Elachistocleis, Parker, 1927

Novaes-E-Fagundes, Gabriel, Lyra, Mariana L., Loredam, Vinicius S. A., Carvalho, Thiago R., Haddad, Célio F. B., Rodrigues, Miguel T., Baldo, Diego, Barrasso, Diego A., Loebmann, Daniel, Ávila, Robson W., Brusquetti, Francisco, Prudente, Ana L. C., Wheeler, Ward C., Orrico, Victor Goyannes Dill & Peloso, Pedro, 2023, A tale of two bellies: systematics of the oval frogs (Anura: Microhylidae: Elachistocleis), Zoological Journal of the Linnean Society 197, pp. 545-568 : 554-555

TAXONOMY OF ELACHISTOCLEIS – A COMPLEX MATTER

Parker (1927) erected Elachistocleis to allocate Rana ovalis Schneider, 1799 and Engystoma ovale bicolor ( Guérin-Méneville, 1838) . As usual at the time, the descriptions of both species were brief and did not designate type specimens. Lavilla et al. (2003) suggested that these names apply to complexes of species, because species of this genus are morphologically similar, with few obvious external characters that can be used as reliable diagnostic characters (i.e. non-overlapping and fixed). Ultimately, as mentioned throughout our contribution, diagnoses in this genus rely heavily on the ventral colour pattern, head proportions and the presence/absence and extension of body lines, such as a mid-dorsal white line and lines on the hidden surface of hindlimbs ( Parker, 1927; Lavilla et al., 2003; Caramaschi, 2010; Pereyra et al., 2013).

Caramaschi (2010) and Piva et al. (2017) suggested the ventral colour pattern as a diagnostic character for two internal species groups: a group with immaculate bellies (uniformly clear, free of markings) and another with maculate bellies. We found that the groups proposed by the above-cited authors are not monophyletic. For example, Elachistocleis muiraquitan (immaculate-bellied) is not related to the remaining immaculate-bellied species; instead, it was recovered in the E. surinamensis species group, a clade composed of species with predominantly maculate bellies. Furthermore, whereas E. panamensis has a maculate belly ( Dunn et al., 1948; Nelson, 1972), E. araios has an immaculate belly ( Sánchez-Nivicela et al., 2020). Given their phylogenetic position as successive sistertaxa to the remaining species of the genus (but see our comments on the position of E. araios ), the description of E. araios already rendered the immaculate-bellied species group as paraphyletic. Also, it implies that the optimization of the ventral pattern of the Elachistocleis ancestor is ambiguous, given our topology and that of Sánchez-Nivicela et al. (2020). More importantly, we found intraspecific variation of the ventral pattern in some species, such as E. bicolor and E. nigrogularis ( Figs 1 , 4 ). Ontogenetic variation of the ventral pattern has also been recently reported for E. haroi ( Bueno-Villafañe et al., 2020) . Therefore, the available phylogenetic and morphological evidence emphatically rejects the existence of groups diagnosed solely by the colour of their bellies.

It is noteworthy that this discussion seems far from settled. Ventral patterns have been recorded idiosyncratically in the literature. Sánchez-Nivicela et al. (2020) cite E. cesarii as ‘having uniform, immaculate, ventral colouration’, although Toledo et al. (2010) when resurrecting the species clearly state that it has ‘ventral colouration white or yellow with grey marks and reticulations’, and their figure 3d depicts a specimen with a maculate belly. Jowers et al. (2021) state that E. cesarii has a ‘uniform’ belly. This same was said of E. erythrogaster and E. bicolor . While E. cesarii and E. erythrogaster are maculate (the latter comparatively less maculate), E. bicolor bellies usually present no markings whatsoever (but see discussion on E. bicolor below). Given the historical use of the ventral pattern in the taxonomy of Elachistocleis , we attempted to verify as many specimens as possible for this trait. We have not extensively checked our vouchers for other morphological diagnostic traits (male throat colour; presence and shape of the femoral stripe; post-commissural gland). Nevertheless, our preliminary data suggest that none of the above-cited morphological traits is as reliable as suggested in the literature [see also comments in Nunes et al. (2010) and Marinho et al. (2018)].

Data from other sources are scarce. Tadpoles of few species are described (e.g. Rossa-Feres & Nomura, 2006; Pereyra et al., 2013; Schulze et al., 2015; Ferreira & Weber, 2021), but it is difficult to confidently assign these descriptions to lineages recovered because most lack an association with available DNA sequences. Moreover, there is no comprehensive understanding of variation, although Gómez & Kehr (2012) reported some variability in larval morphology related to chemical cues in the presence of predators. Advertisement calls of several Elachistocleis have been described (e.g. Nelson, 1972; Duellman, 1997; Kwet & Di-Bernardo, 1998; Lavilla et al., 2003; Nunes et al., 2010; Toledo, 2010; Marinho et al., 2018; Pansonato et al., 2018; Jowers et al., 2021). Some of these are associated with vouchers present in our phylogeny (see: Marinho et al., 2018; Jowers et al., 2021). These are not adequately distributed in our topology and most are described from a few specimens and localities. Published data have already shown that even slight increases in sample size may affect recorded variation (see: Marinho et al., 2018). Finally, some localities present at least two sympatric species of Elachistocleis . For instance, both E. nigrogularis and E. surinamensis are syntopically found at the type locality of E. nigrogularis ( Jowers et al., 2021) , and both E. bicolor and E. erythrogaster are found at the type locality of E. erythrogaster ( Kwet & Di-Bernardo, 1998) (also see Fig. 3 ). This makes it even more difficult to associate larvae with adults and advertisement calls of unvouchered specimens to named species without molecular data.

A fine line

A white or light yellow mid-dorsal line is present in many individuals of Elachistocleis , and other genera of Gastrophryninae, such as Chiasmocleis Méhelÿ, 1904 (e.g. Peloso et al., 2014), Ctenophryne Mocquard, 1904 (e.g. Duellman, 1978), Hamptophryne A.L. Carvalho, 1954 (e.g. Parker, 1927; Duellman, 1978), Hypopachus [e.g. Cope (1889); see also figure 5 in Greenbaum et al. (2011)], Dasypops Miranda-Ribeiro, 1924 and Stereocyclops Cope, 1870 (PP, personal observation). Many authors have noticed important intraspecific or intrapopulational variation (i.e. being present in some individuals and absent in others) in this character in other genera (e.g. Chiasmocleis, Peloso et al., 2014 ; Ctenophryne, Zweifel & Myers, 1989 ; Stereocyclops, GNF , personal observation) and in Elachistocleis (e.g. Nelson, 1972; Toledo, 2010; Toledo et al., 2010; Marinho et al., 2018). Notwithstanding, the occurrence of this line and the variation of its extension when present (e.g. from snout to vent or from post-cephalic fold to vent) were used in diagnoses of several Elachistocleis species (e.g. Caramaschi & Jim, 1983; Caramaschi, 2010; Nunes-de-Almeida & Toledo, 2012; Piva et al., 2017).

We found a remarkable variation in the occurrence of the mid-dorsal white line in and among species ( Fig. 1 ), similar to the variation reported for Chiasmocleis ( Peloso et al., 2014) . This variation challenges the reliability of the trait as a diagnostic feature and weakens the known diagnoses of several currently recognized species (see below). It was impossible to confidently evaluate the extension of the mid-dorsal line, thus we only scored the presence or the absence of the line. Hence, we recommend caution when using the mid-dorsal line in the systematics of Elachistocleis , especially for diagnostic purposes.

Figures

Figure 1. Species delimitation in Elachistocleis. The tree shown is from the maximum likelihood (ML) analysis with only 16S + COI from only Elachistocleis samples (tree rooted in E. araios). Numbers on branches are SH-aLRT/UFBoot support values. Asterisks represent support above 80/95. The terminals of the tree are named as: original labelling of

Figure 3. Geographic location of sampled specimens (symbols) and type-localities of the nominal species (numbers). 1, Elachistocleis araios; 2, E. bicolor (here illustrated as the city of Buenos Aires; see Discussion); 3, E. bumbameuboi; 4, E. carƲalhoi; 5, E. cesarii; 6, E. corumbaensis; 7, E. erythrogaster; 8, E. haroi; 9, E. helianneae; 10, E. magna; 11, E. matogrosso; 12, E. muiraquitan; 13, E. nigrogularis; 14, E. panamensis; 15, E. pearsei; 16, E. piauiensis; 17, E. sikuani; 18, E. skotogaster; 19, E. surinamensis; 20, E. surumu; 21, E. tinigua.

Figure 4. Variation of ventral colour pattern among sample vouchers of different species of Elachistocleis of the bicolor group. A, B, E. helianneae; MPEG 29416, 29419. C, E. haroi; IIBPH 2947. D–F, E. bicolor; CFBH 10117, AAGUFU 5953, IIBPH 0285. G, H, E. nigrogularis; CAS 231811, 245101.