Brachycephalusdacnis, Toledo & Botelho & Carrasco-Medina & Gray & Ernetti & Gama & Lyra & Blackburn & Nunes & Muscat, 2024

Toledo, Luís Felipe, Botelho, Lucas Machado, Carrasco-Medina, Andres Santiago, Gray, Jaimi A., Ernetti, Julia R., Gama, Joana Moura, Lyra, Mariana Lucio, Blackburn, David C., Nunes, Ivan & Muscat, Edelcio, 2024, Among the world ’ s smallest vertebrates: a new miniaturized flea-toad (Brachycephalidae) from the Atlantic rainforest, PeerJ 12, pp. e 18265-e 18265 : 7-19

Brachycephalusdacnis sp. nov. ( Fig. 1 ).

Brachycephalus sp. ( Bornschein et al., 2021).

LSID: urn:lsid:zoobank.org:act:18F93070-2CAE-4817-8CCD-48C0C5DA5210 .

Etymology –The specific epithet name ‘ dacnis’ honors the Projeto Dacnis private reserve and NGO that has supported biodiversity research since 2010 in the municipalities of São José dos Campos, Miracatu and Ubatuba (where the new species was discovered), state of São Paulo, Brazil. The name is used as an invariable noun in apposition to the generic name.

Holotype – Adult male ( ZUEC-AMP 24982 ) collected and recorded (FNJV 51341) at Projeto Dacnis reserve , municipality of Ubatuba, state of São Paulo, Brazil, on 18 May 2021 by A. Mariano ( Fig. 1 ; Table 1 ). We also sequenced the DNA of this individual.

Paratypes – Eleven adults, only considered those confirmed by DNA and/or call traits: 03–09 August 2019 by R. Mitsuo and R.C. Menegucci (HCLP-A 267–68; heard, but not recorded); 10 March 2021 by E. Muscat and L.M. Botelho ( ZUEC-AMP 24978 ; sequenced) ; 05 May 2021 by L.M. Botelho ( ZUEC-AMP 24984 ; sequenced); 18 May 2021 by A. Mariano ( ZUEC-AMP 24979 ; recorded); 24 May 2021 by E. Muscat and L.M. Botelho ( ZUEC-AMP 24981 ; recorded and sequenced); 21, 22 February 2022 by L.M. Botelho ( ZUEC-AMP 25270 ; 25272; 25274–75; sequenced); 14 April 2022 by E. Muscat, L. M. Botelho, L.A.O.S Ferreira (adult male; ZUEC-AMP 25612 ; Fig. 2 ; recorded) . All paratypes were collected at the same locality as the holotype. Sex determination was not possible in most cases, except for the individual that was dissected (a female; see below) and the calling males.

Diagnosis –The new species is assigned to the genus Brachycephalus because of its miniature body size, fewer phalanges and toes than a typical frog, fingers and toe tips not expanded but apically pointed, and toes lacking circumferential grooves. The new species can be diagnosed from its congeners by the following combination of characters: (1) “leptodactyliform” body shape; (2) adult body length (SVL) smaller than 1 cm; (3) distinct and functional toes II and V; (4) presence of vestigial fingers Iand IV; (5) distinct iris; (6) absence of dark markings on the skin over the pectoral region; (7) dark black or pale brown marbled venter with small white blotches in preserved specimens; (8) and advertisement call composed of one or two multi-pulsed (3–7 pulses) note with dominant frequency between 8.01 and 8.44 kHz, note duration between 0.03– 0.08 s (when isolated), up to 0.41 s (when in pairs), and absence of attenuated notes.

Holotype description –Body slender and leptodactyliform ( Fig. 1 ; SVL = 7.55 mm); head as wide as long (1.78 mm; Table 1 ), and one-fourth of total body length (HW/SVL = HL/ SVL = 23.6%); snout short and slightly mucronate in dorsal view and between rounded and vertical in lateral view; nostrils oval, not protuberant and directed anterolaterally; canthus rostralis distinct; loreal region slightly concave; lips nearly sigmoid; eyes slightly protruding in dorsal and lateral view, and directed anterolaterally, 27.5% ED/HW and ED/HL (as HW = HL); tympanum indistinct; vocal sac not externally expanded; vocal slits present; tongue longer than wide, with posterior half free; choanae small and rounded; vomerine teeth absent. Forearm more than twice (2.3 times) length of upper arm length; hand length shorter than forearm or upper arm lengths; hands with all fingers distinct; fingers Iand IV externally vestigial; fingers II and III robust; tips of fingers Iand IV (when visible) rounded, fingers II and III pointed; relative finger length I ≤ IV <II <III; subarticular tubercles absent; inner and outer metacarpal tubercles absent. Leg relatively long, total leg length (TBL + THL + FL = 12.1 mm) 160% of SVL (7.6 mm); shank slightly shorter than thigh (SHL/SVL = 40.6% and THL/SVL = 47.5%); foot longer than tarsus and the shank; toe I externally absent; toes II, III, IV and V externally distinct and functional. Skin smooth without dermal ossifications. Measurements of holotype presented in Table 1 .

Color in life of the holotype –The general coloration of the dorsal body surface is yellowish-brown. Adark brown stripe extends laterally from the tip of the snout to the flanks. Dorsal surfaces of arms and legs interrupted by dark brown stripes. An x-shaped or dorsal mark is present. The lateral view of the body is dark brown; the region overlying the upper jaw is dark brown with distinct white spots. The ventral body surface is brown with small white blotches. The ventral surface of the hands and feet are dark brown with white blotches over each phalanx of the fingers and toes. The pupil is black, and the iris is bronze.

Color variation –Color and ornamentation patterns in this species vary considerably among individuals. Some frogs have a leaf-like dorsal coloration, while others have a rock-like dorsal coloration ( Fig. 3 ). We did not observe significant variation in the ventral pattern of the individuals.

Color in preservative of the holotype (less than 1 year in preservative)–The general coloration of the dorsal surface is reddish-brown ( Fig. 1 ). The white spots overlying the upper jaw and ventral region can become pale cream; no linea masculina is visible. The stripes on the arms and legs become more evident. The extremities of the fingers and toes become pale cream.

Internal anatomy: organs (ZUEC-AMP 25273; female)–After cutting the skin, the transparent abdominal muscles were observed to have a slight dark-brown coloration (melanocytes). Beneath the musculature, the liver and small intestine were also observed with similar dark-brown pigmentation. However, it was more prominent in the intestine than in the ventral musculature and liver ( Fig. 4A ). After removing the abdominal muscles and opening the pericardium, we observed the heart to have melanocytes scattered throughout the atria, ventricle, conus arteriosus, and truncus arteriosus. After cutting the pectoral girdle in the medial region of the weakly mineralized procoracoid cartilages, we observed that the dark pigmentation extends from the cranial region of the heart to the carotid and systemic arteries. We did not see this pigmentation in the pulmocutaneous artery ( Figs. 4B, 4C ). Aslight extension of this dark pigmentation occurs in the musculature lateral to the systemic and carotid arteries ( Figs. 4B, 4C ). Near the bifurcation of the internal and external jugular arteries and between the systemic and carotid arteries, there were two black, ovoid parathyroid glands on each side ( Fig. 4C ). Less black pigmentation was observed in the thoracic region; however, a few small spots were observed close to the white eggs and in the oviduct. The oviduct was transparent and between the left and right oviduct two mature eggs were observed, and the oviducts joined near the pubis ( Figs. 4D, 4E ).

Internal anatomy: skeleton (ZUEC-AMP 24978; undetermined sex). The skeleton of B. dacnis is broadly similar to that of other small species in the genus, including by lacking osteoderms and having unornamented dermal skull bones that are not co-ossified to the skin. This individual ( Fig. 5 ) is osteologically mature with ossified mesopodials and complete, ossified distal long bones; the specimen is missing the right leg, which was removed for DNA analysis.

The skull is compact, slightly longer than wide, and lacking ornamented dermal roofing bones. The frontoparietals, sphenethmoids, prootics, exoccipitals, and parasphenoid are incompletely synostosed with each other and often have clear separations between the bones. There are well-defined and bony margins to both the optic fenestrae and prootic foramina. The premaxillae are broad, clearly separated from one another, and have teeth; each has a robust pars dentalis, and a robust alary process that is approximately as tall as wide and widely separated from the nasal. In ventral view, the maxillae are nearly straight and bear odontoids on the anterior third. The quadratojugals are thin and angled dorsally, and do not articulate with the maxillae. The pterygoids are slender, each with a long, straight, and forked anterior ramus that approaches but does not articulate with the adjacent maxilla, a short, subtriangular posterior ramus adjacent to but not articulating with the ventral ramus of the squamosal, and a short, broad medial ramus that is well separated from the prootic. Distinct neopalatines are not observable and likely not synostosed to the sphenethmoid. Distinct, small c-shaped vomers are present bordering the anterior margins of the choanae. Tall, thin, and curved septomaxillae are present at the anterior margin of the nasal capsule. The parasphenoid is triradiate with a long, rectangular anterior ramus. The squamosals are robust, and each has a prominent zygomatic ramus that is expanded ventrally as a sheet-like flange along its anterior margin and a long, slender posterior ramus that is adjacent but not fused to the prootic. The portions of the prootic are not completely synostosed and the broad fenestra ovalis contains a small, irregularly shaped operculum; a stapes (or columella) is not present. The posteromedial processes of the hyoid are ossified and slender. The arytenoid cartilages are not obviously mineralized.

There are eight distinct, procoelous, non-imbricating presacral vertebrae that arenot fused to one another. The atlas lacks transverse processes and cotyles that are widely separated, a distance greater than the posterior diameter of the atlas centrum. Transverse processes of presacral II are distinctly shorter than those of III–VIII. The sacrum is procoelous with stout transverse processes. Asesamoid at the distal transverse process of the sacrum is present at its articulation with the ilium. The urostyle bears a tall dorsal ridge that decreases in height posteriorly along the anterior two-thirds of the urostyle.

The pectoral girdle is not heavily ossified. The coracoids are expanded medially but do not meet at the midline. The coracoids are thin and articulate at the midline with a small, irregularly shaped element that may represent the omosternum. The scapula is short with a well-developed anterior process and a prominent supraglenoid foramen. The epicoracoid and procoracoid cartilages are weakly mineralized and do not form solid elements synostosed to the coracoid and clavicle.

The pelvic girdle is a robust element composed of synostosed ilium, pubis, and ischium; the publis and ischium appear to be incompletely ossified. The circular acetabulum is incompletely ossified but has well-defined margins. The shaft of the ilium is stout and mostly straight in both lateral and dorsal views, and has both a well-developed dorsal protuberance and a weakly developed dorsal crest. There is a broad ventral acetabular expansion comprising both the ilium and ossified pubis, but it is incompletely ossified.

The forearm is somewhat shorter than the humerus. The distal carpals (Element Yand II–V) are fused. The radiale and ulnare are large and subequal in size. The phalangeal formula for the manus is 1–2–3–1. There appears to be a single, minute ossified prepollex. There is one small palmar sesamoid. The tips of the terminal manual phalanges are arrow-shaped in digits II and III but blunt and globular in Iand IV. The tibiofibula and femur are similar in length, and there is a small ossified fabella sesamoid near their articulation. There are two large distal tarsals and an ossified tarsal sesamoid near their articulation with the tibiofibula. The phalangeal formula for the pes is 1–2–3–4–2 and there is a single small ossified prehallux. Medially, there are two small plantar sesamoids, whereas laterally there are two large plantar sesamoids. The tips of the terminal pedal phalanges are arrow-shaped in digits I–IV but blunt in V.

Calls and calling activity –The new species is vocally active throughout the year, with peaks of activity when the humidity was higher than 90% (> 50 males calling). The calling activity was mostly detected during daytime, with a higher number of individuals calling between 6:00 and 11:00 AM, from 3:00 to 6:00 PM, or anytime of the day after heavy rains. From August to April, we also heard calls during the night. Males frequently ceased calling for a few minutes after detecting the researcher’ s approach.

The advertisement call of B. dacnis ( Fig. 6B ) can be simple or complex, respectively containing one or two multipulsed (three to seven pulses) notes, emitted in mean intervals of 2.1 ± 0.32 s (1.67–2.69). Out of the 17 recordings made, only two presented calls with two notes. The mean note duration was 0.05 ± 0.01 s (0.03–0.08), and the call with two notes could reach up to 0.41 s. The call frequency varied from 6.89 (minimum measured frequency) to 8.96 (maximum measured frequency) kHz, while the mean dominant frequency was 8.22 ± 0.1 kHz (8.01–8.44) ( Table 2 ).

Phylogenetic inference and genetic distances –Our phylogenetic analysis supported the distinction of B. dacnis from all the other members of the genus. All the specimens are clustered within a strongly supported clade (posterior probability, PP = 1.0). The new species is recovered as the sister species of B. hermogenesi , and this clade sister of B. sulfuratu s + B. pernix species group ( Figs. 6C; S 1 , Table S4). The phylogenetic inference recovered B. pulex sister to two diverse clades, one containing (( B. puri + two putative distinct species) + ( B. didactylus + ( B. clarissae + ( B. vertebralis species group + B. ephippium species group)) and the other containing (( B. hermogenesi + B. dacnis ) + ( B. sulfuratus + B. pernix species group) ( Figs. 6C, S 1 , Table S4). The 16S uncorrected genetic distance between the new species and B. hermogenesi ranged from 6.8% to 7.8% (mean 7.1%; Table S1 ). These distances are higher than distances found between species of B. ephippium and B. vertebralis species groups (3.2% to 5.9%, Table S1 ). The relationships between species within each of the species groups containing the bright colored and bufoniform species ( B. ephippium , B. vertebralis and B. pernix species groups) are in general low supported ( Fig. S1 ). Furthermore, some species were not recovered as monophyletic (i.e., B. pitanga and B. guarani ; B. mariaeterezae and B. olivaceus ; Fig. S1 ) and the genetic distances estimated were extremely low, less than 1% ( Table S1 ).

Comparison with other species – Brachycephalus dacnis is distinguished from all species of the B. ephippium ( B. atelopoide , B. darkside , B. ephippium , B. garbeanus , B. ibitinga , B. margaritatus, and B. rotenbergae ), B. vertebralis ( B. alipioi , B. bufonoides , B. crispus , B. guarani, B. herculeus , B. nodoterga , B. pitanga , B. toby , and B. vertebralis ), and B. pernix ( B. actaeus , B. albolineatus , B. auroguttatus , B. boticario , B. brunneus , B. coloratus , B. curupira , B. ferruginus , B. fuscolineatus , B. izecksohni , B. leopardos , B. mariaeterezae , B. mirissimus , B. olivaceus , B. pernix , B. pombali , B. quiririensis , B. tabuleiro , B. tridactylus , and B. verrucosus ) species groups by the combination of its small adult body size, a leptodactyliform body form (vs. bufoniform), and a dark-brown coloration (vs. vibrant coloration such as greenish, yellowish, orange or reddish) ( Ribeiro et al., 2015; Lyra et al., 2021). Brachycephalus dacnis is also distinguished from species of the B. ephippium and B. vertebralis species groups by its smooth skin, distinct frontoparietals, nasals, vomers, and parasphenoid that arenot synostosed via a bony sphenethmoid ( Figs. 5 , 7 ), presence of a quadratojugal, robust anterior process of the pterygoid, lack of hyperossification in the skull, distinct coracoid and clavicle and lack of mineralized epicoracoid cartilages, lack of vertebral fusions, lack of bony plates in the axial skeleton, and presence of phalanges on all digits ( Lyra et al., 2021; Clemente-Carvalho et al., 2009, 2012; Condez et al., 2014, 2016; Haddad et al., 2010; Pombal, 2010). From the other six leptodactyliform species, a.k.a. flea-toads ( B. didactylus , B. clarissae , B. hermogenesi , B. pulex , B. puri , and B. sulfuratus ), B. dacnis is distinguished by skin texture, toe morphology, coloration, and some skeletal traits. From B. clarissae , B. dacnis is distinguished by having a smooth dorsal texture (granular in B. clarissae ), a distinct and functional toe V (vestigial in B. clarissae ), lacking synostosis among cranial bones (sphenthmoids, parasphenoid, and frontoparietals synostosed in B. clarissae ; Figs. 5 , 7 ), phalanges on pedal digits Iand V (lacking in B. clarissae), lacking mineralized epicoracoid cartilages (heavily mineralized and fused with pectoral bones in B. clarissae ), and by having a distinct iris (indistinct in B. clarissae ) ( Folly et al., 2022). Brachycephalus dacnis has a distinct, and functional toe V (like B. sulfuratus and B. hermogenesi ), whereas it is vestigial in B. didactylus ( Haddad et al., 2010; Giaretta & Sawaya, 1998). Neither B. dacnis nor B. hermogenesi lack phalanges on the toes, whereas both B. didactylus (pedal digit count: 0, 2, 3, 4, 0) and B. sulfuratus (0, 2, 3, 4, 2) lack phalanges on one or more digit. In B. didactylus , there is a trapezoidal chest mark, which is absent in B. dacnis ( Almeida-Silva et al., 2021) , as well as a fully synostosed braincase ( Fig. 7 ), a wire-like anterior process of the pterygoid, no quadratojugal, heavily mineralized epicoracoid cartilages fused with the pectoral bones, fused presacral vertebrae VI, VII, and VIII, and no phalanges on manual digit Ior IV. From B. sulfuratus , B. dacnis is distinguished by the absence of an inverted V-shaped chest mark ( Condez et al., 2016).

Although this trait varies in B. sulfuratus (see Bornschein et al., 2021), it still can distinguish some individuals. From B. sulfuratus , B. dacnis is distinguished by its thin quadratojugal ( Fig. 7 ), and medially expanded coracoid. From B. puri , B. dacnis is distinguished by having a distinct toe II (reduced in B. puri ), by its dark black or pale brown marbled venter with small white blotches (uniformly brown in B. puri ), and by sometimes having an X-shaped dorsal mark (inverted V-shaped in B. puri ) ( Almeida-Silva et al., 2021). From B. pulex , B. dacnis is distinguished by the presence of vestigial fingers Iand IV (absent in B. pulex ), the presence of distinct and functional toe II and V (absent or vestigial, respectively, in B. pulex ), and lacking a chest mark (presence of an inverted V-shaped chest mark in B. pulex ) ( Almeida-Silva et al., 2021). Brachycephalusdacnis and B. hermogenesi are morphologically similar, including in their skeletons ( Figs. 5 , 7 ), but can be recognized as distinct through acoustic and genetic traits.

The advertisement call of B. dacnis is distinguished from B. hermogenesi and B. sulfuratus by generally having fewer notes per call (one or two notes per call in B. dacnis , 2–6 in B. hermogenesi , 4–6 in B. sulfuratus ), shorter note duration (≤ 0.08 s in B. dacnis ; ≥ 0.17 s in B. hermogenesi , and ≥ 1.5 s in B. sulfuratus ), and shorter call duration (≤ 0.08 s in B. dacnis in calls with one note and ≤ 0.41 s in calls with two notes; ≥ 0.41 s in B. hermogenesi, and ≥ 0.13 s in B. sulfuratus ). Brachycephalus dacnis generally presents fewer pulses per note (3–7) than B. sulfuratus (7–11; Condez et al., 2016), and higher than B. hermogenesi (1–3) (see also Bornschein et al., 2021). The absence of attenuated notes (sensu Bornschein et al., 2021) in the calls of B. dacnis , distinguishes it from B. hermogenesi (which can present attenuated advertisement call notes) (see Bornschein et al., 2021). The dominant frequency of the advertisement call of B. dacnis overlaps with that of B. hermogenesi, but it is higher than that of B. sulfuratus . However, the minimum and maximum frequencies of the call of B. dacnis are lower and do not overlap with those of B. hermogenesi . Call parameters of these three species are presented in Table 2 .

Natural history –We observed Brachycephalus dacnis occupying areas with dense vegetation, tall canopy, reduced sunlight incidence, a thick layer of litter, and moist soil. These areas lie between small hills, forming a valley. The leaf litter along the slopes lies in small clusters at the base of trees and creeping plants. These areas have a small stream nearby with a substrate composed of rock and sand. Water flow intensity varies with precipitation and can soak the adjacent soil. The new species also occurs in an area characterized by thinner vegetation, mainly shrubs, with more solar incidence, and a thick litter layer, and consequently, less moisture. In these sites, finding frogs was difficult because of their small size, cryptic coloration, cryptozoic habits, and the behavior of ceasing calling with any disturbance in their surroundings. After thorough searches, we found the individuals either between juçara palm (Euterpes edulis) roots or under rotten trunks. In these areas, B. dacnis was syntopic with B. hermogenesi . When approached, some individuals occasionally exhibited mouth-gaping ( Fig. 3C ). We also observed a

0.7 mm Brachycephalus cf. dacnis individual (not collected–but likely a B. dacnis , as no B. hermogenesi was found in the same area) jumping 21.8 cm (31 times its SVL) (ZUEC-VID 796).

We identified the stomach contents of two individuals. During dissection, we found one individual (ZUEC-AMP 25273) with four items in its stomach, identified to order: one Collembola, one Coleoptera , one Hymenoptera (Hexapoda), and one Acari (Sarcoptiformes) ( Fig. S2 ). In the microCT scan of another individual (ZUEC-AMP 24978), we identified one item, an Anobiinae beetle ( Coleoptera : Ptinidae ) ( Fig. S2 ).

Distribution – Brachycephalus dacnis is known from its type locality, at Projeto Dacnis private reserve in Ubatuba, state of São Paulo (SP), Brazil, and two additional localities, based on acoustic records previously published in Bornschein et al. (2021): Corcovado, municipality of Ubatuba, SP, and Trilha do Corisco, municipality of Paraty, Rio de Janeiro ( Fig. 6 ).

Figures

Figure 1 Brachycephalus dacnis holotype (ZUEC-AMP 24982) adult male, SVL = 7.55 mm. (A) Dorsal and (B) ventral views. Full-size DOI: 10.7717/peerj.18265/fig-1

Figure 2 Brachycephalus dacnis paratype (ZUEC-AMP 25612) adult male, SVL = 7.89 mm. (A) Dorsal view, (B) ventral view, (C) ventral view of hand, (D) ventral view of feet. Full-size DOI: 10.7717/peerj.18265/fig-2

Figure 3 Brachycephalus dacnis paratype individuals found at the Projeto Dacnis private reserve, municipality of Ubatuba, state of São Paulo, Brazil. (A) ZUEC-AMP 25272; (B) ZUEC-AMP 25274; (C and D) ZUEC-AMP 25275 displaying mouth-gaping behavior and the same individual on top of one herpetologist’s fingertip. Full-size DOI: 10.7717/peerj.18265/fig-3

Figure 4 Ventral dissection of Brachycephalus dacnis (ZUEC-AMP 25273) with the highlighted structures. (A) Abdominal muscles, (B) Liver, (C) Small intestine; (D) Atrium, (E) Ventricle, (F) Conus arteriosus, (G) Truncus arteriosus, (H) Carotid artery, (I) Systemic artery, (J) Parathyroid gland, (K) Eggs, (L) Oviduct. Full-size DOI: 10.7717/peerj.18265/fig-4

Figure 5 Skeletal anatomy of Brachycephalus dacnis (ZUEC-AMP 24978; undetermined sex) based on CT-scans. (A) The entire skeleton, with insets of (B) the skull (in dorsal, ventral, and right lateral views), (C) right hand and (D) left foot (both in dorsal view), (E) vertebral column (in dorsal view), (F) pectoral girdle (in ventral view), and (G) pelvic girdle (in right lateral view). Abbreviations: c, carpus; cl, clavicle; cor, coracoid; dt, distal tarsals; ex, exoccipital; fp, frontoparietal; il, ilium; isc, ischium; m, maxilla; mdI, manual digit I; mdV, manual digit V; min, mineralized cartilage; n, nasal; op, operculum; pd I, pedal digit I; pd V, pedal digit V; pm, premaxilla; pr, prootic; ps, parasphenoid; pt, pterygoid; pub, pubis; qj, quadratojugal; sac, sacrum; sc, scapula; ses, sesamoid; sm, septomaxilla; sp, sphenethmoid; sq, squamosal; ur, urostyle; v, vomer. Full-size DOI: 10.7717/peerj.18265/fig-5

Figure 6 Geographic distribution, advertisement calls and phylogeny. (A) Type localities of Brachycephalus dacnis (red star) and B. hermogenesi (black star), and additional distribution records of B. dacnis (red triangles, based on call recordings published in Bornschein et al., 2021). This map was created using the free and open-source software QGIS, utilizing publicly available base data sources. (B) Brachycephalus dacnis holotype’s advertisement call (FNJV 51341), recorded at Projeto Dacnis private reserve (type locality), Ubatuba, São Paulo (left) and B. hermogenesi topotype’s advertisement call (FNJV 51343), recorded at Picinguaba (type locality), Ubatuba, São Paulo (right). (C) Schematic representation of the Bayesian phylogenetic inference of Brachycephalus based on the mitochondrially encoded 16S rDNA fragment. Black dots indicate fully supported clades (posterior probability, PP => 99). White dots indicate PP> 0.95. Node supports below pp = 0.95 are not shown. Complete tree in the supplement (Fig. S1). Full-size DOI: 10.7717/peerj.18265/fig-6

Figure 7 Selected skulls of other Brachycephalus species. Cranial elements labeled are discussed in comparisons of B. dacnis to other species. Top row is the leptodactyliform “flea toads”, including B. hermogenesi (sister species of B. dacnis), B. clarissae, B. sulfuratus, and B. didactylus. Bottom row contains representatives of the three named species groups: pernix group (B.albolineatus, B. pernix), vertebralis group (B. vertebralis), and ephippium group (B. ephippium). Museum catalog numbers are listed below each species name. Abbreviations: fp, frontoparietal; n, nasal; ps, parasphenoid; pt, pterygoid; qj, quadratojugal; sp, sphenethmoid; sq, squamosal; v, vomer. Full-size DOI: 10.7717/peerj.18265/fig-7

Tables

Table 1 Measurements of the holotype and adult paratypes (one adult that was dissected for internal anatomy was not measured) of Brachycephalus dacnis. Values presented as mean ± standard deviation (minimum−maximum) in millimeters.

Trait	Holotype	Adult paratypes (n = 10)
SVL	7.55	8.29 ± 1.03 (6.95–9.90)
HL	1.78	2.59 ± 0.45 (1.87–3.18)
HW	1.78	2.97 ± 0.55 (1.82–3.63)
END	0.37	0.48 ± 0.23 (0.18–1.02)
IOD	1.42	1.48 ± 0.32 (0.84–2.10)
ED	0.49	0.94 ± 0.24 (0.50–1.30)
ND	0.31	0.27 ± 0.08 (0.16–0.40)
UAL	1.30	1.57 ± 0.22 (1.30–2.02)
FAL	2.95	2.13 ± 0.30 (1.80–2.70)
HAL	1.00	1.28 ± 0.23 (0.92–1.63)
TBL	3.59	3.82 ± 0.48 (2.98–4.50)
THL	4.00	3.77 ± 0.51 (3.25–5.00)
FL	4.50	6.10 ± 0.41 (5.60–6.75)
fL	2.75	3.16 ± 0.47 (2.35–4.00)

Table 2 Advertisement call traits described for three flea-toads species from Brachycephalus genus. Call traits present as mean ± standard deviation (range).

	Brachycephalus dacnis six individuals (including the holotype) Dacnis, Ubatuba	Brachycephalus dacnis Corcovado, SP and Trilha do Corisco, RƖ	Brachycephalus hermogenesi five individuals Ubatuba	Brachycephalus hermogenesi Several locations	Brachycephalus sulfuratus four topotypes São Francisco do Sul	Brachycephalus sulfuratus Several locations
Call duration (s)	0.05 ± 0.01	-	0.56 ± 0.14	-	1.70 ± 0.10	-
	(0.03–0.41)		(0.41–0.78)		(1.50–2.30)
Inter-call interval (s)	2.1 ± 0.06	-	2.70 ± 0.15	-	4.5 ± 1.7	-
	(1.67–2.69)		(2.52–2.91)		(3.1–7.4)
Number of notes per call	1–2	–	2–6	–	4–6	–
Presence of attenuated notes *	No	No	Yes	Yes	No	No
Note duration (s)	The same as call	-	0.25 ± 0.04	-	0.17 ± 0.01	-
	duration		(0.17–0.27)		(0.13–0.21)
Inter-note interval (s)	n/a	-	2.70 ± 0.15	-	n/a	-
			(2.52–2.91)
Note repetition rate	19.58 ± 1.84	-	4.12 ± 0.48	-	0.2 ± 0.1	-
(notes/s)	(12.55–29.41)		(3.58–4.88)		(0.2–0.3)
Number of pulses per note	3–7	1–16	1–2	2–3	7–9	2–14
Minimum frequency	7.27 ± 0.12	-	7.94 ± 0.08	-	4.9 ± 0.3	-
(Frequency 5%) (kHz)	(6.89–7.50)		(7.84–8.01)		(4.5–5.2)
Dominant frequency	8.22 ± 0.10	-	8.38 ± 0.18	-	6.6 ± 0.1	-
(kHz)	(8.01–8.44)		(8.18–8.61)		(6.5–6.7)
Maximum frequency	8.62 ± 0.03	-	8.91 ± 0.19	-	9.1 ± 0.2	-
(Frequency 95%) (kHz)	(8.44–8.96)		(8.70–9.13)		(9.0–10.7)
Frequency bandwidth	1.35 ± 0.12	-	0.97 ± 0.18	-	n/a	-
(90%) (kHz)	(0.94–1.89)		(0.78–1.29)
References	Present study	Bornschein et al. (2021)	Present study	Bornschein et al. (2021)	Condez et al. (2016)	Bornschein et al. (2021)