Contributions to Zoology 89 (2020) 14-73 CTOZ

brill.com/ctoz

Descriptive skeletal anatomy of Blommersia transmarina (Amphibia: Anura: Mantellidae) from the Comoro Islands

Javier H. Santos-Santos Department of Biogeography and Global Change (BGC-MNCN-CSIC), Museo Nacional de Ciencias Naturales, C/José Gutiérrez Abascal 2, 28006, Madrid, Spain Department of Animal Biology, University of Barcelona, Avenida Diagonal 645, 08028, Barcelona, Spain [email protected]

Mireia Guinovart-Castán Department of Biogeography and Global Change (BGC-MNCN-CSIC), Museo Nacional de Ciencias Naturales, C/José Gutiérrez Abascal 2, 28006, Madrid, Spain

David R. Vieites Department of Biogeography and Global Change (BGC-MNCN-CSIC), Museo Nacional de Ciencias Naturales, C/José Gutiérrez Abascal 2, 28006, Madrid, Spain [email protected]

Abstract

Mantellid frogs present an extensive adaptive radiation endemic to Madagascar and Comoros, being the subfamily Mantellinae the most morphologically and ecologically diverse. The Mantellinae present key innovative evolutionary traits linked to their unique reproductive behavior, including the presence of femoral glands and a derived vomeronasal organ. In addition, previous studies pointed to size differ- entiation in playing an important role in species’ dispersal capacities and shaping of their geographic ranges. Despite the high phenotypic variation observed in this clade, to date an exhaustive morphologi- cal analysis of their anatomy has still not been performed, much less in relation to internal structures. Here, we present a comprehensive skeletal description of a mantellid species, Blommersia transmarina, from the island of Mayotte in the Indian Ocean, which has potentially undergone a process of moderate gigantism compared to other Blommersia species. We describe its intraspecific skeletal variation utilizing non-destructive volume renderings from μCT-scans, and characterize the presence of sexual dimorphism and size covariation in skeletal structures. Notably, we found numerous signs of hyperossification, a novel structure for mantellids: the clavicular process, and the presence of several appendicular sesamoids. Our findings suggest that skeletal phenotypic variation in this genus may be linked to biomechanical function for reproduction and locomotion.

© SANTOS-SANTOS et al., 2020 | doi:10.1163/18759866-20191405 This is an open access article distributed under the terms of the cc-by 4.0 License. Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 15

Keywords

Anura: Mantellidae – computed tomography – hyperossifcation – intraspecific variation – osteology – size variation

Introduction ­nauticus and Blommersia transmarina (Glaw et al., 2019). Mayotte consists of a main island, The Mantellidae represents the most species- Maore, and a small island, Pamanzi; with rich and ecologically diverse anuran fam- transmarine colonization having occurred ily endemic to the islands of Madagascar and exclusively on Maore. B. transmarina appears Mayotte (Glaw & Vences, 2007). It is divided to have likely having ­undergone a process of into three subfamilies: Mantellinae (Laurent, moderate gigantism (Glaw et al., 2019), a com- 1946), Laliostominae (Vences & Glaw, 2001), mon phenomenon in oceanic islands (Daugh- and Boophinae (Vences & Glaw, 2001). Exten- erty et al., 1993; Lomolino, 2005; Li et al., 2011). sive work (Guibé, 1978; Scott, 2005; Glaw & Several recent works studied anatomical Vences, 2006; Vieites et al., 2009) has been car- subsystems of various species within Man- ried out in view of resolving their systematics, tellidae, including external morphology (e.g., biogeography, and phylogenetic relationships Blommersia Vences et al., 2010; Pabijan et al., to elucidate their evolutionary history and 2011; Gephyromantis Scherz et al., 2017a; shed light on the nature of active speciation Guibemantis Vences et al., 2015; Mantidacty- processes. In this respect, variation in body lus Vences et al., 2002; Tysingymantis Glaw size has been found to influence species’ et al., 2006), femoral glands (Vences et al., range sizes and biogeographic setting, linking 2007; Altig,­ 2008), intercalary elements (Man- smaller body sizes to higher clade diversity zano et al., 2007), etc. However, there is cur- and geographically closer and more fragment- rently no study of skeletal data that makes ed ranges (Wollenberg et al., 2011; Pabijan reference to size differentiation, especially in et al., 2012). view that body size is known to influence dis- The Mantellidae is phylogenetically nested persal and evolutionary rates within the genus within Indian frogs and its most recent com- (Wollenberg et al., 2011; Pabijan et al., 2012). mon ancestor is estimated to have dispersed Since B. transmarina is the largest Blommersia from this landmass and colonized the island species identified to date (Glaw et al., 2019), it of Madagascar in the Early Cretaceous be- constitutes the perfect starting point to set a tween 87–76 Mya (Million years ago) (Crottini base to study the processes of miniaturization et al., 2012; Samonds et al., 2013). In addition, in the internal anatomy of mantellids on the there have been two more recent colonization grounds that miniaturization can lead to re- events by oceanic dispersal that are estimated duced ossification of post-metamorphic skel- to have occurred between 8–6 Mya (Vences et etal elements, hyperossification, and presents al., 2003; Crottini et al., 2012) from mainland more structural trade-offs than enlargement Madagascar to the neighboring island of May- (Hanken, 1993; Yeh, 2002; Hall, 2005; Pérez- otte in the Comoros, 300 km off the northwest Ben et al., 2018). coast of Madagascar, involving Boophinae and Here, we present the first comprehensive Mantellinae members of this family, which skeletal description of a species within the have recently been described as Boophis­ Mantellidae (but see Guibé, 1978; Scott, 2005),

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 16 Santos-Santos et al. elaborating precise anatomical descriptions fixed with cotton and submerged in 70% etha- of all skeletal elements, and define the level of nol within 50 mL polypropylene falcon tubes intraspecific phenotypic variation (n = 10) of before CT-scanning in a Nikon XT H-160 sys- highly variable elements. Anatomical descrip- tem [reconstructed voxel size (μm) = 29.5–42 tions are based on high-resolution rendered (isometric); X-ray (kV) = 53–56; X-ray (μA) = skeletons obtained from μCT (i.e., computed 172–188; Projections = 1800; voxels = 1008] at tomography) volume scans of individuals the internal Service of Non-destructive Tech- recently captured in Mayotte, Comoros. Ap- niques of the MNCN-CSIC. The CT-scans proximations of non-calcified structures were were reconstructed with CT Pro 3D software. inferred from additional CT-scans of stained Individual skeletons were volume rendered specimens (see Santos-Santos, 2019). Descrip- and visualized in Avizo® version 9.2 software tions are made with elements referenced in a with volrenWhite and physics colormaps to primary anterior (rostral) to posterior (cau- observe bone densities. Images were taken dal) axis and a secondary proximal (medial) in perspective mode in Avizo® to capture the to distal (lateral) axis. With this anatomical largest amount of skeletal features in each study we aim to create a reference for future view. Additional CT-scans of priorly soft- comparative studies with other mantellids tissue-stained specimens were performed to species, as well as explore their phenotypic corroborate the position of non-calcified ana- evolutionary processes, including miniatur- tomical elements; these were performed with a ization / size variation in the family. custom setup [reconstructed voxel size (μm) = 30–35 (isometric); X-ray (kV) = 80; X-ray (μA) = 375; Projections = 1801; voxels = 1166] at the Material and methods Center for X-ray Tomography (UGCT) in col- laboration with the Evolutionary Morphol- A total of 10 (five male and five female) adult ogy of Vertebrates Lab of Ghent University, Blommersia transmarina individuals were re- Belgium (see Santos-Santos, 2019). The speci- trieved from the collections at the National men MNCN50446 was used as a reference Museum of Natural History (MNCN-CSIC) for the generalized anatomical description. in Madrid, Spain (table 1). Individuals were The .stl file of its skeletal segmentation is

Table 1 Data and external snout–vent length (SVLext) in mm of the Blommersia transmarina individuals used for the current study.

Collection n° Field n° Sex SVLext Locality MNCN50430 DRV6835 F 29.28 Mont Choungui MNCN50431 DRV6848 F 26.5 Mont Combani MNCN50432 DRV6805 F 29.13 Mont Combani MNCN50433 DRV6813 F 30.38 Mont Tsapere MNCN50435 DRV6832 M 25.5 Mont Bénara MNCN50436 DRV6833 M 27.5 Mont Bénara MNCN50437 DRV6841 M 24.64 Mont Bénara MNCN50439 DRV6836 M 27 Mont Bénara MNCN50446 DRV6807 M 29 Mont Tsapere N/A DRV6851 F 24.5 Mont Combani

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 17 available on the Spanish National Research Results Council (CSIC) repository (https://digital.csic .es/handle/10261/190938). Skull In addition, a total of 123 linear distances There are two types of bone in the cranium: (see Appendix 1) were measured on the ren- endochondral (i.e., develops from osteoblasts dered skeletons of all individuals to obtain within cartilage) and dermal or membranous the intraspecific range of variation of skel- (i.e., develops intramembranously inside con- etal characters. Descriptive statistics were nective tissue) (Hall, 2005). The neurocranium obtained and sexual dimorphism and size represents an inverted T-shaped box: the lon- covariation evaluated through analysis of (co) gitudinal branch going anteroposteriorly from variance (α = 0.05) in IBM SPSS Statistics for the nasal to the auditory region, and each tip Windows version 24.0. Of the three snout- of the transverse branch ending at an auditory vent length (SVL) measurements taken (i.e., capsule (fig. 1). SVLext, SVLsk, SVLsum; Appendix 1), we deter- mined which one covaried best with skeletal Endochondral neurocranium distances by including all SVL measurements Sphenethmoid. The sphenethmoid is a central as covariates in univariate three-way analy- element of the skull. It is ossified anteriorly, ses of covariance (ANCOVA). Pelvis length encircling the anterior part of the brain, con- (pelvL) was also explored as a covariate, and tributing to the posterior walls of the nasal included with the best resulting SVL covariate capsule, and forming the anterior margin of from the prior analysis in univariate two-way the optic foramen. The bone extends further ANCOVAs with skeletal measurements to de- ventrolaterally (fig. 2), creating a pair of small termine which covaried best with each skel- alae that penetrate the orbitonasal foramen etal linear distance. slightly to either side of the parasphenoid. On

FIGURE 1 Dorsal view of the Blommersia transmarina skull [MNCN50446 male; volrenWhite (A) and physics colormaps (B), grey [density] values: 12000–55500]. Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 18 Santos-Santos et al.

FIGURE 2 Ventral view of the Blommersia transmarina skull [MNCN50446 male; volrenWhite (A) and physics colormaps (B), grey [density] values: 12000–55500].

the ventral side there are exostoses (Duellman frontoparietal fontanelle. In addition, there & Trueb, 1994), a sign of hyperossification evi- is a diamond-shaped fenestra (i.e., otoccipi- denced by the presence of superficial sculpted tal fenestra) between the prootics anterior to patterns. Anteriorly it presents two bilaterally­ the dorsal fusion point of the otoccipital. The symmetric olfactory foramina that open to three fenestrae are completely covered by the the internasal septum (fig. 3). Posterodor- frontoparietals, except for the posterior tip of sally the sphenethmoid separates medially the otoccipital fenestra. Below the protuber- forming the frontoparietal fontanelle, which ance at the posterolateral corner of the fronto- is completely covered by the frontoparietals, parietal, the prootic presents two prominent except for its anterior tip, and extends dorsally epiotic eminences (Trueb, 1968a) that extend for slightly more than half their length (fig. 1). both antero- and postero-laterally (figs. 1 and Prootics and Exoccipitals. The prootics lie at 4); the latter (i.e., opisthotic sensu Ecker, 1889) a distance posterior to the sphenethmoid and terminates in the prominentia ducti semi- expand laterally to cover the auditory cap- circularis posterioris (Laloy et al., 2013). This sules both anteriorly, forming the posterior structure is narrowly separated from the an- margin of the optic foramen, and posterodor- teromedial margin of the suprascapula. sally (fig. 4). The prootics in B. transmarina The exoccipitals are fused completely present signs of hyperossification evidenced ­ventrally and dorsally in a ring-like structure in an exostosed surface and fusion with each that unites dorsoanteriorly with the pro­ other and with the exoccipitals to form a sin- otics, ventroanteriorly with the parasphenoid, gle massive element (i.e., otoccipital; Clack, flanks the foramen magnum, and forms the 2001). They unite dorsomedially at their poste- occipital condyles posteriorly (fig. 3). Imme- riormost point, where they also fuse with the diately anterior to each occipital condyle is exoccipitals (fig. 1). Each prootic presents an- a large ­circular foramen (i.e., foramen jugu- teriorly a large dorsal fenestra with a rostrolat- lare) that corresponds to the exit of cranial eral extension that is discontinuous with the nerves (C.NN.) IX (glossopharyngeal) and X Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 19

FIGURE 3 Medial view of the right side of the Blommersia transmarina skull [MNCN50446 male; volrenWhite (A) and physics colormaps (B), grey [density] values: 12000–55500].

FIGURE 4 Lateral view of the right side of the Blommersia transmarina skull. The skull presents a dorsoventral inclination in the caudal direction. Its left half and exoccipitals have been clipped to facilitate visualization [MNCN50446 male; volrenWhite (A) and physics colormaps (B), grey [density] values:

12000–55500]. Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 20 Santos-Santos et al.

(vagus), which is separated laterally by a thin greatly between individuals, and due to the bone wall from another slightly more ovoid resolution of our CT-scans (~30 μm) the foramen located under the auditory cap- identification of homologous­ structures sule that corresponds to the exit of C.N. VIII at the nanoscale was not possible (but see (vestibulocochlear). Mason et al., 2015). However, there were Auditory/Otic capsules. The tympanic mem­ some recurrent structures across individu- brane in B. transmarina is located at the level als (fig. 5): The shelf-like projection of the of the otic ramus of the squamosal bone. papilla amphibiorum recess was visible Below the posterior otic ramus of the squa- situated ­rostromedially on the utriculus; in mosal lies the ossified portion (i.e., stylus) some individuals (e.g., MNCN50446) one or of the columella (i.e., middle ear ossicle) two additional round diverticula (the dor- with an expanded proximal end (i.e., foot- sal larger when both present) were visible plate), which articulates medially with the extending medially from the amphibiorum membrane of the oval window (i.e., fenes- recess (Richards, 2006). Medially at the lev- tra ovalis), anterior­ to the operculum (fig. 4). el of the posterior perilymphatic duct and Ventroposteriorly covering the fenestra ovalis separated narrowly lateral to the sacculus lies the operculum, a lightly calcified carti- lies the lagena (Duellman & Trueb, 1994), a laginous element (fig. 2). The Eustachian saddle-shaped pouch with its seat oriented tubes, which connect the middle ear to the toward the sacculus. The endolymphatic sys- buccal cavity, exit the prootics at a postero- tem presents grooves that may correspond to lateral ­fenestra behind the articulation with the periotic canals of the perilymphatic sys- the squamosal otic ramus. The inner ear tem (Mason et al., 2015), the most common is situated more medially within the audi- around its center circumference, and also tory capsule and consists of a membranous frequent, dividing the utriculus roughly into labyrinth suspended within the otic cap- left and right sections. sule by loose connective tissue ­(Duellman Nasal capsule. Blommersia transmarina has & Trueb, 1994). The endolymphatic system a calcified internasal septum (Trueb, 1968b; (i.e., otic labyrinth) is suspended within the Laloy et al., 2013) that projects rostrally from perilymphatic cistern, which connects with the anterior sphenethmoid, taking on a rhom- the perilymphatic sac within the neurocra- boid shape. Exostosed mineralization of this nium through two ­medial ­foramina (fig. 3 cartilaginous element evidences hyperossifi- and supplementary fig. S1); the round win- cation of the B. transmarina skull once more dow (Duellman & Trueb, 1994) being located (figs. 1–3). The central region (i.e., septum anterior to that of the C.N. VIII (Simmons nasi) presents the shape of a hollow rectangu- et al., 1992). The endolymphatic system in B. lar prism invaded by spongy bone and is situ- transmarina consists of a large saccule di- ated between the olfactory foramina of the vided by a central constriction that divides sphenethmoid, extending forward to the level it into two compartments: the dorsal utric- of the posteromedial margin of the nasals ulus (semicircular canals) and the ventral where it terminates in a slightly concave rect- sacculus (sensory epithelium). An endolym- angular tip. From both dorsal corners of the phatic duct leads from the sacculus to the tip of the septum nasi, a calcified tectum nasi endolymphatic torus surrounding the brain extends posterolaterally, terminating at the in the neurocranium (Dempster, 1930). The dorsal anterolateral corners of the spheneth- shape of the saccule ­compartments varied moid. The anterior margin of the tectum nasi

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 21

FIGURE 5 (A) Right lateral view of the Blommersia transmarina otic capsule. The cranium has been clipped laterally, posteriorly, and dorsally to facilitate visualization of the interior of the otic capsule. The anterior portion of the skull is clipped at the articulation of the palatine and alary cartilage. (B) Corresponding left sagittal plane, medial view of the B. transmarina otic capsule. The volume has been clipped to facilitate visualization of the interior of the otic capsule [MNCN50446 male; volrenWhite (A) and physics colormaps (B), grey [density] values: 12000–55500]. articulates with the nasals along three-fourths are located at the anterolateral margin of the of its length on both sides. The posterior re- nasals (Reynolds, 1897). gion of the cup-shaped alary cartilages is nar- Septomaxillae. The septomaxilla is a paired rowly separated, lateral, and ventral to the membrane bone that forms on top of the na- tectum nasi’s posterior borders; it is calcified sal capsule and is situated above the maxilla and fuses anteromedially with the ventral sep- and anterior to the vomer’s rostral tip. In two tum nasi at slightly over half its ventral-rostral individuals (i.e., MNCN50430 and 50437) it length (fig. 4). As a whole, the tectum nasi and was observed to contact the vomer. It has the alary cartilages serve as a roof and floor, re- shape of a ventrolaterally oriented loop with spectively, to create a pair of large concavities a thickened ellipsoidal cap facing the maxilla (i.e., cavum principale; Trueb, 1968b) to either (Trueb, 1968b) (fig. 4). In several individuals side of the septum nasi that ultimately lead to (e.g., MNCN50431, 50433, 50446) an unpaired the olfactory foramen of the sphenethmoid. dense spherical object of unknown nature was The palatine articulates with the ventrolateral observed dorsal to the septomaxilla (fig. 1). border of the alary cartilage, bracing the nasal capsule against the maxilla. The vomer’s cup- Dorsal dermal neurocranium shaped anterior segment delimits the poste- Nasals. The nasal is a lightly ossified paired rior nares, while the exterior anterior nares bone that lies dorsal of the nasal capsule and

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 22 Santos-Santos et al. anterior to the ossified region of the sphen- Maglia et al., 2007) and a cup-shaped trailing ethmoid. It presents an extensive cartilagi- edge (i.e., prechoanal and postchoanal por- nous articulation with the medial margin of tions), whose base aligns almost parallel to the the dorsal ridge of the tectum nasi. Postero- pars palatina of the maxilla. Along the lead- laterally, the nasal has a more heavily ossified ing edge before reaching the tip, the posterior maxillary process (Duellman & Trueb, 1994) segment branches off the anterior segment that terminates above the palatine and in line posteromedially via a peduncle. The posterior with the pars facialis of the maxilla (fig. 2). segment (i.e., dentigerous portion) is cylindri- Frontoparietals. The frontoparietal is a cal, covers a minimal part of the palate, and paired bone that originates dorsal to the bears a dentigerous process three-fourths its sphenethmoid and extends posteriorly to distal length below the nasal capsule. The slightly before the medial fusion of the pro- dentigerous process bears ventrolaterally ori- otics (fig. 1). Additionally, in B. transmarina ented vomerine teeth three-fourths its distal it extends a short distance laterally above the length, as well as presenting vomerine teeth anterior auditory capsule, covering about ¼ on its posterior tip, which terminates immedi- of the posterodorsal medial optic foramen ately posterior to the olfactory foramen of the (i.e., supraorbital flanges), which is yet an- sphenethmoid. The vomer has no visible os- other sign of hyperossification (Duellman & teological connection to either the premaxilla Trueb, 1994). The paired bones articulate at or the maxilla in B. transmarina. The septo- their medial margin, investing the entire dor- maxilla is contiguous with the base of the vo- sal roof of the neurocranium. At the postero- mer’s leading edge above the maxilla (fig. 3). lateral corner above the prootic they present Parasphenoid. The parasphenoid is an in- a dorsal odontoid protuberance presumably verted T-shape bone that invests the braincase resulting again from hyperossification in B. ventrally. Its anterior ramus (i.e., cultriform transmarina (fig. 4). process; Duellman & Trueb, 1994) extends from the posteromedial margin of the sphen- Ventral dermal neurocranium ethmoid to the prootic region, and its pos- In B. transmarina the ventral cartilaginous terolateral alae cover the auditory capsules wall of the neurocranium is calcified. There ventrally. In B. transmarina the parasphenoid are two large bilaterally symmetric foramina is less dense along its cultriform process; anterior to the posterior alae of the paras- however, it is hyperossified posteriorly fusing phenoid that correspond to the exit of C.NN. with the neurocranium and otoccipital pos- V (trigeminal) and VII (facial) (Ecker, 1889; teroventrally (fig. 2). Blommersia transmarina Reynolds, 1897). Slightly anterior and more does not have parasphenoid odontoids. medial to these are another pair of smaller fo- Palatines. The palatine is a slim paired ramina that correspond to the exit of C.N. VI transverse element that braces the upper jaw (abducens). The orbitonasal foramen (fig. 3) against the neurocranium. Its medial end ar- serves as an exit for C.NN. II (optic nerve), III ticulates laterally below the alary cartilages (oculomotor), and IV (trochlear) (Reynolds, of the nasal capsule (fig. 2), and its distal end 1897; Duellman & Trueb, 1994). widens and articulates with the medial lam- Vomers. The vomer is a paired bone located ina horizontalis of the maxilla immediately below the nasal capsule (fig. 2). Its anterior posterior to its pars facialis (fig. 4). In B. trans- segment resembles a shark’s dorsal fin with an marina it is smooth and does not bear a ser- elongated leading edge (i.e., anterior portion; rate bony ridge.

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 23

Pterygoids. The pterygoid is a paired trira- embracing the cartilaginous pseudobasal pro- diate medial brace between the upper jaw, cess of the quadrate. suspensorium, and neurocranium (Duellman & Trueb, 1994). It has the shape of an invert- Lateral dermal neurocranium ed “Y” laying horizontally (figs. 2 and 4). The Squamosal bones. The squamosal in lateral anterior ramus presents a broad articulation view looks like an anteriorly inclined “T” and with the maxilla and the ventral posterior participates in the suspension of the upper ramus articulates laterally with the quad- jaws from the skull (fig. 4). The dorsal ramus rate. The latter presents a caudoventrally ori- (i.e., otic ramus) is oriented dorsomedially, ented process at the ventroposterior edge of articulating posteriorly with the dorsolateral its medial surface (fig. 3). The anterior ramus margin of the prootic above the auditory cap- is thicker along its dorsal edge, presenting a sule at a cartilaginous joint. The anterior otic groove below its lateral margin that extends ramus extends toward the maxilla a short dis- posteriorly to the base of the dorsal posterior tance, terminating in an acuminate tip at the ramus, investing the cartilaginous pterygoid level of the posterior margin of the orbitona- process of the quadrate. The dorsal posterior sal foramen of the neurocranium. The ventral ramus extends medially toward the lateral ramus (i.e., zygomatic ramus) extends ventro- anteroventral auditory capsule; however, it is posteriorly at a slightly over 50° angle with not completely ossified; ossification extends the maxilla and articulates with the quadrato- along its leading edge about half the distance jugal laterally and with the quadrate medially and its lateral face remains open like a saddle, at its ventral end (fig. 6). The medial ­surface

FIGURE 6 Dorsal and internal view of the right Blommersia transmarina suspensorium / jaw articulation [MNCN50446 male; volrenWhite (A) and physics colormaps (B), grey [density] values: 12000–55500].

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 24 Santos-Santos et al. of the zygomatic ramus presents a groove the level of the vomer’s anterolateral tip. An- that continues into the posterior otic ramus, terior to the articulation with the palatine the which invests the cartilaginous otic process of maxilla’s external margin extends into a flat the quadrate. and wide triangular process (i.e., pars facia- lis) directed dorsally toward, and at the level Upper and lower jaws of, the posterior tip of the nasal. The ventral Premaxillae. The premaxillae are located border of the maxilla bears a dental ridge with anteromedially in the head and are united teeth from its anterior tip to slightly before its syndesmotically medially to each other and articulation with the quadratojugal (i.e., pars ­laterally to the maxillae. Anteriorly, the max- dentalis). illae slightly overlap the premaxillae laterally. Quadratojugals. The quadratojugals com- Each premaxilla is like a plate with a curved plete the upper jaw posteriorly. They articu- downward leading edge that bifurcates pos- late with the maxillae anteriorly and with the teriorly in two acuminate branches. From the ventrorostral margin of the zygomatic arch center of the leading edge a less dense verti- of the squamosals posteriorly in an upward cal strut extends posterodorsally. Medially the curve (i.e., pars articularis) (fig. 4). bone presents a less dense circular fenestra. Mentomeckelian bones. The Mentomeck- Traditionally the premaxilla is divided into elian bones are paired bones that form an- three parts (Duellman & Trueb, 1994) (fig. 3): teromedially replacing the Meckel’s cartilage. the pars dentalis has a curved downward They fuse syndesmotically with each other leading edge and bears a dental ridge with medially and have the shape of cylindrical teeth; the pars palatina serves as the lingual hooks that are oriented dorsally and face to- shelf and consists of two posterior acuminate ward each other medially (fig. 2). They form a branches; and the alary process is a dorsal ver- short less dense acuminate process ventrolat- tical strut that serves as a vertical abutment erally, and dorsolaterally they articulate with for the nasal capsule. the dentary (fig. 3). Maxillae. Each maxilla resembles a longi- Dentaries. The dentary is a paired bone that tudinal plank following a parabolic curve in covers the Meckel’s cartilage anterolaterally. the anterior–posterior direction. It overlaps It articulates medially with the dorsolateral the premaxilla slightly anteriorly and articu- margin of the Mentomeckelian bone and in- lates with the quadratojugal posteriorly. The vests the dorsal region of the angulosplenial lamina horizontalis (Roček, 1994) runs along (fig. 3). As in most anurans, in B. transmarina the midline of its medial surface, which termi- the dentary bears no teeth and covers almost nates caudally slightly anterior to the articu- half the anterior length of the angulosplenial; lation with the quadratojugal and posterior both bones are solidly connected and their to the articulation with the pterygoid (fig. 3). differentiation is inconspicuous. This lamina expands medially from the bone’s Angulosplenials. The angulosplenial invests surface in the caudal–rostral direction imme- the ventrolateral, medial, and posterior re- diately at the level of the pars facialis, creat- gions of the Meckel’s cartilage and is situated ing a lingual shelf between the maxilla and its lateral to the ventral margin of the mento- acuminate branch tip that terminates at the meckelian bone. It is an arcuate paired bone height of the septomaxilla (i.e., pars palatina). that extends anteroposteriorly, forming a tight One individual (i.e., MNCN50435) presented sigmoidal curve before its posterior articula- paired circlet processes on the lingual shelf at tion with the quadrate. In its caudal region,

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 25 following the sigmoidal curve, there is a large of the pterygoid at the mandibular articula- groove along the midline of its lateral sur- tion (fig. 6). The quadrate is cuboid in shape. face that opens up and flattens to terminate in an upward facing lobular plate, which em- Hyobranchial apparatus braces ventrally the quadratojugal–quadrate–­ In B. transmarina the hyoid plate (i.e., parahy- pterygoid articulation of the suspensorium oid) is very lightly calcified at its center with (fig. 3). Along the dorsomedial border of its no obvious pattern across individuals (fig. 7). posterior curved region there is a crest that The hyoid plate in this species bears two long terminates in a curved upward odontoid pro- bony posteromedial processes whose ante- cess anterior to the quadratojugal–quadrate– rior ends lie at the level of and dorsal to the pterygoid articulation. Prior to this odontoid anterior process of the omosternum, each process, a bony process extends posterolater- laying roughly on either side just rostral of ally from the crest’s base supporting a large the fork in the omosternum. The posterome- rounded head that curves downward toward dial processes present an expanded anterior a tiny acuminate process that protrudes from epiphysis and terminate as a truncated cylin- the dorsolateral border of the angulosplenial der. Most individuals (excluding MNCN50430, parallel to the odontoid process (fig. 6). The 50431, 50432) have a lateral crest on the poste- head of the bony process that extends from rior margin that terminates slightly rostral of the crest is in line with an edge of the ossified its posterior limit, and which displays varying cuboid quadrate (fig. 6). degrees of development. They are flattened dorsoventrally and exhibit an upward inclina- Suspensorium tion in the anteroposterior direction, termi- Quadrates. The quadrate is the central ele- nating at the level slightly before the scapula–­ ment that suspends and braces the jaws to suprascapula articulation. the skull and is mostly cartilaginous. Its pars articularis (Duellman & Trueb, 1994) presents Pectoral girdle perichondral ossification along its medial Blommersia transmarina presents a firmis- margin and buttresses the quadratojugal with ternal pectoral girdle (Emerson, 1983) in that the squamosal and the posteroventral ramus the epicoracoid cartilages lack horns and are

FIGURE 7 Ventral view of the Blommersia transmarina pectoral girdle. The dorsal coronal plane is clipped at the mandible to facilitate visualization [MNCN50446 male; volrenWhite (A) and physics colormaps (B),

grey [density] values: 12000–55500]. Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 26 Santos-Santos et al. fused medially; the sternum is fused to the along its dorsoventral axis. It articulates with pectoral arch and possesses relatively lengthi- the procoracoid cartilage proximally, the cor- er prezonal (i.e., omosternum) and postzonal acoid ventrally, and with the scapula distal-­ (i.e., mesosternum) elements than in arcifery. dorsally (fig. 7). Ossification of the clavicle Anteroposteriorly the prezonal elements in- penetrates the procoracoid cartilage to various clude a cartilaginous episternum and an os- degrees depending on the individual. Its distal sified omosternum; the postzonal elements end widens cranially and its degree of ossifi- consist in an ossified (meso)sternum and cation also varies across individuals (­table 2). a cartilaginous xiphisternum (Duellman & The clavicle presents a yet undescribed pro- Trueb, 1994). cess positioned cranially at its proximal end, Episternum. The episternum is an anterior distal to the clavicle–coracoid articulation, cartilaginous extension of the omosternum which appears to articulate with the posterior that does not ossify. It envelops the end of branches of the omosternum and serve as an the anterior branch of the omosternum and attachment site for several muscles (e.g., mm. presents a rhomboid shape like a cranially-­ deltoideus and coracoradialis) (fig. 7). directed angled spade. Coracoids. The coracoid is a paired chon- Omosternum. The omosternum is an elon- dral bone with a flattened proximal end and gated bone that bifurcates posterolaterally a rounded distal end. Its distal end articulates and articulates with both clavicles at a process with the scapula at the glenoid cavity, where- along the first third of their proximal end, but as its proximal end articulates medially with never at their medial tip (fig. 7). The length, the epicoracoid cartilage (fig. 7). The degree of width, and degree of ossification of the poste- ossification of its proximal end is variable and rior branches of the omosternum vary across its shaft is smooth. individuals (table 2). Scapulae. The scapula is a dorsoventrally Mesosternum. The (meso)sternum is rela- elongated paired chondral bone that resem- tively lengthened compared to arciferal an- bles an hourglass in shape. At the suprascapu- urans and its shape depends on its level of lar articulation (fig. 8) the corpus scapulae ­ossification. The mesosternum is an elongated (Špinar, 1972) is flat and triangular in shape. It bone that articulates cranially with the medial tapers in the ventral direction and converges ends of both coracoids at the epicoracoid car- into the collum scapulae, which is elevated tilages (fig. 7). It presents either of two con- medially, at approximately half the scapula’s figurations: a wide anterior end that tapers length (table 2). At the glenoid cavity, the toward the center (e.g., MNCN50430, 50432, caput scapulae is divided into the pars acro- 50435, 50436, 50446), or a bifurcated anterior mialis anteriorly, which articulates with the end, with wider branches than those of the clavicle, and the pars glenoidalis posteriorly, omosternum, in which each branch articu- which articulates with the coracoid (Roček, lates anteriorly with one of the coracoids (e.g., 1994). These are of similar width, length, and MNCN50431, 50433, 50437, 50439). One indi- thickness in B. transmarina. In the medial vidual (i.e., DRV6851) has a sternum with simi- view, there is a deep cleft (i.e., sinus intergle- lar widths at the anterior and posterior ends noidalis; Roček, 2005) between the pars acro- that resembles an hourglass. The xiphisternum mialis and pars glenoidalis. does not ossify and consists of a heart-shaped Suprascapulae and Cleithra. The supra- cartilage posterior to the mesosternum. scapula is a paired dermal bone with varying Clavicles. The clavicle is a paired, thin, and degrees of ossification that covers roughly ⅔ mostly straight dermal bone that is flattened of its surface. The Downloadedbony leading from Brill.com10/04/2021 edge of the 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 27

−.86 −.98 −.01 −1.39 −1.96 −1.00 Kurtosis

.

�.06 .83 .44 .75 .65 .94 Asymmetry Blommersia transmarina .7760 .7450 .3371 .9223 .9380 .5103 .4933 .5908 .2599 1.7102 1.1287 sd 2.0697 2.4016 2.3977 2.5957 2.6265 2.8591 2.0457 9.91 9.50 6.65 6.88 6.42 10.43 10.96 10.06 10.62 Mean 27.34 27.96 26.73 26.19 27.59 24.79 29.55 30.68 28.43 9.68 6.54 10.18 27.25 25.40 28.38 Median

0.10 0.38 0.20 0.51 0.40 0.24 partial eta-squared 0.378 0.057 0.190 0.152 0.021 0.048 Sexual dimorphism 7.52 11.76 11.59 max 33.39 30.38 30.07 8.90 9.66 6.07 min 23.81 27.00 24.50 1.46 range 5.88 2.10 2.69 6.26 6.39 5 5 5 5 5 5 5 5 5 5 5 5 N 10 10 10 10 10 10 F M F M F M F M F M F M Descriptive statistics and sexual dimorphism of the 123 skeletal measures taken on 10 specimens (five male / five female) of male / five on 10 specimens (five dimorphism oftaken measures statistics and sexual Descriptive skeletal the 123 and hind limb. column, pelvic girdle, vertebral forelimb, girdle, SVL, skull, pectoral lecture: facilitate to regions anatomical partitioned into are Measures those in the description of follow in bold. Abbreviations shown males and females (α = 0.05) are the in mean length between Significant differences (Appendix 1). measurements

(mm) SVLext SVLsk SVLsum skullL skullW oticsW Table 2

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 28 Santos-Santos et al. .15 .81 .14 .11

-.61 −.25 −.45 Kurtosis

.

.87 .71 .15 .61 -.50 1.10 −.08 Asymmetry Blommersia transmarina .1395 .1304 .1423 .1683 .2006 .0666 .1314 .1279 .0633 .0882 .0717 .0731 .1093 .0901 .1370 .8783 .9177 .3621 .6356 .5811 .3419 sd .97 .96 .79 .79 .79 1.44 1.49 1.39 1.52 1.61 1.43 1.05 1.14 1.01 1.07 9.56 9.54 9.97 9.12 10.11 10.66 Mean .77 1.41 1.50 1.02 1.00 9.87 9.39 Median

0.15 0.30 0.48 0.40 0.00 0.44 0.50 partial eta-squared 0.271 0.102 0.973 0.027 0.049 0.037 0.022 Sexual dimorphism .98 1.66 1.82 1.28 1.15 11.69 max 10.57 .84 .60 .88 1.18 1.33 8.71 9.18 ). min cont. .48 .49 .40 .31 .38 1.86 range 2.51 5 5 5 5 5 5 5 5 5 5 5 5 5 5 10 10 10 10 N 10 10 10 F M F M F M F M F M F M F M Descriptive statistics and sexual dimorphism of the 123 skeletal measures taken on 10 specimens (five male / five female) of male / five on 10 specimens (five taken dimorphism of measures statistics and sexual Descriptive skeletal the 123 and hind limb. column, pelvic girdle, vertebral forelimb, girdle, SVL, skull, pectoral lecture: facilitate to regions anatomical partitioned into are Measures those in the description of follow in bold. Abbreviations shown males and females (α = 0.05) are the in mean length between Significant differences (Appendix 1) ( measurements

(mm) colmL pmaxH pmaxW pmaxD MmbL maxL mandL Table 2

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 29 .14 .36 -.41 -.13 -.28 -.53 Kurtosis -1.76 .53 .06 .43 .48 .02 .97 -.24 Asymmetry .2648 .2149 .1553 .1509 .1423 .1757 .1094 .0900 .1322 .1705 .1618 .2085 .1900 .2610 .1037 .2080 .1275 .1431 .3383 .3623 .1877 sd .97 .93 1.68 1.68 1.67 3.19 3.37 3.00 1.00 1.39 1.42 1.36 1.36 1.39 1.33 2.82 2.98 2.67 5.48 5.67 5.29 Mean .95 1.67 3.17 1.34 1.34 2.86 5.44 Median

0.55 0.00 0.15 0.04 0.03 0.62 0.35 partial eta-squared 0.901 0.014 0.455 0.696 0.639 0.007 0.070 Sexual dimorphism 1.91 3.70 1.11 1.60 1.69 3.18 6.08 max 1.42 .80 2.81 1.19 1.08 2.50 5.06 min .31 .41 .61 .68 .90 .49 1.02 range 6 3 3 6 3 3 5 5 5 5 5 5 5 5 5 5 10 10 10 10 10 N F M F M F M F M F M F M F M (mm) nasL nasW intnasW intnsptH sphH sphW frptlL

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 30 Santos-Santos et al.

.62 .39 -.97 -.29 -.67 2.75 Kurtosis -1.12

.

.46 .51 .25 -.50 -.23 -.39 1.61 Asymmetry Blommersia transmarina .0904 .1012 .0727 .2216 .3035 .1018 .1795 .2267 .0515 .2438 .2056 .1372 .2598 .2166 .3130 .1039 .0739 .1061 .2276 .1739 .2341 sd 1.68 1.71 1.64 2.46 2.52 2.41 2.91 3.00 2.82 3.52 3.69 3.35 4.63 4.68 4.57 1.17 1.22 1.11 1.48 1.59 1.37 Mean 1.68 2.42 2.86 3.48 4.66 1.14 1.52 Median

0.16 0.07 0.25 0.54 0.05 0.31 0.27 partial eta-squared 0.258 0.452 0.137 0.015 0.551 0.094 0.123 Sexual dimorphism 1.86 2.87 3.33 3.88 4.99 1.29 1.77 max ). 1.54 2.11 2.73 3.16 4.15 1.00 1.07 min cont. .32 .76 .60 .73 .84 .30 .70 range 5 5 5 5 5 5 5 5 5 5 5 5 5 5 10 10 10 10 10 10 10 N F M F M F M F M F M F M F M Descriptive statistics and sexual dimorphism of the 123 skeletal measures taken on 10 specimens (five male / five female) of male / five on 10 specimens (five taken dimorphism of measures statistics and sexual Descriptive skeletal the 123 and hind limb. column, pelvic girdle, vertebral forelimb, girdle, SVL, skull, pectoral lecture: facilitate to regions anatomical partitioned into are Measures those in the description of follow in bold. Abbreviations shown males and females (α = 0.05) are the in mean length between Significant differences (Appendix 1) ( measurements

(mm) frptlW proH exoL psphH psphW vomH vomW Table 2 Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 31 .50 .71 -.57 1.19 2.33 1.48 Kurtosis -1.11 -1.04 .85 .25 -.28 1.29 1.29 1.02 1.05 1.09 Asymmetry .3481 .1637 .1757 .1998 .1594 .1306 .1448 .0320 .1280 .1256 .1233 .2991 .2802 .1688 .2573 .2592 .1045 .6036 .5141 .4383 .3014 .3383 .1458 .3322 sd 3.02 3.21 2.83 2.01 2.18 1.84 5.59 5.97 5.21 2.80 2.97 2.64 2.89 3.09 2.69 2.98 3.02 2.94 1.52 1.60 1.43 1.43 1.48 1.39 Mean 2.99 1.93 5.59 2.77 2.83 3.00 1.45 1.40 Median

0.47 0.48 0.44 0.34 0.40 0.06 0.43 0.16 partial eta-squared 0.029 0.027 0.036 0.079 0.048 0.498 0.041 0.252 Sexual dimorphism 3.65 2.60 6.48 3.45 3.51 3.22 1.76 1.70 max 2.61 1.68 4.84 2.44 2.45 2.70 1.38 1.28 min .38 .42 .92 .52 1.05 1.64 1.02 1.05 range 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 N 10 10 10 10 10 10 10 10 F M F M F M F M F M F M F M F M (mm) pltnL ptygdH ptygdW squaH squaW hyoL omostH omostL

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 32 Santos-Santos et al.

.67 .13 -.13 1.32 1.17 1.17 Kurtosis -1.14

.

.89 .24 .56 .84 .74 1.19 1.19 Asymmetry Blommersia transmarina .3683 .4436 .1474 .1494 .1801 .1041 .2605 .3600 .1045 .2651 .2554 .2256 .0894 .0909 .0606 .1086 .1192 .0761 .1465 .1286 .0706 sd .79 .84 .74 .96 .50 .56 .45 .61 .71 .51 3.38 3.56 3.19 3.05 3.12 2.98 1.63 1.76 1.50 1.01 1.06 Mean .73 .99 .50 .58 3.32 2.98 1.63 Median

0.28 0.14 0.08 0.27 0.34 0.25 0.55 partial eta-squared 0.113 0.291 0.430 0.127 0.079 0.144 0.014 Sexual dimorphism .73 .89 4.17 1.12 3.60 2.02 1.17 max .63 .88 .37 .42 ). 2.96 2.68 1.28 min cont. .49 .92 .74 .30 .36 .47 1.21 range 5 5 5 5 5 5 5 5 5 5 5 5 5 5 N 10 10 10 10 10 10 10 F M F M F M F M F M F M F M Descriptive statistics and sexual dimorphism of the 123 skeletal measures taken on 10 specimens (five male / five female) of male / five on 10 specimens (five taken dimorphism of measures statistics and sexual Descriptive skeletal the 123 and hind limb. column, pelvic girdle, vertebral forelimb, girdle, SVL, skull, pectoral lecture: facilitate to regions anatomical partitioned into are Measures those in the description of follow in bold. Abbreviations shown males and females (α = 0.05) are the in mean length between Significant differences (Appendix 1) ( measurements

(mm) clavL clav_pL coraL sterH sterWsup sterWmin sterWinf Table 2

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 33 -.19 -.79 -.74 -.26 1.37 2.54 Kurtosis -1.38 .29 .75 .76 .81 .71 1.32 1.53 Asymmetry .2965 .3817 .1702 .2499 .3402 .1266 .1530 .2041 .0568 .2886 .3604 .1889 .5982 .7231 .1978 .6147 .7281 .2103 .0717 .0697 .0595 sd .55 .58 .51 3.48 3.57 3.38 2.86 2.92 2.80 1.54 1.60 1.49 2.26 2.36 2.17 7.59 7.90 7.27 7.62 7.95 7.29 Mean .53 3.41 2.78 1.52 2.22 7.45 7.49 Median

0.12 0.06 0.15 0.12 0.30 0.32 0.27 partial eta-squared 0.334 0.488 0.273 0.335 0.100 0.086 0.122 Sexual dimorphism .68 3.93 3.40 1.90 2.82 8.64 8.72 max .45 3.08 2.61 1.40 1.94 6.88 6.94 min .85 .80 .50 .88 .23 1.76 1.78 range 5 5 5 5 5 5 5 5 5 5 5 5 5 5 N 10 10 10 10 10 10 10 F M F M F M F M F M F M F M (mm) scapL cleiL supraH supraW humLcap humLtroc dltd_crH

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 34 Santos-Santos et al.

.69 .93 -.49 1.19 3.20 1.84 Kurtosis -1.36

.

.67 .77 .57 1.24 1.02 1.66 1.15 Asymmetry Blommersia transmarina .3373 .4341 .2541 .4171 .5148 .1089 .2341 .3244 .0930 .3162 .4334 .1515 .2586 .3395 .1029 .2502 .3533 .0518 .1628 .2140 .0791 sd 2.18 2.22 2.15 5.51 5.72 5.29 2.32 2.38 2.26 2.70 2.77 2.62 3.11 3.21 3.01 2.57 2.65 2.50 1.43 1.49 1.38 Mean 2.01 5.37 2.30 2.68 3.02 2.50 1.38 Median

0.01 0.29 0.08 0.06 0.16 0.09 0.13 partial eta-squared 0.766 0.106 0.439 0.485 0.246 0.387 0.311 Sexual dimorphism 2.73 6.35 2.81 3.38 3.73 3.13 1.73 max ). 1.86 5.05 1.99 2.31 2.85 2.24 1.21 min cont. .87 .82 .88 .89 .52 1.30 1.07 range 5 5 5 5 5 5 5 5 5 5 5 5 5 5 N 10 10 10 10 10 10 10 F M F M F M F M F M F M F M Descriptive statistics and sexual dimorphism of the 123 skeletal measures taken on 10 specimens (five male / five female) of male / five on 10 specimens (five taken dimorphism of measures statistics and sexual Descriptive skeletal the 123 and hind limb. column, pelvic girdle, vertebral forelimb, girdle, SVL, skull, pectoral lecture: facilitate to regions anatomical partitioned into are Measures those in the description of follow in bold. Abbreviations shown males and females (α = 0.05) are the in mean length between Significant differences (Appendix 1) ( measurements

(mm) dltd_crW radulnL mtcpl1 mtcpl2 mtcpl3 mtcpl4 hand_ph1L1 Table 2

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 35 .39 .74 .39 -.21 -.36 -.63 1.38 Kurtosis .46 .68 .40 .15 .90 .96 1.36 Asymmetry .0823 .1118 .0438 .4643 .6419 .1661 .1671 .2246 .0907 .0830 .1001 .0695 .5391 .7354 .2641 .1902 .2422 .0827 .1019 .1035 .0483 sd .62 .64 .60 .68 .69 .66 4.38 4.51 4.24 1.79 1.83 1.75 5.17 5.30 5.04 1.90 1.98 1.82 1.68 1.75 1.62 Mean .62 .67 4.35 1.80 5.13 1.82 1.64 Median

0.05 0.09 0.07 0.04 0.07 0.20 0.50 partial eta-squared 0.517 0.390 0.475 0.573 0.473 0.201 0.033 Sexual dimorphism .75 .82 5.29 2.10 6.30 2.32 1.88 max .50 .56 3.70 1.55 4.53 1.73 1.56 min .26 .55 .27 .59 .32 1.59 1.77 range 5 5 5 5 5 5 5 5 5 5 5 5 9 4 5 N 10 10 10 10 10 10 F M F M F M F M F M F M F M (mm) hand_tph1L hand_D1 hand_ph2L1 hand_tph2L hand_D2 hand_ph3L1 hand_ph3L2

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 36 Santos-Santos et al.

.80 .26 .44 -.75 2.42 2.17 4.23 Kurtosis

.

.15 .78 -.34 -.08 1.58 1.69 1.02 Asymmetry Blommersia transmarina .0563 .0773 .0311 .5729 .7039 .1606 .1382 .1852 .0749 .1477 .1850 .0908 .0673 .0933 .0385 .5734 .7880 .2324 .9767 .5386 1.2317 sd .77 .78 .76 .76 .77 .76 7.51 7.89 7.20 1.34 1.37 1.30 1.37 1.43 1.32 6.05 6.21 5.89 8.70 9.07 8.33 Mean .78 .75 7.33 1.31 1.38 5.95 8.35 Median

0.03 0.40 0.07 0.14 0.00 0.09 0.16 partial eta-squared 0.657 0.070 0.457 0.296 0.945 0.406 0.255 Sexual dimorphism .86 .89 8.78 1.68 1.64 7.32 max 10.47 .66 .65 ). 7.03 1.18 1.13 5.21 7.74 min cont. .20 .50 .50 .24 1.75 2.11 2.73 range 5 5 9 4 5 5 5 5 5 5 5 5 5 5 5 N 10 10 10 10 10 10 F M F M F M F M F M F M F M Descriptive statistics and sexual dimorphism of the 123 skeletal measures taken on 10 specimens (five male / five female) of male / five on 10 specimens (five taken dimorphism of measures statistics and sexual Descriptive skeletal the 123 and hind limb. column, pelvic girdle, vertebral forelimb, girdle, SVL, skull, pectoral lecture: facilitate to regions anatomical partitioned into are Measures those in the description of follow in bold. Abbreviations shown males and females (α = 0.05) are the in mean length between Significant differences (Appendix 1) ( measurements

(mm) hand_tph3L hand_D3 hand_ph4L1 hand_ph4L2 hand_tph4L hand_D4 column Table 2

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 37 .04 -.66 -.51 -.03 -.41 -.82 -1.14 Kurtosis .59 .61 .88 .63 .48 -.21 1.10 Asymmetry .1266 .1441 .0963 .0769 .0942 .0438 .0596 .0828 .0236 .1106 .1422 .0647 .1549 .1947 .1136 .1487 .1960 .0989 .1657 .2022 .0973 sd .83 .86 .80 .95 .96 .93 .99 .95 .96 1.06 1.11 1.01 1.02 1.00 1.04 1.07 1.10 1.05 1.09 1.03 Mean 1.16 .81 .95 .95 .97 1.09 1.06 1.04 Median

0.17 0.19 0.07 0.11 0.06 0.03 0.18 partial eta-squared 0.240 0.209 0.453 0.343 0.501 0.623 0.216 Sexual dimorphism .96 1.26 1.05 1.21 1.29 1.33 1.40 max .88 .74 .87 .88 .79 .90 .92 min .38 .22 .17 .33 .50 .43 .48 range 5 5 5 5 5 5 5 5 5 5 5 5 5 5 N 10 10 10 10 10 10 10 F M F M F M F M F M F M F M (mm) atlasH V2H V3H V4H V5H V6H V7H

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 38 Santos-Santos et al.

-.96 -.85 -.93 -1.47 Kurtosis -1.23 -1.92 -2.06

.

.21 .16 .47 .50 .71 .28 .25 Asymmetry Blommersia transmarina .1706 .1787 .1325 .0821 .0993 .0564 .4348 .4978 .3896 .4793 .4940 .3713 .4978 .6095 .2879 .4534 .5376 .3449 .3901 .4300 .3491 sd .92 .70 .73 .67 1.00 1.08 4.58 4.68 4.48 5.91 6.14 5.68 5.20 5.41 5.00 5.11 5.26 4.96 5.15 5.26 5.03 Mean .98 .69 4.48 5.81 4.97 5.00 5.05 Median

0.24 0.14 0.06 0.26 0.19 0.12 0.10 partial eta-squared 0.147 0.287 0.493 0.131 0.215 0.332 0.363 Sexual dimorphism .83 1.24 5.19 6.78 6.11 5.79 5.67 max .76 .58 ). min 3.98 5.36 4.65 4.63 4.74 cont. .48 .25 .93 1.21 1.41 1.45 1.16 range 5 5 5 5 5 5 5 5 5 5 5 5 5 5 N 10 10 10 10 10 10 10 F M F M F M F M F M F M F M Descriptive statistics and sexual dimorphism of the 123 skeletal measures taken on 10 specimens (five male / five female) of male / five on 10 specimens (five taken dimorphism of measures statistics and sexual Descriptive skeletal the 123 and hind limb. column, pelvic girdle, vertebral forelimb, girdle, SVL, skull, pectoral lecture: facilitate to regions anatomical partitioned into are Measures those in the description of follow in bold. Abbreviations shown males and females (α = 0.05) are the in mean length between Significant differences (Appendix 1) ( measurements

(mm) V8H sacrumH V2_tpL V3_tpL V4_tpL V5_tpL V6_tpL Table 2

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 39 -.38 -.74 1.20 Kurtosis -2.13 -1.99 -1.41 -1.61 .11 .25 .81 .64 .63 -.07 -.30 Asymmetry .4058 .4475 .4028 .4408 .4585 .4454 .6165 .5889 .5631 .7067 .8395 .4326 .1540 .1748 .0862 .5031 .4577 .4051 .1249 .1473 .0780 sd .94 5.06 5.11 5.00 4.78 4.43 4.71 4.16 8.14 8.44 7.83 1.47 1.55 1.39 4.89 4.61 5.17 1.00 1.05 Mean 4.89 4.67 .96 5.11 4.63 4.28 7.88 1.39 4.86 Median

0.02 0.07 0.22 0.21 0.29 0.34 0.21 partial eta-squared 0.701 0.476 0.167 0.187 0.110 0.075 0.185 Sexual dimorphism 5.55 5.39 5.21 9.51 1.71 5.76 1.22 max .85 min 4.62 4.27 3.47 7.41 1.33 3.89 .93 .38 .38 1.12 1.74 2.10 1.87 range 9 5 4 5 5 5 5 5 5 5 5 5 5 5 5 N 10 10 10 10 10 10 F M F M F M F M F M F M F M (mm) V7_tpL V8_tpL sacrum_tpL uroL uroH uro_crL uro_crH

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 40 Santos-Santos et al.

.25 .05 -.96 -.81 -.64 -.94 Kurtosis -1.13

.

.63 .67 .72 .54 -.15 1.22 1.18 Asymmetry Blommersia transmarina .7354 .8214 .4107 .8862 .4731 .4066 .5002 .1169 .2297 .2972 .0941 .1097 .0776 .0993 .6217 .9421 1.0088 1.2775 1.3972 1.5427 1.6892 sd .76 .84 .71 8.21 8.59 7.82 9.76 3.69 3.96 3.47 1.95 2.08 1.85 Mean 10.21 10.67 13.96 14.69 13.23 16.22 17.03 15.42 .76 7.95 9.96 3.52 1.89 Median 13.88 16.11

0.31 0.30 0.39 0.29 0.40 0.36 0.30 partial eta-squared 0.097 0.103 0.072 0.136 0.067 0.065 0.100 Sexual dimorphism .92 9.45 4.44 2.34 max 11.74 16.09 18.85 .60 ). min 7.38 9.17 3.28 1.74 12.59 14.58 cont. .60 .33 2.07 2.57 1.16 3.50 4.27 range 5 5 5 5 9 4 5 9 4 5 9 4 5 5 5 5 5 N 10 10 10 10 F M F M F M F M F M F M F M Descriptive statistics and sexual dimorphism of the 123 skeletal measures taken on 10 specimens (five male / five female) of male / five on 10 specimens (five taken dimorphism of measures statistics and sexual Descriptive skeletal the 123 and hind limb. column, pelvic girdle, vertebral forelimb, girdle, SVL, skull, pectoral lecture: facilitate to regions anatomical partitioned into are Measures those in the description of follow in bold. Abbreviations shown males and females (α = 0.05) are the in mean length between Significant differences (Appendix 1) ( measurements

(mm) iliL pelvL pelwL epipH epipW femL tibfibL Table 2

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 41 .07 -.52 -.13 1.43 2.16 1.39 1.06 Kurtosis .75 .32 1.03 1.36 1.17 1.07 1.17 Asymmetry .6216 .7146 .2947 .2259 .3037 .1405 .3789 .4891 .1509 .4189 .5367 .1997 .4721 .5568 .2370 .4243 .5045 .2330 .1483 .2025 .0761 sd 8.29 8.62 7.96 2.14 2.17 2.10 4.07 4.22 3.91 5.31 5.47 5.15 6.21 6.44 5.97 5.47 5.67 5.27 1.20 1.23 1.17 Mean 8.16 2.13 4.01 5.23 6.07 5.34 1.18 Median

0.31 0.03 0.19 0.17 0.27 0.24 0.05 partial eta-squared 0.093 0.662 0.209 0.237 0.124 0.153 0.519 Sexual dimorphism 9.42 2.63 4.92 6.21 7.10 6.38 1.47 max .97 min 7.58 1.90 3.64 4.81 5.75 5.04 .73 .50 1.84 1.28 1.40 1.36 1.34 range 5 5 5 5 5 5 5 5 5 5 5 5 5 5 N 10 10 10 10 10 10 10 F M F M F M F M F M F M F M (mm) tiblfiblL mttrsl1 mttrsl2 mttrsl3 mttrsl4 mttrsl5 foot_ph1L1

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 42 Santos-Santos et al.

.86 .41 1.24 2.59 4.08 3.39 1.48 Kurtosis

.

.65 .81 1.59 1.58 1.68 1.03 1.32 Asymmetry Blommersia transmarina .0662 .0876 .0353 .4315 .5827 .2469 .2101 .2795 .0603 .0642 .0836 .0386 .6259 .8114 .2383 .2669 .3258 .1121 .1619 .0756 sd .1994 .49 .51 .47 .57 .59 .55 3.82 3.91 3.74 1.83 1.91 1.74 6.46 6.72 6.20 2.47 2.60 2.35 1.65 1.57 Mean 1.72 .49 .56 3.80 1.75 6.33 2.42 1.59 Median

0.10 0.04 0.18 0.09 0.19 0.26 0.23 partial eta-squared 0.383 0.560 0.226 0.410 0.210 0.133 0.162 Sexual dimorphism .63 .72 4.72 2.32 7.96 3.02 2.01 max .39 .49 ). min 3.26 1.65 5.82 2.20 1.50 cont. .24 .67 .23 .81 .50 1.46 2.14 range 5 5 5 5 5 5 5 5 5 5 5 5 5 5 N 10 10 10 10 10 10 10 F M F M F M F M F M F M F M Descriptive statistics and sexual dimorphism of the 123 skeletal measures taken on 10 specimens (five male / five female) of male / five on 10 specimens (five taken dimorphism of measures statistics and sexual Descriptive skeletal the 123 and hind limb. column, pelvic girdle, vertebral forelimb, girdle, SVL, skull, pectoral lecture: facilitate to regions anatomical partitioned into are Measures those in the description of follow in bold. Abbreviations shown males and females (α = 0.05) are the in mean length between Significant differences (Appendix 1) ( measurements

(mm) foot_tph1L foot_D1 foot_ph2L1 foot_tph2L foot_D2 foot_ph3L1 foot_ph3L2 Table 2

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 43 .39 .63 -.23 -.60 -.35 1.54 Kurtosis -1.12 .96 .89 .98 .83 .50 .94 1.30 Asymmetry .0605 .0814 .0245 .8688 .3688 .3189 .3721 .1902 .2398 .3107 .0769 .1878 .2229 .0728 .0629 .0517 .0364 .5484 1.0907 1.2787 1.4737 sd .60 .62 .58 .67 .72 .63 9.65 3.59 3.73 3.44 2.39 2.49 2.28 1.77 1.87 1.67 Mean 10.03 10.42 14.69 15.55 14.00 .59 .65 9.72 3.50 2.32 1.71 Median 14.05

0.12 0.22 0.24 0.21 0.31 0.58 0.41 partial eta-squared 0.319 0.172 0.154 0.185 0.097 0.017 0.063 Sexual dimorphism .71 .77 4.24 2.85 2.08 max 11.94 17.05 .52 .60 min 9.15 3.16 2.16 1.57 13.50 .19 .70 .51 .17 2.80 1.08 3.54 range 5 5 5 5 5 5 5 5 5 5 9 4 5 9 4 5 N 10 10 10 10 10 F M F M F M F M F M F M F M (mm) foot_tph3L foot_D3 foot_ph4L1 foot_ph4L2 foot_ph4L3 foot_tph4L foot_D4

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 44 Santos-Santos et al.

-.76 -.20 3.39 1.67 Kurtosis

.

.77 .23 1.69 1.48 Asymmetry Blommersia transmarina .2476 .3320 .1103 .2134 .2376 .1223 .0865 .1214 .0464 .9583 .4846 1.2843 sd .61 .64 .58 2.71 2.79 2.63 1.73 1.84 1.62 Mean 10.47 10.92 10.11 .61 2.65 1.63 Median 10.07

0.11 0.30 0.12 0.20 partial eta-squared 0.343 0.099 0.364 0.229 Sexual dimorphism .75 3.31 2.12 max 12.53 .48 ). min 2.47 1.49 9.58 cont. .84 .64 .27 2.95 range 5 5 5 5 9 4 5 9 4 5 N 10 10 F M F M F M F M Descriptive statistics and sexual dimorphism of the 123 skeletal measures taken on 10 specimens (five male / five female) of male / five on 10 specimens (five taken dimorphism of measures statistics and sexual Descriptive skeletal the 123 and hind limb. column, pelvic girdle, vertebral forelimb, girdle, SVL, skull, pectoral lecture: facilitate to regions anatomical partitioned into are Measures those in the description of follow in bold. Abbreviations shown males and females (α = 0.05) are the in mean length between Significant differences (Appendix 1) ( measurements

(mm) foot_ph5L1 foot_ph5L2 foot_tph5L foot_D5 Table 2

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 45

FIGURE 8 (A) Ventral view of the Blommersia transmarina vertebral column. (B) Dorsal view of the B. transma- rina vertebral column. The volumes are clipped anteriorly at the foramen magnum and posteriorly immediately before the sacrum [MNCN50446 male; volrenWhite (A) and physics colormaps (B), grey [density] values: 12000–55500]. suprascapula is denominated the cleithrum. eminence of the prootic, even making con- Ossification invades the suprascapular carti- tact in some individuals (e.g., MNCN50430, lage from the cleithrum and extends toward 50431, 50446; fig. 1, DRV6851). At its distal the anterior medial edge, and posteriorly end the suprascapula also presents vary- and medially, resembling a crab claw (fig. 8). ing levels of ossification and articulates with The anteromedial edge of the cleithrum is the scapula, but a cleft is clear between both ­narrowly separated from the posterior epiotic elements. Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 46 Santos-Santos et al.

Forelimb although they can be differentiated at their Humerus. The humerus head is round and distal end due to the presence of a groove (i.e., slightly elongated. There is a large prominent sulcus longitudinalis; Bolkay, 1919). Proximal- cranial process at its proximal end denomi- ly, the ulna articulates with the humerus at the nated the deltopectoral crest (i.e., crista ven- concave olecranon and with the radius at the tralis; Duellman & Trueb, 1994). The anterior capitulum. Distally, the ulna articulates with ridge of the deltopectoral crest articulates the ulnare and the intermedium (here ulnare’; medially with the humerus head, leaving an see Discussion), while the radius articulates opening (i.e., fenestra) directly below the ar- with the radiale and the centrale (fig. 9). ticulation that also forms the beginning of Hand. Following Fabrezi & Alberch (1996), an intertubercular groove that extends dis- the hand consists of five mesopodial elements tally below the deltopectoral crest to the lat- (i.e., ulnare, radiale, Element Y, Distal Carpal eral border of the diaphysis (fig. 7). The crista 5-4-3, and Distal Carpal 2), four metapodial el- lateralis (Špinar, 1972; Roček, 2005) is greatly ements (i.e., metacarpals), the prepollex, the reduced and limited to the proximal portion phalanges, and intercalary cartilage elements. of the humerus immediately following the The ulnare articulates postaxially with the greater tubercle of the humerus head; other- lateral border of the ulna’s distal epiphysis. It wise the diaphysis is smooth. The diaphysis has the shape of a laterally compressed square presents a slight sigmoidal curve that follows with two projections along its posterior mar- the line of the deltopectoral crest and ends at gin, one dorsal, one ventral, and with another the distal lateral epicondyle. projection extending medially from its dorso- Radioulna. The radius and ulna of tetra- anterior border. Ventral to the ulnare lies the pods are fused into a single bone in anurans, compound bone Distal Carpal 5-4-3. This bone

FIGURE 9 Dorsoanterior view of the Blommersia transmarina carpus. [MNCN50446 male; volrenWhite (A) and physics* colormaps (B), grey [density] values: 12000–55500]. *This individual is relatively less developed than MNCN50446, and as such is less densely ossified. Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 47 articulates distally with Metacarpals IV–II and ­elongated than the rest and Digit I shorter. displays a lateral apophysis on its anterodor- Preaxial to Digit I is the prepollex (Fabrezi & sal border. Its ventral side extends below the Alberch, 1996). Between the terminal and sub- metacarpals to constitute the palmar surface, terminal phalanges there is an intercalary ele- which presents two points of contact: postaxi- ment that can be found in all digits (of both ally at the level between Metacarpals IV–III hands and feet) believed to be related to the and preaxially below Metacarpal II. Above the mechanical requirements of arboreal habits latter, slightly lower, point of contact, Distal (Manzano et al., 2007). It is fully mineralized, Carpal 5-4-3 presents a traverse tunnel along wedge-shaped, and presents biaxial articula- the anterior–posterior axis, plausibly a rem- tions both distally with the subterminal pha- nant from the secondary fusion of embryonic lanx (sellaris type) and proximally with the Distal Carpal 3 into this compound bone (Fab- terminal phalanx (plane type) (Manzano et rezi & Alberch, 1996). The radiale articulates al., 2007). The terminal phalanges are differ- preaxially with the posterior margin of the entiated from the other phalanges by lacking a distal end of the radius. It is cuboid in shape terminal epiphysis; they are straight or slight- and presents two projections: one dorsome- ly curved downward with a V-shaped tip, lack dially that contacts that of the ulnare, and apophyses, and have an ellipsoidal proximal another ventromedially that articulates with epiphysis (Manzano et al., 2007). The other the dorsal surface of Distal Carpal 5-4-3 at phalanges are elongate and articulate at their the level of the postaxial border of Metacar- proximal end with their corresponding digit’s pal II. The former results in the formation of metacarpal. Both the non-terminal phalanges two medial fenestrae between both carpals: and the metacarpals present a denser diaphy- a small dorsal one below the distal end of sis than epiphysis (fig. 10). the radioulna and another larger ventral one that ends at the dorsal surface of Distal Car- Vertebral column pal 5-4-3. Element Y (Shubin & Alberch, 1986) The vertebral column (fig. 8) is composed of extends ventrolaterally from the radiale’s pos- nine vertebrae [eight presacrals (I–VIII) and terior margin. This bone articulates medially the sacrum (IX)] and the urostyle (i.e., fused with Distal Carpal 5-4-3, and distally along its postsacrals; coccyx). Each vertebra is com- dorsal margin with Metacarpal I and along its posed of a centrum and a neural arch, which ventral margin with the reduced Distal Carpal ossify separately and then fuse. The neural 2. Preaxially and dorsal to Element Y lies the arches in B. transmarina are non-­imbricate, prepollex. It appears to be composed of three decrease slightly in height posteriorly, and elements, the most distal of which is a less join medially above the spinal cord in a dense elongated acuminate process. Distal blunt neural spine, or crest, which serves as Carpal 2 articulates distally with Metacarpal a muscle and ligament attachment site. Each I, while the second segment of the prepollex vertebra bears two pairs of processes located articulates postaxially and only along its an- dorsally on the neural arches for articulation terolateral border with this same element. with adjacent vertebrae: the prezygapophyses There are four Digits (I–IV), with the meta- anteriorly and the postzygapophyses posteri- carpal of Digit I lacking the nuptial tubercle orly. In addition, a pair of transverse process- distinctive of anurans that perform amplexus es emanate laterally from the neural arches; (Altig, 2008). The phalangeal formula is 2– these rise from a more ventral position on 2–3–3 (Digits I–IV), Digit III being more presacral II to a more dorsal one on presacral

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 48 Santos-Santos et al.

FIGURE 10 Medial/preaxial view of the Blommersia transmarina intercalary elements and terminal phalanges [MNCN50446 male; volrenWhite (A) and physics colormaps (B), grey [density] values: 12000–55500].

VII and are expanded in presacrals II–IV due Kluge & Farris, 1969; perichordal with chorda- to the attachment of musculature for suspen- central additions: Wake, 1970; Gardiner, 1983). sion of the pectoral girdle. Ribs are absent in The centra in this species are dorsoventrally B. transmarina. depressed (characteristic of stegochordal and In addition to the vertebral articulations, epichordal development) with a less dense the successive monospondylous vertebral core (probably a remnant of the notochord, or centra (Duellman & Trueb, 1994) articulate hypochord) toward its posterior end (charac- via condyloid joints. These articulations are teristic of perichordal development). The hy- displasiocoelous in B. transmarina, meaning pochord remnants are absent in the atlas and all articulations are procoelous (i.e., centrum increase in size caudally. concave anteriorly and intervertebral carti- Presacral I (Atlas). The first vertebra (i.e., lage associated with the posterior end of each presacral I) is known as the atlas and is modi- centrum; allows movement in two planes) fied anteriorly at its articulation with the except for that of the presacral VIII, which is skull, bearing two cup-shaped atlantal cotyles ­amphicoelous (i.e., centrum biconcave and (instead of prezygapophyses) that form con- separated by intervertebral cartilage that may dyloid joints with the occipital condyles of or may not be independent of adjacent centra). the skull. Transverse processes are absent and The centra of B. transmarina do not fit some individuals present an incompletely os- perfectly into the categories for the devel- sified neural spine (e.g., MNCN50431, 50432). opmental classification of anuran vertebral Presacral II. The transverse process of pre- centra (i.e., ectochordal, holochordal, stego- sacral II is located relatively more ventrally on chordal: Griffiths, 1959; perichordal, epichord- the neural arch than presacrals IV-VIII. It is dor- al: Mookerjee,­ 1931; Mookerjee & Das, 1939; soventrally flattened along its full length and

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 49 its leading edge displays a slight ventrodorsal posterolateral direction and ends in a truncat- inclination. It branches from the neural arch ed cylinder. Its distal extremity is thickened anterolaterally. Additionally, the transverse to different degrees, more so posteriorly, de- process is relatively shorter than in the rest of pending on the individual. presacral vertebrae (table 2). In two individu- Presacrals V–VIII. The transverse processes als (i.e., MNCN50431, 50437) the neural spines of presacrals V–VII branch from the neural of the atlas and presacral II are imbricate. arch in a posterolateral direction, present- Presacral III. Presacral III displays the lon- ing a gradient of decreasing magnitude from gest transverse processes of all vertebrae presacral IV to presacral VII; the presacral (table 2). The transverse process branches VIII breaks this trend, branching once more laterally from the neural arch lower than in slightly in an anterolateral direction. Unlike presacral II. Additionally, its leading edge presacrals II–IV, the transverse processes of presents a steeper ventrodorsal inclination posterior presacrals V–VIII are not expanded, relative to presacral II. The transverse process but are narrower and rod-like in shape. of presacral III displays the largest dispar- Sacral vertebra. The sacrum is a single spe- ity of all vertebrae; there are three different cialized vertebra from which the pelvic girdle characters that appear developed to differ- is suspended (fig. 11). It is located posterior to ent degrees depending on the individual and the presacrals and anterior to the urostyle, or with no particular pattern (e.g., sexual di- coccyx. The sacrum bears a pair of prezyg- morphism, symmetry, …): (A) The develop- apophyses but lacks postzygapophyses, pre- ment of a protuberance nearly three-fourths senting two posterior condyles instead. Its the distal distance on the tranverse process’s transverse processes are expanded and articu- anterodorsal edge ranges from a small pro- late ventrolaterally with the ilia. In B. transma- tuberance (e.g., MNCN50446, DRV6851) to a rina the sacral diapophyses are thick, rod-like, large anteriorly directed process with a trian- dorsoventrally compressed cylinders. These gular profile (e.g., MNCN50432, 50435). From are located relatively more dorsally than in the tip of this process, the transverse process the rest of vertebrae nearly at the level of its flattens dorsoventrally to its extreme in the neural spine, which leaves the anterior tip of lateral direction. (B) Another protuberance, the ilial shafts in line with the transverse pro- displaying significantly less variation in size cesses of the presacral vertebra. and position, is located on the posterodorsal Urostyle. The urostyle is a long straight edge of the transverse process slightly more shaft that has a bicondylar articulation at its proximal than (A). (C) At the distal extrem- proximal end with the sacral vertebra (fig. 11). ity of the transverse process some individuals It presents a dorsal urostylic crest that occu- (e.g., MNCN50433, 50436) present light calci- pies slightly less than ⅔ of its shaft (table 2). fications that can contact ossified portions of The crest starts anteriorly at a large tubercle, the posterior suprascapula. which can be ossified to varying degrees de- Presacral IV. Relative to presacrals II– pending on the individual, and progressively III, presacral IV’s transverse process is less decreases in height in the caudal direction. ­dorsoventrally flattened, and branches more Where the urostylic crest ends, the shaft re- dorsally on the neural arch in a similar man- duces in density, flattens dorsoventrally, and ner to the succeeding vertebrae (presacrals continues until slightly before the caudal bor- V–VIII). It branches from the neural arch in a der of the ischium.

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 50 Santos-Santos et al.

FIGURE 11 Dorsal view of the Blommersia transmarina pelvic girdle. The volume is inclined slightly to the right and clipped anteriorly immediately after presacral VIII [MNCN50446 male; volrenWhite (A) and phys- ics colormaps (B), grey [density] values: 12000–55500].

Pelvic girdle the sagittal-hinge morphotype being charac- The pelvis comprises three paired elements teristic of long-distance jumpers (i.e., > eight that unite in a medial symphysis: the ilium, body lengths; Zug, 1978; Emerson, 1979) since the ischium, and the pubis. The anterior half it allows dorsoventral excursion of the pelvic of the pelvic wheel is formed by the expanded girdle along the vertical plane. This morpho­ posterior portion of the ilium, whereas the type in B. transmarina is characterized by pos- posterior half is comprised by the ischium; terolaterally oriented sacral diapophyses with the cartilaginous pubis is limited to its ventral round distal borders and well developed joint portion, presenting varying degrees of ossifi- capsules at each ilial shaft articulation, pre- cation and fusion (fig. 12). senting a narrow transverse ligament deep to Ilium. The ilial shafts articulate at their an- the back musculature instead of a groove-like terior end with the ventral side of the distal articulation characteristic of Emerson’s (1982) ends of the sacral transverse processes (fig. 11). type I. In some individuals (e.g., MNCN50430, In B. transmarina this articulation is of type 50436, 50437), there are sesamoid bones wi­ IIB (Emerson, 1982) or of the sagittal-hinge thin the ligament of the articulation that align type (Reilly & Jorgensen, 2011). The configu- with the tips of the sacral diapophyses. ration of the ilial–sacral articulation largely The ilial shafts of B. transmarina have large determines the locomotor mode of a species; crests along their full length, which originate

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 51

FIGURE 12 Right lateral view of the Blommersia transmarina pelvic girdle. The volume is clipped anteriorly at presacral VIII [MNCN50446 male; volrenWhite (A) and physics colormaps (B), grey [density] values: 12000–55500].

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 52 Santos-Santos et al.

FIGURE 13 Ventrolateral view of the right Blommersia transmarina hind limb [MNCN50446 male; volrenWhite (A) and physics colormaps (B), grey [density] values: 12000–55500]. anteriorly lateral to the ilial–sacral articula- to distinguish­ (fig. 12). However, the presence tion and terminate at a well-developed lat- of a bony plate, suggestive of an ‘epipubis’ eral protuberance (i.e., tuber superius; Roček, (i.e., De Villiers, 1925: chondrification of the 1994) immediately before the ilium expands posterior linea alba; Ritland, 1955; Ročková to form the anterior part of the acetabulum & Roček, 2005: praepubis), at the posterior (ilial p in fig. 12). The crest increases in height border of the acetabulum may indicate its ar- posteriorly and presents two grooves: one ticulation with the ischium. At the proposed medially along three-fourths its length due articulation, the posterior origin of the plate to the medial curvature of the crest tip, and presents two lateral processes that invade the one laterally below the posterior protuber- acetabulum medially. The articulation of the ance that extends ¼ its length in the rostral ischium with the ilium is equally difficult to direction. The anterior and posterior expan- distinguish; however, we note a thickened sions of the ilium (i.e., pars ascendens and bony cap on the tip of the posterior-most bor- pars descendens, respectively; Roček, 1994) der of the pelvic wheel where other authors are prominent, slightly more posteriorly, propose the location of this articulation (Eck- and present varying degrees of ossification er, 1889; Duellman & Trueb, 1994). depending on the individual. There are two concavities between the ilial shafts: one pos- Hind limb terior to the acetabulum and another ante- Femur. The femur is a long bone with a smooth rior to it that is relatively deeper (ilial cav in diaphysis that presents a slight sigmoidal fig. 11). curve (fig. 13). The femur head is round and Ischium and Pubis. The ischium comprises articulates with the acetabulum of the pelvis the posterior half of the pelvic wheel and can (fig. 12). At its distal end the bone widens and vary considerably in shape due to the me- terminates in a truncated surface with two chanical stresses exerted by the musculature condyles (i.e., medial and lateral) that articu- in this area (Duellman & Trueb, 1994). The late with the tibiofibula. pubis is mostly calcified in B. transmarina, Tibiofibula. The tibia and fibula are com- more densely posteriorly, and its articula- pletely fused into a single bone (i.e., the tib- tions with the ischium and ilium are difficult iofibula). Its head bears two condyles with a Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 53

Table 3 Statistical results of one-way, two-way, and three-way analyses of covariance for the 123 skeletal meas- ures taken of Blommersia transmarina. Measures are partitioned into anatomical regions to facilitate lecture: SVL, skull, pectoral girdle, forelimb, vertebral column, pelvic girdle, and hind limb. Significant results (α = 0.05) are shown in bold. Abbreviations follow those indicated in the description of the measurements (Appendix 1). ANCOVAS 3-way 1-way 1-way 2-way SVLext SVLsk SVLsum SVLsk pelvL SVLsk pelvL

SVLext - - - .002 .000 .374 .035 SVLsk - - - - .000 - - SVLsum - - - .000 .000 .805 .132 skullL .658 .009 .429 .000 .000 .035 .634 skullW .840 .073 .365 .000 .000 .579 .057 oticsW .276 .029 .381 .000 .000 .386 .377 colmL .188 .740 .851 .008 .005 .965 .417 pmaxH .013 .002 .009 .001 .001 .591 .449 pmaxW .175 .012 .181 .000 .000 .393 .358 pmaxD .087 .079 .139 .004 .005 .604 .686 MmbL .500 .555 .699 .141 .770 .225 .422 maxL .504 .000 .039 .000 .000 .007 .321 mandL .150 .001 .062 .000 .000 .118 .640 nasL .825 .381 .911 .004 .003 .945 .374 nasW .058 .726 .493 .112 .660 .288 .537 intnasW .349 .956 .297 .135 .264 .252 .154 intnsptH .414 .720 .355 .434 .634 .297 .242 sphH .796 .283 .761 .004 .009 .224 .717 sphW .681 .768 .461 .023 .016 .965 .468 frptlL .176 .010 .088 .001 .001 .485 .574 frptlW .090 .875 .654 .044 .013 .200 .061 proH .021 .007 .009 .024 .028 .626 .879 exoL .342 .131 .708 .000 .000 .965 .135 psphH .746 .198 .517 .049 .043 .919 .653 psphW .006 .173 .030 .040 .020 .589 .250 vomH .649 .928 .566 .424 .557 .493 .394 vomW .454 .489 .545 .115 .067 .284 .529 pltnL .787 .440 .727 .120 .068 .268 .502 ptygdH .849 .136 .821 .002 .001 .816 .355 ptygdW .658 .019 .608 .000 .000 .037 .896 squaH .168 .011 .137 .000 .000 .172 .939 squaW .936 .009 .301 .000 .005 .010 .103 hyoL 1.000 .864 .434 .007 .007 .706 .636 omostH .899 .228 .682 .001 .001 .600 .451

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 54 Santos-Santos et al.

Table 3 Statistical results of one-way, two-way, and three-way analyses of covariance for the 123 skeletal measures taken of Blommersia transmarina. Measures are partitioned into anatomical regions to facilitate lecture: SVL, skull, pectoral girdle, forelimb, vertebral column, pelvic girdle, and hind limb. Significant results (α = 0.05) are shown in bold. Abbreviations follow those indicated in the description of the measurements (Appendix 1) (cont.). ANCOVAS 3-way 1-way 1-way 2-way SVLext SVLsk SVLsum SVLsk pelvL SVLsk pelvL omostL .578 .955 .265 .013 .018 .472 .997 clavL .485 .587 .942 .005 .003 .932 .309 clav_pL .794 .976 .227 .005 .002 .935 .299 coraL .777 .497 .603 .001 .000 .909 .185 sterH .307 .214 .338 .001 .004 .192 .760 sterWsup .769 .275 .928 .016 .009 .863 .358 sterWmin .407 .446 .532 .065 .028 .361 .148 sterWinf .458 .358 .726 .074 .063 .574 .661 scapL .881 .792 .310 .003 .003 .661 .574 cleiL .323 .083 .307 .003 .003 .630 .605 supraH .928 .667 .661 .019 .007 .527 .173 supraW .521 .105 .901 .000 .000 .190 .878 humLcap .136 .021 .528 .000 .000 .482 .083 humLtroc .125 .009 .406 .000 .000 .362 .060 dltd_crH .713 .172 .503 .030 .019 .867 .417 dltd_crW .201 .498 .360 .094 .064 .411 .723 radulnL .115 .005 .056 .000 .001 .364 .660 mtcpl1 .432 .195 .602 .003 .002 .758 .470 mtcpl2 .182 .084 .199 .005 .006 .573 .744 mtcpl3 .868 .210 .968 .008 .004 .889 .319 mtcpl4 .681 .225 .812 .003 .002 .934 .342 hand_ph1L1 .488 .270 .209 .000 .000 .398 .525 hand_tph1L .932 .078 .633 .003 .010 .090 .394 hand_D1 .769 .157 .968 .001 .001 .650 .414 hand_ph2L1 .901 .242 .772 .001 .001 .317 .838 hand_tph2L .685 .301 .960 .105 .141 .767 .486 hand_D2 .472 .131 .494 .003 .004 .852 .442 hand_ph3L1 .623 .110 .857 .000 .000 .507 .245 hand_ph3L2 .483 .029 .561 .000 .000 .050 .789 hand_tph3L .631 .142 .835 .041 .059 .438 .811 hand_D3 .961 .111 .870 .001 .001 .476 .588 hand_ph4L1 .879 .233 .924 .016 .027 .363 .807 hand_ph4L2 .205 .287 .301 .008 .017 .233 .654 hand_tph4L .681 .206 .532 .031 .041 .512 .947 hand_D4 .888 .202 .964 .004 .005 .853 .462 column .794 .711 .133 .000 .000Downloaded .914from Brill.com10/04/2021.100 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 55

ANCOVAS 3-way 1-way 1-way 2-way SVLext SVLsk SVLsum SVLsk pelvL SVLsk pelvL atlasH .259 .369 .367 .018 .042 .171 .444 V2H .850 .445 .467 .001 .002 .403 .803 V3H .525 .584 .904 .017 .008 .718 .270 V4H .685 .885 .241 .001 .000 .987 .200 V5H .367 .361 .168 .007 .002 .618 .159 V6H .784 .579 .127 .008 .005 .904 .329 V7H .755 .330 .193 .027 .007 .262 .073 V8H .450 .093 .529 .000 .000 .724 .045 sacrumH .332 .151 .471 .002 .000 .340 .037 V2_tpL .454 .846 .229 .009 .018 .291 .743 V3_tpL .922 .175 .449 .000 .000 .220 .741 V4_tpL .361 .257 .053 .000 .000 .472 .273 V5_tpL .931 .673 .058 .000 .000 .916 .127 V6_tpL .692 .385 .033 .002 .001 .820 .161 V7_tpL .998 .119 .017 .012 .005 .658 .236 V8_tpL .697 .045 .005 .005 .001 .686 .166 sacrum_tpL .650 .700 .350 .001 .000 .898 .244 uroL .267 .609 .371 .000 .000 .337 .013 uroH .275 .019 .005 .000 .000 .313 .096 uro_crL .672 .043 .084 .876 .860 .202 .203 uro_crH .286 .330 .620 .000 .000 .208 .003 iliL_lft .668 .071 .263 .000 .000 .923 .002 pelvL .175 .013 .963 .000 - - - pelwL .856 .019 .524 .000 .001 .062 .631 epipH .345 .102 .312 .006 .009 .465 .888 epipW .630 .480 .597 .211 .141 .366 .576 femL .196 .003 .123 .000 .000 .029 .632 tibfibL .109 .007 .124 .000 .000 .133 .850 tiblfiblL .064 .007 .035 .001 .003 .281 .995 mttrsl1 .639 .234 .670 .009 .013 .445 .991 mttrsl2 .408 .076 .361 .002 .005 .216 .777 mttrsl3 .257 .032 .177 .002 .004 .239 .864 mttrsl4 .132 .002 .033 .000 .002 .126 .728 mttrsl5 .225 .022 .160 .001 .001 .360 .761 foot_ph1L1 .928 .130 .786 .003 .006 .282 .885 foot_tph1L .681 .103 .857 .009 .021 .191 .550 foot_D1 .819 .158 .722 .004 .008 .890 .320 foot_ph2L1 .885 .198 .949 .001 .001 .723 .386 foot_tph2L .711 .211 .910 .041 .710 .301 .603 foot_D2 .603 .085 .561 .001 .002 .987 .276 Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 56 Santos-Santos et al.

Table 3 Statistical results of one-way, two-way, and three-way analyses of covariance for the 123 skeletal measures taken of Blommersia transmarina. Measures are partitioned into anatomical regions to facilitate lecture: SVL, skull, pectoral girdle, forelimb, vertebral column, pelvic girdle, and hind limb. Significant results (α = 0.05) are shown in bold. Abbreviations follow those indicated in the description of the measurements (Appendix 1) (cont.). ANCOVAS 3-way 1-way 1-way 2-way SVLext SVLsk SVLsum SVLsk pelvL SVLsk pelvL foot_ph3L1 .856 .278 .510 .001 .000 .951 .170 foot_ph3L2 .988 .046 .785 .000 .000 .449 .462 foot_tph3L .661 .031 .694 .002 .007 .091 .438 foot_D3 .578 .038 .513 .000 .000 .684 .288 foot_ph4L1 .660 .108 .908 .000 .000 .819 .054 foot_ph4L2 .958 .134 .768 .000 .000 .871 .148 foot_ph4L3 .232 .002 .963 .000 .000 .032 .618 foot_tph4L .905 .006 .217 .001 .005 .133 .649 foot_D4 .530 .012 .321 .000 .000 .560 .218 foot_ph5L1 .919 .383 .883 .006 .003 .879 .282 foot_ph5L2 .722 .012 .668 .000 .000 .162 .959 foot_tph5L .116 .094 .388 .009 .029 .126 .406 foot_D5 .787 .068 .660 .001 .001 .576 .422 thick bony ridge that crosses three-fourths of ­metaphyses, and only a short distance along their medial surface. The diaphysis is smooth the most distal part of their smooth diaphy- and straight, presenting the nutrient foramen ses. The tibiale presents a sesamoid bone (i.e., halfway along its length. Its epiphysis also ossification of the cartilago sesamoides; Nuss- presents two condyles, which articulate with baum, 1982) on its proximal ventral epiphysis, the tibiale and fibulare. and a pointed, medially-oriented apophysis Foot. The foot consists of six mesopodial el- on its distal epiphysis (fig. 13). ements (i.e., tibiale, fibulare, centrale, hallux– Sole. Ventral to the fibulare’s distal epiphy- tarsal, and two distal tarsals), five metapodial sis is an ovoid plantar (Ponssa et al., 2010) os elements (i.e., metatarsals), the prehallux, the sesamoideum (Nussbaum, 1982). Its longitu- phalanges, intercalary cartilage elements, and dinal axis lies in the proximal–distal direc- a heterotopic sesamoid (Hall, 2005). There are tion and only its proximal margin is in con- five Digits (I–V) with relatively more elongat- tact with the fibulare (fig. 13). Its distal border ed metapodial elements and phalanges than ends ventral to the separation of the proximal the hand. The phalangeal formula is 2–2–3– epiphyses of Metatarsals V and IV. In addi- 4–3 (Digits I–V). tion, its distal segment presents preaxially Tibiale and Fibulare. The preaxial tibiale either a rounded apophysis or an additional and postaxial fibulare are mesopodial ele- sesamoid bone (e.g., ossification of the carti- ments of the hind limb that are extremely lago plantaris; Hoyos, 2003). The centrale (aka elongated relative to their homologues in naviculare; Howes, 1888) articulates along its other tetrapods. They are fused both proxi- complete surface preaxially on the tibiale’s mally and distally at their epiphyses and distal epiphysis. It presents an extensive Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 57 ventral apophysis­ below the tibiale that ex- skeletal SVL measure was the best size predic- tends distally at its base, producing a sort of tor, being the only significant covariable in fulcrum. The centrale articulates postaxially most cases and presenting the lowest p-value with Metatarsal I, and with the prehallux pre- when more than one covariable was signifi- axially before the branching of its apophysis. cant (table 3). One-way ANCOVAs showed Embryonic Distal Tarsals 2 and 3 fuse into a that all variables covaried with SVLsk and diagonally-oriented ellipsoidal bone along pelvL except mentomeckelian bone length, the preaxial–postaxial axis (i.e., cuboideum; nasal width, inter-nasal width, internasal sep- Howes, 1888). The cuboideum’s preaxial mar- tum height, vomer height and width, palatine gin lies below Metatarsal II, articulating with length, sternum inferior widths, deltopectoral this element only at its dorsal margin (fig. 13), crest width, hand Digit II terminal phalanx whereas it articulates with Metacarpal III length, urostyle crest length, and epipubis along its complete distal postaxial margin. width. In addition, sternum minimum width In between the dorsal segment of the cen- did not covary with SVLsk, and hand Digit trale and the preaxial head of the cuboideum, III and foot Digit II terminal phalanx lengths lies the hallux–tarsal (Howes, 1888); a small did not covary with pelvL (table 3). Two-way nodule of calcified cartilage. The fused distal ANCOVAs, including both SVLsk and pelvL, epiphyses of the tibiale and fibulare articulate resulted in SVLsk showing significant covaria- preaxially with the cuboideum and postaxial- tion with skull length, maxilla length, ptery- ly with the metatarsals of Digits IV and V. The goid width, squamosal width, hand Digit III morphological characteristics of the interca- terminal phalanx length, femur length, and lary elements in the foot are the same as those foot Digit IV terminal phalanx length; and for the hand (see above). with pelvL showing significant covariation with exterior snout–vent length, presacral Intraspecific variation and statistical VIII height, sacrum height, urostyle length, analyses urostyle crest height, and ilial shaft length A multivariate ANOVA on the 123 skeletal (table 3). measures detected sexual dimorphism in B. transmarina (Pillai F = 2874.45; p = 0.014; par- Figure abbreviations tial eta-squared = 1), with females present- a alae ing larger values for all measured variables ant anterior (table 2). Between-subject effects showed ap apophysis sexual dimorphism in the following variables c. cartilage (table 2): skull length, skull width, premax- C.N. cranial nerve illa width, premaxilla depth, maxilla length, cav cavity mandible length, nasal length, sphenethmoid cr crest/crista width, parasphenoid height, palatine length, dor dorsal pterygoid height and width, squamosal width, epio. e. epiotic eminence omosternum height, sternum inferior length, f foramen hand Digit III phalanx 2 length, and foot Digit ff fontanelle/fenestra IV terminal phalanx length. med medial Univariate three-way ANCOVAs includ- p process/protuberance ing all three snout–vent measurements (i.e., pl plate SVLext, SVLsk, SVLsum) determined that the pp pars

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 58 Santos-Santos et al. pst posterior pltn palatine r ramus ptyd pterygoid so.fl. supraorbital flange qdj quadratojugal ven ventral qu quadrate zygo zygomatic rad radius radle radiale a.c. alary cartilage s.n. septum nasi agspl angulosplenial sc scapula amph.r. amphibiorum recess sph sphenethmoid at atlas sptm septomaxilla clei cleithrum sq squamosal clv clavicle ssc suprascapula cntl centrale st sternum col columella t.n. tectum nasi crc coracoid t.ph. terminal phalanx d dentary uln ulna dcpl distal carpal ulne ulnare elym endolymph uro urostyle E.t. Eustachian tube vom vomer exo exoccipital fem femur fp frontoparietal Discussion hms humerus hyob hyobranchial Blommersia transmarina osteological int.el. intercalary element peculiarities lg lagena There are many aspects of hyperdiverse tropi- Mm.b. mentomeckelian bone cal radiations for which we lack data that mtcrp metacarpal could help us to understand the patterns and mttrl metatarsal processes related to species’ diversification mx maxilla and speciation mechanics. These include de- n / nas nasal tailed data on anatomical structures as well as o.c. otic capsule phenotypic variation (i.e., intraspecific dispar- omst omosternum ity), which have been key to diagnose species op operculum since Linnaeus. However, despite the fact that oto otoccipital osteological and anatomical data have been phy parahyoid at the core of species descriptions in the clas- plym perilymph sical literature, for most species we currently preh prehallux only rely on data of their external morphology prem premaxilla (Vences et al., 2002; Glaw et al., 2006; Glaw & prep prepollex Vences, 2007). This is the case in the hyperdi- prezyga prezygapophysis verse family Mantellidae, where the majority pro prootic of the recently described species mostly inte- psph parasphenoid grate external morphology, genetic data, and pstzyga postzygapophysis bioacoustics (Vieites et al., 2009; Vences et al.,

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 59

2010, 2015; Scherz et al., 2017b), with few exam- Reilly & Jorgensen, 2011; Soliz et al., 2017). ples including internal structures such as the A better understanding of the functional sig- presence or absence of vomerine teeth in the nificance of skeletal traits could significantly skull or bone structures in hands or feet (Man- improve the ecomorphological interpretation zano et al., 2007; Kamermans & Vences, 2009). of the interactions between these characteris- Here, we present the thus far most detailed tics and locomotor mode and/or habitat use osteological description of a mantellid frog (Arnold, 1983), and help to understand the with two main purposes: (1) characterize the evolution of the mantellid adaptive radiation. skeletal (cranial and postcranial) anatomy of The clavicular process in B. transmarina a mantellid species with its intraspecific varia- is located at the proximal end of the clavicle tion, and (2) provide a reference base for future and articulates with the posterior branches studies concerning the comparative anatomy, of the omosternum (fig. 7). This structure has phenotypic variation, and taxonomic relation- not been reported so far in any anuran spe- ships between the members of this family. cies (e.g., Emerson, 1988; Duellman & Trueb, The genus Blommersia comprises 10 de- 1994; Soliz et al., 2018); however, some have scribed species from Madagascar and one suggested that the clavicle may be involved in candidate species from the Comoros. All spe- impact absorption during landing (Emerson, cies vary from 14 to 30 mm snout–vent length 1984, 1988). In CT-scans of stained specimens and are within the smaller-size range of the (J.H.S.-S., pers. obs.) this process was ob- mantellid frog spectrum. However, B. trans- served to be embedded within the mm. del- marina from the Comoros, the one for which toideus and coracoradialis, which flex at the we here describe its osteology, is larger than shoulder and elbow joint, respectively. Fur- the rest and has likely undergone a process of thermore, we cannot exclude the possibility moderate gigantism (Glaw et al., 2019), which that the pectoral and forelimb musculature in is commonly found in island taxa (Daugherty B. ­transmarina or other mantellines presents a et al., 1993; Lomolino, 2005; Li et al., 2011). This more elaborate architecture than that of other makes it an interesting species to study its frog taxa, involving supplementary musculo- musculoskeletal system as a prior to compara- tendon connections between the clavicular tive morphological analyses with other Blom- process and the deltopectoral crest of the hu- mersia and mantellid frogs. So far, there are no merus or other forelimb elements. To test this studies on the evolution of the musculoskele- idea, we are working on a segmentation of tal system in mantellids, nor on their jumping the pectoral girdle’s and forelimbs’ muscular performance and kinematics, all of which are architecture in B. transmarina. However, we directly linked to dispersal capacity, habitat have also observed the presence of the cla- selection, and speciation. This lack of infor- vicular process in several more species within mation precludes inferring conclusions about the genus Blommersia (e.g., B. wittei, B. sarotra, the relationships between their functional B. grandisonae, B. domerguei, B. blommersae). morphology and adaptive processes such Thus, for the time being, we put forward a dif- as dispersion, size evolution, or phenotypic ferent locomotor function than those suggest- plasticity and adaptation to specific habitats. ed by Emerson (1988), which is in line with Postcranial morphology in anurans has been the unique reproductive behavior of Mantel- linked numerous times to a specific locomo- linae frogs (Glaw & Vences, 2006): As part of tor mode and/or habitat use (Zug, 1978; Em- their distinctive reproductive behavior (Blom- erson, 1979, 1982; Nauwelaerts et al., 2007; mers-Schlösser, 1975), Mantellinae frogs cling

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 60 Santos-Santos et al. vertically in parallel on leaves or walls during found in B. transmarina. Such skeletal ele- fertilization. This position does not involve ments have been reported in other anuran the frog grasping a solid border to ascend with groups (Nussbaum, 1982; Hoyos, 2003) and are its forelimbs, thus we hypothesize that the said to be biomechanically advantageous in presence of the clavicular process may pro- that they maintain the shape of and strength- vide some sort of biomechanical structural en tendons, as well as increase the mechanical support for the musculature involved in the advantage of force translation at articulations adhesion of the frog to the leaf’s surface. Until (Olson, 2000; Summer & Koob, 2002; Abdala segmentation of the involved musculature is et al., 2018). Other developmental studies complete, we cannot make any inferences on have revealed that certain­ sesamoids differen- the muscular mechanism involved since most tiate before a mature tendon tissue is recog- studies on frog adhesion focus on the digit nizable, thus showing their origination not to pads, which are also present in B. transmarina be solely due to extrinsic factors such as me- (see below). chanical loading. If we consider the categories This species also has several sesamoids of sesamoid distribution (i.e., omnipresent (i.e., appendicular ossicles; Hall, 2007) and and fluctuating; Hall, 2007), these different a heterotopic bone (i.e., cartilage plantaris circumstances relating to their presence sug- ap in fig. 13) on the sole. The most extensive gest that the origin of phylogenetically ubiq- review of sesamoids in anurans (Ponssa et uitous sesamoids may be more linked to al., 2010) presents many missing data for the ­conservative genetic and developmental stim- Mantellidae family on which we can now shed uli, while extrinsic factors may dominate over light with the inclusion of an additional genus genetic ones more in those whose occurrence (i.e., Blommersia). Of the sesamoids recog- is irregular (Ponssa et al., 2010). In Neobatra- nized by Ponssa et al. (2010) in Mantellidae, chia, the cartilago sesamoides is considered we confirm in B. transmarina the presence an autapomorphy, while the plantar sesamoid of the cartilago sesamoides embedded in the and those lateral to the sacral diapophyses are­ m. plantaris profundus at the articulation of considered to be plesiomorphic in Anura the tibiofibula with the tibiale. The presence (Ponssa et al., 2010). The presence of the lat- of this sesamoid was thought to have phylo- ter in several additional Blommersia species, genetic relevance (Nussbaum, 1982); however, with scarce morphological differen­ces be- this hypothesis has been put in doubt since tween them (only one individual of B. wittei Hoyos (2003). Other sesamoids observed in and one of B. domerguei differed), suggests­ B. transmarina that had not been described that these sesamoides may have arisen in before for the family are those lateral to the the genus in response to the same extrinsic sacral diapophyses, only in some individu- factors, possibly involving biomechanical als, and that on the plantar surface of the adaptations. foot. The latter is located where Hoyos (2003) In addition to heterotopic elements, B. considered the cartilago plantaris to be; how- transmarina presents two peculiar apophy- ever, in B. transmarina this heterotopic­ ele- ses in the sole that may also have a locomo- ment possesses an additional apophysis or tor function: a pointed apophysis on the associated os sesamoides (fig. 13). Ponssa tibiale’s distal epiphysis (absent in lightly et al. (2010) described three bony sesamoids in ­ossified MNCN50431 and B. blommersae) and the adult embedded in the plantar aponeu- a fulcrum that extends ventrally from the cen- rosis tendon that may correspond to those trale (Fig. 13). In a similar manner the Distal

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 61

Carpal 5-4-3 presents a lateral apophysis and ­terminal phalanges in B. transmarina, but for ventral expansions that result in two points the moment have not confirmed the presence of contact (fig. 9). The sesamoid-traction of their actuating muscles; however, there is epiphysis hypothesis (Parsons, 1904; Hall, no obvious correlation between the presence 2007) postulates that certain bony projections of intercalary elements or bifurcated terminal where tendons or ligaments insert develop phalanges and climbing habits: These char- independent of the limb itself, ­suggesting acters exist in several taxa, such as the man- that these apophyses represent sesamoids tellid genus Aglyptodactylus, which possess that have been incorporated onto the long these elements and are completely terrestrial. bones of the appendicular skeleton. This As such, these elements appear to not cause may most likely be the case for the tibiale’s any disadvantages to non-arboreal frogs and apophysis, and even for that of Distal Car- to not be under strong selective pressure. In pal 5-4-3. The bony projections on the pal- B. transmarina the intercalary elements are mar and sole surface may serve as a fulcrum strongly mineralized and we suspect they for better control and mechanical advan- aid the frogs during their reproduction when tage during jumping locomotion (Biewener, they hang vertically. The intercalary element 2003) analogous to the function hooves serve in anurans seems to be a morphological mountainous ungulates on steep slippery novelty that appeared early within the Neo- slopes. batrachia and is interpreted as an autopo- In contrast, the presence of intercalary el- morphy for the Rhacophoroidea (Frost et al., ements has been hypothesized to be related 2006). to arboreal lifestyles, with distal phalanges, Furthermore, B. transmarina presents sev- intercalary elements, muscles, and digit pads eral osteological peculiarities; the first that acting as integrated units to enhance climb- jumps into view is its evident signs of hyper- ing abilities (Burton, 2004; Fabrezi et al., ossification. Several regions of the skull (i.e., 2017). These are functionally integrated into internasal septum, prootics, sphenethmoid, a complex system, with distinctive morpho- and parasphenoid) present dermal sculpting, structural patterns between Hyloides and or exostosis, that evidence this phenomenon Ranoides (Manzano et al., 2007), related to and have been observed affecting the structure different types of movements produced by a of the skull in different ways in other anuran similar set of muscles and tendons. This sys- genera (Duellman & Trueb, 1994; Yeh, 2002; tem allows the angular movements necessary Clemente-Carvalho et al., 2009; Evans et al., for the attachment and detachment of the 2014). Other signs of hyperossification in der- adhesive digit pad (Hanna & Barnes, 1991). mal elements of B. transmarina include the fu- Kamermans & Vences (2009) found evidence sion of the prootics and exoccipitals into the for concerted evolution between climbing otoccipital and the production of extensive habits and bifurcated terminal phalanges in marginal flanges, such as that of the frontopa- that this architecture has two additional artic- rietals’ supraorbital flange or posterior protu- ulations that allow to move the digit without berance (fig. 1). Hyperossification is unrelated detaching the digit pad. These movements with the absolute size of a species (Duellman are produced by the action of the mm. exten- & Trueb, 1994), and accordingly we found dif- sores breves ­profundus and distalis, whose fering levels of ontogenetic ­ossification across presence is variable in Mantellidae (Kamer- the 10 individuals of B. transmarina in this mans & ­Vences, 2009). We observed bifurcated study. The degree of ossification of several

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 62 Santos-Santos et al. skeletal elements varied independently from Intraspecific osteological variation in size (i.e., SVL), sex, and each other, obfuscat- Blommersia transmarina ing the functional purpose of hyperossifica- tion in these structures, but suggesting that When describing the variation within a spe- their calcification may depend more on en- cies, it is important to distinguish individual vironmental cues rather than developmental from sexual and ontogenetic variation. The ones (i.e., genetic). For example, the devel- inclusion of 10 individuals of B. transmarina opment of bony crests on the humerus has in its osteological description has made it pos- been associated to hypertrophied muscula- sible to discriminate the different status and ture in male frogs (Duellman & Trueb, 1994). directionality of ontogenetic ossification in In B. transmarina the deltopectoral crest is several skeletal elements (e.g., skull, vertebral not only well developed in both sexes, but transverse processes, and pelvic wheel), es- also displays a continuation on the humerus tablish the ubiquity of appendicular ossicles, head separated by a fenestra of varying size. discriminate between individual artifacts and Thus, we hypothesize that ­hyperossification the general characteristics of the species, and of this structure may play an important role in determine what structures present sexual di- the musculoskeletal functional performance morphism. Characterization of the observed of both males and females in the Blommer- intraspecific variation leads us to interpret the sia genus. Furthermore, hyperossification of variables with positive values of asymmetry certain skeletal elements evokes the idea of (>1; table 2) to indicate skeletal structures in B. structural compensation due to increasing transmarina that are under a continuous pro- individual size or improved mechanical per- cess of ontogenetic ossification since all sam- formance at more developed stages (Pérez- pled individuals are adults and annual cohort Ben et al., 2018). This is especially true for survival in anurans decreases with increasing structures that serve as muscle attachment age. In B. transmarina we also observed an sites such as the bony crest on the hyobran- overall sexual dimorphism with females being chial’s posteromedial processes or the level relatively larger in all measurements (table 2). of development of diverse processes on the However, univariate analyses show that this transverse processes of presacral vertebrae II– does not hold true for snout–vent length and IV. However, the asymmetry in the degree of that most sexually dimorphic characters are development of the latter observed in several located in the head with the exception of the individuals (i.e., MNCN50430, 50435, 50437) omosternum, sternum, and a couple of pha- suggests that hyperossification­ of these skel- langes (table 2). Increased size of the head etal structures does not alter their functional in females may indicate an increased cranial performance since an asymmetrical musculo- volume and vomeronasal organ. This would skeletal lever system would obviously be bio- be a logical evolutionary outcome in view of mechanically unstable in a bilateral animal. the transmission of volatile compounds from This ambiguity in the role of the observed the males’ femoral glands to the females’ vo­ hyperossification still leaves unresolved the meronasal organ during courtship (Nowack & issue if this process has any (biomechanical) Vences, 2016; Nowack et al., 2017). Larger ster- adaptive value in non-burrowing frogs other na and phalanges may likewise aid the female than that of water conservation due to exos- biomechanically in maintaining her vertical tosis (Ruibal & Shoemaker, 1984; Evans et al., position during the reproductive act (Altig, 2014). 2008) by increasing muscle attachment area

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 63 and potentially providing longer input levers skull and limb elements, while pelvic length relative to output levers. was a better predictor of the elements of the The observation of sexual dimorphism vertebral column and pelvic girdle ­(table 3). distributed throughout the skeleton of B. These observations in size covariation suggest transmarina (table 2) allows us to test for size that the anuran skeletal architecture is orga- covariation of skeletal characteristics. The nized in modular developmental units. In this traditional length measurement in anurans context, Soliz et al. (2017) determined that the has always been (external) snout–vent length; girdles and vertebral column in hylids adjusted however, the precision and repeatability of to different evolutionary modes of evolution this measure is highly dependent on the ob- that may well reflect a morphological con- server. This is because the structure of the pel- tinuum between locomotor modes (Fabrezi vic girdle in most anurans can compress under et al., 2017). This may also hold true for other vertical load and their vertebral column is flex- anurans like B. transmarina that we have de- ible, thus different measurement habits across scribed here, although further research includ- observers can lead to different values for the ing additional species of the genus is required same individuals. With this in mind we tested to develop this hypothesis further. three different snout–vent measurements to evaluate which one was a better predictor of Modern methods in anatomical size covariation. Our results (table 3) indicate reconstructions that the snout–vent length directly measured Despite the fact that osteological data have on the volume-rendered skeleton of the frog been a substantial source for phylogeneti- in its resting position is a better size covari- cally informative characters, the vast major- ate than either exterior snout–vent length or ity of phylogenies produced these days rely the summed skeletal snout–vent length (Ap- on genetic/genomic­ data. Nevertheless, the pendix 1). That the SVLsum measurement ­availability and expansion of μCT is facilitat- calculated by summing up the skeletal length ing the generation of massive amounts of distances of each anatomical region is a worse ­osteological data from vouchers deposited in predictor than the SVL measured directly natural history collections, and it is likely to be- on the rendered skeleton, suggests that the come more common as the technology further “crouched” resting position of B. transmarina develops shortening CT generation times (Cole is correlated with size variation in the entire et al., 2018). In this view, there are currently sev- anatomy of a particular individual and may eral ongoing projects and open resources that play a role in the allometric scaling of jumping together are making osteological data across performance in this species (Choi et al., 2000; almost all vertebrate genera readily avail- Azizi & Roberts, 2010; Gillis, 2010). Another able for phylogenetic research (Cross, 2017). proposed length measurement for anurans However, osteological data are sometimes that focuses on their locomotor function is hard to interpret and compare as they can be that of pelvic length. When compared to skel- ­influenced by the age of the specimen (i.e., etal snout–vent length, pelvic length presents ontogeny), allometric and/or heterochronic very similar size covariation with skeletal el- processes, or phenotypic plasticity in response ements (table 3); however, each covariable to local ecological conditions; in addition, the showed higher predictive power than the impact of all these factors varies across time, other in specific anatomical regions: skeletal phylogeny, and different parts of the skeleton. snout–vent length was a better predictor of the Despite this, osteological data are routinely

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 64 Santos-Santos et al. used in paleontological studies to describe fos- ­inferences on the development of the car- sils and infer their ­phylogenetic relationships pus (fig. 9), whose nomenclature has been a (e.g., Kluge & Farris, 1969; Hall, 2007; Biton et ­controversial topic in anurans until the works al., 2013; Laloy et al., 2013; Gao & Chen, 2017; by Fabrezi and Alberch (e.g., 1996). They pro- ­Marjanovic & Laurin, 2019). posed the absence of centralia in the anuran Another potential issue with osteological carpus because these arise from the interme- data generated from μCT is the processing dium. The intermedium in anurans has a du- and rendering of the CT-scans. Due to the dif- bious homology with that in other tetrapods ferential ossification of skeletal structures and because its embryonic origin is by segmenta- the small size of many frog species, the thresh- tion off the ulnare rather than branching off olds between ossified and non-ossified struc- the ulna. In B. transmarina we observed a me- tures are usually hard to establish. This has dial projection off the ulnare that corresponds caused discrepancies in the skeletal descrip- to the embryonic structure Fabrezi (1992) de­ tion when interpreting surface renders ver- signated to be homologous with the interme- sus volume renders (Scherz et al., 2017c). The dium, or the ulnare’. We believe this structure production of surface models requires the is a chondrogenetic projection rather than a ­establishment of an isosurface density value separate element in B. transmarina in view for the rendering, and this threshold may of the existing large fenestrae in-between not be the most adequate for the complete the ulnare and radiale in this species (fig. 9). anatomy of the sample, creating artifacts and In the same way, Distal Carpal 5 is not a cen- obfuscating the precise description of certain trale since it arises from a single embryonic morphological structures. The development condensation (Fabrezi & Alberch, 1996). In B. of software tools that could allow to differen- transmarina this element seems to segment tially surface render distinct modules of the twice, subsequently fusing primarily with Dis- vo­lume, and/or advances in the computation- tal Carpal 4 and secondarily with Distal Car- al weighting of isosurfaces that could allow pal 3 as evidenced in the presence of a tunnel to equalize differences between those gene­ between both elements (Distal Carpal 5-4 and rated between different anatomical modules Distal Carpal 3) in the compound bone Distal for their assemblage, may lead in the future Carpal 5-4-3 (Fabrezi & Alberch, 1996). The to more precise surface model interpretation. radiale in B. transmarina also presents two In contrast to isosurfaces that are generated projections (fig. 9), indicating two indepen- based on a particular density value, volume dent embryonic cartilaginous condensations: rendering allows for a dynamic visualization the (posterior) radiale dorsomedially and the of the model along a user-defined density (anterior) radiale’ ventromedially (Fabrezi & curve with different options of transparency ­Alberch, 1996). The latter is considered a new and coloration for predefined subranges of element characteristic of the hand ontogeny grey values. This allows exploring anatomical of some anuran species since the distal (pos- structures along their complete density gra- terior) end of the radiale gives rise to the El- dient and to render tissues that correspond ement Y and prepollex (Shubin & Alberch, to different subranges of grey values within 1986). The Element Y is a carpal element found the sample differentially, in consequence in- exclusively in amphibians that arises from one creasing the contrast between them during to three embryonic condensations. In B. trans- visualization. marina this element appears to originate from Density differentiation in the CT-scans two independent embryonic chondrifications of B. transmarina allowed us to make some that fuse primarily Downloadedlater on (fromfig. 9 Brill.com10/04/2021). 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 65

We are far from understanding how large relationship with locomotor modes. J. Morphol., tropical hyperdiverse endemic radiations, 279, 895–903. like those of frogs from Madagascar, evolved Altig, R. (2008) Notes on the breeding biology of and diversified. These Malagasy radiations are four species of mantellid frogs from Madagas- particularly interesting as they involve several car. Trop. Zool., 21, 187–194. cases of miniaturization as well as a wide array Arnold, S.J. (1983) Morphology, performance and of morphologies occupying different ecologi- fitness. Am. Zool., 23, 347–361. cal niches. While from the molecular perspec- Azizi, E. & Roberts, T.J. (2010) Muscle performance tive we have a good position to understand during frog jumping: influence of elasticity on phylogenetic relationships, from the anatomi- muscle operating lengths. Proc. R. Soc. B-Biol. cal one we are lacking a work of comparative Sci., 277, 1523–1530. morphology. Selection acted on the morphol- Biewener, A.A. (2003) Animal Locomotion. Oxford ogy of every species, allowing for adaptations University Press, Oxford, 285 pp. and the colonization of novel environments, Biton, R., Geffen, E., Vences, M., Cohen, O., Bailon, and future research on the anatomy of Mala- S., Rabinovich, R., Malka, Y., Oron, T., Boistel, R., gasy frog radiations will help to understand Brumfeld, V. & Gafny, S. (2013) The rediscovered how these species radiated and colonized all Hula painted frog is a living fossil. Nat. Com- the environments in which they are present. mun., 4, 1959. Blommers-Schlösser, R.M. (1975) A unique case of mating behavior in a Malagasy tree frog, Ge- Acknowledgements phyromantis liber (Peracca, 1893), with obser- We would like to thank Dominique Adriaens, vations on the larval development (Amphibia, for extensive feedback on the anatomical de- Ranidae). Beaufortia, 23, 15–25. scription, and Barbara de Kegel, for helping in Bolkay, S. (1919) Elements of the comparative os- the realization of the staining protocol, from teology of the tailless batrachians. Glas. Zemal. the Evolutionary­ Morphology of Vertebrates Muz. u Bosni i Hercegovini, 31, 277–358. Lab at Ghent University, and the members Burton, T.C. (2004) Muscles of the pes of hylid of the UGCT that aided us in the CT process- frogs. J. Morphol., 260, 209–233. ing of stained specimens. We thank the Unité Choi, I.-H., Shim, J.H., Lee, Y.S. & Ricklefs, R.E. Biodiversité-DEAL of Mayotte for research au- (2000) Scaling of jumping performance in thorization. This research was funded by the ­anuran amphibians. J. Herpetol., 2000, 222–227. ­Spanish Ministerio de Economía y Competi- Clack, J. (2001) The otoccipital region—origin, on- tividad (FEDER, EU) CGL2013-40924-P, and togeny and the fish-tetrapod transition. In: E. Agencia Española de Investigación CGL2017- Ahlberg (Ed) Major Events in Early Vertebrate 89898-R (AEI/FEDER, EU) to David Vieites. Evolution, pp. 400–413. CRC Press, London. Clemente-Carvalho, R.B., Antoniazzi, M.M., Ja­ red, C., Haddad, C.F., Alves, A.C., Rocha, H.S., Supplementary material Pereira, G.R., Oliveira, D.F., Lopes, R.T. & dos Supplementary material is available online at: Reis, S.F. (2009) Hyperossification in miniatur- https://doi.org/10.6084/m9.figshare.9699731 ized toadlets of the genus Brachycephalus (Am- phibia: Anura: Brachycephalidae): microscopic structure and macroscopic patterns of varia- References tion. J. Morphol., 270, 1285–1295. Abdala, V., Ponssa, M.L., Tulli, M.J., Fabre, A.C. & Cole, J.M., Symes, D.R., Lopes, N.C., Wood, J.C.,

Herrel, A. (2018) Frog tendon structure and its Poder, K., Alatabi,Downloaded S., Botchway,from Brill.com10/04/2021 S.W., Foster, 09:02:24AM via free access

204346 66 Santos-Santos et al.

P.S., Gratton, S., Johnson, S., Kamperidis, C., Emerson, S.B. (1988) Testing for historical patterns Kononenko, O., De Lazzari, M., Palmer, C.A., of change: a case study with frog pectoral gir- Rusby, D., Sanderson, J., Sandholzer, M., Sarri, dles. Paleobiology, 14, 174–186. G., Szoke-Kovacs, Teboul, L., Thompson, J.M., Evans, S.E., Groenke, J.R., Jones, M.E., Turner, A.H. Warwick, J.R., Westerberg, H., Hill, M.A., Norris, & Krause, D.W. (2014) New material of Beelze- D.P., Mangles, S.P. & Najmudin, Z. (2018) High- bufo, a hyperossified frog (Amphibia: Anura) resolution μCT of a mouse embryo using a com- from the Late Cretaceous of Madagascar. PLOS pact laser-driven X-ray betatron source. Proc. ONE, 9, e87236. Natl. Acad. Sci. U.S.A., 115, 6335–6340. Fabrezi, M. (1992) El carpo de los anuros. Alytes, 10, Cross, R. (2017) The inside story on 20,000 verte- 1–29. brates. Science, 357, 742–743. Fabrezi, M. & Alberch, P. (1996) The carpal ele- Crottini, A., Madsen, O., Poux, C., Strauss, A., Vie- ments of anurans. Herpetologica, 1996, 188–204. ites, D.R. & Vences, M. (2012) Vertebrate time- Fabrezi, M., Goldberg, J. & Chuliver-Pereyra, M. tree elucidates­ the biogeographic pattern of a (2017) Morphological variation in anuran limbs: major biotic change around the K–T boundary constraints and novelties. J. Exp. Zool. B, 328, in Madagascar. Proc. Natl. Acad. Sci. U.S.A., 109, 546–574. 5358–5363. Frost, D.R., Grant, T., Faivovich, J., Bain, R.H., Haas, Daugherty, C.H., Gibbs, G.W. & Hitchmough, R. A, et al. (2006) The amphibian tree of life. Bull. (1993) Mega-island or micro-continent? New Am. Mus. Nat. Hist., 297, 1–291. Zealand and its fauna.Trends Ecol. Evol., 8, Gao, K-Q. & Chen, J. (2017) A new crown-group frog 437–442. (Amphibia: Anura) from the Early Cretaceous De Villiers, C. (1925) On the development of the of Northeastern Inner Mongolia, China. Am. “epipubis” of Xenopus. Annals of the Transvaal Mus. Novitates, 3876, 1–39. Museum, 11, 129–135. Gardiner, B.G. (1983) Gnathostome vertebrae and Dempster, W.T. (1930) The morphology of the am- the classification of the Amphibia. Zool. J. Linn. phibian endolymphatic organ. J. Morphol., 50, Soc. Lond., 79, 1–59. 71–126. Gillis, G.B. (2010) Frog muscles start stretched. J. Duellman, W.E. & Trueb, L. (1994) Biology of Am- Exp. Biol., 213, vi. phibians. John Hopkins University Press, Balti- Glaw, F. & Vences, M. (2006) Phylogeny and genus- more, 696 pp. level classification of mantellid frogs (Amphib- Ecker, A. (1889) The Anatomy of the Frog. Claren- ia, Anura). Org. Divers. Evol., 6, 236–253. don Press, Oxford, 484 pp. Glaw, F. & Vences, M. (2007) A Field Guide to the Emerson, S.B. (1979) The ilio-sacral articulation in Amphibians and Reptiles of Madagascar, Third frogs: form and function. Biol. J. Linn. Soc., 11, Edition. Vences and Glaw Verlag, Cologne, 496 153–168. pp. Emerson, S.B. (1982) Frog postcranial morphology: Glaw, F., Hoegg, S. & Vences, M. (2006) Discovery identification of a functional complex. Copeia, of a new basal relict lineage of Madagascan 1982, 603–613. frogs and its implications for mantellid evolu- Emerson, S.B. (1983) Functional analysis of frog tion. Zootaxa, 1334, 27–43. pectoral girdles.The epicoracoid cartilages. J. Glaw, F., Hawlitschek, O., Glaw, K. & Vences, M. Zool., 201, 293–308. (2019) Integrative evidence confirms new en- Emerson, S.B. (1984). Morphological variation in demic island frogs and transmarine dispersal of frog pectoral girdles: testing alternatives to a amphibians between Madagascar and Mayotte traditional adaptive explanation. Evolution, 38, (Comoros archipelago). The Science of Nature,

376–388. 106, 19. Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 67

Griffiths, I. (1959) The phylogeny of Sminthillus Maglia, A.M., Pugener, L.A. & Mueller, J.M. (2007) limbatus and the status of the Brachycephali- Skeletal morphology and postmetamorphic dae (Amphibia Salientia). J. Zool., 132, 457–487. ontogeny of Acris crepitans (Anura: Hylidae): Guibé, J. (1978) Les Batraciens de Madagascar. Bon- a case of miniaturization in frogs. J. Morphol., ner Zoologische Monographien, 11, 1–144. 268, 194–223. Hall, B.K. (2005) Bones and Cartilage: Developmen- Manzano, A.S., Fabrezi, M. & Vences, M. (2007) tal and Evolutionary Skeletal Biology. Elsevier/ Intercalary elements, treefrogs, and the early Academic Press, London, 792 pp. differentiation of a complex system in the Neo- Hall, B.K. (2007) Fins into Limbs: Evolution, Deve­ batrachia. Anat. Rec., 290, 1551–1567. lopment, and Transformation. The University of Marjanovic, D. & Laurin, M. (2019) Phylogeny of Pa- Chicago Press, Chicago, 433 pp. leozoic limbed vertebrates reassessed through Hanken, J. (1993) Adaptation of bone growth to revision and expansion of the largest published miniaturization of body size. In: B.K. Hall (Ed) relevant data matrix. PeerJ, 6, e5565. Bone. Vol. 7: Bone Growth—B, pp. 79–104. CRC Mason, M.J., Segenhout, J.M., Cobo-Cuan, A., Qui- Press, Boca Raton, FL. ñones, P.M. & van Dijk, P. (2015) The frog inner Hanna, G. & Barnes, J.P. (1991) Adhesion and de- ear: picture perfect? J. Assoc. Res. Otolaryngol., tachment of the toe pads of tree frogs. J. Exp. 16, 171–188. Biol., 155, 103–125. Mookerjee, H.K. (1931) IV. On the development of Howes, G.B. (1888) On the carpus and tarsus of the the vertebral column of Anura. Phil. Trans. R. Anura. J. Zool., 56, 141–182. Soc. B, 219, 165–196. Hoyos, J.M. (2003) Additions to our knowledge of Mookerjee, H.K. & Das, S.K. (1939) Further inves- anuran sesamoids. Herpetol. Rev., 34, 112. tigation on the development of the vertebral Kamermans, M. & Vences, M. (2009) Terminal column in Salientia (Anura). J. Morphol., 64, phalanges in ranoid frogs: morphological diver- 167–209. sity and evolutionary correlation with climbing Nauwelaerts, S., Ramsay, J. & Aerts, P. (2007) Mor- habits. Alytes, 26, 117–152. phological correlates of aquatic and terrestrial Kluge, A.G. & Farris, J.S. (1969) Quantitative phy- locomotion in a semi-aquatic frog, Rana escu- letics and the evolution of anurans. Syst. Biol., lenta: no evidence for a design conflict. J. Anat., 18, 1–32. 210, 304–317. Laloy, F., Rage, J.-C., Evans, S.E., Boistel, R., Lenoir, Nowack, C. & Vences, M. (2016) Ontogenetic deve­ N. & Laurin, M. (2013) A re-interpretation of the lopment of the derived olfactory system of the Eocene anuran Thaumastosaurus based on mi- mantellid frog Mantidactylus betsileanus. Anat. croCT examination of a ‘mummified’ specimen. Rec., 299, 943–950. PLOS ONE, 8, e74874. Nowack, C., Peram, P.S., Wenzel, S., Rakotoarison, Laurent, R. (1946) Mises au point dans la taxono- A., Glaw, F., Poth, D., Schulz, S. & Vences, M. mie des ranidés. Revue de Zoologie et de Bota- (2017) Volatile compound secretion coincides nique Africaines, 39, 336–338. with modifications of the olfactory organ in Li, Y., Xu, F., Guo, Z., Liu, X., Jin, C., Wang, Y. & mantellid frogs. J. Zool., 303, 72–81. Wang, S. (2011) Reduced predator species rich- Nussbaum, R.A. (1982) Heterotopic bones in the ness drives the body gigantism of a frog species hind limbs of frogs of the families Pipidae, Ra- on the Zhoushan Archipelago in China. J. Anim. nidae and Sooglossidae. Herpetologica, 1982, Ecol., 80, 171–182. 312–320. Lomolino, M.V. (2005) Body size evolution in insu- Olson, W. (2000) Phylogeny, ontogeny, and func- lar vertebrates: generality of the island rule. J. tion: extraskeletal bones in the tendons and

Biogeogr., 32, 1683–1699. joints of HymenochirusDownloaded from boettgeri Brill.com10/04/2021 (Amphibia: 09:02:24AM via free access

204346 68 Santos-Santos et al.

Anura: Pipidae). Zool.-Anal. Complex Syst., 103, recent phylogenetic divergence using novel 15–24. non-destructive morphometric methods. Ph.D. Pabijan, M., Gehring, P.-S., Koehler, J., Glaw, F. dissertation, University of Barcelona, Barce- & Vences, M. (2011) A new microendemic frog lona, 320 pp. species of the genus Blommersia (Anura: Man- Samonds, K.E., Godfrey, L.R., Ali, J.R., Goodman, tellidae) from the east coast of Madagascar. S.M., Vences, M., Sutherland, M.R., Irwin, M.T. & Zootaxa, 2978, 34–50. Krause, D.W. (2013) Imperfect isolation: factors Pabijan, M., Wollenberg, K. & Vences, M. (2012) and filters shaping Madagascar’s extant verte- Small body size increases the regional differ- brate fauna. PLOS One, 8, e62086. entiation of populations of tropical mantel- Scherz, M.D., Vences, M., Borrell, J., Ball, L., Nomen- lid frogs (Anura: Mantellidae). J. Evol. Biol., 25, janahary, D.H., Parker, D., Rakotondratsima, M., 2310–2324. Razafimandimby, E., Starnes, T., Rabearivony, J. Parsons, F. (1904) Observations on traction epiphy- & Glaw, F. (2017a) A new frog species of the sub- ses. J. Anat. Physiol., 38, 248. genus Asperomantis (Anura, Mante­llidae, Ge- Pérez-Ben, C.M., Schoch, R.R. & Báez, A.M. (2018) phyromantis) from the Bealanana District of Miniaturization and morphological evolution northern Madagascar. Zoosyst. Evol., 93, 451. in Paleozoic relatives of living amphibians: a Scherz, M.D., Razafindraibe, J.H., Rakotoarison, quantitative approach. Paleobiology, 44, 58–75. A., Dixit, N.M., Bletz, M.C., Glaw, F. & Vences, Ponssa, M.L., Goldberg, J. & Abdala, V. (2010) Sesa- M. (2017b) Yet another small brown frog from moids in anurans: new data, old issues. Anat. high altitude on the Marojejy Massif, northeast- Rec., 293, 1646–1668. ern Madagascar (Anura: Mantellidae). Zootaxa, Reilly, S.M. & Jorgensen, M.E. (2011) The evolution 4347, 572–582. of jumping in frogs: morphological evidence for Scherz, M.D., Hawlitschek, O., Andreone, F., Ra- the basal anuran locomotor condition and the kotoarison, A., Vences, M. & Glaw, F. (2017c) A radiation of locomotor systems in crown group review of the taxonomy and osteology of the anurans. J. Morphol., 272, 149–168. Rhombophryne serratopalpebrosa species group Reynolds, S.H. (1897) The Vertebrate Skeleton. Cam- (Anura: Microhylidae) from Madagascar, with bridge University Press Warehouse, London, comments on the value of volume rendering of 559 pp. micro-CT data to taxonomists. Zootaxa, 4273, Richards, C.L. (2006) Has the evolution of com- 301–340. plexity in the amphibian papilla influenced Scott, E. (2005). A phylogeny of ranid frogs (Anura: anuran speciation rates? J. Evolution Biol., 19, Ranoidea: Ranidae), based on a simultaneous 1222–1230. analysis of morphological and molecular data. Roček, Z. (1994) Taxonomy and distribution of Ter- Cladistics, 21, 507–574. tiary discoglossids (Anura) of the genus Latonia Shubin, N.H. & Alberch, P. (1986) A morphogenetic (v. Meyer, 1843). Geobios, 27, 717–751. approach to the origin and basic organization Roček, Z. (2005) Late Miocene amphibia from Ru­ of the tetrapod limb. In: M.K. Hecht, B. Wallace dabánya. Palaeontographia Italica, 90, 11–29. & G.T. Prance (Eds) Evolutionary Biology, pp. Ročková, H. & Roček, Z. (2005) Development of 319–387. Springer, Boston. the pelvis and posterior part of the vertebral Simmons, D., Bertolotto, C. & Narins, P. (1992) In- column in the Anura. J. Anat., 206, 17–35. nervation of the amphibian and basilar papil- Ruibal, R. & Shoemaker, V. (1984) Osteoderms in lae in the leopard frog: reconstructions of single anurans. J. Herpetol., 1984, 313–328. labeled fibers. J. Comp. Neurol., 322, 191–200. Santos-Santos, J.H. (2019) Ecomorphological dis- Soliz, M.C., Tulli, M.J. & Abdala, V. (2017) Osteologi-

crimination of vertebrate sister species with cal postcranial traitsDownloaded in hylid from anurans Brill.com10/04/2021 indicate 09:02:24AM a via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 69

morphological continuum between swimming Vences, M., Pabijan, M., Kohler, J. & Glaw, F. (2010) and jumping locomotor modes. J. Morphol., 278, Two syntopic and microendemic new frogs of 403–417. the genus Blommersia from the East coast of Soliz, M.C., Ponssa, M.L. & Abdala, V. (2018) Com- Madagascar. Afr. J. Herpetol., 59, 133–156. parative anatomy and development of pectoral Vences, M., Jovanovic, O., Safarek, G., Glaw, F. & and pelvic girdles in hylid anurans. J. Morphol., Koehler, J. (2015) A new arboreal frog of the ge- 279, 904–924. nus Guibemantis from the Southeast of Mada- Špinar, Z.V. (1972) Tertiary Frogs From Central Eu- gascar (Anura: Mantellidae). Zootaxa, 4059, rope. Academia, Prague, 286 pp. 569–580. Summers, A.P. & Koob, T.J. (2002) The evolution of Vieites, D.R., Wollenberg, K.C., Andreone, F., tendon—morphology and material properties. Köhler, J., Glaw, F. & Vences, M. (2009) Vast un- Comp. Biochem. Phys. A, 133, 1159–1170. derestimation of Madagascar’s biodiversity evi- Trueb, L. (1968a) Variation in the tree frog Hyla lan- denced by an integrative amphibian inventory. casteri. Copeia, 1968, 285–299. Proc. Natl. Acad. Sci. U.S.A., 106, 8267–8272. Trueb, L. (1968b) Cranial osteology of the hylid Wake, D. (1970) Aspects of vertebral evolution in frog, Smilisca baudini. University of Kansas Pub­ the modern Amphibia. Forma Functio, 3, 33–60. li­ca­tions, Museum of Natural History, 18, 11–35. Wollenberg, K.C., Vieites, D.R., Glaw, F. & Vences, Vences, M. & Glaw, F. (2001) When molecules claim M. (2011) Speciation in little: the role of range for taxonomic change: new proposals on the and body size in the diversification of Malagasy classification of Old World treefrogs. Spixiana, mantellid frogs. BMC Evol. Biol., 11, 217. 24, 85–92. Yeh, J. (2002). The effect of miniaturized body size Vences, M., Andreone, F., Glaw, F. & Mattioli, F. on skeletal morphology in frogs. Evolution, 56, (2002) New dwarf species of Mantidactylus 628–641. from northwestern Madagascar (Anura: Man- Zug, G.R. (1978) Anuran Locomotion—Structure tellidae). Copeia, 2002, 1057–1062. and Function, 2: Jumping Performance of Semi- Vences, M., Vieites, D.R., Glaw, F., Brinkmann, H., aquatic, Terrestrial, and Arboreal Frogs. Smith- Kosuch, J., Veith, M. & Meyer, A. (2003) Multiple sonian Contributions to Zoology, 276. Smithso- overseas dispersal in amphibians. Proc. R. Soc. nian Institution Press, Washington. B-Biol. Sci., 270, 2435–2442. Vences, M., Wahl-Boos, G., Hoegg, S., Glaw, F., received: 16 April 2019 | revised and accepted: Oliveira, E.S., Meyer, A. & Perry, S. (2007) Mo- 20 AuguSt 2019 lecular systematics of mantelline frogs from editor: J.W. arntzen Madagascar and the evolution of their femoral glands. Biol. J. Linn. Soc., 92, 529–539.

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 70 Santos-Santos et al.

Appendix 1 Premaxilla width (pmaxW): in the ventral face of the coronal plane; distance between the medial Descriptions of measurements. All measurements and lateral tips of the premaxilla’s pars palatina. were performed on the left side or paired element of the skeleton unless otherwise specified. (Para) Premaxilla depth (pmaxD): in the ventral face of sagittal planes may correspond to translations of the coronal plane; distance from the premaxilla’s the median plane to clip irrelevant structures for a anterior medial margin to the caudal (medial) tip given measurement. of its pars palatina.

Snout–vent length external (SVLext): snout– Mentomeckelian bone length (MmbL): in the vent length measured externally on the preserved ventral face of the coronal plane; distance along specimen. the mentomeckelian bone’s midline from its me- dial margin to its lateral margin. Snout–vent length skeletal (SVLsk): in the right face of the sagittal plane; distance from the rostral Upper jaw length (maxL): distance along the up- tip of the premaxilla to the caudal-most margin of per jaw’s surface from the anteromedial margin the pelvis. of the premaxilla to the caudal-most tip of the quadratojugal. Summed Snout–vent length (SVLsum): The sum of skull length (skullL), column length (column), Lower jaw length (mandL): distance along the and pelvis length (pelvL). lower jaw’s surface from the anteromedial margin of the mentomeckelian bone to the caudal-most Skull length (skullL): in the ventral face of the margin of the angulosplenial plate. coronal plane; linear distance from the anterior margin of the left premaxilla to the caudal border Nasal length (nasL): in the left face of the sagittal of the left exoccipital. plane; maximum distance along the nasal’s lateral border from its cranial tip to the caudal termina- Skull width (skullW): in the ventral face of the tion of its maxillary process. coronal plane; maximum distance between the left and right jaw articulations. Nasal width (nasW): in the dorsal face of the coro- nal plane; transverse distance from the left nasal’s Otic to Otic capsule distance (oticsW): in the most-medial articulation point with the internasal ventral face of the coronal plane; maximum trans- septum to its lateral margin. verse distance between the lateral margins of the left and right otic capsules. Inter-nasal width (intnasW): in the dorsal face of the coronal plane; transverse distance between the Columella (colmL): distance along the columella’s medial calcified articulation points of the left and midline from its lateral tip to its medial articula- right nasals with the tectum nasi. tion with the oval window. Internasal septum height (intnsptH): in the ven- Premaxilla height (pmaxH): in the left face of the tral face of the coronal plane; distance from the sagittal plane; distance from the premaxilla’s an- left cranial margin of the internasal septum to teroventral margin (i.e., pars dentalis) to the dorsal the caudal tip of the left vomer (i.e., left olfactory tip of its alary process. foramen).

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 71

Sphenethmoid height (sphH): in the ventral face Vomer width (vomW): in the ventral face of the of the coronal plane; distance from the left vomer’s coronal plane; length of the anterolateral margin caudal tip (i.e., left olfactory foramen) to the caudal- of the vomer’s anterior portion. most margin of the sphenethmoid’s left lateral ala. Palatine length (pltnL): distance along the pala- Sphenethmoid width (sphW): in the ventral face tine’s caudal margin from its medial articulation of the coronal plane; distance from the caudal- with the alary cartilage to its lateral articulation most margin of the sphenethmoid’s left lateral ala with the maxilla. to the caudal-most margin of its right ala. Pterygoid height (ptygdH): in the right face of Frontoparietal length (frptlL): in the dorsal face the sagittal plane; distance from the tip of the left of the coronal plane; distance from the posterolat- pterygoid’s dorsoposterior ramus to the pterygoid’s eral margin of the tectum nasi to the caudal tip of ventral margin along its ventral border. the posterior frontoparietal protuberance. Pterygoid width (ptygdW): in the right face of Frontoparietal width (frptlW): in the dorsal face the sagittal plane; distance along the left ptery- of the coronal plane; maximum distance between goid’s surface from its anterior-most articulation the left frontoparietal’s medial and lateral margins. with the maxilla to its caudal-most tip at the jaw articulation. Prootic height (proH): in the dorsal face of the coronal plane; distance from the anterior border of Squamosal height (squaH): in the left face of the the anterior epiotic eminence to the caudal-most sagittal plane; distance along the squamosal’s zy- margin of the prootic at the prominentia ducti gomatic ramus midline from its ventrocaudal tip semicircularis posterioris. to the dorsal border of the squamosal’s otic ramus

Exoccipital length (exoL): in the ventral face of Squamosal width (squaW): in the left face of the the coronal plane; transverse distance from the sagittal plane; distance along the squamosal’s caudolateral border of the left exoccipital to that otic ramus midline from its anterior to posterior of the right exoccipital. margins.

Parasphenoid height (psphH): in the ventral face Hyoid bone length (hyoL): distance along the left of the coronal plane; distance from the lateral in- hyoid’s midline from its cranial tip to its caudal tersection of the parasphenoid with the left ala end. of the sphenethmoid to the cranial margin of the origin of the parasphenoid’s left posterolateral ala. Omosternum superior branch height (omostH): in the ventral face of the coronal plane; distance Parasphenoid width (psphW): in the ventral face along the omosternum’s midline from its cra- of the coronal plane; transverse distance from the nial edge to its caudal border at its posterior anterior, lateral-most corner of the parasphenoid’s bifurcation. left posterior ala to that of its right. Omosternum branch length (omostL): in the Vomer height (vomH): in the ventral face of the in the ventral face of the coronal plane; distance coronal plane; distance from the caudal tip of the along the left omosternum branch’s midline from vomer’s dentigerous process to its branching point its articulation with the clavicular process to the with the postchoanal portion along its lateral edge. lateral border of theDownloaded opposing from branch’s Brill.com10/04/2021 bifurcation. 09:02:24AM via free access

204346 72 Santos-Santos et al.

Clavicle length (clavL): in the ventral face of the Suprascapula height (supraH): in the dorsal face coronal plane; distance along the clavicle’s mid- of the coronal plane; anteroposterior distance line from its proximal articulation with the epi- along the articulation with the scapula. coracoid cartilage to its distal articulation with the scapula. Suprascapular caudal-medial projection (supraW): in the dorsal face of the coronal plane; Distance to the clavicular process (clav_pL): in distance from the posterolateral articulation with the ventral face of the coronal plane; distance from the scapula to the most medial point of the supra- the medial origin of the clavicle to the clavicular scapula’s ossification. process. Humerus length capitulum (humLcap): linear Coracoid length (coraL): in the ventral face of the distance from the proximal end of the humerus coronal plane; distance along the coracoid’s mid- head to the distal capitulum of the humerus (i.e., line from its proximal articulation with the epi- part of the epiphysis that articulates with the coracoid cartilage to its distal articulation with the radius). humerus head in the glenoid cavity. Humerus length trochlea (humLtroc): linear dis- Sternum height (sterH): in the ventral face of the tance from the proximal end of the humerus head coronal plane; distance along the sternum’s mid- to the distal trochlea of the humerus (i.e., part of line from its cranial border to its caudal edge. the epiphysis that articulates with the ulna).

Sternum superior width (sterWsup): in the ven- Deltopectoral crest height (dltd_crH): in the ven- tral face of the coronal plane; maximum trans- tral face of the coronal plane; maximum height of verse distance between the sternum’s anterior left the deltopectoral crest from its anterior edge to its and right lateral edges. posterior base.

Sternum minimum width (sterWmin): in the ven- Deltopectoral crest width (dltd_crW): in the ven- tral face of the coronal plane; minimum transverse tral face of the coronal plane; distance from the distance between the sternum’s left and right lat- proximal end of the base of the deltopectoral crest eral edges. where it touches the humeral head to its distal end on the humerus. Sternum inferior width (sterWinf): in the ventral face of the coronal plane; maximum transverse Radioulna length (radulnL): distance from the distance between the sternum’s posterior left and proximal end of the olecranon (articulation with right lateral edges. the humerus) to the distal end of the ulna (articu- lation with the carpals). Scapula length (scapL): distance from the scapu- la’s ventral point of articulation with the humerus Metacarpal lengths (mtcplX): distance along in the glenoid cavity to its caudal edge that articu- the metacarpal’s midline from its proximal end lates with the suprascapula. ­(articulation with the carpals) to its distal end (ar- ticulation with the phalanges). Cleithrum length (cleiL): in the dorsal face of the coronal plane; distance along the cranial edge of Phalanx lengths (hand_phXL/foot_phXL): dis- the suprascapula from its proximal border to its tance along the phalanx’s midline from its proxi- distal articulation with the scapula. mal end to its distal end.Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Descriptive skeletal anatomy of Blommersia transmarina 73

Terminal phalanx lengths (hand_tphL/foot_ Ilial shaft length (iliL): in the left face of the tphL): distance from the terminal phalanx’s left sagittal plane; distance from the left ilial shaft’s proximal end to its left projection on its distal end. ­dorsocranial tip at the ilial–sacral articulation to the caudal edge of the ilial protuberance immedi- Digit lengths (hand_DX/foot_DX): sum of the ately before the acetabulum. metacarpal/metatarsal length (mtcpl/mttrsl), pha- lanx lengths (hand_phXL/foot_phXL), and termi- Pelvis length (pelvL): in the left face of the sagittal nal phalanx length (hand_tphL/foot_tphL) for the plane; maximum distance from the dorsoanterior same limb digit. tip of the ilial shaft to the caudalmost border of the pelvic wheel at the ischium. Column (column): sum of all vertebral body heights. Pelvic wheel length (pelwL): in the left face of the sagittal plane; maximum linear distance between Vertebral body heights (atlasH/VXH/sacrumH): the anterior and posterior edges of the pelvis. in the ventral face of the coronal plane; distance along the midline of the vertebra’s centrum from Epipubis height (epipH): in the ventral face of the its cranial border to its caudal margin. coronal plane; distance along the epipubis’ mid- line from its anterior edge to its posterior edge. Distances between vertebral transverse pro- cesses (VX_tpL/sacrum_tpL): in the ventral face Epipubis width (epipW): in the ventral face of the of the coronal plane; distance from the tip of the coronal plane; distance from epipubis’ left lateral left transverse process to the tip of that of the right edge to its right lateral border. along their midline. Femur length (femL): distance along the femur’s Urostyle length (uroL): in the left face of the sag- midline from its proximal end (articulation with ittal plane; distance along the urostyle’s midline the acetabulum) to its distal end (articulation with from its articulation with the sacral vertebra to its the tibiofibula). caudal tip. Tibiofibula length (tibfibL): distance along the Urostyle height (uroH): in the left face of the sag- tibiofibula’s midline from its proximal end (articu- ittal plane; distance from the urostyle’s dorsal bor- lation with the femur) to its distal end (articula- der to its ventral border at its cranial tip. tion with the tibiale and fibulare).

Urostyle crest length (uro_crL): in the left face of Tibiale and Fibulare length (tiblfiblL): distance the sagittal plane; distance from the base of the along the fibulare’s midline from its most proximal cranial origin of the urostyle crest to its posterior end (articulation with the tibiofibula) to its most end. distal end (articulation with the metatarsals).

Urostyle crest height (uro_crH): in the left face Metatarsal lengths (mttrslX): distance along the of the sagittal plane; maximum distance from metatarsal’s midline from its proximal end (articu- the dorsal edge of the urostyle crest to its ventral lation with the tarsals) to its distal end (articula- edge. tion with the phalanges).

Downloaded from Brill.com10/04/2021 09:02:24AM via free access

204346 Contributions to Zoology 89 (2020) 245 CTOZ

brill.com/ctoz

Erratum

Santos-Santos, J.H., Guinovart-Castán, M. & Vieites, D.R. (2020) Descriptive skeletal anatomy of Blommersia transmarina (Amphibia: Anura: Mantellidae) from the Comoro Islands. Contrib. Zool. 89, 14–73. doi:10.1163/18759866-20191405

Unfortunately, the image of Figure 5 on p. 21 of this article contains the incorrect image. The correct Figure 5 should look as below. The figure caption was also slightly corrected to increase its readability.

Figure 5 (A) Right lateral view of the Blommersia transmarina otic capsule. The cranium has been clipped laterally, posteriorly, and dorsally to facilitate visualization of the interior of the otic capsule. The anterior portion of the skull is clipped at the articulation of the palatine and alary cartilage. (B) Corresponding left sagittal plane, medial view of the B. transmarina otic capsule. The volume has been clipped to facilitate visualization of the interior of the otic capsule [MNCN50446 male; volrenWhite (Left) and physics colormaps (Right), grey [density] values: 12000–55500].

We would like to apologise for any inconvenience caused. UNCORRECTED PROOF

© SANTOS-SANTOS et al., 2019 | doi:10.1163/18759866-08902001 This is an open access article distributed under the terms of the cc-by 4.0 License. Downloaded from Brill.com10/04/2021 09:02:24AM via free access