Leptodactylus wagneri (Peters 1862)

Leptodactylus wagneri (Peters, 1862)

Plectromantis wagneri Peters, 1862:232 [type locality: “der Westseite der Anden in Ecuador (Peters, 1862:233)” in error, catalog entry for the holotype “Pastassa;” holotype destroyed, formerly in ZSM collection, Munich].

Leptodactylus wagneri.—Nieden, 1923:479 [first apparent association of wagneri with Leptodactylus].

DIAGNOSIS.—Leptodactylus wagneri occurs along the amazonian flanks of the Andes and is known to occur with or in the same general region as the following Leptodactylus species with toe fringes (excepting L. ocellatus and its closest relatives): L. bolivianus, colombiensis, dantasi, diedrus, griseigularis, leptodactyloides, pascoensis, and petersii. Leptodactylus wagneri does not reach the same size as bolivianus (wagneri females 52–82 mm SVL, males 39–61 mm SVL; bolivianus females to 88 mm SVL, males to 94 mm SVL). Few L. wagneri have distinct posterior lip stripes, and all discernible lip stripes in wagneri extend from the posterior corner of the eye; many bolivianus have light stripes on the entire upper lip including under the eye. Leptodactylus wagneri is larger than colombiensis (colombiensis females 38–62 mm SVL, males 36–56 mm SVL), and most wagneri have long dorsolateral folds whereas only some colombiensis do. Leptodactylus wagneri do not have distinct light belly spots; L. dantasi does. Leptodactylus wagneri is larger than diedrus (diedrus females 34–48 mm SVL, males 30–40 mm SVL), and the ventral and posterior thigh patterns merge in wagneri whereas the patterns abut in diedrus. Leptodactylus wagneri is larger than griseigularis (griseigularis females 39–58 mm SVL, males 35–51 mm SVL). The most common condition for dorsolateral folds in wagneri is long, and the most common belly pattern is moderately mottled. The most common fold condition in griseigularis is moderate length, and the most common belly pattern is the lightly mottled state. Leptodactylus wagneri is larger than leptodactyloides (leptodactyloides females 35–56 mm SVL, males 28–48 mm SVL). Most L. wagneri have long dorsolateral folds; few L. leptodactyloides have long folds. Many L. wagneri have boldly mottled bellies, whereas the bellies of leptodactyloides characteristically are finely mottled. In addition to the bold pattern, the bellies of most L. wagneri are moderately mottled and some are extensively mottled; the bellies of pascoensis are lightly to moderately, but never boldly, mottled. Characteristically, L. wagneri is larger than petersii (petersii females 31–51 mm SVL, males 27–41 mm SVL), and no petersii have long dorsolateral folds.

ADULT CHARACTERISTICS (N = 82 females, 110 males).—Light posterior lip stripes distinct (27%), indistinct (54%), or indiscernible (19%), when discernible, stripe extending from posterior corner of eye; light posterior thigh stripes rarely very distinct (1%), often distinct (32%), indistinct (28%), or indiscernible (39%); dorsolateral folds rarely (apparently) absent (2%) or short (1%), often medium length (21%), usually long (75%); toe tips rarely narrow (2%), usually either just swollen (43%) or swollen (54%), rarely just expanded (1%); male black thumb spines rarely small size (1%) or medium–small (5%), usually medium (68%), sometimes medium–large (18%), occasionally large (8%); chin/throat sometimes with light spots on a darker background (27%); belly rarely lightly mottled (3%), usually either moderately mottled (49%) or extensively mottled (48%), mottle usually in a strikingly bold pattern (57%).

Females 52.3–81.7 mm SVL (x =65.5 ± 6.5), males 39.1–60.7 mm SVL (x = 51.6 ± 4.7); female head length 32%–40% SVL (x = 36 ± 2%), male head length 34%–41% SVL (x = 37 ± 1%); female head width 31%–37% SVL (x = 34 ± 1%), male head width 33%–38% SVL (x = 36 ± 1%); female tympanum diameter 6%–8% SVL (x = 7 ± 0%), male tympanum diameter 6%–9% SVL (x = 8 ± 0%); female thigh length 39%–51% SVL (x = 45 ± 3%), male thigh length 40%–50% SVL (x = 45 ± 2%); female shank length 45%–55% SVL (x = 51 ± 2%), male shank length 46%–54% SVL (x = 55 ± 2%); female foot length 47%–60% SVL (x = 55 ± 3%), male foot length 49%–61% SVL (x = 55 ± 2%).

LARVAL CHARACTERISTICS.—Unknown.

ADVERTISEMENT CALL.—Unknown.

KARYOTYPE.—Unknown.

DISTRIBUTION.—Most specimens are from the amazonian slopes of the Andes in southern Colombia, Ecuador, and northern Peru; there are a few records from lowland Amazon localities in Ecuador, Peru, and Colombia, and a single specimen is known from the low Pacific slopes of the Andes in Colombia (Figure 45; Appendix 2).

Distributions

Understanding the distribution patterns of the podicipinus-wagneri complex is still at an early stage. This is due in part to unresolved systematic problems and to undercollected regions where members of the complex occur. Until the status of the Venezuelan Andes OTUs are resolved, the distributions of L. colombiensis, leptodactyloides, and sabanensis can not be understood adequately. Similarly, if the geographic samples currently included in L. griseigularis prove to contain more than one species, the distribution patterns involved will change significantly. Additional collecting efforts could change the distribution of L. diedrus as documented in this paper (Figure 36). There are two species for which the systematic understanding and collecting efforts are adequate such that additional data are unlikely to change the distribution patterns described in this paper: L. natalensis and podicipinus. Before commenting further on the distributions of these latter two species, discussion is required for a locality from which I believe the provenance of specimens in museum collections is problematical.

Borys Malkin collected the specimens from Igarapé Belém, Rio Solimões, Amazonas, Brazil, that are the basis of all Leptodactylus records for that locality. The Malkin specimens for Igarapé Belém are in the American Museum of Natural History (AMNH) and Museu de Zoologia da Universidade de São Paulo (MZUSP) collections. The following members of the podicipinus-wagneri complex are recorded from Igarapé Belém: L. diedrus (Figure 36), leptodactyloides (Figure 38), petersii (Figure 42), podicipinus (Figure 43), and wagneri (Figure 45). The locality is well within the known distributional limits of L. leptodactyloides and petersii, and it forms part of the known distributional limit of L. diedrus, but unremarkably so (Figure 36). Igarapé Belém is an outlier locality for L. wagneri, but the presence of wagneri (as currently understood) in other lowland Peruvian localities that are somewhat geographically removed from the Andean slopes suggests either that all of the lowland localities are in error (probably due to identification errors) or that the Igarapé Belém record may be valid for wagneri (Figure 45). However, in the case of L. podicipinus, Igarapé Belém is well removed from the rest of the podicipinus localities (Figure 43), and I believe this record represents an error. Leptodactylus podicipinus occurs in open formations. It is not a particularly difficult species to collect. Collections are available between Igarapé Belém and the other known localities for podicipinus; if podicipinus occurs in the mapped hiatus (Figure 43), it should have been collected by now. There is an additional member of the L. melanonotus group in Malkin's collection from Igarapé Belém: L. pustulatus. There are no habitat data available for pustulatus to my knowledge, but the other known localities for pustulatus all occur in open vegetation domains, geographically distant from Igarapé Belém (Figure 46). Leptodactylus pustulatus is arguably the most distinctive species in the melanonotus group, i.e., there is no doubt about identification for this species. The distributions of both L. podicipinus and pustulatus provide, in my opinion, conclusive evidence that the locality data for Igarapé Belém are in error for those two species at least. The problem is to discern which records are valid for Igarapé Belém.

There is no question that Malkin collected frogs from Igarapé Belém. The likeliest answer is that Malkin combined, by mistake, collections made from another locality with those from Igarapé Belém. Malkin did not individually field tag his specimens. Malkin collected at Igarapé Belém from 8 to 28 April 1966. He also collected at Barra do Tapirapés, Mato Grosso, Brazil, from 29 December 1965 to 16 January 1966, from which locality both L. podicipinus and pustulatus are known to occur. Both of these collections were sold to the museums simultaneously by Malkin. Dr. P.E. Vanzolini indicates that there is a possibility of specimen mixing at the MZUSP between the Malkin collections labelled as from Estirón, Peru (whence both L. diedrus and wagneri also are known), and Igarapé Belém, but not from collections labelled as from Igarapé Belém and Barra do Tapirapés (pers. comm.). I conclude that Malkin did mix some specimens from Barra do Tapirapés with those from Igarapé Belém in both the AMNH and MZUSP collections. The evidence is not as strong that Malkin also mixed in some specimens from Estirón, Peru, with the Igarapé Belém collection, but the possibility has interesting consequences in our understanding of distributions. For further work, Igarapé Belém is excluded as a locality for Leptodactylus podicipinus and pustulatus, but questionably included as a locality for L. diedrus, leptodactyloides, petersii, and wagneri.

The distribution of Leptodactylus natalensis follows the Atlantic Forest Morphoclimatic Domain from its northern extent, ending in the State of Rio de Janeiro (Figure 39). Given the level of collecting effort in the states of Rio de Janeiro and São Paulo, that southern distributional limit is believable. Interestingly, Serra dos Cavalos, Pernambuco, one of the northern localities that on a map lies outside the Atlantic Forest Morphoclimatic Domain (Ab'Sáber, 1977) is a “brejo,” containing a mesic forest that was connected with the Atlantic Forest vegetation during the last, wettest phase of the glacial cycle (see Vanzolini, 1981, for a general discussion of brejos in northeastern Brazil).

The distribution of L. podicipinus suggests a basic adaptation of podicipinus to the open formation vegetations of Argentina (northeast), Bolivia, Paraguay, and Brazil with limited invasion of the southern Amazon basin only along the major river systems of the Rio Madeira and Amazonas. The limited distribution of L. podicipinus in Amazonia suggests that it may be a relatively recent invader in that region.

Differentiation

Variation of types of characters among taxa of the podicipinus-wagneri complex is in itself variable. For example, the same adult morphological character that has discrete states and that is consistently different between two species may well demonstrate a continuous variation between two other species. Also, some taxa seem better defined by advertisement call differences, others by larval morphologies. The differentiation patterns of the character complexes analyzed for this paper can be summarized as follows.

Size is among the adult morphological features that shows the most interspecific and least intraspecific variation. Size differences among taxa usually are not discrete, however, including taxa that have extensive sympatric distributions. Measurement data, as well as the morphological features of dorsolateral folds, toe tips, and size of male thumb spines, demonstrate intraspecific variation that sometimes equals that found interspecifically, but usually each species has a distinctive (but not discretely distinct) distribution of character states. Adult pattern characters also often vary as much intraspecifically as interspecifically, but some species may have distinct character states. For example, the L. podicipinus with distinct light belly spots are distinct from all L. leptodactyloides, which never have distinct light belly spots; however, some individual L. podicipinus without distinct belly spots have the same belly patterns found in some L. leptodactyloides.

Although few larval samples have been analyzed for members of this complex, the available data suggest that interspecific variation may be more discrete than for adult morphological characters. Features of size, tooth row morphology, and tail pattern appear to characterize each species (for those species for which identifications are not in doubt).

Advertisement calls are quite species-specific in Leptodactylus species in general (Heyer and Straughan, 1976; Heyer, 1978, 1979). However, advertisement calls of the podicipinus-wagneri complex are not as distinct as those found in the other species groups. For one thing, individuals of the podicipinus-wagneri complex demonstrate a broad array of vocalization types, the functions of which are not well understood at present. At the least, individual males are capable of producing distinct advertisement and aggressive calls, but it is not clear which are which for all species. Theoretically, one would expect advertisement calls to demonstrate the greatest interspecific differences because males should attract only females of their same species, but they may defend their calling site against any intruding male independent of species. More work needs to be done to evaluate the call functions so that calls with the same functions are compared among species. It is possible that members of this complex are using the advertisement call more as a location signal than as a species coding signal. Typically, the calls of these frogs are quieter than those of other Leptodactylus species and have wide broadcast frequency bands. The latter feature facilitates point location of the sound source and the quieter call may be a consequence of producing a call with wider broadcast frequency bands.

Karyotypes have been described for only three members of the podicipinus-wagneri complex. The karyotypes of griseigularis and podicipinus are quite similar, but they are rather distinct from that of natalensis, differing in number of telocentric pairs and location of secondary constrictions (Bogart, 1974). Karyotype analysis of additional species might be fruitful in terms of understanding the patterns of evolutionary differentiation occurring in the complex.

There appears to be a modest degree of habitat differentiation among members of the complex. Some species, such as L. leptodactyloides, pallidirostris, and wagneri occur both in open vegetation formations and closed forests. Others appear to demonstrate greater habitat specialization: Leptodactylus podicipinus and validus only occur in open vegetation formations or situations; L. diedrus, pascoensis, and petersii occur exclusively (or almost so) in closed forest habitats. Whether these habitat distributions are related to processes of species differentiation is unknown at present.

Comment on Relationships

Morphological data are inadequate to perform a robust cladistic analysis for the members of the podicipinus-wagneri complex in at least two ways. All of the morphological data that could be used to evaluate relationships are the data presented in this paper. I know of no other morphological data that could be used. These data are very difficult, if not impossible, to categorize in distinct, polarized states. That is the first inadequacy. The second is that for a cladistic analysis to be robust, one should have more states than taxa. At present, that is not the case for the frogs in question. When more complete data sets are available for larvae, karyotypes, and behavior, then a cladistic analysis may well be instructive.

At this time, in order to evaluate relationships in Leptodactylus (in general), I prefer to obtain genetic-estimate data from appropriate molecular analyses. Materials are being gathered toward this end.

Unresolved Problems and Their Consequences in Understanding Evolutionary Processes

Data are insufficient to resolve all species boundaries questions in the podicipinus-wagneri complex at this time. Some of these problems have serious consequences relative to our understanding of distribution patterns and speciation processes within this complex. The unresolved problems are discussed in descending order of severity of consequences, in my opinion.

The series of populations occurring along the Venezuelan Andes and in the coastal mountains of Venezuela represent the largest unresolved problem. There is considerable morphological variation among the few available samples for these populations, and it is not at all clear how many species of the complex occur in the montane slope regions of Venezuela. A second-level problem is whether any of the Venezuela montane slope populations are conspecific with either L. colombiensis or leptodactyloides. The distributions of either L. colombiensis or leptodactyloides would be significantly expanded if either occurred along the Venezuelan montane slopes. This problem must be resolved before we can understand the role of mountain building and geographic isolation in the differentiation and speciation processes of the taxa that occur in northern South America. Data needed to resolve this problem are new intensive collections all along the Venezuelan montane slopes, concentrating not only on obtaining adequate series of specimens to analyze adult morphological variation but also on obtaining recordings of calls, larval samples, and tissue samples for molecular analysis. The data at hand are so inadequate that sampling needs to be done as if no materials existed to aid in resolution of the problem.

The second major unresolved problem involves the taxa considered as L. validus and pallidirostris in this paper. The current recognition of L. validus as a species occurring on the Lesser Antilles, Trinidad, and Tobago, but not on the mainland of South America runs counter to almost all other distribution patterns for the islands involved. As Trinidad and Tobago are continental islands and the Lesser Antilles are oceanic islands, the expected distribution patterns for a single species occurring on any combination of Trinidad, Tobago, and the Lesser Antilles are as follows: (1) the species occurs on the mainland, Trinidad and Tobago, and the Lesser Antilles; (2) the species occurs on the mainland and Trinidad and Tobago but not on the Lesser Antilles; (3) the species occurs only on Trinidad and/or Tobago; (4) the species occurs only on the Lesser Antilles. Thus, in terms of validus, one would expect that either (1) validus should be restricted to the Lesser Antilles, and the Trinidad and Tobago populations represent a closely related, but distinct, species that either is found only on Trinidad and Tobago or also occurs on the mainland of South America; or (2) validus also occurs on the mainland of South America. The adult morphological analyses indicate that there is slight differentiation of the Lesser Antilles populations from the Trinidad and Tobago populations. If the tadpoles analyzed have been correctly associated with species, then the larval data clearly support alternative (1). The mainland species that would be expected to be conspecific with validus (either as defined in this paper, or only the Trinidad and Tobago populations) is pallidirostris. The available call data are consistent with a single species, pallidirostris plus validus, occurring on the mainland, Trinidad and Tobago, and the Lesser Antilles; the adult morphological data are not. This conflict of call and adult morphological data could be due to combining of two (or more?) species within pallidirostris. There is no doubt in my mind, in comparing pallidirostris from the Guianas with validus from the islands, that different species are represented. There are no call data on hand for pallidirostris from the Guianas. The available calls for pallidirostris all come from Venezuela. There is certainly color pattern variation differences between at least some of the Venezuelan populations I have included in pallidirostris (including those for which call data are available); the taxonomic decision I took was conservative, and I would not be surprized if at least some of the Venezuelan populations I included in pallidirostris are distinct at the species level and in fact are conspecific with the populations on Trinidad and Tobago I assigned to validus. The available adult morphological data are probably sufficient; that is, analysis of additional adult morphological data is not likely to resolve this problem. The problem should be resolvable with either call data or molecular analyses. Call data from throughout the entire range of pallidirostris would be needed to determine whether two or more species have been included (available call data are adequate for validus). Molecular analysis of samples from validus from the Lesser Antilles and Trinidad and Tobago, and samples of pallidirostris at least from the Guianas and Venezuela proximate to Trinidad would be required. Adequate tissue samples are available for the various populations of validus as recognized in this paper; analysis of those should determine whether the Lesser Antilles populations are conspecific with the Trinidad and Tobago populations. Available tissue samples for pallidirostris are inadequate to resolve the relationships between pallidirostris and validus.

It is unlikely that Leptodactylus nesiotus occurs only on Trinidad. Read (1986) pointed out that frogs that are known only from the Icacos peninsula on Trinidad occur in the (adjacent) Orinoco floodplain of Venezuela (e.g., Adenomera hylaedactyla, Leptodactylus macrosternum) or the llanos of Venezuela (e.g., Hyla miniscula). The latter distribution is zoogeographically difficult to explain, but it would be worth field time to explore the Orinoco floodplain and the llanos in the State of Yaracay, Venezuela, for the presence of Leptodactylus nesiotus.

A series of geographically isolated populations along the Amazonian flanks of the Andes in Peru and Bolivia has been included in a single species, L. griseigularis. There is some differentiation among these populations as indicated by adult morphological data. Isolation of populations of L. griseigularis likely has been due to mountain-building activities. In this situation, where populations have become geographically isolated without subsequent contact, advertisement calls may not have differentiated to include species-coding information. It is likely that understanding of differentiation within L. griseigularis will depend entirely on molecular analytic techniques. To my knowledge, tissue samples are available for a couple of geographically close localities in Peru. Additional tissue samples throughout the entire range of L. griseigularis are needed.

The final problem worth comment is whether the Region 3 representatives of L. petersii are in fact conspecific with the other region representatives of petersii. All specimen-related data suggests that they are; however, the lack of habitat data for any of the Region 3 petersii raises the question of conspecificity. Almost all species of L. petersii with habitat data are from closed forests. Extensive closed forests do not exist as such in the area where the Region 3 representatives of petersii occur; the only forests are gallery forests along streams. My prediction is that the Region 3 petersii were collected in and are restricted to gallery forest habitats. Probably a single new collection of any population of Region 3 petersii with habitat and call data would resolve this problem.

The above problems are detailed for two reasons. First, although considerable progress has been made in understanding variation in the podicipinus-wagneri complex, I do not wish to leave the impression that all problems have been resolved. Second, by summarizing the problems, I hope to encourage others to gather the data needed for their resolution and to report their results.

Appendix 1

OTU (Operational Taxonomic Unit) Assignments

Part A: Single-taxon and sympatric species population OTUs. Because of the large number of OTUs designated in the “Single taxon and sympatric species analyses,” these OTUs are organized by country localities alphabetically.

Single-taxon and sympatric

species Population analyses OTU names Taxa/Region analyses OTU Names Final (species) assignment

Leptodactylus wagneri (Peters)

MATERIAL.—USNM 241307 (two specimens dissected, stage 37, SVL 10.8 mm; stage 38, SVL 9.9 mm). Collected from pond at Curuçá, Amazonas, Brazil, 9 November 1975.

REFERENCE.—Kenny (1969) described the external anatomy (as Leptodactylus podicipinus petersi).

GENERAL

Leptodactylus wagneri (common name: Wagner's white-lipped frog) is a species of frog in the family Leptodactylidae. It is found in northern South America (Brazil, Colombia, Ecuador, and Peru).

Leptodactylus wagneri are found in a variety of habitats, usually not far from water: clearings, marshes and swamps, primary forest, secondary forest, terra firme forest, flooded forest, streams, and lakes. Eggs are laid in foam nests in standing water.

Male Leptodactylus wagneri grow to a snout–vent length of 39–61 mm (1.5–2.4 in) and females to 52–82 mm (2.0–3.2 in).