The freshwater turtle fauna of South America is dominated by species of the family Chelidae. Even though there are 21 described species, the phylogeny of the chelids is poorly understood and in Brazil, chelid taxonomy is rife with inconsistencies in species descriptions and identifications. Recently, Cunha et al. (2019) discussed why Mesoclemmys heliostemma should not be considered a valid species because its specimens are a color variant of Mesoclemmys raniceps; both color morphs may hatch from a single clutch of eggs. The genus Mesoclemmys has ≥ 10 described species, widely distributed in northern South America (Brazil, Bolivia, Colombia, Ecuador, French Guiana, Guyana, Peru, Suriname, Trinidad, and Venezuela [Turtle Taxonomy Working Group 2017]). Species of Mesoclemmys inhabit the Amazonia, Cerrado, Pantanal, Caatinga, and Atlantic Forest biomes in Brazil. The diversity of chelonian species is linked to the paleogeographic history of the environment (Pereira et al. 2017), as well as phylogeographic differences, and preference for different habitats reflects on the genetic diversity (Vargas-Ramírez et al. 2012). Moreover, with a vast heterogeneity in the ecosystems and biomes of the South American continent, there are numerous localities where we still do not fully understand the diversity of chelonians or their abundance or distribution (Brito et al. 2019a, 2019b; Vargas-Ramírez et al. 2020), making the expansion of taxonomy and diversity of the group a distinct possibility, via either the description of new species or insight into phylogenetic rates of diversification.

Chelid turtles of the Amazonian region have a great diversity in Brazil, with the genus Chelus having 2 species, one of which was recently described (Chelus fimbriata and Chelus orinocensis), 4 toad-headed turtles (M. raniceps, M. wermuthi, M. nasuta, and M. gibba), 2 side-necked turtles (Phrynops geoffroanus and P. tuberosus), 1 twist-necked turtle (Platemys platycephala), and 1 red-headed side-necked turtle (Rhinemys rufipes). Both classical approaches, with morphological evidence, as well as modern approaches, with molecular evidence, are essential for the advancement of a stable taxonomy of living turtles (Holley et al. 2019).

In a global scenario with significant losses of biodiversity (Stanford et al. 2020), describing a new species allows us to better understand the fauna of turtles and advance our knowledge of their phylogenetic relationships and ecology. This fact looms large when it comes to a region like the Amazon, considered a hotspot for turtles (Mittermeier et al. 2015; Ennen et al. 2020), because of great diversity and high endemism. Accelerating changes in the natural landscape of the Amazon, mainly due to anthropogenic issues, are having a great impact on turtles on both local and global scales.

Herein we describe a new species, bringing together phenotypic differences and genetic data, and discuss the geographic range and basic ecology of the new species. This is the first time that a taxonomic description of a species for the genus Mesoclemmys is based on strong and robust scientific evidence of both morphological variables and molecular analyses, with morphometric analyses for sympatric species and a determination of phylogenetic relationships for 18 species of the family. We also compare morphometric measurements of adults and juveniles of sympatric congeners and provide information and images of the nonossified region of the plastron, present in the new species and previously reported only in Mesoclemmys hogei. Both M. hogei and our new species do not have the plastron ossified to the carapace, rather there is a cartilaginous connection allowing for kinetic shell expansion during oviposition. We have not observed this condition in other species of Mesoclemmys. It is important to mention that a thorough taxonomic review is needed, including all species of the group, using classic and modern techniques to clarify the relationships of all extant chelids.

METHODS

Morphology. — We recorded 84 standard quantitative morphological characters and 30 qualitative characters on specimens of five Mesoclemmys species, with measurements described and specimens listed in Appendix 1. We used analog calipers (Marberg® for taking smaller measurements and Haglof® for taking larger measurements) to the nearest 0.1 mm, reported as means ± 1 standard deviation (SD). Measurements were taken for carapace length (CL), plastron length (PL), carapace height (CH), carapace width (CW), plastron width (PW), gular scute length left (GSLL), gular scute width left (GSWL), intergular scute length (ISL), intergular scute width (ISW), gular scute length right (GSLR), gular scute width right (GSWR), humeral scute length (HSL), humeral scute width (HSW), pectoral scute length (PSL), pectoral scute width (PSW), abdominal scute length (ASL), abdominal scute width (ASW), femoral scute length (FSL), femoral scute width (FSW), anal scute length (AnSL), anal scute width (AnSW), head length (HL), head height (HH), and head width (HW).

To test specimens of the 5 species of Mesoclemmys for differences in morphometric data, we performed 1-way analysis of variance (ANOVA) tests on the mathematical ratios of CW, PL, CH, HW, PSL, and ASL to CL, all in millimeters as straight-line lengths. Tests were first performed demonstrating that the data were parametric and had equal variances (Levene test). The sample sizes were not the same for the five species, so we used Scheffé comparisons as a post hoc test to identify significant interspecific differences.

To compare the mathematical ratios CW/CL and HW/ CL, we used 38 specimens of 4 Mesoclemmys species (M. gibba, M. raniceps, M. perplexa, and M. vanderhaegei) with specimens of our new species in a linear regression. All statistical relationships and analyses were performed according to Zar (2014), with Statistica® 7.0 Software. We recorded the plastral formula adapted from Lovich and Ernst (1989). We compared the characteristics of the new turtle species with those of related Mesoclemmys species.

All specimens were measured by the first author. Five live turtles were collected in the presence of the first author and 3 live animals were collected by nonscientific collectors; all were collected during the rainy season (March and May 2018; January and March 2019). Specimens were collected manually with the use of portable traps. All animals were photographed individually. All 8 specimens analyzed were similar with regard to their morphological characteristics and color patterns.

Four of the 8 specimens are adults, defined by the lack of growth rings on the carapace or plastron scutes and the presence of secondary sexual dimorphism characters, such as tail size and thickness. Dissection of the turtles showed mature ovaries and testes.

Radiographic images were produced on Compact 500 VMI equipment, with 100 mA, 42 kV light intensity, 10 sec of exposure time, without contrast, in live animals. Images of 2 adults and 1 juvenile were produced on 6 February 2020.

Genetic Sampling, Extraction, Amplification, and Sequencing of DNA. — We collected 7 samples from blood tissue preserved in absolute EtOH. We obtained total DNA with Promega's Wizard Genomic kit, according to the manufacturer's protocol. We amplified fragments of the COI and 16S gene by polymerase chain reaction (PCR). For the PCRs, a final volume of 15 µL was used, containing about 30 ng of genomic DNA, 2.4 µL of dNTPs (1.25 mM), 1.5 µL of 10× Buffer (200 mM Tris-HCl, 500 mM KCl), 1 µL of MgCl₂ (25 mM), 1 µL of each primer (0.2 µM), and 1 U of Taq DNA polymerase. The amplification protocol was initiated with 4 min of denaturation at 95°C, followed by 35 cycles of 3 stages: 1) denaturation at 95°C for 30 sec; 2) annealing at a specific temperature of 47°C to COI and 53°C to 16S; and 3) extension at 72°C for 60 sec. After completion of the 35 cycles, there was a final extension stage at 72°C for 10 min. We then purified the PCR products using polyethylene glycol and ethanol (Paithankar and Prasad 1991). The sequencing reactions were run using the BigDye Terminator Sequencing kit v. 3.1 (Life Technologies) and the reaction products were separated and visualized using an ABI 3500xl automatic sequencer (Life Technologies).

Sequence Alignment, Phylogenetic Analyses, and Genetic Distances. — We incorporated into our database all species of Mesoclemmys available on GenBank, as well as species of genera closely related to Mesoclemmys. The DNA sequences were aligned using the online program Mafft (Katoh et al. 2019). We used the software PartitionFinder (Kimura 1980; Lanfear et al. 2012) to test different partition schemes and select the most appropriate evolutionary model (Table 1). For Partition Finder analyses, we allowed linked branch lengths and used the Bayesian information criterion. Our analysis suggested that the best scheme for our data set was to separate it into 4 partitions. Phylogenetic reconstruction was performed in W-IQ-TREE online (Trifinopoulos et al. 2016), considering the scheme and model estimated in PartitionFinder. In MEGA (Kumar et al. 2018), we estimated the genetic distances between the taxa. Additionally, we used Automatic Barcode Gap Discovery (ABGD; Puillandre et al. 2012) to test the possibility of a new species among the analyzed samples.

Table 1. Evolutionary model, including the partitions, evolutionary models, and numbers of base pairs.

View Fullscreen

RESULTS

Morphological Data. — There was a significant difference in the CW/CL ratio among the four described species studied (abbreviated below by the first three letters of their specific epithets) and the new species (abbreviated sp. nov.; ANOVA F_4,37 = 98.66, p < 0.01). Scheffé's comparisons showed that per ≠ ran, per ≠ sp. nov., van ≠ ran, van ≠ sp. nov., van ≠ gib, and van ≠ per. A similar result was found for the PL/CL ratio (ANOVA F_4,37 = 7.55, p < 0.01; gib ≠ ran, gib ≠ sp. nov., and per ≠ gib). Significant interspecific differences were also found for CH/CL (ANOVA F_4,37 = 6.79, p < 0.01; sp. nov. ≠ gib, ran ≠ per, gib ≠ per, and van ≠ per); HW/ CL (ANOVA F_4,37 = 16.78, p < 0.01; gib ≠ sp. nov, per ≠ gib, van ≠ sp. nov, and van ≠ gib); PSL/CL (ANOVA F_4,37 = 4.94, p < 0.01; van ≠ sp. nov.); and ASL/CL (ANOVA F_4,37 = 9.26, p < 0.01; gib ≠ ran and gib ≠ sp. nov.). There also was a difference in the relationship of the ratios CW/CL and HW/CL among M. raniceps, M. gibba, M. perplexa, M. vanderhaegei, and M. sp. nov. (ANOVA F_4,37 = 20.26, p < 0.01; Fig. 1).

View Figure

• Download as Powerpoint
• Download Figure

Regarding the anatomy of the new species, we produced radiographs for adult animals (male and female) and a hatchling. The new species has an unossified cartilaginous region in the region of the bridge that connects the carapace and the plastron (Fig. 2). Ours is the first report of this anatomical condition for Amazon chelids.

View Figure

• Download as Powerpoint
• Download Figure

Genetic Data. — Our results produced 2 databases that have 640 base-pairs (bp) for the COI gene and 549 bp for 16S. These fragments are homologous to the sequences from GenBank that we used in our analyses. With data for COI analyzed separately because of their distinct rate of evolution, Mesoclemmys sp. nov. and M. raniceps were separated by 1.5% and both were separated by > 12.0% from other Mesoclemmys (Mesoclemmys hogei, M. vanderhaegei, and M. tuberculata; Table 2). Considering 16S gene sequences (Table 3), the new species was genetically different from M. nasuta by 2.4% and by > 4.5% from other species of the family (Phrynops hilarii, Phrynops williamsi, and Platemys platycephala).

Table 2. Matrix of percent genetic distances based on COI sequences among Mesoclemmys jurutiensis and other species of the genus and family with sequences deposited and available in GenBank.

View Fullscreen

Table 3. Matrix of percent genetic distances based on 16S ribosomal sequences among Mesoclemmys jurutiensis and other species of the genus and family with sequences deposited and available in GenBank.

View Fullscreen

Our phylogenetic analysis of maximum likelihood grouped the individuals analyzed in a monophyletic clade with M. raniceps that had 93% bootstrap support (Fig. 3). Species of the genus Mesoclemmys were not all recovered as a monophyletic group, as M. tuberculata and M. vanderhaegei were grouped in a clade with Acanthocelys radiolata and P. platycephala, albeit with low support value.

View Figure

• Download as Powerpoint
• Download Figure

SYSTEMATICS

Order: Testudines Batsch, 1788 Suborder: Pleurodira Cope, 1864 Family: Chelidae Gray, 1825 Subfamily: Chelinae Gray, 1825 Genus: MesoclemmysGray, 1873Mesoclemmys jurutiensis sp. nov. (Fig. 4)

• English names: Lesser Amazon Toad-headed Turtle, Black-headed Amazon Turtle; Portuguese name: Perema preta da Amazônia; Indigenous name: Acaçú.

• Holotype. — INPA-H41283 (Instituto Nacional de Pesquisas da Amazônia), whole, liquid-preserved adult male (CL = 179.0 mm; Table 4; Fig. 5a–b); collected from a small pond about 1 m deep, 15 m in diameter, and 50 m in circumference. Interior pond in dense evergreen rainforest, bottom of pond covered with leaves and low emergent aquatic plants (type locality, Fig. 6). Hand-collected with a dip net 26 January 2019 by Fábio A.G. Cunha. Pond is in the region of the Capiranga Community, Juruti-Velho Village, in Juruti municipality, State of Pará, Brazil (WGS84, 02°36′16.20″S, 56°24′40.80″W; 90 m elev. asl).

• Paratypes. — INPA-H41284, whole, liquid-preserved adult female (CL = 173.0 mm); hand-collected with a dip net. Same data as the holotype.

• INPA-H41285, whole, liquid-preserved juvenile, probably female (CL = 117.5 mm; Fig. 5e–f); collected crossing temporary unpaved dirt road, 8 March 2019, Pacoval Community, in Juruti municipality, State of Pará, Brazil (WGS84, 02°30′23.00″S, 56°07′0.79″W; 98 m elev. asl) by Gledson R.P. Machado.

• INPA-H41286, whole, liquid-preserved adult male (CL = 178.0 mm; Fig. 5c–d); hand-collected from mud on an unpaved road by Márcio R.B. Burlamaque on 9 May 2018 in Aveiro municipality, State of Pará, Brazil (WGS 84, 03°20′20.54″S, 55°59′56.70″W; 84 m elev. asl).

• INPA-H41381, whole, liquid-preserved juvenile, probably female (CL = 141.0 mm). Same data as the holotype.

• INPA-H41379, whole, liquid-preserved adult female (CL = 228.0 mm) and INPA-H41380, whole, liquid-preserved juvenile, probably female (CL = 137.8 mm); both collected 9 March 2018 by Fábio A.G. Cunha in a small rainwater-filled pond in the region of Três Bocas Community, in Juruti municipality, State of Pará, Brazil (WGS84, 02°19′41.64″S, 56°5′33.30″W; 73 m elev. asl).

• One hatchling (CL = 85.0 mm); in the live collection of Fábio A.G. Cunha (pit-tag no. 963.007100036948), collected 18 March 2019 crossing temporary mud road, in Capitão Community, in Juruti municipality, State of Pará, Brazil (WGS84, 02°25′43.60″S, 56°22′12.54″W; 59 m elev. asl) by Antônio, a resident of Capitão Community, and donated to Fábio A.G. Cunha, to be deposited upon its death in the Scientific Collection of Reptiles and Amphibians of the Instituto Nacional de Pesquisas da Amazônia.

• Diagnosis. — Mesoclemmys jurutiensis is one of the smaller Amazon members of the genus; only M. gibba is smaller. The largest specimen collected was an adult female (CL = 228.0 mm). Head uniformly black, medium-sized (mean HW 41.2 ± 4.7 mm; HW/CL 21.3%), triangular, flattened dorso-ventrally, widening posteriorly. Eyes surrounded by 2 rows of triangular and hexagonal scales of similar shape but of distinct sizes with flat surface, large, and positioned anteriorly. Tympanum large and superficial. Neck granular dorsally, ventrally with 2 large and distinct barbels prominent on the chin. Rigid oval shell, dark reddish brown in color, with subtle markings on the posterior marginal scutes, slightly flattened with subtle keel in the fourth and fifth vertebral scutes. Nuchal scute distinct. Rigid plastron with vibrant burnt-yellow coloration at the edges and black in the central region; marginal seams black ventrally. Large femoral scutes with rounded outer ends. Anal scutes of the plastron with marked V-shaped invagination. Black front and rear feet with 3 conical large black scales on tibia, the final one most developed. Inguinal region grayish yellow coloration. Emits a musky odor.

• Description of Holotype. — The holotype is an adult male in excellent condition with the following morphometric characters: CL = 179.0 mm; CW = 127.7 mm; CH = 47.0 mm; nuchal scute length = 19.8 mm; first through 12th right marginal scute lengths = 20.0, 19.8, 14.5, 12.4, 12.3, 12.7, 16.6, 19.4, 20.9, 19.7, 19.0, and 20.5 mm; first through fifth vertebral scute lengths = 37.3, 24.8, 24.6, 26.4, and 24.7 mm; length of the first left costal scute = 30.7 mm; length of the first right costal scute = 30.7 mm; first through fourth right costal scute maximum lengths = 39.6, 28.5, 29.9, and 27.8 mm; first through fourth right costal scute minimum lengths = 14.3, 22.9, 25.1, and 15.8 mm; maximum length of the nuchal scute = 5.8 mm; first through fifth vertebral scute maximum anterior widths = 56.6, 36.1, 33.3, 29.5, and 38.5 mm; first through fifth vertebral scute maximum posterior widths = 25.1, 26.8, 21.6, 11.9, and 26.7 mm; first through fourth costal scute posterior widths = 37.2, 39.4, 25.9, and 23.6 mm; PL = 148.1 mm; PW = 104.4 mm; GSLL = 25.0 mm; GSWL = 22.3 mm; ISL = 33.9 mm; ISW maximum = 18.6 mm; ISW minimum = 13.7 mm; GSLR = 25.0 mm; GSWR = 21.7 mm; HSL = 16.8 mm; HSW = 39.1 mm; PSL = 19.0 mm; PSW = 52.6 mm; ASL = 18.4 mm; ASW = 32.3 mm; FSL = 25.5 mm; FSW = 25.0 mm; AnSL = 18.9 mm; AnSW = 20.4 mm; length of the bridge at the marginals seam = 45.0 mm; length of the bridge at the plastron seam = 36.3 mm; plastral formula (longest > shortest): 3-5-7-4-1-6-2 (i.e., humeral > abdominal > anal > pectoral > gular > femoral > intergular); HL = 44.4 mm; HW = 38.3 mm; HH = 21.4 mm; inter orbital distance = 9.2 mm.

• Head medium-sized and flat (HH/HL = 0.48). Head slightly triangular and short, base of the head is proportionally larger, HW/HL = 0.86 and HW/ CL = 0.21. Large eyes oriented dorsally. Snout pointed, with 2 long light yellow barbels attached just posterior to the apices of the jaw. Lateral and dorsal region of the head with flat, smooth texture, without the presence of granules; interocular region with rows of plates with regular shapes of different sizes. Neck 58.4 mm in length (from the posterior border of the skull to the base of the carapace), with numerous small black granular scales. Tail with 11 white granular scales and 9 black uniformly spaced scales. White wart-like scales on the otherwise dark gray skin. Carapace oval with a subtle keel in the region of the fourth and fifth vertebral scutes, without a groove in the carapace. Carapace has a slight upturning of the marginal scutes posteriorly. A radiograph showed that the bridge connecting the carapace to the plastron is not ossified.

• Coloration in Life. — Uniformly dark shell with reddish dark brown tones. Central region of the plastron dark (shades of dark brown to black), with the outer edges of the left and right gular shields, intergular, and anal shields with burnt-yellow coloration. Ventral aspect of the burnt-yellow and darker marginal scutes with black seams. Totally black head, with lighter mandibular region and black nostrils. Predominantly black neck with grayish inner region, becoming lighter at the base of the attachment to the carapace. Two yellow barbels below the chin. Black front and rear feet, becoming lighter with shades of gray in the femoral region, becoming completely light (yellowish white) in the inguinal region.

• Etymology. — The specific epithet jurutiensis refers to a small municipality in the interior of the Brazilian Amazon, Juruti, in western Pará, because most of the specimens were found in this area. It is a Latinized word referencing the city. The name Juruti (Yuru-ty) is derived from the indigenous name for the tree trunk Tupi, which means strong neck, alluding to the aspect of the dove that sings in this region and was common in the epoch of the colonization of this city. It is one of the 144 cities in the state of Pará, in northern Brazil, with a rich cultural and natural history.

• Ecology, Diet, and Conservation Status. — This species inhabits small and medium temporary rainwater pools in the center of the ombrophilous forest and temporary puddles of unpaved roads. Predominantly a semi-aquatic species, cohabiting with other species of the families Chelidae (Mesoclemmys gibba, 1 individual captured, and P. platycephala, 3 individuals captured) and Kinosternidae (Kinosternon scorpioides; 31 individuals captured). Tadpoles were flushed from the stomachs of an adult female (3 mL) and a juvenile (1.9 mL) M. jurutiensis. Like other freshwater turtle species from the Brazilian Amazon, M. jurutiensis potentially suffers threats from changes to the ecosystem, such as habitat destruction, habitat fragmentation, river damming for hydroelectric construction, major mining projects, and global climate change. As with other small species in this genus distributed in the Brazilian Amazon, M. jurutiensis has yet to suffer threats from hunting or illegal capture for human consumption or the pet trade.

• Geographic Range. — To date, the known range of M. jurutiensis includes only the State of Pará, Brazil, in the central region of the Brazilian Amazon (Fig. 7). We documented the species from 4 distinct collection regions: 1) municipality of Juruti, a municipality located in the western region of the state of Pará, at the extreme limit of the State of Amazonas (municipality of Parintins); 2) in the northwest in the municipalities of Terra-Santa and Faro, both in Pará State; 3) east in the municipality of Santarém, Pará; and 4) southeast in the municipality of Aveiro, Pará. One of the individuals of M. jurutiensis was captured in the municipality of Aveiro. The species has a known geographic range of 2183 km².

View Figure

• Download as Powerpoint
• Download Figure

Table 4. Morphometric data (absolute values, in millimeters) and relationship between morphometric measurements and carapace length (%CL) for holotype INPA-H41283, paratype INPA-H41284, paratype INPA-H41285, and paratype INPA-H41286 of Mesoclemmys jurutiensis, from Juruti, Pará, Brazil.

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

DISCUSSION

Although the Amazon is considered a turtle hotspot (Mittermeier et al. 2015) and a priority area for chelonian conservation (Buhlmann et al. 2009; Rhodin et al. 2018; Ennen et al. 2020), there are still gaps in our knowledge of the natural history, distribution, taxonomy, and ecology of Brazilian Amazonian chelids. Inconsistency in the identification of species of Mesoclemmys and incorrect taxonomic descriptions are commonplace (Cunha et al. 2019). We are presenting for the first time a taxonomic description based on classical taxonomy (i.e., morphometrics and morphology) combined with molecular genetics (COI and 16S sequences).

Regarding the absence of ossification in the region of the bridge, the present study is the first record of this anatomical condition for species of chelids from the Amazon. Behavioral and anatomical studies should be conducted in order to investigate this characteristic; it is potentially an adaptation for the deposition of eggs during nesting.

Related Species. — Mesoclemmys jurutiensis is morphologically different from other extant species (Fig. 8; Table 5). Among the sympatric congeners, M. jurutiensis resembles Mesoclemmys gibba in body size, but M. gibba has a relatively higher carapace and differs significantly in head size; in addition, the two species have a very different color pattern.

View Figure

• Download as Powerpoint
• Download Figure

Table 5. Morphometric data (means ± SD, in millimeters) and relationship between morphometric measurements with carapace length (%CL) for adults, juveniles, and hatchling of Mesoclemmys jurutiensis; for adults and hatchlings of Mesoclemmys raniceps (hatchling data from Cunha et al. 2019); for adults and hatchlings of Mesoclemmys gibba (hatchlings data from Ferronato et al. 2010). * = absolute value; NR = not reported.

View Fullscreen

With regard to related species, M. gibba hatchlings (artificially incubated) from a tributary of the Madeira River, Novo Aripuanã municipality, have a similar mottled black and green dorsal head pattern, described by Ferronato et al. (2010), whereas the hatchlings and juveniles of M. gibba have a moss-green head with irregular spots and brown streaks. In hatchling specimens of M. gibba, the tympanum is light orange with a black center and the orange coloration extends anteriorly to the apex of the jaws and posteriorly onto the ventral scales of the neck, which is mottled orange-black, as are the ventral surfaces of the limbs (Cunha et al. 2019). The hatchling M. jurutiensis specimen is different from any other Mesoclemmys hatchling we have seen.

Perhaps M. gibba hatchlings lose this coloration after a few months. Carapace length of the specimens of Cunha et al. (2019) was 41.0–49.0 mm after 2 wks of growth and they weighed 10.0–18.0 g at hatching, compared with data reported for the Madeira River specimens (54.0–63.0 mm CL and 20.0–30.0 g), with no caruncle or umbilical scar (Ferronato et al. 2010), suggesting that the latter turtles were not recent hatchlings. Ferronato et al. (2010) also described the hatchlings as having a black postorbital streak, but the hatchlings in their photos do not have this coloration; their heads are completely mottled in black and green, including the tympanum. Mesoclemmys jurutiensis has a completely black head different from the heads of the M. gibba hatchlings we have seen, both for those captured in the wild and for those artificially incubated.

Further regarding sympatric congeners, M. jurutiensis differs from M. raniceps mainly in body size, where adult individuals of M. raniceps can reach 315 mm. Another difference between the species is in coloration. As already mentioned, M. jurutiensis has a completely black head and neck, both juveniles and adults. In turn, M. raniceps has a head and neck in shades of dark gray with irregular white spots on the dorsal and ventral regions of the head. There is a pair of relatively small barbels for M. raniceps.

Regarding the color of the carapace, M. raniceps presents in dark tones, with regions dotted in an irregular way with white to pale yellow tones. In turn, M. jurutiensis has a reddish brown carapace. Another, striking difference between species is found in the size of the abdominal plastral scute, where for M. raniceps it is proportionally larger than in M. jurutiensis. The characteristics mentioned above make it easy to distinguish M. jurutiensis from M. raniceps and M. gibba.

For M. nasuta, the biggest difference is found in the color of individuals, both adults and hatchlings, because they have a bicolor coloration pattern with a bright, vibrant yellow patch on the ventral region of the head, passing through the mandible and extending to the chin, at the height of the eyes, in a uniform manner well-defined, from the beginning of the nostril to the neck. The dorsal part of the head is dark brown. Also, M. nasuta is considered a megacephalic species and M. jurutiensis is not.

Mesoclemmys jurutiensis is distinguished from other Brazilian Mesoclemmys, such as M. perplexa and M. tuberculata of the caatinga region in northeastern Brazil (South Atlantic Basin and São Francisco Basin) and M. vanderhaegei of the Pantanal and Cerrado in central Brazil (upper Paraguay River Basin), by having flat scales in the periocular region, smaller plastron shields proportionally, especially the humeral and pectoral shields, by the mean size of adult individuals, and by the structure and color of the head, especially in the tympanic region (Fig. 9).

View Figure

• Download as Powerpoint
• Download Figure

Genetic Data. — Clearly, the genus Mesoclemmys needs more in-depth studies that review its taxonomy and the phylogenetic relationships among species. Although the molecular results present M. raniceps as the species with the lowest rate of genetic divergence, there is a large and conspicuous morphological divergence of M. jurutiensis from M. raniceps (Gray 1855) in size and shape of the head and shape and color of the jaw. In addition, M. jurutiensis differs from M. raniceps in body size, plastron and carapace shape, and coloration. We bring here a detailed description of a new species of Mesoclemmys because we conducted this work with a integrative analysis between traditional morphology and molecular analyzes.

Our phylogenetic hypothesis corroborates the results produced by Guillon et al. (2012), Holley et al. (2019), and Colston et al. (2020), in the sense of placing C. fimbriata in a separate clade, placing Mesoclemmys hogei in a clade with Phrynops spp., placing M. tuberculata and M. vanderhaegei together, and placing P. platycephala in a clade with the paraphyletic group of Acanthochelys spp.

For Chelidae, the divergence between species is more marked at the morphological than the genetic level. Divergences between species are relatively low, with some species in this family showing divergences of < 2% (Reid et al. 2011). Recent evolutionary radiation added to the low rates of evolution of mtDNA already observed in turtles, potentially can cause the low levels of interspecific divergences (Fujita et al. 2004; Reid et al. 2011; Vargas-Ramírez et al. 2012; Kehlmaier et al. 2019). Additionally, our species delimitation performed via ABGD also recovered the samples collected in Juruti as a taxon distinct from the species so far recognized.

We present a set of genetic and phenotypic information to introduce a species new to science. Mesoclemmys jurutiensis exhibits divergence from nearby species and is accompanied by different and preserved phenotypic characters. In addition, the isolation of M. jurutiensis, the presence of adult, juvenile, and hatchling individuals, and the habitat where they were found support the validity of the new species.