ABSTRACT
Anolis lizards have been a model of study in ecomorphology in the Caribbean islands because species with the same type of microhabitat share similar morphological features. But despite their great diversity, little is known about continental species. We analyzed the relationship between the anatomical characteristics of the appendicular skeleton and the locomotor performance of two Anolis species found in Colombia that have different use of habitat. Anolis heterodermus a strictly arboreal species was compared with Anolis tolimensis that inhabits the lower strata of vegetation. These two species differ in their body plan, not only in body shape and external morphological features, but also in the skeleton and appendicular musculature. The results highlight the muscle and bone specializations associated with the use of habitat in this genus, such as the presence of more robust bones to enlarge the surface of muscle insertion, the thickening and loss of carpal parts, thickening of tendons associated with the manus, and greater development of muscle mass in the forelimbs by A. heterodermus with respect to A. tolimensis. These differences are related to the use of the microhabitat and the locomotor style of each species.
INTRODUCTION
The Anolis lizards have been a model of study in ecomorphology in the Caribbean islands since the species that share the same type of microhabitat also share similar morphological features (Losos, 1992; Beuttell & Losos, 1999). Thus, the ecomorphs describe the correlated evolution of the morphological and ecological features in species that occupy the same space with a diversity of microhabitats (Williams, 1972; Poe & Anderson, 2019). Although equivalent processes have been observed in continental lizards, little has been studied about the morphology of continental Anolis and the various specieś relation to ecology, habitat use and the interactions that regulate these processes. In 2016, Moreno-Arias and Calderón-Espinosa described morphological diversity in 51 Anolis species from northern South America, defining as a result ten different morphological groups (morphotypes) determined by similar characteristics such as body shape, size, proportion of limbs and subdigital lamellae, possibly originated through an adaptive radiation pattern similar to that observed in other studies in the Caribbean islands (Losos & Miles, 2002; Losos, 2009). Later studies also support that adaptive radiation could explain the origin of similar Anolis morphotypes of continental and insular environments (Poe et al. 2018; Poe & Anderson, 2019).
Colombia has the greatest Anolis diversity in the world with 78 described species (Uetz & Hošek, 2019), among which we find species with different life styles. Anolis heterodermus Duméril 1851 is a strictly arboreal species that uses the thin branches of vegetation and lives in ecosystems of scrub and Andean forests (Miyata, 1983; Vargas-Ramírez & Moreno-Arias, 2014), and is classified as morphotype MT4 (Moreno-Arias & Calderón-Espinosa, 2016); along with other strictly arboreal continental lizards of compressed bodies and short limbs that correspond ecologically to the ecomorph “twig” in studies carried out on insular lizards (Miyata, 1983; Torres-Carvajal et al. 2010; Vargas-Ramírez & Moreno-Arias, 2014). On the other hand, Anolis tolimensis Werner 1916 is a species that inhabits the lower strata of Andean forest, using mainly the soil and the lower parts of trunks and branches (Ardila-Marín et al. 2008), and it is classified as morphotype MT2, along with other small lizards with a cylindrical or depressed body and long limbs, corresponding ecologically to the “grass-bush” ecomorph in studies carried out in the Caribbean islands (Moreno-Arias & Calderón-Espinosa, 2016). Consequently, these species are a model for comparative functional studies, given the morphological differences in the shape of the bodies, the sizes and proportions of their limbs as well as their habitat use. It is likely that these variations are also represented in their musculoskeletal characteristics and, therefore, in their locomotor performance.
In Anolis, as in other lizards, the correlation between limbs dimensions, the substrate and the locomotion characteristics of species has been analyzed; and a strong tendency of the organisms is recognized with related corporal plans using a habitat of equivalent forms (Losos & Sinervo, 1989; Losos & Irschick, 1996; Losos et al. 1997; Irschick & Losos, 1998, 1999; Vanhooydonck et al. 2005, 2006b). Herrel et al. (2008) analyzed the relationship between the characteristics of the appendicular skeleton and the locomotor style of Anolis valencienni and Anolis sagrei from the Caribbean islands. These species are assigned to different ecomorphs, and the authors concluded that there are differences in the musculature and the bone elements associated with the locomotor performance of each species. Likewise, studies in continental Anolis species support the relationship between morphological traits and locomotor performance (Velasco & Herrel, 2007; Moreno-Arias, 2014).
Therefore, taking into account the context of the morphological diversity in continental Anolis, we set the following objectives: (1) characterize the appendicular morphology of A. heterodermus and A. tolimensis; (2) evaluate the differences between the locomotor style and performance of these two species; (3) compare the morphology and locomotor performance between the continental Anolis species belonging to the MT4 and MT2 morphotypes with the insular species belonging to the “twig” and “grass-bush” ecomorphs.
METHODS
We obtained specimens of Anolis heterodermus and Anolis tolimensis from the Reptile Collection of the Instituto de Ciencias Naturales of the National University of Colombia, Bogotá. For each species the musculature was described. We dissected 6 individuals of A. heterodermus (ICN 6248, 6250, 6251, 6256, ICN-R 13170, 13171) and 6 individuals of A. tolimensis (ICN 12836, ICN-R 12831, 12832, 12833, 12834, 12835). We identified each muscle of the anterior and posterior limbs and recorded its origin, insertion and the presence of tendons, following the nomenclature of Snyder (1954), Zaaf et al. (1999) and Herrel et al. (2008). Additionally, we classified the muscles into functional groups following Herrel et al. (2008). We extracted the muscles and stored them in vials with 70% alcohol, and we weighed them by functional groups using a PRECISA XB220A ± 0.01g analytical balance. The muscles for A. heterodermus were recorded in detail in Tables 1 and 2, and we listed the differences with A. tolimensis.
The description of the appendicular skeleton of A. heterodermus and A. tolimensis was based on two adult specimens of each species (ICN-R-6287, 6251, 13036, 13027). We prepared the skeletons using the technique of differential staining for bone and cartilage from Wassersug (1976). The osteological description follows the Krause nomenclature (1989), and the description of the sesamoids follows Jerez et al. (2010). Observations were made with a NIKON C-LEDS stereoscope, and specimens were photographed with a LEICA M125 motorized stereomicroscope.
For the locomotor performance tests we obtained live specimens of the two species. We captured 30 adult specimens of Anolis heterodermus (17 males: SVL = 65.7 mm average ± 4.80 mm SD and 13 females: SVL = 64.5 ± 5.69 mm) from the municipality of Tabio, Cundinamarca (Colombia) between April and May 2016. The individuals were transported to the Evolutionary Ecology Lab of the National University of Colombia in Bogotá, where they were kept in terrariums with branches and plant material at room temperature (17-19 ° C), thanks to the temperature in Bogotá is equivalent to the operational temperature of the species in its natural environment (Méndez-Galeano & Calderón-Espinosa, 2017). Individuals were fed twice a day with small insects and the terrariums were sprayed with water twice a day to keep the environment humid. Once the performance tests were performed, the individuals were released at the same capture site.
Likewise, we captured 30 adult individuals of A. tolimensis (11 males: SVL = 47.5 ± 2.28 mm and 19 females: SVL = 50.2 ± 2.22 mm) in the municipality of Silvania, Cundinamarca (Colombia) between May and July of 2017. These were temporarily stored in wet canvas bags with vegetation, while the performance tests were carried out at the capture site; and they were immediately released at the same site.
For all individuals of the two species that were tested for performance we recorded sex (presence of hemipenia, body size), weight (balance OHAUS CL201 ± 0.1g), body dimensions (snout-vent length [SVL] and tail length [Lco], forelimb length: humerus [Lbr], radius-ulna [Labr], metacarpus and length of the longest digit not including the claw; hindlimb length: femur [Lm], tibia [Lp], metacarpus [Ldp] and length of the longest toe without including the claw). We used a digital electronic caliper (REDLINE M ± 0.02 mm) and always took the lengths of the limbs of the right side.
For the performance tests we filmed each individual from the side view using a NIKON D3300 camera at 60 frames per second (60p) as the animals ran on two cylindric platforms of different diameters (10 mm and 80 mm) that were 2m long and arranged at an inclination of 45°. The animals were encouraged to run as fast as possible clapping behind them or touching their tails with a thin paintbrush. At least three sequences per individual were obtained and analyzed for each platform.
For all the analyses we performed ANOVA and Kruskal-Wallis statistical tests according to the parametric or non-parametric nature of the data. Following the methodology proposed by Sokal & Rohlf (1995) we used the Shapiro-Wilks test to evaluate normality and the Levene test for homoscedasticity.
Initially we evaluated differences in snout-vent length (SVL) and other body measurements (LCo, Lbr, Labr, Lm, Lp and Ldp) between species. We also analyzed the differences between the proportions of the hindlimbs with respect to the forelimbs (Lep/Lea) in total length and in each of its components (arm-thigh and forearm-leg). This analysis was done with a correction for size (SVL) in order to eliminate the error produced by the difference in the size of the species.
In order to establish which were the most important muscle groups during the locomotion of each species and their differences, we did an analysis of variance from the weights recorded by the functional groups of each species, corrected with the total weight of each.
We analyzed the videos using the software TRACKER (version 4.10.0; Brown, 2017) examining from the moment the lizard starts to move until it leaves the frame of the scene (pixel error margin <20%). Based on this data, we quantified the average speed of movement obtained by analyzing the position of the tip of the snout (distance traveled in the frame) with respect to time. We also quantified the length of the step, that is, the distance that the body moves forward during the support of the hindlimb and finally the step frequency, which is the number of steps per second. These variables were obtained by analyzing the frames in which the right foot was in contact with the substrate.
To analyze the locomotor performance, a correlation analysis was performed (Pearson test for the parametric samples and Spearman test for the non-parametric samples) taking in account the locomotion variables with respect to the size of the individuals, in order to evaluate whether these variables are determined by size. From this result, we corrected the data with the SVL values of each individual. We carried out a bidirectional multivariate analysis of nonparametric variance (two-way PERMANOVA (Anderson, 2001)) to establish the level of interaction of the species and platforms on the locomotion variables evaluated. In addition, we did a univariate comparison using the ANOVA and Kruskal-Wallis tests to establish the influence of the species and the surface among the locomotion variables studied. Following the methodology proposed by Herrel et al. (2008), only the movements made after the third step in each event were considered, that corresponds to the acceleration stage. We performed all the statistical tests using the software Past 3.15 (Hammer et al. 2001).
RESULTS
Morphometrics
Anolis heterodermus is a medium-sized lizard (SVL = 65.78 ± 4.96 mm) with a head length of almost 23% of its body (LaCa = 22.14 ± 0.72 mm). Its tail is long (LCo = 85.59 ± 9.63 mm) in relation to the length of its body. It is not sexually dimorphic in terms of size (Moreno-Arias, 2014). The species has a compressed and robust body. The forelimbs are longer and more robust (26.96 ± 3.82 mm) than the hindlimbs (22.71 ± 2.94 mm) as can be seen in Figure 1-A.
Anolis tolimensis is a small lizard (SVL = 49.36 ± 2.56 mm) with a head length of almost 26% of its body (LaCa = 13.33 ± 0.84 mm). It has a very long tail (LCo = 97.11 ± 17.42 mm), twice the length of the body. It is sexually dimorphic with females larger than males (Ardila-Marín et al. 2008). The males have a thin and compressed body, while the females are robust and have a cylindrical body. This species has short and thin forelimbs (22.06 ± 1.38 mm) as compared to the hindlimbs that are approximately twice the length (40.70 ± 2.10 mm) of the forelimbs as shown in Figure 1-B.
Anolis heterodermus is a significantly larger species than A. tolimensis (Kruskal-Wallis: H1.56 = 39.96, P = 2.581E-10). In general, the two species varied significantly in tail length (Kruskal-Wallis: H1.54 = 40.79, P = 1.695E-10), forearm length (Kruskal-Wallis: H1.56 = 5.999, P = 0.01431), thigh length (ANOVA: F1.56 = 105.9, P = 1.61E-14), leg length (Kruskal-Wallis: H1.56 = 41.15, P = 1.409E-10) and foot length (Kruskal-Wallis: H1.56 = 42.36, P = 7.605E-11). The exception was the arm length (Kruskal-Wallis: H1.56 = 1.185, P = 0.2763) which is similar in the two species.
An analysis of the differences between the proportions of the limbs with respect to the body of each species shows that A. tolimensis has a Lep / Lea ratio significantly higher than that of A. heterodermus (Kruskal-Wallis: H1.56 = 36.31, P = 1.68E-09), both in the total length of the limb and in each of its components (Thigh-Arm: ANOVA: F1.56 = 23.07, P = 1.21E-05 and Leg-Forearm: Kruskal-Wallis: H1.56 = 39.96, P = 2.59E-10). Therefore, both limbs of A. tolimensis are longer proportionally than A. heterodermus.
Appendicular skeleton
Anolis heterodermus
Pectoral girdle: it has a complete girdle (Figure 2-A, C). The clavicle is depressed and widened in almost all its extension, only compressed in the lateral end; it extends from the midline to the acromial process region. The interclavicle, T-shaped and depressed, has the medial process very wide anteriorly, but narrow and acute posteriorly; this extends to 1/3 of the pre-sternum; the lateral processes are very wide, with the distal end straight and extend up to ¾ of the length of the clavicle, and do not contact the coracoid. The epicoracoid, laminar and narrow, closes the coracoid fenestra anteriorly, and lies on the meso-sternum posteriorly. The coracoid presents the mesocoracoid thin and long; whereas, the metacoracoid is broad and slightly concave at the base and extends anteriorly where it becomes narrower and slightly convex; also, it presents the coracoid fenestra and the coracoid foramen. The scapula is quadrangular and presents a large process on the anterior edge very broad at the base and pointed, which forms the lateral edge of the scapula-coracoid fenestra. The suprascapula is broad at the base, and dorsally widens, forming almost a semicircle. The pre-sternum is rhomboidal, broad towards the anterior region, and narrow in the posterior region, where it articulates with two pairs of ribs; the meso-sternum is formed by two thin and long bars and articulates with two pairs of ribs.
Pelvic girdle: it has a complete, wide and robust girdle (Figure 3-A, C). The pubis is wide and exhibits a very wide and pointed prepubic process, which extends anteriorly from the base of the pubis. The epipubis is rhomboidal and is fused to the pubis. The ischium is wide and quadrangular, exhibits the ischiadic process posteriorly, which is very wide at the base and pointed. The ischiadic symphysis is ossified. The hipoischium is thin, short and cartilaginous. The obturator fenestra is broad and cordiform. Finally, the ilium is very compressed and wide in all its extension; presents a short process at the base of the anterior edge, which is broad at the base and pointed; it is positioned at 45° with respect to the longitudinal axis and articulates with the sacral vertebrae.
Forelimb: constituted by humerus, radio-ulna, carpals, metacarpals and five digits (Figure 4-A). The carpal elements of the proximal series are very robust, with a quadrangular ulnar bone, an almost rectangular radial and a triangular central. Regarding to the distal series, only carpals 2 to 5 are observed, carpal 1 being absent; the carpal 4 stands out, since it is longer and wider than the other carpals. The diaphyses of the metatarsals are thin, while the proximal epiphyses are widened, especially the one of metacarpal I; metacarpal III is the longest, and decreases in size III, II, IV, I, V. The phalangeal formula is 2-3-4-5-3. The terminal phalange is a claw, with a ventral and proximal process, which is wide, low and sharp. The sesamoid elements are the ulnar patella, the distal supraphalangial sesamoids, the palmar, the pisiform, and the sesamoid anterior to the pisiform. In general, the bones of the stylopodium and the zeugopodium are robust and thick.
Hindlimb: is conformed by femur, tibia-fibula, tarsus, metatarsus and five toes (Figure 5-A, C). The tarsus has the proximal tarsal with a conspicuous dorsal concavity, and the distal tarsals III and IV. In the metatarsals, the diaphyses are thin and the epiphyses are slightly widened; metatarsal IV is the longest of the series, which decreases in sequence IV, III, II, I, V. The phalangeal formula is 2-3-4-5-4. The terminal phalange is a claw, with a short proximal process and pointed on the lower edge. The sesamoid elements are the tibial lunula, the dorsal tarsal sesamoid and the distal supraphalangeals. In general, the bones of the zeugopodium and stylopodium are thick and robust, which is very noticeable in the femur that shows the proximal end of the diaphysis quite widened.
Anolis tolimensis
Pectoral girdle: it has a complete girdle (Figure 2-B, D). The clavicle extends from the midline to the region of the acromial process; it is depressed, wide in the medial region and narrow towards the lateral ends. The T-shaped interclavicula has a wide, pointed medial process and extends to the middle region of the sternum; the lateral processes are wide, thinning towards the ends and ending in acute form, and do not contact the coracoid. The epicoracoid is laminar and narrow, lies on the sternum, surrounds the coracoid and closes anteriorly with the coracoid fenestra. The coracoid has two processes that limit the coracoid fenestra: the anterior and thin mesocoracoid, and the metacoracoid wide at the base and slightly convex anteriorly; presents the coracoid foramen. The scapula is rectangular and exhibits a process on the anterior edge, wide at the base and pointed, which forms the lateral border of the scapula-coracoid fenestra. The suprascapula is wide at the base, but dorsally it only widens slightly. The pre-sternum is rhomboidal and articulates with two pairs of ribs; The meso-sternum is constituted by two long and thin bars, which articulates with three pairs of ribs.
Pelvic girdle: it has a complete and narrow appearance (Figure 3-B, D). The pubis is narrow in the anterior middle region and wide in the posterior middle region; in this last region it presents the prepubic process wide at the base, pointed, and extending anteriorly. At the base of the pubis it encounters the obturator foramen. Anteriorly, the pubis articulates with the epipubis, that haves a hexahedral shape. The ischium is rectangular, and very wide towards the middle region; presents a short, narrow and acute sciatic process. The obturator fenestra is narrow and cordiform. The ilium is wide, compressed and presents a narrow, short and blunt process at the base and towards the front edge.
Forelimb: constituted by humerus, radio-ulna, carpus, metacarpus and phalanges (Figure 4-B). The carpus presents the radial, the ulnar, the central and the carpals 1-5; with carpals 4 and 5 large and the other small. The metapodium is made up of five metacarpals; where III is the longest of the series that decreases in sequence III, IV, II, V, I. The phalangeal formula is 2-3-4-5-3. The terminal phalange is a sharp claw, with a short and blunt ventral process. The sesamoid elements are the ulnar patella, pisiform, palmar and distal supraphalangial. In general, the bones of the stylopodium and zeugopodium are thin and gracile.
Hindlimb: it is complete (Figure 5-B, D). The tarsus presents the proximal tarsal, the distal tarsals III and IV. The metatarsal is composed of five metatarsals, where the IV is the longest of the series that decreases in sequence IV, III, II, I, V. The phalangeal formula is 2-3-4-5-4. The terminal phalange is an acute claw, with a short proximal and blunt process in the lower margin. The elements of the sesamoids are the tibial patella, the tibial lunula, the dorsal tarsal sesamoid and the supraphalangial distal. In general, the bones of the stylopodium and zeugopodium are thin, the femur has a curved appearance and the fibula is very thin with respect to the tibia.
Muscles and distribution of muscle mass
We identified all the muscles of the anterior and posterior limbs in the two species, detailing their origin and insertion (Tables 1, 2). We grouped the muscles into functional groups, following the proposal of Herrel et al. (2008). Once corrected by body mass (Table 3), the functional groups showed that the two species do not differ significantly in the total muscle mass (Kruskal-Wallis: H1.10 = 0.9231, P = 0.3367). We observed significant differences regarding the total weight of the muscles of each limb (Table 4), where the hindlimb had greater weight than the forelimb in the two species (forelimb ANOVA: F1.10 = 522, P = 5.819E-10, hindlimb Kruskal-Wallis: H1.10 = 8.308, P = 0.003948). However, in A. heterodermus the mass of the forelimb corresponds to 83% of the mass of the hindlimb, while in A. tolimensis the forelimb mass is only 43% of the hindlimb.
Based on the average muscle mass, some functional groups stand out. The humeral retractor muscles are the heaviest in the forelimb of both species, followed by the elbow extensors and the elbow flexors. Likewise, these same functional groups are heavier in A. heterodermus than in A. tolimensis. In the hindlimb the two species differ in the functional group with a greater muscle mass. For A. heterodermus the femur adductors followed by the knee flexors and the knee extensors are the heaviest muscles groups of the hindlimb. But for A. tolimensis the heaviest muscle groups are the femur retractors, followed by the femur adductors and the knee flexors (Table 3).
In the forelimb there are significant differences between the two species only in some functional groups such as the humeral retractors (ANOVA: F1.10 = 28.69, P = 0.0003207), elbow flexors (ANOVA: F1.10 = 31.98, P = 0.0002109) and elbow extensors (ANOVA: F1.10 = 39.67; P = 0.00008928), which present higher mass in A. heterodermus than in A tolimensis.
The results of the analysis of variance show significant differences in all the muscle groups of the hindlimb between both species, except for the femoral protractors (ANOVA: F1.10 = 0.1808, P = 0.6797). A. heterodermus has greater muscle mass in the femur adductors (ANOVA: F1.10 = 21.33, P = 0.0009523), knee flexors (Kruskal-Wallis: H1.10 = 8.308, P = 0.003948) and other hindlimb muscles (ANOVA: F1.10 = 11.52, P = 0.006841), while A. tolimensis shows greater development in femur retractors (Kruskal-Wallis: H1.10 = 8.308, P = 0.003948), femoral abductors (ANOVA: F1.10 = 13.7, P = 0.004103), knee extensors (ANOVA: F1.10 = 12.22, P = 0.00577), ankle flexors (ANOVA: F1.10 = 266.9, P = 0.00000001536) and ankle extensors (ANOVA: F1.10 = 73.79, P = 0.000006281).
Locomotor performance
In the applied correlation analysis, all the variables were correlated with body size (all P< 0.05). The respective corrections were made in such a way that the variance due to this factor was eliminated a priori. The two-way PERMANOVA showed that both the species and the platform had a significant effect on locomotor performance (Species: F1.116 = 225.01, P = 0.0001. Platform: F1.116 = 16.201; P = 0.0001); and the interaction effects were significant (Interaction: F1.116 = 10.547, P = 0.0004).
The ANOVA and Kruskal-Wallis tests showed significant differences in most of the comparisons in the variables analyzed (Table 5). The exception was in the comparisons of step frequency between both platforms for A. tolimensis (ANOVA: F1.58 = 1.878; P = 0.1758) and the step frequency between species for the wide platform (ANOVA: F1.58 = 1.099; P = 0.2988). This established that A. tolimensis did not vary the step frequency between different substrates, whereas A. heterodermus decreased this frequency on narrow substrates (Figure 6).
We found that velocity and the step length are the most important variables that differentiate these species in their locomotor style, since A. tolimensis shows a greater step length for any substrate and it reaches speeds of up to 4.6 times higher than A. heterodermus. However, both species increase the step length and increase their average speed significantly on broader substrates (Figure 6).
DISCUSSION
Morphometric characteristics
In the general form of the body, differences can be found between the two species that were analyzed. A. heterodermus has a broad and robust head, short limbs with respect to its SVL and a compressed body, while A. tolimensis has a smaller head, long limbs with respect to its SVL and a cylindrical body. Both species differ in most of their body dimensions, except in the length of the humerus and radio-ulna. This could be related to the fact that A. heterodermus, which uses narrow surfaces, has a compressed and robust body like other arboreal lizards (Molnar et al. 2017), while A. tolimensis with a cylindrical body could be more stable on wider surfaces (Losos & Irschick, 1996; Zaaf et al. 1999; Herrel et al. 2008; Abdala et al. 2009).
These proportions coincide to the morphotypes assigned by Moreno-Arias & Calderón-Espinosa, (2016): the relative proportion of the hindlimb corresponds to less than 50% (approximately 35% in A. heterodermus) of its SVL, which is the morphotype MT4; and the relative proportion of the hindlimb is more than 80% (approximately 83% in A. tolimensis) of its SVL so it corresponds to the MT2 morphotype. Thus, the lizards located in the MT4 morphotype are those that exhibit the most extreme morphology, with short legs and arms associated with the use of very narrow surfaces and slow movements (Moreno-Arias & Calderón-Espinosa, 2016).
However, although the length and proportions of the limbs have been previously associated with habitat use and type of perch (Losos, 1990b, 1992), other morphological characteristics of A. heterodermus may be the product of ecological factors of the ecosystems where it has evolved. This species is the only Anolis in Colombia that is distributed at more than 3200 meters above sea level in the páramo (Miyata, 1983; Vargas-Ramírez & Moreno-Arias, 2014), which experiences drastic changes in temperature during the day, as well as high solar radiation and high water stress during the dry season (Wollenberg et al. 2014). The highland species generally has larger bodies and larger scales on the body to better withstand water stress (Soulé & Kerfoot, 1966), both characteristics present in A. heterodermus; and this explains the size differences with A. tolimensis that lives in low altitudes in Colombia.
Appendicular Skeleton
The two species analyzed differ anatomically, both in the girdles and in the limbs. In the complex pectoral-sternum girdle A. heterodermus has a clavicle that is noticeably wider than in A. tolimensis, which increases the bone surface for greater muscle fiber insertion of the Mm. clavodeltoideus pars superficialis and pars profundus; these originate in the clavicle and interclavicle and are inserted in the anterior part of the humerus, promoting a greater force for the movement of the arm upwards and to the front (protraction and humeral abduction according to Herrel et al. (2008)) that A. heterodermus must develop in the narrow branches where it moves.
A. heterodermus has a concave coracoid at the base that is markedly convex anteriorly, whereas in A. tolimensis it is slightly convex at the base. In this bone the muscles of the Mm coracobrachialis longus and brevis, M. coracohumeralis posterior and M. supracoracoideus all originate; these are inserted into the humerus and allow the arms to approach the body (retraction and humeral adduction according to Herrel et al. (2008)). These differences of the coracoides are related to the greater mass of these muscles for A. heterodermus compared to A. tolimensis, allowing greater traction in climbing and in vertical movements. Also, in the coracoid region the biceps muscles (M. biceps I) are located. These muscles possess long and strong tendons, which in A. heterodermus are involved in keeping the body close to the grip surface to avoid possible falls, for example, due to wind and rain.
Anolis heterodermus has an elongated sternum that is consistent with the morphology reported by Herrel et al. (2008) for A. valencienni. The sternum of A. tolimensis is broad and short, as described in A. sagrei. The meso-sternum of A. heterodermus is long and thin and articulates only with two pairs of ribs instead of three as in A. tolimensis. These modifications are reported by Beuttell & Losos (1999) for other Anolis in the Carribean islands with climbing habits; and this characteristic could be related to a broad bone surface for the insertion of the humeral retractor muscles, such as M. pectoralis pars superficialis, which originates along the entire sternal plate, acting to resist gravity when climbing on vertical surfaces.
In the pelvic girdle we see a particular pattern between the continental and island species so far reported: A. heterodermus has a pelvic girdle with very broad elements, compared to A. sagrei and A. valencienni that exhibit relatively widened girdles (Herrel et al. 2008); and A. tolimensis has thinner bones than the other three species. Therefore, the species analyzed in this work differ from their ecomorphological equivalents in the Caribbean islands in the pelvic girdle, probably related to phylogenetic factors, but a comparative analysis with a greater number of species is required.
The widening of the pelvic girdle suggests a greater surface for muscle insertion related to the capacity of moving forward. For example, muscles associated with the pubis and ischium can lead to a greater pushing force in A. heterodermus compared to A. tolimensis, since they exhibit a greater mass (e.g. femoral adductor: M. pubofemoralis pars ventralis, M. anterior ischiofemoralis, M. adductor femoris, and femoral protractors: M. pubofemoralis pars dorsalis int., and ext., M. ischiofemoralis dorsalis anterior).
The knee extensors that originate in the ilium (e.g. M. iliofibularis and M. A. pars dorsalis, among others), could be involved on a horizontal surface in pushing the body forward, while on a vertical surface they would participate, in addition to the push, to overcoming gravity while climbing (James et al. 2007; Herrel et al. 2008). Thus, in A. heterodermus a wide pelvic girdle does not imply the use of muscles for more efficient movement on the ground. In contrast, in A. tolimensis a long ilium and a narrow pelvic girdle half laterally compressed, allows each leg to move easier in a sagittal plane (Russell & Bauer, 2008), and this represents an advantage for running.
The magnitude of the changes in A. heterodermus could be due to phylogenetic factors, since the Phenacosaurus clade has traditionally been associated with structural changes at the level of bones (Dunn, 1944; Lazell, 1969); whereas the absence of these characters in A. tolimensis could be due to the evolutionary history of their independent lineage.
Anolis heterodermus exhibits limbs with robust bones when compared to A. tolimensis. This contrasts with A. valencienni and other arboreal lizards that have thin and gracile bones, associated with the loss of body mass to counteract weight on inclined surfaces (Beuttell & Losos, 1999; Mattingly & Jayne, 2004, 2005; Herrel et al. 2008). It has been shown that the relative length and thickness of elements of the appendicular skeleton are associated with differences in locomotor abilities (Lemelin & Schmitt, 1998; Kirk et al. 2008; Patel, 2010; Almécija et al. 2015). The differences in A. heterodermus with respect to A. tolimensis could be due to a musculoskeletal specialization of the limbs, which allows a better performance of the grip (Zaaf et al. 1999; Moro & Abdala, 2004).
Other differences in the appendicular skeleton are observed in the autopodium. A. heterodermus has carpal elements and proximal epiphyses of metacarpals that are wider than A. tolimensis. A greater development of the proximal epiphyses of the metacarpals as a muscle insertion area is also a characteristic that is observed in arboreal species such as Polychrus acutirostris (Moro & Abdala, 2004) and Chamaeleo calyptratus (Molnar et al. 2017).
The absence of the first carpal bone in A. heterodermus was also seen, and although this feature has not been previously reported in other Anolis lizards, the reduction and modification of bone elements in the manus (within which there is a reduction of the palmar sesamoid and the central bone that are both characteristics present in both A. heterodermus and A. tolimensis) are linked to a more secure grip (Abdala et al. 2009; Fontanarrosa & Abdala, 2013; Molnar et al. 2017). According to Fontanarrosa & Abdala (2016), these characters are shared with other groups of lizards (Iguana, Physignatus, Tropidurus, Anisolepis, Techadactylus, Gonatodes and Homonota), and can be strongly linked to climbing and gripping skills. Thus, although A. tolimensis is not a strict climber, it has all the characteristics necessary for efficient climbing.
In the same way, the differences found in the tarsal elements, where the proximal tarsal bone of A. heterodermus has a very pronounced concave region with respect to A. tolimensis, could be related to a muscular and tendinous arrangement associated with the grip habit that requires greater muscular insertion surfaces in highly specialized lizards (Zaaf et al. 1999; Moro & Abdala, 2004; Abdala et al. 2009).
Other studies have found a strong phylogenetic signal associated with body characteristics such as the SVL and the proportion of the limbs (Moreno-Arias & Calderón-Espinosa, 2016; Poe & Anderson, 2019), and this could also have a phylogenetic incidence in the differences observed in the anatomy of the appendicular skeleton between A. heterodermus and A. tolimensis, associated with the evolutionary history of each lineage. This should be evaluated in a comparative analysis of the continental Anolis.
Muscles and distribution of muscle mass
The results of this work show how the distribution of muscle mass varies between the two species in most functional groups. The two species differ considerably in terms of the total weight of the muscles in each limb. Although both species concentrate most of their muscle mass in the hindlimbs, A. heterodermus shows a 45% of total muscle mass in the forelimbs, while A. tolimensis only concentrates 30%. This distribution in both species is expected, since the hindlimb is typically dominant in the locomotion of lizards and is the one that generates the impulse for movement (Snyder, 1954; Reilly & Delancey, 1997a, 1997b; Irschick & Jayne, 1999).
However, it is clear that A. heterodermus distributes the muscle mass of the limbs almost equally, given a greater relevance of the forelimbs with respect to A. tolimensis and terrestrial lizards that distribute their weight mainly towards the hindlimbs (Herrel et al. 2008; Abdala et al. 2009). This may be due in part to the difference in body dimensions, since A. heterodermus has short hindlimbs, which reduce the total muscle surface.
Both species have the same total muscle mass with respect to their size, which indicates that the only difference is in the distribution of weight in the functional groups. Some functional groups show marked differences between both species, such as the retractor muscles of the femur, which are heavier in A. tolimensis. This could mean an increase in acceleration by having the ability to pick up the leg faster in a race event and with this would increase the frequency of steps in the movement phase (Mattingly & Jayne, 2004). This could also be seen in the performance tests, since A. tolimensis showed considerably higher speeds (4.6 times faster on average) than does A. heterodermus, where this functional group is less developed.
Likewise, A. heterodermus has heavier extensor muscles of the elbow and the abductors and adductors of the humerus when compared to A. tolimensis, and this could represent an advantage for supporting the weight of the body on vertical surfaces (Molnar et al. 2017). However, these modifications could reduce the possibility of moving quickly on both wide and narrow surfaces (Zaaf et al. 1999; Mattingly & Jayne, 2004, 2005).
When comparing the muscle mass of the functional groups of the species analyzed in the present work with the data previously published by Snyder (1954) for Iguana iguana, Crotaphytus collaris and Sceloporus undulatus, which are exclusively arboreal, terrestrial and semi-arboreal respectively, and Herrel et al. (2008) for island Anolis, we concluded that terrestrial lizards using broad substrates such as A. tolimensis, A. sagrei and C. collaris in general have similar muscle mass distributions. But they have certain variations in the femoral retractors, femoral adductors and ankle extensors (Figure 7), which may be associated with specific variations in the use of the microhabitat and its frequency of use. For example, A. tolimensis might use the soil less frequently than A. sagrei. The arboreal lizards show greater differences in most of the functional groups (Figure 7), accounting for the morphological variation at muscular level present in the species of this habit. Despite this, A. heterodermus was found to have more similarities with A. valencienni than with I. iguana, possibly due to its phylogenetic distance.
The myology and distribution of musculature in the Squamata are apparently very conservative (Diogo & Abdala, 2010; Molnar et al. 2017). Anolis and other closely related clades have no important differences in muscle distribution among lizards of terrestrial or arboreal habits (Lécuru, 1968; Russell, 1988; Zaaf et al. 1999; Moro & Abdala, 2004; Herrel et al. 2008), except in specific muscles such as carpal flexors, which are modified in some species in the insertion and with tendons that are more or less short between digits III and IV (Molnar et al. 2017; Lowie et al. 2018). In this case, both A. heterodermus and A. tolimensis do not exhibit significant differences in the appendicular muscle arrangement except for the muscles associated with the carpus in A. heterodermus, which have thicker and more conspicuous tendons than in A. tolimensis. This is especially true in the muscles M. pronator accesorius and M. extensor digitorum longus pars profundus, a feature not previously reported by Herrel et al. (2008) and Moro & Abdala (2004).
Similar to bone modifications in the carpus, these changes could be associated with specializations for gripping narrow surfaces (Fontanarrosa & Abdala, 2013; Molnar et al. 2017), as an additional modification in the tendinous pattern P of the palmar surface, where the tendon plate in the hand is reduced and the muscles of the digits derive from tendons accompanying the phalanges, a pattern found in both species. The P pattern has been associated with clades such as Polychrus, Chamaeleo and Anolis that grip to narrow surfaces (Moro & Abdala, 2004; Abdala et al. 2009). This would indicate that, although A. heterodermus is a specialist in grasping narrow branches, A. tolimensis also has the ability to move through this environment, since they exhibit the P configuration of the palmar surface. In a comparison between ground Anolis, A. tolimensis and A. auratus that have morphological and ecological similarities (although not in an equivalent way) (Moreno-Arias & Calderón-Espinosa, 2016), they share the same P configuration (personal obs), which could suggest that this arrangement associates both phylogenetic and functional factors, supporting the idea that Anolis are natural climbers.
In addition to this, as reported by Molnar et al. (2017) for chameleons, A. heterodermus has greater development of the humeral protractors with respect to A. tolimensis, which allows a more efficient movement of the shoulder in a parasagittal plane and favors the movement of the pectoral girdle in the direction of the body (Peterson, 1984) and, therefore, may imply an improvement in climbing. It is evident how A. heterodermus has tended to super-specialization of the forelimb, mainly of muscles and bones associated with the manus, which represents a significant advantage in the movement associated with the microhabitat that it occupies.
There are similarities in some functional groups used for common movements, such as the other forelimb and hindlimb muscles groups used for rotation and pronation among others, and the wrist flexors, which are apparently conserved in lizards (Lécuru, 1968; Russell, 1988; Zaaf et al. 1999; Moro & Abdala, 2004).
Locomotor performance
The analyses show a relationship between the movement pattern of the two species with respect to the substrate, tending to have slower movements on the thin platform, similar to that found by Herrel et al. (2008). However, although both species move more slowly on thin surfaces, the data obtained in the analysis of A. heterodermus are considerably more homogeneous than those obtained from A. tolimensis (Figure 6). It is clear that A. tolimensis uses fast reaction movements with higher acceleration velocities, while A. heterodermus makes long and harmonic movements, moving at a much more constant rate. In similar tests applied to the arboreal lizard Anolis equistris, the results show that individuals do not change in the parameters of steps per second on both substrates (Abdala et al. 2009), and this could determine that the twig Anolis have a relatively low sprint speed, but its locomotion is safer (Losos & Sinervo, 1989; Losos & Irschick, 1996; Vanhooydonck et al. 2006b).
As verified by Herrel et al. (2008), the significantly lower reaction speed in A. heterodermus with respect to A. tolimensis can be attributed to the importance of maintaining the center of mass close to the surface at locomotion. When a lizard performs movements on wide surfaces, the center of gravity remains constant between steps due to the additional range of movement that exists laterally. However, when the movement is over narrow surfaces, having a higher reaction speed for explosive propulsion could be less optimal, since the lateral displacement induces a change of the center of mass away from the surface (Spezzano & Jayne, 2004). Thus, the more harmonic movements of A. heterodermus could mean an advantage that allows them to stay on narrow surfaces in a more efficient manner. In contrast, by more frequently using wider surfaces for foraging and breeding activities, A. tolimensis will necessarily depend on a high reaction rate to perform better on hot surfaces and as an antipredation strategy (Mattingly & Jayne, 2004).
Although the differences observed from the performance tests carried out between A. heterodermus and A. tolimensis agree with the results obtained by Herrel et al. (2008) in A. sagrei and A. valencienni, we must take into account the differential habitat use by A. tolimensis, which moves on wide surfaces most of the time and occasionally on narrow surfaces (Ardila-Marín et al. 2008; Moreno-Arias & Calderón-Espinosa, 2016). Due to the evolution of subdigital lamellae, Anolis lizards are arboreal in different degrees; and although they show great diversity in the use of microhabitat, many of them retain the ability to climb in specific situations (Losos, 2009).
The continental species differ ecologically from the Caribbean species (Irschick et al. 1997; Schaad & Poe, 2010). Although Anolis morphotype MT4 shows strong similarity with its respective ecomorph in the Caribbean, more than any other group (Schaad & Poe, 2010), it is not so with the MT2 morphotype, which is difficult to quantitatively compare with any of the insular ecomorphs given the ecological differences in the species (Moreno-Arias & Calderón-Espinosa, 2016). Thus, the species analyzed in this study do not exhibit completely contrasting life habits (e.g. A. tolimensis using branches occasionally) (Ardila-Marín et al. 2008), and this causes the differences observed in the locomotion patterns between A. heterodermus and A. tolimensis to be smaller in magnitude than those of A. valencienni and A. sagrei observed by Herrel et al. (2008).
CONCLUSIONS
The results obtained in this article show that there are anatomical differences in the skeleton and the appendicular musculature of Anolis species that allow us to understand the locomotor mode of the species studied in relation to their use of the habitat. Although the species studied are phylogenetically separated, a morphological pattern is found that highlights the muscle specializations associated with habitat use in this genus.
It is evident that A. heterodermus exhibits specializations of the forelimb, associated with the muscles and bones of the manus (mainly of the carpus), which favors movement in the microhabitat that it occupies. Although A. heterodermus shows such specializations, many of the changes associated with the efficiency of vertical movement (climbing and grasping) were also observed in A. tolimensis, which shows how, although the Anolis species are associated with a specific microhabitat, they have the ability to move in arboreal habitats to a greater or lesser degree, supporting the idea that they are natural climbers.
The data provided by Herrel et al. (2008) and the present study, show how an understanding of the morphology of the locomotor apparatus can help to explain the evolutionary correlation of morphology, movement, locomotor style (advance characteristics) and habitat use in Anolis lizards. This demonstrating that morphological modifications based on habitat use go beyond simple external differences in the size and shape of the limbs.
AUTHOR CONTRIBUTIONS
All the authors worked equally in the elaboration of this manuscript.
CONFLICT OF INTERESTS
The authors declare that they have no conflict of interests.
ACKNOWLEDGEMENTS
Thanks to Dr. Martha Calderón, curator of Reptiles Collection from the Instituto de Ciencias Naturales (ICN-UNAL), for the loan of the specimens, for her statistical advice and for her contributions in the revision of the manuscript. We thank Dr. Nelsy Pinto for the information on some specimens and Dr. Pedro Sánchez for his advice in the statistical processing of data. We also thank Laboratorio de Equipos Ópticos Compartidos of Departamento de Biología, Universidad Nacional de Colombia, Sede Bogotá, and Dr. Luis Carlos Montenegro for the loan of the necessary equipment for the elaboration of this work. We thank El Recodo farm and its administrative committee, the Villa Marcela Poultry Farm and the Castañeda Family, which allowed the captured specimens to be obtained. Finally, thanks to Miguel Méndez, Johana Muñoz and Miller Castañeda for their help in data collection and field work, Maria José Espejo for the anatomical English corrections and Dr. Thomas R. Defler for the English editing.