2017 Vol. 12 N. 1. June PRESS ISSN FIRENZE. Poste Italiane S.p.A. - Spedizione in Abbonamento Postale 70% DCB Firenze

Transcription

1 Iscritto al Tribunale di Firenze con il n 5450 del 03/11/2005 Poste Italiane S.p.A. - Spedizione in Abbonamento Postale 70% DCB Firenze June 2017 Vol. 12 N. 1 Acta Herpetologica ISSN FIRENZE UNIVERSITY PRESS

2 Acta Herpetologica Acta Herpetologica è la rivista ufficiale della Societas Herpetologica Italica (S.H.I.), un associazione scientifica che promuove la ricerca erpetologica di base e applicata, la divulgazione delle conoscenze e la protezione degli Anfibi e Rettili e dei loro habitat. Acta Herpetologica is the official journal of the Societas Herpetologica Italica (S.H.I.), a scientific association that promotes basic, applied, and conservation researches on Amphibians and Reptiles. Direttore responsabile (Editor): Marco Mangiacotti, DSTA, Università di Pavia, Via Taramelli 24, Pavia, Italia Redattori (Associate Editors): Aaron Bauer, Department of Biology, Villanova University, 800 Lancaster Avenue, Villanova, PA 19085, USA Adriana Bellati, Dipartimento di Scienze della Terra e dell Ambiente Università degli Studi di Pavia, Italy Paolo Casale, Dept. of Biology and Biotechnologies Charles, Darwin, University of Rome La Sapienza, Viale dell Università 32, I Roma, Italy Francesco Ficetola, Laboratoire d Ecologie Alpine LECA, Université Grenoble-Alpes. F Grenoble, France Ernesto Filippi, via Aurelia 18, I Ariccia, Roma, Italy Uwe Fritz, Museum of Zoology, Senckenberg Dresden, A.B. Meyer Building, Dresden, Germany Fabio Maria Guarino, Dipartimento di Biologia, Università degli Studi di Napoli Federico II, Via Cinthia, Napoli, Italy Sandra Hochscheid, Stazione Zoologica Anton Dohrn, Villa Comunale 1, I Napoli, Italy Daniele Pellitteri-Rosa, Dipartimento di Scienze della Terra e dell Ambiente, università degli Studi di Pavia, Italy Marco Sannolo, CIBIO-InBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos, da Universidade do Porto, e reside no Campus Agrário de Vairão, Vairão, Portugal Giovanni Scillitani, Dipartimento di Biologia, sezione di Biologia animale ed ambientale, Università degli studi Aldo Moro, Bari, Italy Rocco Tiberti, Parco Nazionale Gran Paradiso, Degioz 11, 1101 Valsavarenche, Aosta Consiglio direttivo S.H.I. (S.H.I. Council): Presidente (President): Roberto Sindaco Vice Presidente (Vice-President): Sandro Tripepi Segretario (Secretary): Dalila Giacobbe Tesoriere (Treasurer): Gulia Tessa Consiglieri (Council members): Gentile Francesco Ficetola, Lucio Bonato, Luciano di Tizio Sito ufficiale S.H.I. (Official S.H.I. website): Modalità di associazione Le nuove domande di associazione sono esaminate periodicamente dal Consiglio Direttivo; solo successivamente i nuovi soci riceveranno la comunicazione di accettazione con le modalità per regolarizzare l iscrizione (ulteriori informazioni sul sito: La quota annuale di iscrizione alla S.H.I. è di 35,00. I soci sono invitati a versare la quota di iscrizione sul conto corrente postale n intestato a: SHI Societas Herpetologica Italica. In alternativa è possibile effettuare un bonifico bancario sul Conto Corrente Postale: n. conto intestatario: SHI Societas Herpetologica Italica IBAN: IT-54-K Membership The S.H.I. Council will examine periodically new applications to S.H.I.: if accepted, new Members will receive confirmation and payment information (for more information contact the official website: Annual membership fee is (Euro). Payments are made on the postal account of SHI Societas Herpetologica Italica no , or by bank transfer on postal account no IBAN: IT-54-K to SHI Societas Herpetologica Italica. Versione on-line:

3 Acta Herpetologica Vol. 12, n. 1 - June 2017 Firenze University Press

4

5 Acta Herpetologica 12(1): 3-17, 2017 DOI: /Acta_Herpetol Morphological variation and sexual dimorphism in Liolaemus wiegmannii (Duméril & Bibron, 1837) (Squamata: Liolaemidae) from Uruguay Joaquín Villamil 1, *, Arley Camargo 2, Raúl Maneyro 1 1 Laboratorio de Sistemática e Historia Natural de Vertebrados, Instituto de Ecología y Ciencias Ambientales, Facultad de Ciencias, UdelaR, Iguá 4225, Montevideo, Uruguay; *Corresponding author. joakorep@gmail.com 2 Programa de Desarrollo Universitario, Centro Universitario de Rivera, UdelaR, Ituzaingó 667, Rivera, Uruguay Submitted on: 2016, 20 th April; revised on: 2016, 1 st December; accepted on: 2017, 26 th February Editor: Aaron M. Bauer Abstract. Intraspecific morphological variation is a relatively common pattern among lizards, where several selective factors have been suggested as responsible for this phenomenon. For instance, geographic variation could result from natural selection along with historical processes, whereas sexual dimorphism has usually been attributed to sexual selection, natural selection, and niche segregation. Liolaemus wiegmannii is a diurnal lizard distributed in the center, center-east and north-west of Argentina, as well as on the shores of south-west and south Uruguay. Information about morphological variation in this species is almost entirely limited to differences in mid-body scales between populations in the north and center of Argentina and some sex-based morphometric variation. Herein, we studied the geographic and sexual morphological variation of Liolaemus wiegmannii from Uruguay to test the hypothesis of morphological isolation by distance and morphological structuring by geographic barriers (rivers), as well as exploring the occurrence of sexual dimorphism in morphometry and lepidosis. Neither geographic distance nor rivers seem to play an important structuring role on the external morphology of Liolaemus wiegmannii in Uruguay. Multiple multivariate analyses support the hypothesis that most of the external morphological variation is probably due to sexual dimorphism. Natural and sexual selection acting on females and males, respectively, are the most plausible mechanisms underlying the dimorphism observed in this species. Keywords. Morphometry, lepidosis, sexual dimorphism, Liolaemus, Uruguay. INTRODUCTION Intraspecific morphological variation can arise as the result of selective pressures, phenotypic plasticity or historical factors acting at a population level. Geographic variation in morphology is quite frequent among lizards and has been attributed to both genetic and non-genetic factors (Ballinger, 1983; Dunham et al., 1988; Qualls and Shine, 1998). Intersexual variation is also common, and at least three hypotheses have been proposed in order to explain its causes in lizards: sexual selection (Maynard Smith, 1987; Carothers, 1984; Braña, 1996; Cox et al., 2003); fecundity advantage (Tinkle et al., 1970; Braña, 1996; Fairbairn, 1997; Blanckenhorn, 2005) and trophic niche segregation (Schoener, 1967; Pianka and Huey, 1978). Sexual selection as the main factor behind sexual dimorphism predicts higher reproductive success in larger males or in those that are more attractive to females. Consequently, those structures involved in male-male combat or related to female choice of male quality often become conspicuous (Carothers, 1984; Andersson, 1994; Andersson and Simmons, 2006). Alternatively, the fecun- ISSN (print) ISSN (online) Firenze University Press

6 4 J. Villamil, A. Camargo, R. Maneyro dity advantage hypothesis states that natural selection will favor larger females, and therefore, predicts a proportional increase of fecundity with body size (Tinkle et al., 1970; Kozlowski, 1989; Braña, 1996; Fairbairn, 1997; Zamudio, 1998; Cox et al., 2003; Blanckenhorn, 2005; Du et al., 2005). Otherwise, sexual dimorphism could also be a consequence of competition for trophic resources, where the resulting morphological differences between sexes allow a more efficient exploitation of the trophic niche (Schoener, 1967; Pianka and Huey, 1978; Herrel et al., 1999). With about 240 recognized species, Liolaemus is one of the most widely distributed and species rich lizard genera worldwide (Lobo et al., 2010; Etheridge and Frost, 2010; Breitman et al., 2011, 2013; Avila et al., 2013). Liolaemus is distributed exclusively in South America, occurring in Bolivia, Paraguay, Peru, Chile, Argentina, Brazil and Uruguay, spanning different environments from the Andes Mountains to the Atacama Desert and from the Pacific Ocean shores to the Atlantic Ocean coasts (Lobo, 2001; Avila, 2003; Pincheira-Donosso et al., 2008). The systematics of the genus is quite complex, with several sections, series and groups being recognized (Schulte et al., 2000; Espinoza et al., 2004; Cruz et al., 2005; Fontanella et al., 2012; Olave et al., 2014). Liolaemus wiegmannii (Duméril and Bibron, 1837) is a member of the L. wiegmannii group, a species group that has been suggested as monophyletic based on molecular and morphological data (Etheridge, 2000; Schulte et al., 2000; Espinoza et al., 2004; Avila et al., 2006; Abdala, 2007; Pincheira- Donoso et al., 2008; but see Olave et al., 2014). It is a diurnal terrestrial lizard, which occupies a variety of habitats throughout its extensive and fragmented range, although it is often found in sand dunes and sandy soils (Etheridge, 2000). Its distribution in Argentina includes the Provinces of Río Negro, Buenos Aires, La Pampa, Entre Ríos, Santa Fé, Córdoba, San Luis, Mendoza, San Juan, Catamarca, Tucumán, Jujuy and Salta, whereas in Uruguay, this species occurs in the sandy shores of the Departments of Río Negro, Soriano, Colonia, San José, Montevideo, Canelones, Maldonado, and Rocha, to the west of Valizas Creek (Cei, 1986, 1993; Carreira et al., 2005; Carreira and Maneyro, 2013; Párraga, 2011, Avila et al., 2013; Stellatelli et al., 2014). Nonetheless, numerous authors have pointed out that Liolaemus wiegmannii is probably a species complex containing several disjunct populations that may represent separate species (Avila, 2003; Morando, 2004; Avila et al., 2006; Avila et al., 2009; Aiassa and Gorla, 2010; Olave et al., 2014). Beyond its systematic complexity, the morphological variation of Liolaemus wiegmannii across its whole range has been poorly studied. Cei (1979), and Avila and Martori (1996), reported some geographic variation in the number of mid body scales from Argentina. Etheridge (2000) published data on snout-vent length, dorsal scales, supralabials, infralabials and precloacal pores. In addition, Avila et al. (2009) provided some morphological information about L. wiegmannii. However, the last two studies did not provide a geographically explicit analysis of the morphological variation. Some authors have noted sexual dichromatism in Liolaemus wiegmannii, with males exhibiting conspicuous orange and blue scales in the reproductive season that in general are absent in females, although, a tenuous orange coloration can be occasionally observed on females (Cei, 1986, 1993; Etheridge, 2000; Carreira et al., 2005; Avila et al., 2009). Recently, Cabrera et al. (2013) studied the sexual size dimorphism of several species of the Liolaemus laurentii group (sensu Abdala 2007), and showed that Liolaemus wiegmannii has sexual differences in humerus length and axila-groin distance. However, sexual dimorphism in lepidosis and other morphometric traits have not been considered yet. In addition, because L. wiegmannii is a species complex, it is possible that the individuals analyzed by Cabrera et al. (2013) from Argentina could belong to a species different to the one occurring in Uruguay. In this work, we studied the intraspecific morphological variation of Liolaemus wiegmannii throughout its range in Uruguay using meristic (lepidosis) and morphometric data. We hypothesized that the morphological variation between localities should show a geographic pattern due to isolation-by-distance. Alternatively, considering the influence of rivers on the variation of terrestrial organisms (Pound and Jackson, 1981; Gascon et al. 1998; Pellegrino et al., 2005; Geghring et al., 2012), three main rivers were considered as potential geographic barriers that could influence the structure morphological variation. Finally, we expected to find significant differences in lepidosis and other morphological traits between sexes, in addition to those sexually-dimorphic traits found in previous studies. MATERIALS AND METHODS Morphological variation of Liolaemus wiegmannii was studied through a set of eight classic morphometric and nine meristic variables (lepidosis) (Table 1) from 134 specimens housed in the vertebrate collection of the Faculty of Sciences, University of the Republic and the herpetological collection of the National Museum of Natural History (Montevideo, Uruguay) (Appendix 1). The geographic coverage of the specimens includes 40 localities spanning almost the whole range of this species in Uruguay (Fig. 1).

7 Variation and dimorphism in Liolaemus wiegmannii 5 Table 1. Morphometric and meristic variable abbreviations and their corresponding meanings. MorphometricVariables Detail Meristic Variables Detail SVL Snout-vent length MBSc Scales around mid-body HL Head length DSc Dorsal scales HW Head width VSc Ventral Scales huml Humerus length PP Precloacal pores antbl Forearm length Lam3 Subdigital lamellae of third finger FL Femur length Lam4 Subdigital lamellae of fourth toe TibL Tibia length InfLab Infralabial scales A-G Axilla-groin distance SupLab Supralabial scales LorLab Lorilabials scales Fig. 1. (a) Distribution of Liolaemus wiegmannii after Etheridge (2000) and Avila et al. (2009). (b) Localities of Uruguay used for geographic analysis with permanova. A and B: west and east of Rosario River (1) localities; C: east Santa Lucía River (2) localities; D: east Maldonado Stream (3) localities. Variables were chosen based on Etheridge (2000), Verrastro et al. (2003), and Avila et al. (2009), and the lepidosis terminology followed Smith (1946). Morphometric measurements were taken with a digital caliper to the nearest 0.01 mm, whereas for meristic variables a stereoscopic microscope was used. Sex was determined based on the shape of the cloaca, which is squaredshaped in males and rounded in females (Cabrera et al., 2013). In order to remove the effect of size from morphometric variables, each variable was transformed as Z =Y i (SVL / SVL i ) b (1) following Lleonart et al. (2000), where Z represent the transformed value of the variable Y, which is the variable affected by size, represented in this case as the snout-vent length (SVL). The exponent b is the slope of the linear regression between logy and logsvl. This transformation completely removes all the information related to size by scaling individuals to the same size and adjusting their shape to that they would have at the new size according to allometry (Lleonart et al., 2000). To check that no size effect persisted after transformation, slopes of linear regression between each transformed variable and SVL were evaluated trough a Student s t test implemented in R (R Core Team 2016) (see appendix 2). Under a successful size correction a slope of zero is expected. Principal Component Analyses were computed separately for morphometric (excluding SVL) and meristic variables through a variance-covariance matrix with the purpose of understanding the structure of the morphological variation and the contribution of each variable to the components that explain most of the variation observed. The assumption of multivariate normality was evaluated through Mardia (Mardia, 1970) and Omnibus (Dornik and Hansen, 2008) tests implemented in PAST 3.07 (Hammer et al., 2001). For testing the hypothesis of isolation by distance, a Mantel test between a morphological and geographical distance matri-

8 6 J. Villamil, A. Camargo, R. Maneyro ces was conducted. The original dataset was divided into four matrices (one for each sex and class of variable, i.e. morphometric and meristic), which were independently analyzed using Mahalanobis and Correlation distances for morphometric and meristic data respectively and geographic distances obtained from coordinates. Correlation distances were obtained with Past software as 1-r, where r is the Pearson s coefficient. To test for statistical differences between sexes and geography, morphometric and meristic variables were analyzed separately through a permanova (Anderson, 2001), which was computed based on Mahalanobis and Correlation distances respectively, and one million permutations. This analysis uses a multivariate statistic analogous to Fisher s F-ratio constructed from sums of squared distances within and between groups and provides a p value that is calculated through permutations (Anderson, 2001). For testing geographic variation independently from sex, geographic arrangements were tested separately for males and females. Three main rivers were considered as potential barriers (Rosario River, Santa Lucía River, and Maldonado Stream), and therefore localities were grouped into four groups limited by these courses (Fig. 1b). Moreover, differences between localities west and east of the Santa Lucía River, and between all localities were also tested. Sexual dimorphism was also explored through a Discriminant Function Analysis (DFA), considering morphometric and meristic measurements separately. Besides the discriminant function, Mann-Whitney and Student s t-test were performed to evaluate which variables show significant differences between sexes. All the statistical analyses, except Student s t-test on slopes, were implemented in PAST 3.07 software (Hammer et al., 2001). Descriptive statistics RESULTS Based on the descriptive statistics for each untransformed variable, there are several sexually-dimorphic variables (Table 2). Considering both sexes, individuals of L. wiegmannii from Uruguay exhibit a maximum snoutvent length of mm. Although males reached larger maximum sizes than females, both mean and median of females were higher than males (SVL male mean: mm; SVL female mean: mm; SVL male median: mm; SVL female median: mm) (Table 2). Regarding lepidosis, the ranges observed were particularly broad except for the number of lorilabial scales, which was 2 for all specimens analyzed except one (not shown in Table 2). The number of scales around the mid-body (MBSc) was similar for both sexes (males: 41 58; females: 42 58), whereas the range of the number of dorsal scales (DSc) was broader in females (males: 41 59; females: 43 65). The number of ventral scales (VSc) shows overlapping ranges between sexes, although females exhibit a higher maximum value (males: 40 59; females: 45 64). The number of precloacal pores ranged between 0 and 7 for males and between 0 and 6 for females. Although these intervals are quite similar, the median of males was 5 whereas that for females was 0. In addition, the frequency of males without precloacal pores (data not shown) was markedly lower than for females (males: 0.097; females: 0.325). The number of subdigital lamellae on both the third finger and the fourth toe had very similar intervals among sexes. The range (4-7) and the median (5) of supralabials were equal between sexes, whereas the infralabials differ in range (males: 5-8; females: 5-7), but have the same median (6). Ordination analysis Considering the transformed morphometric variables and excluding SVL, the PCA shows that 81% of the variance is comprised by the first three components with the first principal component (PCA1) accounting for 47% of the variation. On the other hand, 84% of the meristic variance is explained by the first three components with a 50% of variation accounted for in the first principal component (Table 3). For morphometric variables, the correlation coefficients show that PCA1 has a very strong positive correlation with the axilla-groin distance (Fig. 2A), whereas PCA2 is mostly positively correlated with the tibia and femur lengths, and in a lesser way, with the head length (Fig. 2B). Regarding to meristic characters, PCA1 has a high positive correlation with ventral and dorsal scales. To a lesser extent, PCA1 shows a negative correlation with the number of precloacal pores, where the magnitude of this correlation is about half of the absolute value of the maximum correlation observed in PCA1 (Fig. 2C). Moreover, PCA2 is mainly linked with the number of scales around midbody, with which it shows a very strong positive correlation (Fig. 2D). The bidimensional projection of the first two principal components shows a substantial overlap of sexes for both morphometric (Fig. 3A) and meristic data (Fig. 3B). However, from a morphometric point of view, it is possible to note that females tend to be located towards the positive values of PCA1 and negative values of PCA2, whereas males show the opposite tendency (Fig. 3A). Taking into account the variables that have the strongest correlation with these two components (Fig. 2A and B), this tendency might reflect that females have longer axilla-groin distances than males whereas males tend to have longer femur, tibia and head lengths. For meristic variables, females tend to be located toward the positive values of the PCA1 while males are principally located in the region of negative values (Fig 3B). Considering the

9 Variation and dimorphism in Liolaemus wiegmannii 7 Table 2. Descriptive statistics for morphometric and meristic variables considered for L. wiegmannii from Uruguay. Min: minimum value; Max: maximum value; Me: median; X : mean; S Ẍ : standard error of the mean; S 2 n-1 : variance; S n-1 : standard desviation. SVL: snout-vent length; HL: head length; HW: Head width; huml: humerus length; antbl: forearm length; FL: femur length; TibL: Tibial length; A-G: axillagroin distance; MBsc: scales around mid-body; DSc: dorsal scales; VSc: ventral scales: PP: precloacal pores; Lam3: subdigital lamellaeof third finger; Lam4: subdigital lamellaeof fourth toe; SupLab: supralabial scales; InfLab: infralabial scales. SVL HL HW huml antbl FL TibL A-G MBSc DSc VSc PP Lam3 Lam4 SupLab InfLab Min Max Me , X S Ẍ S 2 n S n Table 3. Principal components explaining most of the morphological variation of Liolaemus wiegmannii from Uruguay. Variables Principal Component Eigenvalue % of variance explained Morphometrics Meristic an opposite pattern: the sex with longer axilla-groin distance has a shorter head and tibia, and vice versa. For meristic variables, the DFA classified correctly 85% of the specimens with similar percentages between sexes (Table 4). In the discriminant function for these variables (DF = 0.046MBSc DSc VSc 1.40PP Lam Lam SupLab 0.029InfLab) the number of ventral scales (VSc) and the numbers of precloacal pores are the most important variables for discriminating between sexes. Again, the sex with higher values for one of these variables has lower values for the other. correlation of each variable in PCA1 (Fig. 2C), this pattern might reflect that females have higher numbers of dorsal and ventral scales than males, whereas males have more precloacal pores than females. Dispersion among PCA2 axes is particularly high, so meristic data for this component do not show a conspicuous pattern among sexes (Fig. 3B). DFA based on transformed morphometric variables correctly classified 72% of the analyzed individuals (75% of males and 71% of females, Table 6). The discriminant function obtained (DF = 0.198Zhl 0.029Zhw 0.027ZhumL 0.028ZantbL 0.12Zfl 0.24Ztl Za-g) shows that the axilla-groin distance is the most influential variable on the equation; therefore, it is the morphometric variable that best discriminates between sexes. Other less important variables for morphometric sex discrimination are tibia and head length, which show Multivariate tests Multivariate normality of the entire dataset was rejected with a 95% confidence (Table 5). According to the non-parametric multivariate analysis of variance (permanova) there are significant differences between sexes for both morphometric (F = 6.31; P < 0.001) and meristic (F = 54.85; P < 0.001) data, whereas this analysis rejected geographic differences for all arranges of localities considering main rivers as barriers and also between localities individually treated (Table 6). Mantel tests did not find significant correlations and therefore rejected the hypothesis of isolation by distance for both males (morphometric data: R = 0.09, P = 0.895; meristic data: R = 0.08, p = 0.910) and females (morphometric data: R = 0.06, P = 0.159; meristic data: R = 0.01, P = 0.557).

10 8 J. Villamil, A. Camargo, R. Maneyro Fig. 2. Correlation coefficients of each variable in the first two principal components. A: morphometrics variables in the first principal component; B: morphometric variables in the second principal component; C: meristic variables in the first principal component; D: meristic variables in the second principal component. SVL: snout-vent length; HL: head length; HW: Head width; huml: humerus length; antbl: forearm length; FL: femur length; TibL: Tibial length; A-G: axilla-groin distance; MBsc: scales around mid-body; DSc: dorsal scales; VSc: ventral scales: PP: precloacal pores; Lam3: subdigital lamellae of third finger; Lam4: subdigital lamellae of fourth toe; SupLab: supralabial scales; InfLab: infralabial scales. Univariate tests Univariate tests (Mann-Whitney or t-test) show that those variables that are most influential for discriminating between sexes according to the DFA, also have highly significant sexual differences (P < and P < 0.001) (i.e., axilla-groin distance; tibia length; head length; number of ventral scales, and number of precloacal pores) (Table 7). In addition, there are also significant differences (P < 0.05) in snout-vent length (SVL) and number of subdigital lamellae of the fourth toe (Lam4). However, it is possible that an asymmetric distribution of juveniles and adults among sexes could be influencing this result for SVL. Unfortunately, given that data about minimum size at sexual maturity are only available for females (Ramirez Pinilla 1991; Martori and Aun 1997), the limit between juveniles and adults cannot be established for both sexes.

11 Variation and dimorphism in Liolaemus wiegmannii 9 Table 4. Confusion matrix of the Discriminant Function Analysis for morphometric and meristic variables of Liolaemus wiegmannii from Uruguay. Values in the third and fourth column represent the number of individual assigned to each sex by the discriminant function. Male Female Total Correctly assigned (%) Morphometric variables Meristic variables Male Female Total Male Female Total Table 5. Multivariate normality test results for the morphology of Liolaemus wiegmannii from Uruguay. All analyses were implemented in Past Test Parameter Coefficient Statistic d.f. P(normal) Mardia Skewness x10-15 Dornik & Hansen omnibus Kurtosis x x10-13 Fig. 3. Bidimensional projection of the morphological data of Liolaemus wiegmannii in the first two principal components. A: morphometric variables; B: meristic variables. Squares represent males whereas black full circles are females. DISCUSSION Intraspecific and geographic variation Table 6. permanova results for morphometric and meristic data of Liolaemus wiegmannii from Uruguay considering three levels of variation. A+B/C+D tests differences between west and east Santa Lucía River localities. A/B/C/D evaluates differences between all the groups of localities divided by Rosario River, Santa Lucía River and Maldonado Stream and Localities considers the hypothesis of differences between each locality. Groups tested Data F P A+B/C+D Morphometric Meristic A/B/C/D Morphometric Meristic Localities Morphometric Meristic Considering both males and females, specimens of Liolaemus wiegmannii from Uruguay reach a maximum snout-vent length of 61 mm with females tending to be larger than males. This observed size is in the range of those reported in the literature for individuals from Argentina (Etheridge 2000; Avila et al. 2009). Cei (1979) and Avila and Martori (1996) reported geographic variation in size and number of scales around mid-body in individuals from Argentina. The highest values were observed in Bahía Blanca (56-60), whereas the lowest were found in Tucumán (46-48), while Mendoza (48-50) had intermediate ranges. The range of this variable in individuals from Uruguay (41-58) is markedly broader than that observed in each one of these populations, being comparable with the whole interval reported for this species (Etheridge 2000; Avila et al. 2009). Unlike the number of scales around mid-body, information about other meristic variables in literature is scarce

12 10 J. Villamil, A. Camargo, R. Maneyro Table 7. Morphological differences of L. wiegmannii between sexes, tested for each variable through a Mann-Whitney or t test, depending on the normality of each variable according to Shapiro-Wilk test. Z indicates transformed variables following the method proposed by Lleonart et al. (2000). Variable Shapiro-Wilk Test Mann-Whitney Test t Test SVL* W=0.95; P= U=2381; P=0.039 Z-HL** W=0.98; P=0.015 U=2198; P= Z-HW W=0.99; P=0.55 t=0.49; P=0.62 Z-humL W=0.99; P=0.18 t=0.36; P=0.72 Z-antbL W=0.99; P=0.78 t=0.48; P=0.64 Z-FL W=0.98; P=0.017 U=2598; P=0.20 Z-TibL** W=0.99; P=0.39 t=2.99; P= Z-A-G*** W=0.99; P=0.65 t=-6.72; P=3.45x10-10 MBSc W=0.97; P= U=2888; P=0.82 DSc W=0.97; P= U=2513; P=0.11 VSc*** W=0.99; P=0.40 t=-7.98; P=3.30x10-13 PP*** W=0.81; P=8.24x10-13 U=1280; P=2.95x10-10 Lam3 W=0.86; P=1.12x10-10 U=2817; P=0.60 Lam4* W=0.95; P=5.52x10-5 U=2339; P=0.023 SupLab W=0.78 ; P=6.23x10-14 U=2727; P=0.36 InfLab W=0.78; P=7.96x10-14 U=2866; P=0.72 P<0.05*; P<0.01**; P<0.001***. and without specific geographic information, hampering the comparison with our data. Nevertheless, it can be observed that lepidosis in Uruguay shows broad ranges for almost all variables that overlap with the known variation across the whole distribution of the species (Etheridge, 2000). Most of the morphological variation observed in Uruguay seems to be related to differences between sexes because most variables highly correlated with PCA1 and PCA2 are the same that statistically discriminate sexes. However, other variables are also highly correlated with the first two principal components (Fig. 2), but do not show significant differences between sexes (Table 7), suggesting that there is some morphological variation that is not explained by sexual dimorphism and accounts for the overlap seen in PCA space (Fig. 3). Nevertheless, this additional variation could not be attributed to geographic variability, given that morphological differences among groups of localities were rejected and there was no correlation between morphological and geographic distances. The absence of geographically-structured morphological variation in Uruguay contrasts with the differences in mid body scales found by Cei (1979) and Avila and Martori (1996) among several isolated populations in Argentina, which could actually represent separate species (Avila, 2003; Morando, 2004; Avila et al., 2006; Avila et al., 2009; Olave et al., 2014). Sexual dimorphism Sexual dimorphism in L. wiegmannii from Uruguay is strongly supported by several independent multi- and univariate analyses, which show that males have a longer head and tibia, and present a higher number of precloacal pores than females. On the other hand, females have a longer axilla-groin distance and exhibit a higher number of ventral scales. Based on seven morphometric variables, Cabrera et al. (2013) found that males and females differ only in axilla-groin distance and humerus length. Unlike the axilla-groin distance, the humerus length is not sexually dimorphic in specimens from Uruguay. Indeed, the humerus length is one of the variables with the lowest correlation coefficient with PCA1 and PCA2, and the least influential on the morphometric discriminant function. These discrepancies might suggest differences on the distribution of morphological variation among populations from Argentina and Uruguay. Among Liolaemus, head dimensions and axilla-groin distance are sexually dimorphic in most of the species analyzed, suggesting that it is a relatively common pattern among these lizards (Villavicencio et al., 2003; Verrastro, 2004; Laspiur and Acosta, 2007; Cabrera et al., 2013; Astudillo et al., 2015). Taking into account the main hypotheses regarding the causes of sexual dimorphism in lizards, sexual selection might adequately explain the differences in head and

13 Variation and dimorphism in Liolaemus wiegmannii 11 tibia length and the number of precloacal pores found in this work. Longer heads in males of L. wiegmannii could suggest the existence of agonistic encounters between them, where individuals with longer heads would have an advantage in combats getting greater access to females (Carpenter and Fergusson, 1977; Carothers, 1984; Anderson and Vitt, 1990; Herrel et al., 1999; Verrastro, 2004; Huyghe et al., 2005; Vanhooydonck et al., 2010). In this sense, although scarcely documented, some male-male agonistic encounters have been observed among Liolaemus species (Halloy, 1996; Martins et al., 2004; Verrastro, 2004; Labra et al., 2007; Cabrera et al., 2013; Halloy et al., 2013). Additionally, males of some species of Liolaemus show hierarchical structure according to the size of their home range, suggesting the occurrence of male-male interactions, and therefore a potential evidence of sexual selection (Halloy and Robles, 2002; Frutos and Belver, 2007; Robles and Halloy, 2009; Cabrera et al., 2013). Moreover, Martins et al. (2004) has pointed out that the head and limbs are structures that play an important role in communication in Liolaemus lizards, and consequently, it is possible that sexual differences in these traits could be related to their use in this context. Head-bobbing seems to be a common behavior for signaling among Liolaemus males, both in aggressive interactions and courtship (Martins et al., 2004; Labra et al., 2007; Halloy, 2012; Halloy et al., 2013; Vicente and Halloy, 2015). In addition, several species displaying these behaviors, also show sexual dimorphism in head size, where males are always the sex with larger heads (Villavicencio et al., 2003; Laspiur and Acosta, 2007; Cabrera et al., 2013; Astudillo et al., 2015). This might suggest, at least for some species, a possible relationship between head-bobs and sexual dimorphism in head size that deserve further study. Although little is known about the behavior of L. wiegmannii, Achaval and Olmos (2007) have mentioned that head movements are observed during courtship, suggesting the possibility that head size differences could be related to its use in courtship or for warning other males during territory defense. Alternatively, sexual dimorphism in head length observed in L. wiegmannii could also be the result of trophic niche segregation. However, Vanhooydonck et al. (2010) pointed out that bite force and consequently head dimension differences in male Liolaemus lizards is better explained by sexual selection rather than natural selection (i.e., niche segregation). Therefore, considering the phylogenetic context, head length sexual dimorphism found in L. wiegmannii is more likely to be the result of sexual selection than trophic niche segregation. Even so, given that information about the intersexual diet variation in L. wiegmannii remains scarce (see Aun et al., 1999), the role of niche segregation on sexual dimorphism still demands further research. Several lizard species show a strong positive correlation between hindlimb length and sprint speed (Snell et al., 1988; Losos, 1990; Miles, 1994; Bauwens et al., 1995; Bonine and Garland, 1999). Additionally, some studies have suggested that variation in sprinting ability can affect survival probabilities within populations of reptiles (Christian and Tracy, 1981; Jayne and Bennett, 1990; Miles, 2004, Vervust et al., 2007). In this sense, Bauwens et al. (1995) pointed out that the evolution of longer hindlimbs relative to body size is one of the main factors driving the evolution of high maximum sprinting ability in lizards. In this context, considering the notorious sexual differences in coloration observed in L. wiegmannii, where males are clearly less cryptic than females and thus, probably more detectable, it is reasonable to think that predation might play a more important role as a selection pressure in males, favoring faster runners, and therefore longer tibia. Alternatively, it should be taken into account that sprint speed could be also important for a more effective territory defense, where faster males are expected to defend a larger territory and/or more females through exclusion of slower rival males that still can usurp mates (Husak et al., 2006, 2008; Peterson and Husak, 2006). Consequently, if territorial defense occurs in Liolaemus wiegmannii, as in other species of the genus, and if the longer tibia (and thus, hindlimb) of males provide them an advantage for faster movements around territory boundaries, sexual selection could also have favored a longer tibia in males. On the other hand, given that among Liolaemus only the forelimbs are known to be implicated in communication (Martins et al., 2004; Halloy and Castillo, 2006; Halloy, 2012), it seems unlikely that differences in hindlimb dimensions are related to a communication use in this species. Precloacal pores allow for the external exudation of chemical secretions by integumentary glands (Valdecantos et al., 2014), which in Liolaemus species have been suggested as important for stimulating copulation (Rocha 1996). Other authors think that these secretions play a role in territorial defense and recognition contexts similar to those in which head-bob displays are observed (Labra and Niemeyer, 1999; Martins et al., 2004). In addition, the use of chemical secretions by some male lizards allows them to reduce the cost of territory defense (Labra and Niemeyer, 1999 and references therein). In this scenario, the presence of more and bigger precloacal pores in males of Liolaemus wiegmannii could suggest that these might play a more important role in males than females, for instance, for territorial defense. If this is the case, it is possible that sexual selection could be underly-

14 12 J. Villamil, A. Camargo, R. Maneyro ing the observed dimorphism in the number of precloacal pores. Moreover, the presence of precloacal pores in both male and female individuals suggests that the role of these glands on communication is not only for male territorial defense. It is known that chemical recognition of females during reproductive season occurs in Liolaemus tenuis, a species with precloacal pores present in both sexes (Labra and Niemeyer, 1999). Therefore, although the dimorphism in the number of precloacal pores found in this study may have evolved for male-male defense, we cannot discard their role in male-female communication, or even both roles as the drivers of sexual selection. In many lizard species, females are under fecundity selection for larger abdomen size, suggesting that this is probably the most frequent mechanism underlying the widespread pattern of sexual dimorphism in axilla-groin distance (Scharf and Meiri, 2013). An increase in axillagroin distance, and therefore in abdomen size, provides to females the possibility of enlarging the space for storing eggs and, consequently, increasing their fecundity, which easily becomes a target of natural selection (Tinkle et al., 1970; Kozlowski, 1989; Braña, 1996; Fairbairn, 1997; Zamudio, 1998; Cox et al., 2003; Blanckenhorn, 2005; Du et al., 2005). Based on the fecundity advantage hypothesis, it is reasonable to think that the larger abdomen found in females of L. wiegmannii is a result from fecundity selection. In addition, it is possible that the enlargement of the female s abdomen could have led to an increase in the number of ventral scales observed in this sex. However, there is a reduction in the size of the ventral scales of females toward the cloacae, which might also contribute to the sexual differences found. Finally, it is important to see that although sexual dimorphism seem to be a widespread pattern among lizards, where head (larger in males) and abdomen (larger in females) are the most frequent traits that differ between sexes, studies that test explicitly the hypotheses behind this phenomenon remain scarce (Scharf and Meiri, 2013). In this sense, the potential mechanisms proposed here for explaining the occurrence of sexual dimorphism in Liolaemus wiegmannii, should be taken as tentative. Further research is needed to explore and test explicitly the role of fecundity, behavior, and niche segregation on the sexual dimorphism observed on this species. ACKNOWLEDGEMENTS We are grateful to the research and innovation agency of Uruguay (ANII) for financial support through the grant FCE We also acknowledge financial support from National System of Investigators (SNI-ANII) and PEDECIBA. We thank Diego Arrieta and Melitta Meneghel for allowing us to access to the collections of the National Museum of Natural History (MNHN) and the Faculty of Sciences (University of the Republic), respectively. REFERENCES Abdala, C.S. (2007): Phylogeny of the boulengeri group (Iguania: Liolaemidae, Liolaemus) based on morphological and molecular characters. Zootaxa 1538: Achaval, F., Olmos, A. (2007): Anfibios y Reptiles del Uruguay, Tercera Edición. Biophoto, Montevideo. Aiassa, D., Gorla, N. (2010): Cariotipos de Liolaemus quilmes y Liolaemus wiegmannii y comparación con otros taxones del grupo boulengeri. Multequina 19: Anderson, M.J. (2001): A new method for non parametric multivariate analysis of variance. Austral Ecol. 26: Anderson, R.A, Vitt, L.J. (1990): Sexual selection versus alternative causes of sexual dimorphism in teiid lizards. Oecologia 84: Andersson, M. (1994): Sexual Selection. Princeton University Press, Princeton. Andersson, M., Simmons, L.W. (2006): Sexual selection and mate choice. Trends Ecol. Evol. 21: Astudillo, G.V., Acosta, J.C., Villavicencio, H.J., Córdoba, M.A. (2015): Ecología trófica y dimorfismo sexual del lagarto endémico Liolaemus eleodori (Iguania: Liolaemidae) del Parque Nacional San Guillermo. San Juan. Cuad. Herpetol. 29: Aun, L., Martori, R., Rocha, C. (1999): Variación estacional de la dieta de Liolaemus wiegmannii (Squamata: Tropidiiridae) en un agroecosistema del sur de Córdoba, Argentina. Cuad. Herpetol. 13: Avila, L.J. (2003): A new species of Liolaemus (Squamata: Liolaemidae) from northeastern Argentina and southern Paraguay. Herpetologica 59: Avila, L.J., Martori, R. (1996): Variación geográfica de Liolaemus wiegmannii (Duméril & Bibron) (Sauria: Tropiduridae) en Argentina. Resúmenes del IV Congreso Latinoamericano de Herpetología, Santiago, Chile. Avila, L.J., Morando, M., Sites, J.W. Jr. (2006): Congeneric phylogeography: hypothesizing species limits and evolutionary processes in Patagonian lizards of the Liolaemus boulengeri group (Squamata: Liolaemini). Biol. J. Linn. Soc. 89: Avila, L.J., Morando, M., Perez, D.R., Sites, J.W. Jr. (2009): A new species of Liolaemus from Añelo sand dunes,

15 Variation and dimorphism in Liolaemus wiegmannii 13 northern Patagonia, Neuquén, Argentina, and molecular phylogenetic relationships of the Liolaemus wiegmannii species group (Squamata. Iguania. Liolaemini). Zootaxa 2234: Avila, L.J., Martínez, L.E., Morando, M. (2013): Checklist of lizards and amphisbaenians of Argentina: an update. Zootaxa 3616: Ballinger, R.E. (1983): Life-history variations. In: Lizard Ecology: Studies of a Model Organism, pp Huey, R.B., Pianka, E.R., Schoener, T.W., Eds. Harvard University Press, Cambridge. USA. Bauwens, D., Garland, T. Jr., Castilla, A.M., Van Damme, R. (1995): Evolution of sprint speed in lacertid lizards: morphological, physiological and behavioral covariation. Evolution 1995: Blanckenhorn, W.U. (2005): Behavioral causes and consequences of sexual size dimorphism. Ethology 111: Bonine, K.E., Garland, T. Jr. (1999): Sprint performance of phrynosomatid lizards, measured on a high-speed treadmill, correlates with hindlimb length. J. Zool. 248: Braña, F. (1996): Sexual dimorphism in lacertid lizards: male head increase vs female abdomen increase? Oikos 75: Breitman, M.F., Avila, L.J., Sites, J.W. Jr, Morando, M. (2011): Lizards from the end of the world: Phylogenetic relationships of the Liolaemus lineomaculatus section (Squamata: Iguania: Liolaemini). Mol. Phylogenet. Evol. 59: Breitman, M.F., Morando, M., Avila, L.J. (2013): Past and present taxonomy of the Liolaemus lineomaculatus section (Liolaemidae): is the morphological arrangement hypothesis valid? Zool. J. Linn. Soc. 168: Cabrera, M.P., Scrocchi, G.J., Cruz, F.B. (2013): Sexual size dimorphism and allometry in Liolaemus of the L. laurenti group (Sauria: Liolaemidae): Morphologic lability in a clade of lizards with different reproductive modes. Zool. Anz. 252: Carothers, J.H. (1984): Sexual selection and sexual dimorphism in some herbivorous lizards. Am. Nat. 1984: Carpenter, C.C., Ferguson, G.W. (1977): Variation and evolution of stereotyped behavior in reptiles. In: Biology of the Reptilia, 7, pp Gans, C., Tinkle, D.W., Eds. Academic Press, London, New York and San Francisco. Carreira, S., Maneyro, R. (2013): Guía de Reptiles del Uruguay. Ediciones de la Fuga, Montevideo. Carreira, S., Meneghel, M., Achaval, F. (2005): Reptiles de Uruguay. DI.R.A.C, Facultad de Ciencias, Universidad de la República, Montevideo. Cei, J.M. (1979): Nota preliminar sobre la distribución de Liolaemus wiegmannii (Dumeril & Bibron) (Sauria: Iguanidae). Publ. Ocas. Inst. Biol. Animal. 14: 1-4. Cei, J.M. (1986): Reptiles del centro, centro-oeste y sur de la Argentina. Herpetofauna de las zonas áridas y semiáridas. Monogr. 4. Mus. Reg. Sci. Nat., Torino. Cei, J.M. (1993): Reptiles del noroeste, nordeste y este de la Argentina, Herpetofauna de las selvas subtropicales, Puna y Pampas. Monogr. 14. Mus. Reg. Sci. Nat., Torino. Christian, K.A., Tracy, C.R. (1981): The effect of the thermal environment on the ability of hatchling Galapagos land iguanas to avoid predation during dispersal. Oecologia 49: Cox, R., Skelly, M.S.L., John Alder, H.B. (2003): A comparative test of adaptive hypotheses for sexual size dimorphism in lizards. Evolution 57: Cruz, F.B., Fitzgerald, L.A., Espinoza, R.E., Schulte, J.A. (2005): The importance of phylogenetic scale in tests of Bergmann s and Rapoport s rules: lessons from a clade of South American lizards. J. Evol. Biol. 18: Du, W., Ji, X., Shine, R. (2005): Does body volume constrain reproductive output in lizards? Biol. Lett. 22: Doornik, J.A., Hansen, H. (2008): An omnibus test for univariate and multivariate normality. Oxford Bull. Econ. Stat. 70: Dunham, A.E., Miles, D.B., Reznick, D.N. (1988): Life history patterns in squamate reptiles. In: Biology of the Reptilia, vol. 16, pp Gans, C., Huey, R.B., Eds. AR Liss, New York, USA. Espinoza, R.E., Wiens, J.J., Tracy, C.R. (2004): Recurrent evolution of herbivory in small, cold-climate lizards: breaking the ecophysiological rules of reptilian herbivory. Proc. Natl. Acad. Sci. U.S.A. 101: Etheridge, R. (2000): A review of lizards of the Liolaemus wiegmannii group (Squamata: Iguania: Tropiduridae) and a history of morphological change in the sanddwelling species. Herpetol. Monogr. 14: Etheridge, R., Frost, D.R. (2010): Catalogues of the pleurodont iguanian families: Liolaemidae. Version 1 (May 19, 2010). Available on: < > Fairbairn, D.J. (1997): Allometry for sexual size dimorphism: pattern and process in the coevolution of body size in males and females. Annu. Rev. Ecol. Syst. 28: Fontanella, F.M., Olave, M., Avila, L.J., Morando, M. (2012): Molecular dating and diversification of the South American lizard genus Liolaemus (subgenus

16 14 J. Villamil, A. Camargo, R. Maneyro Eulaemus) based on nuclear and mitochondrial DNA sequences. Zool. J. Linn. Soc. 164: Frutos, N., Belver, L.C. (2007): Dominio vital de Liolaemus koslowskyi Etheridge (Iguania: Liolaemini) en el Noroeste de la provincia de La Rioja. Argentina. Cuad. Herpetol. 21: Gehring, P.S., Pabijan, M., Randrianirina, J.E., Glaw, F., & Vences, M. (2012): The influence of riverine barriers on phylogeographic patterns of Malagasy reed frogs (Heterixalus). Mol. Phylogenet. Evol. 64: Gascon, C., Lougheed, S.C., Bogart, J.P. (1998): Patterns of genetic population differentiation in four species of Amazonian frogs: a test of the riverine barrier hypothesis. Biotropica 30: Halloy, M. (1996): Behavioral patterns in Liolaemus quilmes (Tropiduridae), a South American lizard. Bull. Md. Herpetol. Soc. 32: Halloy, M. (2012): Visual display variations in neotropical lizards, Liolaemus quilmes (Iguania: Liolaemidae): relation to sex and season. Herpetol. J. 22: Halloy, M., Robles, C. (2002): Spatial distribution in a neotropical lizard, Liolaemus quilmes (Liolaemidae): site fidelity and overlapping among males and females. Bull. Maryland. Herpetol. Soc. 38: Halloy, M., Robles, C., Salica, M.J., Semhan, R., Juárez Heredia V., Vicente N. (2013): Estudios de comportamiento y ecología de lagartijas de los géneros Liolaemus y Phymaturus (Iguania: Liolaemini). Cuad. Herpetol. 27: Halloy, M., Castillo, M. (2006): Forelimb wave displays in lizard species of the genus Liolaemus (Iguania: Liolaemidae). Herpetol. Nat. Hist. 9: Hammer, Ø., Harper, D.A.T., Ryan, P.D. (2001): PAST: Paleontological statistics software package for education and data analysis. Palaeontol. Electron. 4: issue1_01.htm Herrel, A., Spithoven, L., Van Damme, R., De Vree, F. (1999): Sexual dimorphism of head size in Gallotia galloti: testing the niche divergence hypothesis by functional analyses. Funct. Ecol. 13: Husak, J.F., Fox, S.F., Lovern, M.B., Van Den Bussche, R.A. (2006): Faster lizards sire more offspring: sexual selection on whole-animal performance. Evolution 60: Husak, J.F., Fox, S.F., Van Den Bussche, R.A. (2008): Faster male lizards are better defenders not sneakers. Anim. Behav. 75: Huyghe, K., Vanhooydonck B., Scheers, H., Molina Borja, M., Van Damme, R. (2005): Morphology, performance and fighting capacity in male lizards, Gallotia galloti. Funct. Ecol. 19: Jayne, B.C., Bennett, A.F. (1990): Selection on locomotor performance capacity in a natural population of garter snakes. Evolution 1990: Kozlowski, J. (1989): Sexual size dimorphism: a life history perspective. Oikos 1989: Labra, A., Niemeyer, H.M. (1999): Intraspecific chemical recognition in the lizard Liolaemus tenuis (Tropiduridae). J. Chem. Ecol. 25: Labra, A., Carazo, P., Desfilis, E., Font, E. (2007): Agonistic interactions in a Liolaemus lizard: structure of head bob displays. Herpetologica 63: Laspiur, A., Acosta, J.C. (2007): Dimorfismo sexual de Liolaemus cuyanus Cei & Scolaro, 1980 (Iguania: Liolaemidae) en una población de San Juan, Argentina. Rev. Per. Biol. 14: Lleonart, J., Salat, J., Torres, G.J. (2000): Removing allometric effects of body size in morphological analysis. J. Theoret. Biol. 205: Lobo, F. (2001): A phylogenetic analysis of lizards of the Liolaemus chiliensis group (Iguania: Tropiduridae). Herpetol. J. 11: Lobo, F., Espinoza, R.E., Quinteros S. (2010): A critical review and systematic discussion of recent classification proposals for liolaemid lizards. Zootaxa 2549: Losos, J.B. (1990): The evolution of form and function: morphology and locomotor performance in West Indian Anolis lizards. Evolution 44: Mardia, K.V. (1970): Measures of multivariate skewness and kurtosis with applications. Biometrika 57: Martins, E.P., Labra, A., Halloy, M., Thompson, J.T. (2004): Large-scale patterns of signal evolution: an interspecific study of Liolaemus lizard headbob displays. Anim. Behav. 68: Martori, R., Aun, L. (1997). Reproductive and fat body cycle of Liolaemus wiegmannii in Central Argentina. J Herpetol. 31: Maynard Smith, J. (1987): Sexual selection: a classification of models. In: Sexual Selection: Testing the Alternatives, pp Bradbury J.W., Andersson, M., Eds, J. Wiley & Sons, Chinchester. Miles, D.B. (1994): Covariation between morphology and locomotor performance in sceloporine lizards. In: Lizard Ecology: Historical and Experimental Perspectives, p Vitt L.J., Pianka E.R., Eds, Princeton University Press, Princeton. Miles, D.B. (2004): The race goes to the swift: fitness consequences of variation in sprint performance in juvenile lizards. Evol. Ecol. Res. 6: Morando, M. (2004): Sistemática y filogenia de grupos de especies de los géneros Phymaturus y Liolaemus (Squamata: Tropiduridae: Liolaeminae) del oeste y

17 Variation and dimorphism in Liolaemus wiegmannii 15 sur de Argentina. Unpublished doctoral dissertation. Universidad Nacional de Tucumán, San Miguel de Tucumán, Argentina. Olave, M., Avila, L.J., Sites, J.W. Jr., Morando, M. (2014): Multilocus phylogeny of the widely distributed South American lizard clade Eulaemus (Liolaemini. Liolaemus). Zool. Scripta 252: Párraga, M.R. (2011): Nueva localidad para Liolaemus wiegmannii (Duméril & Bribon. 1837) en la Provincia de Salta (Argentina). Cuad. Herpetol. 25: Pellegrino, K., Rodrigues, M.T., Waite, A.N., Morando, M., Yassuda, Y.Y., Sites, J.W. Jr. (2005): Phylogeography and species limits in the Gymnodactylus darwinii complex (Gekkonidae, Squamata): genetic structure coincides with river systems in the Brazilian Atlantic Forest. Biol. J. Linn. Soc. 85: Peterson, C.C., Husak, J.F. (2006): Locomotor performance and sexual selection: individual variation in sprint speed of collared lizards (Crotaphytus collaris). Copeia 2006: Pianka, E.R., Huey, R.B. (1978): Comparative ecology, niche segregation, and resource utilization among gekkonid lizards in the southern Kalahari. Copeia 1978: Pincheira-Donoso, D., Scolaro, J.A., Sura, P. (2008): A monographic catalogue on the systematics and phylogeny of the South American iguanian lizard family Liolaemidae (Squamata. Iguania). Zootaxa 1800: Pounds, J.A., Jackson, J.F. (1981): Riverine barriers to gene flow and the differentiation of fence lizard populations. Evolution 1981: Qualls, F.J., Shine, R. (1998): Geographic variation in lizard phenotypes: importance of the incubation environment. Biol. J. Linn. Soc. 64: R Core Team (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL Ramirez Pinilla, M. P. R. (1991). Reproductive and fat body cycle of the lizard Liolaemus wiegmanni. Amphibia-Reptilia 12: Robles, C., Halloy, M. (2009): Home Ranges and reproductive strategies in a neotropical lizard. Liolaemus quilmes (Iguania: Liolaemidae). S, Amer. J. Herpetol. 4: Rocha, C.F.D. (1996): Sexual dimorphism in the sand lizard Liolaemus lutzae of southeastern Brazil. In: Herpetologia Neotropical 2, p Péfaur J. E., Ed, Universidad de los Andes, Conselho de Publicaciones, Mérida, Venezuela. Scharf, I., Meiri, S. (2013): Sexual dimorphism of heads and abdomens: different approaches to being large in female and male lizards. Biol. J. Linn. Soc. 110: Schoener, T.W. (1967): The ecological significance of sexual dimorphism in size in the lizard Anolis conspersus. Science 155: Schulte, J.A., Macey, R.J., Espinoza, R.E., Larson, A. (2000): Phylogenetic relationships in the iguanid lizard genus Liolaemus: multiple origins of viviparous reproduction and evidence for recurring Andean vicariance and dispersal. Biol. J. Linn. Soc. 69: Smith, H.M. (1946): Handbook of lizards: Lizards of the United States and of Canada. Cornell University Press, Ithaca. Snell, H.L., Jennings, R.D., Snell H.M., Harcourt, S. (1988): Intrapopulation variation in predator-avoidance performance of Galapagos lava lizards: the interaction of sexual and natural selection. Evol. Ecol. 2: Stellatelli, O.A., Bo M.S., Madrid, E., Vega L.E., Block, C. (2014): Nueva localidad para Liolaemus wiegmannii (Duméril & Bibrón. 1837) en la Provincia de Río Negro (Argentina). Cuad. Herpetol. 28: Tinkle, D.W., Wilbur H.M., Tilley, S.G. (1970): Evolutionary strategies in lizard reproduction. Evolution 1970: Valdecantos, S., Martínez, V., Labra, A. (2014): Comparative morphology of Liolaemus lizards precloacal glands. Acta Herpetol. 9: Vanhooydonck, B., Cruz, F.B., Abdala, C.S., Azócar, D.L.M., Bonino, M.F., Herrel, A. (2010): Sex specific evolution of bite performance in Liolaemus lizards (Iguania: Liolaemidae): the battle of the sexes. Biol. J. Linn. Soc. 101: Vervust, B., Grbac, I., Van Damme, R. (2007): Differences in morphology, performance and behaviour between recently diverged populations of Podarcis sicula mirror differences in predation pressure. Oikos 116: Verrastro, L. (2004): Sexual dimorphism in Liolaemus occipitalis (Iguania. Tropiduridae). Iheringia. Sér. Zool. 94: Verrastro, L., Veronese, L., Bujes, C., Dias Filho, M.M. (2003): A new species of Liolaemus from southern Brazil (Iguania: Tropiduridae). Herpetologica 59: Vicente, N., Halloy, M. (2015): Male headbob display structure in a neotropical lizard. Liolaemus pacha (Iguania: Liolaemidae): relation to social context. Herpetol. J. 25: Villavicencio, H.J., Acosta, J.C., Cánovas, M.G., Marinero, J.A. (2003): Dimorfismo sexual de Liolaemus pseudoanomalus (Iguania: Liolaemidae) en el centro-oeste de Argentina. Rev. Esp. Herpetol. 17: Zamudio, K.R. (1998): The evolution of female-biased sexual size dimorphism: a population-level comparative study in horned lizards (Phrynosoma). Evolution 1998:

18 16 J. Villamil, A. Camargo, R. Maneyro APPENDIX 1 Individuals of Liolaemus wiegmannii used in the analysis and their respective localities from Uruguay. ZVC: vertebrate collection of the Faculty of Sciences, University of the Republic; MNHN: collection of the Museum of Natural History, Uruguay. *Excluded from geographical analysis. Locality Individuals Arazatí, San José ZVC 303, ZVC 324, ZVC , ZVC 1223, ZVC 6876, ZVC 6877, MNHN 5838, MNHN 5839 Artilleros, Colonia MNHN 3346, MNHN 3348 Balneario Argentino, Canelones ZVC 2498 Bello Horizonte, Canelones MNHN 5662 Boca del Cufré, San José ZVC 3761 Boca del Mauricio, San José ZVC 6869, ZVC 6870, ZVC Boca del San Salvador, Soriano MNHN 154 Brisas del Plata, Colonia ZVC 2105 ZVC 1499, ZVC 1863, ZVC 2504, ZVC 3387, ZVC 1839, ZVC 1932, ZVC 1497, ZVC 1497, ZVC 1864, Cabo Polonio, Rocha ZVC 646, ZVC 1224, ZVC 6621, ZVC 6625, ZVC 6626, ZVC , MNHN 3425, MNHN 3427, MNHN 5671 Carrasco Stream, Canelones ZVC 1939 Carrasco, Montevideo ZVC 3549, ZVC 5153, MNHN 149, MNHN 165, MNHN 3304 Coast of Rio Negro River, in ZVC 774* front of Villa Soriano, Soriano Cuchilla Alta, Canelones MNHN 3361 El Pinar, Canelones ZVC 5050, ZVC 6859 La Floresta, Canelones MNHN 168 La Paloma, Rocha ZVC 914, MNHN 3127 Lagomar, Canelones ZVC 6023, ZVC 6227, ZVC 3712, ZVC 2824, ZVC 2825, ZVC 3711, ZVC 3764 Laguna de Garzón, Rocha MNHN 157, MNHN 3338, MNHN 3339 Laguna de Rocha ZVC 5366 Laguna del Diario MNHN 3344 Las Cañas, Río Negro MNHN 3308* Las Vegas, Canelones ZVC 6860, ZVC 6861 Lomas de Carmelo, Colonia ZVC Los Titanes, Canelones ZVC 2435 Malvin, Montevideo MNHN 1064, MNHN 1065, MNHN 3349, ZVC 595, ZVC 876 ZVC 585 Manantiales, Maldonado ZVC 2442 Médanos de Solymar, Canelones ZVC 3570 Nueva Palmira, Colonia ZVC 1967, ZVC 1968, ZVC 1969 Pajas Blancas, Montevideo ZVC 4357, ZVC 4361, ZVC 4362, ZVC Pando, Canelones MNHN 196, MNHN 199 Pinamar, Canelones ZVC 1342 Playa Pascual, San José MNHN 2445, MNHN 3310, MNHN 3311, MNHN 3312, MNHN 3313, MNHN 3314, MNHN 3337, ZVC 5152, ZVC , ZVC 6871 Portezuelo, Maldonado MNHN 176 Punta Espinillo, Montevideo ZVC Punta Negra, Maldonado MNHN 266, MNHN 3306, MNHN 3340, MNHN 3342, MNHN 3343, ZVC 6308 Rio de la Plata, 3 km east of Martin Chico, Colonia ZVC 2164 San Gregorio Stream, San José ZVC 970 San Luis, Canelones ZVC 3307 Tigre Stream, San José ZVC West side of Cerro de Montevideo, Montevideo ZVC 1385

19 Variation and dimorphism in Liolaemus wiegmannii 17 APPENDIX 2 Results of the Student s t-test on the slopes of linear regressions between each transformed variable and snout-vent length (SVL). Z indicates transformed variables; hl: head length, hw: head width, huml: humerus length, antbl: forearm length, fl: femur length, tibl: tibia length, a-g: axilla-groin distance. Zhl vs SVL Zhw vs SVL ZhumL vs SVL ZantbL vs SVL Zfl vs SVL Ztibl vs SVL Za-g vs SVL Slope x x t value p(slope=0)

20

21 Acta Herpetologica 12(1): 19-27, 2017 DOI: /Acta_Herpetol Sex does not affect tail autotomy in lacertid lizards Panayiotis Pafilis 1, *, Kostas Sagonas 2, Grigoris Kapsalas 1, Johannes Foufopoulos 3, Efstratios D. Valakos 4 1 Department of Zoology and Marine Biology, School of Biology, National and Kapodistrian University of Athens, Panepistimioupolis, Ilissia 15784, Greece. *Corresponding author. ppafil@biol.uoa.gr 2 School of Biological and Chemical Sciences, Queen Mary University of London, Mile End Road, London E1 4NS, United Kingdom 3 School of Natural Resources and Environment, Dana Building, 430 East University, University of Michigan, Ann Arbor, MI , USA 4 Department of Animal and Human Physiology, School of Biology, National and Kapodistrian University of Athens, Panepistimioupolis, Ilissia 15784, Greece Submitted on 2017, 26 th January; revised on 2017, 16 th February; accepted on 2017, 22 nd February Editor: Marco Mangiacotti Abstract. Caudal autotomy is one of the most effective and widespread defensive mechanisms among lizards. When predators grasp the tail, lizards are able to shed it from the point of the attack and further. Numerous factors have been reported to affect tail-shedding performance such as temperature, age, predation pressure, intraspecific competition etc. Interestingly, the impact of sex on tail loss remains greatly understudied. Here, we analyzed tail autotomy performance, simulated in the lab, in 12 species of lacertid lizards belonging to five genera (Algyroides, Anatololacerta, Hellenolacerta, Ophisops, Podarcis). Our aim was to investigate whether sex affects caudal autotomy and/or the duration of post-autotomic tail movement. We failed to detect any effect of sex on tail loss in the species examined. Also, we did not find any sexual impact on the duration of tail movement after autotomy, with a single exception. Our findings suggest that autotomy serves as a defensive tactic equally in both sexes and is used in the same extent. Keywords. Predation, intraspecific competition, defense, Greece. INTRODUCTION Autotomy, the self-amputation of a body limb, is rather rare among vertebrates: the behavior is restricted to reptiles (Cooper and Alfieri, 1993; Hoare et al., 2006), salamanders (Marvin, 2010; Romano et al., 2010) and very few mice (Seifert et al., 2012). Undoubtedly, lizards are the champions of autotomy (Arnold, 1984; Bellairs and Bryant, 1985). Most families of the suborder shed their tail in response to mechanical stimuli exerted by a predator s attack (Arnold, 1987; Downes and Shine, 2001; Bateman and Fleming, 2009). The identity of the predator may vary. It could be any occasional or specialized saurophagous predator (e.g., snakes, birds, mammals) as traditional theory predicts (Pianka, 1970; Turner et al., 1982; Cooper et al., 2004), or a conspecific, triggered by intraspecific competition, as recent literature suggests (Pafilis et al., 2009b; Donihue et al., 2016; Itescu et al., 2017). In either case the result remains the same: the shed tail trashes vigorously, fuelled by anaerobic metabolism, to distract the predator while the tailless lizard escapes (Dial and Fitzpatrick, 1983; Pafilis et al., 2005). Many factors are known to affect caudal autotomy such as temperature, age and body shape (Arnold, 1984; Daniels, 1984; Pafilis and Valakos, 2008; Fleming et al., 2013). Previous studies on other aspects of autotomy did not report sexual effects on the trait, without though focusing on this particular issue (Chapple and Swain, 2004; Pafilis et al., 2005; Brock et al., 2015, but see Itescu et al., 2017). However there are several clues indicat- ISSN (print) ISSN (online) Firenze University Press

22 20 P. Pafilis et alii ing that tail autotomy could be sexually biased (Pérez- Mellado et al., 1997; Bateman and Fleming, 2009). Male lizards, particularly those belonging to territorial species, expose themselves in their effort to defend their territory (Kaiser and Mushinsky, 1994; Salvador and Veiga, 2001). Thus, typical alien predators (non conspecifics) have better chances to prey on the more conspicuous males that patrol or oversee their territory, a fact that could lead to higher autotomy rates (Bateman and Flemming, 2011; Marshall et al., 2016). Also, intramale competition may increase tail loss. Male lizards are much more aggressive against their peers compared to females (Kwiatkowski and Sullivan, 2002; Lailvaux and Irschick, 2007; McEvoy et al., 2013). Ergo, agonistic encounters between males are more common and may end up to tail loss (Cooper and Vitt, 1987; Bateman and Fleming, 2009; Cooper et al., 2015). Nonetheless, a byproduct of this increased intraspecific competition could be a higher propensity of males to hold on to tails more strongly. On the other hand, females are expected to lessen their ability to autotomize because of the high energetic demands of vitellogenesis and offspring production (Dial and Fitzpatrick, 1981; Hare and Miller, 2010). Species with high reproductive output (massive clutches) restrict or even completely lose autotomic abilities to offset the high costs of caudal autotomy (Pafilis and Valakos, 2008). Also, intrafemale competition is minimal since females do not defend territories (Braña, 1996; Moreira et al., 2006). Moreover, in some lizard families where males maintain harems, females interact frequently and do not compete (Zamudio and Sinervo, 2000). Hence females do not attack each other and incidents of tail loss are rare (Cooper et al., 2015). Tail autotomy comes with many disadvantages, such as degradation of social status (Fox et al., 1990; Salvador et al., 1995), loss of caudal fat that many species store in their tail (Roig et al., 2000; Chapple and Swain, 2002a; Cencetti et al., 2011), alterations in locomotion (Chapple and Swain, 2002b; Cromie and Chapple, 2012; Savvides et al., 2017, but see Kelehear and Webb, 2006), reduction of the immune function (Slos et al., 2009) and impaired reproduction (Fox and McCoy, 2000; Chapple et al., 2002). Nonetheless, in the case of intraspecific predation there is a clear advantage for the conspecific predator. By shedding and consuming a conspecific tail, males kill two birds with one stone: they may reduce their rival s ability to mate (Fox and Rostker, 1982; Martín and Salvador, 1993) and gain an energetically rich meal (McConnachie and Whiting, 2003; Cooper et al., 2015). Hence, intraspecific predation comes with a strong advantage, particularly favorable beneficial for males. In this study we aimed to clarify whether sex influences caudal autotomy performance. To this end we simulated tail shedding in the lab in 12 species of lacertid lizards. First, we hypothesized that since males are more exposed to predation (inter- or intraspecific) due to their particular social role, they would demonstrate higher rates of tail loss. In the case of insular species though, conditions are more complicated. Predation is more relaxed on the islands (MacArthur and Wilson, 1967; Whittaker and Fernández- Palacios, 2007) and this drives to higher lizard densities (Rodda and Dean-Bradley, 2002; Buckley and Jetz, 2007) that trigger more intense intraspecific competition (Knell, 2009; Raia et al., 2010). Relaxed predation advocates lower autotomy rates (traditional theory) whereas intraspecific competition suggests higher ones (recent approach), which would be even higher among males due to more frequent intraspecific agonistic encounters (Mougeot et al., 2003; Kokko and Rankin, 2006; Cooper et al., 2015). Second, we expected that post-autotomy duration of tail movement between males and females would not differ, as this feature appears to be conservative among species (Pafilis et al., 2005; Pafilis et al., 2008). Study species MATERIAL AND METHODS We examined the rates of caudal autotomy in 12 lacertid lizards assigned in five genera: Algyroides (A. moreoticus and A. nigropunctatus), Hellenolacerta (H. graeca), Ophisops (O. elegans) and Podarcis (P. cretensis, P. peloponnesiacus, P. erhardii, P. gaigeae, P. milensis, P. muralis and P. tauricus) (Table 1 for sample sizes individuals in total). The focal species are distributed in different locations and habitats in mainland and insular Greece (Fig. 1). All of them are small, diurnal insectivorous predators, with snout vent length (SVL) varying from 55 up to 85 mm (Valakos et al.; 2008). For each individual we recorded SVL, sex and the condition of the tail (intact or regenerated). For the purposes of this study we worked exclusively with adult individuals with intact tails. Captured lizards were transferred to the laboratory facilities of the Department of Biology at the University of Athens. All animals were housed individually in vitreous terraria ( cm) with sand and artificial shelters and were held at 30 o C under a controlled photoperiod with fluorescent tube lighting (12 h light: 12 h dark). An incandescent heat lamp (60 Watts) above each terrarium allowed lizards to thermoregulate for eight hours per day. Lizards were fed every other day with mealworms coated with supplementary vitamins and minerals and had access to water ad libitum. Predation simulation and postautotomy tail movement Prior to the experimental procedure food was withheld from lizards for two days (Pafilis et al., 2009a). We simulated predation using the method proposed by Pérez-Mellado et al.

23 No sexual effect on tail autotomy 21 Fig. 1. Map of the collecting sites in mainland and insular Greece, NE Mediterranean Basin. (1997). Since body temperature may affect caudal autotomy (Bustard, 1968; Daniels, 1984), lizards were allowed to thermoregulate for two hours in a specially designed terrarium ( cm) with two ice bags at one end and two heating lamps (100 W and 60 W) at the other end that provided a thermal gradient ranging from 10 to 50 o C (Van Damme et al., 1986). After achieving its preferred body temperature, each lizard was placed in a terrarium with cork substrate in order to maintain good traction during predation simulation. We used a pair of calipers to simulate the bite of a predator and grasped the tail 20 mm behind the cloaca for 15 sec. To standardize pressure, the calipers were closed to half the original diameter of the tail (Pérez-Mellado et al. 1997). Lizards were free to react and were not restrained. If autotomy occurred, we recorded the duration of movement of the shed, thrashing tail from the moment of autotomy to complete cessation of movement (no continuous twitches for 20 sec) using a digital timekeeper.

24 22 P. Pafilis et alii Table 1. Values for snout-vent length (SVL; mm) in males and females for all species and t-test results (P > 0.05, ns; P 0.05, *; P 0.01, **; P 0.001, ***). Means ± standard deviation; sample size in parenthesis. Species Males Females t-test Algyroides moreoticus ± 1.25 (22) ± 1.85 (15) *** Algyroides nigropunctatus ± 3.13 (26) ± 2.44 (18) *** Anatololacerta oertzeni ± 1.91 (26) ± 2.10 (20) *** Hellenolacerta graeca ± 2.24 (23) ± 2.31 (17) *** Ophisops elegans ± 2.18 (24) ± 1.78 (17) *** Podarcis cretensis ± 2.11 (17) ± 2.14 (14) *** Podarcis peloponnesiacus ± 2.23 (46) ± 2.04 (42) *** Podarcis erhardii ± 4.69 (40) ± 5.05 (29) * Podarcis gaigeae ± 3.44 (19) ± 4.36 (18) *** Podarcis milensis ± 2.09 (24) ± 2.24 (16) *** Podarcis muralis ± 2.22 (28) ± 1.92 (13) ** Podarcis tauricus ± 2.89 (18) ± 2.41 (14) *** Statistics We examined the normality of our data using the Kolmogorov-Smirnov and Lilliefors normality tests. Whenever parametric assumption was not met, data were log-transformed. T-test was used to compare the SVL between two sexes. Chisquare test followed by Fischer exact test was used to compare autotomy performance between the two sexes in each species. We used analysis of variance (ANOVA) to compare the duration of tail movement between the two sexes in each species. In order to eliminate the influence of SVL on the duration of tail movement we repeated the above-mentioned analysis using the SVL as covariate (ANCOVA). All comparisons were conducted independently for each species. All statistical analyses were conducted in R (R Development Core Team, 2015). RESULTS The comparison of body length revealed significant sexual dimorphism for all species (Table 1), with males being larger than females (the opposite pattern was revealed only in one species, Podarcis muralis). Chisquare test showed that rates of caudal shedding (all P > 0.05; Table 2) did not differ between sexes within each species (Fig. 2). The same lack of sexual differences was detected regarding the post-autotomy duration of tail movement (all P > 0.05; Table 3). DISCUSSION Caudal autotomy is quite common among lizards and a growing body of literature continuously provides new insights. However, the impact of sex on tail loss remains rather obscure. Here, we examined the possible Table 2. Differences in tail autotomy rates between males and females for all species and Chi-square (χ 2 ) test results (df = degree of freedom). Species χ 2 df P Algyroides moreoticus Algyroides nigropunctatus Anatololacerta oertzeni Hellenolacerta graeca Ophisops elegans Podarcis cretensis Podarcis peloponnesiacus Podarcis erhardii Podarcis gaigeae Podarcis milensis Podarcis muralis Podarcis tauricus sexual effect by investigating aspects of tail autotomy. Our results clearly refuted our first working hypothesis. We found no sexual differences in the rates of tail shedding in the 12 focal species. Our analyses revealed that insular species also conformed to this pattern. In line with our second prediction, we found no difference in the duration of the post-autotomic movement between sexes. Sex had no effect on caudal loss. Though several differences arose from the rates of tail shedding among species unveiling striking differences (e.g., 80% in A. nigropunctatus compared to 32% in P. gaigeae), intraspecies analyses yield a uniform pattern in tail autotomy performance between males and females. Contrary to our initial prediction, the higher exposure to predators and the

25 No sexual effect on tail autotomy 23 Fig. 2. Rates of laboratory autotomy (dark bars for males and light bars for females). Black diamonds denote island species. Table 3. Values for the duration of post-autotomy tail movement (min) in males and females for all species and ANOVA and ANCOVA results (P values in parenthesis): Means ± standard deviation; sample sizes are the same reported in Table 1. Species Males Females ANOVAs ANCOVAs Algyroides moreoticus 5.19 ± ± 0.31 F 1,35 = 2.88 (0.099) F 1,34 = (0.851) Algyroides nigropunctatus 5.29 ± ± 2.44 F 1,42 = 4.04 (0.051) F 1,41 = 3.68 (0.062) Anatololacerta oertzeni 6.15 ± ± 0.29 F 1,44 = 3.89 (0.055) F 1,43 = 0.09 (0.762) Hellenolacerta graeca 6.55 ± ± 0.23 F 1,38 = 3.72 (0.061) F 1,37 = 2.78 (0.104) Ophisops elegans 4.95 ± ± 0.23 F 1,39 = 3.43 (0.072) F 1,38 = 1.1, (0.292) Podarcis cretensis 5.54 ± ± 0.27 F 1,29 = 2.86 (0.101) F 1,28 = 3.09 (0.089) Podarcis peloponnesiacus 6.48 ± ± 0.34 F 1,86 = 3.63 (0.060) F 1,85 = 1.95 (0.167) Podarcis erhardii 6.14 ± ± 0.05 F 1,67 = 3.09 (0.083) F 1,66 = 2.76 (0.101) Podarcis gaigeae 5.71 ± ± 0.23 F 1,35 = 3.44 (0.076) F 1,34 = 3.51 (0.069) Podarcis milensis 6.46 ± ± 0.25 F 1,38 = 3.76 (0.060) F 1,37 = 3.81 (0.060) Podarcis muralis 6.23 ± ± 0.29 F 1,39 = 0.86 (0.360) F 1,38 = 1.23 (0.273) Podarcis tauricus 5.45 ± ± 0.21 F 1,30 = 3.32 (0.079) F 1,42 = 0.56 (0.459) different levels of intrasexual aggressiveness were not transformed into higher autotomy rates for males (Fig. 2). Our findings suggest that both sexes employ caudal autotomy at the same extent, at least among lacertids. The few studies that have assessed the impact of sex on tail autotomy provide contradictory results. On the one hand, Itescu et al. (2017) reported that male geckos (Mediodactylus kotshyi) had higher autotomy frequencies

26 24 P. Pafilis et alii than females in 31 different populations. These authors attributed the higher male autotomic rates to the more intense intraspecific competition. On the other hand, Fox et al. (1998) found that males of the phrynosomatid lizard Uta stansburiana shed their tail less easily compared to females and retain it more strongly as they approach sexual maturity. The latter researchers ascribed the tendency of males to avoid autotomy to the significance of the tail in reproductive success (Fox et al., 1998). The above indicate that phylogeny might have a distinct role on the sexual differentiation of autotomy. Coming to lacertids, although aggressive interactions with conspecific are well known (Castilla and Van Damme, 1996; Salvador and Veiga, 2001; Cooper et al., 2015), they do not seem to account for sexual differences in tail shedding performance according to our results. At this point we have to stress out an important caveat in our study: both sampling in the field and experimental procedure in the lab took place during the nonreproductive period. Reproduction triggers major shifts in lizards (Bauwens and Thoen, 1981; Brodie, 1989). Future mothers avoid exposing themselves to open areas so as to avoid predation, and adopt a more cryptic behavior (Shine, 1980; Karasov and Anderson, 1984; Braña, 1993). On the contrary, males are much more active and aggressive during the same period as they protect their territory and fight against rivals for access to females (Martín and Forsman, 1999; Salvador and Veiga, 2001; Troncoso-Palacios and Labra, 2012). Most probably, inter- and intraspecific predation pressure during the reproductive period would differ because of the dramatic behavioral shifts that both sexes undergo. Reassessment of tail autotomy performance during this period would shed further light on the impact of reproduction (e.g., Cooper et al., 2009). Despite the two contradicting drivers of tail autotomy prevailing on islands (low predation and high intraspecific competition), islanders followed the same pattern with mainland species and showed no sexual differences in tail autotomy performance (Fig. 2). However, island size might play a role. Lizard densities are higher on islands thanks to ecological release (Buckley and Jetz, 2007; Novosolov et al., 2016) and this applies to Mediterranean lacertids as well (Chondropoulos and Lykakis, 1983; Adamopoulou, 1999; Scalera et al., 2004). The highest densities, though, have been reported from very small islets (Castilla and Bauwens, 1991; Pérez-Mellado et al., 2008; Pafilis et al., 2013). Intraspecific competition peaks on these islets and very often includes consumption of conspecific limbs such as tails (Raia et al., 2010; Donihue et al., 2016; Lymberakis et al., 2016). Cannibalism on these islets may be as common that lizards may change their physiology and morphology to cope with this extreme intraspecific competition (Pafilis et al., 2016). Such antagonistic encounters are much more common and intense among males (Brock et al., 2015; Cooper et al., 2015). Our study was carried out on large islands that might host abundant populations (varying from 76 to 396 individuals per hectare) but certainly not as dense as those on the small Mediterranean islets. Also, we have to clarify that these large islands had considerably many predators, fewer though compared to the mainland (Pafilis et al., 2009a). Repeating the experiment on predatorfree islet populations might yield useful new insights. The duration of post-autotomy movement did not differ between sexes, in accordance with our second prediction. At this point, we have to stress out that though there was marginal significance in the duration of tail thrashing post autotomy between the sexes, this difference was eliminated by taking into account body size. Shed tails thrashed between five and seven minutes (Table 3), receiving values that fall within the same range with other Greek lacertids (Pafilis et al., 2005; Pafilis et al., 2009a). Our findings come to corroborate previous reports on Podarcis lizards (Pafilis et al., 2005; Pafilis et al., 2008). Duration of movement after tail shedding is very important for the successful escape from predators (Dial and Fitzpatrick, 1983; Cooper et al., 2004). We believe that this importance is reflected in the lack of sexual differentiation in post-autotomy movement. We report that tail shedding performance and post-autotomic duration are not affected by sex. Future research that would repeat the experiment during the reproductive period and include gravid females would provide further interesting results. Also, experimental work on small islets would shed light on the impact of intense intramale competition on autotomy rates. ACKNOWLEDGEMENTS Lizards were captured during the non-reproductive period (July to October, 2000 and 2001), handled and housed in accordance with Greek National Law (Presidential Decree 67/81). At the end of the experimental procedures all animals were released in the habitats they derived from. REFERENCES Adamopoulou, C. (1999): Structure and function of epigeic animal communities with emphasis in the lizard Podarcis milensis (Sauria: Lacertidae), in insular

27 No sexual effect on tail autotomy 25 ecosystems of the Aegean. Ph.D Thesis. National and Kapodistrian University of Athens, Athens, Greece. Arnold, E.N. (1984): Evolutionary aspects of tail shedding in lizards and their relatives. J. Nat. Hist. 18: Arnold, E.N. (1987): Caudal autotomy as a defense. In: Biology of the Reptilia, Vol. 16. Ecology B, pp Gans, C., Huey, R.B., Eds, Alan R. Liss, Inc., New York. Bateman, P.W., Fleming, P.A. (2009): To cut a long tail short a review of lizard caudal autotomy studies carried out over the last 20 years. J. Zool. 277: Bateman, P.W., Fleming, P.A. (2011): Frequency of tail loss reflects variation in predation levels, predator efficiency, and the behaviour of three populations of brown anoles. Biol. J. Linn. Soc. 103: Bauwens, D., Thoen, C. (1981): Escape tactics and vulnerability to predation associated with reproduction in the lizard Lacerta vivipara. J. Anim. Ecol. 50: Bellairs, A.d.A., Bryant, S.V. (1985): Autotomy and regeneration in reptiles. In: Biology of the Reptilia, Vol. 15. Development B, pp Gans, C., Billett, F., Eds, John Wiley & Sons, New York. Braña, F. (1993): Shifts in body temperature and escape behaviour of female Podarcis muralis during pregnancy. Oikos 66: Braña, F. (1996): Sexual dimorphism in lacertid lizards: male head increase vs female abdomen increase? Oikos 75: Brock, K.M., Bednekoff, P.A., Pafilis, P., Foufopoulos, J. (2015): Evolution of antipredator behavior in an island lizard species, Podarcis erhardii (Reptilia: Lacertidae): The sum of all fears? Evolution 69: Brodie, E.D. (1989): Behavioral modification as a means of reducing the cost of reproduction. Am. Nat. 124: Buckley, L.B., Jetz, W. (2007): Insularity and the determinants of lizard population density. Ecol. Lett. 10: Bustard, H.R. (1968): Temperature dependent tail autotomy mechanism in gekkonid lizards. Herpetologica 24: Castilla, A.M., Bauwens, D. (1991): Observations on the natural history, present status, and conservation of the insular lizard Podarcis hispanica atrata on the Columbretes archipelago, Spain. Biol. Conserv. 58: Castilla, A.M., Van Damme, R. (1996): Cannibalistic propensities in the lizard Podarcis hispanica atrata. Copeia 1996: Cencetti, T., Poli, P., Mele, M., Zuffi, M.A.L. (2011): Preliminary results on tail energetics in the Moorish gecko, Tarentola mauritanica. Acta Herpetol. 6: Chapple, D.G., McCoull, C.J., Swain, R. (2002): Changes in reproductive investment following caudal autotomy in viviparous skinks (Niveoscincus metallicus): lipid depletion or energetic diversion? J. Herpetol. 36: Chapple, D.G., Swain, R. (2002a): Distribution of energy reserves in a viviparous skink: Does tail autotomy involve the loss of lipid stores? Austral Ecol. 27: Chapple, D.G., Swain, R. (2002b): Effect of caudal autotomy on locomotor performance in a viviparous skink, Niveoscincus metallicus. Funct. Ecol. 16: Chapple, D.G., Swain, R. (2004): Inter-populational variation in the cost of autotomy in the metallic skink (Niveoscincus metallicus). J. Zool. 264: Chondropoulos, B.P., Lykakis, J.J. (1983): Ecology of the Balkan Wall Lizard, Podarcis taurica ionica (Sauria: Lacertidae) from Greece. Copeia 1983: Cooper, W.E., Alfieri, K.J. (1993): Caudal autotomy in the eastern garter snake, Thamnophis s. sirtalis. Amphibia-Reptilia 14: Cooper, W.E. Jr., Dimopoulos, I., Pafilis, P. (2015): Sex, age, and population density affect aggressive behaviors in island lizards promoting cannibalism. Ethology 121: Cooper, W.E. Jr., Pérez Mellado, V., Vitt, L.J. (2004): Ease and effectiveness of costly autotomy vary with predation intensity among lizard populations. J. Zool. 262: Cooper, W.E. Jr., Vitt, L.J. (1987): Deferred agonistic behavior in a long-lived scincid lizard Eumeces laticeps. Oecologia 72: Cooper, W.E. Jr., Wilson, D.S., Smith, G.R. (2009): Sex, reproductive status, and cost of tail autotomy via decreased running speed in lizards. Ethology 115: Cromie, G.L., Chapple, D.G. (2012): Impact of tail loss on the behaviour and locomotor performance of two sympatric Lampropholis skink species. PLoS ONE 7: e Daniels, C.B. (1984): The importance of caudal lipid in the gecko Phyllodactylus marmoratus. Herpetologica 40: Dial, B.E., Fitzpatrick, L.C. (1981): The energetic costs of tail autotomy to reproduction in the lizard Coleonyx brevis (Sauria: Gekkonidae). Oecologia 51: Dial, B.E., Fitzpatrick, L.C. (1983): Lizard tail autotomy: function and energetics of postautotomy tail movement in Scincella lateralis. Science 219: Donihue, C.M., Brock, K.M., Foufopoulos, J., Herrel, A. (2016): Feed or fight: testing the impact of food availability and intraspecific aggression on the functional ecology of an island lizard. Funct. Ecol. 30:

28 26 P. Pafilis et alii Downes, S., Shine, R. (2001): Why does tail loss increase a lizard s later vulnerability to snake predators? Ecology 82: Fleming, P.A., Valentine, L.E., Bateman, P.W. (2013): Telling tails: selective pressures acting on investment in lizard tails. Physiol. Biochem. Zool. 86: Fox, S.F., Conder, J.M., Smith, A.E. (1998): Sexual dimorphism in the ease of tail autotomy: Uta stansburiana with and without previous tail loss. Copeia 1998: Fox, S.F., Heger, N.A., Delay, L.S. (1990): Social cost of tail loss in Uta stansburiana: lizard tails as status-signalling badges. Anim. Behav. 39: Fox, S.F., McCoy, J.K. (2000): The effects of tail loss on survival, growth, reproduction, and sex ratio of offspring in the lizard Uta stansburiana in the field. Oecologia 122: Fox, S.F., Rostker, M.A. (1982): Social cost of tail loss in Uta stansburiana. Science 218: Hare, K.M., Miller, K.A. (2010): Frequency of tail loss does not reflect innate predisposition in temperate New Zealand lizards. Naturwissenschaften 97: Hoare, J.M., Pledger, S., Keall, S.N., Nelson, N.J., Mitchell, N.J., Daugherty, C.H. (2006): Conservation implications of a long term decline in body condition of the Brothers Island tuatara (Sphenodon guntheri). Anim. Conserv. 9: Itescu, Y., Schwarz, R., Meiri, S., Pafilis, P. (2017): Intraspecific competition, not predation, drives lizard tail loss on islands. J. Anim. Ecol. 86: Kaiser, B.W., Mushinsky, H.R. (1994): Tail loss and dominance in captive adult male Anolis sagrei. J. Herpetol. 28: Karasov, W.H., Anderson, R.A. (1984): Interhabitat differences in energy acquisition and expenditure in a lizard. Ecology 65: Kelehear, C., Webb, J.K. (2006): Effects of tail autotomy on anti-predator behavior and locomotor performance in a nocturnal gecko. Copeia 2006: Knell, R.J. (2009): Population density and the evolution of male aggression. J. Zool. 278: Kokko, H., Rankin, D.J. (2006): Lonely hearts or sex in the city? Density-dependent effects in mating systems. Phil. Trans. R. Soc. B. 361: Kwiatkowski, M.A., Sullivan, B.K. (2002): Mating system structure and population density in a polygynous lizard, Sauromalus obesus (= ater). Behav. Ecol. 13: Lailvaux, S.P., Irschick, D.J. (2007): The evolution of performance based male fighting ability in Caribbean Anolis lizards. Am. Nat. 170: Lymberakis, P., Valakos, E.D., Sagonas, K., Pafilis, P. (2016): The castaway: characteristic islet features affect the ecology of the most isolated European lizard. Acta Herpetol. 11: MacArthur, R.H., Wilson, E.O. (1967): The theory of island biogeography. Princeton University Press, Princeton, New Jersey. Marshall, K.L.A., Philpot, K.E., Stevens, M. (2016): Microhabitat choice in island lizards enhances camouflage against avian predators. Sci. Rep. 6: Martín, J., Forsman, A. (1999): Social costs and development of nuptial coloration in male Psammodromus algirus lizards: an experiment. Behav. Ecol. 10: Martín, J., Salvador, A. (1993): Tail loss reduces mating success in the Iberian rock-lizard, Lacerta monticola. Behav. Ecol. Sociobiol. 32: Marvin, G.A. (2010): Effect of caudal autotomy on aquatic and terrestrial locomotor performance in two desmognathine salamander species. Copeia 2010: McConnachie, S., Whiting, M.J. (2003): Costs associated with tail autotomy in an ambush foraging lizard, Cordylus melanotus melanotus. Afr. Zool. 38: McEvoy, J., While, G.M., Sinn, D.L., Wapstra, E. (2013): The role of size and aggression in intrasexual male competition in a social lizard species, Egernia whitii. Behav. Ecol. Sociobiol. 67: Moreira, P.L., López, P., Martín, J. (2006): Femoral secretions and copulatory plugs convey chemical information about male identity and dominance status in Iberian rock lizards (Lacerta monticola). Behav. Ecol. Sociobiol. 60: Mougeot, F., Redpath, S.M., Leckie, F., Hudson, P.J. (2003): The effect of aggressiveness on the population dynamics of a territorial bird. Nature 421: Novosolov, M., Rodda, G.H., Feldman, A., Kadison, A.E., Dor, R., Meiri, S. (2016): Power in numbers. Drivers of high population density in insular lizards. Global Ecol. Biogeogr. 25: Pafilis, P., Anastasiou, I., Sagonas, K., Valakos, E.D. (2013): Grazing by goats on islands affects the populations of an endemic Mediterranean lizard. J. Zool. 290: Pafilis, P., Foufopoulos, J., Poulakakis, N., Lymberakis, P., Valakos, E.D. (2009a): Tail shedding in island lizards [Lacertidae, Reptilia]: decline of antipredator defenses in relaxed predation environments. Evolution 63: Pafilis, P., Meiri, S., Foufopoulos, J., Valakos, E. (2009b): Intraspecific competition and high food availability are associated with insular gigantism in a lizard. Naturwissenschaften 96:

29 No sexual effect on tail autotomy 27 Pafilis, P., Meiri, S., Sagonas, K., Karakasi, D., Kourelou, E., Valakos, E.D. (2016): Body size affects digestive performance in a Mediterranean lizard. Herpetol. J. 26: Pafilis, P., Pérez-Mellado, V., Valakos, E. (2008): Postautotomy tail activity in the Balearic lizard, Podarcis lilfordi. Naturwissenschaften 95: Pafilis, P., Valakos, E.D. (2008): Loss of caudal autotomy during ontogeny of Balkan Green Lizard, Lacerta trilineata. J. Nat. Hist. 42: Pafilis, P., Valakos, E.D., Foufopoulos, J. (2005): Comparative postautotomy tail activity in six Mediterranean lacertid lizard species. Physiol. Biochem. Zool. 78: Pérez-Mellado, V., Corti, C., Lo Cascio, P. (1997): Tail autotomy and extinction in Mediterranean lizards. A preliminary study of continental and insular populations. J. Zool. 243: Pérez-Mellado, V., Hernández-Estévez, J.Á., García-Díez, T., Terrassa, B., Ramón, M.M., Castro, J., Picornell, A., Martín-Vallejo, J., Brown, R. (2008): Population density in Podarcis lilfordi (Squamata, Lacertidae), a lizard species endemic to small islets in the Balearic Islands (Spain). Amphibia-Reptilia 29: Pianka, E.R. (1970): Comparative autecology of the lizard Cnemidophorus tigris in different parts of its georgraphic range. Ecology 51: R Development Core Team (2015): R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Raia, P., Guarino, F.M., Turano, M., Polese, G., Rippa, D., Carotenuto, F., Monti, D.M., Cardi, M., Fulgione, D. (2010): The blue lizard spandrel and the island syndrome. BMC Evol. Biol. 10: 289. Rodda, G.H., Dean Bradley, K. (2002): Excess density compensation of island herpetofaunal assemblages. J. Biogeogr. 29: Roig, J.M., Carretero, M.A., Llorente, G.A. (2000): Reproductive cycle in a Pyrenean oviparous population of the common lizard (Zootoca vivipara). Neth. J. Zool. 50: Romano, A., Amat, F., Rivera, X., Sotgiu, G., Carranza, S. (2010): Evidence of tail autotomy in the European plethodontid Hydromantes (Atylodes) genei (Temmick and Schlegel, 1838) (Amphibia: Urodela: Plethodontidae). Acta Herpetol. 5: Salvador, A., Martín, J., López, P. (1995): Tail loss reduces home range size and access to females in male lizards, Psammodromus algirus. Behav. Ecol. 6: Salvador, A., Veiga, J.P. (2001): Male traits and pairing success in the lizard Psammodromus algirus. Herpetologica 57: Savvides, P., Stavrou, M., Pafilis, P., Sfenthourakis, S. (2017): Tail autotomy affects bipedalism but not sprint performance in a cursorial Mediterranean lizard. Sci. Nat. 104: 3. Scalera, R., Capula, M., Fornasari, L., Zava, B., Bombi, P., Mariottini, P., Bologna, M.A. (2004): Population structure, genetics and conservation of the Maltese wall lizard, Podarcis filfolensis, on Linosa Island (Reptilia, Lacertidae). Ital. J. Zool. 71: Seifert, A.W., Kiama, S.G., Seifert, M.G., Goheen, J.R., Palmer, T.M., Maden, M. (2012): Skin shedding and tissue regeneration in African spiny mice (Acomys). Nature 489: Shine, R. (1980): Costs of reproduction in reptiles. Oecologia 46: Slos, S., De Block, M., Stoks, R. (2009): Autotomy reduces immune function and antioxidant defence. Biol. Lett. 5: Troncoso-Palacios, J., Labra, A. (2012): Is the exploratory behavior of Liolaemus nitidus modulated by sex? Acta Herpetol. 7: Turner, F.B., Medica, P.A., Jennrich, R.I., Maza, B.G. (1982): Frequencies of broken tails among Uta stansburiana in southern Nevada and a test of the predation hypothesis. Copeia 1982: Valakos, E.D., Pafilis, P., Lymberakis, P., Maragou, P., Sotiropoulos, K., Foufopoulos, J. (2008): The amphibians and reptiles of Greece. Edition Chimaira, Frankfurt, Germany. Van Damme, R., Bauwens, D., Verheyen, R.F. (1986): Selected body temperatures in the lizard Lacerta vivipara: variation within and between populations. J. Therm. Biol. 11: Whittaker, R.J., Fernández-Palacios, J.M. (2007): Island biogeography: ecology, evolution, and conservation. Oxford University Press, Oxford. Zamudio, K.R., Sinervo, B. (2000): Polygyny, mateguarding, and posthumous fertilization as alternative male mating strategies. Proc. Natl. Acad. Sci. USA 97:

30

31 Acta Herpetologica 12(1): 29-36, 2017 DOI: /Acta_Herpetol Fire salamander (Salamandra salamandra) males activity during breeding season: effects of microhabitat features and body size Raoul Manenti*, Andrea Conti, Roberta Pennati Dipartimento di Bioscienze, Università degli Studi di Milano, Via Celoria, Milano (Italy).*Corresponding author. Submitted on 2016, 15 th March; revised on 2016, 17 th November; accepted on 2016, 19 th November Editor: Rocco Tiberti Abstract. After metamorphosis, fire salamander is considered fully terrestrial, usually inhabiting wooded areas around aquatic habitats. It is often reported that only females go back to water for laying the larvae. The aim of this study is to assess if sites where males are active during the breeding seasons have specific features among microhabitat determinants and distance from the breeding sites. In the autumns of 2013 and 2014, we surveyed 26 transects and 72 plots around six isolated breeding sites in North-Western Italy. During rainy nights, we recorded males position and distance from breeding pools, while during daytime we characterized the environmental features of the plots. Males detection probability was relatively high (mean ± SE: 81.0 ± 4.3%). Several males (15% of the observations) were encountered inside breeding pools where females were laying larvae. Males occurrence was positively related to plots closer to breeding pools and higher leaf litter depth. Larger males were found closer to the breeding pools. This case study shows that the distribution of fire salamander males during the breeding season depends on the breeding sites. Keywords. Breeding pond, ecology, mating behaviour. INTRODUCTION The assessment of ecological relationships between amphibians and their terrestrial habitat are very important to understand proper management actions and enhance conservation policies. Among European amphibians with extensive terrestrial habits there is the fire salamander, Salamandra salamandra. This is a widespread species in Europe, that has been extensively studied in its use of different breeding sites, while studies on terrestrial habitat requirements were less frequent (Denoël, 1996; Carafa and Biondi, 2004; Catenazzi, 2016). For breeding, S. salamandra is generally linked to small lotic environments (Manenti et al., 2009), but it can also use a wide variety of habitats like ponds (Egea-Serrano et al., 2006; Denoël and Winandy, 2014; Caspers et al., 2015) and hypogeous (i.e., subterranean) springs or pools (Ianc et al., 2012; Manenti et al., 2015). Moreover, some populations evolved complete viviparity, skipping the aquatic larval stage (Velo-Anton et al., 2007). Adults are strictly terrestrial and generally live in broadleaved forests (Joly, 1968). Even if it is generally reported that only females migrate to wetlands to give birth to the aquatic larvae while males do not go back to water after metamorphosis (Joly, 1968; Nöllert and Nöllert, 1992; Denoël, 1996; Lanza et al., 2009; Ficetola, 2012), it is known that the species lives around water bodies without strong displacements from them (Ficetola, 2012). In general the species seems to prefer deciduous mixed woodlands, especially those mostly dominated by the common beech Fagus sylvatica (Lanza et al., 2009; Balogova and Uhrin, 2014). Studies performed in the Alps show that the fire salamander prefers areas which are characterized by a mean annual precipitation > 600 ISSN (print) ISSN (online) Firenze University Press

32 30 R. Manenti, A. Conti, R. Pennati mm (Schauer et al., 2012). Salamanders tend to be active at temperature as low as 1 C, with a thermal optimum in terms of trade-off between thermoregulation and metabolic costs at 8 C (Catenazzi, 2016). Estivation occurs at temperatures above 16 C (Catenazzi, 2016). S. salamandra seems to avoid dry forests (especially spruce monocultures) and its ecology in floodplains and coniferous forests needs further studies (Lanza et al., 2009). Recent studies provided evidences of the importance that a complex ensemble of determinants, linked to both terrestrial and aquatic habitats, plays in allowing the occurrence of stable populations (Ficetola et al., 2011). However few studies are available in terms of microhabitat use in the forest environment. These data could allow to gain a better knowledge on proper forests habitat management for conservation purposes. Adults are considered to be territorial even if some superimpositions of home ranges may occur (Lanza et al., 2009) and during winter individuals may aggregate in shelters (Baumgart, 1981; Balogova and Uhrin, 2014). The adult activity usually shows two peaks during the year, coinciding with high rainfall periods: one during spring, and one in autumn (Joly, 1968; Lanza et al., 2009). These are also the two seasons in which both mating and larvae laying occur, especially South to Alps (Romeo et al., 2015). Females usually lay larvae during night and may stay around their breeding sites for several nights, as females repeatedly give birth to only some of the larvae they retain (Joly, 1986); moreover in these occasions they likely mate for the following year s brood (Steinfartz et al., 2006). Mating happens on land habitats, most often in spring and in autumn months, and usually at night (Lanza et al., 2009). The reproductive males are reported to show a frequent posture during the courtship, in particular they rise on their anterior limbs and show a quick bucco-pharyngeal respiration. This posture can be frequently observed in the males of S. salamandra (Bruno, 1973; Catenazzi, 1998) and also in males of other congeneric species as S. lanzai and S. atra, also in absence of females and in more or less plane and open areas (Manenti and Pennati pers. obs.). In S. salamandra the display of this posture has been interpreted both as a territorial and competitive behaviour among males (Bruno, 1973), and as a strategy allowing a more effective perception or attraction of the females (Lanza et al., 2009). The advantage of an affective female detection could also favour a non-random choice by the males of the microhabitat where being active during the mating seasons. However, specific ethological and ecological studies are missing and, at present, it is impossible to say if males spatial patterns are linked to specific reproductive behaviours. The general hypothesis of the present study is that, to increase their chances of mating, males should stay close to the females as soon as they are ready to mate, i.e., after they laid their larvae. Our specific hypotheses are that getting closer to the breeding ponds should be the most effective spatial strategy for males and that they should compete for the territory patches closer to the breeding sites. We therefore tested if the sites where males are active during the breeding seasons have specific features among some microhabitat determinants and distance from the breeding site. Moreover, we want to assess if male size (as a proxy of their reproductive quality) have a role in determining the location of activity sites. These information could increase our understanding of the spatial and reproductive ecology of S. salamandra and confirm that males compete for the occupancy of some sites/ microhabitats and could be useful to perform further ethological studies on their reproductive and territorial behaviour. MATERIAL AND METHODS During the months of October and November 2013 and 2014, we performed surveys around 6 distinct breeding sites of the fire salamander located in the districts of Lecco and Monza e Brianza, Lombardy, North-western Italy (Fig. 1). The study area is characterized by hilly reliefs with a good cover of broadleaved woodlands dominated by Castanea sativa and Quercus robur. It is crossed by a dense hydrographic network characterized by different typologies of water bodies such as creeks, streams, brooks, resurgences and rivers. Breeding sites were characterised by isolated pools, fed by small springs, with an area ranging from 5 to 15 m 2. Each isolated pool corresponded to a different locality. The maximum distance between the 6 sites studied was 14 Km (average 6.9 Km). Linear transect surveys We looked for fire salamander males during night surveys in rainy nights (between 9:00 pm and 12:00 pm), during which we also check for larvae depositions by females. These surveys were performed along 26 linear straight transects (4 linear transect around 4 pools and 5 linear transect around other two pools) beginning at distances comprised between 110 and 75 m (average ± SE: 97.2 ± 5.3 m) from the breeding sites and ending in them. The starting positions of the transects were randomly chosen during the daytime of the first field survey, and distinctive marks were placed to distinguish them to each other. Then we walked in the direction of the breeding site measuring the covered distance with a measuring strip and tracking and mapping the transect with a GPS Garmin Eterx 10 (precision 3 m). The width of the surveyed area around the transect was 5 meters (2.5 m on each side). For each survey season (October- November 2013 and 2014), we performed at least two surveys

33 Fire salamander males activity around breeding sites 31 Fig. 1. Localization of the breeding sites (triangles) of fire salamander around which we positioned linear transects and plots.numbers indicate the site ID; some of the sites are superimposed due to geographic proximity. for transect (range 2-4 surveys per transect, average ± SE: 2.76 ± 0.11 times). Surveys for each transect lasted from 30 to 45 minutes. During night surveys, we caught all the fire salamanders that we encountered. We placed the individuals in small holed bags in the same point of observation following protocols used for other salamanders (Lunghi et al., 2015) in order to do the measures successively, without affecting the position of the individuals occurring in the successive parts of the transect. The point of encounter was registered with the same GPS Garmin Etrex 10. At each survey we re-recorded the starting point of the transects and the pool position, tracking all the transects length. After the capture of all the observed individuals we took a picture of each individual and registered their position. Each individual was photographed on a plastic millimetre paper to allow successive accurate measures of the total length, and weighed with a precision dynamometer PESOLA (precision ± 0.1 g). The total lengths (precision ± 1 mm) of the males was obtained through the software ImageJ. In 2014 we also took a picture of the dorsal patterns to allow the individual identification of the salamanders. For each salamander, we used GPS data to record the distance from the pool, and we measured the maximum slope inclination of the capture point with a digital inclinometer Digital Angle Bevel Box 360 (precision ± 0.2 ) placed over the middle of a flat wooden board of cm. Sex determination was performed on the basis of their cloaca features. As males develop a swollen cloaca when they are sexually mature (Raffaëlli, 2007), the individuals shorter than the smallest recognizable male were considered juveniles. Sample areas/plots surveys Along each transect, we chose a total of 72 circular plots with a 2.5 m ray, separated each other by at least 5 meters, and placed within 30 m from the breeding pools (average number ± SE of plots per transect was 2.7 ± 0.2, range: 2-5). The number of plots was chosen to specifi-

34 32 R. Manenti, A. Conti, R. Pennati cally cover the part of the transects placed closer to the breeding sites, and their position was randomly chosen within a buffer of 30 m around each breeding site to gather detailed information on the microhabitats. In these plots, we recorded five environmental variables during daytime surveys: litter leaf depth, maximum slope inclination, number of trees with a diameter > 50 cm, distance from the breeding pool, bush cover. Leaf litter depth may be linked to shelter availability, humidity level of the soil and prey availability: it was measured placing a rigid ruler in the point of the plot with the maximum litter depth and measuring its height from the soil. Maximum slope may be important to explain salamander distribution (Manenti et al., 2011; Werner et al., 2014): inclination was measured placing digital inclinometer over a rigid wooden board (20 20 cm) in the point of the plot with the maximum inclination. The trees diameter was assessed using a measuring strip: all the trees with a diameter > 50 cm occurring in the plots were counted, in order to detect the possible effect of the occurrence of large and old trees that are often shelters for salamanders (Apodaca and Godwin, 2015). The distance of the plot from the breeding sites was assessed using measuring strips with a precision of 0.01 m. The bush cover was estimated following the point-intercept method described by Dodd (2010), using a 2 m long pole placed vertically in the plot center and numbering the times it intercepts any piece of vegetation. The position of each plot was recorded through GPS and each plot was marked with small, temporary but visible during night marks over trees or rocks. Statistical analyses Linear transect surveys We built a Linear Mixed Model (LMM) to assess factors influencing the distance from the breeding sites at which males were found during night samplings. We limited this analysis to the males observations of 2014 (n = 80), because the identification of each individual (based on their dorsal pattern) was performed only in the second year. We added the males square-root transformed total length as dependent variable, and log-transformed distance from the breeding site, and log-transformed slope inclination as independent variables. As we found recaptures, we added the individuals identity as random effect. Variables transformation was performed to increase normality (tested through a Shapiro test). Since there is a strong exponential linear dependence between weight (W) and total length (L) of males (W = 0.02 L 2.67 ; R 2 = 0.82), we considered only total lengths to analyse the effects of male size on their distribution. The variables significance was assessed through a Wald F test (Bolker et al., 2008). Sample areas/plots surveys We used PRESENCE 5.5 to assess in each plot the probability of detection per visit of males and females (Hines, 2006). Accounting for the probability of detection it is important because a non-detection record (equivalent to a recorded absence) represents a lack of evidence that the species occurs in a specific site (Gomez-Rodriguez et al., 2012). We used multiple season models (Mac- Kenzie et al., 2003) to understand the predictors (survey and environmental variables) that satisfactorily described the collected records while accounting for imperfect detection (MacKenzie et al., 2002). Meaning that we considered the story of detection and non-detection per visit per plot considering both Autumn 2013 and Autumn On the basis of this history, PRESENCE returns a probability to have correctly assessed the presence/ absence of the species in the plots, such as a probability of occurrence (MacKenzie et al., 2003). To disentangle the role of the predictors it is necessary to build models comprising all their possible combinations and to select the best reliable model. We used the Akaike s information criterion (AIC values) to select the best model explaining males probability of occurrence in the plots (Rolls, 2011; Lele et al., 2013). The model with the lowest AIC and the highest weight was considered the best model describing species detectability (Burnham and Anderson, 2002). Since a model with no covariates for neither occupancy nor detectability (model ψ(.) p(.)) showed that the detection probability of males was relatively high (mean ± SE: 81.0 ± 4.3%), when building the other occupancy models, we only varied the predictor variables for occupancy and always used a model with constant detection probability (i.e., model ψ(predictor variables) p(.)). Thanks to the relative small number of variables considered, we were able to test all the possible combinations among them. In order to reduce the retention of overly complex models we excluded from the candidate set those models that were more complicated versions of any model with a lower QAIC value (Richards et al., 2011). This approach reduces model uncertainty, and improves the performance of model selection (Richards et al., 2011). We present all the candidate models with some supporting information, e.g. delta-aic < 4, QAIC weight > 0.1. To assess the relative role of the variables composing the best model we used a Wald F test (Bolker et al., 2008) with car package in R environment. To asses which environmental factors affect males sites choice, we preferred to use the estimates of males probability of occupancy at a given plot (calculated with

35 Fire salamander males activity around breeding sites 33 PRESENCE) rather than the observed naïve absence/ presence. We considered this value as the dependent variable of a Generalized Linear Mixed Model (GLMM) with a quasi-binomial error distribution using the log-transformed microhabitat features recorded in the plots as independent variables. We included pool locality as random factor, to account for the fact that several plots can surround the same pool. All the analyses were performed by R software using the packages lmertest, lme4 and car (R Development Core Team, 2013). Linear transects surveys RESULTS During the two sampling seasons, we caught 213 salamanders 147 of which were males. Males total length was 17.1 cm on average (maximum 21.6 cm, minimum 8 cm, SD = 2.39), while weight was 32.5 g on average (maximum 45 g, minimum 4 g, SD = 10.0). Females were 17.5 cm long on average (maximum 21.2 cm, minimum 8.5 cm, SD = 3.3) and weighed 43.4 g on average (maximum 78 g, minimum 6 g, SD = 18.7). In 67% of night surveys, we observed ponding females near the pools borders or on floating elements in the pools. The 15 % of the observed males were encountered inside breeding pools where females were laying larvae. Average males distance from the pools was 21.2 m (maximum 106 m, minimum 0 m, SE = 3.44). We detected a significant effect of males size on their distribution. In particular, on average, longer males were closer to the breeding pools (Table 1). Sample areas/plots surveys Detection probability on the surveyed plots was higher for males (p = 0.81 ± 4.3% SE) than for females (p = 0.54 ± 6.2% SE). The plots in which males were Table 1. Results of LMMs analysis showing the relationship between the size of the encountered males considered as dependent variable and, slope inclination and distance from the breeding pools of the encountering point considered together as independent variables (NumDF = degrees of freedom in the numerator; DenDf = degrees of freedom in the denominator). Variable Estimate NumDf DenDf F P Distance from the breeding site Slope inclination encountered showed a higher mean litter depth (1,5 cm ± 0.2 SE vs. 0.9 cm. ± 0.1 SE), slower slope inclination (5.48 ± 1.2 SE vs ± 1.7 SE) and lower number of trees (1.46 ± 0.2 SE vs ± 0.3 SE). The best occupancy model was that including distance to breeding pools, slope inclination, number of trees and leaf litter depth (Table 2). The occupancy probability of males was higher in plots closer to breeding pools (F = 22.29; P < 0,001) and with higher leaf litter depth (F = 10.11; P < 0,01). There was a non-significant negative tendency to both slope inclination and trees density. DISCUSSION Our results provide novel data on the fire salamander males spatial activity around breeding sites. In particular, our study provides evidence that during the breeding season males use microhabitats with specific features like deeper leaf litter and closer to breeding sites. Generally the use of terrestrial habitats by the salamanders of the genus Salamandra has been investigated to study their site fidelity to a restricted area and their relatively small home ranges (Schulte et al., 2007; Ficetola, 2012). Most of the information about the use of the terrestrial environment is at the landscape scale and underlines the importance of suitable woodland habitats in riparian, mountainous and agricultural areas (Manenti et al., 2009; Tanadini et al., 2012; Manenti et al., 2013); while studies on terrestrial microhabitat preference in the genus Salamandra are lacking. Our study was carried out when females reach the breeding sites to lay the larvae and focuses on adult males, to understand their microhabitat choice when they likely look actively for mating. In the fire salamander, copulation may occur at almost any period of the year except during the winter (Francis, 2002); it hap- Table 2. Model selection of the best fitted site-occupancy models (only models with ΔQAIC 4 are shown). The symbol (.) indicates a constant parameter with no covariate and K is the number of parameters in the model. ΔAIC is the difference between the QAIC score of the model and the best ranked model and w is the Akaike model weight. Model K QAIC ΔQAIC w ψ(distance + leaf litter depth + slope + n. of trees) p(.) ψ(distance + n. of trees) p(.) ψ(distance + leaf litter depth) p(.) ψ(distance) p(.)

36 34 R. Manenti, A. Conti, R. Pennati pens generally during night, and both in spring and autumn it is quite frequent (Rivera et al., 1999; Lanza et al., 2009). The higher probability of finding males was found in the plots where the trees were less numerous and slope inclination lower, even if the tendency was not significant. A similar result has been already reported by Denoël (1996) who observed a high encounter rate for footpaths and open areas. It is likely that this choice may be linked to the possibility to enhance mating opportunities through a more effective females detection. Some previous studies showed the existence of a role played by the slope inclination since the species generally visits more frequently sites with inclinations lower than 35 (Manenti et al., 2011; Werner et al., 2014) and, especially in mountainous landscapes, salamanders are active on paths and microhabitats with lower inclinations (Pantuso et al., 2015). We found males in plots with a lower mean slope inclination, and this variable was included in the best occupancy model. No information is available on the role played by leaf litter; even if its depth may be easily linked to prey occurrence, shelters availability and optimal humidity. The sex ratio observed during our study (69% males) is similar to those found by Klewen (1985) and Schulte et al. (2007) in autumn. The observed differences between sexes are correlated with general differences in activity patterns (Schulte et al., 2007), with females that likely are less active outside shelters than males which probably remain active to maximize their mating opportunities. Our results clearly show that males are more active in microhabitats close to the breeding sites. This fact may be incidental to the use of the breeding sites by females, with males exploiting the pools and their surroundings to enhance the probabilities to mate. Although other explanations may exist, as males could prefer the surroundings of the water bodies because humidity or prey availability could be higher. This is the first report demonstrating that the activity of adult fire salamanders of both sex depends to the vicinity of the water, while it was generally believed that after metamorphosis only females go back to the water bodies (Lanza, 1983; Joly, 1986; Nöllert and Nöllert, 1992; Ficetola, 2012). Moreover, our results reveal that there are differences among males in their distribution around the breeding sites. In particular, larger males exploit the microhabitats closest to the pools thus increasing their probability to meet the females. Thus, a competition and a territoriality for the sites is likely to occur. The existence of possible territoriality in S. salamandra at least during the breeding seasons has been reported (Catenazzi, 1998). Moreover in the close relative S. algira the existence of fights between males has been reported (Bogaerts and Donaire- Barroso, 2005). Males that are likely to be looking for mating display a typical posture rising on their anterior limbs; this position has also been reported as a fighting posture (Bruno, 1973) but may be interpreted as an olfactory strategy allowing a more effective perception of the females presence, or as a behavioural display to attract females, related to male quality (Wells, 2007; Lanza et al., 2009). We did not precisely quantified the males exhibiting this posture because we took any care to avoid disturb and fear responses but this aspect may be of particular ethological interest and further researches should be conducted to relate males behaviour and microhabitats choice; not only in the fire salamander, but also in its fully terrestrial congeneric species like S. atra and S. lanzai in which the same behaviour is very common (Andreone, 1992) and the knowledge of the ecological and ethological factors favouring active mate searching by males may be of importance for proper habitat managements. In general our study underlines that factors facilitating females perception like vicinity to the breeding sites drive the male distribution patterns. The extensive exploitation of terrestrial habitats surrounding the breeding pools also has implications for planning proper conservation actions for S. salamandra, as the maintenance of favourable microhabitats allowing females perception may enhance mating possibilities for males. The terrestrial environment surrounding water bodies appears to play a fundamental role in the ecology and ethology of the fire salamander. AKNOWLEDGMENTS We are grateful to Rocco Tiberti and two anonymous reviewers for precious suggestions to previous version of the manuscript. We thank Chiara Galliani for the professional English revision of the text. The research was approved by the Lombardy Region Authority and authorized complying with the regional law 10/2008, p. n. F , started on January , ended on December Adults were manipulated as less as possible, and released in the collecting point according to the permission prescriptions. REFERENCES Andreone, F. (1992): Observations on the territorial and reproductive behaviour of Salamandra lanzai and considerations about its protection (Amphibia: Salamandridae). British Herp. Soc. Bull. 39:

37 Fire salamander males activity around breeding sites 35 Apodaca, J.J., Godwin, J.C. (2015): Effects of buffering key habitat for terrestrial salamanders: implications for the management of the federally threatened red hills salamander (Phaeognathus hubrichti) and other imperiled plethodontids. Forests 6: Balogova, M., Uhrin, M. (2014): Patterns of wintering of fire salamanders (Salamandra salamandra) in an artificial underground roost. N. West. J. Zool. 10: Baumgart, G. (1981): Observations sur l ibernation de quelques amphibians dans les anciennes mines vosgiennes: Salamandre tachetee (Salamandra salamandra Linneé, 1758), Grenouille rousse (Rana temporaria Linné, 1758) et Crapaud commun (Bufo bufo Linné, 1758). Aquarama 58: Bogaerts, S., Donaire-Barroso, D. (2005): Rival combats in the North Moroccan Fire salamander Salamandra algira tingitana. Zeitschrift für Feldherpetologie 12: Bolker, B.M., Brooks, M.E., Clark, C.J., Geange, S.W., Poulsen, J.R., Stevens, M.H.H., White, J.S. (2008): Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol. Evol. 24: Bruno, S. (1973): Anfibi d Italia: Caudata. Natura 64: Burnham, K.P., Anderson, D.R. (2002): Model selection and multimodel inference: a practical informationtheoretic approach. Springer Verlag, New York. Carafa, M., Biondi, M. (2004): Application of a method for individual photographic identification during a study on Salamandra salamandra gigliolii in central Italy. It. J. Zool. 71: Caspers, B.A., Steinfartz, S., Krause, E.T. (2015): Larval deposition behaviour and maternal investment of females reflect differential habitat adaptation in a genetically diverging salamander population. Behav. Ecol. Sociobiol. 69: Catenazzi, A. (1998): Ecologie d une population de Salamandre tachetée au Sud des Alpes. MSc thesis. Université de Neuchâtel, Neuchâtel. Catenazzi, A. (2016): Ecological implications of metabolic compensation at low temperatures in salamanders. PeerJ 4: e2072. Denoël, M. (1996): Phénologie et domaine vital de la salamandre terrestre Salamandra salamandra terrestris (Amphibia, Caudata) dans un bois du Pays de Herve (Belgique). Cahiers Etol. 16: Denoël, M., Winandy, L. (2014): Fire salamander (Salamandra salamandra) in Larzac plateau: low occurrence, pond-breeding and cohabitation of larvae with paedomorphic palmate newts (Lissotriton helveticus). Acta Herpetol. 9: Dodd, C.K.J. (2010): Amphibian Ecology and Conservation. Oxford University Press, Oxford. Egea-Serrano, A., Oliva-Paterna, F.J., Torralva, M. (2006): Breeding habitat selection of Salamandra salamandra (Linnaeus, 1758) in the most arid zone of its European distribution range: application to conservation management. Hydrobiologia 560: Ficetola, G.F., Marziali, L., Rossaro, B., De Bernardi, F., Padoa-Schioppa, E. (2011): Landscape-stream interactions and habitat conservation for amphibians. Ecol. Appl. 21: Ficetola, G.F., Manenti, R., De Bernardi, F., Padoa-Schioppa, E. (2012): Can patterns of spatial autocorrelation reveal population processes? An analysis with the fire salamander. Ecography 35: Francis, E.T.B. (2002): The anatomy of the salamander. Reprint Society for the Study of Amphibians and Reptiles. Ithaca, New York. Gomez-Rodriguez, C., Bustamante, J., Diaz-Paniagua, C., Guisan, A. (2012): Integrating detection probabilities in species distribution models of amphibians breeding in Mediterranean temporary ponds. Divers. Distrib. 18: Ianc, R., Cicort-Lucaciu, A.S., Ilies, D., Kovacs, E.H. (2012): Note on the presence of Salamandra salamandra (Amphibia) in caves from Padurea Craiului Mountains, Romania. N. West. J. Zool. 8: Joly, J. (1968): Données écologiques sur la salamandre tachetée Salamandra salamandra. Annales des Sciences Naturelles Zoologie - Paris 12: Joly, P. (1986): La reproduction de la salamandre terrestre. In: Traité de Zoologie, p Grassé, P.P., Ed, Masson, Paris. Klewen, R. (1985): Untersuchungen zur Ökologie und Populationsbiologie des Feuersalamanders (Salamandra salamandra terrestris Lacépède 1788) an einer isolierten Population im Kreise Paderborn. Abh. Westfäl. Mus. Nat.kd. 47: Lanza, B. (1983): Anfibi e rettili (Amphibia-Reptilia). Guide per il riconoscimento delle specie animali delle acque interne italiane. CNR, Roma. Lanza, B., Nistri, A., Vanni, S. (2009): Anfibi d Italia. Ministero dell Ambiente e della Tutela del Territorio e del Mare, ISPRA. Grandi & Grandi Editori, Modena. Lele, S.R., Merrill, E.H., Keim, J., Boyce, M.S. (2013): Selection, use, choice and occupancy: clarifying concepts in resource selection studies. J. Anim. Ecol. 82: Lunghi, E., Manenti, R., Ficetola, G.F. (2015): Seasonal variation in microhabitat of salamanders: environmental variation or shift of habitat selection? PeerJ 3: e1122.

38 36 R. Manenti, A. Conti, R. Pennati MacKenzie, D.I., Nichols, J.D., Hines, J.E., Knutson, M.G., Franklin, A.B. (2003): Estimating site occupancy, colonization, and local extinction when a species is detected imperfectly. Ecology 84: Manenti, R., Ficetola, G.F., De Bernardi, F. (2009): Water, stream morphology and landscape: complex habitat determinants for the fire salamander Salamandra salamandra. Amphibia-Reptilia 30: Manenti, R., De Bernardi, F., Ficetola, G.F. (2013): Pastures vs forests: do traditional pastoral activities negatively affect biodiversity? The case of amphibians communities. N. West. J. Zool. 9: Manenti, R., Pennati, R., Ficetola, G.F. (2015): Role of density and resource competition in determining aggressive behaviour in salamanders. J. Zool. 296: Manenti, R., Ficetola, G.F., Marieni, A., De Bernardi, F. (2011): Caves as breeding sites for Salamandra salamandra: habitat selection, larval development and conservation issues. N. West. J. Zool. 7: Nöllert, A., Nöllert, C. (1992): Die Amphibien Europas. Kosmos, Stuttgart. Pantuso, F., Delle Monache, D., Ficetola, G.F., Manenti, R. (2015): Selezione del microhabitat frequentato dagli adulti di salamandra pezzata (Salamandra salamandra) durante l attività epigea nell area dell Alpe del Vicerè comune di albavilla (CO). In: X Congresso Nazionale Societas Herpetologica Italica, p Doria, G., Poggi, R., Salvidio, S., Tavano, M., Eds, Ianieri Edizioni, Pescara. R Development Core Team. (2013): R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. Raffaëlli, J. (2007): Les Urodeles du monde. Penclen Edition, Condé-sur-Noireau. Richards, S.A., Whittingham, M.J., Stephens, P.A. (2011): Model selection and model averaging in behavioural ecology: the utility of the IT-AIC framework. Behav. Ecol. Sociobiol. 65: Rivera, X., Simon, J.G., Melero, J.A., Urjos, N., Vilagrosa, A. (1999): El ciclo de actividad de Salamandra salamandra (L.) en una poblaci6n de Catalunya (Nordeste de la Peninsula lberica). But. Soc. Cat. Herp. 14: Rolls, R.J. (2011): The role of life-history and location of barriers to migration in the spatial distribution and conservation of fish assemblages in a coastal river system. Biol. Conserv. 144: Romeo, G., Giovine, G., Ficetola, G.F., Manenti, R. (2015): Development of the fire salamander larvae at the altitudinal limit in Lombardy (north-western Italy): effect of two cohorts occurrence on intraspecific aggression. N. West. J. Zool. 11: Schauer, J., Meikl, M., Gimeno, A., Schwarzenbacher, R. (2012): Larval monitoring of fire salamanders within a Sparkling Science Project. Eco. Mont. 4: Schulte, U., Kusters, D., Steinfartz, S. (2007): A PIT tag based analysis of annual movement patterns of adult fire salamanders (Salamandra salamandra) in a Middle European habitat. Amphibia-Reptilia 28: Steinfartz, S., Stemshorn, K., Kuesters, D., Tautz, D. (2006): Patterns of multiple paternity within and between annual reproduction cycles of the fire salamander (Salamandra salamandra) under natural conditions. J. Zool. 268: 1-8. Tanadini, M., Schmidt, B.R., Meier, P., Pellet, J., Perrin, N. (2012): Maintenance of biodiversity in vineyard-dominated landscapes: a case study on larval salamanders. Anim. Conserv. 15: Velo-Anton, G., Garcia-Paris, M., Galan, P., Rivera, A.C. (2007): The evolution of viviparity in holocene islands: ecological adaptation versus phylogenetic descent along the transition from aquatic to terrestrial environments. J. Zoolog. Syst. Evol. Res. 45: Wells, K.D. (2007): The Ecology and Behaviour of Amphibians. The University of Chicago Press, Chicago. Werner, P., Lötters, S., Schimdt, B.R. (2014): Analysis of habitat determinants in contact zones of parapatric European salamanders. J. Zool. 292:

39 Acta Herpetologica 12(1): 37-48, 2017 DOI: /Acta_Herpetol Call variation and vocalizations of the stealthy litter frog Ischnocnema abdita (Anura: Brachycephalidae) Pedro Carvalho Rocha 1,2, *, João Victor A. Lacerda 2,3, Rafael Félix de Magalhães 2,3, Clarissa Canedo 4, Bruno V. S. Pimenta 5, Rodrigo Carrara Heitor 6, Paulo Christiano de Anchieta Garcia 2,3 1 Programa de Pós-Graduação em Biologia Animal, Laboratório de Paleontologia e Osteologia Comparada, Departamento de Biologia Animal, Universidade Federal de Viçosa , Viçosa, MG, Brazil. *Corresponding author. p.rocha1990@gmail.com 2 Laboratório de Herpetologia, Departamento de Zoologia, Instituto de Ciências Biológicas, Universidade Federal de Minas Gerais , Belo Horizonte, MG, Brazil. 3 Programa de Pós-Graduação em Zoologia, Instituto de Ciências Biológicas, Universidade Federal de Minas Gerais , Belo Horizonte, MG, Brazil. 4 Departamento de Zoologia, Universidade do Estado do Rio de Janeiro , Rio de Janeiro, RJ, Brazil. 5 Bicho do Mato Meio Ambiente Ltda. (Bicho do Mato Instituto de Pesquisa) , Belo Horizonte, MG, Brazil. 6 Prefeitura Municipal de Espera Feliz/MG, Secretaria Municipal de Meio Ambiente e Defesa Civil , Espera Feliz, MG, Brazil. Submitted on 2016, 19 th May; revised on 2016, 16 th December; accepted on 2017, 21 st February Editor: Fabio Maria Guarino Abstract. Ischnocnema abdita is a small-sized litter frog belonging to the I. verrucosa species series and only known for mountainous areas of southeastern Espírito Santo State, Brazil, in the Municipalities of Santa Teresa (type locality), Cariacica and Mimoso do Sul. In this paper, we describe the calls and provide estimates of within-male variation of I. abdita from its type locality and from a recently discovered population in the region of Alto Caparaó, Municipality of Espera Feliz, Minas Gerais State, Brazil. Additionally, we also performed a GMYC analysis of molecular assignment that recovered the haplotypes of I. abdita from its type locality and from the new record (Alto Caparaó) under the same taxonomical entity. Our bioacoustical analysis revealed two distinct types of calls, herein referred as A and B calls. The A call was observed in both populations, whereas the B call was only recorded at Alto Caparaó. Despite the apparent similarity in the A calls from both localities, we observed differences in all traits analyzed. Moreover, each call trait expressed variation within males. The peak frequency never exceeded 5% variation and it was classified as static in both populations. Temporal parameters, such as call duration and interval between calls were classified either as dynamic or intermediate, with variations ranging from % within males. Although number of pulses per note was a dynamic trait at the type locality, it did not vary in both types of call recorded at Alto Caparaó. Keywords. Systematics, Taxonomy, Bioacoustics, molecular assignment, general mixed Yule-coalescent (GMYC) model. ISSN (print) ISSN (online) INTRODUCTION Mate attraction has been associated with anuran vocalizations for more than a century (Courtis, 1907; Miller, 1909). The specificity of these calls is known for over fifty years (Blair, 1955, 1958; Martof, 1961) and their taxonomic role has been assessed several times (Wells, 1977; Gerhardt, 1982; Cocroft and Ryan, 1995; Robillard et al., 2006). With the recent advances in technology and the popularization of high quality recorders, ethological studies of mating calls increased rapidly in the past decades (Gerhardt, 1998; Rand, 2001; Gerhardt and Huber, 2002; Bruyninckx, 2015). Moreover, the degrees of call variation within and between species is now documented Firenze University Press

40 38 Pedro C. Rocha et alii for several species (Castellano and Giacoma, 2000; Tárano, 2001; Castellano et al., 2002; Kaefer and Lima, 2012; Klymus et al., 2012; Carvalho et al., 2015; Jansen et al., 2016; Miranda et al., 2016), including the genus Ischnocnema Reinhardt and Lütken (i.e., I. guentheri and I. henselii Kwet and Solé, 2005; I. izecksohni Taucce et al., 2012). The genus Ischnocnema currently includes 33 species distributed in central and eastern Brazil and northern Argentina, most species being associated with the Brazilian Atlantic Forest (Canedo and Haddad, 2012; Frost, 2016). Although traditionally grouped into five species series (i.e., I. guentheri, I. lactea, I. parva, I. ramagii and I. verrucosa; Hedges et al., 2008), some species of the I. guentheri, I. lactea, I. parva and I. verrucosa series were clustered with different groups in the most recent phylogenetic study of the genus (Canedo and Haddad, 2012). Among these species, Ischnocnema abdita Canedo and Pimenta, 2010 and I. bolbodactyla (Lutz, 1925), which had previously been assigned to the I. lactea species series (Hedges et al., 2008; Canedo et al., 2010; Canedo and Pimenta, 2010), were relocated to the I. verrucosa series. Consequently, the I. lactea and I. verrucosa groups lost their morphological diagnoses (i.e., I. lactea species series was previously recognized by at least the outer digital discs of fingers moderate to large, whereas digital discs in species of the I. verrucosa series are small; Hedges et al. 2008). Currently, these two species series together are composed of 18 species (Canedo and Haddad, 2012; Padial et al., 2014), but the calls of only seven of them have been described up to date (i.e., Ischnocnema verrucosa series: I. bolbodactyla, I. juipoca (Sazima and Cardoso, 1978), I. penaxavantinho Giaretta et al., 2007; I. lactea series: I. nigriventris (Lutz, 1925), I. randorum (Heyer, 1985) and I. vizottoi Martins and Haddad, 2010). Ischnocnema abdita is a small-sized litter frog belonging to the I. verrucosa species series and only known for mountainous areas of southeastern Espírito Santo State, Brazil, in the Municipalities of Santa Teresa (type locality), Cariacica and Mimoso do Sul (Canedo and Pimenta, 2010; Canedo and Haddad, 2012). Aside from its calling microhabitat (i.e. hidden at the base of bushes), there is no further information on the natural history of this species. The present study aims to (1) report a newly discovered population and new State record for I. abdita from the region of Alto Caparaó, Municipality of Espera Feliz, Minas Gerais State, Brazil; (2) investigate the taxonomic identity of this population through the analysis of molecular assignment; (3) describe and compare the calls of the newly discovered population and those from the type locality (i.e., Municipality of Santa Teresa, State of Espírito Santo); and (4) evaluate the degree of variation in bioacoustic traits within and between populations. MATERIALS AND METHODS Hypothesis test on molecular assignment We sampled four individuals (MZUFV 15919, 15920, and 15923) from the Municipality of Espera Feliz, Serra do Caparaó region, Minas Gerais State (20 38 S, W, 921 m a.s.l.) and used the mitochondrial partial sequence of 16S rrna, amplified with the primers 16sAR (5 -CGCCTGTTTAT- CAAAAACAT-3 ; Palumbi et al., 1991) and 16sWilk2 (3 -GAC- CTGGATTACTCCGGTCTGA-5 ; Wilkinson et al., 1996), plus M13 tail. This marker was chosen based on its good performance as barcode for amphibians (Vences et al., 2005). The choice also considered availability of GenBank sequences for species of the I. verrucosa species group for comparative purposes. Our dataset comprised 17 sequences from five species, which also included samples of I. abdita from its type locality, plus an outgroup (I. izecksohni). Fragments were pre-aligned using ClustalW algorithm (Larkin et al., 2007) implemented with MEGA7 software (Kumar et al., 2016). Gaps open were penalized 10 times more than gaps extension (see Giribet and Wheeler, 1999) and final alignment was handmade. Establishment of primary homologies in the regions of ambiguous alignment of rrna is not trivial (Gillespie, 2004); therefore, we opted for the exclusion of one of these regions with 85pb and final alignment had 511pb. Since Ischnocnema abdita from its type locality showed a distinct haplotype from that of I. cf. abdita from Caparaó and Espera Feliz, we did not discard the distinct species hypothesis. Hence, we tested against the same species hypothesis. Intraspecific distances were estimated only to I. abdita and I. juipoca, since only one individual represented each of the other species. An ultrametric and full bifurcated mitochondrial gene tree was generated using a four-step procedure. In the first step, we excluded repeated haplotypes, maintaining only one copy. Then, we selected a model from a set of best models of DNA evolution estimated by jmodeltest for dataset (Darriba et al., 2012). The selection was made using the corrected Akaike information criteria, in which all models with ΔAICc < 2 were considered significantly supported (Burnham and Anderson, 2002; Burnham and Anderson, 2004). GTR+G was the secondbest model selected (ΔAICc = 0.497) and chosen to be applicable to all subsequent analyses. Third step was the estimation of a topology reliable and free of polytomies. For this purpose, we performed a maximum likelihood analysis on RAxML (Stamatakis, 2014) choosing the best tree under GTRGAMMA model, and performing a bootstrap support test with 1000 replicates. This analysis was repeated five times using distinct random seeds to verify topology, branch lengths and likelihood congruencies in estimated trees. One of these trees was chosen since all of them were virtually indistinguishable. In the last step, we made this gene tree ultrametric in MrBayes (Ronquist et al., 2012) leaving mutation rates analysis adjustable. The search was made under five independent runs, 10 Markov chains with default heating value and 10 6 generations each run. Trees were sampled each 5000 generations and the final consensus tree was calculated from the last 75% retained trees. The convergence of runs was graphically evaluated in Tracer 1.6

41 Calls of Ischnocnema abdita 39 software (Rambaut et al., 2014). This approach ensured a final tree with little variance in branch lengths. The hypothesis test per se was made submitting the ultrametric tree to a maximum likelihood general mixed Yulecoalescent test, implemented in GMYC (Fujisawa and Barraclough, 2013) using single threshold method. This method was designed to single-locus and is intended to identify the limit between tokogeny and phylogeny, fitting the branches in a gene tree under models about these relationship patterns. GMYC performs a log-likelihood ratio (LR) test of fitted multiple species model against a null hypothesis of one species in gene tree (Fujisawa and Barraclough, 2013). Hypothesis was validated though calculations of intraspecific and interspecific K-2p distances (Kimura, 1980) in MEGA7 software (Kumar et al., 2016), with taxonomical units identified by GMYC as references. We expected interspecific distances greater than intraspecific ones (Hebert et al., 2003). Bioacoustical analysis We analyzed 36 calls of four individuals (MNRJ 34902, 34904, and 34906; paratypes) recorded at Estação Biológica de Santa Lúcia, Municipality of Santa Teresa, Espírito Santo State, Brazil, type locality of Ischnocnema abdita (19 57 S, W; 650 m a.s.l.), on 14 and 16 January 2004, between 15:15 and 15:50. Calls were recorded with a Panasonic RQ-L31 cassette tape recorder coupled to a Leson SM-48 cardioid microphone. Vocalizations were digitized using software Avisoft SASLab Light, version 4.39, at a sampling rate of Hz and a resolution of 16 bits. We also recorded 95 calls of four individuals of Ischnocnema abdita from Serra do Caparaó region, Municipality of Espera Feliz, Minas Gerais State (20 38 S, W, 921 m a.s.l.), on 10 March Although we collected some calling individuals at this locality (i.e., MZUFV 15919, 15920, 15921, and 15923), we were unable to relate them to any sound recording. Recordings were made with a Tascam DR-40 digital recorder, at sampling rate of Hz and a resolution of 24 bits. Voucher specimens are deposited at Museu Nacional do Rio de Janeiro (MNRJ), Municipality of Rio de Janeiro, Rio de Janeiro State, and at Museu de Zoologia João Moojen (MZUFV), Universidade Federal de Viçosa, Municipality of Viçosa, Minas Gerais State, Brazil. Call measurements were made with the software Raven Pro 1.5 (Bioacoustics Research Program, 2014). Spectrogram was generated using window size = 512 samples, overlap = 70%; hop size = 3.49 ms; DFT size = 1024 samples and; and grid spacing = 43.1 Hz. Sound graphics were obtained using Seewave (Sueur et al., 2008) package of R platform (R Core Team, 2015) with the following settings: FFT = 512 samples and 70% overlap. Parameters measured were call duration (CD), call rate (CR), interval between calls (CI), number of notes per call (NN), note duration (ND), note rate (NR; given as notes/min), interval between notes (NI), number of pulses per note (PN), pulse rate (PR; given as pulses/sec), dominant frequency range (DF) and peak frequency (PF). Temporal parameters were measured directly from the oscillogram. Following Rocha et al. (2016), the DF represent the most energetic band of the call and is given in range from the lowest value of Frequency 5% to the highest value of Frequency 95%. The PF was acquired through the parameter Peak Frequency and represents the frequency that is coincident with the peak of energy within the call. Further call terminology follows that of Toledo et al., 2015b. Results are presented as mean ± standard deviation and range. We performed t-tests for comparative issues. Normality and Levene s homogeneity test of variance were conducted for all sets of variables. When variables distribution deviate from normality curve, we log 10 -transformed them (results preceded by *), and when variables had heterogeneous variances, we conducted tests with separate variance estimates (results preceded by ). All statistical tests were conducted through Statistica v Ischnocnema abdita is one of the species of the I. lactea and I. verrucosa series that have been reallocated between groups in recent papers (Heinicke et al., 2007; Hedges et al., 2008; Canedo and Haddad, 2012). Furthermore, the species I. manezinho Garcia, 1996 and I. sambaqui Castanho and Haddad, 2000, originally allocated tentatively in the I. lactea species series (Garcia, 1996; Castanho and Haddad, 2000), were not included in the phylogeny from Hedges et al. (2008) and are not currently assigned to any species series (Canedo and Haddad, 2012). Therefore, we decided to compare the call of I. abdita from the type locality with the species from both I. lactea and I. verrucosa groups plus the species I. manezinho and I. sambaqui. Moreover, the calls of I. penaxavantinho and I. sambaqui were described as a single multi-pulsed note emitted at irregular intervals (Giaretta et al., 2007; Castanho and Haddad, 2000; respectively). However, we considered their calls as sequences of notes (notes = temporally discrete vocalization units composing the call; sensu Toledo et al., 2015b) based on their resemblance with other calls described for Ischnocnema. Within-male variation Estimates of within-male variation were made through coefficient of variation [(CV = SD/Mean) x 100]. CV was calculated for each male and results are expressed as means for each trait following previous authors (Gerhardt, 1991; Tárano, 2001; Carvalho et al., 2013; Miranda et al., 2016). Call traits with low variability (i.e. usually less than 5%) were classified as static; and those with relatively high variability (i.e. usually more than 12%) were classified as dynamic. Parameters with values between 5% and 12% were considered intermediate (Gerhardt, 1991; Miranda et al., 2016). Molecular identification RESULTS The GMYC analysis returned six species, with a confidence interval from three to 10. Haplotypes of Ischnocnema abdita from its type locality, Caparaó and Espera Feliz were recovered under the same taxonomical entity (Fig. 1). The null hypothesis of no distinct species in the

42 40 Pedro C. Rocha et alii Fig. 1. Ultrametric gene tree of unique 16S haplotypes. GMYC identified lineages are alternating continuous and dashed branches. Numbers above nodes are the bootstrap supports estimated in maximum likelihood analysis. Information among parenthesis is the municipality where haplotypes were collected and GenBank accession numbers, respectively. ES = Espírito Santo State, MG = Minas Gerais State, SP = São Paulo State, RJ = Rio de Janeiro State. global gene tree, including the outgroup, was not rejected (LR test: P = 0.055). This is probably a consequence of the low number of individuals sampled per species (Fujisawa and Barraclough, 2013). The six recovered entities were I. izecksohni (outgroup), I. abdita, I. bolbodactyla, I. juipoca, I. cf. penaxavantinho and I. verrucosa (Fig. 1). Intraspecific distances of Ischnocnema abdita were (s.e. = 0.001, n = 6) and of I. juipoca ones (s.e. = 0.002, n = 7). The mean of interspecific distances, excluding outgroup, was ± (minimum between I. abdita and I. bolbodactyla; maximum between I. bolbodactyla and I. juipoca); such value was approximately 32 times greater than the mean of intraspecific distances. This result supports the GMYC assignment. Bioacoustical analysis Santa Teresa (type locality) A single type of call was observed in the recordings (Table 1). It is a call composed of a single type of pulsed note (Fig. 2A) emitted in sequences of 4 21 calls at mean rate of 6.2 ± 2.95 calls per min (CR = calls/min; n = 3 individuals) and intervals of 9.0 ± 5.1 sec between calls (CI = sec; n = 30 calls). Each note had three pulses on average (PN = 2.7 ± 0.4, 2 3) emitted at mean rate of 36.3 ± 4.3 pulses per second (PR = pulses/sec) and duration of 76.2 ± 13.5, ms (ND = CD). Dominant frequency ranged from 2781 to 4046 Hz with the peak frequency around 3500 Hz (PF = 3543 ± 126.9, Hz). Espera Feliz (Fig. 3) Recordings from this locality had two types of calls (Table 1). The most common type of call (hereafter referred as A call: 58.9%, n = 56 calls) was composed of a single type of pulsed note that is similar to the notes recorded at the type locality (Fig. 2B) and it is likely to be the advertisement call of Ischnocnema abdita (sensu Toledo et al., 2015b). This type of call was emitted in sequences of 1 16 calls at mean rate of 2.4 ± 1.1 calls per minute (CR = calls/min; n = 6), and interval between calls was 43.6 ± 16.3 (CI = sec; n = 39). Each note had mean duration of 41.5 ± 4.0 ms (ND = CD = ms) and two pulses on average (PN = 2.01 ± 0.13; 2 3) emitted at a mean rate of 48.6 ± 4.9 pulses per second (PR = pulses/sec). Dominant frequency ranged from 2670 to 4306 Hz, with peak frequency around 3200 Hz (PF = 3250 ± 163.5; Hz). The second and less common type of call (B call: 40.1%, n = 38) is usually composed of two pulsed notes, each note similar to the A call (Fig. 2C; Table 1). Although we also heard calls with three notes in the field, we were unable to record them. The B call was emitted in sequences of 1 13 calls emitted at mean rate of 2.5 ± 0.9 calls per minute (CR = calls/min; n = 3). Mean duration of B call was ± 34.7 ms (CD = ms) and interval between calls was 29.9 ± 14.5 seconds on average (CI = sec; n = 22). The first note was longer than the second one (ND 1 = 48.2 ± 13.6, ms; ND 2 = 35.7 ± 2.6; ms) and had two or three pulses, whilst the second note always had two pulses. First note had lower pulse rate than the second (PR 1 = 50.7 ± 6, ; PR 2 = 56.2 ± 4.3, pulses/ sec). Mean interval between notes was ± 30 ms (NI = ms) and note rate was 7.8 ± 1.04 notes per

43 Calls of Ischnocnema abdita 41 Table 1. Call traits of Ischnocnema abdita from the Municipality of Santa Teresa, Espírito Santo State, Brazil (n = 36 calls from four individuals) and from the Municipality of Espera Feliz, Minas Gerais State, Brazil (n = 95 calls from four individuals). See materials and methods section for trait acronym definitions. Santa Teresa Espera Feliz A call A call B call NN CD (ms) 76.2 ± 13.5 (54-101) n = ± 4.0 (33-49) n = ± 34.7 ( ) n = 39 CR (calls/min) CI (s) ND (ms) 6.2 ± 2.95 ( ) n = ± 5.1 ( ) n = ± 13.5 (54-101) n = ± 1.1 ( ) n = ± 16.3 ( ) n = ± 4.0 (33-49) n = 54 NR (notes/s) NI (ms) PN PR (pulses/s) DF (Hz) PF (Hz) 2.7 ± 0.4 (2-3) n = ± 4.3 ( ) n = n = ± ( ) n = ± 0.13 (2-3) n = ± 4.9 ( ) n = n = ± ( ) n = ± 0.9 ( ) n = ± 14.5 ( ) n = ± 13.6 (31-68) 35.7 ± 2.6 (30-42) n = ± 1.04 ( ) n = ± 30 ( ) n = ± 0.5 (2-3) 2 ± 0 (2-2) n = ± 6 ( ) 56.2 ± 4.3 ( ) n = n = ± ( ) 3454 ± ( ) n = 38 second on average (NR = notes/sec). Dominant frequency ranged from 2756 to 3962 Hz in the first note and from 2670 to 4134 Hz in the second one. However, there was no significant difference in the peak frequency between notes of B call. (PF 1 = 3462 ± 144.2, ; PF 2 = 3454 ± 129.4, Hz). Statistical analysis The A call from the type locality (i.e., most common type of call, composed of a single type of pulsed note) was emitted at a higher rate (*t = 2.587; df = 7; P = 0.036) and with shorter intervals than the calls recorded at Espera Feliz (t = ; df = 67; P < 0.001). Notes at the type locality had longer duration (t = ; df = 84; P < 0.001), one more pulse per note ( t = 9.099; df = 36.6; P < 0.001) and pulses emitted at lower rate than at Espera Feliz (t = ; df = 84; P < 0.001). Ultimately, the peak of energy was higher at the type locality ( t = 9.093; df = 73.1; P < 0.001). The B call (i.e., less common type of call, composed of two pulsed notes) was observed only in the recordings from Espera Feliz. Within the call, the first note was longer ( t = 5.584; df = 40.9; P < 0.001) and had a lower pulse rate (* t = 4.884; df = 66.2; P < 0.001). There were no significant differences in both notes peak frequency (t = 0.252; df = 74; P = 0.801). Despite the apparent similarity in the oscillogram and spectrogram of A and B calls from Espera Feliz (Fig. 2B C), we observed several differences between the notes of the different calls (Table 1). The B call is emitted

44 42 Pedro C. Rocha et alii Fig. 2. Sonogram (above) and oscillogram of unvouchered calls of Ischnocnema abdita from (A) Municipality of Santa Teresa, Espírito Santo State, Brazil; (B) A call and (C) B call from the Municipality of Espera Feliz, Minas Gerais State, Brazil.

45 Calls of Ischnocnema abdita 43 at similar rate than observed in the A call from Espera Feliz (t = 0.101; df = 7; P = 0.922), although with shorter intervals between calls (t = 3.272; df = 59; P = 0.001). The note from the A call has shorter duration than the first note from B call ( t = 2.982; df = 42.8; P < 0.005), and longer than the second note from B call ( t = 8.218; df = 89.8; P < 0.005). The note from the A call is composed of two pulses (three pulses observed once), whilst the first note from the B call has two or three pulses, and the second note was always composed of two pulses (similarly to the A call). On the other hand, mean pulse rate was the same for A call and first note of B call (t =1.808; df = 91; P = 0.07), but lower than that of the second note of B call ( t = 7.873; df = 85.8; P < 0.001). Ultimately, the peak frequency of A call was lower than those of both the first (t = 6.403; df = 89.0; P < 0.001) and the second ( t = 6.638; df = 88.1; P < 0.001) notes of the B call. Comparison between species Fig. 3. Unvouchered male of Ischnocnema abdita from Espera Feliz, Minas Gerais State, Brazil. Table 2. Within-male variation in the call traits of Ischnocnema abdita from the Municipality of Santa Teresa, Espírito Santo State, Brazil (n = 4 males) and from the Municipality of Espera Feliz, Minas Gerais State, Brazil (n = 4 males). Results are expressed as mean coefficient of variation (CV) followed by range in parenthesis. Locality - Call type Santa Teresa Espera Feliz - A call Espera Feliz - B call Call traits Average CV (%) Trait type Call duration 16.1 ( ) Dynamic Interval between calls 54.9 ( ) Dynamic Pulses per note 17.8 ( ) Dynamic Pulse rate 10.6 ( ) Dynamic Peak frequency 1.8 ( ) Static Call duration 5.3 ( ) Intermediate Interval between calls 31.5 ( ) Dynamic Pulses per note 0% Static Pulse rate 5.3 ( ) Intermediate Peak frequency 3.6 ( ) Static Call duration 9.1 ( ) Intermediate Interval between calls 49.6 ( ) Dynamic Note duration 7.0 ( ) 6.5 ( ) Intermediate Interval between notes 13.6 ( ) Dynamic Pulses per note 0% 0% Static Pulse rate 7.3 ( ) 6.6 ( ) Intermediate Peak frequency 2.2 ( ) 1.4 ( ) Static The high emission of the A call by Ischnocnema abdita, associated with its similarity to the advertisement call of I. bolbodactyla (sister species of I. abdita according to Canedo & Haddad, 2012), led us to compare this call with the advertisement calls described for the I. lactea and I. verrucosa groups, plus the species I. manezinho and I. sambaqui (see Materials and Methods for details). The A call of I. abdita is distinguished from the calls of I. juipoca, I. penaxavantinho, I. sambaqui and I. manezinho by having less notes per call (Table 3). It can be distinguished from the call of I. randorum by the shorter call duration and from those of I. nigriventris and I. vizzotoi by having a pulsed note structure. It can also be distinguished from the call of I. bolbodactyla by the longer note duration. Within-male variation To a certain degree, each call trait analyzed expressed variation (Table 2). Peak frequency never exceeded 4% variation, and was classified as static in both populations. Interval between calls had variations exceeding 65% in some individuals and was always classified as a dynamic trait. Call duration and number of pulses per note were classified as dynamic traits in the individuals from Santa Teresa. On the other hand, the number of pulses per note presented no variation (i.e., CV = 0%) in both types of call from Espera Feliz and were classified as static. Call duration and pulse rate were intermediate parameters in both types of call from Espera Feliz. Furthermore, duration of both notes was also an intermediate trait in the B call from Espera Feliz.

46 44 Pedro C. Rocha et alii Table 3. Comparative traits of the advertisement call of species from the Ischnocnema lactea and I. verrucosa species series, sensu Canedo and Haddad (2012)tRNA-Val, and 16S. See materials and methods section for definition of trait acronyms. *Cited as pulses. Species I. abdita I. bolbodactyla I. juipoca I. penaxavantinho I. randorum I. nigriventris I. vizzotoi I. sambaqui I. manezinho Species series I. verrucosa I. verrucosa I. verrucosa I. verrucosa I. lactea I. lactea I. lactea NN * * CD (ms) CR (calls/min) ND (ms) 76.2 ± 13.5 (54-101) 6.2 ± 2.95 ( ) 76.2 ± 13.5 (54-101) 32 ± 2 (29-37) ± 10.2 (38-72) 592 ± 29.7 ( ) ± 57.2 ( ) ± ± 2 (29-37) NI (ms) PN 2.7 ± 0.4 (2-3) DF (Hz) PF (Hz) 3543 ± ( ) Reference This paper ± 12 ( ) 3-4 Unpulsed Unpulsed 3-8 Unpulsed Unpulsed Unpulsed Unpulsed (first 5 notes) (last notes) Pombal & Cruz, 1999 Sazima & Cardoso, 1978 Giaretta et al., 2007 Heyer, 1985 Berneck et al., 2013 Martins & Haddad, 2000 Castanho & Haddad, 2000 Castanho & Haddad, 2000

47 Calls of Ischnocnema abdita 45 DISCUSSION We observed two distinct types of calls emitted by Ischnocnema abdita. The A call was observed in populations from both Santa Teresa (type locality) and Espera Feliz (new record) and was the most common type of call. Although we did not observe any behavior that could associate the described calls (i.e. A and B calls) with the attraction of mates, the high emission of the A call should possibly indicate that this type of call corresponds to the advertisement call of I. abdita (sensu Toledo et al., 2015b). Moreover, we observed significant differences in both temporal and spectral traits of the A call between localities. However, our molecular analysis supported the assignment of the newly discovered population to I. abdita in comparison with topotypical specimens. Our analysis on within-male variation showed that spectral traits are less variable than temporal traits in both populations. Variation in temporal traits is often related to environmental condition (e.g., temperature) and the social context of the call (e.g., presence of a female or other males), whereas the variation in the spectral traits is linked to the calling apparatus of a frog (Gerhardt, 1991; Gerhardt & Huber, 2002). On the other hand, spectral parameters are frequently associated with species recognition (i.e., static traits), whereas temporal traits (i.e., dynamic traits) have a greater influence on the meaning and attractiveness of the signal (Ryan and Rand, 1990; Gerhardt, 1991, 1992; Tárano, 2001; Gerhardt, 2005). Among the genus Ischnocnema, only three species had studies on intraspecific call variation (Kwet and Solé, 2005; Taucce et al., 2012; this paper) and they all found variation that agrees with previous studies (e.g., Castellano and Giacoma, 2000; Tárano, 2001; Castellano et al., 2002; Klymus et al., 2012; Carvalho et al., 2013; Grenat et al., 2013; Carvalho et al., 2015; Miranda et al., 2016). Intraspecific variability related to isolated and overlapping populations were described by Blair (1955) before Blair (1958) himself proposed the specificity of anuran mating calls. As stated by Castellano and Giacoma (2000), intraspecific call variation in frogs is likely the rule rather than the exception. Despite the great overlap in the dominant frequency ranges, we observed significant differences in the peak frequency from both populations. Variation in spectral traits is often associated with the calling apparatus of a frog (i.e., vocal fold). Within the same species, for instance, larger individuals are prone to have larger vocal folds, which causes dominant frequencies to be lower (Gerhardt & Huber, 2002). Thus, the differences observed are likely related to morphological variation between populations. On the other hand, there is no information on air temperature during the recordings and the cryptic behaviour of Ischnocnema abdita did not allow us to observe calling behaviour in the field. Therefore, we had no means of explaining the observed variation without further tests under the same environmental conditions and explicit social context. Under the phylogenetic hypothesis from Canedo and Haddad (2012) (see also Padial et al., 2014), a general bioacoustical pattern for Ischnocnema lactea and I. verrucosa series is unclear. The I. verrucosa series includes species whose calls are either composed of a single type of pulsed note (e.g., I. bolbodactyla) or multiple unpulsed notes (e.g., I. juipoca). The same is observed for the I. lactea series, with representatives that display advertisement calls composed of a single type of unpulsed notes (e.g., I. vizzotoi) and of sequences of pulsed notes (e.g., I. randorum). These results somewhat agree with the current lack of diagnostic morphological characters for those groups (Canedo and Haddad, 2012). This struggle to relate bioacoustical characters to synapomorphies is also observed in previous studies (e.g., Cannatella et al., 1998), whilst others have observed a strong phylogenetic signal in the calls (e.g., Robillard et al., 2006; Erdtmann and Amézquita, 2009). As shown by Robillard et al. (2006), the biomechanics involved in the sound production should also be taken into consideration in order to better understand homology between call units produced (see also McLister et al., 1995). Furthermore, our results point to intraspecific call variation that may be related to geographic and genetic structuration (see also Miranda et al., 2016). We encourage future researches to further investigate this, since geographic and genetic structuration may contribute to allopatric speciation (Turelli et al., 2001; Uyeda et al., 2009) and mating calls may evolve fast, leading to reproductive isolation in response to structuration (Panhuis et al., 2001). ACKNOWLEDGEMENTS We thank the Programa de Pós-graduação em Ciências Biológicas (Zoologia), Museu Nacional/UFRJ for offering the field course História Natural de Anfíbios - Campo in which the first recordings of this work took place; and the lecturers of the course: Prof. Dr. J.P. Pombal Jr. and Prof. Dr. C.A.G. Cruz. JVAL thanks Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG) for the fellowship grant (RDP ), CC thanks Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for fellowship grant and financial support (PAPD-RJ

48 46 Pedro C. Rocha et alii E-26/ /2011). RFM thanks CAPES for PhD followship grant. PCAG thanks Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) for research fellowship. We also thank Instituto Chico Mendes de Conservação da Biodiversidade (ICMBio) for the collecting permit number REFERENCES Bioacoustics Research Program (2014): Raven Pro: Interactive Sound Analysis Software (Version 1.5). Ithaca, NY: The Cornell Lab of Ornithology. Available from Blair, W.F. (1955): Mating Call and stage of speciation in the Microhyla olivacea-m. carolinensis complex. Evolution 9: Blair, W.F. (1958): Mating call in the speciation of Anuran Amphibians. Am. Nat. 92: Bruyninckx, J. (2015): Trading twitter: Amateur recorders and economies of scientific exchange at the Cornell Library of Natural Sounds. Soc. Stud. Sci. 45: Burnham, K.P., Anderson, D.R. (2002): Model selection and multimodel inference: a practical informationtheoretic approach. New York, NY: Springer-Verlag. Burnham, K.P., Anderson, D.R. (2004): Multimodel inference: understanding AIC and BIC in model selection. Sociol. Method. Res. 33: Canedo, C., Haddad, C.F.B. (2012): Phylogenetic relationships within anuran clade Terrarana, with emphasis on the placement of Brazilian Atlantic rainforest frogs genus Ischnocnema (Anura: Brachycephalidae). Mol. Phylogenet. Evol. 65: Canedo, C., Pimenta, B.V.S. (2010): New species of Ischnocnema, (Anura, Brachycephalidae) from the Atlantic Rainforest of the state of Espírito Santo, Brazil. South. Am. J. Herpetol. 5: Canedo, C., Pimenta, B.V.S., Leite, F.S.F., Caramaschi, U. (2010): New species of Ischnocnema (Anura: Brachycephalidae) from the State of Minas Gerais, Southeastern Brazil, with comments on the I. verrucosa species series. Copeia 2010: Cannatella, D. C., Hillis, D. M., Chippindale, P. T., Weigt, L., Rand, A. S., Ryan, M. J. (1998): Phylogeny of frogs of the Physalaemus pustulosus species group, with an examination of data incongruence. Syst. Biol. 47: Carvalho, T.R., Martins, L.B., Teixeira, B.F.V., Godinho, L.B., Giaretta, A.A. (2015): Intraspecific variation in acoustic traits and body size, and new distributional records for Pseudopaludicola giarettai Carvalho, 2012 (Anura, Leptodactylidae, Leiuperinae): implications for its congeneric diagnosis. Pap. Avulsos Zoo. 55: Carvalho, T.R., Teixeira, B.F.V., Martins, L.B., Giaretta, A.A. (2013): Intraspecific variability of the advertisement call of Chiasmocleis albopunctata (Anura: Microhylidae): Note structure as an additional diagnostic character within the genus. Herpetol. Notes. 6: Castanho, L.M., Haddad, C.F.B. (2000): New species of Eleutherodactylus (Amphibia: Leptodactylidae) from Guaraqueçaba, Atlantic Forest of Brazil. Copeia 3: Castellano, S., Cuattom B,, Rinella, R., Rosso, A., Giacoma, C. (2002): The advertisement call of the European treefrogs (Hyla arborea): a multilevel study of variation. Ethology 108: Castellano, S., Giacoma, C. (2000): Morphometric and advertisement call geographic variation in polyploid green toads. Biol. J. Linn. Soc. 70: Cocroft, R., Ryan, M.J. (1995): Patterns of advertisement call evolution in toads and chorus frogs. Anim. Behav. 49: Courtis, S.A. (1907): Response to Toads to Sound Stimuli. Am. Nat. 41: Darriba, D., Taboada, G.L., Doallo, R., Posada, D. (2012): jmodeltest 2: more models, new heuristics and parallel computing. Nat. Method. 9: 772. Erdtmann, L., Amézquita, A. (2009): Differential evolution of advertisement call traits in Dart Poison Frogs (Anura: Dendrobatidae). Ethology 115: Frost, D.R. (2016): Amphibian Species of the World: an Online Reference. Version 6.0 (accessed on 11 October 2016). Am. Museum Nat. Hist. New York, USA. index.html. Fujisawa, T., Barraclough, T.G. (2013): Delimiting species using single-locus data and the generalized mixed Yule-coalescent approach: a revised method and evaluation on simulated data sets. Syst. Biol. 62: Garcia, P.C.A. (1996): Nova espécie de Eleutherodactylus Duméril & Bibron, 1891 (Amphibia, Anura, Leptodactylidae) do Estado de Santa Catarina, Brasil. Biociências 4: Gerhardt, H.C. (1982): Sound pattern recognition in some North American treefrogs (Anura: Hylidae): Implications for mate choice. Integr. Comp. Biol. 22: Gerhardt, H.C. (1991): Female mate choice in treefrogs: Static and dynamic properties. Anim. Behav. 42: Gerhardt, H.C. (1992): Multiple messages in acoustic signals. Semin. Neurosci. 4:

49 Calls of Ischnocnema abdita 47 Gerhardt, H.C. (1998): Acoustic signals of animals: recording, field measurements, analysis and description. In: Animal acoustic communication, pp Springer Berlin Heidelberg. Gerhardt, H.C. (2005): Acoustic spectral preferences in two cryptic species of grey treefrogs: Implications for mate choice and sensory mechanisms. Anim. Behav. 70: Gerhardt, H.C., Huber, F. (2002): Acoustic communication in insects and anurans: common problems and diverse solutions. University of Chicago Press. Giaretta, A.A., Toffoli, D., Oliveira, L.E. (2007): A new species of Ischnocnema (Anura: Eleutherodactylinae) from open areas of the Cerrado Biome in southeastern Brazil. Zootaxa 1666: Gillespie, J.J. (2004): Characterizing regions of ambiguous alignment caused by the expansion and contraction of hairpin-stem loops in ribosomal RNA molecules. Mol. Phylogenet. Evol. 33: Giribet, G., Wheeler, W.C. (1999): On gaps. Mol. Phylogenet. Evol. 13: Grenat, P.R., Valetti, J.A., Martino, A.L. (2013): Intra-specific variation in advertisement call of Odontophrynus cordobae (Anura, Cycloramphidae): a multilevel and multifactor analysis. Amphibia-Reptilia 34: Hebert, P.D.N, Cywinska, A., Ball, S.L., Waard, J.R. (2003): Biological identifications through DNA barcodes. P. Roy. Soc. Lond. B Bio. 270: Hedges, S.B., Duellman, W.E., Heinicke, M.P. (2008): New World direct-developing frogs (Anura: Terrarana): molecular phylogeny, classification, biogeography, and conservation. Zootaxa 1737: Heinicke, M.P., Duellman, W.E., Hedges, S.B. (2007): Major Caribbean and Central American frog faunas originated by ancient oceanic dispersal. Proc. Natl. Acad. Sci. U.S.A. 104: Heyer, W.R. (1985): New species of frogs from Boracéia, São Paulo, Brazil. P. Biol. Soc. Wash. 98: Jansen, M., Plath, M., Brusquetti, F., Ryan, M.J. (2016): Asymmetric frequency shift in advertisement calls of sympatric frogs. Amphibia-Reptilia 37: Kaefer, I,L,, Lima, A.P. (2012): Sexual signals of the Amazonian frog Allobates paleovarzensis: geographic variation and stereotypy of acoustic traits. Behaviour. 149: Kimura, M. (1980): A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16: Klymus, K.E., Humfeld, S.C., Gerhardt, H. (2012): Geographical variation in male advertisement calls and female preference of the wide ranging canyon treefrog, Hyla arenicolor. Biol. J. Linn. Soc. 107: Kumar, S., Stecher, G., Tamura, K. (2016): MEGA7: Molecular Evolutionary Genetics Analysis Version 7.0 for Bigger Datasets. Mol. Biol. Evol. 33: Kwet, A., Solé, M. (2005): Validation of Hylodes henselii Peters, 1870, from Southern Brazil and description of acoustic variation in Eleutherodactylus guentheri (Anura: Leptodactylidae). J. Herpetol. 39: Larkin, M.A., Blackshields, G., Brown, N.P., Chenna, R., McGettigan, P.A., McWilliam, H., Valentin, F., Wallace, I.M., Wilm, A., Lopez, R., Thompson, J.D. (2007): Clustal W and Clustal X version 2.0. Bioinformatics 23: Lutz, A. (1925). Batraciens du Brésil. Comptes Rendus et Mémoires Hebdomadaires des Séances de la Société de Biologie et des ses Filiales. Paris 93: Martins, I.A., Haddad, C.F.B. (2010): A new species of Ischnocnema from highlands of the Atlantic Forest, Southeastern Brazil (Terrarana, Brachycephalidae). Zootaxa 2617: Martof, B.S. (1961): Vocalization as an isolating mechanism in frogs. Am. Midl. Nat. 65: McLister, D., Stevens, E.D., Bogart, J.P. (1995): Comparative contractile dynamics of calling and locomotor muscles in three hylid frogs. J. Exp. Bio. 198: Miller, N. (1909): The American Toad (Bufo lentiginosus americanus, LeConte) a study in dynamic biology. Am. Nat. 43: Miranda, R.B., Abrunhosa, P.A., Silva, R.H. (2016): Serenading for ten thousand years : The mating call of insular populations of the green treefrog Aplastodiscus eugenioi (Anura: Hylidae). Trop. Conserv. Sci. 9: Padial, J.M., Grant, T., Frost, D.R. (2014): Molecular systematics of terraranas (Anura: Brachycephaloidea) with an assessment of the effects of alignment and optimality criteria. Zootaxa 3825: Palumbi, S.R., Martin, A., Romano, S., McMillan, W.O., Stice, L., Grabowski, G. (1991): The Simple Fool s Guide to PCR, Version 2.0, privately published document compiled by S. Palumbi. Dept. Zoology, Univ. Hawaii, Honolulu, HI. Panhuis, T.M., Butlin, R., Zuk, M., Tregenza, T. (2001): Sexual selection and speciation. Trends Ecol. Evol. 16: R Core Team (2015): R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available from: Rambaut, A., Suchard, M.A., Xie, D., Drummond, A.J. (2014): Tracer (Version 1.6) [Computer software]. Available from

50 48 Pedro C. Rocha et alii Rand, A.S., (2001): A history of frog call studies 405 BC to In: Anuran communication. Smithsonian Institution Press, Washington. Robillard, T., Höbel, G., Gerhardt, C.H. (2006): Evolution of advertisement signals in North American hylid frogs: Vocalizations as end-products of calling behavior. Cladistics 22: Rocha, P.C., Thompson, J.R., Leite, F.S.F., Garcia, P.C.A. (2016): The advertisement call of Bokermannohyla flavopicta Leite, Pezzuti & Garcia, 2012 (Anura: Hylidae) from the mountains of Chapada Diamantina, Bahia, Brazil. Zootaxa 4061: Ronquist, F., Teslenko, M., van der Mark, P., Ayres, D.L., Darling, A., Höhna, S., Larget, B., Liu, L., Suchard, M.A., Huelsenbeck, J.P. (2012): MrBayes 3.2: efficient bayesian phylogenetic inference and model choice across a large model space. Syst. Biol. 61: Ryan, M.J., Rand, A.S. (1990): The sensory basis of sexual selection for complex calls in the tungara frog, Physalaemus pustulosus (sexual selection for sensory exploitation). Evolution 44: Sazima, I., Cardoso, A.J. (1978): Uma espécie nova de Eleutherodactylus do sudeste Brasileiro (Amphibia, Anura, Leptodactylidae). Rev. Bras. Biol. 38: Stamatakis, A. (2014): RAxML Version 8: A tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30: Sueur J., Aubin, T., Simonis, C. (2008): Seewave: a free modular tool for sound analysis and synthesis. Bioacoustics 18: Tárano, Z. (2001): Variation in male advertisement calls in the Neotropical frog Physalaemus enesefae. Copeia 2001: Taucce, P.P., Leite, F.S., Santos, P.S., Feio, R.N. Garcia, P.C.A. (2012): The advertisement call, color patterns and distribution of Ischnocnema izecksohni (Caramaschi and Kisteumacher, 1989) (Anura, Brachycephalidae). Pap. Avulsos Zool. 52: Toledo, L.F., Llusia, D., Vieira, C.A., Corbo, M., Márquez, R. (2015a): Neither convergence nor divergence in the advertisement call of sympatric congeneric Neotropical treefrogs. Bioacoustics 24: Toledo, L.F., Martins, I.A., Bruschi, D.P., Passos, M.A., Alexandre, C., Haddad, C.F. (2015b): The anuran calling repertoire in the light of social context. Acta Ethol. 18: Turelli, M., Barton, N.H., Coyne, J.A. (2001): Theory and speciation. Trends Ecol. Evol. 16: Uyeda, J.C., Arnold, S.J., Hohenlohe, P.A., Mead, L.S. (2009): Drift promotes speciation by sexual selection. Evolution 63: Vences, M., Thomas, M., Van der Meijden, A., Chiari, Y., Vieites, D.R. (2005): Comparative performance of the 16S rrna gene in DNA barcoding of amphibians. Front. Zool. 2: 1. Wells, K.D. (1977): The social behaviour of anuran amphibians. Anim. Behav. 25: Wilkinson, J.A., Matsui, M., Terachi, T. (1996): Society for the Study of Amphibians and Reptiles Geographic Variation in a Japanese Tree Frog (Rhacophorus arboreus) Revealed by PCR-Aided. J. Herpetol. 30:

51 Acta Herpetologica 12(1): 49-54, 2017 DOI: /Acta_Herpetol Feeding ecology of two sympatric geckos in an urban area of Northeastern Brazil José Guilherme G. Sousa 1, Adonias A. Martins Teixeira 2, *, Diêgo Alves Teles 2, João Antônio Araújo-Filho 2, Robson Waldemar Ávila 3 ¹ Programa de Pós-Graduação em Ecologia e Recursos Naturais, Departamento de Ciências Biológicas, Universidade Federal do Ceará, Campus Universitário do Pici, CEP , Fortaleza, Ceará, Brazil ² Programa de Pós-Graduação em Ciências Biológicas (Zoologia), Departamento de Sistemática e Ecologia DSE, Centro de Ciências Exatas e da Natureza CCEN, Universidade Federal da Paraíba UFPB, Cidade Universitária, Campus I, CEP , João Pessoa, PB, Brazil. *Corresponding author. adoniasteixeira01@gmail.com ³ Laboratório de Herpetologia, Departamento de Ciências Biológicas, Universidade Regional do Cariri, Rua Cel. Antônio Luiz, 1161, Campus do Pimenta, , Crato, Brazil; and Departamento de Ciências Biológicas, Universidade Federal do Ceará, Campus Universitário do Pici, CEP , Fortaleza, Ceará, Brazil Submitted on 2016, 31 st May; revised on 2016, 11 th November; accepted on 2016, 16 th December Editor: Sandra Hochsheid Abstract. The diets of two sympatric gecko species, Hemidactylus mabouia and Phyllopezus pollicaris, were studied from an urban area of the Crato municipality, Northeastern Brazil. While the house gecko H. mabouia is an introduced species widely distributed in North, Central, and South America, the Brazilian gecko P. pollicaris is a native species distributed along the great diagonal of open formations of South America. The diets of both species were mainly composed by arthropods, Diptera was the most important item for both species, corroborating others studies with lizards in urban areas. Male and female adults of both H. mabouia and P. pollicaris use similar microhabitats which can explain the high sexual and interspecific trophic niche overlap. In these populations from an urban area of the Crato municipality, the alien H. mabouia seems to have not negatively affected the trophic niche of the native P. pollicaris. Keywords. Trophic ecology, diet composition, Gekkota, Hemidactylus mabouia, invasive lizard, Phyllopezus pollicaris, body size. INTRODUCTION Gekkota is a species-rich clade of lizards with a great potential to invade new habitats, frequently introduced in anthropogenically disturbed areas and/or urban habitats (Hanley et al., 1998; Carranza and Arnold, 2006; Gamble et al., 2008). Its success in colonizing and establishing populations in new habitats is due in part to its great plasticity, which led to a wide distribution of geckos, especially Hemidactylus, in the old and new world (Rocha et al., 1994; Hanley et al., 1998; Vences et al., 2004; Gamble et al., 2008; Meshaka, 2011). The tropical house gecko Hemidactylus mabouia is native to Africa and has successfully colonized South, Central, and North America (Rocha et al., 2011). In Brazil, H. mabouia is frequently found in human-altered areas, but also in pristine habitats in the Amazon, Atlantic Forest, Cerrado, and Caatinga (Vanzolini, 1978; Vanzolini et al., 1980; Araújo, 1991; Rocha et al., 2000; Rocha et al., 2002; Rocha et al., 2011; Albuquerque et al., 2013). Several studies have demonstrated the role of Hemidactylus spp. in niche displacement of resident geckos by means of direct exploitative competition, indirect competition, aggression, and predation (Meshaka, 1995; ISSN (print) ISSN (online) Firenze University Press

52 50 J.G.G. Sousa et alii Meshaka and Moody, 1996; Meshaka, 2000; Meshaka et al., 2005; Hoskin, 2011; Hughes et al., 2015). Buildings in Brazil are colonized by many native gecko species, but the co-occurrence with H. mabouia caused exclusion by competition in some species, such as Thecadactylus rapicauda (Vitt and Zani, 1997). However, Phyllopezus pollicaris, a medium-sized gecko inhabiting boulder slabs in the dry formations of South America, is also frequently found in human habitations syntopic with H. mabouia (Vanzolini et al., 1980; Vitt, 1995; Sousa et al., 2010; Recoder et al., 2012). Plasticity of feeding habits, associated with ecological and biological conditions, has been considered the main cause of H. mabouia s invasion success (Zamprogno and Teixeira, 1998; Bonfiglio et al., 2006). Thus, diet studies can provide valuable information on the importance of prey types on the mechanism by which Hemidactylus geckos interact with other lizard species (Belver and Avila, 2001). Herein, we present data on the diets of H. mabouia and P. pollicaris and test if the invasive species has affected the native lizard, considering the trophic niche, in order to answer the following questions: i) is there sexual dimorphism between H. mabouia and P. pollicaris and between species? ii) how do the diets of these two geckos differ regarding composition? iii) what are the average niche breadths and the niche overlaps between these species? iv) is niche overlap between the studied species caused by chance? MATERIALS AND METHODS Lizard specimens were collected from March to November 2011 in human habitations in the city of Crato (7 14 S and W, datum WGS84), Ceará state, Northeastern Brazil. The regional climate is predominantly tropical, hot, and sub-humid (average annual temperatures vary from 24 to 26 C). The rainy period occurs between January and May, with an average annual rainfall of 1,091 mm (IPECE, 2013). The lizards were collected manually, humanely killed with lethal doses of 2% lidocaine, fixed in 10% formaldehyde, and preserved in 70% ethyl alcohol. The specimens were sacrificed for another study on parasitism which has been already published (Sousa et al., 2014). Snout-vent length (SVL), jaw width (JW), mouth length (ML), and head length (HL) of each lizard were measured using a digital caliper (± 0.01 mm; Table 1). Voucher specimens were deposited in the Coleção Herpetológica da Universidade Regional do Cariri - URCA. To test for morphological variation due to the sex of each species, between adults of H. mabouia and P. pollicaris in body shape (using all morphometric variables), and between each individual variable, we performed a multivariate discriminant function analysis with residuals of morphological variables of each sampled lizard species. Residuals were obtained from a simple linear regression following Sousa and Ávila (2015). Discriminant function analysis was executed using Statistica version 10.0 (Statsoft, 2011). Lizards were dissected in the laboratory and their stomach contents were analyzed under a stereomicroscope. Prey items were identified and classified to the order level; length and width were measured with a digital caliper (precision 0.01mm). The percentage of lizards with empty stomachs was calculated and only lizards with identifiable items were used. The volume of each prey item was estimated by the ellipsoid formula: V = 4 3 π L 2 W 2 2, where L = length and W = width. The Importance Value Index (I) was estimated by the following formula (Powell et al., 1990): I = F%+ N%+V%, 3 where F% is the relative frequency, N% is the numerical percentage, and V% is the volumetric percentage of each prey item. Trophic niche breadth (both numerical and volumetrically) was calculated by the inverse of the Simpson s diversity index (Simpson, 1949): n 2 B =1/ P i, where p is the numeric or volumetric proportion of prey category i and n is the number of categories (Pianka, 1973). The value of niche breadth varies from 1 to n, where the lowest values represent a more specialized diet and the highest values indicate a generalist diet. Differences in number and volume of prey items between males and females were tested by nonparametric Mann-Whitney U test using Statistica version 10.0 (Statsoft-Inc, 2011). Trophic niche overlap between sexes and species was calculated using the Pianka s overlap index (Pianka, 1973): ik = n i=1 n i=1 i=1 PijPik, (Pij 2 )(Pik 2 ) where Pij and Pik are the rate consumption of prey category i, with j and k representing the compared species and sexes. Pianka s overlap index varies from zero (no overlap) to one (complete overlap). Finally, we compared the observed niche overlap values for H. mabouia and P. pollicaris against a null model (1,000 interactions) generated by the randomization algorithm 3 (RA3) (Lawlor, 1980). The use of RA3 seems adequate because it retains the niche breadth of each species, but randomizes which particular resource states are used, allowing the species to potentially use other resource states (Winemiller and Pianka, 1990). Niche overlaps and null models were calculated using EcosimR - R code for null model analysis (Gotelli and Ellison, 2013). Differences between proportions of each prey category among the two gecko species were tested by a non-parametric

53 Feeding ecology of two sympatric geckos 51 Table 1. Morphometric data (mean ± standard deviation) and results of discriminant function analysis. Hm= Hemidactylus mabouia; Pp= Phyllopezus pollicaris. M= Males; F= Females; Body Shape= pooled variables; SVL= Snout-Vent Lenght; JW= Jaw width; ML= Mouth lenght; HL= Head lenght; λ = Wilks' Lambda. Hm Pp Hm Pp M F M F Adults Adults SVL 53.4 ± ± ± ± ± ± 11.6 JW 9.6 ± ± ± ± ± ± 2.0 ML 10.2 ± ± ± ± ± ± 2.1 HL 14.2 ± ± ± ± ± ± 2.7 λ (M vs F) P λ (M vs F) P λ (Hm vs Pp) Body shape <0.0000* SVL * JW * ML HL P multivariate analysis of similarity (ANOSIM), using the Bray- Curtis similarity coefficient and 9,999 permutations in the software PAST 3.0 (Hammer et al., 2001). A similarity percentage analysis (SIMPER) was performed to determine which preys were responsible for dissimilarity in diets between the two gecko species. Data matrix for ANOSIM and SIMPER were standardized (as a percentage) to minimize the discrepancy between samples. RESULTS We collected 58 H. mabouia, 10 adult males (mean ± standard deviation) (SVL = ± 9.76 mm), 21 adult females (SVL = ± 6.21 mm), and 27 juveniles (SVL = ± 5.36 mm). Of the species P. pollicaris, we collected 100 specimen, 38 adult males (SVL = ± 9.06 mm), 54 adult females (SVL = ± mm), and 8 juveniles (SVL =34.54 ± 4.32 mm). Of these, 18 (31.03%) H. mabouia (2 males, 7 females, and 9 juveniles) and 45 (45%) P. pollicaris had empty stomachs (12 males, 27 females, and 6 juveniles). There were no significant sexual differences in morphometric variables of both species (Table 1). On the other hand, adults of P. pollicaris were significantly larger than H. mabouia in body shape, with SVL and JW of P. pollicaris greater than those of H. mabouia (Table 1). The diet of both species consisted of 10 arthropod prey items: Diptera, Hymenoptera, and Coleoptera were the most important items for H. mabouia and Diptera, Coleoptera, and Hemiptera for P. pollicaris (Table 2). Hemidactylus mabouia had numerical and volumetrically niche breadth values of and 1.979, respectively, which was smaller than the niche breadth of P. pollicaris: 4.05 and (Table 2). The average number of prey items per stomach (only those individuals whose stomachs were not empty) was similar for both species: 3.86 ± 8.22 for H. mabouia and 3.47 ± 4.52 for P. pollicaris, with no significant difference (U = 1,017.5; P = 0.387). However, the average number of prey categories per stomach was 1.54 ± 0.61 and 1.24 ± 0.51 for H. mabouia and P. pollicaris, respectively, and there was a significant difference (U = 769.5; P = 0.014). Males and females of P. pollicaris ingested more prey items (4.26 ± 6.02 and 3.18 ± 3.07, respectively) than those of H. mabouia (2.2 ± 2.69 and 1.68 ± 1.21), but there were no statistical differences in prey items ingested by females of P. pollicaris and H. mabouia (U= 230.5; P= 0.836) neither between males of both species (U = 32; P= 0.081). Also, there were no significant differences in number (U = 117; P = 0.385) or volume (U = 111; P = 0.3) of the prey items between H. mabouia males and females, between P. pollicaris males and females (Number: U = ; P= 0.323; Volume U = 605; P = 0.25), or in the number between adult individuals of both species (U = 995; P = 0.21). On the other hand, there was a significant difference in volume between adults of P. pollicaris ( ± ) and H. mabouia and ( ± ) (U = 697; P < 0.001). Niche overlap between the sexes was in H. mabouia and in P. pollicaris; and between the two gecko species, niche overlap was The high niche overlap between adult individuals of both studied species could be due to chance (average overlap simulated = ; P = 0.99). There was no significant difference

54 52 J.G.G. Sousa et alii Table 2. Diet composition and numerical and volumetric niche breadth of Hemidactylus mabouia and Phyllopezus pollicaris from an urban area of Crato city, Ceará state, Brazil. n= total of analyzed individuals. F = frequency of occurrence, N = number of preys, V= volume (mm³) and I = relative importance index. Category H. mabouia (n = 58) P. pollicaris (n = 100) F F % N N% V V % I F F % N N% V V % I Araneae Blatodae Coleoptera Diptera H. mabouia shed skin Vertebrate eggshell Hemiptera Hymenoptera Isoptera Insect larvae Lepidoptera Ortoptera Total Niche breadth Empty stomachs between the proportions in prey use (ANOSIM, R = ; P = ). The results of the SIMPER analysis also showed a small dissimilarity between diet composition (32.69%), with Diptera, Hemiptera, and Coleoptera as prey groups that contribute most to dissimilarities between the two geckos, explaining 32.2, 26.7 and 11.1% of the variation. DISCUSSION The diets of the sympatric H. mabouia and P. pollicaris in urban habitats of Northeastern Brazil were very similar, with both species consuming 10 prey categories. Diptera, Coleoptera, Hemiptera, and Hymenoptera were the most frequently ingested preys; Diptera were the most important prey for both species. However, Hymenoptera and Coleoptera were the second most important prey for H. mabouia and P. pollicaris, respectively. These insects are commonly found in urban habitats, mainly close to streetlights (Robinson, 2005). Food composition of urban populations of lizards may differ from that of pristine habitats (Hódar and Pleguezuelos, 1999). Moreover, the opportunistic behavior of these two gecko species, plus changes in diversity and abundance of prey, may contribute to differences between study sites (Zamprogno and Teixeira, 1998). Vitt (1995), studying the diet of H. mabouia and P. pollicaris at Caatinga habitats from Exu, found that insect larvae were the most important prey for both species. In the coastal plain of the Espírito Santo state, Zamprogno and Teixeira (1998) found Araneae, Homoptera, and Isopoda as the most important prey for H. mabouia, while Apterygota, Araneae, and Orthoptera were the main items for P. pollicaris at Cerrado habitats in the state of Tocantins (Recoder et al., 2012). Albuquerque et al. (2013) found cannibalistic habits for H. mabouia, where H. mabouia was the most important item, followed by Formicidae and Hemiptera; for P. pollicaris, the most important items were Coleoptera, Araneae, and Homoptera in periantropic environments in the state Mato Grosso do Sul. Our results corroborate the study of Bonfiglio et al. (2006), in an urban area in Rio Grande do Sul, who found that Diptera were the most important item in the diets of the two species studied here. That convergence in diet composition may be due to the fact that Diptera is often present at urban environments, mainly close to streetlights as mentioned above. In this case, perhaps geckos could act as a control of Diptera in cities. Energetic balance in a given population of lizards can be estimated by the proportion of empty stomachs, where nocturnal lizards have a higher proportion of empty stomachs (24.1%) than diurnal lizards (10.5%); the mean percentage of empty stomachs for nocturnal geckos is 21.2% (Huey and Pianka 1981). In the present study, H. mabouia (31.01%) and P. pollicaris (42.6%) showed higher proportions of empty stomachs than stated by Huey and Pianka (1981). This may be explained by the time

55 Feeding ecology of two sympatric geckos 53 when the lizards were captured, i.e., in the first hours of the night. The lack of sexual differences in the diets of the species we studied is probably a result of similar use of space and time, which led to the use of the same prey types for both sexes and higher dietary overlap (see Rocha and Anjos, 2007). Dietary overlap was higher even when we paired both species (0.91), and the null model showed that high niche overlap between them is due to chance, indicating that the two species may eventually compete for prey. However, P. pollicaris presented higher prey volume and niche breadth than H. mabouia (probably due to its larger body) and despite the expected, H. mabouia was apparently not significantly affected by the trophic niche of P. pollicaris. Dissimilarity between the diets of these two geckos was about 30% in the present study, but this pattern can vary according to prey availability and/or greater time of colonization by H. mabouia. These factors can make this invasive species compete more intensely, damaging the spatial and/or trophic niche of native species (see Rodder et al., 2008; Rocha et al., 2011). Additionally, further studies are necessary to understand the role of competition for food and/or space in the coexistence of native and invader lizard species in disturbed and natural areas in Brazil. ACKNOWLEDGMENTS We thank the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior CAPES for providing a fellowship to JGGS, AAMT, and JAAF, the Conselho Nacional de Desenvolvimento Científico (CNPq) for a research grant to RWA (process # /2015-6) and a fellowship to DAT. We also thank Sandra Hochscheid and two anonymous reviewers for their valuable suggestions. The specimens were collected under the license of the Instituto Chico Mendes de Conservação da Biodiversidade-ICMBio REFERENCES Albuquerque, N.R.D., Costa-Urquiza, A.D.S., Soares, M.P., Alves, L.S., Urquiza, M.V.S. (2013): Diet of two sit-and-wait lizards, Phyllopezus pollicaris (Spix, 1825) (Phyllodactylidae) and Hemidactylus mabouia (Moreau de Jonnès, 1818) (Gekkonidae) in a perianthropic area of Mato Grosso do Sul, western Brazil. Biota Neotrop. 13: Araújo, A.F. (1991): Structure of a white sand-dune lizard community of coastal Brazil. Rev. Brasil. Biol. 51: Belver, L.C., Avila L.J. (2001): Ritmo de actividad diaria y estacional de Cnemidophorus longicaudus (Squamata: Teiidae: Teiinae) en el Norte de La Rioja, Argentina. Bol. Soc. Biol. Concepción Chile 72: Bonfiglio, F., Balestrin, R.L., Cappellari, L.H. (2006): Diet of Hemidactylus mabouia (Sauria, Gekkonidae) in urban area of southern Brazil. Biociências (On-line) 14: Carranza, S., Arnold, E. (2006): Systematics, biogeography, and evolution of Hemidactylus geckos (Reptilia: Gekkonidae) elucidated using mitochondrial DNA sequences. Mol. Phylogenet. Evol. 38: Gamble, T., Bauer, A.M., Greenbaum, E., Jackman, T.R. (2008): Evidence for Gondwanan vicariance in an ancient clade of gecko lizards. J. Biogeogr. 35: Gotelli, N.J., Ellison, A.M. (2013): EcoSimR. Version Computer software. EcoSim/EcoSim.html. (accessed on ). Hammer, Ø., Harper, D.A.T., Ryan, P.D. (2001): PAST: Paleontological Statistics Software Package for education and data analysis. Palaeontol. Electron. 4: 1-9. Hanley, K.A., Petren, K., Case, T.J. (1998): An experimental investigation of the competitive displacement of a native gecko by an invading gecko: no role for parasites. Oecologia 115: Hódar, J., Pleguezuelos, J. (1999): Diet of the Moorish gecko Tarentola mauritanica in an arid zone of southeastern Spain. Herpetol. J. 9: Hoskin, C.J. (2011): The invasion and potential impact of the Asian House Gecko (Hemidactylus frenatus) in Australia. Austral Ecol. 36: Huey, R.B., Pianka, E.R. (1981): Ecological consequences of foraging mode. Ecology 62: Hughes, D.F., Meshaka We, J.R., Van Buurt, G. (2015): The superior colonizing gecko Hemidactylus mabouia on Curaçao: Conservation implications for the native gecko Phyllodactylus martini. J. Herpetol. 49: IPECE. (2013): Crato. Anuário Estatístico do Ceará. Perfil Básico dos Municípios. (accessed on ). Meshaka We, J.R. (1995): Reproductive cycle and colonization ability of the Mediterranean gecko (Hemidactylus turcicus) in south-central Florida. Fla. Sci. 58: Meshaka We, J.R., Moody, B.A. (1996): The old world tropical housegecko (Hemidactylus mabouia) on the Dry Tortugas. Fla. Sci. 59: Meshaka We, J.R. (2000): Colonization dynamics of two exotic geckos (Hemidactylus garnotii and H. mabouia) in Everglades National Park. J. Herpetol. 34: Meshaka We, J.R., Smith, H.T., Severson, R., Severson, M.A. (2005): Spatial picture of a gecko assemblage in flux. Fla. Sci. 68:

56 54 J.G.G. Sousa et alii Meshaka We, J.R. (2011): A runaway train in the making: the exotic amphibians, reptiles, turtles, and crocodilians of Florida. Monograph 1. Herpetol. Conserv. Biol. 6: Pianka, E.R. (1973): The structure of lizard communities. Annu. Rev. Ecol. Syst. 4: Powell, R., Parmelee, J.S., Rice, M.A., Smith, D.D. (1990): Ecological observations of Hemidactylus brooki haitianus Meerwath (Sauria: Gekkonidae) from Hispaniola. Caribbean J. Sci. 26: Recoder, R., Junior, M.T., Camacho, A., Rodrigues, M.T. (2012): Natural history of the tropical gecko Phyllopezus pollicaris (Squamata, Phyllodactylidae) from a sandstone outcrop in Central Brazil. Herpetol. Notes 5: Robinson, W.H. (2005): Urban insects and arachnids: a handbook of urban entomology. Cambridge University Press. New York. Rocha, C.F.D., Anjos, L.A. (2007): Feeding ecology of a nocturnal invasive alien lizard species, Hemidactylus mabouia Moreau de Jonnès, 1818 (Gekkonidae), living in an outcrop rocky area in southeastern Brazil. Braz. J. Biol. 67: Rocha, C.F.D., Dutra, G., Vrcibradic, D., Menezes, V. (2002): The terrestrial reptile fauna of the Abrolhos Archipelago: species list and ecological aspects. Braz. J. Biol. 62: Rocha, C.F.D., Vrcibradic, D., Araújo, A., Esteves, F., Lacerda, L. (2000): Ecofisiologia de répteis de restingas brasileiras. In: Ecologia de Restingas e Lagoas Costeiras, pp Esteves, F.V., Lacerda, L.D., Eds, NUPEN-UFRJ, Macaé. Rocha, C.F.D., Anjos, L.A., Bergallo, H.G. (2011): Conquering Brazil: the invasion by the exotic gekkonid lizard Hemidactylus mabouia (Squamata) in Brazilian natural environments. Zoologia 28: Rocha, C.F.D., Nascimento, L.B., Bernardes, A.T., Cotta, G.A. (1994): Introdução a ecologia de lagartos brasileiros. Herpetologia no Brasil 1. Pontifícia Universidade Católica de Minas Gerais/PUC-MG, Belo Horizonte. Rödder, D., Solé, M., Böhme, W. (2008): Predicting the potential distributions of two alien invasive Housegeckos (Gekkonidae: Hemidactylus frenatus, Hemidactylus mabouia. North-West J. Zool. 4: Simpson, E.H. (1949): Measurement of diversity. Nature 163: 688. Sousa, J.G.G., Ribeiro, S.C., Roberto, I.J., Teles, D.A., Almeida, W.O. (2010): Ocorrência de pentastomídeos (Metameria: Ecdysozoa) no lagarto Phyllopezus pollicaris (Spix, 1825). Cad. Cult. Cienc. 2: Sousa, J.G.G., Brito, S.V., Ávila, R.W., Teles, D.A., Araujo-Filho, J.A., Teixeira, A.A.M., Anjos, L.A., Almeida, W.O. (2014). Helminths and Pentastomida of two synanthropic gecko lizards, Hemidactylus mabouia and Phyllopezus pollicaris, in an urban area in Northeastern Brazil. Braz. J. Biol. 74: Sousa, J.G.G., Ávila, R.W. (2015). Body size, reproduction and feeding ecology of Pleurodema diplolister (Amphibia: Anura: Leiuperidae) from Caatinga, Pernambuco state, Northeastern Brazil. Acta Herpetol. 10: Statsoft, Inc. (2011): Statistica for Windows [Computer program manual]. Versão Tulsa: StatSoft. Vanzolini, P.E. (1978): On South American Hemidactylus (Sauria, Gekkonidae). Arq. Zool. Estado São Paulo 17: Vanzolini, P.E., Ramos-Costa, A.M.M., Vitt, L.J. (1980): Réptéis das Caatingas. Academia Brasileira Ciéncias, Rio de Janeiro. Vences, M., Wanke, S., Vieites, D.R., Branch, W.R., Glaw, F., Meyer, A. (2004): Natural colonization or introduction? Phylogeographical relationships and morphological differentiation of house geckos (Hemidactylus) from Madagascar. Biol. J. Linn. Soc. 83: Vitt, L.J. (1995): The ecology of tropical lizards in the caatinga of northeast Brazil. Oklahoma Museum of Natural History, University of Oklahoma. Vitt, L.J., Zani, P.A. (1997): Ecology of the nocturnal lizard Thecadactylus rapicauda (Sauria: Gekkonidae) in the Amazon region. Herpetologica 53: Zamprogno, C., Teixeira, R. (1998): Hábitos alimentares da lagartixa de parede Hemidactylus mabouia (Reptilia, Gekkonidae) da planície litorânea do norte do Espírito Santo, Brasil. Braz. J. Biol. 58:

57 Acta Herpetologica 12(1): 55-63, 2017 DOI: /Acta_Herpetol A pattern-based tool for long-term, large-sample capture-markrecapture studies of fire salamanders Salamandra species (Amphibia: Urodela: Salamandridae) Jeroen Speybroeck 1, *, Koen Steenhoudt 2,$ 1 Research Institute for Nature and Forest, Kliniekstraat 27, 1070 Brussels, Belgium. *Corresponding author. jeroenspeybroeck@ inbo.be, jeroen.speybroeck@hotmail.com 2 Waterschaapstraat 30, 1570 Galmaarden, Belgium $ Programming and database development was solely done by K.S. Submitted on January, 11 th 2017; revised on March, 12 th 2017; accepted on March, 22 nd 2017 Editor: Uwe Fritz Abstract. Solid population studies depend on reliable capture-mark-recapture (CMR) methodology. The available methods for such studies on amphibians are often invasive, unsuitable for long-term studies, time-consuming and/ or expensive. We present a new software tool, connected to a MS Access database, ManderMatcher, for CMR study of fire salamanders (Salamandra salamandra and related species) by means of a robust matching algorithm using 44 pattern characteristics. Metadata related to standardised counts (e.g., weather and lunar variables) as well as a myriad of individual sighting variables can be entered and saved as well. Application of the presented method to a database of 9,397 sighting records gathered over a period of eight years, as well as a random sample validation, demonstrate the accuracy of the applied matching algorithm. Differences with other methods are discussed. The program is made freely available for download and widespread application is encouraged, especially given the contemporary context of a fungal disease threatening survival of fire salamander populations. Keywords. Salamandra salamandra, pattern recognition, capture-mark-recapture, software, methodology, free download INTRODUCTION Long-term demographic studies depend on reliable capture-mark-recapture (CMR) methodology to model population size and abundance, survival and detection rate, as well as to provide data on a myriad of life history features such as individual growth, reproductive cycles, longevity, dispersal and migration (e.g. Ferner, 1979; Schmidt et al., 2002; Schmidt, 2004; Amstrup et al., 2005). A number of methods are available for practical execution of such studies in amphibians. Toe clipping as well as the use of tattoos or passive or active transponders, however, all require a certain degree of (at best temporary) damage to the subject s body, and are thus commonly labelled as invasive and, at least potentially, damaging (e.g. Davis and Ovaska, 2001; Le Galliard et al., 2011). Furthermore, given the regenerative capacity of certain taxa, namely salamanders, toe clipping becomes unsuitable for long-term studies (e.g., Ferner, 1979; Davis and Ovaska, 2001) and may in certain cases even reduce recapture and survival rates (McCarthy et al., 2009). Use of transponders sets additional limitations such as being costly and fairly time-consuming in the field (Ott and Scott, 1999), as well as in some cases having low retention rates in smaller individuals and species (Ryan et al., 2014), while detection and recapture rates may differ ISSN (print) ISSN (online) Firenze University Press

58 56 Jeroen Speybroeck, Koen Steenhoudt according to the size of the used tags (Ousterhout and Semlitsch, 2014). Species with an individually varying pattern of colour spots offer the possibility of recognising animals without the need to apply invasive techniques. A number of (more or less) automated image analysis tools has been developed, such as WildID (Bolger et al., 2012), AmphIdent (Drechsler et al., 2015), I³S (Van Tienhoven et al., 2007, with newer, improved versions having been successfully applied to the newt Triturus carnifex by Sannolo et al., 2016, as well as to reptiles, as reviewed by Sacchi et al., 2016), APHIS (Moya et al., 2015) etc. While these claim to offer more robust recognition procedures and automatic pattern recognition, they usually require labour-intensive image processing steps in order to select a certain area of interest. Thus, they do not necessarily guarantee faster processing than alternatives based on interpretative pattern coding. One (AmphIdent) is even fairly expensive, especially for non-institutional users. Additionally, while pattern shape becomes of greater importance, spot shapes may alter drastically given the position of the animal within the picture. As such, higher demands are set on the quality of the pictures, resulting in more time spent properly positioning live and fairly hard to immobilise animals while collecting data. The use of non-invasive recognition techniques to study fire salamander Salamandra salamandra evolved throughout the last decades, making use of the typical pattern of yellow spots on a black background in this species. Earlier authors used time-consuming, direct comparison of photographs to recognise individuals (e.g., Feldmann, 1971; Klewen, 1985; Seifert, 1991). Kopp- Hamberger (1998) coded different S. salamandra pattern types into an alphanumerical code, while not ruling out occurrence of type doubles, thus causing obvious problems in larger samples. Carafa and Biondi (2004) elaborated on this by combining a colour typology with a software application storing data into an MS Access database for the Italian subspecies S. s. gigliolii, using a sample of 233 individual animals. Given the limited description of the applied characteristics in their paper and the fact that the software was not made freely available, use of and detailed insight into this methodology remained unavailable to a wider audience. A different approach was presented by Šukalo et al. (2013), using the number of glandular pores located within the boundaries of the yellow spots. In many subspecies of S. salamandra, however, the number, size and shape of these spots is known to change throughout an individual s lifetime (Eiselt, 1958; Klewen, 1991; Mutz, 1992; Bogaerts, 2002). Becoming stable at adulthood only (Klewen, 1991), pattern changes may be common, limiting indeed the use of colour patterns for individual recognition at earlier life stages. While the degree of colour pattern alteration during early terrestrial life may vary considerably between S. salamandra subspecies (Bogaerts, 2002), a relatively stable pattern is said to be reached after at least two or three years (Eiselt, 1958). This leads to changes in pore count within the span of a single year in some individuals, even in animals larger than 12 cm (Speybroeck, unpublished data). Thus, the method of Šukalo et al. (2013) seems of limited value for studies spanning more than a couple of months at most, and a methodology combining a multitude of characteristics seems preferable. As such, use of an individual-specific code which can be applied to easily collected and processed images, which is sufficiently stable over time and can be queried by an automated matching algorithm applied to a database, is of merit. We present a new software tool, connected to a MS Access database, ManderMatcher, for CMR study of Salamandra salamandra and other spotted species of the genus by means of a robust matching algorithm using 44 pattern characteristics. We demonstrate how this tool can serve study of large sample sizes involving a high number of different individuals. MATERIAL AND METHODS Salamandra salamandra and related spotted species (S. algira, S. corsica and S. infraimmaculata) are characterised by a black background colour covered with a variable amount of most commonly yellow (but in certain taxa rather orange or red) spots or stripes (Thiesmeier, 2004; Beukema et al., 2016b; Speybroeck et al., 2016). The top of the head is usually covered with yellow on the eyes and parotoid glands, while additional spots may be present on the snout tip. The tail may be crossed by spots as well. Toes may be black or yellow. This pattern layout offers potential for coding each animal s pattern into a series of numbers. These numbers can be supplied to an algorithm that matches each pattern with pattern codes from previously collected data. Pattern code The spot pattern is coded into 44 characteristics. Each of those indicates the number, presence or mutual relation between the yellow spots within well-defined areas of the salamander body: head (11 characteristics), back and tail (19) and toes (14). In case of doubt, characteristics can be left blank and will not be taken into account by the matching algorithm. As patterns may vary over time, reliable use is advised for animals with a total length of more than 14 cm only. Matching may work on smaller animals, depending on the stability of that animal s pattern as well as the portion of characteristics that are left blank by the user. Thus, while no reliable assessment of subadult and juvenile population size can be made with this

59 A pattern-based tool for CMR studies 57 (or any) pattern-based method, it may allow certain other studies which do not require tracing every single individual (e.g. growth analysis). In contrast to criteria used in some of the other cited methods, each of the 44 characteristics has a clear definition, allowing a limited array of entered values only and leaving little room for variation caused by interpretation. All are discussed and illustrated in the manual of the freely available program, serving as supplementary material to this paper (available for download, along with the program itself, at hylawerkgroep.be/jeroen/index.php?id=85). For demonstration to new users, a version containing already some data is also provided. The ManderMatcher website also offers data exploration tools for data stored in the program's database). Program A software tool, ManderMatcher, was developed to allow CMR recognition and storage of other visit and sighting related data. The program was written in the VB.NET (Visual Basic) language, using the freely available basic version of Visual Studio Express. Practical use is discussed at length in the manual (see above link). The program basically consists of three main windows. The first is the Visits window, available after launch as the left tab (Fig. 1). It lists the visits and allows to enter metadata related to that visit. Visits are time-restricted count events, covering e.g. a standardised transect or habitat surface. The second window is the Sightings window (Fig. 2), available as the second tab, next to the Visits window. The pivot animal, shown at the upper left, will be matched with the entire database of available sightings at the right on the basis of the pattern code that was entered for it. At the lower left, criteria can be specified to filter and restrict the database entries which are returned at the lower right. The as such selected entries are sorted by their resemblance to the pivot animal, with the most similar entry at the top. The third window is accessed through the buttons for editing one of the two sightings shown at either the upper left or Fig. 1. Visits window for edit and creation of visits, and entry of environmental data in the program ManderMatcher.

60 58 Jeroen Speybroeck, Koen Steenhoudt Fig. 2. Sightings window for recapture recognition and database querying in the program ManderMatcher. Fig. 3. Window for entry of individual sighting records in the program ManderMatcher.

61 A pattern-based tool for CMR studies 59 the upper right. This window is used for entry of new sightings, as well as for editing of existing sightings. At the left, the CMR pattern code is entered, while the upper fields allow to store several other variables. Hitting the OK button at the right after data entry/edit leads back to the Sightings window and initiates the matching process. In the cited available methods (see Introduction), a list of observations ordered by decreasing match likelihood is the endpoint of the matching process. Unlike any other, however, our method offers filter criteria to reduce this list (by entering features which are considered to be stable) as an additional tool to detect recaptures. This is particularly useful for matching younger specimens, which are likely to have undergone certain pattern changes. Additional variables Besides CMR matching, the program offers several tools which are relevant for ecological study of salamanders. A first set involves data related to the individual sighting, based on characteristics of the animal. We refer for a full list to the provided download link and the manual of the program, and only provide a general outline here. Biometric fields are available weight and length measurements (total length and snout-vent length). The latter can be measured within the program itself if a calibration object is properly inserted into the picture. Ecological aspects can be documented in the shape of discrete variables to discern patterns in e.g. activity and habitat use throughout salamander life history. Several fields allow further specification of details concerning the encounter location. All individual-related data is entered through the individual sighting entry window (Fig. 3). A second set of variables, which are entered through the Visits window (Fig. 1), consists of metadata related to the count event (visit), including an array of environmental variables which can be registered both at start and end of the count event. All are defined and discussed in the program s manual. Database All data is saved into a MS Access database, which offers ample opportunity to query the entered data and export for analysis. Application The program was applied to S. s. terrestris data collected in a forest in the town of Merelbeke, Flanders, Belgium. Sampling was performed along a standardised transect of 1.3km. From March 2008 to January 2017, 169 sampling events were executed by night during activity-provoking weather conditions. All encountered post-metamorphic fire salamanders were captured, photographed and released. Validation To validate the applied search algorithm, a random sample of 100 sightings of adult salamanders was drawn from the database. One-by-one visual photograph comparison was conducted. Application RESULTS A total of 9,397 sighting records were collected during standardised transect sampling events. Of these, Fig. 4. Recapture rate by cumulative number of sighting records of adult animals, by visit (circles) and for entire accumulating dataset (triangles) with non-parametric, locally weighted regression smoother (loess) and 95% confidence interval. Point size relates to absolute number of encountered adults within each visit (encounter event) for both point types (circles and triangles), presented as a continuous variable.

62 60 Jeroen Speybroeck, Koen Steenhoudt 6,899 are sightings of animals larger than 14 cm, here treated as adults. These adult sightings consist of (re) captures of 1,477 different individual adult salamanders. The overall percentage of animals captured at least twice, as obvious from the growing, cumulative database is shown as a function of the growing total of adult individual observations by the triangles in Figure 4, each data point representing a capture event. The cumulated dataset recapture rate amounts to 62.4% in the entire database at the final instance of data collection. The circles in the graph, however, not showing the percentage of recaptures in the cumulated database but the percentage within the data of a single visit, reach far higher values and, for instance, correspond to 91% of the adult individuals captured during the last visit (at the extreme right in the graph). Validation The random sample comprised 89 adult individuals, of which 80 were encountered only once within the sample, 7 were encountered twice, and 2 individuals were encountered three times. The exact same result was obtained by application of the program, even though matching of these 100 randomly selected sightings was done against up to 9,397 records, thus making successful match finding more complicated than if 100 new records were to be entered into an empty database. While matching through individual picture comparison took about 4 hours, matching 100 new records one by one with an unstructured collection of 9,397 photographs would take an impractical amount of time. While the speed of using ManderMatcher differs depending on the size of the database and recapture rate within the data, using the matching algorithm and finding a match or not (excluding entry of other data) takes a user which is sufficiently acquainted with the program an average of 2 minutes or less per record. The uniqueness of certain code combinations can be calculated theoretically and from our data on S. s. terrestris. As an example, we consider the toe characteristics. Fourteen toe characteristics (with potential values 0 and 1) corresponds to 2 14 or 16,384 theoretical combinations. Using our data to assess the number of actual combinations, we used a subset of 1,597 observations for which all toe characteristics were determined. These observations relate to 737 individual animals. Within these 737 individuals, 488 different toe colour combinations occur, of which 52.2% are found in a single individual only. The most commonly encountered combination (i.e., all black toes) is shared by 5.4% of the individuals. DISCUSSION While the actual recapture numbers shown for the example population greatly depend on the applied count methodology and, possibly, the nature of the investigated population, they serve as illustration that with sufficient effort high recapture success can be achieved. Manual validation showed no mismatches in comparison with application of the algorithm. After the program s initial release, as substantiated by this paper, tools for exploratory analysis are under development (Speybroeck and Steenhoudt, in prep.). These include analysis of population dynamics through space and time and mapping features, but also analysis of pattern variability, allowing the recognition variables to serve a second purpose, apart from CMR matching. Several non-invasive methods exist for CMR recognition of individuals. Some may require a lot of time in the field and/or behind the desk. We have demonstrated how ManderMatcher allows effective matching of flexiblebodied fire salamanders at a swift rate, requiring limited handling time per record in the field, no photograph processing at the desk and easy and fast record entry. During a high number of recapture events (169), the method was successfully applied to a dataset of 9,397 records and 1,477 adult individuals, which is clearly larger than any to which other methods were applied (e.g., 1,606 records of 1,321 individuals in Drechsler et al., 2015; 852 photos of 324 individuals in Sannolo et al., 2016). More importantly, the dataset spans a period of eight years. This is particularly relevant to assess how changes of pattern over time may affect recognising individual animals. Using data from a single season does not offer assurance that the quality of the matching process remains sufficiently high over time. For example, Sannolo et al. (2016) used data collected over a four month period of a single year. While these authors claim the ventral spot pattern of Triturus carnifex to be stable over time (citing Arntzen and Wallis, 1999, who however do not provide proof for this statement), this seems unlikely, as Drechsler et al. (2015) showed this to be incorrect for the closely related T. cristatus. Only by using data from a period of time which is long enough to span a relevant portion of adult life and growth of the considered taxon (e.g., three years in Drechsler et al. 2015, eight years in our study), the impact of pattern changes (whether they exist and/ or are relevant or not) on the use of the recognition algorithm can be assessed. While the dorsal pattern of adult fire salamanders is traditionally considered to remain stable over time (e.g., Eiselt, 1958; Klewen, 1991), changes may occur (Speybroeck, unpublished data). Application of ManderMatcher to data collected over eight years has

63 A pattern-based tool for CMR studies 61 shown that the algorithm is sufficiently robust to prevent such changes from causing matching problems, which is likely to be an advantage of using numeric coding of the pattern, instead of image analysis tools basing matching on the extent of pigmentation. ManderMatcher does not require photographs to be taken in a standard way (e.g., with the animal perfectly stretched), which is particularly time-consuming when dealing with salamanders, such as the species we studied, and newts, including the subject species of Drechsler et al. (2015) and Sannolo et al. (2016). Finally, in contrast to all other tools, our program is a comprehensive research tool, including tools for length measurement and a myriad of variables related to each individual and each record. While the pattern coding was so far only applied to S. s. terrestris, the character definitions offer sufficient precision and resolution for use with most spotted taxa within the Salamandra genus (S. s. salamandra, S. s. almanzoris, S. s. terrestris, S. s. longirostris, S. s. bejarae, S. s. morenica and most S. s. bernardezi) as well as S. algira, S. corsica and S. infraimmaculata). Two exceptions, however, may exist. Populations with a high portion of animals with poorly defined spots (e.g., populations of S. s. gallaica, S. s. crespoi, and part of S. s. bernardezi (i.e., populations formerly assigned to S. s. alfredschmidti but placed in synonymy by Beukema et al., 2016a)) may offer coding difficulties, whereas taxa with a high portion of animals with low numbers of spots (e.g., S. s. fastuosa, S. s. gigliolii) will have identical values for a number characteristics across many individual animals. However, considering the frequency distribution of actual toe colour combinations in our data, we have demonstrated how powerful the algorithm may be. Only 5.4% of the individuals share the most common combination, allowing to drastically reduce the number of individuals to be compared by using the toe characteristics alone. Further use will proof applicability to other taxa/populations, and it seems likely that for some of the taxa with less welldefined dorsal spots (e.g., S. s. gallaica) toe colour may provide sufficient discerning power. The program is made freely available for use by any researcher, while the authors applaud getting in touch with them for potential collaboration, applying the same methodology across fire salamander taxa. While this method is presented for a single taxon (as was the case for the initial publication of most other methods, e.g., Van Tienhoven et al., 2007; Drechsler et al., 2015), the presented approach has potential to be used to develop similar tools for use with other species, applying a modified code with relevant characteristic definitions. It is particularly useful for species which have a pattern that can be coded to a sufficient number of characteristics. How high this number has to be in order to allow precise matching, depends on the sample size and the frequency distribution of possible values of each characteristic. As such, using a smaller number of characteristics may be sufficient for smaller datasets. For the species at hand, population studies may serve as a crucial baseline at this point in time, given the imminent threat presented by lethal fungal infection caused by Batrachochytrium salamandrivorans (Martel et al., 2013, 2014). After decimating the Dutch population (Spitzen-van der Sluijs et al., 2013), severe fungus-caused mortality has occurred at several sites in Belgium and the fungus has also been detected in Germany (Spitzen-van der Sluijs et al., 2016). Thus, the availability of an easy-to-use, inexpensive CMR methodology for study of both infected and non-infected populations is of great value. ACKNOWLEDGEMENTS Wouter Beukema provided valuable comments to an earlier draft of the manuscript. The Flemish Agency for Nature and Forest (Agentschap Natuur en Bos, ANB) is thanked for providing the necessary permits. REFERENCES Amstrup, S.C., McDonald, T.L., Manly, B.F.J. (2005): Handbook of Capture-Recapture Analysis. Princeton University Press, Princeton. 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: Beukema, W., Nicieza, A., Lourenço, A., Velo-Antón, G. (2016a). Colour polymorphism in Salamandra salamandra (Amphibia: Urodela), revealed by a lack of genetic and environmental differentiation between distinct phenotypes. J. Zool. Syst. Evol. Res. 54: Beukema, W., Speybroeck, J., Velo-Antón, G. (2016b). Quick Guide Salamandra. Curr. Biol. 26: R689-R700. Bogaerts, S. (2002): Farbkleidentwicklung bei einigen Feuersalamanderarten und unterarten. Amphibia 1: 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: Carafa, M., Biondi, M. (2004): Application of a method for individual photographic identification during a

64 62 Jeroen Speybroeck, Koen Steenhoudt study on Salamandra salamandra gigliolii in central Italy. Ital. J. Zool. Supp. 2: 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: Drechsler, A., Helling, T., Steinfartz, S. (2015): Genetic fingerprinting proves cross-correlated automatic photoidentification of individuals as highly efficient in large capture-mark-recapture studies. Ecol. Evol. 5: Eiselt, J. (1958): Der Feuersalamander Salamandra salamandra (L.), Beiträge zu einer taxonomischen Synthese. Abh. Ber. Naturk. Vorgesch. Magdeburg 10: Feldmann, R. (1971): Felduntersuchungen an westfälischen Populationen des Feuersalamanders, Salamandra salamandra terrestris LACÉPÈDE, Dortmunder Beitr. Landeskunde, Naturw. Mitt. 5: Ferner, J.W. (1979): A review of marking techniques for amphibians and reptiles. Soc. for the Study of Amphibians and Reptiles Herp. Circ. 9: Klewen, R. (1985): Untersuchungen zur Ökologie und Populationsbiologie des Feuersalamanders (Salamandra salamandra terrestris LACÉPÈDE 1788) an einer isolierten Population im Kreise Paderborn. Abh. Westf. Mus. f. Naturkunde 47: Klewen, R. (1991): Die Landsalamander Europas. Ziemsen Verlag, Stuttgard. Kopp-Hamberger, M. (1998): Eine Methode zur individuellen Erkennung von Feuersalamandern (Salamandra salamandra terrestris) abhand des Zeichnungsmusters. Salamandra 3: Le Galliard, J.-F., Paquet, M., Pantelic, Z., Perret, S. (2011): Effects of miniature transponders on physiological stress, locomotor activity, growth and survival in small lizards. Amphibia-Reptilia 32: Martel, A., Spitzen-van der Sluijs, A., Blooi, M., Bert, W., Ducatelle, R., Fisher, M.C., Woeltjes, A., Bosman, W., Chiers, K., Bossuyt, F., Pasmans, F. (2013): Batrachochytrium salamandrivorans sp. nov. causes lethal chytridiomycosis in amphibians. Proc. Natl. Acad. Sci. USA. 110: Martel, A., Blooi, M., Adriaensen, C., Van Rooij, P., Beukema, W.M., Fisher, M.C., Farrer, R.A., Schmidt, B.R., Tobler, U., Goka, K., Lips, K.R., Muletz, C., Zamudio, K.R., Bosch, J., Lötters, S., Wombwell, E., Garner, T.W.J., Cunningham, A.A., Spitzen-van der Sluijs, A., Salvidio, S., Ducatelle, R., Nishikawa, K., Nguyen, T.T., Kolby, J.E., Van Bocxlaer, I., Bossuyt, F., Pasmans, F. (2014): Recent introduction of a chytrid fungus endangers Western Palearctic salamanders. Science 346: 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: Moya, Ó., Pep-Luis, M., Sergio, M., José-Manuel, I., Andreu, R., Antonio, R., Giacomo, T. (2015): APHIS: A new software for photo-matching in ecological studies. Ecol. Inform. 27: Mutz, T. (1992): Vergleich der Farbkleidentwicklung bei Unterarten des Feuersalamanders. Urodela Info 3: 4-5. Ott, J.A., Scott, D.E. (1999): Effects of toe-clipping and PIT-tagging on growth and survival in metamorphic Ambystoma opacum. J. Herpetol. 33: Ousterhout, B.H., Semlitsch, R.D. (2014): Measuring terrestrial movement behavior using passive integrated transponder (PIT) tags: effects of tag size on detection, movement, survival, and growth. Behav. Ecol. Sociobiol. 68: Ryan, K.J., Zydlewski, J.D., Calhoun, A.J.K. (2014): Using passive integrated transponder (PIT) systems for terrestrial detection of blue-spotted salamanders (Ambystoma laterale) in situ. Herpetol. Conserv. Biol. 9: 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. Dodd, K.C., Ed, Oxford University Press, Oxford. Sannolo, M., Gatti, F., Mangiacotti, M., Scali, S., Sacchi, R. (2016): Photo-identification in amphibian studies: a test of I³S Pattern. Acta Herpetol. 11: Schmidt, B.R., Schaub, M., Anholt, B.R. (2002): Why you should use capture-recapture methods when estimating survival and breeding probabilities: on bias, temporary emigration, overdispersion, and common toads. Amphibia-Reptilia 23: Schmidt, B.R. (2004): Declining amphibian populations: the pitfalls of count data in the study of diversity, distributions, dynamics, and demography. Herpetol. J. 14: Seifert, D. (1991): Untersuchungen an einer ostthüringischen Population des Feuersalamanders (Salamandra salamandra). Artenschutzreport Heft 1: Speybroeck, J., Beukema, W., Bok, B., Van Der Voort, J. (2016): Field guide to the amphibians and reptiles of Britain and Europe. Bloomsbury Publishing, London. Spitzen-van der Sluijs, A., Martel, A., Asselberghs, J., Bales, E.K., Beukema, W., Bletz, M.C., Dalbeck, L., Goverse, E., Kerres, A., Kinet, T., Kirst, K., Laudelout, A., Marin da Fonte, L.F., Nöllert, A., Ohlhoff, D., Sabino-Pinto, J., Schmidt, B.R., Speybroeck, J., Spikmans, F., Steinfartz, S., Veith, M., Vences, M., Wagner, N.,

65 A pattern-based tool for CMR studies 63 Pasmans, F., Lötters, S. (2016): Expanding distribution of lethal amphibian fungus Batrachochytrium salamandrivorans in Europe. Emerg. Infect. Dis. 22: Spitzen-van der Sluijs, A., Spikmans, F., Bosman, W., De Zeeuw, M., Van der Meij, T., Goverse, E., Kik, M., Pasmans, F., Martel, A. (2013): Rapid enigmatic decline drives the fire salamander (Salamandra salamandra) to the edge of extinction in the Netherlands. Amphibia-Reptilia 34: Šukalo, G., Dorđević, S., Golub, D., Dmitrović, D., Tomović, L. (2013): Novel, non-invasive method for distinguishing the individuals of the fire salamander (Salamandra salamandra) in capture-mark-recapture studies. Acta Herpetol. 8: Thiesmeier, B. (2004): Der Feuersalamander. Laurenti Verlag, Bielefeld. Van Tienhoven, A.M., Den Hartog, J.E., Reijns, R.A., Peddemors, V.M., (2007): A computer-aided program for pattern-matching of natural marks on the spotted raggedtooth shark Carcharias taurus. J. Appl. Ecol. 44:

66

67 Acta Herpetologica 12(1): 65-77, 2017 DOI: /Acta_Herpetol Skeletal variation within the darwinii group of Liolaemus (Iguania: Liolaemidae): new characters, identification of polymorphisms and new synapomorphies for subclades Linda Díaz-Fernández*, Andrés S. Quinteros, Fernando Lobo Instituto de Bio y Geociencias del NOA, Consejo Nacional de Investigaciones Científicas y Técnicas, Universidad Nacional de Salta, Av. 9 de Julio 14, 4405 Rosario de Lerma, Salta, Argentina *Corresponding author. lindadiazfernandez@gmail.com Submitted on: 2016, 19 th August; revised on: 2016, 27 h October; accepted on: 2017, 12 th March Editor: Aaron M. Bauer Abstract. Fifty-five skeletal characters (continuous and discrete) were analyzed for species of the L. darwinii group: L. albiceps, L. chacoensis, L. grosseorum, L. irregularis, L. koslowskyi, L. lavillai, L. ornatus, L. quilmes, plus L. inacayali (L. telsen group) and L. scapularis (L. wiegmannii group). We report polymorphic intraspecific variation that has not previously been taken into account and we describe 21 new characters that provide original information across the group. We detected several morphological synapomorphies for the darwinii group and subclades. The enclosure of Meckel s cartilage by a dentary outgrowth on lingual side of lower jaw (a synapomorphy of the subgenus Liolaemus sensu stricto and of the Phymaturus patagonicus group) also occurs within the L. darwinii group. The morphology of maxillary teeth with three conspicuous cusps may be a potential synapomorphy of the subgenus Eulaemus. The morphology of maxillary teeth may have adaptive value. Characters that were studied in other groups of lizards were informative for Liolaemus. Keywords. Cranial skeleton, postcranial skeleton, L. boulengeri group, evolution. INTRODUCTION Osteological characters are used for phylogenetic analyses and reconstructions and for investigating environmental adaptations because skeletons exhibit a very wide range of morphological variation at supraspecific levels. Detailed examples that illustrate their use in systematic and phylogenetic studies within Iguania are found in Etheridge (1965, 1967), de Queiroz (1987), Estes et al. (1988), Etheridge and de Queiroz (1988), Lang (1989) and Frost and Etheridge (1989), among others. Although this information was primarily used in phylogenetic reconstructions, it is also important for future works in studies of comparative biology, as it can document evolutionary changes of characters in the context of any hypothesized selective regime. Lizards of the Family Liolaemidae extend from the Andes through Bolivia, Peru and Chile to the coast of Tierra del Fuego in Argentina (Donoso-Barros, 1966; Cei, 1986, 1993; Lobo and Quinteros, 2005; Abdala, 2007). Currently, this family has 298 species (Uetz, 2016) and comprises three genera: Ctenoblepharys, Liolaemus, and Phymaturus (Etheridge, 1995; Frost et al., 2001; Schulte et al., 2003; Espinoza et al., 2004). Within the subgenus Eulaemus of Liolaemus, there is the species-rich L. boulengeri group (Etheridge, 1995; Abdala, 2007), which is characterized by having a group of enlarged scales at the back of the thigh, so it is also called the patch group. As a result of taxonomic and phylogenetic studies of the L. boulengeri group, numerous subgroups within it have been proposed (Avila et al., 2006; Abdala, 2007). One of these is the Liolaemus darwinii clade, which was ISSN (print) ISSN (online) Firenze University Press

68 66 L. Díaz-Fernández, A.S. Quinteros, F. Lobo defined by Etheridge (1993) based on the presence of posterior teeth with straight-edged crowns and marked sexual dichromatism in which males exhibit a more colorful dorsal color pattern than females. This group was inferred to be monophyletic in different analyses, based on morphological and/or molecular characters (Abdala, 2007; Avila et al., 2006; Fontanella et al., 2012; Olave et al., 2014). Following Abdala s total evidence hypothesis (2007), the L. darwinii group consists of two clades (L. grosseorum and L. ornatus clades) and of five species basal to these ones (L. abaucan, L. espinozai, L. koslowskyi, L. quilmes and L. uspallatensis). The L. ornatus clade comprises 6 species (L. albiceps, L. crepuscularis, L. calchaqui, L. irregularis, L. lavillai and L. ornatus), which have a wide distribution in the Puna, Montes de Sierras y Bolsones and the northern part of the ecoregion of Monte de Llanuras y Mesetas of Argentina (Burkart et al., 1999). Lizards in this group are distinguished by being viviparous and having a large number of precloacal pores in females. Information about osteological features of liolaemid lizards appears in detailed descriptions of the head skeleton of Liolaemus lutzae, L. occipitalis, and L. signifer (Beurman and Vieira, 1980; Simoes-Lopes and Krause, 1988). In addition, studies of the appendicular skeleton of L. occipitalis (Keller and Krause, 1986) and the skeleton of L. lutzae and L. multiformis simonsii (Beurman and Vieira, 1980) have been published. Osteological character states for Liolaemus and Phymaturus in the context of iguanian phylogenetic analysis at the generic level were recorded by Etheridge and de Queiroz (1988). Lobo and Abdala (2001; 2002) described the variation found in the skeleton of 24 species. They demonstrated the phylogenetic information contained in those characters, recovering main clades and subclades of Liolaemus formally recognized in the literature. Additional skeletal characters were reported recently by Núñez et al. (2003) in the description of two new taxa (L. manueli and L. torresi). Da Silva and Verrastro (2007) described the axial skeleton of L. arambarensis and González-Marín and Hernando (2013) described the postcranial osteology of L. azarai. In the present contribution we report new characters and polymorphisms and we provide additional informative characters for the Liolaemus darwinii group and subclades therein. We report the evolution of some characters of special interest such as the enclosure of Meckel s cartilage (a character traditionally studied in iguanian lizards), the morphology of maxillary teeth, the cartilaginous extremity of cervical rib IV, and the bladelike process on the posterior distal tibia mentioned by Etheridge (1995) and Lobo and Abdala (2001). MATERIALS AND METHODS A total of 30 adult specimens of the Liolaemus darwinii group were studied (Appendix A). The species included were L. albiceps, L. chacoensis, L. grosseorum, L. irregularis, L. koslowskyi, L. lavillai, L. ornatus, and L. quilmes. We also included skeletons of L. scapularis (representing the L. wiegmannii group) and L. inacayali (L. telsen group) for outgroup comparisons, and examined characters described in Lobo and Abdala (2001, 2002) for 24 taxa belonging to different groups of Liolaemus. We studied specimens deposited in biological collections and collected some additional specimens of L. ornatus, L. scapularis, and L. quilmes, which were sacrificed by injection with 10% sodium pentothal. They were fixed in 10% formalin and finally preserved in 70% ethanol. They are deposited in the herpetological collection of the Museo de Ciencias Naturales of the Universidad Nacional de Salta (MCN), Instituto de Bio y Geociencias del NOA, and the Fundación Miguel Lillo (FML). The skeletons were prepared following the technique of differential staining of bone and cartilage (Wassersug, 1976). This allows the observation of cartilaginous structures that are not visible in the dry skeletons. Measurements were taken using digital calipers, 0.01 mm precision, under a stereoscopic microscope. In the case of smaller structures, the measurements were taken using a micrometer eyepiece. Nomenclature of bones, processes and foramina follow de Queiroz (1982), Keller and Krause (1986), Frost (1992), Etheridge (2000), Lobo and Abdala (2001), Lobo (2001, 2005) and Torres-Carvajal, (2004). In total, 55 characters, 19 continuous and 36 discrete, of the cranial and postcranial skeleton were examined. The discrete characters were coded as non-polymorphic binary, polymorphic binary, non-polymorphic multistate, and polymorphic multistate. We add our data matrix to that of Abdala s (2007) Total Evidence analysis and performed a new phylogenetic analysis. Character evolution mapping and the optimization of new characters were performed using TNT (Goloboff et al., 2003) over the resulting tree (same topology as recovered by Abdala 2007; Fig. 1). We follow Abdala (2007) because it includes the taxa studied here (instead of Fontanella et al., 2012 or Olave et al., 2014; whose studies lacked many of the species of interest). Data on diet were taken from the literature (Aun and Martori, 1998; Espinoza et al., 2004; Semhan et al., 2013). RESULTS All morphological variation observed in the skeletons of Liolaemus is summarized in the following list of char-

69 Skeletal variation in the Liolaemus darwinii group 67 acters, indicating the state of the character in each case. Variation of discrete and continuous characters is presented in Tables 1 and 2, respectively. Updated list of osteological characters (modified from Lobo and Abdala, 2002) 1. Number of scleral ossicles: (0) 15; (1) 14; (2) 13. Polymorphic multistate. 2. Bones forming parietal foramen: (0) formed mainly by frontal bone; (1) mainly by parietal bone; (2) both bones participate approximately equally. Polymorphic multistate. 3. Shape of parietal foramen: (0) with regular edges; (1) with irregular edges. Polymorphic binary. 4. Ceratohyal process: (0) gradually widened; (1) abruptly widened; (2) hook-shaped. Polymorphic multistate. 5. Distal ending of ceratobranchial II: (0) narrow; (1) widened. Polymorphic binary. 6. Anterior process of arytenoid: (0) reaches the level of the anterior process of the cricoid; (1) does not reach that level. Polymorphic binary. 7. Number of tracheal rings. Continuous (Table 2). 8. Number of incomplete tracheal rings / total number of rings. Continuous (Table 2). 9. Number of pterygoid teeth. Continuous (Table 2). 10. Number of maxillary teeth. Continuous (Table 2). 11. Number of modified anterior maxillary teeth (heterodonty): in most species the anterior-most teeth of the maxilla are conical, elongated, and exhibit only one cusp. Continuous (Table 2). 12. Maxillary tooth morphology I (Fig. 2): (0) crowns with their anterior and posterior margins divergent, expanded crowns; (1) crowns with anterior and posterior margins straight. Non-polymorphic binary. 13. Maxillary tooth morphology II: (0) crowns without differentiated cusps, (1) three conspicuous cusps (all species studied here). Species of the L. nigromaculatus group (subgenus Liolaemus sensu stricto) were reported as having broad maxillary teeth without secondary cusps (Lobo, 2001). Non-polymorphic binary. 14. Meckel s groove (Fig. 3): (0) open; (1) fused, Meckel s cartilage is hidden by an extensive outgrowth of the dentary bone. This last character state was described as an apomorphy of the L. chiliensis group (sub- Table 1. Discrete osteological characters and their variation within the Leiolaemus darwinii group. Intraspecific variation indicated by in brackets surrounding the relevant carácter states. Polymorphism in characters 1, 4, and 20, have not previously been reported L. albiceps [02] 0 [01] 1 0 [01] [01] [12] 0 L. chacoensis [12] [02] [01] [01] [01] [01] 0 [12] 0 L. grosseorum [01] [01] 1 [12] 0 L. inacayali [12] [01] L. irregularis 1 0 [01] 1 0 [01] [01] [01] L. koslowskyi [12] [012] [01] 1 [01] [01] 1 [01] 0 [01] 1 [12] 0 L. lavillai [01] L. ornatus 1 [12] [123] 0 L. quilmes 1 [02] 1 [01] 0 [01] [01] 1 0 [01] [01] L. scapularis 2 [02] [01] [12] L. albiceps [01] 0 1 [012] 0 [01] [01] L. chacoensis [01] [01] 0 2 [01] [01] 1 1 [01] [01] L. grosseorum [01] L. inacayali L. irregularis [01] [12] [01] 0 0 L. koslowskyi [01] [01] [01] L. lavillai [01] L. ornatus [01] [01] [01] 1 0 L. quilmes [02] 1 [01] [01] [01] [01] [01] 1 0 L. scapularis [01] [01] 0 0

70 68 L. Díaz-Fernández, A.S. Quinteros, F. Lobo Table 2. Continuous osteological characters (see text for character descriptions). Above: min-max. below: mean + standard deviation. Chars Liolaemus albiceps (N=4) Liolaemus chacoensis (N=4) Liolaemus grosseorum (N=2) Liolaemus inacayali (N=1) Liolaemus irregularis (N=4) Liolaemus koslowskyi (N=3) Liolaemus lavillai (N=2) Liolaemus ornatus (N=4) Liolaemus quilmes (N=5) Liolaemus scapularis (N=2) (61; 6) (52; 0) (61; 9) (62; 9) (59; 6.3) (55; 7.8) (55; 10.6) (52; 5) (56; 2.1) (0.3; 0.04) (0.3; 0.1) (0.2; 0.01) (0.3; 0.03) (0.2; 0.05) (0.2; 0.1) (0.3; 0.08) (0.3; 0.07) (0.2; 0.06) (7; 1.4) (17; 1) 0-0 (0; 0) 3-7 (4; 2) (3.3; 0.8) (0.8; 0.1) (0.21; 0.03) (2.2; 0.5) (0.05; 0.02) (0.13; 0.06) (0.29; 0.06) (0.08; 0.04) (0.84; 0.09) 8-7 (7.25;0.5) -4 (5.7;1.5) 6-6 (6; 0) (21; 0.8) 0-4 (2.3; 1.7) (17.6; 2) 0-0 (0; 0) 2-4 (3; 1.4) (0.4; 0.04) (0.7; 0.08) (0.21; 0.02) (1.8; 0.3) (0.04; 0.04) (0.19; 0.07) (0.29; 0.03) (0.09; 0.03) (0.82; 0.17) 6-3 (4.7;1.3) 4-4 (4;0) 5-6 (5.7; 0.5) (20.5; 1) 0-3 (1.5; 2.1) (14; 1.4) 0-0 (0; 0) 2-4 (3; 1.4) (0.7; 0.04) (0.2; 0.02) (3.1; 0.5) (0.02; 0) (0.1; 0.04) (0.34; 0) (0.1; 0) (0.78; 0.02) 5-3 (4;1.4) 5-5 (5;0) 6-6 (6; 0) (19; 1.4) 0.7? (0.05; 0) (0.19; 0) (0.27; 0) (0.07; 0) 1-1 (1; 0) 6-6 (6; 0) 4-4 (4;0) 6-6 (6; 0) (17; 0) 5-6 (5.8; 0.7) (18; 0) 0-0 (0; 0) 3-4 (3.5; 0.6) (0.38; 0.02) (0.8; 0.06) (0.27; 0.03) (2.3; 0.3) (0.05; 0.01) (0.11; 0.07) (0.34; 0.03) (0.08; 0.03) (0.88; 0.14) 6-8 (6.7; 0.9) 4-6 (4.5; 1) 6-6 (6; 0) (23; 1.1) 3-5 (4.3; 1.2) (16; 0) 0-0 (0; 0) 2-3 (2.7; 0.6) (0.36; 0.02) (0.8; 0.08) (0.24; 0.02) (2.6; 0.9) (0.06; 0.04) (0.12; 0.07) (0.31; 0.03) (0.08; 0.01) (0.66; 0.14) 5-12 (9;3.6) 6-3 (4.67;1.63) 4-6 (4.7; 1.1) (20; 0) 4-4 (4; 0) (14; 0) 0-0 (0; 0) 3-4 (3.5; 0.7) (0.39; 0.01) (0.78; 0.04) (0.26; 0.02) (2.6; 0.05) (0.05; 0.01) (0.09; 0.02) (0.33-0) (0.07; 0.01) (0.84; 0.04) 5-8 (6.5; 2.1) 4-5 (4.5; 0.71) 6-6 (6; 0) (18; 0) 2-6 (3; 2) (16; 0) 0-0 (0; 0) (5; 0.4) (0.37; 0.01) (0.76; 0.01) (0.29; 0.01) (2.7; 0.9) (0.05; 0.02) (0.11; 0.06) (0.28; 0.05) (0.09; 0.01) (0.86; 0.07) 7-3 (4.7;1.7) 8-7 (7.25;0.5) 5-6 (5.25; 0.5) (20.75; 0.5) 1-3 (2.6; 0.9) (17; 1.1) 1-1 (1; 1) 2-2 (2; 0) (0.39; 0.02) (0.78; 0.05) (0.24; 0.04) (2.3; 0.5) (0.05; 0.01) (0.14; 0.06) (0.35; 0) (0.06; 0) (0.86; 0.03) 5-5 (5;0) 7-7 (7;0) 5-6 (5.75; 0.5) (19; 1.4) 2-4 (3; 1.4) (13; 1.4) 2-3 (2.5; 0.7) 3-4 (3.5; 0.7) (0.4; 0.04) (0.21; 0.02) (2.1; 0.5) (0.04; 0.02) (0.13; 0.06) (0.35; 0.04) (0.08; 0.02) (0.88; 0.06) 5-11 (6.6;2.5) 5-4 (4.8;0.5) 3-6 (15.4; 1.3) (19.8; 1.6) genus Liolaemus) and it is the condition exhibited by the Phymaturus patagonicus group according to Etheridge (1995), Lobo et al (its Fig. 7D). Nonpolymorphic binary. 15. Cervical rib III: this rib, when present, is very small and remains cartilaginous. (0) present; (1) absent. Polymorphic binary. 16. Cartilaginous extremity of cervical rib IV (Fig. 4A-B): (0) bifurcated; (1) not bifurcated. Polymorphic binary. 17. Number of postxiphisternal elongated ribs: According to Etheridge (1995), Liolaemus species exhibit the most common pattern of postxiphisternal ( inscriptional ribs posterior to xiphisternals ), with their free endings bearing an elongated cartilage. Continuous (Table 2). 18. Posterior process of the sternum: (0) present; (1) absent (all species studied here). State (0) reported only for the L. pictus group by Lobo and Abdala (2001). Non-polymorphic binary.

71 Skeletal variation in the Liolaemus darwinii group 69 Fig. 1. Phylogenetic relationships recovered for the Liolaemus darwinii group. Fig. 3. Evolution of Meckel s groove within the Liolaemus darwinii group. (A) Character states optimized on the recovered topology, (B) Meckel s groove of lower jaw closed, (C) open Meckel s groove (C). Scale = 2 mm. Fig. 2. Evolution of morphology of maxillary crowns and diet within the Liolaemus darwinii group. (A) Character states optimized on the topology recovered, (B) maxillary teeth with straight crowns, (C) maxillary teeth with expanded crowns. Dietary data for L. inacayali and L. lavillai were not found in the literature. Scale = 1 mm. 19. Clavicles: (0) without fenestra; (1) with fenestra. Polymorphic binary. 20. Sternal fenestra (Fig. 4C-D) located in the posterior half of the sternum over the posterior half of interclavicle: (0) single; (1) divided, as described by Etheridge (2000) for species of the L. wiegmannii group. Polymorphic binary. 21. The posterior end of hipoischium: (0) expanded; (1) unexpanded. Polymorphic binary. 22. Bladelike process on posterior distal tibia: (0) absent; (1) present. The presence of the bladelike process on the posterior distal tibia was proposed as a synapomorphy of the L. montanus group by Etheridge (1995) and as a synapomorphy of the subgenus Eulaemus (including the L. anomalus group according to Lobo et al., 2010). Non-polymorphic binary. 23. Caudal vertebrae without chevron : (0) caudal vertebra I; (1) caudal vertebrae I and II; (2) caudal vertebrae I-III; (3) caudal vertebrae I-IV. Polymorphic multistate. 24. Skull height / skull length. Continuous (Table 2). 25. Skull width / skull length. Continuous (Table 2). 26. Lateral rami of interclavicle / skull length. Continuous (Table 2). 27. Diameter of major coracoid fenestra / diameter of major scapular fenestra. Continuous (Table 2). 28. Preischial length / skull length. Continuous (Table 2). 29. Xiphisternal rod length / skull length. Continuous (Table 2). 30. Clavicle length / skull length. Continuous (Table 2). 31. Maximum clavicle width / skull length. Continuous (Table 2). 32. Sternal width / sternal length. Continuous (Table 2). 33. Orientation of the pubis: (0) facing forward, forming an acute angle with vertebral column; (1) perpendicular to the vertebral column. State (0) only reported for L. pseudoanomalus (Lobo and Abdala, 2001). Non-polymorphic binary. 34. Membranes over coracoid fenestrae: (0) without ossification; (1) with ossification. Non- polymorphic binary.

72 70 L. Díaz-Fernández, A.S. Quinteros, F. Lobo Fig. 4. Skeletal characters exhibiting variation within the Liolaemus darwinii group. (A-B) Cartilaginous extremity of cervical rib IV: bifurcate (A) in L. scapularis (MCN2431) and non-bifurcate (B) in L. ornatus (MCN 3545). Scale = 2 mm. (C-D) Sternal fenestra: single (C) in L. ornatus (MCN 3546) and divided (D) in L. quilmes (MCN 3524). Scale = 2 mm. (E-F) Temporal fenestra: open, without contact between postorbital and squamosal (E) in L. ornatus (MCN 3548) and closed, with contact between postorbital and squamosal (F) in L. koslowskyi (MCN 574). Scale = 1mm. New characters found in the present contribution 35. Temporal fenestra (Fig. 4E-F): (0) open (without contact between postorbital and squamosal); (1) closed (contact between postorbital and squamosal). Non-polymorphic binary. 36. Posterior edge of parietal (Fig. 5A-B): (0) convex; (1) forming a straight margin. Non- polymorphic binary. 37. Posfrontal shape (Fig. 5C-D): (0) triangular; (1) elongated, not triangular. Polymorphic binary. 38. Premaxillary shape (Fig. 5E-F): (0) nasal spine narrow and pars dentalis wide; (1) nasal spine wide and pars dentalis narrow (modified from Frost, 1992). Polymorphic binary. Fig. 5. Skeletal characters exhibiting variation within the Liolaemus darwinii group. (A-B) Posterior edge of parietal: convex (A) in L. irregularis (MCN 2443) and forming a straight margin (B) in L. quilmes (MCN 3527). Scale = 1mm. (C-D) Posfrontal shape triangular (C) in L. quilmes (MCN 3527) and elongated (D) in L. inacayali (MCN 500). Scale = 1mm. (E-F) Premaxillary shape: nasal spine narrow and pars dentalis wide (E) in L. albiceps (MCN 402) and nasal spine wide and pars dentalis narrow (F) in L. koslowskyi (MCN 573). Upper arrow indicates the width of the nasal spine and bottom arrows indicate the width of the area of premaxillary teeth. Scale = 2 mm. 39. Otic ramus of squamosal (Fig. 6A-B: (0) otic ramus located over the superior fossa of quadrate; (1) otic ramus inserted in the superior fossa of quadrate. Non-polymorphic binary. 40. Number of labial foramina (lateral view of maxilla). Continuous (Table 2). 41. Disposition of labial foramina (maxilla): (0) L-shaped; (1) forming two parallel rows; (2) forming a unique series in a single line. Polymorphic multistate. 42. Number of mental foramina of dentary (lateral view). Continuous (Table 2).

73 Skeletal variation in the Liolaemus darwinii group 71 Fig. 6. Skeletal characters exhibiting variation within the Liolaemus darwinii group. (A-B) Otic ramus of squamosal located outside to the superior fossa of quadrate (A) in L. irregularis (MCN 2436) and otic ramus inserted in the superior fossa of quadrate (B) in L. grosseorum (MCN 508). Arrow indicates the insertion of the squamosal in the superior fossa of the quadrate. Scale = 2mm. (C-D) Lower jaw dentition homodont (C) in L. inacayali (MCN 500) and heterodont (D) in L. quilmes (MCN 3525). Scale = 1mm. (E-F) First chevron condition: incomplete (E) in L. quilmes (MCN 3524) and complete (F) in L. albiceps (MCN 402). Scale = 3mm. (G-H). Length of metatarsal IV with respect to toe V: reaches phalanx III (G) in L. irregularis (MCN 2431) and reaches phalanx II (H) in L. inacayali (MCN 500). Scale = 1 mm. 43. Number of premaxillary teeth. Continuous (Table 2). 44. Numbers of dentary teeth. Continuous (Table 2). 45. Lower jaw dentition (Fig. 6C-D): (0) homodont; (1) heterodont. Polymorphic binary. 46. First chevron shape (Fig. 6E-F): Chevron bones appear on anterior caudal vertebrae (Hoffstetter and Gasc, 1969). In Liolaemus the first chevron can appear on caudal vertebra III or IV. (0) incomplete; (1) complete. Polymorphic binary. 47. Length of metatarsus IV with respect to toe V (Fig. 6 G-H): (0) reaches phalanx II; (1) reaches phalanx Fig. 7. Skeletal characters exhibiting variation within the Liolaemus darwinii group. (A-B) Presence of open intercalated tracheal rings: present (A) in L. lavillai (MCN 4351) and absent (B) in L. grosseorum (MCN 509). Scale = 1 mm. (C-D) Lateral processes of the cricoid: not pronounced (C) in L. irregularis (MCN 2443) and pronounced (D) in L. albiceps (MCN 457). Scale = 1mm. (E-F) Sternal fenestra shape: symmetrical not widened (E) in L. albiceps (MCN 457) and wider in the posterior half (F) in L. chacoensis (MCN 599). Scale = 1mm. III (modified from Arias, 2012). Non-polymorphic binary. 48. Length of IV metacarpal with respect to finger V: (0) reaches phalanx I; (1) reaches phalanx II. Non-polymorphic binary. 49. Hipoischial fenestra: (0) absent; (1) present. Polymorphic binary. 50. Ischial fenestra (located close to the posterior margin of the ischium): (0) absent; (1) present. Polymorphic binary. 51. Number of sternal ribs: (0) three; (1) four. Polymorphic binary. 52. Number of branches of the xiphisternal rib: (0) three; (1) two; (2) none. Polymorphic binary.

74 72 L. Díaz-Fernández, A.S. Quinteros, F. Lobo Fig. 8. Evolution of cartilaginous extremity of cervical rib IV. Potential synapomorphy of the Liolaemus darwinii group. Recovered Topology of the L. darwinii phylogeny. Liolaemus crepuscularis is excluded because of the lack of information on this character state. Fig. 9. Evolution of bladelike process on posterior distal tibia in Liolaemus. Note the secondary loss in the L. anomalus group. Tree modified from Schulte et al. (2000). 53. Open intercalated tracheal rings (Fig. 7A-B): (0) present; (1) absent. Polymorphic binary. 54. Lateral processes of the cricoid (Fig. 7C-D): (0) not pronounced; (1) pronounced. Non-polymorphic binary. 55. Shape of sternal fenestra (Fig. 7E-F): (0) symmetrical, not widened; (1) wider in the posterior half. Polymorphic binary. DISCUSSION In this contribution, twenty one new characters for the genus Liolaemus were studied (characters from 35 to 55). We also report additional states for characters previously described (Table 1). First, the number of scleral ossicles was previously reported by Lobo and Abdala (2001) as binary polymorphic (13 or 14 ossicles), whereas here we found a new state for L. albiceps (15 ossicles), we consequently coded the character as polymorphic multistate. Second, the ceratohyal process was coded in Lobo and Abdala (2001) as non-polymorphic multistate (gradually widened, abruptly widened, and hookshaped), whereas here we found in L. quilmes a polymorphism (gradually widened and abruptly widened). Third, according to Lobo and Abdala (2001), the sternal fenestra can be divided or single, without polymorphism. In this study we report a polymorphism in L. quilmes, which may have a divided or undivided sternal fenestra. The specimens of L. cf. quilmes included in Lobo and Abdala (2001) correspond to L. crepuscularis (Abdala and Diaz Gómez, 2006), a species distinct from that included in this contribution as L. quilmes. Maxillary teeth with three conspicuous cusps were found in all specimens studied. This is consistent with Lobo and Abdala (2001), who found this character state in all specimens of the subgenus Eulaemus, whereas the species of the L. nigromaculatus group (belonging to the subgenus Liolaemus sensu stricto) show the maxillary teeth without differentiated cusps. This evidence allows us to consider tricuspid maxillary teeth as a potential synapomorphy of Eulaemus. The number of tracheal rings for species of the Liolaemus boulengeri group reported by Lobo and Abdala (2001) ranges from 48 to 67. We extend this range to A potential synapomorphy for the Liolaemus darwinii group (Fig. 8) involves the morphology of the terminal cartilage of cervical rib IV, which is narrow and not bifurcated in Liolaemus albiceps, L. chacoensis, L. grosseorum, L. irregularis, L. koslowskyi, L. lavillai, L. ornatus, and L. quilmes (though polymorphic in the latter). Lobo and Abdala (2001) found this character to be polymorphic in L. koslowskyi. Etheridge (1993) defined the Liolaemus darwinii complex based on the possession of maxillary teeth crowns with straight edges as an exclusive character of this group. Liolaemus scapularis (a member of the L. wiegmannii group), also shows tooth crowns with straight edges, so this character state is not exclusive to the L. darwinii complex, as Etheridge (1993) proposed. Moreover, our results indicate variation within this group (Fig. 2A) and according to recognized relationships (Abdala, 2007), this character has changed in the terminal subclade of the L. darwinii group (the L. ornatus group). Therefore, it can be considered a synapomorphy of the L. ornatus group (Fig. 2A-C). Optimizing the character states of maxillary tooth crowns and the diet in the tree recovered, we found that an insectivorous diet and the straight-edged maxillary tooth crowns change together along the tree (Fig. 2A).

75 Skeletal variation in the Liolaemus darwinii group 73 This can be interpreted as supporting a possible relationship between diet and tooth crown shape. Tooth morphology can reflect ecological adaptations and exhibit derived traits which may distinguish alimentary specializations (Hotton, 1965). We found that L. scapularis, L. quilmes, L. lavillai, L. grosseorum and L. chacoensis have straight crowns and are insectivorous. Lobo and Abdala (2001) cited straight crowns for L. crepuscularis (their L. cf. quilmes). Semhan et al. (2013) reported that L. crepuscularis feeds mainly on insects, but can fluctuate to omnivorous or herbivorous diets through the year based on prey availability. All other species studied here show expanded crowns, and their diet can be characterized as omnivorous or herbivorous (L. albiceps). The only exception to this association is L. koslowskyi, which has expanded crowns and an insectivorous diet (Aun and Martori 1998). Aun and Martori (1998) do not mention the season of the study, so it is not known if this taxon can change its diet as does L. crepuscularis. Phymaturus is the sister clade of Liolaemus, and has a strictly herbivorous diet (Lobo et al., 2010). Species of Phymaturus have teeth with expanded crowns (Lobo and Quinteros, 2005). The same phenomenon can be observed in nonliolaemid lizards. Hotton (1965) found that herbivorous lizards (Dipsosaurus, Sauromalus, and Ctenosaura) have highly cuspidate and antero-posteriorly widened teeth (similar to the expanded crowns of humans). In lizards that mainly feed on ants (Phrynosoma), Hotton (1965) described pointed and conical teeth, and in lizards that feed on bees and wasps (Urosaurus and Callisaurus), he described thin, cylindrical, sharp teeth. Herrel et al. (2004) observed that lacertids with omnivorous diets show teeth with wider crowns, whereas insectivorous species had slender and pointed teeth. In agreement with these results, we found that dentition seems to vary with diet. Nevertheless, further studies are needed in Liolaemus in order to confirm the hypothesis of correlation between straight crowns and insectivorous diet. The exposure or not of Meckel s groove in liolaemid lizards was used by Etheridge (1995) as a character in his taxonomic proposal. He proposed this character as a synapomorphy of the Liolaemus chiliensis (subgenus Liolaemus) group which have a fused channel, and also for the Phymaturus patagonicus group within the sister genus of Liolaemus (Etheridge,1995; Lobo and Quinteros, 2005; Lobo et al., 2010). Lobo and Abdala (2001) viewed the groove as a potential synapomorphy of the L. darwinii group. Here, we found that Meckel s groove exhibits an additional change (reversal), being open in the L. ornatus group. Moreover, Lobo and Abdala (2001) found this character state for L. crepuscularis, a basal member of the L. ornatus group. Therefore, we can conclude that the open Meckel s groove can be considered a synapomorphy of the L. ornatus group (Fig. 2A-C). The ancestral state would be the open Meckel s groove, already present in Ctenoblepharys, the basal genus of the Family Liolaemidae, and preserved in the Phymaturus palluma group. The hypothesis of the open Meckel s groove as the ancestral state is supported by the presence of this character state in other families related to Liolaemidae (e.g., Leiosauridae and Opluridae according to Pyron et al., 2013 and Reeder et al., 2015), but exhibiting polymorphism in many Iguanian families (Frost and Etheridge, 1989) such as Leiocephalidae (Etheridge, 1966) and Phrynosomatidae (Etheridge, 1964). The bladelike process on the posterior distal tibia was described by Etheridge (1995), as a synapomorphy of the Liolaemus montanus group. In his proposal, Etheridge (1995) did not include the L. anomalus group inside the L. montanus group. In recent analyses (Espinoza et al., 2004; Abdala 2007), the L. anomalus group is inferred as more closely related to the L. boulengeri group. These two groups together are called the L. boulengeri series (included inside the L. montanus section in Schulte et al., 2000). The L. anomalus group lacks this tibial process, which we consider to be a secondary loss (Fig. 9). Here we found that every member of the L. darwinii group studied has the bladelike process on the distal tibia. Characters that were studied in other groups of lizards were informative for Liolaemus, including shape of the premaxilla (Frost, 1992; Tropidurids); squamosalquadrate joint (modified from Frost, 1992); metacarpal of IV finger reach the I or II phalange of V finger (modified from Arias, 2012; teiids); and presence of open tracheal rings (Lobo and Quinteros, 2005; Lobo et al., 2010; Phymaturus). Here, we found the same variation in the shape of the premaxilla described by Frost (1992) for tropidurines: (0) narrow nasal spine - wide area of premaxillary tooth attachment and (1) broad nasal spine- narrow premaxillary tooth area. Frost (1992) found no relationship between the width of the area of premaxillary teeth and number of teeth in it. We found similar results for the Liolaemus species studied here, where the number of premaxillary teeth is constant regardless of the width of the area in which they are inserted. Also, Frost (1992) described two states for the squamosal-quadrate articulation. These states are related to the width of the superior fossa of the quadrate, which may be relatively small or enlarged. In the Liolaemus species studied, it was observed that the superior fossa of the quadrate corresponds to the relatively enlarged state according to Frost (1992). All species except two (L. albiceps and L. irregularis) exhibit one of the states proposed

76 74 L. Díaz-Fernández, A.S. Quinteros, F. Lobo by Frost (1992): the otic ramus of the squamosal contacts (but is not inserted in) the posterior edge of the superior fossa of the quadrate. In contrast, L. albiceps and L. irregularis share the same state, with the otic ramus of the squamosal inserted in the medial part of the fossa (Fig. 6 B), thus reinforcing the hypothesis that they are sister species. The character observed in teiids is related to the relative length of metacarpals, metatarsals and digits. Arias (2012) coded a character related to the length of toe V into three states: length of toe V exceeding that of metatarsal IV, equaling that of metatarsal IV, and not reaching the length of metatarsal IV. In all Liolaemus species studied here, the toe V always exceeds metatarsal IV in length, but there is variation as to which phalanx of toe V metatarsal IV reaches (Fig. 6 G-H). We recorded this character differently for one of the outgroup taxa (L. inacayali, where metatarsal IV reaches phalanx II) with respect to the L. darwinii group (in which metatarsal IV reaches phalanx III). Since we do not have samples of other members of the L. telsen group, we were not able to determine if this is a potential synapomorphy for that group. Similar variation to that described above for the hindlimbs was described for the forelimbs in Liolaemus. The length of metacarpal IV with respect to finger V, exhibited two states: reaches phalanx I or reaches phalanx II. Within the L. darwinii group, the character state in which finger V reaches phalanx II occurs independently in L. quilmes and L. chacoensis. It is clear that variation between fore and hindlimbs is independent. The presence of intercalated open tracheal rings was only found in the members of the Liolaemus ornatus group. Therefore, this character state can be considered as a possible synapomorphy of this group. Nevertheless, the intercalated open tracheal rings were also found to be polymorphic in L. scapularis and in L. quilmes, species basal to the L. ornatus group. This character state was primarily proposed for Phymaturus (Lobo and Quinteros, 2005), but no polymorphisms were found in this genus. While newly described characters have not yet been observed in many taxa, a general idea of the polarity of the optimized character was obtained. The study of new sources of variation and the distribution of characters described here in other taxa, as well as in the other two genera of Liolaemidae (Phymaturus and Ctenoblepharys) would allow us to hypothesize about relationships within this iguanian family. In this study we note that in the Liolaemus darwinii group the most informative characters are taken from the regions of the ribs, sternal plate, pectoral girdle, snout, jaw, larynx and hyoid arches. The literature shows that the osteology has been more thoroughly studied in the families Phrynosomatidae, Tropiduridae and Corytophanidae within Iguania (Reeve, 1952; Etheridge, 1964; Presch, 1969; Etheridge and de Queiroz, 1988; Frost and Etheridge, 1989; Frost, 1992; McGuire, 1996; Reeder and Wiens, 1996; Torrez Carvajal, 2007; Frost et al., 2011) and osteological characters are observed mainly from the cranial skeleton. This shows a bias in the focus of skeletal variation studies, making it difficult to determinate the levels of variation across the whole anatomy of lizards. Even if the literature emphasizes the idea of the importance of using osteological characters in different phylogenetic analyses and morphological descriptions (Conrad, 2008; Gauthier et al., 2012; Reeder et al., 2015), it is clear that there is not enough information yet to fully understand the patterns of morphological diversity across all existing iguanian families. ACKNOWLEDGEMENTS We thank Roberto Sanchez, Soledad Valdecantos, Alejandra Paz, Jessica Monroig, Thomas Hibbard, Matias Quipildor, Silvio Salcedo, Sabrina Portelli, Romina Semhan, Cristian Abdala, Alejandro Laspiur, Adolfo Juarez, Mario Fernandez and Melisa Diaz Fernandez for helping us with field or lab work. Patricia Garcia and T. Hibbard improved the English style. We also thank the Ministerio de Ambiente y Producción Sustentable (Secretaria de Ambiente) of Salta Argentina and Y. Bonduri for assisting with scientific collection permits (Resolution Nº 815/13). We thank E. Lavilla and S. Kretzschmar (Instituto de Herpetología, Fundación Miguel Lillo, Tucumán Argentina) for allowing us to study FML materials under their care. This study was supported by grants (FL) from CONICET Consejo Nacional de Investigaciones Científicas y Técnicas of Argentina (PIP 2841) and CIUNSA Consejo de Investigaciones de la Universidad Nacional de Salta, Argentina (CIUNSA 2036), and Graduate Fellowship, from Consejo Nacional de Investigaciones Científicas y Técnicas (LDF). We also thank two anonymous reviewers for their careful reading of our manuscript and their insightful comments and suggestions. REFERENCES Abdala, C.S. (2007): Phylogeny of the boulengeri group (Iguania: Liolaemidae, Liolaemus) based on morphological and molecular characters. Zootaxa 1538: Arias, F. (2012): Relaciones filogenéticas en la tribu Teiini. (Squamata: Teiidae). Evaluación de la monofilia del género Cnemidophorus y análisis de su estructura

77 Skeletal variation in the Liolaemus darwinii group 75 filogenética. Tesis Doctoral. Universidad Nacional de Tucumán, Argentina. Aun, L., Martori, R. (1998): Reproducción y dieta de Liolaemus koslowskyi Etheridge Cuad. Herpetol. 12: Avila, L.J., Morando, M., Sites, J.W. (2006): Congeneric phylogeography: hypothesizing species limits and evolutionary processes in Patagonian lizards of the Liolaemus boulengeri group (Squamata: Liolaemini). Biol. J. Linn. Soc. 89: Beurman, F.M.E., Vieira, M.I. (1980): Sobre a osteología craniana de Liolaemus occipitalis Boulenger, 1885, L. lutzae Mertens, 1938 e L. multiformis simonsii Boulenger, 1902 (Lacertilia, Iguanidae). Iheringia Sér. Zool. 56: Burkart, R., Bárbaro, N.O., Sánchez, R.O., Gómez, D.A. (1999): Ecorregiones de la Argentina. Administración de Parques Nacionales, Buenos Aires. Cei, J.M. (1986): Reptiles del centro, centro-oeste y sur de la Argentina. Herpetofauna de las zonas áridas y semiárida. Monogr. Mus Reg. Sci. Nat. Torino 4. Cei, J.M. (1993): Reptiles del noroeste, nordeste y este de la Argentina. Monogr. Mus Reg. Sci. Nat. Torino 14. Conrad, J.L. (2008): Phylogeny and systematics of Squamata (Reptilia) based on morphology. Bull. Am. Mus. Nat. Hist. 310: Da Silva, C.M., Verrastro, L. Veronese, B., Dias, F. (2007): Descrição do esqueleto axial de Liolaemus arambarenis (Iguania, Liolaemidae): Regiões pré-sacral e sacral. Rev. Bras. Zoo. 24: de Queiroz, K. (1982): The scleral ossicles of sceloporine iguanids: A reexamination with comments on their phylogenetic significance. Herpetologica 38: de Queiroz, K. (1987). Phylogenetic systematics of iguanine lizards: A comparative osteological study. Univ. Calif. Publ. Zool. 118: Donoso-Barros, R. (1966): Reptiles de Chile. Ediciones de la Universidad de Chile, Santiago. Espinoza, R.E., Wiens, J.J., Tracy, C.R. (2004): Recurrent evolution of herbivory in small, cold-climate lizards: Breaking the ecophysiological rules of reptilian herbivory. Proc. Natl. Acad. Sci. USA 101: Estes, R., de Queiroz, K., Gauthier, J. (1988): Phylogenetic relationships within Squamata. In: Phylogenetic relationships of the lizard families, pp Estes, R., Pregill G., Eds, Stanford University Press, Stanford. Etheridge, R.E. (1964): The skeletal morphology and systematics relationships of sceloporine lizards. Copeia 1964: Etheridge, R.E. (1965): The abdominal skeleton of lizards in the family Iguanidae. Herpetologica 21: Etheridge, R.E. (1966): The systematic relationships of West Indian and South American lizards referred to the iguanid genus Leiocephalus. Copeia 1966: Etheridge, R.E. (1967): Lizard caudal vertebrae. Copeia 1967: Etheridge, R.E. (1993): Lizards of the Liolaemus darwinii complex (Squamata: Iguania: Tropiduridae) in northern Argentina. Boll. Mus. Regionale Sci. Nat. 11: Etheridge, R.E. (1995): Redescription of Ctenoblepharys adspersa Tschudi, 1845, and the taxonomy of Liolaeminae (Reptilia: Squamata: Tropiduridae). Am. Mus. Novit. 3142: Etheridge, R.E. (2000): A review of the Liolaemus wiegmannii group (Squamata, Iguania, Tropiduridae), and a history of morphological change in the sand dwelling species. Herpetol Monogr. 14: Etheridge, R.E., de Queiroz, K. (1988): A phylogeny of Iguanidae. In: Phylogenetic relationships of the lizard families, pp Estes, R., Pregill, G., Eds. Stanford University Press, Stanford. Fontanella, F.M., Olave, M., Avila, L.J., Sites Jr, J.W., Morando, M. (2012): Molecular dating and diversification of the South American lizard genus Liolaemus (subgenus Eulaemus) based on nuclear and mitochondrial DNA sequences. Zool. J. Linn. Soc. 164: Frost, D.R. (1992): Phylogenetic analysis and taxonomy of the Tropidurus group of lizards (Iguania: Tropiduridae). Am. Mus. Novit. 3033: Frost, D.R., Etheridge, R. E. (1989): A phylogenetic analysis and taxonomy of iguanian lizards (Reptilia: Squamata). Misc. publ. Univ. Kans. Mus. Nat. Hist. 81: Frost, D.R., Rodrigues, M.T., Grant, T., Titus, T.A. (2001). Phylogenetics of the lizard genus Tropidurus (Squamata: Tropiduridae: Tropidurinae): Direct optimization, descriptive efficiency, and sensitivity analysis of congruence between molecular data and morphology. Mol. Phylogenet. Evol. 21: Gauthier, J.A., Kearney, M., Maisano, J.A., Rieppel, O., Behlke, A.D.B. (2012). Assembling the squamate tree of life: perspectives from the phenotype and the fossil record. Bull. Peabody Mus. Nat. Hist. 53: Goloboff, P., Farris, J., Nixon, K. (2003): T.N.T.: Tree Analysis Using New Technology. Available from: González-Marín, A., Hernando, A. (2013): Notes on the postcranial osteology of the sand lizard Liolaemus azarai (Squamata: Liolaemidae). Phyllomedusa 12: Herrel, A., Vanhooydonck, B., Van Damme, R. (2004): Omnivory in lacertid lizards: adaptive evolution or constraint? BMC Evol. Biol. 17:

78 76 L. Díaz-Fernández, A.S. Quinteros, F. Lobo Hoffstetter, R., Gasc, J.P. (1969): Vertebrae and ribs of modern reptiles. In: Biology of the Reptilia, Morphology A I, pp Gans, C., Ed. Academic Press, London and New York. Hotton, N. (1965): A survey of adaptative relationships of dentition to diet in the North American Iguanidae. Am. Midl. Nat. 53: Keller, C., Krause, L. (1986): The appendicular skeleton of Liolaemus occipitalis Boulenger 1885 (Sauria, Iguanidae). Rev. Bras. Biol. 46: Lang, M. (1989): Phylogenetic and biogeographic patterns of basiliscine iguanians (Reptilia: Squamata: Iguanidae). Bonn. Zool. Monogr. 28: Lobo, F. (2001): A phylogenetic analysis of lizards of the Liolaemus chiliensis group (Iguania: Tropiduridae). Herpetol. J. 11: Lobo, F. (2005): Las relaciones filogenéticas en el grupo chiliensis de Liolaemus (Iguania: Liolaemidae). Sumando nuevos caracteres y taxa. Acta Zool. Lilloana 49: Lobo, F., Abdala, C., Valdecantos, S. (2012): Morphological diversity and phylogenetic relationships within a South-American clade of iguanian lizards (Liolaemidae: Phymaturus). Zootaxa 3315: Lobo, F., Abdala, C. (2001): Variación morfológica en el esqueleto de Liolaemus (Iguania: Liolaemidae). Búsqueda y descripción de caracteres. Cuad. Herpetol. 15: Lobo, F., Abdala, C. (2002): La información cladística de un set de datos morfológicos en lagartos del género Liolaemus (Iguania: Liolaemidae). Cuad. Herpetol. 16: Lobo, F., Quinteros, A.S. (2005): A morphological approach on the phylogenetic relationships within the genus Phymaturus (Iguania: Liolaemidae). The description of four new species of Argentina. Pap. Avulsos. Zool. 45: Lobo, F., Espinoza, R.E., Quinteros, A.S. (2010): A critical review and systematic discussion of recent classification proposals for liolaemid lizard. Zootaxa 2549: McGuire, J.A. (1996): Phylogenetic systematics of crotaphytid lizards (Reptilia: Iguania: Crotaphytidae). Bull. Carnegie Mus. Nat. Hist. 32: Núñez, H., Navarro, J., Garín, C., Pincheira-Donoso, D., Meriggio, V. (2003): Phrynosaura manueli y Phrynosaura torresi, nuevas especies de lagartijas para el norte de Chile (Squamata: Sauria). Bol. Mus. Nac. Hist. Nat. 52: Olave, M., Avila, L.J., Sites Jr., J.W., Morando, M. (2014): Multilocus phylogeny of the widely distributed South American lizard clade Eulaemus (Liolaemini, Liolaemus). Zool. Scr. 43: Presch, W. (1969): Evolutionary osteology and relationships of the horned lizard genus Prynosoma (Family Iguanidae). Copeia 1969: Pyron, R.A., Burbrink, F. T., Wiens, J.J. (2013): A phylogeny and revised classification of Squamata, including 4161 species of lizards and snakes. BMC Evol. Biol. 13: Reeder, T., Wiens, J.J. (1996): Evolution of the lizard family Phrynosomatidae as inferred from diverse types of data. Herpetol. Monogr. 10: Reeder, T.W., Townsend, M., Mulcahy, D.G., Noonan, B. P., Wood, P.L., Sites Jr., J.W., Wiens, J.J. (2015): Integrated analyses resolve conflicts over squamate reptile phylogeny and reveal unexpected placements for fossil taxa. PLoS ONE 10: Reeve, W.L. (1952): Taxonomy and Distribution of the Horned Lizard Genus Phrynosoma. Univ. Kans. Sci. Bull. 34: Schulte, J.A. II, Macey, J.R., Espinoza, R.E., Larson, A. (2000): Phylogenetic relationships in the iguanid lizard genus Liolaemus: multiple origins of viviparous reproduction and evidence for recurring Andean vicariance and dispersal. Biol. J. Linn. Soc. 69: Schulte, J.A. II, Valladares, J.P., Larson, A. (2003): Phylogenetic relationships within Iguanidae inferred using molecular and morphological data and a phylogenetic taxonomy of iguanian lizards. Herpetologica 59: Semhan, V.R., Halloy, M., Abdala, C.S. (2013): Diet and reproductive states in a high altitude Neotropical lizard, Liolaemus crepuscularis (Iguania: Liolaemidae). South Am. J. Herpetol. 8: Simoes-Lopes, P.C.A., Krause, L. (1998): Osteologia do sincranio de Liolaemus occipitalis Boulenger, 1885 (Sauria, Iguanidae). Rev. Bras. Zool. 5: Torres-Carvajal, O. (2004): The abdominal skeleton of tropidurid lizards (Squamata: Tropiduridae). Herpetologica 60: Uetz, P. (editor), The Reptile Database, accessed July 31, Wassersug, R.J. (1976): A procedure for differential staining of cartilage and bone in whole formalin fixed vertebrates. Stain Technol. 51:

79 Skeletal variation in the Liolaemus darwinii group 77 APPENDIX A Specimens examined. The acronyms used were MCN (Museo de Ciencias Naturales de la Universidad Nacional de Salta) and FML (Fundación Miguel Lillo). Liolaemus albiceps (n=4): ARGENTINA: Salta: Los Andes: Camino al Acay desde Estación Muñano, 5-6 km, (24 18'31.6''S; 66 09'7''W) MCN 402, 407, 1452, Liolaemus chacoensis (n=4): ARGENTINA. MCN 503, 504, 505, 599. No data. Liolaemus grosseorum (n=2): ARGENTINA: Mendoza: San Rafael: Orillas del embalse el Nihuil, MCN 508, 509. Liolaemus inacayali (n=1): ARGENTINA: Rio Negro: Ing. Jacobacci: 25 de Mayo, MCN 500. Liolaemus irregularis (n=4): ARGENTINA: Salta: Los Andes: Aprox. a 11KM al SE de San Antonio de los Cobres por ruta 51 (24 8'53.44''S; 66 8'21.37''W). MCN 2431, 2436, 2443, Liolaemus kingii (n=1): ARGENTINA: Santa Cruz. MCN 565. No data. Liolaemus koslowskyi (n=3): ARGENTINA: La Rioja: Castro Barros: 6 km. E. De Anillaco (28 47 S; W). MCN 573, 574, 576. Liolaemus lavillai (n=2): ARGENTINA: Jujuy: extremo norte del Parque nacional los cardones, oeste de la Recta de Tin Tin. ( S; W). MCN 2688, Liolaemus multicolor (n=1): ARGENTINA: Jujuy: Abra Pampa: Cochinoca. FML Liolaemus ornatus (n=4): ARGENTINA: Jujuy: Castro Tolay: A 7 km de S de Rio las Burras. MCN 3545, 3546, 3547, Liolaemus pseudoanomalus (n=1): ARGENTINA: La Rioja: Castro Barros: 6 km. E. De Anillaco (28 47 S; W). MCN 526. Liolaemus quilmes (n=5): ARGENTINA: Salta: Cachi: Cachi: A 7 Km Sur de Palermo entre Cachi adentro y Palermo, ( ,9 S; ,2 W). MCN 3524, 3525, 3526, 3527, Liolaemus scapularis (n=2): ARGENTINA: Salta: Cafayate: Los Médanos. MCN 253, 283.

80

81 Acta Herpetologica 12(1): 79-88, 2017 DOI: /Acta_Herpetol Marking techniques in the Marbled Newt (Triturus marmoratus): PIT-Tag and tracking device implant protocols Hugo Le Chevalier 1, Olivier Calvez 2, Albert Martínez-Silvestre 3, Damien Picard 4, Sandra Guérin 4,5, Francis Isselin-Nondedeu 6, Alexandre Ribéron 1, Audrey Trochet 1,2, * 1 CNRS, ENFA, UMR5174 EDB (Laboratoire Evolution et Diversité Biologique), Université Paul Sabatier, 118 route de Narbonne, Toulouse F-31062, France. *Corresponding author. trochet.audrey@wanadoo.fr 2 Station d Ecologie Théorique et Expérimentale, UMR 5321, Moulis F-09200, France 3 CRARC Avinguda Maresme s/n Masquefa, Barcelona, Spain 4 GECCO Groupe Ecologie et Conservation des Vertébrés, Université d Angers, 2 Boulevard Lavoisier, F Angers, France 5 Centre d Ecologie Fonctionnelle et Evolutive; UMR 5175; 1919 route de Mende, F Montpellier Cedex 5, France 6 Departement Aménagement et Environnement Ecole Polytechnique de l Université François Rabelais de Tours, CNRS ; UMR 7324 CITERES équipe IPAPE; Allée Ferdinand de Lesseps, F-Tours, France Submitted on 2016, 10 th October; revised on 2016, 31 th October; accepted on 2016, 5 th December Editor: Marco Sannolo Abstract. Individual marking has become essential for studying population dynamics and ecological requirements. However, marking small-bodied species such as amphibians is becoming a challenge in the last decades. Amphibian surveys may require to mark manually individuals, using toe clipping, polymers and pigments, or passive integrated transponders (PIT-tags). Even if ethics committees have recently recommend avoiding toe clipping in amphibians, the use of PIT-tags led to controversial results because low tag retention reported in some studies. Here, we describe a protocol of potentially life-long PIT-tag marking in a protected species, the marbled newt Triturus marmoratus. In addition, we also detailed a second procedure of surgery for the implantation of transmitters needed in radio-tracking surveys. During both procedures, we found that the newt phase (either aquatic or terrestrial) strongly affected the anesthesia duration. Indeed, newts in aquatic phase were more quickly anesthetized than newts under terrestrial phase. We then recommend to pay attention of this physiological particularity when performing this kind of procedure. Improving our knowledge on ecological requirements and population dynamics of this species is crucial for management and conservation plans, and could be extended to other large newts. Keywords. Anesthesia, transmitter implantation, Triturus marmoratus, PIT-tagging, newts, skin permeability INTRODUCTION Individual marking has become an essential method for studying ecological requirements, population dynamics or colonization rates (McCarthy and Parris, 2004). Two different methods are widely used for monitoring species: capture-marking-recapture (CMR) and radiotracking surveys. The CMR method is a powerful method for estimating population parameters such as abundance (Thompson et al., 1998), survival, recruitment, and population growth rate (Lettink and Armstrong, 2003). For CMR studies, marking of individuals can be performed using several techniques (Ferner, 2007), such as color pattern (non-invasive method often used in some salamanders and anurans species; see Delarze et al., 2000; Lama et al., 2011; Ribeiro and Rebelo, 2011; Waye, 2013; Elgue et al., 2014), the use of passive integrated transponders (PIT-tags; Jehle and Hödl, 1998), toe clipping, or the use ISSN (print) ISSN (online) Firenze University Press

82 80 H. Le Chevalier et alii of polymers and pigments (Brannelly et al., 2014). Toe clipping is a technique commonly applied to amphibians, but may be useful for a limited time due to their regeneration ability (Andreone, 1986). Indeed, time intervals for regeneration range between less than one year to several years or not observable regenerations, depending on the species (Ferner, 2007). This technique could also be a source of stress and involves tissue damage and a risk of infection in many species and may also interfere with behaviour and movement pattern of individuals (Parris and McCarthy, 2001; McCarthy and Parris, 2004; Funk et al., 2005). Photo matching is a relatively low-costly and non-invasive powerful technique allowing individuals to be recognize using photographic identification of external body markings (Arntzen et al., 2004). This technique had been used in a variety of species, including newts (Drechsler et al., 2015; Mettouris et al., 2016; see also Diego-Rasilla and Luengo, 2002) but photo-identification might be difficult to apply in some cases (for instance in the marbled newt during the aquatic phase, females are very dark colours with very low contrast). Observers can correctly match 100% of photo pairs by eye (Morrison et al., 2016), but this process is highly time-consuming, and was considered as relatively not appropriate for studies on large populations (Arntzen et al., 2004). Since several years, many automatic recognition softwares have been developed (Wild-ID, Bolger et al., 2012; APHIS, Moya et al., 2015; Hotspotter, Crall et al., 2013; AMPHI- DENT, Matthé et al., 2008; I 3 S Pattern, van Tienhoven et al., 2007) to reduce the time of individual identification. Many studies demonstrated that, using automated processes for photographic re-identification, pattern mapping is a successful approach for the identification of individuals, even in large populations (Drechsler et al., 2015; Mettouris et al., 2016). However in some cases, these tools failed to match many image pairs (Morrison et al., 2016), and might induce significant bias in the CMR analysis. PIT-tagging is a relatively new marking technique (Christy, 1996) providing rapid recognition of individuals during recapture sessions and limiting recognition errors (using recorder scanning device for instance). Also, some scanning devices can read the PIT-tags at several centimeters of the marked individual, which reduce the stress induced during the manipulation. Along with the demonstration of drawbacks of this method in amphibians (Tracy et al., 2011; Brannelly et al., 2014), such as low tag retention among marked individuals and the expense, ethics committees have recently recommend avoiding toe clipping in amphibians. Indeed, other recent studies showed pertinent results using PIT-tagging for amphibian monitoring (Connette and Semlitsch, 2011; Heard et al., 2012) with no significant effects of marking on survival rate or body condition (Perret and Joly, 2002). Miniaturization process of PIT-tags and similar marking techniques has increased these last years, being less invasive, and the size should continue to decrease in the future (Gibbons and Andrews, 2004; Cooke et al., 2013). PIT-tagging (relatively costly and rapid recognition) and photo identification (low-costly and non-invasive) are often compared for studying the benefits and disadvantages of each technique (Arntzen et al. 2004), and to identify the more appropriate method to apply on amphibian populations. The choice of technique depends on financial and human costs, but also on the study species, study duration and sampling proportion (Arntzen et al. 2004). The use of photo-identification technique is well documented in the literature (Drechsler et al., 2015; Mettouris et al., 2016; Sannolo et al., 2016) but clear and detailed protocols about the use of PIT-tagging still remain rare, in particular in newts. Moreover, it is noteworthy that individuals tested here were captured for locomotion and mobility experimental tests under controlled conditions (see below) and that each individual from host captive wildlife within French establishments authorized must be marked with PIT-tagging since August 2004 (art. R ; 10 August 2004). Radio-tracking surveys are also commonly used in many species, even in small-bodied urodeles (Jehle and Arntzen, 2000; Ribéron and Miaud, 2000; Rittenhouse and Semlitsch, 2006). For radio-tracking studies, an active transmitter is either attached to the different parts of the animal body, depending on the taxon (stuck on carapace, neck or wings) or surgically implanted into the coelomic cavity. Then, tracked individuals can then be located using a receiver. Contrary to CMR surveys, radiotracking allows precise and remote detection of individuals allowing determination of habitat use and daily movements. Here, we describe the protocol of a successful PITtag marking in the marbled newt. We also detail a surgical procedure for the implantation of transmitters in this species. Details on individual morphology, anesthesia duration, and the effect of individual phase (terrestrial or aquatic phase) are explored. Studies on ecological requirements and population dynamics of this species are crucial for management and conservation plans. Technological progress over the last few years in transmitter size allows the use of small transmitters without an external part. Developing an efficient protocol for both PIT-tagging and implantation of transmitters could strongly be useful and applicable in other large newt species.

83 Marking techniques in the Marbled Newt 81 Study species MATERIALS AND METHODS 70 days at their site of capture after locomotion and mobility experimental tests (data not shown here). The marbled newt (Triturus marmoratus) is a protected species listed on both the Bern Convention and the European Habitat Directive (annex III and annex IV respectively; least concern on the IUCN RedList), with declining populations (Arntzen et al., 2009). During the breeding season (from March to May), this species lives in a variety of temporary or permanent water sources, such as well-vegetated ponds, pools, ditches and streams (Nöllert and Nöllert, 2003). Newts have a permanent permeable skin. But during the breeding season their skin is more permeable, allowing for respiration, but also resulting in significant water loss. A nearby water source therefore constitutes critical habitat for this species during reproduction (Wells, 2007). After breeding, the skin of newts undergoes a seasonal change in osmotic permeability and becomes less permeable (Wells, 2007). Newts become more terrestrial and move out of water to winter and feeding habitats, notably wet habitats such as forests or under stones, to hibernate. Within this terrestrial phase, the mobility of the marbled newt is estimated to be less than 1 kilometer (Jehle, 2000; Jehle and Arntzen, 2000). Anesthesia procedure Both procedures were performed in sterile conditions using diluted Chlorexidine 0.75%. Prior to each procedure, individuals were rinsed and placed in individual boxes. Considering the average weight of newts (mean ± SD: ± 3.32 g; min-max: g; Table 1) and the concentration of Lidocaine/Prilocaine in EMLA ointment (5% Lidocaine 2.5% and Prilocaine 2.5%; Astra- Zeneca GmbH Laboratories, Germany, EMLA), the final dosage for anesthesia was 450 mg/kg, by percutaneous absorption of the ointment in a cutaneous squared surface of 1cm 1 cm. We applied one spot of cutaneous anesthetic cream on their left flank, until the muscular system was relaxed and animals stopped moving. We considered animals to be surgically anesthetized (only cardiac impulse was present) when individuals lost the withdrawal reflex (i.e. no response when pinching digits and tail; Fig. 1a) and righting reflex (i.e. unable to right themselves when put on their back; Fig. 1b; Mitchell, 2009). Animals were rinsed before the procedures to remove excess cream. Anesthesia and recovery durations were recorded for both protocols. Newt sampling We captured a total of 46 marbled newts (22 females and 24 males) from two sites to minimize the impact on newt populations and also to test if radio-tracking survey data were similar between individuals from different landscapes (see Trochet et al., 2017). We sampled marbled newts in southern France, in the department of Ariège (n = 30; 12 females and 18 males; N, E) between the 1 st and the 29 th April 2015, and in the department of Gers (n = 16; 10 females and 6 males; N, E) between the 24 th April and the 6 th of May Individuals were caught using a landing net and transported to the Station d Ecologie Théorique et Expérimentale CNRS ( N, E; Moulis, FRANCE), an ecological research station of the National Center for scientific research situated in the foothills of the Pyrenees. Animals were housed in groups of 6 to 8 individuals in aquaria of cm and kept at a temperature of 20 C, and were fed with mealworms. For each individual, we measured the snoutvent length (SVL) using a caliper and the body mass (BM) with an electronic scale (precision 0.01 g). PIT-tagging was performed between the 2 nd April and the 11 th May 2015, during the aquatic phase of the breeding season when newts had large crest in males and highly permeable skin (Table 1). During this aquatic phase, the skin is smoother and less coloured. Transmitter implantations were then performed between the 28 th May and the 19 th June 2015 during the postbreeding migration when newts were in the terrestrial phase, recognizable by the small crest in males and low permeable skin (individuals more coloured and grainy skin; Table 1). All individuals were pit-tagged, with 24 individuals also implanted with transmitters. Animals were released after 40 to PIT-tagging protocol (n = 46) During anesthesia, the newt skin was firstly disinfected before surgery with diluted Chlorexidine 0.75%. We then placed a PIT-tag (RFID Standards ISO & type FDX-B, mm, khz from BIOLOG-ID, FR; Fig. 1c) into the dorsal side using an injector, previously disinfected with diluted Chlorexidine 0.75%. The mean mass of PIT-tags was 0.03 g, representing approximately 0.26% (range min-max: %) of the newt body mass. The needle was inserted on the left side, at the site of anesthetic application, from the bottom of the back and pushed up under the skin to install the PIT-tag lateral to the hepatic area between the posterior and anterior limbs. The injection site was immediately disinfected after injection with Chlorexidine 0.75%. PIT-tags were not removed from individuals before releasing. Transmitter implantation (n = 24) During anesthesia, a longitudinal incision of about 1 centimeter, matching the width of the transmitter, was made on the right flank using surgical scissors in two steps, for the skin and then muscle tissue. The transmitter (V1 10A ultimate lite implants: battery life around 54 days; BIOTRACK, UK) was placed into the abdominal cavity and fitted between the internal organs (Fig. 2a). The mean mass of transmitters was 1.77 ± 0.04 g, representing approximately 16% of the newt body mass (average ± 2.23 g), a proportion considered unlikely to impact displacements (Madison and Farrand, 1998). The muscle tissue and the skin were then pulled over the transmitter and sutured. The muscle tissue was sutured with running subcuticular clo-

84 82 H. Le Chevalier et alii Table 1. Individual data on the 46 marbled newts used for both PIT-tagging and transmitter implantations. ID: individual number (corresponding to the seven last digits of PIT-tag number); SVL: snout-vent length (cm); BM: body mass (g). Capture PIT-tags Transmitter implantations Morphology Site Date ID Date Transmitter frequency Surgery date Sex SVL BM Release date Ariege 01/04/ /04/ /06/2015 M /06/2015 Ariege 01/04/ /04/ /06/2015 F /06/2015 Ariege 26/04/ /04/ /06/2015 M /06/2015 Ariege 26/04/ /04/ /06/2015 F /06/2015 Ariege 26/04/ /04/2015 M /06/2015 Ariege 26/04/ /04/ /06/2015 M Dead Ariege 28/04/ /04/ /05/2015 F /06/2015 Ariege 28/04/ /04/ /05/2015 M /06/2015 Ariege 28/04/ /04/ /06/2015 F /06/2015 Ariege 28/04/ /04/ /06/2015 F /06/2015 Ariege 28/04/ /04/ /06/2015 M /06/2015 Ariege 28/04/ /04/2015 F /06/2015 Ariege 28/04/ /04/2015 M /06/2015 Ariege 28/04/ /04/2015 M /06/2015 Ariege 28/04/ /04/2015 M /06/2015 Ariege 28/04/ /04/2015 M /06/2015 Ariege 28/04/ /04/ /06/2015 F /06/2015 Ariege 28/04/ /04/ /06/2015 M /06/2015 Ariege 28/04/ /04/ /06/2015 F /06/2015 Ariege 28/04/ /04/ /06/2015 M Dead Ariege 29/04/ /04/2015 M /06/2015 Ariege 29/04/ /04/2015 M /06/2015 Ariege 29/04/ /04/2015 M /06/2015 Ariege 29/04/ /04/2015 M /06/2015 Ariege 29/04/ /04/2015 F /06/2015 Ariege 29/04/ /04/2015 F /06/2015 Ariege 29/04/ /04/2015 F /06/2015 Ariege 29/04/ /04/2015 F /06/2015 Ariege 29/04/ /04/2015 M /06/2015 Ariege 29/04/ /04/2015 M /06/2015 Gers 26/04/ /04/ /06/2015 F /06/2015 Gers 26/04/ /04/2015 F /06/2015 Gers 26/04/ /04/2015 F /06/2015 Gers 26/04/ /04/2015 M /06/2015 Gers 06/05/ /05/ /06/2015 M /06/2015 Gers 06/05/ /05/ /06/2015 F /06/2015 Gers 06/05/ /05/ /06/2015 M /06/2015 Gers 06/05/ /05/ /06/2015 F /06/2015 Gers 06/05/ /05/ /06/2015 M /06/2015 Gers 06/05/ /05/ /06/2015 F /06/2015 Gers 06/05/ /05/ /06/2015 M /06/2015 Gers 06/05/ /05/ /06/2015 F /06/2015 Gers 06/05/ /05/ /06/2015 F /06/2015 Gers 06/05/ /05/2015 M /06/2015 Gers 06/05/ /05/2015 F /06/2015 Gers 06/05/ /05/2015 F /06/2015

85 Marking techniques in the Marbled Newt 83 Fig. 1. Anesthesia procedure after application of EMLA cream: loss of pain reflex (a); loss of righting reflex (b); PIT tag implantation (c). Fig. 2. Surgical implantation of transmitter: insertion of transmitter in the abdominal cavity (a); completed suture of muscle tissue using a simple continuous suture pattern (b); skin suture using the horizontal mattress suture technique (c); and completed procedure (d) sure (Fig. 2b) using absorbable material (ETHILON, diameter 3/0). The skin was sutured with simple interrupted suture (3, 4 or 5 surgeon s knots depending on the length of the incision; Fig. 2c, 2d) using absorbable material (ETHILON, diameter 3/0). Sutures were performed using an iris cutting needle (a C-shaped needle 7 mm long) and surgical silk. Transmitters were not removed from individuals before releasing. Recovery phase After each protocol, anesthetized individuals were then placed in a recovery room, into wet boxes, where the ventral portion of the body was placed into water to promote the expulsion of the anesthetic agent. We considered newts to be recovered with the return of the righting reflex. They were then placed into a well-vegetated terrarium, with water, soil substrate, shelter (mud) and food (tubifex worms) and kept under observation for 5 to 10 days. No visible impact on health and behaviour was observed after the PIT-tagging procedure. However after transmitter implantation, triggering of a molt on 14 newts was observed two days after operation. The return of the feeding behaviour occurred in the four days after the operation. Statistical analyses We performed a total of 70 fully-recorded procedures (46 PIT-tag implants, 24 transmitter implants). In order to test if skin permeability (i.e., the date of both operations) could have an effect both on the anesthesia and the recovery durations (both not normally distributed), we used non-parametric Spearman correlation tests. We also compared both the anesthesia and the recovery durations and their variability (i.e., using a coefficient of variation defined as standard variation/mean 100) depending on the two procedures tested here using Wilcoxon tests. We then calculated a condition index (hereafter CI) for each newt estimated by the residuals of the regression of log(bm) on log(svl) (Jakob et al., 1996), a useful index in amphibian studies (Denoël et al., 2002; Bancila et al., 2010) to test if the morphology could be related to both anesthesia and recovery durations in the two different procedures using Spearman correlations. Finally, we tested if the sex had an influence on both anesthesia and recovery durations using Wilcoxon tests. All statistical analyses were performed using R (R Development Core Team, 2014). Anesthesia duration RESULTS Our results showed that the anesthesia duration was shorter among individuals in aquatic phase (mean ± SD: 5.17 ± 3.09 minutes) than among newts in terrestrial phase (mean ± SD: ± 4.49 minutes) and that anesthesia duration was variable as both during PIT-tagging

86 84 H. Le Chevalier et alii as in transmitter implantation procedures (Wilcoxon test: W = 1, P = 1). The anesthesia duration was significantly different between both procedures (Wilcoxon test: W = 125.5, P 0.01) and increased from the first procedures performed (i.e., PIT-tagging) at the beginning of April 2015 (aquatic phase with high permeable skin) to the final procedures (i.e., transmitter implantation) performed at the end of June 2015 (terrestrial phase with less permeable skin; Spearman correlation: r s = 0.39, P < 0.01; Fig. 3). We found non-significant relationship between the anesthesia duration and newts CI (Spearman correlations: r s = -0.08, P = 0.50). Our results also showed no influence of sex on the anesthesia duration (Wilcoxon tests: W = 560.5, P = 1). Recovery duration Our results did not found difference in recovery duration variability between both procedures (Wilcoxon test: W = 0, P = 1). We found a strong difference in the recovery duration between the two procedures (Wilcoxon tests: W = 43, P < 0.01). The recovery duration after PIT-tagging (mean ± SD: ± minutes) was significantly shorter than the recovery duration after transmitter implantation (mean ± SD: ± minutes; Spearman correlation: r s = 0.60, P = < 0.01; Fig. 4). Contrary to the anesthesia duration, newt CI influenced recovery duration (Spearman correlations: r s = -0.27, P = 0.02) where the largest newts woke up faster. Our results showed no influence of sex on the recovery duration (Wilcoxon tests: W = 635, P = 0.36). PIT-tags and transmitter surveys No PIT-tags were reported lost, even several months after implantation. Before releasing, we reported the number of days for which newts kept their PITtags: days in average; range min-max: days. We also noted that none of the tags moved from the insertion site. Newts with transmitters were then radiotracked up to 48 days after release (see Trochet et al., 2017). Only one newt (ID number: ) was found dead in the field during the survey (seven days after release). Fig. 3. Anesthesia duration (in minutes) depending on procedure date related to the number of newts in terrestrial phase (less permeable skin). PIT-tagging procedures are represented by black dots and transmitters implantations are represented by white dots. Fig. 4. Recovery duration (in minutes) for both procedures.

87 Marking techniques in the Marbled Newt 85 DISCUSSION Marking amphibians is a challenge for studying population dynamics. Indeed, their small size and their semiaquatic life cycle increase the difficulty to mark them. Toe clipping is still in use in amphibian monitoring, even if this process is actually criticized (Parris and McCarthy, 2001). As a non-invasive and successful method for individual identification, photo-matching is becoming a powerful technique largely used in amphibian species since many years. But this method can be time-consuming, even using automatic processes for photographic re-identification. Moreover, pattern mapping may be not appropriate for studies on large populations (but see Drechsler et al., 2015; Mettouris et al., 2016). PIT-tagging is a recent advance in animal marking technique and is now frequently used in a variety of species, including amphibians (Ott and Scott, 1999; Cucherousset et al., 2008) but this method is relatively costly and had demonstrated contrasting results. The recent cost reduction and miniaturization of PIT-tags (in this study: PIT-tag length mm, approximately $2.50 each) make this technique particularly interesting in amphibian monitoring surveys. While photo-identification has been explored in several amphibian species, a detailed and powerful method of PITtagging still remains rare, especially in newts. Here, we specified how perform PIT-tagging in the marbled newt to limit PIT-tag loss respecting the welfare of individuals. Also, this kind of marking became mandatory in French establishments authorized to host captive wildlife, and for which a clear protocol is needed. Anesthesia and recovery durations The use of a cutaneous anesthetic agent permits the performance of reliable procedures in a less invasive way. Physiological state should be closely observed during the procedures and dosage recommendations followed. Our findings also demonstrated that the recovery duration in newts after anesthesia was more influenced by the invasiveness of the procedure than by the phase of the individuals. We reported deaths after both marking procedures only for two newts operated on twice after the use of surgical adhesive (see below). Hence, we recommend following these protocols to increase the likely success of surveys, even with the constraint of requiring sterile conditions. PIT-tagging as a robust method for amphibian marking PIT-tag marking has been used in many species, allowing assessment of movement patterns or growth and survival rates (Jehle and Hödl, 1998; Ott and Scott, 1999; Perret and Joly, 2002). This method is often preferred in long-term population surveys (Perret and Joly, 2002; Arntzen et al., 2004) and in behavioral studies (Winandy and Denoël, 2011), even in relatively small-bodied salamanders (Connette and Semlitsch, 2011). Indeed, no significant effect of PIT-tagging on growth and survival has been demonstrated (Ott and Scott, 1999; Perret and Joly, 2002; Connette and Semlitsch, 2011). PIT-tags are permanent and internal transponders, and this technique is generally considered as less invasive than other methods such as toe clipping which might be a stressful marking method for individuals (Parris and McCarthy, 2001). Most studies using PIT-tagging reported PIT-tag loss and failure, or mortality of individuals, but these results were likely caused by improper implantation (Gibbons and Andrews, 2004). Tag loss can also occur just after PITtag injection if the tag exits through the opening caused by the needle (12.1% of tag loss in Feldheim et al., 2002). Most of these studies, which inject PIT-tags without anesthesia, tended to reject the use of these transponders because of low tag retention rate (33.3% tag loss in Brannelly et al., 2014). Here, no PIT-tag expulsion was recorded. Hence, in order to minimize PIT-tag loss, we strongly recommend anesthetizing individuals before PIT-tagging. This results in the immobilization of the individual for several minutes which reduces the chance of expulsion and stress to the individual thereby improving healing. We note that animals can be released just after recovery (i.e., ± minutes) and kept less than one day in captivity. Using the procedure described here, all newts kept their PIT-tags, even after 69 days. We also applied this protocol previously to Bufo bufo (data not shown here) and individuals kept their PIT-tags over 183 days, suggesting that PIT-tags had not been lost by individuals for a long time, and that long-term survey studies should be successful using the protocol described here. Safe protocol for transmitter implantation for radio-tracking surveys Most radio-tracking studies in amphibians showed no significant effect on feeding, body mass or survival rate in implanted individuals (Olders et al., 1985; Madison and Farrand, 1998). In amphibians, the use of internal implants is highly recommended, as external tags may hinder mobility and feeding, despite the use of transmitters with various modes of attachment (Fukuyama et al., 1988; Fiorito et al., 1994; Golay, 1996; Tramontano, 1997). Distances and data obtained by radio-tracking are very useful and can be used both in population dynamics and conservation biology studies. To this day, radio-

88 86 H. Le Chevalier et alii tracking is the only technique that can be used to determine habitat use during seasons where cryptic behavior is displayed, as well as providing indispensable data for conservation management. Using our protocol of implantable transmitters, we observed no infection after operating on newts. An important observation was made concerning the use of surgical adhesive (3M VETBOND Tissue Adhesive, 3M Animal Care Products, United States) for the 10 first operations, which was applied to the sutures at the end of these procedures. Surgical adhesive may be placed over a suture line for added protection from dehiscence and to serve as a waterproof coating (Wright and Whitaker, 2001). It was observed that these newts had ruptured their sutures due to the solidified surgical adhesive. Newts that had ruptured their sutures were then operated on a second time. Two marbled newts died after this second operation (ID numbers and ; Table 1). We believe that these deaths would not have occurred if we not used the surgical adhesive and been obliged to perform a second procedure. As a consequence, the use of surgical adhesive for this particular protocol of transmitter implantation is ill-advised. To summarized, 87.5% animals survived up to 48 days in the wild for successful radio-tracking surveys (see Trochet et al., 2017) post-surgery, until the battery life expired. Transmitter implantation seemed to not impact behaviour of newts, as already observed in other amphibian species (Olders et al., 1985; Madison and Farrand, 1998; Johnson, 2006; Marcec et al., 2016). Radio-tracking studies however have their limitations: the precision of tracking depends on the number of location points and the presence of an observer may influence the behaviour of the studied animal, even if this last effect is reduced (Brown et al. 2013). Moreover, in small-bodied animals such as newts, the main limitation of radio-tracking studies using implantable transmitters is the transmitter size and battery duration, leading to a restricted signal range (Gourret et al., 2011). Despite the reduction of transmitter size and increased battery life, this restriction could become problematic with very mobile species, for example during migration (van Gelder et al., 1986; Sinsch, 1990). CONCLUSION We demonstrated that, in disinfected conditions and following an anesthesia protocol before marking, PIT-tagging can be a very useful marking method, without any tag loss or significant impact on individual health and behavior. Surgical procedures, such as internal transmitter implantation could also be performed under the same anesthesia protocol. Our findings also showed that anesthesia duration was strongly dependent on the phase of newts (aquatic or terrestrial), which is related to certain modifications of the skin, such as permeability. Contrary to the anesthesia duration, the recovery duration was only related to the procedure used (PIT-tagging vs. transmitter implantation). ACKNOWLEDGMENTS Financial support was provided by the Fondation de France and by the French Agence Nationale de la Recherche through the AnaEE project ( Analysis and Experimentation on Ecosystems ; project no ), and partly with the FEDER and AELB funds managed by the EPL (Etablissement Public Loire) and SNB (French National Biodiversity Strategy). This work was supported by the French Laboratory of Excellence project TULIP (ANR-10-LABX-41; ANR-11-IDEX ) and benefited from an Investissement d Avenir grant managed by Agence Nationale de la Recherche (CEBA, ref. ANR- 10-LABX-25-01). Authors have complied with all applicable Animal Care guidelines, and all required French permits (capture permit and authorization for experimentation with animals) have been obtained. AT and HLC collected newts (permit nos and ). DP and SG developed the transmitter implantation protocol in the marbled newt. OC performed surgery (animal experimentation accreditation n A09-1). HLC and AT analyzed the data and wrote the manuscript, which was corrected by all authors. REFERENCES Andreone, F. (1986): Considerations on marking methods in newts, with particular reference to a variation of the belly pattern marking technique. Bull. Brit. Herpetol. Soc. 16: 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: Arntzen, J.W., Jehle, R., Bosch, J., Miaud, C., Tejedo, M., Lizana, M., Martínez-Solano, I., Salvador, A., García- París, M., Recuero Gil, E., Sá-Sousa, P., Marquez, R. (2009): Triturus marmoratus. The IUCN Red List of Threatened Species. Version < org>. Downloaded on 04 May Bancila, R. I., Hartel, T., Plaiasu, R., Smets, J., Cogalniceanu, D. (2010): Comparing three body condition indices in

89 Marking techniques in the Marbled Newt 87 amphibians: a case study of yellow-bellied toad Bombina variegata. Amphibia-Reptilia 31: 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: Brannelly, L.A., Berger, L., Skerratt, L.F. (2014): Comparison of three widely used marking techniques for adult anuran species Litoria verreauxii alpina. Herpetol. Conserv. Biol. 9: Brown, D.D., Kays, R., Wikelski, M., Wilson, R., Klimley, A.P. (2013): Observing the unwatchable through acceleration logging of animal behavior. Anim. Biotelem.1: 20. Christy, M. (1996): The efficacy of using Passive Integrated Transponder (PIT) tags without anaesthetic in free-living frogs. Aust. Zool. 30: Connette, G.M., Semlitsch, R.D. (2011): Successful use of a passive integrated transponder (PIT) system for below-ground detection of plethodontid salamanders. Wildlife Res. 39: 1-6. Cooke, S.J., Midwood, J.D., Thiem, J.D., Klimley, P., Lucas, M.C., Thorstad, E.B., Eiler, J., Holbrook, C., Ebner, B.C. (2013): Tracking animals in freshwater with electronic tags: past, present and future. Anim. Biotelem. 1: Crall, J., Stewart, C., Berger-Wolf, T., Rubenstein, D., Sundaresan, S. (2013): Hotspotter-patterned species instance recognition. Pp in 2013 IEEE Workshop on Application of Computer Vision. Cucherousset, J., Marty, P., Pelozuelo, L., Roussel, J.M. (2008): Portable PIT detector as a new tool for nondisruptively locating individually tagged amphibians in the field: a case study with Pyrenean brook salamanders (Calotriton asper). Wildlife Res. 35: Delarze, R., Ciardo, F., Pellet, J. (2000): Identification individuelle de crapauds sonneurs (Bombina variegata, Anura): application à l estimation des populations. Bull. Murithienne 118: Denoël, M., Hervant, F., Schabetsberger, R., Joly, P. (2002): Short- and long-term advantages of an alternative ontogenetic pathway. Biol. J. Linnean Soc. 77: Diego-Rasilla, F.J., Luengo, R.M. (2002): Celestial orientation in the marbled newt (Triturus marmoratus). J. Ethol. 20: Drechsler, A., Helling, T., Steinfartz, S. (2015): Genetic fingerprinting proves cross-correlated automatic photoidentification of individuals as highly efficient in large capture mark recapture studies. Ecol. Evol. 5: Elgue, E., Pereira, G., Achaval-Coppes, F., Maneyro, R. (2014): Validity of photo-identification technique to analyze natural markings in Melanophryniscus montevidensis (Anura: Bufonidae). Phyllomedusa 13: Feldheim, K.A., Gruber, S.H., De Marignac, J.R.C., Ashley, M.V. (2002): Genetic tagging to determine passive integrated transponder tag loss in lemon sharks. J. Fish Biol. 61: Ferner, J.W. (2007): A review of marking and individual recognition techniques for Amphibians and Reptiles. Society for the study of Amphibians and Reptiles Press, Salt Lake City. Fiorito, S., Cazzanti, P., Rolando, B., Castellano, S., Rolando, A., Giacoma, C. (1994): Post-breeding dispersal by adult common frogs Rana temporaria studied by radio-tracking. Studi Trentini Sci. Nat. Acta Geol. 71: Fukuyama, K., Kusano, T., Nakane, M. (1988): A radio tracking study of the behavior of females of the frog Buergeria buergeri (Rhacophoridae, Amphibia) in a breeding stream in Japan. Jpn. J. Herpetol. 12: Funk, W.C., Donnelly, M.A., Lips, K.R. (2005): Alternative views of amphibian toe-clipping. Nature 433: 193. Gibbons, J.W., Andrews, K.M. (2004): PIT tagging: simple technology at its best. BioScience 54: Golay, N. (1996): Die Kreuzkröte (Bufo calamita) Laur. als Pionierart. Thesis Medizinische Biologie, University of Basel, Switzerland. Gourret, A., Alford, R.A., Schwarzkopf, L. (2011): Very small, light dipole harmonic tags for tracking small animals. Herpetol. Rev. 42: Heard, G.W., Scroggie, M.P., Malone, B.S. (2012): The life history and decline of the threatened Australian frog, Litoria raniformis. Austral. Ecol. 37: Jakob, E., Marshall, S., Uetz, G. (1996): Estimating fitness: a comparison of body condition indices. Oikos 77: Jehle, R. (2000): The terrestrial summer habitat of radiotracked crested and marbled newts (Triturus cristatus and T. marmoratus). Herpetol. J. 10: Jehle, R., Arntzen, J.W. (2000): Post-breeding migrations of newts (Triturus cristatus and T. marmoratus) with contrasting ecological requirements. J. Zool. 251: Jehle, R., Hödl, W. (1998): Pits versus patterns: effects of transponders on recapture rate and body condition of Danube crested newts (Triturus dobrogicus) and common spadefoot toads (Pelobates fuctus). Herpetol. J. 8: Johnson, J.R. (2006): Success of intracoelomic radiotransmitter implantation in the treefrog, Hyla versicolor. Lab. Animal. 35: Lama, F.D., Rocha, M.D., Andrade, M.Â., Nascimento, L.B. (2011): The use of photography to identify individual tree frogs by their natural marks. South. Am. J. Herpetol. 6:

90 88 H. Le Chevalier et alii Lettink, M., Armstrong, D.P. (2003): An introduction to using mark-recapture analysis for monitoring threatened species. Department of Conservation Technical Series. 28A: Madison, D.M., Farrand, L. (1998): Habitat use during breeding and emigration in radio-implanted tiger salamanders, Ambystoma tigrinum. Copeia. 1998: Marcec, R., Kouba, A., Zhang, L., Zhang, H., Wang, Q., Zhao, H., Jiang, W., Willard, S. (2016): Surgical implantation of coelomic radiotransmitters and postoperative survival of chinese giant salamanders (Andrias davidianus) following reintroducion. J. Zoo. Wildl. Med. 47: Matthé, M., Schönbrodt, T., Berger, G. (2008): Computergestützte Bildanalyse von Bauchfleckenmustern des Kammmolchs (Triturus cristatus). Z. Feldherpetol. 15: McCarthy, M.A., Parris, K.M. (2004): Clarifying the effect of toe clipping on frogs with Bayesian statistics. J. Appl. Ecol. 41: Mettouris, O., Megremis, G., Giokas, S. (2016): A newt does not change its spots: using pattern mapping for the identification of individuals in large populations of newt species. Ecol. Res. 31: Mitchell, M.A. (2009): Anesthetic considerations for amphibians. J. Exot. Pet. Med. 18: Morrison, T.A., Keinath, D., Estes-Zumpf, W., Crall, J.P., Stewart C.V. (2016): Individual Identification of the Endangered Wyoming Toad Anaxyrus baxteri and Implications for Monitoring Species Recovery. J. Herpetol. 50: 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: Nöllert, A., Nöllert, C. (2003): Guide des amphibiens d Europe: biologie, identification, répartition. Delachaux et Niestlé, Paris. Olders, J.H.J., van Gelder, J.J., Krammer, J. (1985): A thermo-sensitive transmitter for radiotracking small animals. Neth. J. Zool. 35: Ott, J.A., Scott, D.E. (1999): Effects of toe-clipping and PIT-tagging on growth and survival in metamorphic Ambystoma opacum. J. Herpetol. 33: Parris, K.M., McCarthy, M.A. (2001): Identifying effects of toe-clipping on anuran return rates: the importance of statistical power. Amphibia-Reptilia 22: Perret, N., Joly, P. (2002): Impacts of tattooing and pittagging on survival and fecundity in the Alpine Newt (Triturus alpestris). Herpetologica 58: R Development Core Team. (2014): R: A language and environment for statistical computing. Vienna (Austria): R Foundation for Statistical Computing. ISBN , URL Ribeiro, J., Rebelo, R. (2011): Survival of Alytes cisternasii tadpoles in stream pools: a capture-recapture study using photo-identification. Amphibia-Reptilia 32: Ribéron, A., Miaud, C. (2000): Home range and shelter use in Salamandra lanzai (Caudata, salamandridae). Amphibia-Reptilia 21: Rittenhouse, T.A.G., Semlitsch, R.D. (2006): Grasslands as movement barriers for a forest-associated salamander: migration behavior of adult and juvenile salamanders at a distinct habitat edge. Biol Conserv. 131: Sannolo, M., Gatti, F., Mangiacotti, M., Scali, S., Sacchi, R. (2016): Photo-identification in amphibian studies: a test of I3S Pattern. Acta. Herpetol. 11: Sinsch, U. (1990): Migration and orientation in anuran amphibians. Ethol. Ecol. Evol. 2: Thompson, W.L., White, G.C., Gowan, C. (1998): Monitoring vertebrate populations. 1st edition. Academic Press, San Diego. Tracy, C.R., Christian, K.A., McArthur, L.J., Gienger, C.M. (2011): Removing the rubbish: Frogs eliminate foreign objects from the body cavity through the bladder. Biol. Lett. 7: Trochet, A., Le Chevalier, H., Calvez, O., Barthe, L., Isselin-Nondedeu, F., Picard, D., Debelgarric, M., Pégourié, N., Rocher, R., Ribéron, A. (2017): Post-breeding movements in marbled newts (Caudata, Salamandridae): a comparative radiotracking study in two habitat types. Herpetologica 73: 1-9. Tramontano, R. (1997): Continuous radio tracking of the common frog, Rana temporaria. In: Herpetologica Bonnensis, pp Bohme, W., Bischoff, W., Ziegler, T., Eds, Bonn. van Gelder, J.J., Aarts, H.M.J., Staal, H.W.M. (1986): Routes and speed of migrating toads (Bufo bufo): A telemetric study. Herpetol. J. 1: van Tienhoven, A.M., Den Hartog, J.E., Reijns, R.A., Peddemors, V.M. (2007): A computer-aided program for pattern-matching of natural marks on the spotted raggedtooth shark Carcharias taurus. J. Appl. Ecol. 44: Waye, H.L. (2013): Can a tiger change its spots? A test of the stability of spot patterns for identification of individual tiger salamanders (Ambystoma tigrinum). Herpetol Conserv. Biol. 8: Wells, K.D. (2007): The ecology and behavior of amphibians. University of Chicago Press, Chicago. Winandy, L., Denoël, M. (2011): The use of visual and automatized behavioral markers to assess methodologies: a study case on PIT-tagging in the Alpine newt. Behav. Res. 43: Wright, K.N., Whitaker, B.R. (2001): Amphibian medicine and captive husbandry, 1st edition. Krieger Publishing Company, Malabar.

91 Acta Herpetologica 12(1): 89-93, 2017 DOI: /Acta_Herpetol Evidence for directional testes asymmetry in Hyla gongshanensis jindongensis Qing Gui Wu 1, Wen Bo Liao 2, * 1 Ecological Security and Protection Key Laboratory of Sichuan Province, Mianyang Normal University, Mianyang , China 2 Key Laboratory of Southwest China Wildlife Resources Conservation (Ministry of Education), China West Normal University, Nanchong , Sichuan, China. *Corresponding author. Liaobo_0_0@126.com Submitted on 2016, 30 st January; revised on 2016, 30 st November; accepted on 2016, 6 th December Editor: Giovanni Scillitani Abstract. The compensation hypothesis predicts that one testis may grow more for compensating for a reduced function in the other testis, thus exhibiting a directional asymmetry in testis size. In this study, we tested the prediction of the compensation hypothesis in the Chinese endemic Tree Frog Hyla gongshanensis jingdongensis in a population in Kegong Reserve site of Yunan Province in western China. For fifty-three male samplings, we found that the left testis size was significantly bigger than the right testis, which exhibited a significantly directional testis asymmetry, consistent with the evidence that mainly the left testis is functional with the right testis having a compensatory role, i.e. the left testis would increase in size if the right testis became non-functional. However, the relative testes size and the degree of testes asymmetry were not correlated with body condition in this species, suggesting that the testes asymmetry can not reflect male quality: high-quality individuals would not have more asymmetric testes. Keywords. Jingdong tree frog, Hylidae, body condition, compensatory function, directional testes asymmetry. The compensation hypothesis predicts that one testis may grow more for compensating for a reduced function in the other testis (Møller, 1994). As a result, testis size exhibits a directional asymmetry where the testis on one side, often the left, is bigger than the other side, the right one in bird because the right testis increases in size to compensate for any reduction in function of the left (Møller, 1994; Birkhead et al., 1997). Hence, the directional testes asymmetry (DTA) is assumed where the left testis size increases if the right testis serves a compensatory role in frogs due to developmental stress, such as growth season length (Hettyey et al., 2005). In recent years, the testis asymmetry has been confirmed in some anurans species (Hettyey et al., 2005; Zhou et al., 2011; Liu et al., 2011; Mi et al., 2012). Testis size asymmetry is widely used as a measure of male body condition (Møller, 1994; Hettyey et al., 2005). As a result, the degree of directional asymmetry in testes ISSN (print) ISSN (online) size reflects male quality: high-quality individuals have more asymmetric testes. Because possibilities for energy acquisition were correlated with male quality, the energy acquisition affected the degree of directional testes asymmetry (Møller, 1994). For instance, male individual with good condition develops higher degree of directional testes asymmetry than that with poor condition in the common frog, Rana temporaria (Hettyey et al., 2005). By contrast, the degree of testes size asymmetry is not correlated with body condition in the Guenther s frog, Hylarana guentheri (Liu et al., 2011). Our primary aim in this study was to test the compensation hypothesis in a frog. A prerequisite for testing the hypothesis is good evidence that a particular highquality male is attractive to females (Jin et al., 2016a). We studied the Jingdong Tree Frog (Hyla gongshanensis jingdongensis), a species endemic to Sichuan and Yunnan province in China. The species is restricted to a nar- Firenze University Press

92 90 Q. Gui Wu, W. Bo Liao row range of altitudes, ranging from 1500 to 2470 m (Fei and Ye, 2001). Males have well-developed vocal sacs and attract their mates by their calls during the breeding period (Fei and Ye, 2001). Besides, little information on testes size asymmetry in H. g. jingdongensis is available. Here, we tested the prediction of the compensation hypothesis that the difference in size of the left and right testis arose in H. g. jingdongensis. Specially, we examined whether the directional testes asymmetry is occurred in all samplings, and whether there is a positive correlation between body condition and the degree of the directional testes asymmetry. The population is located at Kegong Reserve site in Bama Snow Mountain Nature Reserve of Yuanan province, western China (27 33'N, 99 19'E; a.s.l m). The Reserve is characterized by a subtropical climate with a strong seasonality due to the high altitude which has an annual average temperature of C (Liao et al., 2015; Liao et al., 2016a). All individuals were caught by hand at night from March to May in We collected a total of 53 males and brought them to the laboratory. Until processing, males were kept individually in a rectangular tank ( m; L W H) with a water depth of 25 cm at room temperature. Two day after collection, body size (snout vent length, SVL) of each frog was measured to the nearest 0.1 mm using a caliper and body mass was weighted to the nearest 0.1 g using an electronic balance. We used the single-pithing to sacrifice all animals and then anatomized them (Liao et al., 2016b; Mai et al., 2017; Lüpold et al., 2017). Left and right testes were removed, and weighed to the nearest 0.1 mg using an electronic balance. Following the suggestion of Møller and Swaddle (1997), we calculated the degree of the directional testis size asymmetry as DTA = (left-testis mass right-testis mass)/0.5(left-testis mass + right-testis mass). Relative testes size was calculated as the ratio of observed testes mass to that predicted by the allometric regression equation between testes mass and body size. Body condition was measured as the ratio of observed body mass to that predicted by linear regression by entering male body mass as a dependent variable and body size as an independent variable. Body size, body mass and testes mass were log 10 - transformed to achieve normality. We compared the sizes of the left and right testes using a paired t-test. We performed line regressions to test the relationships between both body mass and body size and testes mass. We also analyzed the relationships between male body condition and both relative testis mass and the degree of the testis asymmetry using Pearson correlation analysis. All analyses were conducted using SPSS 21.0 (Statistical Product and Service Solutions Company, Chicago, USA). Testis mass was positively correlated with body mass (Fig. 1A; F 1, 52 = , r 2 = 0.360, P < 0.001, log (testis mass (mg)) = log (body mass (g)) ). An increase in testes mass with body mass was larger than predicted by the power law (β > 1), providing an evidence for an allometric relationship. Testis mass was also positively correlated with SVL (Fig. 1B; F 1, 52 = 9.414, r 2 = 0.156, P = 0.003, log (testis mass (mg)) = log (SVL (mm)) ). The relative testes size was not correlated with body condition (Fig. 2; Pearson s correlation coefficient: r = , n = 53, P = 0.558). The Jingdong Tree Frog exhibited a directional testes asymmetry: the left testis (4.5 mg ± SD 2.7) was significantly bigger than the right one (3.5 mg ± SD 1.9) in 71.69% of all individuals (Paired t-test: t = 3.67, df = 53, P = 0.001). Left testis mass was positively correlated with right testis mass (r = 0.723, n = 53, P < 0.001). We found a non-significant correlation between the degree of the directional testes asymmetry and body condition (Fig. 3; Pearson s correlation coefficient: r = , n = 53, P = 0.304). Fig. 1. Relationships between (A) body mass and testes mass, (B) SVL and testes mass. Displayed values are log-transformed data.

93 Testes asymmetry in a tree frog 91 Fig. 2. Relationship between body condition and relative testes size in a H. g. jingdongensis population. Fig. 3. Relationship between body condition and the degree of testes asymmetry in a H. g. jingdongensis population. Our results uncovered positive correlations between testis mass and both body mass and SVL in H. g. jingdongensis in a population. Furthermore, the left testis was significantly bigger than the right one, which exhibited a directional testes size asymmetry. However, body condition was non-significantly correlated with both the relative testes size and the degree of the testes asymmetry. Previous studies have shown that the directional testes asymmetry is common in animal taxa (anurans: Hettyey et al., 2005; Liao and Lu, 2010; Zhou et al., 2011; Liu et al., 2011; Liu et al., 2012; Jin et al., 2016b; Chen et al., 2016; birds: Wright and Wright, 1944; Møller, 1994; Birkhead et al., 1997; mammals: Yu, 1998). For H. g. jingdongensis left testis was bigger than right testis in most individuals, demonstrating a directional testes asymmetry. Meanwhile, most individuals had at least one testis smaller than the median for the population, which was considered as natural selection for compensation, consistent with the compensation hypothesis (Møller, 1994). Furthermore, a larger left testis and a smaller right testis might be advantageous because constraints during embryonic development results in the higher efficiency in sperm production of the left testis (Kempenaers et al., 2002; Liao and Lu, 2012). However, if a bigger sperm production on left testicle influences for the decrease on sperm production on right testicle, that otherwise would maintain the same rate of production, this condition does not represent an advantage on testes asymmetry. There is a heritability evidence of male body condition that sexual selection supports females making use of the genetic correlation between body condition and sperm traits in frogs (Byrne et al., 2003). As a result, male individuals with good body condition have larger testis size than that with poor body condition to obtain mates. In the study, we found that the relative testis size was not correlated with male body condition, suggesting that males in good condition did not exhibit the higher ability of sperm competition. Hettyey and Roberts (2007) found that sperm-depletion after mating led to a non-significant correlation between body condition and testis mass. In our study, a non-significant correlation between body condition and relative testes size in H. g. jingdongensis might result from sperm depletion. Hence, we showed a consistent hypothesis for the lack of body condition and asymmetrical directional testicles. Previous studies have shown that the degree of the testes asymmetry may not be a good measure of body condition in birds (Birkhead et al., 1997; Birkhead et al., 1998; Kimball et al., 1997; Kempenaers et al., 2002). By contrast, the house sparrow and the barn swallow exhibit a positive correlation between testes asymmetry and body condition (Møller, 1994). For anurans, some species exhibit a positive correlation between body condition and testes asymmetry while other species do not exhibit correlation between them (Hettyey et al., 2005; Zhou et al., 2011; Liu et al., 2011). For instance, the degree of the testes asymmetry is significantly correlated with body mass in R. temporaria (Hettyey et al., 2005). However, Zhou et al. (2011) found that the degree of the testes asymmetry is negatively correlated with body size in Pelophylax nigromaculata. In this study, the degree of the testes asymmetry did not co-vary with male body condition, suggesting that the degree of the directional testes asymmetry may not be a good indicator of male quality. Similar results have been observed in two species of anurans (Hylarana guentheri, Liu et al., 2011; Rana omeimontis, Liu et al., 2012). Although the degree of the testis asymmetry in H. g. jingdongensis did not reflect good body

94 92 Q. Gui Wu, W. Bo Liao condition, males developed larger left testes and smaller right testes for supporting the compensation hypothesis. In conclusion, consistent with the prediction of the compensation hypothesis, our analyses of difference in left and right testes in H. g. jingdongensis showed a directional testes asymmetry, suggesting that the left testis exhibited functional and the right testis had a compensatory role. However, body condition was not significantly correlated with the degree of the testes asymmetry, suggesting that directional asymmetry in testes size may not be a good measure of male quality. ACKNOWLEDGEMENTS Financial support was provided by the Open Scientific Research Foundation of Ecological Security and Protection Key Laboratory of Sichuan Province, Mianyang Normal University (ESP1405), the Project of Mianyang Normal University (QD2015B07 and 2015A01). The reported experiments comply with the current laws of China concerning animal experimentation, and permission to collect frogs was received from the Ethical Committee for Animal Experiments in China Council on Animal Care (CCAC) guidelines. All experiments involving the sacrifice of these live animals were approved by the Animal Ethics Committee at Mianyang Normal University. REFERENCES Birkhead, T.R., Buchanan, K.L., Devoogd, T.J., Pellatta, E.J, Székelyd, T., Catchpoleb, C.K. (1997): Song, sperm quality and testes asymmetry in the sedge warbler. Anim. Behav. 53: Birkhead, T.R., Fletcher, F., Pellatt, E.J. (1998): Testes asymmetry, condition and sexual selection in birds: an experimental test. Proc. Roy. Soc. B 265: Byrne, P.G., Simmons, L.W., Roberts, J.D. (2003): Sperm competition and the evolution of gamete morphology in frogs. Proc. R. Soc. B 270: Chen, C., Huang, Y.Y., Liao, W.B. (2016): A comparison of testes size and sperm length between Polypedates megacephalus populations at different altitudes. Herpetol. J. 26: Fei, L., Ye, C.Y. (2001): The Colour Handbook of Amphibians of Sichuan. China Forestry Publishing House, Beijing, China. Hettyey, A., Laurila, A., Herczeg, G., Jönsson, K.I., Kovács, T., Merilä, J. (2005): Does testis weight decline towards the Subarctic? A case study on the common frog, Rana temporaria. Naturwissenschaften 92: Hettyey, A., Roberts, J.D. (2007): Sperm traits in the quacking frog (Crinia geogiana), a species with plastic alternative mating tactics. Behav. Ecol. Sociobiol. 61: Jin, L., Mi, Z.P., Liao, W.B. (2016a): Altitudinal variation in male reproductive investments in a polyandrous frog species (Hyla gongshanensis jingdongensis). Anim. Biol. 66: Jin, L., Yang, S.N., Liao, W.B., Lüpold, S. (2016b): Altitude underlies variation in the mating system, somatic condition and investment in reproductive traits in male Asian grass frogs (Fejervarya limnocharis). Behav. Ecol. Sociobiol. 70: Kempenaers, B., Peer, K., Vermeirssen, E.L.M., Robertson, R.J. (2002): Testes size and asymmetry in the tree swallow Tachycineta bicolor: A test of the compensation hypothesis. Avian Sci. 3: Kimball, R.T., Ligon, D.J., Merola-Zwartjes, M. (1997): Testicular asymmetry and secondary sexual characters in red junglefowl. Auk 114: Liao, W.B., Lu, X. (2010): Age structure and body size of the Chuanxi Tree Frog Hyla annectans chuanxiensis from two different elevations in Sichuan (China). Zool. Anz. 248: Liao, W.B., Lu, X. (2012): Adult body size = f (initial size + growth rate age): explaining the proximate cause of Bergman s cline in a toad along altitudinal gradients. Evol. Ecol. 26: Liao, W.B., Liu, W.C., Merilä, J. (2015): Andrew meets Rensch: Sexual size dimorphism and the inverse of Rensch s rule in Andrew s toad (Bufo andrewsi). Oecologia 177: Liao, W.B., Luo, Y., Lou, S.L., Lu, D., Jehle, R. (2016a): Geographic variation in life-history traits: growth season affects age structure, egg size and clutch size in Andrew s toad (Bufo andrewsi). Front. Zool. 13: 6. Liao, W.B., Lou, S.L., Zeng, Y., Kotrschal, A. (2016b): Large brains, small guts: The expensive tissue hypothesis supported in anurans. Am. Nat. 188: Liu, W.C., Huang, Y, Liao, Y.M. (2012): Testes asymmetry of Chinese endemic frog (Rana omeimontis) in relation to body condition and age. N-W. J. Zool. 8: Liu, Y.H., Liao, W.B., Zhou, C.Q., Mi, Z.P., Mao, M. (2011): Asymmetry of testes in Guenther s Frog, Hylarana guentheri (Anura: Ranidae). Asian Herpetol. Res. 2: Lüpold, S., Jin, L., Liao, W.B. (2017): Population density and structure drive differential investment in pre- and

95 Testes asymmetry in a tree frog 93 post-mating sexual traits in frogs. Evolution 71: Mai, C.L., Liao, J., Zhao, L., Liu, S.M., Liao, W.B. (2017): Brain size evolution in the frog Fejervarya limnocharis does neither support the cognitive buffer nor the expensive brain framework hypothesis. J. Zool. Lond. 302: Mi, Z.P., Liao, W.B., Jin, L., Chen, J., Wu, H. (2012): Testis asymmetry and sperm length in Rhacophorus omeimontis. Zool. Sci. 29: Møller, A.P. (1994): Directional selection on directional asymmetry: Testes size and secondary sexual characters in birds. Proc. Roy. Soc. B 258: Møller, A.P., Swaddle, J.P. (1997): Asymmetry, developmental stability, and evolution. Oxford: Oxford University Press. Wright, P.L., Wright, M.H. (1944): The reproductive cycle of the male Red-winged Blackbird. Condor 46: Yu, Z.H. (1998): Asymmetrical testicular weights in mammals, birds, reptiles and amphibians. Internat. J. Androl. 21: Zhou, C.Q., Mao, M., Liao, W.B. (2011): Testis asymmetry in the dark-spotted frog Pelophylax nigromaculata. Herpetol. J. 21:

96

97 Acta Herpetologica 12(1): , 2017 DOI: /Acta_Herpetol The advertisement call of Pristimantis subsigillatus (Anura, Craugastoridae) Florina Stănescu 1,4, Rafael Márquez 2, Paul Székely 3,4, *, Dan Cogălniceanu 1,4,5 1 Ovidius University Constanţa, Faculty of Natural and Agricultural Sciences, Al. Universităţii 1, campus B, Constanţa, Romania 2 Fonoteca Zoológica, Dept. de Biodiversidad y Biología Evolutiva. Museo Nacional de Ciencias Naturales (CSIC), José Gutiérrez Abascal 2, E Madrid, Spain 3 Universidad Técnica Particular de Loja, Departamento de Ciencias Biológicas, San Cayetano Alto, calle Marcelino Champagnat s/n, Loja, Ecuador 4 Asociación Chelonia, Str. Pașcani, nr. 5, București, , Romania. *Corresponding author. jpszekely@utpl.edu.ec 5 Universidad Nacional de Loja, CITIAB, Ciudadela universitaria La Argelia, EC Loja, Ecuador Submitted on 2016, 11 th August; revised on 2016, 18 th October; accepted on 2016, 26 th October Editor: Adriana Bellati Abstract. We describe for the first time the advertisement call of Pristimantis subsigillatus from southern Ecuador. Our study provides a detailed quantitative characterization of the advertisement call of P. subsigillatus, filling a gap in our knowledge of this genus, the most speciose among vertebrates. Males called perched on vegetation m above ground, always during mild rain. The advertisement call is composed of a single note with a duration of ms, with an initial short pulse (3-10 ms) followed by a longer tonal component. Call rates ranged between 4-12 calls/ min. The dominant frequency varied between khz. Keywords. Advertisement call, Anura, Craugastoridae, acoustics. Vocalizations play an important role in the biology and evolution of anurans since acoustic parameters may encode information about species, sex, fitness, reproductive availability, territoriality or distress (Gerhardt and Huber, 2002; Wells and Schwartz, 2006). Acoustic parameters can be genetically and environmentally shaped, therefore deciphering the information within vocal repertoires and analysing their variability in time and space is vital to tracing and understanding evolutionary patterns (Duellman and Trueb, 1994; Gerhardt, 1994; Cocroft and Ryan, 1995; Wells and Schwartz, 2006). Four main types of calls are currently described in anurans (Duellman and Trueb, 1994): advertisement, reciprocation, release and distress, and their functionality is relatively well known. Advertisement calls represent one of the most conspicuous and important components of anuran communication, acting as a premating isolating mechanism and thus providing valuable data for phylogenetic studies (Cocroft and Ryan, 1995; Vences and Wake, 2007). With more than 470 species, direct-developing frogs of the genus Pristimantis make up for almost one third of the anuran species of Ecuador (Ron et al., 2011), and the genus is by far the most diverse among terrestrial vertebrates (Hedges et al., 2008; Frost, 2016). Species of the genus Pristimantis have a high phenotypic variability within populations and scant morphological differences among species (Crawford and Smith, 2005). This frequently causes misidentifications even in museum specimens (Padial et al., 2008). Advertisement calls can be a valuable tool in describing or distinguishing among species when morphological characters are not sufficient in species identification (e.g., Köhler and Lötters, 1999; Reichle et al., 2001). ISSN (print) ISSN (online) Firenze University Press

98 96 F. Stănescu et alii Pristimantis subsigillatus (Fig. 1) was described by Boulenger (1902) from Salidero, Esmeraldas Province, Ecuador and is considered a locally abundant species in western Ecuador and south-western Colombia at elevations of m a.s.l. (Lynch, 1980; Lynch and Duellman, 1997; Frenkel et al., 2013). This species is encountered most frequently in forests, at night, perching on vegetation at heights of m above the ground; males usually call during rainy nights from vegetation at 2-5 m height (Lynch and Duellman, 1997). The advertisement call was described by Lynch (1980) as a single sharp explosive tweet or as a single, clear, bell-like note. However, to the best of our knowledge, there is no quantitative characterization of these calls. Our paper provides the first quantitative description of the advertisement call in P. subsigillatus. DC recorded the advertisement calls of three P. subsigillatus males from Buenaventura Ecological Reserve, EL Oro Province, Ecuador (Table 1). Calls were recorded with an Olympus LS-11 Linear PCM Recorder and a RØDE NTG2 condenser shotgun microphone, at 44.1 khz sampling frequency and 16-bit resolution, in wav file format. Air temperature and humidity were measured with a data logger (Lascar Electronics, model EL-USB- 2-LCD, accuracy: ± 0.5 C; ± 5%). The three focal males were captured following the recording sessions after being photographed. The snout-vent length (SVL) was measured to the nearest 0.1 mm with dial callipers, and then the animals were anaesthetized with benzocaine, fixed with formalin 10% and preserved in 70% alcohol. The specimens were deposited in the amphibian collection of Museum of Zoology of the Pontificia Universidad Católica del Ecuador (specimen ID is provided in Table 1). Following Toledo et al. (2015) recordings were deposited in Fonoteca Zoológica - at Museo Nacional de Ciencias Naturales (CSIC), Madrid, Spain (records ID are provided in Table 1). Individuals were identified as P. subsigillatus based on the characters described by Lynch and Duellman (1997), Fig. 1. Male Pristimantis subsigillatus calling (Photo credit: José Seoane Rodríguez). especially as having skin on dorsum finely shagreened, that of venter areolate, without dorsolateral folds, discoidal fold prominent, tympanic membrane and annulus evident, snout subacuminate in dorsal view, truncate or protruding in profile, snout bearing a small papilla at tip, canthus rostralis relatively sharp, upper eyelid lacking tubercles, outer fingers bearing broad discs, heel lacking or bearing very small tubercle and fifth toe much longer than the third (Lynch, 1980; Lynch and Duellman, 1997). Despite the fact that identification of Pristimantis species based only on morphological characters is challenging, we are confident that the recorded individuals belong to P. subsigillatus. We have conducted a thorough inventory of amphibians in the area and are familiar with the species found there (e.g., Székely et al. 2016), and P. subsigillatus cannot be confounded with other species of the genus present based on morphological characters. At present, P. subsigillatus is not assigned to any species group after it was removed from the former, nonmonophyletic, Pristimantis unistrigatus group (Padial et al., 2014). Pristimantis subsigillatus is most similar to P. nyctophylax from which it is distinguished by the absence Table 1. Information regarding the recordings used in this study. Air temperature = Temp; humidity = H; distance from the tip of the microphone and the focal male = Dist. Specimen ID/ FonoZoo ID SVL (mm) Coordinates Altitude (m) Date Time (h) Temp ( C) H (%) Dist (cm) QCAZ47284/ QCAZ62538/ QCAZ62543/ S W S W S W September : September : September :

99 Call of Pristimantis subsigillatus 97 of observable tubercles on the eyelids and heels, presence of small terminal papilla at the tip of the snout, presence of numerous supernumerary plantar tubercles and of the inner tarsal tubercle. Additionally, the eye of P. nyctophylax is very distinctive in having orange or red sclera (Lynch and Duellman, 1997). Also, this species occurs usually at higher elevations ( m a.s.l.) compared to P. subsigillatus. We used Raven Pro 1.4 software ( cornell.edu/raven) to analyse 54 advertisement calls. We measured the temporal parameters from the oscillograms and the spectral parameters from spectrograms obtained through Hanning function, at a window size of 1024 samples, and 50% overlap. We measured the duration, rise time proportion, dominant frequency (Df), and aggregate entropy (Entropy) of the calls and component parts. We computed rise time proportion, as the ratio between the period from the onset of the sound and the moment of maximum amplitude within the analysed call, and the call duration (Márquez et al., 2005). We also computed the dominant frequency modulation within a call (DfM) as the difference between the dominant frequencies of the two component parts (Márquez et al,. 1996). The aggregate entropy provides a measure of the overall disorder in the sound, by analysing the distribution of energy in the spectrogram; zero entropy values characterize single pure tones, while higher entropy values correspond to greater disorder in the acoustic spectrum (Charif et al., 2010). Following Gerhardt (1991), we computed a coefficient of within-individual variation (CV% = SD/mean X 100) in order to differentiate between static (CV% < 5) and dynamic (CV% > 12) call parameters. The three P. subsigillatus males were calling after sunset perched on leaves or branches up to 2.5 m above ground, during mild rain. In each case, several other males were calling in the background. The advertisement call was composed of a single note with two consecutive parts: an initial short pulsed part, followed by a longer tonal part with harmonics (Fig. 2). Call rates ranged Table 2. Quantitative description of P. subsigillatus advertisement calls. Mean ± SD (above the line) and min-max values (below the line) are provided for all acoustic parameters analysed. n = call sample size, Df = Dominant frequency, DfM = Dominant frequency modulation. Specimen Measurement Duration (ms) Rise time (%) Df (Hz) DfM (Hz) Entropy (u) (n = 9) Call 80 ± ± ± ± ± Part 1 7 ± ± ± ± Part 2 73 ± ± ± ± (n = 15) Call 63 ± ± ± ± ± Part 1 3 ± ± ± ± Part 2 60 ± ± ± ± (n = 30) Call 70 ± ± ± ± ± Part 1 7 ± ± ± ± Part 2 63 ± ± ± ± All (n = 54) Call 70 ± ± ± ± ± Part 1 6 ± ± ± ± Part 2 64 ± ± ± ±

100 98 F. Stănescu et alii Table 3. Within-individual coefficient of variation of the analyzed acoustic parameters. n = call sample size, Df = Dominant frequency, DfM = Dominant frequency modulation. CV values < 5% are highlighted in grey (static parameters sensu Gerhardt, 1991). Specimen Measurement Duration Rise time Df DfM Entropy (n = 9) Call Part Part (n = 15) Call Part Part (n = 30) Call Part Part between 4-6 calls/min (individual 47284) to 9-12 calls/ min (individuals and 62543). Descriptive statistics of the analysed acoustic parameters and their intra-individual variation are presented in detail in Tables 2 and 3. The first part of the note was composed of 2 to 4 pulse groups in individual 47284, while in the other two individuals it consisted of a single pulse group. In individual 62543, the pulse was bell-shaped (Fig. 2C). The second part of the note was always longer than the first, and had a harmonic structure, with the fundamental harmonic corresponding to the dominant frequency. The dominant frequency was higher in the second part and therefore the calls presented an upward frequency modulation pattern. Call duration, part 2 duration and dominant frequency showed low within individual variability (CV < 5%), and therefore can be considered static acoustic parameters, useful in distinguishing among individuals of the population. The first part of the call presented the highest acoustic variability. Our study provides the first detailed characterization of the advertisement call of P. subsigillatus, filling a gap in our knowledge of this poorly studied group. The calls described in this study are markedly different from the calls of Pristimantis eugeniae, P. nyctophylax and P. phoxocephalus which are considered similar species by Lynch and Duellman (1997). P. eugeniae has calls with a much higher dominant frequency ( Hz) and lack the pulsed initial part of the call, P. nyctophylax has calls with lower dominant frequencies, around 2100 Hz. As for P. phoxocephalus, the recordings posted in Amphibia- WebEcuador show a high variation of dominant frequencies, from 1900 Hz to more than 3000 Hz, suggesting that maybe more than one taxon was included in these recordings. Advertisement calls are an important tool in describing and characterizing anuran species, since they act as a premating isolating mechanism (Ryan, 1988). Based on detailed acoustic analyses, taxonomic relationships can be elucidated (e.g., Padial et al., 2008). As the most speciose vertebrate genus, Pristimantis species have been subject to taxonomic debates for a long time (e.g., Hedges et al., 2008; Padial et al., 2014). In August 2016, the total number of Pristimantis species recognized was 480, of which 121 new species were described during the last decade only (Frost, 2016). Most descriptions do not include calls and are often based only on museum specimens. Recent reviews based only on molecular data could not shed light on the systematics of this genus (e.g., Hedges et al., 2008; Padial et al., 2014). In order to facilitate taxonomic work, this large genus has been subdivided into several species groups (Lynch and Duellman, 1997). There is a huge variation in the call parameters of the frogs within the genus Pristimantis (Padial et al., 2007), and acoustic parameters could prove to be an important tool in the revision of this genus. ACKNOWLEDGEMENTS The Secretaría de Educación Superior, Ciencia, Tecnología e Innovación, Republic of Ecuador (SENESCYT) provided funding for Paul Székely and Dan Cogălniceanu through the Prometeo Project. The Jocotoco Foundation provided access to Buenaventura Forest Reserve. Florina Stănescu, Paul Székely and Dan Cogălniceanu received additional support from the SYNTHESYS Project. The SYNTHESYS Project is financed by the European Community Research Infrastructure Action under the FP7 Capacities Specific Programme. Partial call analyses and call deposit were funded by project TATANKA (CGL ), Ministerio de Ciencia e Innovación,

101 Call of Pristimantis subsigillatus 99 Fig. 2. Oscillogram and spectrogram views of the advertisement calls of the three males of Pristimantis subsigillatus: specimen (A), (B), and (C). Spectrogram window size: 256 samples, 50% overlap, hop size 128 samples.

102 100 F. Stănescu et alii FCW (CGL E). We are grateful to Dr. Santiago Ron, Pontificia Universidad Católica de Ecuador for his support. Collecting permits were granted to Pontificia Universidad Católica de Ecuador as No IC-FAU- DNB/MA and No IC-FAU-DNB/MA. REFERENCES Boulenger, G.A. (1902): Descriptions of new batrachians and reptiles from north-western Ecuador. Ann. Mag. Nat. Hist. 9: Charif, R.A., Waack, A.M., Strickman, L.M. (2010): Raven Pro 1.4 User s Manual. Cornell Lab of Ornithology, Ithaca, New York. Cocroft, R.B., Ryan, M.J. (1995): Patterns of advertisement call evolution in toads and chorus frogs. Anim. Behav. 49: Crawford, A.J., Smith, E.N. (2005): Cenozoic biogeography and evolution in direct-developing frogs of Central America (Leptodactylidae: Eleutherodactylus) as inferred from a phylogenetic analysis of nuclear and mitochondrial genes. Mol. Phylogenet. Evol. 35: Duellman, W.E., Trueb, L. (1994): Biology of Amphibians. Johns Hopkins University Press, Baltimore. Frenkel, C., Yánez-Muñoz, M.H., Guayasamín, J.M., Varela-Jaramillo, A. Ron, S.R. (2013): Pristimantis subsigillatus. In: Ron, S. R., Guayasamin, J. M., Yanez-Muñoz, M. H., Merino-Viteri, A., Ortiz, D. A. y Nicolalde, D. A. (2014). AmphibiaWebEcuador. Version Museo de Zoología, Pontificia Universidad Católica del Ecuador. Electronic Database accessible at zoologia.puce.edu.ec/vertebrados/anfibios/fichaespecie.aspx?id=1479, accessed 27 December Frost, D.R. (2016): Amphibian Species of the World: an Online Reference. Version 6.0 American Museum of Natural History, New York. Electronic Database accessible at accessed 5 August Gerhardt, H.C. (1991): Female mate choice in tree frogs: static and dynamic acoustic criteria. Anim. Behav. 42: Gerhardt, H.C. (1994): The evolution of vocalization in frogs and toads. Annu. Rev. Ecol. Syst. 25: Gerhardt, H.C., Huber, F. (2002): Acoustic communication in insects and anurans: common problems and diverse solutions. University of Chicago Press, Chicago. Hedges, S.B., Duellman, W.E., Heinicke, M.P. (2008): New World direct-developing frogs (Anura: Terrarana): molecular phylogeny, classification, biogeography, and conservation. Zootaxa 1737: Köhler J., Lötters S. (1999): New species of the Eleutherodactylus unistrigatus group (Amphibia: Anura: Leptodactylidae) from montane rain forest of Bolivia. Copeia 1999: Lynch, J.D. (1980): Systematic status and distribution of some poorly known frogs of the genus Eleutherodactylus from the Chocoan lowlands of South America. Herpetologica 36: Lynch, J.D., Duellman, W.E. (1997): Frogs of the genus Eleutherodactylus in Western Ecuador: systematics, ecology, and biogeography. Special Publication 23. Natural History Museum, University of Kansas, Lawrence. Márquez, R., De la Riva, I., Bosch, J. (1996): Advertisement calls of three glass frogs from the Andean forests (Amphibia: Anura: Centrolenidae). Herpetol. J. 6: Márquez, R., Penna, M., Marques, P., Do Amaral, J.P.S. (2005): Diverse types of advertisement calls in the frogs Eupsophus calcaratus and E. roseus (Leptodactylidae): a quantitative comparison. Herpetol. J. 15: Padial, J.M., Castroviejo-Fisher, S., Koehler, J., Domic, E., De la Riva, I. (2007): Systematics of the Eleutherodactylus fraudator species group (Anura: Brachycephalidae). Herpetol. Monogr. 21: Padial, J.M., Köhler, J., Muñoz, A., De la Riva, I. (2008): Assessing the taxonomic status of tropical frogs through bioacoustics: geographical variation in the advertisement calls in the Eleutherodactylus discoidalis species group (Anura). Zool. J. Linn. Soc-Lond. 152: Padial, J.M., Grant, T., Frost, D.R. (2014): Molecular systematics of Terraranas (Anura: Brachycephaloidea) with an assessment of the effects of alignment and optimality criteria. Zootaxa 3825: Reichle, S., Lötters, S., De la Riva, I. (2001): A new species of the discoidalis group of Eleutherodactylus (Anura, Leptodactylidae) from inner-andean dry valleys of Bolivia. J. Herpetol. 35: Ron, S.R., Guayasamin, J.M., Menendez-Guerrero, P.A. (2011): Biodiversity and Conservation status of amphibians of Ecuador. In: Amphibian Biology 9, Status of decline of amphibians: Western Hemisphere: Uruguay, Brazil, Colombia, and Ecuador, pp Heatwole, H., Barrio-Amorós, C.L., Wilkinson, J.W., Eds, Surrey Beatty & Sons, Baulkham Hills. Ryan, M.J. (1988): Constraints and patterns in the evolution of anuran acoustic communication. In: The evolution of the amphibian auditory system, pp Fritzch, B., Ryan, M., Wilczynski, W., Walkowiak, W.,

103 Call of Pristimantis subsigillatus 101 Hetherington, T., Eds, John Wiley & Sons, New York. Székely, P., Cogălniceanu, D., Székely D., Páez, N., Ron, S.R. (2016): A new species of Pristimantis from southern Ecuador (Anura, Craugastoridae). ZooKeys 606: Toledo, L.F., Tipp, C., Márquez, R. (2015): The value of audiovisual archives. Science 347: Vences, M., Wake, D. (2007): Speciation, species boundaries and phylogeography of amphibians. In: Amphibian Biology 7, pp Heatwole, H., Tyler, M.J., Eds, Surrey Beatty & Sons, Chipping Norton. Wells, K.D., Schwartz, J.J. (2006): The behavioural ecology of anuran communication. In: Hearing and sound communication in amphibians, pp Narins, P., Feng, A.S., Fay, R.R., Eds, Springer, New York.

104

105 Acta Herpetologica 12(1): , 2017 DOI: /Acta_Herpetol Nematodes infecting Anotosaura vanzolinia (Squamata: Gymnophthalmidae) from Caatinga, northeastern Brazil Bruno Halluan S. Oliveira¹, Adonias A. Martins Teixeira¹, *, Romilda Narciza M. Queiroz¹, João A. Araujo-Filho¹, Diêgo A. Teles¹, Samuel V. Brito 2, Daniel O. Mesquita¹ 1 Programa de Pós-Graduação em Ciências Biológicas (Zoologia), Departamento de Sistemática e Ecologia DSE, Centro de Ciências Exatas e da Natureza CCEN, Universidade Federal da Paraíba UFPB, Cidade Universitária, Campus I, CEP , João Pessoa, PB, Brazil. *Corresponding author. adoniasteixeira01@gmail.com 2 Centro de Ciências Agrárias e Ambientais, Universidade Federal do Maranhão, Boa Vista, CEP Chapadinha, MA, Brazil. Submitted on 2016, 30 th August; revised on 2016, 13 th December; accepted on 2016, 21 st December Editor: Paolo Casale Abstract. The present study investigated the composition of helminth parasites of the completely unknown lizard Anotosaura vanzolinia (Squamata: Gymnophthalmidae) and evaluated the effects of sex, host size, and seasonality on endoparasite abundance in two areas of Caatinga, northeast Brazil. We collected 110 lizards between May 2013 to June 2014 and found 173 nematodes (overall prevalence: 16.3%), with 49 nematodes infecting seven adult males (prevalence: 25%), 84 nematodes infecting six adult females (Prevalence: 23%), and 40 nematodes infecting five juveniles (prevalence: 8.9%), where one nematode was in the lungs and 172 were in the gastrointestinal tracts. We identified all nematodes as Oswaldocruzia brasiliensis, representing a new record for the host and for Gymnophthalmidade, showing overall intensity of infection ± SD of 9.6 ± 5.2. Furthermore, abundance of endoparasites was related to the rainy season and sex, but not to host body size (SVL). Keywords. Parasites, Lizards, Neotropical, Oswaldocruzia brasiliensis, Nematodes. In the last decade, there has been an increasing number of studies about endoparasites affecting reptiles. The last survey in South America recorded about 155 helminth species infecting lizards and amphisbaenians (Ávila and Silva, 2010). Although the number of these studies in Neotropical regions is still increasing, some lizard families are more deeply studied, such as Leiosauridae (Vrcibradic et al., 2008; Barreto-Lima and Anjos, 2014; Dorigo et al., 2014); Phyllodactylidae (Ávila et al., 2010a; Sousa et al., 2010); Teiidae (Brito et al., 2014b), and Tropiduridae (Almeida et al., 2009; Anjos et al., 2012; Pereira et al., 2012). In contrast, studies about Gymnophthalmidae are scarce and mainly regard new host records (Ávila and Silva, 2010). Currently, parasitological studies are only available for ten Gymnophthalmidae species in Neotropical regions: Cercosaura ISSN (print) ISSN (online) eigenmanni and C. oshaughnessyi (Bursey and Goldberg, 2004); Alopoglossus angulatus (Goldberg et al., 2007b); Leposoma osvaldoi and Neusticurus ecpleopus (Goldberg et al., 2007a); Iphisa elegans elegans (Ávila et al., 2010b); Bachia scolecoides and Cercosaura ocellata ocellata (Ávila and Silva, 2011); and Micrablepharus maximiliani (Brito et al., 2014a). Gymnophthalmid lizards, popularly known as microteiids, are small (40 to 150 mm snout-vent length) and commonly distributed in Neotropical regions from South Mexico to Argentina, including some islands of Central and South America, totalling 220 species in 48 genera (Presch, 1980). In Brazil, there are about 93 gymnophthalmid species, distributed in 33 genera. One of them is Anotosaura, which comprises two species, A. vanzolinia and A. collaris (Costa and Bérnils, 2015). Firenze University Press

106 104 B.H.S. Oliveira et alii Fig. 1. Collecting sites in Paraíba State (A) in Northeast Brazil, Campina Grande municipality (B). Study areas (C): São José da Mata district (SJM) and Complexo Aluízio Campos Forest Park (CAC). Anotosaura vanzolinia is a small lizard restricted to semi-arid regions in northeastern Brazil, locally known as Caatinga, with fossorial habits, occurring in sites with accumulated leaf litter, its diet is comprised basically of microarthropods associated with sites where these lizards live (Oliveira and Pessanha, 2013). However, there are no reports about parasites of A. vanzolinia, possibly because A. vanzolinia is a rare lizard in Caatinga. Therefore, this study aims to evaluate the effect of sex, host size, and seasonality on the composition and abundance of helminths associated with the lizard Anotosaura vanzolinia, using a considerable sample from Caatinga, northeast of Brazil. We collected Anotosaura vanzolinia specimens during expeditions in the Complexo Aluízio Campos Forest Park ( S, W), a Caatinga area with an altitude of approximately 500 m, with shrubby vegetation represented mainly by Bromeliaceae and Cactaceae, a larger number of rock outcrops, and accumulated leaf litter (Alves et al., 2010; Silva et al., 2010) and in the São José da Mata district ( S, W), located between the arboreal formations locally known as Brejo and Caatinga, with an altitude of approximately 700 m. The site is settled between the Agreste and Sertão, being probably one of the last remains of the transitional arboreal vegetation of Paraiba, with typical plant species from Caatinga and the Atlantic Forest (Barbosa et al., 2007). Both sites are located in Paraiba, northeastern Brazil (Fig. 1). Climate is tropical, with a mean temperature of 22.9ºC; highest precipitation occurs between March and August (99 mm) and lowest between September and February (29 mm) (Climate-Date, 2016). We collected 110 lizards between May 2013 to June 2014 (Table 1) by hand and using pitfall-traps (six sets in each area), constructed with four buckets (20 L each), totalling 24 buckets per area, arranged in three lines of three meters each, forming angles of 120 from a same central point and connected by a plastic fence fixed with staples on wooden stakes.

107 Nematodes infecting the lizard A. vanzolinia from Brazil 105 Table 1. Number of lizards collected in each sampled area, months of the year and season. CAC = Complexo Aluízio Campos Forest Park; SJM = São José da Mata District. Months CAC Locality SJM Wet season Dry season Wet season Dry season May 8 1 Jun 9 1 Jul 4 0 Aug 8 2 Sep 7 2 Oct 9 6 Nov 4 2 Dec 7 1 Jan 16 1 Feb 7 1 Mar 6 0 Apr 7 1 Table 2. Parasitological data from population of Anotosaura vanzolinia from Caatinga, Northeast Brazil, infected by Oswaldocruzia brasiliensis. P% = prevalence; SD = standard deviation; I = intestine, L = lung. Host sex and maturity Host sample size SVL/SD (mm) P% (Host infected) Mean intensity of infection/ SD Site infection Males ± 2 25 (7) 7 ± 3.8 I Females ± (6) 14 ± 8.5 I,L Juveniles ± (5) 8 ± 3.4 I Total (18) 9.6 ± We killed the lizards with a lethal injection of 2% lidocaine hydrochloridec and measured snout-vent length (SVL) with digital calipers. Subsequently, we sexed them, preserved them in 10% formalin, and stored them in 70% alcohol. All lizards were kept in the Coleção Herpetológica da Universidade Federal da Paraíba (CHUFPB). In the laboratory, we removed the respiratory and gastrointestinal tracts and analyzed them in a stereomicroscope to search for endoparasites. The endoparasites found were cleared with Hoyer s solution, counted, identified, stored in 70% alcohol, and kept in the Coleção de Invertebrados Paulo Young, in Universidade Federal da Paraíba (UFPB- NEM: 0001; 0002). The following infection rates were calculated according to Bush et al.(1997), where parasite abundance is defined as the total number of parasites found in a sample (individual host, host population and / or host community); mean intensity of infection is the total number of parasites found in a sample, divided by the number of hosts infected with that parasite; prevalence (P%) is the number of host infected with one or more individuals of a particular parasite species divided by total host number. Throughout the text, means appear ± 1 SD. To analyze the influence of SVL on endoparasite abundance, we performed simple linear regression (excluding juveniles), using the Statistica Software, version 8.0 (Statsoft, 2007). In addition, we performed a generalized linear model (GLM), adopting Poisson s distribution, using the Software R, package R commander (R core team, 2008) to evaluate whether parasite abundance was influenced by host sex, seasonality and interaction between sex and seasonality. We examined 110 Anotosaura vanzolinia specimens, of which 26 were adult females (SVL: 42.6 ± 2.3), 28 adult males (SVL: 37.9 ± 2), and 56 juveniles (SVL: 26.7 ± 5.2). We found 173 nematodes (overall prevalence: 16.3%), with 49 nematodes infecting seven adult males (prevalence: 25%), 84 nematodes infecting six adult females (prevalence: 23%), and 40 nematodes infecting five juveniles (Prevalence: 8.9%), where one nematode was in the lungs and 172 were in the gastrointestinal tracts. We identified all nematodes as Oswaldocruzia brasiliensis, representing a new host record and the first record for Gymnophthalmidade, showing overall intensity of infection of 9.6 ± 5.2 (Table 2). Simple linear regression revealed a non-significant relationship between SVL and endoparasite abundance (F 1,12 = 4.24; R 2 = 0.26; P = 0.061). We observed a significant variation between the abundance of endoparasites and sex of the host (GL = 3; Z value = ; P = ), where females had more parasites than males (Fig. 2). In addition, abundance of endoparasites was higher in the rainy season (GL = 3; Z value = ; P <0.0001) (Fig. 2). However, abundance of endoparasites was not influenced by interaction sex-season (GL= 4; Z value = 0.024; P = 0.981) (Fig. 2). Several factors can influence parasitism by helminths in reptiles (Aho, 1990), especially host size (Poulin and George Nascimento, 2007), sex (Galdino et al., 2014), season (Brito et al., 2014), and life cycle of the parasite (Araujo-Filho et al., 2016). Usually, larger hosts (mass and body size) have the capacity to provide shelter to a greater number of parasites, consequently, more resources for parasite development and reproduction (George Nascimento et al., 2004). However, in the present study, we did not observe a significant relation between SVL and parasite abundance in A. vanzolinia. For lizards, this hypothesis is supported mainly by Tropiduridade: Tropi-

108 106 B.H.S. Oliveira et alii Fig. 2. Mean abundance of nematodes between seasons for species of lizard A. vanzolinia, males (triangle) and female (circle). durus torquatus (range of intensity: 22.1 ± 20.2) (Ribas et al., 1998); T. hispidus (range of intensity: 8.5 ± 1) (Anjos et al., 2012); T. torquatus (range of intensity: 8.2 ± 6.9) (Pereira et al., 2012); and Teiidae species: Ameiva ameiva (range of intensity: 7.2 ± 7.3) (Ribas et al., 1998); A. festiva (intensity of infestation: 21, range: 1-115) (Ramírez- Morales et al., 2012). Nevertheless, besides body size and host mass, other factors can explain the variation in parasite abundance, such as ecology (Aho, 1990), physiology (Poulin and Mouillot, 2004; Poulin, 2007), behaviour, and phylogeny (Poulin and Mouillot, 2005; Patterson et al., 2008; Brito et al., 2014a). Furthermore, the small size of A. vanzolinia (SVL = 42 mm) can constrain the diversification of its parasitic fauna, hindering niche differentiation and microhabitat segregation by endoparasite competitors (Kuris et al., 1980; Ávila et al., 2010a) thus explaining the presence of only a single nematode species infecting the host studied. This observation was also recorded in other small lizard species, such as Liolaemus lutzae (SVL = 61 mm), Aspronema dorsivittatum (SVL = 64 mm), and Phyllopezus lutzae (SVL = 42 mm), which exhibit up to two endoparasites species (Rocha, 1995; Rocha et al., 2003; Ávila et al., 2010a). We found significant differences in the abundance of nematodes in relation to host sex, where females were more parasitized than males. According to Poulin (1996), physiological, morphological, and behavioral differences between males and females may explain the differences of infection rates between sexes. However, due to the absence of research that verified possible differences in the ecology between the sexes for the lizard A. vanzolinia, we can not explain what factors may have influenced the increase of the parasite abundance in females in the present study. The nematode Oswaldocruzia brasiliensis has a monoxenous life cycle (Lent and Freitas, 1935), and congeneric species infect mainly frogs (Well Slimane and Durette- Desset, 1996), lizards (Bursey et al., 2006), and snakes (Durette-Desset et al., 2006). According to Anderson (2000), monoxenous nematodes can easily be influenced by variations in temperature and humidity. Our results support this hypothesis, considering that A. vanzolinia presents a greater endoparasite abundance in the rainy season, mainly because of its fossorial habits, increasing the possibility of infection by O. brasiliensis through accidental ingestion of eggs (in the faeces of other hosts, substrate in general) or by active penetration of infective larvae (via the skin and / or mucosa).this can also explain the higher parasite abundance in females, because they lay eggs during the rainy season and in humid microhabitats to avoid egg desiccation (Oliveira and Pessanha, 2013). The present study revealed that the parasitic fauna of A. vanzolinia from Caatinga consists of a single nematode, Oswaldocruzia brasiliensis, representing the first record of infection to the host and to Gymnophthalmidae. Furthermore, endoparasite abundance is related to the rainy season and sex, but is not correlated with host body size (SVL). ACKNOWLEDGMENTS We thank the Fundação Universitária de Apoio ao Ensino, à Pesquisa e à Extensão (FURNE) for providing collecting permits at Complexo Aluízio Campos, Mr. Guilherme Leão, who not only granted access to his property in São José da Mata, but also encouraged our survey, and Mr. Vandeberg Ferreira Lima for helping with statistical analyses. AAMT thank Coordenação de Aperfeiçoamento da Pessoa de Nível Superior CAPES for

109 Nematodes infecting the lizard A. vanzolinia from Brazil 107 a research fellowship. DAT, JAAF and DOM thank the Conselho Nacional de Desenvolvimento Científico e Tecnológico CNPq for a research fellowship (303610/2014-0).The study was approved by the Instituto Chico Mendes de Conservação da Biodiversidade with the permit to collect the animals (process ). We thank Proof- Reading-Service.com Ltda, for revision professional of English version of the manuscript. REFERENCES Aho, J.M. (1990): Helminth communities of amphibians and reptiles: comparative approaches to understanding patterns and processes. In: Parasite Communities: Patterns and Processes, p Esch, G.W., Bush, A.O., Aho, J.M., Eds, Dordrecht, Springer Netherlands. Almeida, W.O., Ribeiro, S.C., Santana, G.G., Vieira, W.L.S., Anjos, L.A., Sales, D.L. (2009): Lung infection rates in two sympatric Tropiduridae lizard species by pentastomids and nematodes in northeastern Brazil. Braz. J. Biol. 69: Alves, L.S., Albuquerque, H.N., Barbosa, J.S., Aguiar, C.B. (2010): Ações Socioeducativas e ambientais no Complexo Aluízio Campos. Rev. Bras. de Inf. Cient. 1: Anjos, L.A., Ávila, R.W., Ribeiro, S.C., Almeida, W.O., Silva, R.J. (2012): Gastrointestinal nematodes of the lizard Tropidurus hispidus (Squamata: Tropiduridae) from a semi-arid region of northeastern Brazil. J. Helminthol. 81: Araujo-Filho, J.A., Brito, S.V., Lima, V.F., Pereira, A.M.A., Mesquita, D.O., Albuquerque, R.L., Almeida, W.O. (2016): Influence of temporal variation and host condition on helminth abundance in the lizard Tropidurus hispidus from north-eastern Brazil. J. Helminthol. 26: 1-8. Ávila, R.W., Anjos, L.A., Gonçalves, U., Freire, E.M.X., Almeida, W.O., Silva, R.J. (2010a): Nematode infection in the lizard Bogertia lutzae (Loveridge, 1941) from the Atlantic forest in north-eastern Brazil. J. Helminthol. 84: Ávila, R.W., Silva, R.J. (2010): Checklist of helminths from lizards and amphisbaenians (Reptilia, Squamata) of South America. J. Venom. Anim. Toxins. Incl. Trop. Dis. 16: Ávila, R.W., Silva, R.J. (2011): Helminths of Lizards (Reptilia: Squamata) from Mato Grosso State, Brazil. Comp. Parasitol. 78: Ávila, R.W., Strussmann, C., Silva, R.J. (2010b): A New Species of Cosmocercoides (Nematoda: Cosmocercidae) from a Gymnophthalmid Lizard of Western Brazil. J. Parasitol. 96: Barbosa, A.R., Nishida, A.K., Costa, E.S., Cazé, A.L.R. (2007): Abordagem etnoherpetológica de São José da Mata Paraíba Brasil.Rev. Biol. Ciênc. Terra 7: Barreto-Lima, A.F., Anjos, L.A. (2014): Occurrence of Strongyluris oscari (Nematoda; Heterakidae) in Enyalius bilineatus (Squamata: Leiosaurinae) from the Brazilian Atlantic Forest. Herpetol. Notes 7: Brito, S.V., Corso, G., Almeida, A.M., Ferreira, F.S., Almeida, W.O., Anjos, L.A., Mesquita, D.O., Vasconcellos, A. (2014a): Phylogeny and micro-habitats utilized by lizards determine the composition of their endoparasites in the semiarid Caatinga of Northeast Brazil. Parasitol. Res. 113: Brito, S.V., Ferreira, F.S., Ribeiro, S.C., Anjos, L.A., Almeida, W.O., Mesquita, D.O., Vasconcellos, A. (2014b): Spatial-temporal variation of parasites in Cnemidophorus ocellifer (Teiidae) and Tropidurus hispidus and T. semitaeniatus (Tropiduridae) from Caatinga areas in Northeastern Brazil. Parasitol. Res. 113: Bursey, C.R., Goldberg, S.R. (2004): Cosmocerca vrcibradici n. sp.(ascaridida: Cosmocercidae), Oswaldocruzia vitti n. sp.(strongylida: Molineoidae), and other helminths from Prionodactylus eigenmanni and Prionodactylus oshaughnessyi (Sauria: Gymnophthalmidae) from Brazil and Ecuador. J. Parasitol. 90: Bush, A.O., Lafferty, K.D., Lotz, J.M., Shostak, A.W. (1997) Parasitology meets ecology on its own terms: Margolis et al. revisited. J. Parasitol. 83: Climate-Date (2016): Dados climáticos para cidades mundiais. Available at: Last accessed on January Costa, H.C., Bérnils, R.S. (2015): Répteis brasileiros: Lista de espécies. Versão Available at: sbherpetologia.org.br/, Last accessed on April Dorigo, T.A., Maia-Carneiro, T., Almeida-Gomes, M., Siqueira, C.C., Vrcibradic, D., Van Sluys, M. (2014): Diet and helminths of Enyalius brasiliensis (Lacertilia, Iguania, Leiosauridae) in an Atlantic Rainforest remnant in southeastern Brazil. Braz. J. Biol. 74: Galdino, C.a.B., Ávila, R.W., Bezerra, C.H., Passos, D.C., Melo, G.C., Zanchi, D. (2014): Helminths infection patterns in a lizard (Tropidurus hispidus) Population from a Semiarid Neotropical Area: Associations Between Female Reproductive Allocation and Parasite Loads. J. Parasitol. 100: George Nascimento, M., Munoz, G., Marquet, P.A., Poulin, R. (2004): Testing the energetic equivalence rule with helminth endoparasites of vertebrates. Ecol. Lett. 7:

110 108 B.H.S. Oliveira et alii Goldberg, S.R., Bursey, C.R., Caldwell, J.P., Vitt, L.J., Costa, G.C. (2007a): Gastrointestinal helminths from six species of frogs and three species of lizards, sympatric in Pará State, Brazil. Comp. Parasitol. 74: Goldberg, S.R., Bursey, C.R., Vitt, L.J. (2007b): Parasite communities of two lizard species, Alopoglossus angulatus and Alopoglossus atriventris, from Brazil and Ecuador. Herpetol. J. 17: Kuris, A.M., Blaustein, A.R., Alio, J.J. (1980): Hosts as Islands. Am. Nat. 116: Oliveira, B.H.S., Pessanha, A.L.M. (2013): Microhabitat use and diet of Anotosaura vanzolinia (Squamata: Gymnophthalmidae) in a Caatinga area, Brazil. Biota Neotrop. 13: Patterson, B.D., Dick, C.W., Dittmar, K. (2008): Parasitism by bat flies (Diptera: Streblidae) on neotropical bats: effects of host body size, distribution, and abundance. Parasitol. Res. 103: Pereira, F.B., Sousa, B.M., Lima, S.S. (2012): Helminth Community Structure of Tropidurus torquatus (Squamata: Tropiduridae) in a Rocky Outcrop Area of Minas Gerais State, Southeastern Brazil. J. Parasitol. 98: Poulin, R. (1996): Sexual Inequalities in Helminth Infections: A Cost of Being a Male? Am. Nat. 147: Poulin, R. (2007): Evolutionary Ecology of Parasites, New Jersey, Princeton University Press. Poulin, R., George Nascimento, M. (2007): The scaling of total parasite biomass with host body mass. Int. J. Parasitol. 37: Poulin, R., Mouillot, D. (2004): The relationship between specialization and local abundance: the case of helminth parasites of birds. Oecol. 140: Poulin, R., Mouillot, D. (2005): Combining phylogenetic and ecological information into a new index of host specificity. J. Parasitol. 91: Presch, W. (1980): Evolutionary history of the South- American Microteiid lizards (Teiidae, Gymnophthalminae). Copeia 1980: Ramírez-Morales, R., Lislevand, T., Retana-Salazar, A., Solhøy, T., Roth, S. (2012): Ectoparasite loads of the Central American whiptail lizard Ameiva festiva (Squamata: Teiidae). Herpetol. J. 22: Ribas, S.C., Rocha, C.F.D., Teixeira-Filho, P.F., Vicente, J.J. (1998): Nematode infection in two sympatric lizards (Tropidurus torquatus and Ameiva ameiva) with different foraging tactics. Amphibia-Reptilia 19: Rocha, C.F.D. (1995): Nematode parasites of the brazilian sand lizard, Liolaemus lutzae. Amphibia-Reptilia 16: Rocha, C.F.D., Vrcibradic, D., Vicente, J.J., Cunha-Barros, M. (2003): Helminths infecting Mabuya dorsivittata (Lacertilia, Scincidae) from a high-altitude habitat in Itatiaia National Park, Rio de Janeiro state, Southeastern Brazil. Braz. J. Biol. 63: Silva, M.A., Barbosa, J.S., Albuquerque, H.N. (2010): Levantamento florístico das plantas espontâneas e suas potencialidades fitoterapêuticas: um estudo no Complexo Aluízio Campos - Campina Grande - PB. Rev. Bras. de Inf. Cient. 1: Sousa, J.G.G., Ribeiro, S.C., Roberto, I.J., Teles, D.A., Almeida, W.O. (2010): Ocorrência de pentastomídeos (Metameria: Ecdysozoa) no lagarto Phyllopezus pollicaris (Spix, 1825). Cad. Cult. Cienc. 2: Statsoft, Inc. (2007): Statistica for Windows [data analysis software system], Version 8.0. Tulsa: StatSoft. Vrcibradic, D., Anjos, L.A., Vicente, J.J., Bursey, C.R. (2008): Helminth parasites of two sympatric lizards, Enyalius iheringii and E. perditus (Leiosauridae), from an Atlantic Rainforest area of southeastern Brazil. Acta Parasitol. 53:

111 Acta Herpetologica 12(1): , 2017 DOI: /Acta_Herpetol Where is my place? Quick chorus structure assembly in the European tree frog Michal Berec Faculty of Agriculture, University of South Bohemia, Studentská 13, České Budějovice, Czech Republic. seznam.cz Submitted on 2016, 4 th October; revised on 2016, 24 th November; accepted on 2017, 3 rd April Editor: Giovanni Scillitani Abstract. Lek mating systems are characteristic of anurans that use oviposition sites that cannot be easily monopolized by individual males. The dynamics of the chorus structure in leks is not well known. Here, we examine the relationship between the movement activity of individual males during the breeding season and their size. According to our observations, the site fidelity of males is not completely random, with the larger males moving significantly shorter distances than smaller males. However, this difference applies only to the distance between the first and second captures. Whether higher site fidelity contributes to higher mating success needs further investigation. Keywords. Vocalization, chorus, movement, size, tree frog. Individuals of the majority of amphibian species live their own solitary lives throughout the year except during the reproductive period, when they gather in defined spaces to find appropriate sexual partners. Males of some of these species do not occupy territories during the reproductive period but rely on physical characteristics during short and intensive bouts (scramble competition) to overpower mating competitors (Duellman and Trueb, 1994). Males of the other species divide the space into defended segments and lure individual females by acoustic or visual stimuli (Wells, 2007). In lek mating systems, males typically occupy and defend small areas in which they vocalize (Höglund and Alatalo, 1995). Whenever intruders of the same sex approach too closely, they fight against them. Females choose their mates by moving both within and between several aggregations of males (Höglund and Alatalo, 1995). Surprisingly, there are only a handful of studies describing lek structure and dynamics in anurans (Emlen, 1976, Tárano, 2009), although many species of frogs and toads form typical lek aggregations (Wells, 2007). This lack of research is surprising, as the positions of individual males within the lek can strongly influence their mating success (Arita and Kaneshiro, 1985; Kokko et al. 1998; Howard et al. 2011; but see Sæther et al for the opposite results). At the same time, the size of individual males is the most important determinant of attractiveness in many anurans (Wells, 2007 and references therein). Here, we report the results of our survey of movement dynamics within a lek system of European tree frog (Hyla arborea). We specifically focused on the relationship between the movement of males during the reproductive season and their size (as a proxy for age). Our goal was to examine whether leks in this tree frog species show spatiotemporal structure. Despite its possible relevance to complex frog reproductive behaviour, this subject has not previously been analysed in detail. We investigated a population of European tree frog (Hyla arborea) in Velký Vávrovský pond (1.5 ha, 390 m a.s.l., average depth 40 cm, maximum depth 120 cm, N, E) at the outskirts of České ISSN (print) ISSN (online) Firenze University Press

112 110 M. Berec Budějovice in South Bohemia, Czech Republic. This pond is used for carp production and is also a breeding site used by other anurans (Bombina bombina, Bufo bufo, Rana dalmatina, Pelophylax kl. esculentus). This pond is nearly square in shape. Three of its four sides are overgrown by willow shrubs (Salix sp.), and the fourth side is paved. The littoral vegetation consists of sparse clumps (less than 3% of water area) of Carex sp., Typha angustifolia, and Alisma plantago-aquatica. The frogs were therefore easily observed and caught. Tree frog males were collected during the breeding season, beginning at the onset of calling activity every day from April 13 to June 17, 2005, excluding the days with unsuitable conditions (pouring rain, storms). Each night, all males in the chorus were captured and measured. As all males called within four metres from the shore, we walked slowly around the pond and entered the water only to catch males to reduce disturbance to the chorus. All males were caught by hand and individually marked by toe-clipping, and the snout-urostyle length (SUL) was measured to the nearest millimetre using a vernier calliper. Only one digit per limb was toeclipped, with a maximum of three digits per individual. For toe-clipping, we used scissors disinfected in alcohol. No analgesic was used. We did not observe any signs of pain or distress in the frogs when handled. Additionally, we did not observe any negative effect on survival (we have used this method repeatedly in previous studies). Some males resumed calling a few minutes after toe clipping (see Funk et al., 2005 for comments on the toeclipping method). The position of each male was determined using coloured tags with numerical codes positioned along the breeding-site shoreline at 2 m intervals. When an individual frog was recaptured on a subsequent night or nights, we measured its position to the nearest centimetre from the previous position, using a measuring tape in cases of distances within a radius of five metres. For greater distances, we measured the distance covered by using the marks on the map and rounded the value to the nearest ten centimetres. We did not use GPS or similar system because the measurement error with such devices is at least as large as that with our approach. All statistical tests were performed using the software package Statistica 10.0 (StatSoft, 2012). All distances were log-transformed before analysis. In total, 188 males were captured and individually marked. During the season, 44% of males were captured only once, 22% were captured twice, 15% were captured three times, 13% were captured four times, 3% were captured five times, and 3% were captured six times. The number of captures of the same individual was not dependent on the SUL of males (Kruskal-Wallis ANO- VA by ranks: H (5, n = 188) = 0.581, P = 0.988). Meanwhile, the SUL of males captured did not depend on the day of the season (R = 0.082; R 2 = 0.007; F 1,186 = 1.266; P = 0.262). In contrast, the SUL of tree frog males partially determined the movement pattern during the breeding season. Our analysis revealed significant relationships between SUL and distances between subsequent captures (R = ; R 2 = 0.146; F 1,222 = ; P < 0.001; Fig. 1). However, only the distance between the 1 st and 2 nd capture was clearly associated with the above mentioned relationship. Here, smaller males covered longer distances between the first and the second capture than did Fig. 1. Relationship between the SUL of males and the average distance between all pairs of subsequent captures. Fig. 2. Relationship between the SUL of males and the distance between the first and second captures.

113 Movement dynamics of tree frogs 111 Table 1. Mean distances ± standard deviation between two subsequent captures, as well as results of regression analyses between the SUL of males and distances between subsequent captures. Capture Mean distance ± SD (cm) Range (cm) R R 2 F d.f. p 1 st -2 nd 801 ± ,104 < nd -3 rd 614 ± , rd -4 th 541 ± , th -5 th 414 ± , larger males (Fig. 2). None of the subsequent movements showed a significant association with SUL (Table 1). Although the distance covered by males steadily decreased between subsequent captures (Table I), this effect was not statistically significant (Kruskal-Wallis ANOVA by ranks: H (4 n = 171) = 0.127, P = 0.998). In general, larger males usually spend more nights in a lek, which is related to their mating success in lekbreeding anurans (Pröhl, 2003; Friedl and Klump, 2005; Castellano et al. 2009; Botto and Castellano, 2016). Contrary to this, our data did not support this pattern, as the number of captures (i.e., nights when the male attended the lek) was not dependent on the SUL of males. Here, the data suggest another scenario, in which males of different sizes spend similar amounts of time (numbers of nights) in a lek. However, the size of males determined their movement behaviour. Male size is generally the most important determinant of mating success in explosive competitors (Holliday and Tejedo, 1995), whereas the situation is much more complex in lek-breeding anurans (Höglund and Alatalo, 1995). Studies of lekking non-anuran vertebrates suggest that individual position in the lek can be very important (Kokko et al., 1998; Howard et al., 2011). Surprisingly, despite the large number of lek-breeding anurans, hardly any studies linking mating success of males and their individual positions within a lek have been published. As many lek-breeding anurans establish nonrandom spacing patterns during reproduction events (Whitney and Krebs, 1975; Brenowitz et al., 1984; Tárano, 2009), different outcomes from different positions in the lek could be predicted. The lek is re-established every night in lek-breeding anurans, including European tree frog (Grosse, 2009). According to our data, the site fidelity of males is not completely random and is determined by their size during part of the breeding season. In fact, the larger males (i.e., older sensu Kyriakopoulou- Sklavounou and Grumiro, 2002) moved significantly shorter distances than smaller (younger) males. However, this difference applies only to the distance between the first and second captures. After that, no differences were observed in movement activity between subsequent captures. Emlen (1976) f ound a different pattern of chorus formation in bullfrogs (Lithobates catesbeianus). He described leks in bullfrogs as both spatially and temporally ephemeral but subsequently found that larger (older) males occupied more centrally located territories. On the other hand, the locations of individual males changed rapidly, and they moved from one aggregation to another during the breeding season. The fact that larger males occupied much stable area implies question of the advantage of this behaviour. This can be the result of their previous experience as larger males are presumed to be older ones (Friedl and Klump, 1997). Thus, larger males have better knowledge about the locality from previous reproductive seasons and should improve their position according to their previous mating success. Simultaneously, experienced males can also recognize better mating positions (e.g., determined by abundance and structure of vegetation cover, distance from the shore, or oviposition site, etc.) in shorter time than smaller ones. Moreover, physical superiority of larger males enables the territories to be hold for longer period. More intensive movement activity of smaller males can therefore be only the consequence of their effort to find better mating opportunities. Site fidelity has been documented in several anuran species. Various authors explain this behaviour as advantageous for male-male competition (Davis, 1987; Brenowitz and Rose, 1994; Marshall et al., 2003), search strategies of females (Fellers, 1979; Murphy and Gerhardt, 2002), female attraction or choice (Bradbury and Gibson, 1983; Grafe, 1997), female paths (Grafe, 1997), or sound transmission (Narins and Hurley, 1982; Wells and Schwartz, 1982; Marshall et al., 2003). Using dynamic programming, the theoretical model by Switzer (1993) of sequential settlement decisions predicted site fidelity to be positively correlated to age. Unfortunately, previous studies did not relate the observed behaviour directly to the size of males. The reason why larger tree frog males chose and returned to the same positions more precisely than smaller males is not obvious at this time and deserves further investigation.

114 112 M. Berec ACKNOWLEDGEMENTS Permission for conducting the study was issued by the Ministry of Environment of the Czech Republic (permission number 1211/32/SOP/E/05/456/Cho). I would like to thank Irena Šetlíková, Luděk Berec, and two anonymous reviewers for valuable comments on the earlier versions of the manuscript. Research was financially supported by the GAJU 081/2016/Z. REFERENCES Botto, V., Castellano, S. (2016): Attendance, but not performance, predicts good genes in a lek-breeding treefrog. Behav. Ecol. 27: Bradbury, J.W., Gibson, R.M. (1983): Leks and mate choice. In: Mate Choice, pp Bateson, P., Ed, Cambridge University Press, Cambridge. Brenowitz, E.A., Rose, G.J. (1994): Behavioural plasticity mediates aggression in choruses of the Pacific treefrog. Anim. Behav. 47: Brenowitz, E.A., Wilczynski, W., Zakon, H.H. (1984): Acoustic communication in spring peepers. J. Comp. Physiol. A 155: Castellano, S., Zanollo, V., Marconi, V. Berto, G. (2009): The mechanisms of sexual selection in a lek-breeding anuran, Hyla intermedia. Anim. Behav. 77: Davis, M.S. (1987): Acoustically mediated neighbor recognition in the North American bullfrog, Rana catesbeiana. Behav. Ecol. Sociobiol. 21: Duellman, W.E., Trueb, L. (1994): Biology of amphibians. John Hopkins University Press. Emlen, S.T. (1976): Lek organization and mating strategies in the bullfrog. Behav. Ecol. Sociobiol. 1: Fellers, G.M. (1979): Aggression, territoriality, and mating behaviour in North American treefrogs. Anim. Behav. 27: Friedl, T.W., Klump, G.M. (1997): Some aspects of population biology in the European treefrog, Hyla arborea. Herpetologica 53: Friedl, T.W., Klump, G.M. (2005): Sexual selection in the lek-breeding European treefrog: body size, chorus attendance, random mating and good genes. Anim. Behav. 70: Funk, W.C., Donnelly, M.A., Lips, K.R. (2005): Alternative views of amphibian toe-clipping. Nature 433: Grafe, T.U. (1997): Costs and benefits of mate choice in the lek-breeding reed frog, Hyperolius marmoratus. Anim. Behav. 53: Grosse, W.R. (2009): Der Laubfrösche. Edition Chimaira, Westarp Wissenschaften. Halliday, T., Tejedo, M. (1995): Intrasexual selection and alternative mating behaviour. In: Amphibian biology 2, pp Heatwole H., Sullivan, B. K., Eds, Surrey Beatty and Sons, Chipping Norton, NSW. Höglund, J., Alatalo, R.V. (1995): Leks. Princeton University Press, Princeton (NJ). Howard, D.R., Lee, N., Hall, C.L., Mason, A.C. (2011): Are centrally displaying males always the centre of female attention? Acoustic display position and female choice in a lek mating subterranean insect. Ethology 117: Kokko, H., Lindström, J., Alatalo, R.V., Rintamäki, P.T. (1998): Queuing for territory positions in the lekking black grouse (Tetrao tetrix). Behav. Ecol. 9: Kyriakopoulou-Sklavounou, P., Grumiro, I. (2002): Body size and age assessment among breeding populations of the tree frog Hyla arborea in northern Greece. Amphibia-Reptilia 23: Marshall, V.T., Humfeld, S.C., Bee, M.A. (2003): Plasticity of aggressive signalling and its evolution in male spring peepers, Pseudacris crucifer. Anim. Behav. 65: Murphy, C.G., Gerhardt, H.C. (2002): Mate sampling by female barking treefrogs (Hyla gratiosa). Behav. Ecol. 13: Narins, P.M., Hurley, D.D. (1982): The relationship between call intensity and function in the Puerto Rican coqui (Anura: Leptodactylidae). Herpetologica 38: Pröhl, H. (2003): Variation in male calling behaviour and relation to male mating success in the strawberry poison frog (Dendrobates pumilio). Ethology 109: Sæther, S.A., Baglo, R., Fiske, P., Ekblom, R., Höglund, J., Kålås, J.A. (2005): Direct and indirect mate choice on leks. Am. Nat. 166: StatSoft Inc. (2012): STATISTICA (data analysis software system), version 9.1 (Version 9.1). Available from Switzer, P.V. (1993): Site fidelity in predictable and unpredictable habitats. Evol. Ecol. 7: Tárano, Z. (2009): Structure of transient vocal assemblages of Physalaemus fischeri (Anura, Leiuperidae): Calling site fidelity and spatial distribution of males. S. Am. J. Herp. 4: Wells, K.D. (2007): The Ecology and Behavior of Amphibians. The University of Chicago Press, Chicago. Wells, K.D., Schwartz J.J. (1982): The effect of vegetation on the propagation of calls in the Neotropical frog Centrolenella fleischmanni. Herpetologica 38: Whitney, C.L., Krebs, J.R. (1975): Spacing and calling in Pacific tree frogs, Hyla regilla. Can. J. Zool. 53:

115 Acta Herpetologica 12(1): , 2017 DOI: /Acta_Herpetol A possible mutualistic interaction between vertebrates: frogs use water buffaloes as a foraging place Piotr Zduniak 1, *, Kiraz Erciyas-Yavuz 2, Piotr Tryjanowski 3 1 Department of Avian Biology and Ecology, Faculty of Biology, Adam Mickiewicz University in Poznań, Umultowska 89, Poznań, Poland. *Corresponding author. kudlaty@amu.edu.pl 2 Ornithology Research Center, Ondokuz Mayis University, Kurupelit Samsun, Turkey 3 Institute of Zoology, Poznań University of Life Sciences, Wojska Polskiego 71C, Poznań, Poland Submitted on 2017, 23 rd March; revised on 2017, 19 th April; accepted on 2017, 20 th April Editor: Marco Mangiacotti Abstract. Mutualisms shape biodiversity by influencing the ecology and the evolution of populations and communities. For example, among many others, birds commonly forage in association with large mammals, including livestock, but so far no similar relationship has been described for amphibians. In this note we describe the association between the Marsh Frog (Pelophylax ridibundus) and the Anatolian Water Buffalo (Bubalus bubalis) in Turkey and provide possible explanations for the existence of direct relations between these representatives of two vertebrate classes. We hope that our note stimulates future research on this subject. Keywords. Bubalus bubalis, interaction, Pelophylax ridibundus. Interspecific interactions such as mutualism are main processes that shape biodiversity by influencing the ecology and the evolution of populations and communities (Bascompte, 2009; Sazima et al., 2010). Among many others, birds commonly forage in association with large mammals, including livestock, wild ungulates and pachyderms (Dean and MacDonald, 1981; Sazima, 2011). In particular, associations with cattle and buffaloes are one of the best known and occur in many geographical regions (Bradshaw and White, 2006; Sazima, 2011). However, searching through different kinds of scientific information sources, we did not find any records on similar relationships between amphibians and large mammals. Thus, the purpose of this note is to describe the association of the Marsh Frogs (Pelophylax ridibundus) and the Anatolian Water Buffaloes (Bubalus bubalis) in Turkey and to provide a possible explanation for the existence of direct relations between these representatives of two vertebrate classes. The observations were carried out in the Kızılırmak delta (N Turkey) that stretches along the Black Sea coast from 41 30'N and 41 45'N to 35 43'E and 36 08'E. The delta is one of the largest wetlands in the Middle East and an important area for migratory birds (Erciyas-Yavuz et al., 2015). The delta is located at an altitude between 0 to 15 m above sea level and its total surface area is about 56,000 ha. While 70% of the delta is intensively used by people, the remainder is a natural habitat including open water, freshwater and semi-saline lakes, marsh vegetation, sand dunes, woodland, and farmland, including pastures for the local buffalo race (Barış et al., 2005). Anatolian Water Buffaloes play a major role in structuring the vegetation in the Kızılırmak delta, including its permanent and ephemeral wetlands and adjacent coastal sand dune systems. They stay outside during April-November and the rest of the year they are kept in farms (Sullivan et al., 2016). To confirm the rarity and novelty of our observations, we searched information on possible relationships ISSN (print) ISSN (online) Firenze University Press

116 114 P. Zduniak, K. Erciyas-Yavuz, P. Tryjanowski between frogs and buffaloes, domestic cattle and similar large mammals using the internet-based search engines of Thomson-Reuters (Web of Science, Zoological Record) and Scopus databases, Google Scholar and Google Books. Additionally, we searched for amphibian and mammal (generally, and specifically frog and buffalo) records in previously undescribed sources including internet web searches for trip reports, images and videos carried by Google, Google images, Flickr and YouTube. We searched not only the English and Latin but also included numerous other languages. Browsing internet graphic sources and searching for the sitting frogs on buffaloes, four photographs were found. Three were from south Asia and one from Hungary. In all cases these were single buffaloes with the head protruding out of the water, where frogs were present but no insects were observed. No data about the described phenomenon were found in other sources including scientific on-line databases. Real data in the field were also collected. The random and independent 12 observations of Anatolian Water Buffalo took place during October 3 rd and 10 th We recorded on 10 occasions the situation, when Marsh Frogs were present on buffaloes (Tab. 1). Frogs were recorded both on resting and standing mammals in different part of their body including the head. In most cases frogs hunted flies (Fig. 1). Further random observations carried out in the next year confirmed that the described phenomenon only occurs in autumn, when the density of frogs is much higher than during spring (pers. obs.). Our observations indicate an association between frogs and buffaloes and may have a biological meaning. The observed behaviour was not incidental or loosely structured. Frogs foraged on buffaloes in a similar manner as birds on large mammals (Heatwole, 1965; Dean and MacDonald, 1981; Yosef and Yosef, 1991; Sazima, 2011). The food habits of Marsh Frogs are generalist and the species may change its diet in response to local variation in frequency of available prey items, mainly insects, but sometimes also fish, amphibians, and small mammals (Çiçek and Mermer, 2007; Mollov et al., 2010). Because the diet of the species also includes fly species, and many of them are parasites or main disease vectors for large mammals, including buffaloes (Bengis et al., 2002; Altintas, 2004), the recorded association between frogs and buffaloes can be considered as a possible mutualistic interaction (Bascompte, 2009; Sazima et al., 2010). Mutualistic interactions between large mammals and birds may originate and intensify rapidly under specific local conditions (Bradshaw and White, 2006). Such situations probably also occurred in the Kızılırmak delta, where frogs were present on buffaloes only in autumn, when the density of amphibians is high compared to spring. However, because we did not study the diet of the frogs, we cannot prove that water buffaloes have tangible benefit from the presence of the frogs. Hence, the observed relationship can be also interpreted as commensalism (Dickman, 1992), where only frogs benefit from the buffaloes without affecting them. The reasons why the associations between frogs and free ranging water buffaloes were not previously reported are unclear. This could be due to frog and buffalo density in one specific area, but also other local factors like density of insects or habitat, like small water bodies constructed by buffaloes during body washing, grazing and excrements (Jansen and Healey, 2003; Hartel and von Wehrden, 2013; Musitelli et al., 2016). Photographs from south Asia and Europe with frogs resting (not foraging) on buffaloes suggest that the occurrence of frogs on buffaloes may be occasional and accidental and the probability of such observations may depend on frog and buffalo densities. However, our observations where in most cases frogs hunt flies present on buffaloes confirm the exist- Table 1. Basic data about the following observations of frogs on buffaloes carried out in the Kızılırmak delta in October No. observation Date Air temperature (C ) No. buffaloes No. buffaloes with frogs No. frogs on buffaloes Mean no. frogs per buffalo Overall mean ± SD 19.5 ± ± ± ± ± 8.5

117 Interaction between frogs and buffaloes 115 Fig. 1. Photographs of the interaction between frogs and buffaloes. Clock-wise: (1) general view of the habitat in the Kızılırmak delta; (2) foraging and resting buffaloes; (3) sitting buffaloes with many frogs on the fur; and (4) foraging frogs and flies on buffalo fur. ence of an interspecific interaction between amphibians and large mammals. An additional explanation for the observed phenomenon could be the use of buffaloes by frogs as an efficient heat source, which can be important for heterothermic amphibians especially at low ambient temperatures (Sinsch, 1984). To the best of our knowledge, these are the first observations of an interaction between Marsh Frogs and Water Buffaloes, or even more generally between amphibians and large mammals. However, our data provide the first evidence of such associations. We hope that our note will stimulate further research on this subject. ACKNOWLEDGMENTS We thank M. Kaczmarski, A.P. Møller, Kerim Çiçek, Cemal Özsemir and Nizamettin Yavuz for their help in the field, comments and literature search. We also thank two reviewers for the detailed comments on the manuscript. More pictures documenting frogs behaviour, as well as results of internet search can be send on request. REFERENCES Altintas, N. (2004): Parasitic zoonotic diseases in Turkey. Vet. Ital. 44: Barış, S., Erciyas, K., Gürsoy, A., Özsemir, C., Nowakowski, J.K. (2005): Cernek - a new bird ringing station in Turkey. Ring 27: Bascompte, J. (2009): Disentangling the web of life. Science 325: Bengis, R.G., Kock, R.A., Fischer, J. (2002): Infectious animal diseases: the wildlife/livestock interface. Rev. Sci. Tech. O.I.E. 21: Bradshaw, C.J.A., White, W.W. (2006). Rapid development of cleaning behaviour by Torresian crows Corvus orru on non-native banteng Bos javanicus in northern Australia. J. Avian Biol. 37: Çiçek, K., Mermer, A. (2007): Food composition of the marsh frog, Rana ridibunda Pallas, 1771, in Thrace. Turk. J. Zool. 31: Dean, W.R.J., MacDonald, I.A.W. (1981): A review of African birds feeding in association with mammals. Ostrich 52:

118 116 P. Zduniak, K. Erciyas-Yavuz, P. Tryjanowski Dickman, C.R. (1992): Commensal and mutualistic interactions between terrestrial vertebrates. Trends Ecol. Evol. 7: Erciyas-Yavuz, K., Zduniak, P., Barış Y.S. (2015): Spring and autumn migration of the red-breasted flycatcher through the Kizilirmak delta, Turkey. Curr. Zool. 61: Hartel, T., von Wehrden, H. (2013): Farmed areas predict the distribution of amphibian ponds in a traditional rural landscape. PLoS ONE 8: e Heatwole, H. (1965): Some aspects of the association of cattle egrets with cattle. Anim. Behav. 13: Jansen, A., Healey, M. (2003): Frog communities and wetland condition: relationships with grazing by domestic livestock along an Australian floodplain river. Biol. Cons. 109: Mollov, I., Boyadzhiev, P., Donev, A. (2010): Trophic role of the Marsh Frog Pelophylax ridibundus (Pallas, 1771) (Amphibia, Anura) in the aquatic ecosystems. Bulg. J. Agri. Sci. 16: Musitelli, F., Romano, A., Møller, A.P., Ambrosini, R. (2016): Effects of livestock farming on birds of rural areas in Europe. Biodivers. Conserv. 25: 615. Sazima, C., Guimarães, P.R., Dos Reis, S.F.,Sazima, I. (2010): What makes a species central in a cleaning mutualism network? Oikos 119: Sazima, I. (2011): Cleaner birds: a worldwide overview. Rev. Bras. Ornitol. 19: Sinsch, U. (1984): Thermal influences on the habitat preference and the diurnal activity in three European Rana species. Oecologia 64: Sullivan, G.T., Ozman-Sullivan, S.K., Lumaret, J.P., Baxter, G., Zalucki M.P., Zeybekoğlu, Ü. (2016): Dung beetles (Coleoptera: Scarabaeidae) utilizing water buffalo dung on the Black Sea coast of Turkey. Turk. J. Zool. 40: Yosef, R., Yosef. D. (1991): Tristram s Grackles groom Nubian Ibex. Wilson Bull. 103:

119 Acta Herpetologica 12(1): , 2017 DOI: /Acta_Herpetol Good vibrations: a novel method for sexing turtles Donald T. McKnight 1,2, *, Hunter J. Howell 3, Ethan C. Hollender 1, Day B. Ligon 1 1 Department of Biology, Missouri State University, Springfield, Missouri, USA. *Corresponding author. donald.mcknight@ my.jcu.edu.au 2 College of Science and Engineering, James Cook University, Townsville, Queensland, Australia 3 Department of Biological Sciences, Towson University, Towson, Maryland, USA Submitted on December 19th, 2016; revised on April 18th, 2017; accepted on April 21th, 2017 Editor: Ernesto Filippi Abstract. The ability to accurately determine the sex of individuals is important for research and conservation efforts. While most species of turtle exhibit secondary sexual dimorphisms that can be used to reliably infer sex, there are some species that are very difficult to sex, and even within many dimorphic species, it is not uncommon to encounter individuals that appear to exhibit both male and female secondary sex characteristics. Therefore, we tested the novel method of using a vibrator to sex turtles by stimulating male turtles to evert their penises. We tested this method on males of four species (three families) with known sexual dimorphisms: spiny softshell turtles (Apalone spinifera; n = 14), western chicken turtles (Deirochelys reticularia miaria; n = 17), Mississippi mud turtles (Kinosternon subrubrum hippocrepis; n = 10), and common musk turtles (Sternotherus odoratus; n = 9). The method accurately sexed 100% of A. spinifera, 64.7% of D. r. miaria, 80.0% of K. s. hippocrepis, and 55.6% of S. odoratus. Despite the low success rates in some species, there are situations in which this method will be useful for researchers working with species that are difficult to sex using external morphological characteristics. Keywords. Apalone, chelonia, Deirochelys, Kinosternon, penis, Sternotherus, vibrator. The ability to accurately differentiate males and females is important for ecological studies, and for many turtle species this is a relatively simple task. Turtles often exhibit a variety of secondary sexual dimorphisms in traits such as size, color, claw length, plastron shape, and pre-cloacal tail length (Gibbons and Lovich, 1990; Readel et al., 2008). Nevertheless, some species lack obvious dimorphisms, and dimorphisms may vary among populations (Iverson, 1985; Rowe, 1997). Further, even for species that are strongly dimorphic, it is not uncommon to encounter individuals that appear to have some characteristics of males and some characteristics of females, thus making them difficult to sex (McKnight, pers. obs.). Several methods to overcome these problems are available, such as measuring testosterone levels (Owens et al., 1978; Rostal et al., 1994), laparoscopy (Wibbels et al., 1989; Ligon et al., 2009), and cloacoscopy (Ligon et al., 2013); however, these methods are often invasive, time-consuming, and difficult to implement in the field. Recently, two methods have been published for inducing penile erections in male turtles, thus allowing males and females to be differentiated. While penile eversion is a common method for sexing squamates, it has received little attention in turtles. Although this is a promising technique, the methods that have been proposed so far appear to be species-specific and have only been applied to common snapping turtles (Chelydra serpentina), whose penis can be everted by gently bouncing a turtle up and down (De Solla et al., 2001; Dustman, 2013), and Cotinga River toadhead turtles (Phrynops tuberosus), whose penis can be everted by immobilizing the neck and limbs (Rodrigues et al., 2014). ISSN (print) ISSN (online) Firenze University Press

120 118 D.T. McKnight et alii Vibrators may provide a more broadly applicable method of penile eversion. Lefebvre et al. (2013) found that a vibrator could be used to induce ejaculation in male turtles, and ejaculation was preceded by a visible erection. Therefore, this method may be valuable as a means of sexing turtles. We examined its utility on four species of freshwater turtle representing three different families. To test our method of using a vibrator to induce erections in males, we used species that can be sexed using external sexual dimorphisms such as size, color, and tail morphometrics. Thus, we could test the efficiency of the method by seeing how frequently it induced an erection in individuals that were known to be males. The four species that we used were: western chicken turtles (Deirochelys reticularia miaria; family: Emydidae), Mississippi mud turtles (Kinosternon subrubrum hippocrepis; family: Kinosternidae), common musk turtles (Sternotherus odoratus; family: Kinosternidae), and spiny softshell turtles (Apalone spinifera; family: Trionychidae). We captured them using hoop nets placed in ponds in southeastern Oklahoma (detailed trapping methods in McKnight et al., 2015). Once a male turtle was captured, we attempted to induce an erection by applying an 18 cm, variable-speed, silver bullet vibrator to its shell and tail. We vibrated turtles for 10 min or until an erection was achieved, and we recorded the amount of time that it took to induce an erection. Trials were scored as unsuccessful if an erection had not been induced by the end of the 10-minute trial period. Our preliminary trials indicated that turtles needed to be fairly relaxed and willing to extend their limbs and tails before the method would be effective. Therefore, for the sake of time, we limited our trials Fig. 1. A male spiny softshell turtle (Apalone spinifera) being vibrated on the tail. to turtles that were already active at the time of capture. Although this is a drawback of our method, most turtles quickly acclimate to being handled, and a few moments of holding a turtle is generally sufficient for it to extend its limbs (McKnight, pers. obs.). During our trials, we held turtles vertically, with their plastrons facing the researcher that was operating the vibrator. Then, we gently applied the tip of the vibrator to the plastron, carapace, and tail (Fig. 1). We moved it among those regions based on the turtles responses (i.e., if a turtle responded by tightly pulling its limbs and tail against its body, we moved to a different area). For each species, erections generally occurred when the vibrator was placed on the tail itself, but it was often necessary to first vibrate areas other than the tail, because starting with the tail generally resulted in the turtles pulling the tail tightly against the body, rather than extending it. Therefore, we started with the plastron or carapace, and moved to the tail once a turtle had fully extended its tail. In general, turtles appeared to respond best when only the tip of the vibrator was touching them and when the vibrator had fresh batteries and was set on the fastest setting. Also, they seemed to respond best when the tip was held firmly against them (rather than allowing it to bounce), but not be pressed hard against them. Both allowing it to bounce and pressing it too hard generally resulted in turtles holding their limbs and tail tightly against the body, rather than relaxing. Additionally, it was often useful to move the vibrator around in small, slow, steady circles. As a general rule, we tried to hold the vibrator against the tail whenever possible (including following the tail if the turtle is waving it from side to side), but if this caused the turtle to retract its tail, then we moved the vibrator to a different position until the tail was extended again. Finally, sometimes males only protracted their penises briefly and quickly retracted them, rather than maintaining an erection. Therefore, it was necessary to watch the cloaca closely. Although this was the general pattern, each species responded differently, so we had to adapt our protocol based both on the species and individual responses, and it will be necessary to test different positions and techniques when trying this method on a new species. For A. spinifera it was generally not necessary to spend time on parts of the body other than the tail. They usually extended their tails immediately and would allow us to hold the vibrator against their tails. They did frequently wave their tails from side to side, forcing us to move the vibrator with the tail, but they generally did not hold the tail against the body in a stressed position. In contrast, K. s. hippocrepis, S. odoratus, and D. r. miaria usually held their tails against their bodies initial-

121 A novel method for sexing turtles 119 ly and required stimulation to other parts of their body before they would relax and extend their tails. For K. s. hippocrepis and S. odoratus, this generally involved moving the vibrator in slow, small circles on the abdominal and pectoral scutes (the diameter of the circle was only 1-2 cm). Deirochelys r. miaria was similar, but it was usually necessary to vibrate slightly higher (more on the pectoral scutes than abdominal scutes). Also, sometimes they responded to being vibrated on the carapace (usually on the first vertebral scute). Finally, Lefebvre et al. (2013) reported that, when inducing male turtles to ejaculate, vibrating turtles on their heads was often effective, however, that method generally did not work in our study. This further illustrates the differences among species and highlights the need for testing several different locations and methods when vibrating a species for the first time. In addition to the differences in the techniques necessary for stimulating the different species, our success rates also varied among species (Table 1). The method was the most successful for A. spinifera (100%), followed by K. s. hippocrepis (80.0%). It was less successful for D. r. miaria (64.7%) and S. odoratus (55.6%). We compared the success rates among species using a Fisher s exact test, and this showed that there was a statistically significant difference (P = 0.026). The median time required to induce an erection also varied among the species, but it was lowest in A. spinifera and K. s. hippocrepis. Because the utility of this method varied among species, it will need to be validated on a species by species basis. Despite the low success rate in some species, we think that this method has potential to be useful in several situations. First, based on our success employing this technique on A. spinifera and K. s. hippocrepis, it should be useful for some species or populations that are difficult to sex. However, it will first need to validated using individuals of a known sex (such as individuals in a captive population that have been sexed by other methods). If it is successful on those known individuals, then it will provide a cheap and non-invasive way of sexing that species in the field. Second, even for species that can usually be sexed via secondary sexual dimorphisms, it is not uncommon to Table 1. Sample sizes and results for the species that we vibrated. Both Median time and Time range represent the time for trials that were successful. Unsuccessful trials were aborted after 600 s. Deirochelys reticularia miaria Kinosternon Sternotherus subrubrum odoratus hippocrepis Apalone spinifera N % successful Median time (s) Time range (s) find individuals that possess some characteristics of both males and females, thus making them difficult to sex. This method can be applied to those individuals even if it has not been validated for that species. In other words, if the method has been validated, then the outcome of vibrating the turtle can be used to assign the sex as either male or female; however, if it has not been validated for that species, inducing an erection would allow the turtle to be sexed as a male, and failure to induce an erection would simply leave the turtle with an unassigned sex code. Using a vibrator in this manner had already been helpful for sexing problematic individuals in our own research (McKnight, pers. obs.). Third, it is often desirable to collect or monitor several individuals of a known sex (e.g., for movement studies). This is another situation where the method can be used even for species for which vibrating has a low or undetermined success rate. For example, if a research endeavor requires ten males of a species that is difficult to sex, an investigator could simply vibrate turtles until they had ten with erect phalli. Finally, in situations where a researcher is working with a species, subspecies, or population for which secondary sexual dimorphisms are unknown or questionable, this method can be used to help validate a secondary sexual dimorphism. It is often possible to identify some individuals as females by palpating turtles for the presence of eggs (Zuffi et al., 1999) or employing a sonogram to look for evidence of enlarged follicles, and using the method we described to vibrate turtles will allows some males to be identified. Therefore, the combination of these two methods would allow researchers to easily compare the morphometrics of known males and known females to identify secondary sexual dimorphisms. Indeed, this final method has proved useful in our research. At the outset of this project, we were not certain if our populations of D. r. miaria (a subspecies that has been poorly studied) were sexually dimorphic. We had a few known females (identified by the presence of eggs or enlarged follicles), but the majority of individuals appeared to be males (with a few immature females), resulting in a strongly skewed sex ratio (8:1 M:F [adult sex ratio], 4.7:1 [including suspected immature females]). Based on our extensive trapping, the sex ratio did not appear to be a trapping artifact, but it was skewed enough that we were not confident that published sexual dimorphisms were correct for this subspecies at this location (Ernst and Lovich, 2009). However, by using the vibrator method to confirm that a subset of the suspected males were actually males, we were able to plot regressions (Fig. 2), which then allowed us to use secondary sexual dimorphisms to confidently assign sex to turtles in our populations.

122 120 D.T. McKnight et alii the Missouri State University Institutional Animal Care and Use Committee (IACUC protocol no ). We conducted the research under Oklahoma Department of Wildlife Conservation scientific collecting permits #5269, 5950, and 5610 and adhered to the ASIH/HL/SSAR Guidelines for Use of Live Amphibians and Reptiles. REFERENCES Fig. 2. Secondary sexual dimorphisms in western chicken turtles (Deirochelys reticularia miaria). A) Prior to vibrating several males, we had identified females by the presence of eggs, but because of the skewed sex ratio and large size of the confirmed females, we were not confident that the relative precloacal tail length was reliable for this species. B) The vibrator allowed us to confirm the sex of several males, thus establishing that this is a true sexual dimorphism (precloacal tail length was not recorded for two of the vibrated males). M = males, F = females, J = juveniles, (vibr) = confirmed via the vibrator, (eggs) = confirmed via the presence of eggs or enlarged follicles. Regression lines are only shown for confirmed males and confirmed females. In conclusion, although this method may not be a silver bullet for sexing all problematic turtle species, it is reliable for some species, and it has value even for species with low or undetermined rates of efficiency. It is cheaper, easier to implement in the field, and less invasive than many of the alternative techniques, and it has already proved useful in our own research. Therefore, we think that it will enhance other research projects as well. ACKNOWLEDGEMENTS We would like to thank the Oklahoma Nature Conservancy and several private land owners for granting us access to their properties and assisting us with our research. The research presented here was conducted as part of a large survey effort that was funded by an Oklahoma Department of Wildlife Conservation State Wildlife Grant and the Delta Foundation and approved by De Solla, S.R., Portelli, M., Spiro, H., Brooks, R.J. (2001): Penis displays of snapping turtles (Chelydra serpentina) in response to handling: Defensive or displacement behavior. Chelonian Conserv. Biol. 4: Dustman, E.A. (2013): Sex identification in the common snapping turtle (Chelydra serpentina): A new technique and evaluation of previous methods. Herpetol. Rev. 44: Ernst, C.H., Lovich, J.E. (2009): Turtles of the United States and Canada, 2 nd ed. Johns Hopkins University Press, Baltimore, Maryland, USA. Gibbons, J.W., Lovich, J.E. (1990): Sexual dimorphism in turtles with emphasis on the slider turtle (Trachemys scripta). Herpetol. Monogr. 4: Iverson, J.B. (1985): Geographic variation in sexual dimorphism in the mud turtle Kinosternon hirtipes. Copeia 1985: Lefebvre, J., Carter, S., Mockford, S.W. (2013): New experimental method for semen extraction in freshwater turtles. Herpetol. Rev. 44: Ligon, D.B., Bidwell, J. Lovern, M.B. (2009): Temperature effects on hatchling growth and metabolic rate in the African spurred tortoise, Geochelone sulcata. Can. J. Zool. 87: Ligon, D.B., Backues, K., Fillmore, B.M., Thompson, D.M. (2013): Ontogeny of gonad and genital morphology in juvenile alligator snapping turtles (Machrochelys temminckii). Herpetol. Conserv. Biol. 9: McKnight, D.T., Harmon, J.R., McKnight, J.L., Ligon, D.B. (2015): Taxonomic biases of seven methods used to survey a diverse herpetofaunal community. Herpetol. Conserv. Biol. 10: Owens, D.W., Hendrickson, J.R., Lance, V., Callard, L.P. (1978): A technique for determining sex of immature Chelonia mydas using radioimmunoassy. Herpetologica 34: Readel, A.M., Dreslik, M.J., Warner, J.K., Banning, W.J., Phillips, C.A. (2008): A quantitative method for sex identification in emydid turtles using secondary sexual characters. Copeia 2008: Rodrigues, J.F.M., Soares, D.de O., Silvia, J.R.F. (2014): Sexing freshwater turtles: penile eversion in Phrynops