Introduction

Reptile hatchlings usually grow rapidly and in many cases during development, there is variation in body size between sexes. In several species of lizards, males develop larger and more robust bodies than females, thus sexual size dimorphism seems to be most common in several species (^{Braña, 1996}; ^{Olsson et al., 2002}; ^{Bustos-Zagal et al., 2011}), while it is true that variation in body size is related to sexual selection, fecundity, and differential resource use between the sexes (^{Kratochvil & Frynta, 2002}). It has also been suggested that this may occur due to differences in energy allocation during growth (^{Smith & Ballinger, 1994}; ^{Haenel & John-Alder, 2002}), or even by the development of an increasingly aggressive behavior, which makes it possible to defend large territories with a lot of food and shelters available, as has been observed in herbivorous lizards (^{Carothers, 1984}).

In Ctenosaura pectinata, as occurs in other species of Iguanidae (e.g., I. iguana, C. pectinata ^{Casas- Andreu & Valenzuela-López, 1984}; C. similis ^{Fitch & Henderson, 1977}; Dipsosaurus dorsalis ^{Dibble et al., 2008}; Cyclura n. nubila ^{Gonzáles, 2018}), morphological variation related to sex is also common and usually recognizable in adult individuals, which is possible due to the changes in the coloration pattern and the variation in the body size. Adult males have larger heads and are more robust than females (^{Fitch & Henderson, 1978}). However, morphological variation in juvenile stages is a subject that has received little attention and should be addressed, due to changes that occur as a product of ontogenetic development (^{Durtsche, 2000}). Sex recognition in C. pectinata hatchlings a few days after hatching is difficult, due to the leaf green color of most of the body and the apparent little variation in morphology. The first attempt to identify the sex in juveniles showed that only after 21 months of age (1.7 years), males and females can be recognized by the development of the dorsal crest, the femoral pores, the bulge of the hemipenes, the total weight and the length of the head (^{Arcos-García et al., 2005}), which undoubtedly indicates that they are subadult individuals. Generating information to facilitate sex recognition in juvenile stages is useful when it comes to managing Ctenosaura pectinata hatchlings in confinement conditions; above all, to estimate if there are differences in development rates related to sex. The divergence in growth rates between individuals in wild populations, and in those kept under captive conditions, has been recorded in Crocodylus moreletti (^{Serna-Lagunes et al., 2010}), and it has been observed that confinement limits the development of individuals, as frequently occurs in C. pectinata farms (personal observation).

The objective of the study was to differentiate the sexes of C. pectinata offspring through morphological variation. This work also includes a key with the most divergent morphological characters with the purpose of making the identification of both sexes easier. Our test hypothesis that with the use of several morphological characters and their processing with traditional statistical techniques and multivariate analyses, it is possible to differentiate males and females in early stages of development.

Materials and methods

As part of a series of periodic visits to an iguanarium located in Morelos State, México (2 km south of Yautepec at 18° 49' 20" N, 99° 05' 37" W; 1,105 m elev.), we obtained morphometric data on 127 C. pectinata hatchlings (authorization Semarnat No. SGPA/DGVS/03490), produced by 4 females with an average size of 285.3 mm of snout-vent length (SVL) (± 21.5, range: 240.0 - 320.0), average clutch size = 31.1 eggs (±8.6), developed in an environment of tropical deciduous forest and semitropical climate (Fig. 1). To differentiate the sex of hatchlings, the following criteria according to ^{Solorzano and Canales (2009)} were considered: males with ovoid-shaped abdominal testicles connected to tubular spermatic cords; females with elongated abdominal ovaries and connected to wavy oviducts. The morphological measurements were obtained from dead individuals that only lived the first week after hatching, and were the following: SVL, tail length (TL), head length (HL, distance in a straight line between the anterior edge of the snout and the anterior edge of the eardrum), head depth (HD, vertical distance in a straight line between the lateral border of the throat and the parietal region), head width (HW, distance between the two edges of the head at the height of the postocular region), forelimb length (FLL, distance obtained by adding the length of the humerus, length of the radius, length of the metacarpus, and length of the phalanges of the longest toe), hindlimb length (HLL, linear distance obtained by adding the length of the femur, length of the tibia, length of the metatarsus, and length of the phalanges of the longest toe) (^{Hamilton & Sullivan, 2005}; ^{Herrel et al., 2007}), the weight (g) and the physical robustness (Weight/SVL)*100. All measurements were obtained with a digital vernier (0.01 mm), and the weight was recorded with a scale (precision 0.01 g).

Figure 1 Artificial nesting, incubation and hatching of Ctenosaura pectinata in captivity. A) Nest for incubating eggs; B) Eclosion of hatchlings; C) and D) Newly hatched hatchlings. SVL: snout-vent length, TL: tail length, HL: head length, HD: head depth, HW: head width, FLL: forelimb length, HLL: hindlimb length. Photos by Anna Theresa Dietrich Frick.  

Our data were confirmed with the additional review of gonads of 25 preserved individuals of similar size and age than hatchlings analyzed, hosted in the herpetological collection of the Universidad Autónoma del Estado de Morelos (INE.MOR.REP.119.1101).

The normal distribution of all the data was verified with the Kolmogorov-Smirnov test and the homoscedasticity of the variables between the sexes was estimated with a Levene test (Table 1). The differences between characters that differentiate males from females were obtained with one-way analysis of variance (ANOVA). To estimate whether the variation in the data really could be attributable to sex, a multivariate analysis of variance was used (MANOVA) with the variables obtained. Subsequently, the reliability of the variables between the sexes due to use of individual morphological characters was estimated with the coefficient of variation CV = (standard Deviation/mean) 100%, where CV ≤ 5.0% indicates that there is no variation and the use of the character is reliable, while CV ≥ 5.0% indicates variation (^{Castro-Franco et al., 2011}). To estimate whether all the variables contribute in the same magnitude to the separation of the sexes, a discriminant analysis was used (^{Johnson, 2000}), which allows the identification of separation of groups with greater precision, considering the simultaneous effect of all analyzed variables. To increase the certainty in the separation of the sexes, an analysis of covariance was used (ANCOVA), using the variable HL as the dependent variable, SVL as the predictor variable, and sex as the categorical variable. These variables were used because it is known that the head and the body regularly have allometric relationships (^{Sánchez & Gutiérrez, 2020}); males have large heads (^{Fitch & Henderson, 1978}) and ANCOVA is a robust test that allows to identify if the variation in the structures can be attributable to sex. An alpha of P ≤ 0.05 was used in all statistical tests.

Results

The ANOVA (Table 2) showed that males were 3.3 mm longer than females in SVL character, 1.29 mm longer in the TL character 2.91 mm FLL and 6.35 mm in HLL. These characters were also those that showed a coefficient of variation below 10% which reveals that they are reliable characters to separate the sexes Although it is true that there were no significant differences in HL character (F1,125= 0.11, p= 0.74560; Table 2), it also behaved as a character with little variation in both sexes (males CV= 8.77%, females CV= 7.08%). The characters with the greatest variation in the two sexes were the weight (CV= 13.91% males, CV= 12.58% females), and the physical robustness (CV= 14.36% males, CV= 12.66% females). The characters with the greatest variation only in females were head width (CV= 10.32% females, CV= 7.71% males), and forelimb length (CV= 14.64% females, CV= 7.83% males; Table 2). The MANOVA showed variation in the data because sex (λ Wilks= 0.3885, F98= 17.1, p= 0.0000). Of the nine characters examined, snout-vent length, tail length, forelimb length, and hindlimb length presented statistically significant differences between the sexes (Table 2). Discriminant analysis also showed that SVL (λ Wilks= 0.621095), HL (λ Wilks= 0.595876) and HLL (λ Wilks= 0.7707) are the characters that best contribute to the separation of sexes due to the little variation they have (Table 3). Additionally, ANCOVA showed that head length varies depending on the SVL of hatchlings (F1,121= 44.46, p= 0.0000) and the variation is attributable to sex (F1,121= 5.20, p= 0.0242). The same happened with the hindlimb length, which varies depending on the SVL (ANCOVA F1,121= 4.92, p= 0.0283) and the variation is also due to the sex (F1,121= 97.96, p= 0.0000).

Discussion

The traditional way of evaluating sexual dimorphism in lizards has usually been through the variation in the body size of males and females (^{Fitch, 1978}), and this has been observed in several families (e.g., Phrynosomatidae ^{Fitch, 1978}, ^{Lacertidae-Braña, 1996}, ^{Liubisavliević et al., 2008}; Scincidae Olsson et al., Anguidae ^{Dashevsky et al., 2013}; Chamaeleonidae ^{Stuart Fox & Moussalli, 2007}). In Iguanidae, species such as C. similis which is phylogenetically close to C. pectinata it has been observed that the SVL and the size of the head is larger in males than in females (^{Fitch & Henderson, 1977}). The variation in the size of the hatchlings within the litters in I. iguana has been explained as an adaptative strategy to survive drought (^{Van Marker & Albers, 1993}) or to facilitate the adaptation of various phenotypes in the face of unpredictable situations (^{Castro-Franco et al., 2011}). All this observed variation has been explained by the effect of environmental factors (extrinsic) such as the differential allocation of energy during growth the development of aggressive behaviors in adults to protect large territories, sexual selection, fecundity, and the differential use of microhabitats (^{Carothers, 1984}; ^{Haenel & John-Alder, 2002}; ^{Kratochvil & Frynta. 2002}). Only in recent years has it been observed that intrinsic factors such as the formation of chromosome races and genetic hybridization (Sceloporus grammicus ^{Cruz-Elizalde et al., 2023}; ^{Lozano et al., 2023}; C. pectinata ^{Zarza et al., 2009}; ^{Zarza et al., 2019}) also contribute to the variation in the body size of males and females. This reveals that the information generated on the subject has been focused on explaining the variation in body size of adults as a mechanism that offers evolutionary advantages of larger over smaller individuals. However, the difference between the sexes is not always due to variation in a single character such as SVL. We have evidence that morphological variation considering various characters can also occur in hatchlings which has been a virtually ignored topic. The data presented in this work show that dimorphism in C. pectinata can be identified in juvenile stages, considering the additive effect of various morphological characters, among which the snout vent length and the length of the forelimbs and hindlimbs stand out. This pattern of elongated bodies with long limbs in hatchlings is consistent with the same characters that define typical dimorphism in adult males of this species (^{Fitch & Henderson, 1977}). This explains why in a previous study (^{Arcos-García et al., 2005}) it was not possible to distinguish the sexes considering few characters, analyzed separately. Sex differences in SVL, and adult limb length have also been recorded in Sphenodon punctatus tuatara (^{Herrel et al., 2009}). We suggest that the variation in body size between males and females of C. pectinata emerges as a character that is linked to sex, which produces a certain level of predisposition in hatchlings, so that the body and other structures such as the length of the head and legs, develop faster in males than in females, and behaves as an allometric function of the SVL (^{Sánchez & Gutiérrez, 2020}). Variation in sex linked morphological characters (large heads in males vs large abdomen in females) has been observed in various groups of lizards (^{Braña, 1996}; ^{Scharf & Meiri, 2013}), and it has been suggested that a specific sex is the mechanism that triggers the variation in morphology between the sexes. In parthenogenetic species where there is obviously only one sex body and head size should vary little between individuals, which has been observed in lizards of the genus Aspidoscelis where little variation in body size has been recorded (CV= 4.75% ^{Hernández-Gallegos et al., 2015}). In conclusion, we suggest that more research is needed to verify if sexual dimorphism in body size and limb length also occurs in thermal sex dependent species, or if morphological variation does not exist in unisexual species. The use of multivariate analysis to document sex morphological differences of ranching specimens, is a useful tool considering the additive value of the variables involved.

KEY TO THE IDENTIFICATION OF SEXES IN NEONATES OF Ctenosaura pectinata USING MORPHOMETRIC CHARACTERS

1.A Body size (SVL) large (x= 52.53 mm ±3.21), three times longer than the head length.….....2A

B Body size (SVL) relatively small, no more than 50.00 mm (x= 49.23 mm ±4.33), and no more than three times the length of the head …….………………………………………………………………………….2B

2.A Large hind legs (x= 44.43 mm ±2.68), almost twice the length …...………………………………..Male

B Relatively small hind legs (x= 38.08 mm, ±3.40), no more than 1.5 times the length of the front legs………………………………………………………………………...………………………………………………Female