Morphometric differentiation and sexual dimorphism in Limnomedusa macroglossa (Duméril & Bibron, 1841) (Anura: Alsodidae) from Uruguay

Intersexual morphological differences within a species occur in many traits, including body size and shape. Many processes that cause geographic variability in morphology have been proposed: population structure, phenotypic plasticity (environmental effects on development), and natural and/or sexual selection. Several hypotheses can explain patterns of sexual dimorphism in anurans, including natural or intra/inter-sexual selection, and differences in life history strategies between sexes. Limnomedusa macroglossa is considered a habitat specialist restricted to rocky outcrops in Brazil, Argentina, Paraguay, and Uruguay. We evaluated the extent of sexual (size and shape) dimorphism in L. macroglossa from Uruguay based on morphometrics and secondary sexual characteristics, while taking into account geographic variation. Sexual dimorphism in body size of adults was found, but multivariate analyses did not demonstrate the existence of significant differences in shape. There were also significant differences in body size and hind leg measurements among six hydrographic basins as a result from the phenotypic plasticity correlated with local temperature, representing a clinal variation along the latitudinal gradient of Uruguay. The sexual dimorphism found in body size is probably the consequence of higher growth rates and/or late sexual maturity in females, which favors larger body size for accommodating larger ovaries, and thus, higher reproductive output.


INTRODUCTION
Morphology is one of the main components of the phenotype that can be studied through qualitative as well as quantitative characteristics. In particular, morphology can be assessed via morphometrics to quantitatively describe, analyze and interpret morphological variation within and between species (Kaliontzopoulou, 2011;Rohlf, 1990). Morphological quantitative traits are usually polygenically inherited and show considerable plasticity in relation to environmental factors (Babik and Rafinski, 2000). Furthermore, plasticity can lead to geographic variability in morphology. In that sense, many processes have been proposed, such as: biogeographical barriers that partially isolate populations, effects of environmental parameters (precipitations and temperature) on growth rates, and action of sexual selection resulting in sexual dimorphism (Schäuble, 2004). Body size is a strongly plastic morphological trait (Green, 2015) fundamental in physiological and ecological contexts. Traditionally, snout-vent length (SVL) has been used as the gold standard to measure body size in frogs (Kupfer, 2007). Among anurans, analyses of intraspecific geographical variability in morphology have often revealed extensive variation in body size (Schäuble, 2004). Due to the limited dispersal ability and high philopatry in frogs, it is common to find intraspecific differences in morphology among geographically separated populations, particularly in body size, caused by genetic divergence among isolated populations (Baraquet et al., 2012;Castellano et al., 2000). In addition to geographic distance, landscape features could account for spatial morphological variation. For instance, hydrographic basins could act as physical barriers promoting isolation and spatial structuring among populations as a result of changes in altitude, slope and landscape features among basins. Moreover, climate and food availability may also vary geographically, leading to differences in the ability to grow, resulting in morphological variation (Lovich and Gibbons, 1992;Hartmann, 2016).
Another source of intraspecific variation could be sexual dimorphism; the occurrence of morphological differences between individuals of different sex within a species, may affect several traits like body size, shape and sometimes, secondary sexual characteristics (Wells, 2007). Several factors can influence sexual dimorphism including female reproductive strategy, sexual selection, and competition for resources (Fathinia et al., 2012). Sexual dimorphism may have important consequences for animal ecology, and is a key aspect for understand the evolution of life history traits (Kupfer, 2007). In particular, sexual size dimorphism (SSD) is defined as the difference in body length or mass of sexually mature males and females (Fairbairn, 1997;Kupfer, 2007;Nali et al., 2014). Several evolutionary processes have been proposed to explain patterns of sexual dimorphism in anurans. On one hand, the usually biased SSD in favor of females (Shine, 1979) is hypothesized as the result of a fecundity advantage driven by natural selection: bigger females can harbor more eggs, and then produce larger clutches (Arak, 1988;Wells, 2007). Whereas in males, natural selection operates against of bigger body sizes, because higher vulnerability of prolonged breeders to predators increase their cost of reproduction in terms of survival at small body sizes (Camargo et al., 2008). On the other hand, some authors argue that sexual dimorphism is a consequence of sexual selection. In this sense, Darwin envisioned that sexual selection depends on the struggle between males to access females, and recognized two mechanisms: intrasexual selection, through competition between members of the same sex (usually males) for access to mates, where large males defeat small ones in aggressive encounters and displace them from territories; and intersexual (epigamic) selection, where members of one sex (usually females) choose members of the opposite sex, by comparing traits of potential mates and select those that are more attractive (Darwin, 1871;Shine, 1979;Woolbright, 1983;Arak, 1988;Lovich and Gibbons, 1992). However, some authors proposed that sexual dimorphism is a function of differences in life history strategies between the sexes, as well as the result of a variety of selective forces. In this sense, SSD can be explained in terms of disparate age structure between sexes in reproductive populations; that is, females were larger because they were older than the males, which mature earlier at smaller size. In fact, Monnet and Cherry (2002) found that age differences between breeding males and females appear to have a major influence on the extent of dimorphism. Female anuran fecundity appears to be correlated with body size in all anuran species in which this phenomenon has been investigated, and, as anurans display indeterminate growth (Halliday and Verrell, 1986), this could be expected to lead to faster growth rates and delayed reproduction in females (Monnet and Cherry, 2002).
Limnomedusa Fitzinger 1843, is the most basal genus within the family Alsodidae (Frost et al., 2006;Pyron and Wiens, 2011). The only species of the genus, "rapids frog" Limnomedusa macroglossa (Duméril and Bibron, 1841), is a generalist insect predator of medium to large size, with shades of brown-and-gray and conspicuous glands in the back, and an immaculate white belly (Maneyro and Carreira, 2012). As secondary sexual characteristics, males present a single vocal sac and dark nuptial pads on their fingers. It is a habitat specialist, with a restricted distribution in rocky outcrops of basaltic origin and superficial soils, with or without vegetation (Maneyro and Carreira, 2012). Regarding its geographic variation in Uruguay, larval dispersion appears to be connecting separate major basins via watercourses, although it is also likely that adults disperse between habitat patches by land. As a corollary, an isolation pattern by distance was established, which maintains population stability and genetic diversity in northern populations (Fernández, 2016). Recently, de Olivera et al. (2018) found a correlation between body size and ovarian mass in populations of Limnomedusa macroglossa from Uruguay, suggesting a fecundity advantage for larger females since they can accommodate larger ovaries. Moreover, they also reported a prolonged pattern of reproduction for this species, which is usually associated with higher levels of intra/ inter-sexual selection (Wells, 2007). Further, in populations from Rio Grande do Sul state, SSD has been found, where females attain larger SVL than males, and they also classified the pattern of reproduction as prolonged, although highly seasonal (Kaefer et al., 2009).
Its geographic distribution includes the south of Brazil (from Paraná to Rio Grande do Sul), the northeast of Argentina (Misiones and Entre Ríos), the southeast of Paraguay (Alto Paraná), and almost the entire Uruguayan territory (Frost, 2020). However, despite being a relatively common species, geographical variation in morphology has not been investigated in L. macroglossa overall distribution. This circumstance is relevant since most of the distribution range occurs in Uruguay, and thorough evaluation of the morphological variation across such distribution is necessary given its latitudinal, environmental gradient. Lastly, in reference to its conservation status, is categorized nationally and globally as Least Concern according to the IUCN criteria (Silvano et al., 2004;Maneyro et al., 2019).
The aim of this work was to evaluate the occurrence of sexual (size and shape) dimorphism in Limnomedusa macroglossa based on morphometrics and secondary sexual characteristics across populations from Uruguay.
We hypothesized that: (1) Sexual dimorphism and minimum size at sexual maturity (MSSM) are important life history traits due of their value in reproductive output of a species. Besides, most anuran females have larger body sizes than males (female biased SSD) and, thus, females usually reach sexual maturity at larger sizes.
(2) Isolation pattern by distance triggered by hydrographic basins favor geographical differences in morphology.
From which the following predictions emerge: (1.1) we expect that L. macroglossa present sexual size dimorphism with females larger than males, and in fact, with females reaching MSSM at bigger sizes than males.
(2.1) Finally, hydrographic basins, due to environmental differences, will favor a greater morphological differentiation in L. macroglossa between than within basins.
All collected individuals were euthanized using topic lidocaine and intraperitoneal injection of sodium pentobarbital (0.5 ml of a 0.2 g/ml solution), fixed with 10% formaline, and preserved in 70% ethanol, following the experimental protocol "Euthanasia method for amphibians and reptiles in the field" approved by the Institutional Animal Care and Use Committee (IACUC), Faculty of Science, University of the Republic. Individuals were euthanized with the purpose of being genetically studied by Fernández (2016), therefore in this work, those individuals were reused. All the specimens are stored in the Vertebrate Zoology Collection (ZVC-B) of the Faculty of Sciences, University of the Republic.
We measured eleven morphometric variables using a digital calliper (0.01 mm precision) by a single observer for consistency : snout-vent length (SVL), mandibular width (MW), head length (HL), inter-orbital distance (IOD), eye diameter (ED), inter-narial distance (IND), eye-nostril distance (END), arm length (ARML), tibia length (TiL), tarsus length (TaL) and metatarsus length (MtL). We followed the methodology of Duellman (1970) to obtain the measurements of SVL, IOD, ED, IND, TiL, and MtL, as well as Napoli (2005) for END, and Greene and Funk (2009) for ARML. Finally, we measured TaL as the straight length of the tarsus, MW as the straight line between oral commissures, and HL as the straight line distance from the posterior edge of the skull to the tip of the snout. All individuals were measured twice to ensure accuracy and all measurements were taken on the right side of the body (Fig. 2).
For each individual, sex and maturation status (juvenile/ adult) were determined by gonadal analysis. Additionally, males were considered mature by the presence of nuptial pads in their fingers. Finally, to infer MSSM on each sex, we pooled all individuals from all basins and register the size of the adult male/ female with the lowest SVL.

Data analysis
Using the morphometric variables, we tested for sexual dimorphism and quantified morphometric variation through univariate and multivariate analyses while taking into account geographic distribution.
To remove allometric effects of body size in the sexual dimorphism analyzes we applied the transformation proposed by Lleonart et al. (2000), which scales all individuals to same size and adjust their shape to that they would have in the new size.
For all the variables we tested the normality (Lilliefors' test) and homogeneity of the variance (Levene's test) of raw and transformed data. A priori, the raw data did not reject the hypotheses of normality neither homogeneity of variances. Although, with the transformed data, there were rejected. No outlier individuals were found in the analyzed sample.
We performed a t-test to evaluate for a significant difference in SVL between males and females. Sexes were also compared through one-way perMANOVA using Euclidean similarity index. As differences in body size between sexes are not always related to SVL and can involve body parts used in various behavioral contexts (Kupfer, 2007), we conducted multivariate analyses. Differences in shape between males and females were examined through a Principal Component Analysis (PCA) using the Variance-Covariance matrix, and a Hierarchical Cluster with Unweighted Pair-Group Average algorithm and Euclidean similarity index with 9999 pseudoreplicates.
SVL, MtL and TiL variables were log-transformed to estimate standardized mayor axis (SMA) regression slopes. This method estimates the line of best fit (slope) when both variables are measured with error . We estimated the SMA relationship between SVL and MtL/TiL. We tested for significant allometry assuming the null hypothesis that the slope was equal to 1 (isometry), performed slope comparisons between sexes, tested for shifts along the common SMA slope and in elevation of slope between sexes using Wald test, with 1000 iterations and critical P-value to 0.05.
To evaluate morphometric variation, we analyzed differences among basins by one-way perMANOVA (using Bonferroni correction for P-values) and PCA, based on raw measured variables, because our goal was also to evaluate the effect of body size, and we box plot SVL, TiL and MtL variables according to basins. In addition, we calculated the average leg length (= TiL + TaL + MtL) among individuals belonging to each basin. Finally, we tested for significant differences in the leg length and SVL among basins through t-test. The latitude vs. SVL relationship was evaluated through regression analysis (using Reduced Major Axis algorithm). In these analyses we used 35 mature males and 30 mature females because coordinate data were not available for all individuals.

Sexual dimorphism
In total, we examined 180 specimens of which 102 were juveniles, 34 mature females and 44 mature males. We found dark nuptial pads in the first, second, and sometimes, the third fingers of all mature males (Fig. 3). We found that females longer than 49.82 mm and males longer than 41.29 mm in SVL were sexually mature (i.e., nuptial pads in males and fully-developed oocytes in females). Taking this into account, we set the MSSM in females at 49.82 ± 0.01 mm and in males at 41.29 ± 0.01 mm.
Significant differences in mature body size were found between sexes. Mature females had an average SVL (56.99 ± 4.27 mm) significantly higher than that of mature males (49.69 ± 4.73 mm) (t = 7.04, P < 0.001).  Furthermore, significant differences were found in means of all other variables, with females reaching larger measurements than males (Table 1). When all the morphometric variables were introduced in a nonparametric per-MANOVA test, the comparison of sexes was not significant (F = 0.349, P = 0.865).
The PCA of size-adjusted measurements showed a total of ten components, with 69.09 % of the variance comprised by the first two components, with PC1 accounting for 54.14% and PC2 14.95% of the total variation. The bi-dimensional projection of the first two components exhibited a substantial overlap of sexes (Fig. 4A). The loadings indicate that PC1 has a strong positive correlation with TiL (0.61) and MtL (0.58), and the lowest with ED, IND, END and IOD (Fig. 4B), whereas PC2 is moderately correlated with TaL (0.35) and ARML (0.33), while MtL stood out with a very strong negative correlation of -0.76 (Fig. 4C).
The dendrogram obtained, with hierarchical clustering, was a good representation of the data matrix given the obtained coefficient of cophenetic correlation (CCC = 0.74) and showed a topology of females and males in congruence with PCA and perMANOVA analyses, determining the absence of morphometric, sexual shape dimorphism. However, the low bootstrap percentages (<50%) do not indicate high support for most of the similarity relationships.

Morphometric differentiation
Since multivariate analyses performed previously did not reveal significant differences between sexes, we pooled males and females within each basin in subsequent analyses. When all the morphometric variables were analyzed through perMANOVA test, significant differences were found among the six hydrographic basins evaluated in this work (F = 2.553, P < 0.05). The pairwise comparisons showed significant differences between Río Negro and Río de la Plata basins ( Table 2).
The PCA of original measured variables (including SVL variable), showed a total of eleven components, with 96.4% of the variance comprised by the first two components, with PC1 accounting for 93.96% and PC2 2.47% of the total variation. The bi-dimensional projection of PC1 vs. PC2 showed the convex polygons grouping indi- viduals from different basins with an elevated degree of overlap (Fig. 5A). According to Greene and Funk (2009), in PCA of morphological data, the first axis (PC1) is usually associated with size, and the remaining axes describe orthogonal axes of variation in shape. Indeed, we found that, the first axis has a strong positive correlation with body size and a moderate correlation with a few hind leg measurements: SVL (0.69), TiL (0.45) and MtL (0.33). Meanwhile, head measurements (IOD, ED, IND and END) showed the weakest correlation (Fig. 5B). The second axis (shape axis) has a strong positive correlation with MtL (0.56), and a moderate correlation with MW (0.32) and TiL (0.31), while SVL stood out with a strong negative correlation -0.67 (Fig. 5C).
Given the considerable contribution of SVL, TiL and MtL variables in size and shape axis of PCA, we did a box plot according to the hydrographic basins in order to show the differences between them. Significant differences were found between hind leg length of individuals from Río de la Plata and Río Negro basins (t = 3.533, P < 0.001), being those of Río de la Plata basin the longest hind leg (82.12 ± 2.21 mm, n = 12), while those of Río Negro basin were the shortest legs (72.55 ± 1.55 mm, n = 25) (Fig. 6A,B), reaching a difference of 11,65%. Regarding SVL, we found a similar pattern, reaching higher values in Río de la Plata basin (56.31 ± 1.57 mm, n = 12) and lower ones in Río Negro basin (50.12 ± 1.06 mm, n = 25; t = 3.302, P < 0.002; Fig. 7A), reaching a difference   of 11% between groups. Additionally, a significant correlation between latitude and SVL was found (r = 0.60, F = 4.35, P < 0.001; Fig. 7B).

Allometric regressions
We performed a SMA analysis with the variables that showed the highest correlations with PC1 and PC2 in PC analyzes. We found a significant SMA relationship between SVL and MtL in females [b=0.98, 95% confidence interval (CI) = 0.  (Table 3).
In all cases, the test for common slope across sexes indicated that there are no significant differences in common slope between males and females. When testing for shifts along the common slope, we found significant shifts in all relationships with higher values in females (SVL-MtL relationship: W = 42.952, P < 0.01; SVL-TiL relationship: W = 47.729, p<0.01). The test for shift in elevation was only significant in the SVL vs. MtL relationship in favor of males (W = 4.411, P < 0.05), but the difference in elevation was rather small and close to our resolution limit (0.01 mm).

DISCUSSION
In this study we determined the minimum size at sexual maturity (MSSM) and described morphometric and intersexual differences in Limnomedusa macroglossa. We showed that females and males differ in MSSM, presence of dark nuptial pads in males (a sexually dimorphic characteristic) and body size, while no differences were found in body shape. Nuptial pads can be observed during the breeding season in response to increases in circulating levels of androgens, but later regress during the non-breeding period, although without resembling to a female-like morphology (Wells, 2007). Some authors argued that well-developed nuptial pads are associated with breeding in water to prevent the female´s escape during amplexus (Duellman and Trueb, 1986). However, according to Savage (1961), nuptial pads also allow the male to hold the female while defending her against rival males.
The MSSM is the size at which an individual has all the morphological and physiological conditions to begin to breed, and along with sexual dimorphism, are important life history traits to understand population changes through time. Life history theory explains the variation in MSSM between sexes through natural selection mechanisms, mainly related with adult mortality rates (Tolosa et al., 2014). We found that females of Limnomedusa macroglossa reach sexual maturity around 49.82 mm, while males reach it at a smaller size of about 41.29 mm. This difference between the sexes can be explained by sexual selection: selection for mating effort in males to defend territories, in detriment of larger males due to the high energetic expenditures and risks of mortality, and parental effort in females to produce more eggs to maximize their reproductive output, which favors females with a larger size; both processes have been pointed out as potential explanations for sexual maturation at different ages (Howard, 1981).
There was sexual dimorphism in size in Limnomedusa macroglossa with females being larger than males, as it occurs in more than 90% of anurans species (Shine, 1979). These results agree with those found in a population of L. macroglossa in southern Brazil based on SVL only (Kaefer et al., 2009). Taking into account the main hypotheses regarding the causes of sexual dimorphism in anurans, natural and/or sexual selection might adequately explain the differences in body sizes between females and males found in this work. Given the available data until date (Kaefer et al., 2009;de Olivera et al., 2018) and our results, it seems that the preference for larger females evolved because they produce more oocytes per clutch (Bionda et al., 2011) or bigger eggs (the fecundity advantage hypothesis), whereas in males, natural selection works against of bigger body sizes due to the existence of possible differential predation, since the long reproductive period exposes and makes them more vulnerable to predators (Camargo et al., 2008). Furthermore, intra/intersexual selection could be playing an important role in the differentiation between males and females, through malemale competition and/or female choice (Darwin, 1871;Shine, 1979;Woolbright, 1983;Arak, 1988). Although, in our field work, we did not observe such behaviors, we cannot rule out their existence, since it has been reported that its a species with a prolonged reproduction pattern (Kaefer et al., 2009;de Olivera et al., 2018) which is usually associated with more territorial males, choosy females, and overall higher levels of sexual selection (Wells, 2007). Finally, the age structure in the reproductive populations may also be operating between sexes (Halliday and Verrell, 1986;Monnet and Cherry, 2002). Thus, the sexual dimorphism in size found in L. macroglossa could be the result of distinct, possibly opposing, selective forces that trade-off differently in each sex.
In addition to size, anurans exhibit other forms of sexual dimorphism, including: the proportions and muscular development of the forelimbs (related with clasping behavior), skin color, texture and glands (visual, tactile and chemical cues for sex recognition), fangs and tusks (related with combat), abdominal and laryngeal muscles, and lung capacities (calling behavior) (Wells, 2007;Bell and Zamudio, 2012) and head morphology (feeding strategies) (Khoshnamvand et al., 2018).
No differences were found in shape between sexes, but significant differences were found among basins. Some variables related with the hind legs showed the highest contributions to overall shape differentiation. A functional interpretation of the differentiation in the hindlimb length found in L. macroglossa could be that  zaola and Laurila, 2009;Drakulic et al., 2016;Zamora-Camacho, 2018;Zamora-Camacho and Aragón, 2019), as well as jumping distance increases as the individual grows larger (Zug, 1978). Meanwhile, other studies revealed that locomotor performance is negatively affected at larger sizes (Moreno-Rueda et al., 2020), relatively large differences (>10%) in leg length can affect the jumping efficiency (Emerson, 1978;Babik and Rafinski, 2000). Differences in jumping ability could be occurring in L. macroglossa because our results showed differences greater than 10% in body size and leg length in individuals from Río de la Plata basin compared to those from the Río Negro basin. Alternatively, differences in the hindlimb length may be the result of unequal growth and developmental rate during the larval and juvenile stages (Emerson et al., 1988;Babik and Rafinski, 2000). Because amphibians are ectotherms and depend on water, they show phenotypic responses to changes in environmental factors. In this sense, some phenotypic plasticity can be attributed to environmental factors such as the duration of the larval period and its relation to size as a function of temperature (Vences et al., 2002). A general Temperature-Size rule for ectotherms states that higher temperatures increase developmental rates, at the cost of smaller size (Drakulic et al., 2016) and conversely, at low temperatures develop more slowly, so they metamorphose at larger body sizes (Harkey and Semlitsch, 1988). Moreover, some studies replace the idea of temperature and relate body size to latitude, predicting that body size within species increases with latitude (Lindsey, 1966;Schäuble, 2004). In this study we found that individuals which had the longest legs were from Río de la Plata basin, which correlates with the colder climate in the studied distribution (InUMet, 2020). On the other hand, the individuals which had shortest legs were found in Río Negro basin, where the temperature is significantly higher (InUMet, 2020). So, we can expect that differences in environmental temperature during the larval period may have been responsible for the variation in the hindlimbs length in L. macroglossa. This trend has already been reported in other studies (Atkinson, 1994(Atkinson, , 1995Angilletta et al., 2004). Furthermore, our results are consistent with the intraspecific version of Bergmann's rule. It relates to geographic variation in the body sizes of animals (Blackburn et al., 1999) which has been briefly stated by Mayr as: The smaller-sized geographic races of a species are found in the warmer parts of the range, the larger-sized races in the cooler districts (Ray, 1960).
In this study we report a clinal variation in the relative leg length and body size of Limnomedusa macroglossa along a latitudinal gradient in Uruguay. The body size dimorphism likely reflects differences in growth rates of males and females. In organisms with indeterminate growth, body size is a result of a trade-off between growth and reproduction (Camargo et al., 2008). Therefore, females of Limnomedusa macroglossa appear to delay sexual maturity, while maintaining the same body shape and proportions as the males, reaching larger sizes at maturity, based on the combined evidence of distinct MSSM and the body size shift along the common isometric slopes of males and females. This difference in size could be adaptive for females, since a larger body size would increase the volume of the abdominal cavity, being able to accommodate larger ovaries (de Olivera et al., 2018) and consequently, increasing their reproductive output [the so called fecundity advantage hypothesis (Darwin, 1871)]. Therefore, sexual dimorphism in L. macroglossa could be determined by differential growth rate between the sexes, since the growth rates are usually asymptotic after maturation and sexes generally mature at different ages (sexual bimaturity) (Monnet and Cherry, 2002;Kupfer, 2007;Wells, 2007), or it may be the result of difference in the age distributions of males and females (Howard, 1981). Therefore, the sexual dimorphism found in body size is probably the consequence of higher growth rates and/or late sexual maturity in females of Limnomedusa macroglossa, which favors a larger body size and larger ovaries, and consequently, higher reproductive output.

CONCLUSIONS
In conclusion, our data on MSSM and SSD of Limnomedusa macroglossa from Uruguay may contribute to the knowledge of the life history traits of this species.
Our results show that females attained sexual maturity at larger sizes than males with a marked female biased sexual size dimorphism. These traits are driven by a tradeoff between natural and sexual selection on each sex: parental effort in females does favor larger sizes to maximize their reproductive output, because bigger females can accommodate more eggs in their abdominal cavity. Meanwhile, mating effort in males does not favor large sizes due to energetic expenditures and risk of mortality during the long breading season, because bigger males invest most of their energy in search and calling behavior and have high mortality rates due to predation risk.
We also report a clinal variation in the relative leg length and body size of Limnomedusa macroglossa along a latitudinal gradient in Uruguay. Individuals with longest legs and bigger body sizes were from Río de la Plata basin, meanwhile individuals with shortest legs and smaller body sizes were those found in Río Negro basin. These differences could be explained by phenotypic plasticity in age and size at metamorphosis when separate populations are exposed to different environmental conditions (Ruthsatz et al., 2018). Studies demonstrated a plastic response of metamorphic traits in anuran larvae to changes in environmental conditions such as temperature. With increasing temperature time to metamorphosis may be reduced and metamorphosis occurs at smaller body sizes (Vences et al., 2002). Then, this may be occurring in L. macroglossa, since Río de la Plata basin is correlated with the colder climate in the studied distribution, meanwhile Río Negro basin is correlated with a warmer one.
All the evidence gathered in this work and its interpretations show that sexual dimorphism found in body size is likely the consequence of higher growth rates and/or late sexual maturity in females of Limnomedusa macroglossa, which favors a larger body size and bigger ovaries, and consequently, higher reproductive output. Examination of adult females and males, already in progress, will soon allow a more in depth understanding of L. macroglossa reproductive biology in Uruguay.

APPENDIX 2
Thirty four mature females (♀) of Limnomedusa macroglossa used in the analyses and their respective morphometric measurements, basin and latitude/longitude from Uruguay. ZVC-B: vertebrate collection of the Faculty of Sciences, University of the Republic. SVL: snout-vent length; MW: mandibular width; HL: head length; IOD: inter-orbital distance; ED: eye diameter; IND: inter-narial distance; END: eye-nostril distance; ARML: arm length; TiL: tibia length; TaL