DIVERSIFICAÇÃO EVOLUTIVA EM ANFÍBIOS AMAZÔNICOS

INSTITUTO NACIONAL DE PESQUISAS DA AMAZÔNIA - INPA
PROGRAMA DE PÓS-GRADUAÇÃO EM ECOLOGIA
DIVERSIFICAÇÃO EVOLUTIVA EM ANFÍBIOS AMAZÔNICOS
(Allobates, DENDROBATOIDEA): CARACTERES GENÉTICOS,
MORFOLÓGICOS E COMPORTAMENTAIS
IGOR LUIS KAEFER
Manaus, Amazonas
Outubro, 2012
IGOR LUIS KAEFER
DIVERSIFICAÇÃO EVOLUTIVA EM ANFÍBIOS AMAZÔNICOS
(Allobates, DENDROBATOIDEA): CARACTERES GENÉTICOS,
MORFOLÓGICOS E COMPORTAMENTAIS
ORIENTADORA: Dra. ALBERTINA PIMENTEL LIMA
CO-ORIENTADORA: Dra. IZENI PIRES FARIAS
Tese apresentada ao Instituto Nacional de
Pesquisas da Amazônia como parte dos
requisitos para obtenção do título de
Doutor em Biologia (Ecologia).
Manaus, Amazonas
Outubro, 2012
ii
BANCA EXAMINADORA DO TRABALHO ESCRITO
Nome (instituição)
Parecer
Andrew Crawford (Smithsonian Tropical Research Institute)
Não emitido
Heike Pröhl (Stiftung Tierärztliche Hochschule Hannover)
Aprovado
Jeffrey Podos (University of Massachusetts)
Aprovado
Vanessa Verdade (Universidade Federal do ABC)
Aprovado
Walter Hödl (Universität Wien)
Aprovado
BANCA EXAMINADORA DA DEFESA PÚBLICA DA TESE
Nome (instituição)
Parecer
Camila Ribas (Instituto Nacional de Pesquisas da Amazônia)
Aprovado
Marcelo Gordo (Universidade Federal do Amazonas)
Aprovado
Marina Anciães (Instituto Nacional de Pesquisas da Amazônia)
Aprovado
iii
K11
CDD 19. ed. 597.8044
Kaefer, Igor Luis
Diversificação evolutiva em anfíbios amazônicos (Allobates,
Dendrobatoidea): caracteres genéticos, morfológicos e comportamentais /
Igor Luis Kaefer.--- Manaus : [s.n.], 2012.
xviii, 131 f. : il. color.
Tese (doutorado) --- INPA, Manaus, 2012
Orientador : Albertina Pimentel Lima
Coorientador : Izeni Pires Farias
Área de concentração : Ecologia
1. Amazônia. 2. Anura. 3. Bioacústica. 4. Biogeografia.
5. DNA mitocondrial. 6. Filogeografia. I. Título.
CDD 19. ed. 597.8044
Sinopse:
Mecanismos históricos e geográficos responsáveis pela diversificação evolutiva foram
avaliados em três espécies de Allobates ao longo de suas áreas de distribuição na
Amazônia brasileira. Foram analisados aspectos como o grau de estereotipia e a
plasticidade dos sinais sexuais. Foram estudadas as relações genealógicas, a história
filogeográfica das populações, e a divergência relativa em caracteres genéticos (DNA
mitocondrial), morfológicos (medidas morfométricas externas) e comportamentais
(caracteres acústicos). Por fim, foi avaliado o papel de caracteres fenotípicos em atribuir o
pertencimento de indivíduos a seus respectivos agrupamentos genéticos.
Palavras-chave: Amazônia, Anura, bioacústica, biogeografia, DNA mitocondrial,
filogeografia.
iv
AGRADECIMENTOS
Sou grato à Albertina P. Lima não somente pela orientação, mas por ter me permitido e
incentivado a procurar novos horizontes na pesquisa, muito além daqueles inicialmente
traçados no projeto de tese.
À Izeni P. Farias pelo incondicional apoio em todas as etapas e pelos ensinamentos em
genética molecular que deverão fazer diferença ao longo do meu futuro acadêmico.
Ao Pedro Ivo Simões e Luciana K. Erdtmann que, além de grandes amigos e colegas de
laboratório, participaram em diversas etapas dessa caminhada e representam para mim
exemplos a serem seguidos na conduta científica.
Ao prof. Miguel Vences, que permitiu que eu concluísse a redação desta tese nas
dependências da Technische Universität Braunschweig.
Ao Instituto Nacional de Pesquisas da Amazônia (INPA). Primeiramente, por reunir
pesquisadores que me inspiraram a conhecer mais sobre esse vasto bioma. Além disso, por
fornecer infraestrutura e apoio logístico para a condução dos meus estudos.
Ao Programa de Pós-Graduação (PPG) em Ecologia do INPA. Em especial, às coordenadoras,
Flávia Costa e Claudia Keller, pela dedicação ao curso. Aos professores do Programa, cujo
legado vai além do que nos foi apresentado durante as disciplinas.
Aos colegas e amigos do PPG Ecologia, em especial aqueles ingressos em 2008. Hoje já nos
encontramos “dispersos” pelo mundo. Porém, continuamos unidos pelo forte sentimento de
amizade que se consolidou ao longo dos anos de convivência na floresta e na cidade.
À Andresa Melo, Beverly Franklin, Érica Magalhães, Giselle Campos, Itamara da Gama,
Rosirene Farias, Walmira da Paz, Maria Carmozina de Araújo e Rafael de Fraga por terem
descomplicado diversos trâmites burocráticos e pelas conversas animadas na hora do café.
O trabalho de campo, com amostragens pontuais ao longo da Amazônia Central, envolveu
dezenas de famílias que gentilmente me acolheram em suas casas e cederam suas
embarcações para a procura dos Allobates. Agradeço em especial à D. Irene Melo, Moisés S.
Melo, Arthur O. Nascimento, Priscila P. Correa, Raimundo N. Amorim e aos diversos outros
“Raimundos” que me auxiliaram nessa empreitada. Também agradeço aos motoristas Srs.
Lourival dos Santos e Arnoldo Souza pelas idas e vindas aos sítios de coleta. Pela companhia
v
em campo, sou grato à Aída Rodrigues, Ana Carolina Mello, Anelise Montanarin, Bruno M.
Tsuji-Nishikido, Francisco C. de Freitas e Maria Aparecida Carvalho. Pela linda e proveitosa
viagem pelos Rios Negro e Unini, agradeço ao Niro Higuchi e toda a equipe do Laboratório
de Manejo Florestal do INPA.
Sou grato à Universidade Federal do Amazonas através do Laboratório de Evolução e
Genética Animal (LEGAL) pela possibilidade de realizar as análises moleculares
apresentadas nesta tese. À grande família do LEGAL, com a qual tanto pude aprender: Mário
Nunes, Adriano Cantuária, Concy Santos, Deyla Oliveira, Fabíola Rodrigues, Maria Lizarazo,
Gabriela de Pinho, Carla Bantel, Valéria Machado, Kelmer Passos, Jaqueline Fortuna,
Roberta Canton, Jessica Souza, Roberta Oliveira, Natasha Meliciano, Olavo Colatreli, Rafaela
Cardoso, Fabio Muniz, Adam Leão, Nicole Dutra, Marina Anciães e Tomas Hrbek. Em
especial, à Waleska Gravena, Daniela Leroy e Edvaldo Mota por terem me acompanhado em
meus passos iniciais dentro do lab.
Aos pesquisadores que dedicaram seu tempo para avaliar uma ou mais das diversas etapas da
execução deste projeto, desde sua proposição junto ao curso até a avaliação do documento
final: Adrian Garda, Andrew Crawford, Camila Ribas, Heike Pröhl, Ilse Walker, Jeffrey
Podos, Marcelo Gordo, Marcelo Menin, Marina Anciães, Mario Cohn-Haft, Maristerra
Lemes, Vanessa Verdade e Walter Hödl. Agradeço também pelas construtivas contribuições
recebidas de Adolfo Amézquita e Janet Reid, bem como aquelas provenientes dos editores e
revisores anônimos dos periódicos aos quais os capítulos foram submetidos para publicação.
Ao Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) pela bolsa de
doutorado a mim concedida e pelo financiamento do projeto por meio dos processos CTAmazônia 553997/2006-8 e 575572/2008-6.
Agradeço imensamente aos amigos e colegas de faculdade que, a despeito de todas as
dificuldades decidiram, junto comigo, realizar o sonho de estudar na distante Amazônia:
Anderson Bueno, Cristian Dambros, Daiani Kochhann e Mauro Ortiz. Estou certo de que
nossa bravura rendeu e renderá excelentes frutos.
À Sonia Cechin, da Universidade Federal de Santa Maria (UFSM), por ter me contagiado com
sua paixão pelo estudo dos anfíbios e répteis. Aos colegas do Laboratório de Herpetologia da
UFSM pelo incentivo e companheirismo que se estendem até os dias atuais.
vi
Aos amigos de longa data que, mesmo distantes fisicamente, têm me apoiado e continuado a
fazer parte do meu dia-a-dia. Aos novos amigos que conquistei durante minha estadia em
Manaus, e que acabaram por tornar-se minha família manauara. Muito obrigado por fazerem
parte da minha vida! Espero tê-los comigo sempre!
E, principalmente, agradeço à minha família por ter me apoiado em todas as minhas decisões,
e por constantemente expressar seu orgulho pelos meus estudos. Felizmente, o amor que nos
une independe de quaisquer barreiras físicas.
vii
“Ansioso, esperei o amanhecer: a natureza, aqui,
além de misteriosa é quase sempre pontual."
Milton Hatoum
viii
RESUMO
Esse estudo objetivou elucidar mecanismos históricos e geográficos responsáveis pela
diversificação evolutiva em três espécies de Allobates ao longo de suas áreas de distribuição
na Amazônia brasileira. No primeiro capítulo, avaliou-se a potencial importância dos
caracteres acústicos do canto de anúncio de Allobates paleovarzensis na discriminação entre
indivíduos e populações. Além disso, foram acessados os efeitos de fatores individuais,
ambientais e geográficos na variabilidade dos componentes deste sinal sexual. Nenhum dos
caracteres acústicos analisados foi indicado como potencial sinal para reconhecimento social
ou seleção sexual na espécie estudada. Lado do rio e distância geográfica não afetaram
significativamente a variabilidade acústica, indicando que forças evolutivas estabilizadoras
locais devem ser importantes no processo de diferenciação do sinal sexual. O segundo
capítulo apresentou a caracterização do grau de variabilidade de cada componente acústico do
sinal sexual entre populações e entre as espécies irmãs e alopátricas Allobates nidicola e A.
masniger. Em adição, testaram-se os efeitos do tamanho corporal, da temperatura ambiental, e
da distância geográfica na variabilidade acústica do sistema de estudo. Propriedades
espectrais foram mais distintivas que propriedades temporais entre populações e entre
espécies, com maior distinção ao nível interpopulacional quando comparado ao nível
interespecífico. Houve acentuada plasticidade em caracteres temporais do canto de anúncio. O
efeito do Rio Madeira como barreira foi significativo entre todas as variáveis acústicas
analisadas. A ausência de efeitos de isolamento por distância indica que pressões seletivas
estabilizadoras locais devem ser mais importantes que a deriva genética na evolução da
diferenciação do canto de anúncio. No terceiro capítulo, investigaram-se mecanismos
evolutivos e fatores geográficos levando à especiação, avaliando-se as relações genealógicas,
a história filogeográfica e a divergência de caracteres genéticos, morfológicos e
comportamentais entre populações das três espécies. Foi observada significativa estruturação
genética populacional. A distribuição da variabilidade genética nos dois sistemas de estudo
exibiu assinaturas congruentes com a transposição de grandes rios amazônicos. A efetividade
do baixo Rio Madeira como uma barreira ao fluxo gênico foi maior do que a do médio
Amazonas. Uma grande porção da diferenciação genética foi correlacionada com distância
geográfica linear, evidenciando a importante atuação da deriva genética na diversificação do
marcador mitocondrial estudado. Os caracteres acústicos e morfológicos não responderam às
distâncias lineares entre populações. Em adição, análises de função discriminante falharam
em classificar corretamente os indivíduos analisados em seus agrupamentos genéticos de
acordo com seus fenótipos. Esse resultado é surpreendente levando-se em consideração a alta
estruturação genética observada, e sugere que limitações evolutivas podem estar impedindo a
diferenciação fenotípica. Em geral, a diversificação evolutiva refletiu efeitos de isolamento
por distância e de barreiras vicariantes, e a contribuição relativa de cada fator diferiu entre os
dois sistemas estudados.
ix
ABSTRACT
Evolutionary diversification in Amazonian amphibians (Allobates, Dendrobatoidea):
genetic, morphological and behavioral characters
This study aimed to elucidate the historical and geographical mechanisms responsible for
evolutionary diversification in three species of Allobates throughout their distributional areas
in Brazilian Amazonia. In the first chapter, the potential importance of the acoustic characters
of the advertisement call of Allobates paleovarzensis in discriminating among individuals and
populations was evaluated. In addition, I assessed the effects of individual, environmental and
spatial factors on the variability of the components of this sexual signal. None of the acoustic
characters analized was indicated as a potential cue for social recognition or sexual selection
in this species. River side and geographical distance did not significantly affect the acoustic
variability, indicating that stabilizing local evolutionary forces might be important in the
process of sexual signal differentiation. The second chapter presents the characterization of
the degree of variability of each component of the acoustic sexual signal among populations
and between the allopatric sister species Allobates nidicola and A. masniger. Additionally, I
tested the effects of body size, environmental temperature, and geographical distance on the
acoustic variability within the study system. Spectral traits were more effective than temporal
properties in distinguishing among populations and between species, with greater difference
at the interpopulation level when compared to the interspecific level. The temporal characters
of the advertisement call were highly plastic. The effect of the Madeira River as a barrier was
significant for all acoustic traits analyzed. The absence of isolation-by-distance effects
indicates that local stabilizing selection pressures might be more important than genetic drift
in the evolution of the advertisement call differentiation. In the third chapter, I investigated
evolutionary mechanisms and geographical factors leading to speciation, by assessing the
genealogical relationships, the phylogeographic history, and the divergence in genetic,
morphological and behavioral traits among populations belonging to the three species. There
was significant population genetic structure. The distribution of genetic variability in the two
study systems exhibited signatures consistent with the transposition of large Amazonian
rivers. The effectiveness of the lower Madeira River as a barrier to gene flow was higher than
that of the middle Amazon. A large portion of the genetic differentiation was correlated with
linear geographical distance, showing the important role of genetic drift in the diversification
of the mitochondrial marker studied. The acoustic and morphological traits did not respond to
linear distances between populations. In addition, discriminant function analyses failed to
correctly classify the individuals analyzed in their genetic clusters based on their phenotypes.
This result is surprising in view of the high genetic structure observed, and suggests that
evolutionary constraints may be preventing phenotypic differentiation. In general, the
evolutionary diversification reflected effects of isolation by distance and of vicariant barriers,
and the relative contribution of each factor differed between the two systems studied.
x
SUMÁRIO
LISTA DE TABELAS
xi
LISTA DE FIGURAS
xvi
INTRODUÇÃO GERAL
1
OBJETIVOS
4
Capítulo I – Artigo: Sexual signals of the Amazonian frog Allobates
5
paleovarzensis: geographic variation and stereotypy of acoustic traits
Capítulo II – Artigo: Beyond the river: underlying determinants of population
30
acoustic signal variability in Amazonian direct-developing Allobates (Anura:
Dendrobatoidea)
Capítulo III – Artigo: The early stages of speciation in Amazonian forest
52
frogs: phenotypic conservatism despite strong genetic structure
SÍNTESE
106
REFERÊNCIAS BIBLIOGRÁFICAS
109
ANEXOS
125
xi
LISTA DE TABELAS
Capítulo I
Tabela 1. Efeito da temperatura e do tamanho corporal nas propriedades acústicas de
22
Allobates paleovarzensis (n = 123 machos). Coeficientes de determinação das
regressões lineares (r2), estatísticas F e valores de significância (p) são apresentados.
Para tamanho corporal (CRC), propriedades acústicas foram ajustadas com relação à
temperatura. Valores de p significativos após correção de Bonferroni estão marcados
com asteriscos.
Tabela 2. Estatística descritiva de caracteres do canto de anúncio baseada em médias
23
gerais dos parâmetros acústicos calculadas a partir dos valores médios de 123 machos
registrados em dez localidades ao longo da distribuição de Allobates paleovarzensis,
representando variação natural entre 23.1 and 28.0 oC. Coeficientes de variação médio
intraindividual, médio intrapopulacional, e geral (CVwi, CVwp, CVo, %) de dez
propriedades do canto de anúncio (ajustadas para 25.4oC por meio de regressão linear)
de Allobates paleovarzensis. As propriedades estão classificadas por tipo com base
nos CVs médios intraindividuais de acordo com Gerhardt (1991).
Tabela 3. Análise hierárquica de (co-)variância de caracteres do canto de anúncio
24
(variáveis dependentes) de Allobates paleovarzensis. População (aninhada em lado do
rio) e lado do rio são as variáveis categóricas independentes. No modelo b, tamanho é
usado como covariável. Valores de p significativos após correção de Bonferroni estão
marcados com asteriscos.
Capítulo II
Tabela 1. Estatística descritiva dos caracteres acústicos analisados, com base em
43
médias individuais das propriedades do canto de 100 indivíduos machos, ajustadas
para 25.9 °C utilizando-se regressão linear. Coeficientes de variação (%, apresentados
como média ± DP) intrapopulacional (CVwp), intraespecífico (CVws), e geral (CVo)
de seis propriedades do canto de anúncio de Allobates nidicola e A. masniger.
Tabela 2. Efeito da temperatura e do tamanho corporal nos caracteres acústicos dos
indivíduos estudados (n = 100 machos). Coeficientes de determinação ajustados das
regressões lineares (r2), estatísticas F e valores de significância (p) são apresentados.
Para tamanho corporal, estatísticas foram calculadas utilizando-se caracteres acústicos
44
xii
ajustados para temperatura. Valores de p significativos após correção de Bonferroni
estão marcados com asteriscos.
Tabela 3. Análise hierárquica de (co-)variância de caracteres do canto de anúncio
45
(variáveis dependentes) de A. nidicola e A. masniger. População (aninhada em lado do
rio) e lado do rio (espécies) são as variáveis categóricas independentes. No modelo b,
tamanho é usado como covariável. Valores de p significativos após correção de
Bonferroni estão marcados com asteriscos.
Tabela 4. Testes de Mantel avaliando correlações entre distâncias acústica (DAco), de
46
tamanho corporal (DSVL), e geográfica (DGeo) de Allobates amostrados em dez
localidades na Amazônia Central. Por causa dos efeitos do rio, os testes também
foram restritos aos lados esquerdo e direito do Rio Madeira (cinco localidades cada).
Valores de p significativos após correção de Bonferroni estão marcados com
asteriscos.
Capítulo III
Tabela 1. Sítios de estudo na Amazônia brasileira com respectivos tamanhos
85
amostrais e coordenadas geográficas. Localidades tipo estão indicadas com asteriscos.
Os tamanhos amostrais correspondem ao número de Allobates amostrados para cada
classe de caracter.
Tabela 2. Análises de variância molecular (AMOVA) baseadas em fragmentos do
86
DNAr 16S mitocondrial. A distribuição relativa da variabilidade genética é
apresentada de acordo com os níveis hierárquicos de estruturação genética em cada
sistema de estudo de Allobates.
Tabela 3. Índices de fixação FST par-a-par (matriz inferior esquerda) e distâncias
87
genéticas médias Kimura-2-parâmetros (matriz superior direita). Os valores foram
calculados entre as localidades de amostragem de cada sistema de estudo de Allobates
na Amazônia brasileira. Valores de FST significativos estão marcados com asteriscos.
Tabela 4. Estatística descritiva de parâmetros de polimorfismo genético e resultados
de testes de neutralidade realizados em cada agrupamento genético de Allobates.
Análise Bayesiana de Estruturação Genética foi utilizada para estimar a composição
dos agrupamentos genéticos. n = tamanho amostral; h = número de haplótipos; S =
número de sítios segregantes; π = distância média par-a-par (± um desvio padrão)
entre amostras pertencentes ao mesmo agrupamento; D = D de Tajima; Fs = Fs de Fu;
88
xiii
R2 = R2 de Ramos-Onsins & Rozas. Valores significativos em testes de neutralidade
estão marcados com asteriscos.
Tabela 5. Testes de expansão demográfica realizados de acordo com os agrupamentos
89
genéticos de Allobates indicados pela BAPS. Valores foram inferidos a partir da soma
do desvio quadrado (SSD) entre distribuições de mismatch observadas e esperadas, e
do índice de raggedness de Harpending (Hri). Os testes não foram conduzidos no
agrupamento 4 de Allobates paleovarzensis devido ao pequeno número de amostras
que o compõe. n = número de amostras. Os valores não significativos de p de ambos
os testes indicam que a hipótese nula de expansão populacional através do tempo não
pode ser rejeitada.
Tabela 6. Testes de Mantel simples e parciais avaliando correlações entre distâncias
90
geográficas, fenotípicas e genéticas de Allobates entre as localidades estudadas na
Amazônia brasileira. Testes de Mantel simples são apresentados como “Matriz 1” X
“Matriz 2” e testes de Mantel parciais são apresentados como “Matriz 1” X “Matriz
2”.“Matriz Covariável”. GenD = distância genética; GeoD = distância geográfica;
MorD = distância morfológica; AcoD = distância acústica; AcoDwa = Distância
acústica sem ajuste de tamanho corporal; River = lado do rio (variável binária).
Correlações significativas são indicadas com asteriscos.
Recurso Online 1. Medidas do canto de anúncio de indivíduos de Allobates
paleovarzensis em cada localidade de estudo na Amazônia brasileira. Valores são
apresentados como média (acima) e desvio padrão (abaixo). Os caracteres do canto
são: Repetição da nota (NR, em notas/s); Duração da nota (ND, em s); Intervalo entre
notas, como o intervalo silencioso entre duas notas consecutivas de um canto (InI, em
s); Repetição do canto (CR, em cantos/s); Duração do canto (CD, em s); Intervalo
entre cantos, como o intervalo silencioso entre dois cantos consecutivos (IcI, em s);
Frequência máxima (pico), como a frequência de maior intensidade calculada para
toda a nota por meio da função power spectrum do programa Raven Pro 1.3 (MF, em
Hz); Frequência mais baixa (LF, em Hz); Frequência mais alta (HF, em Hz);
Modulação da nota, como a diferença entre as frequências mais alta e mais baixa da
nota (NM, em Hz). Os códigos das localidades de amostragem, os tamanhos amostrais
e as respectivas coordenadas geográficas são apresentados na Tabela 1.
98
xiv
Recurso Online 2. Medidas do canto de anúncio de indivíduos de Allobates nidicola
99
e A. masniger em cada localidade de estudo na Amazônia brasileira. Valores são
apresentados como média (acima) e desvio padrão (abaixo). Os caracteres do canto
são: Duração da nota (ND, em s); Intervalo entre notas (IN, em s); Frequência mais
baixa (LF, em Hz); Frequência mais alta (HF, em Hz); Modulação da nota, calculada
como a diferença entre as frequências mais alta e mais baixa (NM, em Hz); e
Frequência pico (PF, em Hz). Os códigos das localidades de amostragem, os tamanhos
amostrais e as respectivas coordenadas geográficas são apresentados na Tabela 1.
Recurso Online 3. Medidas morfométricas (em mm) de indivíduos de Allobates
100
paleovarzensis em cada localidade de estudo na Amazônia brasileira. Valores são
apresentados como média (acima) e desvio padrão (abaixo). Os caracteres
morfométricos são: Comprimento rostro-cloacal (SVL); Comprimento da cabeça, da
articulação da maxila até a ponta do focinho (HL); Largura da cabeça ao nível da
articulação da maxilla (HW); Comprimento do focinho, do canto anterior do olho até
a ponta do focinho (SL); Distância do olho à narina, do canto anterior do olho ao
centro da narina (EN); Distância entre narinas (IN); Comprimento do olho, do canto
anterior ao posterior (EL); Distância interorbital (IO); Diâmetro do tímpano (TYM);
Comprimento do antebraço, da extremidade proximal do tubérculo palmar até a borda
exterior do cotovelo flexionado (FAL); Comprimento do braço, da sua inserção no
tronco até a borda exterior do cotovelo flexionado (UAL); Comprimento da mão, da
extremidade proximal do tubérculo palmar até a ponta dos dedos I, II, III e IV
(HAND1, HAND2, HAND3 e HAND4); Largura do disco no dedo III da mão
(WFD); Comprimento da tíbia, da borda externa do joelho flexionado até o calcanhar
(TL); Comprimento do pé, da extremidade proximal do tubérculo metatarsal até a
ponta do dedo IV (FL); Comprimento do fêmur (LL); Diâmetro do tubérculo palmar
(DPT); Largura do tubérculo tenar (WTT); Largura do disco no dedo IV do pé
(WTD); Largura do dedo III da mão (WPF). Os códigos das localidades de
amostragem, os tamanhos amostrais e as respectivas coordenadas geográficas são
apresentados na Tabela 1.
Recurso Online 4. Medidas morfométricas (em mm) de indivíduos de Allobates
nidicola e A. masniger em cada localidade de estudo na Amazônia brasileira. Valores
são apresentados como média (acima) e desvio padrão (abaixo). As medidas
morfométricas são descritas no Recurso Online 3. Os códigos das localidades de
amostragem, os tamanhos amostrais e as respectivas coordenadas geográficas são
apresentados na Tabela 1.
101
xv
Recurso Online 5. Distribuição dos haplótipos de rDNA 16S de Allobates entre 20
102
localidades amostradas na Amazônia brasileira. Números de coleta de espécimes
testemunho (INPA-H) e números de acesso do GenBank são apresentados. As
localidades amostradas estão numeradas de acordo com a Tabela 1.
Recurso Online 7. Matriz de classificação da análise de função discriminante,
104
baseada em dados fenotípicos de machos de Allobates paleovazensis pertencentes a
quarto grupos genéticos. Grupos estão codificados de C1 a C4, de acordo com a
Tabela 4. Caracteres de forma corporal foram usados para discriminar indivíduos
morfologicamente. Medidas acústicas foram ajustadas para temperatura e para
temperatura e tamanho corporal simultaneamente por meio de regressões lineares. O
número e a percentagem de indivíduos corretamente atribuídos a cada grupo genético
são indicados.
Recurso Online 8. Matriz de classificação da análise de função discriminante,
baseada em dados fenotípicos de machos de Allobates nidicola e A. masniger
pertencentes a sete grupos genéticos. Grupos estão codificados de C1 a C7, de acordo
com a Tabela 4. Caracteres de forma corporal foram usados para discriminar
indivíduos morfologicamente. Medidas acústicas foram ajustadas para temperatura e
simultaneamente para temperatura e tamanho corporal por meio de regressões
lineares. O número e a percentagem de indivíduos corretamente atribuídos a cada
grupo genético são indicados.
105
xvi
LISTA DE FIGURAS
Capítulo I
Figura 1. Distribuição geográfica das localidades de amostragem na Amazônia
26
Central. 1 = Careiro, 2 = Janauacá, 3 = Manaquiri, 4 = Hiléia, 5 = Novo Airão, 6 =
Anamã, 7 = Anori, 8 = Codajás, 9 = Unini, 10 = Barcelos.
Figura 2. (A) Oscilograma e (B) espectrograma de três cantos de anúncio multinotas,
27
e (C) espectro de força de uma única nota de um macho de Allobates paleovarzensis
em sua localidade tipo em Careiro da Várzea, Brasil.
Figura 3. Coeficientes de variação intraindividuais (CVwi) para dez propriedades
28
acústicas do canto de anúncio de Allobates paleovarzensis. Retângulos cinza indicam
os 50% centrais dos dados (interquartil), e linhas horizontais representam os valores
medianos. Pontos representam valores discrepantes. Linhas verticais indicam o 10º e o
90º percentis.
Figura 4. Variação em propriedades do canto (ajustadas para 25.4 oC por meio de
29
regressão linear) e tamanho do corpo ao longo da distribuição de Allobates
paleovarzensis na Amazônia Central. Retângulos cinza indicam os 50% centrais dos
dados (interquartil), e linhas horizontais representam os valores medianos. Pontos
representam valores discrepantes. Linhas verticais indicam o 10º e o 90º percentis.
Capítulo II
Figura 1. Machos de A. nidicola (acima) e A. masniger (abaixo) vocalizando em suas
48
respectivas localidades tipo na Amazônia brasileira.
Figura 2. Localidades de amostragem visitadas neste estudo: 1 Km 12 da Estrada de
49
Autazes, localidade tipo de A. nidicola; 2 BR-319, km 260; 3 PPBio Manaquiri; 4 BR319 Tupana; 5 Vila Gomes; 6 Borba; 7 Novo Aripuanã; 8 Estrada para Apuí; 9
Jacareacanga; 10 PARNA da Amazônia, localidade tipo de A. masniger.
Figura 3. Representação gráfica das medidas acústicas adotadas neste estudo.
Oscilograma (a) e espectrograma (b) de uma sequência de cantos de um macho
topotípico de A. masniger (CRC = 18.5 mm; INPA-H 28196), registrado em
temperatura do ar de 23.7 °C. Espectro de força (c) de um único canto deste indivíduo.
As frequências mais alta e mais baixa foram medidas a 20 dB abaixo da intensidade
50
xvii
pico (linha pontilhada horizontal), valor no qual a energia do sinal pode ser
claramente distinguida do ruído de fundo. A modulação da nota foi calculada como a
diferença entre as frequências mais alta e mais baixa do canto.
Figura 4. Coeficientes de variação intrapopulacionais (CV, em percentagem) para
51
seis propriedades acústicas do canto de anúncio dos indivíduos estudados. Retângulos
cinza indicam os 50% centrais dos dados (interquartil), e linhas horizontais
representam os valores medianos. Linhas verticais indicam o 10º e o 90º percentis.
Note que os coeficientes de variação mais baixos são relacionados a caracteres
espectrais do canto.
Capítulo III
Figura 1. Localização geográfica dos pontos de amostragem (A) e rede de haplótipos
92
(B) para Allobates paleovarzensis. A rede foi construída a partir de 142 sequências de
DNAr 16S. O tamanho e a cor de cada elipse indicam a frequência e origem
geográfica dos indivíduos que apresentam aquele haplótipo. Pontos brancos e barras
transversais representam haplótipos intermediários não amostrados (faltantes). As
localidades amostradas estão numeradas de acordo com a Tabela 1.
Figura 2. Localização geográfica dos pontos de amostragem (A) e rede de haplótipos
93
(B) para Allobates nidicola e A. masniger. A rede foi construída a partir de 134
sequências de DNAr 16S. O tamanho e a cor de cada elipse indicam a frequência e
origem geográfica dos indivíduos que apresentam aquele haplótipo. Pontos brancos e
barras transversais representam haplótipos intermediários não amostrados (faltantes).
As localidades amostradas estão numeradas de acordo com a Tabela 1.
Figura 3. Gráfico de barras da Análise Bayesiana de Diferenciação Genética. (A) 142
sequências individuais de DNAr 16S de Allobates paleovarzensis de dez localidaes de
amostragem. (B) 134 sequências individuais de DNAr 16S de Allobates nidicola e A.
masniger de outras dez localidades. Cores distintas representam cada grupo genético
estimado. Indivíduos são exibidos de acordo com as localidades de amostragem, as
quais estão numeradas de acordo com a Tabela 1.
94
xviii
Figura 4. Árvore datada dos sistemas (A) paleovarzensis e (B) nidicola-masniger
95
baseada em fragmentos do gene DNAr 16S. Somente haplótipos únicos foram
incluídos. Estimativas de idade e intervalos de confiança (barras horizontais) das
divergências foram obtidos via BEAST (Bayesian Evolutionary Analysis Sampling
Trees). As probabilidades posteriores dos nós mais antigos são apresentadas.
Terminais foram coloridos de acordo com suas localidades de origem, as quais estão
representadas nas Figuras 1 e 2. Terminais pretos e brancos são grupos externos. As
barras verticais pretas e cinzas indicam a origem de cada haplótipo de acordo com o
lado do rio.
Recurso Online 6. Distribuições de mismatch obtidas a partir de diferenças par-a-par
de sítios nucleotídicos entre sequências de DNAr 16S que compõem os grupos
genéticos dos sistemas (A) paleovarzensis e (B) nidicola-masniger. Os grupos foram
determinados via Análise Bayesiana de Diferenciação Genética. Diferenças par-a-par
não foram calculadas para o Grupo 4 de Allobates paleovarzensis devido ao pequeno
número de amostras que o compõe.
103
1
INTRODUÇÃO GERAL
A integração entre distintas áreas do conhecimento tem permitido um melhor
entendimento a respeito de padrões e mecanismos de diferenciação entre populações de
diversos organismos (Crisci et al., 2003; Santini et al., 2012). A análise conjunta de diferentes
classes de caracteres permite elucidar mecanismos evolutivos levando à especiação, uma vez
que a história inferida a partir de sequências de DNA proporciona um contexto demográfico
histórico para avaliar a divergência em diversos atributos fenotípicos (Wiens, 2008; Campbell
et al., 2010; Guillot et al., 2012). Importantes avanços podem ser obtidos por meio da
investigação do padrão geográfico no qual a variabilidade de uma determinada classe de
caracteres é distribuída entre populações, assim como da ordem em que novidades evolutivas
aparecem (West-Eberhard, 1989; Dall, 1997; Panhuis et al., 2001). Tal cenário multicaráter
tem impulsionado esforços para a investigação das forças envolvidas na promoção e
manutenção da diversificação intraespecífica e do processo de formação de novas espécies.
Em adição à variedade de classes de marcadores, estudos considerando diferentes planos
organizacionais contribuem significativamente para a compreensão da diversificação
evolutiva, uma vez que esta pode ser acessada a partir de distintos níveis da hierarquia
biológica (Castellano et al., 2002; Diniz-Filho et al., 2008). Assim, espera-se que estudos
dirigidos à variabilidade de caracteres aos níveis de indivíduo, população e espécie possam
proporcionar respostas complementares a respeito da multiplicidade de forças evolutivas
envolvidas nesse processo.
A compreensão dos processos de especiação requer o estudo de fatores geográficos
que podem sustentar a origem e diversificação de linhagens, tais como vicariância e
isolamento por distância (Hutchison & Templeton, 1999; Zink et al., 2000). Uma vez que
efeitos de isolamento por distância são inferidos a partir de correlações entre distâncias
geográficas e de caracteres fenotípicos ou genotípicos, estas podem ser afetadas pela presença
de rios como barreiras, os quais podem levar a dissimilaridades desproporcionalmente
maiores do que aquelas produzidas pela distância linear. Na Amazônia, investigações sobre o
padrão de distribuição de espécies (e.g., Wallace, 1852; Haffer, 1969; Ayres & CluttonBrock, 1992; Hayes & Sewlal, 2004) e avaliações genealógicas em fina escala (e.g., CohnHaft, 2000; Funk et al., 2007; Solomon et al., 2008; Amézquita et al., 2009) têm inferido o
efeito desses dois fatores históricos em organismos terrestres, embora a contribuição relativa
atribuída a cada fator tenha variado. Assim, até o presente momento, nenhum padrão
2
geográfico ou força paleoambiental emergiu como uma explicação geral para a diversificação
da maioria dos grupos animais (Antonelli et al., 2010; Ribas et al., 2011).
Nesse contexto, anfíbios anuros figuram como excelentes modelos de estudo em
variação geográfica. A alta abundância dos anfíbios propicia amostragens densas ao longo de
suas áreas de distribuição, e o sinal sexual dos anuros (canto de anúncio) permite a utilização
de caracteres acústicos como marcadores populacionais (Wycherley et al., 2002; Lougheed et
al., 2006). No nível individual, caracteres acústicos são utilizados pelos anuros para
reconhecimento e escolha sexual (Gerhardt & Huber, 2002). Diferenças nesses sinais podem
aumentar dramaticamente o potencial evolutivo para especiação devido à possibilidade de
isolamento pré-zigótico (Panhuis et al., 2001; Pröhl et al., 2006; Boul et al., 2007). A
natureza multidimensional dos sinais sexuais torna os caracteres acústicos potencialmente
sujeitos a diferentes determinantes (Gerhardt, 1991; Prestwich, 1994; Gerhardt & Huber,
2002). Assim, ao decompormos esses sinais em suas propriedades, podemos medir o grau de
variabilidade de caracteres quantitativos, e acessar seu potencial para seleção sexual e
reconhecimento social, assim como entender o papel de determinados caracteres acústicos em
processos microevolutivos (e.g., Friedl & Klump, 2002; Gasser et al., 2009; Bee et al., 2010).
Entretanto, estudos bioacústicos comparativos devem considerar covariáveis que afetam o
sinal sexual, mas que não necessariamente carregam sinal filogenético. Visto que caracteres
acústicos apresentam plasticidade induzida pela temperatura ambiental e pelo tamanho do
corpo (Prestwich, 1994), faz-se necessária a detecção e a consideração desses efeitos para a
obtenção de medidas mais precisas de diferenciação acústica quando comparamos indivíduos,
populações ou espécies (Castellano et al., 2002).
Pelo fato de muitas vezes serem prontamente perceptíveis, caracteres morfológicos
têm servido historicamente como os marcadores mais amplamente utilizados para o estudo de
variação entre populações. Tal característica conferiu grande valor à morfologia no
desenvolvimento de conceitos e diagnoses em nível específico (Cronquist, 1978; Coyne &
Orr, 2004). Atualmente, a importância relativa desses caracteres em estudos evolutivos tem
diminuído em face do desenvolvimento de outros marcadores, em especial os de origem
molecular (Avise, 2004). Entretanto, uma abordagem interdisciplinar não deve prescindir da
utilização de dados morfológicos, quando disponíveis, para a identificação de processos
evolutivos e, frequentemente, para a proposição de rearranjos taxonômicos decorrentes de
análises em nível populacional (Dayrat, 2005; Buckley, 2009).
3
Ao longo das últimas décadas, testemunhou-se um desenvolvimento expressivo no
estudo do arranjo espacial de linhagens genéticas, com ênfase no entendimento de como o
isolamento genético no tempo e espaço desencadeia eventos de especiação e cria assinaturas
em larga escala na diversidade genética dentro de espécies e em conjuntos de espécies
proximamente relacionadas (Avise, 2009). Desde o cunho do termo filogeografia em 1987
(Avise et al., 1987), essa disciplina tem sido beneficiada por grandes avanços metodológicos
e conceituais na investigação de processos ecológicos e históricos que influenciam e explicam
a distribuição geográfica de organismos (Buckley, 2009; Avise, 2009). Atualmente, esta área
do conhecimento vem desenvolvendo métodos sofisticados dentro de um sistema rigoroso em
termos de obtenção de marcadores e de ferramentas estatísticas para o teste de hipóteses
definidas a priori (e.g., Nielsen & Beaumont, 2009; Emerson et al., 2010). Assim, a
filogeografia tem promovido uma verdadeira revolução no estudo da biodiversidade por
tornar-se um componente fundamental em biogeografia, e por estabelecer uma conexão entre
a investigação de processos micro e macroevolutivos, atuando assim como uma extensão
geográfica e temporal da genética de populações (Riddle & Hafner, 2004; Riddle, 2009).
Nos três capítulos apresentados a seguir, procuro elucidar mecanismos evolutivos e
fatores geográficos levando à diversificação evolutiva em dois sistemas de estudo. Para isso,
associei a utilização de diferentes classes de marcadores em um contexto que contemplasse os
níveis organizacionais intraindividual, intraespecífico e interespecífico. Como modelos de
estudo, utilizei três espécies recentemente descritas de Allobates, anuros diurnos e terrestres
pertencentes à família Aromobatidae (Grant et al., 2006): Allobates paleovarzensis Lima et al.
(2010), Allobates nidicola (Caldwell & Lima, 2003) e Allobates masniger (Morales, 2002).
Os estudos aqui apresentados constituem as primeiras avaliações de cunho comportamental
e/ou evolutivo envolvendo estas espécies. Foram amostradas 20 populações de Allobates em
duas áreas adjacentes da bacia amazônica, nos estados brasileiros do Amazonas e Pará, em
uma região caracterizada por grandes porções de florestas tropicais. Estas áreas compreendem
seções de dois grandes rios do bioma, o Amazonas e o Madeira, possibilitando assim avaliar a
contribuição relativa desses rios e do isolamento por distância como barreiras biogeográficas
para os organismos estudados. As hipóteses apresentadas em cada capítulo são testadas por
meio de enfoques populacionais utilizando marcadores genéticos, morfológicos e
comportamentais. Assim, o conjunto dos capítulos objetiva uma abordagem integrativa e
transdisciplinar para as questões apresentadas.
4
OBJETIVOS
Os objetivos gerais de cada capítulo foram os seguintes:
Capítulo I – Acessar a potencial importância dos caracteres acústicos do canto de anúncio de
Allobates paleovarzensis na discriminação entre indivíduos e populações ao longo da área de
ocorrência da espécie, e avaliar o efeito de fatores individuais, ambientais e geográficos na
variabilidade dos componentes do sinal sexual.
Capítulo II – Caracterizar o grau de variabilidade de cada componente acústico do sinal
sexual entre populações e entre espécies irmãs e alopátricas (Allobates nidicola e A.
masniger), e testar o efeito do tamanho corporal, da temperatura ambiental, e da distância
geográfica na variabilidade acústica do sistema de estudo.
Capítulo III – Elucidar mecanismos evolutivos e fatores geográficos que levam à especiação,
avaliando as relações genealógicas, a história filogeográfica e a divergência de caracteres
genéticos, morfológicos e comportamentais entre populações de três espécies de anfíbios
anuros da Amazônia Central.
18
Capítulo 1
Kaefer, I.L. & Lima, A.P. 2012. Sexual signals
of the Amazonian frog Allobates paleovarzensis:
geographic variation and stereotypy of acoustic
traits. Behaviour 149:15-33.
6
1
2
3 Sexual signals of the Amazonian frog Allobates paleovarzensis: geographic variation and
stereotypy of acoustic traits
4
5
6 Short title: Acoustic signal variability in Amazonian frogs
Igor Luis Kaefer1) & Albertina Pimentel Lima
7
8 Coordenação de Pesquisas em Ecologia, Instituto Nacional de Pesquisas da Amazônia, CP
9 478, 69011–970, Manaus, Amazonas, Brazil.
10
1)
Corresponding author’s e-mail address: [email protected]
11
12 Summary
13 Because of its close relationship with the process of evolutionary differentiation, it is expected
14 that geographic variability in acoustic sexual traits should be greater among than within
15 populations. This is particularly expected in organisms with typically high population genetic
16 structure and low dispersal abilities, such as anuran amphibians. We studied the acoustic traits
17 of the advertisement call in the small-sized dendrobatoid frog Allobates paleovarzensis
18 through its range in Central Amazonia. We accessed the variability of call traits from the
19 within-male to the among-population levels, and evaluated the degree of stereotypy of the call
20 characteristics. Call variability had comparable magnitudes within and among populations,
21 and was independent of the degree of stereotypy of call measurements. Therefore, none of the
22 call traits stood out as a potential cue for discrimination between populations. Spectral call
23 measurements were static and strongly related with body size, which explained between 30
24 and 35% of the variation of these acoustic traits. Temporal characters of the notes were
25 dynamic and influenced by environmental temperature (e.g., 27% of note rate variation),
26 whilst temporal measurements of the entire calls were not related to the co-factors analysed.
27 Both spectral and temporal call traits varied among populations and between sides of the
28 Amazon River. Our results also indicate that body size and sampling site jointly affected the
29 variability of the call traits. However, geographic distances among populations and the river
7
30 barrier had no significant effect on the overall acoustic variation, indicating that local
31 stabilising selective forces may be important in the process of call differentiation.
32 Key words: acoustic communication, advertisement call, Amazon River, biogeography,
33 Dendrobatoidea, intraspecific variation, signal stereotypy.
34
35 Introduction
36
Variability in secondary sexual characters is a pivotal component that underlies the
37 evolutionary framework in which selective forces take place, and changes in mate recognition
38 systems may even precede the evolution of other indicators of specific status (Ryan et al.,
39 1996; Panhuis et al., 2001). Therefore, analyses of geographic variation in intraspecific
40 communication systems have the potential to detect the development of differentiation among
41 populations, allowing researchers to derive testable predictions about the evolutionary
42 significance of such divergences (Gerhardt & Huber, 2002).
43
An acoustic signal conveys information about its sender, such as species identity, body
44 condition, and mate quality, and these characteristics have been decoded from the signals in a
45 variety of animal groups, such as monkeys (Fischer et al., 2004), bats (Behr et al., 2007),
46 crickets (Verburgt et al., 2011) and frogs (Chakraborty et al., 2010). Given that these signals
47 are often related to sexual recognition and other aspects of pre-mating reproductive isolation
48 (Gerhardt & Huber, 2002), their decomposition into their spectral and temporal properties
49 should provide clues regarding their potential in sexual discrimination among populations
50 (e.g., Pröhl et al., 2007; Gomez et al., 2011) and even individuals (e.g., Gasser et al., 2009;
51 Bee et al., 2010; Melendez & Feng, 2010). Bioacoustic traits of individuals can be classified
52 as static or dynamic according to the continuum of coefficients of variation observed
53 (Gerhardt, 1991). Previous results suggested that female frogs impose more stabilising
54 selection on stereotyped (static) properties, while directional selection should be imposed on
55 more dynamic call traits, which show higher coefficients of variation (Gerhardt, 1991). In
56 addition, the degree of stereotypy has been proposed to be related to the spatial pattern of
57 variation of the different properties of the call (Castellano et al., 2000).
58
Spatial separation is a potential predictor of both gene flow and degree of
59 environmental similarity between populations (Lougheed et al., 2006). Geographic signal
8
60 differentiation is frequently observed in frogs, on both small (e.g., Boul et al., 2007;
61 Rodríguez et al., 2010) and large scales (e.g., Heyer & Reid, 2003; Amézquita et al., 2009).
62 Empirical patterns of spatial distribution of acoustic intraspecific variability have been
63 attributed to different evolutionary forces which could act synergistically: correlations
64 between acoustic and geographic distances are associated with neutral evolutionary processes
65 such as genetic drift (Pröhl et al., 2007; Amézquita et al., 2009), while vicariant effects are
66 invoked when differentiation in calls is congruent with the crossing of geographic barriers
67 (Simões et al., 2008; Rodríguez et al., 2010). Intraspecific signal divergence may reflect
68 adaptation to structural features of the environment that differ between habitats (Ryan et al.,
69 1990; Ryan & Wilczynski, 1991), or may be associated with female preferences; the latter
70 may be interpreted as evidence of speciation caused by sexual selection (Boul et al., 2007;
71 Guerra & Ron, 2008).
72
As a highly energy-demanding activity, the emission of advertisement calls has
73 morphological and environmental constraints acting on ecological time scales, which are
74 regarded as co-factors and should be controlled in comparative bioacoustic studies (Prestwich,
75 1994). They constitute an additional source of variation, of which the extent of interference
76 varies between species and among the different acoustic traits considered (Gerhardt, 1991;
77 Gerhardt & Huber, 2002). Studies observed a correspondence between the degree of call trait
78 variability and the co-factors responsible for their variation (e.g., Castellano & Giacoma,
79 1998; Castellano et al., 2000), and our knowledge on the evolution of call trait stereotypy
80 should be benefited by a better comprehension of this association.
81
Because of its close relationship – either as cause or outcome – with the process of
82 evolutionary differentiation, it is expected that geographic variability in acoustic sexual traits
83 should be greater among than within populations (e.g., Castellano et al., 2002; Smith &
84 Hunter, 2005). This is particularly expected in organisms with typically high population
85 genetic structure and low dispersal abilities, such as many anuran amphibians (Lougheed et
86 al., 2006; Vences & Wake, 2007). The potential for a given call trait to act as a cue for
87 discrimination (through either social recognition or sexual selection) can be evaluated by
88 comparing the ratios between among- and within-population coefficients of variation (Pröhl et
89 al., 2007). Therefore, we analysed the variability in advertisement call traits throughout the
90 distribution of the Amazonian lowland frog Allobates paleovarzensis (Dendrobatoidea:
91 Aromobatidae) in order to access the variability of sexual signal in different levels. These are
9
92 small-sized, diurnal, leaf-litter frogs that inhabit seasonally flooded forests of Central
93 Amazonia (Lima et al., 2010). Dendrobatoid frogs are frequently used as models in both
94 exploratory and experimental studies of acoustic communication, although no such studies
95 have been conducted with any member of the dull-colored A. trilineatus species group (sensu
96 Morales, 2002). Related with comparing the within- and among-population variability in
97 acoustic sexual traits, we aimed to evaluate the degree of stereotypy of each call trait by
98 assessing coefficients of variation for individual frogs, test the association between
99 environmental temperature and male body size with acoustic characters, and test for the effect
100 of a river barrier leading to call differentiation in the study system.
101
102 Material and methods
103
Data acquisition
104
We recorded advertisement calls of Allobates paleovarzensis in ten localities (Figure
105 1), which will be regarded as individual populations in this study. Sampling sites were located
106 in the state of Amazonas, Brazil, on both sides of the Amazon River, and were at least 28 km
107 distant from each other. This study extends the known distribution of this species, for which
108 the confirmed occurrence was formerly restricted to its type locality in the municipality of
109 Careiro da Várzea (Lima et al., 2010). Therefore, this study comprises the entire known range
110 of A. paleovarzensis. Fieldwork was conducted between January 2009 and June 2010. In total,
111 we recorded advertisement calls of 123 males, with a mean of 12 males per population (SD =
112 3.8, range = 2–15; 12–15 when excluding Unini). Each male had its advertisement call
113 recorded for about three minutes with a Marantz PMD660 digital recorder (44.1 kHz
114 sampling rate; 16-bit resolution) and a Sennheiser K6/ME66 directional microphone
115 positioned approx. 1 m from the calling site. After each recording, we measured
116 environmental (air) temperature with a digital thermometer, and captured the frog, to measure
117 body size (snout-vent length, SVL, in mm) with a digital caliper. Voucher specimens were
118 deposited in the Herpetological Collection of the Instituto Nacional de Pesquisas da
119 Amazônia, in Manaus, Brazil.
120
Call measurements were obtained from recordings in Raven Pro 1.3 software (Charif
121 et al., 2008). The advertisement call of Allobates paleovarzensis consists of groups of single
10
122 notes separated by silent intervals (Lima et al., 2010; see Figure 2). For each recording, we
123 chose 10 calls with less background noise for detailed analyses. Note measurements,
124 including the spectral traits, were taken from the middle or next note of each call. Spectral
125 measurements were taken after a Fast Fourier Transform with frequency resolution of 82 Hz
126 and 2048 points. We considered ten call traits in our analyses: Note rate (NR, in notes/s);
127 Note duration (ND, in s); Internote interval, as the silent interval between two consecutive
128 notes of a call (InI, in s); Call rate (CR, in calls/s); Call duration (CD, in s); Intercall interval,
129 as the silent interval between two consecutive calls (IcI, in s); Maximum (peak) frequency, as
130 the frequency of higher intensity calculated for an entire note by a power spectrum function of
131 Raven Pro 1.3 (MF, in Hz); Lowest frequency (LF, in Hz); Highest frequency (HF, in Hz);
132 Note modulation, as the difference between the highest and the lowest frequencies of the call
133 (NM, in Hz). LF and HF were measured at 20 dB below the peak intensity, the value at which
134 the signal energy could still be clearly distinguished from background noise (Lima et al.,
135 2010).
136
137
Data analysis
138
The data set was tested for normality. When appropriate, we log-transformed data
139 prior to carrying out the parametric statistical analyses. Given that certain call traits are
140 affected by environmental temperature and male body size (Gerhardt & Huber, 2002), we
141 tested to what extent these two co-variables affect each of the call parameters in the species
142 studied. For temperature, linear regression analyses were conducted. For all within- and
143 among-population analyses we used temperature-adjusted acoustic data, by calculating overall
144 regression coefficients (b) for all call parameters at the mean temperature of 25.4 oC (SD =
145 1.08, range = 23.1–28.0). We applied the equation Y adj = y – (b*tcalling site) + (b*tmean) where
146 Y adj = adjusted value of call measurement, b = regression coefficient and t = temperature. To
147 test for morphological correlates of call traits, we performed linear regression analyses
148 between body size and temperature-adjusted call measurements. The adjusted r2 values
149 obtained were used to estimate the percentage of variation explained by the independent
150 variable.
151
We calculated coefficients of variation (CV = 100%*SD/ ) for each call trait within
152 individuals, within populations, and over all individuals combined. For each focal male, we
153 determined the individual mean (
wi)
and standard deviation (SDwi) of each call property (n =
11
154 10 calls/male), and we used these values to calculate a within-individual (i.e., within155 recording) coefficient of variation (CVwi = 100%*SDwi/
wi)
(Bee et al., 2010). By averaging
156 all within-individual CVs, we obtained the mean within-individual CV (CVwi). This measure
157 allows us to assess the degree of stereotypy of each call property: call traits with a mean CVwi
158 less than 5% were considered static; those with a mean CVwi greater than 12% were
159 considered dynamic; and those with a mean CVwi between 5% and 12% were classified as
160 intermediate (Gerhardt, 1991). In addition, we calculated within-population CVs based on the
161 mean (
wp)
and standard deviation (SDwp) over all call-recorded individuals in each
162 population (CVwp = 100%*SDwp/
wp).
By averaging all within-population CVs, we obtained
163 the mean within-population CV (CVwp) (Pröhl et al., 2007). Overall CVs (CVo) were obtained
164 from the grand mean and standard deviation based on averaging all the individual means from
165 our sample (CVo = 100%*grand SD/grand
). This calculation has also been regarded as
166 among-individual (Bee et al., 2010) and among-population coefficients of variation (Pröhl et
167 al., 2007). The Unini locality was excluded from the analysis of geographic variation because
168 of the small sample size. Based on these measurements, we determined the ratio between the
169 overall and mean within-population variation as CVo/CVwp.
170
To compare the difference in calls (with temperature-adjusted traits) between sides of
171 the Amazon River and among populations, nested analyses of variance (ANOVA, populations
172 nested on river sides) were conducted. We used two different models for each call trait: (a) a
173 nested ANOVA without co-factor and (b) a nested ANOVA with size as co-factor, thus
174 constituting a nested ANCOVA. These two approaches were used in order to compare the F
175 statistics obtained through the inclusion of body size as a co-factor in the analyses: if an
176 interaction occurs between co-factors and populations (sampling sites), we would expect a
177 decline in F statistics due to worsening of the model. All statistical analyses, unless otherwise
178 stated, were conducted in SYSTAT 12.0 (Wilkinson, 2007).
179
We tested the correlation between acoustic/body size distances among sampling sites
180 and respective linear geographic distances by applying Mantel tests. In addition, partial
181 Mantel tests were conducted to check for correlations between acoustic/body size distances
182 and their division by the Amazon River channel, while controlling for effects of linear
183 geographic distance among populations. For this, we constructed a binary correspondence
184 matrix, designating the values “0” and “1” for localities within the same and between opposite
185 sides of the Amazon River, respectively. Acoustic distance matrices were obtained from the
186 dataset of temperature-adjusted call measurements by calculating pairwise Euclidean
12
187 distances between their scores on first and second components (which together explained
188 79.0% of the total acoustic variation) produced by a principal components analysis (PCA). As
189 the scores of the first principal component were correlated with body size (linear regression r2
190 = 0.51, F = 9.39, p = 0.017), we regressed them against corresponding mean SVL
191 measurements for each population, and used the residuals as new, size-independent acoustic
192 variables from which new Euclidean distances were calculated. Mantel tests were conducted
193 in ZT (Bonnet & van de Peer, 2002) using permutation of the null models (Anderson &
194 Legendre, 1999), and applying 10000 randomisations.
195
196 Results
197 Effect of temperature and body size on acoustic traits
198
There was a high temperature dependency of all temporal traits related to the notes of
199 the advertisement call. This relation was not observed in temporal call traits related to the
200 entire call, nor in spectral characters (Table 1). Males produced shorter and more closely
201 spaced notes with increasing temperature. Environmental temperature affected one spectral
202 property (note modulation), which was highly correlated with the temporal trait of note
203 duration (rs = 0.53, p < 0.001, n = 123). The size of the frogs was related only to spectral
204 parameters of the advertisement call. There were significant negative relationships between
205 body size and the maximum, lowest and highest frequencies. Male SVL explained between 30
206 and 35% of the variation in spectral traits (Table 1).
207
208 Stereotypy of acoustic traits
209
According to the within-male coefficients of variation observed, the most dynamic
210 traits were related to temporal characteristics of the call. Call duration (range = 0.75–8.73 s)
211 and inter-call interval (range = 1.02–5.51 s) showed the highest variability between
212 individuals and populations. Temporal traits of the notes showed lower coefficients of
213 variation, and spectral measurements were highly stereotyped (Table 2, Figure 3). The lowest
214 and highest frequencies of the call were the most static properties. Overall (or among-
13
215 population) and within-population coefficients of variation were similar, indicating a
216 comparable magnitude in the variability of call traits regardless of grouping (Table 2).
217
218 Geographic variability in acoustic traits
219
Call properties differed between the sampling localities, although measurements
220 overlapped (Figure 4). Seven of the ten temperature-adjusted call traits differed among
221 populations, and five call traits differed between sides of the Amazon River (Table 3).
222 However, male body size also differed among the sampling sites (Nested ANOVA F = 3.68; p
223 = 0.001) and between river sides (Nested ANOVA F = 4.23; p = 0.042). Therefore, analyses of
224 variance were conducted with a new model, which included body size as covariate (nested
225 ANCOVA). As a result, we still found inter-population and between-river differences in most of
226 the call traits analysed. The differences were distributed among both temporal (related to the
227 notes and calls) and frequency traits (Table 3). The inclusion of body size as a co-variate
228 resulted in generally lower F statistics for the proposed models, indicating that interactive
229 effects between morphology and geography (sampling site) take place in the system.
230
Body-size differences were correlated with linear geographic distances (Mantel test r
231 = 0.669, p = 0.020) and not with the side of the Amazon River (Partial Mantel test r = -0.128,
232 p = 0.289). However, acoustic differences, after temperature and body size adjustments, did
233 not show correlations with linear geographic distances between population pairs (Mantel test r
234 = -0.029, p = 0.653) and river sides (Partial Mantel test r = -0.089, p = 0.449).
235
236 Discussion
237
Variability in acoustic sexual signal traits throughout the range of the Amazonian frog
238 Allobates paleovarzensis showed comparable magnitudes within and among populations, and
239 was independent of the degree of stereotypy of the acoustic characters. The relative
240 importance of a given call trait as a potential cue for discrimination can be assessed by
241 comparing the ratios between among- and within-population coefficients of variation (Pröhl et
242 al., 2007). Higher between- than within-population CV ratios were found in the polymorphic
243 Oophaga pumilio, a dendrobatoid species studied on a similar geographic scale (Pröhl et al.,
14
244 2007). In a study comprising most of Brazilian Amazonia, the neotropical hylid frog
245 Dendropsophus leucophyllatus also showed a much lower within- than between-population
246 call variability, although this taxon may consist of a complex of cryptic species (Lougheed et
247 al., 2006). Therefore, our findings suggest that none of the call traits analysed is indicated as a
248 potential cue for social recognition or sexual selection between populations of A.
249 paleovarzensis.
250
Ecological and behavioural traits characteristic of the genus may be related to the
251 unexpectedly low overall-/within-population CV ratios observed in Allobates paleovarzensis.
252 Here, we hypothesise three non-exclusive causes: (1) Absence of local selective forces
253 causing signal differentiation. Its presence is expected in scenarios involving acoustic
254 character displacement (Lemmon, 2009). At all the localities visited in this study, no other
255 species of the ecologically similar A. trilineatus species group was found to occur with A.
256 paleovarzensis (Kaefer I. L., unpubl. data). Therefore, we believe that no ecological
257 interspecific selective pressures would be causing important call differentiation (i.e., higher
258 than between-populations) through the range of the species studied. However, it is important
259 to note that the composition of acoustically coactive sympatric frog species was not important
260 for intraspecific call differentiation in the congener A. femoralis (Amézquita et al., 2006). (2)
261 High call trait variability among males. Studies on A. femoralis suggest that between-male
262 call variation is pronounced and sufficient to allow statistical distinction of individual males
263 (Gasser et al., 2009). This individual signature in call traits could also be present in A.
264 paleovarzensis and lead to the high within-population levels of call variability observed. (3)
265 The use of visual cues to recognise and select mates. Interpopulation variation in calls may
266 have less importance in systems where visual cues are employed in mating interactions. It is
267 likely that A. femoralis uses visual components in mate recognition and choice (Luna et al.,
268 2010; Montanarin et al., 2011), as was observed in the dendrobatoid Oophaga pumilio (Maan
269 & Cummings, 2009). In addition, the observation of multimodal communication (with the
270 visual component as the secondary channel) during agonistic interactions in the genus lends
271 further support to this hypothesis (Narins et al., 2003; Narins et al., 2005). Therefore, future
272 investigation involving playback experiments on intraspecific mate recognition in A.
273 paleovarzensis should provide us with elements to disentangle the candidate causes for the
274 observed pattern of similar within- and among-population call variability.
15
275
Only spectral properties of the advertisement call of Allobates paleovarzensis were
276 significantly affected by male body size. This effect is often observed in frogs, and is due to
277 the mass and resonance of the laryngeal sound-producing structures (McClelland et al., 1996),
278 although some species also show temporal properties influenced by male size (e.g., Zweifel,
279 1968; Castellano et al., 2002; Gasser et al., 2009). The high coefficients of determination
280 observed indicate that all static acoustic properties of the call are under strong morphological
281 constraint, and also that these call properties may be honest indicators of their control factors.
282 Therefore, we should expect that females might impose a stabilising selection on spectral
283 traits of A. paleovarzensis, although we cannot disentangle the selective forces driving body
284 size and spectral features in order to evaluate their relative importance in call evolution.
285
As a neuromuscular mechanism, the regulation of the temporal properties of calls for
286 anurans and other ectotherms is affected by environmental temperature (Prestwich, 1994). In
287 Allobates paleovarzensis, all temporal traits related to individual notes, but none related to
288 calls, were affected with this co-factor. This apparently lower sensitivity of calls to
289 temperature may be related to the lower stereotypy (i.e., higher coefficients of variation) of
290 call parameters in relation to those of the notes, making them subject to higher “confounding”
291 intrinsic variability. In fact, Gerhardt & Huber (2002, p. 15) reported a trend for shorter
292 components of the call to be more static, and our data support this observation. Additionally,
293 the social environment composed by co- and heterospecific acoustic signals at the time of the
294 recording may constitute a third source of variation that may affect call traits (Wells &
295 Schwartz, 2007). Therefore, given that call characters were not related to morphological or
296 temperature constraints, it seems likely that, in addition to intrinsic factors, their display may
297 be subject to social regulation.
298
Both spectral and temporal call traits varied among populations and between sides of
299 the Amazon River. Intraspecific geographic variation in acoustic signals is frequently
300 observed in Amazonian frogs, and highlights the neglected within-species diversity of frogs in
301 the biome, which is also home to a large number of undescribed species (Fouquet et al., 2007;
302 Angulo & Icochea, 2010). Differentiation in Amazonian anurans has been tested regarding
303 river barrier effects, and provided mixed results associated with geological, idiosyncratic river
304 characteristics (e.g., Gascon et al., 1996, 1998; Lougheed et al., 1999; Funk et al., 2007;
305 Simões et al., 2008; Tsuji-Nishikido et al., 2012). Among these characteristics, width has
306 been indicated as determining the effectiveness of a river barrier in promoting speciation in
16
307 Amazonian birds (Hayes & Sewlal, 2004) and mammals (Patton & da Silva, 1998). We are
308 not aware of any other study investigating the Amazon, the largest river of the biome, as a
309 promoter of phenotypic differentiation in frogs. Not all acoustic properties of A.
310 paleovarzensis showed congruent patterns of divergence, and both river and geographic
311 distance effects were not significant in the overall acoustic variability when summarising all
312 call traits in principal components. The Amazon River was found to be most effective as a
313 dispersal barrier for forest birds along its lower, wider portion, between its confluence with
314 the Negro River and its delta, where it meets the Atlantic Ocean (Hayes & Sewlal, 2004).
315 According to this proposal, our study area lies in a portion where the Amazon River is
316 expected to exert intermediate effects as a vicariant barrier. Future studies should be designed
317 to test the effect of the river on intraspecific differentiation along different portions of its
318 course. In relation to isolation-by-distance effects, a correlation between geographic and
319 acoustic distances would be indicative of neutral processes acting in phenotypic
320 differentiation. Therefore, our results indicate that local stabilising selective forces may be
321 important in call divergence.
322
There was a positive relation between body size and geographic distances among
323 populations, and as expected, spectral variables of the call followed a similar pattern of
324 variation. However, the different call properties showed divergent patterns of geographic
325 variability through the range of the species, as also observed for the dendrobatoid Oophaga
326 pumilio in Costa Rica and Panama (Pröhl et al., 2007). In addition, we could not associate the
327 divergent geographic patterns of call variation to the trait’s degree of stereotypy, as observed
328 for green toads from Central Asia (Castellano et al., 2000), which suggests that call variation
329 in some anuran species may not follow this pattern.
330
331 Acknowledgements
332 We are grateful to Ana C. Mello, Anelise Montanarin, Arthur O. Nascimento, Moisés S.
333 Melo, Priscila P. Correa, and Raimundo N. Amorim for fieldwork assistance; to Anderson S.
334 Bueno, Luciana K. Erdtmann, and Pedro I. Simões for valuable suggestions during this study;
335 to Janet W. Reid for the English revision; and to two anonymous reviewers whose
336 observations greatly improved this paper. This work is part of the PhD thesis of ILK. We also
337 thank the Brazilian Conselho Nacional de Desenvolvimento Científico e Tecnológico for
338 financial support (CNPq - CT Amazônia 575572/2008-6). CNPq provided a research
17
339 fellowship to ILK and a productivity grant to APL. This study was allowed by RAN340 ICMBio/IBAMA (licences 13777-2, 18516-1 and 20065-2).
341
342 References
343 Amézquita, A., Hödl, W., Lima, A.P., Castellanos, L., Erdtmann, L. & Araújo, M.C. (2006).
344 Masking interference and the evolution of the acoustic communication system in the
345 Amazonian dendrobatid frog Allobates femoralis. – Evolution 60: 1874-1887.
346 Amézquita, A., Lima, A.P., Jehle, R., Castellanos, L., Ramos, O., Crawford, A.J., Gasser, H.
347 & Hödl, W. (2009). Calls, colours, shape, and genes: a multi-trait approach to the study of
348 geographic variation in the Amazonian frog Allobates femoralis. – Biol. J. Linn. Soc. 98: 826349 838.
350 Anderson, M.J. & Legendre, P. (1999). An empirical comparison of permutation methods for
351 tests of partial regression coefficients in a linear model. – J. Stat. Comp. Simul. 62: 271-303.
352 Angulo, A. & Icochea, J. (2010). Cryptic species complexes, widespread species and
353 conservation: lessons from Amazonian frogs of the Leptodactylus marmoratus group (Anura:
354 Leptodactylidae). – Syst. Biodiv. 8: 357-370.
355 Bee, M.A., Cook, J.M., Love, E.K., O’Bryan, L.R., Pettitt, B.A., Schrode, K. & Vélez, A.
356 (2010). Assessing acoustic signal variability and the potential for sexual selection and social
357 recognition in Boreal Chorus Frogs (Pseudacris maculata). – Ethology 116: 564-576.
358 Behr, O., von Helversen, O., Heckel, G., Nagy, M., Voigt, C.C. & Mayer, F. (2007):
359 Territorial songs indicate male quality in the sac-winged bat Saccopteryx bilineata
360 (Chiroptera, Emballonuridae). – Behav. Ecol. 17: 810-817.
361 Bonnet, E. & Van de Peer, Y. (2002). ZT: a software tool for simple and partial Mantel tests.
362 – J. Stat. Soft. 7: 1-12.
363 Boul, K.E., Funk, W.C., Darst, C.R., Cannatella, D.C. & Ryan, M.J. (2007). Sexual selection
364 drives speciation in an Amazonian frog. – Proc. R. Soc. Lond., B. 274: 399-406.
18
365 Castellano, S., Cuatto, B., Rinella, R., Rosso, A. & Giacoma, C. (2002). The advertisement
366 call of the European treefrogs (Hyla arborea): a multilevel study of variation. – Ethology 108:
367 75-89.
368 Castellano, S. & Giacoma, C. (1998). Stabilizing and directional female choice for male calls
369 in the European green toads. – Anim. Behav. 56: 275-287.
370 Castellano, S., Giacoma, C. & Dujsebayeva, T. (2000). Morphometric and advertisement call
371 geographic variation in polyploid green toads. – Biol. J. Linn. Soc. 70: 341-360.
372 Chakraborty, M., Mangiamele, L.A. & Burmeister, S.S. (2010). Neural activity patterns in
373 response to interspecific and intraspecific variation in mating calls in the Túngara Frog. –
374 PLoS One 5: e12898.
375 Charif, R.A., Waack, A.M. & Strickman, L.M. (2008). Raven Pro 1.3 User’s Manual. –
376 Cornell Laboratory of Ornithology, Ithaca, New York.
377 Fischer, J., Kitchen, D.M., Seyfarth, R.M. & Cheney, D.L. (2004). Baboon loud calls
378 advertise male quality: acoustic features and their relation to rank, age, and exhaustion. –
379 Behav. Ecol. Sociobiol. 56: 140-148.
380 Fouquet, A., Gilles, A., Vences, M., Marty, C., Blanc, M. & Gemmell, N.J. (2007).
381 Underestimation of species richness in Neotropical frogs revealed by mtDNA analyses. –
382 PLoS One 32: e1109.
383 Funk, W.C., Caldwell, J.P., Peden, C.E., Padial, J.M., De la Riva, I. & Cannatella, D.C.
384 (2007). Tests of biogeographic hypotheses for diversification in the Amazonian Forest Frog,
385 Physalaemus petersi. – Mol. Phyl. Evol. 44: 825-837.
386 Gascon, C., Lougheed, S.C. & Bogart, J.P. (1996). Genetic and morphological variation in
387 Vanzolinius discodactylus: a test of the river hypothesis of speciation. – Biotropica 28: 376388 387.
389 Gascon, C., Lougheed, S.C. & Bogart, J.P. (1998). Patterns of genetic population
390 differentiation in four species of Amazonian frogs. A test of the Riverine Barrier Hypothesis.
391 – Biotropica 30: 104-119.
19
392 Gasser, H., Amézquita, A. & Hödl, W. (2009). Who is calling? Intraspecific call variation in
393 the aromobatid frog Allobates femoralis. – Ethology 115: 596-607.
394 Gerhardt, H.C. (1991). Female mate choice in treefrogs: static and dynamic acoustic criteria.
395 – Anim. Behav. 42: 615-635.
396 Gerhardt, H.C. & Huber, F. (2002). Acoustic Communication in Insects and Anurans:
397 Common Problems and Diverse Solutions. – The University of Chicago Press, Chicago.
398 Gomez, D., Richardson, C., Théry, M., Lengagne, T., Léna, J., Plénet, S. & Joly, P. (2011).
399 Multimodal signals in male European treefrog (Hyla arborea) and the influence of population
400 isolation on signal expression. – Biol. J. Linn. Soc. 103: 633-647.
401 Guerra, M.A. & Ron, S. (2008). Mate choice and courtship signal differentiation promotes
402 speciation in an Amazonian frog. – Behav. Ecol. 19: 1128-1135.
403 Hayes, F.E. & Sewlal, J.N. (2004). The Amazon River as a dispersal barrier to passerine
404 birds: effects of river width, habitat and taxonomy. – J. Biogeogr. 31: 1809-1818.
405 Heyer, R.W. & Reid, Y.R. (2003). Does advertisement call variation coincide with genetic
406 variation in the genetically diverse frog taxon currently known as Leptodactylus fuscus
407 (Amphibia: Leptodactylidae)? – An. Acad. Bras. Ciênc. 75: 39-54.
408 Lemmon, E.M. (2009). Diversification of conspecific signals in sympatry: geographic overlap
409 drives multidimensional reproductive character displacement in frogs. – Evolution 63: 1155410 1170.
411 Lima, A.P., Caldwell, J.P., Biavati, G. & Montanarin, A. (2010). A new species of Allobates
412 (Anura: Aromobatidae) from paleovárzea forest in Amazonas, Brazil. – Zootaxa 2337: 1-17.
413 Lougheed, S.C., Austin, J.D., Bogart, J.P, Boag, P.T. & Chek, A.A. (2006). Multi-character
414 perspectives on the evolution of intraspecific differentiation in a neotropical hylid frog. –
415 BMC Evol. Biol. 2006: 6-23.
416 Lougheed, S.C., Gascon, C., Jones, D.A., Bogart, J.P. & Boag, P.T. (1999). Ridges and rivers:
417 a test of competing hypotheses of Amazonian diversification using a dart-poison frog,
418 Epipedobates femoralis. – Proc. R. Soc. Lond., B. 266: 1829-1835.
20
419 Luna, A.G., Hödl, W. & Amézquita, A. (2010). Colour, size and movement as visual
420 subcomponents in multimodal communication by the frog Allobates femoralis. – Anim.
421 Behav. 79: 739-745.
422 Maan, M.E. & Cummings, M.E. (2009). Sexual dimorphism and directional sexual selection
423 on aposematic signals in a poison frog. – Proc. Natl. Acad. Sci. U. S. A. 106: 19072-19077.
424 McClelland, B.E., Wilczynski, W. & Ryan, M.J. (1996). Correlations between call
425 characteristics and morphology in male cricket frogs (Acris crepitans). – J. Exp. Biol. 199:
426 1907-1919.
427 Melendez, K.V. & Feng, A.S. (2010). Communication calls of little brown bats display
428 individual-specific characteristics. – J. Acoust. Soc. Am. 128: 919-923.
429 Montanarin, A., Kaefer, I.L. & Lima, A.P. (2011). Courtship and mating behaviour of the
430 brilliant-thighed frog Allobates femoralis from Central Amazonia: implications for the study
431 of a species complex. – Ethol. Ecol. Evol. 23: 141-150.
432 Morales, V.R. (2002 “2000”). Sistemática y biogeografia del grupo trilineatus (Amphibia,
433 Anura, Dendrobatidae, Colostethus), con descripción de once nuevas especies. – Publ. Asoc.
434 Amigos Doñana 13: 1-59.
435 Narins, P.M., Grabul, D.S., Soma, K.K., Gaucher, P. & Hödl, W. (2005). Cross-modal
436 integration in a dart-poison frog. – Proc. Natl. Acad. Sci. U. S. A. 102: 2425-2429.
437 Narins, P.M., Hödl, W. & Grabul, D.S. (2003). Bimodal signal requisite for agonistic
438 behavior in a dart poison frog, Epipedobates femoralis. – Proc. Natl. Acad. Sci. U. S. A. 100:
439 577-580.
440 Panhuis, T.M., Butlin, R., Zuk, M. & Tregenza, T. (2001). Sexual selection and speciation. –
441 Trends Ecol. Evol. 16: 364-371.
442 Patton, J.L. & da Silva, M.N.F. (1998). Rivers, refuges, and ridges: The geography of
443 speciation of Amazonian mammals. – In: Endless Forms: Species and Speciation (Howard,
444 D.J. & Berlocher, S.H., eds). Oxford University Press, Oxford, p. 202-213.
445 Prestwich, K.N. (1994). The energetics of acoustic signaling in anurans and insects. – Am.
446 Zool. 34: 625-643.
21
447 Pröhl, H., Hagemann, S., Karsch, J. & Höbel, G. (2007). Geographic variation in male sexual
448 signals in Strawberry Poison Frogs (Dendrobates pumilio). – Ethology 113: 825-837.
449 Rodríguez, A., Nuez, D. & Alonso, R. (2010). Intraspecific variation in the advertisement call
450 of the Cloud-Forest Frog Eleutherodactylus glamyrus (Anura: Eleutherodactylidae). – J.
451 Herpetol. 44: 457-466.
452 Ryan, M.J., Cocroft, R.B., & Wilczynski, W. (1990). The role of environmental selection in
453 intraspecific divergence of mate recognition signals in the cricket frog, Acris crepitans. –
454 Evolution 44: 1869-1872.
455 Ryan, M.J., Rand, A.S. & Weigt, L.A. (1996). Allozyme and advertisement call variation in
456 the tungara frog, Physalaemus pustulosus. – Evolution 50: 2435-2453.
457 Ryan, M.J. & Wilczynski, W. (1991). Evolution of intraspecific variation in the advertisement
458 call of a cricket frog (Acris crepitans, Hylidae). – Biol. J. Linn. Soc. 44: 249-271.
459 Simões, P.I., Lima, A.P., Magnusson, W.E., Hödl, W. & Amézquita, A. (2008). Acoustic and
460 morphological differentiation in the frog Allobates femoralis: relationships with the upper
461 Madeira River and other potential geological barriers. – Biotropica 40: 607-614.
462 Smith, M.J. & Hunter, D. (2005). Temporal and geographic variation in the advertisement call
463 of the Booroolong Frog (Litoria booroolongensis: Anura: Hylidae). – Ethology 111: 1103464 1115.
465 Tsuji-Nishikido, B.M., Kaefer, I.L., Freitas, F.C., Menin, M. & Lima, A.P. (2012). Significant
466 but not diagnostic: differentiation through morphology and calls in the Amazonian frogs
467 Allobates nidicola and A. masniger. – Herpetol. J. 22: 105-114.
468 Vences, M. & Wake, D.B. (2007). Speciation, species boundaries and phylogeography of
469 amphibians. – In: Amphibian Biology, Systematics (Heatwole, H. & Tyler, M.J., eds). Surrey
470 Beatty & Sons, Chipping Norton, p. 2613-2671.
471 Verburgt, L., Ferreira, M. & Ferguson, J.W.H. (2011). Male field cricket song reflects age,
472 allowing females to prefer young males. – Anim. Behav. 81: 19-29.
22
473 Wells, K.D. & Schwartz, J.J. (2007). The Behavioral Ecology of Anuran Communication. –
474 In: Hearing and Sound Communication in Amphibians (Narins, P.M., Feng, A.S., Fay, R.R. &
475 Popper, A.N., eds). Springer-Verlag, New York, p. 44-86.
476 Wilkinson, L. (2007). SYSTAT: the system for statistics. SYSTAT inc., Evanston.
477 Zweifel, R.G. (1968). Effects of temperature, body size, and hybridization on mating calls of
478 toads, Bufo a. americanus and Bufo woodhousii fowleri. – Copeia 1968: 269-285.
22
Table 1. The effect of temperature and body size on call traits in Allobates paleovarzensis (n
= 123 males). Linear regression coefficients of determination (r2), F statistics and significance
values (p) are presented. For body size (SVL), call traits were temperature-adjusted.
Significant p values after Bonferroni adjustment are marked with asterisks.
Temperature
Body size
Trait
r2
F
P
r2
F
P
NR
0.27
45.14
<0.001*
0.00
0.87
0.353
ND
0.18
28.39
<0.001*
0.00
0.05
0.818
InI
0.21
33.08
<0.001*
0.00
1.56
0.213
CR
0.00
0.09
0.768
0.00
0.61
0.437
CD
0.00
0.91
0.341
0.02
4.11
0.045
IcI
0.00
0.00
0.981
0.00
0.19
0.664
MF
0.00
0.35
0.554
0.30
54.33
<0.001*
LF
0.00
1.13
0.289
0.35
66.55
<0.001*
HF
0.00
1.44
0.232
0.34
63.74
<0.001*
NM
0.11
15.99
<0.001*
0.01
2.47
0.118
23
Table 2. Descriptive statistics of call traits based on overall acoustic parameter means calculated from mean values of 123 males recorded from
ten localities throughout the range of Allobates paleovarzensis, representing natural variation between 23.1 and 28.0 oC air temperature. Mean
within-individual, mean within-population and overall coefficients of variation (CVwi, CVwp, CVo, %) of ten traits of the advertisement call
(adjusted to 25.4oC using linear regression) of Allobates paleovarzensis. Traits are classified by type based on mean within-individual CVs
according to Gerhardt (1991).
Trait
Mean ± SD
Range
CVwi
CVwp
CVo
CVo/CVwp
Type
NR
5.98 ± 1.37
3.56–10.14
11.1 ± 6.4
17.1 ± 4.2
19.7
1.15
Intermediate
ND
0.04 ± 0.005
0.02–0.06
6.8 ± 3.6
10.5 ± 2.6
11.3
1.08
Intermediate
InI
0.14 ± 0.04
0.07–0.29
16.0 ± 10.2
21.7 ± 3.8
26.3
1.21
Dynamic
CR
0.27 ± 0.07
0.12–0.44
25.0 ± 14.0
23.8 ± 6.6
26.7
1.12
Dynamic
CD
2.31 ± 1.03
0.75–8.73
31.8 ± 13.5
33.2 ± 8.7
44.8
1.35
Dynamic
IcI
2.04 ± 0.75
1.02–5.51
28.1 ± 19.3
29.8 ± 9.7
34.3
1.15
Dynamic
MF
4534.15 ± 192.47
4045.17–5092.62
1.2 ± 1.1
3.6 ± 0.7
4.1
1.12
Static
LF
4147.06 ± 160.87
3645.14–4545.05
0.9 ± 0.7
3.1 ± 0.7
3.7
1.19
Static
HF
4832.32 ± 187.83
4255.77–5316.92
1.0 ± 0.6
3.1 ± 0.6
3.6
1.16
Static
NM
685.26 ± 104.98
462.00–968.51
6.5 ± 3.4
13.4 ± 2.4
14.3
1.07
Intermediate
24
Table 3. Nested analysis of (co-)variance of call traits (dependent variables) of Allobates
paleovarzensis. Population (nested in river side) and river side are the independent categorical
variables. In model b, size is used as the co-variate. Significant p values after Bonferroni
adjustment are marked with asterisks.
a. Nested ANOVA
b. Nested ANCOVA (body size as co-variate)
Trait
F Pop
P
F River
P
F Pop
P
F River
P
NR
5.21
<0.001*
8.57
0.004*
5.04
<0.001*
8.09
0.005
ND
2.75
0.011
0.38
0.534
2.69
0.012
0.31
0.576
InI
7.41
<0.001*
2.86
0.093
7.09
<0.001*
2.68
0.104
CR
3.66
0.001*
16.79
<0.001*
3.75
0.001*
17.61
<0.001*
CD
3.34
0.002*
0.00
0.966
3.23
0.003*
0.10
0.747
IcI
1.48
0.182
1.20
0.275
1.53
0.164
0.88
0.348
MF
1.96
0.066
19.20
<0.001*
0.62
0.731
14.30
<0.001*
LF
3.40
0.002*
27.70
<0.001*
1.65
0.126
22.77
<0.001*
HF
3.70
0.001*
16.80
<0.001*
2.52
0.019
11.97
<0.001*
NM
3.17
0.004*
0.37
0.540
3.20
0.003*
0.78
0.37
25
Figure captions
Figure 1. Geographic distribution of sampling locations in Central Amazonia. 1 = Careiro, 2 =
Janauacá, 3 = Manaquiri, 4 = Hiléia, 5 = Novo Airão, 6 = Anamã, 7 = Anori, 8 = Codajás, 9 =
Unini, 10 = Barcelos.
Figure 2. (A) Oscillogram and (B) spectrogram of three multinote advertisement calls, and (C)
power spectrum of a single note of a male Allobates paleovarzensis at its type locality in
Careiro da Várzea, Brazil.
Figure 3. Within-individual coefficients of variation (CVwi) for ten acoustic properties of the
advertisement call of Allobates paleovarzensis. Grey boxes indicate the middle 50% of the
data (interquartile range), and horizontal lines represent the median values. Dots represent
outlying values. Vertical lines indicate the 10th and 90th percentiles.
Figure 4. Variation in call traits (adjusted to 25.4 oC using linear regression) and body size
across the range of Allobates paleovarzensis in Central Amazonia. Grey boxes indicate the
middle 50% of the data (interquartile range), and horizontal lines represent the median values.
Dots represent outlying values. Vertical lines indicate the 10th and 90th percentiles.
26
Fig. 1
27
Fig. 2
28
Fig. 3
29
Fig. 4
30
Capítulo 2
Kaefer, I.L.; Tsuji-Nishikido, B.M. & Lima,
A.P. 2012. Beyond the river: underlying
determinants of population acoustic signal
variability in Amazonian direct-developing
Allobates (Anura: Dendrobatoidea). Acta
Ethologica, 15:187-194.
31
1
2
3
Beyond the river: underlying determinants of population acoustic signal variability in
4
Amazonian direct-developing Allobates (Anura: Dendrobatoidea)
5
Igor L. Kaefer1, Bruno M. Tsuji-Nishikido2 and Albertina P. Lima1
6
7
1
Programa de Pós-Graduação em Ecologia, Instituto Nacional de Pesquisas da Amazônia, Av.
8 Efigênio Sales, 2239, CP 478, CEP 69011–970, Manaus, Amazonas, Brazil.
9
2
Programa de Pós-Graduação em Diversidade Biológica, Universidade Federal do Amazonas,
10 Av. General Rodrigo Octavio, 3000, CEP 69077-000, Manaus, Amazonas, Brazil.
11 Corresponding author: Igor L. Kaefer. E-mail address: [email protected]
12
13 Abstract The multidimensional nature of animal signals makes acoustic traits potentially
14 subject to different determinants. The Amazonian frogs Allobates nidicola and Allobates
15 masniger have an allopatric distribution, occurring along the left and right sides of the
16 Madeira River, respectively. These are two sister, phenotypically similar species whose eggs
17 are deposited and develop entirely in a terrestrial nest. In this study, we analyzed 2,000
18 advertisement calls recorded from ten localities across Central Amazonia, in order to
19 understand the role of determinants of acoustic signal variability at the population and species
20 levels. We assessed, through nested analyses of variance, the differentiation of six characters
21 of this sexual signal among populations and between interfluves. Moreover, we measured the
22 degree of variability and the extent of temperature- and body size-induced plasticity in call
23 traits. We also tested for isolation-by-distance effects in phenotypic differentiation through
24 Mantel tests. Coefficients of variation were higher among than within populations for all call
25 measurements. Spectral call properties were more distinctive than temporal traits among
26 populations and species. Advertisement call traits showed strong temperature-induced
27 plasticity (e.g., 45% of the variation in note duration). In contrast, the effects of body size
28 were restricted to frequency-related characters. The river barrier effect was significant among
29 all the acoustic variables analyzed even after controlling for male body size. Geography
30 (sampling locality) and body size also jointly affected call variability. No correlation between
32
31 geographical and acoustic distances among populations was observed, suggesting that local
32 stabilizing selective pressures have an important role in the evolution of call differentiation.
33 Keywords advertisement call, geographical variation, isolation by distance, river barrier
34 effect, sexual signal, signal plasticity.
35
36 Introduction
37 Each of the properties of an acoustic signal typically shows a characteristic range of variation
38 at different levels of analysis (e.g., within populations, among populations, between species)
39 (Castellano et al. 2002). Moreover, it is expected that the multidimensional nature of animal
40 signals will make acoustic traits potentially subject to different determinants (Gerhardt 1991;
41 Gerhardt and Huber 2002). By decomposing acoustic signals into their properties, we might
42 be able to assess the agents responsible for call variation and to account for their effects
43 depending on the question asked in a particular study system. Additionally, by measuring the
44 degree of variability of quantitatively variable traits of frog calls in different levels, we can
45 evaluate their potential for sexual selection and social recognition, and also understand the
46 roles of particular acoustic characters in microevolutionary processes (e.g., Friedl and Klump
47 2002; Gasser et al. 2009; Bee et al. 2010).
48
Call traits show temperature- and body size-induced plasticity (Prestwich 1994).
49 Environmental temperature is often regarded as a cofactor in acoustic studies, and it is
50 necessary to account for its effect, especially for ectotherms, in order to obtain more accurate
51 measures of call differentiation when comparing individuals, populations, or species. The
52 extent to which temperature affects call traits varies considerably in anurans, and tropical
53 species, which undergo more subtle temperature variations, are less studied than are species in
54 temperate regions (see Gerhardt and Huber 2002; Wells 2007). Another cofactor often
55 controlled in comparative acoustic studies is male body size, which is known to affect
56 predominantly frequency-related characters of the call (Ryan 1988).
57
Geography plays an important role in the evolution and differentiation of acoustic
58 signals (Wycherley et al. 2002), but its effect may not be equivalent for all the properties of
59 the call. It has been suggested that spatial variation takes place according to the degree of
60 variability of a given acoustic character, with static properties exhibiting a more pronounced
33
61 variation among populations (Castellano et al. 2000). Moreover, relationships between
62 geography and covariables such as body size might produce interactive effects between
63 determinants of call differentiation (Pröhl et al. 2007; Kaefer and Lima 2012).
64 Biogeographical studies in Amazonia are revealing the effects of river barriers on call
65 divergence for different frog species, both in overall acoustic variability (Simões et al. 2008;
66 Tsuji-Nishikido et al. 2012) and in specific properties of the call (Kaefer and Lima 2012).
67 However, isolation by distance is also recognized as an important mechanism in phenotypic
68 divergence in Amazonian frogs (e.g., Lougheed et al. 2006; Pröhl et al. 2006; Amézquita et
69 al. 2009). Because isolation-by-distance effects are inferred through correlations between
70 phenotypic and distance measures, they can be masked by the presence of river barriers,
71 which might lead to acoustic distances that are disproportionally greater than those produced
72 by linear distance. Therefore, testing for the occurrence of isolation-by-distance effects might
73 provide clues to inform the selection/drift debate regarding the evolutionary forces driving
74 differentiation.
75
In an earlier study, we demonstrated the usefulness of advertisement calls in
76 discriminating between the similar Amazonian species Allobates nidicola Caldwell and Lima
77 2003 and A. masniger Morales 2002 (Tsuji-Nishikido et al. 2012). This study indicated that
78 the lower portion of the Madeira River is an important biogeographical barrier involved in
79 phenotypic differentiation, and suggested that A. nidicola and A. masniger have an allopatric
80 distribution, occurring along the left and right sides of the river, respectively. Molecular
81 genetic data confirm this proposed distribution and provide evidence that they constitute sister
82 species (Kaefer et al. unpublished data). A. nidicola and A. masniger are two diurnal, ground83 dwelling species whose eggs are deposited and develop entirely in a terrestrial nest, and their
84 adults share all the available morphological diagnostic characters (Morales 2002; Caldwell
85 and Lima 2003) (Fig. 1). Tsuji-Nishikido et al. (2012), as in most taxonomic studies, limited
86 themselves to summarizing the call measurements by applying principal components
87 analyses, thus overlooking the individual contribution of each call trait to the differentiation
88 between populations or species. Moreover, possible effects of isolation by distance were not
89 addressed. Therefore, in this study we characterized the underlying determinants of
90 population acoustic variability, by exploring this study system through the following
91 questions: (1) What is the degree of variability of each of the call traits among populations
92 and between species? (2) To what extent do temperature and body size induce call trait
93 variability? (3) Is the river barrier effect significant among all the different acoustic variables?
34
94 (4) Does isolation by distance play a significant role in call divergence, beyond the known
95 effects of the river as a barrier?
96
97 Material and methods
98 We analyzed a total of 2,000 calls from 100 males (20 calls per male) recorded at ten sites
99 located in the Brazilian states of Amazonas and Pará, including the type localities of A.
100 nidicola and A. masniger (Fig. 2). Ten individuals were sampled in each locality, which were
101 regarded as populations in this study. Data were collected between December 2009 and
102 March 2010, during the rainy season, which encompasses the breeding periods of both
103 species. Recordings were made using a Marantz PMD660 digital recorder (44.1 kHz sampling
104 rate; 16-bit resolution) and an AKG 568 EB directional microphone positioned 1-1.5 m in
105 front of the calling male. After each recording, we measured environmental (air) temperature
106 with a digital thermometer and captured the frog to measure body size (snout-vent length,
107 SVL, in millimeters) with a digital caliper. The advertisement call of both species is
108 composed of a single note repeated rhythmically over time, thus forming sequences of notes
109 lasting from 2 to 5 min. A comparative graphical representation of the advertisement call of
110 both species can be found in Tsuji-Nishikido et al. (2012). Call measurements were taken on
111 Raven 1.2 (Window: Blackman; Discrete Fourier Transform: 1,024 samples; 3 dB filter
112 bandwidth: 80.0 Hz) (Charif et al. 2004). The measurement used to represent each individual
113 in the statistical analysis was the mean value of the parameters obtained from 20 notes and
114 their respective intervals. We considered the following traits: note duration (ND, seconds);
115 note modulation (NM, Hertz); interval between notes (IN, seconds); lowest frequency (LF,
116 Hertz); highest frequency (HF, Hertz) and peak frequency (PF, Hertz). Call measurements are
117 illustrated in Fig. 3.
118
Our data set was checked for normality prior to the parametric statistical analyses. In
119 order to assess the effect of temperature on call properties, we performed linear regression
120 analyses for all populations combined. For all within- and among-population analyses, we
121 used temperature-adjusted acoustic data by calculating overall regression coefficients (b) for
122 all call parameters at the mean temperature of 25.9 oC (SD = 1.77, range = 23.1–31.2) through
123 the equation Y adj = y – (b*tcalling
site)
+ (b*tmean) where Y adj = adjusted value of call
124 measurement, b = regression coefficient, and t = temperature. The relationship between
35
125 temperature-adjusted acoustic traits and male body size was also assessed by linear regression
126 analyses. The adjusted r2 values obtained were used to estimate the percentage of variation
127 explained by the independent variable.
128
Coefficients of variation (CV = SD*100/mean) for each call trait were calculated
129 within populations (CVwp), within species (CVws), and over all individuals analyzed (CVo).
130 Our objectives were to evaluate the degree of variability of each call measurement and to
131 determine which call parameters might be most distinctive among populations and between
132 species. Within-population CVs are based on the mean and SD over all call-recorded
133 individuals in each population (n = 10). We obtained the mean within-population CV (CVwp)
134 by averaging all within-population CVs. Within-species CVs are based on the mean and SD
135 over all call-recorded individuals in each side of the river. We obtained the mean within136 species CV (CVws) by averaging the two within-species CVs. Overall CVs (CVo) were
137 obtained from the grand mean and SD over all populations and were initially calculated for
138 individuals (n = 100) of all populations combined (Pröhl et al., 2007).
139
To compare the difference in calls (with temperature-adjusted traits) between opposite
140 sides of the Madeira River (species) and among populations, we conducted nested analyses of
141 variance (ANOVA), with populations nested in river sides. Two different models for each call
142 trait were used: (a) a nested ANOVA without cofactor and (b) a nested ANOVA with size as
143 cofactor (nested ANCOVA). These approaches allowed us to compare the F statistics
144 obtained through the inclusion of body size as a cofactor in the analyses: if an interaction
145 takes place between cofactors and populations, a decline in F statistics is expected due to
146 worsening of the model.
147
The correlation between linear geographical/body size distances among sampling sites
148 and the respective acoustic distances was evaluated through Mantel tests (Smouse et al.
149 1986). A body size distance matrix was obtained through the calculation of Euclidean
150 distances between mean SVL measurements for each sampling locality. The acoustic distance
151 matrix was calculated by conducting a principal components analysis from the data set of
152 temperature-adjusted call measurements. We obtained pairwise Euclidean distances between
153 all possible pairs of populations by using the scores of the mean acoustic measurements in the
154 first and second principal components (which together explained 71.4% of the total acoustic
155 variation). Mantel tests were conducted in ZT (Bonnet and van de Peer 2002), using
36
156 permutation of the null models (Anderson and Legendre 1999), and applying 10,000
157 randomizations.
158
159 Results
160 The coefficients of variation, which indicate the degree of variability of the call traits, differed
161 when within-population, within-species, and overall variability were considered. Considering
162 the within-species and overall perspectives, most of the call traits analyzed showed
163 intermediate coefficients of variation. Regardless of grouping, spectral acoustic traits were the
164 less variable (Fig. 4). They also showed the highest CVo/CVwp and CVo/CVws ratios, which
165 suggests that spectral characters are more distinctive than temporal traits among populations
166 and species (Table 1).
167
Environmental temperature significantly affected all call measurements except note
168 modulation. Body size induced variation of all spectral properties of the call (Table 2).
169 Acoustic traits differed statistically among the populations. The F statistics of spectral trait
170 differentiation decreased when body size was included as a covariate. However, among171 population differentiation was still significant for the highest frequency after this procedure.
172 The lower Madeira River showed a strong effect on acoustic differentiation for all acoustic
173 traits considered, even when the effects of male body size were controlled (Table 3).
174
When all localities sampled were considered, the Mantel test revealed a strong
175 association between the advertisement call and body size differences between population
176 pairs. Geographical distances were not correlated with acoustic or body size distances. When
177 the Mantel tests were restricted to the left and right sides of the Madeira River (species), no
178 correlation regarding the traits analyzed was found between population pairs (Table 4).
179
180 Discussion
181
What is the degree of variability of each of the call traits among populations and
182 between species?
37
183
Coefficients of variation were higher in the overall perspective and reached its lower
184 values within populations for all call traits analyzed (i.e., CVwp < CVws < CVo). Our results
185 showed that spectral call properties were more distinctive than temporal traits among
186 populations and species, with higher distinctiveness at the among-population level compared
187 to the between-species level (i.e., CVo/CVws < CVo/CVwp). Therefore, our understanding of
188 the effect of river barriers in speciation processes would be greatly benefited by future
189 approaches that evaluate the degree of stereotypy of call traits at the within-individual level,
190 as well as the role of spectral properties of the advertisement call in mate recognition and
191 choice between individuals from opposite river sides.
192
To what extent do temperature and body size induce call trait variability?
193
In spite of the small range of variation in air temperature characteristic of tropical
194 environments (23.1 to 31.2 oC in this study), there was strong temperature-induced plasticity
195 in call traits. Note duration had 45% of its variation induced by this variable. Spectral
196 characters of the call were also significantly affected by air temperature, although to a lesser
197 extent (6–7%). Significant effects on frequency-related call measurements were also observed
198 in Oophaga pumilio throughout its range (Pröhl et al. 2007), but not in Allobates femoralis in
199 a single locality (Gasser et al. 2009). As ectotherms, frogs likely have spectral call traits that
200 are indirectly subject to the strong physiological constraints imposed by the environmental
201 and consequently the body temperature (Prestwich 1994).
202
The effects of body size on the advertisement call of the individuals were restricted to
203 frequency-related characters, as is most commonly reported for anurans (Gerhardt and Huber
204 2002). Body size-induced plasticity was pronounced in these measurements, ranging from 41
205 to 43%. Because of the invariant nature of these call properties, it was suggested that spectral
206 characters of the advertisement call may be important indicators of male size (e.g., Bee et al.
207 2000; Smith and Hunter 2005). Body-size effects are frequently removed in comparative
208 population analyses. However, as ontogenetic differences may be successfully controlled by
209 this procedure, natural variability in adult SVL, which might be important and even
210 diagnostic in comparative studies, is overlooked by applying body-size corrections (e.g.,
211 Castellano et al. 2000; Pröhl et al. 2007). Therefore, despite the difficulties in differentiating
212 between the two sources of variability related to body size, the description of these
213 relationships is important to the investigation of the morphological constraints of acoustic
214 variation. Interestingly, the properties of note modulation and internote interval were little
38
215 affected by temperature and the size of the frogs. This suggests that additional underlying
216 factors, such as physiological state, body condition, and social environment (Castellano and
217 Giacoma 1998; Tárano 2001; Wells and Schwartz 2007), might be responsible for the residual
218 (i.e., unexplained) variability in the call traits analyzed.
219
Is the river barrier effect significant among all the different acoustic variables?
220
The river barrier effect was significant among all the different acoustic variables
221 analyzed, even after the effects of male body size were controlled. The inclusion of SVL as a
222 covariate lowered the F statistics obtained in the analysis of spectral measurements,
223 suggesting an interactive effect between geography and body size on call differentiation. Our
224 results in the nested analysis of variance complement the findings of Tsuji-Nishikido et al.
225 (2012) by addressing the role of the lower Madeira River in affecting each of the call traits.
226 The effect of Amazonian rivers has been observed in a range of organisms, evidencing its
227 important role in evolutionary processes in the biome (e.g., Ayres and Clutton-Brock 1992;
228 Peres et al. 1996; Hayes and Sewlal 2004; Simões 2010). In contrast, the among-population
229 differentiation regarding call traits was severely reduced after we accounted for body-size
230 effects and was restricted to the properties of note duration, interval between notes, and
231 highest frequency. Thus, the overall ANOVA results show a predominant effect of the
232 Madeira River in addition to the among-population differences in the acoustic signals studied.
233
Does isolation by distance play a significant role in call divergence beyond the known
234 effects of the river as a barrier?
235
No correlation between geographic distance (a potential predictor of both degree of
236 environmental similarity and gene flow) and acoustic (sexual signal) dissimilarities was
237 revealed by the Mantel tests, and the observed pattern held when the analyses were restricted
238 to each side of the river. This apparent absence of isolation-by-distance effects was also
239 observed for call differentiation in Allobates paleovarzensis in Central Amazonia (Kaefer and
240 Lima 2012), but not for the neotropical frogs Dendropsophus leucophyllatus and A. femoralis
241 on wider geographical scales (Lougheed et al. 2006; Amézquita et al. 2009). This suggests
242 that local stabilizing selective pressures may be more important than neutral effects, such as
243 genetic drift, in the evolution of call differentiation in the A. nidicola/masniger system.
244 Regarding the two geographical factors presumed to be driving phenotypic differentiation
245 among the populations studied – river and linear geographical distance – our results give
39
246 further support to the relationship between sexual signal differentiation and the Madeira River
247 as a barrier, as indicated by Tsuji-Nishikido et al. (2012). A molecular assessment of
248 population variability, together with data on phenotypic variation, will seek to assess the
249 relationship between genetic and phenotypic divergence between A. nidicola and A. masniger,
250 a study system that shows the potential to elucidate various aspects regarding the initial stages
251 of allopatric speciation.
252 Acknowledgements We are grateful to Francisco C. de Freitas for fieldwork assistance,
253 Marcelo Menin for the loan of recording equipment, Pedro Ivo Simões for valuable
254 suggestions during this study, and to Janet W. Reid and anonymous reviewers for comments
255 on earlier versions of this manuscript. This work is part of the Ph.D. thesis of ILK. The
256 Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brazil) provided
257 financial support (CT Amazônia 553997-2006-8 and 575572-2008-6), a research fellowship
258 to ILK, and a productivity grant to APL. The Coordenação de Aperfeiçoamento de Pessoal de
259 Nível Superior (CAPES, Brazil) provided a research fellowship to BMT. This study was
260 approved by RAN-ICMBio/IBAMA (permit # 21950-1).
261
262 References
263 Amézquita A, Lima AP, Jehle R, Castellanos L, Ramos O, Crawford AJ, Gasser H, Hödl W
264
(2009) Calls, colours, shape, and genes: a multi-trait approach to the study of geographic
265
variation in the Amazonian frog Allobates femoralis. Biol J Linn Soc 98:826-838
266 Anderson MJ, Legendre P (1999) An empirical comparison of permutation methods for tests
267
of partial regression coefficients in a linear model. J Stat Comp Simul 62:271-303
268 Ayres JM, Clutton-Brock TH (1992) River boundaries and species range size in Amazonian
269
primates. Am Nat 140:531-537
270 Bee MA, Cook JM, Love EK, O’Bryan LR, Pettitt BA, Schrode K, Vélez A (2010) Assessing
271
acoustic signal variability and the potential for sexual selection and social recognition in
272
boreal chorus frogs (Pseudacris maculata). Ethology 116:564-576
40
273 Bee MA, Perrill SA, Owen PC (2000) Male green frogs lower the pitch of acoustic signals in
274
defense of territories: a possible dishonest signal of size? Behav Ecol 11:169-177.
275 Bonnet E, Van de Peer Y (2002) ZT: a software tool for simple and partial Mantel tests. J Stat
276
Soft 7:1-12
277 Caldwell JP, Lima AP (2003) A new Amazonian species of Colostethus (Anura:
278
Dendrobatidae) with a nidicolous tadpole. Herpetologica 59:219-234
279 Castellano S, Cuatto B, Rinella R, Rosso A, Giacoma C (2002) The advertisement call of the
280
European treefrogs (Hyla arborea): a multilevel study of variation. Ethology 108:75-89
281 Castellano S, Giacoma C (1998) Stabilizing and directional female choice for male calls in
282
the European green toads. Anim Behav 56:275-287
283 Castellano S, Giacoma C, Dujsebayeva T (2000) Morphometric and advertisement call
284
geographic variation in polyploid green toads. Biol J Linn Soc 70:341-360
285 Charif RA, Clark CW, Fristrup KM (2004) Raven 1.2 user’s manual. Cornell Laboratory of
286
Ornithology, Ithaca
287 Friedl TWP, Klump GM (2002) The vocal behaviour of male European treefrogs (Hyla
288
arborea): implications for inter- and intrasexual selection. Behaviour 139:113-136
289 Gasser H, Amézquita A, Hödl W (2009) Who is calling? Intraspecific call variation in the
290
aromobatid frog Allobates femoralis. Ethology 115:596-607
291 Gerhardt HC (1991) Female mate choice in treefrogs: static and dynamic acoustic criteria.
292
Anim Behav 42:615-635
293 Gerhardt HC, Huber F (2002) Acoustic communication in insects and anurans: common
294
problems and diverse solutions. The University of Chicago Press, Chicago
295 Hayes FE, Sewlal JN (2004) The Amazon River as a dispersal barrier to passerine birds:
296
effects of river width, habitat and taxonomy. J Biogeogr 31:1809-1818
297 Kaefer IL, Lima AP (2012) Sexual signals of the Amazonian frog Allobates paleovarzensis:
298
geographic variation and stereotypy of acoustic traits. Behaviour 149:15-33
41
299 Lougheed SC, Austin JD, Bogart JP, Boag PT, Chek AA (2006) Multi-character perspectives
300
on the evolution of intraspecific differentiation in a neotropical hylid frog. BMC Evol
301
Biol 2006:6-23
302 Morales VR (2002) Sistematica y biogeografía del grupo trilineatus (Amphibia, Anura,
303
Dendrobatidae, Colostethus), con descripción de once nuevas especies. Publ Asoc
304
Amigos Doñana 13:1-59
305 Peres CA, Patton JL, da Silva MNF (1996) Riverine barriers and gene flow in Amazonian
306
saddle-back tamarins. Folia Primatol 67:113-124
307 Prestwich KN (1994) The energetics of acoustic signaling in anurans and insects. Am Zool
308
34:625-643
309 Pröhl H, Hagemann S, Karsch J, Höbel G (2007) Geographic variation in male sexual signals
310
in Strawberry Poison Frogs (Dendrobates pumilio). Ethology 113:825-837
311 Pröhl H, Koshy RA, Mueller U, Rand AS, Ryan MJ (2006) Geographic variation of genetic
312
and behavioral traits in northern and southern túngara frogs. Evolution 60:1669-1679
313 Ryan, MJ (1988) Energy, calling, and selection. Am Zool 28:885-898
314 Simões PI (2010) Diversificação do complexo Allobates femoralis (Anura, Dendrobatidae)
315
em florestas da Amazônia brasileira: desvendando padrões atuais e históricos. Ph.D.
316
Thesis, Instituto Nacional de Pesquisas da Amazônia
317 Simões PI, Lima AP, Magnusson WE, Hödl W, Amézquita A (2008) Acoustic and
318
morphological differentiation in the frog Allobates femoralis: Relationships with the
319
upper Madeira River and other potential geological barriers. Biotropica 40:607-614
320 Smith MJ, Hunter D (2005) Temporal and geographic variation in the advertisement call of
321
the booroolong frog (Litoria booroolongensis: Anura: Hylidae). Ethology 111:1103-1115
322 Smouse PE, Long JC, Sokal RR (1986) Multiple regression and correlational extensions of
323
the Mantel test of matrix correspondence. Syst Zool 35:627-632
324 Tárano Z (2001) Variation in male advertisement calls in the neotropical frog Physalaemus
325
enesefae. Copeia 2001:1064-1072
42
326 Tsuji-Nishikido BM, Kaefer IL, Freitas FC, Menin M, Lima AP (2012) Significant but not
327
diagnostic: Differentiation through morphology and calls in the Amazonian frogs
328
Allobates nidicola and A. masniger. Herpetol J 22:105-114
329 Wells KD (2007) The ecology and behavior of amphibians. The University of Chicago Press,
330
Chicago
331 Wells KD, Schwartz JJ (2007) The behavioral ecology of anuran communication. In: Narins
332
PM, Feng AS, Fay RR, Popper AN (eds) Hearing and sound communication in
333
amphibians. Springer, New York, pp 44-86
334 Wycherley J, Doran S, Beebee TJC (2002) Male advertisement call characters as
335
phylogeographical indicators in European water frogs. Biol J Linn Soc 77:355-365
43
Tables
Table 1 Descriptive statistics of the call traits analyzed, based on mean individual call properties of 100 individual males, adjusted to 25.9 °C
using linear regression. Within-population (CVwp), within-species (CVws), and overall (CVo) coefficients of variation (%, given as mean ± SD)
of six traits of the advertisement call of Allobates nidicola and A. masniger.
Trait
Mean ± SD; A. nidicola
Mean ± SD; A. masniger
CVwp
CVws
CVo
CVo/CVwp
CVo/CVws
ND (s)
0.048 ± 0.005
0.050 ± 0.005
7.1 ± 2.3
9.4 ± 0.6
9.6
1.36
1.01
NM (Hz)
354.7 ± 46.7
396.9 ± 51.8
12.5 ± 2.6
13.1 ± 0.1
14.2
1.14
1.08
IN (s)
0.290 ± 0.048
0.339 ± 0.086
16.2 ± 5.4
20.9 ± 6.0
23.3
1.44
1.11
LF (Hz)
4048.9 ± 255.6
4299.3 ± 231.1
4.5 ± 1.9
5.8 ± 0.6
6.5
1.45
1.12
HF (Hz)
4403.9 ± 253.7
4699.4 ± 207.2
3.8 ± 1.3
5.0 ± 0.9
6.0
1.58
1.18
PF (Hz)
4222.5 ± 244.3
4497.6 ± 209.7
4.0 ± 1.4
5.2 ± 0.8
6.1
1.50
1.16
44
Table 2 The effect of temperature and body size on call traits of the individuals studied (n =
100 males). Adjusted linear regression coefficients of determination (r2), F statistics and
significance values (p) are presented. For body size, statistics were calculated using
temperature-adjusted call traits. Significant p values after Bonferroni adjustment are marked
with asterisks
Temperature
Body size
Trait
r2
F
P
r2
F
P
Note duration
0.45
82.00
<0.001*
-0.003
0.62
0.562
Note modulation
-0.003
0.69
0.589
-0.01
0.00
0.985
Internote interval
0.07
8.62
0.004*
0.03
3.60
0.057
Lowest frequency
0.07
8.15
0.005*
0.41
69.64
<0.001*
Highest frequency
0.06
7.84
0.006*
0.43
74.59
<0.001*
Peak frequency
0.07
8.61
0.004*
0.42
73.10
<0.001*
45
Table 3 Nested analysis of (co-)variance of call traits (dependent variables) of A. nidicola and
A. masniger. Population (nested in river side) and river side (species) are the independent
categorical variables. In model b, size is used as covariate. Significant p values after
Bonferroni adjustment are marked with asterisks
a. Nested ANOVA
b. Nested ANCOVA (size)
Trait
F Pop
P
F River
P
F Pop
P
F River
P
ND
8.62
<0.001*
7.72
0.006*
8.67
<0.001*
7.49
0.007*
NM
1.34
0.232
18.79
<0.001*
1.30
0.249
24.08
<0.001*
IN
8.60
<0.001*
19.99
<0.001*
8.53
<0.001*
10.22
0.001*
LF
6.09
<0.001*
37.35
<0.001*
2.14
0.039
9.66
0.002*
HF
8.02
<0.001*
64.01
<0.001*
3.58
0.001*
22.41
<0.001*
PF
6.58
<0.001*
53.13
<0.001*
2.55
0.014
16.92
<0.001*
46
Table 4 Mantel tests evaluating correlations between acoustic (DAco), body size (DSVL),
and geographical (DGeo) distances of Allobates sampled at ten localities in Central
Amazonia. Because of river effects, tests were also restricted to the left and right sides of the
Madeira River (five localities each). Significant values after Bonferroni correction are marked
with an asterisk.
Model
r
p
DAco X DGeo
-0.15
0.312
DAco X DSVL
0.642
0.001*
DSVL X DGeo
-0.193
0.205
DAco X DGeo
-0.484
0.108
DAco X DSVL
0.409
0.191
DSVL X DGeo
0.152
0.366
DAco X DGeo
-0.355
0.166
DAco X DSVL
0.45
0.091
DSVL X DGeo
-0.419
0.058
All localities
Left side (A. nidicola)
Right side (A. masniger)
47
Figures
Fig. 1 A. nidicola (top) and A. masniger (bottom) males vocalizing at their respective type
localities in Brazilian Amazonia
Fig. 2 Sampling sites visited in this study: 1 Km 12 of Autazes Road, type locality of A.
nidicola; 2 BR-319, km 260; 3 PPBio Manaquiri; 4 BR-319 Tupana; 5 Vila Gomes; 6 Borba;
7 Novo Aripuanã; 8 Road to Apuí; 9 Jacareacanga; 10 PARNA da Amazônia, type locality of
A. masniger
Fig. 3 A graphical representation of the call measurements adopted in this study. Oscillogram
(a) and spectrogram (b) of a call sequence of a topotypic A. masniger male (SVL = 18.5 mm;
INPA-H 28196), recorded at air temperature of 23.7 °C. Power spectrum (c) of a single call of
this individual. Highest and lowest frequencies were measured at 20 dB below the peak
intensity (horizontal dotted line), the value at which the signal energy could still be clearly
distinguished from background noise. Note modulation was calculated as the difference
between the highest and the lowest frequencies of the call
Fig. 4 Within-population coefficients of variation (CV, in percent) for six acoustic properties
of the advertisement call of the individuals studied. Gray boxes indicate the middle 50% of
the data (interquartile range), and horizontal lines represent the median values. Vertical lines
indicate the 10th and 90th percentiles. Note that the lowest coefficients of variation are related
to spectral traits of the call
48
Fig. 1
49
Fig. 2
50
Fig. 3
51
Fig. 4
52
Capítulo 3
Kaefer, I.L.; Tsuji-Nishikido, B.M.; Mota, E.P.;
Farias, I.P. & Lima, A.P. The early stages of
speciation in Amazonian forest frogs:
phenotypic conservatism despite strong genetic
structure. Submetido à Evolutionary Biology
53
1
2
3 The early stages of speciation in Amazonian forest frogs: phenotypic conservatism
4 despite strong genetic structure
5
6 Igor L Kaefer 1*, Bruno M Tsuji-Nishikido 2, Edvaldo P Mota 2, Izeni P Farias 2, Albertina P
7 Lima 1
8
9
1
Programa de Pós-Graduação em Ecologia, Instituto Nacional de Pesquisas da Amazônia, CP
10 478, 69011-970, Manaus, Amazonas, Brazil
11
2
Instituto de Ciências Biológicas, Universidade Federal do Amazonas, 69077-000, Manaus,
12 Amazonas, Brazil
13
14 *Corresponding author
15 Email: [email protected]
16 Telephone: 55 (92) 81018404
17 Fax: 55 (92) 36431909
18
19
20
21
22
23
24
54
25 Abstract
26 Phylogeographic perspectives incorporating multiple classes of characters, especially those
27 relating to sexual signals, are promising for the elucidation of recent evolutionary
28 mechanisms driving speciation. Here, forest frogs were used as a model system to access
29 distinct stages in the process of evolutionary differentiation. We studied 280 individuals
30 assigned to three species: Allobates paleovarzensis, A. nidicola and A. masniger. Samples
31 were collected at 20 localities arranged in two study systems, along the middle Amazon and
32 the lower Madeira Rivers, in Central Amazonia. Mantel tests, analyses of molecular variance,
33 and the spatial distribution of haplogroups indicated that the distribution of genetic variability,
34 as inferred from a mitochondrial DNA marker, was determined by a combination of isolation35 by-distance effects and the transposition of large Amazonian rivers. These two factors had
36 contrasting relative influences in each of the study systems, which also differed regarding the
37 estimated time of the major cladogenetic events. Pronounced population genetic structure was
38 observed. However, multivariate discriminant function analyses showed that the phenotypic
39 (morphological and acoustic) divergence was loosely related with genetic differentiation and
40 did not successfully predict assignment of individuals to genetic groups. The observed
41 distribution of genetic variability showed the important role of genetic drift in the
42 diversification of the mitochondrial marker studied. The phenotypic conservatism observed
43 among populations was surprising when considering the high genetic structuring observed,
44 and indicates a prevailing role of stabilizing selective forces in the process of sexual signal
45 and morphological differentiation.
46 Keywords
47 genetic drift, integrative phylogeography, isolation by distance, mitochondrial DNA, river
48 barrier, sexual signals.
49
50 Introduction
51
Historically isolated populations are primer candidates for the study of the origin of
52 differences in genetic and phenotypic characters because they would have been potentially
53 exposed to long term genetic drift or divergent selection pressures, without the homogenizing
54 influences of gene flow (Avise 2000; Nosil et al. 2009). Therefore, it is expected that spatially
55
55 structured populations, which experience restricted demographic connectivity, might show
56 concordant divergences among different sets of characters. Nevertheless, decoupled patterns
57 of population differentiation can be observed, indicating that either stabilizing or directional
58 selection pressures acting differentially on specific characteristics may lead to distinct patterns
59 of the distribution of variation among populations and have important implications for the
60 speciation process (Santos et al. 2003; Adams et al. 2009; Lemmon 2009).
61
Recent theoretical and analytical advances enabled by the integration of distinct
62 disciplines have led to a better understanding of patterns and mechanisms of differentiation
63 among populations of diverse organisms (Crisci et al. 2003; Santini et al. 2012). Nonetheless,
64 the relationship between genotypic and phenotypic differentiation in the early stages of
65 speciation still remains poorly understood, directing questions to ecological and historical
66 mechanisms that promote or maintain variation at this level of the biological hierarchy (Diniz67 Filho et al. 2008). Important advancements may be obtained by investigating the geographic
68 pattern in which genetic variability is distributed among populations, as well as the order
69 according to which ecological and phenotypic differentiation appears (West-Eberhard 1989;
70 Dall 1997; Panhuis et al. 2001).
71
The extent and shape of the spatial structure of intraspecific genealogical relationships
72 are typically unknown for natural populations, especially in megadiverse tropical regions,
73 where not even the species boundaries are clearly defined (e.g., Groeneveld et al. 2009;
74 Vieites et al. 2009; Hubert et al. 2012). Beyond the challenge of delimiting species, a better
75 comprehension of the speciation process requires the study of geographical factors which may
76 underpin the origin and diversification of lineages within species, such as vicariance and
77 isolation by distance (Hutchison and Templeton 1999; Zink et al. 2000). In Amazonia, species
78 distribution patterns (e.g., Wallace 1852; Haffer 1969; Ayres and Clutton-Brock 1992) and
79 fine-scale genealogical structure assessments (e.g., Cohn-Haft 2000; Funk et al. 2007;
80 Solomon et al. 2008; Amézquita et al. 2009) have inferred the effect of these two historical
81 factors in terrestrial organisms, although the relative contribution of each of them varied
82 greatly, perhaps due to idiosyncratic particularities of the study systems. Still, no clear
83 geographical pattern or palaeoenvironmental force has emerged as a general explanation for
84 the diversification of most of the animal groups (Antonelli et al. 2010; Ribas et al. 2011).
85
Phylogeographic perspectives involving more than one class of character are
86 particularly promising for the elucidation of recent evolutionary mechanisms driving
56
87 differentiation, given that the history inferred from DNA sequences may provide a historical
88 demographic context to evaluate divergence in various phenotypic attributes (Wiens 2008,
89 Campbell et al. 2010; Guillot et al. 2012). These appraisals might be conducted with any kind
90 of genetically based marker, not necessarily of molecular origin (Avise 2004). This is because
91 many phenotypic attributes carry phylogenetic and even phylogeographic signals (Wycherley
92 et al. 2002; Erdtmann and Amézquita 2009; Goicochea et al. 2009) that might, together with
93 gene trees, reciprocally illuminate the history of populations and the mechanisms responsible
94 for evolutionary divergence. Integrative, multicharacter approaches are particularly important
95 for our knowledge on tropical organisms, which urgently need detection of significant
96 evolutionary units for conservation, assessment of cryptic diversity, and determination of
97 species limits (Moritz et al. 2000; Dayrat 2005; Wiens 2007; Crawford et al. 2010).
98
Phylogeographic structure is commonly observed among species of frogs, and is often
99 credited to their low vagility when compared to other vertebrates (Avise 2000; Vences and
100 Wake 2007). Frogs are suitable for population-level multi-character evolutionary approaches
101 because of their high abundance, which allows dense sampling along their distributional
102 ranges, and their conspicuous advertisement calls, that can be used as population markers
103 (Wycherley et al. 2002; Lougheed et al. 2006). Given that anurans use advertisement calls for
104 sexual recognition and mate choice, differences in these signals dramatically increase the
105 evolutionary potential for speciation due to the possibility of pre-zygotic isolation (Panhuis et
106 al. 2001; Pröhl et al. 2006; Boul et al. 2007).
107
Allobates (Anura, Dendrobatoidea, Aromobatidae) frogs were used here as a model
108 system to access distinct stages of the evolutionary differentiation process through the
109 examination of morphological, acoustic and genetic characters. We studied individuals
110 assigned to three nominal species: Allobates paleovarzensis Lima et al. (2010), A. nidicola
111 (Caldwell and Lima 2003) and A. masniger (Morales 2002). Sampling was performed on 20
112 localities arranged in two distinct geographic study systems. The choice of these systems was
113 based on studies of phenotypic variability that suggested that these populations may have
114 experienced initial, but distinct stages in the process of allopatric speciation (Kaefer and Lima
115 2012; Kaefer et al. 2012; Tsuji-Nishikido et al. 2012). This is the first population genetic
116 assessment of the three aforementioned species. This study focuses on the using of gene trees
117 and phenotypic data in order to disentangle most recent events that led to the spatial
118 organization of the multi-trait variation currently observed.
57
119
We expected that that a neutral, stochastic process plays a central role in the
120 emergence of evolutionary differentiation which may lead to the formation of new species.
121 Thus, we tested the premise that the distribution of genetic variability is determined by a
122 combination of the effects of isolation by distance (i.e., geographic distance among individual
123 samples) and the transposition of vicariant barriers (Amazonian rivers). Our null hypothesis
124 was that the phenotypic differentiation is related to the genetic pattern observed. Therefore,
125 we expected that phenotypic divergence would be correlated with genetic differentiation and
126 not incongruent with the groupings indicated by molecular analyses. In order to test this
127 hypothesis, we accessed the genealogical relationship among individual samples through
128 phylogeographic analyses, including parameters related to historical demography inferred
129 from a neutrally evolving DNA marker. We also evaluated the extent of population
130 differentiation in advertisement calls and morphology by their correlation with geographic
131 and genetic distances, as well as the roles of both acoustic and morphological traits in
132 predicting the assignment of individuals to genetic groups.
133
134 Methods
135 Study area
136
This study was conducted in two adjacent areas of the Central Amazon lowlands, in
137 the Brazilian states of Amazonas and Pará, on a region characterized by wide tracts of tropical
138 rainforests. These areas are crossed by the two major rivers of the biome: the Amazon, at its
139 middle course (Fig. 1A), and the Madeira, at its lower portion (Fig. 2A). The Madeira River is
140 the main southern tributary of the Amazon, which represents the largest river system on Earth.
141 The white, sediment-rich water channels of these rivers are characterized by intense
142 sedimentary dynamics in the Central Amazon, permitting the formation of marginal lakes,
143 islands and channels and by extensive floodplains covered with seasonally flooded forests
144 (várzeas) on its riverbanks (Sioli 1984; Irion and Kalliola 2010). The region also has
145 extensive plateaus harboring non-flooded rainforest (terra firme), as well as transitional
146 forests, called palaeo-várzeas, that are ancient floodplains of the Amazon River and its
147 tributaries (Irion and Kalliola 2010).
58
148 The paleovarzensis system
149
The Amazonian palaeo-várzea forests constitute the habitat for Allobates
150 paleovarzensis, a diurnal aromobatid frog which deposits its eggs in the forest floor and
151 whose tadpole development must be completed in water bodies. This is a recently described
152 species which distribution encompasses its type and adjacent localities south of the Amazon
153 River, and also localities to the north, reaching the municipality of Barcelos, at the right
154 margin of the Negro River (Kaefer and Lima 2012). Our sampling design includes all known
155 areas of occurrence of this species, including its type locality (Fig. 1A).
156 The nidicola-masniger system
157
Allobates nidicola and Allobates masniger are diurnal leaf-litter frogs that occur in
158 terra firme forests of Central Amazonia, having been described without reference to each
159 other. Recently, it was found that these species share all the available diagnostic characters,
160 including the direct, terrestrial development of endotrophic larvae into froglets (Tsuji161 Nishikido et al. 2012). Nevertheless, this study detected significant differences in morphology
162 and calls between populations from both type localities and from opposite sides of the lower
163 Madeira River, the main biogeographic barrier in the studied area. This report also suggests
164 that they constitute allopatrically distributed sister species. In order to better understand the
165 speciation process occurring in this study system, we analyze these two species jointly,
166 according to the sample design presented in Tsuji-Nishikido et al. (2012) (Fig. 2A).
167 Data collection
168
Data were collected at 20 sites (10 per study system), which will be referred to as
169 populations hereafter. The fieldwork procedures employed for the acquisition of the
170 phenotypic data analyzed here are fully described in recent studies which evaluated the
171 taxonomic identity and the variation in phenotypic characters of individuals belonging to both
172 study systems (Kaefer and Lima 2012; Kaefer et al. 2012; Tsuji-Nishikido et al. 2012). The
173 collection of specimens constituted the first assessment of the taxonomic identity of
174 individuals through the distributional ranges of the three species involved. The number of
175 individuals analyzed for each data set, and the geographic coordinates of the sample sites, are
176 provided in Table 1. Muscular tissue samples from each individual were extracted and
177 preserved in 95 percent ethanol prior to fixation of voucher specimens in formol. Tissue
178 samples were housed at the Coleção de Tecidos de Genética Animal of the Universidade
59
179 Federal do Amazonas (CTGA – UFAM), Manaus, Brazil. Studied individuals were deposited
180 in the Herpetological Collection of the Instituto Nacional de Pesquisas da Amazônia, in
181 Manaus (INPA-H).
182
Acoustic data. Call parameters from individuals of the paleovarzensis (10
183 calls/individual) and nidicola-masniger (20 calls/individual) systems were obtained from
184 recordings with Raven Pro 1.3 (Charif et al. 2008) and Raven 1.2 (Charif et al. 2004),
185 respectively. Temporal and spectral call traits analysed for each species are described on
186 Online Resource 1 and Online Resource 2. These measurements were averaged in order to
187 represent each of the populations in the statistical analyses. The call structure of each of the
188 species, including graphic representations, and the detailed description of call traits analyzed
189 are provided elsewhere (Kaefer and Lima 2012; Kaefer et al. 2012; Tsuji-Nishikido et al.
190 2012). Summary statistics of the call measurements are given on Online Resource 1 and
191 Online Resource 2.
192
Morphological data. External morphometric measurements (in mm) were taken from
193 the left side of the specimens, with a stereoscopic microscope with a graduated ocular lens
194 (precision 0.10 mm) and a digital caliper (to the nearest 0.01 mm). All individuals were
195 analyzed according to the twenty-three measurements presented and described in detail in
196 Online Resource 3. Among these measurements, snout-vent length (SVL) was used
197 exclusively as a covariate in analyses of phenotypic differentiation. Descriptive statistics for
198 the morphometric measurements in each sampling locality are given in Online Resource 3 and
199 Online Resource 4.
200
Molecular data. Genomic DNA was extracted using a standard cetyl trimethyl
201 ammonium bromide (CTAB) protocol. For the amplification of 16S ribosomal RNA gene, the
202 primers 16Sar and 16Sbr were used according to Palumbi (1996). This gene is the most
203 commonly used marker for amphibian systematics (Fouquet et al. 2007), including studies on
204 the superfamily Dendrobatoidea (Vences et al. 2000; Grant et al. 2006; Santos et al. 2009). It
205 shows high rates of amplification success (Vences et al. 2005), having satisfactory
206 performance at the intraspecific level, as well as among closely related species (Lougheed et
207 al. 2006; Vences et al. 2005; Simões 2010).
208
Polymerase Chain Reaction (PCR) amplifications were performed in 15 µL reaction
209 volumes containing 5.8 µL ddH2O, 1.5 µL of 25mM MgCl2, 1.5 µL of 10X amplification
210 buffer (75mM Tris HCl, 50 mM KCl, 20 mM (NH4)2SO4), 1.5 µL of a 2 µM solution of each
60
211 primer, 1.5 µL of 10 mM dNTPs (2.5mM each dNTP), 0.7 µL of Taq DNA polymerase 2.5
212 U/µL (Biotools, Spain) and 1 µL of template DNA (about 40 ng/ µL). The temperature profile
213 consisted of 1) preheating at 90 °C for 60 s, 2) denaturation at 92 °C for 60 s, 3) primer
214 annealing at 50 °C for 50 s, 4) extension at 72 °C for 90 s, and 5) a final extension at 72 °C
215 for 5 min. Steps 2-4 were repeated 35 times. Amplicons were purified with exonuclease and
216 alkaline phosphatase (Fermentas Life Sciences, Canada), and sequenced using forward
217 primers and ABI BigDye Terminator Cycle Sequencing Kit protocols, as indicated by the
218 manufacturer. The resulting single-stranded products were read in an ABI 3130xl automatic
219 capillary sequencer. The DNA sequences had their homologous regions aligned in BioEdit
220 (Hall 1999), where a preliminary alignment was generated by the ClustalW algorithm
221 (Thompson et al. 1994), and were finally checked and corrected by eye in comparison with
222 the original chromatograms. The final alignments were 527 bp long for Allobates
223 paleovarzensis, and 539 bp long for A. nidicola and A. masniger. Newly determined
224 sequences were deposited in GenBank (Online Resource 5).
225 Data analysis
226
Population analyses. The genealogical relationships among individual samples and
227 populations were estimated through haplotype networks using statistical parsimony
228 (Templeton et al. 1992) in TCS 1.21 (Clement et al. 2000). Gaps were considered as 5th
229 character state, and the connection limit was fixed at 23 steps to allow association among all
230 sampled haplotypes. We estimated the relative partitioning of genetic variation (within
231 populations, among populations, and between riversides) through standard molecular analysis
232 of variance (AMOVA; Excoffier et al. 1992), in Arlequin version 3.5 (Excoffier and Lisher
233 2010). The relative genetic structuring among sampling localities was estimated through the
234 calculation of distance-based fixation indices (FST; Wright 1951), equivalent to NST of Lynch
235 and Crease (1992). The statistical significance of FST values was obtained after 1023
236 haplotype permutations in Arlequin. Genetic differentiation among sampling localities was
237 estimated via Kimura 2-parameter corrected genetic distances (Kimura 1980) in MEGA
238 version 5 (Tamura et al. 2011).
239
The most likely number of genetic clusters formed by the sampled mitochondrial DNA
240 sequences was inferred through Bayesian analysis of population structure, in BAPS version 5
241 (Corander et al. 2008). Based on nucleotide frequencies, the model attempts to create k groups
242 of individuals, such that those allocated in the same group resemble each other genetically as
61
243 much as possible. The upper limit to the number of clusters was set from one to the number of
244 sample localities on each study system. The log-likelihood values of the best models were
245 used to select the most probable cluster arrangement.
246
Within each of the genetic clusters inferred by the Bayesian approach, we calculated
247 haplotype and nucleotide diversities, as well as the number of segregating sites, and
248 performed the neutrality tests Tajima’s D, Fu’s Fs, and Ramos-Onsis & Rozas’ R2 (Tajima
249 1989; Fu 1997; Ramos-Onsins and Rozas 2002). The statistical significance of these tests was
250 estimated through coalescent simulations with 10000 replicates in DnaSP 5.10 (Librado and
251 Rozas 2009). In addition to the neutrality tests, past demographic expansion events were
252 inferred from the distribution of the observed pairwise nucleotide site differences (mismatch
253 distributions) in each cluster, as calculated in DnaSP 5.10. We used coalescent simulation
254 procedures in order to test the validity of the estimated expansion models. The parameters of
255 these models were re-estimated in order to obtain their empirical confidence intervals, and the
256 empirical distribution of the sum of squared deviations between the observed and the
257 expected mismatch values (SSD; Schneider and Excoffier 1999) and the raggedness index
258 (Hri; Harpending 1994). These calculations were obtained through 10000 parametric
259 bootstrap replicates in Arlequin.
260
Divergence time estimation. Divergence times were estimated using BEAST 1.6.2
261 (Bayesian Evolutionary Analysis Sampling Trees; Drummond and Rambaut 2007). We
262 analyzed only unique haplotypes recovered from the population analyses conducted in
263 Arlequin. An available 16S sequence of Allobates talamancae (GenBank AY843577) was
264 considered as an external group for both systems. In addition, 16S sequences of A. trilineatus
265 (GenBank DQ502118) and A. brunneus (GenBank EU342522) were used as outgroups for the
266 paleovarzensis and nidicola-masniger systems, respectively. As priors (calibration points), we
267 used average means and confidence intervals of the most recent common ancestors of both the
268 external groups and the studied systems obtained from Santos et al. (2009). Models of
269 nucleotide substitution were estimated for the reduced mitochondrial DNA datasets in MEGA
270 version 5 (Tamura et al. 2011). For the two study systems, we applied an uncorrelated
271 lognormal relaxed clock model, a lognormal prior distribution, a randomly generated starting
272 tree, and length of chain of 10 million generations with samples taken every 10 thousand
273 generations, and discarding 10% of the trees as burn-in, resulting in 900 sampled trees in the
274 Markov Chain Monte Carlo (MCMC). The stationarity of the posterior distributions for all
275 model parameters, including medians and 95% Highest Posterior Density intervals (HPD) of
62
276 the nodes was checked on Tracer 1.5 (Rambaut and Drummond 2007). From the MCMC
277 output, we generated the final consensus tree using Tree Annotator 1.6.2 (Drummond and
278 Rambaut 2007).
279
Phenotypic differentiation. In order to test whether the genetic clusters assigned via
280 Bayesian analysis of population structure were phenotypically distinguishable, we used
281 discriminant function analyses (DFA) with acoustic and morphometric traits as predictor
282 variables. The discriminant function was used to calculate the probability of classification of
283 each individual collected in its genetic cluster (grouping variable) by a Jackknifed
284 classification matrix. This approach included all DNA-sequenced individuals with available
285 measurements regarding acoustic and/or morphometric traits. Acoustic data were both
286 temperature and temperature-SVL adjusted through linear regression. In order to account for
287 the effect of body size on morphometric measurements, we used 12 morphological ratios as
288 predictor variables (Verdade and Rodrigues 2007): LL/SVL, HAND3/SVL, FL/SVL,
289 HW/HL, EN/HL, EL/HL, TYM/HL, TYM/EL, IN/HW, HAND1/HAND2, HAND2/HAND3,
290 HAND4/HAND1 (see Online Resource 3 for acronyms).
291
Geographic correlates of phenotypic and genetic variation. We tested correlations
292 between the linear geographic distances (measured in km) between sampling sites and genetic
293 and phenotypic distances between the populations of each of the study systems by applying
294 Mantel tests on distance matrices derived from genetic, acoustic and morphological data sets
295 (Mantel 1967). In addition, we conducted partial Mantel tests to evaluate correlations between
296 genetic/phenotypic distances among populations of each study system and their separation by
297 the river channel while controlling for effects of geographic distance between sampling sites
298 (Smouse et al. 1986; Telles et al. 2001). We constructed a binary correspondence matrix,
299 designating the values “0” and “1” for localities within the same and between opposite sides
300 of the river, respectively.
301
Average genetic distances (16S) between sampling localities were calculated
302 according to the Kimura 2-parameter (K2P) model in MEGA version 5. Acoustic distance
303 matrices were obtained from the dataset of temperature-adjusted call measurements by
304 calculating pairwise Euclidean distances between all possible pairs of populations by using
305 the scores of the mean acoustic measurements on first and second components (which
306 together explained 79.0% of the total acoustic variation in A. paleovarzensis and 79.3% in A.
307 nidicola/masniger) produced by a principal components analysis (PCA). The principal
63
308 components were obtained from the arithmetic means of the call traits among all sampled
309 individuals for each of the localities, and were used in order to reduce the number of
310 independent phenotypic variables. Morphological distance matrices were generated from the
311 external measurements (except SVL) through the same procedure. The first two components
312 explained 64.3% of the total morphological variation in A. paleovarzensis and 71.1% in A.
313 nidicola/masniger. As body size was correlated with the scores of the first principal
2
314 component of both acoustic (linear regression r = 0.51, F = 9.39, p = 0.017 in A.
315 paleovarzensis, and r2 = 0.72, F = 23.78, p = 0.001 in A. nidicola/masniger) and
2
316 morphological data sets (linear regression r
= 0.83, F = 43.65, p < 0.001 in A.
317 paleovarzensis, and r2 = 0.95, F = 174.09, p < 0.001 in A. nidicola/masniger), we regressed
318 them against corresponding mean SVL measurements for each population, and used the
319 residuals as new, size-independent acoustic variables from which new Euclidean distances
320 were calculated. Body size adjustments are useful for controlling possible ontogenetic effects
321 that might affect the outcome of comparative results. However, they can mask important call
322 differentiation related to body size dissimilarities among populations and species, which were
323 reported in the studied systems (Kaefer and Lima 2012; Kaefer et al. 2012). Therefore, we
324 also calculated acoustic distance matrices without applying SVL corrections. Mantel tests
325 were conducted in ZT (Bonnet and Van der Peer 2002) using permutation of the null models
326 (Anderson and Legendre 1999), and applying 10 000 randomizations.
327
328 Results
329 Population analyses
330
We obtained 276 16S rDNA sequences (Table 1), corresponding to 55 different
331 haplotypes (Online Resource 5). The distribution of haplotypes among the sampled localities
332 and the large number of mutations separating the genealogical groups showed a pronounced
333 genetic structure among populations, with restricted haplotype sharing across riverbanks in A.
334 paleovarzensis, while absent in the nidicola-masniger system (Fig. 1B and 2B). The AMOVA
335 indicated that a small fraction of the overall genetic variation can be found within sampling
336 localities, being concentrated among populations in the paleovarzensis system (60.37%) and
337 between opposite riverbanks in the nidicola-masniger system (59.74%) (Table 2). The FST
338 values obtained for both study systems revealed an overall high and significant population
64
339 structuring (Table 3). Genetic distances were more pronounced between populations of the
340 nidicola-masniger system (reaching 5.2%) than those observed in A. paleovarzensis (reaching
341 1.7%), thus lending further support to the interspecific nature of the former.
342
Bayesian analysis of genetic structure supported the partition of the individuals of A.
343 paleovarzensis into four genetic clusters (log ML value = -525.13; probability = 0.99), while
344 seven groups were estimated for the nidicola-masniger system (log ML value = -1182.18;
345 probability = 1.00). Co-occurrence of genetic clusters within sampling localities was nearly
346 absent. The distribution of sampling localities exerted a clear effect on the determination of
347 the genetic groupings, especially in A. nidicola and A. masniger, where six populations
348 corresponded to completely segregated clusters (Fig. 3). The three neutrality tests indicated
349 historical demographic changes in two clusters composed by individuals from almost the
350 entire range of A. paleovarzensis. In contrast, in the nidicola-masniger system the cluster
351 composed by the locality Road to Apuí was the only one to show significant departures from
352 historical demographic equilibrium according to the three neutrality tests, including the R2,
353 which is the most appropriate for small samples (Ramos-Onsins and Rozas 2002) (Table 4).
354 The tests of demographic growth based on the sum of squared deviation (SSD) between
355 observed and expected mismatch distributions, and Harpending’s raggedness index (Hri) did
356 not reject the demographic expansion hypothesis, giving support for the significant values
357 observed in the neutrality tests (Table 5, Online Resource 6).
358 Divergence time estimation
359
The Bayesian evolutionary analysis indicated the time of the first divergence between
360 the clades of A. paleovarzensis occupying the north and south sides of the middle Amazon
361 River to be on Pleistocene (median = 1.24 million years (mya); HPD = 0.54, 2.23). The
362 median time of the most recent common ancestor for the species was estimated at 2.29 mya
363 (HPD = 1.23, 3.81), and most of the intraspecific cladogenetic events in A. paleovarzensis are
364 indicated as having occurred during the last 1 million years (Fig. 4A). The split of the
365 nidicola-masniger system between the left and right sides of the lower Madeira River and the
366 time of the most recent common ancestor for this study system did not have occurred after
367 Early Pliocene (HPD = 6.88, 15.33; median = 10.65 mya;), with the subsequent cladogenetic
368 events evenly distributed from the time of this divergence to the present (Fig. 4B).
65
369 Phenotypic differentiation
370
Discriminant function analyses using genetic clusters as grouping variables was able
371 to correctly classify only 50.0% of the individuals of Allobates paleovarzensis according to
372 temperature-adjusted acoustic characters, and 43.0% of the individuals with both temperature
373 and body size-adjusted traits. Based on morphological characters, the percentage of correct
374 membership assignment of A. paleovarzensis was even poorer: 38.0% (Online Resource 7).
375 Individuals of A. nidicola and A. masniger were more successfully classified than those of A.
376 paleovarzensis in their respective genetic clusters based on acoustic characters adjusted for
377 temperature (59.0%), and both temperature and body size (48.0%). Morphological characters
378 also performed worse than acoustic traits within the nidicola-masniger system (41.0% of
379 individuals correctly classified; Online Resource 8).
380 Geographic correlates of phenotypic and genetic variation
381
The Mantel tests showed a significantly positive correlation between geographic and
382 genetic distances in both study systems, indicating a strong role of stochastic processes in the
383 process of evolutionary diversification. In contrast, correlations between phenotypic (acoustic
384 and morphological) and geographic distances were not significant (Table 6). We did not
385 observe signal of the Amazon River as a barrier promoting differentiation in A.
386 paleovarzensis in genetic or phenotypic distances. On the other hand, the Partial Mantel tests
387 revealed a role of the lower Madeira River as a vicariant barrier on the nidicola-masniger
388 system. Significant correlations were found between side of river and both overall genetic
389 differentiation and uncorrected acoustic distances.
390
391 Discussion
392 Pronounced population genetic structure
393
The high levels of genetic structure at the mitochondrial level found among
394 populations, especially in A. nidicola and A. masniger, are notable even considering earlier
395 reports on this subject, which indicate that panmictic scenarios are the exception among
396 amphibian populations (Crawford 2003; Zeisset and Beebee 2008). The main life history
397 characteristics pointed out to promote and maintain high genetic structure levels in anurans,
398 such as small body size and strong site attachment (Vences and Wake 2007) are widely
66
399 reported among Allobates (Pröhl 2005; Kaefer et al. in press). In addition, the studied species
400 are reported to be territorial (Lima et al. 2010; Caldwell and Lima 2003), and to exhibit direct
401 larval development in the nidicola-masniger system (Tsuji-Nishikido et al. 2012). Given that
402 water bodies are believed to constitute the main extraterritorial resource for dendrobatoid
403 species (Pröhl 2005), the absence of tadpole transport behavior may further restrict the use of
404 space of adult frogs, as well as concentrate offspring close to parental individuals,
405 contributing to the limited gene flow among lineages. The effects of this presumed high
406 phylopatry in A. nidicola and A. masniger may also be observed from the results of the
407 analysis of molecular variance, where a low within-population molecular variability was
408 observed for this study system (only 2.02 % of the total variation).
409
The configuration of genetic clusters recovered from the Bayesian analyses showed
410 significant genetic structuring over short geographic distances, resembling the pattern
411 observed in the haplotype networks. The star-shaped network of A. paleovarzensis, together
412 with the results of the neutrality tests, indicated that this species experienced recent
413 demographic changes across most of its range. Such patterns were less pronounced in the
414 molecular data of A. nidicola and A. masniger. Although the population growth tests did not
415 reject the population expansion hypothesis, the highly structured haplotype network and the
416 non-significant results in neutrality tests in most of the genetic clusters did not support past
417 demographic expansion events within A. nidicola and A. masniger. A mixed result in
418 neutrality tests was also observed through the range of A. femoralis along the Madeira River,
419 and might be related to demographic expansion of determinate populations which experienced
420 dispersal events, probably allowed by historical river course vagility (Simões 2010).
421 Spatio-temporal contexts of the inferred vicariant events
422
Both study systems exhibited signatures in their genetic structures that are congruent
423 with the transposition of large Amazonian rivers. The effectiveness of the lower Madeira
424 River as a barrier to gene flow in Allobates was observed to be higher than that of the middle
425 Amazon River. This observation is supported by the (1) absence of haplotype sharing among
426 sides of the Madeira River (which was restricted to different banks of the Amazon); (2)
427 accountability for most (60.37%) of the genetic variability in A. nidicola and A. masniger (the
428 river barrier was the second major factor explaining molecular variation within A.
429 paleovarzensis); (3) effect on overall variability of acoustic characters (significant differences
430 between interfluves were restricted to determinate call traits in A. paleovarzensis; (Kaefer and
67
431 Lima 2012); (4) Existence of reciprocally monophyletic clades on opposite sides of the
432 Madeira river (were paraphyletic between sides of the Amazon). This last characteristic
433 satisfies even the most stringent predictions of the classic river barrier hypothesis (Gascon et
434 al. 1996; Gascon et al. 1998), indicating the completeness of the speciation process in the
435 nidicola-masniger system according to both the evolutionary (sensu Wiley 1978) and
436 phylogenetic (Rosen 1978) species concepts. Under neutrality, the usual chronology of
437 evolutionary events leading towards speciation begins with the stage of polyphyly in a gene
438 tree, with successive paraphyly stages before the eventual achievement of the status of
439 reciprocal monophyly (de Queiroz 2007). According to this model, the monophyly observed
440 in the gene tree of the nidicola-masniger system is expected to arise from old geographic
441 barriers, in which diverging populations have experienced the time necessary to achieve such
442 status. Nevertheless, non-monophyly (both paraphyly and polyphyly) can be observed in
443 many of the currently recognized species (Funk and Omland 2003; McKay and Zink 2009).
444
Idiosyncratic characteristics of the sampling design and the history of the studied
445 species can be related the differential roles of the rivers as barriers that have been reported
446 (Colwell 2000). In this study, these system-specific characteristics may include: (1) Different
447 sampling designs: while the effect of the Madeira River as a barrier was accessed in its lower
448 portion, where it is believed to play a significant role as barrier to gene flow (Simões 2010),
449 A. paleovarzensis occurs along the middle section of the Amazon River, where it was
450 observed to exert intermediate effects as a vicariant barrier (Hayes and Sewlal 2004). (2)
451 Distinct study models: although belonging to the same genus, aspects of the natural history of
452 the studied species, such as dispersal abilities, might be responsible for distinct
453 microevolutionary
arrangements (see discussion on genetic structure above). (3)
454 Asynchronous divergence times: besides the large credibility intervals estimated by the
455 Bayesian analyses, the median dates for the split of clades inhabiting opposite sides of the
456 river in both study systems indicates that these events took place in different geological
457 epochs. The cladogenic event separating haplotypes of A. paleovarzensis inhabiting the north
458 and south of the Amazon River probably occurred during the Pleistocene, being younger that
459 the divergence time estimated between A. nidicola and A. masniger, which might have not
460 occurred after the early Pliocene. According to these findings, we infer that A. paleovarzensis
461 might not have undergone the level and duration of genetic isolation required for population
462 differentiation to cross the threshold necessary to met the criteria of speciation under
463 monophyly- and phenotype-based species concepts (Coyne and Orr 2004).
68
464
Our timetree suggests that the lineages of A. paleovarzensis occupying opposite
465 riverbanks of the middle Amazon River diverged during the late phase of its formation, when
466 its fan experienced an expressive increase in sedimentation rates and high vagility in the
467 course of its drainage (Rossetti et al. 2005; Figueiredo et al. 2009). Therefore, although the
468 Amazon River initiated as a transcontinental watercourse between 11.8 and 11.3 Mya ago, we
469 believe that sedimentation-induced channel course dynamics, which are evidenced by
470 Quaternary palaeochannels and meander relicts (Costa et al. 2001), allowed dispersal events
471 across its middle section. The late Miocene split between A. nidicola and A. masniger at the
472 portion of the Madeira River where it meets the Amazon seems to match the onset of the
473 latter as a channel running from the Andes to the Atlantic Ocean (Hoorn 1994; Figueiredo et
474 al. 2009). The estimated age of this evolutionary split is in accordance with studies that
475 indicated that most of the speciation events in Amazonia largely predated the Pleistocene
476 (Hoorn et al. 2010; Ribas et al. 2011). For amphibians, the origin of most of the current
477 Amazonian species is estimated to have occurred during late Miocene (Santos et al. 2009;
478 Funk et al. 2012; this study), with estimates reaching the late Oligocene (Elmer et al. 2007;
479 Fouquet et al. 2012).
480
The present appraisal of the effects of rivers as barriers must be interpreted cautiously
481 as they represent byproducts of a multi-character approach conducted in species with
482 restricted distributions. Investigation on this subject should involve species (or species
483 groups) that can be found throughout the course of a river to be assessed, thus allowing for a
484 multiple paired sampling design (e.g., Silva and Patton 1998; Simões 2010). In addition,
485 nuclear markers should be considered in future genetic studies in order to access their
486 potential in disentangle chronological aspects on the history of the studied populations (Zhang
487 and Hewitt 2003). In spite of the limitations of our data to provide accurate details on the
488 temporal and spatial mechanisms by which these vicariant barriers operate, the results
489 presented here and elsewhere (Kaefer and Lima 2012; Kaefer et al. 2012; Tsuji-Nishikido et
490 al. 2012) support the notion that the Amazon and Madeira Rivers do play a role in the
491 distribution of the genetic and phenotypic variability of the studied systems. Our conclusions
492 add to a growing body of studies that supported the effectiveness of the Amazon (e.g.,
493 Capparella 1988; Hayes and Sewlal 2004), Madeira (e.g., Cohn-Haft 2000; Simões et al.
494 2008; Fernandes et al. 2012) and other Amazonian rivers (e.g., Funk et al. 2007; Ribas et al.
495 2011; Fouquet et al. in press) as vicariant barriers promoting and maintaining evolutionary
496 diversity. In this context, the various studies conducted at the Juruá River, another southern
69
497 tributary of the Amazon, which rejected the river barrier hypothesis (e.g., Patton et al. 1994;
498 Gascon et al. 1996; 1998; 2000; Lougheed et al. 1999) seem to reflect the ineffectiveness of
499 this particular meandering river as a barrier to gene flow rather than reject the river barrier
500 hypothesis for the entire biome.
501 Stochastic and selective forces in evolutionary differentiation
502
In addition to the genetic discontinuities associated with the transposition of the large
503 Amazonian rivers, results from both study systems suggested that a large part of their genetic
504 variation is correlated with geographic distance. This was more evident in A. paleovarzensis,
505 for which Mantel tests indicated a pronounced correlation between genetic and geographic
506 distances together with a non-significant role of the river barrier in genetic differentiation.
507 Distance effects have been reported from phylogeographic studies involving Amazonian frogs
508 (Funk et al. 2007; Amézquita et al. 2009; Simões 2010) and suggest an important role for
509 isolation by distance and genetic drift on their evolutionary diversification. Divergent sexual
510 selection was observed to occur in Amazonian frog species where genetic divergence and
511 reduced gene flow were associated with differential female preferences (Boul et al. 2007;
512 Guerra and Ron 2008). Therefore, a combination of stochastic and selective forces seems to
513 have contributed to the richness of frog lineages observed in Amazonia. This assumption is in
514 line with the idea that neutral and selective processes are non-exclusive explanations for the
515 evolution of a given character (Hansen and Orzack 2005).
516
The acoustic and morphological traits analyzed in this study did not respond to linear
517 geographic distance among sampling localities, as observed with molecular characters. In
518 addition, the discriminant function analyses revealed that a large fraction of the individuals
519 could not be correctly classified in their respective genetic clusters according to their
520 phenotypes. This misclassification may be attributed to a single marker-effect, given that
521 mitochondrial genealogies may exhibit patterns of variation different that those obtained from
522 the analysis of nuclear markers (e.g., Sequeira et al. 2011; Turmelle et al. 2011, but see Creer
523 et al. 2004; Zink and Barrowclough 2008). Nonetheless, the significant correlations between
524 our mitochondrial data and the main geographic barriers of the study areas (river and
525 distance) suggest that the 16S marker was sensitive to the recent history of connectedness
526 among these populations. Therefore, it is reasonable to explore other explanations to the
527 relationship between the evolution of genetic and phenotypic characters.
70
528
The results of this study, together with the observation that there is low among-
529 population variability in phenotypic characters (Kaefer and Lima 2012; Kaefer et al. 2012),
530 indicate a prevailing role of stabilizing selective forces in phenotypic differentiation. In this
531 case, evolutionary constraints may be preventing segregation in morphology and calls among
532 populations. Additionally, the time elapsed since the allopatric condition may not have been
533 sufficient for the accumulation of significant differences.
534
Supporting the first explanation is the observation that well diverged lineages of frogs
535 generally exhibit low phenotypic differentiation (Cherry et al. 1978; Wiens 2008), and that
536 the functional importance of these characters imposes evolutionary constraints (Lougheed et
537 al. 2006). According to the phylogenetic niche conservatism hypothesis, an important and
538 often neglected component of allopatric speciation is the tendency of lineages to maintain
539 their ancestral ecological niche (Wiens 2004). Our results appear to support the occurrence of
540 this phenomenon, as it predicts that closely related lineages are more ecologically similar that
541 would be expected based on their phylogenetic relationships (Losos 2008). This evolutionary
542 conservatism, in which diversification can occur with little phenotypic change (Adams et al.
543 2009) might be related, for example, to the historical failure of researchers in providing
544 reasonable estimates of species richness for many anuran groups based on the classic
545 morphological and biological species concepts (Camargo et al. 2006; Fouquet et al. 2007;
546 Angulo and Reichle 2008; Funk et al. 2012). In both studied systems, bioacoustic characters
547 did not echo fine-scale phylogeographic structure, supporting the notion that mate-recognition
548 signals are subject to severe constraining or diversifying pressures and therefore are not
549 closely related to geographic distance (Gerhardt and Huber 2002). In fact, behavioral traits are
550 expected to be relatively evolutionarily malleable (Blomberg et al. 2003). However,
551 exceptions to this pattern were reported (Ryan et al. 1996; Wycherley et al. 2002; Pröhl et al.
552 2007; Amézquita et al. 2009), depicting the large array of evolutionary forces underlying
553 geographic divergence, as well as the idiosyncratic nature of their prevalence.
554
In favour of the second interpretation is the observation that the only the more
555 differentiated system, composed by A. nidicola and A. masniger, exhibited significant
556 differences in summarized acoustic parameters between riversides. It suggests that the
557 diversifying effects of genetic drift assumed prominence in ancient, isolated lineages simply
558 because of the accumulation of mutations over a longer time span. This interpretation also
559 assumes that, under non-selective evolutionary forces, neutrally evolving genotypic characters
560 diverge first, and that more time is needed for phenotypic differences to arise.
71
561 Conclusions
562
This study indicated that the distribution of genetic variability in three species of
563 neotropical frogs is determined by a combination of the effects of isolation-by-distance and
564 the transposition of large Amazonian rivers acting as vicariant barriers. Thus, we confirm our
565 expectation that genetic drift plays a central role in the emergence of evolutionary
566 differentiation that may lead to the formation of new species. In addition, the current
567 phenotypic divergence was loosely related with genetic differentiation and did not
568 successfully predict assignment of individuals to genetic groups. These results argue against
569 the hypothesis that the phenotypic differentiation is related to the genetic groupings that
570 compose the study systems, and are surprising when considering the high levels of genetic
571 structuring observed. Together with recent studies conducted on these model organisms
572 (Kaefer and Lima 2012; Kaefer et al. 2012; Tsuji-Nishikido et al. 2012), we suggest a
573 prevailing role of stabilizing selective forces in phenotypic differentiation. In this case,
574 evolutionary constraints, which have no effect on neutrally evolving molecular markers, may
575 be preventing phenotypic differentiation.
576
577 Ethical standards
578
This study complies with the current Brazilian laws and was allowed by RAN-
579 ICMBio/IBAMA (licences 13777-2, 18516-1, 20065-2, and 21950-1).
580
581 Conflict of interest
582
The authors declare that they have no conflict of interest.
583
584 Acknowledgements
585
We thank Anelise Montanarin, Francisco C. de Freitas, Irene da S. Melo, Maria A.
586 Carvalho, Moisés da S. Melo and Raimundo N. Amorim for fieldwork assistance; Daniela
587 Leroy e Waleska Gravena for help in laboratory procedures; Adolfo Amézquita, Adrian
72
588 Garda, Andrew J. Crawford, Janet W. Reid, Jeffrey Podos, Luciana K. Erdtmann, Marcelo
589 Menin, Marina Anciães, Pedro Ivo Simões and Tomas Hrbek for valuable suggestions during
590 this study. We also thank the Brazilian Conselho de Desenvolvimento Científico e
591 Tecnológico for financial support (CNPq-CT Amazônia 553997/2006-8 and 575572/2008-6).
592
593 References
594 Adams, D. C., Berns, C. M., Kozak, K. H. & Wiens, J. (2009). Are rates of species
595 diversification correlated with rates of morphological evolution? Proceedings of the Royal
596 Society B, 276, 2729–2738.
597 Amézquita, A., Lima, A. P., Jehle, R., Castellanos, L., Ramos, O., Crawford, A. J., Gasser, H.
598 & Hödl, W. (2009). Calls, colours, shapes, and genes: A multi–trait approach to the study of
599 geographic variation in the Amazonian frog Allobates femoralis. Biological Journal of the
600 Linnean Society, 98, 826–838.
601 Anderson, M. J. & Legendre, P. (1999). An empirical comparison of permutation methods for
602 tests of partial regression coefficients in a linear model. Journal of Statistical Computation
603 and Simulation, 62, 271–303.
604 Angulo, A. & Reichle, S. (2008). Acoustic signals, species diagnosis, and species concepts:
605 the case of a new cryptic species of Leptodactylus (Amphibia, Anura, Leptodactylidae) from
606 the Chapare region, Bolivia. Zoological Journal of the Linnean Society, 152, 58–77.
607 Antonelli, A., Quijada-Mascareñas, A., Crawford, A. J., Bates, J. M., Velazco, P. M & Wüster
608 W. (2010). Molecular studies and phylogeography of Amazonian tetrapods and their relation
609 to geological and climatic models. In C Hoorn, F. P. Wesselingh (Eds.), Amazonia, landscape
610 and species evolution (pp. 386–403). Oxford: Blackwell Publishing.
611 Avise, J. C. (2004). Molecular Markers, Natural History and Evolution. 2nd edition.
612 Sunderland: Sinauer Associates Inc.
613 Avise, J. C. (2000). Phylogeography: the history and formation of species. Cambridge:
614 Harvard University Press.
73
615 Ayres, J. M. & Clutton-Brock, T. H. (1992). River boundaries and species range size in
616 Amazonian primates. The American Naturalist, 140, 531–537.
617 Blomberg, S. P., Garland Jr. T. & Ives, A. R. (2003). Testing for phylogenetic signal in
618 comparative data: behavioral traits are more labile. Evolution, 57, 717–745.
619 Bonnet, E. & Van de Peer, Y. (2002). ZT: a software tool for simple and partial Mantel tests.
620 Journal of Statistical Software, 7, 1–12.
621 Boul, K. E., Funk, W. C., Darst, C. R., Cannatella, D. C. & Ryan, M. J. (2007). Sexual
622 selection drives speciation in an Amazonian frog. Proceedings of the Royal Society B, 274,
623 399–406.
624 Caldwell, J. P. & Lima, A. P. (2003). A new Amazonian species of Colostethus (Anura:
625 Dendrobatidae) with a nidicolous tadpole. Herpetologica, 59, 219–234.
626 Camargo, A., de Sá, R. & Heyer, W. R. (2006). Phylogenetic analyses of mtDNA sequences
627 reveal three cryptic lineages in the widespread neotropical frog Leptodactylus fuscus
628 (Schneider, 1799) (Anura, Leptodactylidae). Biological Journal of the Linnean Society, 87,
629 325–341.
630 Campbell, P., Pasch, B., Pino, J. L., Crino, O. L., Phillips, M. & Phelps, S. M. (2010).
631 Geographic variation in the songs of neotropical singing mice: testing the relative importance
632 of drift and local adaptation. Evolution, 64, 1955–1972.
633 Capparella, A. P. (1988). Genetic variation in Neotropical birds: implications for the
634 speciation process. In H. Ouellet (Ed.), Acta XIX Congressus Internationalis Ornithologici
635 (pp. 1658–1664). Ottawa: University of Ottawa Press.
636 Charif, R. A., Clark, C. W. & Fristrup, K. M. (2004). Raven 1.2 User’s Manual. Ithaca:
637 Cornell Laboratory of Ornithology.
638 Charif R. A., Waack, A. M. & Strickman, L. M. (2008). Raven Pro 1.3 User’s Manual.
639 Ithaca: Cornell Laboratory of Ornithology.
640 Cherry, L. M., Case, S. M. & Wilson, A. C. (1978). Frog perspective on the morphological
641 difference between humans and chimpanzees. Science, 200, 209–211.
74
642 Clement, M., Posada, D. & Crandall, K. A. (2000). TCS: a computer program to estimate
643 gene genealogies. Molecular Ecology, 9, 1657–1659.
644 Cohn-Haft, M. (2000). A case study in amazonian biogeography: Vocal and DNA sequence
645 variation in Hemitriccus flycatchers. PhD thesis. Luisiana State University.
646 Colwell, R.K. (2000). A barrier runs through it… or maybe just a river. Proceedings of the
647 National Academy of Sciences, 97, 13470–13472.
648 Corander, J., Marttinen, P., Sirén, J. & Tang, J. (2008). Enhanced Bayesian modelling in
649 BAPS software for learning genetic structures of populations. BMC Bioinformatics, 9, 539.
650 Costa, J. B. S., Bemerguy, R. L., Hasui, Y. & Borges, M. S. (2001). Tectonics and
651 paleogeography along the Amazon River. Journal of South American Earth Sciences, 14,
652 335–347.
653 Coyne, J. A. & Orr, H. A. (2004). Speciation. Sunderland: Sinauer Associates Inc.
654 Crawford, A. J. (2003). Huge populations and old species of Costa Rican and Panamanian dirt
655 frogs inferred from mitochondrial and nuclear gene sequences. Molecular Ecology, 12, 2525–
656 2540.
657 Crawford, A. J., Lips, K. R. & Bermingham, E. (2010). Epidemic disease decimates
658 amphibian abundance, species diversity, and evolutionary history in the highlands of central
659 Panama. Proceedings of the National Academy of Sciences, 107, 13777–13782.
660 Creer, S., Thorpe, R. S., Malhotra, A., Chou, W. H. & Stenson, A. G. (2004). The utility of
661 AFLPs for supporting mitochondrial DNA phylogeographical analyses in the Taiwanese
662 bamboo viper, Trimeresurus stejnegeri. Journal of Evolutionary Biology, 17, 100–107.
663 Crisci, J. V., Katinas, L., & Posadas, P. (2003). Historical biogeography: An Introduction.
664 Cambridge: Harvard University Press.
665 Dall, S. R. X. (1997). Behaviour and speciation. Trends in Ecology and Evolution , 12, 209–
666 210.
667 Dayrat, B. (2005). Towards integrative taxonomy. Biological Journal of the Linnean Society,
668 85, 407–415.
75
669 de Queiroz, K. (2007). Species concepts and species delineation. Systematic Biology, 56, 879–
670 886.
671 Diniz-Filho, J. A. F., Telles, M. P. C., Bonatto, S. L., Eizirik, E., Freitas, T. R. O., de Marco,
672 P., Santos, F. R., Solé–Cava, A. M. & Soares, T. N. (2008). Mapping the evolutionary
673 twilight zone: molecular markers, populations and geography. Journal of Biogeography, 35,
674 753–763.
675 Drummond, A. J. & Rambaut, A. (2007). BEAST: Bayesian evolutionary analysis by
676 sampling trees. BMC Evolutionary Biology, 7, 214.
677 Elmer K. R., Dávila, J.A. & Lougheed, S.C. (2007). Cryptic diversity and deep divergence in
678 an upper Amazonian leaflitter frog, Eleutherodactylus ockendeni. BMC Evolutionary Biology,
679 7, 247.
680 Erdtmann, L. & Amézquita, A. (2009). Differential evolution of advertisement call traits in
681 Dart–Poison Frogs (Anura: Dendrobatidae). Ethology, 115, 801–811.
682 Excoffier, L. & Lischer, H. E. L. (2010). Arlequin suite ver 3.5: A new series of programs to
683 perform population genetics analyses under Linux and Windows. Molecular Ecology
684 Research, 10, 564–567.
685 Excoffier, L., Smouse, P. & Quattro, J. (1992). Analysis of molecular variance inferred from
686 metric distances among DNA haplotypes: application to human mitochondrial DNA
687 restriction data. Genetics, 131, 479–491.
688 Fernandes, A. M., Wink, M. & Aleixo, A. (2012). Phylogeography of the chestnut tailed
689 antbird (Myrmeciza hemimelaena) clarifies the role of rivers in Amazonian biogeography.
690 Journal of Biogeography, doi:10.1111/j.1365-2699.2012.02712.x
691 Figueiredo, J., Hoorn, C., van der Ven, P. & Soares, E. (2009). Late Miocene onset of the
692 Amazon River and the Amazon deep-sea fan: Evidence from the Foz do Amazonas Basin.
693 Geology, 37, 619–622.
694 Fouquet, A., Gilles, A., Vences, M., Marty, C., Blanc, M. & Gemmell, N. J. (2007).
695 Underestimation of species richness in Neotropical frogs revealed by mtDNA analyses. PLoS
696 One, 10, e1109.
76
697 Fouquet, A., Noonan, B. P., Rodrigues, M. T., Pech, N., Gilles, A. & Gemmell, N. J.:
698 Multiple Quaternary Refugia in the Eastern Guiana Shield Revealed by Comparative
699 Phylogeography of 12 Frog Species. Systematic Biology, in press.
700 Fouquet, A., Recoder, R., Teixeira–Jr, M., Cassimiro, J., Amaro, R. C., Camacho, A.,
701 Damasceno, R., Carnaval, A. C., Moritz, C. & Rodrigues, M. T. (2012). Molecular phylogeny
702 and morphometric analyses reveal deep divergence between Amazonia and Atlantic Forest
703 species of Dendrophryniscus. Molecular Phylogenetics and Evolution, 62, 826–838.
704 Fu, Y. X. (1997). Statistical tests of neutrality of mutations against population growth,
705 hitchhiking and background selection. Genetics, 147, 915–925.
706 Funk, D. J. & Omland, K. E. (2003). Species-level paraphyly and polyphyly: frequency,
707 causes, and consequences, with insights from animal mitochondrial DNA. Annual Review of
708 Ecology, Evolution, and Systematics, 34, 397–423.
709 Funk, W. C., Caldwell, J. P., Peden, C. E., Padial, J. M., de la Riva, I. & Cannatella, D. C.
710 (2007). Tests of biogeographic hypotheses for diversification in the Amazonian forest frog,
711 Physalaemus petersi. Molecular Phylogenetics and Evolution, 44, 825–837.
712 Funk, W. C., Caminer, M. & Ron, S. R. (2012). High levels of cryptic species diversity
713 uncovered in Amazonian frogs. Proceedings of the Royal Society B, 279, 1806–1814.
714 Gascon, C., Lougheed, S. C. & Bogart, J. P. (1996). Genetic and morphological variation in
715 Vanzolinius discodactylus: a direct test of the riverine barrier hypothesis. Biotropica, 28, 376–
716 387.
717 Gascon, C., Lougheed, S. C. & Bogart, J. P. (1998). Patterns of genetic population
718 differentiation in four species of Amazonian frogs: A test of the Riverine Barrier Hypothesis.
719 Biotropica, 30, 104–119.
720 Gascon, C., Malcolm, J. R., Patton, J.L., Silva, M. N. F., Bogart, J.P., Lougheed, S.C., Peres,
721 C.A., Neckel, S. & Boag P.T. (2000). Riverine barriers and the geographic distribution of
722 Amazonian species. Proceedings of the National Academy of Sciences, 97, 13672–13677.
723 Gerhardt. H.C. & Huber, F. (2002). Acoustic Communication in Insects and Anurans:
724 Common Problems and Diverse Solutions. Chicago: University of Chicago Press.
77
725 Goicochea, N., De La Riva, I. & Padial, J. M. (2009). Recovering phylogenetic signal from
726 frog mating calls. Zoologica Scripta, 39, 141–154.
727 Grant, T., Frost, D. R., Caldwell, J. P., Gagliardo, R., Haddad, C. F. B., Kok, P. J. R., Means,
728 B. D, Noonan, B. P., Schargel, W. & Wheeler, W. C. (2006). Phylogenetic systematics of
729 dart–poison frogs and their relatives (Anura: Athesphatanura: Dendrobatidae). Bulletin of the
730 American Museum of Natural History, 299, 1–262.
731 Groeneveld, L. F., Weisrock, D. W., Rasoloarison, R. M., Yoder, A. D. & Kappeler, P. M.
732 (2009). Species delimitation in lemurs: multiple genetic loci reveal low levels of species
733 diversity in the genus Cheirogaleus. BMC Evolutionary Biology, 9, 30.
734 Guerra, M. A. & Ron, S. R. (2008). Mate choice and courtship signal differentiation promotes
735 speciation in an Amazonian frog. Behavioral Ecology, 19, 1128–1135.
736 Guillot, G., Renaud, S., Ledevin, R., Michaux, J. & Claude, J. (2012). A unifying model for
737 the analysis of phenotypic, genetic and geographic data. Systematic Biology, doi:
738 10.1093/sysbio/sys038.
739 Haffer, J. (1969). Speciation in Amazonian Forest Birds. Science, 165, 131–137.
740 Hall, T. A. (1999). BioEdit: a user-friendly biological sequence alignment editor and analysis
741 program for Windows 95/98/NT. Nucleic Acids Symposium Series, 41, 95–98.
742 Hansen, T. F. & Orzack, S.H. (2005). Assessing current adaptation and phylogenetic inertia as
743 explanations of trait evolution: the need for controlled comparisons. Evolution, 59, 2063–
744 2072.
745 Harpending, R. C. (1994). Signature of ancient population growth in a low–resolution
746 mitochondrial DNA mismatch distribution. Human Biology, 66, 591–600.
747 Hayes, F.E. & Sewlal, J.N. (2004). The Amazon River as a dispersal barrier to passerine
748 birds: effects of river width, habitat and taxonomy. Journal of Biogeography, 31, 1809–1818.
749 Hoorn, C. (1994). An environmental reconstruction of the palaeo-Amazon River system
750 (Middle-Late
Miocene,
NW
751 Palaeoecology, 112, 187–238.
Amazonia).
Palaeogeography,
Palaeoclimatology,
78
752 Hoorn, C., Wesselingh, F.P., ter Steege, H., Bermudez, M. A., Mora, A., Sevink, J.,
753 Sanmartín, I., Sanchez-Meseguer, A., Anderson, C. L., Figueiredo, J.P., Jaramillo, C., Riff,
754 D., Negri, F. R., Hooghiemstra, H., Lundberg, J., Stadler, T., Särkinen, T. & Antonelli, A.
755 (2010). Amazonia through time: Andean uplift, climate change, landscape evolution, and
756 biodiversity. Science, 330, 927–931.
757 Hubert, N., Meyer, C. P., Bruggemann, H. J., Guérin, F., Komeno, R. J. L., Espiau, B.,
758 Causse, R., Williams, J. T. & Planes, S. (2012). Cryptic diversity in Indo–Pacific coral–reef
759 fishes revealed by DNA–barcoding provides new support to the centre–of–overlap
760 hypothesis. PloS One, 7, e28987.
761 Hutchison, D. W. & Templeton, A. R. (1999). Correlation of pairwise genetic and geographic
762 distance measures: inferring the relative influences of gene flow and drift on the distribution
763 of genetic variability. Evolution, 53, 1898–1914.
764 Irion, G. & Kalliola, R. (2010). Long–term landscape development processes in Amazonia. In
765 C Hoorn, F. P. Wesselingh (Eds.), Amazonia, landscape and species evolution (pp. 185–197).
766 Oxford: Blackwell Publishing.
767 Kaefer, I. L. & Lima, A. P. (2012). Sexual signals of the Amazonian frog Allobates
768 paleovarzensis: geographic variation and stereotypy of acoustic traits. Behaviour, 149, 15–33.
769 Kaefer, I. L., Montanarin, A., Costa, R. S. & Lima, A. P. Temporal patterns of reproductive
770 activity and site attachment of the Brilliant–thighed Frog Allobates femoralis from Central
771 Amazonia. Journal of Herpetology, in press.
772 Kaefer, I.L., Tsuji-Nishikido, B. M. & Lima, A. P. (2012). Beyond the river: underlying
773 determinants of population acoustic signal variability in Amazonian direct–developing
774 Allobates (Anura: Dendrobatoidea). Acta Ethologica, doi:10.1007/s10211-012-0126-0.
775 Kimura, M. (1980). A simple method for estimating evolutionary rate of base substitutions
776 through comparative studies of nucleotide sequences. Journal of Molecular Evolution, 16,
777 111–120.
778 Lemmon, E. M. (2009). Diversification of conspecific signals in sympatry: geographic
779 overlap drives multidimensional reproductive character displacement in frogs. Evolution, 63,
780 1155–1170.
79
781 Librado, P. & Rozas, J. (2009). DnaSP v5: A software for comprehensive analysis of DNA
782 polymorphism data. Bioinformatics, 25, 1451–1452.
783 Lima, A.P., Caldwell, J. P., Biavati, G. & Montanarin, A. (2010). A new species of Allobates
784 (Anura: Aromobatidae) from paleovárzea forest in Amazonas, Brazil. Zootaxa, 2337, 1–17.
785 Losos, J. B. (2008). Phylogenetic niche conservatism, phylogenetic signal and the relationship
786 between phylogenetic relatedness and ecological similarity among species. Ecology Letters,
787 11, 995–1003.
788 Lougheed, S. C., Austin, J. D., Bogart, J. P., Boag, P.T. & Chek, A. A. (2006). Multi–
789 character perspectives on the evolution of intraspecific differentiation in a neotropical hylid
790 frog. BMC Evolutionary Biology, 6, 23.
791 Lougheed, S. C., Gascon, C., Jones, D. A., Bogart, J. P. & Boag, P. T. (1999). Ridges and
792 rivers: a test of competing hypothesis of Amazonian diversification using a dart–poison frog
793 (Epipedobates femoralis). Proceedings of the Royal Society B, 266, 1829–1835.
794 Lynch, M. & Crease, T. J. (1990). The analysis of population survey data on DNA sequence
795 variation. Molecular Biology and Evolution, 7, 377–394.
796 Mantel, N. (1967). The detection of disease clustering and a generalized regression approach.
797 Cancer Research, 27, 209–220.
798 McKay, B. D. & Zink, R. M. (2009). The causes of mitochondrial DNA gene tree paraphyly
799 in birds. Molecuar Phylogenetics and Evolution, 54, 647–650.
800 Morales, V. R. (2002). Sistematica y biogeografía del grupo trilineatus (Amphibia, Anura,
801 Dendrobatidae, Colostethus), con descripción de once nuevas especies. Publicación de la
802 Asociación Amigos Doñana, 13, 1–59.
803 Moritz, C., Patton, J. L., Schneider, C. J. & Smith, T. B. (2000). Diversification of rainforest
804 faunas: An integrated molecular approach. Annual Review of Ecology and Systematics, 31,
805 533–563.
806 Nosil, P., Harmon, L. J. & Seehausen, O. (2009). Ecological explanations for (incomplete)
807 speciation. Trends in Ecology and Evolution, 24, 145–156.
80
808 Palumbi, S. R. (1996). Nucleic acids II: the polymerase chain reaction. In D. M. Hillis, C.
809 Moritz, B. K. Mable (Eds.), Molecular Systematics (pp. 205–247). Sunderland: Sinauer
810 Associates Inc.
811 Panhuis, T. M., Butlin, R., Zuk, M. & Tregenza, T. (2001). Sexual selection and speciation.
812 Trends in Ecology and Evolution, 16, 364–371.
813 Patton, J. L., Silva, M. N. F. & Malcolm, J. R. (1994). Gene genealogy and differentiation
814 among arboreal spiny rats (Rodentia: Echymidae) of the Amazon Basin: a test of the riverine
815 barrier hypothesis. Evolution, 48, 1314–1323.
816 Pröhl, H. (2005). Territorial behavior in dendrobatid frogs. Journal of Herpetology, 39, 354–
817 365.
818 Pröhl, H., Hagemann, S., Karsch, J. & Höbel, G. (2007). Geographic variation in male sexual
819 signals in Strawberry Poison Frogs (Dendrobates pumilio). Ethology, 113, 825–837.
820 Pröhl, H., Koshy, R. A., Mueller, U. G., Rand, A. S. & Ryan, M. J. (2006). Geographic
821 variation of genetic and behavioral traits in northern and southern túngara frogs. Evolution,
822 60, 1669–1679.
823 Rambaut, A. & Drummond, A. J. (2007). Tracer v. 1.5. [http://beast.bio.ed.ac.uk/Tracer].
824 Accessed 04 May 2012.
825 Ramos–Onsins, S. E. & Rozas, J. (2002). Statistical properties of new neutrality tests against
826 population growth. Molecular Biology and Evolution, 19, 2092–2100.
827 Ribas, C. C., Aleixo, A., Nogueira, A. C. R., Miyaki, C. Y. & Cracraft, J. (2011). A
828 palaeobiogeographic model for biotic diversification within Amazonia over the past three
829 million years. Proceedings of the Royal Society B, 279, 681–689.
830 Rosen, D. E. (1978). Vicariant patterns and historical explanation in biogeography. Systematic
831 Zoology, 27, 159–188.
832 Rossetti, D. F., de Toledo, P. M. & Góes, A. M. (2005). New geological framework for
833 Western Amazonia (Brazil) and implications for biogeography and evolution. Quaternary
834 Research, 63, 78–89.
81
835 Ryan, M. J., Rand, A. S. & Weigt, L. A. (1996). Allozyme and advertisement call variation in
836 the túngara frog, Physalaemus pustulosus. Evolution, 50, 2435–2453.
837 Santini, F., Miglietta, M. P. & Faucci, A. (2012). Where are we now? An introduction to a
838 special issue on speciation. Evolutionary Biology, 39: 141-147.
839 Santos, J. C., Coloma, L. A., Summers, K., Caldwell, J. P., Ree, R. & Cannatella, D. C.:
840 (2009). Amazonian amphibian diversity is primarily derived from Late Miocene Andean
841 lineages. PLoS Biology, 7, e1000056.
842 Santos, S., Hrbek, T., Farias, I. P., Schneider, H. & Sampaio, I. (2006). Population genetic
843 structuring of the king weakfish, Macrodon ancylodon (Sciaenidae), in Atlantic coastal waters
844 of South America: deep genetic divergence without morphological change. Molecular
845 Ecology, 15, 4361–4373.
846 Schneider, S. & Excoffier, L. (1999). Estimation of demographic parameters from the
847 distribution of pairwise differences when the mutation rates vary among sites: Application to
848 human mitochondrial DNA. Genetics, 152, 1079–1089.
849 Sequeira, F., Sodré, D., Ferrand, N., Bernardi, J., Sampaio, I., Schneider, H. & Vallinoto, M.
850 (2011). Hybridization and massive mtDNA unidirectional introgression between the closely
851 related Neotropical toads Rhinella marina and R. schneideri inferred from mtDNA and
852 nuclear markers. BMC Evolutionary Biology, 11, 264.
853 Silva, M. N. F & Patton, J. L. (1998). Molecular phylogeography and the evolution and
854 conservation of mammals. Molecular Ecology, 7, 475–486.
855 Simões, P. I. (2010). Diversificação do complexo Allobates femoralis (Anura, Dendrobatidae)
856 em florestas da Amazônia brasileira: desvendando padrões atuais e históricos. PhD thesis.
857 Instituto Nacional de Pesquisas da Amazônia, Manaus.
858 Simões, P. I., Lima, A. P., Magnusson, W. E., Hödl, W. & Amézquita, A. (2008). Acoustic
859 and morphological differentiation in the frog Allobates femoralis: Relationships with the
860 Upper Madeira River and other potential geological barriers. Biotropica, 40, 607–614.
861 Sioli, H. (1984). The Amazon: Limnology and Landscape Ecology of a Mighty Tropical River
862 and its Basin. Dordrecht: Dr W. Junk Publisher.
82
863 Smouse, P. E., Long, J. C. & Sokal, R. R. (1986). Multiple regression and correlation
864 extensions of the Mantel test of matrix correspondence. Systematic Zoology, 35, 727–732.
865 Solomon, S. E., Bacci, M., Martins, J., Vinha, G. G. & Mueller, U. G. (2008).
866 Paleodistributions and comparative molecular phylogeography of leafcutter ants (Atta spp.)
867 provide new insight into the origins of Amazonian diversity. PLoS One, 3, e2738.
868 Tajima, F. (1989). Statistical method for testing the neutral mutation hypothesis by DNA
869 polymorphism. Genetics, 123, 585–595.
870 Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M. & Kumar, S. (2011). MEGA5:
871 Molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance,
872 and maximum parsimony methods. Molecular Biology and Evolution, 28, 2731–2739.
873 Telles, M. P. C., Silva, R. S. M., Chaves, L. J., Coelho, A. S. G & Diniz–Filho, J. A. F.
874 (2001). Divergência entre subpopulações de cagaiteira (Eugenia dysenterica) em resposta a
875 padrões edáficos e distribuição espacial. Pesquisa Agropecuária Brasileira, 36, 1387–1394.
876 Templeton, A. R., Crandall, K. A. & Sing, C. F. (1992). A cladistic analysis of phenotypic
877 associations with haplotypes inferred from restriction endonuclease mapping and DNA
878 sequence data. III. Cladogram estimation. Genetics, 132, 619–633.
879 Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994). Improved sensitivity of profile
880 searches through the use of sequence weights and gap excision. Computer Applications in the
881 Biosciences, 10, 19–29.
882 Tsuji-Nishikido, B., Kaefer, I. L., Freitas, F. C., Menin, M. & Lima, A. P. (2012). Significant
883 but not diagnostic: Differentiation through morphology and calls in the Amazonian frogs
884 Allobates nidicola and A. masniger. Herpetological Journal, 22, 105–114.
885 Turmelle, A. S., Kunz, T. H. & Sorenson, M. D. (2011). A tale of two genomes: contrasting
886 patterns of phylogeographic structure in a widely distributed bat. Molecular Ecology, 20,
887 357–375.
888 Vences, M., Kosuch, J., Lötters, S., Widmer, A., Jungfer, K., Köhler, J. & Veith, M. (2000).
889 Phylogeny and classification of Poison Frogs (Amphibia: Dendrobatidae), based on
890 mitochondrial 16S and 12S ribosomal RNA gene sequences. Molecular Phylogenetics and
891 Evolution, 15, 34–40.
83
892 Vences, M., Thomas, M., Meijden, A., Chiari, Y. & Vieites, D. R. (2005). Comparative
893 performance of the 16S rRNA gene in DNA barcoding of amphibians. Frontiers in Zoology,
894 2, 5.
895 Vences, M. & Wake, D. B. (2007). Speciation, species boundaries and phylogeography of
896 amphibians. In HH Heatwole, M. Tyler (Eds.), Amphibian Biology. Volume 6: Systematics
897 (pp. 2613–2669). Chipping Norton: Surrey Beatty & Sons.
898 Verdade, V. K. & Rodrigues, M. T. (2007). Taxonomic review of Allobates (Anura,
899 Aromobatidae) from the Atlantic Forest, Brazil. Journal of Herpetology, 41, 566–580.
900 Vieites, D. R., Wollenberg, K. C., Adreone, F., Köhler, J., Glaw, F. & Vences, M. (2009).
901 Vast underestimation of Madagascar’s biodiversity evidenced by an integrative amphibian
902 inventory. Proceedings of the National Academy of Sciences, 106, 8267–8272.
903 Wallace, A. R. (1852). On the monkeys of the Amazon. Proceedings of the Zoological
904 Society of London, 20, 107–110.
905 West-Eberhard, M. J. (1989). Phenotypic plasticity and the origins of diversity. Annual
906 Review of Ecology and Systematics, 20, 249–278.
907 Wiens, J. J. (2004). Speciation and ecology revisited: phylogenetic niche conservatism and
908 the origin of species. Evolution, 58, 193–197.
909 Wiens, J. J. (2007). Species delimitation: New approaches for discovering diversity.
910 Systematic Biology, 56, 875–878.
911 Wiens, J. J. (2008). Systematics and herpetology in the age of genomics. Bioscience, 58, 297–
912 307.
913 Wiley, E. O. (1978). The evolutionary species concept reconsidered. Systematic Zoology, 27,
914 17–26.
915 Wright, S. (1951). The genetical structure of populations. Annals of Human Genetics, 15,
916 323–354.
917 Wycherley, J., Doran, S. & Beebee, T. J. C. (2002). Male advertisement call characters as
918 phylogeographical indicators in European water frogs. Biological Journal of the Linnean
919 Society, 77, 355–365.
84
920 Zeisset, I. & Beebee, T. J. C. (2008). Amphibian phylogeography: a model for understanding
921 historical aspects of species distributions. Heredity, 101, 109–119.
922 Zhang, D. X. & Hewitt, G. M. (2003). Nuclear DNA analyses in genetic studies of
923 populations: Practice, problems and prospects. Molecular Ecology, 12, 563–584.
924 Zink, R. M. & Barrowclough, G. F. (2008). Mitochondrial DNA under siege in avian
925 phylogeography. Molecular Ecology, 17, 2107–2121.
926 Zink, R. M., Blackwell-Rago, R.C. & Ronquist, F. (2000). The shifting roles of vicariance in
927 biogeography. Proceedings of the Royal Society B, 267, 497–503.
85
Tables
Table 1 Study sites in Brazilian Amazonia with respective geographic coordinates and sample
sizes. Type localities are indicated by asterisks. Sample sizes are given as the number of
Allobates sampled for each class of character.
Locality
Allobates paleovarzensis
1. Careiro*
2. Janauacá
3. Manaquiri
4. Hiléia
5. Novo Airão
6. Anamã
7. Anori
8. Codajás
9. Unini
10. Barcelos
Allobates nidicola
1. Autazes Road Km 12*
2. BR-319, km 260
3. PPBio Manaquiri
4. BR-319 Tupana
5. Vila Gomes
Allobates masniger
6. Borba
7. Novo Aripuanã
8. Road to Apuí
9. Jacareacanga
10. PARNA da Amazônia*
Total
Coordinates
Morphology
Acoustics mtDNA
3°22'26.30"S 59°52'6.40"W
3°28'45.00"S 60°12'11.00"W
3°25'58.00"S 60°36'43.80"W
3°11'52.80"S 60°26'33.00"W
3° 6'5.50"S 60°47'16.10"W
3°30'52.10"S 61°27'14.00"W
3°39'39.20"S 61°39'46.10"W
3°46'17.50"S 62°1'13.90"W
1°43'9.50"S 61°54'22.70"W
0°54'57.70"S 62°58'14.30"W
15
12
15
15
14
13
12
12
09
13
15
12
15
15
14
13
12
12
02
13
17
14
15
15
17
13
09
12
18
12
3°28'2.70"S 59°49'9.16"W
4°37'0.43"S 61°14'27.70"W
3°40'31.52"S 60°19'44.85"W
4° 6'2.84"S 60°39'25.40"W
4°19'33.70"S 59°41'0.33"W
10
10
10
10
10
10
10
10
10
10
11
12
14
13
14
4°26'22.00"S 59°37'10.10"W
5° 7'11.10"S 60°15'2.87"W
5°33'12.10"S 60° 7'31.30"W
6°14'55.50"S 57°52'28.50"W
4°32'51.70"S 56°18'13.60"W
10
10
10
10
10
230
10
10
10
10
10
223
15
15
12
14
14
276
86
Table 2 Analyses of molecular variance (AMOVA) based on fragments of the mitochondrial
16S rDNA. The relative distribution of genetic variability is presented according to
hierarquical levels of genetic structure in each of the Allobates study systems.
Source of variation
Between riversides
Among populations within riversides
Within populations
paleovarzensis system nidicola-masniger system
Percentage of variation
20.91
59.74
60.37
38.23
18.71
2.02
87
Table 3 Pairwise FST fixation indexes (lower left matrix) and average Kimura-2-parameter genetic distances (upper right matrix). Values were
calculated between the sampling localities of each of the Allobates study systems in Brazilian Amazonia. Significant FST values are indicated with
asterisks.
Locality
1
2
3
4
5
6
7
8
9
10
1
0.529*
0.630*
0.801*
0.750*
0.711*
0.647*
0.674*
0.901*
0.858*
2
0.005
0.671*
0.791*
0.746*
0.707*
0.636*
0.670*
0.892*
0.844*
3
0.003
0.005
0.789*
0.669*
0.634*
0.528*
0.528*
0.918*
0.918*
Localitiy
1
2
3
4
5
6
7
8
9
10
1
0.399*
-0.018
0.604*
0.917*
0.991*
0.992*
0.986*
0.989*
0.983*
2
0.002
0.354*
0.315*
0.821*
0.975*
0.977*
0.967*
0.972*
0.970*
3
0.000
0.002
0.582*
0.907*
0.990*
0.990*
0.985*
0.988*
0.983*
paleovarzensis system
4
5
6
0.005
0.006
0.006
0.007
0.007
0.007
0.002
0.003
0.003
0.001
0.001
0.097
0.001
0.007
0.062
0.133*
0.131*
0.076*
0.015
0.057
-0.072
0.944*
0.918*
0.904*
0.910*
0.876*
0.851*
nidicola-masniger system
4
5
6
0.004
0.004
0.039
0.004
0.006
0.041
0.004
0.005
0.039
0.008
0.043
0.820*
0.043
0.969*
0.984*
0.972*
0.985*
0.964*
0.960*
0.977*
0.976*
0.967*
0.981*
0.978*
0.964*
0.977*
0.974*
7
0.007
0.009
0.004
0.002
0.003
0.003
0.058*
0.867*
0.794*
8
0.006
0.007
0.003
0.001
0.002
0.002
0.003
0.885*
0.823*
9
0.016
0.017
0.012
0.014
0.015
0.014
0.016
0.014
0.131*
10
0.014
0.016
0.011
0.012
0.013
0.012
0.014
0.012
0.002
-
7
0.037
0.039
0.038
0.041
0.041
0.010
0.974*
0.978*
0.975*
8
0.033
0.035
0.033
0.035
0.033
0.028
0.026
0.965*
0.960*
9
0.042
0.042
0.042
0.043
0.041
0.034
0.032
0.024
0.965*
10
0.050
0.052
0.050
0.052
0.049
0.040
0.038
0.030
0.040
-
88
Table 4 Descriptive statistics of genetic polymorphism parameters and results of neutrality tests performed on each of the genetic clusters of
Allobates. Bayesian Analysis of Genetic Structure was used to estimate the composition of the genetic clusters. n = sample size; h = number of
haplotypes; S = number of segregating sites; π = average (± one standard deviation) pairwise distance between samples pertaining to the same
cluster; D = Tajima’s D; Fs = Fu’s Fs; R2 = Ramos-Onsins & Roza’s R2. Significant values in neutrality tests are indicated with asterisks.
Cluster
Localities
paleovarzensis system
1
2+3+4+5+6+7+8
2
1+2+3
3
9+10
4
6+8
nidicola-masniger system
1
10
2
1+2+3+4
3
5
4
6
5
7
6
8
7
9
n
h
S
π ± 1 SD
D
D 95%
Fs
Fs 95%
R2
R2 95%
79
31
30
2
10
7
8
2
13
6
8
1
0.00146 ± 0.00026
0.00307 ± 0.00030
0.00169 ± 0.00048
0.00190 ± 0.00095
-1.972
0.209
-1.696
-
<0.001*
0.625
0.017*
-
-5.945
-1.074
-4.483
-
0.002*
0.281
0.001*
-
0.038
0.130
0.066
0.500
0.040*
0.590
0.034*
1.000
14
50
14
15
15
12
14
3
7
2
3
3
4
3
4
8
1
2
3
3
2
0.00172 ± 0.00066
0.00210 ± 0.00037
0.00092 ± 0.00016
0.00071 ± 0.00031
0.00075 ± 0.00046
0.00094 ± 0.00039
0.00053 ± 0.00030
-0.876
-1.006
1.212
-1.001
-1.685
-1.629
-1.480
0.239
0.173
0.906
0.186
0.009*
0.015*
0.028*
0.686
-1.569
1.139
-0.917
-0.830
-2.123
-1.475
0.687
0.211
0.824
0.095
0.107
0.007*
0.033*
0.202
0.081
0.247
0.139
0.180
0.144
0.175
0.707
0.253
0.754
0.193
0.462
0.046*
0.300
89
Table 5 Tests of demographic expansion performed according to genetic clusters of Allobates indicated by BAPS. Values were inferred from the
sum of squared deviation (SSD) between observed and expected mismatch distributions, and Harpending’s raggedness index (Hri). Tests were
not conducted on Cluster 4 of Allobates paleovarzensis due to the small number of samples that compose it. n = number of samples. The nonsignificant p values of both tests indicate that the null hypothesis of population expansion through time cannot be rejected.
Cluster
Localities
paleovarzensis system
1
2+3+4+5+6+7+8
2
1+2+3
3
9+10
4
6+8
nidicola-masniger system
1
10
2
1+2+3+4
3
5
4
6
5
7
6
8
7
9
n
Mismatch obs.
Mean
Mismatch obs.
variance
SSD
p (SSD)
Hri
p (Hri)
79
31
30
2
0.771
1.742
0.892
-
0.878
1.382
1.138
-
0.00009
0.00222
0.01234
-
0.976
0.630
0.202
-
0.073
0.051
0.129
-
0.843
0.498
0.241
-
14
50
14
15
15
12
14
0.923
1.127
0.495
0.381
0.400
0.667
0.549
1.272
1.396
0.253
0.277
0.550
0.718
0.361
0.03276
0.00657
0.02076
0.00506
0.01003
0.00110
0.01290
0.220
0.238
0.180
0.577
0.327
0.783
0.328
0.205
0.093
0.244
0.192
0.396
0.104
0.155
0.208
0.192
0.183
0.377
0.560
0.868
0.533
90
Table 6 Simple and partial Mantel tests evaluating correlations between geographic,
phenotypic and genetic distances of Allobates between the studied localities in Brazilian
Amazonia. Simple Mantel tests are presented as “Matrix 1” X “Matrix 2” and partial Mantel
tests are presented as “Matrix 1” X “Matrix 2”.”Covariate matrix”. GenD = genetic distance;
GeoD = geographic distance; MorD = morphological distance; AcoD = acoustic distance;
AcoDwa = Acoustic distance without body size adjustment; River = riverside (binary
variable). Significant correlations are indicated with asterisks.
Model
GenD x GeoD
MorD x GeoD
AcoD x GeoD
AcoDwa x GeoD
GenD x River.GeoD
MorD x River.GeoD
AcoD x River.GeoD
AcoDwa x River.GeoD
paleovarzensis system
r
p
0.761
0.006*
-0.094
0.402
-0.029
0.653
0.327
0.117
0.063
0.379
0.175
0.152
-0.089
0.457
0.014
0.448
nidicola-masniger system
R
P
0.572
<0.001*
0.530
0.094
-0.041
0.456
-0.150
0.311
0.778
0.002*
-0.025
0.425
0.082
0.212
0.388
0.014*
91
Figures
Fig. 1 Geographic location of sampling sites (A) and haplotype network for Allobates
paleovarzensis (B). The network was built from 142 16S rDNA sequences. The size and color
of each ellipse indicate the frequency and geographic origin of the individuals bearing that
haplotype. White dots and transverse bars represent not sampled (missing) intermediate
haplotypes. Sample localities are numbered according to Table 1.
Fig. 2 Geographic location of sampling sites (A) and haplotype network for Allobates
nidicola and A. masniger (B). The network was built from 134 16S rDNA sequences. The size
and color of each ellipse indicate the frequency and geographic origin of the individuals
bearing that haplotype. White dots and transverse bars represent not sampled (missing)
intermediate haplotypes. Sample localities are numbered according to Table 1.
Fig. 3 Barplot from the Bayesian Analysis of Genetic Differentiation. (A) 142 individual 16S
rDNA sequences of Allobates paleovarzensis from ten sampling localities. (B) 134 individual
16S rDNA sequences of Allobates nidicola and A. masniger from other ten sampling
localities. Distinct colors represent each estimated genetic cluster. Individuals are displayed
according to sampling localities, which are numbered as on Table 1.
Fig. 4 Timetree of the (A) paleovarzensis and (B) nidicola-masniger systems based on
fragments of the 16S rDNA gene. Only unique haplotypes were included. Age estimates and
confidence intervals (horizontal bars) of the divergences were obtained via BEAST (Bayesian
Evolutionary Analysis Sampling Trees). Posterior probabilities of the older nodes are
presented. Terminals were colored according to their localities of origin, which are depicted in
Fig. 1 and 2. Black and white terminals are outgroups. The black and grey vertical bars
indicate the origin of each haplotype according to the riverside.
92
Fig. 1
93
Fig. 2
94
Fig. 3
95
Fig. 4
96
Electronic Supplementary Material
Online Resource 1 Advertisement call measurements of individuals of Allobates
paleovarzensis in each study locality in Brazilian Amazonia. Values are presented as mean
(above) and standard deviation (below). Call traits are: Note rate (NR, in notes/s); Note
duration (ND, in s); Internote interval, as the silent interval between two consecutive notes of
a call (InI, in s); Call rate (CR, in calls/s); Call duration (CD, in s); Intercall interval, as the
silent interval between two consecutive calls (IcI, in s); Maximum (peak) frequency, as the
frequency of higher intensity calculated for an entire note by a power spectrum function of
Raven Pro 1.3 (MF, in Hz); Lowest frequency (LF, in Hz); Highest frequency (HF, in Hz);
Note modulation, as the difference between the highest and the lowest frequencies of the call
(NM, in Hz). Sampling locality codes, sample sizes and respective geographic coordinates are
given on Table 1.
Online Resource 2 Advertisement call measurements of individuals of Allobates nidicola and
A. masniger in each study locality in Brazilian Amazonia. Values are presented as mean
(above) and standard deviation (below). Call traits are: Note duration (ND, in s); Interval
between notes (IN, in s); Lowest frequency (LF, in Hz); Highest frequency (HF, in Hz); Note
modulation, calculated as the difference between the maximum and minimum frequencies
(NM, in Hz); and Peak frequency (PF, in Hz). Sampling locality codes, sample sizes and
respective geographic coordinates are given on Table 1.
Online Resource 3 Morphometric measurements (in mm) of individuals of Allobates
paleovarzensis in each study locality in Brazilian Amazonia. Values are presented as mean
(above) and standard deviation (below). Morphometric traits are: Snout-vent length (SVL);
head length, from angle of jaws to tip of snout (HL); Head width at level of angle of jaws
(HW); Snout length, from anterior corner of eye to tip of snout (SL); Eye to nostril distance,
from anterior corner of eye to centre of nostril (EN); Inter-nostril distance (IN); Eye length,
from anterior to posterior corner (EL); Inter-orbital distance (IO); Diameter of tympanum
(TYM); Forearm length, from proximal edge of palmar tubercle to outer edge of flexed elbow
(FAL); Upper arm length, from trunk insertion to outer edge of flexed elbow (UAL); Hand
length, from proximal edge of palmar tubercle to tip of fingers I, II, III and IV (HAND1,
HAND2, HAND3 and HAND4); Width of disc on finger III (WFD); Tibia length, from outer
edge of flexed knee to heel (TL); Foot length, from proximal edge of outer metatarsal tubercle
to tip of toe IV (FL); Femur length (LL); Diameter of palmar tubercle (DPT); Width of tenar
97
tubercle (WTT); Width of disc on toe IV (WTD); Width of finger III (WPF). Sampling
locality codes, sample sizes and respective geographic coordinates are given on Table 1.
Online Resource 4 Morphometric measurements (in mm) of individuals of Allobates nidicola
and A. masniger in each study locality in Brazilian Amazonia. Values are presented as mean
(above) and standard deviation (below). Morphometric traits are described in Online Resource
S3. Sampling locality codes, sample sizes and respective geographic coordinates are given on
Table 1.
Online Resource 5 Distribution of 16S rDNA haplotypes of Allobates among 20 sampled
localities in Brazilian Amazonia. Collection numbers of vouchers (INPA-H) and GenBank
accession numbers are provided. Sampled localities are numbered according to Table 1.
Online Resource 6 Mismatch distributions obtained from pairwise nucleotide site differences
among 16S rDNA sequences composing the clusters of the (A) paleovarzensis and the (B)
nidicola-masniger systems. Clusters were determined via Bayesian Analysis of Genetic
Differentiation. Pairwise differences were not calculated for Cluster 4 of Allobates
paleovarzensis due to the small number of samples that compose it.
Online Resource 7 Classification matrix based on the discriminant function analysis, using
phenotypic data from males of Allobates paleovazensis pertaining to four genetic clusters.
Clusters are coded from C1 to C4, according to Table 4. Body-shape traits were used to
discriminate individuals based on morphology. Acoustic measurements were adjusted for
temperature and for both temperature and body size through linear regressions. The number
and percentage of individuals correctly assigned to each genetic cluster are indicated.
Online Resource 8 Classification matrix based on the discriminant function analysis, using
phenotypic data from males of Allobates nidicola and A. masniger pertaining to seven genetic
clusters. Clusters are coded from C1 to C7, according to Table 4. Body-shape traits were used
to discriminate individuals based on morphology. Acoustic measurements were adjusted for
temperature and for both temperature and body size through linear regressions. The number
and percentage of individuals correctly assigned to each genetic cluster are indicated.
98
Online Resource 1
Locality
1
2
3
4
5
6
7
8
9
10
NR
ND
6.24
±0.99
5.47
±0.85
5.45
±0.90
4.60
±0.83
5.21
±0.87
6.57
±1.20
6.97
±1.30
7.15
±1.74
6.53
±0.74
6.48
±1.50
InI
0.04
±0.01
0.04
±0.00
0.04
±0.00
0.05
±0.01
0.04
±0.01
0.04
±0.01
0.04
±0.00
0.04
±0.00
0.04
±0.00
0.04
±0.01
CR
0.12
±0.04
0.15
±0.03
0.15
±0.04
0.19
±0.04
0.16
±0.04
0.12
±0.02
0.10
±0.02
0.11
±0.03
0.16
±0.06
0.12
±0.03
CD
0.28
±0.08
0.33
±0.07
0.30
±0.05
0.30
±0.05
0.27
±0.06
0.27
±0.06
0.21
±0.04
0.25
±0.06
0.27
±0.12
0.21
±0.08
IcI
1.85
±0.55
1.66
±0.50
1.75
±0.72
2.14
±0.61
2.34
±0.68
2.12
±0.76
3.01
±0.84
2.55
±0.61
2.37
±0.34
3.55
±1.94
MF
2.11
±0.52
1.73
±0.44
1.92
±0.32
1.53
±0.41
1.86
±0.54
2.08
±0.88
2.10
±0.41
2.23
±0.85
4.48
±1.46
2.50
±1.08
4458.43
±142.65
4410.71
±173.41
4472.01
±208.27
4462.54
±209.18
4626.97
±162.68
4611.94
±195.25
4603.46
±130.76
4580.65
±130.81
4132.84
±123.98
4666.92
±145.29
LF
HF
NM
4069.69
4768.18
698.49
±123.69
±164.53
±72.65
4027.73
4709.60
681.87
±140.63
±161.41
±106.08
4109.14
4795.75
686.61
±159.39
±155.80
±103.87
4054.99
4734.83
679.84
±164.57
±192.25
±96.31
4203.95
5005.74
801.80
±106.65
±154.80
±86.85
4234.59
4890.46
655.88
±147.38
±184.53
±91.30
4198.16
4879.78
681.63
±104.90
±147.29
±103.93
4237.81
4900.20
662.39
±122.23
±123.36
±93.38
3769.95
4341.19
571.24
±176.51
±120.80
±55.71
4274.71
4898.46
623.76
±84.27
±111.44
±116.87
99
Online Resource 2
Locality
1
2
3
4
5
6
7
8
9
10
ND
0.04
±0.01
0.05
±0.00
0.05
±0.00
0.05
±0.00
0.05
±0.00
0.05
±0.00
0.05
±0.00
0.05
±0.01
0.04
±0.00
0.06
±0.01
IN
0.27
±0.04
0.30
±0.03
0.24
±0.03
0.29
±0.06
0.31
±0.07
0.35
±0.06
0.42
±0.09
0.41
±0.04
0.28
±0.04
0.28
±0.07
LF
4069.05
±129.50
4103.23
±184.94
4051.22
±181.26
4205.78
±255.10
3686.44
±94.83
4387.85
±324.03
4215.64
±137.96
4440.11
±205.10
4198.56
±279.52
4374.48
±94.77
HF
4421.53
±114.81
4487.66
±127.16
4429.06
±148.95
4541.89
±235.64
4007.51
±139.53
4782.38
±282.56
4628.13
±111.98
4829.18
±187.38
4627.34
±237.70
4752.86
±99.53
NM
354.08
±31.78
365.36
±63.84
377.81
±56.02
336.90
±34.26
342.31
±36.06
392.55
±51.10
396.79
±55.15
385.78
±44.10
428.79
±60.98
377.81
±44.03
PF
4231.73
±145.24
4302.14
±153.46
4231.38
±164.48
4351.32
±240.51
3863.47
±96.02
4581.54
±289.99
4418.89
±112.00
4627.57
±187.86
4416.06
±243.14
4566.55
±92.95
100
Online Resource 3
Locality
1
2
3
4
5
6
7
8
9
10
SVL
20.25
±0.99
20.25
±1.15
19.90
±0.70
20.53
±0.78
19.51
±0.87
19.83
±1.02
19.81
±0.61
20.06
±0.76
21.73
±0.61
19.06
±0.51
HL
6.29
±0.28
6.28
±0.24
6.05
±0.25
5.80
±0.44
5.96
±0.27
6.02
±0.31
6.19
±0.19
5.92
±0.35
6.46
±0.15
5.92
±0.27
HW
6.70
±0.28
6.84
±0.26
6.58
±0.33
6.81
±0.32
6.49
±0.30
6.79
±0.33
6.83
±0.19
6.66
±0.24
7.02
±0.21
6.39
±0.18
SL
2.86
±0.16
2.81
±0.20
2.74
±0.21
2.63
±0.21
2.74
±0.18
2.80
±0.14
3.01
±0.22
2.74
±0.22
2.93
±0.22
2.77
±0.23
EN
2.03
±0.07
1.99
±0.12
1.92
±0.07
1.93
±0.07
1.91
±0.09
1.95
±0.11
1.99
±0.11
1.96
±0.15
2.00
±0.07
1.89
±0.10
IN
2.75
±0.12
2.81
±0.18
2.73
±0.10
2.78
±0.09
2.70
±0.10
2.73
±0.08
2.72
±0.09
2.75
±0.10
2.90
±0.09
2.61
±0.09
EL
2.56
±0.14
2.60
±0.14
2.63
±0.14
2.70
±0.15
2.55
±0.09
2.61
±0.22
2.58
±0.14
2.63
±0.14
2.51
±0.13
2.62
±0.15
IO
5.50
±0.29
5.70
±0.13
5.57
±0.21
5.66
±0.26
5.46
±0.17
5.54
±0.23
5.56
±0.16
5.78
±0.16
6.02
±0.13
5.35
±0.19
TYM
1.14
±0.12
1.03
±0.12
1.03
±0.11
1.01
±0.14
1.10
±0.11
1.05
±0.15
1.08
±0.15
0.95
±0.12
1.16
±0.09
1.05
±0.09
FAL
4.67
±0.29
4.63
±0.31
4.51
±0.18
4.69
±0.19
4.53
±0.18
4.40
±0.23
4.51
±0.28
4.42
±0.20
4.91
±0.33
4.36
±0.16
UAL
5.39
±0.24
5.32
±0.21
5.11
±0.22
4.99
±0.42
5.04
±0.20
5.15
±0.21
5.05
±0.32
5.18
±0.34
5.21
±0.43
4.72
±0.42
HAND1
3.00
±0.25
3.03
±0.31
2.68
±0.29
2.92
±0.24
2.78
±0.26
3.00
±0.26
2.98
±0.26
3.09
±0.19
3.19
±0.27
2.98
±0.18
HAND2
2.72
±0.21
2.77
±0.23
2.60
±0.25
2.73
±0.24
2.59
±0.15
2.70
±0.19
2.63
±0.28
2.73
±0.26
3.00
±0.17
2.65
±0.16
HAND3
3.86
±0.24
3.86
±0.27
3.65
±0.26
3.95
±0.31
3.83
±0.20
3.91
±0.23
3.94
±0.24
3.94
±0.27
4.23
±0.19
3.91
±0.19
HAND4
2.34
±0.13
2.50
±0.17
2.34
±0.09
2.50
±0.15
2.36
±0.16
2.49
±0.18
2.42
±0.15
2.38
±0.23
2.67
±0.50
2.44
±0.19
WFD
0.55
±0.05
0.57
±0.09
0.52
±0.07
0.59
±0.06
0.51
±0.05
0.59
±0.08
0.63
±0.08
0.64
±0.08
0.59
±0.08
0.60
±0.06
TL
9.26
±0.37
9.53
±0.33
9.19
±0.25
9.20
±0.50
8.99
±0.26
9.17
±0.51
9.06
±0.33
9.03
±0.43
10.11
±0.49
9.03
±0.21
FL
7.81
±0.38
7.83
±0.43
7.58
±0.41
8.07
±0.31
7.68
±0.34
7.47
±0.34
7.56
±0.26
7.83
±0.41
8.31
±0.55
7.58
±0.41
LL
8.48
±0.44
9.03
±0.46
8.40
±0.32
8.49
±0.56
8.55
±0.26
8.54
±0.40
8.48
±0.38
8.73
±0.47
9.38
±0.46
8.34
±0.49
DPT
0.71
±0.07
0.78
±0.08
0.77
±0.10
0.75
±0.12
0.78
±0.08
0.74
±0.09
0.75
±0.07
0.75
±0.08
0.81
±0.08
0.72
±0.09
WTT
0.63
±0.12
0.68
±0.08
0.64
±0.07
0.63
±0.13
0.61
±0.13
0.62
±0.07
0.65
±0.05
0.58
±0.08
0.61
±0.11
0.65
±0.11
WTD
0.65
±0.10
0.69
±0.10
0.65
±0.11
0.69
±0.11
0.66
±0.08
0.72
±0.11
0.73
±0.11
0.80
±0.11
0.64
±0.13
0.70
±0.07
WPF
0.54
±0.05
0.58
±0.06
0.55
±0.05
0.45
±0.05
0.52
±0.07
0.55
±0.05
0.59
±0.05
0.58
±0.06
0.50
±0.09
0.52
±0.04
101
Online Resource 4
Locality
1
2
3
4
5
6
7
8
9
10
SVL
19.74
±0.62
19.25
±0.63
19.76
±0.71
18.75
±0.46
20.63
±0.10
18.58
±0.69
19.27
±0.35
18.02
±0.65
19.32
±0.38
20.46
±0.70
HL
7.17
±0.40
6.67
±0.42
6.90
±0.41
6.63
±0.34
7.09
±0.08
6.50
±0.38
6.53
±0.51
6.26
±0.34
6.57
±0.39
7.02
±0.48
HW
7.04
±0.25
6.68
±0.19
6.96
±0.23
6.54
±0.19
7.26
±0.19
6.49
±0.30
6.69
±0.35
6.23
±0.24
6.71
±0.34
7.16
±0.45
SL
2.87
±0.27
2.64
±0.12
2.72
±0.12
2.86
±0.14
2.89
±0.24
2.63
±0.23
2.71
±0.17
2.57
±0.24
2.66
±0.17
2.25
±0.21
EN
2.07
±0.18
1.89
±0.15
2.03
±0.05
1.98
±0.09
2.08
±0.07
1.84
±0.15
1.90
±0.13
1.82
±0.14
1.91
±0.07
1.53
±0.13
IN
2.67
±0.07
2.49
±0.14
2.60
±0.12
2.58
±0.23
2.80
±0.31
2.54
±0.12
2.64
±0.16
2.44
±0.12
2.60
±0.16
2.72
±0.15
EL
2.74
±0.16
2.64
±0.17
2.60
±0.11
2.66
±0.12
2.74
±0.33
2.64
±0.18
2.63
±0.11
2.63
±0.08
2.73
±0.16
2.55
±0.14
IO
2.44
±0.28
2.22
±0.19
2.33
±0.14
2.25
±0.19
2.50
±0.20
2.12
±0.14
2.33
±0.16
2.22
±0.12
2.42
±0.19
2.34
±0.24
TYM
1.05
±0.14
1.05
±0.08
1.15
±0.13
1.00
±0.12
1.22
±0.18
1.06
±0.15
1.01
±0.10
1.02
±0.10
1.03
±0.17
1.02
±0.14
FAL
4.18
±0.24
4.19
±0.17
4.07
±0.17
4.15
±0.17
4.36
±0.25
4.22
±0.27
4.21
±0.27
4.25
±0.25
4.45
±0.20
4.35
±0.18
UAL
4.91
±0.21
4.88
±0.20
4.95
±0.25
4.75
±0.30
5.13
±0.16
4.75
±0.27
4.86
±0.25
4.77
±0.33
5.04
±0.25
5.12
±0.20
HAND1
3.31
±0.15
3.36
±0.13
3.33
±0.17
3.32
±0.18
3.57
±0.10
3.33
±0.19
3.47
±0.19
3.33
±0.19
3.58
±0.19
3.69
±0.15
HAND2
3.13
±0.12
3.10
±0.12
3.17
±0.15
3.10
±0.19
3.20
±0.40
2.91
±0.11
3.05
±0.12
2.95
±0.18
3.16
±0.11
3.39
±0.19
HAND3
4.42
±0.13
4.31
±0.14
4.49
±0.14
4.28
±0.22
4.47
±0.23
4.02
±0.18
4.18
±0.23
4.00
±0.17
4.21
±0.17
4.58
±0.26
HAND4
2.98
±0.18
2.88
±0.17
2.98
±0.20
2.88
±0.15
2.95
±0.37
2.71
±0.21
2.79
±0.16
2.62
±0.16
2.82
±0.15
3.06
±0.17
WFD
0.50
±0.05
0.46
±0.08
0.50
±0.07
0.47
±0.09
0.54
±0.06
0.47
±0.06
0.49
±0.05
0.42
±0.04
0.45
±0.05
0.60
±0.05
TL
8.88
±0.38
8.69
±0.31
8.78
±0.20
8.63
±0.33
8.90
±0.08
8.36
±0.24
8.79
±0.39
8.42
±0.51
8.83
±0.33
9.05
±0.33
FL
7.65
±0.34
7.68
±0.27
7.92
±0.30
7.71
±0.34
7.88
±0.07
7.40
±0.31
7.85
±0.29
7.53
±0.34
7.89
±0.36
7.82
±0.42
LL
8.77
±0.35
8.90
±0.13
8.87
±0.31
8.71
±0.34
9.16
±0.05
8.66
±0.35
9.00
±0.41
8.60
±0.41
9.06
±0.32
8.96
±0.48
DPT
0.63
±0.06
0.64
±0.05
0.62
±0.08
0.65
±0.06
0.66
±0.10
0.60
±0.04
0.68
±0.07
0.59
±0.06
0.68
±0.09
0.64
±0.04
WTT
0.48
±0.06
0.43
±0.06
0.43
±0.06
0.43
±0.04
0.41
±0.08
0.37
±0.04
0.36
±0.08
0.36
±0.05
0.39
±0.06
0.45
±0.07
WTD
0.65
±0.05
0.69
±0.09
0.68
±0.06
0.68
±0.08
0.72
±0.19
0.62
±0.06
0.64
±0.06
0.55
±0.06
0.59
±0.07
0.79
±0.05
WPF
0.40
±0.04
0.36
±0.06
0.36
±0.05
0.38
±0.05
0.40
±0.24
0.35
±0.03
0.40
±0.04
0.31
±0.02
0.33
±0.04
0.46
±0.04
102
Online Resource 5
Locality
paleovarzensis system
Haplotype
1
2
H01
7
1
H02
1
H03
6
H04
2
H05
1
H06
7
H07
4
H08
1
H09
1
H10
H11
H12
H13
H14
H15
H16
H17
H18
H19
H20
H21
H22
H23
H24
H25
H26
H27
H28
H29
nidicola-masniger system
H01
11 1
H02
5
H03
6
H04
H05
H06
H07
H08
H09
H10
H11
H12
H13
H14
H15
H16
H17
H18
H19
H20
H21
H22
H23
H24
H25
H26
-
3
13
1
1
-
4
14
1
-
5
14
3
-
6
12
1
-
7
3
1
2
2
1
-
8
10
1
1
-
9
12
2
1
1
2
-
10
7
3
1
1
Collection #
INPA-H 29176
APL 12625
INPA-H 29071
APL 12631
APL 12632
APL 954
INPA-H 29185
INPA-H 29145
INPA-H 29184
INPA-H 29117
INPA-H 29114
INPA-H 29124
APL 12681
INPA-H 29135
INPA-H 29171
INPA-H 29193
INPA-H 29188
INPA-H 29194
INPA-H 29195
INPA-H 29146
INPA-H 29155
INPA-H 29079
INPA-H 29091
INPA-H 29098
INPA-H 29106
INPA-H 29097
INPA-H 29086
INPA-H 29082
INPA-H 29083
GenBank #
JQ966835
JQ966836
JQ966837
JQ966838
JQ966839
JQ966840
JQ966841
JQ966842
JQ966843
JQ966844
JQ966845
JQ966846
JQ966847
JQ966848
JQ966849
JQ966850
JQ966851
JQ966852
JQ966853
JQ966854
JQ966855
JQ966856
JQ966857
JQ966858
JQ966859
JQ966860
JQ966861
JQ966862
JQ966863
13
1
-
7
1
2
2
1
-
9
5
-
2
12
1
-
1
13
1
-
9
1
1
1
-
2
10
1
1
-
5
8
1
INPA-H 28140
INPA-H 28173
INPA-H 28145
INPA-H 28130
INPA-H 28134
INPA-H 28142
INPA-H 28177
APL 14320
APL 14322
INPA-H 28092
INPA-H 28070
APL 12763
APL 14336
APL 14399
APL 14397
INPA-H 28079
APL 14404
APL 14406
INPA-H 28102
APL 14252
INPA-H 28093
APL 14251
APL 14250
APL 12957
APL 12960
APL 12958
JQ966864
JQ966865
JQ966866
JQ966867
JQ966868
JQ966869
JQ966870
JQ966871
JQ966872
JQ966873
JQ966874
JQ966875
JQ966876
JQ966877
JQ966878
JQ966879
JQ966880
JQ966881
JQ966882
JQ966883
JQ966884
JQ966885
JQ966886
JQ966887
JQ966888
JQ966889
103
Online Resource 6
104
Online Resource 7
C1
C1
C2
C3
C4
Total
27
5
4
1
37
C1
C2
C3
C4
Total
22
6
3
1
32
C1
C2
C3
C4
Total
21
8
5
0
34
Genetic cluster
C2
C3
C4
Temperature-adjusted call traits
11
5
10
9
0
3
2
5
0
0
0
0
22
10
13
Temperature- and body-size-adjusted call traits
15
4
12
7
2
2
2
6
0
0
0
0
24
12
14
Morphology
16
15
1
6
3
0
5
7
1
0
1
0
27
26
2
% correct
51
53
45
0
50
42
41
55
0
43
40
35
39
0
38
105
Online Resource 8
C1
C2
C1
C2
C3
C4
C5
C6
C7
Total
5
2
1
2
0
0
0
10
1
25
0
0
2
0
3
31
C1
C2
C3
C4
C5
C6
C7
Total
3
2
5
1
0
0
0
11
1
23
0
1
2
0
2
29
C1
C2
C3
C4
C5
C6
C7
Total
8
1
0
0
0
0
0
9
0
19
2
1
2
0
1
25
Genetic cluster
C3
C4
C5
C6
Temperature-adjusted call traits
0
1
0
0
2
1
2
0
9
0
0
0
0
1
3
3
0
2
3
2
0
2
1
5
0
0
0
0
11
7
9
10
Temperature- and body-size-adjusted call traits
2
1
0
0
4
3
0
0
4
0
0
1
1
1
2
2
0
2
2
3
0
2
2
4
0
0
0
0
11
9
6
10
Morphology
0
0
0
0
6
3
2
4
6
1
0
0
2
0
4
2
1
2
0
2
0
2
0
3
1
2
2
1
16
10
8
12
C7
% correct
1
6
0
1
1
0
7
16
63
66
90
10
30
63
70
59
1
6
0
2
1
0
8
18
38
61
40
10
20
50
80
48
0
3
1
1
3
3
3
14
100
50
60
0
0
38
30
41
106
SÍNTESE
Em conjunto, os resultados dos três capítulos aqui apresentados ressaltam a
complementaridade de respostas proporcionada pela investigação de distintas classes de
caracteres e de diferentes níveis da hierarquia biológica no estudo de questões em
biogeografia histórica. Os dois primeiros capítulos transitaram do nível intraindividual ao
interespecífico no estudo de como caracteres comportamentais importantes em escolha e
reconhecimento sexual variam no espaço. Em adição, foi possível determinar como fatores
individuais e ambientais afetam esses sinais. No terceiro capítulo, foi possível realizar uma
abordagem comparativa para avaliar a resposta de diferentes classes de marcadores à história
biogeográfica das espécies.
Por meio da amostragem realizada, verificou-se que Allobates paleovarzensis ocorre
em ambas as margens do Rio Amazonas. Sua área de distribuição é maior do que aquela
anteriormente conhecida, estendendo-se até o município de Codajás, a oeste, e até o
município de Barcelos, ao norte. Nenhum dos caracteres acústicos analisados foi indicado
como potencial sinal para reconhecimento social ou seleção sexual na espécie estudada, e
futuras investigações envolvendo experimentos de reconhecimento sexual intraespecífico
deverão elucidar as potenciais implicações do padrão observado. Todas as propriedades
estáticas (no caso, espectrais) do canto de anúncio estiveram sob forte regulação morfológica.
Portanto, esses caracteres podem ser indicadores honestos dos seus fatores controladores.
Essas propriedades também deverão estar sujeitas à seleção estabilizadora por parte das
fêmeas da espécie. Todas as propriedades acústicas temporais relacionadas às notas estiveram
correlacionadas com a temperatura ambiental. Uma vez que caracteres relacionados ao canto
(conjunto de notas) não variaram de acordo com os fatores individuais e ambientais avaliados,
acredita-se que estes sejam sujeitos à regulação social. Tanto caracteres espectrais quanto
temporais diferiram entre lados opostos do Rio Amazonas. Todavia, lado do rio e distância
geográfica não afetaram significativamente a variabilidade acústica quando todos os
parâmetros acústicos foram sumarizados em componentes principais. Esses resultados
indicam que forças evolutivas estabilizadoras locais devem ser importantes no processo de
diferenciação do sinal sexual de A. paleovarzensis.
Com relação ao sistema de estudo composto por populações de Allobates nidicola e A.
masniger, propriedades espectrais do canto de anúncio foram mais distintivas que
propriedades temporais entre populações e entre espécies, com maior distinção ao nível
107
interpopulacional quando comparado ao nível interespecífico. O entendimento do rio como
barreira levando à especiação no sistema de estudo seria beneficiado por abordagens que
avaliem a variabilidade dos caracteres acústicos ao nível intraindividual, bem como o papel de
propriedades espectrais no reconhecimento e escolha sexual por parte de indivíduos de ambas
as margens. Houve acentuada plasticidade em caracteres temporais do canto de anúncio, os
quais tiveram sua variabilidade fortemente induzida pela temperatura ambiental. Caracteres
espectrais também mostraram correlação positiva, embora moderada, com essa variável
ambiental. Os efeitos do tamanho corporal foram restritos às propriedades espectrais do canto
de anúncio. As propriedades modulação da nota e intervalo entre notas foram pouco afetadas
pela temperatura ambiental e tamanho corporal, sugerindo que outros fatores subjacentes
devem estar relacionados à sua regulação. O efeito do Rio Madeira como barreira foi
significativo entre todas as variáveis acústicas analisadas, mesmo após a inclusão de tamanho
corporal como covariável. Foi detectada uma interação entre geografia (localidade) e tamanho
corporal na diferenciação de caracteres espectrais do sinal sexual. A ausência de efeitos de
isolamento por distância indica que pressões seletivas estabilizadoras locais devem ser mais
importantes que a deriva genética na evolução da diferenciação do canto de anúncio no
sistema nidicola/masniger.
Observou-se significativa estruturação genética populacional em ambos os sistemas de
estudo. Tal padrão pode estar relacionado ao pequeno tamanho corporal e à territorialidade
das espécies. A estrutura genética dos dois sistemas de estudo exibiu assinaturas congruentes
com a transposição de grandes rios amazônicos. A efetividade do baixo Rio Madeira como
uma barreira ao fluxo gênico foi maior do que a do médio Amazonas. Entretanto,
características idiossincráticas dos sistemas de amostragem e das espécies estudadas podem
ter contribuído para o resultado observado. Por exemplo, as espécies de desenvolvimento
larval direto (A. nidicola e A. masniger) apresentaram estruturação genética populacional mais
expressiva, bem como um efeito mais pronunciado do rio como barreira em relação à A.
paleovarzensis, cujos girinos completam seu desenvolvimento em corpos d’água. Allobates
nidicola e A. masniger podem ser consideradas espécies distintas sob os conceitos evolutivo e
filogenético de espécie. Populações de Allobates paleovarzensis pertencentes à interflúvios
distintos parecem não ter sido sujeitas ao nível e à duração do isolamento genético necessários
para que a diferenciação populacional atinja os critérios de especiação sob os conceitos de
espécie baseados em monofilia e fenótipo. Além dos efeitos observados de rios como
barreiras biogeográficas, uma grande porção da diferenciação genética foi correlacionada com
108
distância geográfica linear, evidenciando a importante atuação da deriva genética na
diversificação do marcador mitocondrial estudado. Os caracteres acústicos e morfológicos não
responderam às distâncias lineares entre populações. Em adição, análises de função
discriminante falharam em classificar corretamente os indivíduos analisados em seus
agrupamentos genéticos de acordo com seus fenótipos. Esse resultado é surpreendente
levando-se em consideração a alta estruturação genética observada, e sugere que limitações
evolutivas, as quais não agiram sobre marcadores moleculares neutros, podem estar
impedindo a diferenciação fenotípica.
109
REFERÊNCIAS BIBLIOGRÁFICAS
Adams, D.C.; Berns, C.M.; Kozak, K H.; Wiens, J. 2009. Are rates of species diversification
correlated with rates of morphological evolution? Proceedings of the Royal Society B, 276:
2729-2738.
Amézquita, A.; Hödl, W.; Lima, A.P.; Castellanos, L.; Erdtmann, L.; Araújo, M.C. 2006.
Masking interference and the evolution of the acoustic communication system in the
Amazonian dendrobatid frog Allobates femoralis. Evolution, 60: 1874-1887.
Amézquita, A.; Lima, A.P.; Jehle, R.; Castellanos, L.; Ramos, O.; Crawford, A.J.; Gasser, H.;
Hödl, W. 2009. Calls, colours, shapes, and genes: A multi-trait approach to the study of
geographic variation in the Amazonian frog Allobates femoralis. Biological Journal of the
Linnean Society, 98: 826-838.
Anderson, M.J.; Legendre, P. 1999. An empirical comparison of permutation methods for tests of
partial regression coefficients in a linear model. Journal of Statistical Computation and
Simulation, 62: 271-303.
Angulo, A.; Icochea, J. 2010. Cryptic species complexes, widespread species and conservation:
lessons from Amazonian frogs of the Leptodactylus marmoratus group (Anura:
Leptodactylidae). Systematics and Biodiversity, 8: 357-370.
Angulo, A.; Reichle, S. 2008. Acoustic signals, species diagnosis, and species concepts: the case
of a new cryptic species of Leptodactylus (Amphibia, Anura, Leptodactylidae) from the
Chapare region, Bolivia. Zoological Journal of the Linnean Society, 152: 58-77.
Antonelli, A.; Quijada-Mascareñas, A.; Crawford, A.J.; Bates, J.M.; Velazco, P.M; Wüster, W.
2010. Molecular studies and phylogeography of Amazonian tetrapods and their relation to
geological and climatic models. In: Hoorn, C.; Wesselingh, F.P. (Eds). Amazonia, landscape
and species evolution. Blackwell Publishing, Oxford. p. 386-403.
Avise, J.C.; Arnold, J.; Ball, R.M.; Bermingham, E.; Lamb, T.; Neigel, J.E.; Reeb, C.A.;
Saunders, N.C. 1987. Intraspecific phylogeography: the mitochondrial DNA bridge between
population genetics and systematics. Annual Review of Ecology and Systematics, 18: 489-522.
110
Avise, J.C. 2000. Phylogeography: the history and formation of species. Harvard University
Press, Cambridge.
Avise, J.C. 2004. Molecular Markers, Natural History and Evolution. Sinauer Associates Inc.,
Sunderland.
Avise, J.C. 2009. Phylogeography: retrospect and prospect. Journal of Biogeography, 36: 3-15.
Ayres, J.M.; Clutton-Brock, T.H. 1992. River boundaries and species range size in Amazonian
primates. The American Naturalist, 140: 531-537.
Bee, M.A.; Cook, J.M.; Love, E.K.; O’Bryan, L.R.; Pettitt, B.A.; Schrode, K.; Vélez, A. 2010.
Assessing acoustic signal variability and the potential for sexual selection and social
recognition in boreal chorus frogs (Pseudacris maculata). Ethology, 116: 564-576.
Bee, M.A.; Perrill, S.A.; Owen, P.C. 2000. Male green frogs lower the pitch of acoustic signals in
defense of territories: a possible dishonest signal of size? Behavioral Ecology, 11: 169-177.
Behr, O.; von Helversen, O.; Heckel, G.; Nagy, M.; Voigt, C.C.; Mayer, F. 2007. Territorial songs
indicate male quality in the sac-winged
bat
Saccopteryx bilineata (Chiroptera,
Emballonuridae). Behavioral Ecology, 17: 810-817.
Blomberg, S.P.; Garland Jr., T.; Ives, A.R. 2003. Testing for phylogenetic signal in comparative
data: behavioral traits are more labile. Evolution, 57: 717-745.
Bonnet, E.; Van de Peer, Y. 2002. ZT: a software tool for simple and partial Mantel tests. Journal
of Statistical Software, 7: 1-12.
Boul, K.E.; Funk, W.C.; Darst, C.R.; Cannatella, D.C.; Ryan, M.J. 2007. Sexual selection drives
speciation in an Amazonian frog. Proceedings of the Royal Society B, 274: 399-406.
Buckley, D. 2009. Toward an organismal, integrative, and iterative phylogeography. Journal of
Biogeography, 36: 3-15.
Caldwell, J.P.; Lima, A.P. 2003. A new Amazonian species of Colostethus (Anura:
Dendrobatidae) with a nidicolous tadpole. Herpetologica, 59: 219-234.
Camargo, A.; de Sá, R.; Heyer, W.R. 2006. Phylogenetic analyses of mtDNA sequences reveal
three cryptic lineages in the widespread neotropical frog Leptodactylus fuscus (Schneider,
1799) (Anura, Leptodactylidae). Biological Journal of the Linnean Society, 87: 325-341.
111
Campbell, P.; Pasch, B.; Pino, J.L.; Crino, O.L.; Phillips, M.; Phelps, S.M. 2010. Geographic
variation in the songs of Neotropical singing mice: testing the relative importance of drift and
local adaptation. Evolution, 64: 1955-1972.
Capparella, A.P. 1988. Genetic variation in Neotropical birds: implications for the speciation
process. In: Ouellet, H. (Ed). Acta XIX Congressus Internationalis Ornithologici. University
of Ottawa Press, Ottaxa. p. 1658-1664.
Castellano, S.; Cuatto, B.; Rinella, R.; Rosso, A.; Giacoma, C. 2002. The advertisement call of the
European treefrogs (Hyla arborea): a multilevel study of variation. Ethology, 108: 75-89.
Castellano, S.; Giacoma, C. 1998. Stabilizing and directional female choice for male calls in the
European green toads. Animal Behaviour, 56: 275-287.
Castellano, S.; Giacoma, C.; Dujsebayeva, T. 2000. Morphometric and advertisement call
geographic variation in polyploid green toads. Biological Journal of the Linnean Society, 70:
341-360.
Chakraborty, M.; Mangiamele, L.A.; Burmeister, S.S. 2010. Neural activity patterns in response
to interspecific and intraspecific variation in mating calls in the Túngara Frog. PLoS One, 5:
e12898.
Charif, R.A.; Clark, C.W.; Fristrup, K.M. 2004. Raven 1.2 User’s Manual. Cornell Laboratory of
Ornithology, Ithaca, New York.
Charif, R.A.; Waack, A.M.; Strickman, L.M. 2008. Raven Pro 1.3 User’s Manual. Cornell
Laboratory of Ornithology, Ithaca, New York.
Cherry, L.M.; Case, S.M.; Wilson, A.C. 1978. Frog perspective on the morphological difference
between humans and chimpanzees. Science, 200: 209-211.
Clement, M.; Posada, D.; Crandall, K.A. 2000. TCS: a computer program to estimate gene
genealogies. Molecular Ecology, 9: 1657-1659.
Cohn-Haft, M. 2000. A case study in amazonian biogeography: Vocal and DNA sequence
variation in Hemitriccus flycatchers. Tese de Doutorado, Luisiana State University, Baton
Rouge.
112
Colwell, R.K. 2000. A barrier runs through it… or maybe just a river. Proceedings of the National
Academy of Sciences, 97: 13470-13472.
Corander, J.; Marttinen, P.; Sirén J.; Tang, J. 2008. Enhanced Bayesian modelling in BAPS
software for learning genetic structures of populations. BMC Bioinformatics, 9: 539.
Costa, J.B.S.; Bemerguy, R.L.; Hasui, Y.; Borges, M.S. 2001. Tectonics and paleogeography
along the Amazon River. Journal of South American Earth Sciences, 14: 335-347.
Coyne, J.A.; Orr, H.A. 2004. Speciation. Sinauer Associates Inc., Sunderland.
Crawford, A.J. 2003. Huge populations and old species of Costa Rican and Panamanian dirt frogs
inferred from mitochondrial and nuclear gene sequences. Molecular Ecology, 12: 2525-2540.
Crawford, A.J.; Lips, K.R.; Bermingham, E. 2010. Epidemic disease decimates amphibian
abundance, species diversity, and evolutionary history in the highlands of central Panama.
Proceedings of the National Academy of Sciences, 107: 13777-13782.
Creer, S.; Thorpe, R.S.; Malhotra, A.; Chou, W.H.; Stenson, A.G. 2004. The utility of AFLPs for
supporting mitochondrial DNA phylogeographical analyses in the Taiwanese bamboo viper,
Trimeresurus stejnegeri. Journal of Evolutionary Biology, 17: 100-107.
Crisci, J.V.; Katinas, L.; Posadas, P. 2003. Historical biogeography: An Introduction. Harvard
University Press, Cambridge.
Cronquist, A. 1978. Once again, what is a species? In: Knutson, L.V. (Ed). Biosystematics in
Agriculture. Allenheld Osmun, Montclair, New Jersey. p. 3-20.
Dall, S.R.X. 1997. Behaviour and speciation. Trends in Ecology and Evolution, 12: 209-210.
Dayrat, B. 2005. Towards integrative taxonomy. Biological Journal of the Linnean Society, 85:
407-415.
de Queiroz, K. 2007. Species concepts and species delineation. Systematic Biology, 56: 879-886.
Diniz-Filho, J.A.F.; Telles, M.P.C.; Bonatto, S.L.; Eizirik, E.; Freitas, T.R.O.; de Marco, P.;
Santos, F.R.; Solé-Cava, A.M.; Soares, T.N. 2008. Mapping the evolutionary twilight zone:
molecular markers, populations and geography. Journal of Biogeography, 35: 753-763.
113
Drummond, A.J.; Rambaut, A. 2007. BEAST: Bayesian evolutionary analysis by sampling trees.
BMC Evolutionary Biology, 7: 214.
Elmer, K.R.; Dávila, J.A.; Lougheed, S.C. 2007. Cryptic diversity and deep divergence in an
upper Amazonian leaflitter frog, Eleutherodactylus ockendeni. BMC Evolutionary Biology, 7:
247.
Emerson, K.J.; Merz, C.R.; Catchen, J.M.; Hohenlohe, P.A.; Cresko, W.A.; Bradshaw, W.E.;
Holzapfel, C.M. 2010. Resolving postglacial phylogeography using high-throughput
sequencing. Proceedings of the National Academy of Sciences, 107: 16196-16200.
Erdtmann, L.; Amézquita, A. 2009. Differential evolution of advertisement call traits in DartPoison Frogs (Anura: Dendrobatidae). Ethology, 115: 801-811.
Excoffier, L.; Lischer, H.E.L. 2010. Arlequin suite ver 3.5: A new series of programs to perform
population genetics analyses under Linux and Windows. Molecular Ecology Research, 10:
564-567.
Excoffier, L.; Smouse, P.; Quattro, J. 1992. Analysis of molecular variance inferred from metric
distances among DNA haplotypes: application to human mitochondrial DNA restriction data.
Genetics, 131: 479-491.
Fernandes, A.M.; Wink, M.; Aleixo, A. 2012. Phylogeography of the chestnut tailed antbird
(Myrmeciza hemimelaena) clarifies the role of rivers in Amazonian biogeography. Journal of
Biogeography, doi:10.1111/j.1365-2699.2012.02712.x
Figueiredo, J.; Hoorn, C.; van der Ven, P.; Soares, E. 2009. Late Miocene onset of the Amazon
River and the Amazon deep-sea fan: Evidence from the Foz do Amazonas Basin. Geology,
37: 619-622.
Fischer, J.; Kitchen, D.M.; Seyfarth, R.M.; Cheney, D.L. 2004. Baboon loud calls advertise male
quality: acoustic features and their relation to rank, age, and exhaustion. Behavioral Ecology
and Sociobiology, 56: 140-148.
Fouquet, A.; Gilles, A.; Vences, M.; Marty, C.; Blanc, M.; Gemmell, N.J. 2007. Underestimation
of species richness in Neotropical frogs revealed by mtDNA analyses. PLoS One, 32: e1109.
114
Fouquet, A.; Noonan, B.P.; Rodrigues, M.T.; Pech, N.; Gilles, A.; Gemmell, N.J. Multiple
Quaternary Refugia in the Eastern Guiana Shield Revealed by Comparative Phylogeography
of 12 Frog Species. Systematic Biology, in press.
Fouquet, A.; Recoder, R.; Teixeira–Jr, M.; Cassimiro, J.; Amaro, R.C.; Camacho, A.; Damasceno,
R.; Carnaval, A.C.; Moritz, C.; Rodrigues, M.T. 2012. Molecular phylogeny and
morphometric analyses reveal deep divergence between Amazonia and Atlantic Forest species
of Dendrophryniscus. Molecular Phylogenetics and Evolution, 62: 826-838.
Friedl, T.W.P.; Klump, G.M. 2002. The vocal behaviour of male European treefrogs (Hyla
arborea): implications for inter- and intrasexual selection. Behaviour, 139: 113-136.
Fu, Y.X. 1997. Statistical tests of neutrality of mutations against population growth, hitchhiking
and background selection. Genetics, 147: 915-925.
Funk, D.J.; Omland, K.E. 2003. Species-level paraphyly and polyphyly: frequency, causes, and
consequences, with insights from animal mitochondrial DNA. Annual Review of Ecology,
Evolution, and Systematics, 34: 397-423.
Funk, W.C.; Caldwell, J.P.; Peden, C.E.; Padial, J.M.; de la Riva, I.; Cannatella, D.C. 2007. Tests
of biogeographic hypotheses for diversification in the Amazonian forest frog, Physalaemus
petersi. Molecular Phylogenetics and Evolution, 44: 825-837.
Funk, W.C.; Caminer, M.; Ron, S.R. 2012. High levels of cryptic species diversity uncovered in
Amazonian frogs. Proceedings of the Royal Society B, 279: 1806-1814.
Gascon, C.; Lougheed, S.C.; Bogart, J.P. 1996. Genetic and morphological variation in
Vanzolinius discodactylus: a direct test of the riverine barrier hypothesis. Biotropica 28, 376387.
Gascon, C.; Lougheed, S.C.; Bogart, J. P. 1998. Patterns of genetic population differentiation in
four species of Amazonian frogs: A test of the Riverine Barrier Hypothesis. Biotropica, 30:
104-119.
Gascon, C.; Malcolm, J.R.; Patton, J.L.; Silva, M.N.F.; Bogarti, J.P.; Lougheed, S.P.; Peres, C.A.;
Neckel, S.; Boag, P.T. 2000. Riverine barriers and the geographic distribution of Amazonian
species. Proceedings of the National Academy of Sciences, 97: 13672-13677.
115
Gasser, H.; Amézquita, A.; Hödl, W. 2009. Who is calling? Intraspecific call variation in the
aromobatid frog Allobates femoralis. Ethology, 115: 596-607.
Gerhardt, H.C. 1991. Female mate choice in treefrogs: static and dynamic acoustic criteria.
Animal Behaviour, 42: 615-635.
Gerhardt, H.C.; Huber, F. 2002. Acoustic communication in insects and anurans: common
problems and diverse solutions. University of Chicago Press, Chicago.
Goicochea, N.; De La Riva, I.; Padial, J.M. 2009. Recovering phylogenetic signal from frog
mating calls. Zoologica Scripta, 39: 141-154.
Gomez, D.; Richardson, C.; Théry, M.; Lengagne, T.; Léna, J.; Plénet, S.; Joly, P. 2011.
Multimodal signals in male European treefrog (Hyla arborea) and the influence of population
isolation on signal expression. Biological Journal of the Linnean Society, 103: 633-647.
Grant, T.; Frost, D.R.; Caldwell, J.P.; Gagliardo, R.; Haddad, C.F.B.; Kok, P.J.R.; Means, B.D.;
Noonan, B.P.; Schargel, W.; Wheeler, W.C. 2006. Phylogenetic systematics of dart-poison
frogs and their relatives (Anura: Athesphatanura: Dendrobatidae). Bulletin of the American
Museum of Natural History, 299: 1-262.
Groeneveld, L.F.; Weisrock, D.W.; Rasoloarison, R.M.; Yoder, A.D.; Kappeler, P.M. 2009.
Species delimitation in lemurs: multiple genetic loci reveal low levels of species diversity in
the genus Cheirogaleus. BMC Evolutionary Biology, 9: 30.
Guerra, M.A.; Ron, S.R. 2008. Mate choice and courtship signal differentiation promotes
speciation in an Amazonian frog. Behavioral Ecology, 19: 1128-1135.
Guillot, G.; Renaud, S.; Ledevin, R.; Michaux, J.; Claude, J. 2012. A unifying model for the
analysis
of phenotypic,
genetic
and
geographic
data.
Systematic
Biology,
doi:
10.1093/sysbio/sys038.
Haffer, J. 1969. Speciation in Amazonian forest birds. Science, 165: 131-137.
Hall, T.A. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis
program for Windows 95/98/NT. Nucleic Acids Symposium Series 41: 95-98.
116
Hansen, T.F.; Orzack, S.H. 2005. Assessing current adaptation and phylogenetic inertia as
explanations of trait evolution: the need for controlled comparisons. Evolution, 59: 20632072.
Harpending, R.C. 1994. Signature of ancient population growth in a low-resolution mitochondrial
DNA mismatch distribution. Human Biology, 66: 591-600.
Hayes, F.E.; Sewlal, J.N. 2004. The Amazon River as a dispersal barrier to passerine birds: effects
of river width, habitat and taxonomy. Journal of Biogeography, 31: 1809-1818.
Heyer, R.W.; Reid, Y.R. 2003. Does advertisement call variation coincide with genetic variation
in the genetically diverse frog taxon currently known as Leptodactylus fuscus (Amphibia:
Leptodactylidae)? Anais da Academia Brasileira de Ciências, 75: 39-54.
Hoorn, C. 1994. An environmental reconstruction of the palaeo-Amazon River system (Middle–
Late Miocene, NW Amazonia). Palaeogeography, Palaeoclimatology, Palaeoecology, 112:
187-238.
Hoorn, C.; Wesselingh, F.P.; ter Steege, H.; Bermudez, M.A.; Mora, A.; Sevink, J.; Sanmartín, I.;
Sanchez-Meseguer, A.; Anderson, C.L.; Figueiredo, J.P.; Jaramillo, C.; Riff, D.; Negri, F.R.;
Hooghiemstra, H.; Lundberg, J.; Stadler, T.; Särkinen, T.; Antonelli, A. 2010. Amazonia
through time: Andean uplift, climate change, landscape evolution, and biodiversity. Science,
330: 927-931.
Hubert, N.; Meyer, C.P.; Bruggemann, H.J.; Guérin, F.; Komeno, R.J.L.; Espiau, B.; Causse, R.;
Williams, J.T.; Planes, S. 2012. Cryptic diversity in Indo-Pacific coral-reef fishes revealed by
DNA-barcoding provides new support to the centre-of-overlap hypothesis. PloS One, 7:
e28987.
Hutchison, D.W.; Templeton, A.R. 1999. Correlation of pairwise genetic and geographic distance
measures: inferring the relative influences of gene flow and drift on the distribution of genetic
variability. Evolution, 53: 1898-1914.
Irion, G.; Kalliola, R. 2010. Long-term landscape development processes in Amazonia. In: Hoorn,
C.; Wesselingh, F.P. (Eds). Amazonia, landscape and species evolution. Blackwell
Publishing, Oxford. p. 185-197.
117
Kaefer, I.L.; Lima, A.P. 2012. Sexual signals of the Amazonian frog Allobates paleovarzensis:
geographic variation and stereotypy of acoustic traits. Behaviour, 149: 15-33.
Kaefer, I.L.; Montanarin, A.; Costa, R.S.; Lima, A.P. Temporal patterns of reproductive activity
and site attachment of the Brilliant-thighed Frog Allobates femoralis from Central Amazonia.
Journal of Herpetology, in press.
Kaefer, I.L.; Tsuji-Nishikido, B.M.; Lima, A.P. 2012. Beyond the river: underlying determinants
of population acoustic signal variability in Amazonian direct-developing Allobates (Anura:
Dendrobatoidea). Acta Ethologica, doi:10.1007/s10211-012-0126-0.
Kimura, M. 1980. A simple method for estimating evolutionary rate of base substitutions through
comparative studies of nucleotide sequences. Journal of Molecular Evolution, 16: 111-120.
Lemmon, E.M. 2009. Diversification of conspecific signals in sympatry: geographic overlap
drives multidimensional reproductive character displacement in frogs. Evolution, 63: 11551170.
Librado, P.; Rozas, J. 2009. DnaSP v5: A software for comprehensive analysis of DNA
polymorphism data. Bioinformatics, 25: 1451-1452.
Lima, A.P.; Caldwell, J.P.; Biavati, G.; Montanarin, A. 2010. A new species of Allobates (Anura:
Aromobatidae) from paleovárzea forest in Amazonas, Brazil. Zootaxa, 2337: 1-17.
Losos, J.B. 2008. Phylogenetic niche conservatism, phylogenetic signal and the relationship
between phylogenetic relatedness and ecological similarity among species. Ecology Letters,
11: 995-1003.
Lougheed, S.C.; Austin, J.D.; Bogart, J.P.; Boag, P.T.; Chek, A.A. 2006. Multi-character
perspectives on the evolution of intraspecific differentiation in a Neotropical hylid frog. BMC
Evolutionary Biology, 6: 23.
Lougheed, S.C.; Gascon, C.; Jones, D.A.; Bogart, J.P.; Boag, P.T. 1999. Ridges and rivers: a test
of competing hypothesis of Amazonian diversification using a dart-poison frog (Epipedobates
femoralis). Proceedings of the Royal Society B, 266: 1829-1835.
Luna, A.G.; Hödl, W.; Amézquita, A. 2010. Colour, size and movement as visual subcomponents
in multimodal communication by the frog Allobates femoralis. Animal Behaviour, 79: 739745.
118
Lynch, M.; Crease, T.J. 1990. The analysis of population survey data on DNA sequence variation.
Molecular Biology and Evolution, 7: 377-394.
Maan, M.E.; Cummings, M.E. 2009. Sexual dimorphism and directional sexual selection on
aposematic signals in a poison frog. Proceedings of the National Academy of Sciences, 106:
19072-19077.
Mantel, N. 1967. The detection of disease clustering and a generalized regression approach.
Cancer Research, 27: 209-220.
McClelland, B.E.; Wilczynski, W.; Ryan, M.J. 1996. Correlations between call characteristics and
morphology in male cricket frogs (Acris crepitans). Journal of Experimental Biology, 199:
1907-1919.
McKay, B.D.; Zink, R.M. 2009. The causes of mitochondrial DNA gene tree paraphyly in birds.
Molecuar Phylogenetics and Evolution, 54: 647-650.
Melendez, K.V.; Feng, A.S. 2010. Communication calls of little brown bats display individualspecific characteristics. The Journal of the Acoustic Society of America, 128: 919-923.
Montanarin, A.; Kaefer, I.L.; Lima, A.P. 2011. Courtship and mating behaviour of the Brilliantthighed Frog Allobates femoralis from Central Amazonia: implications for the study of a
species complex. Ethology Ecology & Evolution, 23: 141-150.
Morales, V.R. 2002 “2000”. Sistemática y biogeografia del grupo trilineatus (Amphibia, Anura,
Dendrobatidae, Colostethus), con descripción de once nuevas especies. Publicaciones de la
Asociación Amigos de Doñana, 13: 1-59.
Moritz, C.; Patton, J.L.; Schneider, C.J.; Smith, T.B. 2000. Diversification of rainforest faunas: an
integrated molecular approach. Annual Review of Ecology and Systematics, 31: 533-563.
Narins, P.M.; Grabul, D.S.; Soma, K.K.; Gaucher, P.; Hödl, W. 2005. Cross-modal integration in
a dart-poison frog. Proceedings of the National Academy of Sciences, 102: 2425-2429.
Narins, P.M.; Hödl, W.; Grabul, D.S. 2003. Bimodal signal requisite for agonistic behavior in a
dart-poison frog, Epipedobates femoralis. Proceedings of the National Academy of Sciences,
100: 577-580.
119
Nielsen, R.; Beaumont, M.A. 2009. Statistical inferences in phylogeography. Molecular Ecology,
18: 1034-1047.
Nosil, P.; Harmon, L.J.; Seehausen, O. 2009. Ecological explanations for (incomplete) speciation.
Trends in Ecology and Evolution, 24: 145-156.
Palumbi, S.R. 1996. Nucleic acids II: the polymerase chain reaction. In: Hillis, D.M.; Moritz, C.;
Mable, B.K. (Eds). Molecular Systematics. Sinauer & Associates Inc., Sunderland,
Massachusetts. p. 205-247.
Panhuis, T.M.; Butlin, R.; Zuk, M.; Tregenza, T. 2001. Sexual selection and speciation. Trends in
Ecology and Evolution, 16: 364-371.
Patton, J.L.; Silva, M.N.F. 1998. Rivers, refuges, and ridges: The geography of speciation of
Amazonian mammals. In: Howard, D.J; Berlocher, S.H. (Eds). Endless Forms: Species and
Speciation. Oxford University Press, Oxford. p. 202-213.
Patton, J.L.; Silva, M.N.F.; Malcolm, J.R. 1994. Gene genealogy and differentiation among
arboreal spiny rats (Rodentia: Echymidae) of the Amazon Basin: a test of the riverine barrier
hypothesis. Evolution, 48: 1314-1323.
Peres, C.A.; Patton, J.L.; Silva, M.N.F. 1996. Riverine barriers and gene flow in Amazonian
saddle-back tamarins. Folia Primatologica 67: 113-124.
Prestwich, K.N. 1994. The energetics of acoustic signaling in anurans and insects. American
Zoologist, 34: 625-643.
Pröhl, H. 2005. Territorial behavior in dendrobatid frogs. Journal of Herpetology, 39: 354-365.
Pröhl, H.; Hagemann, S.; Karsch, J.; Höbel, G. 2007. Geographic variation in male sexual signals
in Strawberry Poison Frogs (Dendrobates pumilio). Ethology, 113: 825-837.
Pröhl, H.; Koshy, R.A.; Mueller, U.; Rand, A.S.; Ryan, M.J. 2006. Geographic variation of
genetic and behavioral traits in northern and southern túngara frogs. Evolution, 60: 16691679.
Rambaut, A.; Drummond, A.J. 2007. Tracer v 1.5 (http://beast.bio.ed.ac.uk/Tracer). Acesso:
02/03/2012.
120
Ramos-Onsins, S.E.; Rozas, J. 2002. Statistical properties of new neutrality tests against
population growth. Molecular Biology and Evolution, 19: 2092-2100.
Ribas,
C.C.;
Aleixo,
A.; Nogueira,
A.C.R.; Miyaki,
C.Y.;
Cracraft,
J.
2011.
A
palaeobiogeographic model for biotic diversification within Amazonia over the past three
million years. Proceedings of the Royal Society B, 279: 681-689.
Riddle, B.R. 2009. What is modern biogeography without phylogeography? Journal of
Biogeography, 36: 1-2.
Riddle, B.R.; Hafner, D.J. 2004. The past and future roles of phylogeography in historical
biogeography. In: Lomolino, M.V.; Heaney, L.R. (Eds). Frontiers of biogeography: new
directions in the geography of nature. Sinauer Associates Inc, Sunderland. p. 93-110.
Rodríguez, A.; Nuez, D.; Alonso, R. 2010. Intraspecific variation in the advertisement call of the
Cloud-Forest Frog Eleutherodactylus glamyrus (Anura: Eleutherodactylidae). Journal of
Herpetology, 44: 457-466.
Rosen, D.E. 1978. Vicariant patterns and historical explanation in biogeography. Systematic
Zoology, 27: 159-188.
Rossetti, D.F.; de Toledo, P.M.; Góes, A.M. 2005. New geological framework for Western
Amazonia (Brazil) and implications for biogeography and evolution. Quaternary Research,
63: 78-89.
Ryan, M.J. 1988. Energy, calling, and selection. American Zoologist, 28: 885-898.
Ryan, M.J.; Cocroft, R.B.; Wilczynski, W. 1990. The role of environmental selection in
intraspecific divergence of mate recognition signals in the cricket frog, Acris crepitans.
Evolution, 44: 1869-1872.
Ryan, M.J.; Rand, A.S.; Weigt, L.A. 1996. Allozyme and advertisement call variation in the
túngara frog, Physalaemus pustulosus. Evolution, 50: 2435-2453.
Ryan, M.J.; Wilczynski, W. 1991. Evolution of intraspecific variation in the advertisement call of
a cricket frog (Acris crepitans, Hylidae). Biological Journal of the Linnean Society, 44: 249271.
121
Santini, F.; Miglietta, M.P.; Faucci, A. 2012. Where are we now? An introduction to a special
issue on speciation. Evolutionary Biology, 39: 141-147.
Santos, J.C.; Coloma, L.A.; Summers, K.; Caldwell, J.P.; Ree, R.; Cannatella, D.C. 2009.
Amazonian amphibian diversity is primarily derived from Late Miocene Andean lineages.
PLOS Biology, 7: e1000056.
Santos, S.; Hrbek, T.; Farias, I.P.; Schneider, H.; Sampaio, I. 2006. Population genetic structuring
of the king weakfish, Macrodon ancylodon (Sciaenidae), in Atlantic coastal waters of South
America: deep genetic divergence without morphological change. Molecular Ecology, 15:
4361-4373.
Schneider, S.; Excoffier, L. 1999. Estimation of demographic parameters from the distribution of
pairwise differences when the mutation rates vary among sites: Application to human
mitochondrial DNA. Genetics, 152: 1079-1089.
Sequeira, F.; Sodré, D.; Ferrand, N.; Bernardi, J.; Sampaio, I.; Schneider, H.; Vallinoto, M. 2011.
Hybridization and massive mtDNA unidirectional introgression between the closely related
Neotropical toads Rhinella marina and R. schneideri inferred from mtDNA and nuclear
markers. BMC Evolutionary Biology, 11: 264.
Silva, M.N.F.; Patton, J.L. 1998. Molecular phylogeography and the evolution and conservation
of mammals. Molecular Ecology, 7: 475-486.
Simões, P.I. 2010. Diversificação do complexo Allobates femoralis (Anura, Dendrobatidae) em
florestas da Amazônia brasileira: desvendando padrões atuais e históricos. Tese de
Doutorado, Instituto Nacional de Pesquisas da Amazônia, Manaus, Amazonas.
Simões, P.I.; Lima, A.P.; Magnusson, W.E.; Hödl, W.; Amézquita, A. 2008. Acoustic and
morphological differentiation in the frog Allobates femoralis: Relationships with the upper
Madeira River and other potential geological barriers. Biotropica, 40: 607-614.
Sioli, H. 1984. The Amazon: Limnology and Landscape Ecology of a Mighty Tropical River and
its Basin. Dr W. Junk Publisher, Dordrecht.
Smith, M.J.; Hunter, D. 2005. Temporal and geographic variation in the advertisement call of the
booroolong frog (Litoria booroolongensis: Anura: Hylidae). Ethology, 111: 1103-1115.
122
Smouse, P.E.; Long, J.C.; Sokal, R.R. 1986. Multiple regression and correlation extensions of the
Mantel test of matrix correspondence. Systematic Zoology, 35: 727-732.
Solomon, S.E.; Bacci, M.; Martins, J.; Vinha, G.G.; Mueller, U.G. 2008. Paleodistributions and
comparative molecular phylogeography of leafcutter ants (Atta spp.) provide new insight into
the origins of Amazonian diversity. PLoS One, 3: e2738.
Tajima, F. 1989. Statistical method for testing the neutral mutation hypothesis by DNA
polymorphism. Genetics, 123: 585-595.
Tamura, K.; Peterson, D.; Peterson, N.; Stecher, G.; Nei, M.; Kumar, S. 2011. MEGA5:
Molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance,
and maximum parsimony methods. Molecular Biology and Evolution, 28: 2731-2739.
Tárano, Z. 2001. Variation in male advertisement calls in the neotropical frog Physalaemus
enesefae. Copeia, 2001: 1064-1072.
Telles, M.P.C.; Silva, R.S.M.; Chaves, L.J.; Coelho, A.S.G.; Diniz–Filho, J.A.F. 2001.
Divergência entre subpopulações de cagaiteira (Eugenia dysenterica) em resposta a padrões
edáficos e distribuição espacial. Pesquisa Agropecuária Brasileira, 36: 1387-1394.
Templeton, A.R.; Crandall, K.A.; Sing, C.F. 1992. A cladistic analysis of phenotypic associations
with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III.
Cladogram estimation. Genetics, 132: 619-633.
Thompson, J.D.; Higgins, D.G.; Gibson, T.J. 1994. Improved sensitivity of profile searches
through the use of sequence weights and gap excision. Computer Applications in the
Biosciences, 10: 19-29.
Tsuji-Nishikido, B.; Kaefer, I.L.; Freitas, F.C.; Menin, M.; Lima, A.P. 2012. Significant but not
diagnostic: Differentiation through morphology and calls in the Amazonian frogs Allobates
nidicola and A. masniger. Herpetological Journal, 22: 105-114.
Turmelle, A.S.; Kunz, T.H.; Sorenson, M.D. 2011. A tale of two genomes: contrasting patterns of
phylogeographic structure in a widely distributed bat. Molecular Ecology, 20: 357-375.
Vences, M.; Wake, D.B. 2007. Speciation, species boundaries and phylogeography of
amphibians. In: Heatwole, H.; Tyler, M.J. (Eds). Amphibian Biology, Systematics. Surrey
Beatty & Sons, Chipping Norton. p. 2613-2671.
123
Vences, M.; Kosuch, J.; Lötters, S.; Widmer, A.; Jungfer, K.; Köhler, J.; Veith, M. 2000.
Phylogeny and classification of Poison Frogs (Amphibia: Dendrobatidae), based on
mitochondrial 16S and 12S ribosomal RNA gene sequences. Molecular Phylogenetics and
Evolution, 15: 34-40.
Vences, M.; Thomas, M.; Meijden, A.; Chiari, Y.; Vieites, D.R. 2005. Comparative performance
of the 16S rRNA gene in DNA barcoding of amphibians. Frontiers in Zoology, 2: 5.
Verburgt, L.; Ferreira, M.; Ferguson, J.W.H. 2011. Male field cricket song reflects age, allowing
females to prefer young males. Animal Behaviour, 81: 19-29.
Verdade, V.K.; Rodrigues, M.T. 2007. Taxonomic review of Allobates (Anura, Aromobatidae)
from the Atlantic Forest, Brazil. Journal of Herpetology, 41: 566-580.
Vieites, D.R.; Wollenberg, K.C.; Adreone, F.; Köhler, J.; Glaw, F.; Vences, M. 2009. Vast
underestimation of Madagascar’s biodiversity evidenced by an integrative amphibian
inventory. Proceedings of the National Academy of Sciences, 106: 8267-8272.
Wallace, A.R. 1852. On the monkeys of the Amazon. Proceedings of the Zoological Society of
London, 20: 107-110.
Wells, K.D. 2007. The ecology and behavior of amphibians. The University of Chicago Press,
Chicago.
Wells, K.D.; Schwartz, J.J. 2007. The Behavioral Ecology of Anuran Communication. In: Narins,
P.M.; Feng, A.S.; Fay, R.R.; Popper, A.N. (Eds). Hearing and Sound Communication in
Amphibians. Springer-Verlag, New York. p. 44-86.
West-Eberhard, M.J. 1989. Phenotypic plasticity and the origins of diversity. Annual Review of
Ecology and Systematics, 20: 249-278.
Wiens, J.J. 2004. Speciation and ecology revisited: phylogenetic niche conservatism and the
origin of species. Evolution, 58: 193-197.
Wiens, J.J. 2007. Species delimitation: New approaches for discovering diversity. Systematic
Biology, 56: 875-878.
Wiens, J.J. 2008. Systematics and herpetology in the age of genomics. Bioscience, 58: 297-307.
Wiley, E.O. 1978. The evolutionary species concept reconsidered. Systematic Zoology, 27: 17-26.
124
Wilkinson, L. 2007. SYSTAT: the system for statistics. SYSTAT inc., Evanston.
Wright, S. 1951. The genetical structure of populations. Annals of Human Genetics, 15: 323-354.
Wycherley, J.; Doran, S.; Beebee, T.J.C. 2002. Male advertisement call characters as
phylogeographical indicators in European water frogs. Biological Journal of the Linnean
Society, 77: 355-365.
Zeisset, I.; Beebee, T.J.C. 2008. Amphibian phylogeography: a model for understanding historical
aspects of species distributions. Heredity, 101: 109-119.
Zhang, D.X.; Hewitt, G.M. 2003. Nuclear DNA analyses in genetic studies of populations:
practice, problems and prospects. Molecular Ecology, 12: 563-584.
Zink, R.M.; Barrowclough, G.F. 2008. Mitochondrial DNA under siege in avian phylogeography.
Molecular Ecology, 17: 2107-2121.
Zink, R.M.; Blackwell-Rago, R.C.; Ronquist, F. 2000. The shifting roles of vicariance in
biogeography. Proceedings of the Royal Society B, 267: 497-503.
Zweifel, R.G. 1968. Effects of temperature, body size, and hybridization on mating calls of toads,
Bufo a. americanus and Bufo woodhousii fowleri. Copeia, 1968: 269-285.
125
Anexos
Pareceres emitidos pelas bancas examinadoras
da aula de qualificação, da versão escrita da tese
e da defesa pública da tese, respectivamente.
126
127
128
129
130
131