Introduction

Numerous studies on anuran anomalies have reported a higher prevalence rate in hindlimbs (e.g. Rebouças et al. 2019, Meteyer et al. 2000) or forelimbs (e.g. Kaiser 1997, Johnson et al. 2002) compared to the axial skeleton. Ouellet et al. (1997) suggested that anuran hindlimbs have higher anomaly prevalence rates as they are more susceptible to environmental influences during development and, consequently, potentially more affected by exogenous factors. This greater susceptibility arises because anurans usually have free-living larvae, with hindlimbs developing externally and exposed to environmental impacts, while forelimbs develop hidden under the skin, emerging only at the climax of metamorphosis (Gosner 1960, Duellman & Trueb 1986).

However, anurans exhibit a remarkable variety of reproductive modes, displaying differences in oviposition site, developmental mode, and other attributes (Duellman & Trueb 1986, Haddad & Prado 2005, Pombal & Haddad 2007). The most common, and presumably basal, reproductive mode in anurans is characterised by aquatic eggs and exotrophic tadpoles that undergo extensive modification during metamorphosis (Duellman & Trueb 1986). In this reproductive mode, the forelimbs develop exposed to the environment, making them vulnerable to the action of chemical agents and physical or biological factors. Many other anuran species, however, lack a larval stage or have tadpoles that develop with no direct contact with aquatic environments (Duellman & Trueb 1986, Haddad & Prado 2005, Pombal & Haddad 2007). Thus, species will be differentially exposed to the influence of the environment on the development of hindlimbs, according to reproductive mode.

In this study, we compare five species to investigate whether anomalies vary according to developmental mode and whether hindlimbs were more affected by anomalies due to the development process of anuran larvae, as suggested by Ouellet et al. (1997).

Material and Methods

For this study, we used data on osteological anomalies in anuran species, described and illustrated by Baroni et al. (2024). In each case, the specimens were sourced from urban forests within conservation units in Rio de Janeiro, State of Rio de Janeiro, encompassing the south-eastern Brazilian Atlantic Forest domain. These specimens were previously deposited in herpetological collections, with records spanning from 1940 to 2019.

Fig.1.(A)

Aplastodiscus albofrenatus (Hylidae), photo J.P. Pombal, Jr.;(B)Hylodesnasus (Hylodidae), photo J.P. Pombal, Jr.;(C) Adenomera marmorata (Leptodactylidae), photo P.H. Pinna; (D) Brachycephalus ephippium and (E) Ischnocnema guentheri (Brachycephalidae), photos P.H. Pinna. All specimens are from Parque Nacional da Floresta da Tijuca, an urban forest in Rio de Janeiro, Southeastern Brazil.

The dataset comprised five anuran species (Fig. 1) from four families, with reproductive modes defined according to Haddad & Prado (2005), i.e. Aplastodiscus albofrenatus (Hylidae; mode 5 – eggs deposited in nests built underground, exotrophic tadpoles developing in the aquatic environment) and Hylodes nasus (Hylodidae; mode 3 - eggs deposited in underwater chambers, exotrophic tadpoles developing in the aquatic environment); Adenomera marmorata (Leptodactylidae, mode 32 – a nidicolous species with eggs deposited in foam nests within chambers built underground, where endotrophic tadpoles also develop); and Brachycephalus ephippium and Ischnocnema guentheri (both Brachycephalidae, mode 23 – terrestrial eggs and direct development).

Here, to test the hypothesis of Ouellet et al. (1997) that hindlimbs would be more affected by anomalies due to the development process of anuran larvae, we assess whether the prevalence of anomalies was significantly different between species, considering reproductive mode and presence of anomalies. Owing to the low frequency of anomalies, we utilised Fisher's exact test for statistical analysis as it is more suitable for this type of comparison (Siegel 1975). Analyses were performed using R 4.2.1. software (R Core Team 2023), with a P-value < 0.05 considered as statistically significant.

Table 1.

Anomalies in anurans with different developmental modes from forests in the municipality of Rio de Janeiro, southeastern Brazil; summary based on descriptive osteological data from Baroni et al. (2024).

We conducted four sets of comparisons considering different types of anomalies and distinct clusters of species based on their reproductive mode. Initially, we categorised the species into three groups: 1) those with exotrophic tadpoles, 2) those with endotrophic tadpoles, and 3) those with direct development. One set of tests compared these categories regarding the frequency of anomalies, specifically in the hindlimbs, while another set considered all anomalies, regardless of the region of the body in which they were observed. These two datasets (anomalies limited to hindlimbs and total anomalies) were used to compare two reproductive mode categories: 1) species with exotrophic tadpoles developing in aquatic environments (reproductive modes 3 and 5); and 2) species with endotrophic tadpoles developing outside aquatic environments, as well as those with direct development (reproductive modes 32 and 23).

Results

From the final dataset of 622 specimens, anomalies were recorded in 20 individuals, resulting in a prevalence rate of 3.22% (Table 1). In total, we examined 99 specimens of A. albofrenatus (six anomalous, prevalence rate of 6.06%), 171 specimens of H. nasus (eight anomalous, 4.68%), 160 specimens of A. marmorata (three anomalous, 1.88%), 90 specimens of B. ephippium (two anomalous, 2.22%), and 102 specimens of I. guentheri (one anomalous, 0.98%).

Of the total recorded anomalies, 11 were in the vertebral column, mainly in the sacral region (n = 8), representing a prevalence rate of 1.77% (four in A. albofrenatus – vertebral column anomaly prevalence rate 4.04%; three in H. nasus – 1.75%; two in A. marmorata – 1.25%; one in B. ephippium – 1.11%; and one in I. guentheri – 0.98%). Only two specimens had anomalies in the forelimbs, representing just 0.32% of the total prevalence rate. Both affected individuals were H. nasus, resulting in a forelimb anomaly prevalence rate of 1.17% for this species. A single pelvic girdle anomaly was recorded in A. albofrenatus, representing 0.16% of all specimens analysed, a prevalence rate of 1.01% for the species. Regarding hindlimbs, six specimens had anomalies (prevalence rate 0.96%): one specimen of A. albofrenatus (1.01% prevalence rate), three H. nasus (1.75%), one A. marmorata (0.63%), and one B. ephippium (1.11%). No anomalies were recorded in the hindlimbs of I. guentheri.

We found no significant differences in the prevalence of anomalies in hindlimbs when comparing the three reproductive mode categories (P = 0.5839), i.e. exotrophic tadpoles (A. albofrenatus plus H. nasus), endotrophic tadpoles (A. marmorata), and direct developers (B. ephippium plus I. guentheri); nor when considering the two reproductive mode categories (P = 0.4115), i.e. exotrophic tadpoles in aquatic environments (A. albofrenatus and H. nasus) and absence of tadpoles in aquatic environments (summing endotrophic tadpoles and direct developers: A. marmorata, B. ephippium, and I. guentheri). Regarding overall anomalous specimens, we also found no significant difference when comparing the three reproductive mode categories (P = 0.06759); however, when species were distributed into the two reproductive mode categories (exotrophic tadpoles in aquatic environment and absence of tadpoles in aquatic environments), there was a significant difference in overall anomaly prevalence (P = 0.02036).

Discussion

Descriptions and illustrations (x-rays) of all anomalies found in the specimens are available in Baroni et al. (2024). The overall prevalence rate of anomalies was relatively low (3.22%), consistent with populations inhabiting areas with no significant environmental impacts (see also the discussion in Baroni et al. 2024). Ouellet (2000) and Lunde & Johnson (2012) both suggest that a prevalence rate of anomalous individuals between 0-5% would be expected for amphibians in natural populations. Only A. albofrenatus, with 6% anomalous individuals, exceeded these reference values, though just barely. In environments under intense anthropic pressure, such as agricultural areas affected by pesticides, the prevalence of anomalies in anurans is usually much higher and can exceed 60% (Ouellet et al. 1997). As the specimens in this study came from conservation units located within a large metropolitan region, the possibility of anthropic impacts acting on these populations cannot be completely ruled out (Netto 2005). Thus, the low prevalence rate recorded could be an indication of satisfactory environmental quality conditions, despite anthropic pressure in the surrounding areas.

Anomalies may be linked to genetic and environmental factors caused by infectious agents, exposure to contaminants, and other factors yet unknown (Gutiérrez 2018). Morphological analysis alone cannot precisely determine why an anomaly developed (Meteyer et al. 2000, Lunde & Johnson 2012, Henle et al. 2017). However, some specific types of anomaly, including severe morphological changes in the hindlimbs such as polymelia, are often related to parasitism by the trematode Ribeiroia ondatrae (Johnson et al. 2002, Stopper et al. 2002). Studies that exposed larvae to different pesticides under laboratory conditions resulted in severe developmental malformations (e.g. Yoon & Ko 2019, Ko 2020). However, even in environments with minimal anthropic impact, ecological interactions, such as predation or parasitism during the development stages, have been shown to interfere with the frequency of anomalies recorded (Niazi & Ratnasamy 1984). Therefore, it is possible that even in areasnotsubjectedtoextremeenvironmentalchanges, development in aquatic environments favours the exposure of larvae to impacts and interactions that result in differences in anomaly rates related to reproductive mode, making forms of development with less exposure to the aquatic environment more favourable, a possible benefit of derived reproductive modes in relation to the frequency of anomalies in general.

On the other hand, a lower anomaly prevalence rate was expected in this study as the data was focused exclusively on fully metamorphosed specimens (generally adults). When comparing different studies on this topic, it is necessary to consider that life stage is a determining factor in the prevalence of anomalies, and a higher prevalence of anomalous individuals will be expected if larval and newly metamorphosed stages are analysed, compared to the adult stage. Owing to differences in survival rate, the proportion of anomalous individuals tends to progressively decrease over the species' lifetime (Johnson et al. 2002). Lunde & Johnson (2012) proposed that the analysis of anomalies should focus on individuals in the metamorphic stage, shortly before or immediately after metamorphosis, to maximise the recording of anomalies. This approach allows for recording anomalies critical to survival, providing a more accurate survey of anomalies during development. However, it greatly reduces the possibility of analysing internal anomalies, which accounted for half of the records in this study. Nonetheless, in populations with a low prevalence of anomalies, less disabling cases, restricted to parts of the appendages or digits, are the most common (Johnson et al. 2002). Here, we deal with extensive anomalies, such as partial or total absence of hindlimbs (see Figs. 6 and 7 in Baroni et al. 2024), indicating that the possible bias from investigating adult animals may not have been so significant. Moreover, it is also possible that viability of survival, despite the extensive anomaly, may have influenced the most recorded type of sacral anomaly. For example, in all cases in which the sacral diaphysis supported only one side of the pelvic girdle, it appears to remain functional as support on the opposite side was provided by the last presacral vertebrae, with resulting asymmetry rebalanced by the differential length of the ilium (see Fig. 2 of Baroni et al. 2024).

Our analysis does not support the hypothesis previously suggested in the literature (Ouellet et al. 1997) that anurans have higher rates of anomalies in the legs due to their exposure to the environment during development, i.e. leg anomaly prevalence was not significantly different between exotrophic and endotrophic reproductive modes. Additionally, internal axial anomaly prevalence levels were similar to those in the legs. However, there were differences in the overall anomaly prevalence rate according to reproductive mode, suggesting that development in aquatic environments could result in a higher anomaly prevalence rate. Species with exotrophic tadpoles had significantly higher frequencies of overall anomalies than species that developed outside the aquatic environment, including nidicolous species with endotrophic tadpoles and direct developers. These results suggest that reproductive mode may interfere with the occurrence of overall anomalies, and not just anomalies of pelvic appendages (hindlimbs).