True river dolphins are some of the rarest and most endangered of all vertebrates. They comprise relict evolutionary lineages of high taxonomic distinctness and conservation value, but are afforded little protection. We report the discovery of a new species of a river dolphin from the Araguaia River basin of Brazil, the first such discovery in nearly 100 years. The species is diagnosable by a series of molecular and morphological characters and diverged from its Amazonian sister taxon 2.08 million years ago. The estimated time of divergence corresponds to the separation of the Araguaia-Tocantins basin from the Amazon basin. This discovery highlights the immensity of the deficit in our knowledge of Neotropical biodiversity, as well as vulnerability of biodiversity to anthropogenic actions in an increasingly threatened landscape. We anticipate that this study will provide an impetus for the taxonomic and conservation reanalysis of other taxa shared between the Araguaia and Amazon aquatic ecosystems, as well as stimulate historical biogeographical analyses of the two basins.
Citation: Hrbek T, da Silva VMF, Dutra N, Gravena W, Martin AR, et al. (2014) A New Species of River Dolphin from Brazil or: How Little Do We Know Our Biodiversity. PLoS ONE 9(1): e83623. doi:10.1371/journal.pone.0083623
Editor: Samuel T. Turvey, Zoological Society of London, United Kingdom
Received: September 13, 2012; Accepted: October 28, 2013; Published: January 22, 2014
Copyright: © 2014 Hrbek et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Financial support for fieldwork and laboratory analyses was provided by grants CNPq-483408/2009-3 and 482253/2012-6, CNPq/CT Amazônia-575603/2008-9 and CNPq/FAPEAM SISBIOTA-563348/2010-0 to IPF, and the Scott Neotropical Fund and the Society for Marine Mammalogy to WG. TH and IPF were supported by a Bolsa de Pesquisa scholarship from CNPq during the study, ND by a CAPES fellowship and WG by a CNPq fellowship. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
True river dolphins of the genus Inia are widely distributed throughout the Amazon, Orinoco, and Araguaia-Tocantins basins (Fig. 1) . The Araguaia-Tocantins River basin currently does not form part of the Amazon River basin , , the two basins having become disconnected although not completely isolated at the transition of the Pliocene to the Pleistocene . Currently only a narrow canal west of Marajó Island links the Amazon River delta with the Pará River into which the Araguaia-Tocantins River drains. Connectivity between the Araguaia-Tocantins and Amazon basins is further restricted by a series of major rapids on the lower Tocantins River as it transits from the Brazilian Shield to the Amazonian lowland plain, the largest of which have been flooded in the process of creation of the Tucuruí Dam (03° 50′ 0.36″ S; 49° 38′ 15.55″ W). Major barriers such as waterfalls and rapids are thought to delimit Inia taxa as in the case of the upper Madeira River rapids which are thought to delimit I. geoffrensis and I. boliviensis , , . Similarly, the rapids on the upper Orinoco and Negro Rivers together with the Cassiquiare Canal are thought to delimit the Orinoco and Amazon subspecies of I. geoffrensis , , . The Araguaia-Tocantins River system is thus not only almost completely disconnected from the Amazon River basin, the portion of it that flows on the Brazilian Shield is isolated by rapids from the mouth of the river system and the Amazon River delta by a series of rapids. It is therefore possible that Inia individuals upstream of these rapids or even in the entire system represent a biological entity evolutionarily independent of I. geoffrensis in the Amazon basin. To test this hypothesis we analyzed a series of mitochondrial DNA, nuclear DNA and morphological characters of the two known species of Inia as well as individuals from the Araguaia River basin.
Figure 1. Distribution map of all known species and subspecies of Inia.
Black outline denotes the limit of the Amazon basin. Question marks denote uncertainty as to which species occurs in the Tocantins River downstream of the Tucuruí dam which potentially delimits the distributions of I. geoffrensis and I. araguaiaensis sp. nov. Bars on the Madeira River represent a series of rapids that delimit the distribution of I. geoffrensis and I. boliviensis. The single bar on the northern limit of the Amazon basin represents the Casiquiare canal which connects the Amazon and Orinoco basins, and is thought to delimit the I. g. humboldtiana subspecies from I. g. geoffrensis.doi:10.1371/journal.pone.0083623.g001
Materials and Methods
Samples were collected either from carcases, or from live animals using minimally invasive sampling procedures . Protocol for handling and removing small quantities of cutaneous tissue samples from the caudal fluke of live animals was approved by the Committee on the Ethics of Animal Use (Comissão de Ética do Uso de Animais – CEUA) of National Research Institute of the Amazon (INPA).
Samples were obtained from animals in the Amazon-Solimões basin, the Bolivian sub-basin, and the Araguaia River. Samples were maintained in 95% ethanol and deposited in the official tissue collection (Coleção de Tecidos de Genética Animal – CGTA) of the Federal University of Amazonas and the tissue collection (Coleção de Tecidos) of the Instituto Nacional de Pesquisas da Amazônia. The licenses for collections were obtained through IBAMA/ICMBIO No. 11325-1 to IPF, No. 13462-1 to VMS, and No. 16741-2 to ND. In total we sampled 45 I. boliviensis from the Bolivian sub-basin, 44 I. geoffrensis from the Amazon-Solimões basin and 32 Inia sp. individuals from the Araguaia River.
Nuclear Microsatellite DNA Amplification
Total genomic DNA was extracted from skin samples using Qiaquick blood and tissue kit (Hilsden, Germany) following the manufacturer’s protocol. Polymerase Chain Reaction (PCR) amplification was performed on total genomic DNA. Negative controls were run for all reactions.
To characterize Inia individuals, we used 7 microsatellite pairs (Ig2B1, Ig11B1, Ig8H1, Ig10E, Ig11D2, Ig3A1, Ig7F2) developed by Gravena et al.  for I. geoffrensis, and 3 microsatellite pairs (Ttr11, Ttr48, TtruAAT40) developed by Caldwell et al.  and Rosel et al.  for Tursiops truncatus. Genotyping was done according Gravena et al.  on the ABI 3130XL (Life Technologies) automatic sequencer. The program GeneMapper (Life Technologies) was used to extract raw fragment data, and infer fragment sizes.
Mitochondrial DNA Amplification
The complete cytochrome b gene was amplified using standard PCR conditions with the primers IniaGluF.1 (5′-GACCAATGACATGAAAAATCATCG-3′) and IniaProR.1 (5′-CAGCTTTGGGTGTTGGTGGTGA-3′) developed for this study. The cytochrome oxidase I DNA barcode region was amplified using the C_VF1LFt1-C_VRLRt1 primer cocktail . All the regions were amplified in a final volume of 15 µl and contained 7.6 µl of ddH2O, 1.2 of 25 mM MgCl2, 1.2 µl of 10 mM dNTPs, 1.5 µl of 10x buffer (750 mM Tris-HCl – pH 8.8 at 25°C, 200 mM (NH4)2SO4, 0.1% (v/v) Tween 20), 1.5 µl of each primer (2 µM), 0.3 µl of Taq DNA Polymerase (1 U/µl) and 1 µl of DNA (concentration varied between 10 ng and 30 ng). PCR reagents were obtained from Fermentas (Vilnius, Lithuania). PCR reaction consisted of 35 cycles of denaturation at 93°C for 60 seconds, primer annealing at 50°C for 40 seconds, and primer extension at 72°C for 90 seconds, followed by a final extension at 72°C for 5 minutes. PCR products were purified using EXO-SAP (Exonuclease – Shrimp Alcaline Phosphatase) following the manufacturer’s suggested protocol .
PCR products were evaluated on a 1% agarose gel, and then purified using the polyethylene glycol and ethanol precipitation protocol . cytochrome b PCR products were sequenced using the primer IniaCytbF.1 (5′-CTACCCTGAGGACAAATATCATT-3′), IniaCytbF.2 (5′-CCCGACCTACTAGGAGACCC-3′), IniaCytbF.3 (5′-CCCGACTTACTAGGGGACCC-3′) and IniaCytbR.1 (5′-GAAGAAGCGTGTTAGTGTTGC-3′) developed for this study. cytochrome oxidase c I DNA barcode region PCR products were sequenced with M13F(−21) and M13R(−27) primers . Cycle sequencing PCR followed manufacturer’s recommended protocol for BigDye sequencing kit (Life Technologies); primer annealing temperature was 50°C. Cycle sequencing reaction products were resolved on the ABI 3130XL (Life Technologies) automatic sequencer. Sequence products were edited, concatenated and aligned using the program Geneious v5.6.3 . Sequences are deposited in Genbank under accession numbers KF657385– KF657432, KF657593– KF657640, KF802241– KF802276, KF802313– KF802348, KF802349– KF802380 and KF802381– KF802412.
Population Genetic Characterization of Nuclear Microsatellite Data
The program ARLEQUIN version 22.214.171.124  was used to test for linkage disequilibrium between pairs of loci, and to test for Hardy-Weinberg equilibrium. In cases of multiple comparisons, p-values were adjusted using the Holm method  implemented in the stats package of the software R 2.14.1 .
Structure Analyses of Nuclear Microsatellite Data
The existence of biological groups is a necessary prerequisite for inferring the existence of a species under the Biological Species Concept (BSC) . The microsatellite data were therefore analyzed in the program STRUCTURE version 2.3.3 ,  with the goal of assigning individuals to groups, given a specific number groups (K). We used the ‘admixture’ and ‘correlated-allelic-frequencies’ models with a location prior. Assignment space was explored with 1,000,000 MCMC chains, preceded by 100,000 MCMC chains discarded as burn-in. Convergence was examined by viewing profiles of posterior probabilities, and values of α. Each analysis was repeated 20 times from a different randomly selected starting point, raw output was processed in Structure Harvester 0.6.92 . The 20 independent runs were summarized in the program CLUMPP 1.1.2 , and results were visualized in the program DISTRUCT 1.1 . The most likely number of biological groups (K) was inferred using Bayes’ Rule .
Principal Components and Correspondence Analysis of Nuclear Microsatellite Data
The objective of this analysis was to reduce the dimensionality of the microsatellite data using the Ade4 package  in the software R 2.14.1 . The first and second principal components were then plotted against each other, and the dispersion of points within each sample group was visualized by an ellipse encompassing a 67.53% density contour around each group representing the Inia from the Araguaia River, I. geoffrensis and I. boliviensis. Discrimination of the three groups of Inia was further assessed using the correspondence analysis implemented in the Ade4 package of R. Around each group an ellipse encompassing a 67.53% density contour was drawn.
Population Aggregation Analysis and Phylogenetic Analysis of Mitochondrial DNA Data
Evolutionary lineages are diagnosed by discontinuities in character variation among lineages, and correspond to phylogenetic species. The existence of lineages is therefore a necessary and sufficient prerequisite for inferring the existence of a species under the different conceptualizations of the Phylogenetic Species Concept (PSC) –. The existence of lineages in a non tree-based context  was inferred using Population Aggregation Analysis performed at the level of an individual , , while the existence of lineages in a phylogenetic tree-based context ,  was performed using Maximum Parsimony  and Bayesian Inference  analyses. Maximum Parsimony topology was estimated using PaupRat  implemented on the CIPRES Science Gateway website . We used the TBR option to swap topologies, and ran 100 independent replicates. Bayesian Inference topology for molecular data was estimated in MrBayes 3.1.2 ; we ran two replicates, each with four independent chains for 10 million MCMC generations, sampling a topology every 10,000 generations. Searches were done using the General Time Reversible (GTR) model of molecular evolution  with probability of change from one nucleotide state to another assumed to follow a gamma distribution, and a portion of sites were assumed to be invariant. The GTR+G+I model was suggested as the most appropriate model of molecular evolution in jModeltest  using the Akaike Information Criterion (AIC)  to select among alternative models. To improve the accuracy of phylogenetic reconstruction , , we partitioned the data a priori by codon position. The GTR+G+I model of molecular evolution was applied to each codon position, and parameters were assumed to be uncorrelated between partitions. Convergence between the two independent MrBayes runs was assessed by comparing the posterior probability estimates for each analysis using a scatter-plot created by the compare command in AWTY . Convergent topologies were combined for the calculation of posterior clade probabilities. Following Huelsenbeck et al. , the frequency of a clade among the sampled trees was interpreted as the posterior probability of that clade; the posterior probabilities are true probabilities under the assumed model of substitution ; thus we considered clades to be significantly supported when Bayesian posterior probabilities were ≥95% .
Phylogenetic Dating Using Mitochondrial DNA Data
The objective was to estimate the divergence time of the three lineages of Inia from their respective sister taxa. To provide calibration points, we included taxa that resulted from diversification events for which a good fossil record exists, including taxa near the root of our phylogeny. We included additional taxa for the purpose of cross validation of our estimates with those published by Xiong et al.  based on complete mitochondrial genomes. We used prior age distributions so that they encompassed the entire period of the origination of the fossil as reported in literature, unless a more precise fossil age estimate was provided by one of the authors of the fossil literature. Calibrations were performed in the program BEAST 1.7.1 , with divergence times being lognormally distributed. Calibrations were chosen such that the youngest age of the distribution corresponded with the youngest fossil age estimate, and the oldest fossil estimate was encompassed within the 95% credible interval of the lognormal age distribution. Calibrations follow those used by Xiong et al.  with modifications of the 95% credible intervals to truly and exactly encompass the upper and lower fossil ages. The following four calibration points were used:
- The divergence between Mysticeti (baleen whales) and Odontoceti (toothed whales) was calibrated using the earliest record of mysticetes from the Eocene/Oligocene boundary (see  for a review). There has been debate about when the Eocene/Oligocene transition occurred (38–33.5 Mya in ), and therefore we chose a lognormal age prior so that the earlier age corresponded to 33.5 Mya and the older 95% credible interval (CI) encompassed the beginning of the Late Eocene (40 Mya) (standard deviation = 1.068).
- The age of the root of crown Odontoceti was calibrated using the earliest record of a physeterid (sperm whales) – the Late Oligocene Ferecetotherium ; see also . The age prior was lognormally distributed so that the youngest age corresponded to 23.7 Mya, the end of the Late Oligocene, and the older 95% credible interval (CI) encompassed the beginning of the Late Oligocene (30 Mya) (standard deviation = 1.135).
- The divergence between Iniidae (the Amazon River dolphins) and Pontoporidae (La Plata River dolphins) was calibrated using the earliest record of a pontoporid, the Middle Miocene Brachydelphis . The age prior was lognormally distributed so that the youngest fossil age corresponded to 11.2 Mya and the older 95% CI encompassed the beginning of the Middle Miocene (16.6 Mya) (standard deviation = 1.223).
- The divergence between Phocoenidae (porpoises) and Monodontidae (narwhals) was calibrated using the earliest record of a phocoenid, the Late Miocene Salumiphocaena . This fossil is approximately 10–11 million years old , and therefore we chose a lognormal age prior with the younger age corresponding to 10 Mya and the older 95% CI encompassing the beginning of the Late Miocene (11.2 Mya) (standard deviation = 1.061).
The BEAST analysis consisted of 1×108 generations with a random starting tree, the Yule speciation prior, uniform [0, 100] prior distribution for the GTR+G+I substitution rates, default values for remaining priors, random starting topology, and topologies were sampled every 10000 generations. To improve the accuracy of phylogenetic reconstruction , , , we partitioned the data a priori by codon position. The topologies were used to calculate posterior clade probabilities and divergence times. The reliability of the estimated divergence times was further assessed by calculating an expected substitution rate compatible with the observed divergence time.
Cranial Measurement Analyses
Cranial measurements and counts follow da Silva  with data from da Silva  serving as a base for comparison. All measurements were standardized as described in da Silva  while counts were analyzed untransformed. Since only few individuals were available for morphological analyses, no statistical analyses were performed and only means and ranges of equivalent measurements and counts were compared to those taken from I. geoffrensis and I. boliviensis.
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix “http://zoobank.org/”. The LSID for this publication is: urn:lsid:zoobank.org:pub:2F417149-4E97-4245-BF83-A8FA06388BEB.The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central and LOCKSS.
We collected molecular data from 44, 45 and 32 individuals of Inia geoffrensis, I. boliviensis and Inia from the Araguaia River. The 10 microsatellite loci were unlinked, and at Hardy-Weinberg equilibrium within each lineage. In I. geoffrensis and I. boliviensis all 10 loci were polymorphic, while in the Inia from the Araguaia River only eight loci were polymorphic. In I. geoffrensis, I. boliviensis and the Inia from the Araguaia River we observed 12, 10 and 3 alleles in the cytochrome b gene, and 9, 6 and 2 alleles in the cytochrome oxidase c I gene, respectively.
Analyses conducted with the program STRUCTURE ,  assuming the admixture model, which allows individuals to have ancestors from more than one biological group, and interpreted using Bayes’ Rule  identify three groups as the most likely number of biological groups within the genus Inia (Fig. 2). Mean posterior probabilities for a given number of biological groups are listed in Table 1. The three biological groups corresponded to the group from the Araguaia River and to the species I. geoffrensis and I. boliviensis. All but three individuals from the Araguaia River group had a clustering coefficient (q) higher than 99%; the remaining three individuals had a q value greater than 95%, indicating a lack of genetic admixture of the Araguaia River group with the species I. geoffrensis and I. boliviensis. Individuals with clustering coefficients larger than 90% are normally considered to have originated from only one biological population (according to STRUCTURE manual), i.e. are non-admixed.
We analyzed 10 nuclear microsatellite loci sampled from 45 I. boliviensis, 44 I. geoffrensis and 32 I. araguaiaensis sp. nov. individuals. Highest posterior probability is associated with three biological groups; mean LnP(K = 3) = −2041.115.doi:10.1371/journal.pone.0083623.g002
Table 1. Mean posterior probabilities of each a priori number of biological groups of Inia (K) analyzed.doi:10.1371/journal.pone.0083623.t001
Principal component analysis (PCA) of microsatellite data carried out in the R package Ade4  also indicates the presence of three groups of individuals corresponding to Inia individuals from the Araguaia River, and the species I. geoffrensis and I. boliviensis (Fig. 3). Assuming a priori the existence of three groups representing the individuals from the Araguaia River and the species I. geoffrensis and I. boliviensis, correspondence analysis (COA) unambiguously separates these three entities (Fig. 4).
Figure 3. Principal component analysis of the microsatellite data  performed in the Ade4 package in the R vers. 2.14.1 statistical program.
Principal component axes 1 and 2 are plotted on the X and Y axes, respectively, and ellipses denote 67.53% density contours around each sample group.doi:10.1371/journal.pone.0083623.g003
Figure 4. Correspondence analysis of the microsatellite data  performed in the Ade4 package in the R vers. 2.14.1 statistical program.
Factors 1 and 2 are plotted on the X and Y axes, respectively, and ellipses denote 67.53% density contours around each sample group.doi:10.1371/journal.pone.0083623.g004
The Population Aggregation Analysis  of mitochondrial sequence data identified three clusters corresponding to the group from the Araguaia River and to the species I. geoffrensis and I. boliviensis. The group from the Araguaia River is diagnosable by 28 and 9 apomorphic character states in the mitochondrial cytochrome b and the barcode region of the cytochrome oxidase c I gene, respectively, while I. geoffrensis is diagnosable by 18 apomorphic characters in the cytochrome b and 15 in the barcode region of the cytochrome oxidase c I gene, and I. boliviensis is diagnosable by 21 apomorphic characters in the cytochrome b and 9 in the barcode region of the cytochrome oxidase c I gene (Table 2). Phylogenetic analyses under both Maximum Parsimony  and Bayesian Inference  place all individuals of Inia from the Araguaia River into a strongly supported sister clade to I. geoffrensis (Fig. S1). Furthermore, the estimated divergence of the Araguaia River Inia clade from its I. geoffrensis sister lineage is 2.08 million years (95% HPD 0.811–3.744 million years) (Fig. 5); converted into a substitution rate, this divergence results in an expected substitution rate of 1.01×10−8 substitutions per site per year. The clade formed by Inia from the Araguaia basin and I. geoffrensis from the Amazonas basin is sister to I. boliviensis with an estimated time of divergence of 2.87 mya (95% HPD 1.315-4.903 mya) (Fig. 5).
Figure 5. Bayesian phylogenetic analysis and divergence time estimation in BEAST 1.7.1  using the complete mitochondrial DNA cytochrome b gene.
We used four independent fossil calibration points (indicated as orange bars). Numbers at nodes represent estimates of divergence times with highest posterior probability, while bars around each divergence time estimate represent 95% highest posterior density of the estimate.doi:10.1371/journal.pone.0083623.g005
Analysis of osteological and morphometric data indicates a number of differences between the Inia individuals from the Araguaia River and I. geoffrensis and I. boliviensis. In general the width of the braincase across the parietals as well as the greatest width of the maxillaries and the greatest parietal width are larger when compared with I. geoffrensis and I. boliviensis. Specifically diagnostic differences between the Araguaia Inia and I. geoffrensis are in the width of the braincase across the parietals of females (range: 9.6 to 11.5 cm, mean: 10.6 cm, n = 2 vs. range: 11.6 to 13.5 cm, mean: 12.4 cm, n = 9). Principal differences between the Araguaia Inia and I. boliviensis are in the number of teeth per hemimandible (24–28) when compared with I. boliviensis (31–35); larger greatest parietal width in males (range: 8.1 to 9.2 cm, mean: 8.7 cm, n = 2 vs. range: 5.2 to 7.6 cm, mean: 6.7 cm, n = 3); and larger greatest width of maxillaries in females (range: 11.3 to 11.8 cm, mean: 11.6 cm, n = 2 vs. range: 9.3 to 10.4 cm, mean: 9.8 cm, n = 5).
There is, therefore, strong evidence that Inia individuals from the Araguaia River represent a biological group distinct from I. geoffrensis and I. boliviensis. They also form a lineage with an evolutionary trajectory independent of I. geoffrensis and I. bolivensis since the early Pleistocene. Furthermore, there exist morphological, albeit subtle, differences between the Inia individuals from the Araguaia River and individuals of I. geoffrensis and I. boliviensis. Therefore we recognize and designate the population of Inia inhabiting the Araguaia River as a new species.
Cetacea Brisson, 1762.
Odontoceti Flower, 1867.
Iniidae Gray, 1846.
Inia d’Orbigny, 1834.
Inia araguaiaensis sp. Nov.
Hrbek, Farias, Dutra and da Silva 2014 urn:lsid:zoobank.org:act:399CD69C-6452-46C4-BDA9-2416409B50A7.
The specific name refers to the geographic area where the new species is currently known to occur.
Suggested common name.
We propose the common names “Boto-do-Araguaia” and “Araguaian boto” for this species. Boto is the common regional and international name of species of Inia, and Araguaia refers to the geographic distribution of this species.
INPA MA 144 (Fig. 6); Adult male, skeleton partially complete (skull, mandibles, total of 68 teeth, 2 tympanic bulla, 1 sternum, hyoid apparatus [1 basyhyal and 2 stylohyal], a total of 37 vertebrae [7 cervical, 11 thoracic, 3 lumbar, 16 caudal], 8 chevrons, 12 pair of ribs, 1 floating rib) and fragments of skin preserved in alcohol deposited at the Mammal Collection of the National Institute for Amazonian Research (INPA), Manaus, Amazonas, Brazil. The specimen was collected near the entrance of Lake Jurumirim (15° 10′ 19.65″ S; 51° 38′ 48.30″ W), Araguaia River, state of Goiás on 21 September 2009 by Claryana Costa Araújo. The type was found floating dead, in an early stage of decomposition.
Figure 6. The cranium (6a) and the mandible (6b) of the holotype of Inia araguaiaensis sp. nov.; 10 cm ruler shown.doi:10.1371/journal.pone.0083623.g006
INPA MA 145 and INPA MA 146; skulls and mandibles of one adult male (INPA MA 145) with a total of 87 teeth available and one pregnant female (INPA MA 146) with a total of 72 teeth available. These specimens were both collected on 19 August 2010, about 10 km apart, below and above the entrance of Jurumirim Lake, on the Araguaia River. Collectors were V.M.F. da Silva and N. Dutra, and the material was deposited at the Mammal Collection of the National Institute for Amazonian Research (INPA), Manaus, Amazonas, Brazil, together with fragments of skin preserved in alcohol.
Inia araguaiaensis sp. nov. occurs in the central region of Brazil, in the Araguaia River basin (Fig. 1). We recorded it in the lower and middle Araguaia River from Barra do Garças (15° 54′ 02.8″S; 52° 16′ 58.7″W) to the Santa Isabel rapids (06° 09′ 01.1″S; 48° 22′ 14.7″W), and in several tributaries such as Vermelho, Peixe, Crixás-Açú and Água Limpa Rivers, and dos Tigres and Rico Lakes, all in the state of Goiás, and Lake Montaria in the state of Mato Grosso. The species is likely present in the Tocantins River , and potentially in the whole Araguaia-Tocantins River Basin.
Inia araguaiaensis sp. nov. differs from all its congeners by having the following unique characters: 28 diagnostic sites (apomorphies) in the complete mitochondrial DNA cytochrome b gene; 9 diagnostic sites in the cytochrome oxidase c I DNA barcode region; a reduced number of teeth per hemimandible (24–28) when compared with I. boliviensis (31–35); narrower braincase across parietals of females (range: 9.6 to 11.5 cm, mean: 10.6 cm, n = 2) compared to I. geoffrensis females (range: 11.6 to 13.5 cm, mean: 12.4 cm, n = 9); larger greatest parietal width in males (range: 8.1 to 9.2 cm, mean: 8.7 cm, n = 2) compared to I. boliviensis males (range: 5.2 to 7.6 cm, mean: 6.7 cm, n = 3); and larger greatest width of maxillaries in females (range: 11.3 to 11.8 cm, mean: 11.6 cm, n = 2) compared to I. boliviensis females (range: 9.3 to 10.4 cm, mean: 9.8 cm, n = 5).
Inia araguaiaensis sp. nov. shows rather subtle morphological differences when compared to I. geoffrensis and more accentuated differences when compared to I. boliviensis. All species of the genus Inia show two distinct types of teeth, but I. araguaiaensis sp. nov. possesses 24–28 teeth per hemimandible compared to the 25–29 in I. geoffrensis and 31–35 in I. boliviensis. The 19–21 front pairs of teeth in both upper and lower jaws are conical, but the posterior 6–8 pairs (cheek teeth) have one pointed cusp extending from the crown towards the labial side and a depression on the lingual side. The width of the brain-case across parietals as well as the greatest width of maxillaries and the greatest parietal width are larger when compared with the two other species. Inia araguaiaensis sp. nov. is the sister species to I. geoffrensis, sharing their most recent common ancestor 2.08 mya. It forms a biological group distinct from I. geoffrensis and I. boliviensis and there is strong indication of reproductive isolation from these species.
All species concepts applicable to sexually reproducing organisms that conceive of species as an outcome of an evolutionary process implicitly or explicitly assume that species are groups of individuals that can freely exchange genetic information with each other, but are unable to do so with individuals of other species. For example, the Biological Species Concept (BSC) focuses on ongoing patterns of reproduction, using actual or potential reproduction among individuals of the species as a criterion for identifying species limits . On the other hand, the Phylogenetic Species Concept (PSC) focuses on historical patterns of reproduction that would maintain autonomous evolutionary lineages –. Our analyses identify I. araguaiaensis sp. nov. as both an evolutionary lineage and a group of actually interbreeding individuals, satisfying the criteria necessary for the recognition of this taxon as a species under the Biological as well as Phylogenetic Species Concepts.
Analyses conducted with the program STRUCTURE ,  and interpreted using Bayes’ Rule  identify three groups as the most likely number of biological groups within the genus Inia (Fig. 2). These biological groups represent I. araguaiaensis sp. nov., I. geoffrensis and I. boliviensis. The probability that only two biological groups exist, and that I. araguaiaensis sp. nov. individuals belong to the same biological group as I. geoffrensis is 4.22×10−98. Allowing for individuals to have ancestors from more than one biological group results in the inference that individuals of I. araguaiaensis sp. nov. do not have admixed ancestry. This does not necessarily mean that there is no gene flow or that there has been no recent historical gene flow between I. araguaiaensis sp. nov. and either I. geoffrensis or I. boliviensis, but the data do indicate that all three species represent distinct biological populations characterized by a shared system of mating among members of each species (sensu Templeton ), but not among species. Evolutionary distinctness of I. araguaiaensis sp. nov., I. geoffrensis and I. boliviensis is also manifested by diagnostic alleles, distinct allelic frequencies and allelic combinations of each species as observed in the PCA (Fig. 3). Assuming the existence of the three species of Inia a priori, the three species are clearly separable in multivariate allelic space (Fig. 4).
Inia araguaiaensis sp. nov., I. geoffrensis and I. boliviensis are also diagnosable as biological entities evolutionarily independent of each other, i.e. lineages characterized by independent evolutionary trajectories. As such, each species is diagnosable by a series of unique characters. These characters are both molecular (Table 2) and morphological (see Diagnosis above). However, due to morphological conservatism, the majority of the diagnostic characters are molecular.
The three species have a long history of evolutionary independence with strongly supported chronological order of speciation events (Figs. 2 & 3). The earliest species-level divergence resulted in the separation of I. boliviensis from the lineage leading to I. geoffrensis and I. araguaiaensis sp. nov. in the late Pliocene (Fig. 5). This divergence was followed shortly by the divergence of I. geoffrensis from I. araguaiaensis sp. nov. in the early Pleistocene (Fig. 5). Both divergences are associated with major geological events in the Amazon basin.
With highest probability I. araguaiaensis sp. nov. diverged from its I. geoffrensis sister lineage 2.08 mya (95% HPD 0.811–3.744 mya) (Fig. 5). This divergence time corresponds to the last time that the Araguaia-Tocantins basin discharged directly into the Amazon River at the Pliocene/Pleistocene boundary . At the time of the Rossetti and Valeriano study , the Pliocene included the Gelasian stage, which lasted from 2.588 to 1.806 mya , and therefore the authors estimated an isolation of the Araguaia-Tocantins basin at 1.8 mya. The estimated isolation of the Araguaia-Tocantins basin from the Amazon basin 1.8 mya corresponds closely to the estimated divergence of I. araguaiaensis sp. nov. from the Araguaia basin and I. geoffrensis from the Amazon basin at 2.08 mya (Fig. 5), leading us to the conclusion that speciation was vicariant.
Our results also clearly show that I. boliviensis is a distinct species. Inia boliviensis is an independent biological population (Figs. 2 & 3), and the estimated time of divergence of I. boliviensis from its sister clade is 2.87 mya (95% HPD 1.315–4.903 mya) (Fig. 5). This divergence estimate is broadly compatible with the uplift of the Fitzcarrald Arch which commenced to isolate the Bolivian sub-basin from the Amazon basin no earlier than 4 mya , . The observed divergence between I. boliviensis and its sister clade has, therefore, likely been driven by the uplift of the Fitzcarrald Arch, and has lead to vicariant speciation. Similar signatures of vicariant diversification can be observed in other Amazonian aquatic vertebrate groups such as the piranhas of the genera Pygocentrus and Serrasalmus  and peacock basses of the genus Cichla .
Currently much of the scientific community , , including the IUCN (http://www.iucnredlist.org/details/10831/0), consider I. boliviensis to be a subspecies of I. geoffrensis. However, in 2012 the Ad-Hoc Committee on Taxonomy of the Society for Marine Mammalogy (http://www.marinemammalscience.org/) started considering I. boliviensis a full species citing the control region mtDNA data  and nuclear introns data . We confirm this taxonomic reassessment and demonstrate that I. boliviensis is diagnosable under the biological species (Figs. 2 & 3) as well as the phylogenetic species (Table 1 and Fig. 5) criteria.
Similar to the taxonomic uncertainty involving I. boliviensis, one may question whether or not I. araguaiaensis sp. nov. represents a valid taxon. The observed phylogenetic and mating-system patterns could potentially represent intraspecific variation in a strongly genetically structured species. We reject this notion in favor of the existence of a new species in the Araguaia River basin on the grounds that Inia araguaiaensis sp. nov. 1) has no apparent actual or recent history of admixture with other Inia species, 2) has a large number of diagnostic characters that unambiguously discriminate it from other Inia species, and 3) diverged from its sister taxon I. geoffrensis with highest probability 2.08 million years ago, but no less than 0.811 million years ago. This divergence time exceeds most sister-taxon divergence-time estimates in the closely related family Delphinidae , . Therefore, all the presented data corroborate our interpretation that I. araguaiaensis sp. nov. is an evolutionary species, i.e. a lineage with an independent evolutionary trajectory.
True river dolphins are some of the rarest and most endangered of all vertebrates comprising relict evolutionary lineages of high taxonomic distinctness , . Of the four extant species listed by the IUCN, three are listed in one of the IUCN Red List’s threat categories (VU – Pontoporia blainvillei, EN – Platanista gangetica, CR – Lipotes vexillifer) while Inia geoffrensis is DD. Although listed as critically endangered by the IUCN, the Chinese baiji (Lipotes vexillifer), became extinct between 2004 and 2006 . The description of I. araguaiaensis represents the first description of a new true river dolphin since 1918 when the now extinct L. vexillifer was described. Lipotes vexillifer was also the first river dolphin, and probably the first cetacean, to become extinct as a result of anthropogenic activities –. The conservation status of I. araguaiaensis has not been assessed, but merits attention. Inia araguaiaensis appears to have low levels of genetic diversity, and also has a restricted distribution in what we estimate to be an approximately 1500 km stretch of the Araguaia River with additional fragmented habitat in the Tocantins River basin . The number of individuals is also probably low. In an environmental impact study conducted for the licensing of the Santa Isabel hydroelectric project on the lower Araguaia River, a total of 122 dolphins were seen over a 12 week period in 60 km survey stretch of the river . Assuming an average width of 0.5 km , and assuming that no animal was seen more than once, this would represent approximately 1525 individuals in the Araguaia River basin. Araújo and da Silva  estimated, depending on the transect type used, 0.80 to 0.98 indiv/km2 in the Araguaia River suggesting a population of 600 to 735 individuals. Araújo , assuming that I. araguaiaensis occurs throughout the total 2110 km length of the Araguaia River estimated a total of 1033 individuals Extrapolating from the highest I. geoffrensis density of 1.3 indiv/km2 observed in non-protected areas of the Amazon basin (the Napo River in Ecuador) , 975 individual could occur in the Araguaia River basin. Populations of the middle and upper Tocantins River are fragmented by six hydroelectric dams, not including the Tucurui dam, and are likely to have very few individuals. Since the 1960’s the Araguaia River basin has been experiencing significant anthropogenic pressure via agricultural and ranching activities, and the construction of hydroelectric dams, all of which have had negative effects on many biotic and abiotic aspects of the functioning of the Araguaia River ecosystem  which I. araguaiaensis inhabits. Similarly I. araguaiaensis in the Tocantins River is affected by agricultural and industrial development, and is fragmented into isolated populations by hydroelectric dams . Interpreting these data using IUCN criteria, Inia araguaiaensis should minimally be considered vulnerable (VU Red List category).
Bayesian phylogenetic analysis carried out in MrBayes 3.1.2 . Numbers above nodes represent posterior probabilities. We used all 45 I. boliviensis, 44 I. geoffrensis and 32 I. araguaiaensis sp. nov. individuals in this analysis. Intraspecific relationships were collapsed and represented as a triangle with the depth of the triangle representing intraspecific divergence, and height of the triangle number of individuals. Maximum parsimony topology is identical.
We thank Allan Larson and Jonathan Losos for comments and critical reading of the manuscript; IBAMA for granting collecting permits; Claryana Costa Araújo for collecting the holotype; Soldado, Ney, Nildon, Waldenilson and João Pena for assistance in the field; the environmental police of the Goiás state, Santo Antônio Energia, Tenente Coronel Leite and the Brazilian Air Force, APA Meandros do rio Araguaia for logistic support in the field. Computational resources were provided by the University of Puerto Rico High Performance Computational Facility.
Conceived and designed the experiments: TH IPF. Performed the experiments: ND WG. Analyzed the data: TH. Contributed reagents/materials/analysis tools: IPF TH ARM VMFS. Wrote the paper: TH IPF. Performed field work: ND WG VMFS.
- 1. Da Silva VMF (1994) Aspects of the biology of the Amazonian dolphins genus Inia and Sotalia fluviatilis Cambridge, UK: The University of Cambridge.
- 2. Rossetti D de F, Valeriano MM (2007) Evolution of the lowest Amazon basin modeled from the integration of geological and SRTM topographic data. Catena 70: 253–265 doi:10.1016/j.catena.2006.08.009.
- 3. Goulding M, Barthem RB, Ferreira EJG (2003) The Smithsonian Atlas of the Amazon. Washington, DC: Smithsonian Institution Press.
- 4. Banguera-Hinestroza E, Cárdenas H, Ruiz-García M, Marmontel M, Gaitán E, et al. (2002) Molecular identification of evolutionarily significant units in the Amazon River dolphin Inia sp. (Cetacea: Iniidae). J Hered 93: 312–322 doi:10.1093/jhered/93.5.312.
- 5. Pilleri G, Gihr M (1977) Observations on the Bolivian, Inia boliviensis, (D’Orbigny, 1834) and the Amazonian bufeo, Inia geoffrensis (Blainville, 1817), with a description of a new subspecies (Inia geoffrensis humboldtiana). In: Pilleri G, editor. Investigations on Cetacea. Bogotá, Colombia, Vol. 8: 11–76.
- 6. Da Silva VMF, Martin AR (2000) A study of the boto, or Amazon River dolphin (Inia geoffrensis), in the Mamiraua Reserve, Brazil: Operation and techniques. In: Reeves RR, Smith BD, Kasuya T, editors. Biology and Conservation of freshwater Cetaceans in Asia. Occasional Papers of the IUCN, Species Survival Commission No. 23. Gland, Switzerland: IUCN, Vol. 23: 121–131.
- 7. Gravena W, Hrbek T, da Silva VMF, Astolfi-Filho S, Farias IP (2009) Microsatellite loci for population and parentage analysis in the Amazon River dolphin (Inia geoffrensis de Blainville, 1817). Mol Ecol Resour 9: 600–603 doi:10.1111/j.1755-0998.2008.02458.x.
- 8. Caldwell M, Gaines MS, Hughes CR (2002) Eight polymorphic microsatellite loci for bottlenose dolphin and other cetacean species. Mol Ecol Notes 2: 393–395 doi:10.1046/j.1471-8278.2002.00270.x.
- 9. Rosel PE, Forgetta V, Dewar K (2005) Isolation and characterization of twelve polymorphic microsatellite markers in bottlenose dolphins (Tursiops truncatus). Mol Ecol Notes 5: 830–833 doi:10.1111/j.1471-8286.2005.01078.x.
- 10. Ivanova NV, Zemlak TS, Hanner RH, Hebert PDN (2007) Universal primer cocktails for fish DNA barcoding. Mol Ecol Notes 7: 544–548 doi:10.1111/j.1471-8286.2007.01748.x.
- 11. Werle E, Schneider C, Renner M, Völker M, Fiehn W (1994) Convenient single-step, one tube purification of PCR products for direct sequencing. Nucleic Acids Res 22: 4354–4355.
- 12. Paithankar KR, Prasad KS (1991) Precipitation of DNA by polyethylene glycol and ethanol. Nucleic Acids Res 19: 1346.
- 13. Messing J (1983) New M13 vectors for cloning. Methods Enzymol 101: 20–78.
- 14. Drummond AJ, Ashton B, Buxton S, Cheung M, Cooper A, et al.. (2012) Geneious v5.6.3. Available: http://www.geneious.com/.
- 15. Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: A new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour 10: 564–567 doi:10.1111/j.1755-0998.2010.02847.x.
- 16. Holm S (1979) A simple sequentially rejective multiple test procedure. Scand J Stat 6: 65–70.
- 17. R Development Core Team (2011) R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Available: www.R-project.org.
- 18. Mayr E (1942) Systematics and the Origin of Species. New York, NY: Columbia Univ. Press.
- 19. Falush D, Stephens M, Pritchard JK (2003) Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics 164: 1567–1587.
- 20. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155: 945–959.
- 21. Earl DA, VonHoldt BM (2012) STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv Genet Resour 4: 359–361 doi:10.1007/s12686-011-9548-7.
- 22. Jakobsson M, Rosenberg NA (2007) CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23: 1801–1806 doi:10.1093/bioinformatics/btm233.
- 23. Rosenberg NA (2004) DISTRUCT: a program for the graphical display of population structure. Mol Ecol Notes 4: 137–138 doi:10.1046/j.1471-8286.2003.00566.x.
- 24. Pericchi LR (2005) Model selection and hypothesis testing based on objective probabilities and Bayes Factors. In: Dey DK, Rao CR, editors. Bayesian Thinking: Modeling and Computation. Amsterdam, The Netherlands: Elsevier. 115–149. doi:10.1016/S0169-7161(05)25004-6.
- 25. Thioulouse J, Chessel D, Dolédec S, Olivier J-M (1997) ADE-4: a multivariate analysis and graphical display software. Stat Comput 7: 75–83.
- 26. Cracraft J (1983) Species concepts and speciation analysis. Curr Ornithol 1: 159–187.
- 27. Baum DA, Donoghue MJ (1995) Choosing among alternative “phylogenetic” species concepts. Syst Bot 20: 560–573.
- 28. Rosen DE (1979) Fishes from the uplands and intermontane basins of Guatemala: revisionary studies and comparative geography. Bull Am Museum Nat Hist 162: 267–376.
- 29. Davis JI, Nixon KC (1992) Populations, genetic variation, and the delimitation of phylogenetic species. Syst Biol 41: 421–435 doi:10.1093/sysbio/41.4.421.
- 30. Rach J, DeSalle R, Sarkar IN, Schierwater B, Hadrys H (2008) Character-based DNA barcoding allows discriminations of genera, species and populations in Odonata. Proc R Soc London Ser B Biol Sci 275: 237–247 doi:10.1098/rspb.2007.1290.
- 31. Edwards AWF, Cavalli-Sforza LL (1963) The reconstruction of evolution. Ann Hum Genet 27: 105–106.
- 32. Huelsenbeck JP, Ronquist FR (2001) MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17: 754–755.
- 33. Nixon KC (1999) The Parsimony Ratchet, a new method for rapid parsimony analysis. Cladistics 15: 407–414.
- 34. Miller MA, Pfeiffer WC, Schwartz TS (2010) Creating the CIPRES Science Gateway for inference of large phylogenetic trees. Proc Gatew Comput Environ Work: 1–8.
- 35. Tamura K, Nei M (1993) Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol Biol Evol 10: 512–526.
- 36. Posada D (2008) jModelTest: Phylogenetic Model Averaging. Mol Biol Evol 25: 1253–1256 doi:10.1093/molbev/msn083.
- 37. Akaike H (1974) A new look at the statistical model identification. IEEE Trans Automat Contr 19: 716–723 doi:10.1109/TAC.1974.1100705.
- 38. Castoe TA, Doan TM, Parkinson CL (2004) Data partitions and complex models in Bayesian analysis: The phylogeny of gymnophthalmid lizards. Syst Biol 53: 448–469 doi:10.1080/10635150490445797.
- 39. Nylander JAA, Ronquist FR, Huelsenbeck JP, Nieves-Aldrey JL (2004) Bayesian phylogenetic analysis of combined data. Syst Biol 53: 47–57 doi:10.1080/10635150490264699.
- 40. Nylander JAA, Wilgenbusch JC, Warren DL, Swofford DL (2008) AWTY (are we there yet?): a system for graphical exploration of MCMC convergence in Bayesian phylogenetics. Bioinformatics 2008: 581–583 doi:10.1093/bioinformatics/btm388.
- 41. Huelsenbeck JP, Ronquist FR, Nielsen R, Bollback JP (2001) Bayesian inference of phylogeny and its impact on evolutionary biology. Science (80−) 294: 2310–2314 doi:10.1126/science.1065889.
- 42. Rannala B, Yang Z (1996) Probability distribution of molecular evolutionary trees: a new method of phylogenetic inference. J Mol Evol 43: 304–311 doi:10.1007/BF02338839.
- 43. Huelsenbeck JP, Rannala B (2004) Frequentist properties of Bayesian posterior probabilities of phylogenetic trees under simple and complex substitution models. Syst Biol 53: 904–913 doi:10.1080/10635150490522629.
- 44. Xiong Y, Brandley MC, Xu S, Zhou K, Yang G (2009) Seven new dolphin mitochondrial 30 genomes and a time-calibrated phylogeny of whales. BMC Evol Biol 9: 20 doi:10.1186/1471-2148-9-20.
- 45. Drummond AJ, Rambaut A (2007) BEAST: Bayesian Evolutionary Analysis by Sampling Trees. BMC Evol Biol 7: 214 doi:10.1186/1471-2148-7-214.
- 46. Fordyce E, de Muizon C (2001) Evolutionary history of cetaceans: A review. In: Mazin JM, de Buffrenil V, editors. Secondary Adaptation of Tetrapods to Life in Water. Munich, Germany: Verlag Friedrich Pfeil. 169–233.
- 47. Berggren WA, Kent DV, Obradovich JD, Swisher III CC (1992) Toward a revised Paleogene geochronology. In: Prothero DR, Berggren WA, editors. Eocene-Oligocene climatic and biotic evolution. Princeton, NJ, USA: Princeton University Press. 29–45.
- 48. Mchedlidze GA (1970) Nekotorye obshie cherty istorii kitoobraznykh. (Some general characteristics of the evolution of cetaceans). Tbilisi, Gruzinskoi SSR, Georgia: Metsniereba Press, Institut Paleobiologii Akademia Nauka.
- 49. Carroll RL (1988) Vertebrate Paleontology and Evolution. New York, NY: W. H. Freeman and Company.
- 50. De Muizon C (1983) Pliopontos littoralis, un noveau Platanistidae Cetacea du Pliocène de la côte pèruvienne. C R Biol 296: 625–628.
- 51. Barnes LG (1985) Evolution, taxonomy, and antitropical distributions of the porpoises (Phocoenidae, Mammalia). Mar Mammal Sci 1: 149–165 doi:10.1111/j.1748-7692.1985.tb00003.x.
- 52. Brandley MC, Schmitz A, Reeder TW (2005) Partitioned Bayesian analyses, partition choice, and the phylogenetic relationships of scincid lizards. Syst Biol 54: 373–390 doi:10.1080/10635150590946808.
- 53. Falush D, Stephens M, Pritchard JK (2007) Inference of population structure using multilocus genotype data: dominant markers and null alleles. Mol Ecol Notes 7: 574–578 doi:10.1111/j.1471-8286.2007.01758.x.
- 54. Araújo C, Wang JY (2012) Botos (Inia geoffrensis) in the upper reaches of the Tocantins River (central Brazil) with observations of unusual behavior, including object carrying. Aquat Mamm 38: 435–440 doi:10.1578/AM.38.4.2012.435.
- 55. Templeton AR (2006) Population Genetics and Microevolutionary Theory. New York, NY: John Wiley & Sons.
- 56. Gibbard PL, Head MJ, Walker MJC (2009) Formal ratification of the Quaternary System/Period and the Pleistocene Series/Epoch with a base at 2.58 Ma. J Quat Sci 25: 96–102 doi:10.1002/jqs.1338.
- 57. Espurt N, Baby P, Brusset S, Roddaz M, Hermoza W, et al.. (2010) The Nazca Ridge and uplift of the Fitzcarrald Arch: implications for regional geology in northern South America. In: Hoorn C, Wesselingh FP, editors. Amazonia: Landscape and Species Evolution: A Look into the Past. Wiley-Blackwell. 89–100.
- 58. Espurt N, Baby P, Brusset S, Roddaz M, Hermoza W, et al. (2007) How does the Nazca Ridge subduction influence the modern Amazonian foreland basin? Geology 35: 515–518 doi:10.1130/G23237A.1.
- 59. Hubert N, Duponchelle F, Nuñes J, Garcia Davila C, Pauly D, et al. (2007) Phylogeography of the piranha genera Serrasalmus and Pygocentrus: implications for the diversification of the Neotropical ichthyofauna. Mol Ecol 16: 2115–2136 doi:10.1111/j.1365-294X.2007.03267.x.
- 60. Renno J-F, Hubert N, Torrico JP, Duponchelle F, Nunez Rodriguez J, et al. (2006) Phylogeography of Cichla (Cichlidae) in the upper Madera basin (Bolivian Amazon). Mol Phylogenet Evol 41: 503–510.
- 61. Da Silva VMF (2009) Amazon River Dolphin (Inia geoffrensis). In: Perrin WF, Wursig B, Thewissen JGM, editors. Encyclopedia of Marine Mammals. London, UK: Academic Press. 18–20.
- 62. Rice DW (1998) Marine Mammals of the World: Systematics and Distribution. San Diego, CA, USA: The Society for Marine Mammalogy Special Publication 4.
- 63. Ruiz-García M, Caballero S, Martinez-Agüero M, Shostell JM (2008) Molecular differentiation among Inia geoffrensis and Inia boliviensis (Iniidae, Cetacea) by means of nuclear intron sequences. In: Koven VT, editor. Population Genetics Research Progress. Boca Raton, FL, USA: Nova Publishers Inc. 177–203.
- 64. Viltrup JT, Ho SYW, Foote AD, Morin PA, Kreb D, et al. (2011) Mitogenomic phylogenetic analyses of the Delphinidae with an emphasis on the Globicephalinae. BMC Evol Biol 11: 65 doi:10.1186/1471-2148-11-65.
- 65. Cassens I, Vicario S, Waddell VG, Balchowsky H, Van Belle D, et al. (2000) Independent adaptation to riverine habitats allowed survival of ancient cetacean lineages. Proc Natl Acad Sci U S A 97: 11343–11347 doi:10.1073/pnas.97.21.11343.
- 66. Hamilton H, Caballero S, Collins AG, Brownell Jr RL (2001) Evolution of river dolphins. Proc R Soc London Ser B Biol Sci 268: 549–556 doi:10.1098/rspb.2000.1385.
- 67. Turvey ST, Pitman RL, Taylor BL, Barlow J, Akamatsu T, et al. (2007) First human-caused extinction of a cetacean species? Biol Lett 3: 537–540 doi:10.1098/rsbl.2007.0292.
- 68. Turvey ST (2008) Witness to Extinction: How We Failed to Save the Yangtze River Dolphin. Oxford, UK: Oxford University Press.
- 69. Reeves RR, Gales NJ (2006) Realities of Baiji conservation. Conserv Biol 20: 626–628 doi:10.1111/j.1523-1739.2006.00472.x.
- 70. Gesai C (2010) EIA Santa Isabel [Environmental Impact Study Santa Isabel]. Porto Alegre, RG, Brazil.
- 71. Araújo CC (2010) Distribuição e estimativas populacionais do boto Inia geoffrensis (de Blainville, 1817) (Iniidae) no médio rio Araguaia (Brasil central) Instituto Nacional de Pesquisas da Amazônia.
- 72. Araújo CC, da Silva VMF (2010) Abundance and density of the boto Inia geoffrensis in the Araguaia River (central Brazil). XIV Reunião de Trabalho de Especialistas em Mamíferos Aquáticos da América do Sul (RT) 8°Congresso da Sociedade Latinoamericana de Especialistas em Mamíferos Aquáticos (SOLAMAC). Sociedade Latinoamericana de Especialistas em Mamíferos Aquáticos (SOLAMAC). p.1.
- 73. Trujillo F, Gomez-Salazar C, Portocarrero M, Diazgranados MC, Usma JS (2010) Abundance estimation program of South American river dolphins 2006–2009: A summary. In: Trujillo F, Crespo E, van Damme PA, Usma JS, editors. The Action Plan for South American River Dolphins 2010–2020. Bogotá, D.C., Colombia: WWF, Fundación Omacha, WDS, WDCS, Solamac. 167–176.
- 74. Latrubesse EM, Amsler ML, de Morais RP, Aquino S (2009) The geomorphologic response of a large pristine alluvial river to tremendous deforestation in the South American tropics: The case of the Araguaia River. Geomorphology 113: 239–252 doi:10.1016/j.geomorph.2009.03.014.