizard and Table 1 reports some characteristics of all the seven sites. The common wall lizard was observed in 60% of the investigated sites (n = 71, Fig. 1) and was frequently syn- topic with the Horvath’s rock Lizard, where this last spe- cies occurred (Table 1). Species distribution model. The Mean Temperature of Coldest Quarter (BIO11) turned out to be a better predictor of the distribu- tion of the Horvath’s rock lizard (Table 2) compared to other highly correlated temperature variables: Annual Mean Temperature (BIO1), Max. Temperature of Warm- est Month (BIO5), Min. Temperature of Coldest Month (BIO6) and Mean Temperature of Warmest Quarter (BIO10). The Precipitation of Warmest Quarter (BIO 18) turned out to be a better predictor of the distribution (Table 2) compared to other highly correlated precipita- tion variables: Annual Precipitation (BIO12) and Precipi- tation of Coldest Quarter (BIO19). 6 Giuseppe De Marchi et alii The balancing between the simplicity of the model and goodness of fit to the known distribution resulted in a model that did not include Slope, Temperature Sea- sonality (BIO4) and Vegetation type (Corine Land Cov- er), the variables with the lowest permutation impor- tance. As a result, the best model was explained only by Mean Temperature of Coldest Quarter (BIO11), Rainfall Seasonality (BIO15), Precipitation of Warmest Quarter (BIO18), Aspect, Roughness and Bedrocks (Table 2). The best distribution model obtained after fine tun- ing (see methods and table 2) is shown in Fig. 3 (with cloglog output, which is the easiest to conceptualize as it gives an estimate between 0 and 1 of probability of pres- ence). The model fits the distribution (average training AUC for the 100 bootstrap replicate runs is 0.922, stand- ard deviation is 0.009) much better than a random model (AUC = 0.5). The variables that explain the distribution of the species are in order of percent contribution (Table 3): Roughness (33%), Bedrocks (26.5%), Mean Tempera- ture of Coldest Quarter (19.5), Aspect (8.9), Rainfall of the Warmest Quarter (8.5%) and Rainfall seasonality (3.3). Rainfall of the Warmest quarter decreases westward in the Italian Alps (Fig. 4): the 13 westernmost occur- rences are in areas with a rainfall (Average±SD = 338 ± 37mm, n = 13) that is significantly lower (Welch Test = -9.67, P < 0.001) than the remaining occurrences (Aver- age ± SD = 452 ± 53mm, n = 87). DISCUSSION Our field survey has discovered four new sites of Horvath’s rock lizard in Veneto (Table 1 and Fig. 1 and 2): Forra del Piave is close to already known sites; Casoni (in the Bosconero mountain group) and Val Grisol (in the Schiara mountain group) fill the gap between the known westernmost site of Horvath’s lizard and the pre- viously known sites to the east; San Lucano Valley (in Fig. 1. Sites surveyed in 2016-2018. Red dots show the newly discovered occurrences of Iberolacerta horvathi, while orange dots show some confirmed occurrences. 7Westward expansion of the Horvath’s rock lizard Pale di San Martino mountains) extends further west the known distribution by 9 km. Our extensive research in apparently suitable sites west of San Lucano Valley, in particular in the Vette Feltrine mountains and in the Valsugana Valley (Fig. 1 and Fig. 4), has been unable to discover additional sites. This lack of positive results sug- gests that the western limit of the distribution has been discovered. Interestingly, the San Lucano Valley is also the western limit in the southern Dolomite mountains of another reptile of eastern origin, the Nose-horned viper (Vipera ammodytes) (Sindaco et al., 2006), a species that has been found to have a close habitat similarity with Horvath’s rock lizard also in another part of its distribu- tion (Žagar et al., 2013). There is apparently no physio- graphic barrier to the spread of the lizards further to the west. As a matter of fact, a gravel bed mountain river like the Cordevole river (Fig. 4) should have not been a bar- rier to the westward spread of the species as proved by our discovery of the species just west of this river in the San Lucano Valley. In the past, gravel bed mountain riv- ers such as the Piave and the Cordevole (Fig. 4) should have been an even less effective barrier to the spread of the species during the phase of great sedimentation in the valleys that followed the deglaciation (Cordier et al., 2017) and continued up to 7000 BC in the Piave Val- ley (Carton et al., 2009): in this condition, water mostly flowed deep inside the gravel bed (Arscott et al., 2001) and could not stop the spread of lizards. Therefore, the lack of any apparent physiographic barrier lets us sug- gest that an environmental gradient should make the spe- cies less abundant in Veneto, where it is definitely rare, until it disappears. In general, an environmental variable might affect directly the species physiology (e.g., tem- perature) or might act in an indirect way by influencing Fig. 2. Pictures showing the habitats of the newly discovered sites of Horvath’s rock lizard. 1. Forra del Piave, Cadore; 2. Casoni, Canal del Maè; 3. Val del Grisol; 4. Valle di San Lucano, Agordino. In the last site, the first single individual was observed on the rock wall along the road; later, the presence of a more numerous population was documented in the neighbouring rocky cliffs (around 200 m away). 8 Giuseppe De Marchi et alii the availability of food (e.g., rainfall) or the presence of parasites, predators or competitors (Austin, 2002). An important factor that might limit the abundance and the distribution of the Horvath’s rock lizard may be the strong competition with the common wall lizard (Podar- cis muralis) (Carranza et al., 2004). The two species have a similar overall body plan (Žagar et al., 2012) and simi- lar ecological preferences, for example they eat similar food (Richard and Lapini, 1993). However, Horvath’s rock lizards appear more adapted to rocky habitats due to their flat head and body (Corti et al., 2010), which might favour hiding in small crevices. They are more adapted to relatively cold habitats due to a higher poten- tial metabolic activity (Žagar et al., 2015b), even if they do not exhibit the partial development of the egg inside the female body that is found in some other small moun- tain lizards (Ljubisavljević et al., 2012). On the contrary, common wall lizards are favoured in agonistic interac- tions with Horvath’s rock lizards because of their larger head and consequent stronger biting force (Žagar et al., 2017), for example when competing for thermoregulation spots (Žagar et al., 2015a). Common wall lizards may be favoured as well for their higher reproductive poten- tial (Amat, 2008). As a result, Horvath’s rock lizards are confined to rocky habitats in low altitude shady gorges, at mid altitude in beech forest and to higher altitudes com- pared to the common wall lizards (Žagar, 2016). Vegeta- tion type has not turned out to be important in our anal- ysis for predicting the distribution of the Horvath’s rock lizards likely because the Corine Land Cover classifica- tion that is currently available for our study area has been developed only to the “third level”. This means, for exam- ple, that forests are classified only to the level of “conif- erous forests” or “mixed forests”, a level that is likely not enough for a proper classification. Apart from the land cover classification, our analysis of the importance of the variables that constrain the dis- tribution of the Horvath’s lizard (Table 3) largely agrees with the previous results (Cabela et al., 2007; Žagar, 2016). Table 1. Characteristics of the new and confirmed occurrences of Horvath’s lizard surveyed in 2016-2018. The number associated to each site is the same as in Fig.1 and Fig. 2. Site * new + confirmed Date (day, month, year) Coord. (°N, °E) Altitude (m) Aspect Bedrock (from Geoportale nazionale italiano) Vegetation (CNAT of Veneto and Friuli-Venezia Giulia) Observed specimens Syntopic Common Wall Lizard 1. Old road in Forra Piave, Santo Stefano di Cadore (BL), Veneto * 13.08.16 08.08.17 46.53516, 12.50901. 46.53258, 12.49400 825-870 S R67: Neritic and platform dolomites (medium Triassic) Scots pine oriental forest Up to 3 Yes 2. Casoni, Canal del Maè, Forno di Zoldo (BL), Veneto * 25.08.17 27.08.17 01.07.18 08.07.18 46.30444, 12.22972 750-850 SO R58: Pelagic limestones, marly limestone and marl (Jurassic) Thermophilic calciphile beech forest of the Alps Up to 8 Yes 3. Grisol de Entro, Val del Grisol, Longarone (BL), Veneto * 25.07.18 46.27017, 12.22878 825 SE R58: Pelagic limestones, marly limestone and marl (Jurassic) Ostrya carpinifolia thicket 1 Yes 4. Pont, San Lucano Valley, Taibon Agordino (BL), Veneto * 6.8.2018 9.8.2018 17.8.2018 46.30201, 11.91715. 46.30200, 11.91658. 46.30113, 11.91685. 1200- 1250 NE-E R67: Neritic and platform Dolomites (medium Triassic) Calciphile fir forest of the Alps and central-northern Apennines 4 No 5. Monte Tudaio, Vigo di Cadore (BL), Veneto + 12.08.16 46.51413 12.47563 886 S R67: Neritic and platform dolomites (medium Triassic) Scots pine oriental forest 1 Yes 6. Val Zemola, Erto and Casso (PN), Friuli Venezia Giulia + 18.06.17 08.07.18 46.28607, 12.37178 1020 NE R56: Neritic and platform limestones and sometimes dolomites (Jurassic) Thermophilic calciphile beech forest of the Alps Up to 2 No 7. Climbing rock, Listolade, (BL), Veneto + 22.05.16 25.09.16 13.04.17 09.09.18 46.32622, 11.99672 720 SO R67: Neritic and platform Dolomites (medium Triassic) Acidophile scots pine forest Up to 11 Yes 9Westward expansion of the Horvath’s rock lizard Roughness turns out to be important certainly due the preference of the Horvath’s lizard for steep rocky areas. However, the pixel resolution of our analyses (300 m), constrained by the imprecision of the coordinates of the occurrences, is not enough to discriminate between rocky and simply steep slopes. That’s why some areas that are likely unsuitable, like Visentin Mountain just south of Belluno, are predicted as suitable (Fig. 3 and 4). The categorical variable Bedrock confirmed the importance of sedimentary rocks for the presence of the species with the highest suitability (more than 70%) in areas with the dominant bedrocks consisting of carbon- Table 2. AICc scores of different Maxent species distribution models varying in number of variables, features and regularization multipliers (see methods for details). The best model (in bold) is the one with the lowest AICc Model # Variables Features Reg.mult AICc 1 BIO1,BIO4, BIO12, BIO15, Aspect, Asperity, Slope, Corine, Bedrock LQP 1 2459.82 2 BIO5,BIO4, BIO12, BIO15, Aspect, Asperity, Slope, Corine, Bedrock LQP 1 2468.09 3 BIO6,BIO4, BIO12, BIO15, Aspect, Asperity, Slope, Corine, Bedrock LQP 1 2472.79 4 BIO10,BIO4, BIO12, BIO15, Aspect, Asperity, Slope, Corine, Bedrock LQP 1 2465.77 5 BIO11,BIO4, BIO12, BIO15, Aspect, Asperity, Slope, Corine, Bedrock LQP 1 2453.00 6 BIO11,BIO4, BIO18, BIO15, Aspect, Asperity, Slope, Corine, Bedrock LQP 1 2447.56 7 BIO11,BIO4, BIO19, BIO15, Aspect, Asperity, Slope, Corine, Bedrock LQP 1 2463.85 8 BIO11,BIO4, BIO18, BIO15, Aspect, Asperity, Corine, Bedrock LQP 1 2452.31 9 BIO11, BIO18, BIO15, Aspect, Asperity, Corine, Bedrock LQP 1 2447.72 10 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQP 1 2429.10 11 BIO11, BIO15, Aspect, Asperity, Bedrock LQP 1 2430.33 12 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQP 0.05 2424.50 13 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQP 0.1 2431.21 14 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQP 0.5 2424.09 15 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQP 2 2442.47 16 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQP 4 2462.02 17 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQP 10 2529.97 18 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQ 0.05 2423.22 19 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQ 0.1 2422.58 20 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQ 0.5 2424.63 21 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQ 1 2431.17 22 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQ 2 2447.09 23 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQ 4 2459.09 24 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock LQ 10 2530.51 25 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock H 0.5 2795.60 26 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock H 1 2602.81 27 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock H 2 2529.67 28 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock H 4 2532.34 29 BIO11, BIO18, BIO15, Aspect, Asperity, Bedrock H 10 2571.21 Table 3. Relative contributions of the environmental variables to the best Maxent model (see methods for details) for the distribution of Horvath’s rock lizard. Variable Percent contribution (%) Permutation importance (%) Roughness 33.3 26.7 Bedrock 26.5 16.6 Temperature of the coldest quarter 19.5 38 Aspect 8.9 6 Rainfall of the warmest quarter 8.5 3.5 Rainfall seasonality 3.3 9.2 10 Giuseppe De Marchi et alii ates, dolomite and limestone, and with an intermedi- ate suitability (40%) for areas where these three types of rocks are present, even if they are not dominant. A lower suitability (about 28%) was found for other types of sedi- mentary bedrocks. The remaining two types of bedrocks (metamorphic-magmatic and undifferentiated) were pre- dicted as unsuitable. This result likely stems from the need of the lizard for fissured rocks where they can hide from predators and where rock dwelling plants (impor- tant for attracting the invertebrate food) can take root. The inclusion of temperature as an important varia- ble was similarly expected as the species is generally con- sidered cold-adapted and able to live at higher altitudes and in shadier places compared to the frequently syntopic common wall lizard (Žagar, 2016). Why the temperature of the coldest quarter turned out to be the most impor- Fig. 3. The best species distribution model of Horvath’s rock lizard in the Alps and neighbouring Dinaric Chain based on 100 occurrences (black dots) in Italy, Austria and Slovenia. The output is “cloglog”, which gives an estimate between 0 and 1 of the probability of presence. Fig. 4. Rainfall (the lighter areas the higher rainfall) during the warmest quarter. Yellow dots show the 100 occurrences of Horvath’s rock lizard. The light blue line is the 335 mm isohyet. The location of some areas mentioned in the Discussion are added. 11Westward expansion of the Horvath’s rock lizard tant temperature related variable is not clear. It might be related to the duration of the snow cover and a too brief warm season for the species to thrive. Instead, the impor- tance of a physiological limit for survival at low temper- atures would have probably resulted in a higher impor- tance of the minimum winter temperature. Aspect is of course important as the lizards need direct sunlight for warming up and the north aspect is unsuitable (Žagar, 2008a). So, in spite of the low resolu- tion of the occurrences that might mask the exact aspect of the cliff, this variable is a significant percent contribu- tion (8.9%) in explaining the best model of distribution. Our model shows the importance of two rainfall related variables, rainfall seasonality and average rainfall during the warmest quarter. Rainfall seasonality increases westward in the Eastern Alps (data not shown) but why rainfall seasonality might be important for the ecology of the species is not clear yet. Its ability to explain the distri- bution is, at any rate, very low (3.3%). The higher percent contribution of summer rainfall (8.5%) might stem from its likely importance to the growth of rupicolous plants where the Horvath’s rock lizards look for food. Com- mon wall lizards are comparatively less dependent on this hunting ground as they are much more euryecious (Sin- daco et al., 2006). In addition, summer rainfall might be directly important as a source of drinking water as we have observed Horvath’s lizards drinking in wet, dripping parts of the cliffs. Interestingly summer rainfall decreases westwards in the eastern Alps and the 13 westernmost occurrences (almost all in Veneto) are all close to the 335 mm isohyet (Fig. 4). So, the rarity of the species west of the Piave river and its disappearance further west might effectively depend upon a reduced rainfall during the warmest quarter. Two mountain groups, the Monti del Sole and the Vette Feltrine, appear suitable (Fig. 3) and are not separated by any significant physiograghic bar- rier from the four westernmost occurrences (Fig. 4); they should deserve additional field surveys even if a recent research (Cassol et al., 2017) and our data have so far failed to confirm the species in those areas. The major limitation of our study comes from the low precision of the occurrences that did not allow a deeper analysis of the roughness and the inclusion of oth- er factors that can be important, such as solar radiation, which can change radically within a few meters in rocky areas and is certainly important in the context of the competition with the common wall lizard. Future species distribution models based on more precise occurrences (Ernst, 2017) and on finer scaled climatic variables (now at 1 km resolution) can improve our understanding of the ecological niche of the species. Hopefully, phylogeograph- ic studies (Cocca et al., 2016) will soon shed light on the glacial refugia where the Horvath’s lizards survived before colonizing the deglaciating Alps. AKNOWLEDGEMENTS We thank the Italian Ministry of Environment for authorizing the research (prot. 0018347/PNM of 1 Sep- tember 2016). We thank: Michele Cassol for suggesting a useful coauthorship and possible locations to be sur- veyed; Luca Lapini for encouragement and timely infor- mation on a new site he discovered; Gianluca Rassati for providing the unpublished coordinates of two sites; Ste- ven Cross for improving the English. REFERENCES Amat, F. (2008): Exploring female reproductive tactics: trade-offs between clutch size, egg mass and newborn size in lacertid lizards. Herpetological J. 18: 147-153. Asch, K. (2003): The 1:5 Million International Geological Map of Europe and Adjacent Areas: Development and Implementation of a GIS-Enabled Concept: Geologis- ches Jahrbuch SA3. Schweizerbart’sche Verlagsbuch- handlung, Stuttgart. Araújo, M.B., Nogués‐Bravo, D., Diniz‐Filho, J.A.F., Hay- wood, A.M., Valdes, P.J., Rahbek, C. (2008): Quater- nary climate changes explain diversity among reptiles and amphibians. Ecography 31: 8-15. Arscott, D.B., Tockner, K., Ward, J.V. (2001): Thermal heterogeneity along a braided floodplain river (Taglia- mento River, northeastern Italy). Can. J. Fish. Aquat. Sci. 58: 2359-2373. Austin, M.P. (2002): Spatial prediction of species distribu- tion: an interface between ecological theory and sta- tistical modelling. Ecol. Model. 157: 101-118. Bonato L. (2011): Anfibi e Rettili nelle Dolomiti: contrib- uto per una sintesi faunistica. In: Atti del 2° Conveg- no aspetti naturalistici della provincia di Belluno, pp. 85-105. Tipografia Piave, Belluno. Bonato, L., Fracasso, G., Pollo, R., Richard, J., Semenzato, M. (eds) (2007): Atlante degli Anfibi e dei Rettili del Veneto. Associazione Faunisti Veneti. Nuovadimen- sione Ed., Portogruaro. Cabela, A., Grillitsch, H., Tiedemann, F. (2002): New records of Lacerta horvathi MÉHELY, 1904. Carinthia (Austria). Herpetozoa 15: 190-192. Cabela, A., Grillitsch, H., Tiedemann, F. (2007): Habi- tatpraferenzen von Podarcis muralis (Laurenti, 1768) und Iberolacerta horvathi (Méhely, 1904) bei gemein- samem Vorkommen. Herpetozoa 19: 149-160. 12 Giuseppe De Marchi et alii Cafuta, V. (2010): First record of Horvath`s rock lizard (Iberolacerta horvathi) in the Karavanke Mts., Slove- nia. Nat. Slov. 12: 127-128. Carton, A., Bondesan, A., Fontana, A., Meneghel, M., Miola, A., Mozzi, P., Primon, S., Surian, N. (2009): Geomorphological evolution and sediment transfer in the Piave River system (northeastern Italy) since the Last Glacial Maximum. Géomorphologie 15: 155-174. Cassol, M., Romanazzi, E., Di Cerbo, A.R., Vettorazzo, E. (2017): Atlante degli Anfibi e dei Rettili del Parco Nazionale Dolomiti Bellunesi. Collana Rapporti 10, Parco Nazionale delle Dolomiti Bellunesi, Feltre. Carranza, S., Arnold, E.N., Amat, F. (2004): DNA phy- logeny of Lacerta (Iberolacerta) and other lacertine lizards (Reptilia: Lacertidae): did competition cause long‐term mountain restriction? Syst. Biodiv. 2: 57-77. Cocca, W., Žagar, A., Carretero, M.A., Krofel, M., Luznik, M., Podnar, M., Tvrtkovic, N., Crottini, A. (2016): Phylogeography of the Horvath’s Rock Lizard Iberol- acerta horvathi, across the Alpine-Dinaric Mountain range. - Abstracts of the 9th Symposium on the Lac- ertids of the Mediterranean Basin & 1st Symposium on Mediterranean Lizards, 20-23 June 2016 at Limas- sol. pp. 47. Nicosia. Cordier, S., Adamson, K., Delmas, M., Calvet, M., Har- mand, D. (2017): Of ice and water: Quaternary fluvial response to glacial forcing. Quaternary Sci. Rev. 166: 57-73. Corti, C., Capula, M., Luiselli, L., Razzetti, E., Sindaco, R. (Eds) (2010): Reptilia. Fauna d’Italia, XLV, Ed. Calde- rini, Bologna. Crochet, P.A., Chaline, O., Surget-Groba, Y., Debain, C., Cheylan, M. (2004): Speciation in mountains: phy- logeography and phylogeny of the rock lizards genus Iberolacerta (Reptilia: Lacertidae). Molecular Phylo- genet. Evol. 30: 860-866. De Marchi, G., Boz, B., Bombieri, G., Richard, J. (2019): Proposta di un approccio metodologico rigoroso per la ricerca di specie di problematica determinazione: Il caso della lucertola di Horvath, Iberolacerta horvathi (Squa- mata: Lacertidae). In: Atti 8° Convegno dei Faunisti Veneti (2018), Bonato, L., Spada, A., Cassol, M. (Eds.). Boll. Mus. St. Nat. Venezia. 68 (Suppl): 68-72. Dullinger, S., Willner, W., Plutzar, C., Englisch, T., Schratt‐Ehrendorfer, L., Moser, D., Ertl, S., Essl, F., Niklfeld, H. (2012): Post‐glacial migration lag restricts range filling of plants in the European Alps. Global Ecol. Biogeogr. 21: 829-840. Elith, J., Graham, C.H., Anderson, R.P, Dudik, M., Fer- rier, S., Guisan, A., Hijmans, R.J., Huettmann, F., Leathwick, J.R., Lehmann, A., Li, J., Lohmann, L.G., Loiselle, B.A., Manion, G., Moritz, C., Nakamura, M., Nakazawa, Y., Overton, J.M., Peterson, A.T., Phillips, S.J., Richardson, K., Scachetti-Pereira, R., Schapire, R.E., Soberon J., Williams, S., Wisz, M.S., Zimmer- mann, N.E. (2006): Novel methods improve predic- tion of species’ distributions from occurrence data. Ecography 29: 129-151. Ernst, K., Gollmann, G., Plutzar, C., Schweiger, S. (2017): Species distribution modeling and field survey on Iberolacerta horvathi (Méhely, 1904) in Austria: pre- dicted and actual habitats considering topographical, geological and bioclimatic parameters. Abstract book of the 19th Ordinary General Meeting of SEH, Salz- burg: 69 Hijmans, R.J., van Etten, J. (2016): raster: Geographic Data Analysis and Modeling. R packageversion 2.5-8. URL https://CRAN.R-project.org/package=raster. Ivy-Ochs, S., Kerschner, H., Maisch, M., Christl, M., Kubik, P.W., Schlüchter, C. (2009): Latest Pleistocene and Holocene glacier variations in the European Alps. Quaternary Sci. Rev. 28: 2137-2149. Joger, U., Fritz, U., Guicking, D., Kalyabina-Hauf, S., Nagy, Z.T., Wink, M. (2007): Phylogeography of west- ern Palaearctic reptiles–Spatial and temporal spe- ciation patterns. Zoologischer Anzeiger-A Journal of Comparative Zoology 246: 293-313. Krofel, M., Cafuta, V., Planinc, G., Sopotnik, M., Šalamun, A., Tome, S., Vamberger, M., Žagar, A. (2009): Razširjenost plazilcev v Sloveniji: pregled podatkov, zbranih do leta 2009. Nat. Slov. 11: 61-99. Lapini, L., (2016): Una nuova popolazione di Lucertola di Horvath Iberolacerta horvathi (Méhely 1904) sulle Dolomiti (Reptilia: Lacertidae, Italia nord-orientale, regione Veneto). Gortania 37: 83-87. Lapini, L. (2017): Novità e conferme: la lucertola di Hor- vath (Iberolacerta horvathi) sulle Dolomiti. Frammen- ti, Conoscere e tutelare la natura bellunese 7: 57-62. Lapini, L., Dal Farra, A. (1994): Lacerta horvathi Mehely, 1904 sulle Dolomiti (Reptilia, Lacertidae). Boll. Mus. Civ. Stor. Nat. Venezia 43: 205-208. Lapini, L., Dall’Asta, A., Luiselli, L., Nardi, P. (2004): Lac- erta horvathi in Italy: a review with new data on dis- tribution, spacing strategy and territoriality (Reptilia, Lacertidae). B. Zool. 71 (S2): 145-151. Le Galliard, J.F., Massot, M., Baron, J.P., Clobert, J. (2012): Ecological effects of climate change on Euro- pean reptiles. In: Brodie, J.F., Post, E.S., Doak, D.F. (Eds.). Wildlife conservation in a changing climate. pp.179-203. University of Chicago Press, Chicago. Ljubisavljević, K., Glasnović, P., Kalan, K., Kryštufek, B. (2012): Female reproductive characteristics of the Horvath’s rock lizard (Iberolacerta horvathi) from Slo- venia. Arch. Biol. Sci. 64: 639-645. 13Westward expansion of the Horvath’s rock lizard Merow, C., Smith, M.J., Silander Jr, J.A. (2013): A prac- tical guide to MaxEnt for modeling species’ distribu- tions: what it does, and why inputs and settings mat- ter. Ecography 36: 1058-1069. Morales, N.S., Fernández, I.C., Baca-González, V. (2017): MaxEnt’s parameter configuration and small samples: are we paying attention to recommendations? A sys- tematic review. PeerJ 5: e3093. Ortega, Z., Mencía, A., Pérez-Mellado, V. (2016): Are mountain habitats becoming more suitable for gen- eralist than cold-adapted lizards thermoregulation? PeerJ 4: e2085. Ortner, O. (2006): Reptilienkartierung im Oberen Gailtal und Lesachtal 2005; Naturwissenschaftlicher Verein für Kärnten, 2006; Klagenfurt; 221-230. Osojnik, N., Žagar, A., Carretero, M.A., Garcia-Munoz, E., Vrezec, A. (2013): Ecophysiological dissimilarities of two sympatric lizards. Herpetologica 69: 445-454. Phillips, S.J. (2006): A brief tutorial on Maxent. AT&T Research. Available at: (https://biodiversityinformat- ics.amnh.org/open_source/maxent/Maxent_tutori- al2017.pdf ) Phillips, S.J., Anderson, R.P., Dudík, M., Schapire, R.E., Blair, M.E. (2017): Opening the black box: an open- source release of Maxent. Ecography 40: 887-893. Phillips, S.J., Anderson, R.P., Schapire, R.E. (2006): Maxi- mum entropy modeling of species geographic distri- butions. Ecol. Model. 190: 231-259. Pinkert, S., Dijkstra, K.D.B., Zeuss, D., Reudenbach, C., Brandl, R., Hof, C. (2018): Evolutionary processes, dispersal limitation and climatic history shape current diversity patterns of European dragonflies. Ecography 41: 795-804. Rassati, G. (2010): Contributo alla conoscenza della dis- tribuzione della Lucertola di Horvath Iberolacerta horvathi e della Lucertola dei muri Podarcis muralis in Friuli Venezia Giulia e in Veneto. Atti Mus. Civ. Stor. Nat. Trieste 54: 133-146. Rassati, G. (2012): Contributo alla conoscenza della dis- tribuzione di alcune specie di Amphibia e di Reptilia in Friuli Venezia Giulia e in Veneto. Atti Mus. Civ. Stor. Nat. Trieste 55: 91-135. Rassati, G. (2017): Cohabitation of rupicolous insec- tivorous species belonging to different classes, aves and reptilia on the same cliff face. Uccelli d’Italia 42: 93-96. Richard, J., Lapini, L. (1993): Trophic niche overlap in syntopic populations of Lacerta horvathi and Podarcis muralis (Reptilia, Lacertidae). Atti Mus. Civ. Stor. Nat. Trieste 45: 151-157. Shcheglovitova, M., Anderson, R.P. (2013): Estimating optimal complexity for ecological niche models: a jackknife approach for species with small sample siz- es. Ecol. Model. 269: 9-17. Sindaco, R., Doria, G., Razzetti, E., Bernini, F. (2006): Atlante degli Anfibi e Rettili d’Italia–Atlas of Amphib- ians and Reptiles in Italy. Edizioni Polistampa, Firenze. Speybroeck, J., Beukema, W., Bok, B., Van Der Voort, J. (2016): Field guide to the amphibians and reptiles of Britain and Europe. Bloomsbury Publishing, London. Svenning, J., Skov, F. (2007): Could the tree diversity pat- tern in Europe be generated by postglacial dispersal limitation? Ecol. Let. 10: 453-60 . Tiedemann, F. (1992): Zur Verbreitung der Kroatischen Gebirgseidechse, Lacerta horvathi Méhely, 1904, in Österreich (Squamata: Sauria: Lacertidae). Herpeto- zoa 5: 67-69. Tormen, G., Tormen, F., De Col, S., (1998): Atlante deg- li Anfibi e Rettili della provincia di Belluno. Gruppo natura bellunese. Atti Convegno aspetti naturalistici della provincia di Belluno pp. 285-314. Tip. Niero, Belluno. Warren, D.L, Seifert, S.N. (2011): Ecological niche mod- eling in Maxent: the importance of model complexity and the performance of model selection criteria. Ecol. Appl. 21: 335-342. Willner, W., Di Pietro, R., Bergmeier, E. (2009): Phytoge- ographical evidence for post-glacial dispersal limita- tion of European beech forest species. Ecography 32: 1011-1018. Wilson, M.F., O’Connell, B., Brown, C., Guinan, J.C., Gre- han, A.J. (2007): Multiscale terrain analysis of multi- beam bathymetry data for habitat mapping on the continental slope. Mar. GeoD. 30: 3-35. Žagar, A. (2008a): Pomen presvetlitev za plazilce (Reptil- ia) v gozdni krajini. Univerza v Ljubljani, Biotehniška fakulteta, Oddelek za biologijo, Ljubljana. Žagar, A. (2008b): The lowest altitudinal record of Hor- vath’s Rock Lizard (Iberolacerta horvathi) in Slovenia. Nat. Slov. 10: 59-61. Žagar, A. (2016): Altitudinal distribution and habitat use of the common wall lizard Podarcis muralis (Lin- naeus, 1768) and the Horvath’s rock lizard Iberolacer- ta horvathi (Méhely, 1904) in the Kocevsko region (S Slovenia). Nat. Slov. 18: 47-62. Žagar, A., Carretero, M.A., Krofel, M., Luznik, M., Pod- nar, M., Tvrtkovic, N. (2014): Reptile survey in Dinara Mountain (Croatia) revealed the southern- most known population of Horvath’s rock lizard (Iberolacerta horvathi). Nat. Croat. 23: 235. Žagar, A., Carretero, M.A., Osojnik, N., Sillero, N., Vrezec, A. (2015a): A place in the sun: interspecific interference affects thermoregulation in coexisting liz- ards. Behav. Ecol. Sociobiol. 69: 1127-1137. 14 Giuseppe De Marchi et alii Žagar, A., Carretero, M. A., Vrezec, A., Drašler, K., Kaliontzopoulou, A. (2017): Towards a functional understanding of species coexistence: ecomorpho- logical variation in relation to whole‐organism perfor- mance in two sympatric lizards. Funct. Ecol. 31: 1780- 1791. Žagar, A., Kos, I., Vrezec, A. (2013): Habitat segregation patterns of reptiles in Northern Dinaric Mountains (Slovenia). Amphibia-Reptilia 34: 263-268. Žagar, A., Osojnik, N., Carretero, M.A., Vrezec, A. (2012): Quantifying the intersexual and interspecific morphometric variation in two resembling sympatric lacertids: Iberolacerta horvathi and Podarcis muralis. Acta Herpetol. 7: 29-39. Žagar, A., Simčič, T., Carretero, M.A., Vrezec, A. (2015b): The role of metabolism in understanding the altitudi- nal segregation pattern of two potentially interacting lizards. Comp. Biochem. Phys. A 179: 1-6. Žagar, A., Trilar, T., Carretero, M.A. (2011): Horvath’s rock lizard, Iberolacerta horvathi (Méhely, 1904), feeding on a scorpion in Slovenia. Herpetology Notes 4: 307-309. Zauner, J. (2018): Die Kroatische Gebirgseidechse (Iber- olacerta horvathi) im Nordwesten Sloweniens, eine Bewohnerin von Festungsanlagen aus dem Ersten Weltkrieg. Die Eidechse 29: 65-74. Zeng, Y., Low B.W., Yeo, D.C.J. (2016): Novel methods to select environmental variables in MaxEnt: A case study using invasive crayfish. Ecol. Model. 341: 5-13. Acta Herpetologica Vol. 15, n. 1 - June 2020 Firenze University Press Has the West been won? A field survey and a species distribution model of Iberolacerta horvathi in the Alps Giuseppe De Marchi1,*, Giovanni Bombieri1, Bruno Boz1, Fausto Leandri2, Jacopo Richard3 First record of underwater sound produced by the Balkan crested newt (Triturus ivanbureschi) Simeon Lukanov Identification of biologically active fractions in the dermal secretions of the genus Bombina (Amphibia: Anura: Bombinatoridae) Iryna Udovychenko1,*, Oleksandra Oskyrko1, Oleksii Marushchak2, Tetiana Halenova1, Oleksii Savchuk1 Don’t tread on me: an examination of the anti-predatory behavior of Eastern Copperheads (Agkistrodon contortrix) Andrew Adams1,2,*, John Garrison2, Scott Mcdaniel2, Emily Bueche2, Hunter Howell2,3 An overview of research regarding reservoirs, vectors and predators of the chytrid fungus Batrachochytrium dendrobatidis Jarid Prahl, Thomas P. Wilson*, David Giles, J. Hill Craddock Genetic characteristics of an introduced population of Bombina bombina (Linnaeus, 1761) (Amphibia: Bombinatoridae) in Moselle, France Jean-Pierre Vacher1, 2,*, Damien Aumaître3, Sylvain Ursenbacher2,4 Adaptive significance of the transparent body in the tadpoles of ornamented pygmy frog, Microhyla ornata (Anura, Amphibia) Santosh M. Mogali*, Bhagyashri A. Shanbhag, Srinivas K. Saidapur Comparison of complement system activity amongst wild and domestic animals María S. Moleón1,2,*, Mark E. Merchant3, Hugo H. Ortega4, Pablo A. Siroski1,2 Effects of temperature and food level on plasticity of metamorphic traits in Bufo gargarizans gargarizans larvae Tong Lei Yu*, Yan Ting Han