1 Mar Transcontinental Biogeografia Histórica das jararacas do grupo Bothrops alternatus (Serpentes, Viperidae) Daniela Petrim de Araújo 2 Daniela Petrim de Araújo Biogeografia Histórica das jararacas do grupo Bothrops alternatus (Serpentes, Viperidae) São José do Rio Preto 2015 Câmpus de São José do Rio Preto 3 4 Daniela Petrim de Araújo Biogeografia Histórica das jararacas do grupo Bothrops alternatus (Serpentes, Viperidae) Dissertação apresentada como parte dos requisitos para obtenção do título de Mestre em Biologia Animal, junto ao Programa de Pós-Graduação em Biologia Animal, do Instituto de Biociências, Letras e Ciências Exatas da Universidade Estadual Paulista “Júlio de Mesquita Filho”, Campus de São José do Rio Preto. Orientador: Prof. Dr. Ricardo J. Sawaya São José do Rio Preto 2015 9 10 “Prefiro ser parente de um macaco que de um homem que usa eloquência para destruir a verdade.” Thomas Huxley 11 Aos meus pais, pelo amor, apoio, por tudo. 12 AGRADECIMENTOS Gostaria de agradecer: - primeiramente aos meus pais, Antonio e Ivanil, por acreditarem e respeitarem esse sonho sem contestar, por entenderem a minha ausência, pelo amor, por acreditarem em mim quando nem mesmo eu acreditava, pelas comidas maravilhosas que me esperavam em cada visita, pelas engenhocas que facilitaram a minha vida, pelo apoio financeiro; - ao Ricardo Basso, pela paciência, por me fazer rir, pela ajuda quando o Windows criava vida própria, pelos chocolates, pelo carinho, por aceitar minha ausência; - ao Ricardo J. Sawaya, pela orientação, pelas discussões, ensinamentos, críticas, sugestões conselhos, desafios, paciência, confiança; - ao Cristiano Nogueira, pela co-orientação, pelas dicas valiosas do ArcGIS, pelas planilhas de distribuição das espécies; - ao Silvio Nihei, Pryscilla Moll e Cristiano Nogueira, pelas discussões sempre ricas, pelas sugestões, pela valiosa contribuição; - ao Fausto, pela ajuda com ArcGIS e shapefiles; - ao Vinicius X. da Silva, por gentilmente permitir que eu visitasse a coleção “Alfred Russel Wallace”; - ao Valdir Germano, pelas dicas de identificação das espécies; - ao Marcelo Duarte, pela ajuda com bibliografia antiga; - à Fernanda Magno, por ceder gentilmente alguns dados de localidade de Bothrops alternatus; - aos meus amigos paulistanos, Natália, Cristian, Paulo, Karina, Patrícia, Lucas, Jaqueline, Amanda, Sabrina, que tornaram o mestrado e a vida em São Paulo mais leve, pelas risadas sinceras, pela amizade; - à FAPESP (Proc. nº 2013/03155-5), pelo apoio financeiro, sem o qual este trabalho não seria possível. 13 RESUMO Um dos objetivos centrais da biogeografia histórica é a reconstrução da história de um determinado táxon ou um grupo baseado em hipóteses filogenéticas. A ideia de que as relações filogenéticas fornecem informações da distribuição geográfica passada é antiga, e apenas se tornou possível com a Sistemática Filogenética de Willi Hennig. Análises de Dispersão e Vicariância (DIVA) e modelos de Dispersão, Extinção e Cladogênese (DEC) são métodos amplamente difundidos para a reconstrução da distribuição geográfica ancestral de espécies. Essas análises incluem eventos anagenéticos e cladogenéticos. Recentemente, esses métodos foram implementados em um pacote do programa R denominado BioGeoBEARS. Além dos eventos anagenéticos e cladogenéticos já implementados nas versões originais de DEC e DIVA, o BioGeoBEARS apresenta a possibilidade de incluir a dispersão em saltos, ou jump dispersal, que tem se mostrado interessante mesmo quando consideradas espécies ou grupos que habitam o continente. O grupo Bothrops alternatus é um clado monofilético recuperado em diferentes filogenias do gênero Bothrops e inclui seis espécies de jararacas que habitam principalmente áreas de vegetação aberta: B. alternatus, B. itapetiningae, B. fonsecai, B. cotiara, B. ammodytoides e B. jonathani. Os objetivos deste trabalho são: 1) inferir a origem do grupo B. alternatus; 2) compreender a diversificação e evolução da distribuição geográfica das espécies atuais e seus ancestrais hipotéticos; 3) analisar se a diversificação ocorreu a partir de áreas abertas ou florestais; e 4) entender como processos geológicos e 14 climáticos podem ter influenciado a distribuição atual do grupo. Revisamos e georreferenciamos 3.265 registros de distribuição geográfica de espécimes depositados em coleções zoológicas e dados da literatura. Utilizamos uma filogenia datada que inclui cinco das seis espécies do grupo, pois B. jonathani é uma espécie rara e sem dados moleculares disponíveis até o momento. As unidades biogeográficas nas quais as espécies atuais ocorrem foram delimitadas por meio da intersecção da distribuição geográfica das espécies e superfícies eletrônicas (shapefiles) de províncias e unidades de relevo do Brasil. As distribuições geográficas ancestrais foram reconstruídas a partir dos modelos paramétricos DIVA-like, DIVA-like+j, DEC e DEC+j, implementados no pacote BioGeoBEARS. A distribuição ancestral foi estimada em dois cenários diferentes, com e sem as probabilidades relativas de dispersão entre as áreas, divididas em quatro intervalos de tempo (time slices). Os melhores modelos para os dois cenários incluem o parâmetro relativo à dispersão por salto (jump dispersal) e sugerem que a origem do grupo ocorreu tanto em áreas abertas quanto florestais. No primeiro cenário, a estimativa do modelo DIVA-like+j indica origem do grupo nas unidades Chaco-Pampa, Cerrado e Araucaria. O modelo DEC+j, melhor avaliado no segundo cenário (com time slices) indica origem do grupo no Cerrado e Araucaria. A dispersão por salto (jump dispersal) foi o processo mais importante para a estimativa global de diversificação do grupo nos dois cenários. Assim como a origem, a diversificação do grupo ocorreu de forma simultânea em áreas abertas e florestais. Possivelmente, o Mar Transcontinetal que invadiu o sul da América do Sul durante o Mioceno entre 23 e 13.8 milhões de anos, está 15 relacionado aos eventos de diversificação grupo B. alternatus. Durante a origem e diversificação do grupo, o clima da América do Sul estava se tornando mais seco e a vegetação florestal estava sendo substituída por vegetação aberta, o que deve ter favorecido a expansão e a diversificação do grupo para os biomas localizados mais ao sul da América do Sul. Palavras-chave: Distribuição geográfica ancestral, Bothrops, Dispersão por salto, BioGeoBEARS, Vegetação aberta, Florestas, Floresta de Aráucaria, Cerrado, Mar Trasncontinental. 16 ABSTRACT One central aim of historical biogeography is the history reconstruction of a taxon or clade based on phylogenetic hypotheses. The idea that the phylogenetic relationships provide information on past geographical distribution is old, and it was only possible with the Willi Hennig’s Cladistic method. Dispersal Vicariance Analysis (DIVA) and Models of Dispersal-Extinction-Cladogenesis (DEC) are among the most popular methods to reconstruct ancestral geographical distributions. Both analysis include anagenetic and cladogenetic processes. Both methods were recently implemented in the R software package BioGeoBEARS. Besides the anagenetic and cladogenetic processes of the original DEC and DIVA, BioGeoBEARS includes a parameter related to jump dispersal, that has been shown interesting even for non-insular species. The Bothrops alternatus group corresponds to a monophyletic clade, recovered in different phylogenies of the genus Bothrops and includes six lancehead snakes species that mainly inhabit open vegetation areas: B. alternatus, B. itapetiningae, B. fonsecai, B. cotiara, B. ammodytoides and B. jonathani. Our aims here are: 1) to infer the area origin of the B. alternatus group; 2) to comprehend the diversification and evolution of the geographic distribution of present species and their hypothetical ancestrals; 3) to analyze if diversification occurred in open vegetation or forested areas; and 4) to understand how geological and climatic processes may have influenced the present distribution of the group. We revised and georeferenced 3,265 records of geographic distributions of species via zoological collection records and 17 literature data. We used a phylogeny including five out six species of the group, because B. jonathani is a rare species lacking molecular data. The biogeographic units in which the species occur were delimited by intersection of geographic distribution of the species and shapefiles of provinces and Brazilian relief units. We then reconstructed the ancestral geographic distributions by means of parametric models DIVA-like, DIVA-like+j, DEC and DEC+j, implemented in BioGeoBEARS package. Ancestral distributions were estimated in two different scenarios, with and without relative probability of dispersal and the possibility to disperse among areas in four time slices. The best models on both scenarios included jump dispersal and the origin of the group both in open and forested areas. In the first scenario, DIVA-like+j indicated the origin of the group in Chaco-Pampa, Cerrado and Araucaria units. The DEC+j modelwas the best evaluated model for the second scenario (with time slices), indicating the origin of the group on the Cerrado and Araucaria. The most important process in global estimates for each model for both scenarios was jump dispersal. As well as the origin, the diversification of the group must have occurred simultaneously in open and forested areas. The Transcontinental Sea invading the southern South America during the Miocene between 23 and 13,8 million years is possibly related to the events of the diversification of B. alternatus. During the origin and diversification of the group, the climate of South America was becaming drier and the forests have being replaced by open vegetation, which must have favoured the expansion and diversification of the group towards the southern South American biomes. 18 Keywords: Ancestral range, Bothrops, Jump dispersal, BioGeoBEARS, Open landscape, Forest, Araucaria forest, Cerrado, Transcontinental Sea. 19 SUMÁRIO Introdução Geral.........................................................................................................................1 Objetivos.........................................................................................................................9 Referências....................................................................................................................11 Manuscrito – Historical Biogeography of lancehead snakes of Bothrops alternatus group (Serpentes, Viperidae)...................................................................18 Abstract.........................................................................................................................19 Introduction...................................................................................................................21 Methods.........................................................................................................................24 Results...........................................................................................................................27 Discussion.....................................................................................................................30 References….................................................................................................................35 Appendix.......................................................................................................................42 Supporting Information.................................................................................................46 Conclusões Gerais.....................................................................................................................51 Apêndice...................................................................................................................................52 1 Introdução Geral Biogeografia histórica A biogeografia pode ser definida como o estudo da distribuição geográfica dos organismos (Crisci, 2001). Essa simples definição, entretanto, mascara a sua complexidade, que engloba a geografia, a geologia e a biologia (Crisci, 2001). A Biogeografia tem sido tradicionalmente dividida em duas linhas de pesquisa, a biogeografia ecológica e a histórica (Lomolino et al., 2010). Um dos primeiros autores a tratá-la dessa forma foi o botânico de Candolle, por volta de 1820, que afirmava que a biogeografia ecológica poderia ser atribuída a causas físicas que atuam no presente, e a biogeografia histórica seria relacionada a causas físicas do passado. Assim, a biogeografia histórica está interessada em compreender os processos que ocorreram há milhares ou milhões de anos atrás (Crisci, 2001). Um dos objetivos centrais da biogeografia histórica é a reconstrução da história de um determinado táxon ou grupo baseado em hipóteses filogenéticas (Cox & Moore, 2005). A ideia de que as relações filogenéticas fornecem informações sobre a distribuição geográfica passada é antiga (Ronquist & Sanmartín, 2011). Porém, apenas após a Sistemática Filogenética de Willi Hennig que a biogeografia filogenética se tornou possível, e a partir de então muitos métodos foram propostos (cf. Ronquist & Sanmartín, 2011). Dois métodos amplamente difundidos para a reconstrução de distribuições geográficas ancestrais são a Análise de Dispersão e Vicariância (Dispersal-Vicariance Analysis - DIVA) e os modelos de Dispersão, Extinção e Cladogênese (Dipersal-extinction cladogenesis -DEC) (Yu et al., 2010; Matzke, 2013a). O DIVA é uma análise de parcimônia baseada em eventos que se tornou bastante popular por não requerer conhecimento prévio sobre a história da área de ocorrência das 2 espécies (Nylander et al., 2008). A análise requer apenas uma filogenia e as distribuições geográficas das espécies atuais (Ronquist, 1997). A otimização das áreas ancestrais é feita nos nós internos da filogenia com base em uma matriz tridimensional de custos associados aos eventos. Dispersões e extinções assumem o valor de um, enquanto a vicariância tem custo zero (Ronquist, 1997). A reconstrução ancestral ótima é aquela que apresenta o menor custo, o que acaba favorecendo as hipóteses vicariantes em detrimento de eventos de dispersão e extinção (Kodandaramaiah, 2010). Outra limitação da análise é que a posição relativa das áreas não é reconhecida. Portanto, todas as combinações de áreas são possíveis para compor as distribuições ancestrais potenciais mesmo que sejam áreas disjuntas (Ronquist & Sanmartin, 2011). Uma das condições para a análise DIVA é a disponibilidade de uma filogenia completamente bifurcada, sem a presença de politomias (Nylander et al., 2008). O método ignora a incerteza na inferência filogenética, assumindo que a mesma não apresenta erro (Sanmartín & Ronquist, 2004; Yu et al., 2010). Porém, diferentes filogenias podem fornecer diferentes reconstruções geográficas (Nylander et al., 2008). Outra fonte de incerteza está relacionada à otimização da área ancestral, que sugere várias distribuições igualmente parcimoniosas para os nós ancestrais, especialmente quando a distribuição das espécies atuais é ampla (Ronquist, 1997). Isso tende a aumentar em direção à raiz, onde a ambiguidade das distribuições é maior (Ronquist, 1997). Além disso, os eventos de extinção praticamente nunca são recuperados nas reconstruções (Ronquist, 1997). O processo de vicariância é favorecido, fazendo com que os nós basais apresentem distribuição ampla, e a dispersão é enviesada para os ramos terminais da filogenia (Ree et al., 2005; Nylander et al., 2008; Kodandaramaiah, 2010). A fim de diminuir esse efeito, aconselha-se que o número máximo de áreas ancestrais seja restrito, o que corresponde normalmente ao mesmo número de áreas que a espécie com distribuição mais ampla ocupe (Ronquist & Sanmartín, 2011). 3 Duas propostas foram sugeridas para melhorar a abordagem da análise DIVA. A primeira integra a abordagem Bayesiana no algoritmo de reconstrução de distribuições ancestrais (Nylander et al., 2008), e é denominada Bayes-DIVA. Esta análise fornece reconstruções de distribuições ancestrais mais parcimoniosas, mas pode considerar as incertezas de hipóteses filogenéticas (Nylander et al., 2008). Caso haja várias soluções ótimas para um único nó, o método indica o conjunto de solução mais simples, ou mais restrito, favorecendo alternativas igualmente parcimoniosas (Nylander et al., 2008). A segunda abordagem é a Análise Estatística de Dispersão e Vicariância (Statistical Dispersal- Vicariance Analysis) denominada S-DIVA (Yu, et al., 2010). Além da incerteza filogenética, o S-DIVA fornece suporte estatístico para as reconstruções de distribuições geográficas ancestrais (Yu et al., 2010). O modelo de Dispersão, Extinção e Cladogênese (DEC) é um método paramétrico baseado em probabilidades, ou maximum likelihood (Ree et al., 2005; Ree & Smith, 2008), portanto mais complexo e flexível em relação ao DIVA (Kodandaramaiah, 2010). Os autores do DEC enfatizam a importância de trazer a história da Terra para o seu modelo (Ree et al., 2005). A reconstrução ancestral, nesse caso, requer, além de uma filogenia datada, ou com idades de divergência, o conhecimento das conexões entre as áreas e possíveis probabilidades relativas de dispersão entre elas em diferentes períodos de tempo ou time slices (Ree & Smith, 2008; Ree & Sanmartín, 2009; Kodandaramaiah, 2010). Isso possibilita a remoção de áreas disjuntas, reduzindo o conjunto de áreas permitidas para dispersão (Ree & Sanmartín, 2009). Consequentemente, a probabilidade atribuída à dispersão entre áreas contínuas é maior do que entre áreas que apresentem barreiras para dispersão (Kodandaramaiah, 2010). Em contraste ao DIVA, a dispersão e a extinção no DEC são tratadas como processos estocásticos que causam expansão ou retração da distribuição geográfica, respectivamente (Ree & Sanmartín, 2009). 4 Embora essencialmente diferentes, sendo um baseado em parcimônia e o outro em maximum likelihood, DIVA e DEC, respectivamente, apresentam muitos aspectos em comum (Ree & Sanmartín, 2009). As reconstruções ancestrais podem estimar distribuições ancestrais amplas em ambos os métodos (Kodandaramaiah, 2010), além de produzir resultados muito semelhantes (Clayton et al., 2009). Os eventos de extinção são mal previstos pelos dois métodos, porém o DIVA dificilmente sugere este evento (Kodandaramaiah, 2010; Ronquist & Sanmartín, 2011). Alguns eventos cladogenéticos são diferentes para o DEC e o DIVA (Ree & Sanmartín, 2011). As duas principais diferenças estão nos eventos de vicariância e especiação peripátrica (Ronquist & Sanmartín, 2011). O DIVA prevê a vicariância clássica, na qual um ancestral amplamente distribuído gera duas espécies em que cada espécie herda exatamente o mesmo número de áreas (Ronquist & Sanmartín, 2011). O DEC apenas prevê a vicariância em que uma das espécies-filha herda uma área mais ampla que a outra, além da especiação peripátrica (ou subset sympatry), em que uma espécie-filha herda toda a área de seu ancestral e a outra herda apenas uma pequena porção (Ronquist & Sanmartín, 2011). Entretanto, os eventos de duplicação (Ronquist & Sanmartín, 2011) ou de simpatria a partir de uma área ancestral pequena (Matzke, 2013a) são previstos igualmente nos dois métodos, enquanto que a especiação simpátrica a partir de um ancestral amplamente distribuído não é previsto em ambos (Ronquist & Sanmartín, 2011). Discussões sobre a importância dos processos de vicariância e dispersão voltaram com grande força na literatura recente (de Queiroz, 2005; Cowie & Holland, 2006; Santos, 2007; Heads, 2012). As propostas vicariantes reinaram durante os anos de 1970 e 1980 como o processo mais importante e responsável por explicar os padrões atuais de distribuição geográfica das espécies (Cowie & Holland, 2006; Santos, 2007). A dispersão era considerada apenas um evento aleatório responsável unicamente por causar ruído nos sistemas vicariantes (Morrone & Crisci, 1995; Humpries & Parenti, 1999). Além disso, as hipóteses baseadas em 5 dispersão eram consideradas desnecessárias (Rosen, 1976), e incapazes de oferecer explicações generalizadas (Cowie & Holland, 2006). Entretanto, a partir do aumento expressivo do conhecimento sobre biologia molecular e filogenias com datas de divergência, notou-se que alguns clados são muito jovens para terem suas histórias interpretadas exclusivamente por vicariância (de Queiroz, 2005). Isso fortaleceu as ideias dispersionistas, gerando a inclusão de processos de dispersão nas ferramentas da biogeografia histórica, tais como o DIVA (Santos, 2007) e o DEC. O principal problema com as ideias dispersalistas seria que elas não poderiam ser falseadas. Entretanto, isso pode ser contornado quando as hipóteses vicariantes são falseadas e a dispersão é suportada. Além disso, hipóteses específicas relacionadas à dispersão podem incluir a predição das datas de divergência e áreas das linhagens irmãs, podendo então estar sujeitas a testes de hipóteses (de Queiroz, 2005). Um novo pacote do programa R denominado BioGeoBEARS, BioGeographic Bayesian and likelihood Evolutionary Analysis with R Scripts, vem contribuindo recentemente para o fortalecimento de processos de dispersão relacionados à diversificação de espécies ou unidades filogenéticas (Matzke, 2013a,b). O BioGeoBEARS implementa o DEC e mais duas análises, uma baseada no DIVA, denominada DIVA-like, e outra no BayArea, denominada BayArea-like (Matzke, 2013b). O BayArea é uma análise bayesiana de inferência da história biogeográfica que tem como principal vantagem a possibilidade de inclusão de maior número de áreas (Landis et al., 2013), que podem ser computacionalmente impossíveis em análises DEC. A principal novidade dos modelos gerados pelo BioGeoBEARS é um parâmetro de especiação cladogenética conhecido como evento fundador, dispersão por salto ou jump dispersal (Matzke, 2013b). O jump dispersal é um tipo de dispersão caracterizada por ocorrer em grandes distâncias (Lomolino et al., 2010). Um exemplo clássico é o das ilhas Krakatau, na Indonésia. Em 1883, uma erupção vulcância dizimou três ilhas e destruiu dois- terço da maior ilha do arquipélago. Após 50 anos, as ilhas estavam novamente cobertas por 6 uma densa floresta, com 31 espécies de aves e várias espécies de invertebrados (Lomolino et al., 2010). Embora o jump dispersal seja comumente aceito em sistemas insulares, alguns autores ainda não aceitam a sua importância, preferindo explicações alternativas como land bridges ou stepping-stone islands (Lomolino et al., 2010). Há também evidências de dispersão por longas distâncias no continente (Crudden, 1966). Os modelos apresentados pelo BioGeoBEARS tem chamado a atenção para a importância dos eventos de jump dispersal tanto em ilhas quanto em continentes. A comparação entre modelos DEC com e sem a inclusão de jump dispersal mostram que os modelos que incluem o parâmetro de dispersão por longas distâncias são normalmente melhor avaliados com poucas exceções (Matzke, 2014; N. Matzke, com. pess.). As estimativas de distribuições ancestrais em passeriformes e saguis na América do Sul revelaram que os modelos que incluem jump dispersal apresentam melhor avaliação (cf. Batalha-Filho et al., 2014; Buckner et al., 2015). O gênero Bothrops As serpentes do gênero Bothrops pertencem à família Viperidae e subfamília Crotalinae, e incluem atualmente 48 espécies distribuídas desde o México até a Argentina (Campbell & Lamar, 2004; Carrasco et al., 2012). O Brasil apresenta a maior diversidade do gênero (Martins et al., 2001), abrigando 27 espécies (Costa & Bérnils, 2014). As jararacas apresentam grande diversidade quanto à morfologia e ecologia (Campbell & Lamar, 2004; Martins et al., 2001; Carrasco et al., 2012). São serpentes que ocupam uma ampla variedade de habitats, desde florestas tropicais e subtropicais a regiões áridas ou semiáridas, ilhas e altitudes acima dos 3000 metros acima do nível do mar (Amaral, 1921; Campbell & Lamar, 2004; Carrasco et al., 2010; Barbo et al., 2012; Carrasco et al., 2012). 7 De acordo com alguns trabalhos, os crotalíneos se originaram na Ásia e provavelmente alcançaram as Américas por um único evento de dispersão. A dispersão teria ocorrido da Ásia para a América do Norte através do Estreito de Bering, ainda quando ele conectava a Ásia e o Alasca (Parkinson, 1999; Parkinson et al., 2002; Wüster et al., 2002; Wüster et al., 2008). Os crotalíneos provavelmente alcançaram o Novo Mundo através do Estreito de Bering durante o Oligoceno tardio - início do Mioceno (28 a 23 milhões de anos antes do presente; Ma ap), ainda quando o Estreito de Bering estava coberto por florestas decíduas (Sanmartín et al., 2001). Os crotalíneos do novo mundo divergiram do seu grupo irmão há aproximadamente 24 milhões de anos Ma ap, e provavelmente a partir de um evento de cladogênese, que ocorreu há cerca de 18 Ma ap, os crotalíneos do Novo Mundo divergiram, originando o gênero Bothrops (Wüster et al., 2008). A partir da colonização da América do Norte, ocorreram várias trocas de Bothrops entre as Américas do Norte, Central e do Sul (Parkinson, 1999; Parkinson et al., 2002; Wüster et al., 2002; Wüster et al., 2008). Não há consenso em relação às idades de divergência e de invasão das Américas, mas, a hipótese de dispersão da Ásia para a América do Norte pelo Estreito de Bering está em acordo em diferentes trabalhos. Alguns autores sugerem que a colonização tenha ocorrido no Cretáceo tardio - início do Terciário (70 a 61 Ma ap;,Parkinson et al., 2002) enquanto outros sugerem que tenha ocorrido entre o Oligoceno tardio e início do Mioceno (28 a 23 Ma ap;,Wüster et al., 2008). As filogenias recentes evidenciam a presença de alguns grupos monofiléticos no gênero Bothrops (Wüster et al., 2002; Fenwick et al., 2009; Carrasco et al., 2012). Embora ainda não haja consenso sobre as relações entre e dentro de alguns grupos, o grupo B. alternatus é recuperado nas diferentes propostas filogenéticas. Nas filogenias mais completas e que incluem dados morfológicos, o grupo B. alternatus apresenta seis espécies: B. alternatus, B. itapetiningae, B. fonsecai, B. cotiara, B. ammodytoides e B. jonathani (Fenwick et al., 2009; Carrasco et al., 2012). Essas espécies estão amplamente distribuídas na América 8 do Sul, ocupando as regiões central, sudeste e sul do Brasil, sul da Bolívia, Uruguai, Paraguai e as regiões norte e central da Argentina (Campbell & Lamar, 2004; este estudo). Bothrops alternatus é a espécie do grupo com maior amplitude de distribuição, ocupando desde o sul, sudeste e parte da região central do Brasil, Paraguai e Uruguai até o norte da Argentina (Hoge & Belluomini, 1964; Hoge & Romano, 1971; Hoge & Romano- Hoge, 1981; Melgarejo, 2003; Campbell & Lamar 1989; 2004). Bothrops itapetiningae está presente na região central e sudeste do Brasil (Amaral, 1977; Hoge & Romano-Hoge, 1981; Campbell & Lamar, 1989; 2004). Bothrops ammodytoides ocorre do norte ao sul da Argentina, (Cei, 1986; 1993; Giraudo & Scrocchi, 2002; Campbell & Lamar, 2004; Scrocchi et al., 2006, Carrasco et al., 2010) apresentando a distribuição mais austral entre as serpentes (Carrasco et al., 2010). Bothrops fonsecai apresenta distribuição no sudeste do Brasil (Hoge & Belluomini, 1964; Peters & Orejas-Miranda, 1970; Hoge & Romano, 1971; Hoge e Romano-Hoge, 1981; Campbell & Lamar, 1989; 2004), enquanto, Bothrops cotiara, ocorre no sul do Brasil e em uma pequena área do norte da Argentina, em Missiones (Gomes, 1913; Hoge & Romano, 1971; Hoge & Romano-Hoge, 1981; Campbell & Lamar, 1989; 2004). Bothrops jonathani, está restrita à região central e sul da Bolívia e noroeste da Argentina (Harvey, 1994; Carrasco et al., 2009). Em geral, as espécies do grupo ocorrem em hábitats abertos (Martins et al., 2001), como o Cerrado, os Pampas e o Chaco. Bothrops alternatus é encontrada em formações estépicas, e preferencialmente em áreas alagáveis ou ripárias, bem como em áreas alteradas pelas ações antrópicas (Campbell & Lamar, 1989, 2004; Mesquita & Brites, 2003; Sawaya et al., 2008). Bothrops itapetiningae é endêmica do Cerrado (Leão et al., 2014), e ocorre mais frequentemente em formações campestres mais abertas e preservadas, tais como campos sujos e campos limpos (Sawaya et al., 2008; Leão et al., 2014). Bothrops cotiara é considerada como típica de campos de áreas montanhosas associados à floresta ombrófila mista (Campbell 9 & Lamar, 1989; 2004; Martins, 2002). Porém, essa espécie é encontrada dentro ou na borda da mata e raramente em áreas abertas dos campos vizinhos (cf. Bérnils, 2009). O habitat de Bothrops fonsecai é caracterizado como de campo e mata de Araucária (Campbell & Lamar, 1989; 2004; Martins, 2002). Bothrops ammodytoides vive em áreas rupestres e arenosas da Argentina (Carrasco et al., 2010). Das espécies desse grupo, Bothrops jonathani é a que ocorre nas maiores altitudes, em áreas savânicas e rupestres da Cordilheira dos Andes (Harvey, 1994; Carrasco et al., 2009). Este trabalho se propôs a investigar a evolução da distribuição geográfica e do uso dos habitats abertos e florestais pelas espécies do grupo Bothrops alternatus. Poucos trabalhos de biogeografia enfocam a história de diversificação de serpentes (mas veja, Pook et al., 2000; Parkinson, 1999; Parkinson et al., 2002; Wüster et al., 2002; Quijada-Mascareñas et al., 2007; Wüster et al., 2008; Machado et al., 2014). Compreender como ocorreu a diversificação do grupo B. alternatus pode esclarecer muito da história do gênero na América do Sul, já que corresponde a um grupo basal. Outra característica interessante do grupo é que as suas espécies habitam áreas de vegetação aberta, enquanto a grande parte das espécies de jararacas está associada habitats florestas (Martins et al., 2001). Objetivos O objetivo geral deste trabalho é inferir a origem e compreender os processos relacionados à diversificação do grupo Bothrops alternatus na América do Sul. São propostas as seguintes questões específicas: 1) Onde o grupo se originou? 2) Como ocorreu a diversificação e evolução da distribuição geográfica das espécies atuais e seus ancestrais hipotéticos? 3) O grupo se diversificou a partir de áreas abertas ou florestais? 10 4) Como processos geológicos e climáticos podem ter influenciado a distribuição atual do grupo? 11 REFERÊNCIAS Amaral, A. (1977) Serpentes do Brasil. Iconografia Colorida. São Paulo: Ed. Melhoramentos e Editora da Universidade de São Paulo. 247 p. Amaral, A. (1921) Contribuição para o conhecimento dos ofídios do Brasil. A Parte II. Biologia de uma nova espécie. Anexos das Memórias do Instituto Butantan, 1, 39-44. Barbo, F.E.; Grazziotin, F.G.; Sazima, I.; Martins, M. & Sawaya, R.J. (2012) A new and threatened insular species of lancehead from southeastern Brazil. Herpetologica, 68, 418- 429. Batalha-Filho, H.; Pessoa, R.O.; Fabre, P.H.; Fjsedsa, J.; Irestedt, M.; Ericson, P.G.P.; Silveira, L. & Miyaki, C.Y. (2014) Phylogeny and historical biogeography of gnateaters (Passeriformes, Conopophagidae) in South America forests. Molecular Phylogenetics and Evolution, 79, 422-432. Bérnils, R.S. (2009) Composição e padrões de distribuição de Caenophidia (Squamata, Serpentes) das Serras Atlânticas e Planaltos do sudeste da América do Sul. Tese apresentada ao Programa de Pós-graduação em Ciências Biológicas (Zoologia). Universidade Federal do Rio de Janeiro. Rio de Janeiro. Buckner, J.C.; Alfaro, J.W.L.; Rylands, A.B. & Alfaro, M.E. (2015) Biogeography of the marmosets and tamarins (Callitrichidae). Molecular Phylogenetics and Evolution, 82, 413- 425. Campbell, J.A. & Lamar, W.W. (2004) The venomous reptiles of the western hemisphere. Cornell University Press, USA, volumes I e II. Campbell, J. A. & Lamar, W. W. (1989) The Venomous Reptiles of Latin America. Ithaca: Cornell University Press, Ithaca, New York. 12 Carrasco, P.A.; Mattoni, G.I.; Leynaud, G.C. & Scrocchi, G.J. (2012) Morphology, phylogeny and taxonomy of South American bothropoid pitvipers (Serpentes, Viperidae). Zoologia Scripta, 41, 109-124. Carrasco, P.A.; Leynaud, G.C. & Scrocchi, G.J. (2010) Rediscription of the southernmost snake species, Bothrops ammodytoides (Serpentes: Viperidae: Crotaline). Amphibia- Reptilia, 31, 323-338. Carrasco, P.A.; Harvey, M.B. & Saravia, A.M. (2009) The rare Andinean pitviper Rhinocerophis jonathani (Serpentes: Viperidae: Crotalinae): rediscription with comments on its systematics and biogeography. Zootaxa, 2283, 1-15. Cei, J.M. (1993). Reptiles del nordeste y este de Argentina. Herpetofauna de las selvas subtropicales, Puna y pampas. Museo Regionale di Scienze Naturali. Torino, 14: 1-949. Cei, J.M. (1986). Reptiles del Centro, Centro-oeste y Sur de la Argentina. Herpetofauna de las zonas áridas y semiáridas. Museo Regionale di Scienze Naturali Torino, 14: 1-527. Clayton, J.W.; Soltis, P.S. & Soltis, D.E. (2009) Recent long-distance dispersal overshadows ancient biogeographical patterns in a pantropical angiosperm family (Simaroubaceae, Sapindales). Systematic Biology, 58, 396-410. Costa, H.C. & Bérnils, R.S. (org.) (2014) Répteis Brasileiros: Lista de espécies. Versão 2014. Disponível em http://www.sbherpetologia.org.br/. Sociedade Brasileira de Herpetologia. Acessada em 05/01/2015. Cox, C.B. & Moore, P.D. (2005) Biogeography: an ecological and evolutionray approach, Blackwell, Oxford, UK, 7th edition. Cowie, R.H. & Holland, B.S. (2006) Dispersal is fundamental to biogeography and the evolution of biodiversity on oceanic islands. Journal of Biogeography, 33, 193-198. Crisci, J.V. (2001) The voice of historical biogeography. Journal of Biogeography, 28, 157- 168. 13 Cruden, R.W. (1966) Birds as agents of long-distance for disjunct plant groups of the temperate western hemisphere. Evolution, 20, 517-532. De Queiroz, A. (2005) The resurrection of oceanic dispersal in historical biogeography. Trends in Ecology and Evolution, 20, 68-73. Fenwick, A. M., Gutberlet Jr, R. L., Evans, J. A. & Parkinson, C. L. (2009) Morphological and molecular evidence for phylogeny and classification of South American pitvipers, genera Bothrops, Bothriopsis, and Bothrocophias (Serpentes: Viperidae). Zoological Journal of the Linnean Society, 156, 617-640. Giraudo, A.R., & Scrocchi, G.J. (2002) Argentinean snakes: an annotated checklist. Smithson. Herpetological Information Service, 132, 1-53. Gomes, J.F. (1913) Uma nova cobra venenosa do Brasil. Annais Paulistas de Medicina e Cirurgia, 3, 65-67. Harvey, M.B. (1994) A new species of a montane pitviper (Serpentes: Viperidae: Bothrops) from Cochabamba, Bolivia. Proceedings of the Biological Society of Washington, 107, 60- 66. Heads, M.J. (2012) Molecular panbiogeography of the tropics. University of California Press, Berkeley. Hoge, A.R. & Romano-Hoge, S.A.L.W. (1981) Poisonous snakes of the world. Part I. Check list of the pit vipers, Viperoidea, Viperidae, Crotalinae. Memória do Instituto Butantan, 42/43, 179-309. Hoge, A. R. & Romano, S. A. (1971) Sinopse das serpentes peçonhentas do Brasil (Serpentes, Elapidae e Viperidae). Memórias do Instituto Butantan, 36, 109-208. Hoge, A. R. & Belluomini, H. E. (1964) Notas sobre Bothrops fonsecai (Hoge & Belluomini), Bothrops alternatus (Duméril, Bibron et Duméril) e Bothrops cotiara (Gomes). Memórias do Instituto Butantan, 30, 97-102. 14 Humpries, C.J. & Parenti, L.R. (1999) Cladistic biogeography. Oxford University Press, Oxford, 2ª edição. Kodandaramaiah, U. (2010) Use of dispersal-vicariance analysis in biogeography – a critique. Journal of Biogeography, 37, 3-11. Landis, M.J.; Matzke, N.J.; Moore, B.R. & Huelsenbeck, J.P. (2013) Bayesian analysis of biogeography when the number of areas is large. Systematic Biology, 62, 789-804. Leão, S.M.; Pelegrin, N.; Nogueira C.C. & Brandão, R.A. (2014) Natural history of Bothrops itapetiningae Boulenger, 1907 (Serpentes: Viperidae: Crotalinae), an endemic species of the Brazilian Cerrado. Journal of Herpetology, 48, 324-331. Lomolino, M.V.; Ridlle, B.R.; Whittaker, R.J.; Brown, J.H. (2010) Biogeography. Sinauer Associates, Sunderland, Massachussets, 4ª edição. Martins, M.; Marques, O.A.V. & Sazima, I. (2002) Ecological and phylogenetics correlstes of feedind habits in neotropical pitvipers of the genus Bothrops. Biology of the pitvipers (ed. by G. Schett, M. Höggren, & H.W. Greene), pp. 1-22. Carmel: Biological Sciences Press. Martins, M.; Araujo, M.S.; Sawaya, R.J. & Nunes, R. (2001) Diversity and evolution of macrohabitat use, body size and morphology in a monophyletic group of Neotropical pitvipers (Bothrops). Journal of Zoology, 254, 529-538. Machado, T.; Silva, V.X. & Silva M. J.J. (2014) Phylogenetic relationships within Bothrops neuwiedi group (Serpentes, Squamata): Geographically highlly-structured lineages, evidence of introgressive hybridization and Neogene/Quaternary diversification. Molecular Phylogenetics and Evolution, 71, 1-14. Matzke, N.J. (2014). Model selection in historical biogeography reveals that founder-event speciation is a crucial process in island clades. Systematic Biology, 63, 951-970. Matzke, N.J. (2013a) Probabilistic historical biogeography: new models for founder-event speciation, imperfect detection, and fossils allow improved accuracy and model-testing. 15 Ph.D. thesis, Department Integrative Biology and Designated Emphasis in Computational and Genomic Biology, University of California, Berkeley, 1-240. Matzke, N.J. (2013b) Probabilistic historical biogeography: new models for founder-event speciation, imperfect detection, and fossils allow improved accuracy and model-testing. Frontiers of Biogeography, 5, 242-248. Melgarejo, A. R. (2003) Serpentes peçonhentas do Brasil.. Animais Peçonhentos no Brasil (ed. by J.L.C. Cardoso, F.O.S França; F.H. Wen; C.M.S. Málaque & V. Haddad Jr), p. 42- 70. Sarvier, São Paulo. Mesquita, D.O. & Brites, V.L.C. (2003) Aspectos taxonômicos e ecológicos de uma população de Bothrops alternatus Duméril, Bibron & Duméril, 1854 (Serpentes, Viperidae) das regiões do Triângulo e Alto Paranaíba, Minas Gerais. Biologia Geral e Experimental, 2, 33-38. Morrone, J.J. & Crisci, J.V. (1995) Historical biogeography: introduction to methods. Annual Review of Ecology and Systematics, 26, 373-401. Nylander, J.A.A.; Olsson, U.; Alström, P. & Sanmatín, I. (2008) Accounting for Phylogenetic uncertainty in biogeography: a Bayes approach to Dispersal-Vicariance Analysis of the Thrushes (Aves: Turdus). Systematic Biology, 57, 257-268. Parkinson, C.L., Campbell, J.A. & Chippindale, P.T. (2002). Multigene phylogenetic analysis of pitviper, with comments on their biogeography. Biology of the Vipers (ed.G.W. Schuett, M. Höggren, M.E. Douglas & H.W. Greene), pp. 93-110. Eagle Mountain Publishing, Eagle Mountain, UT. Parkinson, C.L. (1999) Molecular systematics and biogeographical history of pitvipers as determined by mitochondrial ribosomal DNA sequences. Copeia, 1999, 576-586. Peters, J. A. & Orejas-Miranda, B. (1970) Catalogue of the Neotropical Squamata Part I. Snakes. United States National Museum Bulletin, 297. 16 Pook, C.E.; Wüster, W. & Thorpe, R.S. (2000) Historical biogeography of the western rattlesnake (Serpentes: Viperidae: Crotalus viridis), inferred from mithocondrial DNA sequence information. Molecular Phylogenetics and Evolution, 15, 269-282. Quijada-Mascareñas, J.A.; Ferguson, J.E.; Pook, C.E.; Salomão, M.G.; Thorpe, R.S. & Wüster, W. (2007) Phylogeographic patterns of trans-Amazonian vicariants and Amazonian biogeography: the Neotropical rattlesnake (Crotalus durissus complex) as an example. Journal of Biogeography, 34, 1296-1312. Ree, R.H. & Sanmartín, I. (2009) Prospects and challenges for parametric models in historical biogeographical inference. Journal of Biogeography, 36, 1211-1220. Ree, R.H. and Smith, S.A. 2008. Maximum likelihood inference of geographic range evolution by dispersal, local extinction, and cladogenesis. Systematic Biology, 57, 4-14. Ree, R.H.; Moore, B.R.; Webb, O.C. & Donoughue, M.J. (2005) A likelihood framework for inferring the evolution of geographic range on phylogenetic trees. Evolution, 59, 2299- 2311. Ronquist, F. & Sanmartín, I. (2011) Phylogenetic methods in biogeography. Annual Review of Ecology, Evolution and Systematics, 42, 441-464. Ronquist, F. (1997) Dispersal-vicariance analysis: a new approach to the quantification of historical biogeography. Systematics Biology, 46, 195-203. Rosen, D.E. (1976) A vicariance model of Caribbean biogeography. Systematic Zoology, 24, 431-464. Sanmartín, I. & Ronquist, F. (2004) Southern hemisphere biogeography inferred by event- based models: Plant versus animal patterns. Systematic Biology, 53, 216-243. Sanmartín, I.; Enghoff, H. & Ronquist, F. (2001) Patterns of animal dispersal, vicariance and diversification in the Holartic. Biological Journal of the Linnean Society, 73, 345-390. 17 Santos, C.M.D. 2007. On basal clades and ancestral areas. Journal of Biogeography, 34, 1470-1471. Sawaya, R.J., Marques, O. A. V. & Martins, M. (2008) Composition and natural history of Cerrado snake assemblage at Itirapina, São Paulo state, southeastern Brazil. Biota Neotropica, 8: 127-149. Scrocchi, G.J., Moreta, J.C. & Kretzschmar, S. (2006) Serpientes del Noroeste Argentino. Ed. Fundación Miguel Lillo, Tucumán, Argentina. Wüster W.; Peppin, L.; Pook, C.E. & Walker, D.E. (2008) A nesting of vipers: Phylogeny and historical biogeography of the Viperidae (Squamata: Serpentes). Molecular Phylogenetics and Evolution, 49, 445-459. Wüster, W.; Salomão, M. G.; Quijada-Mascarenãs, J.A. & Thorpe R. S. (2002) Origins and evolution of the South American pitviper fauna: evidence from mitochondrial DNA sequence analysis. Biology of the Vipers (ed.G.W. Schuett, M. Höggren, M.E. Douglas & H.W. Greene), pp. 111-128. Eagle Mountain Publishing, Eagle Mountain, UT. Yu, Y.; Harris, A.J. and He, X. (2010) S-DIVA (statistical Dispersal-Vicariance Analysis): A tool for inferring biogeographic histories. Molecular Phylogenetics and Evolution, 56: 848- 850. 18 Historical biogeography of lancehead snakes of Bothrops alternatus group (Serpentes, Viperidae) Daniela P. Araújo1,5, Priscylla M. Arruda2, Cristiano C. Nogueira2, Silvio S. Nihei3 & Ricardo J. Sawaya4,5 1 Programa de Pós-Graduação em Biologia Animal, Instituto de Biociências, Letras e Ciências Exatas, Universidade Estadual Paulista (UNESP), 15054-000, São José do Rio Preto, São Paulo, Brazil. 2 Museu de Zoologia da Universidade de São Paulo (MZUSP), Laboratório de Herpetologia. Av. Nazaré, 481, 04263-000, São Paulo, São Paulo, Brazil. 3 Departamento de Zoologia, Instituto de Biociências, Universidade de São Paulo. Rua do Matão, Travessa 14, 101, Cidade Universitária, 05508-900, São Paulo, São Paulo, Brazil. 4 Departamento de Ciências Biológicas, Universidade Federal de São Paulo (UNIFESP), 09972–270, Diadema, São Paulo, Brazil. Corresponding author: sawayarj@gmail.com 5 Corresponding authors: aradany007@yahoo.com.br sawayarj@gmail.com Running Head: Historical Biogeography of Bothrops alternatus group Word count: 6.371 19 ABSTRACT Aim. To infer the origin and to analyze the biogeographical processes underlying the evolution of geographic distribution of lancehead snakes of Bothrops alternatus group. To discuss the possible ecological and historical factors related to the diversification of this clade in open and forested landscapes. Location. Central and southern South America. Methods. We revised and georeferenced point-locality records of species of Bothrops alternatus group by examination of voucher specimens housed in zoological collections and revision of literature. We used a phylogeny including five species of the group. The biogeographical units were delimited through intersection of species occurrence and biogeographical provinces and Brazilian relief units. Ancestral distributions were inferred through Dispersal-Vicariance-like (DIVA-like) analysis and Dispersal-Extinction- Cladogenesis models (DEC), and both models including one additional parameter of jump dispersal, implemented in BioGeoBEARS package. Ancestral ranges were estimated in two scenarios, with and without relative probabilities of dispersal and the possibility to disperse among in four time slices of 23, 12, 7 and 0.19 million years before present. Results. The best evaluated model when not considered time slices was DIVA-like+j, suggesting the origin of the group in Chaco-Pampa, Cerrado and Araucaria units. The DEC+j was the best evaluated model when considering time slices, predicting the origin of the group in Cerrado and Araucaria units. Both models suggest a mixed origin of the clade in open and forested areas. The most important process for both models was jump dispersal. Main conclusions. The diversification of the group must have occurred simultaneously in open and forested areas. Drier climates and replacement of forests by open vegetation in Miocene, as well as the Transcontinental Sea in southern South America, must have favored the expansion and diversification of the group specially in open vegetation. 20 Keywords. Ancestral range, Bothrops, Jump dispersal, BioGeoBEARS, Open landscape, Forest, Araucaria forest, Cerrado, Transcontinental Sea. 21 INTRODUCTION The presence or absence of a particular species in a certain place is related to its ecological requirements and evolutionary history (Endler, 1982). The historical biogeography is concerned with evolutionary processes that happened over millions of years, usually on large areas or even in global scale (Crisci, 2001). Different methods in historical biogeography allow the reconstruction of the evolutionary history of certain taxa or clade based on phylogenetic hypotheses (Cox & Moore, 2005). The phylogenetic relationships have been used as important tools to explain distributional patterns as well as to reconstruct ancestral geographical distributions (Ree & Sanmartin, 2009). The Dispersal-Vicariance Analysis (DIVA) and Dipersal-extinction-cladogenesis models (DEC) are among the most popular methods to reconstruct ancestral ranges (Yu et al., 2010; Matzke, 2013a). DIVA is a method based on parcimony that do not require prior knowledge on history of areas inhabited by species (Nylander et al., 2008). The optimization of ancestral ranges is based on a cost three-dimensional matrix matched to the processes, where dispersal and extinction cost one each, while vicariance have no costs, which favors vicariant speciation patterns (Ronquist, 1997). All possible area combinations in DIVA are equally allowed in ancestral reconstructions even when they are disjunct (Ronquist & Sanmartin, 2011). Dipersal-extinction-cladogenesis model is a parametric maximum likelihood framework (Ree et al., 2005; Ree & Smith, 2008) similar to the DIVA in some features (Ronquist & Sanmartin, 2011). But the available knowledge about the Earth’s history could be inserted in the model and ancestral range reconstructions, as connections among areas could be accounted for different relative dispersal rates in different time slices (Ree et al., 2005; Ree & Smith, 2008; Ree & Sanmartín, 2009; Kodandaramaiah, 2010). This method treats dispersal and extinction as stochastics processes, responsible for expansion and 22 contraction of distributional ranges, respectively (Ree & Sanmartín, 2009). The ancestral reconstruction in both DIVA and DEC may recover widespread ancestral ranges (Kodandaramaiah, 2010), and furthermore may produce similar results (Clayton et al., 2009). Dispersal has been considered as noise, stochastic in nature, and meaningless to explain general biogeographical patterns (Morrone & Crisci, 1995; Humpries & Parenti, 1999). However, the growth of the knowledge in molecular biology and divergence dates in phylogenies, indicate that some clades are too young to have their histories explained exclusively for vicariance (de Queiroz, 2005). And dispersal has been considered recently a more important process even to explain distributional patterns of species occurrence that once have been attributed to vicariance (de Queiroz, 2005). That strengthened the dispersionist ideas and methods in historical biogeography, that increasingly have consider dispersal in their frameworks (Santos, 2007). The BioGeographic Bayesian and likelihood Evolutionary Analysis with R Scripts (BioGeoBEARS) is an R package which estimates ancestral ranges and implements a new parameter on the analysis, the jump dispersal (Matzke, 2013b). This package runs DEC and more two models based on DIVA and Bayarea, respectively (Matzke, 2013b). The addiction of this new parameter has been contributing to the discussion on the importance of jump dispersal, which is a kind of dispersal usually in long distances (Lomolino et al., 2010). Although common in islands, some authors do not accept it and prefer alternative explanations as land bridges and stepping-stones islands (Lomolino et al., 2010). There are also evidence of dispersal through long distances in the continent (Crudden, 1966), and BioGeoBEARS implementations have shown that jump dispersal could be relevant both in continents and islands (Matzke, 2014). The lancehead snakes of genus Bothrops belongs to the Viperidae family and Crotalinae subfamily, and includes presently 48 species distributed throughout Americas, 23 from Mexico to Argentina (Campbell & Lamar, 2004; Carrasco et al., 2012). Recent phylogenies provide the division of the genus in monophyletic groups (Wüster et al., 2002; Fenwick et al., 2009; Carrasco et al., 2012). The B. alternatus group is recovered in all these phylogenies. Most of the lancehead snakes are forest dwellers (Martins et al., 2001; Campbell & Lamar, 2004; Carrasco et al., 2012). However, species of the B. alternatus group inhabit mainly open vegetation areas from the Brazilian savanna to the Patagonia (Hoge & Belluomini, 1964; Campbell & Lamar, 2004; Carrasco et al., 2009; Carrasco et al., 2010; Leão et al., 2014) The Bothrops genus is morphological and ecological diverse (Martins et al., 2001) and this might be the result of a single adaptive radiation occurred in South America (Carrasco et al., 2012). Dispersal events seem to have been important to the origin of Crotalinae snakes in Americas. The subfamily probably originated in Asia with a single dispersal to North America from northeastern Asia to Alasca through the Bering Land Bridge (Parkinson, 1999; Parkinson et al., 2002; Wüster et al., 2002; Wüster et al., 2008). The dispersal from Asia to Alasca probably occurred during the late Oligocene - early Miocene (28 to 23 Million years before present; Ma bp) when Bering Land Bridge was covered by deciduous forests (Sanmartín et al., 2001). The colonization of the New World was marked by faunal exchanges among North, Central and South America (Parkinson, 1999; Parkinson et al., 2002; Wüster et al., 2002; Wüster et al., 2008). We aim herein to infer the area origin of B. alternatus group and to understand the roles of biogeographical processes responsible for the diversification of this clade in South America. We also propose to discuss how geological and climatic processes may have influenced the present distribution of the group. 24 METHODS Study group and phylogenetic hypotheses Recent phylogenetic studies have supported the existence of monophyletic species groups within Bothrops genus, as Bothrops alternatus group (Fenwick et al., 2009; Carrasco et al. 2012). This group comprises six species (B. alternatus, B. ammodytoides, B. cotiara, B. fonsecai, B. itapetiningae and B. jonathani) and is considered as basal within Bothrops. Most of its species inhabits open landscapes (Martins et al., 2001). Bothrops fonsecai and B. cotiara, however, inhabit Araucaria angustifolia and montane forests in southern and southeastern Brazil. Whereas Bothrops fonsecai can be found in grasslands and ecotones between forests and open habitats, B. cotiara is found inside forests or in their ecotones (Bérnils, 2009). The phylogeny adopted here (L. Alencar, com. pess.) comprises 264 species of the Viperidae family, and includes three subfamilies, Azemiopinae, Viperinae and Crotalinae. The genus Bothrops belongs to the subfamily Crotalinae and this phylogeny contains data for 36 species of Bothrops genus and three forms of Bothrocophias. The phologeny were constructed from 11 genes sequences, six mitochondrial (12S, 16S, cytb, cox1, nd2, nd4) and five nuclear (bndf, c-mos, jun, nt3, rag1) genes, available in GenBank and Barcode of Life Database. Phylogenetic relationships and divergence times among species were estimated simultaneously in a Bayesian framework implemented in software BEAST v1.7.5. (Drummond & Rambaut, 2007). According to this phylogeny, only five species were sampled for the B. alternatus group: B. alternatus, B. ammodytoides, B. cotiara, B. fonsecai and B. itapetiningae. There is no molecular data available for B. jonathani, a poorly known and rare species with only some specimens deposited in zoological collections (Carrasco et al., 2009). 25 Dataset and biogeografical units The dataset was built from locality information of preserved specimens deposited in scientific collections and literature data. The species were mapped in ArcGIS 9.3 (ESRI, 2008) on a modified shapefile of biogeographical provinces (Morrone, 2014) and one of the Brazilian relief (IBGE, 2002). This modification of biogeographical provinces corresponded to the fusion of Chaco and Pampa in a Chaco-Pampa unit, and Prepuna and Monte in Prepuna-Monte unit (Fig. 1). The fusion was providential to avoid wide distributions including numerous biogeographical units (see Ronquist, 1997; Ree et al., 2005; Matzke, 2013a). We have choosen only one area of the Brazilian relief shapefile, corresponding to Mantiqueira, because it better fit to the B. fonsecai distribution in the Parana province (Fig. 1). The biogeographical units of interest were defined through the intersection of the species occurrence points and areas of the shapefiles. We considered five biogeographical units: Prepuna-Monte (P; Prepuna and Monte provinces); Chaco-Pampa (C; fusion of Chaco and Pampa provinces); Cerrado (B); Araucaria forest (A); and Mantiqueira (M; Fig. 1). Bothrops alternatus has the most widespread range, distributed in three units, Chaco-Pampa, Cerrado and Araucaria forest, from central Brazil to central Argentina (Fig. 1). Botrhops ammodytoides is endemic from Argentina and is distributed in Prepuna-Monte and Chaco- Pampa. Bothrops itapetiningae inhabits the Cerrado in central, southeastern and southern Brazil. Bothrops fonsecai is found in highlands in the southeast Brazil, only in Mantiqueira unit. Bothrops cotiara inhabits southern Brazil and a small portion of the Missiones in Argentina, occurring only in Araucaria forest (Fig. 1). 26 Ancestral area estimation The origin and diversification of the B. alternatus group was inferred in BioGeoBEARS package (BioGeographic Bayesian and likelihood Evolutionary Analysis with R Scripts) implemented in the software R (Matzke, 2013b; R Core Team, 2013). BioGeoBEARS allows probabilistic inference of ancestral geographic ranges (Matzke, 2013a). It includes fully parameterized models implementing the most popular historical biogeography reconstruction methods. We run DIVA-like, DEC and both of these models including the parameter jump dispersal (DIVA-like+j and DEC+j) (Matzke, in review; Matzke, 2013a). The DEC model is the same reproduced by LAGRANGE (Ree et al., 2005; Ree et al., 2008). The DIVA-like is a likelihood version of the parsimonious DIVA (Ronquist, 1997). In DIVA-like analysis two parameters might be changed. One is the parameter s, controlling the probability of subset sympatry, for which we employed zero, because the original DIVA does not present this event (Matzke, 2013a). The other is the parameter mx01v, controlling the relative probability of different descendent range size, for which we employed 0.5, meaning equal probability to all descendent range sizes during vicariance (Matzke, 2013a). In BioGeoBEARS, models usually include two free parameters, the extinction and dispersal, while DIVA-like+j and DEC+j presents the additional parameter j corresponding to jump dispersal or a founder event speciation (Matzke, 2013 a,b). The jump dispersal was set as a free parameter, i.e. were estimated from the analysis. Additionally, BioGeoBEARS allows comparisons of the different models, through a likelihood ratio test and Akaike Information Criterion (Matzke, 2013b). We test herein two different scenarios for DEC, DEC+j, DIVA-like and DIVA-like+j. In the first framework, we only constrained the ancestral maximum range size to three, corresponding to the range of B. alternatus, the most widespread species of the group. 27 However, dispersal was equally allowed among all five areas. In the second framework, we constrained both the ancestral maximum range size to three and dispersal relative probabilities in a stratified analysis. Furthermore, we only allowed dispersal between neighbor areas. The time period was divided in four slices according to the southern South America history (see discussion): 0.19 Ma, 7 Ma, 12 Ma and 23 Ma. The incursion of the Transcontinental sea was restricted to the extreme south of South America 23 Ma before present (Le Roux, 2011). For this time slice, dispersal between Prepuna-Monte and Chaco-Pampa, Chaco-Pampa and Cerrado, and Chaco-Pampa and Araucaria forest were allowed. In the second time slice, 12 Ma, dispersal between Prepuna-Monte and Chaco-Pampa, Chaco-Pampa and Cerrado, and Chaco-Pampa and Araucaria forest were not allowed, because the Transcontinental sea expanded to the north of Argentina flooding those areas (see Le Roux, 2011). In the third time slice, 7 Ma, we allowed again the dispersal among areas divided by the Transcontinental sea, as it recued (Le Roux, 2011). In those three slices the dispersal between Araucaria forest and Mantiqueira was disallowed, as Araucaria only has started to expand in 0,19 Ma, when the Mantiqueira and Araucaria forest were probably in contact, due to its expansion in highlands in southern Brazil (Ledru & Stevenson, 2012). In this last time slice we allowed dispersal between Mantiqueira and Araucaria forest. RESULTS In the first scenario, with dispersal equally allowed among all five areas and only the ancestral range constrained to the maximum of three areas, models DEC, DEC+j and DIVA- like+j showed similar ancestral range reconstructions. DEC model presented generally wide ancestral ranges for all nodes (see Appendix S1 in Supporting information). As the original DIVA, DEC provides wide ancestral ranges, especially on the root of the tree (Ronquist, 1997; Ree et al., 2005). However, for DIVA-like the ancestral ranges were not so widespread. 28 The DIVA-like model estimated a disjunct range for the ancestral of the group, corresponding to Prepuna-Monte and Cerrado. The other three models support a widespread ancestral range for the group, in Chaco-Pampa, Cerrado and Araucaria. The two models with jump dispersal only differ on estimation of the ancestral of B. fonsecai and B. cotiara, with DIVA-like+j supporting an ancestral range for those forest dweller species in Mantiqueira, and DEC+j in Araucaria. The likelihood ratio test and the Akaike Information Criterion (AIC) compared the results of similar models (e.g. DEC versus DEC+j) and the four models, respectively. The likelihood ratio favors DEC+j when compared to DEC (p = 0.008; Table 1), and DIVA-like+j when compared to DIVA-like (p = 0.029). Both models with jump dispersal presented best fit when comparing to models without this parameter (see the likelihood ratio tests and AIC values in Table 1). However, DIVA-like+j provided the best fit to the data (AIC = 25.2; see Table 1). We then have chosen the DIVA- like+j ancestral range estimation to discuss the origin and diversification of B. alternatus group, but for alternative results from other models see Supporting Information. The B. alternatus group originated from a widespread ancestral which lived in open vegetation areas including Chaco-Pampa and Cerrado, and the forested area Araucaria (Figure 2). The diversification of the group is related to events of vicariance, dispersal and jump dispersal. The extinction process was not predicted by DIVA-like+j model. A vicariant event followed by dispersal towards the southernmost area probably separated the common ancestral of the group and is related to the origin of B. ammodytoides, in Prepuna-Monte and Chaco-Pampa, and the ancestral of the other four species. This ancestral must be distributed in the same area of the ancestral of the group and has originated B. alternatus, inhabiting the same widespread area, and the ancestral of B. itapetiningae, B. fonsecai and B. cotiara, restricted to the Cerrado area, through vicariance followed by dispersal events. This ancestor, by jump dispersal, gave origin to B. itapetiningae, an endemic Cerrado species, and the 29 ancestral of the B. fonsecai and B. cotiara, supposed to occur in Mantiqueira. The ancestor of these two species, which are the only forest dweller of the group, must have originated B. fonsecai and B. cotiara, through in Mantiqueira and Araucaria forests, respectively. In the second scenario, the relative probabilities of dispersal and the possibility to disperse among areas in four time slices, and the maximum ancestral range constrained to three areas, DEC and DIVA-like models provided exactly the same ancestral range estimation for all nodes, as well as both models including jump dispersal (Appendix 1). According to those models, only the ancestor of B. fonsecai and B. cotiara lived in three areas and was the most widespread distributed ancestral. Both models including jump dispersal (see DIVA- like+j in Fig. 2 S1 of Supplementary Information, and DEC+j in Fig. 2 S1) predicted ancestrals with maximum range sizes of two areas. The likelihood ratio test unfavour DEC in favor to DEC+j (p = 0.026; see Table 1). However, there is no difference between DIVA-like and DIVA-like+j (p = 0.091; Table 1). Comparison among the four models indicates DEC+j as the best fit model (AIC = 31.07; Table 1). In this second scenario, we consider DEC+j to discuss the origin and diversification of the B. alternatus group, but for alternative results from other models see Supplementary Information . According to this second scenario, B. alternatus group must have originated from an ancestral distributed in Cerrado and in Araucaria, which would be a mixture of open and forested areas (Figure 2). An event of jump dispersal would be related to the cladogenensis between the B. ammodytoides, distributed in Prepuna-Monte and Chaco-Pampa, and the ancestral of other four species, distributed in the same areas, Cerrado and Araucaria. This ancestral has originated B. alternatus, expanding its range to the Chaco-Pampa, and the ancestral of B. itapetiningae, B. fonsecai and B. cotiara through jump dispersal, as this ancestor would be restricted to the Mantiqueira forests. This ancestral from the Mantiqueira 30 must have originated B. itapetiningae in another event of jump dispersal, and the ancestral of B. fonsecai and B. cotiara through a dispersal event, and distributed in Cerrado and Mantiqueira. From this last ancestor, an extinction event in Cerrado would be related to the present distribution of B. fonsecai in Mantiqueira (see Fig. 2), and another jump dispersal to the distribuiton of B. cotiara in Araucaria forest. DISCUSSION The estimation of the origin and diversification of the B. alternatus group in two scenarios showed different best fit models. The best fit models in both scenarios included jump dispersal as a most evident process. As demonstrated for other clades, the addiction of jump dispersal seems to improve models not only for island but also for continental clades (Batalha-Filho et al., 2014; Buckner et al., 2015; Matzke, in review). DIVA-like+j predicted vicariance and common dispersal in the first scenario, without time slices considered, however it did not forecast extinction, which is also observed for the parsimonious DIVA (Ronquist, 1997) . The origin of the clade might have been associated with an ancestor widespread distributed in Chaco-Pampa, Cerrado and Araucaria (first scenario, DIVA-like+j), or only in Cerrado and Araucaria (second scenario, DEC+j and considering time slices). Although different, both models predict an ancestor living in open and forested areas. The B. alternatus group must have originated in the middle Miocene (12 Ma) when southern South America was passing through important changes in climate, geology and vegetation structure. The elevation of the Andean Cordillera (65 to 3 Ma) has contributed to affect climate in South America forming a barrier to atmospheric circulation in Southern Hemisphere (Blisniuk et al., 2005). A cycle of drier and cooler climate was being established and vegetation was getting drier and xeric (Gottsberger & Silberbauer, 2009). 31 Another important historical event in South America on that time was the Transcontinental Sea, which successively expanded and contracted, during the early Paleocene to middle Miocene (65 to 13,8 Ma; Le Roux, 2011). The largest area covered by the Transcontinental Sea occurred during the early - middle Miocene (23 to 13.8 Ma; Le Roux, 2011), when B. alternatus group has originated. This event was probably importantly related to the first two cladogenetic events, especially because it splitted southern South America, spreading over eastern Argentina, western Uruguay, southern Paraguay and southeastern Bolivia (see Le Roux, 2011; Ortiz-Jaureguizar & Cladera, 2006). Then, the Transcontinental Sea could be responsible for splitting Chaco-Pampa from Prepuna-Monte on south, and Cerrado and Araucaria on north. We considered the invasion of the Transcontinental Sea for the analysis in the second scenario, disallowing dispersal in one time slice (12 to 7 Ma), from Chaco-Pampa to Prepuna-Monte and from Chaco-Pampa to Araucaria and Cerrado. This may have been related to the predicted jump dispersal of DEC+j in second scenario, when comparing to the DIVA-like+j of first scenario. The ancestral range estimation of DEC+j favored jump dispersal in most nodes of the phylogeny. For DIVA- like+j, jump dispersal was predicted for the two cladogenetic events that originated B. itapetiningae and B. fonsecai and B. cotiara, respectively. The first cladogenetic event must have occurred around 12.5 Ma, and the ancestor of the group gave origin to the B. ammodytoides, in Prepuna-Monte and Chaco-Pampa, and the ancestor of the remaining four species was distributed in the same area of the ancestor for both scenarios (see Figure 2). During this time, besides the disjunction of the Prepuna-Monte and Chaco-Pampa by the Transcontinental Sea, these areas have presented a shrubby and herbaceous landscape (Palazzesi & Barreda, 2012). The transformation of Patagonia, in Monte-Prepuna area, from a predominant forested area to shrubby and herbaceous habitats has occurred in between late Oligocene and early Miocene (28 to 23 Ma; Barreda & 32 Palazzesi, 2007). Open vegetation has dominated the landscape during the late Miocene (around 10 Ma), but grasslands probably only has emerged much latter in Patagonia, on the Quaternary (from 2.6 Ma; Palazzesi & Barreda, 2012). The latter domination by grasslands formation in South America reflects the opposed belief that the radiation of the hypsodont mammals have occurred in an environment dominated by grasses and thus latter grasslands formation (Palazzesi & Barreda, 2012). Nowadays, the Patagonia presents xeric, rocky and sandy areas where Bothrops ammodytoides inhabits (Carrasco et al., 2010) and that must be a different environment from its origin in the middle Miocene (12.5 Ma). During the second cladogenetic event (around 9 Ma), Chaco-Pampa was still isolated by the Transcontinental Sea, and only regressed in the late Miocene, around 7 Ma (Le Roux, 2011). In the second scenario, this disjunction was probably related to the main difference in ancestor range estimation between the scenarios. The predictions show the ancestor of the remaining species except B. ammodytoides living in open and forested areas in Cerrado and Araucaria (in the second scenario), and the addition of Chaco-Pampa in the first scenario, probably because of this absence of dispersal restriction to Chaco-Pampa. The restriction to dispersal in time slices probably reduced the tendency of widespread ancestral range estimation and disfavored disjoint distributions. And then this ancestor has originated B. alternatus and the ancestor of B. itapetiningae, B. cotiara and B. fonsecai, by jump dispersal in the second scenario. Bothrops alternatus is the most widespread species of the group, distributed in Chaco- Pampa, Cerrado and Araucaria. It is probably the most habitat generalist species of the group, inhabiting grasslands, wetlands or riparian areas in the Cerrado (Sawaya et al., 2008) or its ecotones with Atlantic forest ecotones as observed in Araucaria. This feature may have been crucial to B. alternatus reach a so widespread range presently. One of the most striking difference in ancestral range estimation for the different scenarios is the distributional area of 33 the ancestor of B. itapetiningae, B. cotiara and B. fonsecai. In the first scenario, it inhabited the Cerrado, the Brazilian savanna, whereas in the second scenario it occurred in a forested area, the Mantiqueira. This ancestor, originated B. itapetiningae, a habitat specialist of open Cerrado grasslands (Leão et al. 2014), and the ancestor of B. cotiara and B. fonsecai by either jump or simple dispersal in the first and second scenarios, respectively. The last cladogenetic event is related to the origin of B. fonsecai and B. cotiara, the forest dweller species of the group. The ancestor of those species, according to the first scenario was a forest species, living in Mantiqueira, and an event of jump dispersal might have originated B. cotiara restricted to Araucaria and B. fonsecai in Mantiqueira. However, for the second scenario, this ancestor inhabited both Cerrado and Mantiqueira. An extinction might have occurred in Cerrado and a jump dispersal originated B. cotiara, restricted to Araucaria and B. fonsecai, in Mantiqueira presently. Although only the second scenario includes a constraint in dispersal between Mantiqueira and Araucaria, both scenarios suggest jump dispersal related to this cladogenetic event. Data on the history of Araucaria angustifolia are scarce during middle Miocene to middle Pliocene periods (13 to 2 Ma). However, there is data available for the Late Oligocene (28 Ma) for Araucariaceae, one of the three dominant families in northwestern Patagonia (Barreda and Palazzesi, 2007). And for the late Miocene (7 Ma), there are records of A. angustifolia in the Parana Formation (Barreda & Palazzesi, 2007). However, there is no data available on the presence of A. angustifolia in Araucaria and Mantiqueira units until Pleistocene (around 190.000 years ago). We then can not ascertain if Araucaria and Mantiqueira units were as they are in the present covered by Araucaria moist forests. The expansion and contraction of Araucaria forest must have occurred more recently than the last cladogenetic event of the B. alternatus group, in the last glacial-interglacial cycle (Ledru & Stevenson, 2012). The Araucaria expanded twice during this period, from 190,000 to 130,000 34 and from 80.000-45.000 years ago (Kershaw et al., 2007). And in the second scenario, we only allowed dispersal between Mantiqueira and Araucaria after 190,000 years ago. Both scenarios suggest jump dispersal in the cladogeneses of B. fonsecai and B. cotiara which is in accordance to the A. angustifolia history in southern South America and unconforming to the assumption that these species were originated in a vicariant event during the contraction of the A. angustifolia forests. Both scenarios suggested the origin of the B. alternatus group simultaneously in forested and open vegetation. The diversification of the group was supported by different processes, especially in the first scenario, however the jump dispersal was the most important and predicted in both scenarios. The addiction of the dispersal probabilities in time slices may have favored jump dispersal in the second scenario, especially disallowing the dispersal among the areas disjoint by the Transcontinental Sea. The climatic and vegetation structure changes may have influenced the diversification of the group. ACKNOWLEDGEMENTS D.P.A. thanks FAPESP for Masters fellowships. R.J. Sawaya thanks FAPESP (2011/ 50206- 9), CNPq, INCTTox and FADA-UNIFESP for financial support. We also thank Laura Alencar for kindly provide the unpublished phylogeny for Bothrops alternatus group, and M. Fernanda Calió and Fausto Barbo for very useful comments on earlier versions of the manuscript. This work part of D.P.A. master’s dissertation at Universidade Estadual Paulista (UNESP), Campus São José do Rio Preto. The authors declare no conflict of interest. 35 REFERENCES Barreda, V. & Palazzesi, L. (2007) Patagonian vegetation turnovers during the Paleogene- Early Neogene: Origin f arid-adapted floras. The Botanical Review, 73, 31-50. Batalha-Filho, H.; Pessoa, R.O.; Fabre, P.H.; Fjsedsa, J.; Irestedt, M.; Ericson, P.G.P.; Silveira, L. & Miyaki, C.Y. (2014) Phylogeny and historical biogeography of gnateaters (Passeriformes, Conopophagidae) in South America forests. Molecular Phylogenetics and Evolution, 79, 422-432. Bérnils, R.S. (2009) Composição e padrões de distribuição de Caenophidia (Squamata, Serpentes) das Serras Atlânticas e Planaltos do sudeste da América do Sul. Tese apresentada ao Programa de Pós-graduação em Ciências Biológicas (Zoologia). Universidade Federal do Rio de Janeiro. Rio de Janeiro. Blisniuk, P.M.; Stern, L.A.; Chamberlain, C.P.; Idleman, B. & Zeitler, P.K. (2005) Climatic and ecologic changes during Miocene surface uplift in the Southern Patagonian Andes. Earth and Planetary Science Letters, 230, 125-142. Buckner, J.C.; Alfaro, J.W.L.; Rylands, A.B. & Alfaro, M.E. (2015) Biogeography of the marmosets and tamarins (Callitrichidae). Molecular Phylogenetics and Evolution, 82, 413- 425. Campbell, J.A. & Lamar, W.W. 2004. The venomous reptiles of the western hemisphere. Cornell University Press, USA, volumes I e II. Carrasco, P.A.; Mattoni, G.I.; Leynaud, G.C. & Scrocchi, G.J. (2012) Morphology, phylogeny and taxonomy of South American bothropoid pitvipers (Serpentes, Viperidae). Zoologia Scripta, 41, 109-124. Carrasco, P.A.; Leynaud, G.C. & Scrocchi, G.J. (2010) Rediscription of the southernmost snake species, Bothrops ammodytoides (Serpentes: Viperidae: Crotaline). Amphibia- Reptilia, 31, 323-338. 36 Carrasco, P.A.; Harvey, M.B. & Saravia, A.M. (2009) The rare Andinean pitviper Rhinocerophis jonathani (Serpentes: Viperidae: Crotalinae): rediscription with comments on its systematics and biogeography. Zootaxa, 2283, 1-15. Clayton, J.W.; Soltis, P.S. & Soltis, D.E. (2009) Recent long-distance dispersal overshadows ancient biogeographical patterns in a pantropical angiosperm family (Simaroubaceae, Sapindales). Systematic Biology, 58, 396-410. Cox, C.B. & Moore, P.D. (2005) Biogeography: an ecological and evolutionray approach, Blackwell, Oxford, UK, 7th edition. Crisci, J.V. (2001) The voice of historical biogeography. Journal of Biogeography, 28, 157- 168. Cruden, R.W. (1966). Birds as agents of long-distance for disjunct plant groups of the temperate western hemisphere. Evolution, 20, 517-532. De Queiroz, A. (2005) The resurrection of oceanic dispersal in historical biogeography. Trends in Ecology and Evolution, 20, 68-73. Drummound, A.J.; Rambaut, A. (2007) BEAST: Bayesian evolutionary analysis by sampling tress. BMC Evolutionary Biology, 7, 1-8. Endler, J. A. (1982) Problems distinguishing Historical from Ecological factors in Biogeography. American Zoologist, 22, 441-452. Fenwick, A. M., Gutberlet Jr, R. L., Evans, J. A. & Parkinson, C. L. (2009) Morphological and molecular evidence for phylogeny and classification of South American pitvipers, genera Bothrops, Bothriopsis, and Bothrocophias (Serpentes: Viperidae). Zoological Journal of the Linnean Society, 156, 617-640. Gottsberger, G. & Silberbauer-Gottsberger, I. (2009). Tropical savannas – Introduction. Tropical biology and conservation management (ed. by K.D. Claro, P.S. Oliveira, & V. 37 Rico-Gray). International Comission on Tropical Biology and Natural Resources, volume X. Hoge, A. R. & Belluomini, H. E. (1964) Notas sobre Bothrops fonsecai (Hoge e Belluomini), Bothrops alternatus (Duméril, Bibron et Duméril) e Bothrops cotiara (Gomes). Memórias do Instituto Butantan, 30, 97-102. Humpries, C.J. & Parenti, L.R. (1999) Cladistic biogeography. Oxford University Press, Oxford, 2ª edição. IBGE. (2002) Mapa dos compartimentos de Relevo do Brasil. Brasília, 4ª edição. Kershaw, A.P.; Bretherton, S.C. & Van de Kaars, S. (2007) A complete pollen record of the last 230 Ka from Lynch’s Crater, north east Australia. Palaeography, Palaeoclimatology, Palaeoecology, 25, 23-45. Kodandaramaiah, U. (2010) Use of dispersal-vicariance analysis in biogeography – a critique. Journal of Biogeography, 37, 3-11. Leão, S.M.; Pelegrin, N.; Nogueira, C.C. & Brandão, R.A. (2014) Natural history of Bothrops itapetiningae Boulenger, 1907 (Serpentes: Viperidae: Crotalinae), an endemic species of the Brazilian Cerrado. Journal of Herpetology, 48, 324-331. Ledru, M. & Stevenson, J. (2012) The rise and fall of the genus Araucaria: A southern hemisphere climatic connection. Peopled landscapes: (Terra Australis 34) archaeological and biogeographic approaches to landscapes (ed. by S.G. Haberle & B. David), pp 241- 253. Blackwell Scientific Publications, Oxford. Le Roux, J.P. (2011) A review of Tertiary climate changes in southern South America and the Antarctic Peninsula. Part 1: Oceanic conditions. Sedimentary Geology, 247-248, 1-20. Lomolino, M.V.; Ridlle, B.R.; Whittaker, R.J. & Brown, J.H (2010) Biogeography. Sinauer Associates, Sunderland, Massachussets, 4ª edição. 38 Martins, M.; Araujo, M.S.; Sawaya, R.J. & Nunes, R. (2001) Diversity and evolution of macrohabitat use, body size and morphology in a monophyletic group of Neotropical pitvipers (Bothrops). Journal of Zoology, 254, 529-538. Matzke, N.J. (In review). Model selection reveals differences in cladogenesis processes operating in island versus continental clades. Journal of Biogeography. Matzke, N.J. (2014) Model selection in historical biogeography reveals that founder-event speciation is a crucial process in island clades. Systematic Biology, 63, 951-970. Matzke, N.J. (2013a) Probabilistic historical biogeography: new models for founder-event speciation, imperfect detection, and fossils allow improved accuracy and model-testing. Ph.D. thesis, Department Integrative Biology and Designated Emphasis in Computational and Genomic Biology, University of California, Berkeley, 1-240. Matzke, N.J. (2013b) Probabilistic historical biogeography: new models for founder-event speciation, imperfect detection, and fossils allow improved accuracy and model-testing. Frontiers of Biogeography, 5, 242-248. Morrone, J.J. (2014) Biogeographical regionalization of the Neotropical region. Zootaxa, 3782, 1-110. Morrone, J.J. & Crisci, J.V. (1995) Historical biogeography: introduction to methods. Annual Review of Ecology and Systematics, 26, 373-401. Nylander, J.A.A.; Olsson, U.; Alström, P. & Sanmatín, I. (2008) Accounting for Phylogenetic uncertainty in biogeography: a Bayes approach to Dispersal-Vicariance Analysis of the Thrushes (Aves: Turdus). Systematic Biology, 57, 257-268. Ortiz-Jaureguizar, E. & Cladera, G.A. (2006) Paleoenvironmental evolution of Southern South America during the Cenozoic. Journal of Arid Environments, 66, 498-532. Palazzesi, L. & Barreda, V. (2012) Fossil pollen records reveal a late rise of open-habitat ecosystems in Patagonia. Nature Communications, 1294, 1-6. 39 Parkinson, C.L., Campbell, J.A. & Chippindale, P.T. (2002) Multigene phylogenetic analysis of pitviper, with comments on their biogeography. Biology of the Vipers (ed.G.W. Schuett, M. Höggren, M.E. Douglas & H.W. Greene), pp. 93-110. Eagle Mountain Publishing, Eagle Mountain, UT. Parkinson, C.L. (1999) Molecular systematics and biogeographical history of pitvipers as determined by mitochondrial ribosomal DNA sequences. Copeia, 1999, 576-586. R Core Team. (2013) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Viena, Austria. ISBN 3-900051-07-0, URL http://www.R- project.org/. Ree, R.H. & Sanmartín, I. (2009) Prospects and challenges for parametric models in historical biogeographical inference. Journal of Biogeography, 36, 1211-1220. Ree, R.H. & Smith, S.A. 2008. Maximum likelihood inference of geographic range evolution by dispersal, local extinction, and cladogenesis. Systematic Biology, 57, 4-14. Ree, R.H.; Moore, B.R.; Webb, O.C. & Donoughue, M.J. (2005) A likelihood framework for inferring the evolution of geographic range on phylogenetic trees. Evolution, 59, 2299- 2311. Ronquist, F. & Sanmartín, I. (2011) Phylogenetic methods in biogeography. Annual Review of Ecology, Evolution and Systematics, 42, 441-464. Ronquist, F. (1997) Dispersal-vicariance analysis: a new approach to the quantification of historical biogeography. Systematics Biology, 46, 195-203. Sanmartín, I.; Enghoff, H. & Ronquist, F. (2001) Patterns of animal dispersal, vicariance and diversification in the Holartic. Biological Journal of the Linnean Society, 73, 345-390. Santos, C.M.D. (2007) On basal clades and ancestral areas. Journal of Biogeography, 34, 1470-1471. 40 Sawaya, R. J.; Marques, O.A.V. & Martins M. (2008) Composition and natural history of Cerrado snake assemblage at Itirapina, São Paulo state, southeastern Brazil. Biota Neotropica, 8, 127-149. Wüster W.; Peppin, L.; Pook, C.E. & Walker, D.E. (2008) A nesting of vipers: Phylogeny and historical biogeography of the Viperidae (Squamata: Serpentes). Molecular Phylogenetics and Evolution, 49, 445-459. Wüster W.; Salomão, M.G., Quijada-Mascarenãs, J.A. & Thorpe R.S. (2002) Origins and evolution of the South American pitviper fauna: evidence from mitochondrial DNA sequence analysis. Biology of the Vipers (ed.G.W. Schuett, M. Höggren, M.E. Douglas & H.W. Greene), pp. 111-128. Eagle Mountain Publishing, Eagle Mountain, UT. Yu, Y.; Harris, A.J. and He, X. (2010) S-DIVA (statistical Dispersal-Vicariance Analysis): A tool for inferring biogeographic histories. Molecular Phylogenetics and Evolution, 56, 848- 850. 41 BIOSKETCHES Daniela Petrim de Araújo is a Master in Animal Biology from Universidade Estadual Paulista, Brazil. Her research interests include biogeography, conservation, and natural history of South American snakes. Priscylla Arruda and Silvio S. Nihei research interests include Systematics of Diptera and Historical Biogeography. Cristiano Nogueira is a post-doctoral fellow at the Museu de Zoologia da Universidade de São Paulo, Brazil. His main research interest is Neotropical biogeography and conservation. Ricardo J. Sawaya is a professor of biogeography, zoology and ecology. His research interests include geographic distribution on both local and regional scales, ecophylogenetics, conservation and evolutionary history of organisms. Author contributions: D.P.A and R.J.S. conceived the ideas; C.N. organized the geographic distribution database; D.P.A. performed the analysis and wrote the first draft of the manuscript; all authors contributed with interpretation and discussion of the results, writing the final version of the manuscript. 42 Appendix TABLE Table 1: Comparisons of the likelihood between similar models (e.g. DEC versus DEC+j) and Akaike Information Criterion (AIC) amoung all models. The addiction of +j in the models means jump dispersal and the numbers correspond to the scenarios. The dispersal (d), extinction (e) and jump dispersal (j) values are available for each model. Model lnL p d e j AIC DEC (1) -13.5 0.038 0.036 0 30.9 DEC+j (1) -9.99 0.008 0.014 0 0.39 25.9 DIVA (1) -11.9 0.028 0 0 27.9 DIVA+j (1) -9.6 0.029 0.011 0 2 25.2 DEC (2) -15.0 0.073 0.040 0 34.0 DEC+j (2) -12.5 0.026 0.056 0.028 3 31.1 DIVA (2) -14.1 0.075 0.029 0 32.1 DIVA+j (2) -12.6 0.091 0.059 0.029 1.53 31.25 43 FIGURE CAPTIONS Figure 1: Geographic distribution of species of Bothrops alternatus group in some biogeografical units of South America, based on 3.265 records from scientific collections and literature data: B. alternatus (A); B. ammodytoides (B); B. itapetiningae (C), and B. fonsecai and B. cotiara. Figure 2: Reconstructions of ancestral geographic ranges of species of Bothrops alternatus group, implemented in BioGeoBEARS software in two scenarios. A - the best fit model estimation (DIVA-like+j) for the first scenario, with constrained ancestral maximum range size to three areas and dispersal equally allowed among areas; B - the best fit model estimation (DEC+j) for the second scenario, with constrained ancestral maximum range size to three areas and variable dispersal relative probabilities according to a stratified analysis in four time slices (0.19 Ma, 7 Ma, 12 Ma and 23 Ma; see details in Methods); C - Biogeographical units and correspondent colours of areas indicated in each node of reconstructions. 44 Figure1 45 Figure 2 46 SUPPORTING INFORMATION Historical Biogeography of lancehead snakes of the Bothrops alternatus group (Serpentes, Viperidae) Appendix S1 Figure 1 S1: The estimation of dispersal relative probabilities among five areas (P – Prepuna- Monte, C- Chaco-Pampa, B – Cerrado, M – Mantiqueira and A – Araucaria) in a stratified analysis in four time slices (0.19, 7, 12 and 23 Ma). The number 1 shows the allowed dispersal and the number 0 is when dispersal is not allowed. Figure 2 S1: Reconstructions of ancestral geographic ranges of species of Bothrops alternatus group, implemented in BioGeoBEARS software for the first scenario, with constrained ancestral maximum range size to three areas and dispersal equally allowed among areas. A - DEC model; B - DEC+j model; C - DIVA-like model. Figure 3 S1: Reconstructions of ancestral geographic ranges of species of Bothrops alternatus group, implemented in BioGeoBEARS software for the second scenario, with constrained ancestral maximum range size to three areas and variable dispersal relative probabilities according to a stratified analysis in four time slices (0.19 Ma, 7 Ma, 12 Ma and 23 Ma; see details in Methods). A - DEC model. B - DIVA-like model; C - DIVA-like+j model. 47 P C B M A 1 1 0 0 0 1 1 1 0 1 0 1 1 1 1 0 0 1 1 1 0 1 1 1 1 P C B M A 1 1 0 0 0 1 1 1 0 1 0 1 1 1 1 0 0 1 1 0 0 1 1 0 1 P C B M A 1 0 0 0 0 0 1 0 0 0 0 0 1 1 1 0 0 1 1 0 0 0 1 0 1 P C B M A 1 1 0 0 0 1 1 1 0 1 0 1 1 1 1 48 0 0 1 1 0 0 1 1 0 1 END 49 DEC model DEC+j model DIVA-like Ma (Million years) Ma (Million years) Ma (Million years) Figure 1 S1 50 DEC model DIVA-like model DIVA-like+j Ma (Million years) Ma (Million years) Ma (Million years) Figure 2 S1 51 CONCLUSÕES GERAIS O grupo Bothrops alternatus se originou a partir de um ancestral que vivia em áreas abertas e florestais, que incluem o Cerrado, o Chaco, o Pampa e as florestas de Araucária. Os melhores modelos de reconstrução de distribuição geográfica para dois cenários, com e sem fatias de tempo e diferentes probabilidades de dispersão entre áreas, produziram estimativas da distribuição ancestral semelhantes. Os ancestrais mais próximos à raiz da filogenia, para o cenário sem fatias de tempo, apresentaram distribuições mais amplas, com distribuição de até três áreas, o máximo permitido para a análise. O melhor modelo para o segundo cenário, com restrições de dispersão principalmente relativas à presença do Mar Transcontinental que fragmentou áreas no sul da América do Sul, apresentou ancestrais com distribuição mais restrita, ocorrendo em até duas áreas, até mesmo na raiz da filogenia. Uma das principais diferenças entre os dois cenários corresponde às distribuições do ancestral de B. itapetiningae, endêmica do Cerrado, e B. fonsecai e B. Cotiara, as únicas espécies que habitam áreas florestais. No primeiro cenário, este ancestral ocupava o Cerrado, enquanto no segundo cenário 2, ele era distribuído na Mantiqueira. Durante a origem e diversificação do grupo, a América do Sul sofreu mudanças climáticas e de estrutura da vegetação. O clima estava se tornando mais seco e as florestas estavam sendo substituídas por uma vegetação arbustiva. O Mar Transcontinental que invadiu o sul do continente, fragmentou Prepuna-Monte do Chaco- Pampa e do Cerrado e Araucaria deve ter causado eventos de dispersão por saltos, processo que parece ser mais importante em relação aos processos de dispersão simples, extinção e vicariância observados na história de diversificação deste grupo de serpentes. 52 APÊNDICE Figura 1: Filogenia dos gêneros Bothrops e Bothrocophias de acordo com L. Alencar (com. pess.). 53 Figura 2: Proposta de regionalização retirada de Morrone (2014). 54 Figura 3: Mapa dos compartimentos de Relevos do Brasil retirado de IBGE (2002), com destaque para a área em verde que corresponde a uma das unidades biogeográficas utilizadas neste trabalho: Mantiqueira. 55