Acta Herpetologica 11(1): 63-68, 2016 ISSN 1827-9635 (print) © Firenze University Press ISSN 1827-9643 (online) www.fupress.com/ah DOI: 10.13128/Acta_Herpetol-17198 Photo-identification in amphibian studies: a test of I3S Pattern Marco Sannolo1,*, Francesca Gatti2, Marco Mangiacotti3,4, Stefano Scali3, Roberto Sacchi4 1 CIBIO, Research Centre in Biodiversity and Genetic Resources, InBIO, Universidade do Porto, Campus Agrário de Vairão, Vila do Conde, Portugal. *Corresponding author. E-mail: marco.sannolo@gmail.com 2 Via Garibaldi, 20090 Vimodrone, Milano, Italy 3 Museo Civico di Storia Naturale di Milano, Corso Venezia, 55, 20121 Milano, Italy 4 Dipartimento di Scienze della Terra e dell’Ambiente, Università di Pavia, Via Taramelli 24, 27100, Pavia, Italy Submitted on 2015, 21st October; revised on 2016, 8th March; accepted on 2016, 9th March Editor: Rocco Tiberti Abstract. Photo-identification is used for individual recognition of several animal species. It gives the possibility to take photos of large species from a distance or to avoid invasive marking techniques in small animals. For amphib- ians, the use of non-invasive marking methods is even more relevant in the light of their global decline. Here we use the photo-identification data from a population of Triturus carnifex to validate the photo-identification software I3S Pattern. This recently developed utility has never been applied to amphibians. The software proved to be efficient and accurate for individual recognition for this species. Contrarily to the previous releases of the I3S family, I3S Pattern is particularly suitable for amphibians characterized by a complex individual pattern of large blotches or irregular spots, which are not readily identified by eye. Keywords. Triturus carnifex, computer vision, capture-mark-recapture, photo-identification, SURF algorithm, non- parametric MANOVA. INTRODUCTION Individual recognition is important in ecological studies such as Capture-Mark-Recapture designs (CMR). This kind of research is of primary importance because it provides data on variables of interest like abundance, density, habitat use and dispersion, among others (Seber, 1973; Pradel, 1996; Beirinckx et al., 2006; Mazerolle et al., 2007). CMR studies require the individual marking of captured animals, and rely on the assumption that the marking method unequivocally identifies each individual during the study period (Williams et al., 2002; Mazerolle et al., 2007). Photo-identification (hereafter photo-ID) is a well- established technique for individual recognition in CMR studies: it has been successfully applied to invertebrates (Frisch and Hobbs, 2007; Caci et al., 2013), fish (Speed et al., 2007; Van Tienhoven et al., 2007), amphibians (Gam- ble et al., 2008; Ribeiro and Rebelo, 2011; Zaffaroni Caor- si et al., 2012; Bendik et al., 2013; Moya et al., 2015), rep- tiles (Perera and Perez-Mellado, 2004; Sacchi et al., 2010) and mammals (Hammond et al., 1990; Kelly, 2001). This approach is based on the identification of individually distinct and permanent features of a certain body region (e.g., patterns of colouration, scars), which are little vari- able over time or at least during the study period. Photo- ID has long been performed through by-eye comparison of photos. In the case of large photo-ID catalogues, this approach can be time-consuming and prone to recogni- tion errors (Arntzen et al., 2004) The steady development of digital cameras and com- puters allowed overcoming these issues, and in the past thirty years various software and algorithms had been developed and are now available (e.g., I3S, van Tienhoven 64 Marco Sannolo et alii et al., 2007; Wild-ID, Bolger et al., 2012; Mantha Matcher, Town et al., 2013; MYDAS, Carter et al., 2014; APHIS, Moya et al., 2015). Nevertheless, these utilities were often designed ad hoc for one or few species to maximize their efficiency, therefore their application to other species requires preliminary validation (Sacchi et al., 2016). Using photo-ID as a non-invasive marking tech- nique is particularly recommended for amphibians since they are facing a worldwide decline (Blaustein and Wake, 1990; Pounds et al., 2006) and many other mark- ing techniques could be harmful (Bloch and Irschick, 2004; Heemeyer et al., 2007; Waddle et al., 2008) and ethically debated (May, 2004). However, many amphib- ian species possess complex skin patterns, often charac- terized by large blotches and irregular spots. On the one hand such skin properties allow the unique identification of each individual. On the other hand, if the study spe- cies is characterized by extremely complex pattern, error rate might increase sensibly. Therefore, there is the need of software specifically developed to perform such task. In particular, we focused our attention on I3S Pattern (http://www.reijns.com/i3s/), the latest release of the I3S family (van Tienhoven et al., 2007) because of five main reasons: i) it has been specially designed for those species characterized by big and irregular spots or blotches, such as many amphibians species; ii) it is apparently robust to non-standard photographic settings (as in the case of many field works); iii) it requires little image pre-process- ing procedure (saving operator time); iv) being freeware, it could be addressed to a broad audience among herpe- tologists; v) it has not yet been validated on amphibians. I3S Pattern employs the algorithm SURF (Speeded- Up Robust Features) developed in the field of computer vision (Bay et al., 2007). This free utility allows a fast comparison of the image of the individual to be identi- fied with all the previously stored images and ranks the images from the best to the worst match. To do this, it requires the user to process each image only once by a two steps protocol where: i) setting three homologous “reference points” that allow translating all images in the same two-dimensional reference system (see also Sacchi et al., 2010 for further details); ii) defyining a rectangu- lar area from which the software will automatically search and extract the “interest points” on which a unique pro- file of the image is built (Bay et al., 2007). The chosen area should obviously refer to the same body region, but it has not necessarily be perfectly replicated in each image. The two steps require only few seconds, and a “.fgp” file containing the pattern (i.e., position of refer- ence points and interests points) is created and stored. When a new image is processed and its pattern extract- ed, the software compares it with the database and com- putes the dissimilarity score between the new and the stored patterns (see software manual for computational details; http://www.reijns.com/i3s/download/I3S Pattern. pdf ). At the end of the comparison process, the software shows side by side the new image and all possible match- es ranked by increased dissimilarity score. All proposed matches can be visually inspected to assess if the match is correct. We used the photo-ID data from an ongoing long- term study on the Italian crested newt Triturus carnifex to test the performance of I3S Pattern. T. carnifex is a good model species to test the software because adult newts show a ventral pattern of irregular spots and large blotches that is unique for each individual and stable over time (Arntzen and Wallis, 1999). Therefore, the aims of this research are to assess i) if I3S Pattern successfully rec- ognizes Triturus carnifex individuals, and ii) its potential application to CMR herpetological studies. MATERIAL AND METHODS Study species and sampling sessions We sampled 324 T. carnifex in the Groane Regional Park (Lombardy, Northern Italy, 45°38’N 9°6’E), in a natural pond located in a typical moor area (Gatti and Sannolo, 2014). Newts were caught using funnel traps during nine sampling sessions every week between March and June 2014. We meas- ured the snout-to-vent-length (SVL) and the tail length (TL) of each newt and we noted their sex (Table 1). We took photos of the ventral pattern of each newt in standard conditions: each newt was put on graph paper at 20 cm from the lens (using a Nikon D-90 with an 18-55 mm lens) while keeping its body as straight as possible; each newt was kept wet, and we conducted the whole operation with wet hands and as fast as possible, in order to avoid the injuries and minimize any stress. To simulate a recapture, after one hour we took a second photo of each indi- vidual using the same procedure. Between the two photo ses- sions, newts were kept in individual 1L plastic box filled with water. Following measurements and photo identification, newts were released at the exact point of capture. Two authors (MS and FG) identified by eye every individual captured during all field sessions, checking for recaptures and providing a double- checked reference against which the software performances were evaluated (Arntzen et al., 2004; Bendik et al., 2013). All images were then processed using I3S Pattern. Statistical validation of the software We first evaluated the effect of the two primary sources of error in which the operator could incur, following Sacchi et al. (2010): the error in selecting reference points and research area (Drep1), and the prospective error due to non-perfect align- ment and positioning of the newt under the camera (Drep2). 65Photo-identification in amphibian studies Drep1 is, therefore, the dissimilarity score between the same image processed two times, while Drep2 is the dissimilarity between two images of the same individual taken in one ses- sion of sampling (i.e., a simulated recapture). We also calculat- ed a third measure, Dpop, that is the mean dissimilarity score of each newt with all other individuals. If the software is able to discriminate consistently among individuals, Drep2 should be lower than Dpop. We also expected Drep1 being always greater than zero and Drep2 always greater than Drep1 since it combines the errors of digitalization with the variability due to both newt positioning under the camera and photo quality (e.g., light conditions). The software allows adjusting nine addi- tional parameters for a fine tuning of the algorithm. Changing the number of interest points sampled in each photo might be the single parameter that affect more heavily the calculated distances among individuals. Thus, we focused on the effect of the number of interest points used to calculate the dissimilarity score. The default number is 35 and has been optimized on sea turtles. Therefore, we resampled the database also at 25, 30, 40, 45 interest points. To assess if the number of interest points might affect the dissimilarity score, we applied a non-parametric distance-based MANOVA (Anderson, 2001; McArdle and Anderson, 2001) using the three scores (Drep1, Drep2 and Dpop) as dependent variables and the number of interest points, the sex, and their interaction as the independents. For this test, we used 30 adult newts, 15 males, and 15 females. The NP-MANOVA was per- formed using the function “adonis” of the “vegan” R package (Oksanen et al., 2013) in R ver. 3.1.0 (R Core Team, 2014). The P-values were assessed by setting the number of permutations to 999, and we tested the assumption of homogeneity of dis- persions among distance (Anderson, 2001; Warton et al., 2012) using the functions “betadisper” and “permutest” (Anderson, 2006; Borcard et al., 2011) of the “vegan” R package. Field validation of the software The ability of the software to identify correctly each indi- vidual in the context of a CMR study was then assessed using the newts’ images obtained during the nine session of field cap- ture-mark-recapture. We processed 852 photos relative to 324 adults of T. carnifex sampled in the field between March and May 2014 (Table 2). Among them, 55 individuals were captured at least twice (34 males and 21 females). For these individu- als, we randomly selected one photo and searched in the whole database for the matching. We noted if the software correctly matched the individual, the assigned rank, and the dissimilar- ity score between the sample image and the first correct match in the database. This score is equivalent to Drep2 calculated in the previous test. While the false positive rate (FAR) is usually very low and easy to control using a countercheck method (by- eye comparison), the false rejection rate (FRR) is a major issue with photo-identification (Bolger et al., 2012). FRR has there- fore been calculated as the ratio between the number of correct matches on the number of total attempts. We accepted a match as correct if it was at least one of the first 10 listed images. RESULTS Descriptive statistics for the calculated distances are reported in Table 1. The NP-MANOVA detected a sig- nificant main effect of both sex (P < 0.001) and interest points on distance (P < 0.001), but not of the interaction term (P = 0.64). Sex has a significant effect on Drep2 (Kruskal-Wallis test, KW = 13.5, P < 0.001, df = 1) and on Dpop (KW = 4.4, P < 0.001, df = 1); female dissimi- larity scores were always greater than male ones (Table 1). The effect of interest points was significant only at Dpop level (KW = 122, P < 0.001, df = 4). Increasing the number of interest points, the difference between Drep2 and Dpop clearly decreased reaching its minimum at 45 interest points, but it was always highly significant (KS test at 45 interest points = 1, P < 0.001, one-tailed, Fig. 1). The three scores were clearly separated (Fig. 1). In particular, Dpop was always greater than Drep2, which in turn is greater than Drep1. This result means that the Table 1. Mean dissimilarity scores ± standard deviation in relation to sex and number of interest points. Note that females’ scores are always larger than males’ ones. no. of points 25 30 35 40 45 Drep1 males 0.85 ± 0.44 0.78 ± 0.46 0.72 ± 0.52 0.72 ± 0.46 0.79 ± 0.41 females 0.95 ± 0.76 0.94 ± 0.70 0.77 ± 0.44 0.85 ± 0.52 0.83 ± 0.43 Drep2 males 9.73 ± 5.51 8.24 ± 3.50 7.83 ± 3.85 7.99 ± 4.19 7.39 ± 3.54 females 13.4 ± 5.47 11.8 ± 4.97 10.7 ± 4.85 9.88 ± 3.86 8.91 ± 3.17 Dpop males 48.8 ± 5.97 36.4 ± 3.74 30.3 ± 3.27 25.3 ± 3.06 22.2 ± 2.62 females 57.6 ± 9.65 42.1 ± 6.34 35.2 ± 5.61 27.9 ± 3.73 23.9 ± 2.56 Table 2. Mean ± standard deviation snout-to-vent-length (SVL) and tail length of Triturus carnifex divided by sex and age. Sex n SVL TL Males 195 6.56 ± 0.38 4.69 ± 0.28 Females 129 6.89 ± 0.39 5.51 ± 0.37 Juveniles 12 4.64 ± 0.36 3.73 ± 0.35 66 Marco Sannolo et alii software can correctly distinguish among different newts. Both Drep1 and Drep2 decreased across groups (Table 1), but the difference was not significant. Dpop showed an exponential decrease for the mean values. The sex had a significant effect in the model, while the interaction term sex×no. of points was not significant, which means either that the sex effect was constant across groups or that the test lacked enough statistical power. In Table 1 it is shown that female means and variances were always larger than male ones. When the I3S was applied in the context of a CMR study, it was able to correctly match all 55 recaptured newts out of 324 individuals during the nine sampling sessions. In 42 cases the software proposed the correct match as the first one in the list, while for the other 13 cases, the correct match was between the second and the fifth position in the list. The FRR of the I3S was 0.24, but it rapidly dropped to 0 after the first five matches were visually inspected. The mean distance between the tested image and the first correct match (equivalent to Drep2), was 9.55 ± 3.17 SD. DISCUSSION According to our results, I3S Pattern performs reli- able images classification and can be applied to photo-ID of T. carnifex. As expected from a good classifier, Dpop was greater than Drep2, and this latter greater than Drep1. This outcome means that a greater dissimilarity was cal- culated when comparing the images of two different indi- viduals with respect to images representing the same newt (Table 1, Fig. 1). The ability of the software to assign a smaller score to the pair of images representing the same newt should translate into a correctly proposed match as output. Indeed, we found that I3S Pattern always suggested the correct match as one of the first five candidates, and most often as the first one. Moreover, this software proved to be robust against false negatives. In fact, one of the cen- tral issues using a new software or algorithm is that the rate of misidentification is unknown, especially the false negative rate (Sacchi et al., 2010, 2016), which could sys- tematically bias the estimate of demographic parameters (Davis and Ovaska, 2001; McCarthy et al., 2009). Our estimation of FRR, instead, varied from 0 to 0.24 depend- ing on how restrictive was the selection of the false nega- tive. These values are similar to those estimated by other studies (Bolger et al., 2012). A second outcome is that the number of interest points had a significant effect on the dissimilarity score: increasing the number of interest points from the default setting of 35 to 45 did not lead to an increase in per- formance. Therefore, at least for this species, 35 interest points covered the vast majority of inter-individual vari- ation, managing to tell apart different newts. Apparently, increasing the number of interest points led to sample newt belly in uninformative areas, making more similar two images that represented different individuals. For example, Fig. 2 (on the right newt) summarizes as reflect- ed light may be recognized as a coloration pattern by the software and sampled as interest points. On the contra- ry, reducing the interest points led to increasing the dis- tance between Drep2 and Dpop. So, for this species, and maybe for other similar ones (sharing the same pattern), it is possible that reducing the number of interest points could improve the performance of I3S Pattern, leading to the sampling of highly informative areas only. Sex had a significant effect in our analysis, and both the mean and the variance were greater for females than for males (Tab. 1). This result might be due to phenotyp- ic differences in male and female bellies. Indeed, in our population the female pattern is less contrasted, and the blotches are often less sharp in comparison with male ones. If this difference is consistent among the sexes, as we believe, it is possible that I3S Pattern repeatedly calcu- lated a greater dissimilarity scores for female. In the light of the results of the present study, we are confident that I3S Pattern and its underlined algorithm SURF could be successfully applied not only to species characterized by a regular and clearly distinctive pattern, but also to species keeping highly complex and irregular patterns. This is the case for many species of Bufonidae, Fig. 1. Dissimilarity scores compared across groups (no. of interest points). Note that Drep1 and Drep2 are stable across groups while both mean and variance of Dpop tend to decrease with increasing number of interest points. 67Photo-identification in amphibian studies Ranidae, Salamandridae and Ambystomatidae, which exhibit various cryptic or aposematic patterns that are not easily matched by eye (Vitt and Caldwell, 2013). ACKNOWLEDGEMENTS The present study received no financial support from any funding agency in the public, commercial or not-for- profit sectors. The research was conducted with respect to the current Italian laws in relation to amphibians capture, marking and detention (Aut.Min.Conc. Prot. 0008880/ PNM). The authors would thank Gariboldi L., Wayne, B., and the “Elba Emperor”, Zuffi M.A.L., for their help and advice during the research period. Two anonymous reviewers provided valuable comments that improve an earlier version of the manuscript. REFERENCES Anderson, M.J. (2001): A new method for non-paramet- ric multivariate analysis of variance. Austral Ecol. 26: 32-46. Anderson, M.J. (2006): Distance-based tests for homogene- ity of multivariate dispersions. Biometrics 62: 245-253. Arntzen, J.W., Wallis, G.P. (1999): Geographic variation and taxonomy of crested newts (Triturus cristatus superspecies): morphological and mitochondrial DNA data. Contrib. Zool. 68: 181-203. Arntzen, J.W., Goudie, I.B.J., Halley, J., Jehle, R. (2004): Cost comparison of marking techniques in long-term population studies: PIT-tags versus pattern maps. Amphibia-Reptilia 25: 305-315. Bay, H., Ess, A., Tuytelaars, T., Van Gool, L. (2007): Speeded-Up Robust Features (SURF). Comput. Vis. Image Und. 110: 346-359. Beirinckx, K., Van Gossum, H., Lajeunesse, M.J., Forbes, M.R. (2006): Sex biases in dispersal and philopatry: insights from a meta-analysis based on capture-mark- recapture studies of damselflies. Oikos 113: 539-547. Bendik, N.F., Morrison, T.A., Gluesenkamp, A.G., Sand- ers, M.S., O’Donnell, L.J. (2013): Computer-assisted photo identification outperforms visible implant elastomers in an endangered salamander, Eurycea tonkawae. PLoS ONE 8: e59424. Blaustein, A.R., Wake, D.B. (1990). Declining amphibian populations: a global phenomenon? TREE 5: 203-204 Bloch, N., Irschick, D.J. (2004): Toe-clipping dramatically reduces clinging performance in a pad-bearing lizard (Anolis carolinensis). J. Herpetol. 37: 293-298. Borcard, D., Gillet, F., Legendre, P. (2011): Numerical ecology with R. New York, Springer. Bolger, D.T., Morrison, T.A., Vance, B., Lee, D., Farid, H. (2012): A computer-assisted system for photographic mark–recapture analysis. Methods Ecol. Evol. 3: 813- 822. Caci, G., Biscaccianti, A.B., Cistrone, L., Bosso, L., Garon- na, A.P., Russo, D. (2013): Spotting the right spot: computer-aided individual identification of the threat- ened cerambycid beetle Rosalia alpina. J. Insect Con- serv. 17: 787-795. Carter, S.J.B., Bell, I.P., Miller, J.J., Gash, P.P. (2014): Auto- mated marine turtle photograph identification using artificial neural networks, with application to green turtles. J. Exp. Mar. Bio. Ecol. 452: 105–110. Davis, T.M., Ovaska, K. (2001): Individual recognition of amphibians: effects of toe clipping and fluorescent tagging on the salamander Plethodon vehiculum. J. Herpetol. 35: 217-225. Frisch, A.J., Hobbs, J.A. (2007): Photographic identifica- tion based on unique, polymorphic colour patterns: a novel method for tracking a marine crustacean. J. Exp. Mar. Biol. Ecol. 351: 294-299. Gamble, L., Ravela, S., McGarigal, K. (2008): Multi-scale features for identifying individuals in large biological Fig. 2. Two extremes: on the left an image with 25 interest points, on the right the same image with 45 interest points. While sampling 25 points seems to exclude several, possibly important spots, sam- pling 45 points appears to add useless points from uninformative areas. 68 Marco Sannolo et alii databases: an application of pattern recognition tech- nology to the marbled salamander Ambystoma opa- cum. J. Appl. Ecol. 45: 170-180. Gatti, F., Sannolo, M. (2014): Two cases of polydactyly in the Italian crested newt, Triturus carnifex. Herpetol. Notes 7: 477-478. Hammond, P.S., Mizroch, S.A., Donovan, G.P. (1990): Individual recognition of cetaceans: use of photo- identification and other techniques to estimate pop- ulation parameters. Cambridge, The International Whaling Commission. Heemeyer, J.L., Homyack, J.A., Haas, C.A. (2007): Reten- tion and readability of visible implant elastomer marks in the eastern red-backed salamanders Pletho- don cinereus. Herpetol. Rev. 38: 425-428. Kelly, M.J. (2001): Computer-aided photograph matching in studies using individual identification: an example from Serengeti cheetahs. J. Mammal. 82: 440-449. May, R.M. (2004). Ecology: ethics and amphibians. Nature 431: 403. Mazerolle, M.J., Bailey, L.L., Kendall, W.L., Royle, J.A., Converse, S.J., Nichols, J.D. (2007). Making great leaps forward: accounting for detectability in herpetological field studies. J. Herpetol. 41: 672-689. McArdle, B.H., Anderson, M.J. (2001): Fitting multivariate models to community data: a comment on distance- based redundancy analysis. Ecology 82: 290-297. McCarthy, M.A., Weller, W.F., Parris, K.M. (2009): Effects of toe clipping on survival, recapture, and return rates of Jefferson Salamanders (Ambystoma jeffersonianum) in Ontario, Canada. J. Herpetol. 43: 394-401. Moya, O., Mansilla, P.L., Madrazo, S., Igual, J.M., Rotger, A., Romano, A., Tavecchia, G. (2015): APHIS: a new software for photo-matching in ecological studies. Ecol. Inform. 27: 64-70. Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R.B., Simpson, G.L., Solymos, P., Henry, M., Stevens, H., Wagner, H. (2013). vegan: Community Ecology Package. R package version 2.0- 10. http://CRAN.R-project.org/package=vegan. Perera, A., Perez-Mellado, V. (2004): Photographic identi- fication as a non invasive marking technique for Lac- ertid lizard. Herpetol. Rev. 35: 349-350. Pounds, J.A., Bustamante, M.R., Coloma, L.A., Consue- gra, J.A., Fogden, M.P.L., Foster, P.N., La Marca, E., Masters, K.L., Merino-Viteri, A., Puschendorf, R., Ron, S.R., Sánchez-Azofeifa, G.A., Still, C.J., Young, B.E. (2006): Widespread amphibian extinctions from epidemic disease driven by global warming. Nature 439: 161-167. Pradel, R. (1996): Utilization of capture-mark-recapture for the study of recruitment and population growth rate. Biometrics 52: 703-709. R Development Core Team (2014): R: A Language and Environment for Statistical Computing. Vienna, Aus- tria, R Foundation for Statistical Computing. Avail- able at http://www.R-project.org. Ribeiro, J., Rebelo, R. (2011): Survival of Alytes cisternasii tadpoles in stream pools: a capture-recapture study using photo-identification. Amphibia-Reptilia 32: 365-374. Rondinini, C., Battistoni, A., Peronace, V., Teofili, C. (2013): Lista Rossa IUCN dei Vertebrati Ital- iani Roma, Comitato Italiano IUCN e Ministero dell’Ambiente e della Tutela del Territorio e del Mare. Sacchi R., Scali S., Mangiacotti M., Sannolo M., Zuffi M.A.L. (2016): Digital identification and analysis. In: Reptile ecology and conservation: a handbook of techniques, in press. Dodd K.C. (ed), Oxford Univer- sity Press, Oxford. Sacchi, R., Scali, S., Pellitteri-Rosa, D., Pupin, F., Gentilli, A., Tettamanti, S., Cavigioli, L., Racina, L., Maiocchi, V., Galeotti, P., Fasola, M. (2010): Photographic iden- tification in reptiles: a matter of scales. Amphibia- Reptilia 31: 489-502. Seber, G.A. (1973): Estimating animal abundance and related parameters. New York, Hafner. Speed, C.W., Meekan, M.G., Bradshaw, C.J.A. (2007): Spot the match-wildlife photo-identification using information theory. Front. Zool. 4: 1-11. Town, C., Marshall, A., Sethasathien, N. (2013). Manta Matcher: automated photographic identification of manta rays using keypoint features. Ecol. Evol. 3: 1902-1914. van Tienhoven, A.M., Den Hartog, J.E., Reijns, R.A., Ped- demors, V.M. (2007): A computer-aided program for pattern-matching of natural marks on the spotted rag- gedtooth shark Carcharias taurus. J. Appl. Ecol. 44: 273-280. Vitt, L.J., Caldwell, J.P. (2013): Herpetology. An Introduc- tory Biology of Amphibians and Reptiles. 3rd Edi- tion., Academic Press. Waddle, J.H., Rice, K.G., Mazzotti, F.J., Percival, H.F. (2008): Modeling the effect of toe clipping on treefrog survival: beyond the return rate. J. Herpetol. 42: 467-473. Warton, D.I., Wright, S.T., Wang, Y. (2012): Distance- based multivariate analyses confound location and dispersion effects. Methods Ecol. Evol. 3: 89-101. Williams, B.K., Nichols, J.D., Conroy, M.J. (2002): Analy- sis and management of animal populations San Diego, London, Academic Press. Zaffaroni Caorsi, V., Rocha Santos, R., Grant, T. (2012): Clip or Snap? An evaluation of toe-clipping and pho- to-identification methods for identifying individual southern red-bellied toads, Melanophryniscus cam- baraensis. SAJH 7: 79-84. Acta Herpetologica Vol. 11, n. 1 - June 2016 Firenze University Press Feeding Habits of Mesoclemmys vanderhaegei (Testudines: Chelidae) Elizângela Silva Brito1,*, Franco Leandro Souza2, Christine Strüssmann3 Morphology, ecology, and behaviour of Hylarana intermedia, a Western Ghats frog Ambika Kamath1, Rachakonda Sreekar2,3 A new account for the endangered Cerrado Rocket Frog Allobates goianus (Bokermann, 1975) (Anura: Aromobatidae), with comments on taxonomy and conservation Thiago Ribeiro de Carvalho1,2,*, Lucas Borges Martins1,2, Ariovaldo Antonio Giaretta1 Molecular phylogenetics of the Pristimantis lacrimosus species group (Anura: Craugastoridae) with the description of a new species from Colombia Mauricio Rivera-Correa, Juan M. Daza* Population structure and activity pattern of one species of Adenomera Steindachner, 1867 (Anura: Leptodactylidae) in northeastern Brazil Maria Juliana Borges-Leite1,*, João Fabrício Mota Rodrigues2, Patrícia De Menezes Gondim1, Diva Maria Borges-Nojosa1 Vocal repertoire of Scinax v-signatus (Lutz 1968) (Anura, Hylidae) and comments on bioacoustical synapomorphies for Scinax perpusillus species group Marco Antônio Peixoto1,*, Carla Silva Guimarães1, João Victor A. Lacerda2, Fernando Leal2, Pedro C. Rocha2, Renato Neves Feio1 Amendment of the type locality of the endemic Sicilian pond turtle Emys trinacris Fritz et al. 2005, with some notes on the highest altitude reached by the species (Testudines, Emydidae) Federico Marrone*, Francesco Sacco, Vincenzo Arizza, Marco Arculeo Photo-identification in amphibian studies: a test of I3S Pattern Marco Sannolo1,*, Francesca Gatti2, Marco Mangiacotti3,4, Stefano Scali3, Roberto Sacchi4 No evidence for the ‘expensive-tissue hypothesis’ in the dark-spotted frog, Pelophylax nigromaculatus Li Zhao, Min Mao, Wen Bo Liao* No short term effect of Clinostomum complanatum (Trematoda: Digenea: Clinostomatidae) on survival of Triturus carnifex (Amphibia: Urodela: Salamandridae) Giacomo Bruni1,*, Claudio Angelini2 Yeasts in amphibians are common: isolation and the first molecular characterization from Thailand Srisupaph Poonlaphdecha, Alexis Ribas* Introduction of Eleutherodactylus planirostris (Amphibia, Anura, Eleutherodactylidae) to Hong Kong Wing Ho Lee1, Michael Wai-Neng Lau2, Anthony Lau3, Ding-qi Rao4, Yik-Hei Sung5,* Correlation between endoscopic sex determination and gonad histology in pond sliders, Trachemys scripta (Reptilia: Testudines: Emydidae) David Perpiñán1, Albert Martínez-Silvestre2,3, Ferran Bargalló3, Marco Di Giuseppe4, Jorge Orós5, Taiana Costa6,* Book Review: John W. Wilkinson. Amphibian Survey and Monitoring Handbook. Pelagic Publishing Franco Andreone Book Review: Susan Newman. Frogs amphibians and their threatened environment. Discovery and expression through Art K-3. Frogs are green Franco Andreone