Similarities and differences in adult tortoises: a morphological approach and its implication for reproduction and mobility between species Marco A. L. Zuffi1, Annalisa Plaitano1 1Museo di Storia Naturale e del Territorio, Università di Pisa, Via Roma 79, 56011 Calci (Pisa), Italy. Corresponding author. E-mail: marcoz@museo.unipi.it Abstract. Sexes in Chelonia display marked differences. Sexual size dimorphism (SSD) is important in evolutionary biology. Different sexual strategies result in species specif- ic selection. Biometric variation in male and female tortoises of two species is studied. Eighteen biometrics were measured in 75 museum specimens (20 Testudo graeca; 55 T. hermanni). Nine of 18 parameters in T. hermanni and two of 18 in T. graeca were sexually dimorphic. Multivariate analyses (principal component analysis) highlighted two components, with bridge length the first and anal divergence the second com- ponent. The bridge length can be used to separate sexes and species. Males of both species were most different, whereas females of two species overlapped in body shape measurements. We hypothesise that female similarity could be a by-product of repro- ductive biology and sexual selection that optimise individual fitness. Keywords. Testudo hermanni, Testudo graeca, biometry, morphological model, reproductive strategies. INTRODUCTION The evolution of the complex structures that define extinct and living chelonian spe- cies was studied by Lee (1996). A progression of anatomical patterns appeared in the evo- lution of Chelonia. Physical structures range from complete ossification with a broadly developed carapace (i.e. Testudo, Geochelone) to reduction in bony shell (i.e. Chelydra, Apalone), or even to a poorly ossified carapace (i.e. Trionyx; Ernst et al., 1994). Sexual size dimorphism (SSD) (Darwin, 1874) is a very often debated aspect in reptil- ian species, particularly in recent years (Berry and Shine, 1980; Bonnet et al., 2001; Ols- son et al., 2002; Rubolini et al., 2006; Zuffi et al., 2006). Sexes in Chelonia often display marked morphological differences, with the largest dimensions often occurring in females (Berry and Shine, 1980; Ernst and Barbour, 1989; Ernst et al., 1994). Several hypotheses Acta Herpetologica 2(2): 79-86, 2007 ISSN 1827-9643 (online) © 2007 Firenze University Press 80 M.A.L. Zuffi and A. Plaitano have been advanced to explain the patterns of carapacial, scute or colour variation in tur- tle sexes. Habitat, diet, and habitus provide important correlations with carapace, head and skull shape (e.g. Claude et al., 2003, 2004; Zuffi et al., 2007), while different degrees of sexual dimorphism were recently found in the European species of the Testudo (i.e. T. graeca, T. hermanni and T. marginata) by Willemsen and Hailey (2003). Despite the wide interest in the evolution, selection and adaptive forces that mold carapacial morphology of chelonian species, few papers have dealt with morphology and function (Bonnet et al., 2001; Claude et al., 2003, 2004; Willemsen and Hailey, 2003; Zuffi et al., 2007). Broad patterns have become clearer, but more detailed comparisons at the specific level are needed. As data set for a preliminary comparison to examine details of interspecific sexual dimorphism in tortoises, we selected two similar, Palaearctic species of Testudo (but see Lap- parent de Broin et al., 2006 and their proposal to change genus into Eurotestudo), T. graeca and T. hermanni. In both these species, there are pronounced sexual differences in body size dimensions, and their courtship patterns have been described on average (Buskirk et al., 2001; Cheylan, 2001), but no real inspection of relationships between morphology and func- tion has been yet performed (but see Bonnet et al., 2001 for the Testudo horsfieldii). We present a brief study on the sexual dimorphism in T. hermanni and T. graeca, and we examine SSD within a species and between species. Our main idea is to test if morpho- logical differences between sexes and between species may be helpful in explaining dif- ferent evolutionary and reproductive strategies for each species (i.e. reproductive systems, movement). We wish to emphasize that this is a preliminary study and that further analy- ses might need to be conducted when possible. MATERIALS AND METHODS We examined 71 museum specimens (9 males and 9 females of T. graeca; 26 males and 27 females of T. hermanni) in the Museum of Zoology “La Specola” of the University of Florence, Italy and at the Civic Museum of Natural Science “Doria” of Genoa, Italy. We made 18 measurements of shell (Soldani, 2000), of the limbs and of the head and tail. Measurements were made with a calliper (accuracy ± 0.1 mm) or with a tape measure (all the measurements in millimetres). Weights were taken with an analogic dynamometre (accuracy: ± 25 g). For each we record: taxon, sex, age class (adult or juvenile, according to external morphol- ogy of plastron and tail length, Cheylan, 2001), season, year, locality and region of capture and/or death, and preservation matter (i.e. in liquid, mounted), matter of collecting (alive; dead). Animals born and dead in captivity were not considered. Abbreviations for morphometrical parameters are: carlen (carapace length from nucal to sopracaudal scute), carwid (carapace width between 6th and 7th marginal scute), plalengmin (minimum length between jugal and anal scute), plalengmax (maxi- mum length between giugal scute and an estreme margin of the anal scute), plawid (piastron width between abdominal and pectoral scutes), brilen (bridge length), carhei (carapace height between plastron base and the 3rd vertebral plate), anadiv (length of anal divergence), latcirc (latitudinal circumference), loncirc (longitudinal circumference), fore (forelimb length), hind (hindleg length), bmass (body mass in grams), healen (head length), heawid (head width), heahei (head height), tailen (total tail length), cloaca (distance between cloaca and tip of the tail). We excluded one juvenile that we could not determine its sex (Cheylan, 2001). 81Similarities and differences in adult tortoises Statistic analyses were performed with SPSS 13.0 statistical package for Windows. All param- eters were tested for normality. Parameters that were not normally distributed were log transformed prior to analyses. For each of selected morphological variables, we used an analysis of covariance (ANCOVA) to examine dependent variables for the influence of sex, species and their interactions, using the log transformed carapace length as covariate. We, thus, tested the differences between sexes for each species, the differences between males of the two species, and finally between the females of the two species. As all selected variables were highly intercorrelated, we followed up this univariate analysis with a principal component analysis (PCA) in order to identify the categorical variables responsible for the variation we measured (see Fowler and Cohen, 1993). RESULTS Nine parameters out of 18 in T. hermanni and two out of 18 in T. graeca were sexu- ally dimorphic (see Table 1 for details). In particular, males had relatively shorter bridg- es at any particular carapace length than did females (Fig. 1A, B). Comparisons of sexes by species were significant (T. hermanni: ANOVA, F1 = 194.653, P < 0.0001, ANCOVA interaction sex × lncarapace length: F2,49 = 281.87, P < 0.0001, Radj = 0.917; and T. graeca: ANOVA, F1 = 74.751, P < 0.0001; ANCOVA, F = 43.802, P < 0.0001, interaction sex × lncarapace length F2,15 = 46.538, P < 0.001, Radj = 0.843). Both species had a marked sexual dimorphism. Two parameters were identified by PCA as the most important morphometric features of carapace: the bridge length (first component) and the anal divergence (second compo- nent, Table 2). While females widely overlapped, males were markedly different in mor- phological space (Fig. 2). DISCUSSION Sexual size dimorphism is very evident in T. hermanni (Fig. 2a), but less evident for our much smaller sample of T. graeca (Fig. 2b). Nonetheless, males attain smaller body sizes than do females. Furthermore, we found that the bridge length, a parameter not pre- viously recognized as a sexually dimorphic one (Bonnet et al., 2001; Claude et al., 2003, 2004; Willemsen and Hailey, 2003), plays an important role in separating sexes and spe- cies. Bridge length should be carefully considered in future comparative studies. Males of the Hermann’s tortoises have a more marked anal divergence and smaller carapace and plastron than do females. In males of the Moorish tortoise we observed a marked anal divergence as well, but a larger carapace than females. It is interesting to underline how the bridge length appears to work in defining both sexual and interspecific differences. The bridge length is inversely proportional to the degree of mobility (Soldani, 2000). Mobile species (i.e. aquatic turtles such as Emys, Tra- chemys, Lyssemys, Pelomedusa and others) have smaller bridges when compared with ter- restrial taxa (Soldani, 2000; Zuffi, unpubl.). Besides, within a particular taxon (i.e. genus 82 M.A.L. Zuffi and A. Plaitano Ta bl e 1. M ea n va lu es ( ± st an da rd d ev ia tio n) o f th e 18 m or ph ol og ic al v ar ia bl es m ea su re d in b ot h se xe s of t he T es tu do h er m an ni a nd T . g ra ec a (A N C O VA , s ee m et ho ds fo r de ta ils ) (i n bo ld s ig ni fic an t v al ue s) . V ar ia bl e Te st ud o he rm an ni df F P Te st ud o gr ae ca df F P m al es fe m al es m al es fe m al es ca rl en 12 9. 5± 21 .3 14 5. 9± 34 .7 1, 51 4. 28 0. 04 3 15 5. 09 ±2 4. 22 14 2. 67 ±2 3. 89 1, 16 1. 2 0. 29 ca rw id 10 0. 6± 14 .8 10 9. 9± 24 .3 1, 51 2. 78 2 0. 10 1 11 3. 67 ±1 9. 72 10 4. 67 ±1 7. 02 1, 16 1. 07 4 0. 31 5 pl al en gm in 89 .8 1± 14 .9 1 11 9. 27 ±2 2. 77 1, 50 14 .4 42 <0 .0 00 1 12 9. 76 ±2 3. 74 11 9. 56 ±1 7. 46 1, 14 0. 90 4 0. 35 8 pl al en gm ax 11 0. 18 ±1 7. 52 12 9. 99 ±2 6. 91 1, 50 9. 72 0 0. 00 1 14 1. 98 ±2 5. 46 13 5. 51 ±2 7. 81 1, 14 0. 23 4 0. 63 6 pl aw id 90 .6 8± 13 .2 8 97 .4 7± 16 .6 5 1, 51 2. 61 3 0. 11 2 94 .7 2± 15 .0 8 90 .1 9± 13 .7 9 1, 16 0. 44 3 0. 51 5 br ile n 49 .5 2± 6. 30 64 .0 4± 11 .9 3 1, 50 29 .4 15 <0 .0 00 1 67 .9 9± 11 .8 1 65 .9 4± 8. 89 1, 16 0. 17 2 0. 68 4 ca rh ei 67 .6 2± 10 .3 3 76 .5 1± 14 .2 1 1, 50 6. 63 1 0. 01 3 78 .3 2± 11 .6 5 75 .6 3± 7. 65 1, 15 0. 30 9 0. 58 6 an ad iv 39 .2 9± 9. 16 29 .4 2± 7. 00 1, 50 19 .2 38 <0 .0 00 1 35 .4 4± 8. 52 25 .7 8± 4. 54 1, 14 7. 30 3 0. 01 7 la tc ir c 33 3. 04 ±5 3. 28 36 6. 44 ±7 0. 36 1, 50 3. 68 0. 06 1 38 9. 83 ±6 1. 19 34 7. 51 ±4 8. 84 1, 14 2. 23 0. 15 8 lo nc ir c 27 5. 12 ±4 1. 39 30 0. 41 ±5 6. 00 1, 50 3. 38 4 0. 07 2 31 3. 00 ±4 9. 95 28 4. 57 ±3 2. 58 1, 14 1. 69 2 0. 21 4 fo re 42 .1 2± 7. 74 44 .1 5± 9. 87 1, 28 0. 39 1 0. 53 7 46 .0 7± 8. 24 43 .0 4± 6. 52 1, 14 0. 63 2 0. 44 hi nd 44 .3 1± 8. 41 42 .1 1± 12 .9 0 1, 20 0. 23 5 0. 63 3 45 .1 1± 6. 82 43 .0 1± 11 .2 4 1, 14 0. 21 5 0. 65 bm as s 46 8. 91 ±2 48 .2 5 70 4. 80 ±4 40 .7 1 1, 46 5. 09 6 0. 02 9 67 9. 44 ±5 59 .5 1 31 2. 86 ±1 57 .1 6 1, 14 2. 79 3 0. 11 7 he al en 28 .1 1± 6. 57 28 .7 7± 5. 86 1, 36 0. 10 8 0. 74 4 28 .7 3± 2. 71 26 .4 4± 4. 47 1, 12 1. 34 0. 27 he aw id 19 .3 5± 3. 02 20 .4 7± 2. 58 1, 36 1. 46 7 0. 23 4 19 .4 7± 1. 5 17 .6 9± 1. 71 1, 12 4. 33 7 0. 05 9 he ah ei 18 .6 3± 3. 56 19 .9 0± 2. 89 1, 35 1. 38 5 0. 24 7 16 .8 1± 2. 83 14 .6 6± 2. 64 1, 12 2. 18 4 0. 16 5 ta ile n 49 .5 5± 11 .3 8 31 .9 1± 10 .1 1 1, 47 33 .0 21 <0 .0 00 1 40 .2 0± 11 .2 6 24 .5 7± 4. 04 1, 9 6. 90 4 0. 02 7 cl oa ca 25 .6 1± 6. 77 14 .1 0± 5. 65 1, 34 30 .8 67 <0 .0 00 1 -- -- - -- - -- - -- - 83Similarities and differences in adult tortoises or family), a bigger and longer bridge is typical of females. Females have less space for their limbs than do males (Lebboroni and Chelazzi, 1991; Andreu et al., 2000; Corti and Zuffi, 2003). In terrestrial tortoises, males must be able to quickly right themselves during male- male interactions for access to the females (Bonnet et al., 2001). The righting response is is inversely related to bridge length and positively correlated with space for the legs (Sol- dani, 2000; Corti and Zuffi, 2003; Zuffi, unpubl.). However, a short bridge may also be less effective in protecting males from predation. Females apparently sacrifice mobility for protection, thus their relatively long bridges. However, we also found interspecific variation in these sexually dimorphic traits. Such differences may be related to differences in species specific courtship behaviours. For instance, male T. hermanni chases the female, bites her hind legs and back of the carapace. He mounts directly on the female carapace, performing such an action for long time before copulation. In contrast, male T. graeca usually courts females with strong Fig. 1. Morphometric differences in carapace length versus bridge length (log transformed variables) in males and females: (a) in T. hermanni, (b) in T. graeca. Table 2. Principal Component analysis in Testudo hermanni and T. graeca (in bold the most informative variables for each component). Variable PC1 PC2 brilen 0.982 6.719*10-2 plalengmin 0.967 0.215 plalengmax 0.936 0.281 carhei 0.876 0.384 latcirc 0.867 0.485 loncirc 0.861 0.495 carlen 0.856 0.467 carwid 0.763 0.547 plawid 0.746 0.581 anadiv 0.162 0.973 (a) (b) 84 M.A.L. Zuffi and A. Plaitano blows from the gular shields on the posterior side of female carapace and with a pre- mounting behaviour before actual copulation. After a typical general scheme, similar to the T. graeca one, the T. hermanni male climbes on the female’s carapace and gives her repeated blows with the great spur of his tail tip. This happens whether the female is sta- tionary or moving. We hypothesize that T. hermanni males need to be more mobile than T. graeca males. Our study is a preliminary one and only experimental studies will serve to elucidate the interaction of behavioural and morphological patterns. Nonetheless, our finding that sexual dimorphic characters vary interspecifically is important. Mobility is a critical skill in tortoises and we show that females have effectively sacrificed this when compared to males. The selection producing such an outcome is little understood. Further studies should experimentally examine interspecific variation in these traits by comparing degrees of mobility between males and females and between species of tortoises. ACKNOWLEDGMENTS We wish to express our gratitude to John L. Tucker and to an anonymous referee which with their suggestions and criticism haved greatly improved this MS. Fig. 2. Principal component analysis of carapacial and sexual differences in T. hermanni and T. graeca. Small circles, T. hermanni; large circles, T. graeca; black circles, males; grey circles, females. 85Similarities and differences in adult tortoises REFERENCES Andreu, A.C., Díaz-Paniagua, C., Keller, C. (2000): La Tortuga Mora (T. graeca L.) en Doñana. Asociación Herpetológica Española. Monog. Herpetol. 5: pp. Berry, J.F., Shine, R. (1980): Sexual size dimorphism and sexual selection in turtle (Order Testudines). Oecologia (Berl.) 44: 185-191. Bonnet, X., Lagarde, F., Henen, B.T., Corbin, J., Nagy, K.A., Naulleau, G., Balhoul, K., Chastel, O., Legrand, A., Cambag, R. (2001): Sexual dimorphism in steppe tortois- es (Testudo horsfieldii): influence of the environment and sexual selection on body shape and mobility. Biol. J. Linn. Soc. 72: 357-372. Buskirk, J.R., Keller, C., Andreu, A.C. (2001): Testudo graeca Linnaeus, 1758-Maurische Landschildkröte. In: Handbuch der Reptilien und Amphibien Europas, Fritz, U., Ed, p. 125-177. AULA, Verlag. Cheylan, M. (2001): Testudo hermanni Gmelin, 1789-Griechische Landschildkröte. In: Handbuch der Reptilien und Amphibien Europas, Fritz, U. Ed, p. 179-289. AULA,Verlag. Claude, J., Paradis, E., Tong, H., Auffray, J.C. (2003): A geometric morphometric assess- ment of the effects of the environment and cladogenesis on the evolution of the tur- tle shell. Biol. J. Linn. Soc. 79: 485-501. Claude, J., Pritchard, P.C.H., Tong, H.J., Paradis, E., Auffray, J.C. (2004): Ecological corre- lates and evolutionary divergence in the skull of turtles: a geometric morphometric assessment. Syst. Biol. 53: 933-948. Corti, C., Zuffi, M.A.L. (2003): Aspects of population ecology of Testudo hermanni her- manni from Asinara Island, NW Sardinia (Italy, Western Mediterranean Sea): pre- liminary data. Amphibia-Reptilia 24: 441-447. Darwin, C. (1874): The descent of Man, and selection in relation to sex. Rand, McNally and Co., Chicago. Ernst, C.H., Barbour, R.W. (1989): Turtles of the World. Smithsonian Istitution Press, Washington and London. Ernst, C.H., Barbour, R.W., Lovich, J.E. (1994): Turtles of the United States and Canada. Smithsonian Institution Press, Washington and London. Forsman, A., Shine, R. (1995): Sexual size dimorphism in relation to frequency of repro- duction in turtles (Testudines: Emydidae). Copeia 1995: 727-729. Fowler, J., Cohen, L. (1993): Statistica per ornitologi e naturalisti. Muzzio Editore, Padova. Hailey, A., Loumbourdis, N.S. (1990): Population ecology and conservation of tortoises: demographic aspects of reproduction in Testudo hermanni. Herpetol. J. 1: 425-434. Lapparent de Broin, F., Bour, R., Param, J.F., Perälä, J. (2006): Eurotestudo, a new genus for the species Testudo hermanni Gmelin, 1789 (Chelonii, Testudinidae). C. R. Palevol. 5: 803-811. Lebboroni, M., Chelazzi, G. (1991): Activity patterns of Emys orbicularis L. (Chelonia Emydidae) in Central Italy. Ethol. Ecol. Evol. 3: 257-268. Lee, M.S.Y. (1996): Correlated progression and the origin of turtles. Nature 379: 812-815. Olsson, M., Shine, R., Wapstra E., Ujvari, B., Madsen, M. (2002): Sexual dimorphism in lizard body shape: the roles of sexual selection and fecundity selection. Evolution 56: 1538-1542. 86 M.A.L. Zuffi and A. Plaitano Rubolini, D., Pupin, F., Sacchi, R., Gentilli, A., Zuffi, M.A.L., Galeotti, P., Saino, N. (2006): Sexual dimorphism in digit length ratios in two reptile species. Anatom. Rec. 288A: 491-497. Soldani, A. (2000): Il dimorfismo sessuale nei Cheloni e ipotesi sul successo riproduttivo. Ms. Thesis, University of Pisa. Willemsen, R.E., Hailey, A. (2003): Sexual dimorphism of body size and shell shape in European tortoises. J. Zool., Lond. 260: 353-365. Zuffi, M.A.L., Celani, A., Foschi, E., Tripepi, S. (2007): Geographical patterns of reproduc- tive plasticity in the European pond turtle, Emys orbicularis. J. Zool, Lond. 271: 218- 224. Zuffi, M.A.L., Odetti, F., Batistoni, R., Mancino, G. (2006): Morphometrics and genetics in in the study of variability pattern of the European pond turtle, Emys orbicularis. Ital. J. Zool. 73: 363-372.