Living on the edge: habitat selection of Hierophis viridiflavus Stefano Scali, Marco Mangiacotti, Anna Bonardi Museo Civico di Storia Naturale, Milano, C.so Venezia 55, I-20121 Milano (Italy). Corresponding author. E-mail: stefano.scali@comune.milano.it Submitted on 2007, 15th December; revised on 2008, 27th February; accepted on 2008, 1st March. Abstract. We analysed habitat choices of H. viridiflavus in a continental area of north- ern Italy and compared our results with those reported from central Italy by other authors. We used two different field techniques, visual encounter surveys (VES) and radio tracking (RT), and both pointed out a clear preference for edges, while uniform habitats (like mature woods or meadow) were avoided. The same pattern of habitat use is documented for other whip snakes and can be related to the high thermal quality of edges, although other factors could not be ruled out (e.g., prey/shelters abundance). Nevertheless, other researches on Mediterranean populations do not show such a pref- erence, suggesting that at lower latitude habitat thermal quality is not the main con- straint and H. viridiflavus can behave as a habitat generalist. Finally, comparing the two field techniques, we find that VES partially overestimates the importance of edge use, suggesting that caution should be used about its utilization to obtain information in this kind of research. Keywords. Hierophis viridiflavus, habitat use, radio tracking, visual encounter sur- veys, latitude effect. INTRODUCTION Habitat selection data are necessary both for wildlife management/conservation poli- cies and for basic ecological knowledge. In the former case, herpetologists focus their efforts mainly on endangered or threatened species in order to define adequate conserva- tion strategies (e.g., Weatherhead and Charland, 1985; Prior and Shilton, 1996; Nilson et al., 1999; Kingsbury and Coppola, 2000; Webb and Shine, 2000; Filippi and Luiselli, 2003); in the latter case, they obviously prefer species characterized by high-density populations, because they are easily studied (Madsen, 1984; Luiselli and Rugiero, 1990; Luiselli et al., 1994; Larsson, 1995; Luiselli and Capizzi, 1997). In this scenario, Hierophis viridiflavus (Lacépède, 1789) represents an intermediate case: even though it is often very common, particularly in Italy (Vanni and Nistri, 2006), it is protected by European law (included in Acta Herpetologica 3(2): 85-97, 2008 ISSN 1827-9643 (online) © 2008 Firenze University Press 86 S. Scali, M. Mangiacotti and A. Bonardi Annex IV of Habitat Directive 92/43/EEC) because of its relatively restricted geographic range (Naulleau, 1997). As a typical member of the “whip snakes clan” (sometimes called “racers”), H. viridiflavus is a diurnal, thermophilic, generalist predator (Capizzi et al., 1995; Rugiero and Luiselli, 1995; Capula et al., 1997). It can make long-range movements (Ciofi and Chelazzi, 1991; Bonnet and Naulleau, 1996a; Bonnet et al., 1999) and it frequently lives also near human settlements (Luiselli and Capizzi, 1997; Filippi, 2003). All these fea- tures potentially allow it to exploit different habitats, including fragmented and partially anthropic ones. So researchers have both a protected and common snake to study, but despite these favourable condition, habitat preferences of H. viridiflavus had never been investigated in detail and available publications report this kind of ecological information only as mar- ginal note (Luiselli and Rugiero, 1990; Scali and Zuffi, 1994; Capizzi et al., 1995; Capula et al., 1997; Luiselli and Capizzi, 1997; Filippi and Luiselli, 2006). The lack of targeted studies prevents from robust conclusions. The first problem con- cerns the adaptive meaning of habitat selection. Previous studies on H. viridiflavus ecol- ogy were conducted in Mediterranean areas (mainly near Rome), characterised by a hot temperate climate (Cs), according to the Köppen-Geiger classification (Hufty, 1980), with mean annual temperature (MAT) between 14.5 °C and 16.9 °C, and mean temperature in the coldest month (MCM) between 6 °C and 9.9 °C. Climatic conditions are very impor- tant in the ecology of ectothermic vertebrates, and their variations could heavily influ- ence the activity of a thermophilic species, such as H. viridiflavus, and, as a consequence, influence its habitat choices (Shine and Madsen, 1996; Luiselli and Zimmermann, 1997; Webb and Shine, 1998, 2000; Blouin-Demers and Weatherhead, 2001a, 2002; Carfagno and Weatherhead, 2006). The Italian distribution of H. viridiflavus intersects very differ- ent climatic zones: from the temperate subtropical climate (CS) in the southern regions (MAT>17 °C; MCM>10 °C) to the temperate cool climate (Cf ) in the northern part of the peninsula or in the Apennines (670% of the surface), and no herbaceous layer. 4.09 Brambles Thick cluster of brambles (Rubus spp.) with no other vegetation layers 0.29 Wetland Permanent still water. Depth from 30 to 130 cm. Presence of water-lily, reeds, Carex spp. and rush 0.14 Anthropic structure The clay quarry, the kiln and their related buildings. All these structures were abandoned and with some pioneer plants. 3.47 Edge 10 m-wide buffers along habitats interface. Bushy, with several open spots, scarce or absent tree coverage, and often with handmade structures (bricks, ruins, iron bars and other similar materials) 11.61 88 S. Scali, M. Mangiacotti and A. Bonardi used TW4 radio tags (weight 3.0 g; size 22×13×7 mm) and a Mariner Radar Ltd. M57 receiver with a three elements Yagi antenna (Biotrack, Wareham, Dorset, UK). Transmitters weight never exceeded 3.5% of snake weight. Radio tracking was conducted from July to October in 1998 and from March to October in 1999. Snake location was determined once in the morning (between 0800 h and 1000 h) and once in the afternoon, about five hours later. Data analyses To analyse VES data, we superimposed a grid on the area map, identifying 555 cells of 30×30 m (SU = sample unit), being 30 m the mean daily distance moved by radiotelemetered snakes. For- ty-seven were excluded because less than 30% of their surface belonged to the study area, leading to a sample of 508 SUs. We described habitat composition of each SU overlaying the vegetation map to the grid, using ESRI ArcView 3.2 GIS software. In this way, each SU was characterised as the area covered by each habitat type. Each cell was also classified as “present” code (pSU) if at least one snake was observed in it, otherwise as “absent” code (aSU). All the habitat variables were included in the logistic regression (LR) analysis, using the back- ward stepwise likelihood-ratio method, with the presence/absence of snakes as the dependent vari- able. Model significance was assayed by model chi-square test, while its accuracy was tested by com- paring SUs values predicted by the model with observed ones (Field, 2000). Single variable effect on the model was deduced by the regression coefficients (B) sign: a positive B-value means that the probability to find a snake in a given SU increased, and vice versa. To assess the contribution of each variable to the model, and so B-values significance, we tested the change in deviance (-2 log likeli- hood or -2LL) when the estimator was removed (Field, 2000). To describe habitat preferences of radio tracked snakes, we used a type III experimental design, as defined by Thomas and Taylor (1990), where both the resource use and its availability are identified for each individual. We chose the compositional analysis (CA) to analyse data, overcoming any autocorrelation problem (Aebischer et al., 1993; Otis and White, 1999; Garshelis, 2000; Mills- paugh and Marzluff, 2001). For each snake we calculated the proportional use of each habitat (no. of Table 2. Characteristics of implanted snakes. ID Sex Body mass (g) SVL (cm) Year no. of days no. of fixes 12 M 235 95 1998 16 23 13 M 230 96 1998 13 27 14 F 230 90 1998 30 39 15 M 290 89 1999 37 55 16 F 200 82 1999 42 67 18 M 330 88 1999 36 58 27 M 300 93 1999 11 20 28 F 155 79 1999 10 15 30 M 90 64 1999 10 19 32 F 200 83 1999 14 22 33 M 465 101 1999 31 19 89Habitat selection of Hierophis viridiflavus fixes in habitat i/total no. of fixes). To estimate habitat availability, first we calculated the minimum convex polygon (MCP) for each individual; this home range was increased by a 30 m-wide buffer (mean daily movement calculated for all snakes) to obtain a better approximation of available area. In fact, the MCP is calculated on the basis of most external fixes, so a snake could hypothetically move 30 m daily in any direction starting from one of those points. Secondly, we calculated habitat availability for each snake as area covered by ith habitat/total area within extended MCP (EMCP) (Moore and Gillingham, 2006). Habitat use data were transformed using arbitrarily the category “mature wood” as reference, accordingly with the formula proposed by Aebischer et al. (1993): di = ln(xui/xai) - ln(xuj/xaj) where, given a snake, xui is the proportional use of the ith habitat; xai is the available proportion of the ith habitat; xuj is the proportional use of the reference habitat (j); xaj is the available proportion of the reference habitat. When xui was equal to 0, its value was arbitrarily changed to 0.0001 to allow computation (Aebischer et al., 1993). For the test of overall selection we calculated Wilks’ lambda statistic (Λ), using MANOVA techniques, as suggested in Millspaugh and Marzluff (2001). We used snake ID as a fixed factor and five habitats (meadow, mature wood, sparse/shrubby wood, anthropic structures, edge) as dependent variables. Brambles and wetland were omitted because they occurred in a small portion of the study area and their proportional use was marginal. To obtain test signifi- cance we compared –n × ln × (Λ) to χ2 distribution with D-1 degrees of freedom (n is the number of snakes; D is the number of habitat types) (Aebischer et al., 1993). To assess which habitats are preferred, we ranked them on the basis of the average di values ( – di) across all the eleven individuals. Obviously di is equal to zero for the reference category “mature wood”. To test ranking significance, first we compared – di with the reference value using a one-sample t test to verify if the use of each habitat was significantly different from the use of the reference one. Finally, we verified the relative use of each habitat, comparing all pairs of them with a paired t test (Aebischer et al., 1993; Carfagno and Weatherhead, 2006). Field techniques comparison We compared VES and radio-racking results to verify their reliability. A problem arose from the non-independence of radio tracking data, so we decided to compare pSUs obtained simultane- ously by visual surveys and radio tracking data (called “cross” pSU and coded as “0”) with pSUs obtained by radio tracking alone (called “tracking” pSU and coded as “1”). We then conducted a LR using pSU type (cross or track) as the dependent variable and the significant habitat descriptors coming from the previous regression as covariates: if the model is significant, than the two tech- niques give different results; on the contrary the two methods are equivalent. RESULTS Habitat selection A total of 172 snakes were observed using VES and a distribution map with regard to the vegetation was drawn (Fig. 1). Overlapping the 30 × 30m grid to the area map, we obtained 59 pSUs and 449 aSUs. The visual comparison (Fig. 2) of the mean values of each 90 S. Scali, M. Mangiacotti and A. Bonardi vegetation type among pSUs, aSUs and the whole study area shows that pUSs have larger surfaces covered by edges, sparse wood and anthropic structures than aSUs and the whole area. Before attempting LR we randomly extracted 59 aSUs to obtain balanced samples. The model built by LR was highly significant (χ2 = 59.707, df = 3, P < 0.001) and included three variables: wood with shrubby layer (B = 2.49 × 10-3; χ2 = 5.988, df = 1, P < 0.05); anthropic structures (B = 1.90 × 10-3; χ2 = 2.894, df = 1, P < 0.10) and edge (B = 7.73 × 10-3; χ2 = 55.826, df = 1, P < 0.001). All of them had a positive effect on the probability of finding snakes in a SU (positive B-values), but the most significant one was the edge. The model correctly classified 81.4% of pSUs and 72.9% of aSUs, with an overall accuracy of 77.1% and an increase of 27.1% in comparison to the null model. The proportional use of each habitat for each radio tagged snake and the overall habi- tat availability are summarized in Table 3 and in Fig. 3. The mean values show a preference for edges, while mature wood, meadow, sparse/shrubby wood and anthropic structure were less used. In particular, it is interesting to note from Fig. 3 that MCP-availability is quite different from the total area one, showing that selection may occur also at this level. Fig. 1. Spatial relations among habitat features and snakes distribution both by VES (filled circles) and RT (open circles). See legend for habitat types and Table 1 for their detailed description. 91Habitat selection of Hierophis viridiflavus Data in Table 3 were used to calculate the differences in log ratios (di), setting mature wood as the reference category. Wetland and brambles were excluded from the analyses because their availability is too scarce (Bingham and Brennan, 2004). We conducted a MANOVA on the di matrix, and we obtained a significant Wilks’ lambda statistic (Λ = Fig. 3. Comparison among use and two different levels of availabilities for the eleven radio tracked snakes. Bars represents respectively: percentage coverage of each habitat type in the whole study area (total area availability); mean percentage of habitat availability calculated on the basis of buffered individual MCP (mean individual availability); mean percentage of fixes occurring in each habitat types (mean habitat use). Fig. 2. Comparison among mean percentages of each habitat types calculated respectively for the whole area (Total Area), for the absent SUs (aSU; n = 449) and for the present SUs (pSU; n = 59). See text for more detailed definitions. 92 S. Scali, M. Mangiacotti and A. Bonardi 5.70 × 10-2; χ2 = 31.441, df = 4, P < 0.001), showing a non-random habitat selection. So, the habitats were ranked on the basis of –di from the least to the most selected as follows: i) sparse wood, ii) meadow, iii) anthropic structures, iv) mature wood, and v) edge. The pairwise comparison among all the habitats gave significant result only for edges, which were preferred over all other categories (all P < 0.01). Techniques comparison LR used to discriminate between “cross” and “tracking” pSUs gave significant result (χ2 = 12.877, df = 3, P < 0.01). Among the three previously selected variables, only “edge” had a correspondent B-value significantly different from zero (B = -4.28 × 10-3; χ2 = 9.070, df = 1, P < 0.01). A negative value means that an increase in edge surface produced a decrease in the probability that a pSU is a “tracking” unit (coded as “1”). In other words: VES detects that snakes prefer edges to greater degree. The model correctly classified 71.4% of “cross” pSUs (20 out of 28) and 63.0% of “tracking” pSUs (10 out of 17). DISCUSSION Habitat selection Our study demonstrates that H. viridiflavus prefers edges and does not appreciate homogeneous habitats, such as meadows and woods. LR extracted three main variables Table 3. Percentage use and availability of each habitat types for radio-tracked snakes (in parenthesis the sex). Individual snake ID Use/availability (%) Meadow Mature wood Sparse wood/ shrubby layer Brambles Wetland Anthropic structure Edge 12 (M) 0.00/12.95 4.35/2.39 4.35/28.84 0.00/0.00 0.00/0.00 0.00/12.45 91.30/43.38 13 (M) 0.00/4.97 0.00/13.12 44.44/25.41 0.00/0.00 0.00/0.00 0.00/0.00 55.56/56.50 14 (F) 5.13/4.67 10.26/46.21 0.00/15.07 0.00/0.00 0.00/0.24 0.00/0.00 84.62/33.81 15 (M) 0.00/10.10 5.46/29.69 3.64/12.93 0.00/0.98 0.00/0.96 0.00/3.33 90.91/42.02 16 (F) 1.47/25.90 0.00/2.17 0.00/4.68 4.41/3.64 0.00/1.65 1.47/7.36 92.64/54.61 18 (M) 0.00/19.55 3.45/2.86 5.17/12.43 3.45/3.07 0.00/1.64 0.00/2.82 87.93/57.63 27 (M) 0.00/15.58 0.00/3.33 0.00/20.47 0.00/1.01 0.00/1.11 0.00/6.00 100.0/52.50 28 (F) 6.67/31.33 0.00/1.47 0.00/5.63 0.00/0.00 0.00/1.70 0.00/2.33 93.33/57.53 30 (M) 10.53/18.17 0.00/0.01 0.00/11.85 0.00/0.94 0.00/2.16 0.00/14.52 89.47/52.35 32 (F) 4.35/0.35 0.00/1.66 0.00/53.38 0.00/0.00 0.00/0.00 4.35/3.26 91.30/41.36 33 (M) 0.00/0.36 15.79/45.66 0.00/19.04 0.00/0.00 0.00/0.00 0.00/0.00 84.21/34.94 93Habitat selection of Hierophis viridiflavus that directly influence the occurrence probability of the European whip snake: edges, anthropic structures and sparse woods with a shrubby layer. All these variables have a positive effect on the snakes presence. Also CA pointed out a clear preference for edges, and a lesser use of all other habitats. The different importance of anthropic structures and sparse/shrubby woods assessed by the two analyses could seem ambiguous, but it could be explained by their different approaches: CA only takes into consideration the occurrence habitat (e.g., the edge) while LR weighs the habitat composition giving indirect informa- tion about adjacent habitats that form the edge (e.g., an edge between meadow and mature wood). In our study anthropic structures and sparse/shrubby woods are common elements in defining edges used by snakes (see Fig. 1). Habitat use in Mediterranean populations of H. viridiflavus has both similarities and differences with our results. While the minor utilization of forest and woodland is con- firmed, in Mediterranean areas snakes show a strong preference for open habitats, rep- resented by grassy pastures and bushlands, often associated with tall grass (Capula et al., 1997; Filippi and Luiselli, 2006; Luiselli, 2006). The importance of handmade structures is also pointed out in Capula et al. (1997) which reported that over 70% of the observations were done in dry-stone walls and rocky sites surrounded by spiny shrubs. Dry-stone walls are not typical of northern Italy plains but their role could be performed by other build- ings and by ruins such as those associated to the kiln in the present study. These habitats, simulating rocky or semi-natural zones, guarantee the presence of many basking sites, shel- ters and prey abundance (Scali and Zuffi, 1994; Capizzi et al., 1995; Rugiero and Luisel- li, 1995; Webb and Shine, 2000). In the meanwhile, they provide low disturbance sites, because the buildings were abandoned, and only small portions of the kiln were occasion- ally used. Under this assumption, our results are similar to those obtained in central Italy. The main difference remains the intensive use of grassy pastures in Latium, while open and uniform habitats, such as meadows, were not suitable for this species in northern Italy. This difference could correspond to a specific selection pattern or to a difference in habi- tat characterization: the lack of the edge category in the past researches might lead to a subjective assignment of the observations to “pure” habitat types, increasing its importance. Analogous problems are reported in Carfagno and Weatherhead (2006) in their compara- tive studies of habitat selection of the black rat snake (Elaphe obsoleta). The use of edges as habitat category is not usual, but should be more considered by researches, because data from our study and from Carfagno and Weatherhead (2006) confirm their importance for snakes. According to Blouin-Demers and Weatherhead (2002), edges show the highest thermal quality, because they are located at the interface of cool habitats (e.g., forests) and warm habitats (e.g., grasslands, or other open habitats). Thus, edges provide the best oppor- tunities for behavioural thermoregulation of all the habitats, and snakes can invest less in thermoregulation than in areas of low thermal quality. Racers are generalist predators, that need a more precise body temperature control to allow greater sprint speeds than more specialized species (Shine, 1980; Carfagno and Weatherhead, 2006). So, it is not surpris- ing that H. viridiflavus prefers edges in colder habitats, where it is more difficult to achieve the preferred body temperature. Similar patterns of habitat choice are documented also for Coluber constrictor (Plummer and Congdom, 1994; Carfagno and Weatherhead, 2006). The differences in the use of open habitats between northern and central Italy could be explained considering that Mediterranean areas have the optimal thermal characteris- tics for the thermophilic H. viridiflavus. So, this species can behave as a habitat generalist 94 S. Scali, M. Mangiacotti and A. Bonardi in those areas, avoiding only wetlands and cultivations (Filippi and Luiselli, 2006) that do not satisfy other ecological needs (e.g., prey or refuge density). On the opposite, in colder areas, the European whip snake could have stronger thermal constraints, that force it to use only optimal habitats, such as edges. Techniques comparison The two field techniques (VES and RT) provide comparable results, highlighting the same main habitat choice descriptor (edge). The main difference is that VES partially overestimates the importance of edges that are characterized by discontinuities in veg- etation cover which make sightings more likely than in other habitats (Kenward, 1987; Manly et al., 1993; Heyer et al., 1994; Bonnet and Naulleau, 1996; Whiting et al., 1996). This problem suggests caution about VES results. Detailed researches on habitat use by colubrid snakes are only available for radio tracking studies (Weatherhead and Char- land, 1985; Plummer and Congdon, 1994; Keller and Heske, 2000; Blouin-Demers and Weatherhead, 2001a, b, 2002; Rodriguez-Robles, 2003; Carfagno and Weatherhead, 2006), because this technique guarantees a high detail level and a homogeneity of used habitat sampling. Snakes have always very cryptic habits, so VES cannot be considered a bias- free research technique for them. The methodological consequence of our data is that VES could be successfully used for habitat selection studies provided that: the target species do not spend most of their time underground or in heavy undergrowth; selected areas can be easily and completely surveyed; the research effort is homogeneous among all the available habitats (Manly et al., 1993; Garshelis, 2000; Millspaugh and Marzluff, 2001). ACKNOWLEDGEMENTS The research was partially financed by MURST. We would like to thank F. Guidali for the logistic support, I. Turchini for field assistance, Roberto Sacchi and David Greenberg for their useful suggestions about data analysis and Parco Groane direction for giving us the permission of study in the park. We are also grateful to Xavier Bonnet for manuscript improvement. REFERENCES Aebischer, N.J., Robertson, P.A., Kenward, R.E. (1993): Compositional analysis of habitat use from animal radio-tracking data. Ecology 74: 1313-1325. Bingham, R.L., Brennan, L.A. (2004): Comparison of type I error rates for statistical analy- ses of resource selection. J. Wildl. Manag. 68: 207-212. Blomberg, S., Shine, R. (1996): Reptiles. In: Ecological census techniques: a handbook, p. 218-226. Sutherland, W.J., Ed, Cambridge University Press, Cambridge (UK). Blouin-Demers, G., Weatherhead, P.J. (2001a): An experimental test of the link between foraging, habitat selection and thermoregulation in black rat snakes Elaphe obsoleta obsoleta. J. Anim. Ecol. 70: 1006-1013. 95Habitat selection of Hierophis viridiflavus Blouin-Demers, G., Weatherhead, P.J. (2001b): Habitat use by black rat snakes (Elaphe obsoleta obsoleta) in fragmented forests. Ecology 82: 2882-2896. Blouin-Demers, G., Weatherhead, P.J. (2002): Habitat-specific behavioural thermoregula- tion by black rat snakes Elaphe obsoleta obsoleta. Oikos 97: 59-68. Bonardi, A., Mangiacotti, M., Scali, S., Guidali, F. (2001): Un radio-collare per chi non ha collo? Metodologie e problematiche nel radio-tracking su serpenti. Pianura 13: 120-122. Bonnet, X., Naulleau, G. (1996a): Are body reserves important for reproduction in male dark green snakes (Colubridae: Coluber viridiflavus). Herpetologica 52: 137-146. Bonnet, X., Naulleau, G. (1996b): Catchability in snakes: consequences for estimates of breeding frequency. Can. J. Zool. 74: 233-239. Bonnet, X., Naulleau, G., Shine, R. (1999): The dangers of leaving home: dispersal and mortality in snakes. Biol. Cons. 89: 39-50. Capizzi, D., Luiselli, L., Capula, M., Rugiero, L. (1995): Feeding habits of a Mediterrane- an community of snakes in relation to prey availability. Rev. Ecol. (Terre et Vie) 50: 353-363. Capula, M., Filippi, E., Luiselli, L., Trujillo-Jesus, V. (1997): The ecology of the Western Whip Snake (Coluber viridiflavus Lacépède, 1789) in Mediterranean Central Italy. Herpetozoa 10: 65-79. Carfagno, G.L.F., Weatherhead, P.J. (2006): Intraspecific and interspecific variation in use of forest-edge habitat by snakes. Can. J. Zool. 84: 1440-1452. Ciofi, C., Chelazzi, G. (1991): Radiotracking of Coluber viridiflavus using external trans- mitters. J. Herpetol. 25: 37-40. Ciofi, C., Chelazzi, G. (1994): Analysis of homing pattern in the colubrid snake Coluber viridiflavus. J. Herpetol. 28: 477-484. Field, A. (2000): Discovering statistics using SPSS for Windows. SAGE publications Ltd., London. Filippi, E. (2003): The effects of timbering on a snake community of a Mediterranean area of central Italy. Amphibia-Reptilia 24: 75-79. Filippi, E., Luiselli, L. (2003): Aspects of the ecology and conservation of the Meadow viper, Vipera ursinii, in the Duchessa Mountains Natural Park Latium, central Italy: Herp. Bull. 86: 5-9. Filippi, E., Luiselli, L. (2006): Changes in community composition, habitats and abundance of snakes over 10+ years in a protected area in Italy: conservation implications. Her- petol. J. 16: 29-36. Garshelis, D.L. (2000): Delusions in habitat evaluation: measuring use, selection, and importance. In: Research techniques in animal ecology. Controversies and conse- quences, p. 111-164. Boitani, L., Fuller, T.K., Eds, Columbia University Press, New York, Chichester, West Sussex. Heyer, W.R., Donnelly, M.A., McDiarmid, R.W., Hayek, L.A.C., Foster, M.S. (1994): Meas- uring and monitoring biological diversity. Standard methods for Amphibians. Smith- sonian Institution Press, Washington and London. Hufty, A. (1980): La climatologia. Newton-Compton, Roma. Keller, W.L., Heske, E.J. (2000): Habitat use by three species of snakes at the Middle Fork Fish and Wildlife Area, Illinois. J. Herpetol. 34: 558-564. 96 S. Scali, M. Mangiacotti and A. Bonardi Kenward, R. (1987): Wildlife radio tagging: equipment, field techniques and data analysis. Academic Press, London. Kingsbury, B.A., Coppola, C.J. (2000): Hibernacula of the copperbelly water snake (Nerodia erythrogaster neglecta) in Southern Indiana and Kentucky. J. Herpetol. 34: 294-299. Larsson, A.M. (1995): Activity range and habitat selection by radio-tracked Smooth Snakes (Coronella austriaca) on the Swedish west coast. Memo. Soc. Fauna Flora Fenn. 71: 103-105. Luiselli, L. (2006): Ecological modelling of convergence patterns between European and African “whip” snakes. Acta Oecol. 30: 62-68. Luiselli, L., Capizzi, D. (1997): Influences of area, isolation and habitat features on distri- bution of snakes in Mediterranean fragmented woodlands. Biodiv. Conserv. 6: 1339- 1351. Luiselli, L., Capula, M., Rugiero, L., Anibaldi, C. (1994): Habitat choice by melanistic and cryptically coloured morphs of the adder, Vipera berus. Boll. Zool. 61: 213-216. Luiselli, L., Rugiero, L. (1990): On habitat selection and phenology in six species of snakes in Canale Monterano (Tolfa Mountains, Latium, Italy) including data on reproduc- tion and feeding in Vipera aspis francisciredi (Squamata: Viperidae). Herpetozoa 2: 107-115. Luiselli, L., Zimmermann, P. (1997): Thermal ecology and reproductive cyclicity of the snake Natrix tessellata in south-eastern Austria and central Italy: a comparative study. Amphibia-Reptilia 18: 383-396. Madsen, T. (1984): Movements, home range size and habitat use of radio-tracked Grass snakes (Natrix natrix) in southern Sweden. Copeia 1984: 707-713. Manly, B.F.J., McDonald, L.L., Thomas, D.L. (1993): Resource selection by animals. Statis- tical design and analysis for field studies. Chapman & Hall, London. Millspaugh, J.J., Marzluff, J.M. (2001): Radio tracking and animal populations. Academic Press, San Diego and London. Moore, J.A., Gillingham, J.C. (2006): Spatial ecology and multi-scale habitat selection by a threatened rattlesnake: the eastern massasauga, Sistrurus catenatus catenatus. Copeia 2006: 742-751. Naulleau, G. (1997): Coluber viridiflavus Lacépède, 1789. In: Atlas of Amphibians and Reptiles in Europe, p. 342-343. Gasc, J.-P., Cabela, A., Crnobrnja-Isailovic, J., Dol- men, D., Grossenbacher, K., Haffner, P., Le scure, J., Martens, H., Martinez Rica, J.P., Maurin, H., Oliveira, M.E., Sofianidou, T.S., Veith, M., Zuiderwijk, A., Eds, Societas Europaea Herpetologica and Museum National d’Histoire Naturelle (IEGB/SPN) Paris. Nilson, G., Andren, C., Ioannidis, Y., Dimaki, M. (1999): Ecology and conservation of the Milos viper, Macrovipera schweizeri Werner, 1935. Amphibia-Reptilia 20: 355-375. Otis, D.L., White, G.C. (1999): Autocorrelation of location estimates and the analysis of radiotracking data. J. Wildl. Manag. 63: 1039-1044. Pignatti, S. (1997): Flora d’Italia. Edagricole, Bologna. Plummer, M.V., Congdon, J.D. (1994): Radiotelemetric study of activity and movements of racers (Coluber constrictor) associated with a Carolina bay in South Carolina. Copeia 1994: 20-26. 97Habitat selection of Hierophis viridiflavus Prior, K.A., Shilton, C.M. (1996): Post-hibernation mortality in black rat snakes, Elaphe o. obsoleta. J. Herpetol. 30: 275-278. Raunkiaer, C. (1934): The life form of plants and statistical plant geography. Clarendon Press, Oxford. Rodriguez-Robles, J.A. (2003): Home ranges of gopher snakes Pituophis catenifer, (Colu- bridae) in Central California. Copeia 2003: 391-396. Rugiero, L., Luiselli, L. (1995): Food habits of the snake Coluber viridiflavus in relation to prey availability. Amphibia-Reptilia 16: 407-411. Scali, S., Zuffi, M. (1994): Preliminary report on a reptile community ecology in a subur- ban habitat of northern Italy. Boll. Zool. 61: 73-76. Shine, R. (1980): Ecology of eastern Australian whipsnakes of the genus Demansia. J. Her- petol. 14: 381-389. Shine, R., Madsen, T. (1996): Is thermoregulation unimportant for most reptiles? An example using water pythons (Liasis fuscus) in tropical Australia. Physiol. Zool. 69: 252-269. Thomas, D.L., Taylor, E.J. (1990): Study designs and tests for comparing resource use and availability. J. Wildl. Manag. 54: 322-330. Vanni, S., Nistri, A. (2006): Hierophis viridiflavus Lacépède, 1789. In: Atlante degli anfibi e dei rettili d’Italia/Atlas of Italian amphibians and reptiles, p. 544-547. Sindaco, R., Doria, G., Razzetti, E., Bernini, F., Eds, Societas Herpetologica Italica and Edizioni Polistampa, Firenze. Weatherhead, P.J., Charland, M.B. (1985): Habitat selection in an Ontario population of the snake, Elaphe obsoleta. J. Herpetol. 19: 12-19. Webb, J.K., Shine, R. (1998): Thermoregulation by a nocturnal elapid snake (Hoplocephalus bungaroides) in southeastern Australia. Physiol. Zool. 71: 680-692. Webb, J.K., Shine, R. (2000): Paving the way for habitat restoration: can artificial rocks restore degraded habitats of endangered reptiles? Biol. Cons. 92: 93-99. Whiting, M.J., Dixon, J.R., Greene, B.D. (1996): Measuring snake activity patterns: the influence of habitat heterogeneity on catchability. Amphibia-Reptilia 17: 47-54.