View
2
Download
0
Category
Preview:
Citation preview
INSTITUTO NACIONAL DE PESQUISAS DA AMAZÔNIA – INPA
PROGRAMA DE PÓS-GRADUAÇÃO EM ECOLOGIA
Variação diária e espectral da atividade vocal de três espécies da
família Tinamidae em florestas de terra firme da Amazônia
GABRIEL BELEIA MCCRATE
Manaus, Amazonas
Outubro, 2013
i
GABRIEL BELEIA MCCRATE
Variação diária e espectral da atividade vocal de três espécies da família Tinamidae em
florestas de terra firme da Amazônia brasileira
Dr. GONÇALO FERRAZ
Dissertação apresentada ao
Instituto Nacional de Pesquisas da
Amazônia como parte dos requisitos
para obtenção do título de Mestre em
Biologia (Ecologia).
Manaus, Amazonas
Outubro, 2013
ii
Banca examinadora do trabalho escrito
Nome e Instituição Parecer
Dr. José Luis Tella Aprovado com correções
Estación Biológica de Donoña
Dr. Pedro Ivo Simões Aprovado com correções
Instituto Nacional de Pesquisas da Amazônia
Dr. Edson Varga Lopes Aprovado com correções
Universidade Federal do Oeste do Pará
Banca examinadora da defesa oral pública
Dr. Mario Cohn-Haft
Instituto Nacional de Pesquisas da Amazônia
Dr. Pedro Ivo Simões
Instituto Nacional de Pesquisas da Amazônia
Dr. Fabricio Baccaro
Universidade Federal do Amazonas
Aprovado por maioria
iii
iv
AGRADECIMENTOS
Agradeço,
Aos meus pais, Adelaide Beleia e Andrew McCrate que sempre incentivaram meu
interesse pelas ciências e pela natureza.
Ao Gonçalo Ferraz pela orientação, pela oportunidade de investigar as gravações
amazônicas de um modo ainda não explorado.
Aos meus amigos, colaboradores e colegas do Laboratório de Ecologia de Populações.
Vocês sabem quem são e como foram de extrema importância para o desenvolvimento deste
trabalho
Aos pesquisadores Mario Cohn-Haft, Cintia Cornelius, José Luis Tella, Pedro Ivo Simões
pelos conselhos e auxilio na concepção deste trabalho. Christian Borges Andretti, Claudeir
Sinopse:
Estudamos o comportamento vocal de três espécies da família Tinamidae e os usos dos
nichos acústico e a segregação temporal da atividade vocal. Utilizamos a probabilidade
de detecção de três espécies de Tinamidae para determinar os comportamentos vocais.
Utilizamos gravadores autônomos para a coleta de dados em 212 pontos amostrais nos
anos de 2010 e 2011. Aspectos sobre a segregação temporal e espectral do nicho acústico
são discutidos.
Palavras chave: Bioacústica, Nicho acústico, Atividade vocal, Tinamidae
v
Vargas foram de imensa ajuda na identificação das espécies mais difíceis e das vocalizações mais
incomuns que gravamos.
Ao Projeto de Dinâmica Biológica de Fragmentos Florestais e ao Programa de Pós
Graduação em Ecologia pelo apoio logístico e intelectual, respectivamente.
Aos moradores da, agora extinta, Casa Preta que me receberam em meu primeiro dia de
Manaus e sempre foram excelentes companhias, João Vitor “JB” Campos e Silva, Rafael
“Fumaça” Assis, José “Zéca” Purri, Arnold “Luguito” Lugo e Bernardo “Berna” Flores.
Ao CNPq pela bolsa de estudo.
vi
‘I listened to the dawn chorus,
Birds were greeting the new day. Were they singing praise to Horus,
As they had in Egypt’s ancient way. Or were they glad to be alive,
As darkness turned to blessed light.
Feeding their fledglings so to thrive, That they may too enjoy this thrilling sight.
Awakening to the early morning dew, Feeling the warmth of the new sun.
A privilege granted to but a few,
For many the daylight do but shun. I have listened to the bird’s greetings,
When all around was calm and still.
To me they were happy meetings, That still gives me a tremendous thrill.’
George Bernard Shaw
RESUMO
As aves comunicam-se principalmente através de sinais acústicos. Sinais acústicos são
especialmente importantes em ambientes como a floresta amazônica, onde sinais visuais de longe
alcance são impedidos pela vegetação. Sinais acústicos de espécies de aves que apresentam
características espectrais semelhantes também podem sofrer com interferência acústica. Portanto,
as aves devem repartir o espaço acústico que é finito. Para evitar à sobreposição espectral e/ou
temporal as espécies adaptam suas vocalizações para utilizar frequências diferentes ou emitem
sinais acústicos em diferentes momentos do coro matinal. Utilizamos gravadores autônomos para
gravar o coro matinal durante duas estações de seca durante os anos de 2010 e 2011 no Projeto de
Dinâmica Biológica de Fragmentos Florestais (PDBFF). Usamos estimativas das probabilidades
de detecção ao longo da manhã de três espécies de aves da família Tinamidae ao longo do coro
matinal. Este estudo investiga como as espécies Tinamus major, Crypturellus variegatus e
C.brevirostris utilizam o espaço acústico e qual o nível de sobreposição temporal e espectral das
vii
atividades vocais. Nossos resultados indicam que estas espécies utilizam o espaço acústico de
forma semelhante e há sobreposição temporal entre estas espécies. Demonstramos que a
sobreposição temporal é maior que esperado ao acaso e que há diferenças significativas entre os
parâmetros vocais analisados. Fizemos estimativas das probabilidades de detecções das três
espécies ao longo do coro matinal para demonstrar que houve diferentes intensidades de
sobreposição temporal da atividade vocal entre as três espécies. Nosso método utiliza uma
resolução temporal e espacial ainda pouco utilizada para estudos bioacústicos, portanto nossos
resultados reforçam que a utilização de gravadores autônomos é uma excelente ferramenta para
futuros estudos de bioacústica.
ABSTRACT
Birds are known to share information through conspicuous acoustic signals. Acoustic signals are
especially important in habitats where vegetation makes long range visual signals ineffective.
However, species that share similar characteristics of acoustic signals might also experience
acoustic interference. Species that are closely related phylogenetically can also overlap in
ecological and behavioral characteristics, enhancing the chance of acoustic interference. In order
to avoid acoustic signal overlap, there might be present behavioral adaptations such as adjusting
their vocal activity to avoid syntopic congeners. In this study, we used autonomous recording
units (ARUs) to sample the dawn chorus activity of three species of tinamous (Family
Tinamidae) during the dry season for two consecutive years at the Biological Dynamics of Forest
Fragments Project in central Amazon, Brazil. We applied a method that accounts for imperfect
detection in order to estimate the time of peak vocal activity for each species, while considering
effects from the sampling bias. This study investigates how the species Tinamus major,
viii
Crypturellus variegatus and C. brevirostris utilize the acoustic space in time, and evaluate the
degree of overlap in spectral and temporal traits of their calls. Our results suggests that these
species use the acoustic space in a similarly, broadly overlapping in relation to the timing of
vocal activity. However, these species differ significantly in characteristics of their calls. Our
sampling and analytical methods had a temporal and spatial resolution that few bioacoustics
studies have utilized; therefore we reinforce the body of publications which make use of ARUs
and the usage of detection probabilities to better understand the dynamics of the dawn chorus.
Keywords: Tinamidae, acoustical niche, central amazon, temporal segregation, autonomous recorders, imperfect
detection
SUMÁRIO
Resumo......................................................................................................................................vi
Abstract....................................................................................................................................vii
Apresentação ……....................................................................................................................8
Artigo ......................................................................................................................................13
Resumo................................................................................................................................14
Abstract…………………………………………………………………………………...15
Introduction.........................................................................................................................17
Methods...............................................................................................................................20
Sampling area......................................................................................................................20
Sampling design..................................................................................................................21
Results.................................................................................................................................26
Discussion ..........................................................................................................................27
ix
Acknowledments.................................................................................................................29
Literature Cited...................................................................................................................30
Tables..................................................................................................................................36
Figures.................................................................................................................................37
Conclusões ..............................................................................................................................42
Anexos…………………………………………………………....……………………….….43
8
APRESENTAÇÃO
O coro matinal das aves, o comportamento de concentrar a atividade vocal durante o amanhecer,
é um fenômeno comum, cíclico e conspícuo em todos os ambientes onde vivem as aves. A
comunicação é qualquer transmissão de um sinal, independente de sua natureza (química, táctil,
visual ou auditiva) que incita uma resposta entre dois ou mais indivíduos. De maneira geral,
entre as aves a comunicação é feita principalmente através de sinais acústicos; chamados e
cantos. Os sinais acústicos possuem funções como discriminar entre espécies, atrair parceiros,
estabelecer e defender territórios e também revelar informações como saúde, localização e
identidade (Wiley e Richards 1978; Marler e Slabbekoorn 2004; Platzen e Magrath 2004).
Compreender a função dos cantos e chamados permite um melhor entendimento sobre a ecologia
das aves.
A informação contida dentro do sinal acústico é o que Vielliard (1987) cunhou como
“canto funcional” e que pode fornecer a identidade do indivíduo e sua disposição
comportamental. (McCracken e Sheldon 1997; Mathevon et al. 2008). Porém, essa é a realidade
de ambientes com sazonalidade bem definida diferentemente do que ocorre em ambientes
tropicais onde as espécies vocalizam durante todo o ano. Para este estudo, cantos e chamados
serão tratados como sinais acústicos ou vocalizações sem fazer distinção entre os dois. Em um
ambiente com baixa luminosidade e vegetação densa, como na floresta Amazônica os sinais
visuais de longo alcance são pouco eficientes e a comunicação através de sinais acústicos se faz
mais relevante quando comparada com sinais visuais (Reis de Magalhães 1996; Planqué e
Slabbekoorn 2008). Em um ambiente como a floresta da Amazônia central onde já foram
registradas aproximadamente 400 espécies de aves (Cohn-Haft et al. 1997) existe uma paisagem
9
sonora onde há a possibilidade efeitos da interferência acústica hetero-específica (Planqué e
Slabbekoorn 2008; Luther 2009). Espécies que compartilham características acústicas
semelhantes estão sujeitas à interferência entre seus sinais acústicos e a intensidade desta
interferência é proporcional à similaridade dos sinais (Marler e Slabbekoorn 2004). A
interferência entre sinais acústicos de espécies congêneres possuem consequências negativas
como, por exemplo, podem ocasionar hibridizações e a interações antagonistas entre indivíduos.
(Hoi e Sageder 1991; Wiley 1994; de Kort 2002).
Para que um indivíduo de uma espécie possa reconhecer um sinal acústico é necessário
que seja capaz de discriminar o sinal acústico do ruído de fundo (Wiley 1994). O ruído de fundo
é todo som produzido no ambiente que interfere no reconhecimento do sinal acústico pelo
indivíduo receptor do sinal. O ruído de fundo tem origem em sons abióticos como, por exemplo,
o som do vento ao passar pela vegetação e da chuva, e também de origem bióticas como, por
exemplo, as vocalizações de outras espécies e o constante estridular de insetos. Quando espécies
são próximas filogeneticamente podem apresentar características morfológicas, ecológicas e
comportamentais semelhantes. Logo, a interferência entre sinais acústicos pode ser mais comum
entre espécies que ocorrem em simpatria (Tobias e Seddon 2009). Espécies diferentes que
compartilham características acústicas, potencialmente diminuem a probabilidade de serem
reconhecidas com sucesso entre indivíduos da mesma espécie. Portanto, quando duas ou mais
espécies vocalizam simultaneamente podem ser vistas como competidoras pelo “espaço acústico”
(Nelson e Marler 1990). Em 1987, Krause propôs a hipótese do nicho acústico, sugerindo que as
espécies que se comunicam através de sinais acústicos devem competir por um espaço acústico
finito; as espécies sofrem pressão seletiva para que seus sinais acústicos não sofram
sobreposições temporais ou espectrais. O espaço acústico é definido como um espaço finito
10
multidimensional com os eixos definidos pelo espectro (e.g. frequências máximas e mínimas,
amplitude e harmônicos complexos) e pelo tempo (e.g. duração da vocalização, duração de notas)
que são características que determinam a estrutura do sinal assim como os padrões diários da
atividade vocal (Marler 1960). Com o aumento no número de espécies vocalmente ativas de um
mesmo ambiente, cada espécie seria beneficiada se o limitado espaço acústico fosse repartido
para que não houvesse sobreposição temporal ou espectral (Krause 1987).
O fenômeno do coro matinal é o evento comportamental mais evidente da atividade vocal
das aves que é comum em todos os habitats em que vivem (Henwood e Fabrick 1979). Durante o
coro matinal há um rápido aumento no número de vocalizações logo antes do sol nascer com uma
diminuição significativa após algumas horas, embora a maior parte espécies vocalizem durante o
coro matinal diversos estudos apontam que existem diferenças nos momentos iniciais de
participação no coro entre espécies (McNamara et al. 1987; Berg et al. 2006; Luther 2008).
Luther (2008) investigou a segregação temporal na atividade vocal de quatro espécies de
Thamnophilidae e encontrou evidências que todas as espécies possuem picos de atividade vocal
significativamente diferentes durante o coro matinal. Padrões temporais de atividade são
indicativos de uso do ambiente por uma espécie e é considerado como uma dimensão de nicho
(Pianka 1973). Deste modo, espécies que coabitam e compartilham dos mesmos recursos, como
presas, podem reduzir a competição se forem ativas em diferentes horas do dia (Cody 1974).
Vários estudos propuseram que as aves tem seu comportamento vocal adaptado ao
ambiente acústico para aumentar a eficiência da transmissão do sinal (Wiley and Richards 1975,
Slabbekoorn et al. 2002, Tobias et al. 2010). Entretanto, estes estudos não representam a
unanimidade e outras hipóteses sobre a divergência vocal das aves também existem. Podos
11
(2001) propõe que as vocalizações são uma adaptação morfológica devido a diferenças do
tamanho corporal influenciada por restrições ambientais (e.g. tamanho da presa, guilda trófica).
Apesar da incerteza sobre o processo evolutivo de divergência das vocalizações, já foi
demonstrado que as aves mantem as características vocais das espécies ancestrais (ten Cate 2004;
Weckstein 2005).
Um sinal acústico sofre com dois problemas físicos inerentes a propagação do som
através do ar: a reverberação e atenuação. A reverberação refere-se à energia do sinal acústico
perdida quando o som encontra obstáculos no ambiente (e.g troncos, galhos e folhas) que acarreta
em distorções na frequência e amplitude do sinal acústico que é reverberado em forma de ecos; a
atenuação refere-se ao decréscimo da amplitude do sinal através da dissipação e absorção da
energia do sinal acústico pelo ar e vegetação. A degradação do sinal acústico é a soma destes dois
fatores que interferem na transmissão do sinal acústico, alterando as características espectrais do
sinal a partir do momento que ele é emitido até chegar ao receptor (Wiley e Richards 1982). A
degradação limita a distância em que o sinal acústico possuí significado biológico, ou seja, a
distância máxima em que um sinal acústico pode percorrer sem perder o significado da
informação transmitida. A degradação diminui a energia do sinal até que este desapareça no ruído
de fundo (Mathevon et al. 2004).
A ecologia das espécies da família Tinamidae foi objeto de estudo de poucos trabalhos,
principalmente as espécies que habitam florestas, esta por serem de hábitos crípticos e de difícil
visualização (Cabot 1992; Brennan 2004). Uma forma eficiente de se obter informações
ecológicas relevantes é através dos sinais acústicos, pois as espécies que vivem no solo de
florestas densas, como os Tinamidae, possuem sinais que tendem a ser caracterizados por baixas
12
frequências (<2.5 KHz); tipicamente possuem uma largura de banda estreita com muito pouca
modulação (Brandes 2008). Estes tipos de sinais acústicos possuem vantagens em florestas
densas por que sofrem menos com a degradação acústica produzida pela reverberação (Bartelli e
Tubaro 2002, Slabbekoorn et al 2002).
Sabendo que as espécies Tinamus major, Crypturellus variegatus e C.brevirostris
compartilham recursos como hábitat, alimentos e o espaço acústico nós investigamos como as
vocalizações dessas espécies são temporal e espectralmente distribuídas no coro matinal da
Amazônia central, com as predições a priori i) o pico de atividade vocal destas três espécies ao
longo do coro matinal irão apresentar atividades vocais segregadas temporalmente. ii) espécies
que apresentam maior sobreposição espectral serão mais segregadas temporalmente.
O atual conhecimento sobre o nicho acústico teve grandes contribuições da ecologia de
insetos quando Gagala e Riede (1995), Riede (1997) sugeriram que comunidades de grilos
evoluíram para evitar a sobreposição temporal e espectral, assim como em cigarras (Schmidt et al
2013). Em anfíbios a partição do nicho acústico também é documentada (Sinsch et al. 2012, Both
e Grant 2012). O presente estudo tem como objetivo contribuir com o conhecimento neste
assunto e expandir a compreensão do nicho acústico de três espécies da família Tinamidae da
Amazônia central.
13
Capítulo 1
McCrate, G.; e Ferraz, G.. Spectral and Temporal
Niche Partioning of Three Tinamidae Species in
Central Amazonia.
Manuscrito formatado para Acta Amazonica
14
Spectral and Temporal Niche Portioning of three Tinamidae Species in Central Amazonia. 1
Gabriel Beleia MCCRATE1,*
e Gonçalo FERRAZ2 2
3
1Programa de Pós Graduação em Ecologia, Instituto Nacional de Pesquisas da Amazônia – INPA, Manaus - AM, 4
Brazil 5
6
2Departamento de Ecologia, Universidade Federal do Rio Grande do Sul, - UFGRS, Porto Alegre, Brazil 7
8
*Correponding authorgmccrate@gmail.com 9
10
Resumo: O coro matinal é o evento diário mais conspícuo produzido pelas aves e é comum em 11
todos os ambientes em que vivem. Apesar da aparente aleatoriedade na emissão de sinais 12
acústicos pelas aves, estas adaptaram-se para que a interferência acústica entre sinais de espécies 13
distintas fosse minimizado. Há duas prováveis formas de evitar a sobreposição de sinais 14
acústicos: evitar a sobreposição temporal, ou seja, cada espécie ajusta o horário de início do coro 15
matinal específico. A segunda forma seria evitar a sobreposição espectral, ou seja, as espécies 16
sofrem pressões seletivas que alteram as características espectrais dos sinais acústicos. O 17
objetivo deste estudo é compreender como três espécies de aves (Família Tinamidae) se 18
segregam temporal e espectralmente no espaço acústico em uma área da Floresta Amazônica 19
dentro do Projeto de Dinâmica Biológica de Fragmentos Florestais, localizado ao norte de 20
Manaus, Brasil. Utilizamos gravadores autônomos para gravar 3 horas diárias do coro matinal 21
durante o período da seca (Jun-Out) nos anos de 2010 e 2011. Testamos a hipótese de que 22
espécies que possuem sinais acústicos mais semelhantes teriam que possuir picos de atividade 23
vocal com um intervalo de tempo maior do que espécies que apresentam sinais acústicos com 24
maiores diferenças espectrais. Fizemos estimativas das probabilidades de detecções das três 25
15
espécies ao longo do coro matinal para demonstrar que houve diferentes intensidades de 26
sobreposição temporal da atividade vocal entre as três espécies. No entanto, as análises 27
espectrais demonstraram que não existe sobreposição temporal entre nenhum das espécies. 28
Nosso método utiliza uma resolução temporal e espacial ainda pouco utilizada para estudos 29
bioacústicos, portanto nossos resultados reforçam que a utilização de gravadores autônomos é 30
uma excelente ferramenta para futuros estudos de bioacústica. 31
Palavras chave: Tinamidae, nicho acústico, segregação temporal, gravadores autônomos, 32
detecção imperfeita 33
__________________________________________________________________________ 34
ABSTRACT 35
Birds are known to share information through conspicuous acoustic signals. Acoustic signals are 36
especially important in habitats where vegetation makes long range visual signals ineffective. 37
However, species that share similar characteristics of acoustic signals might also experience 38
acoustic interference. Species that are closely related phylogenetically can also overlap in 39
ecological and behavioral characteristics, enhancing the chance of acoustic interference. In order 40
to avoid acoustic signal overlap, there might be present behavioral adaptations such as adjusting 41
their vocal activity to avoid syntopic congeners. In this study, we used autonomous recording 42
units (ARUs) to sample the dawn chorus activity of three species of tinamous (Family 43
Tinamidae) during the dry season for two consecutive years at the Biological Dynamics of Forest 44
Fragments Project in central Amazon, Brazil. We applied a method that accounts for imperfect 45
detection in order to estimate the time of peak vocal activity for each species, while considering 46
effects from the sampling bias. This study investigates how the species Tinamus major, 47
16
Crypturellus variegatus and C. brevirostris utilize the acoustic space in time, and evaluate the 48
degree of overlap in spectral and temporal traits of their calls. Our results suggests that these 49
species use the acoustic space in a similarly, broadly overlapping in relation to the timing of 50
vocal activity. However, these species differ significantly in characteristics of their calls. Our 51
sampling and analytical methods had a temporal and spatial resolution that few bioacoustics 52
studies have utilized; therefore we reinforce the body of publications which make use of ARUs 53
and the usage of detection probabilities to better understand the dynamics of the dawn chorus. 54
55
Keywords: Tinamidae, acoustical niche, central amazon, temporal segregation, autonomous recorders, imperfect 56
detection 57
58
59
60
61
62
63
64
65
66
67
17
Introduction 68
Many invertebrates and vertebrates taxa, including birds, use acoustics signals such as songs and 69
calls, to attract mates, establish and defend territories and to share information about signaler 70
location, health and identity (Wiley and Richards 1978; Marler and Slabekoorn 2004; Platzen 71
and Magrath 2004). The information contained within the acoustic signal, what Viellard (1987) 72
coined as the functional call, can include identity and indications of behavioral disposition 73
(Mathevon et al. 2008; McCracken and Sheldon 1997). In an environment with low-light 74
conditions and dense vegetation which obstructs long range visual communication, acoustic 75
communication is more efficient compared to visual signals (Reis de Magalhães 1996 ,Planqué 76
and Slabbekoorn 2008).In a forest which supports a highly diverse bird community, such as the 77
central lowland forest of the Amazon, where approximately 400 species of terrestrial bird species 78
co-occur (Cohn-Haft et al. 1997) there is much potential for several species to share acoustic 79
signal characteristics that could be a source of heterospecific interference (Planqué and 80
Slabbekoorn 2008; Luther 2009). Species which share similar spectral characteristics of 81
acoustical signals are subject to interference and masking of their signal, the degree of this 82
masking depends on the similarity of the competing signals (Marler and Slabbekoorn 2004). 83
Such signal interference between species might lead to unwarranted hybridization and 84
antagonistic interactions between individuals (Wiley 1994; de Kort 2002). 85
For species to recognize intraspecific signals, the receiver should be able to discriminate 86
the signal from the background noise (Wiley 1994). Background noise originates from abiotic 87
sounds, such as wind and rain, and also from biotic sounds, such as the vocalization of other 88
species and stridulating insects. Species that are closely phylogenetic related may share physical, 89
ecological and behavioral characteristics, hence acoustic interference maybe more common in 90
18
these species (Tobias and Seddon 2009). Species that share common spectral characteristics 91
potentially decrease the chances of successfully transmitting a signal. Therefore, when two or 92
more species sing simultaneously that can be seen as competition for “acoustic space” (Nelson 93
and Marler 1990). Krause (1987) proposed the acoustic niche hypothesis, stating that animal 94
species which communicate through acoustic signals must compete for a finite acoustic space; 95
their vocalizations would be under selective pressure to avoid interspecific temporal and/or 96
spectral overlap. The acoustic space is defined as a multidimensional finite space with the axis´ 97
defined by spectral (e.g. maximum and minimum frequencies, amplitude and complex 98
harmonics) and temporal (e.g. song length, note duration) characteristics that determine the 99
signal structure as well as the diel patterns of vocal activity (Marler 1960). With the increase in 100
number of vocally active species at each site, each species would benefit from temporally and/or 101
spectrally partitioning their vocalization within a limited acoustic space so as to avoid acoustic 102
overlap (Krause 1987). 103
The regular phenomenon of the dawn chorus is the most conspicuous behavioral event of 104
vocal activity in birds and it is common to all habits in which birds live (Henwood and Fabrick 105
1979). During the dawn chorus there is a rapid increase of vocalizations before the sunrise with a 106
significant decrease later in the morning, although many species sing during the dawn chorus 107
each species starts vocalizing at different times (McNamara et al. 1987; Berg et al. 2006; Luther 108
2008). Luther (2008) investigated temporal segregation in vocal activity of four species of 109
Thamnophilidae and he oberserved that each species has a significantly different peak of vocal 110
activity within the dawn chorus. Temporal activity patters are indications of species use of the 111
environment and it is considered to be a niche dimension (Pianka 1973). Thus, species that co-112
19
occur in the same habitat and share the same resources, such as prey, may reduce competition by 113
being active at different times of the day (Cody, 1974). 114
Several studies have proposed that birds adapt their vocal behavior to their acoustic 115
environment to maximize efficiency signals transmission (Wiley and Richards 1978; 116
Slabbekoorn et al. 2002; Tobias et al. 2010). Neotropical ground dwelling birds which inhabit 117
dense forests, such as most species from the Tinamidae family, have signals which tend to be 118
characterized by low frequencies; typically song notes are concentrated in a very narrow 119
frequency bandwidth with very little modulation (Brandes 2008). These signal types are 120
advantageous in a dense forest because they suffer less acoustic degradation produced by 121
reverberation (Bartelli and Tubaro 2002; Slabbekoorn et al 2002). Signal degradation by 122
reverberation limits the distance in which an acoustic signal is still biologically relevant 123
(Mathevon et al. 2008). Indeed the Tinamidae produce acoustic signals that concentrate energy 124
in a narrow bandwidth, which increases transmission distance. The signal emitted close to the 125
forest floor and has its effective distance increased by extending the call duration filling the 126
inter-note silences with reverberations (Morton 1975; Slabbekoorn et al. 2002). Unlike oscine 127
passerines, Psitacidae and Trochelidae; Tinamidae do not learn their songs by listening to their 128
co-specifics. Recent studies (Fandino and Kroodsma, unpub. data in Marler and Slabbekoorn 129
2004) suggest that vocal learning might have evolved recently and independently in the 130
Cotingidae family. But further investigation is needed to affirm that this is true. Vielliard (2004) 131
documented that Tinamidae species have genetically innate calls; this was confirmed by raising 132
chicks in acoustic isolation from other birds. In the same study Vielliard (2004) speculated that 133
because calls emitted by Tinamus and Crypturellus have irregular durations, it is only the 134
frequency component of the signal, which is biologically meaningful. 135
20
Knowing that the species Tinamus major, Crypturellus variegatus and C.brevirostris 136
share resources such as space and food, we investigate how these species are temporally and 137
spectrally distributed along the dawn chorus in central Amazon, with the a priori prediction i) 138
that the detection probability, used as a surrogate to determine the peak vocal activity of the three 139
species along the dawn chorus will exhibit a segregated vocal activity to minimize acoustic 140
interference and maximize signal efficiency. ii) Species which have more similar vocal 141
characteristics will be more temporally segregate to avoid signal overlap. 142
The current understanding of acoustical niche has had major contributions from insect 143
ecology (Gogala and Riede 1995, Riede 1997) where there is enough evidence that co-occurring 144
cicadas have evolved to avoid temporal and spectral overlap, as well as crickets (Schmidt et al 145
2013). In frogs acoustic and temporal niche portioning has been documented (Sinsch et al. 2012, 146
Both and Grant 2012). The present study has the objective to increase the knowledge available 147
within this subject and to expand the understanding of the acoustic niche in Amazonian bird 148
communities. 149
150
151
Methods 152
Sampling Area 153
Our study was conducted at the Biological Dynamics of Forest Fragments Project area (BDFFP; 154
2°30´S, 60°W), located 80km north of Manaus, Amazonas, Brazil. The BDFFP consists of an 155
experimental fragmented landscape within a matrix of continuous primary evergreen terra firme 156
21
forest and areas that have been clear cut for pastures. Due to the nutrient poor soils and low 157
productivity pastures have been gradually abandoned and regrown into different stages of 158
secondary forests (capoeiras) with ages varying between 16 to 26 years (Mesquita et al. 2001). 159
Primary forest has an average canopy height of 35m with emerging tree reaching 45m. Rainfall 160
is seasonal in central Amazonia with a pronounced dry season during June-October, rainfall 161
ranges from 1900mm to 3500mm (Bierregaard et al. 2001). The study area is roughly 40km from 162
east to west and 10km from south to north, and the seven camps distributed in area facilitating 163
the extension of our sampling effort in areas of continuous and secondary growth forests. 164
Sampling design 165
To obtain the acoustic signals of our focal species we used thirty autonomous recording units 166
(ARUs; Model SM2, Wildlife Acoustics Concord, Massachusetts, USA). The option of using 167
ARUs was made due to the larger number of replicates that can be made simultaneously in 168
multiple sampling points and thus allowing sampling a larger area compared to the traditional 169
methods in the same time frame (Cerqueira et al. 2013).. Recordings were made in stereo with 170
two omnidirectional microphones SMX-II (Frequency response = 20Hz - 20,000Hz). The ARUs 171
were programmed using Song Meter Configuration Utility (Wildlife Acoustics, Concord, MA) to 172
begin recording 40 minutes before sunrise and continuously record for three hours in total. The 173
SM2 ARU allows adding the geographical position of the sampling point, in which case all 174
recordings are corrected daily for the small incremental changes of sunrise due to Earth´s axis 175
movement. Audio data was stored as uncompressed WAVE files with a sampling rate of 16 KHz 176
and 16 bits. The ARUs were attached to a tree at a height of approximately 1.5m with the 177
microphones perpendiculars to the tree. The recordings were made during the dry seasons of 178
2010 and 2011. During the first year we sampled 151 points in 91 field days (Jun.-Oct.). For the 179
22
second year we sampled the same 151 points plus 71 sampling points in 73 consecutive days 180
(Jun.-Aug.; Figure 1). Sampling order was done based on campsites that were further apart 181
avoiding sampling campsites next to each other to minimize spatial and temporal correlations 182
during the sampling period of both years. This was an effort to minimize correlation with 183
sampling order and east-west spatial gradient (Camargo 2011). 184
Each point was sampled for five up to seven consecutive days. Rempel et al. (2013) 185
investigating the efficiency of different brands of ARU´s concluded that the active sampling 186
radius of the SM2 is roughly over 100m for acoustic signals within the same frequency range 187
and amplitude as the Tinamidae. Thus, to minimize the odds of the same individual bird being 188
recorded simultaneously in two ARUs; the sampling points were separated by a distance of 189
400m. After concluding the field work all recordings were archived into a 1 Terabyte external 190
hard drive. Each of the three hours recordings was then divided into non-overlapping five 191
minutes tracks, which were then randomized and given a unique number that did not contain any 192
information about where or when the recording was made, this was done to not bias the 193
identification of the recorded species. Following this procedure, we sampled five of the five 194
minute tracks from each 3h recording for each day and sampling point, totaling over 8900 five 195
minute tracks. All recorded species in the five minute tracks were identified by a team of six 196
people using Raven Pro 1.4 (Charif et al. 2010) aided by a reference collection of vocalization 197
recordings from BDFFP (Naka et al. 2008). We also obtained occasional assistance from local 198
experts when acoustic signals could not be easily identified. After listening to all five minute 199
tracks they were rearranged in a matrix as to represent the true dawn chorus of every sampling 200
point. Recorded acoustic signals that were distinctively clear from background noise were treated 201
23
as high quality (HQrec) recording. The acoustic signals recording marked as HQrec were 202
archived separately and analyzed for the spectral parameters. 203
Following the species identification a spectrogram (Figure 3) for each of the HQrec was 204
made using Raven Pro 1.4 (Chariff et al. 2010) and further analyzed. In order to avoid simple 205
pseudo replication (Hurlbert 1984) of the spectral characteristics of acoustic signal analysis, only 206
one signal from each HQrec was used. Spectral measurements were extracted using the 207
following parameters: Hann window type, FFT window of 2048 with an overlap of 75% and hop 208
size of 512. For the extraction of spectral characteristics from the acoustic signals we adopted the 209
method from Zollinger et al. (2012). A species can be detected from a recording even when there 210
is a low Signal to Noise ratio (SNR), such as when its far enough to be recorded but not close 211
enough for the signal to be accurately measured (Zollinger et al. 2012, Rempel et al. 2013), thus 212
only calls which were clear above background noise were used. Frequency bandwidth was 213
calculated as the difference from the highest frequency and the lowest frequency after a -24dB 214
mark was set from the maximum power of each signal (Podos, 1997). 215
To test our hypothesis about temporal segregation of vocal activity we used a hierarchical 216
model which includes imperfect detection (Macknzie et al. 2002).In a hierarchical model the 217
biological components and sampling methods are related with each other, but are specified in 218
separate levels. To determine the peak of vocal activity of each species we included a linear and 219
quadratic function of time, measured in minutes before and after the sunrise, which enables us to 220
incorporate the probability of detection during each morning. The upper part of the model 221
(Figure 2) describes the sampling process (detection or not of each species) while the lower part 222
describes the biological state (species present or not) at each site. We treated each of the five 223
minute tracks of each sampling point as a unique visit and used each dawn chorus as temporal 224
24
replicates to estimate the detection probability of each species during the analyses (MacKanzie et 225
al. 2002). 226
Figure 2 shows a schematic of our model used to determine the peak of vocal activity for 227
each species. The i index designates the species while j represent the sampling point. Within the 228
biological part of the model, the latent variable (unobserved) Z represents the true occupancy, 229
such that zi=1 means that in this location an acoustic signal was detected for species i and zi=0 230
means no acoustic signals were detected. The values of zi are obtained with a Bernoulli 231
distribution with success probability of ψ. Therefore, ψ represents the occurrence probability at 232
dawn chorus k, which can be a value between 1 and 0. The parameter ψ is a logistical function of 233
the environment, described as logit (ψ) = α where α is the intercept. Because our emphasis is on 234
understanding the temporal patterns of vocal activity, our model does not include environmental 235
covariates. 236
In the sampling process, the indices j and k are used because the sampling conditions can 237
vary between locations (j) and dawn chorus (k). The product of µit = Zi *Pi,j,k is the detection 238
probability that has value 0 when zi=0 and pij when zi=1. The pij is a detection probability on the 239
site being occupied (acoustic signal was present), it is a logistical function which accounts for 240
the time of detection of each vocalization: logit (pit) = a1 + a2 * hourjk + a3 * hour2
jk; where a1 is 241
the intercept and a2 is a linear effect and a2 is a quadratic effect of time on vocal activity. The 242
linear and quadratic effects of time of detection were standardized for computational 243
convenience. Our observational data (µjk) were of species i at sampling point j during the dawn 244
chorus k, which equals µjk = 1 when the acoustic signal was detected and µijk = 0 when there 245
were no acoustic signal detection. 246
25
In a temporally segregated community species should be able to secure an exclusive 247
access to a particular resource, being it time or space and facilitate coexistence between species. 248
Most randomization models that are used to analyze resource partitioning are categorical in 249
nature (such as food or habitat) and a continuous resource asks a different analytical approach 250
(Winemiller and Pianka 1990). A null model approach can efficiently use the patterns of species 251
distribution to gain greater understanding of how species partition time. Time, being cyclical in 252
nature, requires random shifts of entire activity patters within a time frame, which encompasses 253
the species activities (Castro-Arellano et al. 2010). ROSARIO is a temporal overlap null model 254
that then can be compared with the empirical temporal overlap data (Castro–Arellano et al. 255
2010). ROSARIO algorithm method analyzes species (rows)-by-time (columns) matrices that are 256
populated with frequency occurrences (e.g. acoustic signal detections). The index is symmetrical, 257
approaching zero when species are completely segregated and one when completely overlapping. 258
The ROSARIO algorithm used to analyze overlapping temporal activities comes with the freely 259
available software TimeOverlap 260
(http://hydrodictyon.eeb.uconn.edu/people/willig/Research/activity%20pattern.html). In each 261
iteration the algorithm shifts the distribution of occurrences of each species and compares the 262
overlap with the randomly generated data, using 10.000 randomizations to create a null 263
distribution of overlap values for the Czechanowski index (Aguiar et al. 2012) 264
A Welch t-test was performed on the bandwidth of species pairs to determine if each 265
species bandwidth use of the acoustic space is significantly different from each other. Because of 266
the uneven number of HQrec from each species we selected the same number of HQrec from the 267
species that had the least number of HQrec (n=16 from Crypturellus brevirostris) 268
269
26
Results 270
We analyzed 747 hours of recordings (total of 8964 five-minute tracks) from the 2010 and 2011 271
field seasons. Despite this effort Crypturellus brevirostris had the lowest detection being 272
detected in only 22 occasions. Tinamus major and Crypturellus variegatus were detected in 174 273
and 704 occasions, respectively. There was a substantial concentration of vocalizations during 274
different time periods in the dawn chorus for Tinamus major and Crypturellus variegatus, while 275
C. brevirostris calls were detected more evenly throughout the mornings. The species that had 276
the earliest peak of vocal activity was T. major within the Tinamidae. A Czechanowski index of 277
temporal niche overlap was obtained for each species pair using the ROSARIO algorithm 278
Crypturellus brevirostris and Crypturellus variegatus had the lowest temporal overlap of all 279
species pairs (Czechanowski index = 0.04) and Tinamus major and Crypturellus variegatus had 280
the largest temporal overlap (Czechanowski index = 0.74). Tinamus major and Crypturellus 281
brevirostris obtained an intermediate value of temporal niche overlap (Czechanowski index = 282
0.17). A t-test was conducted with the Czechanowski index difference of each species pair 283
(Table 1) and it is calculated as the proportion of randomization that resulted in overlap that is 284
equal or less than the empirical (observed) temporal overlap value, when a priori expectations 285
are of segregated temporal activities (Castro–Arellano et al. 2010); therefore temporal overlap 286
was greater than expected by chance between all species pairs. 287
The models that best explained the detection (hence vocal activity) of all species were 288
models which represented as a quadratic function of time (Table 2). The quadratic effect on 289
detection was comparatively small in C.brevirostris where the detection probability remained 290
relatively constant throughout the morning. As expected all species pairs reported a significant 291
difference between the spectral characteristics of all species pairs. 292
27
Discussion 293
In this study we examined how three species of allopatric Tinamous species share the acoustic 294
space in the cyclical event of the dawn chorus. Acoustic interference is expected to be more 295
common in species that share acoustic signal characteristics, foraging strata and time 296
participating in the dawn chorus. All three species were vocally active throughout the dawn 297
chorus, albeit with slightly temporal shifts at time of peak signaling. Although the acoustic 298
characteristics were significantly different, similar species sharing ecological requirements might 299
also evolve other cues for co-specific identification(Cody 1969) and the similarity of the acoustic 300
characteristics of the signals converge to better transmit the signal though the environment 301
(Seddon et al. 2005). 302
. The vocal activity of Neotropical birds has been found to correlate with foraging height 303
and relative eye size. When enough sunlight shines through the forest canopy and reaches the 304
forest floor enabling for the species to start foraging the vocalizations cease or decrease 305
substantially (Berg et al. 2006; Ross and Hall, 2007). Berg et al. (2006) concluded that foraging 306
height and time of first song are positively correlated for non-passerines indicating that ground-307
dwelling species were vocally active earlier than other Neotropical birds. Our results are indirect 308
evidence of this, in which T.major being the largest of the tinamous in the Central Amazon and 309
also with the largest relative eye size (±1kg; Cabot 1992; Berg et al 2006) it does, in fact, have 310
an earlier onset of vocal activity than the other two smaller species 311
Active temporal avoidance has been shown in birds with acoustic signals with spectral 312
overlap, so that one species will avoid vocalizing whilst another species is vocalizing (Planqué 313
and Slabekoorn, 2008). Though C.brevirostris was the least detected species, it is part of the 314
28
acoustic community, and its low detection rate can conceivably cause a weak influence in the 315
vocal activity of the other two species. The two species with higher detection rates, T.major and 316
C.variegatus (Figure 5) had a period during the dawn chorus when T.major peaks in vocal 317
activity while C.variegatus shows a decrease in vocal activity, which suggests that T.major 318
might inhibit C.variegatus. Acoustic signal inhibition between species has been well documented 319
in interactions among anurans (Littlejohn and Martin 1969; Schwartz and Wells 1983, 1984), 320
insects and anurans (Paez et al. 1994), in intraspecific interactions of nightingales and corncrakes 321
(Brumm 2006; Ręk et al.2011), two owl species (Lourenço et al. 2013) and marine mammals 322
(Rankin et al. 2012). An experimental investigation with the combined use of signal playbacks 323
and ARUs may provide further insight in the interactions between T.major and C.variegatus. 324
Brooks and Pando-Vasquez. (2004) suggested that T.major and C.variegatus are species 325
that are vocally active during crepuscular and night time hours only, not being detected by 326
acoustic signals during the dawn chorus. Considering the nocturnal detections of these species 327
(Brooks and Pando-Vasquez 2004) and the present results, it is possible that these species might 328
present a bimodal distribution of vocal activity, having two distinct peaks one at the dawn chorus 329
and another during night time (J. Bonnanomi, unpub. data) . Therefore, partitioning the acoustic 330
space into two distinct periods of vocal activity; one nocturnal while roosting and during the 331
dawn chorus. Nocturnal acoustic signaling by diurnal species has been the focus of few studies 332
and all in temperate climate (La 2012), while most neotropical studies focus on the ecology of 333
nocturnal species such as owls (Sberze et al. 2012; Lourenço et al. 2013) Further investigation of 334
nocturnal acoustic signaling by diurnal neotropical birds is needed to explain this behavior. 335
Because of the significant temporal overlap in vocal activity for all species and non-336
significant spectral overlap our results further support the generality that species with little 337
29
frequency modulation in signals it is the frequency bandwidth that is more important for signal 338
detection and recognition of co-specifics (Vielliard 2004; Brandes 2008).It has also been 339
proposed that active temporal overlap of acoustic signals can be considered a form of aggression 340
among individuals trying to maintain a territory (Naguib and Mennill 2010). Since there have 341
been no studies on the territory size of forest dwelling Tinamidae, one can only speculate that the 342
temporal overlap in this study is a form aggression among individuals. 343
Our second hypothesis that birds with greater similarity of vocal characteristics would 344
peak in vocal activities further apart temporally was not conclusively supported with the dawn 345
chorus activity, as T.major and C.variegatus have greater overlap in frequencies (Figure 4) and 346
have greater temporal overlap. Both species did have slightly segregated peaks (Figure 5) 347
roughly 40 minutes apart. Acoustic niche hypothesis studies have historically concentrated in the 348
northern hemisphere (Wiley and Richards, 1977, Kroodsma et al. 1982) and more recently have 349
been done in the Neotropical (Planqué and Slabbekoorn 2007; Luther 2008; 2009); with several 350
species sharing similar frequencies during the same time period of the dawn chorus. Further 351
studies are still needed within the Tinamidae family including further investigations in natural 352
history and the influence of territoriality effects individuals distributed in the environment as 353
well as accounting for other environmental, morphological and physiological constraints which 354
can influence the acoustic signaling in this bird family. 355
356
Acknowledgements 357
We thank Ulisses Camargo, Marconi Cerqueira, Christian Andretti and Claudeir Vargas helping 358
with field work and identifying the species. We also thank BDFFP logistical team for their 359
30
support in both years. This research was only made possible by the scholarship provided by 360
CNPq for GM, FD and JB. This is contribution #### of the BDFFP technical series. 361
Literature Cited 362
Aguiar, C. M. L.; Santos, G. M.; Martins, C. F.; Presley, S. J. 2012, Trophic niche breadth and 363
niche overlap in a guild of flower-visiting bees in a Brazilian dry forest. Apidologie 44:153-364
162. 365
Baptista, L.F.; Schumann, K. 1990 Song learning in the anna hummingbird (Calypte anna). 366
Ethology, 84:15-26. 367
Bartelli, S.; Tubaro, P. L. 2002, Body mass and habitat correlates of song structure in a primitive 368
group of birds. Biological Journal of the Linnaean Society, 77:423-430 369
Bierregaard, R.O., Gascon, C., Lovejoy,T.E., Mesquita, R.C.G. (Editors). 2001, The biological 370
dynamics of forest fragments project the study site, experimental design, and research 371
activity: Lessons from Amazonia: the ecology and conservation of a fragmented forest 31–372
42. New Heaven, MA: Yale University Press. 373
Berg, K. S.; Brumfield, R. T.; Apanius, V.2006. Phylogenetic and ecological determinants of the 374
Neotropical dawn chorus. Proceedings of the Royal Society B: Biological Sciences, 375
1589:999-1005. 376
Brandes, S.T. 2008. Automated sound recording and analysis techniques for bird surveys and 377
conservation. Bird Conservation International, 18:163-173. 378
Brennan, P. L. R. 2004. Techniques for studying the behavioral ecology of forest-dwelling 379
tinamous ( tinamidae ). Ornitologia Neotropical 15(sup.):329–337. 380
Brooks, D.; Pando-Vasquez, L., 2004, Resource separation in a Napo-Amazonian tinamou 381
community. Ornitologia Neotropical, 15:323-328 382
Brumm, H. 2006, Signaling through acoustic windows: nightingales avoid interspecific 383
competition by short-term adjustment of song timing. Journal of Comparative Physiology 384
A, 192(12), 1279-1285. 385
Cabot, J. 1992 Family Tinamidae (Tinamous) In: Elliott, A.; del Hoyo; Sargatal, J. (Ed.) 386
Handbook of the birds of the world, v.1. Lynx Edicions, Spain p.696. 387
31
Camargo, U.M. 2012. O uso de clareiras por aves da Amazônia central: Uma abordagem 388
quantitativa considerando a detecção imperfeita. Dissertação de Mestrado, Instituto 389
Nacional de Pesquisas da Amazônia, Manaus, Amazonas. 42p. 390
Castro-Arellano, I.; Lacher, T.E.; Willig, M.R. Rangel, T. 2010, Assessment of assemblage-391
wide temporal niche segregation using null models. Methods in Ecology and Evolution, 392
1:311-318 393
Cerqueira, M.C., Cohn-Haft, M., Vargas, C.F., Nader, C.E., Andretti, C.B., Costa, T.V.V., 394
Sberze, M., Hynes, J.E. & Ferraz, G. (2013) Rare or elusive? A test of expert knowledge 395
about rarity of Amazon forest birds. Diversity and Distributions, 17:710-720. 396
Charif, R.A; Waack,A.M., Strickman, L.M. 2010, Raven Pro 1.4 User’s Manual. Cornell Lab of 397
Ornithology, Ithaca, NY. 398
Cody, M. L. 1974. Competition and the structure of bird communities, V.7. Princeton University 399
Press. Princeton, New Jersey p.326 400
Cohn-Haft, M.; Whitaker, A.; Stouffer,P.C. 1997, A New Look at the “Species-poor” Central 401
Amazon: The Avifauna North of Manaus, Brazil. Ornithological Monographs, 48:205-235 402
De Kort, S. R.; Den Hartog, P. M.; Ten Cate, C. 2002, Diverge or merge? The effect of 403
sympatric occurrence on the territorial vocalizations of the vinaceous dove Streptopelia 404
vinacea and the ring‐necked dove S. capicola. Journal of avian biology, 33(2), 150-158. 405
Gogala, M.; Riede, K. L. A. U. S. 1995. Time sharing of song activity by cicadas in Temengor 406
Forest Reserve, Hulu Perak, and in Sabah, Malaysia. Malayan Nature Journal, 48(3-4), 407
297-305. 408
Henwood, K.; Fabrick, A. 1979, A quantitative analysis of the dawn chorus: temporal selection 409
for communicatory optimization. American Naturalist 114:260–274 410
Hoi, H.; Sageder, G. 1991, Parental investment and territorial/sexual behavior in male and 411
female reed warblers: are they mutually exclusive?. Ethology, 88(3), 249-255. 412
Hurlbert, S. H. 1984. Pseudoreplication and the design of ecological field experiments. 413
Ecological monographs, 54:187-211. 414
Krause, B.L. 1987. The niche hypothesis: How animals taught us to dance and sing. Whole Earth 415
Review, 57:1-6 416
La, V. T. 2012. Diurnal and nocturnal birds vocalize at night: a review. The Condor, 114:245-417
257. 418
32
Littlejohn, M. J.; Martin, A. A. 1969. Acoustic interaction between two species of leptodactylid 419
frogs. Animal Behaviour, 17:785-791. 420
Lourenço, R.; Goytre, F.; del Mar Delgado, M.; Thornton, M.; Rabaça, J. E.; Penteriani, V. 2013. 421
Tawny owl vocal activity is constrained by predation risk. Journal of Avian Biology, 422
44:461-468 423
Luther, D. 2008. Signaler: receiver coordination and the timing of communication in Amazonian 424
birds. Biology letters, 6:651–654 425
Luther, D. 2009. The influence of the acoustic community on songs of birds in a neotropical rain 426
forest. Behavioral Ecology, 20:864-871 427
Marler, P. 1960 Bird songs and mate selection. In: Lanyon, W.E.; Tovolga, W.N. (ed.) Animal 428
sounds and communication, vol. 7. American Institute of Biological Sciences. p. 348-367 429
Washington D.C., USA 430
Marler, P. R.; Slabbekoorn, H. 2004. Nature's music: the science of birdsong. Academic Press. 431
Waltham, Massachusetts, p.513 432
Mathevon, N.; Aubin, T.; Dabelsteen, T.; Vielliar, J. M. 2004. Are communication activities 433
shaped by environmental constraints in reverberating and absorbing forest habitats?. Anais 434
da Academia Brasileira de Ciências,76:259-263 435
Mathevon, N.; Aubin, T.; Vielliard, J.M.E.; Silva, M.L. 2008. Singing in the rain forest: How a 436
tropical bird song transfers information. Plos one, 3:2-7 437
McCracken K.G; Sheldon F.H. 1997 Avian vocalizations and phylogenetic signal. Proceedings 438
of the National Academy of Sciences , 94:3833–3836 439
McNamara, J. M.; Mace, R. H.; Houston, A. I. 1987. Optimal daily routines of singing and 440
foraging. Behavioral Ecology and Sociobiology, 20:399–405 441
Mesquita, R. C. G.; Ganade,G.; Ickes,K.; Williamson,B.; 2001 Alternative successional 442
pathways in the Amazon basin. Journal of Ecology, 89:1-10 443
Morton, E. S. 1975 Ecological sources of selection on avian sounds. American Naturalist, 444
109:17–34 445
Naka, L.N.; Stouffer, P.C.; Cohn-Haft, M.; Marantz, C.A.; Whittaker, A.; Bierregard, R O. 2008 446
Voices of the Brazilian Amazon, Vol 1. Birds of the terra firme forests north of Manaus: 447
Guianan area of endemism. 4 Audio Compact Discs 448
33
Nelson, D. A.; Marler, P. 1990 The perception of bird song and an ecological concept of signal 449
space. In Stebbins, W.C.; Berkley, M.A. (eds.) Comparative perception: complex signals, 450
v. 2. p.443–478. Wiley New York, NY 451
Pepperberg, I. M. 2007. Emergence of linguistic communication: Studies on Grey parrots. In: 452
Lyon, C.; Nehaniv, C. L.; Cangelosi, A. (Eds.). Emergence of communication and language. 453
Emergence of communication and language, pp. 355-386). Springer, London, UK. 454
Pianka, E. R. 1973. The structure of lizard communities. Annual Review of Ecology and 455
Systematics, 4:53-74. 456
Podos, J.1997. A performance constraint on the evolution of trilled vocalizations in a songbird 457
family (Passeriformes: Emberizidae). Evolution, 51:537- 551. 458
Podos, J. 2001. Correlated evolution of morphology and vocal signal structure in Darwin's 459
finches. Nature, 6817:185-188. 460
Planqué, R.; Slabbekoorn, H. 2008 Spectral Overlap in Songs and Temporal Avoidance in a 461
Peruvian Bird Assemblage. Ethology, 114:262-271 462
Platzen, D.; Magrath, R. D. 2004 Parental alarm calls suppress nestling vocalization. 463
Proceedings of The Royal Society Biological sciences, (271):1271-1276 464
Rankin, S.; Archer, F.; Barlow, J. 2012. Vocal activity of tropical dolphins is inhibited by the 465
presence of killer whales, Orcinus orca. Marine Mammal Science, 29:679-690 466
Reis De Magalhães, J.C 1996 Sobre alguns tinamídeos florestais brasileiros. Boletim do Centro 467
de Estudos Ornitológicos, 10h16min-26 468
Ręk, P.; Osiejuk, T. S.; Budka, M. 2011. Functionally similar acoustic signals in the corncrake 469
Crex crex transmit information about different states of the sender during aggressive 470
interactions. Hormones and behavior,60:706-712. 471
Rempel, R. S.; Francis, C. M.; Robinson, J. N.; Campbell, M. 2013 Comparison of audio 472
recording system performance for detecting and monitoring songbirds. Journal of Field 473
Ornithology, 84:86–97 474
Riede, K. 1997. Bioacoustic monitoring of insect communities in a Bornean rainforest 475
canopy.In: Stork, N.E.; Adis, J.A.; Didham, R.K.(eds.) Canopy Arthropods, p442-452, 476
Chapman and Hall, London, United Kingdom 477
Schmidt, A. K.; Römer, H.; Riede, K. 2013. Spectral niche segregation and community 478
organization in a tropical cricket assemblage. Behavioral Ecology, 24:470-480 479
34
Schwartz, J. J.; Wells, K. D. 1983. An experimental study of acoustic interference between two 480
species of Neotropical tree frogs. Animal Behaviour, 31:181-190. 481
Schwartz, J. J.; Wells, K. D. 1984. Vocal behavior of the Neotropical tree frog Hyla 482
phlebodes. Herpetologica, 40:452-463. 483
Sinsch, U.; Lümkemann, K.; Rosar, K.; Dehling, J. M. 2012. Acoustic niche partitioning in an 484
anuran community inhabiting an Afromontane wetland (Butare, Rwanda). African Zoology, 485
47:60-73. 486
Slabbekoorn, H.; Ellers, J.; Smith, T. 2002 Birdsong and sound transmission: the benefits of 487
reverberations. The Condor, 104:564-573 488
Tobias, J. A.; Aben, J.; Brumfield, R. T.; Derryberry, E. P.; Halfwerk, W.; Slabbekoorn, H.; 489
Seddon, N. 2010. Song divergence by sensory drive in Amazonian 490
birds. Evolution, 64:2820-2839. 491
Tobias, J. A.; Seddon, N. 2009. Signal design and perception in Hypocnemis antbirds: evidence 492
for convergent evolution via social selection. Evolution, 63:3168-89 493
Vielliard, J.M.E.1987. Uso da bioacústica na observação de aves In: Coelho, E.P. (Ed.) II 494
Encaminhamentos Nacional dos Anilhadores de Aves, UFRJ, Rio de Janeiro p.98-121. 495
Vielliard, J. M. 2004. A diversidade de sinais e sistemas de comunicação sonora na fauna 496
brasileira. Seminário Música Ciência Tecnologia, 1(1). 497
Weckstein, J. D. 2005. Molecular phylogenetics of the Ramphastos toucans: implications for the 498
evolution of morphology, vocalizations, and coloration. The Auk, 122:1191-1209. 499
Wildlife Acoustics, 2009. Wildlife Acoustics, Inc.—Bioacoustics Software and Field Recording 500
Equipment. http://www.wildlifeacoustics.com/ (accessed 01.04.13.) 501
Wiley, R. H. 1994 Errors, exaggeration, and deception in animal communication. In: Real, L.A. 502
(ed.) Behavioral mechanisms in evolutionary Ecology, p.157–189. University of Chicago 503
Press, Chicago, Illinois 504
Wiley, R. H.; Richards, D. G. 1978 Physical constraints on acoustic communication in the 505
atmosphere: Implications for the evolution of animal vocalizations. Behavioral Ecology and 506
Sociobiology, 3:69-94. 507
Wiley, R.H; Richards, D.G. 1982. Adaptations for acoustic communication in birds: sound 508
transmission and signal detection In: Kroodsma, D.E.; Miller, H.E. Acoustic 509
Communication in birds,, pp. 131-181. Academic Press. New York,USA 510
35
Winemiller, K. O.; Pianka, E. R. 1990. Organization in natural assemblages of desert lizards and 511
tropical fishes. Ecological Monographs, 60:27-55. 512
Zollinger, S. A.; Podos, J.; Nemeth, E.; Goller, F.; Brumm, H. 2012. On the relationship 513
between, and measurement of, amplitude and frequency in birdsong. Animal Behaviour. 514
84:1-9 515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
36
Tables 530
Table 1 Results (p-values) of the t-test using the Czechanowski index value for temporal 531 activity pattern between the Tinamidae birds at BDPFF 532
T.major C.variegatus C. brevirostris
Tinamus major - 0.684 0.345
Crypturellus variegatus - ˗ 0.435
Crypturellus brevirostris - - -
533
534 Table 2. . Summary of best model selection(∆AIC < 2) obtained by fitting all models in the a 535 priori model set to data from each of the three species of Tinamidae. The table shows models 536
applied for each species. Bold values are models with ∆AIC < 2. Columns to the right of Model 537 show twice the negative log-likelihood, the number of covariates (k) and AIC model weight 538 (ω).STHR and STHR
2 are the linear and quadratic effect of time of detection. 539
Species
Model -2Log (L) K ∆ AIC ω
T. major
Ψ(.) p(STHR + STHR²) 1461.75 4 0.00 0.7161
Ψ(.) p(STHR) 1465.60 3 1.85 0.2839
Ψ(.) p(.) 1664.74 2 189.99 0
C. variegatus Ψ(.) p(STHR + STHR²) 4566.39 4 0.00 1.00
Ψ(.) p(STHR) 1553.57 3 7.91 0.0118
Ψ(.) p(.) 1720.95 2 173.29 0
C. brevirostris Ψ(.) p(STHR + STHR²) 403,79 4 0.00 0,5089
Ψ(.) p(STHR) 406.02 3 0.28 0.4424
Ψ(.) p(.) 412.48 2 4.69 0.048
540
541
542 543 544
545
546
547
37
Figures548
549
Figure 1. Map of BDFFP located 80km north of Manaus in the Brazilian Amazon. The sampling points marked by the black dots were 550 sampled during the 2010 and 2011 dry seasons, the sampling point which were sampled only in 2011 are represented by the white circles. 551
The continuous forest is the dark gray area while the second growth is the lighter shade of gray. The access road (dotted line) and BR-174 552
highway (continuous line) are also shown. 553
554
38
555
yi,t | µi,jk ~ Bern (µi,kj) 556
557
558
Sampling µit = Zi,j * pi,j,k 559
560
logit (pi,j,k) = a1 + a2i * hourj,k + a3i * hour2
j,k 561
562
Zi,j|Ψi ~ Bern (Ψi) 563
Biological 564
565
logit (Ψi) = α 566
Figure 2. The model used to estimate the acoustic signal detections in the dawn chorus. The 567 bottom bracket shows the biological process of detection (Ψ) at site j and species i; the Zi.j which 568 is the latent variable indicates the true detection of an acoustic signal. The top bracket presents 569
Pi,j which is the conditional detection probability of species i at sampling point j at dawn chorus 570 k; the variables hourj,k and hour
2j,k are the linear and quadratic effects of time of detection in the 571
detection probability; yi,j is the observational component (data) 572
573
574
575
576
577
39
578
579
580
581
Tinamus major
Crypturellus variegatus
C. brevirostris
Figure 3 Waveforms and spectrograms of the Tinamidae acoustic signals recorded with
the SM2 ARUs. Spectrograms generated with Hann window and FFT size of 2048. Both
T.major and C.variegatus have considerably shorter acoustic signals than C.brevirostris.
The bottom figure represents a power spectrum of the Tinamus major call, notice that the
frequency is now in the X axis and the Y axis is the amplitude. The frequency bandwidth
is measured as the width of the selection box and the peak frequency being represented
highest point in the signal. The maximum power (horizontal line) is the guideline to
which we subtracted 24dB (bracket) to measure the signal above the background noise.
- 24dB
40
582
Figure 4. Beanplots of the peak frequency and frequency bandwidth utilized by each species. 583
The shape of the plot shows the distribution of the data for each species. The horizontal lines 584
indicate the median. 585
41
Figure 5. The top graphs show the linear and quadratic effect on the detection
probability of each species before and after sunrise. The bottom figures are
density histograms of vocalizations for each species. The blue density
histogram shows the sampling effort for each five minute tracks recorded with
the SM2 ARU´s
42
CONCLUSÕES
Todas as espécies estudas apresentaram algum grau de sobreposição temporal ao logo do coro
matinal na Amazônia central. As espécies T.major e C.variegatus apresentaram a maior
sobreposição temporal ao longo do coro matinal, enquanto a C.variegatus e C.brevirostris
tiveram a menor sobreposição temporal. Os sinais das três espécies são significativamente
diferentes entre si, corroborando a hipótese que a frequência é o componente mais importante
para o reconhecimento intraespecífico em espécies com sinais acústicos com frequências de
pouca modulação. Ao comparar nossos dados obtidos em campo com as informações
encontradas na literatura deparamos com informações conflitantes que argumentam que estas
espécies possuem atividade vocal restrito ao período noturno Podemos concluir que o nicho
acústico destas espécies não é restrito somente ao coro matinal e que elas provavelmente
apresentam uma distribuição bimodal de atividade vocal. No entanto, nossa suposição a priori de
que elas seriam temporalmente segregadas durante o coro matinal não foi corroborada. Porém,
cada espécie possui um pico de atividade vocal, o que é evidência que nosso método é eficaz
para este tipo de estudo sobre a atividade vocal em comunidades de aves, tanto na Amazônia
como em outros ambientes. Encontramos apoio para a hipótese do nicho acústico na questão de
que as espécies são segregadas espectralmente, indicando que há uma pressão seletiva para que
as vocalizações sejam divergentes a fim de evitar a sobreposição espectral. Futuros estudos
devem investigar como outros fatores morfológicos, ecológicos e comportamentais podem
influenciar a atividade vocal das aves. A investigação sobre a territorialidade destas espécies
também se faz necessária já que não foram feitos estudos sobre a influência de indivíduos
vizinhos sobre o comportamento vocal entre espécies quando há potencialmente sobreposição de
territórios.
43
ANEXOS*
* Pareceres das bancas examinadoras da aula de qualificação, da versão escrita e da defesa oral,
respectivamente.
44
45
46
47
48
Recommended