Adenomera saci, Carvalho, Thiago Ribeiro De & Giaretta, Ariovaldo Antonio, 2013

Carvalho, Thiago Ribeiro De & Giaretta, Ariovaldo Antonio, 2013, Taxonomic circumscription of Adenomera martinezi (Bokermann, 1956) (Anura: Leptodactylidae: Leptodactylinae) with the recognition of a new cryptic taxon through a bioacoustic approach, Zootaxa 3701 (2), pp. 207-237 : 218-230

Adenomera saci , new species

Figures 2 C–D, 5–8

Holotype. AAG-UFU 1339, adult male, collected at the Parque Nacional da Chapada dos Veadeiros (14°10'S, 47°47'W; approximately 1030 a.s.l.), Municipality of Alto Paraíso de Goiás, northern State of Goiás, central Brazil, on 10–15 November 2012, by T. R. de Carvalho, B. F. V. Teixeira, and L. B. Martins.

Paratopotypes. Five adult males: AAG-UFU 0108–0109, on 2–3 December 2010, by A. A. Giaretta; AAG- UFU 0762–0763, on 13–19 November 2011, by A. A. Giaretta and K. G. Facure; ZUEC 3287, on 10 January 1974, by W. C. A. Bokermann, all collected at the Chapada dos Veadeiros, northern State of Goiás, central Brazil.

Referred specimens. Non-topotype specimens under the name Adenomera martinezi: Heyer 1973 ; Vitt et al. 2005; Cintra et al. 2009; Moreira et al. 2009; Oda et al. 2009; Silva Jr. et al. 2009; Kopp et al. 2010; Morais et al. 2011; Valdujo et al. 2012.

Diagnosis. Adenomera saci sp. nov. is assigned to the genus ( L. marmoratus species group and Adenomera genus definitions; sensu Heyer 1973, 1974, respectively) by the following set of characters: 1) small body size (up to 34.1 mm; sensu Kok et al. 2007); 2) toes lacking fringing or webbing; 3) adult males lacking thumb spines; 4) first and second fingers of approximately equal length. The new species is diagnosed from the other 15 species of Adenomera by the following combination of characters: 1) presence of four-six symmetrically arranged rows of longitudinal dark-colored spots on dorsum; 2) medium-sized Adenomera (adult male SVL 19.7–22.6 mm); 3) vertebral pin-stripe always present; 4) toe tips unexpanded and unflattened; 5) non-pulsed (a whistle to the human ear), long advertisement call (72–241 ms).

Comparison with other species. Adenomera saci sp. nov. is diagnosed from all congeners (except A. martinezi ) by its distinctive 4–6 symmetrically arranged rows of longitudinal dark-colored spots on dorsum (figs. 2C–D, 5–7).

Adenomera saci sp. nov. (adult male SVL 19.7–22.6 mm; Table 1) can additionally be diagnosed from A. araucaria , and A. nana (combined adult male SVL 16.3–19.4 mm; Kwet & Angulo 2002; Kwet 2007) by its larger size, and from A. lutzi , and A. simonstuarti by its smaller size (combined adult male SVL 25.9–34.1 mm; Kok et al.

2007; Angulo & Icochea 2010). Adenomera saci sp. nov. has always a vertebral pin-stripe, whereas no specimens of A. bokermanni (Heyer 1973) , A. engelsi (Kwet et al. 2009) , and almost no specimens of A. andreae (Heyer 1973) do. Adenomera saci sp. nov. is diagnosed from A. thomei by not presenting a mask-like pattern on the inverted triangle of the interorbital region (Almeida & Angulo 2006); from A. lutzi , by lacking tubercles on forearms, posterior surface of thighs black with distinct yellow, orange or red spotting or mottling, and males with yellow to orangish yellow throat and belly (Heyer 1975; Kok et al. 2007). Adenomera saci sp. nov. has neither expanded nor flattened toe tips, whereas specimens of A. andreae , A. marmorata , and A. nana do (Heyer 1973; Kwet 2007).

The advertisement call (figs. 9–11; Tables 2–3) distinguishes Adenomera saci sp. nov. from A. andreae , A. coca , A. hylaedactyla , A. martinezi (present study), and A. thomei , by its non-pulsed structure, pulsed call structure in all aforementioned species (see Table 3). Adenomera araucaria was reported to possess calls with 5–11 severe amplitude modulations (see Kwet & Angulo 2002), and A. simonstuarti to possess 3–4 pulse-like, strong amplitude modulations (see Angulo & Icochea 2010), whereas Adenomera saci sp. nov. calls lack any strong amplitude modulations forming ‘pulse-like’ sections along their extent (see figs. 9–11). Adenomera saci sp. nov. can also be diagnosed from A. andreae , A. hylaedactyla , A. lutzi , and A. simonstuarti (combined value range 16–71 ms) by its longer call duration (72–241 ms); from A. marmorata (dominant frequency 4.50–5.60 kHz), as well as A. andreae and A. nana [combined value range: 2.30–3.05 kHz (1st harmonic); 4.61–5.44 kHz (2nd harmonic)] by its lower peaks of frequency [1.69–2.25 kHz (1st harmonic); 3.38–4.41 kHz (2nd harmonic)]. Bioacoustic values of A. diptyx from Santa Cruz, Bolivia (Marquez et al. 1995; see Table 3) had a marginal overlap with those of Adenomera saci sp. nov. in call duration, and peaks of sound energy of the first and second harmonics, but different from each other taking mean values into consideration. Furthermore, specimens recorded from Paraguay and Argentina (V.H. Zaracho doctoral thesis) and from the State of Mato Grosso do Sul, western Brazil (T.R. de Carvalho unpubl. data), assigned to A. diptyx had a pulsed call, which can help diagnose this species from the non-pulsed advertisement call pattern of Adenomera saci sp. nov.

Adenomera saci sp. nov. can additionally be diagnosed from A. andreae (Hero 1990; Menin & Rodrigues 2013), A. hylaedactyla (Heyer & Silverstone 1969; Menin et al. 2009), A. marmorata (Heyer et al. 1990) , and Adenomera sp. (Kokubum & Giaretta 2005), by having exotrophic larvae, with tooth rows on oral papillae, and spiracle (fig. 12) (endotrophic larvae with both tooth rows and spiracle absent in all aforementioned species; except A. marmorata larvae, which was reported to have a spiracle). Among Adenomera species with exotrophic larvae, Adenomera saci sp. nov. larvae are distinguished from those of A. thomei (first posterior tooth row with no gaps; Almeida & Angulo 2006) by having a short gap in the first posterior tooth row (fig. 13). Morphological features of A. diptyx mouthparts (see fig. 2 in De la Riva 1995) agree with those of Adenomera saci sp. nov. (fig. 13C) by having marginal papillae with a dorsal gap, and two anterior and three posterior tooth rows sharing the following labial tooth row formula: [LTFR 2(2)3/(1)].

Description of holotype. AAG-UFU 1339 (figs. 2C–D, 5A, 8). Adult male. Snout pointed in dorsal view (fig. 8A), acuminate in lateral view (fig. 8B). Weakly developed shovel-like fleshy ridge on snout tip. Nostrils positioned dorsolaterally, closer to snout tip than to eyes; canthus rostralis indistinguishable; loreal region concave; upper eyelids glandular. Supratympanic folds extending from posterior margin of eyes, passing over tympani, and extending to base of arms. A discrete, elongated post-commissural gland. Vocal sac subgular, internal. Vocal slits present; vomerine teeth in two straight rows posterior to choanae; tongue ovoid, free behind. Relative finger lengths I ~ II ~ IV <III; finger tips rounded, not expanded, and with no webbing or fringing; inner metacarpal tubercle elongated; outer metacarpal tubercle nearly rounded (fig. 8D); subarticular and supernumerary tubercles rounded. No thumb asperities. Dorsum and dorsal surface of shanks glandular; posterior half of body, dorsal surface of shanks, and posterior surface of tarsi with minute tubercles. Vertebral pin-stripe present. Four longitudinal rows of symmetrically arranged spots. Flanks with longitudinal granular rows from posterior margin of tympani to groin. Throat and belly smooth. Ventral surface of thighs areolate. Posterior surface of thighs spotted. Relative toe lengths I <II <V ~ III <IV; toe tips rounded, not expanded (pointed toe tips: character state A; see fig. 1A in Heyer 1973); toes ridged laterally, with no webbing. Inner metatarsal tubercles ovoid, outer conical (fig. 8C); subarticular and supernumerary tubercles rounded. Tarsal folds formed by ill-defined rows of tubercles, extending from inner metatarsal tubercle to approximately 1/2 length of tarsi. Heel smooth.

Measurements of holotype. Morphological measurements (mm) and ratios (%) in relation to SVL (22.5 mm): HL 8.8 (39.1), HW 7.1 (31.6), ED 2.3 (10.2), TD 1.6 (7.1), END 2.0 (8.9), IND 2.1 (9.3), FRL 4.8 (21.3), HAL 5.6 (24.9), TL 9.9 (44.0), SL 10.5 (46.7), TSL 6.8 (30.2), FL 12.3 (54.7).

Coloration of holotype in alcohol (fig. 2C–D). Snout tip has a faded coloration. Tympani reddish brown. Upper jaw with sinuous cream-colored stripes, white spots/blotches, and a white-colored gland. Dorsum with four longitudinal rows of symmetrically arranged black spots barely distinguishable from the very dark brown background. Flanks with cream-colored longitudinal row of granules on a black background. Pale gray vertebral stripe. White-tipped tubercles scattered on posterior half of dorsum, dorsal surface of shanks, and outer surface of tarsi. Dorsal surface of limbs with brown blotches on a very dark brown background. Posterior surface of thighs with white spots on a very dark brown background. Throat and belly cream, and ventral surface of limbs creamcolored, with melanophores, but with no distinctive color pattern.

Coloration of holotype in life (fig. 5A). Iris copper, upper eyelids bordered by an ill-outlined cream stripe. Tympani reddish brown. Dorsum with black longitudinal rows of spots on a moss green and grayish brown background. Spots and blotches on lateral of head and flanks with a whitish silver or yellow coloration. Flanks with cream-colored longitudinal granular rows on a grayish brown background. Vertebral stripe cream, with some portions tending to an orange coloration. Dorsal surface of limbs with black stripes and/or blotches on a brown background. Throat, belly, and ventral surface of limbs cream.

Variation. Morphometric variation is summarized in Table 1. Variation within the type series is restricted to dorsal texture (smooth or glandular), and color pattern (dark gray to very dark brown background), considering that the four to six longitudinal rows of symmetrically arranged black spots are variable due to the dorsal coloration itself, which might become evident or barely indistinguishable in preserved specimens. Posterior surface of thighs vary from a spotted or mottled pattern, and which can form ill-defined and fragmented stripes (AAG-UFU 0 109, 0762–63, 0 809, 1673, 1761–63, 2671, and ZUEC 3287) or a well-defined striped pattern (AAG-UFU 0108). Lower jaw can be dark-colored coinciding with the lateral expansion of vocal sac. Besides the melanophores scattered on ventral surface, specimens AAG-UFU 0 109, 0 809, 1673, 1761–63, have throat, belly, and ventral surface of thighs finely covered with white dots and/or spots.

Non-topotype specimens also vary in dorsal texture (smooth or glandular), and color pattern (dark gray to very dark brown background). Four to six longitudinal rows of symmetrically arranged black spots on dorsum are variable due to the dorsal coloration itself, which might become evident or barely indistinguishable in preserved specimens. Specimens AAG-UFU 0 809, 1745–46, and 1761 have a faded white, well-developed shovel-like fleshy ridge on snout. Dorsal surface of thighs have spotted or mottled pattern in all specimens (AAG-UFU 0 208, 1745– 48).

The adult female AAG-UFU 1748 has a robust body shape, head sub-elliptical in dorsal view, with no shovellike fleshy ridge on snout tip, as in the case of the adult female topotype of A. martinezi (AAG-UFU 1521). Specimens in life have some variation with respect to coloration, such as dorsal coloration (olive green to black), vertebral stripe (cream to brownish orange), granular rows on flanks (cream to yellow) (figs. 5–7).

Advertisement call from the type locality (Chapada dos Veadeiros). Seven males recorded (N = 343 analyzed calls). Advertisement call (fig. 9; Table 2) consists of a non-pulsed signal (a whistle to the human ear) with up to nine harmonics and with a slight ascendant frequency modulation, emitted at a rate of 1–3 calls/second (mean 2.2; SD = 0.3), and at a rate of 111–132 calls/minute (mean 123.2; SD = 10.3). Call duration varies from 90– 139 ms (mean 112.1; SD = 11.0), and intercall interval from 220–513 ms (mean 361.4; SD = 77.4). Dominant frequency coincides with either the first (14%) or the second harmonic (72%) among the recorded individuals, or can alternate between the first and second harmonics (14%) among analyzed calls of the same individual. Peak of sound energy varies from 1.69–2.06 kHz (mean 1.85 kHz; SD = 0.07) in the first harmonic, and/or from 3.38–3.94 kHz (mean 3.67 kHz; SD = 0.18) in the second harmonic. The other harmonics, if present, are increasingly weaker in sound energy.

Bioacoustic comparisons. Dominant frequency of non-topotype populations of A. saci sp. nov. coincides with either the first (75%) or the second harmonic (15%) among the recorded individuals from Cristalina (N = 4; fig. 10); with either the first (57%) or the second harmonic (43%) among the recorded individuals from Pontal do Araguaia (N = 7; figs. 11A–B); and with the second harmonic among the analyzed calls of the individual from Alto Araguaia (N = 1; figs. 11C–D).

Bioacoustic data of topotypes of Adenomera martinezi , and Adenomera saci sp. nov. from the type locality and of comparative non-topotype populations are summarized in Table 2. There was no correlation between the first principal component, taking all three temporal variables tested into account, and the temperature [r = 0.13 (r2 <2%), p = 0.463; N = 34], in other words, temperature did not affect the temporal variables and did not explain the variability found in the calls of topotypes of A. martinezi and all four populations of A. saci sp. nov.

The first three principal components (PCs) accounted for around 96% of the overall bioacoustic variability among the populations of A. saci sp. nov. A scatterplot of bioacoustic variables on the first two principal components (fig. 12) showed that the topotype (Chapada dos Veadeiros) and Cristalina populations were almost completely distinguished from the populations from Pontal do Araguaia and Alto Araguaia along the PC1, an expected result based on bioacoustic differences previously observed by us in the quantitative data comparisons (see Table 2) of populations from Goiás (topotypes and Cristalina) and Mato Grosso (Pontal do Araguaia). The insufficient call sample of the single recorded individual from Alto Araguaia (Mato Grosso) did not allow us to assume that the bioacoustic characteristics of this population are similar to that of Pontal do Araguaia, or different from both Pontal do Araguaia and populations from Goiás (topotypes and Cristalina).

Tadpole (fig. 13). Stage 37. Body tending to ovoid in dorsal view, compressed dorsoventrally. Eyes and nares positioned dorsally, just nearer the eyes than the snout tip, with a slight rim, but lack any papillae or ornamentations (fig. 13B). Oral disc positioned anteroventrally. Spiracle singular, sinistral. Spiracular opening with its inner wall as a slight ridge adhered to body (fig. 13A). Right wall of vent tube displaced dorsally and anteriorly. Dorsal fin originates at body-tail juncture; ventral fin originates at the vent tube. Tail height just greater than body height. Dorsal and ventral fins with the same height. Tail tip with a narrow and pointed shape. Oral disc lacks tooth rows (see variation below). Marginal papillae with a wide dorsal gap surround the oral disc. Presence of keratinized of both upper (arched) and lower jaw (v-shaped) sheaths with serrated cutting edges. Dorsal surface of body dark brown, ventral surface cream. Tail with dark brown blotches on both dorsal and ventral fins, and on tail muscle. Measurements (mm): TL 34.8; BL 10.1; TAL 24.6; IOD 3.4; IND 1.6; TMH 2.4; MTH 3.9; TMW 2.5.

Variation of tadpoles. Two tadpoles (Gosner’s stages 36–37) have well-defined tooth rows: LTRF 2(2)/3(1) (fig. 13C). The second anterior tooth row has a wide medial gap, approximately 1/3 length of the entire row. The first posterior tooth row has a very slight medial gap.

Natural history. Adenomera saci sp. nov. occurs in montane (> 1000 m a.s.l.) rock fields with sandy/muddy soil, or lowland (~ 350 m a.s.l.) grassy fields with sandy/muddy soil that are almost always associated with palm grove marshes (Veredas). Males call within underground chambers or from exposed calling sites, most times among grassy tufts. The species is mainly active during the night, but males may also be heard during daylight in rainy days.

Geographic distribution. The widespread pattern of distribution of Adenomera saci sp. nov. encompasses both rocky Cerrado environments of central Brazil and transitional Amazonian regions of the Araguaia River and Xingu River basins, central-western and northern Brazil (see fig. 14 for a distribution map and references).

Etymology. The proper name ‘saci’ in Portuguese (sa’si, sassy perere, or matintape’re in Tupi indigenous language) stands for a kind of whistling imp, which according to the native Tupi ethnic groups in Brazil, is an annoying prankster to whoever sees him. The name of the species, an invariant noun in apposition, is an allusion to the whistle call pattern emitted by the species that we associate to this well-known character in Brazilian folklore.

Remarks. It is noteworthy to add that the advertisement calls of A. martinezi and A. saci sp. nov. are distinguishable to each other to the ear. Specimens reported from other localities/regions in addition to those discussed earlier in the geographic distribution section (Vitt et al. 2005; Silva Jr. et al. 2009; Morais et al. 2011) compared either to our study areas (Chapada dos Veadeiros, Cristalina, Alto Araguaia, and Pontal do Araguaia) or to localities from where we have bioacoustic information available (Formoso do Araguaia and Santa Terezinha) were assigned to Adenomera saci sp. nov. (see fig. 14). Specimens reported from the Ilha do Bananal, southwestern State of Tocantins, and the Xingu River basin, State of Mato Grosso (Heyer 1973), were unambiguously assigned to Adenomera saci sp. nov. based on digitized recordings of Adão J. Cardoso available at the Fonoteca Neotropical Jacques Vielliard from Formoso do Araguaia, State of Tocantins, and Santa Terezinha, State of Mato Grosso, respectively (sound files FNJV 11223–11224). Taking also the proximity of those localities into account, reinforced by the widespread distribution pattern of the species against a putative ‘Cachimbo endemic’, restricted distribution pattern of A. martinezi , we tentatively assigned the specimens reported from the State of Mato Grosso do Sul (Silva Jr et al. 2009), and the Jalapão and nearby localities, southeastern State of Tocantins (Vitt et al. 2005; Morais et al. 2011) to Adenomera saci sp. nov. as well (fig. 14).

Regardless of the bioacoustic differences (fig. 12; Table 2) and geographic distance (fig. 14) among populations of Adenomera saci sp. nov. discussed earlier, our results are still not conclusive. We believe that these populations ought to benefit from a molecular evidence, as well as additional efforts to increase call sample, combining two datasets to properly address the taxonomic status of populations from Mato Grosso and uncover a potential cryptic diversity among the widely distributed populations currently assigned to Adenomera saci sp. nov. through a bioacoustic evidence.

Examined material. Topotype specimens—GOIÁS: Alto Paraíso de Goiás (CHUNB 42537, 49506–49509, 58405, 58884, 58886); Non-topotype specimens –DISTRITO FEDERAL: Brasília (AAG-UFU 2671; CHUNB 58358); GOIÁS: Colinas do Sul (CHUNB 44706); Cristalina (AAG-UFU 0 809, AAG-UFU 1673, AAG-UFU 1761–1763); MATO GROSSO: Alto Araguaia (CHUNB 65849); Nova Xavantina (CHUNB 64343–64344); Novo Santo Antônio (CHUNB 57787, 57861, 57982, 57985); Pontal do Araguaia (AAG-UFU 0 208, 1745–1748); Ribeirão Cascalheira (ZUEC 7496–7498); Santa Terezinha (ZUEC 7459–7460); MATO GROSSO DO SUL: Cassilândia (CHUNB 58798); TOCANTINS: Caseara (CHUNB 45600); Colinas do Tocantins (CHUNB 50969, 50972–50975, 52822–52825, 53140–53141); Guaraí (CHUNB 52446–52447); Lagoa da Confusão (CHUNB 58135, 58140, 58186); Mateiros (CHUNB 27314–27320, 27322, 27324–27327, 41764, 41766, 42005–42006, 42008–42017); Palmas (CHUNB 11245–11249, 14702, 24266, 24407–24412, 24415–24420, 24435–24436, 24439–24440, 48094–48097); Porto Nacional (CHUNB 47732).

Figures

FIGURE 2. Dorsal (A) and ventral (B) views of an adult male topotype of Adenomera martinezi (AAG-UFU 1517; SVL 23.7 mm); and dorsal (C) and ventral (D) views of the holotype of Adenomera saci sp. nov. (AAG-UFU 1339; SVL 22.5 mm).