Acta Herpetologica 13(1): 33-42, 2018 ISSN 1827-9635 (print) © Firenze University Press ISSN 1827-9643 (online) www.fupress.com/ah DOI: 10.13128/Acta_Herpetol-20871 Electric circuit theory applied to alien invasions: a connectivity model predicting the Balkan frog expansion in Northern Italy Mattia Falaschi1,2,*, Marco Mangiacotti3, Roberto Sacchi3, Stefano Scali2, Edoardo Razzetti4 1 Dipartimento di Scienze e Politiche Ambientali, Università degli Studi di Milano, Via Celoria 26, 20133 Milano, Italy 2 Museo Civico di Storia Naturale di Milano, Corso Venezia 55, 20121 Milano, Italy 3 Dipartimento di Scienze della Terra e dell’Ambiente, Università di Pavia, Via Taramelli 24, 27100 Pavia, Italy 4 Museo di Storia Naturale, Università di Pavia, Piazza Botta 9, 27100 Pavia, Italy *Corresponding author. E-mail: matt_fala@hotmail.it Submitted on 2017, 23rd June; revised on 2017, 15th November; accepted on 13th December 2017 Guest Editor: Sebastiano Salvidio Abstract. Invasive species are a major threat to biodiversity and alien herpetofauna already caused conservation prob- lems in Italy and in the world. Pelophylax kurtmuelleri (Gayda, 1940) is a water frog native to Southern Balkans, that was introduced in NW Italy (Liguria) in 1941, from where it has spread to Piedmont, Lombardy and recently to Emil- ia Romagna. Surprisingly, during its expansion, the Balkan frog seemed not to be able to cross the Po River (with a single exception). So, we investigated the reasons for such directional limitation of range expansion, focusing on the potential role of the Po River. Combining ecological niche and connectivity models, we adopted a two-step process in order to assess: (i) if the habitat suitability differs between the two sides of the Po Plain; (ii) if the Po River could act as a barrier in terms of dispersal. Results showed that the northern side of the plain seems less suitable than the southern one, even if many suitable patches occur along the main left tributaries of the Po River and in the hilly part of the plain. The Po River behaves like a barrier for the Balkan frog, but the only known point on the north bank of the river indicates that it may be able to only slow the dispersal and not to completely stop it. So, the most probable interpretation for the absence of the Balkan frog from the Northern Po Plain, is the combination of habitat suitability and connectivity issues: the former makes less probable that new viable populations can establish in the North, the lat- ter decreases the northwards flux of the frogs. Given these findings, the Balkan frog seems able to spread in the whole Northern Italy and the colonization of the northern Po plain could happen shortly. Keywords. Balkan frog, Pelophylax kurtmuelleri, invasive alien species, connectivity models, Italy. INTRODUCTION Invasive alien species (IAS) are among the main threats to biodiversity at global scale (Clavero and García-Berthou, 2005), notably for freshwater ecosys- tems (Gallardo et al., 2016). Indeed, throughout a vari- ety of interactions with the native biota (e.g., competi- tion for space and resources, Carmona-Catot et al., 2013; predation on native fauna, Roy et al., 2012; genetic loss due to hybridization, Muhlfeld et al., 2014; spread of dis- ease, Gherardi, 2006), IAS have been proved to have the potential to radically affect the invaded ecosystems. Her- petological examples of IAS are available (e.g., Rhinella marina, Shine, 2010; Boiga irregularis, Rodda and Fritts, 1992; Lever, 2003) and alien herpetofauna already caused conservation problems in Italy (e.g., Trachemys scripta; Ficetola et al., 2009; Xenopus laevis, Lillo et al., 2011; Lithobates catesbeianus, Federici et al., 2008; Ficetola and Scali, 2010; Pelophylax ridibundus/kurtmuelleri, Ficetola and Scali, 2010). As a consequence, the issue of IAS has 34 Mattia Falaschi et alii started to be listed among the most important threats to the conservation of the Italian herpetofauna (Scali, 2008). Pelophylax kurtmuelleri (Gayda, 1940) is native to Greece, Albania, Serbia, and Macedonia. It was intro- duced in Italy in western Liguria region in 1941 for food purposes (Lanza, 1962). In the subsequent years, it spread in Liguria and Piedmont (Andreone, 1999). The species has been reported in several localities of South- ern Lombardy between 2005 and 2010 (Razzetti et al., 2010). In 2016, the first occurrences for Emilia Romag- na were reported (http://www.ornitho.it/; accessed on 15/01/2017). This invasive species may potentially threat native water frogs (P. lessonae and P. kl. esculentus) through several negative interactions, as competition for habitat and trophic resources (Ficetola and Scali, 2010), hybridization (Ellstrand and Schierenbeck, 2000; Paga- no et al., 2003), or increasing the spread of pathogens. Indeed, IASs have been proved to work as vectors (Gar- ner et al., 2006) and the presence of the fungal patho- gen Batrachochytrium dendrobatidis has been assessed in P. kurtmuelleri (Bellati et al., 2012). By combining these potential threats P. kurtmuelleri has the potential to nega- tively affect native amphibian populations and it calls for urgent and targeted studies on the real consequences of the introduction of this species in the Italian freshwaters. Beside this need, from a conservation perspective, the knowledge of possible expansion routes of this species could give a substantial benefit in planning future moni- toring and control strategies. Habitat connectivity has been recognized essential by both researchers and nature conservationists (Rödder et al., 2016) and maintaining it has become a crucial con- servation strategy because it determines numerous eco- logical phenomena (McRae et al., 2016). For example, connectivity can mitigate the effects of climate change, by allowing species to shift habitat ranges in a rapidly changing climate (McGuire et al., 2016). Moreover, con- nectivity is fundamental for population survival in a fragmented landscape: isolation makes populations more vulnerable to processes such as inbreeding, genetic drift as well as demographic stochasticity, resulting in higher local extinction probabilities (Fahrig and Merriam, 1985; Taylor et al., 1993). However, the creation of ecological networks could be a double-edged sword, accelerating the spread and establishment of invasive species (Cowley et al., 2015). In this scenario, the dispersal of the Balkan frog in Italy seems to be a suitable case study for analyz- ing the potential effect of connectivity on the spread of an alien species. On the one side, the invasive range of this frog is expanding and every year several new records from previously unoccupied areas are reported (http:// www.ornitho.it/); on the other side, looking at the cur- rent distribution and at the past expansion of this species, it seems that the Po River is acting as a barrier, prevent- ing the Balkan frog to colonize the northern part of the Po plain (Fig. 1A). Indeed, while being widespread South to the Po River, the species is almost absent northwards (with the only exception of a single location in Travacò Siccomario (PV) at the confluence between Po and Tici- no River, E. R. pers. obs.). Nor its absence can be reason- ably explained by a sampling bias, because i) the species can be easily identified by its call; ii) in this region many herpetologists and qualified naturalists usually work, (iii) the new technologies have made very easy to share data on animal occurrence (e.g., ornitho.it, i-naturalist.org), and (iv) targeted field surveys in some areas of Piedmont and Lombardy failed to find the Balkan frog (E.R. and S.S., pers. obs.). Classical Biotic-Abiotic-Mobility theory for species distribution (Soberón and Peterson, 2005) predicts that a species occurs in a given place only if three main condi- tions are fulfilled: the area is appropriate in terms of the fundamental niche (habitat suitability), it meets the suite of biotic conditions necessary for the maintenance of viable populations (e.g., interspecific competition), and it has been accessible to the species over relevant time periods (connectivity). Following this framework and assuming that no major discrepancies occur in the her- petological communities on both sides of the Po Plain (Bernini et al., 2006), we might interpret the absence of the Balkan frog as the consequence of (i) the lack of suit- able areas to the North of the Po River (abiotic factors hypothesis) or (ii) the barrier effect of the Po River itself, which constitutes an almost impermeable obstacle to the northwards dispersal (movement factor hypothesis). So, the present study aims at (i) assessing if the habitat suita- bility differs between the two sides of the Plain, support- ing the abiotic factor hypothesis, and, if not, (ii) evalu- ating if the Po River could be an almost impermeable barrier in terms of dispersal, according to the movement factor hypothesis. To answer these two questions we used a combina- tion of both Ecological Niche Models (ENMs; sensu Sil- lero, 2011) and Connectivity Models (CMs): the former allows us to compare the two sides of the plain in term of availability of suitable habitats; the latter let us testing which level of Po River permeability may translate into an effective reduction in the availability of northwards dispersal routes. Notably, CMs were obtained applying the electrical circuit theory (McRae, 2006; McRae et al., 2008) to the landscape where a species occurs. In this framework, the landscape is treated as a conductance surface where each element (cell in a raster) is linked to 35Predicting the Balkan frog expansion its neighbors and has its own conductance (i.e., perme- ability to current flow). Occurrences become the points in the network where current enters or exits the circuit (McRae, 2006), and current intensity across the land- scape measures the animal potential flux (McRae et al., 2008). The outcomes are current maps representing the most probable connection pathways among suitable areas (McRae, 2006; McRae et al., 2008). Since this theory was introduced in the ecological field, this approach has been used for a wide range of purposes, including the analysis of IAS (Cowley et al., 2015). MATERIAL AND METHODS Occurrence data The occurrence data of P. kurtmuelleri from its invasive range in Italy were obtained from different sources: historical published records, authors’ personal observations, and records by experienced herpetologists (see acknowledgments for a full list). All locations (Fig. 1A) were georeferenced on a 1 × 1 km grid. The study area was delimited to Northern and Central Ita- ly (Fig. 1A). Since records of Balkan frog from new localities are reported almost every week we decided to stop the data collec- tion at the end of 2016. Ecological niche model ENM was fitted using MaxEnt (Phillips et al., 2004)(Phillips et al., 2004), version 3.3.3k (MaxEnt, 2017), which have proved to outperform other methods based on presence-only or pres- ence/background data (Elith et al., 2006). A set of nine raster maps at a 1 × 1 km resolution were arranged in order to account for ecologically meaningful factors: distance from streams, land cover (i.e., the relative coverage inside each 1 × 1 km cell for each land cover class) and elevation (Table 1). All environ- mental variables were derived from maps freely available on the web (ISPRA, 2017): a vector layer of the Italian hydrographic network, a raster layer of the land use (Corine Land Cover) referred to 2012, at 100 × 100 m resolution, and a digital eleva- tion model at 75 × 75 m resolution. To take into account mul- ticollinearity among predictors (which can lead to model over- fitting; Warren and Seifert, 2011), the variance inflation factor (VIF) was obtained for each variable and used as diagnostic sta- tistic: values around one indicate no substantial problems, while the more they depart from one, the more the multicollinearity is serious (e.g., a value of 10 represents an unacceptable value; Faraway, 2004). Since the variable “notirri” (not irrigated crops; Table 1) showed a high VIF (7.80), this predictor was removed. Conversely, all the other predictors showed low multicollinearity (Table 1) and had low bivariate correlation coefficients (Supple- mentary material S1). The availability of environmental factors was determined on the basis of 10 000 background cells randomly drawn from the whole study area. HSM performance was assessed using a cross- fold validation procedure considering the area under the receiv- er operating curve (AUC) as performance score (Merow et al., 2013): the occurrence dataset was randomly split into ten equal parts; each of them was used as the test dataset for the model built with the remnants nine. Accordingly, all records were used nine times for training and one for testing the model, and all the observations are equally weighted in the training and in the testing of the model (Merow et al., 2013). The final HSM was obtained by averaging the prediction of the ten ENMs from each replicate. In order to avoid overfitting, MaxEnt was run allow- ing only “linear”, “product” and “quadratic” features (Warren and Seifert, 2011). Data preparation and modeling were carried out in R environment (R Core Team, 2017), using the raster (Hij- mans, 2016) and dismo (Hijmans et al., 2015) packages. Fig. 1. (A) Map showing the study area and the range expansion of Pelophylax kurtmuelleri from the introduction to 2016. Poly- gons represent the invasive range of this species at different points in time. Dots represent occurrence points. PV: occurrence local- ity north of the Po River; (B) habitat suitability map for Pelophylax kurtmuelleri. 36 Mattia Falaschi et alii To compare the suitability of the northern and southern sides of the Po Plain, we computed a kernel density estima- tions of the distribution of suitability values: we firstly cut HSM with a 20 km buffer around the Po River, and then we fitted and graphically compared two separate Gaussian kernels with fixed bandwidth (0.03; obtained using the native R function “densi- ty”) for the North and South riversides. Connectivity models CMs were built using Circuitscape (Shah and McRae, 2008), version 4.0.5 (Circuitscape.org, 2017), a software designed to model habitat connectivity using the electrical cir- cuit theory (McRae, 2006; McRae et al., 2008). Conductance maps were derived from HSM by rescaling it between 0 (mini- mum suitability) and 1000 (maximum suitability), in order to reduce computational time due to floating points values. Further, to better fit the known ecology of the species, which is never found above 800 meters a.s.l., in both its native and invasive range (Sacchi et al., 2011), the cells at higher eleva- tion were considered absolute barriers. For this purpose, these cells have been assigned an NA value, completely exclud- ing them from the CMs computation. This modified suitabil- ity map was used as the conductance map for the reference model (CMref ), i.e., the model that considers the suitability as the only parameter affecting the potential spread. To evaluate the barrier effect of the Po River, six more CMs were built, by forcing the conductance of the cells intersecting the Po River at increasing values: 0 (absolute barrier, CM0), 1 (minimum permeability, CM1), 10 (CM10), 50 (CM50), 100 (CM100), and 500 (CM500). Circuitscape was run in advanced mode (McRae et al., 2013) to allow for the use of arbitrary current sources and grounds: all the occurrence cells were set as current sources, while the cells at the perimeter of the study areas represent- ed the ground. So, the model was able to predict the possible spread of this invasive species considering all possible direc- tions in one fell swoop (Cowley et al., 2015). When multiple sources are set simultaneously, a possible “repulsive” effect with no ecological basis can be observed. Indeed, current flows from the sources (positive) to the ground (negative) and is prevented from moving between two sources (positive poles). To avoid this problem Circuitscape was run separately for each source cell while leaving ground unchanged. Then, the final model was created by cumulating all the current maps obtained in this way. For this purpose, a custom R function was developed (Supple- mentary material S3). To test possible changes of the predicted spread under dif- ferent Po permeability scenarios, each out of the six connec- tivity maps was compared to the CMref (i.e., the map obtained without manipulating the Po permeability). The comparison was made using Map Comparison Kit software, version 3.2.3 (MCK, 2017), a tool specifically developed for understand- ing the nature and the spatial distribution of the differences between map pairs (Visser and de Nijs, 2006). The software allows for a pixel-by-pixel comparison, whose outcome is, in turn, a map highlighting the location of the differences. Fur- ther, it computes an overall similarity index between a map pair: zero stands for two completely different maps, one stands for an identical pair (Visser and de Nijs, 2006). Given that CMs are numerical, the “fuzzy numerical” algorithm was applied (Visser and de Nijs, 2006). RESULTS Ecological niche model A total of 299 occurrence points were collected (cor- responding to 160 occupied cells). The ENM had a good performance (AUC = 0.807 ± 0.012; mean ± standard error), with the highest variable contribution of altitude (43.6 ± 0.4%), followed by distance from the nearest stream (27.1 ± 0.6%). The response curves showed that suitability scores decrease with increasing altitude and Table 1. List of the environmental variables considered in building HSM. “Source” refers to the original map from which the data were derived or drawn. VIF (variance inflation factor): the VIF values for the whole set of variables (before) and for the subset obtained eliminat- ing “notirri”, the most collinear variable (after). Name Description Range Source VIF before after dem Average altitude 1-4180 m digital elevation model 1.51 1.55 streamdist Minimum distance from a stream 0-37000 m streams vector 1.13 1.11 irri Percentage coverage of irrigated fields 0-100% land use 1.57 1.04 notirri Percentage coverage of non-irrigated fields 0-100% land use 7.80 removed open Percentage coverage of meadows 0-100% land use 1.17 1.01 shrubs Percentage coverage of bushes 0-100% land use 2.19 1.44 urban Percentage coverage of urban areas 0-100% land use 2.17 1.12 wetlands Percentage coverage of wetlands 0-100% land use 1.30 1.02 woods Percentage coverage of woodlands 0-100% land use 4.68 1.28 37Predicting the Balkan frog expansion distance from stream (Supplementary material S2). Suit- ability areas crossed the whole study area (Fig. 1B), close- ly matching the hydrographic network: the suitability increases where the network is dense, the opposite where it rarefies. On average, the southern side of the Po Riv- er appeared more suitable than the northern one, where many unsuitable areas can be found just next to the riv- erside (Fig. 1B). Indeed, the suitability distribution of the two sides of the Po River showed a North-South asym- metry (Fig. 2): the southern side has more high-scoring cells than the northern one, which, in turn, still main- tains some highly suitable areas. The distribution of the observed occurrences agreed with the HSM to the South, but not to the North of the Po River. Connectivity models The reference connectivity model (CMref, Fig. 3A) depicted a situation where the current was able to flow both southwards and northwards of the Po. By manipu- lating the permeability, the obtained CMs behaved in two sharply different ways (Fig. 3): when the Po River was set as an absolute barrier (CM0, Fig. 3B), current mainly flowed southeast, and only a very small amount of cur- rent flowed in the northern side of the map; high cur- rent flow was visible along the entire coast of Liguria up to Northern Tuscany and also starting from Southern Lombardy across all Emilia Romagna, reaching North- ern Marche region. On the opposite, as soon as the Po permeability was increased (CM1, CM10, CM50, CM100, CM500; respectively from Fig. 3C to Fig. 3G), the north- ern part of the map suddenly matched CMref, with cur- rent flowing along all the main Po’s tributaries such as Dora Baltea, Sesia, Ticino, Lambro, Adda, Serio, Mincio, and reaching also other major rivers like the Adige. This convergence could be already observed in CM1 (Fig. 3C), i.e., the model corresponding to the lowest permeability scenario. Differences and similarities were confirmed by the fuzzy numerical comparison performed in MCK: similarity maps (Fig. 4) showed nearly no differences between CMref and maps were Po is not set as an abso- lute barrier (CM1 to CM500; from Fig. 4B to Fig. 4F), while the comparison between CMref to CM0 (Fig. 4A) lead to significant differences in the northern part of the map. The rapid convergence towards CMref was noticea- ble considering the sudden growth of the overall similar- ity indices among CM models and CMref (Fig. 5): when Po permeability increased from 0 to 1 (i.e., the mini- mum allowable permeability value) the similarity index rose from 0.36 to 0.97. Fig. 3. Connectivity models showing the current flow at the vari- ation of Po permeability. (A) Reference model without any modi- fication to Po permeability CMref. Polygons represent the temporal range expansion (see Fig. 1A); (B) Po as an absolute barrier CM0; (C) permeability set at minimum value of one CM1; (D) permeabil- ity set at 10 CM10; (E) permeability set at 50 CM50; (F) permeability set at 100 CM100; (G) permeability set at 500 CM500. Fig. 2. Comparison of the available suitability of the northern and southern side of the Po River. Density estimation was restricted to a 20 km buffer (to the left and to the right of the river) and a fixed bandwidth (0.03) was used in both cases. 38 Mattia Falaschi et alii DISCUSSION The establishment of the Balkan frog in Italy started in 1941 from western Liguria (Lanza, 1962). Since then, the species is still spreading over Northern Italy, even it has not been actually able to occupy the northern side of the Po plain. HSM showed that the northern plain offers less suitable areas than the southern side, at least within the 20 km buffer zones along the riverbed (Fig. 1B and 2). At the same time, CMs pointed out that the river may act as a barrier itself, preventing the species to reach the northern corridors (i.e., the left-side tributaries of the Po River, Fig. 3). Together, these findings suggest that the absence of the Balkan frog from the North Po Plain may be due to the combination of these two factors, rather than to a prevalent effect of one of them. Indeed, the abi- otic factor hypothesis is only partially supported by HSM. According to the known ecology of the species, suit- able areas include lowland and hilly sites, near to streams and rivers, regardless of the habitat types (Sacchi et al., 2011). Even though these patches are more common and dense South to the Po, they undoubtedly occur also on the opposite riverside. Similarly, the movement factor hypothesis is fully supported only when the Po River is set as an absolute barrier (CM0, Fig. 2B). Even a mini- mum permeability would result in a widespread inva- Fig. 5. Growth of the similarity index among CM models and CMref over the increase of Po permeability. Fig. 4. Similarity maps of the fuzzy numerical comparison performed in MCK. Comparison between (A) CMref and CM0, (B) CMref and CM1, (C) CMref and CM10, (D) CMref and CM50, (E) CMref and CM100, (F) CMref and CM500. 39Predicting the Balkan frog expansion sion of the Northern Italy through the main tributaries of the Po River. Further, the occurrence of a record north to the Po River (PV in Fig. 1A), makes the absolute bar- rier hypothesis unrealistic. Assuming that the record was not related to anthropogenic transport, it would indicate that the Balkan frog actually keeps the potential of cross- ing the Po, even though very limited. Consequently, the lack of records of this species in the North may be due to a co-occurrence of both connectivity and suitability con- straints, i.e., the barrier effect of the Po River and the spa- tial structure of the suitable habitat patches in the north- ern Po plain. While river banks are often suitable habitat for many amphibians, the river itself could act as a biogeographical barrier for several reasons, such as large size of the river- bed (Lampert et al., 2003; Li et al., 2009; Geissler et al., 2015); therefore, the Po River might have played a role in mitigating the colonization of the north bank. Further- more, the spatial distribution of suitable habitat patches may also have played a significant role. As mentioned above, suitable patches are more common and denser south to the Po, while northwards they concentrate along the major Po tributaries, which head north-south (Fig. 1B). In this context, colonization of new habitat patches can happen only if migration of individuals occurred in northern direction. Combining these two phenom- ena, the result is that the gap between the northern and southern distribution of Balkan frog may not be there in the future, and the colonization of the northern Po plain is just a matter of time. In this scenario, once colonized the major corridors (i.e., Alpine tributaries, such as Dora Baltea, Sesia, Ticino, Lambro, Adda, Serio, and Mincio, as well as other major rivers like the Adige), the Balkan frog could easily spread in the whole northern Italy. In the last years, connectivity modeling became a widely used tool to study connectivity among popula- tions, either for conservation purposes or to study species movement (Chetkiewicz et al., 2006; McRae et al., 2016). Until now, electrical circuit theory has been applied to ecology mainly to study connectivity in a network of habitat patches already occupied by native species and only one study focused on modeling the spread of alien species (Cowley et al., 2015). In our study, we showed how connectivity models could be a useful tool to deal with alien invasive species. We are aware that a number of points have to be improved. Habitat selection applies both to residency and movements (Chetkiewicz et al., 2006) and for this reason HSM has been recently used for deriving landscape connectivity (Rödder et al., 2016). However, the use of habitat suitability as a proxy for land- scape connectivity has been recently criticized, because animals might be able to move through less suitable are- as, particularly during long-distance movements (Keeley et al., 2017). Furthermore, hybrid models that consider population dynamics should allow to better understand the processes driving biological invasions and make more reliable predictions (Gallien et al., 2010). Anyway, con- nectivity models and Circuitscape software in particular, are still useful for conservation and management relat- ed to alien species, and they can return information on where to intervene to try to contain the phenomenon. About northern Italy invasion by Balkan frog, our con- clusions suggest two possible scenarios: i) Po River is an absolute barrier and is preventing the Balkan frog from spreading in the northern Italy; ii) Po River is not an absolute barrier and is only mitigating the Balkan frog’s breakthrough. So, where should conservation efforts be focused? If Po is an absolute barrier, our suggestion is to understand which features play a role in this mechanism (e.g., permanent water, large riverbed; Lampert et al., 2003), in order to take advantage of those characteristics by preserving or reinforcing those features, thus avoiding the northern Italy invasion. If Po River is not an abso- lute barrier, the situation would be much more critical and the only tool currently at our disposal might be the monitoring of the phenomenon, focusing the effort on the mayor corridors identified by this work (i.e., left Po tributaries) and on the northern side of the plain. What research and nature conservation policy will be able to do in the future will be fundamental to preserve the environment from this invasion. Despite several pos- sible negative interactions between the Balkan frog and Italian indigenous water frogs (Ficetola and Scali, 2010), no specific study has ever been carried out to explore if and how far P. kurtmuelleri is threatening local biodiver- sity. Hybridization is widespread in water frogs (Holsbeek and Jooris, 2010) and many alien amphibians serve as pathogen vector (Garner et al., 2006; Measey et al., 2016). Carrying out specific studies about these last topics is an urgent need in our opinion, whereby, given the magni- tude that this invasion seems to be able to reach, we hope that this manuscript will serve as an incentive for this purpose. ACKNOWLEDGMENTS The authors would like to thank Roberto Sindaco, Claudio Fiorini, Danio Miserocchi, Armando Gari- boldi, Cristina Bruno, Fabio Cavagnini, Franco Bernini, Elvire Laurens, Massimo Gigante, Luca Artoni and Fabio Simonazzi, which made possible the creation of this work, by collecting and sharing with us the field data regarding the presence of the Balkan frog in Italy. We 40 Mattia Falaschi et alii also thank two anonymous reviewers which valuable con- tributed to improve the manuscript. SUPPLEMENTARY MATERIAL Supplementary material associated with this article can be found at manuscript number 20871. REFERENCES Andreone, F. (1999): Rana “ridibunda” Pallas 1771. In: Erpetologia del Piemonte e della Valle d’Aosta, Atlante degli Anfibi e dei Rettili. Monografia XXVI, pp. 190- 191. Andreone, F., Sindaco, R., Eds, Museo Regionale di Scienze Naturali, Torino, Italy. Bellati, A., Razzetti, E., Resteghini, M., Sacchi, R., Pellitteri- Rosa, D., Casiraghi, M., Bernini, F., Galeotti, P., Fasola, M. (2013): First molecular characterization of invasive alien populations of Pelophylax kurtmuelleri (Gayda, 1940) and new records from Italy. In: Atti IX Congres- so Nazionale della Societas Herpetologica Italica (Bari - Conversano, 26-30 Settembre 2012), pp. 287-289. Scillitani, G., Liuzzi, C., Lorusso, L., Mastropasqua, F., Ventrella, P., Eds, Pineta, Conversano, Italy. Bernini, F., Doria, G., Razzetti, E., Sindaco, R. (2006): Atlante degli Anfibi e dei Rettili d’Italia / Atlas of Ital- ian Amphibians and Reptiles. Societas Herpetologica Italica and Edizioni Polistampa, Florence, Italy. Carmona-Catot, G., Magellan, K., García-Berthou, E. (2013): Temperature-specific competition between invasive mosquitofish and an endangered cyprinodon- tid fish. PLoS One 8: e54734. Chetkiewicz, C.B., St. Clair, C.C., Boyce M.S. (2006): Corridors for conservation: integrating pattern and process. Annu. Rev. Ecol. Evol. S. 37: 317-342. Circuitscape.org (2017): Available at: www.circuitscape. org. Date accessed: 20/06/2017. Clavero, M., García-Berthou, E. (2005): Invasive species are a leading cause of animal extinctions. Trends Ecol. Evol. 20: 110. Cowley, D.J., Johnson, O., Pocock, M.J.O. (2015): Using electric network theory to model the spread of oak processionary moth, Thaumetopoea processionea, in urban woodland patches. Landscape Ecol. 30: 905-918. Elith, J., Graham, C.H., Anderson, R.P., Dudík, 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., Townsend Peterson, A., Phillips, S.J., Richardson, K., Scachetti-Pereira, R., Schapire, R.E., Soberón, J., Williams, S., Wisz, M.S., Zimmermann, N.E. (2006): Novel methods improve prediction of species’ distributions from occurrence data. Ecography 29: 129-151. Ellstrand, N.C., Schierenbeck, K.A. (2000): Hybridiza- tion as a stimulus for the evolution of invasiveness in plants? Proc. Natl. Acad. Sci. 97: 7043-7050. Fahrig, L., Merriam, G. (1985): Habitat patch connectiv- ity and population survival. Ecology 66: 1762-1768. Faraway, J. (2004): Linear models with R. Chapman & Hall, Boca Raton, USA. Federici, S., Clemenzi, S., Favelli, M., Tessa, G., Andreone, F., Casiraghi, M., Crottini, A. (2008): Iden- tification of the pathogen Batrachochytrium dendroba- tidis in amphibian populations of a plain area in the Northwest of Italy. Herpetol. Notes 1: 33-37. Ficetola, G.F., Scali, S. (2010): Invasive amphibians and reptiles in Italy. In: Atti VIII Congresso Nazionale Societas Herpetologica Italica (Chieti, 22-26 settembre 2010), pp. 335-340. Di Tizio, L., Di Cerbo, A.R., Di Francesco, N., Cameli, A., Eds, Ianieri Edizioni, Pes- cara, Italy. Ficetola, G.F., Thuiller, W., Padoa-Schioppa, E. (2009): From introduction to the establishment of alien spe- cies: bioclimatic differences between presence and reproduction localities in the slider turtle. Divers. Dis- trib. 15: 108-116. Gallardo, B., Clavero, M., Sánchez, M.I., Vilà, M. (2016): Global ecological impacts of invasive species in aquat- ic ecosystems. Global Change Biol. 22: 151-163. Gallien, L., Münkemüller, T., Albert, C.H., Boulangeat, I., Thuiller, W. (2010): Predicting potential distributions of invasive species: Where to go from here? Divers. Distrib. 16: 331-342. Garner, T.W.J., Perkins, M.W., Govindarajulu, P., Seg- lie, D., Walker, S., Cunningham, A.A., Fisher, M.C. (2006): The emerging amphibian pathogen Batra- chochytrium dendrobatidis globally infects introduced populations of the North American bullfrog, Rana catesbeiana. Biol. Lett. 2: 455-459. Geissler, P., Hartmann, T., Ihlow, F., Rödder, D., Poyark- ov, N.A., Nguyen, T.Q., Ziegler, T., Böhme, W. (2015): The Lower Mekong: an insurmountable barrier to amphibians in Southern Indochina? Biol. J. Linn. Soc. 114: 905-914. Gherardi, F. (2006): Crayfish invading Europe: the case study of Procambarus clarkii. Mar. Freshw. Behav. Physiol. 39: 175-191. Hijmans, R.J. (2016): Raster: Geographic Data Analy- sis and modeling. R Packag. version 2.5-8. https// CRAN.R-project.org/package=raster. 41Predicting the Balkan frog expansion Hijmans, R.J., Phillips, S., Leathwick, J., Elith, J. (2015): dis- mo: Species Distribution Modeling. R package version 1.1-1. https://CRAN.R-project.org/package=dismo. Holsbeek, G., Jooris, R. (2010): Potential impact of genome exclusion by alien species in the hybrido- genetic water frogs (Pelophylax esculentus complex). Biol. Invasions 12: 1-13. ISPRA (2017): Rete del Sistema Informativo Nazionale Ambientale. Available at: www.sinanet.isprambiente.it. Date accessed: 20/06/2017. Keeley, A.T.H., Beier, P., Keeley, B.W., Fagan, M.E. (2017): Habitat suitability is a poor proxy for landscape con- nectivity during dispersal and mating movements. Landscape Urban Plan. 161: 90-102. Lampert, K.P., Rand, A.S., Mueller, U.G., Ryan, M.J. (2003): Fine-scale genetic pattern and evidence for sex-biased dispersal in the túngara frog, Physalaemus pustulosus. Mol. Ecol. 12: 3325-3334. Lanza, B. (1962): On the introduction of Rana ridibun- da Pallas and Rana catesbeiana Shaw in Italy. Copeia 1962: 642-643. Lever, C. (2003): Naturalized reptiles and amphibians of the World. Oxford University Press, New York, USA. Li, R., Chen, W., Tu, L., Fu, J. (2009): Rivers as barriers for high elevation amphibians: a phylogeographic analysis of the alpine stream frog of the Hengduan Mountains. J. Zool. 277: 309-316. Lillo, F., Faraone, F.P., Valvo, M.L. (2011): Can the intro- duction of Xenopus laevis affect native amphibian populations? Reduction of reproductive occurrence in presence of the invasive species. Biol. Invasions 13: 1533-1541. MaxEnt (2017): Maxent software for modeling species niches and distributions. Available at: https://biodiver- sityinformatics.amnh.org/open_source/maxent/. McGuire, J.L., Lawler, J.J., McRae, B.H., Nuñez, T.A., Theobald, D.M. (2016): Achieving climate connectiv- ity in a fragmented landscape. Proc. Natl. Acad. Sci. 113: 7195-7200. MCK (2017): Map Comparison Kit. Available at: http:// mck.riks.nl/software. McRae, B.H., Shah, V.B., Edelman, A. (2016): Cir- cuitscape: modeling landscape connectivity to pro- mote conservation and human health. The Nature Conservancy, Fort Collins, Colorado. McRae, B.H., Shah, V.B., Mohapatra, T.K. (2013): Cir- cuitscape 4 User Guide. The Nature Conservancy. http://www.circuitscape.org. McRae, B.H. (2006): Isolation by resistance. Evolution 60: 1551-1561. McRae, B.H., Dickson, B.G., Keitt, T.H., Shah, V.B. (2008): Using circuit theory to model connectivity in ecology, evolution, and conservation. Ecology 89: 2712-2724. Measey, G.J., Vimercati, G., de Villiers, F.A., Mokhatla, M., Davies, S.J., Thorp, C.J., Rebelo, A.D., Kumschick, S. (2016): A global assessment of alien amphibian impacts in a formal framework. Divers. Distrib. 22: 970-981. Merow, C., Smith, M.J., Silander, J.A. (2013): A practical guide to MaxEnt for modeling species’ distributions: what it does, and why inputs and settings matter. Ecography 36: 1058-1069. Muhlfeld, C.C., Kovach, R.P., Jones, L.A., Al-Chokhachy, R., Boyer, M.C., Leary, R.F., Lowe, W.H., Luikart, G., Allendorf, F.W. (2014): Invasive hybridization in a threatened species is accelerated by climate change. Nat. Clim. Change 4: 620-624. Pagano, A., Dubois, A., Lesbarrères, D., Lodé, T. (2003): Frog alien species: a way for genetic invasion? C. R. Biol. 326: S85-S92. Phillips, S.J., Dudík, M., Schapire, R.E. (2004): A maxi- mum entropy approach to species distribution mod- eling. Proceeding Twenty-First International Confer- ence of Machine Learning, pp. 655-662. R Core Team (2017): R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria. URL https//www.R-pro- ject.org/. Razzetti, E., Pellitteri-rosa, D., Sacchi, R., Tiso, E., Berni- ni, F. (2010): Primi dati sulla presenza di Pelophylax kurtmuelleri (Gayda, 1940) in Lombardia. In: Atti VIII Congresso Nazionale Societas Herpetologica Italica (Chieti, 22-26 settembre 2010), pp. 341-347. Di Tizio, L., Di Cerbo, A.R., Di Francesco, N., Cameli, A., Eds, Ianieri Edizioni, Pescara, Italy. Rodda, G.H., Fritts, T.H. (1992): The impact of the intro- duction of the colubrid snake Boiga irregularis on Guam’s lizards. J. Herpetol. 26: 166-174. Rödder, D., Nekum, S., Cord, A.F., Engler, J.O. (2016): Coupling satellite data with species distribution and connectivity models as a tool for environmental man- agement and planning in matrix-sensitive species. Environ. Manage. 58: 130-143. Roy, H.E., Adriaens, T., Isaac, N.J.B., Kenis, M., Onkelinx, T., Martin, G.S., Brown, P.M.J., Hautier, L., Poland, R., Roy, D.B., Comont, R., Eschen, R., Frost, R., Zindel, R., Van Vlaenderen, J., Nedvěd, O., Ravn, H.P., Gré- goire, J.C., de Biseau, J.C., Maes, D. (2012): Invasive alien predator causes rapid declines of native Europe- an ladybirds. Divers. Distrib. 18: 717-725. Sacchi, R., Razzetti, E., Capula, M. (2011): Pelophylax kurt- muelleri (Gayda, 1940). In: Fauna d’Italia, Vol. XLII, Amphibia, pp. 392-396. Lanza, B., Andreone, F., Bolo- gna, M.A., Corti, C., Razzetti, E., Eds, Bologna, Italy. 42 Mattia Falaschi et alii Scali, S. (2008): Il problema delle introduzioni di anfibi e rettili in Italia e nel mondo / The amphibian and rep- tile introduction problem in Italy and in the World. In: Memorie della Società Italiana di Scienze Naturali e del Museo Civico di Storia Naturale Di Milano Vol- ume XXXVI, pp. 13-14. Galasso, G., Chiozzi, G., Azu- ma, M., Banfi, E., Eds. Shah, V.B., McRae, B.H. (2008): Circuitscape: a tool for landscape ecology. Proceedings of the 7th Python Sci- ence Conference, pp. 62-65. Shine, R. (2010): The ecological impact of invasive cane toadas Bufo marinus in Australia. Quartely Rev. Biol. 85: 253-291. Sillero, N. (2011): What does ecological modelling mod- el? A proposed classification of ecological niche mod- els based on their underlying methods. Ecol. Model. 222: 1343-1346. Soberón, J., Peterson, A.T. (2005): Interpretation of mod- els of fundamental ecological niches and species’ dis- tributional areas. Biodivers. Informatics 2: 1-10. Taylor, P.D., Fahrig, L., Henein, K., Merriam, G. (1993): Connectivity is a vital element of landscape structure. Oikos 68: 571-573. Visser, H., de Nijs, T. (2006): The Map Comparison Kit. Environ. Model. Softw. 21: 346-358. 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. Acta Herpetologica Vol. 13, n. 1 - June 2018 Firenze University Press Species diversity and distribution of lizards in Montenegro Katarina Ljubisavljević1,2,*, Ljiljana Tomović3, Aleksandar Urošević1, Slađana Gvozdenović2, Vuk Iković2, Vernes Zagora2, and Nenad Labus4 The first demographic data and body size of the southern banded newt, Ommatotriton vittatus (Caudata: Salamandridae) Abdullah Altunişik Diet and helminth parasites of freshwater turtles Mesoclemmys tuberculata, Phrynops geoffroanus (Pleurodira: Chelidae) and Kinosternon scorpioides (Criptodyra: Kinosternidae) in a semiarid region, Northeast of Brazil Antonio Marcos Alves Pereira*, Samuel Vieira Brito, João Antonio de Araujo Filho, Adonias Aphoena Martins Teixeira, Diêgo Alves Teles, Daniel Oliveira Santana, Vandeberg Ferreira Lima, Waltécio de Oliveira Almeida Electric circuit theory applied to alien invasions: a connectivity model predicting the Balkan frog expansion in Northern Italy Mattia Falaschi1,2,*, Marco Mangiacotti3, Roberto Sacchi3, Stefano Scali2, Edoardo Razzetti4 First report of Bufo bufo (Linnaeus, 1758) from Sardinia (Italy) Ilaria Maria Cossu1, Salvatore Frau1, Massimo Delfino2,3, Alice Chiodi4, Claudia Corti1,5*, Adriana Bellati4 Natural history and conservation of the Nurse Frog of the Serranía del Perijá Allobates ignotus (Dendrobatoidea: Aromobatidae) in northeastern Colombia Hernán Darío Granda-Rodríguez1,2, Andrés Camilo Montes-Correa3,*, Juan David Jiménez-Bolaño3, Marvin Anganoy-Criollo4,5 Exploring body injuries in the horseshoe whip snake, Hemorrhois hippocrepis Juan M Pleguezuelos*, Esmeralda Alaminos, Mónica Feriche How effectively do European skinks thermoregulate? Evidence from Chalcides ocellatus, a common but overlooked Mediterranean lizard Grigoris Kapsalas, Aris Deimezis-Tsikoutas, Thanos Georgakopoulos, Ismini Gkourtsouli-Antoniadou, Kallirroi Papadaki, Katerina Vassaki, Panayiotis Pafilis Thermal tolerance for two cohorts of a native and an invasive freshwater turtle species Jun Geng, Wei Dang, Qiong Wu, Hong-Liang Lu* Diet of a restocked population of the European pond turtle Emys orbicularis in NW Italy Dario Ottonello1, Fabrizio Oneto1,2, Monica Vignone2, Anita Rizzo2, Sebastiano Salvidio2,* Helminths of the lizard Colobosauroides cearensis (Squamata, Gymnophthalmidae) in an area of Caatinga, Northeastern Brazil Aldenir F. da Silva Neta*, Robson W. Ávila