Priscila Santos Carvalho Systematics and Biogeography of the Hydropsini tribe (Serpentes: Xenodontinae) São José do Rio Preto 2022 Priscila Santos Carvalho Systematics and Biogeography of the Hydropsini tribe (Serpentes: Xenodontinae) Tese apresentada como parte dos requisitos para obtenção do título de Doutora em Biologia Animal, junto ao Programa de Pós-Graduação em Biodiversidade, do Instituto de Biociências, Letras e Ciências Exatas da Universidade Estadual Paulista “Júlio de Mesquita Filho”, Câmpus de São José do Rio Preto. Financiadora: CAPES Orientador: Prof. Dr. Diego José Santana Silva São José do Rio Preto 2022 Carvalho, Priscila Santos C331s Systematics and biogeography of the Hydropsini tribe (Serpentes: Xenodontinae) / Priscila Santos Carvalho. -- , 2022 144 f. : tabs., mapas Tese (doutorado) - Universidade Estadual Paulista (Unesp), Instituto de Biociências Letras e Ciências Exatas, São José do Rio Preto, Orientador: Diego José Santana Silva 1. América do Sul. 2. Dipsadidae. 3. Filogenia molecular. 4. Filogeografia. 5. Sistemática. I. Título. Priscila Santos Carvalho Systematics and Biogeography of the Hydropsini tribe (Serpentes: Xenodontinae) Tese apresentada como parte dos requisitos para obtenção do título de Doutora em Biologia Animal, junto ao Programa de Pós-Graduação em Biodiversidade, do Instituto de Biociências, Letras e Ciências Exatas da Universidade Estadual Paulista “Júlio de Mesquita Filho”, Câmpus de São José do Rio Preto. Financiadora: CAPES Comissão Examinadora Profa. Dra. Fernanda de P. Werneck INPA – Manaus, AM Profa. Dra. Thaís Barreto Guedes UNICAMP – Campinas, SP Profa. Dra. Renata Magalhães Pirani UNR – Reno, USA Profa. Dra. Sara Ruane FMNH – Chicago, USA Prof. Dr. Diego José Santana UFMS – Campo Grande, MS Orientador São José do Rio Preto 04 de novembro de 2022 Às mulheres. Agradecimentos O Diego é um orientador ímpar, um ser humano do bem, engraçado, acolhedor, nerd e querido amigo. Há uns meses atrás ele me emprestou um livro chamado "O dia do Curinga". Que livro do caralho! E várias vezes o enxerguei naquelas citações. Ele é um curinga na vida de muitas pessoas. Me espelho muito nele pra me tornar uma pesquisadora melhor. Diego cuida do laboratório muito bem e mantém o ambiente saudável, alegre, vivo, onde podemos falar sobre tudo e todos. Sempre me senti respeitada, valorizada e querida. Diego, obrigada por cada palavra amiga, pelos ensinamentos, pela paciência comigo e por me respeitar como cientista e mulher. Você é o meu curinga. Agradeço aos curadores das coleções científicas e colegas que cederam as amostras de tecidos ou sequências, possibilitando a execução do meu trabalho. Muito obrigada, Adrian A. Garda (UFRN), Ana L. C. Prudente (MPEG), Frank Burbrink (AMNH), Carla Bessa e Vanessa Arzamendia (FHUC), Gustavo H. C. Vieira e Fagner R. Delfim (CHUFPB), Selvino N. de Oliveira (CHUFSC), Guarino Colli (CHUNB), Diego Baldo (CONICET-UNaM), Felipe Grazziotin (IBSP), Christopher Austin (LSUMZ), Miguel R. Treffaut (USP), Paulo Passos (MNRJ), Selma Torquato (MUFAL), Reuber Brandão (UnB), Omar Torres-Carvajal (PUCE), Santiago C. Fischer (PUC-RS), Rejane M. L. da Silva (UFBA-NOAP), Paulo Garcia (UFMG), Felipe Curcio (UFMT), Luiz R. R. Rodrigues (UFOPA), Laura Verrastro e Márcio B. Martins (UFRGS), Vanda Ferreira (UFMS), Gregory Watkins-Colwell (YPM), Pedro M. S. Nunes e Pedro I. Simões (CHUFPE), María E. López e Enrique Arbeláez-Cortés (IAvH), Yaneth Muñoz-Saba (ICN), Martha P. R. Pinilla (MHN-UIS), Fernando Rojas-Runjaic, Hugo E. Cabral, Henrique Folly, Jorge A. D. Pérez e Leandro A. Silva. Também gostaria de agradecer ao Ricardo Marques, Albedi A. Cerqueira Jr. e Fernando Rojas-Runjaic por ter cedido as fotos usadas no capítulo 1 e na apresentação. Quero agradecer também a galera que entrou nessa loucura, assim como eu, de estudar essas cobrinhas incríveis. Nathalie Citeli, Antonio Moraes-da-Silva, Albedi A. Cerqueira Jr. e os professores, Reuber Brandão, Pedro M. S. Nunes, Felipe Curcio e Paulo Passos. Bora time resolver os babados desses bichos maravilhosos. Agradeço ao Ricardo Koroiva pela ajuda no laboratório de Biologia Molecular, ao Felipe Camurugi pela ajuda nas análises moleculares, a Eliana de Oliveira pelas conversas sobre ciência e a vida, por ter me ajudado desde meu mestrado e por estar sempre disponível a sanar dúvidas sobre métodos filogeográficos. Agradeço imensamente ao Don Shepard pelo apoio na bancada, por preparar minhas placas para o sequenciamento e por ter topado participar dessa empreitada. Aos meus amigos Bruna Carvalho, Jéssica Laurentino, Paulo Ricardo, Cirlene da Cunha e Reinaldo Lima por sempre me incentivarem a seguir meus sonhos, por toda amizade, carinho, conversas loucas. Vocês são fodas! Contem sempre comigo, Aos queridos amigos que compartilhei tantos momentos no MZUSP: Ana Bottallo, Marcela Brasil, Renata Fadel, Renata Montalvão, Rafael Henrique, Raissa Siqueira, Paola Sánchez e Diego Cavalheri. Obrigada por todas as conversas, risadas, viagens e pelo carinho. Vocês fazem parte de uma fase muito importante da minha vida. Aos imensos e queridos amigos do laboratório Mapinguari. Com certeza meus dias foram mais alegres, cheio de bagunça e gritaria com vocês: Sarah Mângia, Felipe Camurugi, Eliana Oliveira, Carla Guimarães, Hugo Cabral, Juan Cuestas-Carrilo, Beatriz Vasconcelos, Ibrahim Nehemy, Márcia Müller, Henrique Nogueira, Diego Cavalheri, Nuno (João Emílio), Sean Keuroghlian-Eaton, Leonardo Castro, Henrique Oliveira. O que seria do Mapinguari sem os estagiários? Valeu demais, Lauany Serafim, Ana Alice Cabral, Rafaela Machado e Pílade Filho, vocês me salvaram em vários momentos nessa reta final insana. Ao longo dessa trajetória eu fiz amizades que nem imaginava. Sempre me sinto querida por você Ma (Marcella Souza), mesmo há centenas de quilômetros de distâncias, quando eu voltar em Juiz de Fora, você poderia me emprestar aquele vestido maravilho todo estampado com cobras lindas? hahahahaha. Que surpresa maravilhosa você, Beatriz Vasconcelos, ter escolhido justamente o Mapinguari pra continuar sua trajetória rumo ao desconhecido mundo do jacaré-paguá. Bia, nem sei o que escrever sobre você, já estou com os olhos marejados, mas sei que seu abraço é o mais aconchegante. Você é uma amiga do caralho! Obrigada por tudo. Te amo! O final da minha trajetória no doutorado se mistura muito com a história da Carlinha. Enfrentamos juntas esses momentos alegres e conturbados. Carlota Joaquina, que fase, né amiga? Hahahaha. Que jornada maluca, mas cheia de risada, Pepinha, Queimadinha, gritaria nós vivemos. Outra surpresa maravilhosa que o Mapinguari me deu. Você é pau pra toda obra hahaha. Me leva logo pra conhecer Viçosa e comer a comida da sua mãe. Te amo! Karol Ceron e Renata Fadel, mais conhecidas como Cacatua berrante e Maridinha hahahahaha. Eu sei, a Karol vai berrar quando ler isso. Vocês também foram a minha família em Campo Grande. Dou risada sozinha quando lembro da nossa convivência. Estou morrendo de saudade de vocês. Amo vocês! Vanessinha, minha companheira de leitura. Agora poderemos discutir com mais frequência nossas aventuras pelos maravilhosos livros que navegamos. Obrigada por ser uma amiga tão parceira. Te amo! Peetinha (Sarah Mângia), minha amiga e irmã que amo tanto que chega a doer. Obrigada por ser meu porto seguro, minha família em Campo Grande, por ter me dado o prazer de ser tia das maravilhosas, Farofinha e Mentirinha. Ainda bem que sobrevivemos aquele campo bizarro na Amazônia e agora podemos rir juntas daquela história e de tantas outras coisas. Te amo! Agradeço a minha família por sempre me apoiar e acreditar nos meus sonhos. Obrigada Mãe, por ser uma mulher tão forte e batalhadora. Amo vocês, Sandra, Alexandre, Nilo, Dani e Júnior. Um agradecimento especial a minha irmã maravilhosa Daniela, que é um ser humano incrível, obrigada irmã por tudo! Você é o meu espelho. Por muito tempo eu achei que estava infeliz, sem perspectiva, não sabia onde as minhas escolhidas estavam me levando. Mas percebi que estava muito frustrada com o momento histórico que estamos vivendo. Porém, também percebi que amo muito o que faço e que por mais que esse processo foi desgaste emocional e fisicamente eu ficava empolgada com cada artigo novo que lia. Gostaria que o processo no início tivesse sido diferente e que esse atual momento político no Brasil não tivesse me desgastado tanto. Portanto, gostaria de agradecer cada estudante, professore, pesquisadore, cientista que resiste a cada dia na expectativa de um mundo melhor. Eu quero deixar claro aqui que sou completamente contra esse atual governo genocida, ecocida, misógino, machista, preconceito, neofascista. O simples ato de estudar e fazer ciência no Brasil é uma resistência. #FORABOLSONARO. Agradeço ao programa de pós-graduação em Biodiversidade (antigo Biologia Animal) da Universidade Estadual Paulista, os coordenadores do PPG Francisco Langeani e Antonio Carlos Lofego e aos funcionários da secretária por todo suporte e apoio, principalmente nesse momento pandêmico. O presente trabalho foi realizado com apoio da Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Código de Financiamento 001. Ano passado eu morri, mas esse ano eu não morro. (Sujeito de sorte - Belchior) Respeita a existência ou espere resistência. (Noite inteira - Pitty) Nada em biologia faz sentido exceto à luz da evolução. (Theodosius Dobzhansky, 1973) Resumo A região Neotropical é caracterizada pela alta diversidade biológica e pela presença de grandes bacias hidrográficas, que exercem papel importante no padrão de distribuição e evolução da biota. Tal diversidade está também associada, dentre vários fatores, ao soerguimento final dos Andes a partir do Mioceno. Apesar dos recentes avanços na sistemática de serpentes, a diversidade de várias linhagens neotropicais permanece pouco compreendida. Aqui, abordamos a sistemática e biogeografia de Hydropsini (Pseudoeryx, Hydrops e Helicops), e apresentamos uma análise filogenética com base na maior amostragem molecular feita até o momento, cerca de 83% da diversidade da tribo (22 das 26 espécies), incluindo 1.080 sequencias de dois genes mitocondriais (16S e Cytb) e quatro nucleares (Cmos, NT3, BDNF e R35). Dividimos a tese em três capítulos. No primeiro capítulo inferimos a filogenia de Hydropsini, cujos resultados mostraram que os três gêneros são válidos e recuperamos a monofiletismo da tribo. A inclusão de mais espécies resultou em relacionamentos contrastantes quando comparados com inferências filogenéticas anteriores. No segundo capítulo, com base na filogenia datada do capítulo 1, conduzimos uma estimativa de área ancestral, a fim de investigar a história biogeográfica dessas serpentes aquáticas. Revelamos que o cenário ancestral mais provável para a diversificação de Hydropsini foi a região amazônica, por volta de 21 milhões de anos, no início do Mioceno. Discutimos como o dinamismo da paisagem durante o Mioceno na região amazônica teve um grande impacto na diversificação de Hydropsini, influenciado principalmente pelo sistema Pebas. Por fim, no terceiro capítulo, nos baseamos nos achados sobre o complexo de Helicops leopardinus (He. leopardinus, He. modestus, He. infrataeniatus e He. tapajonicus) do primeiro capítulo e estudamos a filogeografia deste complexo para compreendermos sua diversidade e estruturação genética em níveis mais recentes de diversificação. O complexo de Helicops leopardinus se originou durante o Pleistoceno (~1,2 Mya), e mostramos que este complexo pode representar uma única espécie com cinco agrupamentos geneticamente estruturados com fluxo gênico de forma desigual entre eles. A diferenciação genética do complexo de He. leopardinus é explicada principalmente pela interação da distância geográfica, variação climática e bacias hidrográficas. Fornecemos novas propostas sobre padrões e processos de diversificação para um complexo de espécies de cobras aquáticas amplamente distribuído ao longo da América do Sul. Palavras-chave: América do Sul. Dipsadidae. Distribuição geográfica. Filogenia molecular. Filogeografia. Sistemática. Xenodontinae. Abstract The Neotropical region is characterized by high biological diversity and the presence of large hydrographic basins, which play an important role in the distribution pattern and evolution of the biota. Such diversity is also associated, among several factors, with the final uplift of the Andes during the Miocene. Despite recent advances in snake systematics, the diversity of several Neotropical lineages remains poorly understood. Here, we address the systematics and biogeography of Hydropsini (Pseudoeryx, Hydrops e Helicops) and present a phylogenetic analysis based on the largest molecular sampling to date, about 83% of the tribe’s diversity (22 of 26 species), including 1,080 sequences of two mitochondrial genes (16S and Cytb) and four nuclear (Cmos, NT3, BDNF, and R35). We organized the thesis into three chapters. In the first chapter, we inferred the phylogeny of Hydropsini, and our results showed that the three genera are valid and we recovered the monophyly of the tribe. The inclusion of more species resulted in contrasting relationships when compared to previous phylogenetic inferences. In the second chapter, based on the dated phylogeny from chapter 1, we conducted an ancestral area estimate in order to investigate the biogeographic history of these aquatic snakes. We reveal that the most likely ancestral scenario for Hydropsini diversification was the Amazon region, around 21 million years ago, in the early Miocene. We discuss how the landscape dynamism during the Miocene in the Amazon region had a great impact on the diversification of Hydropsini, mainly influenced by the Pebas system. Finally, in the third chapter, based on the findings on the Helicops leopardinus complex (He. leopardinus, He. modestus, He. infrataeniatus and He. tapajonicus) from the first chapter, we carried out a phylogeography for these snakes to understand their diversity and genetic structure. The Helicops leopardinus complex originated during the Pleistocene (~1.2 Mya), and we show that this complex may represent a single species with five genetically structured clusters with uneven gene flow between them. The genetic differentiation of the He. leopardinus complex is mainly explained by the interaction of geographic distance, climatic variation, and watersheds. We provide new proposals on diversification patterns and processes for a complex of aquatic snake species widely distributed across South America. Keywords: South America. Dipsadidae. Geographic distribution. Molecular phylogeny. Phylogeography. Systematics. Xenodontinae. FIGURE CAPTIONS CHAPTER 1 Figura 1 – Sampling coverage used in this study for Pseudoeryx, Hydrops, and Helicops throughout South America. 31 Figura 2 – Bayesian phylogenetic inference of Hydropsini based on mtDNA (16S and Cytb) and nuDNA (Cmos, NT3, BDNF, and R35) genes using BEAST. Posterior probabilities (PP) and bootstrap (BS) values are ≥ 0.95 and ≥ 70%, above and below, respectively (we present only well-supported node values). Asterisks (*) indicate different phylogenetic relationships found in the ML tree. Clade A: Hydropsini tribe; Clade B: Pseudoeryx; Clade C: Hydrops; Clade D: Hy. relictualis; Clade E: Hy. caesurus; Clade F: Hy. martii; Clade G: Hy. triangularis; Clade H: Helicops; Clade I: He. leopardinus group; Clade J: He. leopardinus complex; Clade K: He. trivittatus; Clade L: He. carinicaudus group; Clade M: He. danieli; Clade N: He. hagmanni group; Clade O: He. pastazae complex; Clade P: He. polylepis; Clade Q: He. angulatus group. Photos: DJSantana (P. plicatilis, He. angulatus); FJMRojas-Runjac (P. relictualis); AACerqueira Junior (Hy. casesurus); Moraes-da-Silva et al. 2019 (He. boitata) and CHdeONogueira (He. carinicaudus L1). 33 Figura 3 – Bayesian phylogenetic inference of Pseudoeryx and Hydrops based on mtDNA (16S and Cytb) and nuDNA (Cmos, NT3, BDNF, and R35) genes using BEAST. Posterior probabilities (PP) values are ≥0.95 (we present only well-supported node values). Clade B: Pseudoeryx; Clade C: Hydrops; Clade D: Hy. relictualis; Clade E: Hy. caesurus; Clade F: Hy. martii; Clade G: Hy. triangularis. L: indicates distinct lineages. 34 Figura 4 – Bayesian phylogenetic inferences based on mtDNA (16S and Cytb) and nuDNA (Cmos, NT3, BDNF, and R35) genes using BEAST. Posterior probabilities (PP) values are ≥ 0.95 (we present only well-supported node values). (A) Helicops leopardinus group and He. leopardinus complex. Clade I: He. leopardinus group; Clade J: He. leopardinus complex. (B) Helicops trivittatus and He. carinicaudus group. Clade K: He. trivittatus; Clade L: He. carinicaudus group. (C) Helicops danieli, He. hagmanni group and He. pastazae complex. Clade M: He. danieli; Clade N: He. hagmanni group; Clade O: He. pastazae complex. 36 Figura 5 – Bayesian phylogenetic inference of He. polylepis, He. angulatus group and He. pastazae complex based on mtDNA (16S and Cytb) and nuDNA (Cmos, NT3, BDNF, and R35) genes using BEAST. Posterior probabilities (PP) values are ≥0.95 (we present only well-supported node values). Clade P: He. polylepis; Clade Q: He. angulatus group. L: indicates distinct lineages. 37 Figure S1 – Maximum likelihood (ML) phylogenetic inference of Hydropsini tribe based on concatenated mtDNA (16S and Cytb) and nuDNA (Cmos, NT3, BDNF, and R35). Numbers below nodes are bootstrap values (BS) ≥ 70% (we present only well- supported node values). 58 Figure S2 – Geographic distribution of Pseudoeryx and Hydrops samples from this study along six watersheds across South America. 61 Figure S3 – Geographic distribution of Helicops leopardinus group and He. leopardinus complex samples from this study along six watersheds across South America. 62 Figure S4 – Geographic distribution of Helicops trivittatus and He. carinicaudus group samples from this study along of four watersheds across South America. 63 Figure S5 – Geographic distribution of Helicops danieli, He. hagmanni group, Helicops pastazae complex and He. polylepis samples from this study along five watersheds across South America. 64 Figure S6. Geographic distribution of Helicops angulatus group samples from this study along 10 watersheds across South America. 65 CHAPTER 2 Figure 1. Ancestral geographic scenario of the Hydropsini tribe reconstructed under DEC+j model by BioGeoBEARS. The biogeographic areas defined based on the combination of Neotropical ecoregions, with the basin hydrographic, and distribution pattern of Hydropsini species. Single capital letters and colors indicate different biogeographic units used in this study. Mixed letters and colors represent combinations of such units. 78 Figure S1 – Dated Species Tree (*BEAST) of the Hydropsini based on mtDNA (16S and Cytb) and nuDNA (Cmos, NT3, BDNF and R35). Values above the nodes indicate posterior probabilities. The bars represent the 95% HDP. 94 CHAPTER 3 Figure 1 – Geographic distribution of Helicops leopardinus complex samples (H. leopardinus, H. infrataeniatus, H. modestus, and H. tapajonicus) along the major hydrobasins across South America. White symbols indicate samples not used in the BAPS analysis. 110 Figure 2 – Phylogenetic relationships of Helicops leopardinus complex and divergence times by Bayesian Inference in BEAST based on mtDNA 16S and Cytb concatenated genes. Numbers below nodes are posterior probability values; nodes without values indicate PP < 0.95. Bars represent the 95% Highest Posterior Density (HPD) interval for divergence dates. Taxon name, voucher number, and locality are indicated for each terminal. 114 Figure 3 – Bayesian Analysis of Population Structure (BAPS) results: A) spatial clustering of individuals. B) Admixture based on mixture clustering. Each color represents one cluster (among the five clusters) generated by the mitochondrial and nuclear gene dataset. 115 Figure 4 – Network of five clusters indicating the gene flow among them by weighted arrows. 116 Figure 5 – Haplotype networks for each locus within of Helicops leopardinus complex. The haplotype colors correspond to the clusters recovered by BAPS, and the size of its area is proportional to its frequency. Black circles represent unsampled haplotypes and the dashes represent mutational steps. 116 TABLE LIST CHAPTER 1 Table S1 – Taxa, vouchers, locality data, genetic markers and GenBank accession numbers used in this study. Absent acronyms in Sabaj (2022): MTR or MRT: researcher field number Miguel Trefaut U. Rodrigues (University of São Paulo, USP); A: researcher field number Vanessa Arzamendia (Faculty of Humanities and Sciences, National University del Litoral); AAGARDA: researcher field number Adrian Antonio Garda; AF: researcher field number Antoine Fouquet; GGU: researcher field number Giussepe Gagliardi Urrutia (Universidad Nacional de la Amazonía Peruana, Facultad de Ciencias Biológicas); LW: field number of the Federal University of Western Pará; MAP and MAP-T: field and tissue number of the Mapinguari laboratory at the Federal University of Mato Grosso do Sul; MHNBA: Bahia Natural History Museum voucher; RABRANDÃO: researcher field number Reuber Brandão (University of Brasilia) SB: project field number scales of biodiversity; VLF: researcher field number Vanda Lucia Ferreira (Federal University of Mato Grosso do Sul. Other museum acronyms follow Sabaj (2022). L: lineages. 66 Table S2 – List of outgroup taxa and GenBank accession numbers of specimens used in this study. 67 Table S3 – PartitionFinder 2 model of nucleotide substitution. Best-fitting partitioning scheme model of nucleotide substitution for 16S, Cytb, Cmos, NT3, BDNF and R35 genes. 69 CHAPTER 2 Table 1 – Summary of the temporal and spatial main events that influenced the dispersal/vicariance for the Most Recent Common Ancestor (MRCA) of Hydropsini lineages. (A) Transandean; (B) Northwest Amazonia; (C) Southeast Amazonia; (D) Northeast Diagonal; (E) Southwest Diagonal and (F) Atlantic Forest. We follow the stratigraphy International Chronostratigraphic Chart (ICS). 78 Table 2 – Comparison of the BioGeoBEARS model for Hydropsini within six areas based on the log-likelihood (LnL) and the Akaike information criterion (AIC); N, parameters number; d, dispersion rate; e, extinction rate; J, relative probability of speciation between founding events. The best model is shown in bold. 81 Table S1 – Taxa, vouchers, locality data, genetic markers and GenBank accession numbers used in Carvalho 2022 (PSC 2022). L: lineages. Absent acronyms in Sabaj (2022): MTR or MRT: researcher field number Miguel Trefaut U. Rodrigues (University of São Paulo, USP); A: researcher field number Vanessa Arzamendia (Faculty of Humanities and Sciences, National University del Litoral); AAGARDA: researcher field number Adrian Antonio Garda; AF: researcher field number Antoine Fouquet; GGU: researcher field number Giussepe Gagliardi Urrutia (Universidad Nacional de la Amazonía Peruana, Facultad de Ciencias Biológicas); LW: field number of the Federal University of Western Pará; MAP and MAP-T: field and tissue number of the Mapinguari laboratory at the Federal University of Mato Grosso do Sul; MHNBA: Bahia Natural History Museum voucher; RABRANDÃO: researcher field number Reuber Brandão (University of Brasilia) SB: project field number scales of biodiversity; VLF: researcher field number Vanda Lucia Ferreira (Federal University of Mato Grosso do Sul. 95 CHAPTER 3 Table 1. Model selection of what variables influence population genetic diversity using redundancy analyses (RDA) for Helicops leopardinus complex. 117 Table S1 – Taxa, vouchers, locality data, genetic markers and GenBank accession numbers used in this study. Absent acronyms in Sabaj (2022): MTR or MRT: researcher field number Miguel Trefaut U. Rodrigues (University of São Paulo, USP); A: researcher field number Vanessa Arzamendia (Faculty of Humanities and Sciences, National University del Litoral); AAGARDA: researcher's field number Adrian Antonio Garda; AF: researcher field number Antoine Fouquet; GGU: researcher field number Giussepe Gagliardi Urrutia (Universidad Nacional de la Amazonía Peruana, Facultad de Ciencias Biológicas); LW: field number of the Federal University of Western Pará; MAP and MAP-T: field and tissue number of the Mapinguari laboratory at the Federal University of Mato Grosso do Sul; MHNBA: Bahia Natural History Museum voucher; RABRANDÃO: researcher field number Reuber Brandão (University of Brasilia) SB: project field number scales of biodiversity; VLF: researcher field number Vanda Lucia Ferreira (Federal University of Mato Grosso do Sul. 132 Table S2 – PartitionFinder 2 model of nucleotide substitution. Best-fitting partitioning scheme model of nucleotide substitution for 16S, Cytb, Cmos, NT3, BDNF, and R35 genes. 132 Table S3 – Specimens, vouchers, and geographic coordinate of samples used in Bayesian Analysis of Population Structure (BAPS) and which spatial clusters they belong. In bold the specimens that were admixture. 133 Table S4 – Genetic statistics for each locus for cluster Bayesian Analysis of Population Structure (BAPS) of Helicops leopardinus complex. L: length in base pairs (bp); N: sample size; S: number of polymorphic sites; H: number of haplotypes (number of exclusive haplotypes in this lineage); Hd: haplotype diversity; π: nucleotide diversity. neutrality tests Tajima’s D and Fu's F. * Phased sequences; ** sampling were not enough to estimate these parameters; *** genetic variation was not enough to estimate these parameters. 135 SUMMARY 1 GENERAL INTRODUCTION 18 2 CHAPTER 1: Phylogeny and systematic of Hydropsini (Serpentes: Dipsadidae: Xenodontinae) 31 2.1 Introduction 25 2.2 Material and Methods 30 2.3 Results 32 2.4 Discussion 38 2.5 Conclusions 49 2.6 References 51 2.7 Supplementary information 58 3 CHAPTER 2: Spatial-temporal evolution of water snakes in South America driven by Pebas System during the Early Miocene 72 3.1 Introduction 73 3.2 Material and Methods 75 3.3 Results 77 3.4 Discussion 81 3.5 Conclusions 84 3.6 References 85 3.7 Supplementary information 94 4 CHAPTER 3: Phylogeography of Helicops leopardinus complex (Serpentes: Dipsadidae: Xenodontinae) 106 4.1 Introduction 107 4.2 Material and Methods 109 4.3 Results 113 4.4 Discussion 117 4.5 Conclusions 122 4.6 References 122 4.7 Supplementary information 132 5 Final considerations 138 References general introduction 138 18 1. General introduction Hydropsini: A brief systematic history With over 3,900 species encompassing a diversity of morphologies, ecologies, and lifestyles (arboreal, fossorial, terrestrial, aquatic) (Vitt & Caldwell 2013), snakes represent about 35% of squamate diversity (Uetz 2022) and approximately 30% of global snake diversity occurs in the Neotropical region (Guedes et al. 2018). Within the extensive Neotropical diversity of snakes, the family Dipsadidae is one of the largest radiations of Caenophidia (“Advanced Snakes”) (>750 species; sensu Grazziotin et al., 2012). Dipsadids were considered restricted to the New World, but recent placement of Stichophanes and Thermophis as sister to Dipsadidae expanded their distribution into the Old World (e.g., Figueroa et al. 2016). Dipsadidae is morphologically supported by two putative hemipenial synapomorphies (body calyces on the asulcate surface of the lobes and lateral enlarged spines disposed on the sides of the hemipenial body) and molecularly (Zaher et al. 2009). Among dipsadids, the tribe Hydropsini belongs to Xenodontinae Bonaparte, 1845, and is composed by three genera: Helicops Wagler, 1828 (21 species); Hydrops Wagler, 1830 (3 spp); and Pseudoeryx Fitzinger, 1826 (2 spp). This tribe is endemic to South America, occur from Trinidad to Uruguay, and their species are strongly associated and adapted to the aquatic environment (Moraes-da-Silva et al. 2019). Among the variety of lifestyles in snakes, about 5% have aquatic habitats (Pauwels et al. 2008) (e.g., members of Acrochordidae, Boidae, Colubridae, Dipsadidae, Elapidae, Homalopsidae, Natricidae and Viperidae). Thereby, snakes have independently invaded the aquatic environment many times (Pauwels et al. 2008). Interestingly, the Oriental and Neotropical regions have a great hydrological richness and are the two largest regions of aquatic or semiaquatic freshwater snake diversity (Pauwels et al. 2008). Some morphological specializations in aquatic/semiaquatic snakes include position of nostrils on the snout top, valvular nostrils, dorsolaterally oriented eyes (Pauwels et al. 2008; Segall et al. 2016). This morphology adaptations are observed, irrespective of the phylogenetic relationships among species of snakes from different families, suggesting that the aquatic environment does indeed drive the evolution of morphology in snakes, thus driving the evolutionary trajectory of this group of animals (Segall et al. 2016). Among the current taxonomy of Hydropsini, Pseudoeryx has two valid species: P. plicatilis and P. relictualis with allopatric distributions. Though, P. plicatilis presents two subspecies P. p. plicatilis and P. p. mimeticus both with cis Andean distribution. Meanwhile, P. relictualis is a trans Andean species, occurring in Lake Maracaibo, Venezuela (Schargel et al. 2007). Hydrops contains three species: Hy. triangularis, Hy. martii and Hy. caesurus. 19 Hydrops martii and Hy. triangularis have widely overlapping geographic distributions, and are often found in syntopy, especially along the central part of the Amazon basin. Meanwhile, Hy. Caesurus is endemic in the La Plata basin (Argentina, Paraguay and Brazil) (von May et al. 2019). Hydrops differs from Helicops and Pseudoeryx in its smooth dorsal scales (keeled in Helicops) and in its maxillary diastema and color pattern with transverse bands (no diastema and longitudinal lines or dots in Pseudoeryx) (Roze 1957a, b). Lastly, Helicops is the most diverse genus of Hydropsini with 21 known species, diagnosed by the presence of nude flounces on the lobes of the hemipenis (Zaher 1999), and differs from Hydrops and Pseudoeryx by the presence of keels on the dorsal scales, absent in both latter taxa (Roze 1957a, b; Moraes-da-Silva et al. 2019). Helicops has species with a wide and restricted distribution throughout South America (Nogueira et al. 2019) and most occur in the Amazon region (e.g., He. petersi, He. pastazae, He. yacu), while He. leopardinus and He. angulatus are widely distributed. In this genus, there are two species with restricted trans Andean distribution (Moraes-da-Silva et al. 2019; 2021; Citeli et al. 2022). Many species within this tribe have undefined type localities, which makes taxonomic decisions difficult. Some species are only known from a few specimens or their type specimens, as many specimens were collected when tissue collection for molecular analysis was not yet standard practice (e.g., He. yacu, He. hogei). Currently, only 22 specimens have been sequenced representing 12 species out of the 26 recognized. There have been several taxonomic rearrangements in recent decades, mainly due to morphological differences in hemipenis between the three genera (see below) (Roze 1957b; Rossman 1973). Since the three genera share a single internasal scale, eyes and nostrils located in a dorsal position, Roze (1957a, b) indicated a phylogenetic affinity between them, and considered such traits as adaptations of these genera to aquatic habits. The author also suggested that Hydrops and Helicops would be more closely related due to dentition characteristics, qualitative (color pattern) and quantitative (dorsal scales) features, and that Pseudoeryx would be an sister clade of this clade. However, Rossman (1973) based sole on hemipenial evidence considered that such characteristics would be convergent adaptations of the aquatic habit and rejected the hypothesis of Roze (1957a, b). Although he did not place the genera into any family, the author suggested that new evidence beyond the hemipenis might alter his conclusion. Within Xenodontinae, Pseudoeryx and Hydrops are the only genera that do not have hemipenial characteristics that would diagnose them as belonging to this subfamily (Zaher 1999). Despite this, the relationship of the three genera was supported by sharing the presence of a highly 20 developed adductor mandible externus superficialis muscle at their point of origin (Zaher 1999). The taxonomic history of Pseudoeryx and Hydrops is confusing and complex, mainly because the species and subspecies have been described from qualitative and quantitative variations of a limited number of specimens (Roze 1957b; Mertens 1965; Albuquerque 2000; Albuquerque & Lema 2008). Furthermore, in Pseudoeryx the lack of information on type localities makes it difficult to solve taxonomic problems. The two currently subspecies of P. plicatilis (P. p. plicatilis and P. p. mimeticus) show great overlap in qualitative and quantitative characters and occur in sympatry at some areas (Dixon & Soini 1986; Cunha & Nascimento 1993; Schargel et al. 2007). Such overlaps seems to indicate that the subspecies of P. plicatilis do not represent distinct evolutionary lineages and merely depict artificial clusters of geographic/individual variation (Schargel et al. 2007), although, so far this issue has not been resolved. Variations in hemipenial morphology in three specimens of P. plicatilis were observed and there is a possibility that some populations may represent different species (Zaher 1999). Roze (1957b) performed a taxonomic review of Hydrops and described two subspecies for Hy. martii (Hy. m. martii and Hy. m. callostictus) and six for Hy. triangularis (Hy. t. triangularis, Hy. t. fasciatus, Hy. t. venezuelensis, Hy. t. neglectus, Hy. t. bassleri, Hy. t. bolivianus), based on morphological characters. Almost 50 years later, Albuquerque (2000) and Albuquerque & Lema (2008) reviewed the genus with a larger sample and found that there was a great overlap between the subspecies of both Hy. triangularis and Hy. martii, and placed the subspecies of each taxon in synonymy with the nominal species. As it presents the greatest diversity within the tribe, Helicops presents the greatest taxonomic issues. There is polymorphism of several quantitative and qualitative data, and most variables overlap between several species, both in taxa with restricted and wide distributions (Kawashita-Ribeiro et al. 2013; Moraes-da-Silva et al. 2019, 2021; Costa et al. 2016). Briefly, Rossman (1976) coined the pastazae complex (He. pastazae and He. petersi) and the “polylepis section” (He. pastazae, He. petersi, He. polylepis and He. yacu). Soon after, Rossman and Abe (1979) discussed the possible synonymy of He. yacu with He. pastazae, since two additional specimens of He. yacu were collected from the upper Amazon region that resembled He. pastazae. Rossman (2002) reported that he “believes” that He. hogei would be a junior synonym of He. scalaris based on morphology. From the 2000s until now, there have been five descriptions of new species of Helicops (Kawashita-Ribeiro et al. 2013; Costa et al. 2016; Moraes-da-Silva et al. 2019, 2021, 2022). However, several widespread taxa, such as He. angulatus, He. leopardinus, He. hagmanni and He. polylepis 21 may represent species complexes and have never had their taxonomic status tested. Therefore, the relationships between Helicops taxa need strong investigations considering integrative approaches, and with larger sampling data. Our understanding of snake phylogenetic relationships has significantly improved in recent years due to the advancement and development of molecular genetics (Zaher et al. 2009; Pyron et al. 2013; Figueroa et al. 2016; Zaher et al. 2019). Molecular phylogenies confirmed the monophyly of the tribe, establishing its validity, although the intergeneric and interspecific relationships remain unresolved, given that taxonomic and geographic sampling remains limited (e.g., Moraes-da-Silva et al., 2019, 2021, Murphy et al. 2020). With the recent discoveries of new species of Helicops, Costa et al. (2016) and subsequent works began to use a molecular phylogeny approach to validate such species and verify their phylogenetic positioning. From these phylogenies with low taxonomic and geographic sampling, a phylogenetic perspective on the tribe began to be elucidated. Summarily, in the description of He. nentur, Costa et al. (2016) revealed phylogenetic affinity of this taxa with He. carinicaudus, and which was corroborated by subsequent studies (Moraes-da-Silva et al., 2019, 2021, Murphy et al. 2020). Moraes-da-Silva et al., 2019, 2021 described the species He. boitata and He. phantasma, respectively. Such studies recovered the first as a sister rate of He. carinicaudus and He. nentur, while He. phantasma was closely related to He. leopardinus. Moraes-da-Silva et al. (2019) presented for the first time the close relationship of He. leopardinus with He. infrataeniatus and He. modestus. They still revealed He. polylepis as sister taxa of the clade formed by He. gomesi and He. angulatus. Moraes-da- Silva et al. (2021) also included He. hagmanni in this new study that was recovered as a sister taxon of the clade formed by He. polylepis, He. gomesi and He. angulatus with low support value. Murphy et al. (2020) pointed out that He. angulatus is paraphyletic with respect to He. gomesi and revalidated Helicops cyclops Cope, 1868 to solve the problem. However, the study presents some taxonomic inconsistencies that need to be reviewed. Despite the panorama presented here, molecular phylogenetic studies so far have used a limited sample of the diversity of the tribe, in addition to using only species and samples occurring in Brazil. Evolutionary approach The molecular approach, combined with advances in bioinformatics, have helped to understand the evolution of biota, aiding in the development and knowledge of systematics, taxonomy and biogeography (Posada & Crandall 2001; Hickerson et al. 2010; Burbrink & Gehara 2018). Phylogeographic studies focus on the patterns and processes that determine the geographic distribution of genealogical lineages, trying to correlate, for example, geoclimatic 22 events, geographic distance, topography (historical and contemporary events) with the evolutionary history of lineages (Avise 2012; Knowles 2009). Such studies can help in the delimitation of species, in the recognition of new and cryptic species, as well as in the understanding of events such as extinction, dispersal, vicariance, and the demographic history of a species or population (e.g., Ledo et al. 2017; Dal Vechio et al. 2019; Carvalho et al. 2020; Damasceno et al. 2021). Phylogeographic studies of aquatic snakes from other regions such as Asia and North America have revealed that the evolutionary trajectory of the species does not depend only on geomorphological and climatic processes, but also intrinsic traits of each species. Therefore, species phylogenetically and ecologically related show different phylogeographic patterns (e.g., Guiher & Burbrink 2008; Brandley et al. 2010; Lukoschek et al. 2011; McCartney-Melstad et al. 2012; Carvalho et al. in press). Besides, comparative observations are determined by shared ancestry, and testing evolutionary hypotheses requires focusing on closely related taxa (Pyron 2015). Methods based on estimates of reconstruction of ancestral areas demand dated trees to connect the spatial and temporal dimensions (e.g., Ronquist 1997; Matzke 2014). The history of the uplift of the Andes includes several elevation phases that triggered a long and complex history of the landscape and river systems linked to the evolutionary history of the South American biota (e.g., Wesselingh & Salo 2006; Hoorn et al., 2010a, b). Therefore, geological events and landscape remodeling on a continental scale have a strong influence on dispersal and vicariance processes, which reflects on the distribution patterns of organisms (Lundberg et al. 1998). Among the most important of these events are marine introgressions, the formation of the Pebas and Acre systems, changes in drainage connectivity and fusion, and the formation of the current transcontinental Amazon River drainage (Wesselingh & Salo 2006; Hoorn et al., 2010a, b; Rosa et al. 2003). The dispersal of freshwater species requires corridors of aquatic habitat connecting basins, and the range limits of most of these species coincide with the limits of watersheds, although many others have distributions that extend beyond the limits of watersheds between basins (Albert & Reis 2011). However, for species that are not restricted to the aquatic environment, as in the case of aquatic or semiaquatic snakes, which can also forage on the terrestrial environment, hydrographic limits are not necessarily barriers (Carvalho et al. 2020; Carvalho et al. in press). Furthermore, species with a wide distribution are less likely to show phylogeographic breaks caused by geographic barriers (e.g., Avise 2012) and may still show greater genetic differentiation due to local adaptation (Kremer et al. 2012; Berthouly-Salazar et al. 2013). 23 Despite the panorama presented, the molecular phylogenetic studies presented so far for Hydropsini used a limited sample of the diversity of the tribe, using only species and samples that occur in Brazil. Furthermore, given its widely and narrowly distributed taxa and aquatic habitat, Hydropsini is an excellent model to study the spatial distribution of aquatic snake diversity in the South American region. Therefore, here we address the systematics and biogeography of Hydropsini, presenting for the first time a robust multi loci (two mtDNA, four nuDNA) molecular phylogenetic analysis including 1080 sequences from 22 of the 26 species. We also discuss the origin and diversification of the tribe based on a dated tree and geological events. Main results We have divided the thesis into three chapters. In the first chapter we inferred the phylogenetic relations of Hydropsini, which our results showed that the three genera are valid and we recover the monophyly of the tribe. In addition, we have redefined new groupings based on phylogenetic evidence. We recovered paraphyletic lineages in He. carinicaudus, He. leopardinus complex, He. pastazae and He. angulatus and discussed new relationships and taxonomic issues. Within Helicops, the inclusion of the majority of species had an impact on previous proposed phylogenetic positions. In the second chapter, based on the phylogeny of chapter 1, we performed an ancestral area estimation, in order to investigate the biogeographic history of these aquatic snakes. We reveal that the most likely ancestral scenario for Hydropsini diversification was the Amazon region, around 21 Mya in the early Miocene. The mountain ranges of Colombia and Venezuela acted as a vicariant barrier separating the cis and trans Andean taxa. Furthermore, colonization to adjacent watersheds further to east was also caused by distinct dispersal events. We discuss how landscape dynamism during the Miocene in the Amazon region had a great impact on the diversification of Hydropsini, mainly influenced by the Pebas system. Lastly, in the third chapter we based on the findings for Helicops leopardinus complex (He. leopardinus, He. modestus, He. infrataeniatus and He. tapajonicus) in the first chapter, we performed a phylogeography of this species complex in order to better understand its diversity and genetic structure. Helicops leopardinus complex originated during the Pleistocene (~1.2 Mya). We found that this complex may represent a single species with five genetically structured clusters, which show gene flow unevenly shared among them. The genetic differentiation of the He. leopardinus complex is mainly explained by the interaction of geographic distance (IBD), climatic variation (IBE), and 24 hydrographic basins. We provided new insights about diversification patterns and processes for a species complex of a broadly distributed group of watersnakes along South America. 25 2 CHAPTER 1: Phylogeny and systematic of Hydropsini (Serpentes: Dipsadidae: Xenodontinae) Abstract The Hydropsini tribe is the most diverse group of water snakes in South America with three genera (Pseudoeryx, Hydrops and Helicops) and 26 valid species widely distributed in the continent. However, the relationships between genera and species are not well established. Therefore, we provide the most comprehensive phylogenetic analysis among Pseudoeryx, Hydrops and Helicops. We have two goals in this study: (1) present the relationship of the genera and species within Hydropsini; and (2) evaluate the diversity of lineages in Hydropsini and test whether the current taxonomy reflects the phylogeny of the tribe. We incorporate new data for seven previously unsampled taxa in the Hydropsini, using the highest levels of taxon and geographic sampling, a total of 22 species (85% of known diversity), which allowed us to improve the phylogenetic relationships among genera and species of the tribe. The dataset included 260 samples with two mitochondrial (16S, Cytb) and four nuclear genes (Cmos, NT3, BDNF and R35). Our phylogenetic analyses show that the three genera are valid and corroborated the monophyly of Hydropsini, and recover Pseudoeryx relictualis as a Hydrops. Within Helicops, the inclusion of almost all species of the genus had impact on previous notions of its relationship. We recovered paraphyletic lineages in He. carinicaudus, He. leopardinus complex, He. pastazae and He. angulatus. We recovered He. scalaris within the He. angulatus group. While Helicops danieli and He. scalaris are not closely related. We present the first comprehensive phylogeny of Hydropsini and provide new relationships that will assist in insights into the evolution in reproductive mode, aquatic habit and, a valuable resource for future comparative (phenotypic or ecological) evolutionary processes and taxonomic studies. Keywords: Bayesian Inference, Diversification, Evolution, Hydrobasin, Multi-loci, South America, Taxonomy, Watersnake. 26 2.1 Introduction Phylogenies are reconstructions of evolutionary history, and provide an understanding of relationships between species, as well as providing the historical basis for test ecological inferences and evolutionary processes (de Queiroz & Gauthier 1990; Edwards 2009). The understanding of phylogenetic relationships among snakes has progressed significantly in recent years, mainly due to the use of molecular data (e.g., Pyron et al. 2013; Zaher et al. 2019). Considering all extant snake species (~3.971 species), about 14% of them have an aquatic or semiaquatic habit, and inhabit freshwater ecosystems and the greatest diversity of freshwater snakes is located in the Oriental and Neotropical regions (Pauwels et al. 2008; Murphy 2012). About half of all aquatic Neotropical species belongs to the Hydropsini Dowling 1975 tribe, a clade currently composed by 26 snakes allocated in three genera (Pseudoeryx, Hydrops and Helicops) distributed throughout South America with cis-andine (23 species) and trans- andine (three spp) species (Moraes-da-Silva et al. 2019; 2021; 2022; Citeli et al. 2022). Helicops Wagler (1828) represents the most diverse and widely distributed genus within the tribe with 21 species. Hydrops Wagler (1830) and Pseudoeryx Fitzinger (1826) contain three and two species, respectively, and occur mainly in the western region of the continent. Species within this tribe have highly morphological specializations that include position of nostrils on the snout top and dorsolaterally oriented eyes (Roze 1957a; b). Within the tribe, Helicops differs from Hydrops and Pseudoeryx, among other characteristics, by the presence of keels on the dorsal scales, absent in both last taxa, while diastema is present in Helicops and Hydrops and absent in Pseudoeryx (Roze 1957 a; b). Therefore, the systematics of the tribe has been historically based on phenotypic traits, which were later shown to lack phylogenetic support (Costa et al. 2016; Moraes-da-Silva et al. 2019; 2021). Furthermore, the evolution of the reproductive mode within Hydropsini is intriguing, given that there are oviparous and viviparous species, and even Helicops angulatus (Linnaeus, 1758) presents populations with different reproductive modes, that is, it is a species with reproductive bimodality with allopatric distribution, oviparous populations are distributed from north to east-central and northeast, while viviparous populations occur from northwest to west-central South America (Braz et al. 2016; 2018). Ancestral state reconstructions suggest that oviparity is plesiomorphic in Hydropsini, while Hydrops and Pseudoeryx are exclusively oviparous (Braz et al. 2016). Nonetheless, the reproductive mode of some species is still unknown (i.e., He. boitata and He. petersi) (Braz et al. 2016; 2018; Moraes-da-Silva e al. 2022). 27 The knowledge of the diversity in Hydropsini has grown remarkably in recent years, with an increase of species descriptions (e.g., Scrocchi et al. 2005; Schargel et al. 2007; Costa et al. 2016; Moraes-da-Silva et al. 2019; 2021). Pseudoeryx possess two valid species, which P. plicatilis has wide distribution in the Amazon region, Pantanal and also in the Chaco of northern Argentina and Paraguay. Pseudoeryx remained for a long time a monospecific genus, when Schargel et al. (2007) described P. relictualis, endemic to Lake Maracaibo, Venezuela, which is geographically isolated from other plains by the Cordillera de Mérida to the east and south and the Sierra de Perijá to the west. Roze (1957a) highlighted the need for a detailed review of Pseudoeryx, considering the possibility of P. plicatilis harboring two taxa, which P. mimeticus Cope, 1885 could represent a valid taxon, nonetheless it has not been done so far. However, Hoge and Nina (1960–1962), based mainly on ventral scales and coloration pattern, superficially allocated P. mimeticus as a subspecies of P. plicatilis (Linnaeus, 1758). Later, Mertens (1965) described P. p. ecuadorensis (SMF 60435, female) from a single specimen to Ecuador without a specific type locality. However, Dixon and Soini (1986) synonymized it with P. p. mimeticus given that it fits the morphological variation of the latter. Other studies have pointed out several intermediate characters and overlap between P. p. plicatilis and P. p. mimeticus and highlighted that both subspecies lack a taxonomic and geographic resolution (Dixon & Soini 1986; Cunha & Nascimento 1993; Schargel et al. 2007). On Hydrops, Roze (1957b) revised Hy. martii (Wagler, 1824) and Hy. triangularis (Wagler, 1824), and separated both species into two and six subspecies, respectively, based on scale counts, color pattern and geographic distribution. Years later, Albuquerque (2000) and Albuquerque and Lema (2008), supported by a larger sampling, synonymized all subspecies of the genus due to the large overlap of qualitative (i.e., color pattern) and quantitative (i.e., scale count) characters, and raised the subspecies to nominal species. Furthermore, Scrocchi et al. (2005) described Hy. caesurus, an endemic species from the La Plata basin (Argentina, Paraguay and Brazil), while both Hy. triangularis and Hy. martii occur in North South America, predominantly in the Amazon basin. Lastly, Helicops is the most diverse, widespread genus within the Hydropsini tribe, and still, has more species with wide distribution, arising questions on species taxonomy (Rossman 1976; Rossman & Abe 1979; Kawashita-Ribeiro et al. 2013; Moraes-da-Silva et al. 2019; 2021; 2022; Murphy et al. 2020). The knowledge of the diversity in Helicops has grown with an increase of news species description, since species that were described from chance encounters to species that were already suspected of their taxonomic status (e.g., Kawashita-Ribeiro et al. 2013; Costa et al. 2016; Moraes-da-Silva et al. 2019; 2021; 2022). In addition, some taxa have 28 had their taxonomic status contested, for example, Rossman and Abe (1979) assumed that He. yacu could be a synonym of Helicops pastazae Shreve, 1934 because they share several morphological characters, but he refrains to do it, and preferred to wait to analyze a larger number of specimens of the two species. In addition to the taxonomic issues on Helicops, is always recovered a polytomy among He. leopardinus, He. infrataeniatus and He. modestus (Moraes-da-Silva et al. 2021). Furthermore, recent molecular phylogenies have also found He. angulatus paraphyletic (Vidal et al. 2010; Murphy et al. 2020). Helicops cyclops Cope, 1868 was removed from the synonymy of He. angulatus (Murphy et al. 2020). Possibly, may have caused more confusion to the history of this species, given that the authors did not carry out a significant taxonomic revision of this species that shows a continental occurrence. Commonly, species with a wide distribution may present a complex and/or cryptic species (e.g., Geurgas & Rodrigues 2010; Nunes et al. 2012; Ruane et al. 2014). Helicops angulatus, He. leopardinus (Schlegel, 1837), He. polylepis Günther, 1861 and He. hagmanni Roux, 1910 had their taxonomic status questioned because they present wide occurrence and are difficult to delimit based on certain morphological characters, making the systematics of these taxa problematic (Kawashita-Ribeiro et al. 2013; Moraes-da-Silva et al. 2019). Therefore, within Helicops, taxonomic problems range from species complexes to multiple polymorphic or polychromatic species that may represent a single taxon (e.g., Rossman 1976; Rossman & Abe 1979; Rossman 2002a). Consequently, in order to know more precisely the phylogenetic relationships of the species, it is important to understand the real diversity, and to sample the entire distribution of their taxa in a phylogeny. Moreover, we can provide answers about their taxonomic scenario, which is essential to elucidate their phylogenetic relationships. Phylogenetic relationship hypotheses were conflicting within the tribe even before molecular data was available (e. g., Roze 1957a; b; Rossman 1973). Roze (1957a; b) was the first to recognize that Pseudoeryx, Hydrops and Helicops had a close relationship, pointing out that these genera shared, as he mentioned, morphological synapomorphies such as: nostrils located in a dorsal position and a single internasal scale, therefore, considered it to be an adaptation to aquatic habit. However, he did not perform any explicit phylogenetic analysis. Furthermore, Roze (1957a; b) proposed Pseudoeryx as the outgroup among the three genera due to dentition without diastema, no-banded color pattern, and lack of dorsal scale reduction (15-15-15). While the presence of keels and the greater number of rows of dorsal scales in Helicops are characterized as the most derived genus among the three hydropsins. Based on hemipenial characters, Rossman (1973) disagreed that these genera could have some type of phylogenetic proximity, and even concluded that the nostrils and eyes in the dorsal 29 position would be an evolutionary convergence. The hemipenial morphology effectively caused confusion in the systematics of Helicops, Hydrops, and Pseudoeryx (Roze 1957a; Rossman 1973). In fact, the hemipenis of Hydrops and Pseudoeryx have notable morphological differences from any other xenodontines, including Helicops, which made it difficult to establish phylogenetic relationships and, therefore, they were often allocated to separate groups (Roze 1957a; Rossman 1973). In the study on hemipenis of Xenodontinae, Zaher (1999) considered the tribe monophyletic because Pseudoeryx and Hydrops share with Helicops (a clear member of the Xenodontinae) the adductor muscle (externus superficialis) of the mandible well developed at its point of origin. Furthermore, he attributed these hemipenial differences to a secondary loss in Pseudoeryx and Hydrops. Most molecular phylogenies focused on understanding suprageneric relationships have always recovered Hydropsini as monophyletic (e.g., Vidal et al. 2000; Zaher et al. 2009; Grazziotin et al. 2012; Pyron et al. 2013; Figueroa et al. 2016). In some topologies, Pseudoeryx and Hydrops form a clade, which is the sister group of Helicops (Vidal et al. 2000; Pyron et al. 2013; Figueroa et al. 2016; Zaher et al. 2019); moreover, it was observed Pseudoeryx and Hydrops representing two successive outgroups for Helicops (Zaher et al. 2009; Grazziotin et al. 2012), or Hydrops more closely related to Helicops, and Pseudoeryx more distantly (Pyron et al. 2011). Rossman (1976), using morphological data (qualitative and quantitative), reviewed the “pastazae complex” (He. pastazae, He. petersi), described He. petersi, and even suggested a group called “polylepis section” composed of: He. pastazae, He. petersi, He. polylepis and He. yacu, which he claims to be closely related. The affinities of He. hagmanni were considered uncertain even with non-bright scales and a large number of rows with the spotted pattern, as evidence that it could be closely related to the polylepis section, but characteristics such as the maxillary dentition, relatively short tail and number of extremely low subcaudals were sufficient for Rossman (1976) to consider not including the section. The most complete phylogenetic approach of the tribe until now included 13 Hydropsini taxa (Moraes-da-Silva et al. 2021), which comprises species occurring mainly in Brazil. The phylogenetic relationships between the genera and species of Hydropsini remain unclear, and a more complete sample, including more taxa, might help to clarify the phylogenetic relationship of the tribe. A phylogeny with large-scale taxon sampling is invaluable for comparative evolutionary studies and for approaching its taxonomy (e.g., Torres-Carvajal et al. 2017; Arredondo et al. 2020; Bernstein et al. 2021; Melo-Sampaio et al. 2021). Despite progress in understanding the evolutionary history of snakes, species with aquatic habits do not receive much scientific attention, and studies in several fields are still extremely scarce (Carvalho et al. 30 2020). Is widely justified the need for large-scale taxon sampling molecular phylogeny to better understand the relationships among species within Hydropsini. In this study we have two main goals: (1) infer and propose hypotheses on the phylogenetic relationships of the genera and species within Hydropsini based on updated geographic, molecular and taxonomic sampling; and (2) evaluate the diversity of lineages in Hydropsini and explore whether relationships between species represent their current taxonomy. Furthermore, we aim to build a phylogenetic relation of the tribe through molecular approach, and present the first large-scale phylogenetic estimate for Hydropsini. 2.2 Material and Methods Taxon sampling and data acquisition We sampled 13 individuals of Pseudoeryx, 12 of Hydrops and 215 of Helicops from 210 localities throughout South America (Fig. 1; Table S1). We generated 1012 novel DNA sequences from 221 specimens representing 22 species of Hydropsini. We also obtained sequences of Hydropsini from GenBank, as well as outgroup species (Table S1) and we used the species Coluber constrictor for rooted the tree. Our final dataset has 260 terminals, including 20 outgroup species (Table S2), from all currently recognized tribes of the Xenodontinae, besides some species of Dipsadidae and Colubridae (sensu Grazziotin et al. 2012). We included 22 species of the 26 recognized and for the first time, the species Pseudoeryx relictualis, Hydrops caesurus, Helicops danieli, He. pastazae, He. scalaris, He. trivittatus and He. yacu have their phylogenetic relationships evaluated. It was not possible to sample He. apiaka, He. petersi and He. acangussu. We do not consider He. cyclops given taxonomic issues (see Discussion for more information). Molecular data We extracted the DNA from muscle, liver or scales, and we used the phenol-chloroform extraction protocol (Sambrook et al. 1989). We amplified the partial sequences of the two mitochondrial genes (mtDNA): 16S ribosomal (16S) and Cytochrome b (Cytb); as well as four nuclear gene (nuDNA): Oocyte maturation factor Mos (Cmos), Neurotrophin-3 (NT3), Brain- Derived Neurotrophic Factor Precursor (BDNF) and RNA fingerprint protein 35 (R35) using the standard Polymerase Chain Reaction (PCR) technique. PCR amplification protocols and primers are presented in supplementary material 1. We aligned each locus using the Muscle algorithm (Edgar, 2004) in Geneious v9.1.8 (Geneious 2022; https://www.geneious.com). 31 Figure 1. Sampling coverage used in this study for Pseudoeryx, Hydrops, and Helicops throughout South America. Inferring phylogenetic relationships We used PartitionFinder 2 to identify partitioning schemes and the most appropriate nucleotide substitution models (Lanfear et al. 2017). According to our concatenated alignment, we found eight partitions evaluated by BIC (Table S3). To test phylogenetic relationships among species of Hydropsini, we used Bayesian Inference (BI) and Maximum Likelihood (ML) methods. We performed a BI analysis using BEAST v.2.6.3 (Bouckaert et al. 2019) with a Yule process tree prior, a strict clock model, and we used rate of 1.3% substitutions per million years in Cytb that has been previously shown in closely related Dipsadidae snakes (Daza et al. 2009). 32 We ran two analyses for 100 million generations, sampling every 10,000 steps. We checked for stationarity by visually inspecting trace plots and ensuring that all ESS values were above 200 in Tracer v1.7.1 (Rambaut et al. 2018). We combined both runs in LogCombiner v2.6.3 (with a 10% burn-in) allowing us to summarize the results in a maximum clade credibility tree in TreeAnnotator v2.6.3. We implemented a Maximum Likelihood tree inferred in RAxML (Stamatakis 2014) via raxmlGUI 2.0 (Edler et al. 2021) with partitions restricted to GTR-based models and using GTRGAMMA, with rapid bootstrap method and 1000 bootstrap iterations. 2.3 Results The final alignment of our sequences (including gaps) was 426pb for 16S, 851pb for Cytb, 541pb for Cmos, 584pb for NT3, 646pb for BDNF, 633pb for R35. The length of the concatenated alignment was 3581pb. Both BI and ML trees showed a similar topology, except by a clade (see below), with the relationships of most clades well supported, and nodes with high posterior probabilities (PP) > 0.95 and high bootstrap values (BS) > 70% (Fig. 2 and Fig. S1). Hydropsini was recovered as monophyletic with maximum support in both analyses. We recovered Pseudoeryx and Hydrops as sister taxa, with Helicops as sister clade of both. Pseudoeryx was recovered as paraphyletic, Pseudoeryx relictualis was recovered as sister to all other Hydrops. Besides, our analyzes recovered two well established lineages for P. plicatilis with maximum support. The monophyly of Hydrops was well supported in BI (PP=1), and we found two lineages within Hy. triangularis. Lastly, the diverse genus Helicops is monophyletic with robust support values (PP=1; BS=100%) (Fig. 2). Pseudoeryx relictualis was recovered as sister taxa of the remaining Hydrops species in BI (PP=0.94), and as sister species to all other Hydrops species (Fig. 3; Fig S2). To better represent the monophyly of this clade, we therefore propose the relocation of P. relictualis within the genus Hydrops: Hydrops relictualis comb. nov. We emphasize that a taxonomic review is extremely important to understand how to diagnose this species within Hydrops. We recovered Hy. caesurus as sister taxa of Hy. martii and Hy. triangularis (Fig. 3; Fig S2). We recovered two lineages well supported for Hydrops triangularis (PP=1; BS=98%). The L1 from Ecuador, Peru and Brazil and the L2 from French Guiana and Brazil (Fig. 3; Fig S2). 33 Figure 2. Bayesian phylogenetic inference of Hydropsini based on mtDNA (16S and Cytb) and nuDNA (Cmos, NT3, BDNF, and R35) genes using BEAST. Posterior probabilities (PP) and bootstrap (BS) values are ≥ 0.95 and ≥ 70%, above and below, respectively (we present only well-supported node values). Asterisks (*) indicate different phylogenetic relationships found in the ML tree. Clade A: Hydropsini tribe; Clade B: Pseudoeryx; Clade C: Hydrops; Clade D: P. relictualis; Clade E: Hy. caesurus; Clade F: Hy. martii; Clade G: Hy. triangularis; Clade H: Helicops; Clade I: He. leopardinus group; Clade J: He. leopardinus complex; Clade K: He. trivittatus; Clade L: He. carinicaudus group; Clade M: He. danieli; Clade N: He. hagmanni group; Clade O: He. pastazae complex; Clade P: He. polylepis; Clade Q: He. angulatus group. Photos: DJSantana (P. plicatilis, He. angulatus); FJMRojas- Runjac (P. relictualis); AACerqueira Junior (Hy. casesurus); Moraes-da-Silva et al. 2019 (He. boitata) and CHdeONogueira (He. carinicaudus L1). Within Helicops, we recovered two clades in BI: we named the first clade as He. leopardinus group with He. phantasma as sister species from the clade formed by, He. leopardinus, He. infrataeniatus, He. modestus and He. tapajonicus, which were all paraphyletic. Within this later clade, we name it as He. leopardinus complex (He. tapajonicus, He. modestus, He. infrataeniatus and He. leopardinus), which all species are paraphyletic/polyphyletic (Fig. 4A; Fig S3). While the second clade by BI, we recovered He. trivittatus as a sister to He. carinicaudus group with moderate support value (PP=0.90). In ML, 34 excepted for H. leopardinus group, He. trivittatus was recovered as sister taxa of all other species with a well-supported node value (BS=87%) (Fig. 4B; Fig. S4). Figure 3. Bayesian phylogenetic inference of Pseudoeryx and Hydrops based on mtDNA (16S and Cytb) and nuDNA (Cmos, NT3, BDNF, and R35) genes using BEAST. Posterior probabilities (PP) values are ≥0.95 (we present only well-supported node values). Clade B: Pseudoeryx; Clade C: Hydrops; Clade D: Hy. relictualis; Clade E: Hy. caesurus; Clade F: Hy. martii; Clade G: Hy. triangularis. L: indicates distinct lineages. The He. carinicaudus group, which is formed by He. boitata, as sister species of the clade conformed by, He. carinicaudus and He. nentur. In our topology, He. carinicaudus species is paraphyletic, with three lineages and the nominal species belongs to the Lineage 1 occurring in Espírito Santo and northern Rio de Janeiro (Fig. 4B; Fig. S4). Helicops nentur is the sister species of the other two lineages of He. carinicaudus (L2 and L3), which may represent distinct evolutionary lineages, then, we consider here as candidate species. All clades within this group have their nodes well supported (PP≥0.99; BS ≥91%). He. danieli was recovered as a sister species of the He. hagmanni group, He. angulatus group and He. polylepis with moderate support values (PP=0.91; Fig. 2). The sole topological difference between the BI and ML trees was the placement of He. polylepis. In BI (Fig. 2 and 5), He. polylepis is the sister taxa of the He. angulatus group (PP=0.90), while in ML (Fig. S1) it is the sister taxa of He. hagmanni group (BS=87%). The recognized here He. hagmanni group, which was recovered with high support values in both analyses (PP=1.00; BS=91%; Fig. 35 2 and 4C), including He. hagmanni, He. yacu, He. pastazae, with He. hagmanni as sister taxon of the remaining species in the group. Besides, we considered the sample of He. hagmanni (IAVH-CT21340) as a candidate species. In addition, we recovered He. yacu closely related to He. pastazae and belongs to He. pastazae complex (Fig. 4C; Fig. S5). The Helicops angulatus group is composed by two clades (Fig. 5; Fig. S6). The first clade with He. scalaris as sister species of the He. angulatus from northern South America (He. angulatus Lineage 1). The second clade within this group is formed by three remaining lineages of He. angulatus (L2, L3 and L4) and He. gomesi. However, the three He. angulatus lineages are paraphyletic related to He. gomesi. 36 Figure 4. Bayesian phylogenetic inferences based on mtDNA (16S and Cytb) and nuDNA (Cmos, NT3, BDNF, and R35) genes using BEAST. Posterior probabilities (PP) values are ≥ 0.95 (we present only well-supported node values). (A) Helicops leopardinus group and He. leopardinus complex. Clade I: He. leopardinus group; Clade J: He. leopardinus complex. (B) Helicops trivittatus and He. carinicaudus group. Clade K: He. trivittatus; Clade L: He. carinicaudus group. (C) Helicops danieli, He. hagmanni group and He. pastazae complex. Clade M: He. danieli; Clade N: He. hagmanni group; Clade O: He. pastazae complex. 37 Figure 5. Bayesian phylogenetic inference of He. polylepis, He. angulatus group and He. pastazae complex based on mtDNA (16S and Cytb) and nuDNA (Cmos, NT3, BDNF, and R35) genes using BEAST. Posterior probabilities (PP) values are ≥0.95 (we present only well-supported node values). Clade P: He. polylepis; Clade Q: He. angulatus group. L: indicates distinct lineages. 38 2.4. Discussion Tribe monophyly and the relationship among the genera Hydropsini is robustly supported as monophyletic (PP=1.00; BS=98%) (Fig. 2 and Fig. S1). Here, we generated 1.012 novel DNA sequences from 221 specimens belonging to Hydropsini with a wide geographic coverage. We emphasize that this is the first approach that have tested the phylogenetic relationship of Hydrops caesurus, Hy. relictualis, Helicops danieli, He. pastazae, He. scalaris, He. trivittatus and He. yacu. We added for the first time exclusively trans-Andean lineages of Helicops species (He. danieli and He. scalaris). Our tree has 85% of the known diversity of the tribe (22 of the 26 recognized species). It was not possible to sample He. acangussu, He. apiaka, He. hogei and He. petersi, since DNA sample of these species were not available. We consider He. cyclops as synonym of He. angulatus (see Discussion below). Previous molecular studies focusing on supra-generic relationships have demonstrated the monophyly of the Hydropsini, confirming the hypothesis presented by Roze (1957b). However, such studies had a limited species sampling, which makes it difficult to taxonomically identify problematic species, cryptic species and phylogenetic relationships (e.g., Vidal et al. 2000; Zaher et al. 2009; Pyron et al. 2011; 2013; Figueroa et al. 2016). A recent study about the phylogeny of mud snakes (Homalopsidae) with a large taxon and geographic sampling have observed an undescribed diversity and, conversely, several indistinct species under different names (Bernstein et al. 2021). Even though, the Hydropsini monophyly was always unquestionable. Notwithstanding, the relationship between Helicops, Hydrops and Pseudoeryx have been uncertain. For example, Vidal et al. (2000) using mitochondrial DNA sequences recovered Helicops as sister to a clade Hydrops + Pseudoeryx. However, our inference suggests that Pseudoeryx and Hydrops sister taxa, which form a clade that are the sister group of Helicops, a relationship most commonly recovered in other phylogenies as Pyron et al 2013; Figueroa et al. 2016; Zaher et al. 2019; Murphy et al. 2020; Moraes-da-Silva et al. 2021. Zaher (1999) considers that Hydrops and Pseudoeryx do not have two derived hemipenial character states, which is present in all Xenodontinae and also in Helicops, confirming the close relationship between Hydrops and Pseudoeryx. On the other hand, Roze (1957b) proposed that characters such as the presence of a diastema, scales, and color pattern place Hydrops and Helicops more closely related. However, these characteristics pointed out by Roze (1957b) do not support the relationship of a clade formed by Helicops and Hydrops. Therefore, when comparing molecular and morphological inference results, the morphological data for this tribe 39 are incongruent due to the various polymorphic and/or homoplastic characteristics, as they possibly undergo a lot of selective pressure (e.g., Moraes-da-Silva et al. 2019; 2021). Hydropsini Dowling, 1975 (Clade A; Fig. 2) Type-genus: Hydrops Wagler, 1830. Content: (three genera). Pseudoeryx Fitzinger, 1826; Hydrops Wagler, 1830; Helicops Wagler, 1828. Pseudoeryx (Clade B; Fig. 2 and 3) Content: (one species). Pseudoeryx plicatilis (Linnaeus, 1758). Comments: We recovered two lineages for P. plicatilis, which is currently divided into two subspecies (Fig. 3; Fig. S2). Linnaeus described Pseudoeryx plicatilis with the wrong type locality (Ternataeis). Based on a single specimen Cope (1885) described P. mimeticus from Marmoré, Bolivia. Ten years later, Bocourt (1895) described the subspecies Pseudoeryx p. anomalepis (MNHN-RA-0.6165), based on a single specimen from Colombia, according to the original description. Although on the MNHN website (https://www.mnhn.fr/fr/reptiles) inform that the type material of P. p. anomalepis was collected in Napo, Colombia, they still report a second type material (MNHN-RA-1881.519, collection date 1881-12-01). Possibly, M. Wiener collected these two specimens around 1880 and Bocourt used only one to describe P. p. anomalepis. The type locality of this subspecies is imprecise. Bocourt informed only that it is from Colombia, while the MNHN website informs Napo, Colombia. Currently, the Napo department is located in Ecuador and there is no province or department in Colombia named Napo. However, there was a possible confusion between Colombia and Gran-Colombia by the author. Gran Colombia was a country established in 1819 with the union of Venezuela, New Granada, Panama and Ecuador, and was dissolved in 1831, and therefore, with this consolidation Napo belonged to that country. Although at the time of collection (~1880) Gran- Colombia had already been dissolved and the countries (Venezuela, Colombia, Ecuador) were arranged as currently known. Roze (1957a) points out that apparently P. p. anomalepis would be a valid subspecies, but the genus needed a revision. Roze (1957a) stated that two valid species of Pseudoeryx may exist: P. plicatilis and P. mimeticus, but a detailed taxonomic revision study would be necessary. Mertens (1965) described P. p. ecuadorensis based on a single specimen from Ecuador without an exact type locality. The author made a brief taxonomic review of the genus dealing with taxonomic issues involving the other subspecies, besides he synonymized P. p. anomalepis with P. p. plicatilis. Mertens analyzed five 40 individuals from Brazil and Paraguay and four specimens from British Guyana from the British Museum via Boulenger 1894: 186 of P. p. plicatilis. Mertens, based on these nine specimens of P. p. plicatilis and on the plates of Seba (1734: 92, Pl. 57, Fig. 4B) and Jan and Sordelli (1868: 7, Pl. 5 Fig. 2–3; SMF34576, Brazil) considered the male P. plicatilis type from ventral (131) and subcaudal (47) scales. Likewise, Mertens separates the three subspecies through chromatic pattern and geographic distribution, P. p. plicatilis occurring from northern and eastern South America (from Colombia to northern Argentina). On the other hand, the other subspecies inhabited only areas east of the Andes, P. p. mimeticus occurred only in Bolivia and P. p. ecuadorensis only in Ecuador. Furthermore, Mertens considered that it would be very unlikely that these three subspecies would occur in sympatry, and that it would be impossible that these subspecies would be connected by intergrades. However, he based his geographic decisions on a small sampling (nine individuals). With this scenario in place, Mertens (1965) designated that the type locality of P. p. plicatilis is Suriname (“Für Coluber plicatilis LINNAEUS wird als terra typica Holländisch - Guayana designiert”), although many subsequent works have not observed such information (e.g., Peters & Orejas-Miranda 1970; Dixon & Soini 1986; Cunha & Nascimento 1993; Wallach et al. 2014). However, we emphasize that these geographic and phenotypic limits are not corroborated on the basis of molecular data. The overlap of quantitative and qualitative characters among Pseudoeryx subspecies has already been observed in several studies, which makes it difficult to establish geographic and nomenclatural limits (e.g., Hoge & Nina 1960– 1962; Dixon & Soini 1986). Because of this, Dixon and Soini (1986) synonymized P. p. ecuadorensis with P. p. mimeticus. For these reasons, we decided to call Pseudoeryx plicatilis lineage 1 (L1) and Pseudoeryx plicatilis lineage 2 (L2). Pseudoeryx L2 has a distribution near to Suriname, it occurs from French Guiana to Peru, so therefore what Mertens defined does not work (Fig. 3; Fig. S2). Thereby, we will not consider Suriname the type locality of P. p. plicatilis and we emphasize that a taxonomic and systematic review is essential, mainly on the type locality, in order to reveal which of the lineages is really the nominal species and which of the subspecies could be elevated to species level. All this confusion of various subspecies was caused by color pattern and low sampling size. The authors who described the subspecies relied only on a single specimen. A large sample size is critical for integrative and systematic taxonomy. Even our study, being the largest to date, using 12 specimens from 10 different locations, was not possible to resolve all these taxonomic issues. 41 Hydrops (Clade C; Fig. 2 and 3) Content: (four species). Hydrops caesurus Scrocchi, Ferreira, Giraudo, Avila and Motte, 2005; Hy. martii (Wagler, 1824); Hy. relictualis comb. nov. (Schargel, Rivas-Fuenmayor, Barros, Péfaur and Navarette, 2007); Hy. triangularis (Wagler, 1824). Generic replacement of Pseudoeryx relictualis (Clade D; Fig. 2 and 3) Comments: According to our topology, the species Pseudoeryx relictualis is not monophyletic, and given its position, we decided to relocate Pseudoeryx relictualis to the genus Hydrops (Fig. 3). Therefore, named here Hy. relictualis comb. nov. In addition, we highlight some morphological characters present in literature that are shared between species of Hydrops, and that are not found in Pseudoeryx. Hydrops relictualis presents the last two maxillary teeth considerably larger and separated from the others by a diastema and nuchal collar (Schargel et al. 2007). Probably the bandless color pattern was decisive for Schargel et al. (2007) discarding the possibility of being a species belonging to the genus Hydrops. In their study, Schargel et al. (2007) compare P. plicatilis and Hy. relictualis using various coloring characters. In addition, the authors point out that the diastema and the nuchal collar are the two main differences between the species, but such characteristics are precisely observed in Hydrops (Roze 1957b; Albuquerque 2000; Albuquerque & Lema 2008). The hemipenis of Hy. relictualis was not described, since Pseudoeryx and Hydrops have very distinct hemipenial morphologies, this phenotypic evidence may reinforce the relocation of Hy. relictualis within Hydrops. It is worth mentioning that, previously, Hydrops was already synonymized with Pseudoeryx by Bocourt (1895). Even though they are two very different genera, Dunn (1944) described Hy. lehmanni based on a single specimen, which Roze (1957a) later identified as P. plicatilis. Hydrops caesurus and Hydrops martii (Clade E and F; Fig. 2 and 3) We recovered Hy. caesurus and Hy. martii two successive outgroups to Hy. triangularis, a monophyletic clade with maximum support in both analyses. This is the first study that includes Hy. caesurus in a phylogeny. The species was previously called Hy. t. bolivianus, until its formal description as a new species by Scrocchi et al. (2005). Hydrops caesurus is more closely related to Hy. martii than Hy. triangularis. The taxonomic history of Hydrops was confusing mainly because the species and subspecies were described from variations in polymorphic characters of a limited number of specimens (Roze 1957b). Subsequent studies have managed to solve in some way the taxonomic enigma of Hydrops (Albuquerque 2000; Albuquerque & Lema 2008). We reinforce the results of previous studies on the validity of the species Hy. martii and Hy. triangularis (Albuquerque 2000; Albuquerque & Lema 2008). Both species have 42 widely overlapped geographic distributions, and are often found in syntopy (von May et al. 2019). Despite this, our results show that Hy. martii and Hy. triangularis are valid species, and we highlight that Hy. triangularis harbors more than one species under this name. Hydrops triangularis (Clade G; Fig. 2 and 3) Comments: We recovered two well-supported lineages for Hydrops triangularis. Since the type locality of Hy. triangularis is in Lake Tefé, Rio Amazonas, Amazonas, Brazil, it was not possible to designate which lineage it belongs due to the gap in our sampling for the state of Amazonas. Hydrops triangularis L1 occurs in the eastern region of the Amazon and Hy. triangularis L2 is found in the western region (Fig. 3; Fig. S2). The nearest distance from the type locality within our data (eastern and western population) is 1000 and 580 km, respectively. Since we do not have topotypes of the subspecies named by Roze (1957b), it was unreasonable to indicate which possible taxa can be considered for the unnamed lineage. We have a sample (LSUMZ-H 3105) from Loreto, Peru that is ~125 km southwest of the type locality of Hy. t. bassleri, which represents the western clade of Hy. triangularis and another sample MZUSP 19497 from UHE Jirau, Rondônia which is ~300 km southwest of the type locality of Hy. t. bolivianus representing the eastern clade of Hy. triangularis. Helicops (Clade H; Fig. 2, 4, and 5) Content: (22 species). He. acangussu Moraes-da-Silva, Walterman, Citeli, Sales-Nunes and Curcio, 2022; He. angulatus (Linnaeus, 1758); He. apiaka Kawashita-Ribeiro, Ávila and Morais, 2013; He. boitata Moraes-da-Silva, Amaro, Sales-Nunes, Strüssmann, Teixeira, Andrade, Sudré, Recoder, Rodrigues and Curcio, 2019; He. carinicaudus (Wied-Neuwied, 1824); He. danieli Amaral, 1938; He. gomesi Amaral, 1922; He. hagmanni Roux, 1910; He. hogei Lancini 1964; He. infrataeniatus Jan, 1865; He. leopardinus (Schlegel, 1837); He. modestus Günther, 1861; He. nentur Costa, Santana, Leal, Koroiva and Garcia, 2016; He. pastazae Shreve, 1934; He. petersi Rossman, 1976; He. phantasma Moraes-da-Silva, Amaro, Sales-Nunes, Rodrigues and Curcio, 2021; He. polylepis Günther, 1861; He. scalaris Jan, 1865; He. tapajonicus Frota, 2005; He. trivittatus (Gray, 1849); He. yacu Rossman and Dixon, 1975. Comments: Given the great diversity of species within this genus, several clades strongly supported, and aiming to indicate a better phylogenetically supported taxonomic organization, we allocated the species into four groups and two complexes as follows: Helicops leopardinus group, Helicops carinicaudus group, Helicops hagmanni group, Helicops angulatus group. 43 Furthermore, Helicops trivittatus, Helicops danieli and Helicops polylepis were not assigned to any species group. The Helicops leopardinus group (Clade I; Fig. 2 and 4A) Content: (five species). He. infrataeniatus Jan, 1865; He. leopardinus (Schlegel, 1837); He. modestus Günther, 1861; He. phantasma Moraes-da-Silva, Amaro, Sales-Nunes, Rodrigues and Curcio, 2021 and He. tapajonicus Da Frota, 2005. Comments: Our topologies revealed that Helicops with a green back color pattern (smooth or striated) do not form a natural group, which corroborates previous molecular studies (Moraes- da-Silva et al. 2019; 2021). Overall, the color pattern within Hydropsini needs to be carefully analyzed, as it is a polychromatic characteristic. Besides, our results corroborate the findings of Moraes-da-Silva et al. (2021), which recovered He. phantasma as sister taxa of the remaining species of the He. leopardinus complex. The Helicops leopardinus complex (Clade J; Fig. 2 and 4A) Our analyzes revealed that the species He. modestus, He. infrataeniatus, He. tapajonicus and He. leopardinus formed one clade, and we decided named He. leopardinus complex. None of those species was recovered as monophyletic, which could represent a single high polymorphic lineage (i.e., a single species) (Fig. 4A; Fig. S3. However, due to the long now history of the names available for these species, a proper phylogeographic study is paramount to understand the evolutionary history and, consequently, the species taxonomy. To better understand the paraphyly of these species, an integrative taxonomic review must be done to clarify the status of the species within this clade. Helicops leopardinus has priority over the other names, however the type locality of He. leopardinus is unknown, whereby we decide to not make any taxonomic decision, until further information appears. Previously, it was pointed out that the species Helicops leopardinus presented a paraphyletic/polyphyletic group (Moraes-da-Silva 2019; 2021), something we also found in our analysis. In the description of He. tapajonicus (Frota 2005), was pointed out that the resembled with He. leopardinus in the number of dorsal scales, ventral scales and teeth, but which had smaller lobes on the hemipenis and a very distinct color pattern (no spotted). Still, He. tapajonicus is sympatric and syntopic with He. leopardinus, occurring in the same microhabitat in the municipalities of Belterra and Santarém, Pará, Brazil. Thus, possibly the different coloration patterns (polychromatism) may be the only characteristic that differentiates these species. Because it has such plasticity, the color pattern has to be used 44 with caution in this tribe to delimit species. We emphasize the importance of perform phylogeography study of He. leopardinus complex to understand their evolutionary history. Helicops trivittatus (Gray, 1849) (Clade K; Fig. 2 and 4B) Comments: The phylogenetic relationship of He. trivittatus had not been tested before. Helicops trivittatus was described by Gray (1849) as Myron Gray 1849, a genus of water snake found in Asia in the family Homalopsidae, due to incorrect location information. In fact, several species of this family considerably resemble Helicops in external morphology due to evolutionary convergence exerted by the aquatic environment (Rossman 2010). The Helicops carinicaudus group (Clade L; Fig. 2 and 4B) Content: (three species). He. boitata Moraes-da-Silva, Amaro, Sales-Nunes, Strüssmann, Teixeira, Andrade, Sudré, Recoder, Rodrigues and Curcio, 2019; He. carinicaudus (Wied- Neuwied, 1824); He. nentur Costa, Santana, Leal, Koroiva and Garcia, 2016. Comments: Helicops boitata was recovered as a sister species of He. carinicaudus and He. nentur with well support, in concordance with previous studies by Moraes-da-Silva et al. (2019, 2021). Helicops carinicaudus and He. nentur occurs in the south and southeast of Brazil, and their relationship is strong geographically structured (Fig. S4). The other species of the group, He. boitata, is endemic to the Pantanal (Fig. S4) with no phylogenetic proximity to the species that inhabit the same region (He. angulatus, He. leopardinus and He. polylepis) (Moraes-da- Silva et al. 2019). We recovered He. carinicaudus as paraphyletic (Fig. 4B), the nominal species coincides with He. carinicaudus L1, because the type locality is Rio Itapemirim, state of Espírito Santo, Brazil (Fig. S4), while the other two candidate species would not have available names (He. carinicaudus L2 and L3). Although He. carinicaudus has already been considered a subspecies of He. infrataeniatus (Lema 1958), both species are not phylogenetically related. Helicops danieli Amaral, 1938 (Clade M; Fig. 2 and 4C) Comments: Helicops danieli as sister species of the He. hagmanni group, He. polylepis and He. angulatus group (Fig. 4C). We tested for the first time the phylogenetic relationship of He. danieli. We highlight that the Cis-Andean Helicops species, He. scalaris and He. danieli, did not form a clade. Amaral (1937–1938), when he described this species, postulated that He. danieli would be phylogenetic close to He. angulatus and He. scalaris. Medem (1968) and Dugand (1975) mentioned that He. danieli might be synonymous with He. scalaris because 45 they consider that both species are morphologically similar. However, no later studies synonymize them (Rossman 2002a), and according to our results He. danieli is a valid species. The Helicops hagmanni group (Clade N; Fig. 2 and 4C) Content: (five species). He. acangussu Moraes-da-Silva, Walterman, Citeli, Sales-Nunes and Curcio, 2022; He. hagmanni Roux, 1910; He. pastazae Shreve, 1934; He. petersi Rossman, 1976 and He. yacu Rossman and Dixon, 1975. Comments: Helicops hagmanni group was recovered with well support in both analyzes (PP=1.00; BS=100%). Helicops hagmanni species is closely related to He. pastazae complex (Fig. 4C). Helicops polylepis was not recovered in BI as belonging to He. pastazae complex (see discussion of He. polylepis below). Therefore, due to the contrasting result between both methods (BI and ML; Fig. 2, 4C and 5; Fig. S1), we decided to remove He. polylepis from the section defined by Rossman (1976) and include He. hagmanni, renaming it Helicops hagmanni group. Rossman (1976) when coined the polylepis section grouping highlighted that He. hagmanni had some characteristics with the species this section, nevertheless the author decides not allocated the He. hagmanni with this grouping. In previous studies, He. hagmanni has been recovered as a sister species of He. infrataeniatus, He. carinicaudus, He. gomesi and He. angulatus (Grazziotin et al. 2012; Pyron et al. 2013; Figueroa et al. 2016). The most current inference using only 11 species of Helicops (Moraes-da-Silva et al. 2021) recovered He. hagmanni as the sister group of the clade composed by, He. polylepis He. gomesi, He. angulatus. However, these relationships are an artifact of the lack of sampling of other Helicops species, mainly species that also occur outside Brazil. Our results differ substantially from the previous studies, and our inference recovered He. hagmanni sister species to He. pastazae complex. The specimen of He. hagmanni IAVH-CT21340 from Colombia is morphologically similar to He. hagmanni (N. Citeli pers. comm.), but we consider it affinis He. hagmanni, which separate it from the other individuals of He. hagmanni (Fig. S5) because this sample has five of the six genes sequenced, lacking only the Cmos gene, this differentiation is not an artifact of missing data. Furthermore, an integrative taxonomic review is urgent to assess the status of He. hagmanni as, for example, dorsal scales vary between 23–29 in different populations. It was not possible to test the phylogenetic relationship of He. acangussu because we did not have access of tissue samples. Morphologically, He. acangussu is more similar to He. hagmanni, and occurs in sympatry with He. hagmanni, He. leopardinus and He. polylepis and can be confused with these species due of the spotted color pattern (Moraes-da-Silva et al. 2022). Nevertheless, He. 46 acangussu is oviparous, like He. hagmanni, while He. leopardinus and He. polylepis are viviparous. For that reason, we tentatively include He. acangussu in He. hagmanni group. The Helicops pastazae complex (Clade O; Fig. 2 and 4C) Content: (three species) He. pastazae Shreve, 1934; He. petersi Rossman, 1976 and He. yacu Rossman and Dixon, 1975. Comments: This is the first time that the phylogenetic relationship of He. pastazae and He. yacu are considered. It is noteworthy that Rossman and Abe (1979) observed that He. yacu and He. pastazae shared several qualitative and quantitative characteristics, and pointed out that more specimens from northeastern Peru would likely reveal that He. yacu is a subspecies of He. pastazae from the Upper Amazon Basin. Rossman and Dixon (1975) in the description of He. yacu points out that the dorsal color pattern of this species resembles in parts He. hagmanni. We decided to consider He. yacu is a valid species pending an integrative taxonomic review. However, the clade that includes He. yacu and He. pastazae is monophyletic with maximum support values in both analyses. Therefore, we decided to include He. yacu in He. pastazae complex (Fig. 4C; Fig. S5). We opt to name He. pastazae complex in order to keep their actual taxonomy. However, we highlight the need for a taxonomic revision within this clade, given the molecular and variation found. It was not possible to include in our analysis He. pertersi, an endemic species of the upper watershed of the Napo River in eastern Ecuador (Arteaga 2021). This species is commonly confused with He. pastazae due to sympatry and for having several overlapping morphological characteristics (Rossman 1976). The specimens IBSP90604 do not present the morphological diagnosis of He. petersi (N. Citeli pers. comm.). We highlight the need for an integrative systematic review and a more comprehensive sampling of He. hagmanni group to solve the systematics of this group. Helicops polylepis Günther, 1861 (Clade P; Fig. 2 and 5). Comments: Helicops polylepis was the sole species with contrasting placement in the two topologies. In BI it is a sister species of He. angulatus group with moderate support value (PP=0.90) (Fig. 5), while in ML it was recovered with low support value as sister group of He. hagmanni group (BS=35%). Even though the support value is moderate in BI, we decided to abandoned the name polylepis section coined by Rossman (1976). The most current inferences from Hydropsini recovered He. polylepis as the sister group of He. gomesi and He. angulatus with high support values (Moraes-da-Silva et al. 2019). However, the lack of a good sampling 47 probably hampers the solution for the positioning of He. polylepis. Rossman (1976) was the only author to propose a phylogenetic hypothesis between Helicops species and coined the following groups: polylepis section (He. pastazae, He. petersi, He. polylepis and He. yacu) and the He. pastazae complex (He. pastazae and He. petersi). The author considered phenotypic characteristics such as scale count and color pattern to propose their relationships. This relationship was not supported by molecular data among He. polylepis, He. pastazae, He. petersi and He. yacu, and demonstrates that several qualitative and quantitative characters in Hydropsini are plastic and can cause instability in the systematics of the tribe. Helicops hogei Lancini, 1964 Lancini (1964) described He. hogei based on a single specimen (MCNC 462, female) from Rio Autana, Amazonas, Venezuela, and later Pérez-Bravo (1976–1977) reported a second specimen (MBUCV 8849, female) from Rio Cuybini, Cerro La Paloma, Sierra Imataca, Delta Amacuro, Venezuela. Rossman (2002b) contested the taxonomic status of Helicops hogei, which he compared He. scalaris with a photo of the back of He. hogei (provided by Vanda Ferreira and Rubens Yuki) as well as descriptions of the literature (Lancini 1964 and Pérez Bravos 1976– 1977). Apparently, the two females of He. hogei are morphologically indistinguishable from females of He. scalaris. Rossman (2002b) declared to be convinced that He. hogei represents a junior synonym of Helicops scalaris (“To my mind, the evidence seems convincing that Helicops hogei Lancini is, in fact, a junior synonym of He. scalaris Jan”), however, he did not fulfill the taxonomic needs to perform the synonym. The International Code of Zoological Nomenclature (ICZN) demand that the taxonomist be clearer, stating that the researcher is synonymizing the species. Moreover, Rossman reports his concern about the gap in t