Bioacoustics of Hyalinobatrachium mashpi

Characterization. The following combination of characters are found in Hyalinobatrachium mashpi sp. nov. : (1) dentigerous process of the vomer lacking teeth; (2) snout truncate in dorsal view and slightly protruding in lateral view; (3) tympanum oriented almost horizontally; tympanic annulus barely visible, hidden under skin; tympanic membrane differentiated, with coloration similar to that of surrounding skin; (4) dorsal skin shagreen; (5) ventral skin areolate; cloacal ornamentation absent, paired round tubercles below vent absent; (6) parietal peritoneum and pericardium translucent (in life, red heart visible in ventral view); liver, viscera and testes covered by iridophores; (7) liver bulbous; (8) humeral spines absent; (9) hand webbing formula: I (2 + –2 1/2 )—(3 – –3) II (2–2 – )—(3 + –3 1/4 ) III (2–2 + )—(1 3/4 –2) IV; (10) foot webbing moderate; webbing formula: I 1 + —(2–2 + ) II (1–1 1/3 )—2 1/4 III (1 1/3 –1 1/2 ) —(2 + –2 1/4 ) IV (2 1/2 –2 1/3 )—1 V; (11) fingers and toes with thin lateral fringes; ulnar and tarsal folds absent; (12) nuptial excrescence present as a small pad on Finger I (Type V), prepollex not enlarged; prepollical spine not projecting (spine not exposed); (13) when appressed, Finger I longer than II; Species SVL in mm (adult males) Dorsal pattern Hear coloration in life Interorbital bar Type of call Biogeographic distribution/ Country/ Elevation (m a. s.l.) Sources H. adespinosai 20.5–22.2 Pale yellowish green with small pale yellow spots and minute gray to black melanophores Red Absent Pulsed Amazonian slopes of the Andes Ecuador 1,670–1,795 Guayasamin et al. (2019a) H. anachoretus 20.6–21.4 Apple green with small yellow spots and minute melanophores Red Absent Pulsed Amazonian slopes of the Andes Peru 2,001–2,050 Twomey, Delia & Castroviejo-Fisher (2014) H. aureoguttatum 20.4–24.0 Greenish yellow with large, bright yellow spots and, in some populations, dark flecks Usually red, but also red and white, or white Usually absent Tonal Chocó, Pacific slopes of Andes Colombia, Ecuador, Panama 0–1,340 Barrera-Rodriguez & Ruiz-Carranza (1989) , Guayasamin et al. (2020) , this study H. bergeri 20.3–22.4 Apple green with small yellow spots and minute melanophores Mostly white Absent Tonal Amazonian lowlands and Amazonian slopes of the Andes Peru, Bolivia 300–1,980 Castroviejo-Fisher et al. (2009) H. chirripoi 24–27 Greenish yellow with small yellow spots Red Absent Pulsed Central America (Costa Rica, Panama); Chocó (Colombia and Ecuador) 0–320 Savage (2002) , Kubicki (2007) , Guayasamin et al. (2020) H. colymbiphyllum 23–30 Greenish yellow with yellow spots Red Absent Pulsed Central America (Costa Rica, Honduras, Panama); Chocó (Colombia) 0–1,710 McCranie & Wilson (2002) , Savage (2002) , Kubicki (2007) H. esmeralda 18.4–22.3 Greenish yellow with yellow spots and some minute dark dots Red or red and white Absent Unknown Amazonian slopes of the Andes Colombia 1,026–1,700 Ruiz Carranza & Lynch (1998) , Acosta-Galvis (2017) , Twomey, Delia & Castroviejo-Fisher (2014) (Continued) Species SVL in Dorsal pattern Hear Interorbital Type of Biogeographic Sources mm coloration bar call distribution/ (adult in life Country/ males) Elevation (m a. s.l.) H. mashpi sp. 19.7–20.9 Greenish yellow with small and Red Usually Pulsed Pacific slopes of This study nov. diffuse yellow spots present the Andes Ecuador 976–1,137 H. nouns sp. nov. 19.1–21.3 Greenish yellow with small and Red Usually Unknown Pacific slopes of This study diffuse yellow spots present the Andes Ecuador 1,177–1,420 H. pellucidum 20.4–21.4 Greenish yellow with small yellow White Absent Tonal Amazonian Lynch & Duellman spots slopes of the (1973) , Guayasamin Andes et al. (2020) Ecuador, Peru 523–1,740 H. valerioi 19.5–24.0 Greenish yellow with large and Red or red Usually Tonal Central Savage (2002) , Kubicki diffuse yellow spots and white absent América, (2007) , Guayasamin Chocó, Pacific et al. (2020) slopes of the Andes Costa Rica, Colombia, Ecuador 0–1,500 H. vireovittatum 21.5–23.0 Greenish yellow with small yellow Red Present or Tonal Central America Savage (2002) , Kubicki spots. Dark green middorsal absent Costa Rica (2007) , Campos- stripe outlined by yellow 250–1,957 Villalobos et al. paravertebral stripes (2020) H. talamancae 24–26 Greenish yellow with small yellow Red Present or Tonal Central America Kubicki (2006 , 2007 ); spots. Dark green middorsal absent Costa Rica Zamora-Roda, stripe present. 475–1,600 Herrera-Martínez & Salazar (2021) H. yaku 20.8–22.3 Green with small yellow spots and Red Absent Tonal Amazonian Guayasamin et al. minute melanophores; posterior lowlands (2017a) head and anterior half of the Ecuador body with few small dark green 300–360 spots placed middorsally Note: Sources of traits are indicated in the last column. (14) diameter of eye about 2 times wider than disc on Finger III; (15) coloration in life: dorsal surfaces lime green with small light yellow spots; (16) coloration in preservative: dorsal surfaces creamish white, with minute lavender melanophores; (17) eye coloration in life: iris yellow to golden-yellow; pupil surrounded by lavender ring; (18) melanophores absent from fingers and toes, except Toes IV and V; (19) males call from underside of leaves; advertisement call consisting of single pulsed note, with duration of 0.37–0.46 s, peak frequency at 5.25–5.60 kHz, maximum frequency at 5.46–5.81 kHz, and minimum frequency at 4.62–4.92 kHz; (20) males attend egg clutches located on the underside of leaves overhanging streams; clutch size of 31 or 32 embryos ( n = 2); (21) SVL in adult males 19.7–20.9 mm (mean = 20.5; n = 5); females unknown; and (22) enameled tubercles absent from sides of head. Description of the holotype . CJ11642, adult male with SVL 20.6 mm. Head wider than long (head width 39% of SVL; head length 78% of head width). Snout truncate in dorsal view and slightly protruding in lateral view. Loreal region concave, nostrils slightly protuberant, elliptical; internarial region concave; canthus rostralis not well defined. Eyes small, directed anterolaterally, eyes about 50 Ǫ relative to midline (where anterior-facing eyes would be 90 Ǫ relative to midline). Tympanum barely visible, oriented almost horizontally; tympanic membrane differentiated and pigmented as surrounding skin. Dentigerous processes on vomers absent; choanae large, oval, separated widely (distance about the same as between nostrils); tongue round, white in preservative, anterior 4/5 attached to mouth; posterior border of tongue widely notched; vocal slits present, extending along floor of mouth lateral to tongue; enameled glands absent from lower part of upper jaw. Ulnar fold absent; humeral spine absent. Relative length of fingers: II <I <IV <III; finger discs rounded, about the same size as discs on toes, disc on Finger III 42% of eye width; hand webbing reduced between Fingers I–III, moderate between Fingers III and IV, with formula I 2 + —3 – II 2 – —3 1/5 III 2 + —1 3/4 IV. Prepollex concealed; subarticular tubercles round, faint; few small supernumerary tubercles present, palmar tubercle round, of moderate size and difficult to see, thenar tubercle ovoid; nuptial excrescences present as a small pad on external edge of Finger I (Type V). Hind limbs slender, tibia length 55% of SVL; tarsal fold absent; discs of toes round; inner metatarsal tubercle small, outer metatarsal tubercle round, both very difficult to distinguish. Webbing formula of feet: I 1 + —2 II 1—2 1/4 III 1 1/2 —2 + IV 2 1/2 —1 V. In preservative, dorsal skin creamish white, with minute dark lavender melanophores (only visible under the stereomicroscope); dorsal skin shagreen; skin on venter areolate; cloacal opening at level of upper thighs, small and non-enameled cloacal warts present. Parietal peritoneum and pericardium translucent (in life, the red heart is visible ventrally); urinary bladder lacking iridophores; liver, viscera, and tested fully covered by iridophores. Kidneys rounded, approximately bean-shaped; liver bulbous. Coloration in life. Dorsal surfaces apple green to yellowish green with diffuse yellow spots; head with light yellow spots that may form an interorbital bar. Melanophores absent from fingers and toes, except Toes IV and V. Ventrally, parietal peritoneum and pericardium transparent, with a red heart always visible. Gall bladder and urinary bladder covered by translucent peritonea; hepatic and visceral peritonea covered by white iridophores; ventral vein red. Iris yellow, with numerous minute lavender spots. Bones white. Coloration in preservative . Dorsal surfaces creamish white dotted with minute dark lavender melanophores; venter uniform cream, with partial translucence; pericardium translucent; visceral peritoneum covered by iridophores. Iris white with minute lavender melanophores. Melanophores absent from hands and feet, except from some few present on dorsal surfaces of Toes IV and V. Measurements of holotype (in mm) . CJ11642, adult male. SVL = 20.6, femur length = 11.4, tibia length = 11.3, foot length = 9.6, head length = 6.2, head width = 8.0, interorbital distance = 2.4,eye diameter = 2.6, tympanum diameter = 0.6, arm length = 4.2, hand length = 6.3, Finger I length = 4.6, Finger II length = 4.1, width of Finger III = 1.1. Vocalizations ( Figs. 6 , 7 ). We measured call variables from two individuals, each from a different locality, Mashpi Reserve (CJ11642; call code LBE-C-051) and Tayra Reserve (CJ11645; call code LBE-C-052). The call of Hyalinobatrachium mashpi sp. nov. ( Fig. 6 ) Species Call codes Call structure Call duration Inter-call Peak Maximum Minimum (s) interval (s) frequency frequency frequency (kHz) (kHz) (kHz) H. mashpi LBE-C-051, LBE- 1 note per call; pulsed; 0.37–0.46 10.07–17.48 5.25–5.6 5.46–5.81 (5.62 4.62–4.92 (4.79 sp. nov. C-052 8–10 pulses per note (0.43 ± 0.03) (12.8 ± 2.17) (5.38 ± ± 0.11) ± 0.10) two individuals, 0.12) 12 calls H. adespinosai LBE-C-050 1 note per call; pulsed; 0.448–0.646 10.87–30.04 4.87–5.04 5.0–5.25 4.32–4.75 one individual ~12 pulses per note (0.543 ± (20.12 ± 8.77) (4.94 ± (5.11 ± 0.08) (4.57 ± 0.15) 10 calls 0.07) 0.07) H. aureoguttatum LBE-C-053–057 1 note per call; tonal; 0.078–0.087 1.99–5.20 6.63–7.41 7.78–8.90 (8.11 5.16–5.91 six individuals harmonics present (0.082 ± (3.24 ± 0.79) (6.87 ± ± 0.27) (5.62 ± 0.21) 24 calls 0.002) 0.25) H. chirripoi LBE-C-010 1 note per call; pulsed; 0.235–0.274 84.3 4.48 4.99–5.77 (5.38 4.16–4.21 one individual 12–13 pulses per note (0.255 ± (only 2 notes (4.48 ± 0) ± 0.05) (4.19 ± 0.04) two calls 0.03) in recording) H. pellucidum LBE-C-003 1 note per call; tonal 0.1–0.146 1.67–5.35 5.60–5.86 5.86–6.14 5.05–5.32 1 individual (0.129 ± (2.94 ± 0.79) (5.70 ± (6.0 ± 0.06) (5.16 ± 0.07) 41 calls 0.009) 0.06) H. tatayoi LBE-C-058 1 note per call; tonal 0.076–0.276 2.05–21.68 4.45–5.11 4.83–5.40 (5.14 3.30–4.61 (4.24 four individuals (0.143 ± (7.64 ± 4.92) (4.82 ± ± 0.17) ± 0.34) 26 calls 0.04) 1.77) H. valerioi ML 201469, 1 note per call; tonal; 0.065–0.10 1.76–8.00 6.46–7.24 7.22–7.90 4.09–5.88 (5.12 three individuals ML201473, harmonics present (0.079 ± (4.27 ± 1.21) (6.77 ± (7.53 ± 0.17) ± 0.51) 70 calls ML201475 0.01) 0.19) Note: For each variable, data range is followed by the mean and standard deviation in parentheses. consists of a single pulsed (amplitude-modulated) note, which starts with one lower-frequency pulse followed by ~9 more consistent pulses at a slightly higher dominant frequency. We analyzed variables from both individuals: four calls from CJ11645 and eight calls from CJ11642. Calls in our field recordings had a duration of 0.373–0.461 s (mean = 0.425 ± 0.027 SD, n = 12). Time between calls ranged from 10.07–17.48 s (mean = 12.80 ± 2.166 SD, n = 10); intervals between H. mashpi calls were longer when a sympatric glassfrog ( Espadarana prosoblepon ) called in the interim period. Peak frequency was 5.25–5.6 kHz (mean = 5.38 kHz ± 0.12 SD; n = 12), with a maximum frequency of 5.46–5.81 kHz (mean = 5.38 kHz ± 0.11 SD; n = 12) and a minimum frequency of 4.62–4.92 kHz (mean = 4.79 kHz ± 0.10 SD; n = 12). We compared the calls of H. mashpi sp. nov. to those of phenotypically and/or genotypically similar species within the same genus: H. adespinosai , H. aureoguttatum , H. pellucidum , H. tatayoi , and H. valerioi ( Table 2 ). The call of H. adespinosai is a single pulsed (amplitude-modulated) note, consisting of ~12 pulses (mean call duration = 0.54 s ± 0.007 SD, n = 10). Time between calls ranged from 10.87–30.04 s (mean inter-call interval = 20.12 s ± 8.77 SD). Mean peak/fundamental frequency was 4.94 kHz (±0.07 SD; range = 4.87–5.04 kHz), with a mean maximum frequency of 5.11 (±0.08 SD; range = 5.0–5.25 kHz) and a mean minimum frequency of 4.57 (±0.15 SD; range = 4.32–4.75 kHz). The call of H. aureoguttatum ( Fig. 6 ; Table 2 ) consists of a very short, single tonal note (mean call duration = 0.082 s ± 0.002 SD, n = 24). Time between calls ranged from 1.99–5.20 s (mean inter-call interval = 3.24 s ± 0.79 SD, n = 23). Mean peak/fundamental frequency was 6.86 kHz (±0.25 SD; range = 6.55–7.41 kHz; n = 24 calls). Two harmonics are present. We measured call variables from individuals recorded in Canandé (0.5112 N, 79.1343 W; 457 m), Esmeraldas Province, Ecuador, in December 2018 by AVA (LBE- 053–55), and in Mashpi Lodge Reserve (0.17057 N, 78.888 W; 721–723 m) in March 2019 by RMB (LBE-056, 057). The call of H. chirripoi is a single pulsed (amplitude-modulated) note, consisting of ~12 pulses (mean call duration = 0.255 s ± 0.03 SD, n = 2). Since our recording only included two bouts of calling, we were unable to include H. chirripoi in the DAPC analysis. The interval between the two calls was 84.3 s. Peak/fundamental frequency was 4.48 kHz, with a maximum frequency of 4.99–5.77 kHz and a minimum frequency of 4.16–4.21 kHz. We measured call variables from one individual recorded in Reserva Itapoa (0.51307 N, 79.134 W; 321 m), Esmeraldas Province, Ecuador, in July 2016 by JC (LBE-019). The call of H. pellucidum consists of a short, single tonal note (mean call duration = 0.129 s ± 0.009 SD, n = 41). Time between calls ranged from 1.67–5.35 s (mean inter-call interval = 2.94 s ± 0.79 SD). Mean peak/fundamental frequency was 5.70 kHz (±0.06 SD; range = 5.60–5.86 kHz), with a mean maximum frequency of 6.0 (± 0.06 SD, range 5.86–6.14 kHz) and a mean minimum frequency of 5.16 (±0.07 SD, range 5.05–5.32 kHz). We measured call variables from one individual (USNM 286708) recorded at Río Azuela, Napo Province, Ecuador, by Roy McDiarmid on February 23th, 1979. The call of H. tatayoi consists of a short, single tonal note (mean call duration = 0.143 s ± 0.04 SD, n = 26). Time between calls ranged from 2.05–21.68 s (mean inter-call interval = 7.64 s ± 4.92 SD). Mean peak/fundamental frequency was 4.82 kHz (±1.77 SD; range = 4.45–5.11 kHz), with a mean maximum frequency of 5.14 (±0.17 SD, range 4.83–5.40 kHz) and a mean minimum frequency of 4.24 (±0.34 SD, range 3.30–4.61 kHz). We measured call variables from four individuals recorded in Jama Coaque Reserve (0.108264 S, 80.117701 W; 700 m), Manabí Province, Ecuador, in March 2019 by RMB. The call of H. valerioi ( Fig. 6 ) consists of a single tonal note (mean call duration = 0.079 s ± 0.01 SD, n = 70). Time between calls ranged from 1.76–8.00 s (mean inter-call interval = 4.27 s ± 1.2 SD). Mean peak frequency was 6.77 kHz (±0.19 SD; range = 6.46–7.24 kHz). Harmonics are likely present but are difficult to discern in the available recordings. We measured call variables from three individuals recorded in Costa Rica (Limón and Rincón de Oso) by Roy McDiarmid. We used the following recordings from the Macaulay Library at the Cornell Lab of Ornithology: ML212787, ML212788, and ML213430. Results from the discriminant analysis of principal components (DAPC) revealed that the calls of H. mashpi sp. nov. cluster separately, and are thus acoustically distinct from H. adespinosai , H. aureoguttatum , H. pellucidum , H. tatayoi , and H. valerioi ( Fig. 7 ). Overlap occurred between H. aureoguttatum and H. valerioi clusters, as well as between H. tatayoi and H. adespinosai clusters. This suggests that the calls of these pairs may not be adequate for species identification alone; more field recordings with genetic verification of PC1 PC2 PC3 PCA Eigenvalues 23.76 1.48 0.16 PCA Loadings Peak Frequency 0.08 −0.57 −0.09 Call (Note) Duration −0.02 0.023 0.05 Inter-call Interval −0.99 −0.14 0.02 Maximum Frequency 0.11 −0.78 −0.19 Minimum Frequency 0.04 −0.21 0.98 Note: Most variation (99.9%) was retained by three principal components. the calling species are thus recommended for future studies. Nearly all (99.9%) of the variance was retained by three principal components. Table 3 lists the eigenvalues and variable loadings of each principal component. Natural history. Most individuals of Hyalinobatrachium mashpi sp. nov. were found on the underside of leaves among riverine vegetation ( Figs. 8 , 9 ). These frogs are difficult to observe because they are found 3–14 m above ground along steep creeks. Males have been observed calling in the months of April, May, June, August, September, October and November. Males that were guarding egg clutches while calling were observed during the rainy season (18 February 2019; 7 May 2021) and dry season (October 2014, June 2015, and August 2021). Examined egg clutches contain 31–34 eggs ( n = 3). A female with mature eggs visible through the skin was observed on 27 May 27 2015.

Generic placement. The new species is placed in the genus Hyalinobatrachium Ruiz-Carranza & Lynch (1991) , as modified by Guayasamin et al. (2009) , on the basis of morphological and molecular data. The molecular phylogeny ( Fig. 2 ) places the new species within the genus Hyalinobatrachium with high confidence. Phenotypically, the main diagnostic traits of Hyalinobatrachium are: (1) completely transparent ventral parietal peritoneum; (2) digestive tract and bulbous liver are covered by iridophores; (3) absent humeral spines; (4) white bones in life; (5) males call from the undersides of leaves; (6) females place the eggs on the undersides of leaves; (7) males provide extended parental care; and (8) tympanum with an orientation that places it almost on a horizontal plane (instead of a more lateral plane as observed in other glassfrog genera). All the aforementioned characteristics are present in Hyalinobatrachium mashpi sp. nov. We note that we have observed males on the same leaves as egg clutches for continuous days, but additional studies are necessary to confirm that these observations actually represent extended paternal care.