Two new species of Eleutherodactylus from western and central Mexico (Eleutherodactylus jamesdixoni sp. nov., Eleutherodactylus humboldti sp. nov.)

Ethical statement

Tissue samples were donated from natural history museum collections in Mexico that were made with permission from the Secretaría de Medio Ambiente y Recursos Naturales (permit numbers 02729, FAUT-0110, and FAUT-0243). Collections in 2022 followed Animal Use Protocol AUP-2022-00192, issued by the Institutional Animal Care and Use Committee at the University of Texas at Austin.

mtDNA sequence data collection and phylogenetic analysis

DNA was isolated using spin columns (Qiagen, Hilden, Germany) and quantified using a Qubit v1 fluorometer with a broad range dsDNA assay (Thermo Fisher, Waltham, MA, USA). We amplified and sequenced an ~570 base-pair (bp) fragment of the 16S ribosomal RNA gene in the forward and reverse directions using the primers 16Sar and 16Sbr (Palumbi, 1996). Specimens and associated metadata are provided in Table S1. We used Geneious Prime 2020.2.3 (Biomatters Ltd., Auckland, New Zealand) to create assemblies and consensus sequences; bases with >1% chance of an error were automatically trimmed.

To estimate phylogenetic relationships, we assembled a data matrix of 360 16S sequences of 595 characters of which 183 sequences were generated by us and 177 downloaded from GenBank (Table S1). Eleutherodactylus (Syrrhophus) zeus and E. (Syrrhophus) symingtoni were designated as outgroups. Sequences were aligned with MAFFT version 7.450 using the einsi alignment strategy (Katoh et al., 2002; Katoh & Standley, 2013) and then manually adjusted in Mesquite 3.2 (Maddison & Maddison) (Data S1).

We used PAUP* (Swofford, 2003) to extract the shared derived character-states for each species in the 360-tip matrix. Character-state changes that were unambiguous under both ACCTRAN and DELTRAN algorithms were extracted by optimizing characters on the best likelihood tree using the command “describe 1/plot = clado labelnode = y diagnose = n apolist = y”.

We pruned the 360-tip matrix to include only unique sequences, yielding 160 sequences for the likelihood analysis (Fig. 1); sequences from all species were retained (Data S1). IQ-TREE v.2.1.2 (Minh et al., 2020) was used to select the optimal model of evolution (-m TEST) under the Bayesian Information Criterion (Kalyaanamoorthy et al., 2017) and to estimate the phylogeny using maximum likelihood. Base frequency parameters (-mfreq F, FO) and rate heterogeneity parameters (-mrate E, G) were estimated. IQ-TREE2 was run ten times to check for consistency of the log-likelihood; all runs yielded the same topology and similar log-likelihood values (within a range of one likelihood unit). Ten thousand ultrafast bootstrap replicates (Minh, Nguyen & von Haeseler, 2013; Hoang et al., 2017) were used to assess branch support. Bootstrap proportions were mapped onto the optimal likelihood tree which was visualized and annotated using the Interactive Tree of Life (iTOL) online tool (Letunic & Bork, 2019).

We calculated mean pairwise p-distances (uncorrected) between species with the ape 5.6-2 package (Paradis & Schliep, 2018) and a custom R script written in RMarkdown (Script S1) using the pairwise deletion for missing data option.

Specimens examined and morphometric analysis

We examined type specimens or photos of type specimens of Eleutherodactylus albolabris (Taylor, 1943), E. maurus (Hedges, 1989), E. modestus (Taylor, 1942), E. nebulosus (Taylor, 1943), E. nitidus (Peters, 1870), E. orarius (Dixon, 1957), E. petersi (Duellman, 1954), E. pipilans (Taylor, 1940a), E. rubrimaculatus (Taylor & Smith, 1945), E. rufescens (Duellman & Dixon, 1959), E. syristes (Hoyt, 1965) and E. teretistes (Duellman, 1958). For more recently described western species we compared specimens to diagnoses (Reyes-Velasco et al., 2015; Grünwald et al., 2018, 2021; Palacios-Aguilar & Santos-Bibiano, 2020).

We photographed preserved adult specimens using a DSLR camera mounted overhead on a copy stand and with a 0.5 mm precision ruler for scale. The ruler was elevated into the same focal plane as the part of the specimen being photographed. Images were imported into ImageJ2 (Rueden et al., 2017) and the scale was set by using the straight-line selection tool to measure 10 mm on the ruler in the image. We used the same tool to measure 17 variables on preserved specimens after Watters et al. (2016), including head width (HW), snout-vent length (SVL), tibia length (TL), interorbital distance (IOD), eye diameter (ED), internarial distance (IND), eye-nostril distance (EN), foot length (FL), tympanum diameter (TD), thigh length (THL), snout length (SL), hand length (HAL), forearm length (FLL), finger III length (Fin3L), finger III disk width (Fin3DW), finger IV length (Fin4L), and finger IV disk width (Fin4DW). One of us (SK) collected all measurements to avoid intermeasurer bias (Lee, 1982). Images are provided in the Dryad digital repository associated with this article (Devitt et al., 2023).

We explored morphological variation using principal components analysis (PCA) on two separate datasets. The E. maurus dataset (Data S2) consisted of seven topotypic E. maurus and eight individuals of its sister taxon (which we describe below as E. humboldti sp. nov.) identified in our phylogenetic analysis. The E. nitidus dataset (Data S3) consisted of five topotypic E. nitidus and three individuals of its sister species (E. jamesdixoni sp. nov.). We first tested whether variables were skewed using the dlookr package for R (Ryu, 2022). Eleven variables in Data S3 (SVL, IND, FL, TD, THL, SL, HAL, FLL, Fin3L, Fin3DW, and Fin4DW) and seven variables in Data S3 (SVL, IOD, IND, EN, THL, SL, and Fin3DW) were skewed so we transformed all measurements on a log base 2 scale. We performed PCA separately for each dataset on the correlation matrix using the R packages tidyverse (Wickham et al., 2019) and broom (Robinson, 2014) (Script S2).

Call analysis

We recorded advertisement calls from topotypic samples of E. maurus (n = 4 individuals), E. nitidus (n = 4), E. humboldti (n = 3) and E. jamesdixoni (n = 3). Advertisement calls were recorded in waveform audio file format (.wav) using a Marantz PMD661 handheld recorder with a Sennheiser shotgun microphone at a sampling rate of 44.1 kHz, held 25–50 cm from the frog. We first filtered background noise from recordings using the noise reduction feature in Audacity 3.2.1 (Audacity Team, 2022). Filtered recordings were then analyzed using Raven Pro 1.6.3 (Cornell Lab of Ornithology, 2022). We selected portions of recordings for analysis where the rate of the calls had stabilized; if a recording contained an intercall interval longer than three times that of the average, we deemed that interval to be a break between calling bouts. Calls were classified as pulsatile (poorly defined energy bursts without intermittent silence) or pulsed (a call consisting of well-defined energy bursts) following Köhler et al. (2017). For all calls, we measured the peak amplitude, call duration, intercall interval duration, call rise time, dominant frequency, and maximum and minimum fundamental frequency. If calls were pulsed, we also measured pulse duration and interpulse interval. Definitions for these characteristics follow Köhler et al. (2017), apart from call rise time which follows Cocroft & Ryan (1995). Descriptive statistics for call characteristics were calculated in R 4.2.2 (R Core Team, 2022). We used Seewave 2.2.0 (Sueur, Aubin & Simonis, 2008) implemented in R to plot the temporal and spectral properties of the signals using short-time Fourier transform (STFT) calculations. To avoid subjectivity in call descriptions and to facilitate comparison with other species, we provide recordings (raw and filtered) and associated output files in the Dryad digital repository for this article (Devitt et al., 2023).

Species delimitation

We view species as separately evolving lineages (Simpson, 1951; Wiley, 1978; de Queiroz, 1998) and the evolutionary processes underlying lineage divergence (natural selection, mutation, gene flow, and genetic drift) as contingent properties that may be used as evidence of independence. For many species, evidence of barriers to gene flow based on genetic data are useful in assessing whether they are evolving separately. We agree with Renner (2016) and Vences (2020) in emphasizing diagnosis over description to accelerate the cataloging of species level diversity, and use DNA-based diagnoses, color images, advertisement call recordings, and geographic distribution to diagnose these species, and include a list of DNA synapomorphies for other named western Syrrhophus species.

New zoological names

The electronic version of this article in Portable Document Format (PDF) will represent a published work according to the International Commission on Zoological Nomenclature (ICZN), and hence the new names contained in the electronic version are effectively published under that Code from the electronic edition alone. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved, and the associated information viewed through any standard web browser by appending the LSID to the prefix http://zoobank.org/. The LSID for this publication is urn:lsid:zoobank.org:pub:C59B3FB8-839F-407C-ADC0-67C1FEA6A5BB. The online version of this work is archived and available from the following digital repositories: PeerJ, PubMed Central SCIE and CLOCKSS.

Alignment

The alignment of 160 unique sequences had 595 characters of which 170 were parsimony-informative and 382 were constant. The best model of molecular evolution was HKY+FO+G4 (three classes of substitutions, AC=AT, AG=CT, CG=GT; unequal base frequencies; base frequencies optimized under likelihood; and discrete gamma model with four rate categories).

Phylogeny

The log-likelihoods of the ten best trees ranged from −4,996.8648 to −4,997.5458. The best tree is shown in Fig. 2 and Fig. S1. Parameter estimates for the best tree are: Rate parameters AC=AT=CG=GT 1.0000, AG=CT 9.2120; base frequencies A: 0.375, C: 0.252, G: 0.114, and T: 0.259; alpha parameter of gamma distribution 0.195.

For almost all species, the tips form monophyletic groups with ultrafast bootstrap support >90; however, the support value for E. petersi is 82. Two clusters of sequences are not monophyletic. The cluster of E. pipilans sequences is paraphyletic to a clade (bootstrap support 100) made up of E. rubrimaculatus and E. nebulosus sequences. The clade including E. pipilans, E. rubrimaculatus, and E. nebulosus has high bootstrap support (94). However, support for the node that makes E. pipilans paraphyletic is 86.

A clade containing Eleutherodactylus grandis (Dixon, 1957), E. angustidigitorum (Taylor, 1940b) (named with suffixes 1, 2, and 3), and E. floresvillelai (Grünwald et al., 2018) has a bootstrap support value of 93, but E. angustidigitorum is paraphyletic with respect to E. grandis and E. floresvillelai. Support for the nodes that make E. angustidigitorum paraphyletic is 92 (E. angustidigitorum 1 + E. grandis) and 74 (E. angustidigitorum 3 + E. floresvillelai).

The level of support for groups of species varies. Support for the six deepest nodes ranges from 91–100. Support for sister-species ranges from 42 (E. dilatus and E. syristes) to 100 (E. modestus and E. pallidus). Support for each of the two new species described below and their sister-species are 96 (E. maurus and E. humboldti sp. nov.) and 78 (E. nitidus and E. jamesdixoni sp. nov.).

Diagnostic characters

The number of unambiguous synapomorphies for a species ranged from 1 (E. petersi) to 16 (E. floresvillelai; Table S2). The number of unambiguous synapomorphies is 10 for E. humboldti, and 5 for E. jamesdixoni; Table S2. Bootstrap support for the clade of nine E. petersi sequences is 82. Bootstrap support values for all clades with >1 diagnostic characters were >90. Of the 195 state changes in Table S2, 155 were transitions and 40 were transversions.

Pairwise distances

Mean distances between the two Cuban outgroup species (E. zeus and E. symingtoni) and ingroup species ranged from 0.104–0.159. The smallest mean between-species distance (uncorrected) was between E. albolabris and E. petersi (0.0203; Table S3). Excluding the E. angustidigitorum sequences, 12 interspecific distance values were <0.03. The two new species that we describe below are each divergent from their sister taxon by mean uncorrected pairwise p-distances of 0.0297 (E. maurus—E. humboldti) and 0.0333 (E. nitidus—E. jamesdixoni).

Morphometric analysis

For the E. maurus dataset, the first two principal components accounted for 65% of the total variance with the first PC explaining about 46%. PC axis one had strong positive loadings for SVL, IND, EN, THL, HAL, FLL, Fin3L Fin3DW, Fin4L, and Fin4DW, while PC axis two had strong positive loadings for HW, SVL, ED and IND, and strong negative loadings for FL and TD (Fig. 3A and Table 1). A biplot of the fitted model showed separation between E. maurus and E. humboldti along the first principal component axis (Fig. 3B).

For the E. nitidus dataset, the first two principal components accounted for 74% of the total variance, with the first PC explaining 57%. PC1 had strong negative loadings for HW, SVL, ED, EN, FL, THL, HAL, FLL, Fin3L, Fin4L and Fin4DW, while PC2 had strong negative loadings for IOD, IND, FLL, and Fin3DW (Fig. 4A and Table 2). A biplot of the fitted model showed two of three E. jamesdixoni samples separating from E. nitidus samples along the first principal component axis (Fig. 4B).

Call analysis

Temporal and spectral call properties are summarized in Table 3 for Eleutherodactylus humboldti and E. maurus. The calls of both species are pulsatile, but E. humboldti has a longer call, intercall interval, and call rise time as well as a lower dominant frequency compared to E. maurus (Fig. 5 and Table 3).

Summary statistics for temporal and spectral call properties are summarized in Table 4 for Eleutherodactylus jamesdixoni and E. nitidus. Calls of E. jamesdixoni are pulsed and longer than calls of E. nitidus, which are pulsatile, but the dominant and fundamental frequencies are similar for both species (Fig. 6 and Table 4).

Raw and filtered call recordings are available in the Dryad digital repository associated with this article (Devitt et al., 2023). Sample calls for each species are also available on AmphibiaWeb (https://amphibiaweb.org/).

Species delimitation

Based on combined evidence from DNA sequences, morphology, and male advertisement calls, we name two new species below. We describe coloration and provide color photos in life and in preservative. We discuss morphological features that may be useful in distinguishing between close relatives and provide call recordings to facilitate identification. The new species—like many new species of terraranan frogs—were first discovered using DNA sequences. To facilitate future species delimitation in this clade, we use molecular synapomorphies for diagnosis in addition to descriptions of morphological features.

Eleutherodactylus (Syrrhophus) humboldti, new species

LSID urn:lsid:zoobank.org:act:8EB140BE-9DAE-46A3-853D-7FF9AB306CB0

Suggested English name: Humboldt’s Peeping Frog

Suggested Spanish name: Rana fisgona de Humboldt

Holotype (Fig. 7)

MZFZ 4505 (field no. TJD 1307), adult male from Mexico, State of México, Temascaltepec Municipality, Rancho El Pinal (19.09433, −100.08279), 2,428 m, collected by Thomas J. Devitt, Antonio Esaú Valdenegro Brito, Rodrigo Gabriel Martínez Fuentes, and Peter Heimes on June 27, 2022.

Paratypes (n = 3; Fig. 7)

MZFZ 4503–4 (field nos. TJD 1305–6), both adult males from the same locality collected by the same individuals on the same date as the holotype; MZFZ 4506 (field no. TJD 1308), adult male from Mexico, State of México, Temascaltepec Municipality, Hotel Las Piñas Avándaro (19.15383, −100.14292), 1,982 m, collected by the same individuals on the same date as the holotype.

Referred specimens (Fig. 7)

IBH 34816–8 (field nos. TJD 833–5), all adult males from Mexico, State of México, Municipality of Temascaltepec, Rancho El Pinal (19.09000, −100.07779), 2,245 m, collected by Uri Omar García Vázquez on June 28, 2011.

Diagnosis and comparisons

Molecular synapomorphies for this species are (character: state change): 147: A => G, 204: C => T, 241: C => T, 243: A => T, 256: T => C, 279: C => A, 281: C => T, 292: G => A, 295: A => G and 358: G => C. Character numbers refer to the position in the alignment (Data S1).

Among species of the western clade of Syrrhophus, E. humboldti is most similar in appearance to E. maurus (Fig. 8); however, our sample sizes for each species are too small to make robust inter- or intrapopulation comparisons. These two taxa are very similar in having a triangular head with a sharp canthus rostralis with a black stripe on the lateral aspect. The stripe fades inferiorly into a grayish brown. This stripe seems to extend across the eye in the sense that the middle and lower part of the iris is black, whereas the upper part of the iris is gold. The stripe extends posterior to the eye across the tympanum (which is slightly paler in color) to the axilla. In the postorbital region, the stripe borders the sharp supratympanic fold.

The dorsum of E. humboldti is brownish tan, and in E. maurus it is either brownish tan or rust-colored with a few scattered dark brown or black markings. Both color morphs of E. maurus occur in the same population at the type locality for that species (Fig. 8). The snout lacks markings, is distinctly paler than the dorsum, and its color is set off abruptly from the dorsal color in the midorbital region; however, there is no distinct pale interorbital line as there is in E. jamesdixoni and E. nitidus.

Both species have scattered tubercles on the dorsal surfaces of the body and limbs, but those in E. maurus are slightly larger. In both a large gland is present in the inguinal region. The skin covering the gland is mottled black and white, with black predominating. The dorsal color extends onto the flanks, fading into a pale gray with irregular blackish blotches. The blotches are larger and more extensive in E. humboldti and the blackish markings are absent midventrally in E. maurus. The color of the dorsal surface of the forearm (but not the upper arm) and hindlimbs is like that of the dorsum. The color of the concealed surfaces differs between the two species: E. humboldti has an orange suffusion of the posterior thigh and upper arm (Fig. 7) that is lacking in E. maurus.

The ventral surfaces of the chin, belly, and limbs are translucent and tan to brownish. The posterior belly and thighs have uniform small tubercles, some of which have white tips.

The fingers of E. humboldti are slenderer and longer than those of E. maurus (Fig. 9). The tips of the third and fourth fingers of E. humboldti are slightly wider than those of E. maurus (E. humboldti: Fin3DW = 0.857 ± 0.099 mm, Fin4DW = 0.806 ± 0.066 mm; E. maurus: Fin3DW = 0.718 ± 0.126 mm; Fin4DW = 0.645 ± 0.095 mm; Fig. 9). The hand of E. maurus is slightly more robust than that of E. humboldti; the subarticular tubercles of E. maurus are larger and more prominent, and the supernumerary tubercles are more numerous. The palmar tubercle of E. maurus is smaller and more rounded (Fig. 9).

Other species of Eleutherodactylus that occur in this general region of Mexico include E. angustidigitorum (Taylor, 1940b), E. floresvillelai (Grünwald et al., 2018), E. grandis, E. nitidus, and E. petersi. Of these species, only E. petersi possesses third and fourth fingers that have expanded tips. The tips of these digits are wider in E. petersi than in E. humboldti and E. petersi has a mottled dorsal color pattern and silvery white spotting on the upper lip (Dixon, 1957) that E. humboldti lacks.

Measurements of the holotype (Fig. 7)

Adult male (MZFZ 4505). Measurements (in mm) are as follows: HW 9.599, SVL 24.638, TL 10.154, IOD 3.406, ED 2.621, IND 1.711, EN 1.814, FL 9.327, TD 0.926, THL 9.727, SL 2.833, HAL 5.63, FLL 6.552, Fin3L 6.114, Fin3DW 0.939, Fin4L 5.005, and Fin4DW 0.859. Vocal slits are present. Advertisement call recordings (raw and filtered) and associated output files are available in the Dryad digital repository for this article (Devitt et al., 2023).

Coloration in life (Fig. 7)

The dorsum is brownish tan. The snout lacks markings and is paler than the dorsum, set off from the dorsal color in the midorbital region. The dorsal color extends onto the flanks, fading into a pale gray with irregular blackish blotches. The skin covering the inguinal gland is mottled black and white, with black predominating. The dorsal surface of the forearm (but not the upper arm) and hindlimbs is colored like that of the dorsum. The concealed surfaces of the posterior thigh and upper arm are suffused with orange. The ventral surfaces of the chin, belly, and limbs are translucent and tan to brownish. The posterior belly and thighs have uniform small tubercles, some of which have white tips.

Coloration in preservative

In preservative, the dorsum fades to gray with dark gray blotches, and the orange suffusion of the concealed surfaces fades to pale yellowish gray. The belly fades to light gray with dark gray blotches and the chin fades to grayish lavender. Dorsal tubercles may not be visible in preservative.

Variation

The type series plus three referred individuals are uniform in coloration (Fig. 7) and size (Data S3).

Distribution and natural history

This species is known from three localities in the Quaternary Valle de Bravo volcanic field in the state of México in Central Mexico. This volcanic field is part of the Mil Cumbres physiographic subprovince within the Eje Neovolcánico physiographic province. The type series was collected at night between 21:00 and 0:00 during a thunderstorm with ambient air temperatures of 9–11 °C. Habitat at the type locality was a disturbed clearing in pine-oak forest. Individuals were calling from low (<1.5 m) vegetation, frequently within dense bunches of non-native grass.

Etymology

We name this species in honor of German scientist Alexander von Humboldt, whose year-long exploration of south-central Mexico at the turn of the 19^th century resulted in the first scientific account of the New World and whose writings continue to influence biogeographers, ecologists and evolutionary biologists seeking to understand the origin and maintenance of biodiversity in mountain regions (Rahbek et al., 2019b, 2019a).

Eleutherodactylus (Syrrhophus) jamesdixoni, new species

urn:lsid:zoobank.org:act:6DB31FC7-7DC2-413B-9934-1D370CF5D5EB

Suggested English name: Dixon’s Peeping Frog

Suggested Spanish name: Rana fisgona de Dixon

Holotype (Fig. 10)

IBH 34852 (field no. TJD 897), adult male from Mexico, State of Sinaloa, Municipality of Concordia, 7.7 road miles west of El Palmito on the highway from Concordia to Durango (23.50418, −105.83701), 1,992 m, collected by Thomas J. Devitt, Andrea Roth Monzón, and Andrés Alberto Mendoza Hernández on July 26, 2011.

Paratypes

IBH 34849, 34851 (field nos. TJD 894 and TJD 896), both adult males from the same locality collected by the same individuals on the same date as the holotype.

Referred specimens

UTA A-5905 (JAC 23704), UTA A-59506 (JAC 23782), TNHC 117001 (JSR 580A), and MZFC 4789 (OFV 325).

Diagnosis and comparisons

Molecular synapomorphies for this species are (character: state change) 204: C => T, 291: C => A, and 360: C => A. Character numbers refer to the position in the alignment (Data S1).

Among species of the western clade of Syrrhophus, E. jamesdixoni (Fig. 10) is most similar in appearance to E. nitidus (Fig. 11). We restrict our comparison to topotypic samples of E. nitidus. The two species are very similar in having a roughly triangular head with a gently angled canthus rostralis. The lateral aspect of the canthus is darker than the dorsal surface of the snout. In E. jamesdixoni a dark greenish black stripe is present, but in E. nitidus this stripe varies from very dark to barely differentiated from the color of the snout (Fig. 11). The stripe fades inferiorly in E. nitidus, but in E. jamesdixoni it is more distinct. This stripe seems to extend across the eye in that the middle and lower part of the iris is dark, whereas the upper part of the iris appears as densely packed gold flecks (this golden region is larger than in E. maurus and E. humboldti). In both, the stripe extends posterior to the eye across the tympanum (which is slightly copper-colored) to the axilla but is much less distinct than in E. maurus and E. humboldti. In the postorbital region the stripe becomes indistinct. The supratympanic fold is rounded but weakly developed.

In E. jamesdixoni the dorsal surfaces of the head, body, and limbs have a contrasting mottled pattern of dark green and tan to pale yellow-orange markings. The dorsal surfaces of E. nitidus are more variable; some have contrasting markings of tan and dark brown (Fig. 11), but others have tan to dark brown or pale greenish dorsal surfaces with few contrasting markings. In both species (and in contrast to E. maurus and E. humboldti), the dorsal surfaces of the snout are patterned like the dorsum. The snout and dorsum are separated by a distinct pale interorbital line.

Both species have tubercles irregularly scattered on the dorsal surfaces of the body and limbs, but those in E. nitidus are slightly larger. In both, a gland is present in the inguinal region; it is smaller and less protuberant than in E. maurus and E. humboldti. The skin covering the gland is mottled black and white, with black predominating. Because of the strong contrast in the pattern, the gland is less obvious in E. jamesdixoni.

In both species the pattern and color of the dorsum extends onto the flanks. In E. nitidus it fades to a pale gray with irregular blackish markings. In E. jamesdixoni the flanks and belly are mottled with dark gray and white, whereas in E. nitidus the belly is pale, translucent, and lacks markings.

The color of the concealed surfaces is similar in both species. An orange suffusion is present on the posterior thigh, anterior thigh and adjacent inguinal region, and upper arm (Figs. 10 and 11).

In E. nitidus the ventral surfaces of the chin, belly, and limbs are translucent and vary from grayish pink, apparently lacking pigment, to dull gray. In both species, the posterior belly and thighs have uniform small tubercles, some of which have white tips. In E. jamesdixoni the belly and portions of the ventral sides of the limbs are mottled with black and white and the chin is darkly pigmented.

The general shape and relative sizes of the fingers, including the expansion of digital tips, are similar in both species (Fig. 12). The subarticular, palmar, and supernumerary tubercles of E. nitidus are slightly more pronounced than those of E. jamesdixoni. The palmar tubercle of E. nitidus is more elongate and conical than that of E. jamesdixoni (Fig. 12).

Other species of Eleutherodactylus that occur within the range of this species include E. interorbitalis (Langebartel & Shannon, 1956), E. modestus, E. pallidus (Duellman, 1958), E. saxatilis (Webb, 1962), E. teretistes and E. wixarika (Reyes-Velasco et al., 2015). Of these species, only E. saxatilis possesses a compact lumbar or inguinal gland. Eleutherodactylus saxatilis is easily distinguished from Eleutherodactylus jamesdixoni by third and fourth fingers that are truncate, with tips about twice the width of the narrowest part of the digit.

Measurements of holotype

Adult male (IBH 34852). Measurements (in mm) are as follows: HW 8.785, SVL 21.961, TL 8.709, IOD 2.585, ED 2.280, IND 0.913, EN 1.667, FL 8.672, TD 0.969, THL 8.635, SL 2.817, HAL 5.254, FLL 6.655, Fin3L 5.432, Fin3DW 0.698, Fin4L 4.38, and Fin4DW 0.563. Vocal slits are present. Advertisement call recordings (raw and filtered) and associated output files are available in the Dryad digital repository for this article (Devitt et al., 2023).

Coloration in life (Fig. 10)

The dorsal surfaces of the head, body, and limbs have a contrasting mottled pattern of dark green and tan to pale yellow-orange markings. A distinct pale interorbital line separates the snout from the dorsum. The belly and portions of the ventral sides of the limbs are mottled with black and white; the chin is darkly pigmented. The skin covering the inguinal gland is mottled black and white, with black predominating. The pattern and color of the dorsum extends onto the flanks. The concealed surfaces of the posterior thigh, anterior thigh and adjacent inguinal region, and upper arm are suffused with orange (Figs. 10 and 11). A greenish black stripe is present along the lateral aspect of the canthus. The lower portion of the iris is dark, whereas the upper part of the iris is golden. The stripe extends posterior to the eye across the copper-colored tympanum to the axilla but becomes indistinct in the postorbital region.

Coloration in preservative (Fig. 10)

In preservative, the dorsum fades to dark gray, and the orange suffusion of the concealed surfaces fades to a light cream color. The belly fades to a cream background color with dark grayish lavender blotches and the chin fades to grayish lavender. Dorsal tubercles may not be visible in preservative.

Variation

The type series is uniform in coloration and size (Data S3).

Distribution and natural history

Eleutherodactylus jamesdixoni is found in the southern Sierra Madre Occidental and westernmost portions of the Eje Neovolcánico and Sierra Madre del Sur physiographic provinces. The type series was collected at night in pine-oak woodland between 21:30 and 23:50 at 15°C. Individuals were calling from low vegetation perched approximately 0.5 m above ground. This species is sympatric with E. saxatilis at the type locality, a saxicolous species found on rock outcrops.

Etymology

We name this species in honor of the late James R. Dixon, Professor Emeritus and Curator Emeritus of amphibians and reptiles at the Biodiversity Research and Teaching Collections at Texas A&M University. Dr. Dixon provided the first and most thorough statistical analysis to date of geographic variation among populations of the former genus Tomodactylus from western Mexico (Dixon, 1957).