UNIVERSIDADE FEDERAL DE MINAS GERAIS Instituto de Ciências Biológicas Programa de Pós-Graduação em Ecologia, Conservação e Manejo da Vida Silvestre TESE DE DOUTORADO PADRÕES E MECANISMOS ESTRUTURADORES DA DIVERSIDADE TAXONÔMICA E FUNCIONAL DE COMUNIDADES DE FORMIGAS NA CADEIA DO ESPINHAÇO FLÁVIO SIQUEIRA DE CASTRO BELO HORIZONTE 2019 Flávio Siqueira de Castro PADRÕES E MECANISMOS ESTRUTURADORES DA DIVERSIDADE TAXONÔMICA E FUNCIONAL DE COMUNIDADES DE FORMIGAS NA CADEIA DO ESPINHAÇO Orientador: Dr. Frederico de Siqueira Neves Co-orientadores: Dr. Ricardo Ribeiro de Castro Solar e Dr. Pedro Giovâni da Silva BELO HORIZONTE 2019 043 Castro, Flávio Siqueira de. Padrões e mecanismos estruturadores da diversidade taxonômica e funcional de comunidades de formigas na Cadeia do Espinhaço [manuscrito] / Flávio Siqueira de Castro. – 2019. 159 f. : il. ; 29,5 cm. Orientador: Dr. Frederico de Siqueira Neves. Co-orientadores: Dr. Ricardo Ribeiro de Castro Solar e Dr. Pedro Giovâni da Silva. Tese (doutorado) – Universidade Federal de Minas Gerais, Instituto de Ciências Biológicas. Programa de Pós-Graduação em Ecologia, Conservação e Manejo da Vida Silvestre. 1. Ecologia. 2. Formigas. 3. Biodiversidade. 4. Campo rupestre. 5. Serra do Cipó. I. Neves, Frederico de Siqueira. II. Solar, Ricardo Ribeiro de Castro. III. Silva, Pedro Giovâni da. IV. Universidade Federal de Minas Gerais. Instituto de Ciências Biológicas. V. Título. Ficha e laborada pela Biblioteca do Instituto de Ciências Biológias da UFMG CDU: 502.7 Ficha elaborada pela Biblioteca do Instituto de Ciências Biológias da UFMG Ficha catalográfica elaborada por Fabiane Cristielle M. Reis – CRB: 6/2680 Sumário Agradecimentos ............................................................................................................................................... 8 Resumo ........................................................................................................................................................... 12 Abstract ......................................................................................................................................................... 13 Introdução Geral........................................................................................................................................... 14 Referências bibliográficas ............................................................................................................................ 20 Capítulo 1: Environmental drivers of taxonomic and functional diversity of ant communities in a tropical mountain .......................................................................................................................................... 30 Abstract ......................................................................................................................................................... 33 Introduction ................................................................................................................................................... 34 Material and Methods .................................................................................................................................. 40 Study area ................................................................................................................................................... 40 Sampling of ants ......................................................................................................................................... 42 Identification of ant species and description of functional traits ................................................................ 43 Taxonomic and functional diversity ........................................................................................................... 47 Environmental variables ............................................................................................................................. 48 Data analysis ............................................................................................................................................... 49 Results ............................................................................................................................................................ 51 Partition of taxonomic and functional diversity .......................................................................................... 53 Effect of elevation and environmental variables on taxonomic and functional diversity ........................... 55 Discussion ...................................................................................................................................................... 62 Acknowledgments ......................................................................................................................................... 66 References ...................................................................................................................................................... 67 Supporting information ................................................................................................................................ 80 Capítulo 2: Snow-free mountaintops are dominated by tiny and dark ants ........................................... 97 Abstract ....................................................................................................................................................... 100 Introduction ................................................................................................................................................. 101 Material and Methods ................................................................................................................................ 108 Study sites ................................................................................................................................................. 108 Sampling design ........................................................................................................................................ 110 Identification of species ............................................................................................................................ 111 Definition of functional traits ................................................................................................................... 111 Macrohabitats variables ............................................................................................................................ 115 Functional diversity metrics ...................................................................................................................... 115 Data analysis ............................................................................................................................................. 117 Results .......................................................................................................................................................... 119 Taxonomic structure ................................................................................................................................. 119 Functional structure (FS) .......................................................................................................................... 120 Effects of environmental filters on ant’s traits - Testing three hypotheses of cuticle colour and body size on elevational and latitudinal gradients ......................................................................................................... 126 Discussion .................................................................................................................................................... 134 Acknowledgments ....................................................................................................................................... 139 References .................................................................................................................................................... 140 Appendix 1 - Supplementary Information ............................................................................................... 154 Conclusão Geral .......................................................................................................................................... 172 Agradecimentos Dedico o trabalho ao meu filho Lucas e minha família Siqueira de Castro. Ao meus pais José Carlos e Stellita, irmãos Rodrigo, Marcelo e Renata agradeço o amor e a parceria eterna. Às minhas cunhadas Lara e Taís, agradeço as horas de descontração em família e por me proporcionarem, junto com meus irmãos, o convívio com meus sobrinhos Lívia, Cauê e Lorena. Agradeço aos avós (in memoriam) e a todos Laender de Castro e Antunes de Siqueira por todos os momentos vividos! Agradeço especialmente tia Beth, que considero uma “mãedrinha”, e tio Claúdio, grande físico e pesquisador, o cara que sempre me incentivou a não desistir da ciência e, principalmente, do doutoramento. Agradeço também aos meus amigos irmãos do Véritas, Gim, Cari, Leandro e Nepá! A todos do Churras Do Véritas e aos filhos da PUC Minas, Queroz, Geraldo, Dorinha, Dudu, Gabriel, Regis, e a todos integrantes do Pé de Cedro! Agradeço ao GSG (Grupo de discussão de Ecologia) e seus integrantes, por todas as discussões e artigos compartilhados durante o doutorado! Agradeço ao meu orientador e amigo Dr. Frederico de Siqueira Neves. Obrigado pela amizade e por me ensinar, aprendi muito nesses quatro anos! E obrigado por sempre insistir para eu voltar para as formigas! Agradeço especialmente o Dr. Lucas Neves Perillo (repetindo suas palavras), “o sócio de tese”. Cheguei com o projeto em andamento e assumi a responsabilidade pelas comunidades de formigas. Foram mais de 50 dias de campo, 5 picos, cachaça, caminhadas, chuva e Aculeata pelas montanhas do Espinhaço, sempre subindo pra direção norte seguindo o Uno verde ou dentro dele. Valeu demais! Aos coorientadores: ao Dr. Ricardo Solar, Bob, pela amizade, discussões e contribuições para a tese; ao Dr. Pedro Giovâni da Silva pela paciência, conversas, análises estatísticas, sugestões de literatura, escrita (no andamento e fechamento da tese) e pelo constante apoio e amizade. Agradeço ao povo das cidades por onde passamos e aconteceram os projetos PELD e Espinhaço: as mineiras Belo Horizonte, Ouro Branco, Ouro Preto, Lavras Novas, Barão de Cocais, Catas Altas, Mariana, Santa Bárbara, Brumal, Cardeal Mota, Jaboticatubas, Itambé do Mato Dentro, Santana do Riacho, Lapinha, Serro, Santo Antônio do Itambé, Capivari, Diamantina, São Gonçalo do Rio Preto, Botumirim, Grão Mogol, Monte Azul, Formosa, Montezuma, Mamonas, Gameleiras, Espinosa, e as baianas Caetité, Abaíra, Catolés, Catolés de Cima, Rio de Contas, Brumadinho, Palmeiras, Capão, Mucugê, Pati, Ruínha, Brejo de Cima, Paiol, Jussiape e Guiné. Em todos esses lugares vivem pessoas do Espinhaço, o povo das montanhas. São eles que tornam cada uma dessas localidades únicas, além das paisagens e rica biodiversidade. No projeto PELD (Capítulo 1), agradeço a Cedro Têxtil, ICMBio-PARNA Serra do Cipó, Pousada Serra Morena e Pouso do Elefante pelo suporte e logística. Agradeço ao pessoal da Reserva Vellozia, especialmente ao professor Dr. GW Fernandes, coordenador do projeto PELD e coautor no primeiro capítulo. Obrigado pela hospitalidade, conversas e contribuições para o trabalho. Agradeço aos amigos que sempre ajudaram em campo ou laboratório, especialmente Rayana Mello, Matheus Couto, Humberto Brant e Marina Catão. Agradeço ao Flávio Camarota e Scott Powel (Cephalotes), Alexandre Ferreira (Pheidole), Rodolfo Probst (Camponotus, Myrmelachista e Octostruma) e ao Mayron Escárraga (Dolichoderinae) pelas identificações das formigas. Agradeço ao Flávio Camarota, Lucas Perillo e Rafael Leitão por todos os valiosos comentários no manuscrito. No projeto Espinhaço (Capítulo 2) agradeço algumas das pessoas que de alguma forma contribuíram no trabalho e para que ele acontecesse: Rodrigo Nescau e Laurinha (Montes Claros). Plínio, Vinicius, Lucas e Luiz (Porteirinha, P.E. Serra Nova). Ao Seo Zé Camilo (Alitôta), João (Dão), Marcos (Prego) e José Custódio Jorge, Alessandre Custódio Jorge, Gandú de Capivara, Chico de Mamonas no Pico da Formosa. À do Zé do Pilão e Dona Maria, Sinvaldo e Marli em Brumadinho (Pico das Almas). À Catarina e Ângela em Rio de Contas. Edmundo, Janho Barbosa de Azevedo, Sr. Antônio e Sr. Melquíades em Catolés (Pico do Barbado) e ao Dê em Guiné (Chapada Diamantina). Agradeço aos amigos: Matteus Carvalho, que, depois de mim (além do Dr. Perillo e Dr. Fred Neves), foi o maior coletor do projeto (Bolo Doido foi em 4 picos!), Heron Hilário, Ivan Monteiro, Caio Marques, Arleu Viana, João Pedro (Jota), Daniela Melo, Luisa Azevedo, André Araújo, Núbia Campos, Frederico Neves, Humberto Brant, Rayana Melo e Jéssica Martins. Muito obrigado pela parceria! Agradeço aos estagiários e integrantes do Laboratório de Ecologia de Insetos LEI (de 2013 aos dias atuais). Sem eles seria impossível processar tanto material: Caio Silveira, Thaís Tavares, Lucas Freitas, Lorenzzo Monteran, Bruna Vaz, Franklin Logan, Júlia Toffalini, Matheus Galvão, Isabella Villani, Laura Braga, Isa Mariah, Bruna Boa Sorte, Mellina Galantini, Maria Eliza Nogueira, Mariana Côrtes, Matheus Belchior, Poliana Gomes, Paola Mitraud, Daniel Vieira, Natalia Santoro, Ana Profeta, Ícaro, Catarina Dias e Bernardo (Museo). Aos amigos e parceiros de laboratório e projetos do lab e da UFMG Luiz Eduardo, Ludmila Hufnagel, Marina Beirão, Paloma Marques, Rodrigo Massara, Tadeu Guerra, Newton Barbosa, Ju Kuchenbecker, Matheus Couto (Jão), Amanda Dias, Fernando Pinho, Paulin, Leo Dias, Arleu Viana, Marcela, Ana Luiza, Vanessa Monteiro e Rayana Mello e à todos os outros integrantes da Villa Parentoni, da Varandinha e Grelo no Asfalto. Agradeço a todos os professores e colegas da ECMVS. Muitos aprendizados e excelente convívio. Obrigado Cris e Fred, secretaria do PPG-ECMVS, por tudo! Agradeço ao Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) por financiar o projeto PELD-Campo Rupetre, Serra do Cipó, e a Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG) por financiar o projeto Espinhaço. Agradeço também a CAPES pela bolsa de doutorado. Agradeço também minha banca de qualificação (Capítulo 1), composta pelos Dr. Milton Barbosa, Dr. Rafael Leitão e Dr. Flávio Camarota, além de Dr. Lucas Perillo como suplente. E, finalizando, agradeço aos membros da banca examinadora, composta pelos professores: Dr. Fernando Silveira, Dr. Danilo Neves, Dr. Ricardo Campo e Dr. Lucas Paolucci, além dos suplentes Dr. Flávio Camarota e Dr. Milton Barbosa. Resumo Montanhas são modelos ideais para o estudo dos padrões e entendimento dos processos que determinam a distribuição das espécies no espaço-tempo. Apresentam condições ambientais extremas para a distribuição da biodiversidade, com comunidades restritas e alta diversidade taxonômica e funcional de espécies. Nesse sentido, utilizo as formigas encontradas nas montanhas da Cadeia do Espinhaço como objetos de estudo. A tese está dividida em dois capítulos. Tem como objetivo determinar os padrões taxonômicos e funcionais e os mecanismos estruturadores das comunidades de Formicidae em gradientes espaço-temporais ao longo da Cadeia do Espinhaço, em diferentes escalas espaciais. Além disso, investigar os aspectos biogeográficos dessas comunidades em campo rupestre e avaliar como comunidades de formigas respondem às diferentes variáveis ambientais. No primeiro capítulo, investigamos padrões de diversidade taxonômica e funcional (diversidades α e β) de formigas em uma paisagem montanhosa da Cadeia do Espinhaço (Serra do Cipó) e os mecanismos associados a esses padrões em diferentes dimensões espaço-temporais. No segundo capítulo, avaliamos os padrões de diversidade funcional das comunidades de formigas e atributos individuais das espécies (cor e tamanho do corpo) em uma extensiva amostragem no campo rupestre ao longo de 12 montanhas em diferentes elevações na Cadeia do Espinhaço. Verificamos quais são os efeitos das variáveis ambientais na estrutura funcional da diversidade das comunidades de formigas e em atributos individuais das espécies em montanhas antigas. Foram testadas três hipóteses macroecológicas associadas à variação clinal de cores do tegumento para verificar o papel da variação da cor do tegumento e do tamanho do corpo em um gradiente geoclimático tropical de elevação e latitude. Descobrimos que a variação da elevação e os efeitos das variáveis geoclimáticos do gradiente de elevação são mais importantes na estruturação das diversidades taxonômica e funcional de formigas do que a variação latitudinal e os efeitos de suas variáveis geoclimáticas. Palavras chaves: Diversidade; Atributos Funcionais; Campo Rupestre; Formigas; Cadeia do espinhaço Abstract Mountains are considered ideal models for the study of patterns and understanding of the processes that determine species distribution in space-time., Exhibit extreme environmental conditions for biodiversity distribution, presenting restricted communities and high taxonomic and functional diversity of species. In this sense, I use the ants found in the mountains of the Espinhaço Range as objects of study. The thesis is divided into two chapters. The goal is to determine the taxonomic and functional patterns and structuring mechanisms of the Formicidae communities in spatiotemporal gradients along the Espinhaço Range, at different spatial scales. In addition, to elucidate biogeographic aspects of these communities in rupestrian field and evaluate how ant communities respond to different environmental variables. In the first chapter, we investigated patterns of taxonomic and functional diversity (α and β diversity) of ants in a mountainous landscape of the Espinhaço Range (Serra do Cipó) and the mechanisms associated with these patterns in different spatio-temporal dimensions. In the second chapter, we evaluated patterns of functional diversity of ant communities and individual species attributes (colour and body size) in an extensive rupestrian field sampling over 12 mountains at different elevations in the Espinhaço Range. We verified the effects of environmental variables on the functional structure of ant community diversity and individual species attributes in ancient mountains. Three macroecological hypotheses associated with clinal tegument colour variation were tested to verify the role of tegument colour and body size variation in a tropical geoclimatic gradient of elevation and latitude. We found that elevation variation and the effects of elevation gradient geoclimatic variables are more important in structuring the taxonomic and functional diversity of ants than latitudinal variation and the effects of their geoclimatic variables. Keywords: Diversity;Traits; Campo Rupestre; Formigas; Cadeia do espinhaço 14 Introdução Geral O entendimento e determinação dos padrões de diversidade, além dos processos envolvidos na estruturação das comunidades, estão entre as principais questões a serem respondidas na ecologia. A utilização de múltiplas escalas e métricas de biodiversidade em resposta aos processos ecológicos é fundamental para auxiliar na resolução dessas questões (Anderson et al., 2011; Barton et al., 2013; Tuomisto, 2010). A utilização das múltiplas escalas de diversidade proposta por Whittaker (1960) se mostra uma abordagem bem adequada para nos ajudar a elucidá-las, focando em padrões de diversidade β (uma medida da variabilidade da composição de espécies entre as amostras), evidenciando a dissimilaridade entre comunidades, ao relacionar tanto a diversidade de espécies na escala local (α) como a diversidade em larga escala (γ). Baselga (2010) trouxe uma abordagem mais focada em componentes da diversidade β (substituição e aninhamento), que pode ser aplicada tanto para a diversidade taxonômica (TD), representada pela riqueza e abundância das espécies, ou funcional (FD), relacionada às funções ecológicas de cada espécie (Baselga, 2013; Jost, 2007; Petchey & Gaston, 2006; Violle et al., 2007). Quando definimos os padrões de diversidade taxonômica consideramos que todas as espécies são diferentes umas das outras, mas desconsideramos que cada uma tem sua função ecológica (Villéger, Grenouillet, & Brosse, 2013). Para a diversidade funcional, os padrões refletem a variação dos atributos funcionais ou atributos funcionais (“traits”) entre as espécies da comunidade (Petchey & Gaston, 2006; Violle et al., 2007). Atributos funcionais são características biológicas mensuráveis dos organismos em nível individual e que tem influência ou relação direta em suas performances (“fitness”) e funções ecológicas (Violle et al., 2007). Analisando os padrões de diversidade com diferentes abordagens (e.g., TD e FD) podemos identificar os processos associados às origens e manutenção da biodiversidade e dos serviços prestados pelas diferentes comunidades ecológicas (Anderson et al., 2011; Barton et al., 2013). Dado o cenário atual de mudanças globais, elucidar os mecanismos que direcionam os padrões de distribuição da biodiversidade é essencial para orientar, 15 por exemplo, futuras ações efetivas de conservação dos ambientes naturais (Cooke et al., 2013; Lassau & Hochuli, 2004). Gradientes ambientais, como gradientes latitudinais e de elevação, são excelentes modelos para entendermos os mecanismos envolvidos na estruturação da biodiversidade (Colwell et al., 2008; Gaston, 2000; Janzen, 1967). No gradiente latitudinal, por exemplo, a riqueza de espécies em geral aumenta em baixas latitudes (Gaston, 2000), além de apresentarem variações nos padrões de diversidade funcional (Stevens et al., 2003; Villéger et al., 2013; Lamanna et al., 2014). No entanto, cada latitude pode apresentar uma composição taxonômica diferente devido à variação nas condições ambientais, com padrões de redundância funcional ao longo do gradiente (Silva & Brandão, 2014) ou diminuição da riqueza funcional em direção às latitudes mais elevadas (Lamanna et al., 2014). Além disso, exibem uma grande variação na riqueza de espécies e nas condições ambientais de uma localidade para outra (por exemplo, longitude e altitude), com uma variedade de topografias e condições climáticas do Sul para o Norte (Gaston, 2000). Da mesma maneira, gradientes de elevação apresentam variação semelhante nos padrões de riqueza de espécies, com diminuição geral da riqueza de espécies devido ao aumento da elevação (Gaston, 2000; Peters et al., 2016; Longino & Branstetter, 2019). Dessa forma, também é esperada uma menor riqueza funcional ou maior redundância funcional com o aumento da elevação (ver Bishop et al., 2014; Tiede et al., 2017). O filtro ambiental no topo das montanhas, devido à variação nas condições geoclimáticas (menores temperaturas, aumente da umidade, radiação solar, além da diminuição de área de ocupação pelas espécies), é uma força severa e seletiva para a ocorrência das espécies de insetos (Bishop et al., 2014; Nunes et al., 2016, 2017; Tiede et al., 2017). Gradientes de elevação são espacialmente heterogêneos e por isso podem abrigar diferentes espécies, uma consequência direta do maior número de micro-habitat, abrigos e locais de forrageamento (Dunn et al., 2010; Körner, 2007; Lessard et al., 2007; Munyai & Foord, 2012). A compreensão de quais mecanismos e processos determinam a estruturação e distribuição das comunidades biológicas em 16 montanhas é de extrema importância, sendo utilizado, por exemplo, como parâmetro para a identificação de mudanças nos padrões de diversidade em função de alterações no clima (Parmesan, 2006; Pecl et al., 2017). Segundo Minx et al. (2017), os topos de montanhas são uma das principais áreas afetadas com o aumento de temperatura global. Com o aquecimento global são esperadas alterações na distribuição das espécies e na estrutura taxonômica e funcional das comunidades nesses gradientes ambientais, em escala local (montanha) ou regional (a cordilheira) (Brousseau, Gravel, & Handa, 2018; Longino & Branstetter, 2018; Parmesan, 2006; Rahbek et al., 2019). As montanhas tropicais, em geral, não possuem uma forte variação anual de temperatura se comparadas às montanhas de regiões temperadas, removendo-se o efeito de baixas temperaturas a ponto de congelamento, importante variável com grande influência na biota de regiões temperadas (Colwell et al., 2008; Janzen, 1967). As montanhas tropicais são consideradas ótimos ambientes para investigarmos quais são os mecanismos estruturadores dos padrões de diversidade das comunidades, no espaço e no tempo, em diferentes escalas espaciais (gradientes de elevação e latitudinal, por exemplo). Além disso, montanhas podem ser consideradas um espaço pequeno o suficiente para permitir que todas as espécies regionais tenham acesso a todas as partes do gradiente (se comparados aos gradientes latitudinais), minimizando os efeitos de limitação de dispersão (Longino & Colwell, 2011). Dentre as cadeias montanhosas tropicais, destaca-se a Cadeia do Espinhaço no Brasil. É uma região tropical megadiversa e extremamente ameaçada por distintas ações antrópicas como mineração, agropecuária e incêndios criminosos (Fernandes et al., 2016; 2018; Domingues & Andrade, 2011). A Cadeia do Espinhaço é a maior cordilheira do Brasil e possui um dos ecossistemas campestres mais antigos e com maior biodiversidade da América do Sul, o campo rupestre, considerado uma paisagem climaticamente tamponada, antiga e infértil (OCBIL, old climate buffered infertile landscape) (Hopper, 2009; Silveira et al., 2016). A ocorrência de espécies restritas a elevações especificas é relativamente comum no campo rupestre (e.g. anfíbios - Leite et al. 2008; 17 aves - Chaves et al. 2015; formigas - Costa et al. 2015; plantas - Mota et al. 2018), tornando-o fonte valiosa de estudos sobre os padrões de distribuição geográfica de espécies ao longo de gradientes de elevação, bem como sobre seus padrões de diversidade funcional e quais os efeitos das variáveis do habitat sobre as comunidades. Formigas exibem uma grande diversidade de estratégias de história de vida (Hölldobler & Wilson, 1990). Vivem em todo o planeta, exceto nos polos e em ambientes acima de 3000 m de elevação ao nível do mar (Bharti et al., 2013; Dunn et al., 2009). Exercem múltiplas funções ecológicas, atuando com predadoras, dispersoras de sementes, herbívoras (cortadeiras) (Baccaro et al., 2015; Hölldobler & Wilson, 1990) e na bioturbação de solos (movimentação de solos por agentes biológicos, com transferência de material biológico e geológico entre solo e superfície), com grande importância na estruturação de solos tropicais (Lavelle, 2002). Também respondem rapidamente às alterações da estrutura da vegetação (Kaspari et al., 2003; Ribas et al., 2003; Solar et al., 2016). Dessa forma, a estrutura da comunidade de formigas pode diferir entre ambientes (por exemplo, campo e floresta) em função das diferenças estruturais dos habitats, como diferenças na complexidade estrutural e heterogeneidade da vegetação; presença ou ausência de serapilheira, e pelas diferentes condições microclimáticas, com variações em sombreamento ou insolação; variação na amplitude de temperatura e umidade (Brühl et al., 1999; Fernandes et al., 2016; Munyai & Foord, 2012). Em regiões tropicais também observamos uma variação sazonal nos padrões de diversidade de espécies de formigas, ocorrendo maior abundância e riqueza de espécies nos períodos mais úmidos e quentes (estação chuvosa) do que nos períodos mais secos e frios (estação seca) (Castro et al., 2012; Esquivel-Muelbert et al., 2017; Leal & Oliveira, 2000; Montine et al., 2014). Em geral, a riqueza de formigas em gradientes montanhosos apresenta o padrão de distribuição linear ou com efeito de pico intermediário (“mid-elevation peak”) com o aumento da elevação (Costa et al., 2015; Longino & Branstetter, 2018; Longino & Colwell, 2011), ocorrendo elevadas taxas de substituição de espécies (turnover) ao longo do gradiente de elevação (Bishop et 18 al., 2015; Brühl et al., 1999; Nowrouzi et al., 2018). Já em gradientes latitudinais, como por exemplo do Cerrado, as comunidades de formigas apresentam um padrão latitudinal inverso, com a diminuição da riqueza de espécies em direção ao nordeste mais seco e quente do Brasil (Vasconcelos et al., 2018). Da mesma forma, comunidades de formigas na Mata Atlântica apresentam um elevado turnover de espécies do sul para o norte, e o padrão de riqueza de espécies invertido , aumentando em direção ao sul, mas com redundância funcional ao longo de todo gradiente, sob forte efeito da diminuição da temperatura em latitudes mais altas (Silva & Brandão, 2014). As formigas constituem um grupo termofílico (“amante do calor”; Hölldobler & Wilson, 1990; Kaspari et al., 2000) e seus amplos padrões na tolerância térmica começaram a ser revelados apenas recentemente (Bishop et al. 2017; Costa et al., 2018; Kaspari et al., 2015; Nowrouzi et al., 2018). Nas regiões tropicais ocorre uma variação sazonal em padrões de diversidade de formigas, geralmente com uma maior abundância e riqueza de espécies no períodos mais quentes e úmido (estação chuvosa) do que no período seco e frio (estação seca) (Castro et al., 2012; Montine et al., 2014; Esquivel-Muelbert et al., 2017; Marques et al., 2017). As comunidades de formigas também respondem à estrutura da vegetação, respondendo positivamente em função do aumento na complexidade e heterogeneidade estrutural do habitat (Kaspari et al., 2003; Solar et al., 2016). Por isso, a composição da comunidade de formigas pode diferir entre os ambientes (por exemplo, campos e florestas) devido a diferenças estruturais em habitat (ou seja, diferenças na complexidade estrutural e heterogeneidade da vegetação e presença ou ausência de serapilheira), além das diferentes condições microclimáticas resultantes, ou seja, maior sombreamento ou insolação, variação nas faixas de temperatura e umidade (Andersen, 2019; Brühl et al., 1999; Fernandes et al., 2016; Lasmar et al., 2020; Munyai & Foord, 2012). Da mesma forma, os padrões de estratégias ou características funcionais podem variar entre diferentes habitats e condições climáticas, como consequência de filtros ambientais (por exemplo, variabilidade climática, heterogeneidade de habitat) (Dunn et al., 2009; Arnan et al., 2018), e pode representar um subconjunto aninhado das estratégias funcionais 19 disponíveis ao longo de um gradiente de elevação (Bishop et al., 2015) ou latitudinal (Silva & Brandão, 2014). Nesses cenários, as variáveis de macrohabitat podem atuar como filtros ambientais, limitando o estabelecimento de espécies incapazes de tolerar condições abióticas de um determinado habitat (Keddy, 1992), o que pode influenciar os padrões de distribuição das espécies e de suas características. Estudos realizados em grandes escalas espaciais para compreender os efeitos dos gradientes ecológicos na distribuição de espécies, como gradientes de elevação e latitude na estrutura taxonômica e funcional de comunidades de formigas, ainda são negligenciados, especialmente em regiões tropicais (Tiede et al., 2017) Visando elucidar lacunas no conhecimento de padrões espaço temporais em montanhas tropicais, temos como objetivo descrever os padrões taxonômicos e funcionais e mecanismos estruturadores das comunidades de Formicidae em gradientes espaço-temporais ao longo da Cadeia do Espinhaço, em diferentes escalas espaciais; em um gradiente de elevação em uma montanha e em um gradiente latitudinal em uma Cadeia de montanhas. Temos o objetivo de elucidar aspectos biogeográficos dessas comunidades no campo rupestre e avaliar como comunidades de formigas respondem às diferentes variáveis ambientais. A tese está organizada em dois capítulos, cada um visando investigar objetivos propostos, na escala local (montanha) e na escala regional (cordilheira) e darão origem a dois manuscritos. 20 Referências bibliográficas Anderson, M. J., Crist, T. O., Chase, J. M., Vellend, M., Inouye, B. D., Freestone, A. L., … Swenson, N. G. (2011). Navigating the multiple meanings of β diversity: a roadmap for the practicing ecologist. Ecology Letters, 14(1), 19–28. doi: 10.1111/j.1461-0248.2010.01552.x Andersen, A. N. (2019). Responses of ant communities to disturbance: Five principles for understanding the disturbance dynamics of a globally dominant faunal group. Journal of Animal Ecology, 88(3), 350–362. doi: 10.1111/1365-2656.12907 Arnan, X., Arcoverde, G. B., Pie, M. R., Ribeiro-Neto, J. D., & Leal, I. R. (2018). Increased anthropogenic disturbance and aridity reduce phylogenetic and functional diversity of ant communities in Caatinga dry forest. Science of the Total Environment, 631–632(March), 429– 438. doi: 10.1016/j.scitotenv.2018.03.037 Baccaro, F. B., Feitosa, R. M., Fernandez, F., Fernandes, I. O., Izzo, T. J., Souza, J. L. P., & Solar, R. (2015). Guia para os gêneros de formigas do Brasil. In Editora INPA. doi: 10.5281/zenodo.32912 Barton, P. S., Cunningham, S. A., Manning, A. D., Gibb, H., Lindenmayer, D. B., & Didham, R. K. (2013). The spatial scaling of beta diversity. Global Ecology and Biogeography, 22(6), 639– 647. doi: 10.1111/geb.12031 Baselga, A. (2010). Partitioning the turnover and nestedness components of beta diversity. Global Ecology and Biogeography, 19(1), 134–143. doi: 10.1111/j.1466-8238.2009.00490.x Baselga, A. (2013). Separating the two components of abundance-based dissimilarity: Balanced changes in abundance vs. abundance gradients. Methods in Ecology and Evolution, 4(6), 552– 557. doi: 10.1111/2041-210X.12029 21 Bharti, H., Sharma, Y. P., Bharti, M., & Pfeiffer, M. (2013). Ant species richness, endemicity and functional groups, along an elevational gradient in the himalayas. Asian Myrmecology, 5(1), 79– 101. doi: 10.1111/j.1600-0706.2010.18772.x Bishop, T. R., Robertson, M. P., van Rensburg, B. J., & Parr, C. L. (2014). Elevation-diversity patterns through space and time: Ant communities of the Maloti-Drakensberg Mountains of southern Africa. Journal of Biogeography. doi: 10.1111/jbi.12368 Bishop, T. R., Robertson, M. P., van Rensburg, B. J., & Parr, C. L. (2015). Contrasting species and functional beta diversity in montane ant assemblages. Journal of Biogeography, 42(9), 1776– 1786. doi: 10.1111/jbi.12537 Bishop, T. R., Robertson, M. P., Van Rensburg, B. J., & Parr, C. L. (2017). Coping with the cold: minimum temperatures and thermal tolerances dominate the ecology of mountain ants. Ecological Entomology, 42(2), 105–114. doi: 10.1111/een.12364 Brousseau, P. M., Gravel, D., & Handa, I. T. (2018). On the development of a predictive functional trait approach for studying terrestrial arthropods. Journal of Animal Ecology, 87(5), 1209–1220. doi: 10.1111/1365-2656.12834 Brühl, C. a., Mohamed, M., & Linsenmair, K. E. (1999). Altitudinal distribution of leaf litter ants along a transect in primary forests on Mount Kinabalu, Sabah, Malaysia. Journal of Tropical Ecology, 15(3), 265–277. doi: 10.1017/S0266467499000802 Castro, F. S. De, Gontijo, A. B., Castro, P. D. T. A., & Ribeiro, S. P. (2012). Annual and seasonal changes in the structure of litter-dwelling ant assemblages (Hymenoptera: Formicidae) in Atlantic semideciduous forests. Psyche, 2012, (Article ID 959715), 12 pages. doi: 10.1155/2012/959715 22 Chaves, A. V., Freitas, G. H. S., Vasconcelos, M. F., & Santos, F. R. (2015). Biogeographic patterns, origin and speciation of the endemic birds from eastern Brazilian mountaintops: A review. Systematics and Biodiversity, 13(1), 1–16. doi: 10.1080/14772000.2014.972477 Colwell, R. K., Brehm, G., Cardelus, C. L., Gilman, A. C., & Longino, J. T. (2008). Global Warming, Elevational Range Shifts, and Lowland Biotic Attrition in the Wet Tropics. Science, 322(5899), 258–261. doi: 10.1126/science.1162547 Cooke, S. J., Sack, L., Franklin, C. E., Farrell, A. P., Beardall, J., Wikelski, M., & Chown, S. L. (2013). What is conservation physiology? Perspectives on an increasingly integrated and essential science. Conservation Physiology, 1(1), 1–23. doi: 10.1093/conphys/cot001 Costa, F. V., Mello, R., Lana, T. C., & Neves, F. S. (2015). Ant fauna in megadiverse mountains: A checklist for the rocky grasslands. Sociobiology, 62(2), 228–245. doi: 10.13102/sociobiology.v62i2.228-245 Costa, Fernanda V., Blüthgen, N., Viana-Junior, A. B., Guerra, T. J., Di Spirito, L., & Neves, F. S. (2018). Resilience to fire and climate seasonality drive the temporal dynamics of ant-plant interactions in a fire-prone ecosystem. Ecological Indicators, 93(October 2017), 247–255. doi: 10.1016/j.ecolind.2018.05.001 Domingues, S. A., Karez, C. S., Biondini, I. V. F., & Andrade, M. Â. (2011). Instrumentos económicos de gestión ambiental en la Reserva de Biosfera de la Serra do Espinhaço. UNESCO Regional Office for Science for Latin America and the Caribbean, p. 36. doi: 10.1109/TDEI.2009.5211872 Dunn, R. R., Agosti, D., Andersen, A. N., Arnan, X., Bruhl, C. A., Cerdá, X., … Sanders, N. J. (2009). Climatic drivers of hemispheric asymmetry in global patterns of ant species richness. Ecology Letters, 12(4), 324–333. doi: 10.1111/j.1461-0248.2009.01291.x 23 Dunn, R. R., Guénard, B., Weiser, M. D., & Sanders, N. J. (2010). Geographic Gradients. In L. Lach, C. L. Parr, & K. L. Abbott (Eds.), Ant Ecology (pp. 38–58). Oxford University Press. Esquivel-Muelbert, A., Baker, T. R., Dexter, K. G., Lewis, S. L., ter Steege, H., Lopez-Gonzalez, G., … Phillips, O. L. (2017). Seasonal drought limits tree species across the Neotropics. Ecography, 40(5), 618–629. doi: 10.1111/ecog.01904 Fernandes, G.W. (2016). The Megadiverse Rupestrian Grassland. In G. W. Fernandes (Ed.), Ecology and Conservation of Mountaintop grasslands in Brazil (pp. 3–14). Springer, Cham. Fernandes, G. Wilson, Almeida, H. A., Nunes, C. A., Xavier, J. H. A., Cobb, N. S., Carneiro, M. A. A., … Beirão, M. V. (2016). Cerrado to Rupestrian Grasslands: Patterns of Species Distribution and the Forces Shaping Them Along an Altitudinal Gradient. In G. W. Fernandes (Ed.), Ecology and Conservation of Mountaintop grasslands in Brazil (pp. 345–377). doi: https://doi.org/10.1007/978-3-319-29808-5_15 Fernandes, G. Wilson, Barbosa, N. P. U., Alberton, B., Barbieri, A., Dirzo, R., Goulart, F., … Solar, R. R. C. (2018). The deadly route to collapse and the uncertain fate of Brazilian rupestrian grasslands. Biodiversity and Conservation, 27(10), 2587–2603. doi: 10.1007/s10531-018-1556- 4 Gaston, K. J. (2000). Global patterns in biodiversity. Nature, 405(6783), 220–227. doi: 10.1038/35012228 Hölldobler, B., & Wilson, E. O. (1990). The ants. Cambridge, Mass, Belknap Press of Harvard University Press. Hopper, S. D. (2009). OCBIL theory: Towards an integrated understanding of the evolution, ecology and conservation of biodiversity on old, climatically buffered, infertile landscapes. Plant and Soil, 322(1), 49–86. doi: 10.1007/s11104-009-0068-0 24 Janzen, D. H. (1967). Why mountain passes are higher in the tropics? The American Naturalist, 101(919), 233–249. Jost, L. (2007). Concepts & Synthesis. Ecology, 88(10), 2427–2439. doi: 10.1890/07-1861.1 Kaspari, M., Clay, N. A., Lucas, J., Yanoviak, S. P., & Kay, A. (2015). Thermal adaptation generates a diversity of thermal limits in a rainforest ant community. Global Change Biology, 21(3), 1092– 1102. doi: 10.1111/gcb.12750 Kaspari, M., Donnell, S. O., & Kercher, J. R. (2000). Energy, Density, and Constraints to Species Richness: Ant Assemblages along a Productivity Gradient. The American naturalist, 155(2), 280-293. Kaspari, M., Yuan, M., & Alonso, L. (2003). Spatial Grain and the Causes of Regional Diversity Gradients in Ants. The American Naturalist, 161(3), 459–477. doi: 10.1086/367906 Keddy, P. A. (1992). Assembly and response rules: two goals for predictive community ecology. Journal of Vegetation Science, 3(2), 157–164. doi: 10.2307/3235676 Körner, C. (2007). The use of “altitude” in ecological research. Trends in Ecology and Evolution, 22(11), 569–574. doi: 10.1016/j.tree.2007.09.006 Lamanna, C., Blonder, B., Violle, C., Kraft, N. J. B., Sandel, B., Šímová, I., … Enquist, B. J. (2014). Functional trait space and the latitudinal diversity gradient. Proceedings of the National Academy of Sciences of the United States of America, 111(38), 13745–13750. doi: 10.1073/pnas.1317722111 Lasmar, C. J., Ribas, C. R., Louzada, J., Queiroz, A. C. M., Feitosa, R. M., Imata, M. M. G., … Domingos, D. Q. (2020). Disentangling elevational and vegetational effects on ant diversity patterns. Acta Oecologica, 102 (November 2019), 103489. doi: 10.1016/j.actao.2019.103489 25 Lassau, S. a., & Hochuli, D. F. (2004). Effects of habitat complexity on ant assemblages. Ecography, 27(2), 157–164. doi: 10.1111/j.0906-7590.2004.03675.x Lavelle, P. (2002). Functional domains in soils. Ecological Research, 17(4), 441–450. doi: 10.1046/j.1440-1703.2002.00509.x Leal, I. R., & Oliveira, P. S. (2000). Foraging ecology of attine ants in a Neotropical savanna: seasonal use of fungal substrate in the cerrado vegetation of Brazil. Insectes Sociaux, 47(4), 376–382. doi: 10.1007/PL00001734 Leite, F. S. F., Juncá, F. A., & Eterovick, P. C. (2008). Status do conhecimento, endemismo e conservação de anfíbios anuros da Serra do Espinhaço, Brasil. Megadiversidade, 4(1), 158– 176. doi: 10.3905/jpm.2003.319889 Lessard, J.-P., Dunn, R. R., Parker, C. R., & Sanders, N. J. (2007). Rarity and Diversity in Forest Ant Assemblages of Great Smoky Mountains National Park. Southeastern Naturalist, 6(sp1), 215–228. doi: 10.1656/1528-7092(2007)6[215:RADIFA]2.0.CO;2 Longino, J. T., & Branstetter, M. G. (2018). The truncated bell: an enigmatic but pervasive elevational diversity pattern in Middle American ants. Ecography, 1–12. doi: 10.1111/ecog.03871 Longino, J. T., & Branstetter, M. G. (2019). The truncated bell: an enigmatic but pervasive elevational diversity pattern in Middle American ants. Ecography, 42(2), 272–283. doi: 10.1111/ecog.03871 Longino, J. T., & Colwell, R. K. (2011). Density compensation, species composition, and richness of ants on a neotropical elevational gradient. Ecosphere, 2(3), art29. doi: 10.1890/ES10-00200.1 Minx, J. C., Callaghan, M., Lamb, W. F., Garard, J., & Edenhofer, O. (2017). Learning about climate change solutions in the IPCC and beyond. Environmental Science and Policy, 77(June), 252– 259. doi: 10.1016/j.envsci.2017.05.014 26 Marques, T., Espiríto-Santo, M. M., Neves, F. S., & Schoereder, J. H. (2017). Ant assemblage structure in a secondary tropical dry forest: The role of ecological succession and seasonality. Sociobiology, 64(3), 261–275. doi: 10.13102/sociobiology.v64i3.1276 Montine, P. S. M., Viana, N. F., Almeida, F. S., Dáttilo, W., Santanna, A. S., Martins, L., & Vargas, A. B. (2014). Seasonality of epigaeic ant communities in a Brazilian atlantic rainforest. Sociobiology, 61(2), 178–183. doi: 10.13102/sociobiology.v61i2.178-183 Morellato, L. P. C., & Silveira, F. A. O. (2018). Plant life in campo rupestre: New lessons from an ancient biodiversity hotspot. Flora: Morphology, Distribution, Functional Ecology of Plants, 238, 1–10. doi: 10.1016/j.flora.2017.12.001 Munyai, T. C., & Foord, S. H. (2012). Ants on a mountain: Spatial, environmental and habitat associations along an altitudinal transect in a centre of endemism. Journal of Insect Conservation, 16(5), 677–695. doi: 10.1007/s10841-011-9449-9 Nowrouzi, S., Andersen, A. N., Bishop, T. R., Simon, ·, & Robson, K. A. (2018). Is thermal limitation the primary driver of elevational distributions? Not for montane rainforest ants in the Australian Wet Tropics. Oecologia, 1(0123456789). doi: 10.1007/s00442-018-4154-y Nunes, Cássio A., Quintino, A. V., Constantino, R., Negreiros, D., Reis Júnior, R., & Fernandes, G. W. (2017). Patterns of taxonomic and functional diversity of termites along a tropical elevational gradient. Biotropica, 49(2), 186–194. doi: 10.1111/btp.12365 Nunes, Cássio Alencar, Braga, R. F., Figueira, J. E. C., Neves, F. de S., & Fernandes, G. W. (2016). Dung Beetles along a Tropical Altitudinal Gradient: Environmental Filtering on Taxonomic and Functional Diversity. PLOS ONE, 11(6), e0157442. doi: 10.1371/journal.pone.0157442 Parmesan, C. (2006). Ecological and Evolutionary Responses to Recent Climate Change. Annual Review of Ecology, Evolution, and Systematics, 37(1), 637–669. doi: 10.1146/annurev.ecolsys.37.091305.110100 27 Pecl, G. T., Araújo, M. B., Bell, J. D., Blanchard, J., Bonebrake, T. C., Chen, I. C., … Williams, S. E. (2017, March 31). Biodiversity redistribution under climate change: Impacts on ecosystems and human well-being. Science, Vol. 355, p. eaai9214. doi: 10.1126/science.aai9214 Petchey, O. L., & Gaston, K. J. (2006). Functional diversity: Back to basics and looking forward. Ecology Letters, 9(6), 741–758. doi: 10.1111/j.1461-0248.2006.00924.x Peters, M. K., Hemp, A., Appelhans, T., Behler, C., Classen, A., Detsch, F., … Steffan-Dewenter, I. (2016). Predictors of elevational biodiversity gradients change from single taxa to the multi-taxa community level. Nature Communications, 7(September 2011). doi: 10.1038/ncomms13736 Rahbek, C., Borregaard, M. K., Antonelli, A., Colwell, R. K., Holt, B. G., Nogues-Bravo, D., … Fjeldså, J. (2019). Building mountain biodiversity: Geological and evolutionary processes. Science, 365(6458), 1114–1119. doi: 10.1126/science.aax0151 Ribas, C. R., Schoereder, J. H., Pic, M., & Soares, S. M. (2003). Tree heterogeneity, resource availability, and larger scale processes regulating arboreal ant species richness. Austral Ecology, 28(3), 305–314. doi: 10.1046/j.1442-9993.2003.01290.x Silva Mota, G., Luz, G. R., Mota, N. M., Silva Coutinho, E., das Dores Magalhães Veloso, M., Fernandes, G. W., & Nunes, Y. R. F. (2018). Changes in species composition, vegetation structure, and life forms along an altitudinal gradient of rupestrian grasslands in south-eastern Brazil. Flora: Morphology, Distribution, Functional Ecology of Plants, 238, 32–42. doi: 10.1016/j.flora.2017.03.010 Silva, R. R., & Brandão, C. R. F. (2014). Ecosystem-wide morphological structure of leaf-litter ant communities along a tropical latitudinal gradient. PLOS ONE, 9(3). doi: 10.1371/journal.pone.0093049 Silveira, F. A. O., Barbosa, M., Beiroz, W., Callisto, M., Macedo, D. R., Morellato, L. P. C., … Fernandes, G. W. (2019). Tropical mountains as natural laboratories to study global changes: A 28 long-term ecological research project in a megadiverse biodiversity hotspot. Perspectives in Plant Ecology, Evolution and Systematics, 38, 64–73. doi: 10.1016/j.ppees.2019.04.001 Silveira, F. A. O., Negreiros, D., Barbosa, N. P. U., Buisson, E., Carmo, F. F., Carstensen, D. W., … Lambers, H. (2016). Ecology and evolution of plant diversity in the endangered campo rupestre: a neglected conservation priority. Plant Soil, 403(1–2), 129–152. doi: 10.1007/s11104-015- 2637-8 Solar, R. R. de C., Barlow, J., Andersen, A. N., Schoereder, J. H., Berenguer, E., Ferreira, J. N., & Gardner, T. A. (2016). Biodiversity consequences of land-use change and forest disturbance in the Amazon: A multi-scale assessment using ant communities. Biological Conservation, 197, 98–107. doi: 10.1016/j.biocon.2016.03.005 Stevens, R. D., Cox, S. B., Strauss, R. E., & Willig, M. R. (2003). Patterns of functional diversity across an extensive environmental gradient: Vertebrate consumers, hidden treatments and latitudinal trends. Ecology Letters, 6(12), 1099–1108. doi: 10.1046/j.1461-0248.2003.00541.x Tiede, Y., Schlautmann, J., Donoso, D. A., Wallis, C. I. B., Bendix, J., Brandl, R., & Farwig, N. (2017). Ants as indicators of environmental change and ecosystem processes. Ecological Indicators, 83, 527–537. doi: 10.1016/j.ecolind.2017.01.029 Tuomisto, H. (2010). A diversity of beta diversities: Straightening up a concept gone awry. Part 1. Defining beta diversity as a function of alpha and gamma diversity. Ecography, 33(1), 2–22. doi: 10.1111/j.1600-0587.2009.05880.x Vasconcelos, H. L., Maravalhas, J. B., Feitosa, R. M., Pacheco, R., Neves, K. C., & Andersen, A. N. (2018). Neotropical savanna ants show a reversed latitudinal gradient of species richness, with climatic drivers reflecting the forest origin of the fauna. Journal of Biogeography, 45(1), 248– 258. doi: 10.1111/jbi.13113 29 Villéger, S., Grenouillet, G., & Brosse, S. (2013). Decomposing functional β-diversity reveals that low functional β-diversity is driven by low functional turnover in European fish assemblages. Global Ecology and Biogeography, 22(6), 671–681. doi: 10.1111/geb.12021 Violle, C., Navas, M. L., Vile, D., Kazakou, E., Fortunel, C., Hummel, I., & Garnier, E. (2007). Let the concept of trait be functional! Oikos, 116(5), 882–892. doi: 10.1111/j.0030- 1299.2007.15559.x Whittaker, R. (1960). Vegetation of the Siskiyou, Oregon and California. Ecological Monographs, 30(3), 279–338. 30 Capítulo 1: Environmental drivers of taxonomic and functional diversity of ant communities in a tropical mountain Flávio Siqueira de Castro, Pedro Giovâni da Silva, Ricardo Ribeiro de Castro Solar, Geraldo Wilson Fernandes e Frederico de Siqueira Neves Artigo submetido para o periódico Insect Conservation and Diversity (Qualis Capes A2, Impact Factor 2.313) Status: Major Revision 31 1 Insect Conservation and Diversity Section: Original Research 2 Environmental drivers of taxonomic and functional diversity of ant communities in a 3 tropical mountain 4 5 Running title: Ant taxonomic and functional diversity 6 7 Flávio Siqueira de Castro1 8 Pedro Giovâni da Silva1 9 Ricardo R. C. Solar1,2 10 G. Wilson Fernandes1,2 11 Frederico de Siqueira Neves1,2,3 12 13 1. Programa de Pós-Graduação em Ecologia, Conservação e Manejo da Vida Silvestre, 14 Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil 15 2. Departamento de Genética, Ecologia e Evolução, Universidade Federal de Minas Gerais, 16 Belo Horizonte, Minas Gerais, Brazil 17 3. Department of Biological Sciences, The George Washington University, Washington, 18 DC, USA 19 20 Correspondence: 21 Flávio Siqueira de Castro 22 32 Programa de Pós-Graduação em Ecologia, Conservação e Manejo da Vida Silvestre, 23 Universidade Federal de Minas Gerais, Av. Antônio Carlos, 6627, Pampulha, CEP 31270-24 901, Belo Horizonte, Minas Gerais, Brazil. 25 E-mail: fsiqueiradecastro@gmail.com 26 27 ORCID ID: 28 Flávio Siqueira de Castro (0000-0002-5533-1355) 29 Pedro Giovâni da Silva (0000-0002-0702-9186) 30 Ricardo R. C. Solar (0000-0001-5627-4017) 31 Geraldo Wilson Fernandes (0000-0003-1559-6049) 32 Frederico de Siqueira Neves (0000-0002-2691-3743) 33 34 Abstract 35 1. We investigated the patterns of taxonomic (TD) and functional (FD) α and β diversities of ants 36 in a mountainous landscape along three dimensions, namely one temporal (seasonal) and two spatial 37 dimensions: between habitats – grassland and forest habitats (horizontal), and among elevation 38 bands (vertical). In addition, we tested the effects of environmental variables (mean elevation and 39 temperature, and normalized difference vegetation index – NDVI) on taxonomic and functional α 40 and β-diversities. 41 2. The β-diversities of the two spatial dimensions are the main components of TD. The α-diversities 42 of the two spatial dimensions exhibit contrasting patterns to drive taxonomic α-diversity, indicating 43 differences in the community at local scale on grassland and forest habitats. The FD is almost 44 entirely represented by the α-diversity component, with very low contribution of β-diversity. 45 3. Regarding environmental drivers, the decrease in temperature caused by increased elevations and 46 seasonal variations had a negative effect on taxonomic α-diversity. There were no effects of 47 environmental variables on functional α-diversity. 48 4. Despite the high turnover of ant species occurring along spatial dimensions, the communities 49 were functionally redundant. The changes in species richness and composition patterns in this 50 mountain were strongly influenced by variables correlated with elevation and habitat structure. 51 5. Species composition changed across all dimensions, but the core traits and functions remained 52 unchanged. Differences observed in the composition of ant communities over relatively short 53 geographic distances highlight the importance to conserve the entire mountain, ensuring the 54 maintenance of the ant diversity and associated ecosystem functions. 55 56 Keywords: turnover; nestedness; ground-dwelling ants; Espinhaço Mountain Range; campo 57 rupestre; rupestrian grassland 58 34 Introduction 59 Understanding how communities are structured in space and time, and the underlying 60 mechanisms driving diversity patterns are among the central goals of ecology and conservation 61 biology (Gaston, 2000; Vellend, 2016). To achieve this goal, it is important to evaluate multiple 62 scales and metrics of biodiversity in response to ecological processes (Tuomisto, 2010; Anderson 63 et al., 2011; Barton et al., 2013). The use of multiple scales of diversity proposed by Whittaker 64 (1960) is an adequate approach to elucidate these questions, focusing on β-diversity (a measure of 65 species composition variability among samples), which links local diversity (α) to broad-scale 66 diversity (γ). Baselga (2010) brought a complimentary process-focused approach to describe β-67 diversity into turnover (species substitution) and nestedness (species loss). Understanding how the 68 total diversity is partitioned into α and β-diversity, as well as how β-diversity is decomposed into 69 its components (turnover and nestedness; Baselga 2010), is extremely important to identify which 70 mechanisms support the observed biodiversity patterns. 71 Taxonomic diversity, mainly based on species richness, considers that all species are 72 different from each other but overlooks that species can play distinct ecological roles (Villéger et 73 al., 2013). Functional diversity, on the other hand, reflects the variety of functional traits among all 74 species in the community (Petchey & Gaston, 2006). Functional traits are measurable biological 75 characteristics of organisms that influence their performance or fitness and ecosystem functioning 76 (Violle et al., 2007). The approaches proposed by Whittaker (1960) and Baselga (2010) have 77 increasingly been applied to both taxonomic (TD) and functional (FD) diversity (Petchey & Gaston, 78 2006; Jost, 2007; Villéger et al., 2013). Analysing diversity patterns with different approaches (e.g., 79 TD and FD) allows to identify the processes associated with the origin and maintenance of 80 biodiversity and the associated services provided by the different species (Anderson et al., 2011; 81 Barton et al., 2013). Thus, elucidating the mechanisms underpinning the patterns of biodiversity 82 35 distribution is key to guide future effective conservation actions (Lassau & Hochuli, 2004; Cooke 83 et al., 2013) given the current scenario of global changes. 84 Mountainous environments are spatially heterogeneous and can harbour many species, as a 85 direct consequence of the greater number of microhabitats, shelters, and foraging sites (Körner, 86 2007; Munyai & Foord, 2012). Elevation gradients can also be considered spatially small (if 87 compared to latitudinal gradients) to allow all species pool to have access to the whole gradient, 88 minimizing the effects of dispersal limitation (Longino & Colwell, 2011). These environments have 89 extreme environmental characteristics (e.g., increasing elevation leading to lower temperature, 90 primary productivity, and species-area ratio) and are considered ideal models for investigating the 91 spatial patterns and processes that determine the distribution of biodiversity (Gaston, 2000; Longino 92 & Colwell, 2011; Smith, 2015; Longino & Branstetter, 2019). Similarly, mountains are also suitable 93 systems for identifying changes in diversity patterns due to changes in climate (Parmesan, 2006; 94 Pecl et al., 2017). Tropical mountains have higher temperatures at low elevations and overall wetter 95 conditions than temperate mountains (McCain & Grytnes, 2010). In addition to the vertical 96 dimension, the occurrence of different habitats (e.g., field and forest, Lasmar et al., 2020) along the 97 mountain that possess a seasonal climate includes two important factors to be addressed (habitat 98 type and seasonality), since forests can buffer daily variations in temperature and maintain the 99 humidity, ultimately changing the patterns of species distribution. The effects of these different 100 factors and their interactions on both taxonomic and functional diversity of communities still need 101 to be comprehensively addressed. 102 Among the tropical mountains, the ancient Espinhaço Range (1100 km length, emerged 103 nearly 640 Ma; Alkmin, 2012) is a tropical region extremely endangered (e.g., mining, agriculture, 104 deforestation) and biodiverse (Silveira et al., 2016; Fernandes et al., 2018). Located in an ecotone 105 region amid three biomes, the Cerrado to the west, the Atlantic Forest to the east (two biodiversity 106 hotspots; Mittermeier et al., 2004), and the Caatinga to the north (Giulietti et al., 1997; Fernandes, 107 36 2016; Silveira et al., 2016), Espinhaço has great strategic importance for the conservation of unique 108 natural environments in Brazil (Domingues et al., 2011; Fernandes et al., 2018). The environments 109 found in the Espinhaço Range are formed mainly by rocky grassland ecosystem, the campo rupestre 110 (Silveira et al., 2016), permeated by forests such as riparian forests and forest islands (capões de 111 mata) (Coelho et al., 2018b), forming vegetation mosaics along the elevational and latitudinal 112 gradients (Giulietti et al., 1997; Silveira et al., 2016). The occurrence of species restricted to specific 113 elevations is relatively common in the campo rupestre (Leite et al., 2008; Chaves et al., 2015; Costa 114 et al., 2015; Mota et al., 2018), making it a valuable system to investigate the patterns of species 115 distribution along elevation gradients. 116 In general, patterns of ant species richness present a linear decrease or a mid-elevation peak 117 with increasing elevation (Longino & Colwell, 2011; Longino & Branstetter, 2019; Lasmar et al., 118 2020), with high species replacement rates (turnover) along the elevation gradient (Brühl et al., 119 1999; Bishop et al., 2015; Nowrouzi et al., 2016). Moreover, in tropical regions a seasonal variation 120 in ant diversity patterns occurs, with usually greater abundance and richness of species in the wetter 121 and hotter periods (rainy season) than in the drier and colder periods (dry season) (Castro et al., 122 2012; Montine et al., 2014; Esquivel-Muelbert et al., 2017; Marques et al., 2017). Ant communities 123 also respond to vegetation structure (Kaspari et al., 2003; Solar et al., 2016) and like so, ant 124 community composition may differ between environments (e.g., grassland and forest) due to 125 structural differences in habitats (i.e., differences in structural complexity and heterogeneity of 126 vegetation, presence or absence of leaf litter) and the resulting different microclimatic conditions 127 (i.e., greater shading or insolation, variation in temperature and humidity ranges) (Andersen, 2019; 128 Brühl et al., 1999; Fernandes et al., 2016; Lasmar et al., 2020; Munyai & Foord, 2012). Similarly, 129 patterns of functional strategies or traits could vary among different habitats and climatic conditions, 130 as a consequence of environmental drivers (e.g., climatic variability, heterogeneity of habitat) 131 37 (Dunn et al., 2009; Arnan et al., 2018) and could represent a nested subset of the available functional 132 strategies along an elevational gradient (Bishop et al., 2015). 133 Here, we aimed to determine the patterns of taxonomic and functional diversity of ants in 134 different spatio-temporal dimensions (horizontal, vertical, or seasonal) along a tropical mountainous 135 landscape, as well as to explore the environmental variables driving such patterns. In this study, 136 three spatio-temporal dimensions were considered (adapted from Basset et al., 2015): one temporal 137 dimension (dry and rainy season) and two spatial dimensions: between habitats within the same 138 elevational range (horizontal dimension), and across different elevations (vertical dimension). Our 139 main prediction was that environmental filters related to elevation (vertical dimension) and habitat 140 (horizontal dimension) would be the main drivers of both taxonomic and functional α and β-141 diversities. In addition, since temperature and humidity rise during the rainy season (Ferrari et al., 142 2016) and ant communities present a seasonal richness pattern, we hypothesized that α and β-143 diversities (TD and FD) would vary along the spatio-temporal gradient studied and we expect they 144 would be higher in the hotter and wetter period, mainly in lower elevations, due to the combined 145 effects of habitat and climate factors occurring along the elevation gradient. Specifically, we expect 146 the following: 147 (i) We expected to find a lower taxonomic diversity (TD) of ants in higher elevations. 148 Diversity is not uniformly distributed throughout space (Whittaker, 1960) and ant richness 149 decreases with increasing elevation (i.e., patterns usually present a linear decrease or a mid-150 elevation peak with increasing elevation) (Smith, 2015; Longino & Branstetter, 2019). In 151 our case, we expected a linear taxonomic decrease as found for other insect groups (i.e., 152 wasps, termites, galling insects, and dung beetles) in the same system (Coelho et al., 2018a; 153 Nunes et al., 2016; 2017; Perillo et al., 2017). Ants are responsive to the habitat openness 154 and to vegetation structure type (Kaspari et al., 2003; Solar et al., 2016; Andersen, 2019), 155 which could result in different communities among grassland and forest. In addition, 156 38 although the vegetation heterogeneity or complexity on campo rupestre could decrease with 157 elevation increases (Mota et al., 2018), the habitat type is kept considering the same 158 vegetation type (forest or grassland). Nonetheless, we expect higher turnover between 159 elevations followed by among habitats. Once ants could be considered an ecological 160 indicator of climatic variation campo rupestre ecosystem (Costa et al., 2018), with greater 161 richness in hot and warm season rather than cold and dry season, we expect to found patterns 162 related to a real effect of seasonal climatic conditions on ant communities. Thus, seasonally, 163 we expect to find greater nestedness since the dry season communities were subsets of the 164 communities found in the rainy season. 165 (ii) Despite the low richness in higher elevations, we also expected that the taxonomic 166 β-diversity would be explained mainly by the turnover pattern, with different ant 167 communities found across the elevational gradient (i.e., high elevations communities will 168 not be subsets of lowland ant communities). Along elevational gradients, there are high 169 species replacement rates (turnover) (Bishop et al., 2015; Nowrouzi et al., 2018). In the 170 campo rupestre, ants and other insects (i.e., wasps, termites, galling insects, and dung 171 beetles) typically exhibit elevational turnover patterns (Coelho et al., 2018a; Nunes et al., 172 2016, 2017; Perillo et al., 2017). Generally, the turnover of ant species in tropical mountains 173 is related to a simplification of the vegetation structure with elevation, and also to a decrease 174 in temperature and increase in rainfall along the elevation (Brühl et al., 1999; Dunn et al., 175 2009), which is the case of the Espinhaço Range (Fernandes et al., 2016). 176 (iii) We expect that functional β-diversity would be mostly driven by nestedness across 177 the elevational gradient, as found for other insects such as ants, dung beetles, galling insects, 178 and termites (Bishop et al., 2015; Coelho et al., 2018a; Nunes et al., 2016, 2017). At higher 179 elevations or under harsher climatic conditions, many organisms (including ants) display a 180 shrank or clustered pattern of phylogenetic diversity due to the species richness decrease 181 39 with an increase in elevation (Smith, 2015; Smith et al., 2014; Machac et al., 2011). Thus, 182 we predict that ant communities there will be poor at higher elevations and the functional 183 diversity could diminish as well and become restricted to particular ant phenotypes in 184 response to harsh environmental conditions in mountain ecosystems (Smith, 2015). 185 186 40 Material and Methods 187 Study area 188 The study was conducted in Serra do Cipó, located in the southern portion of the Espinhaço 189 Range (Fig. 1), dominated by Cerrado vegetation and campo rupestre, with the occurrence of 190 riparian forests, capões de mata (forest islands), and semi-deciduous and deciduous seasonal forest 191 (Giulietti et al., 1997; Silveira et al., 2016). The climate is mesothermic (Cwb in the Köppen 192 classification), with dry winters and rainy summers, mean annual rainfall of 1500 mm and average 193 annual temperature ranging from 17.4 to 19.8ºC (Ferrari et al., 2016; Silveira et al., 2016). 194 Environmental variables, such as temperature and humidity, tend to decrease and rainfall tends to 195 increase as a function of increasing elevation, following the expected patterns for mountainous 196 environments (Ferrari et al., 2016; Silveira et al., 2016; Fernandes et al., 2016). Located in a region 197 of high biodiversity, the Serra do Cipó comprises a private area of environmental protection (APA 198 Morro da Pedreira) and a National Park of integral protection (PARNA Serra do Cipó), besides 199 being part of the Espinhaço Range Biosphere Reserve (Domingues et al., 2011; Fernandes et al., 200 2018). 201 202 203 Fig 1. Map of the study area with the location of each transect at the six sampling sites of PELD/CRSC—Long-Term Ecological Research. Serra do 204 Cipó, Minas Gerais State, Brazil. The three sampling spatio-temporal dimensions considered: (1) temporal dimension (between seasons); (2) 205 horizontal dimension (between habitats – grassland and forest); and (3) vertical dimension (among elevations). 206 The sampled areas are established in the permanent plots of the Long Term Ecological 207 Research Project Campos Rupestres (PELD CRSC/CNPq Project) along a gradient of elevation in 208 the Serra do Cipó National Park, Minas Gerais State, Brazil (19º22’01”S, 43º32’17”W) (Fernandes 209 et al., 2016; Silveira et al., 2019). The elevation gradient has a range of 500 m, ranging from 800 to 210 1300 m (meters above the sea level). The Cerrado vegetation (Neotropical savanna) occurs at 800 211 m and the transition from Cerrado to campo rupestre occurs at 900 m (Silveira et al., 2019). 212 However, between 1000 and 1300 m is where the campo rupestre sensu lato occurs (for more details 213 see Silveira et al., 2019), which are areas with mountainous vegetation, typically rocky grassland 214 and shrubby vegetation, with quartzitic outcrops and sandy, rocky or flooded grasslands, permeated 215 by forest areas with transitional vegetation such as riparian forests (among 800 and 1200 m) and 216 natural forest fragments of Atlantic Forest, capões de mata or Forest Islands (among 1200 and 1300 217 m, see Coelho et al., 2018b). The floristic similarity between riparian forests and capões de mata is 218 also quite high (Coelho et al. 2018b; Meguro et al., 1996). The campo rupestre sensu stricto (where 219 woodlands do not occur) is a very old ecosystem, climatically buffered and infertile landscape 220 (OCBILs), with high concentrations of Al+3 in the soil; possibly the oldest grassland ecosystem to 221 the east of South America (Silveira et al., 2016). 222 223 Sampling of ants 224 Sampling was carried out in the campo rupestre and forest environments (riparian forests 225 and capões de mata), during the dry (July to August, 2012) and rainy (January to February, 2013) 226 seasons, at six sampling sites pre-established by the PELD CRSC/CNPq Project (Silveira et al., 227 2019). Sample sites were distributed along an elevational gradient in six distinct elevation plots 228 among 800 and 1300 m, spaced every 100 m of elevation and geographically distant by at least three 229 kilometres. In each elevation, three linear transects were arranged in the campo rupestre and other 230 three transects were arranged in the forest environment closest to each campo rupestre (riparian 231 43 forests among 800-1200 m and capões de mata at 1300 m), totalling 18 forest transects (Fig. 1). 232 Each transect was 200 m of extension in the north-south direction, distant from each other by 250 233 m, totalling 18 transects along the gradient. In each transect, five pitfall traps were arranged 50 m 234 apart (a plastic pot with 14 cm diameter × 9 cm height with 500 ml of a saline-detergent solution). 235 The spacing of 50 meters between traps is considered enough to avoid interference related to the 236 foraging range of ants belonging to the same colony (Leponce et al., 2004). All pitfall traps 237 remained in the field for 48 hours per survey (Bestelmeyer et al., 2000). Each transect was 238 considered an independent sample replicate (data from traps were pooled) in further analyses (6 239 elevations × 6 transects × 2 seasons, N = 72). 240 241 Identification of ant species and description of functional traits 242 The ants were identified to species and morphospecies by comparison with the Collection 243 of Formicidae from campo rupestre of the Laboratory of Insect Ecology at the Universidade Federal 244 de Minas Gerais, Brazil, and with the help of experts of different ant taxonomic groups. We 245 followed Baccaro et al., (2015) and Ants of Bolton World Catalog (Bolton et al., 2005) 246 classifications. 247 Ant species were described in terms of functional traits that provide information about the 248 ecological functions, linked to diet, nesting, foraging capacity, thermoregulation, and habitat 249 association (Fichaux et al., 2019; Paolucci et al., 2016; Leal et al., 2012; Bishop et al., 2016; Barden, 250 2017; Tiede et al., 2017). Seven traits were described to each species: Weber’s length, femur length, 251 mandible length, colour (mesossoma), polymorphism, integument sculpture, and functional groups 252 (six morphological traits and one ecological trait; Table 1). 253 Table 1. List of morphological and ecological traits measured and their hypothesized ecological 254 functions 255 Traits Measure Abbrev. /Unit Ecological functions Morphological Traits Weber’s length Continuous WL (μm) Proxy for total size, related to habitat complexity (Weber, 1938; Kaspari & Weiser, 1999). Femur length Continuous HFL(μm) Indicator of foraging speed, associated to habitat complexity (Feener et al., 1988; Yates et al., 2014). Mandible length Continuous ML(μm) Indicative of diet (Brandão et al., 2009). Colour (Mesossoma) Continuous V (%) * Thermal melanism: dark individuals has an benefit in cool climates compared to a lighter one (Clusella et al., 2007); Indicative of thermotolerance and, directly related to temperature variation and solar radiation (e.g. ants in cold environments may be darker integument rather than in warm environments with greater UV-B rates) (Bishop et al., 2016). Polymorphism Categorical 1 = monomorphic; 2 = dimorphic; 3 = polymorphic Polymorphism of the workers, attribute related to the ability to develop different tasks in the 45 Traits Measure Abbrev. /Unit Ecological functions colony (e.g., foraging, protection, internal activities of the nest; Wills et al., 2017). Integument Sculpture Ordinal 1 = cuticle smooth/shiny; Protection from desiccation. Thickened cuticles enhanced the dehydration tolerance (Nation, 2008; Terblanche, 2012) 2 = superficial wrinkles/pits; 3 = surface heavily textured Ecological Trait Functional Groups Categorical AA = Army Ants; AD = Arboreal Dominant; AP = Arboreal Predator; AS = Arboreal Subordinate; CO = Cryptic Omnivores; CP = Cryptic Predators; DD = Dominant Dolichoderinae; EO = Epigeic Omnivores; EP = Epigeic Predators; Hatt = High Attini; Latt = Low Attini; Opp = Opportunist; Functional groups based on global-scale responses of ants to environmental stress and disturbance. Also, indicative of ecological tasks, such as nesting, foraging, and diet habits (Andersen, 1995; Leal et al., 2012; Paolucci et al., 2016). All groups were based on the classification used by Paolucci et al., (2016), except for Seed Harvester group (Johnson, 2015) here represented by Pogonomyrmex naegelli, which was not present in this list. 46 Traits Measure Abbrev. /Unit Ecological functions SC = Subordinate Camponotini SH = Seed Harvester; * The HSV cylindrical-coordinate colour model (Smith, 1978), whereas: H = Hue shows the dominant 256 wavelength; S = Saturation, indicates the amount of dominant wavelength (H) present in the colour; 257 and V = Value, defines the amount of bright in the colour. We analysed only the variable V, which 258 measured in % of colour brightness (e.g., white colour presents 100% of bright while black colour has 259 0% of bright) (as proposed by Bishop et al., 2016). 260 261 We followed the guide for identification of functional attributes for ants (The Global Ants 262 trait Database – GLAD; Parr et al., 2017) to perform the morphological measurements, except for 263 the variable “Colour”. Ant colour was obtained from the HSV colour model (Smith, 1978) using 264 only the variable V (colour brightness), as proposed by (Bishop et al., 2016). Differently from 265 Bishop et al., (2016), who considered a predominant colour between head, mesosoma, and gaster, 266 we performed the capture of HSV values of the predominant colour on the mesosoma of each 267 specimen. We measured only the mesosoma because it is an important body part for acquiring heat 268 for body activities (not only for flight) in Hymenoptera, whose warmth transfer occurs via 269 hemolymph from mesossoma to head, gaster, and legs when the thoracic muscles are activated 270 (Terblanche, 2012). All continuous data, except Weber’s length and colour brightness, were divided 271 by Weber’s length to correct for individual body size, because traits were not normally distributed 272 (Arnan et al., 2018; Fichaux et al., 2019). 273 We performed image acquisition using Microscope Digital Camera LC30 OLYMPUS® 274 mounted on a stereomicroscope SZ61 OLYMPUS®. Measurements were made with a digital 275 capture micrometre (accurate to 0.01 mm) provide in the LC Micro 2.2 OLYMPUS® software. We 276 47 take measures of the randomly selected individuals of every species recorded in the dataset. At least 277 six individuals were measured when possible, and whenever it was not possible, we take measures 278 of available individuals (only minor workers were used; N = 2760 images from 920 individuals 279 measured; average = 4.72 individuals per species). Categorical and ordinal morphological traits 280 used (polymorphism and integument sculpture) were attributed using genera/species information 281 available at AntWeb (www.antweb.org) and AntWiki website (www.antwiki.org) (Guénard et al., 282 2017) and by own observations . 283 284 Taxonomic and functional diversity 285 Taxonomic diversity (TD) and functional diversity (FD) were calculated for each of the 72 286 transects. The TD was calculated by species richness in each transect and the functional diversity 287 (FD) was calculated using the mean value functional attribute data for the species in each transect 288 (Table 1; Supplementary Material I Table S1). To obtain functional diversity we calculated a species 289 dissimilarity matrix based on all ant traits using the Gower distance (Villéger et al.,2008) with the 290 “trova” function (i.e., TRait OVerlAp) of the R software (de Bello et al., 2011). Gower distance 291 measures the dissimilarity of traits between species using different types of data: categorical, ordinal 292 or continuous (see de Bello et al., 2011; Nunes et al., 2016 for more details). We then used the 293 dissimilarity matrix to calculate the Rao index, which estimates FD based on species incidence and 294 Gower dissimilarities at each sampling point. We used the “Rao” R function to calculate Rao index 295 considering Jost’s correction (de Bello et al., 2010). 296 297 48 Environmental variables 298 Climatic variables were obtained along the elevation gradient using a meteorological 299 monitoring tower (equipped with an Onset HOBO® U30 data logger) located at every 100 m of 300 elevation on grassland, adjacent to the sampling sites of the PELD CRSC/CNPq Project, between 301 April 2012 and February 2013. The climatic variables measured were the mean air temperature, 302 mean air humidity, and accumulated rainfall for each survey period (dry, July to August/2012, and 303 rainy, January to February/2013, seasons) and for each elevation (Supplementary Material I Table 304 S2). 305 As explanatory variables, the environmental variables used were: (1) average air temperature 306 (ºC), mean air humidity (%), accumulated rainfall (mm) at each elevation (predictive variables of 307 ant activities, which varies in space and time; Kaspari et al., 2015), (2) mean elevation per transect 308 along the elevational gradient; and (3) Normalized Difference Vegetation Index (NDVI): Mean 309 NDVI and Standard Deviation of NDVI, as indirect measures of heterogeneity and structural 310 complexity of vegetation, respectively (Costanza et al., 2011; Flores et al., 2018), for each transect 311 among seasons, across elevations, and habitats. 312 In order to obtain the vegetation cover indices, we calculated the Mean NDVI (with mean 313 values of vegetation cover per environment/sampled elevation) and the Standard Deviation of NDVI 314 for each transect (Costanza et al., 2011; Flores et al., 2018) (Supplementary Material I Table S2). 315 NDVI indexes were calculated from LANDSAT 7 images (with 30 m buffer resolution) (NASA, 316 2009) from the dry season (April and June 2012) and from the rainy season (December 2012 and 317 February 2013). The images are available at the National Institute for Space Research (INPE) 318 database (http://www.dgi.inpe.br/CDSR/). The QGIS 2.18.0 software was used for the geo-319 referencing and atmospheric correction of the images (according to NASA, 2009) and later for 320 49 obtaining the Mean NDVI and Standard Deviation of NDVI for each sampling unit (Costanza et al., 321 2011; Flores et al., 2018). 322 For all variables, the multicollinearity was tested via Pearson correlation to define the 323 variables to be used in the models as explanatory predictors (i.e., all of which with correlation values 324 lower than ± 0.7; Supplementary Material I Table S3). Among these variables we selected mean 325 elevation, mean air temperature, and the indirect measure of vegetation (Mean NDVI). These 326 variables play crucial roles in species diversity patterns and ant community structure (Longino & 327 Branstetter 2019; Fernandes et al., 2016; Kaspari et al., 2015). 328 329 Data analysis 330 We estimated the sufficiency of our samples based on a sample coverage value. The sample 331 coverage measures the inventory completeness and using richness values for a specified sample size 332 or a specified level of sample coverage allows to make statistical comparisons based on these 333 estimates (Chao & Jost, 2012; Chao et al., 2013). The sample-size- and coverage-based rarefaction 334 and extrapolation curves were calculated only for Hill numbers of q = 0 (species richness) by 335 doubling the reference sample sizes and finding a common value of sample coverage between sites 336 to estimate species richness (Chao & Jost, 2012; Chao et al., 2013). We used 50 bootstraps to 337 determine confidence intervals. Analyses were done using iNEXT (Hsieh et al., 2016), available at 338 https://chao.shinyapps.io/iNEXTOnline/. 339 We evaluated patterns along three spatio-temporal dimensions: (1) one temporal dimension 340 (between dry and rainy seasons); and two spatial dimensions: (2) between habitats, campo rupestre 341 and forest environments (horizontal dimension), and (3) between elevation bands (vertical 342 dimension). To test our hypotheses, we estimated the taxonomic and functional diversity of the ant 343 communities for each dimension. The γ-diversity was partitioned into α-diversity (local scale or 344 50 transects) and β-diversity (which was partitioned into βS – between rainy and dry seasons; βH – 345 between habitats; and βV – between elevations), for both TD and FD. 346 We performed these analyses with the “adipart” (additive diversity partitioning, for 347 taxonomic approach) and “hiersimu” (hierarchical null model testing, for functional approach) 348 which works with a statistic returned by a function assessed according to a nested hierarchical 349 sampling design, both of the “vegan” package (Oksanen et al., 2018) in the R software (R Core 350 Team, 2017). The function “adipart” calculates α, β and γ-diversity values for species richness, 351 Shannon and Simpson diversity in a sampling hierarchy (our three dimensions). We used the 352 function “adipart” only to describe the species richness per dimension. The function “hiersimu” 353 works almost in the same way as “adipart”, but it has an additional argument that can be used to 354 partition functional metrics such as Rao’s quadratic entropy (a measure of functional diversity). As 355 proposed by de Bello et al., (2010), we used the “Rao” function of the R software to estimate the 356 FD and perform the partitioning (using the function “hiersimu”) into α, β (for each dimension), and 357 γ considering the Gower distance in the distribution of the species in each transect. To compare the 358 TD and FD partition results, we transformed the values of each diversity component into a 359 percentage of γ-diversity. After that, we also performed the partition of each taxonomic β-diversity 360 (β-TD) and functional β-diversity (β-FD) into turnover and nestedness components to verify if the 361 taxonomic and functional dissimilarities between ant communities in the three spatio-temporal 362 dimensions are explained by replacement or gain/loss of species/functional groups. We used the 363 Jaccard dissimilarity index and the “beta.multi” function of the “betapart” R package (Baselga et 364 al., 2013; Baselga & Orme, 2012). 365 We calculated the variation of the taxonomic (TD) and functional (FD) composition of the 366 ant community (βSOR) for each transect in all spatio-temporal dimensions. Then, we partitioned the 367 TD and FD into the components derived from (βSOR) into species turnover (βSIM) and species 368 gain/loss or nestedness (βNES) using the β-diversity partitioning method components. We used the 369 51 “beta.multi” and “functional.beta.multi” function of the R package “betapart” to partition TD an FD 370 β-diversity (Baselga & Orme, 2012). In this analysis, we used the same functional approach 371 described above. 372 Finally, we constructed two generalized linear mixed models (GLMMs) using the function 373 glmer from “lme4” R package (Bates et al., 2015) to evaluate the effects of all spatio-temporal 374 dimensions (seasonal, horizontal, and vertical) and associated environmental variables on 375 taxonomic (α-TD) and functional (α-FD) richness. We considered ant taxonomic richness (α-TD) 376 and functional richness (α-FD, i.e., Rao index) per transect as response variables, and the three 377 spatio-temporal dimensions (seasonal, horizontal, and vertical) (model 1), mean elevation, mean 378 temperature and mean NDVI (model 2) were used as fixed variables. We used “transect identity” 379 as a random variable to remove the effect of temporal pseudoreplication. The Poisson and Gaussian 380 distributions were used for taxonomic and functional variable responses. In addition, we performed 381 two generalized linear models (GLM) to test if β-TD (species turnover) and β-FD (functional 382 turnover) as response variables are related with spatio-temporal dimensions (model 1) and the same 383 fixed variables presented above (model 2). For each model, we verified whether the error 384 distribution was adequate (Crawley 2013) using the “rdiagnostic” tool of the “RT4Bio” package 385 and consequent approval of the appropriate minimum model through the gradual omission of 386 nonsignificant terms. The binomial distribution corrected for overdispersion was used for 387 taxonomic and functional variable responses. Model adequacy and distribution adjustment were 388 checked using residual analysis (Crawley, 2013). All analyzes were performed using the R software 389 (R Core Team, 2017). 390 Results 391 We recorded 2,548 ant individuals belonging to eight subfamilies, 50 genera, and 195 392 species/morphospecies. Most of the genera sampled belongs to the subfamily Myrmicinae (27 393 52 genera or 55% of the total), followed by Ponerinae (seven), Formicinae, Dolichoderinae, and 394 Dorylinae (four in each). The subfamilies with the highest species richness were Myrmicinae (101 395 species), Formicinae (33 species), Ponerinae (18 species), and Dolichoderinae (16 species). 396 Together, these four subfamilies accounted for 86.2% of the species sampled throughout the system. 397 The genera with most species were Pheidole (36 species), Camponotus (27 species), and Solenopsis 398 and Linepithema (seven species each). Considering all samples, 18 doubletons and 44 singletons 399 (31.8%) were recorded (Supplementary Material I Table S4). 400 Rarefaction-extrapolation accumulation curves suggested high sample coverage for all sites 401 (Fig. 2a, b). The high sample completeness was found for site at 900 m (sample coverage = 0.921, 402 Sobs = 90), followed by site at 800 m (sample coverage = 0.901, Sobs = 128), 1100 m (sample 403 coverage = 0.880, Sobs = 76), 1300 m (sample coverage = 0.871, Sobs = 83), 1200 m (sample coverage 404 = 0.817, Sobs = 85), and 1000 m (sample coverage = 0.800, Sobs = 73). 405 53 406 Fig 2. Rarefaction-extrapolation species accumulation curves of ant richness along the elevational 407 gradient: (a) total sampling; (b) Sampling areas. Colour code: green = 800 m; light blue = 900 m; 408 dark orange = 1000 m; dark blue = 1100 m; lilac = 1200 m; purple = 1300 m. 409 410 Partition of taxonomic and functional diversity 411 The partitioning of taxonomic diversity revealed that the relative contribution of the β-TD 412 component was 89.2% to γ-TD, with 54.3% relative to β-TDV (vertical dimension), 19.2% relative 413 to β-TDH (horizontal dimension), and 15.7% relative to β-TDS (temporal dimension). The 414 54 contribution of α-TD was only 10.8%. In relation to the relative contributions to γ-FD, the total β-415 FD component was 3.2%, being 0.7% relative to β-FDV (vertical dimension), 0.8% relative to β-416 FDH (horizontal dimension) and 1.7% relative to β-FDS (temporal dimension). The main 417 contribution to γ-FD was of the α-FD component, with 96.8% of contribution (Fig. 3). 418 419 Fig 3. The contribution of α, β1 (βS), β2 (βH) and β3 (βV) to both taxonomic diversity (γ-TD) and 420 functional diversity (γ-FD). α is the average diversity, βS is the between-season diversity (temporal 421 dimension), βH is the between-habitat diversity (horizontal dimension), and βV is the between-422 elevation diversity (vertical dimension). 423 424 When we partitioned each β-TD (i.e., β-TDH, β-TDV and β-TDS) into the turnover and 425 nestedness components, we verified that regardless of the spatio-temporal dimension evaluated, β-426 TD was mainly caused by the turnover component (Fig. 4). Despite the β-FD being very small for 427 all dimensions, nestedness was the main cause of the pattern found for temporal dimension (β-FDS), 428 while turnover was the main mechanism for horizontal and vertical dimensions (β-FDH e β-FDV) 429 (Fig. 4). 430 55 431 Fig 4. The relative contribution of turnover and nestedness to taxonomic and functional β-diversity 432 for each spatio-temporal dimension evaluated. For taxonomic diversity (TD): β-TDS = temporal 433 dimension; β-TDH = horizontal dimension; β-TDV = vertical dimension. For functional diversity 434 (FD): β-FDS = temporal dimension; β-FDH = horizontal dimension; β-FDV = vertical dimension. 435 436 Effect of elevation and environmental variables on taxonomic and functional diversity 437 Ant species richness decreased in campo rupestre and increased in forest communities as a 438 function of the increase in elevation (Table 2). The richness was also different between habitats 439 (Table 2), being greater in the campo rupestre (128 spp.) rather than in forest (113 spp.). Evident 440 effects were observed in the interaction of these two explanatory variables on species richness 441 (Table 2). Although we observed evident effects in relation to the vertical dimension, the effects 442 were negative for campo rupestre communities and positive for forest communities (Fig. 5a). There 443 were evident positive effects of the seasonal dimension on species richness, whereas the rainy 444 56 season presented greater species richness than dry season in both habitats (Fig. 5a; Table 2). We 445 observed negative effects of the interaction of seasonal and vertical dimension with decreased on 446 functional richness (α-FD), especially in dry season (Fig. 5b; Table 2). 447 448 Table 2. GLMM model’s results with α-TD (species richness) and α-FD (Rao index) as responses 449 variables. The response variables denote a model with Poisson error distribution for complete 450 models and for minimal adequate models. In all models, sample unit (transect) was the random 451 variable. d.f. = Degrees of freedom; Chisq = Type II Wald chi-square tests; Pr(>Chisq) = P-values 452 (significance codes: *** ≤ 0.001; **≤ 0.01; * ≤ 0.05). 453 Response variables Explanatory variables Chisq. Pr(>Chisq) Model 1 Dimensions α-TD – Richnessa Season 16.8828 0.03976*** d.f.residuals = 64 Mean Elevation 5.4562 0.0195* Habitat 34.4278 0.000004424*** Season : Mean Elevation 0.0000 0.9973 Season : Habitat 0.2426 0.6224 Mean Elevation : Habitat 40.1003 0.0000002413*** α-TD – Richnessb Season 16.8956 0.03949*** d.f.residuals = 66 Mean Elevation 5.4576 0.01948** Habitat 34.4313 0.000004416*** Mean Elevation : Habitat 40.1035 0.0000002409*** α-FD – Raoa Season 0.5445 0.460589 57 d.f.residual = 64 Mean Elevation 2.0474 0.152469 Habitat 2.6022 0.106717 Season : Mean Elevation 10.7650 0.001034** Season : Habitat 0.1252 0.723442 Mean Elevation : Habitat 0.8080 0.368698 α-FD – Raob Season 0.5588 0.4547273 d.f.residual = 67 Mean Elevation 1.6931 0.1931936 Season : Mean Elevation 10.9405 0.0009408*** Model 2 Environmental variables α-TD – Richnessa Mean Elevation 0.0000 0.99973 d.f.residuals = 64 Mean NDVI 0.8621 0.32525 Mean Temperature 17.2647 0.03252 Mean Elevation : Mean NDVI 3.6933 0.05463 Mean Elevation : Mean Temp 0.2590 0.61082 Mean NDVI : Mean Temp 0.1230 0.72583 α-TD – Richnessb Mean Elevation 0.0138 0.90641 d.f.residuals = 66 Mean NDVI 0.8841 0.34709 Mean Temperature 17.4700 0.02919 Mean Elevation : Mean NDVI 3.9692 0.04634 α-FD – Raoa Mean Elevation 0.0004 0.9834 58 d.f.residual = 64 Mean NDVI 0.0006 0.9811 Mean Temperature 0.0004 0.9844 Mean Elevation : Mean NDVI 0.0004 0.9832 Mean Elevation : Mean Temp 0.0026 0.9592 Mean NDVI : Mean Temp 0.0000 0.9983 a Complete model b Minimal adequate model 454 Fig 5. Taxonomic (a) and functional (b) ants’ richness (α-TD and α-FD along an elevational gradient 455 in Serra do Cipó, Minas Gerais State, Brazil. Each point denotes the diversity of an elevation per 456 season (sampling unit). 457 59 We observed negative effects on species richness due to the decreasing temperature as a 458 function of the increase in elevation and between seasons as well (Table 2; Supplementary Material 459 I Table S3). When we considered only mean elevation or mean NDVI, no evident effects were 460 observed on species richness, although we observed clear effects in the interaction of these 461 explanatory variables (Table 2). No evident effects were observed of all environmental variables 462 tested for functional richness (α-FD). 463 In relation to β-TD (taxonomic turnover) and β-FD (functional turnover) patterns, we 464 observed positive effects of the interaction between habitat and elevation for β-TD (Table 3; Fig. 465 6a) and a positive effect of habitat on β-FD patterns(Table 3; Fig. 6a). As found for functional 466 richness (α-FD), no evident effects were observed of environmental variables tested for functional 467 turnover (Table 3) and for taxonomic turnover as well. 468 469 Table 3. GLM model’s results with β-TD (species turnover) and β-FD (functional turnover) as 470 responses variables. The response variables denote a model with binomial error distribution 471 corrected for overdispersion for complete models and for minimal adequate models. d.f. = Degrees 472 of freedom; Chisq = Type II Wald chi-square tests; Pr(>Chisq) = P-values (significance codes: *** 473 ≤ 0.001; **≤ 0.01; * ≤ 0.05). 474 Response variables Explanatory variables Chisq. Pr(>Chisq) Model 1 Dimensions β-TD – Turnovera Season 1.0444 0.3068 d.f.residuals = 17 Mean Elevation 1.1433 0.28496 Habitat 2.5572 0.10979 Season : Mean Elevation 0.025 0.87441 60 Season : Habitat 2.3008 0.12931 Mean Elevation : Habitat 5.0324 0.02488* β-TD – Turnoverb Mean Elevation 1.1216 0.28957 d.f.residuals = 20 Habitat 2.5150 0.11277 Mean Elevation : Habitat 4.9515 0.02607* β-FD – Turnovera Season 0.7948 0.372638 d.f.residuals = 17 Mean Elevation 11.2007 0.000818*** Habitat 0.0406 0.840243 Season : Mean Elevation 0.0181 0.892905 Season : Habitat 2.5217 0.112292 Mean Elevation : Habitat 1.4858 0.222869 β-FD – Turnoverb Habitat 11.3830 0.0007411*** d.f.residuals = 22 Model 2 Environmental variables β-TD – Turnovera Mean Elevation 3.0212 0.08218 d.f.residuals = 17 Mean NDVI 1.7413 0.18698 Mean Temperature 1.4301 0.23174 Mean Elevation : Mean NDVI 1.3237 0.24993 Mean Elevation : Mean Temp 0.0005 0.98233 Mean NDVI : Mean Temp 3.5423 0.05982 61 β-FD – Turnoverb Mean Elevation 0.3338 0.5635 d.f.residuals = 17 Mean NDVI 0.0128 0.9099 Mean Temperature 1.0331 0.3094 Mean Elevation : Mean NDVI 2.5799 0.1082 Mean Elevation : Mean Temp 0.7195 0.3963 Mean NDVI : Mean Temp 0.0510 0.8214 a Complete model b Minimal adequate model 475 62 476 Fig 6. Taxonomic (a) and functional (b) turnover (β-TD and β-FD) of ants along an elevational 477 gradient in Serra do Cipó, Minas Gerais State, Brazil. Each point denotes the β-diversity of an 478 elevation per season (sampling unit). 479 Discussion 480 63 The main contribution to the changes in taxonomic diversity occurred along the vertical 481 dimension (i.e., elevation), followed by the horizontal dimension (i.e., habitats). Total functional β-482 diversity was lower than taxonomic β-diversity and an evident effect of the environmental variables 483 on the species richness due to the decrease in mean temperature with increasing elevation, and lower 484 Mean NDVI values in campo rupestre environments at higher elevations (lower habitat structural 485 heterogeneity). Despite the high turnover of ant species occurring across all spatio-temporal 486 dimensions, the communities are functionally or ecologically redundant, i.e., species change 487 between habitats and elevations, but the main functional characteristics and ecological functions 488 remain basically unchanged. 489 The differences in TD patterns between the horizontal and vertical dimensions can also be 490 explained by the variation in vegetation structure in both dimensions. In the same system, Fernandes 491 et al., (2016) found high β-TD of ants among elevations, and the main explanations were a decrease 492 in the structural heterogeneity and changes in climate (decrease in average temperature) along the 493 habitats and the elevational gradient. In Serra do Cipó, there is a decrease in the structural 494 heterogeneity of the grassland areas with elevation (Santos et al., 2011; Conceição et al., 2016). We 495 observed an evident shift in the climate and vegetation with increasing elevation on campo rupestre, 496 determining changes in composition (high β-TD) and decreases of ant species richness. Moreover, 497 the species richness of forest communities increased as a function of elevation rise. As seen by 498 Lasmar et al. (2020) when evaluating elevational patterns in ants diversity, it is important to 499 highlight the potential bias of different kinds of vegetation types across elevational gradients, 500 showing the importance to preserve them. Probably, this pattern reflects seasonal riverbank 501 conditions, which are subject to fluctuation in water level due to its geomorphological structure and 502 slope. In forest environments between 800 and 1200 m, considered as riparian forests, seasonal 503 flood-prone streams occur (Rosgen, 1994; Galdean et al., 2000). Flooding can act as a physical filter 504 64 for ant communities and literally drown entire colonies, diminishing species richness, and 505 abundance, interfering with the establishment of ground dwelling ants (Ballinger et al., 2007). 506 The weather conditions in the studied area are considered climatically seasonal (Ferrari et 507 al., 2016), but we observed small contribution of the seasonal dimension to changes in species 508 composition and community traits (β-TDS and β-FDS). Since we use a passive sampling method 509 and ants usually have increased activity in higher temperatures (Kaspari et al., 2015), the chance of 510 capturing ants increases in the hotter seasons of the year. Similarly, we observed fewer species in 511 the dry season than in the rainy season (decrease of α-TD) with a consequent decrease in functional 512 richness (α-FD) along the elevational gradient in the dry season. In addition, in the rainy season, we 513 verified a greater richness of specialist predators (cryptic predators), evidencing the influence of the 514 climatic variations on β-FD. 515 A possible explanation for the FD pattern found here is that species are functionally 516 redundant, i.e., although there is a great replacement of species and a decrease in species richness 517 in the vertical dimension, the main ecological functions remained along the gradient. We identified 518 ant communities with similar characteristics and non-random attribute patterns along the elevation 519 and, consequently, we can consider communities ecologically similar along these gradients as 520 demonstrated by Smith (2015), which revision paper included data from a survey in Serra do Cipó 521 (Araújo & Fernandes, 2003). Walker (1992) argues that functional redundancy (or ecological 522 redundancy) occurs when there is no effect of the variation in species composition and richness on 523 FD. Similar patterns of β-TD and β-FD of terrestrial social and non-social insects were observed in 524 the same system of the PELD/CRSC Project (e.g., termites; dung beetles, Nunes et al., 2016, 2017). 525 The explanation for the α-TD patterns found was related to environmental filters, mainly in relation 526 to vertical (negative effect of temperature and humidity variation) and horizontal dimensions 527 (variation of the vegetation heterogeneity). 528 65 When we consider the β-TD decomposition into turnover and nestedness, the vertical and 529 horizontal dimensions (β-TDV and β-TDH) contributed with more than half of the explanation in 530 the variation of species composition along the elevation gradient, being explained by the turnover 531 of species along the vertical and horizontal dimensions. These results confirm the patterns 532 previously found for terrestrial insects in the campo rupestre (Coelho et al., 2018a; da Silva et al., 533 2018; Nunes et al., 2016; 2017; Perillo et al., 2017). Although β-FDV and β-FDH were also 534 explained mainly by trait turnover, it is not possible to state that this pattern actually exists, since β-535 FD was very low and contributed with 3.2% of γ-FD, which was 96.8% explained by the α-536 component, showing a functional redundancy in all spatio-temporal dimensions addressed (Villéger 537 et al., 2013). It may be important to include more functional characteristics related to habitat and 538 thermal tolerance, especially traits with continuous measurements (as proposed by Petchey & 539 Gaston, 2006) to increase the accuracy of FD estimates. 540 In general, the high taxonomic diversity of the communities that we found resemble those 541 found for ants and other taxa in the campo rupestre (Leite et al., 2008; Fernandes et al., 2016; Nunes 542 et al., 2016, 2017; Silveira et al., 2016; Pereira et al., 2017; Perillo et al., 2017). Also, these high 543 taxonomic diversity pattern and the generic composition found for ants in the campo rupestre was 544 very similar with Cerrado (Vasconcelos et al., 2018) and Atlantic Forest biomes. The campo 545 rupestre sensu stricto fits into the OCBIL definition, as being an old, climatically buffered infertile 546 landscape (Hopper et al., 2016; Silveira et al., 2016). It is probably one of the oldest ecosystems of 547 the South American continent and most of the geological formations where they occur are 548 approximately 640 m.y.a. (millions years ago) (Alkmin, 2012). Possibly, ants already existed in the 549 campo rupestre, since the origin of the current ants happened between 115 to 135 m.y.a. (Brady et 550 al., 2006; Branstetter et al., 2017). 551 In this paper, we found that ant communities are very similar functionally despite the 552 changes in local environmental conditions among elevations, but also with compositional 553 66 restrictions regarding the characteristics of spatial dimensions (i.e., habitats and elevation bands). 554 Taxonomic differences in community composition as a consequence of the high β-TD along the 555 gradient over relatively short geographic distances emphasize the importance of conserving the 556 entire mountain, since with the loss of any part of the community, either at higher elevations or not, 557 there will be a loss of diversity, although there is no loss of the main ecological functions (or traits) 558 evaluated in this study. Comparisons of TD and FD patterns demonstrated the high importance of 559 using different diversity facets at distinct scales (Anderson et al., 2011; Barton et al., 2013; Bishop 560 et al., 2015). With information about the patterns of the multiple facets of diversity (TD, FD and 561 including phylogenetic diversity – PD; Cianciaruso et al., 2009; Anderson et al., 2011), we can 562 directly relate our observations and discoveries about biodiversity to applied ecology. For example, 563 in designing public policies or identifying priority areas for biodiversity conservation and cultural 564 diversity, one of the main objectives of the biosphere reserves in the world (Unesco, 1996). 565 566 Acknowledgments 567 We thank all friends involved in data collection during fieldwork, especially Rayana Melo, 568 Humberto Brant, Matheus Couto, and Marina Catão. We also thank Flávio Camarota and Scott 569 Powel (Cephalotes), Alexandre Ferreira (Pheidole), Rodolfo Probst (Camponotus, Myrmelachista, 570 and Octostruma), and Mayron Escárraga (Dolichoderinae) for ant identification. We also thank 571 Milton Barbosa, Flávio Camarota, Lucas Perillo, and Rafael Leitão for all useful comments on our 572 manuscript. We thank the Conselho Nacional de Desenvolvimento Científico e Tecnológico 573 (CNPq) for funding this Long-Term Ecological Research (PELD-CRSC-17 - 441515/ 2016-9), and 574 Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG - CRA - APQ-00311-15) 575 for financial support. GWF and FSN received a Research Productivity Fellowship from CNPq. 576 RRCS is supported by Pesquisa & Desenvolvimento of Agência Nacional de Energia Elétrica and 577 67 Companhia Energética de Minas Gerais (P&D ANEEL/CEMIG, PROECOS project GT-599), 578 FAPEMIG (APQ-00288-17) and CNPq (428298/2018-4) research project grants. This study was 579 financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil 580 (CAPES) - Finance Code 001. FSN, PGdS, and FSC thank the CAPES for grants. We are also 581 grateful to Cedro Têxtil, Reserva Vellozia, ICMBio-PARNA Serra do Cipó, Pousada Serra Morena 582 and Pouso do Elefante and GSG for logistical support. The author(s) declare(s) that there is no 583 conflict of interest regarding the publication of this article. 584 585 References 586 Alkmin, F. F. (2012). Serra do Espinhaço e Chapada Diamantina. In Y. Hasui, C. Carneiro, F. 587 Almeida, & A. Bartorelli (Eds.), Geologia do Brasil (pp. 236–244). São Paulo. 588 Andersen, A. N. (1995). A Classification of Australian Ant Communities, Based on Functional 589 Groups Which Parallel Plant Life-Forms in Relation to Stress and Disturbance. Journal of 590 Biogeography, 22(1), 15–29. doi: 10.2307/2846070 591 Andersen, A. N. (2019). Responses of ant communities to disturbance: Five principles for 592 understanding the disturbance dynamics of a globally dominant faunal group. Journal of 593 Animal Ecology, 88(3), 350–362. doi: 10.1111/1365-2656.12907 594 Anderson, M. J., Crist, T. O., Chase, J. M., Vellend, M., Inouye, B. D., Freestone, A. L., … 595 Swenson, N. G. (2011). Navigating the multiple meanings of β diversity: a roadmap for the 596 practicing ecologist. Ecology Letters, 14(1), 19–28. doi: 10.1111/j.1461-0248.2010.01552.x 597 Araújo, L. M., & Fernandes, G. W. (2003). Altitudinal patterns in a tropical ant assemblage and 598 variation in species richness between habitats. Lundiana, 4(2), 103–109. doi: 599 10.1016/j.epsl.2008.01.003 600 Arnan, X., Arcoverde, G. B., Pie, M. R., Ribeiro-Neto, J. D., & Leal, I. R. (2018). Increased 601 anthropogenic disturbance and aridity reduce phylogenetic and functional diversity of ant 602 68 communities in Caatinga dry forest. Science of the Total Environment, 631–632(March), 429–603 438. doi: 10.1016/j.scitotenv.2018.03.037 604 Baccaro, F. B., Feitosa, R. M., Fernandez, F., Fernandes, I. O., Izzo, T. J., Souza, J. L. P., & Solar, 605 R. (2015). Guia para os gêneros de formigas do Brasil. In Editora INPA. doi: 606 10.5281/zenodo.32912 607 Ballinger, A., Lake, P. S., & Mac Nally, R. (2007). Do terrestrial invertebrates experience 608 floodplains as landscape mosaics? Immediate and longer-term effects of flooding on ant 609 assemblages in a floodplain forest. Oecologia, 152(2), 227–238. doi: 10.1007/s00442-006-610 0648-0 611 Barden, P. (2017). Fossil ants (Hymenoptera: Formicidae): ancient diversity and the rise of modern 612 lineages. Myrmecological News, 24, 1–30. 613 Barton, P. S., Cunningham, S. A., Manning, A. D., Gibb, H., Lindenmayer, D. B., & Didham, R. K. 614 (2013). The spatial scaling of beta diversity. Global Ecology and Biogeography, 22(6), 639–615 647. doi: 10.1111/geb.12031 616 Baselga, A. (2010). Partitioning the turnover and nestedness components of beta diversity. Global 617 Ecology and Biogeography, 19(1), 134–143. doi: 10.1111/j.1466-8238.2009.00490.x 618 Baselga, A., & Orme, C. D. L. (2012). Betapart: An R package for the study of beta diversity. 619 Methods in Ecology and Evolution, 3(5), 808–812. doi: 10.1111/j.2041-210X.2012.00224.x 620 Baselga, A., Orme, D., Villeger, S., Bortoli, J. De, & Leprieur, F. (2013). Package ‘ betapart ’. R 621 package version 15.1. Retrieved from https://cran.r-622 project.org/web/packages/betapart/index.html 623 Basset, Y., Cizek, L., Cuénoud, P., Didham, R. K., Novotny, V., Ødegaard, F., … Leponce, M. 624 (2015). Arthropod distribution in a tropical rainforest: Tackling a four dimensional puzzle. 625 PLoS ONE, 10(12), 1–22. doi: 10.1371/journal.pone.0144110 626 Bates, D., Mächler, M., Bolker, B. M., & Walker, S. C. (2015). Fitting Linear Mixed-Effects Models 627 Using lme4. Journal of Statistical Software, 67(1), 1–48. doi: 10.18637/jss.v067.i01 628 69 Bestelmeyer, B. T., Agosti, D., Alonso, L., Brandão, C. R. F., & Brown Jr., W. L. (2000). Field 629 Techniques for the study of ground-dwelling ants. In Ants: Standard Methods for Measuring 630 and Monitoring Biodiversity (pp. 122–144). 631 Bishop, T. R., Robertson, M. P., Gibb, H., van Rensburg, B. J., Braschler, B., Chown, S. L., … Parr, 632 C. L. (2016). Ant assemblages have darker and larger members in cold environments. Global 633 Ecology and Biogeography, 25(12), 1489–1499. doi: 10.1111/geb.12516 634 Bishop, T. R., Robertson, M. P., van Rensburg, B. J., & Parr, C. L. (2015). Contrasting species and 635 functional beta diversity in montane ant assemblages. Journal of Biogeography, 42(9), 1776–636 1786. doi: 10.1111/jbi.12537 637 Bolton, B., Alpert, G., Ward, P. S., & Naskrecki, P. (2005). Bolton ’ S Catalogue of Ants of the 638 World. Environmental Health, (January), 9145. 639 Brady, S. G., Schultz, T. R., Fisher, B. L., & Ward, P. S. (2006). Evaluating alternative hypotheses 640 for the early evolution and diversification of ants. Proceedings of the National Academy of 641 Sciences, 103(48), 18172–18177. doi: 10.1073/pnas.0605858103 642 Brandão, C. R. F., Silva, R. R., & Delabie, J. H. C. . (2009). Formigas (Hymenoptera). In A. R. 643 Panizzi & J. R. P. . (Org. . Parra (Eds.), Bioecologia e Nutrição de Insetos. Base para o Manejo 644 Integrado de Pragas. (pp. 323-369.). Brasília, DF: EMBRAPA. 645 Branstetter, M. G., Danforth, B. N., Pitts, J. P., Faircloth, B. C., Ward, P. S., Buffington, M. L., … 646 Brady, S. G. (2017). Phylogenomic Insights into the Evolution of Stinging Wasps and the 647 Origins of Ants and Bees. Current Biology, 27(7), 1019–1025. doi: 10.1016/j.cub.2017.03.027 648 Brühl, C. a., Mohamed, M., & Linsenmair, K. E. (1999). Altitudinal distribution of leaf litter ants 649 along a transect in primary forests on Mount Kinabalu, Sabah, Malaysia. Journal of Tropical 650 Ecology, 15(3), 265–277. doi: 10.1017/S0266467499000802 651 Castro, F. S. De, Gontijo, A. B., Castro, P. D. T. A., & Ribeiro, S. P. (2012). Annual and seasonal 652 changes in the structure of litter-dwelling ant assemblages (Hymenoptera: Formicidae) in 653 Atlantic semideciduous forests. Psyche, 2012,(Article ID 959715), 12 pages. doi: 654 70 10.1155/2012/959715 655 Chao, A., & Jost, L. (2012). Coverage-based rarefaction and extrapolation: Standardizing samples 656 by completeness rather than size. Ecology, 93(12), 2533–2547. doi: 10.1890/11-1952.1 657 Chao, A., Wang, Y. T., & Jost, L. (2013). Entropy and the species accumulation curve: A novel 658 entropy estimator via discovery rates of new species. Methods in Ecology and Evolution, 4(11), 659 1091–1100. doi: 10.1111/2041-210X.12108 660 Chaves, A. V., Freitas, G. H. S., Vasconcelos, M. F., & Santos, F. R. (2015). Biogeographic patterns, 661 origin and speciation of the endemic birds from eastern Brazilian mountaintops: A review. 662 Systematics and Biodiversity, 13(1), 1–16. doi: 10.1080/14772000.2014.972477 663 Cianciaruso, M. V., Silva, I. A., & Batalha, M. A. (2009). Diversidades filogenética e funcional: 664 novas abordagens para a Ecologia de comunidades. Biota Neotropica, 9(3), 93–103. doi: 665 10.1590/S1676-06032009000300008 666 Clusella Trullas, S., van Wyk, J. H., & Spotila, J. R. (2007). Thermal melanism in ectotherms. 667 Journal of Thermal Biology, 32(5), 235–245. doi: 10.1016/j.jtherbio.2007.01.013 668 Coelho, M. S., Carneiro, M. A. A., Branco, C. A., Borges, R. A. X., & Fernandes, G. W. (2018)a. 669 Species turnover drives β-diversity patterns across multiple spatial scales of plant-galling 670 interactions in mountaintop grasslands. PLoS ONE, 13(5), 1–16. doi: 671 10.1371/journal.pone.0195565 672 Coelho, M. S., Neves, F. S., Perillo, L. N., Morellato, L. P. C., & Fernandes, G. W. (2018)b. Forest 673 archipelagos: A natural model of metacommunity under the threat of fire. Flora, 238(April), 674 244–249. doi: 10.1016/j.flora.2017.03.013 675 Conceição, A. A., Rapini, A., Carmo, F. F. do, Brito, J. C., Silva, G. A., Neves, S. P. S., & Jacobi, 676 C. M. (2016). Rupestrian Grassland Vegetation, Diversity, and Origin. In G.W. Fernandes 677 (Ed.), Ecology and Conservation of Mountaintop grasslands in Brazil (pp. 105–127). doi: 678 https://doi.org/10.1007/978-3-319-29808-5_6 679 Cooke, S. J., Sack, L., Franklin, C. E., Farrell, A. P., Beardall, J., Wikelski, M., & Chown, S. L. 680 71 (2013). What is conservation physiology? Perspectives on an increasingly integrated and 681 essential science. Conservation Physiology, 1(1), 1–23. doi: 10.1093/conphys/cot001 682 Costa, F. V., Mello, R., Lana, T. C., & Neves, F. S. (2015). Ant fauna in megadiverse mountains: 683 A checklist for the rocky grasslands. Sociobiology, 62(2), 228–245. doi: 684 10.13102/sociobiology.v62i2.228-245 685 Costanza, J. K., Moody, A., & Peet, R. K. (2011). Multi-scale environmental heterogeneity as a 686 predictor of plant species richness. Landscape Ecology, 26(6), 851–864. doi: 10.1007/s10980-687 011-9613-3 688 da Silva, P. G., Lobo, J. M., Hensen, M. C., Vaz-de-Mello, F. Z., & Hernández, M. I. M. (2018). 689 Turnover and nestedness in subtropical dung beetle assemblages along an elevational gradient. 690 Diversity and Distributions, 24(9), 1277–1290. doi: 10.1111/ddi.12763 691 De Bello, F., Lavergne, S., Meynard, C. N., Lepš, J., & Thuiller, W. (2010). The partitioning of 692 diversity: Showing Theseus a way out of the labyrinth. Journal of Vegetation Science, 21(5), 693 992–1000. doi: 10.1111/j.1654-1103.2010.01195.x 694 de Bello, F., Lavorel, S., Albert, C. H., Thuiller, W., Grigulis, K., Dolezal, J., … Lepš, J. (2011). 695 Quantifying the relevance of intraspecific trait variability for functional diversity. Methods in 696 Ecology and Evolution, 2(2), 163–174. doi: 10.1111/j.2041-210X.2010.00071.x 697 Domingues, S. A., Karez, C. S., Biondini, I. V. F., & Andrade, M. Â. (2011). Instrumentos 698 económicos de gestión ambiental en la Reserva de Biosfera de la Serra do Espinhaço. UNESCO 699 Regional Office for Science for Latin America and the Caribbean, p. 36. doi: 700 10.1109/TDEI.2009.5211872 701 Dunn, R. R., Agosti, D., Andersen, A. N., Arnan, X., Bruhl, C. A., Cerdá, X., … Sanders, N. J. 702 (2009). Climatic drivers of hemispheric asymmetry in global patterns of ant species richness. 703 Ecology Letters, 12(4), 324–333. doi: 10.1111/j.1461-0248.2009.01291.x 704 Esquivel-Muelbert, A., Baker, T. R., Dexter, K. G., Lewis, S. L., ter Steege, H., Lopez-Gonzalez, 705 G., … Phillips, O. L. (2017). Seasonal drought limits tree species across the Neotropics. 706 72 Ecography, 40(5), 618–629. doi: 10.1111/ecog.01904 707 Feener, D. H., Lighton, J. R. B., & Bartholomew, G. A. (1988). Curvilinear Allometry, Energetics 708 and Foraging Ecology: A Comparison of Leaf-Cutting Ants and Army Ants. Functional 709 Ecology, 2(4), 509–520. doi: 10.2307/2389394 710 Fernandes, G. Wilson, Almeida, H. A., Nunes, C. A., Xavier, J. H. A., Cobb, N. S., Carneiro, M. A. 711 A., … Beirão, M. V. (2016). Cerrado to Rupestrian Grasslands: Patterns of Species 712 Distribution and the Forces Shaping Them Along an Altitudinal Gradient. In G. W. Fernandes 713 (Ed.), Ecology and Conservation of Mountaintop grasslands in Brazil (pp. 345–377). doi: 714 https://doi.org/10.1007/978-3-319-29808-5_15 715 Fernandes, G. Wilson, Barbosa, N. P. U., Alberton, B., Barbieri, A., Dirzo, R., Goulart, F., … Solar, 716 R. R. C. (2018). The deadly route to collapse and the uncertain fate of Brazilian rupestrian 717 grasslands. Biodiversity and Conservation, 27(10), 2587–2603. doi: 10.1007/s10531-018-718 1556-4 719 Fernandes, G.W. (2016). The Megadiverse Rupestrian Grassland. In G. W. Fernandes (Ed.), 720 Ecology and Conservation of Mountaintop grasslands in Brazil (pp. 3–14). Springer, Cham. 721 Ferrari, L. T., Schaefer, C. E. G. R., Fernandes, R. B. A., Mendonça, B. A. F., Gjorup, D. F., Corrêa, 722 G. R., & E.O., S. (2016). Thermic and Hydric Dynamics of Ironstone (Canga) and Quartzite 723 Rupestrian Grasslands in the Quadrilátero Ferrífero: The Ecological Importance of Water. In 724 G.W. Fernandes (Ed.), Ecology and Conservation of Mountaintop Grasslands in Brazil (pp. 725 72–85). 726 Fichaux, M., Béchade, B., Donald, J., Weyna, A., Delabie, J. H. C., Murienne, J., … Orivel, J. 727 (2019). Habitats shape taxonomic and functional composition of Neotropical ant assemblages. 728 Oecologia, 189(2), 501–513. doi: 10.1007/s00442-019-04341-z 729 Flores, L. M. A., Zanette, L. R. S., & Araujo, F. S. (2018). Effects of habitat simplification on 730 assemblages of cavity nesting bees and wasps in a semiarid neotropical conservation area. 731 Biodiversity and Conservation, 27(2), 311–328. doi: 10.1007/s10531-017-1436-3 732 73 Galdean, N., Callisto, M., & Barbosa, F. (2000). Lotic ecosystems of serra do cipó, southeast brazil: 733 Water quality and a tentative classification based on the benthic macroinvertebrate community. 734 Aquatic Ecosystem Health and Management, 3(4), 545–552. doi: 735 10.1080/14634980008650691 736 Gaston, K. J. (2000). Global patterns in biodiversity. Nature, 405(6783), 220–227. doi: 737 10.1038/35012228 738 Giulietti, A. M., Pirani, J. R., & Harley, R. M. (1997). Centres of plant diversity: a guide and strategy 739 for their conservation. Oxford: Information Press, 397–4004. 740 Guénard, B., Weiser, M. D., Gómez, K., Narula, N., & Economo, E. P. (2017). The Global Ant 741 Biodiversity Informatics (GABI) database: Synthesizing data on the geographic distribution of 742 ant species (Hymenoptera: Formicidae). Myrmecological News, 24, 83–89. 743 Hopper, S. D., Silveira, F. A. O., & Fiedler, P. L. (2016). Biodiversity hotspots and Ocbil theory. 744 Plant and Soil, 403(1–2), 167–216. doi: 10.1007/s11104-015-2764-2 745 Hsieh, T. C., Ma, K. H., & Chao, A. (2016). Introduction to iNEXT ( R package ): Excerpt from 746 iNEXT User ’ s Guide. 1–8. 747 Johnson, R. A. (2015). A taxonomic revision of South American species of the seed-harvester ant 748 genus Pogonomyrmex (Hymenoptera: Formicidae). Part I. Zootaxa, 4029(1), 001–142. doi: 749 10.11646/zootaxa.4029.1.1 750 Jost, L. (2007). Concepts & Synthesis. Ecology, 88(10), 2427–2439. doi: 10.1890/07-1861.1 751 Kaspari, M, & Weiser, M. D. (1999). The size–grain hypothesis and interspecific scaling in ants. 752 Functional Ecology, 13(4), 530–538. doi: 10.1046/j.1365-2435.1999.00343.x 753 Kaspari, Michael, Clay, N. A., Lucas, J., Yanoviak, S. P., & Kay, A. (2015). Thermal adaptation 754 generates a diversity of thermal limits in a rainforest ant community. Global Change Biology, 755 21(3), 1092–1102. doi: 10.1111/gcb.12750 756 Kaspari, Michael, Yuan, M., & Alonso, L. (2003). Spatial Grain and the Causes of Regional 757 Diversity Gradients in Ants. The American Naturalist, 161(3), 459–477. doi: 10.1086/367906 758 74 Körner, C. (2007). The use of “altitude” in ecological research. Trends in Ecology and Evolution, 759 22(11), 569–574. doi: 10.1016/j.tree.2007.09.006 760 Lasmar, C. J., Ribas, C. R., Louzada, J., Queiroz, A. C. M., Feitosa, R. M., Imata, M. M. G., … 761 Domingos, D. Q. (2020). Disentangling elevational and vegetational effects on ant diversity 762 patterns. Acta Oecologica, 102(November 2019), 103489. doi: 10.1016/j.actao.2019.103489 763 Lassau, S. a., & Hochuli, D. F. (2004). Effects of habitat complexity on ant assemblages. 764 Ecography, 27(2), 157–164. doi: 10.1111/j.0906-7590.2004.03675.x 765 Leal, I. R., Filgueiras, B. K. C., Gomes, J. P., Iannuzzi, L., & Andersen, A. N. (2012). Effects of 766 habitat fragmentation on ant richness and functional composition in Brazilian Atlantic forest. 767 Biodiversity and Conservation, 21(7), 1687–1701. doi: 10.1007/s10531-012-0271-9 768 Leite, F. S. F., Juncá, F. A., & Eterovick, P. C. (2008). Status do conhecimento, endemismo e 769 conservação de anfíbios anuros da Serra do Espinhaço, Brasil. Megadiversidade, 4(1), 158–770 176. doi: 10.3905/jpm.2003.319889 771 Leponce, M., Theunis, L., Delabie, J. H. C., & Roisin, Y. (2004). Scale dependence of diversity 772 measures in a leaf-litter ant assemblage. Ecography, 27(2), 253–267. doi: 10.1111/j.0906-773 7590.2004.03715.x 774 Liang, B. (2003). The accuracy of hedge fund returns: Auditing makes a real difference. Journal of 775 Portfolio Management, 29(3), 111–122. doi: 10.3905/jpm.2003.319889 776 Longino, J. T., & Branstetter, M. G. (2019). The truncated bell: an enigmatic but pervasive 777 elevational diversity pattern in Middle American ants. Ecography, 42(2), 272–283. doi: 778 10.1111/ecog.03871 779 Longino, J. T., & Colwell, R. K. (2011). Density compensation, species composition, and richness 780 of ants on a neotropical elevational gradient. Ecosphere, 2(3), art29. doi: 10.1890/ES10-781 00200.1 782 Marques, T., Espiríto-Santo, M. M., Neves, F. S., & Schoereder, J. H. (2017). Ant assemblage 783 structure in a secondary tropical dry forest: The role of ecological succession and seasonality. 784 75 Sociobiology, 64(3), 261–275. doi: 10.13102/sociobiology.v64i3.1276 785 McCain, C., & Grytnes, J. (2010). Elevational gradients in species richness. Encyclopedia of Life 786 Sciences (ELS), (August 2016), 1–10. doi: 10.1002/9780470015902.a0022548 787 Meguro, M., Pirani, J. R., Mello-Silva, R. de, & Giulietti, A. M. (1996). Caracterização florística e 788 estrutural de matas ripárias e capões de altitude da Serra do Cipó, Minas Gerais. Boletim de 789 Botânica Da Universidade de São Paulo, 15(1), 13–29. Retrieved from 790 https://www.revistas.usp.br/bolbot/article/view/57939 791 Mittermeier, R. A., van Dijk, P. P., Rhodin, A. G. J., & Nash, S. D. (2004). Hotspots Revisited: 792 Earth’s Biologically Richest and Most Endangered Ecoregions. Chelonian Conservation and 793 Biology, 14(1), 200. doi: 10.2744/ccab-14-01-2-10.1 794 Montine, P. S. M., Viana, N. F., Almeida, F. S., Dáttilo, W., Santanna, A. S., Martins, L., & Vargas, 795 A. B. (2014). Seasonality of epigaeic ant communities in a Brazilian atlantic rainforest. 796 Sociobiology, 61(2), 178–183. doi: 10.13102/sociobiology.v61i2.178-183 797 Mota, G. S., Luz, G. R., Mota, N. M., Silva Coutinho, E., das Dores Magalhães Veloso, M., 798 Fernandes, G. W., & Nunes, Y. R. F. (2018). Changes in species composition, vegetation 799 structure, and life forms along an altitudinal gradient of rupestrian grasslands in south-eastern 800 Brazil. Flora, 238, 32–42. doi: 10.1016/j.flora.2017.03.010 801 Munyai, T. C., & Foord, S. H. (2012). Ants on a mountain: Spatial, environmental and habitat 802 associations along an altitudinal transect in a centre of endemism. Journal of Insect 803 Conservation, 16(5), 677–695. doi: 10.1007/s10841-011-9449-9 804 NASA. (2009). Landsat 7 science data users handbook. Retrieved June 8, 2017, from National 805 Aeronautics and Space Administration Landsat website: 806 http://glovis.usgs.gov/%0Ahttp://edcsns17.cr.usgs.gov/EarthExplorer/%0Ahttp://www.landc807 over.org/index.shtml%0Ahttp://glovis.usgs.gov/%0Ahttp://edcsns17.cr.usgs.gov/EarthExplor808 er/%0Ahttp://www.landcover.org/index.shtml%0Ahttp://landsathandbook.gsfc.nasa.go 809 Nation, J. L. (2008). Integument: Structure and Function. In J. L. Capinera (Ed.), Encyclopedia of 810 76 Entomology (pp. 2015–2019). doi: 10.1007/978-1-4020-6359-6_1557 811 Nowrouzi, S., Andersen, A. N., Macfadyen, S., Staunton, K. M., Vanderwal, J., & Robson, S. K. A. 812 (2016). Ant diversity and distribution along elevation gradients in the Australian Wet tropics: 813 The importance of seasonal moisture stability. PLoS ONE, 11(4). doi: 814 10.1371/journal.pone.0153420 815 Nunes, Cássio A., Quintino, A. V., Constantino, R., Negreiros, D., Reis Júnior, R., & Fernandes, G. 816 W. (2017). Patterns of taxonomic and functional diversity of termites along a tropical 817 elevational gradient. Biotropica, 49(2), 186–194. doi: 10.1111/btp.12365 818 Nunes, Cássio Alencar, Braga, R. F., Figueira, J. E. C., Neves, F. de S., & Fernandes, G. W. (2016). 819 Dung Beetles along a Tropical Altitudinal Gradient: Environmental Filtering on Taxonomic 820 and Functional Diversity. PLOS ONE, 11(6), e0157442. doi: 10.1371/journal.pone.0157442 821 Oksanen, J., Blanchet, F. G., Kindt, R., Legen-, P., Minchin, P. R., Hara, R. B. O., … Wagner, H. 822 (2018). Package ‘ vegan ’ - Community Ecology Package. R package version 2.5-5. Retrieved 823 from https://cran.r-project.org/web/packages/vegan/index.html 824 Paolucci, L. N., Maia, M. L. B., Solar, R. R. C., Campos, R. I., Schoereder, J. H., & Andersen, A. 825 N. (2016). Fire in the Amazon: impact of experimental fuel addition on responses of ants and 826 their interactions with myrmecochorous seeds. Oecologia, 182(2), 335–346. doi: 827 10.1007/s00442-016-3638-x 828 Parmesan, C. (2006). Ecological and Evolutionary Responses to Recent Climate Change. Annual 829 Review of Ecology, Evolution, and Systematics, 37(1), 637–669. doi: 830 10.1146/annurev.ecolsys.37.091305.110100 831 Parr, C. L., Dunn, R. R., Sanders, N. J., Weiser, M. D., Photakis, M., Bishop, T. R., … Gibb, H. 832 (2017). GlobalAnts: a new database on the geography of ant traits (Hymenoptera: Formicidae). 833 Insect Conservation and Diversity, 10(1), 5–20. doi: 10.1111/icad.12211 834 Pecl, G. T., Araújo, M. B., Bell, J. D., Blanchard, J., Bonebrake, T. C., Chen, I. C., … Williams, S. 835 E. (2017). Biodiversity redistribution under climate change: Impacts on ecosystems and human 836 77 well-being. Science, 355(6332). doi: 10.1126/science.aai9214 837 Pereira, G. C. N., Coelho, M. S., Beirão, M. do V., Braga, R. F., & Fernandes, G. W. (2017). 838 Diversity of fruit-feeding butterflies in a mountaintop archipelago of rainforest. PLoS ONE, 839 12(6), 1–20. doi: 10.1371/journal.pone.0180007 840 Perillo, L. N., Neves, F. D. S., Antonini, Y., & Martins, R. P. (2017). Compositional changes in bee 841 and wasp communities along Neotropical mountain altitudinal gradient. PLoS ONE, 12(7), 1–842 14. doi: 10.1371/journal.pone.0182054 843 Petchey, O. L., & Gaston, K. J. (2006). Functional diversity: Back to basics and looking forward. 844 Ecology Letters, 9(6), 741–758. doi: 10.1111/j.1461-0248.2006.00924.x 845 R Core Team. (2017). R Core Team (2017). R: A language and environment for statistical 846 computing. R Foundation for Statistical Computing, Vienna, Austria. URL Http://Www.R-847 Project.Org/. 848 Rosgen, D. L. (1994). Classification of natural rivers. Catena, 22, 169–199. 849 Santos, M. F., Serafim, H., & Sano, P. T. (2011). Fisionomia e composição da vegetação florestal 850 na Serra do Cipó, MG, Brasil. Acta Botanica Brasilica, 25(4), 793–814. doi: 10.1590/S0102-851 33062011000400007 852 Silveira, F. A. O., Barbosa, M., Beiroz, W., Callisto, M., Macedo, D. R., Morellato, L. P. C., … 853 Fernandes, G. W. (2019). Tropical mountains as natural laboratories to study global changes: 854 A long-term ecological research project in a megadiverse biodiversity hotspot. Perspectives in 855 Plant Ecology, Evolution and Systematics, 38, 64–73. doi: 10.1016/j.ppees.2019.04.001 856 Silveira, F. A. O., Negreiros, D., Barbosa, N. P. U., Buisson, E., Carmo, F. F., Carstensen, D. W., 857 … Lambers, H. (2016). Ecology and evolution of plant diversity in the endangered campo 858 rupestre: a neglected conservation priority. Plant Soil, 403(1–2), 129–152. doi: 859 10.1007/s11104-015-2637-8 860 Smith, A. R. (1978). Color gamut transform pairs. ACM SIGGRAPH Computer Graphics, 12(3), 861 12–19. doi: 10.1145/965139.807361 862 78 Smith, M. A. (2015). Ants, elevation, phylogenetic diversity and community structure. Ecosphere, 863 6(11), art221. doi: 10.1890/ES14-00473.1 864 Solar, R. R. de C., Barlow, J., Andersen, A. N., Schoereder, J. H., Berenguer, E., Ferreira, J. N., & 865 Gardner, T. A. (2016). Biodiversity consequences of land-use change and forest disturbance 866 in the Amazon: A multi-scale assessment using ant communities. Biological Conservation, 867 197, 98–107. doi: 10.1016/j.biocon.2016.03.005 868 Terblanche, J. S. (2012). Thermal relations. In R. F. Chapman, S. J. Simpson, & A. E. Douglas 869 (Eds.), The Insects: Structure and Function (Fifth edit, pp. 588–621). New York: Cambridge 870 University Press. 871 Tiede, Y., Schlautmann, J., Donoso, D. A., Wallis, C. I. B., Bendix, J., Brandl, R., & Farwig, N. 872 (2017). Ants as indicators of environmental change and ecosystem processes. Ecological 873 Indicators, 83, 527–537. doi: 10.1016/j.ecolind.2017.01.029 874 Tuomisto, H. (2010). A diversity of beta diversities: Straightening up a concept gone awry. Part 1. 875 Defining beta diversity as a function of alpha and gamma diversity. Ecography, 33(1), 2–22. 876 doi: 10.1111/j.1600-0587.2009.05880.x 877 Unesco. (1996). La Estrategia de Sevilla y el Marco Estatutario de la Red Mundial (p. 22). p. 22. 878 Vasconcelos, H. L., Maravalhas, J. B., Feitosa, R. M., Pacheco, R., Neves, K. C., & Andersen, A. 879 N. (2018). Neotropical savanna ants show a reversed latitudinal gradient of species richness, 880 with climatic drivers reflecting the forest origin of the fauna. Journal of Biogeography, 45(1), 881 248–258. doi: 10.1111/jbi.13113 882 Vellend, M. (2016). The Theory of Ecological Communities (MPB-57). Retrieved from 883 https://books.google.com.br/books?id=2Yn8CwAAQBAJ 884 Villéger, S., Grenouillet, G., & Brosse, S. (2013). Decomposing functional β-diversity reveals that 885 low functional β-diversity is driven by low functional turnover in European fish assemblages. 886 Global Ecology and Biogeography, 22(6), 671–681. doi: 10.1111/geb.12021 887 Villéger, S., Mason, N. W. H., & Mouillot, D. (2008). New Multidimensional Functional Diversity 888 79 Indices for a Multifaceted Framework in Functional Ecology Published by : Ecological Society 889 of America content in a trusted digital archive. Ecology, 89(8), 2290–2301. doi: 10.1890/07-890 1206.1 891 Violle, C., Navas, M. L., Vile, D., Kazakou, E., Fortunel, C., Hummel, I., & Garnier, E. (2007). Let 892 the concept of trait be functional! Oikos, 116(5), 882–892. doi: 10.1111/j.0030-893 1299.2007.15559.x 894 Walker, B. H. (1992). Biodiversity and Ecological Redundancy Biodiversity and Ecological 895 Redundancy. Conservation Biology, 6(1), 18–23. 896 Weber, N. A. (1938). The biology of the fungus-growing ants. Part IV. Additional new forms. Part 897 V. The Attini of Bolivia. Biología de las hormigas cultivadoras de hongos. Parte IV. Nuevas 898 formas adicionales. Part V. Las Attini de Bolivia. Revista de Entomologia, 7, 154–206. 899 Whittaker, R. (1960). Vegetation of the Siskiyou, Oregon and California. Ecological Monographs, 900 30(3), 279–338. 901 Wills, B. D., Powell, S., Rivera, M. D., & Suarez, A. V. (2017). Correlates and Consequences of 902 Worker Polymorphism in Ants. Annual Review of Entomology, 63(1), 575–598. doi: 903 10.1146/annurev-ento-020117-043357 904 Yates, M. L., Andrew, N. R., Binns, M., & Gibb, H. (2014). Morphological traits: predictable 905 responses to macrohabitats across a 300 km scale. PeerJ, 2, e271. doi: 10.7717/peerj.271 906 907 80 Supporting information 1 Table S1. Species captured and traits used to calculate functional diversity (FD): (1) morphological traits (a) WL: Weber’s length (μm); (b) HFL: Hind femur length (μm); (c) 2 ML: Mandible length (μm); (d) V: measure of color brightness (%); (e) Workers polymorphism: categorical (monomorphic, dimorphic and polymorphic); (f) integument 3 sculpture (ordinal data (smooth/often shine, intermediate and texturized); and (2) ecological traits or life history traits (a): functional groups based on global-scale responses of 4 ants to environmental stress and disturbance (Andersen, 1995; Leal et al,. 2012; Paolucci et al., 2016). AA = Army Ants; AD = Arboreal Dominant; AP = Arboreal Predator; 5 AS = Arboreal Subordinate; CO = Cryptic Omnivores; CP = Cryptic Predators; DD = Dominant Dolichoderinae; EO = Epigeic Omnivores; EP = Epigeic Predators; Hatt = High 6 Attini; Latt = Low Attini; OPP = Opportunist; SC = Subordinate Camponotini SH = Seed Harvester. 7 Species WL HFL ML V Workers’ polymorphism Integument sculpture Functional groups Acromyrmex balzani 2558.66 1.04 0.40 0.54 polymorphic intermediate Hatt Acromyrmex coronatus 1759.29 1.11 0.49 0.38 polymorphic intermediate Hatt Acromyrmex sp2 2235.66 1.13 0.51 0.62 polymorphic intermediate Hatt Acromyrmex sp5 2411.54 1.17 0.50 0.34 polymorphic intermediate Hatt Acromyrmex subterraneus 3107.80 1.12 0.46 0.62 polymorphic intermediate Hatt Anochetus inermis 1378.50 0.69 0.45 0.48 monomorphic intermediate EP Apterostigma gr. pilosum sp1 1281.67 1.08 0.47 0.36 monomorphic intermediate Latt Apterostigma gr. pilosum sp2 1602.61 0.95 0.30 0.28 monomorphic intermediate Latt Apterostigma gr. pilosum sp3 1272.44 0.97 0.30 0.34 monomorphic intermediate Latt Atta sexdens rubropilosa 3788.38 1.41 0.45 0.34 polymorphic intermediate Hatt Brachymyrmex pictus 371.70 0.89 0.33 0.41 monomorphic smooth OPP Brachymyrmex prox. Cordemoyi 475.97 0.89 0.36 0.26 monomorphic smooth OPP Brachymyrmex sp4 451.42 0.75 0.47 0.27 monomorphic smooth OPP Brachymyrmex sp5 704.81 0.87 0.44 0.63 monomorphic smooth OPP Camponotus (Hypercolobopsis) sp3 1775.89 0.74 0.25 0.62 dimorphic intermediate SC Camponotus (Hypercolobopsis) sp6 1732.71 0.82 0.29 0.52 dimorphic intermediate SC Camponotus (Hypercolobopsis) sp7 1742.19 0.82 0.27 0.63 dimorphic intermediate SC Camponotus (Myrmaphaenus) sp1 2452.25 0.83 0.36 0.14 dimorphic intermediate SC Camponotus (Myrmaphaenus) sp5 1434.07 0.88 0.30 0.12 dimorphic intermediate SC Camponotus (Tanaemyrmex) sp2 1931.61 0.87 0.26 0.11 dimorphic intermediate SC Camponotus arboreus 2152.96 0.79 0.27 0.14 dimorphic intermediate SC Camponotus atriceps 2590.19 0.90 0.28 0.46 dimorphic intermediate SC 81 Species WL HFL ML V Workers’ polymorphism Integument sculpture Functional groups Camponotus blandus 2031.42 0.84 0.29 0.12 dimorphic intermediate SC Camponotus cingulatus 2699.53 1.00 0.28 0.38 dimorphic intermediate SC Camponotus crassus 1602.19 0.87 0.32 0.13 dimorphic intermediate SC Camponotus latangulus 1446.18 0.90 0.36 0.38 dimorphic intermediate SC Camponotus lespesii 4288.06 0.97 0.25 0.20 dimorphic intermediate SC Camponotus leydigi 2817.51 0.80 0.28 0.45 dimorphic intermediate SC Camponotus melanoticus 2856.95 0.90 0.26 0.24 dimorphic intermediate SC Camponotus novogranadensis 1608.14 0.73 0.32 0.15 dimorphic intermediate SC Camponotus renggeri 3157.38 0.89 0.28 0.14 dimorphic intermediate SC Camponotus rufipes 3196.31 0.92 0.27 0.14 dimorphic intermediate SC Camponotus senex 1628.06 0.95 0.33 0.15 dimorphic intermediate SC Camponotus sericeiventris 3694.48 0.97 0.26 0.14 dimorphic intermediate SC Camponotus sexguttatus 1889.19 0.84 0.27 0.13 dimorphic intermediate SC Camponotus sp14 1956.54 0.78 0.27 0.18 dimorphic intermediate SC Camponotus sp19 3493.04 0.80 0.30 0.20 dimorphic intermediate SC Camponotus sp20 3174.94 0.85 0.26 0.46 dimorphic intermediate SC Camponotus sp4 3679.38 0.92 0.23 0.35 dimorphic intermediate SC Camponotus sp5 3179.61 0.92 0.27 0.54 dimorphic intermediate SC Camponotus vittatus 2690.20 0.93 0.26 0.61 dimorphic intermediate SC Cardiocondyla cf. obscurior 501.44 0.59 0.37 0.68 monomorphic intermediate AD Carebara urichi 509.08 0.63 0.47 0.60 dimorphic intermediate CO Cephalotes atratus 4140.30 0.86 0.27 0.13 dimorphic intermediate AS Cephalotes pusillus 1338.89 0.70 0.31 0.16 dimorphic intermediate AS Crematogaster acuta 985.77 0.89 0.56 0.30 polymorphic intermediate AD Crematogaster brasiliensis 786.65 1.06 0.50 0.23 polymorphic intermediate AD Crematogaster prox. erecta sp1 758.01 0.83 0.45 0.34 polymorphic intermediate AD Crematogaster prox. obscurata sp1 637.96 0.80 0.40 0.63 polymorphic intermediate AD Crematogaster sp6 799.19 0.98 0.47 0.32 polymorphic intermediate AD Cyphomyrmex gr. rimosus sp1 830.77 0.78 0.43 0.51 monomorphic intermediate Latt Cyphomyrmex gr. rimosus sp2 1002.22 0.88 0.39 0.41 monomorphic intermediate Latt Cyphomyrmex gr. rimosus sp3 858.61 0.83 0.40 0.39 monomorphic intermediate Latt Dolichoderus bispinosus 1911.95 0.94 0.35 0.10 monomorphic intermediate AD Dolichoderus diversus 1886.54 0.90 0.28 0.22 monomorphic smooth AD Dolichoderus lutosus 1391.76 0.77 0.33 0.66 monomorphic smooth AD Dorymyrmex brunneus 1224.75 1.00 0.37 0.39 monomorphic smooth OPP Dorymyrmex goeldii 1198.94 1.02 0.33 0.29 monomorphic smooth OPP Dorymyrmex pyramicus 1015.13 1.03 0.39 0.81 monomorphic smooth OPP 82 Species WL HFL ML V Workers’ polymorphism Integument sculpture Functional groups Dorymyrmex sp5 785.73 1.05 0.36 0.62 monomorphic smooth OPP Eciton cf. vagans 2813.05 0.99 0.35 0.36 polymorphic intermediate AA Eciton mexicanum 2972.83 1.16 1.05 0.51 polymorphic intermediate AA Ectatomma brunneum 3602.40 0.77 0.39 0.17 monomorphic textured EP Ectatomma edentatum 2760.84 0.76 0.41 0.22 monomorphic textured EP Ectatomma opaciventre 4758.47 0.93 0.40 0.53 monomorphic textured EP Ectatomma permagnum 3607.01 0.77 0.41 0.16 monomorphic textured EP Ectatomma planidens 2368.66 0.72 0.41 0.40 monomorphic textured EP Ectatomma tuberculatum 3546.31 0.81 0.45 0.39 monomorphic textured AP Forelius brasiliensis 839.50 1.18 0.43 0.61 monomorphic smooth DD Forelius maranhaoensis 685.37 1.21 0.52 0.35 monomorphic smooth DD Forelius sp2 536.26 0.91 0.42 0.54 monomorphic smooth DD Gnamptogenys gp striatula sp.n. A 1085.81 0.57 0.43 0.46 monomorphic textured EP Gnamptogenys sp3 1889.60 0.66 0.40 0.14 monomorphic textured EP Gnamptogenys striatula 1467.21 0.84 0.38 0.21 monomorphic textured EP Gnamptogenys sulcata 1297.90 0.64 0.41 0.25 monomorphic textured EP Heteroponera sp1 1192.50 0.63 0.28 0.27 monomorphic textured EP Hylomyrma balzani 1076.35 0.76 0.52 0.25 monomorphic textured EO Hylomyrma prox. Reitteri 1164.23 0.80 0.53 0.24 monomorphic textured EO Hylomyrma sp4 1049.22 0.72 0.50 0.28 monomorphic textured EO Hypoponera distinguenda 935.95 0.54 0.36 0.33 monomorphic smooth CP Hypoponera sp1 1144.60 0.58 0.39 0.25 monomorphic smooth CP Hypoponera sp3 1629.02 0.55 0.44 0.23 monomorphic smooth CP Hypoponera sp4 790.79 0.55 0.37 0.39 monomorphic smooth CP Hypoponera sp5 1563.29 0.65 0.34 0.23 monomorphic smooth CP Kalathomyrmex emeryi 860.48 0.68 0.50 0.67 monomorphic intermediate Latt Labidus coecus 1433.98 0.98 0.37 0.30 polymorphic intermediate AA Labidus praedator 1602.77 1.00 0.47 0.23 polymorphic intermediate AA Leptogenys aff. Górgona 1550.62 0.69 0.29 0.15 monomorphic smooth EP Leptogenys crudelis 2933.08 0.68 0.27 0.11 monomorphic smooth EP Linepithema aztecoides 730.18 0.83 0.35 0.36 monomorphic smooth EO Linepithema cerradense 732.48 0.84 0.34 0.65 monomorphic smooth EO Linepithema cf. pulex 864.22 0.67 0.33 0.59 monomorphic smooth EO Linepithema gallardoi 757.60 0.77 0.33 0.39 monomorphic smooth EO Linepithema iniquum 858.67 0.75 0.34 0.38 monomorphic smooth EO Linepithema micans 813.12 0.80 0.36 0.54 monomorphic smooth EO Megalomyrmex sp1 1157.27 0.90 0.34 0.46 monomorphic smooth EO 83 Species WL HFL ML V Workers’ polymorphism Integument sculpture Functional groups Mycetarotes sp1 1063.75 0.89 0.47 0.72 monomorphic intermediate Latt Mycetophylax gr. strigatus sp4 829.21 0.57 0.44 0.47 monomorphic intermediate Latt Mycetophylax gr. strigatus sp5 795.90 1.04 0.74 0.70 monomorphic intermediate Latt Mycetophylax gr. strigatus sp6 730.09 0.81 0.51 0.60 monomorphic intermediate Latt Mycetophylax lectus 712.06 0.81 0.45 0.57 monomorphic intermediate Latt Mycocepurus goeldii 920.92 0.85 0.49 0.39 monomorphic intermediate Latt Myrmelachista catharinae 688.57 0.79 0.32 0.13 monomorphic smooth AS Myrmicocrypta sp1 922.00 0.84 0.45 0.66 monomorphic intermediate Latt Neivamyrmex cf. swainsonii 596.01 0.53 0.30 0.54 polymorphic intermediate AA Neivamyrmex pseudops 1356.80 1.12 0.42 0.46 polymorphic intermediate AA Neivamyrmex sp2 1099.18 0.71 0.35 0.46 polymorphic intermediate AA Neoponera crenata 2138.10 0.74 0.45 0.15 monomorphic intermediate EP Neoponera latinoda 2814.77 0.76 0.48 0.15 monomorphic smooth EP Neoponera villosa 4643.72 0.79 0.48 0.14 monomorphic smooth AP Nomamyrmex esenbeckii 2531.03 0.85 0.38 0.54 polymorphic intermediate AA Nylanderia sp1 783.41 0.95 0.40 0.57 monomorphic smooth OPP Ochetomyrmex semipolitus 495.81 0.69 0.48 0.69 monomorphic intermediate EO Octostruma balzani 576.86 0.70 0.32 0.50 monomorphic textured CP Octostruma iheringi 759.18 0.72 0.31 0.45 monomorphic textured CP Octostruma stenognatha 675.45 0.70 0.30 0.42 monomorphic textured CP Odontomachus bauri 4343.87 0.93 0.47 0.27 monomorphic intermediate EP Odontomachus brunneus 2551.31 0.85 0.45 0.29 monomorphic intermediate EP Odontomachus chelifer 4888.06 0.89 0.46 0.31 monomorphic intermediate EP Odontomachus meinerti 2426.02 0.81 0.46 0.39 monomorphic intermediate EP Oxyepoecus prox. bruschi sp1 566.91 0.64 0.46 0.30 monomorphic intermediate EP Oxyepoecus prox. bruschi sp2 628.51 0.63 0.42 0.37 monomorphic intermediate EP Oxyepoecus sp5 496.43 0.62 0.42 0.39 monomorphic intermediate EP Pachycondyla harpax 2632.19 0.64 0.43 0.15 monomorphic intermediate EP Pachycondyla striata 4269.56 0.73 0.44 0.15 monomorphic intermediate EP Pheidole ambígua 839.68 0.94 0.46 0.50 dimorphic intermediate EO Pheidole capilata 839.16 0.92 0.50 0.27 dimorphic intermediate EO Pheidole dorsata 467.19 0.82 0.50 0.39 dimorphic intermediate EO Pheidole gertrudae 867.83 1.08 0.52 0.41 dimorphic smooth EO Pheidole jelskii 1142.97 1.12 0.39 0.32 dimorphic intermediate EO Pheidole oxyops 1246.70 1.22 0.41 0.32 dimorphic intermediate EO Pheidole prox. Reclusi 811.69 0.93 0.50 0.18 dimorphic intermediate EO Pheidole radoszkowskii 742.21 0.94 0.41 0.55 dimorphic intermediate EO 84 Species WL HFL ML V Workers’ polymorphism Integument sculpture Functional groups Pheidole scolioceps* NA NA NA NA dimorphic intermediate EO Pheidole sensitiva 733.25 1.05 0.48 0.19 dimorphic intermediate EO Pheidole sp1 1167.29 1.05 0.39 0.24 dimorphic intermediate EO Pheidole sp13 457.99 0.77 0.48 0.57 dimorphic intermediate EO Pheidole sp14 728.83 0.82 0.52 0.59 dimorphic intermediate EO Pheidole sp15 991.17 0.88 0.45 0.24 dimorphic smooth EO Pheidole sp17 775.01 0.91 0.51 0.28 dimorphic intermediate EO Pheidole sp18 852.64 1.12 0.33 0.60 dimorphic intermediate EO Pheidole sp2 747.97 0.94 0.43 0.65 dimorphic intermediate EO Pheidole sp20 839.89 0.94 0.42 0.58 dimorphic intermediate EO Pheidole sp24 836.95 1.11 0.43 0.35 dimorphic intermediate EO Pheidole sp25 851.55 0.83 0.51 0.23 dimorphic intermediate EO Pheidole sp29 776.94 1.05 0.42 0.64 dimorphic intermediate EO Pheidole sp32 518.70 0.78 0.50 0.73 dimorphic intermediate EO Pheidole sp39 851.45 0.97 0.41 0.55 dimorphic intermediate EO Pheidole sp40 839.40 1.05 0.49 0.55 dimorphic intermediate EO Pheidole sp45 479.71 0.77 0.50 0.57 dimorphic intermediate EO Pheidole sp5 703.17 0.87 0.53 0.20 dimorphic intermediate EO Pheidole sp51 572.01 0.73 0.52 0.61 dimorphic intermediate EO Pheidole sp52 925.53 0.97 0.43 0.44 dimorphic intermediate EO Pheidole sp53 822.60 1.01 0.42 0.51 dimorphic intermediate EO Pheidole sp6 719.71 0.93 0.41 0.75 dimorphic intermediate EO Pheidole sp7 886.39 1.06 0.42 0.61 dimorphic intermediate EO Pheidole sp8 443.76 0.78 0.51 0.58 dimorphic intermediate EO Pheidole subarmata 529.95 0.85 0.54 0.51 dimorphic intermediate EO Pheidole susannae 1006.22 1.02 0.40 0.49 dimorphic intermediate EO Pheidole termitobla 646.55 0.82 0.51 0.40 dimorphic intermediate EO Pheidole vafra 777.43 1.00 0.42 0.47 dimorphic intermediate EO Pogonomyrmex naegelii 1495.75 0.85 0.50 0.35 monomorphic textured SH Procryptocerus schmitti 1194.11 0.55 0.26 0.13 monomorphic textured AS Pseudomyrmex gr. pallidus sp1 1105.60 0.47 0.29 0.67 monomorphic intermediate AS Pseudomyrmex gracilis 2282.01 0.63 0.34 0.22 monomorphic intermediate AS Pseudomyrmex sp1 2263.75 0.45 0.25 0.63 monomorphic intermediate AS Pseudomyrmex sp2 1532.20 0.51 0.26 0.55 monomorphic intermediate AS Pseudomyrmex sp3 1893.88 0.54 0.30 0.61 monomorphic intermediate AS Pseudomyrmex sp4 1479.88 0.55 0.28 0.26 monomorphic intermediate AS Pseudomyrmex sp7 1232.77 0.48 0.26 0.49 monomorphic intermediate AS 85 Species WL HFL ML V Workers’ polymorphism Integument sculpture Functional groups Pseudomyrmex termitarius 1799.27 0.66 0.39 0.49 monomorphic intermediate AS Rasopone sp2 1688.18 0.54 0.40 0.15 monomorphic intermediate EP Sericomyrmex sp1 1410.22 0.82 0.46 0.42 monomorphic intermediate Hatt Sericomyrmex sp2 1648.22 0.80 0.46 0.86 monomorphic intermediate Hatt Sericomyrmex sp4 1462.01 0.84 0.50 0.51 monomorphic intermediate Hatt Sericomyrmex sp5 1114.70 0.91 0.60 0.49 monomorphic intermediate Hatt Solenopsis (Diplorhoptrum) sp1 409.27 0.68 0.54 0.59 monomorphic smooth CO Solenopsis globularia sp1 648.48 0.63 0.40 0.35 polymorphic smooth EO Solenopsis globularia sp2 641.62 0.63 0.39 0.44 polymorphic smooth EO Solenopsis saevissima 1042.91 0.77 0.40 0.52 polymorphic smooth EO Solenopsis sp2 679.61 0.42 0.48 0.72 polymorphic smooth CO Solenopsis sp7 583.28 0.59 0.44 0.22 polymorphic smooth EO Solenopsis substituta 939.42 0.83 0.40 0.24 polymorphic smooth EO Strumigenys gr. louisianae sp3* NA NA NA NA monomorphic textured CP Strumigenys schulzi 428.46 0.60 0.27 0.64 monomorphic textured CP Trachymyrmex sp1 1538.49 0.81 0.45 0.43 monomorphic intermediate Latt Trachymyrmex sp2 1768.12 0.89 0.43 0.29 monomorphic intermediate Latt Trachymyrmex sp3 1456.10 0.88 0.46 0.55 monomorphic intermediate Latt Trachymyrmex sp4 1092.95 0.79 0.42 0.43 monomorphic intermediate Latt Trachymyrmex sp5 1720.18 0.85 0.46 0.43 monomorphic intermediate Latt Trachymyrmex sp6 1490.50 0.86 0.48 0.44 monomorphic intermediate Latt Tranopelta gilva 560.92 0.61 0.47 0.68 monomorphic intermediate CO Wasmannia affinis 589.15 0.78 0.42 0.59 monomorphic textured EO Wasmannia auropunctata 469.98 0.85 0.42 0.48 monomorphic textured EO Wasmannia lutzi 676.39 0.80 0.38 0.71 monomorphic textured EO *Morphological traits not measured. 8 86 Table S2. Environmental variables used in GLMM’s models and PERMANOVA. Units: Mean Elevation: m a.s.l 9 (above sea level); Mean Air Humidity: relative %; Mean Air Temperature: °C; NDVI Mean: mean; NDVI Std. 10 Dev.: std. dev; Accumulated Rainfall: mm. 11 Season Habitat Elevation band Mean Elevation Mean Air Humidity Mean Air Temperature Mean NDVI Std. Dev NDVI. Accumulated Rainfall Dry Campo rupestre 800 834.74 68.7 19.21 0.35 0.27 0 Dry Forest 800 791.55 68.7 19.21 0.57 0.06 0 Dry Campo rupestre 900 963.51 70.51 18.59 0.16 0.16 0.1 Dry Forest 900 949.62 70.51 18.59 0.29 0.13 0.1 Dry Campo rupestre 1000 1017.11 73.04 17.87 0.44 0.12 0 Dry Forest 1000 1005.52 73.04 17.87 0.30 0.21 0 Dry Campo rupestre 1100 1120.29 75.98 17.01 0.30 0.17 0.3 Dry Forest 1100 1117.94 75.98 17.01 0.29 0.24 0.3 Dry Campo rupestre 1200 1237.21 75.5 17.08 0.23 0.18 0.4 Dry Forest 1200 1115.11 75.5 17.08 0.25 0.23 0.4 Dry Campo rupestre 1300 1281.19 77.19 16.23 0.33 0.10 0.1 Dry Forest 1300 1345.17 77.19 16.23 0.29 0.17 0.1 Rainy Campo rupestre 800 834.74 73.49 23.13 0.41 0.07 126.75 Rainy Forest 800 791.55 73.49 23.13 0.30 0.27 126.75 Rainy Campo rupestre 900 963.51 77.52 22.63 0.28 0.11 250.92 Rainy Forest 900 949.62 77.52 22.63 0.25 0.15 250.92 Rainy Campo rupestre 1000 1017.11 78.1 22.25 0.30 0.16 211.96 Rainy Forest 1000 1005.52 78.1 22.25 0.44 0.06 211.96 Rainy Campo rupestre 1100 1120.29 80.18 21.38 0.31 0.07 267.28 Rainy Forest 1100 1117.94 80.18 21.38 0.41 0.05 267.28 Rainy Campo rupestre 1200 1237.21 82.89 20.83 0.22 0.14 229.29 Rainy Forest 1200 1137.56 82.89 20.83 0.21 0.16 229.29 Rainy Campo rupestre 1300 1281.19 85.14 19.97 0.20 0.13 238.49 Rainy Forest 1300 1345.17 85.14 19.97 0.38 0.12 238.49 12 87 Table S3. Pearson’s correlation coefficient (r) between samples and environmental variables collected on the elevation gradient. An r-value greater than 0.7 was the parameter to consider correlated variables (bold). Mean_Eleva – Mean values of elevation (m a.s.l.); NDVI_Mean - NDVI Mean values; Mean_Hum - Mean values of air humidity or Humidity (%); Mean_Temp - Mean values of air temperature (°C); Pluv - Accumulated rainfall or precipitation (mm); NDVI_Std dev - NDVI Standard deviation values. Mean_Eleva NDVI_Mean Mean_Hum Mean_Temp Pluv NDVI_Std dev Mean_Eleva - NDVI_Mean -0.35 - Mean_Hum 0.72 -0.29 - Mean_Temp -0.45 0.11 0.25 - Pluv 0.13 -0.08 0.74 0.80 - NDVI_Std dev -0.11 -0.47 -0.28 -0.30 -0.43 - 13 88 Table S4. Records of Formicidae species sampling in six elevation cotes, in Rainy and Dry Season and two habitats with different vegetal structures, campo rupestre, and 14 forest at Serra do Espinhaço, Minas Gerais, Brazil. 15 Dry Season Rainy Season Campo Rupestre Forest Campo Rupestre Forest Subfamily/Species 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 Dolichoderinae Dolichoderus bispinosus X Dolichoderus diversus X Dolichoderus lutosus X Dorymyrmex brunneus X X X X X X X X X X X Dorymyrmex goeldii X X X X X X X X X X X X Dorymyrmex pyramicus X X X X X X X X X X X Dorymyrmex sp5 X X X Forelius brasiliensis X X X X Forelius maranhaoensis X X X X X Forelius sp2 X X X X X X X X X Linepithema aztecoides X X X X X X X X X Linepithema cerradense X Linepithema cf. pulex X X X X X X X X Linepithema gallardoi X Linepithema iniquum X X X Linepithema micans X X X X X X X X X X X X X X Dorylinae Eciton cf. vagans X Eciton mexicanum X Labidus coecus X X X X X Labidus praedator X X X X X Neivamyrmex cf. swainsonii X Neivamyrmex pseudops X X X X Neivamyrmex sp2 X Nomamyrmex esenbeckii X X Ectatomminae Ectatomma brunneum X X X X X X X Ectatomma edentatum X X X X X X X X X X X X X X Ectatomma opaciventre X X X X Ectatomma permagnum X X X X Ectatomma planidens X X X X Ectatomma tuberculatum X X X Gnamptogenys gr striatula sp nov A** X X 89 Dry Season Rainy Season Campo Rupestre Forest Campo Rupestre Forest Subfamily/Species 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 Gnamptogenys sp3 X X X Gnamptogenys striatula X X X X Gnamptogenys sulcata X X X X X X Formicinae Brachymyrmex pictus X X X X X X X X X X X X X X X X X X X X X X Brachymyrmex prox. cordemoyi X X X X X X X X X X Brachymyrmex sp4 X X X X Brachymyrmex sp5 X X Camponotus (Hypercolobopsis) sp3 X X Camponotus (Hypercolobopsis) sp6 X Camponotus (Hypercolobopsis) sp7 X X Camponotus (Myrmaphaenus) sp1 X X X X Camponotus (Myrmaphaenus) sp5 X X X X X Camponotus (Tanaemyrmex) sp2 X Camponotus arboreus X X X Camponotus atriceps X X X X X X X X Camponotus blandus X X X X X X X X Camponotus cingulatus X X Camponotus crassus X X X X X Camponotus latangulus X Camponotus lespesii X X X X X X X X X Camponotus leydigi X X X X Camponotus melanoticus X X X X X X X X X X X X X X X X X X X X Camponotus novogranadensis X X X X X Camponotus renggeri X X X X X X X X X Camponotus rufipes X X X X X X X X X X X X X X X X X Camponotus senex X X X X X X X X X X Camponotus sericeiventris X X X X Camponotus sexguttatus X X X Camponotus sp14 X X Camponotus sp19 X Camponotus sp20 X X Camponotus sp4 X Camponotus sp5 X X X X X Camponotus vittatus X X X X Myrmelachista catharinae X X X Nylanderia sp1 X X X X X X Heteroponerinae Heteroponera sp1 X Myrmicinae 90 Dry Season Rainy Season Campo Rupestre Forest Campo Rupestre Forest Subfamily/Species 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 Acromyrmex balzani X X X Acromyrmex coronatus X X X Acromyrmex sp2 X X Acromyrmex sp5 X X Acromyrmex subterraneus X X X X X X X X X X X X X X Apterostigma gr. pilosum sp1 X X X X X X Apterostigma gr. pilosum sp2 X X X X Apterostigma gr. pilosum sp3 X X Atta sexdens rubropilosa X X X X X X X X X X Cardiocondyla cf. obscurior X Carebara urichi X Cephalotes atratus X X Cephalotes pusillus X X X X X X X X X X Crematogaster acuta X X X X X X X X X Crematogaster brasiliensis X Crematogaster prox. erecta sp1 X X X X X X Crematogaster prox. obscurata sp1 X Crematogaster sp6 X Cyphomyrmex gr. rimosus sp3 X X X X X X X Cyphomyrmex gr. rimosus sp1 X Cyphomyrmex gr. rimosus sp2 X X X X X X X X X Hylomyrma balzani X X X X Hylomyrma prox. reitteri X Hylomyrma sp4** X X X Kalathomyrmex emeryi X X X X Megalomyrmex sp1 X Mycetarotes sp1 X Mycetophylax gr. strigatus sp6 X X X X Mycetophylax gr. strigatus sp4 X X Mycetophylax gr. strigatus sp5 X X X X X Mycetophylax lectus X X X X X X Mycocepurus goeldii X X X X X X X X X X X X X X Myrmicocrypta sp1 X X Ochetomyrmex semipolitus X X X X X Octostruma balzani X X Octostruma iheringi X Octostruma stenognatha X Oxyepoecus prox. bruschi sp1 X X X X Oxyepoecus prox. bruschi sp2 X X X Oxyepoecus sp5 X 91 Dry Season Rainy Season Campo Rupestre Forest Campo Rupestre Forest Subfamily/Species 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 Pheidole ambigua X X Pheidole capilata X X X X Pheidole dorsata X X X Pheidole gertrudae X X X X X X X X X X X X Pheidole jelskii X X X X X X X X X X X X X X X Pheidole oxyops X X X X X X X X X X X X X X X X Pheidole prox. reclusi X Pheidole radoszkowskii X X X X X X X X X X X X X X X X X X X X X X X X Pheidole scolioceps X Pheidole sensitiva X X X Pheidole sp1 X X X X X X X Pheidole sp13 X X X X X X Pheidole sp14 X X X Pheidole sp15 X X X Pheidole sp17 X X X Pheidole sp18 X X X Pheidole sp2 X X X X Pheidole sp20 X X X X X Pheidole sp24 X Pheidole sp25 X X Pheidole sp29 X X X X X X X X X X X X X X Pheidole sp32 X X X X X X Pheidole sp39 X X X X X X X X X X Pheidole sp40 X Pheidole sp45 X X X X X X X X X X X Pheidole sp5 X X X X X Pheidole sp51 X Pheidole sp52 X X X X X X X X X Pheidole sp53 X X X X X X Pheidole sp6 X X Pheidole sp7 X X X X X Pheidole sp8 X X X Pheidole subarmata X X X X X X X X X X Pheidole susannae X X X Pheidole termitobla X X X Pheidole vafra X X X X X Pogonomyrmex naegelii X X X X X X X X X X X Procryptocerus schmitti X Sericomyrmex sp1 X X X X X Sericomyrmex sp2 X X 92 Dry Season Rainy Season Campo Rupestre Forest Campo Rupestre Forest Subfamily/Species 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 Sericomyrmex sp4 X Sericomyrmex sp5 X Solenopsis (Diplorhoptrum) sp1 X X X X X X X X X X X X X X X X X X X X X X Solenopsis globularia sp1 X X X X X X X Solenopsis globularia sp2 X X X X Solenopsis saevissima X X X X X X X X X X X X X X X X X X X X X X Solenopsis sp2 X X X Solenopsis sp7 X X X X Solenopsis substituta X X X X X X Strumigenys gr. louisianae sp3 X Strumigenys schulzi X X Trachymyrmex sp1 X X X X X X X Trachymyrmex sp2 X X X Trachymyrmex sp3 X X X X X Trachymyrmex sp4 X X Trachymyrmex sp5 X X X X X X X X X Trachymyrmex sp6 X X Tranopelta gilva X Wasmannia affinis X X X Wasmannia auropunctata X X X X X X X X X X Wasmannia lutzi X X X X X X Ponerinae Anochetus inermis X X X Hypoponera distinguenda X X Hypoponera sp1 X X X X X X X X Hypoponera sp3 X Hypoponera sp4 X Hypoponera sp5 X X Leptogenys aff. gorgona X Leptogenys crudelis X Neoponera crenata X Neoponera latinoda X Neoponera villosa X Odontomachus bauri X X X X X X Odontomachus brunneus X X X Odontomachus chelifer X X X Odontomachus meinerti X X X Pachycondyla harpax X X X X X Pachycondyla striata X X X X X X X X X X X X Rasopone sp2 X X 93 Dry Season Rainy Season Campo Rupestre Forest Campo Rupestre Forest Subfamily/Species 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 800 900 1000 1100 1200 1300 Pseudomyrmicinae Pseudomyrmex gr. pallidus sp1 X X X X Pseudomyrmex gracilis X Pseudomyrmex sp1 X X Pseudomyrmex sp2 X Pseudomyrmex sp3 X Pseudomyrmex sp4 X Pseudomyrmex sp7 X Pseudomyrmex termitarius X X X X X X X X X X X X Species richness 79 51 38 29 39 29 30 30 16 34 21 37 89 51 39 42 41 40 31 40 22 33 40 40 ** Species under description 16 S1. Methods – Composition analysis: PERMANOVA and NMDS 17 18 We used the Permutational multivariate analysis of variance (PERMANOVA) (Anderson, 19 2017) to test the significance of the effects of each spatio-temporal dimension on ant species 20 composition. The effects of each spatio-temporal dimension on the species composition were plotted 21 using the Non-metric multidimensional scaling (NMDS) based on a Jaccard dissimilarity matrix. We 22 performed the same analyzes (PERMANOVA and NMDS) to verify the effects of environmental 23 variables on species composition. We used the “adonis” function for the PERMANOVA and 24 Permdisp analyzes (both with Sorensen index) and the “metaMDS” function for NMDS, “ordihull” 25 for group definition (e.g. fields and forests) and “ordisurf” for adjusting continuous variables (e.g. 26 NDVI, mean temperature) in the NMDS. All functions belong to the “vegan” package (Oksanen et 27 al., 2018) of the R software (R Core Team, 2017). 28 Permanova (Sorensen): 29 Response Variable Df SumsOfSqs MeanSqs F.Model R² Pr(>F) Season 1 0.201 0.201 0.914 0.009 0.484 Habitat 1 4.803 4.803 21.894 0.227 0.001 *** Elevation 1 1.256 1.256 5.724 0.059 0.001 *** Residuals 68 14.919 0.219 0.704 Total 71 21.179 1 30 Permdisp (Sorensen): Habitat and Elevation 31 Df SumSq MeanSq F.value Pr(>F) HABITAT Groups 1 0.095 0.095 14.704 0.000 *** Residuals 70 0.451 0.006 ELEVATION Groups 5 0.03822 0.007645 1.4848 0.2067 Residuals 66 0.33982 0.005149 32 33 95 34 S1. Figure 1. NMDS plot showing the similarity of ants’ species composition. Representation of the 35 species composition of two habitats per season (a) and for each elevation quota(b). The polygon 36 denotes habitats (Field and Forest), which differed in diversity from each other in the PERMANOVA 37 analysis. 38 There are differences in composition between habitats and elevations, but not between 39 seasons. Among habitats, a forest composition is also more heterogeneous than the campo rupestre. 40 References 41 Andersen, A.N. (1995) A Classification of Australian Ant Communities, Based on Functional 42 Groups Which Parallel Plant Life-Forms in Relation to Stress and Disturbance. Journal of 43 Biogeography. 22 (1), pp. 15–29. doi:10.2307/2846070. 44 Anderson, M.J. (2017) Permutational Multivariate Analysis of Variance (PERMANOVA). Wiley 45 StatsRef: Statistics Reference Online [online]. pp. 1–15. Available from: 46 http://doi.wiley.com/10.1002/9781118445112.stat07841doi:10.1002/9781118445112.stat078447 1. 48 Leal, I.R., Filgueiras, B.K.C., Gomes, J.P., Iannuzzi, L. & Andersen, A.N. (2012) Effects of habitat 49 fragmentation on ant richness and functional composition in Brazilian Atlantic forest. 50 Biodiversity and Conservation [online]. 21 (7), pp. 1687–1701. Available from: 51 http://link.springer.com/10.1007/s10531-012-0271-9doi:10.1007/s10531-012-0271-9. 52 (a) (b) 96 Oksanen, J., Blanchet, F.G., Kindt, R., Legen-, P., Minchin, P.R., Hara, R.B.O., Simpson, G.L., Soly-53 , P., Stevens, M.H.H. & Wagner, H. (2018) Package ‘ vegan ’ - Community Ecology Package. 54 R package version 2.5-5. Available from: https://cran.r-55 project.org/web/packages/vegan/index.html. 56 Paolucci, L.N., Maia, M.L.B., Solar, R.R.C., Campos, R.I., Schoereder, J.H. & Andersen, A.N. 57 (2016) Fire in the Amazon: impact of experimental fuel addition on responses of ants and their 58 interactions with myrmecochorous seeds. Oecologia. 182 (2), pp. 335–346. 59 doi:10.1007/s00442-016-3638-x. 60 R Core Team (2017) R Core Team (2017). R: A language and environment for statistical computing. 61 R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/. 62 63 64 65 66 97 Capítulo 2: Snow-free mountaintops are dominated by tiny and dark ants Flávio Siqueira de Castro, Lucas Neves Perillo, Pedro Giovâni da Silva, Ricardo Ribeiro de Castro Solar e Frederico de Siqueira Neves Artigo a ser submetido para o periódico Biological Journal of the Linnean Society, Special Issue on OCBIL Theory. 98 1 Biological Journal of the Linnean Society Section: Special Issue on OCBIL Theory. 2 Snow-free mountaintops are dominated by tiny and dark ants 3 Running title: Snow-free mountaintops’ tiny dark ants 4 Flávio Siqueira de Castro1 5 Lucas Neves Perillo1,2 6 Pedro Giovâni da Silva1 7 Ricardo R. C. Solar1,3 8 Frederico de Siqueira Neves1,3 9 10 4. Programa de Pós-Graduação em Ecologia, Conservação e Manejo da Vida Silvestre, 11 Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil 12 5. Bocaina Biologia da Conservação, Belo Horizonte, MG, Brazil 13 6. Departamento de Genética, Ecologia e Evolução, Universidade Federal de Minas Gerais, 14 Belo Horizonte, Minas Gerais, Brazil 15 16 Correspondence: 17 Flávio Siqueira de Castro 18 Programa de Pós-Graduação em Ecologia, Conservação e Manejo da Vida Silvestre, 19 Universidade Federal de Minas Gerais, Av. Antônio Carlos, 6627, Pampulha, CEP 31270-901, 20 Belo Horizonte, Minas Gerais, Brazil. 21 E-mail: fsiqueiradecastro@gmail.com 22 99 23 ORCID ID: 24 Flávio Siqueira de Castro (0000-0002-5533-1355) 25 Pedro Giovâni da Silva (0000-0002-0702-9186) 26 Lucas Neves Perillo (0000-0003-4291-4452) 27 Ricardo R. C. Solar (0000-0001-5627-4017) 28 Frederico de Siqueira Neves (0000-0002-2691-3743) 29 100 Abstract 30 In this study, we performed surveys in the rainy season in campo rupestre ecosystem in 12 31 mountains at different elevation across the Espinhaço mountain range, from Minas Gerais to 32 Bahia, Brazil. First, we aimed to evaluate the effects of environmental variables on the functional 33 structure of functional diversity of ant communities in snow-free mountaintops. Then, we tested 34 three macroecological hypotheses associated to ants’ integument colour cline variation (thermal 35 melanism hypothesis; melanism desiccation hypothesis; and photo-protection hypothesis) 36 relating them to mean temperature, vapor pressure deficit and solar radiation to determine the 37 role of colour and body size variation in a tropical elevational and latitudinal geo-climatic 38 gradient. We found ants’ communities functionally redundant due to an increase of geographical 39 distance. Despite the redundant functional pattern across the latitudinal gradient, ant species tend 40 to be tinier and darker on mountaintops than ants of lower sites in the campo rupestre. Besides 41 the elevational effects, the main drivers of functional ant diversity acting as environmental filters 42 on ants’ communities were the decrease in mean temperature and vapor pressure deficit. This 43 result is associated to the thermal melanism hypothesis, which occurs in old mountains of 44 Espinhaço mountain range, an old climatic buffered infertile landscape (OCBIL), but not in new 45 mountains in landscapes characterised by fertile soils (young, often disturbed, fertile landscapes 46 - YODFEL). These findings highlight the importance of the climatic variables on ants’ 47 community functional structuring and enhance our understanding of its drivers in the current 48 scenario of the global warming crisis. 49 50 Keywords: traits; functional diversity; ground-dwelling ants; Espinhaço Mountain Range; 51 geographical gradients; campo rupestre; OCBIL. 52 101 Introduction 53 Understanding how communities are structured in space and time, unveiling the 54 mechanisms underpinning such diversity patterns and determining the role of the species to 55 ecosystem functioning are among the central themes of ecology and conservation biology (Gras 56 et al., 2016; Vellend, 2016; Schluter & Pennell, 2017). Patterns of diversity distribution respond 57 to both deterministic (niche-related; e.g., environmental filtering) and non-deterministic (neutral; 58 e.g., dispersal limitation) ecological mechanisms. Deterministic mechanisms take into account 59 the biotic and abiotic interactions between species and between species and their habitat of 60 occurrence (Keddy, 1992). Non-deterministic mechanisms are related to stochastic processes 61 (Gleason, 1926; Eliot, 2007) and we can consider among them the dispersal limitation or 62 dispersion distance, which in many cases is guided by natural physical barriers like rivers, lakes, 63 and mountains (especially considering their mountains range). These barriers can limit species' 64 distribution via landscape attributes and, in the same way, prevent other species from relocating 65 (Måren et al., 2018). 66 The habitat concept is defined by the interactions between its abiotic and biotic 67 components, which determines the patterns of diversity, co-occurrence, and distribution of 68 biological communities (Southwood, 1977). We can only determine where these species are 69 distributed knowing the associations between species and their habitats. However, using only this 70 taxonomic approach, the ecological role of each species and their ecological function on a given 71 ecosystem is broadly neglected (Petchey & Gaston, 2006), especially when the focus is arthropod 72 diversity. Arthropods comprise one of the most abundant animal group in terms of biomass and 73 species richness (Basset et al., 1998), and hence, they play crucial and diverse functional roles in 74 different ecosystems (Losey & Vaughan, 2006). Therefore, the functional diversity approach 75 102 (FD) can be very useful and relevant for little known taxa or communities with many undescribed 76 species, such as ants (Gibb & Parr, 2013; Yates et al., 2014). Functional diversity can be defined 77 as the value and variation of species and their morphological, ecological or behavioural 78 characteristics (e.g., traits), which influence the functioning of communities (Petchey & Gaston, 79 2006). 80 The use of species traits (e.g., morphological, ecological or behaviour attributes) to 81 describe the functional structure and estimate the ecological functions of communities (Dolédec 82 et al., 1996; Vandewalle et al., 2010; Leitão et al., 2018), enable a directly to species responses 83 to changing environmental conditions in local or regional scales. Functional structure (FS) is an 84 important approach to elucidate these differences in species traits across environmental gradients 85 (Mason et al., 2005; Villégere et al., 2008; Mouillot et al., 2013) when considering the 86 multifaceted aspect of the functional diversity: (1) functional richness (FRic) is related to the 87 range of trait combination in the community; (2) functional evenness (FEve) shows how similar 88 one community is to another in terms of ecological functions; (3) functional divergence (FDiv) 89 demonstrates how much a set of attributes can differ from one community to another; and (4) 90 functional originality (FOri), which indicates how original a given attribute or set of them can be 91 in a community structure. Besides, when we argue how environmental filters could acting on 92 biodiversity, β-diversity is among the better approaches to investigate diversity patterns, either 93 using taxonomically or functionally attributes, as well as what factor explain the changes in those 94 patterns (Tuomisto, 2010). The β-diversity approach describes the length of compositional 95 dissimilarities between sites and could reveal the mechanisms that drive these differences 96 (Baselga, 2010; Tuomisto, 2010; Ricotta, 2017). β-diversity can also be used in the functional 97 diversity perspective, describing the length of functional dissimilarities on functional 98 composition, and, in the same way, reveals the mechanisms underlying these differences 99 103 (Swenson et al., 2011). Using these multi-functional indices listed above, plus functional β-100 diversity approach, we could explain the extent of the functional dissimilarity between 101 communities (Mason et al., 2005; Villéger et al., 2011), their responses to environmental 102 changes, and hence, determine which mechanisms act as environmental filters on communities 103 and species traits (Mouillot et al., 2013; Leitão et al., 2018). 104 Large-scale assessments of ecological gradients’ effects such as elevational and 105 latitudinal gradients on the functional structure (FS) of ant communities are yet neglected, 106 especially in tropical regions (Tiede et al., 2017). In these scenarios, macrohabitat variables could 107 act as environmental filters limiting the establishment of species unable to tolerate abiotic 108 conditions of a given habitat (Keddy, 1992), which may drive the species traits’ patterns. Many 109 studies testing macroecological hypotheses could be found for many organisms and they show 110 the potential distribution of organisms due to changes in geo-climatic gradients. For ectothermic 111 organisms, these patterns usually are related to the variation in the tegument colour of species 112 (e.g., cline variation on integument) (Clusella et al., 2007; Clusella-Trullas et al., 2008; Moura 113 et al., 2017), which include studies on ant communities in temperate snow-covered mountains 114 (Bishop et al., 2016) and lowland tropical forest (Law et al., 2019). As pointed by Law et al. 115 (2019), three of these macroecological hypotheses (related to integument colour cline variation) 116 are very useful to determine the role of the colour variation on ectothermic, organism (Figure 117 1a): (1) thermal melanism hypothesis (TMH), which predicts that colder environments (e.g., at 118 higher elevations and latitudes) have individuals with more melanin (i.e., pattern of darker 119 integument) (Clusella Trullas et al., 2007). Darker ants, for example, can assimilate more heat 120 from solar radiation than lighter individuals in a temperate mountain range in higher latitudes 121 (Bishop et al., 2016); (2) melanism desiccation hypothesis (MDH), which points that drier 122 environments with greater vapor pressure deficit (VPD) have darker individuals (Kalmus, 1941; 123 104 Clusella-Trullas et al., 2008). Kalmus (1941) found in Drosophila melanogaster. a positive 124 relationship between the increase of melanisation and body desiccation resistance, by decreasing 125 the cuticular permeability. In a vertical gradient in tropical forest, Law et al. (2019) showed that 126 canopy ants, inhabiting a drier and hotter environment, are darker than understory or 127 ground/subterranean dwelling ants; and (3) photo-protection hypothesis (PPH), which predicts 128 that the melanin provides protection against UV-B radiation being, a driving mechanism behind 129 the integument brightness colour variation. In environments with greater UV-B radiation, a rise 130 in melanisation rates could be found in many ectothermic such as insects (ants, Bishop et al., 131 2016; Law et al., 2019; Drosophila melanogaster, Bastide et al., 2014; carabid beetles, 132 Schweiger & Beierkuhnlein, 2016) and lizards (Clusella-Trullas et al., 2008). Since ultraviolet 133 radiation has deleterious effects on the fitness of ectothermic organisms (such as ants), the 134 increase in melanisation rates could promote protection to the dangerous effects of UV-B 135 radiation. Besides, in many cases body size is also an important attribute for a better fitness of an 136 ectothermic organism (Clusella Trullas et al., 2007; Clusella-Trullas et al., 2008; Bishop et al., 137 2016) and also could be considered a trade-off in the variation of colour patterns (specifically on 138 cline variation of colour from dark to bright) and size variation as well (Clusella Trullas et al., 139 2007; Clusella-Trullas et al., 2008; Schweiger & Beierkuhnlein, 2016). 140 Mountain ecosystems are an important scenario to describe functional diversity and 141 species trait patterns because they present high variation among small (e.g., different isocline 142 band on elevation gradient) and large spatial scales (e.g., different latitudes across the latitudinal 143 mountain range gradient). In the latitudinal gradient, species richness increases at low latitudes, 144 however, each latitude may present a different taxonomic composition due to the variation on 145 environmental conditions (Gaston, 2009; Stein et al., 2014). For instance, ant communities of 146 Cerrado present a reverse pattern with decrease species richness in direction to drier and hotter 147 105 northeast of Brazil (Vasconcelos et al., 2018); ant communities in the Atlantic Forest have high 148 species turnover from south to north of Brazilian coast with an inverted species richness pattern 149 as well (Silva & Brandão, 2014), and ants of campo rupestre, whereas taxonomic diversity 150 decreases due to increased elevation and the turnover rates increases due to increased distance 151 among mountains Latitudinal gradients also present a variation in functional diversity patterns 152 (Stevens et al., 2003; Villéger et al., 2013; Lamanna et al., 2014). They may present a very 153 similar functional pattern between latitude, despite the changes in species composition, with a 154 great functional redundancy across the gradient, as occurs with ants from the Atlantic Forest 155 (Silva & Brandão, 2014). Notably, this gradient exhibits a great variation in species richness and 156 environmental conditions from a locality to another (e.g., longitude and elevation), with a variety 157 of topographies and climatic weathers conditions through the south to north (Gaston, 2009; Dunn 158 et al., 2009; Stein et al., 2014) (Figure 1b). Elevational gradients present similar variation in 159 species richness with decreasing species richness due to increased elevation (Gaston, 2000; 160 Peters et al., 2016; Longino & Branstetter, 2019; Perillo et al. in prep.), and due to low species 161 richness, a lower functional richness and higher functional redundancy is expected (Bishop et al., 162 2014; Tiede et al., 2017; Castro et al. in prep.). As the latitudinal gradient, the elevation gradient 163 causes a decreased species richness at mountaintops due to the variation in geo-climatic 164 conditions, which sometimes become very harsh and selective on species occurrence (Bishop et 165 al., 2014; Nunes et al., 2016, 2017; Tiede et al., 2017; Castro et al. in prep.) (Figure 1c). 166 Therefore, examining how environmental factors at local (between elevations at the same 167 mountain) and regional scales (between mountains across a latitudinal gradient) drive functional 168 diversity patterns is an important contribution to our understanding of how ecological 169 communities change in response to climatic variations locally and regionally. Also, it could 170 106 enhance our knowledge about how species traits changes in response to predicted global climatic 171 variations. 172 Here, we aimed to determine the patterns of functional diversity of ants across elevational 173 and latitudinal gradients and to describe what are the effects of environmental variables on 174 functional structure of ant communities (functional richness, functional evenness, functional 175 divergence, and functional originality) and ant species traits (colour brightness). Thus, we 176 considered as explanatory variables the variations on climatic conditions (i.e., variations of 177 temperature and UV-B radiation, and vapor pressure deficit) across elevational and latitudinal 178 gradients, as well as the geographic distance (i.e., elevation range and latitude). Then, we expect 179 these variables will be important drivers to determine both functional α and β diversity patterns, 180 as seen for taxonomic α and β diversity patterns of Aculeata, ants, wasps, and bees (Bishop et 181 al., 2014; Tiede et al., 2017; Perillo et al in prep). Specifically, we expect that the functional β 182 diversity will be higher as the environments become warmer and drier, that is, approaching the 183 equator or in lower sites. As the taxonomic diversity decreases due to increased elevation and the 184 turnover rates increases due to increased distance among mountains (geographic distance) 185 (Perillo et al. in prep.), we expect a decrease in functional α or functional richness (FRic), 186 functional divergence (FDiv) and functional originality (FOri), and an increase of functional 187 evenness (FEve) due to the decrease on mean temperature (Silva & Brandão, 2014). Because 188 sites with more similar environmental and climatic conditions may have more similar functional 189 diversity than less similar ones (da Silva et al., 2018; Heino et al., 2019), we also expect a positive 190 relationship between functional β diversity and environmental distance due to different climatic 191 conditions between elevation or latitude ranges. 192 107 Additionally, we tested whether these three macroecological hypotheses explain ant’s 193 cuticle brightness at elevational and latitudinal gradients in snow-free mountaintops in the 194 Espinhaço Mountain Range, a megadiverse and endangered old climatically buffered infertile 195 landscape (OCBIL) (Hopper, 2009; Silveira et al., 2016). The Espinhaço Mountain Range is the 196 longest mountain range in Brazil and has one of South America’s oldest and most biodiverse 197 open ecosystems, the campo rupestre (Hopper, 2009; Silveira et al., 2016). Considering both 198 geo-climatic gradients, following the thermal melanism hypothesis (TMH) we expect cuticle 199 colour to get lighter as temperature increases. However, if melanin cline is a driver directly 200 related to dissection protection (MDH) or solar protection (PPH), we expect to find darker ants 201 where the vapor pressure deficit (VPD) rates and UV‐B radiation increase (Figure 1). Since we 202 are dealing with a colour-body size trade-off relationship (Clusella-Trullas et al., 2008; 203 Schweiger & Beierkuhnlein, 2016), we hypothesize that ants’ body size decreases due to the 204 warmer weather conditions as found for mountaintop ants in snow-free mountains (Bishop et al., 205 108 2016).206 207 Fig. 1. Predictions for elevational and latitudinal gradients in abiotic factors (temperature, vapor 208 pressure deficit, VPD, and UV‐B radiation) and in melanism according to the three hypotheses 209 (TMH, thermal melanism; MDH, melanism‐desiccation; PPH, photo‐protection). 210 211 Material and Methods 212 Study sites 213 The study was carried out across the entire extension of Espinhaço mountain range 214 (Figure 2; Appendix 1 – A.1), the biggest mountainous formation that extends for more than 215 1,200 km north-south in the southeast and northeast of Brazil (Giulietti et al., 1997; Fernandes, 216 2016; Silveira et al., 2016). The Espinhaço mountain range is at the ecotone of three Brazilian 217 109 vegetation domains: Cerrado (Brazilian savanna) to the west, Atlantic rainforest to the east, and 218 Caatinga (with scrubby xeromorphic vegetation, such as Dry Forests) in its north eastern (Silveira 219 et al., 2016). Our study was conducted across an extensive latitudinal gradient (ranging from 220 12ºS to 20oS) with samples distributed in elevation ranges from 1,100 to 2,072 m a.s.l. (meters 221 above sea level) where campo rupestre occurs. All sampling were performed in a single 222 ecosystem: campo rupestre, which is among the oldest neotropical grassland formations (Silveira 223 et al., 2016). 224 Fig. 2. (A) Map of 12 sample mountains thought Espinhaço Mountain Range. A: Parque Nacional Chapada Diamantina; B: Pico do Barbado; C: Pico das Almas; D: Pico da Formosa; E: Parque Estadual Serra Nova; F: Botumirim; G: Parque Estadual do Rio Preto; H: Parque Estadual Pico do Itambé; I: Pico do Breu; J: RPPN Santuário do Caraça; 110 K: Pico do Itacolomi, Ouro Preto; L: Ouro Branco. (B) Schematic sampling design: yellow circle corresponds to a sample unit composed by two sample sites, lower and upper in each mountian (N=24 sites with 40 pitfall trap sets). 225 The campo rupestre is a neotropical grassland mosaic in association with vegetation 226 complexes on rocky outcrops (Giulietti et al., 1997; Fernandes, 2016; Silveira et al., 2016), 227 formally classified as an old, climate-buffered, and infertile landscape (OCBIL) (Hopper, 2009; 228 Silveira et al., 2016). This ecosystem mainly occurs in the Espinhaço mountaintop surfaces 229 (mostly above 900 m) within all vegetation domains represented, with high species richness and 230 a considerable number of endemic and threatened plants (e.g., the endemic Trembleya laniflora 231 [Melastomataceae] - Soares and Morellato, 2018; Coccoloba cereifera and C. acrostichoides 232 [Polygonaceae] - (Melo, 2000); and the threatened Philcoxia minensis [Plantaginaceae] and 233 Mitracarpus pusillus [Rubiaceae]), and animal species (e.g., the endemic butterfly Yphthimoides 234 cipoensis [Nymphalidae] and the threatened Strymon ohausi [Lycaenidae] - Rosa et al., 2019). 235 236 Sampling design 237 We defined 12 sample mountains along the Espinhaço Mountain range considering the 238 latitudinal and elevational gradients (mean distance: approx. 67 km; minimum distance: approx. 239 20 km; maximum distance: approx. 216 km apart), selecting only sites within campo rupestre 240 ecosystem (Figure 2; Appendix 1 – A.1). In each mountain, we selected two sample sites at 241 different elevations in campo rupestre: one at the mountain base (lower site: approximately at 242 1000 m), and another near the mountain summit (upper site: ranging from approximately 1300 243 111 to 2100 m). Thus, the minimum and maximum distance between the up-down was 300 to 1100 244 m respectively. 245 In each mountain site (N=12), we installed five sets of pitfall traps separated from each 246 other by 200 m at each elevation site (i.e., five sets on lower site and five at the upper site; Figure 247 2). Each set of traps (50 m length and 20 m width) was composed of four pitfall traps arranged 248 in parcel corners. Each trap remained in the field for 48 hours, totalling 23,040 hours of sampling 249 effort. Each locality was sampled once, during the rainy season (November to February). 250 251 Identification of species 252 Ants were identified to species and morphospecies by comparison with the Collection of 253 Formicidae from campo rupestre of the Laboratory of Insect Ecology at the Universidade Federal 254 de Minas Gerais, Brazil. We followed Baccaro et al., (2015) and Bolton et al.(2005) for ants 255 classifications, and we had help of experts of different ant taxonomic groups (see 256 Acknowledgements).. 257 258 Definition of functional traits 259 We described each ant species in terms of functional traits, which provides information 260 about ants’ ecological functions, such as body size, diet habit, foraging capacity, nesting, 261 thermoregulation, and habitat association (following consolidated Fichaux et al., 2019; Paolucci 262 et al., 2016; Leal et al., 2012; Bishop et al., 2016; Barden, 2017; Tiede et al., 2017). Six traits 263 were described to each species: Weber’s length (WL), hind femur length (HFL), mesossoma 264 colour (% of colour brightness), polymorphism, integument sculpture, and functional groups 265 (five morphological traits and one ecological trait) (Table 1). 266 112 Table 1. List of morphological and ecological traits measured and their related ecological functions. 267 Traits Measure Abbreviation/Unit Ecological functions Morphological traits Weber’s length Continuous WL (μm) Proxy for total size, related to habitat complexity (Weber, 1938; Kaspari & Weiser, 1999). Femur length Continuous HFL (μm) Indicator of foraging speed, associated with habitat complexity (Feener et al., 1988; Yates et al., 2014). Colour (Mesossoma) Continuous V (%) * Thermal melanism: dark individuals has a benefit in cool climates compared to a lighter one (Clusella Trullas et al., 2007); Indicative of thermotolerance and, directly related to temperature variation and solar radiation (e.g. ants in cold environments may be darker integument rather than in warm environments with greater UV-B rates) (Bishop et al., 2016). 113 Traits Measure Abbreviation/Unit Ecological functions Polymorphism Categorical 1 = monomorphic; 2 = dimorphic; 3 = polymorphic Polymorphism of the workers, attribute related to the ability to develop different tasks in the colony (e.g., foraging, protection, internal activities of the nest; Wills et al., 2017). Integument Sculpture Ordinal 1 = cuticle smooth/shiny; Protection from desiccation. Thickened cuticles enhanced dehydration tolerance (Nation, 2008; Terblanche, 2012) 2 = superficial wrinkles/pits; 3 = surface heavily textured Ecological Trait Functional Groups Categorical AA = Army Ants; AD = Arboreal Dominant; AP = Arboreal Predator; AS = Arboreal Subordinate; CO = Cryptic Omnivores; CP = Cryptic Predators; DD = Dominant Dolichoderinae; EO = Epigeic Omnivores; EP = Epigeic Predators; Hatt = High Attini; Latt = Low Attini; Functional groups based on global-scale responses of ants to environmental stress and disturbance. Also, indicative of ecological tasks, such as nesting, foraging, and diet habits (Andersen, 1995; Leal et al., 2012; Paolucci et al., 2016). All groups were based on the classification used by Paolucci et al., (2016), except for the Seed Harvester group 114 Traits Measure Abbreviation/Unit Ecological functions Opp = Opportunist; SC = Subordinate Camponotini SH = Seed Harvester; UT = Underground tender (Johnson, 2015) here represented by Pogonomyrmex naegelli, which was not present in this list. * The HSV cylindrical-coordinate colour model (Smith, 1978), whereas: H = Hue shows the 268 dominant wavelength; S = Saturation, indicates the amount of dominant wavelength (H) present in 269 the colour; and V = Value, defines the amount of bright in the colour. We analysed only the variable 270 V, which measured in % of colour brightness (e.g. white colour presents 100% bright while black 271 colour has 0% of bright) (as proposed by Bishop et al., 2016). 272 273 To perform the morphological measurements, except for the trait “Colour”, we followed 274 the guide for identification of functional attributes for ants (The Global Ants Trait Database – 275 GLAD; Parr et al. 2017). We used the HSV colour model (Smith, 1978) to obtain the ant’s colour, 276 using only the variable V (colour brightness or lightness), as proposed by Bishop et al. (2016). 277 Differently, from Bishop et al. (2016), who considered a predominant colour between head, 278 mesosoma, and gaster, we performed the capture of HSV values of the predominant colour only 279 in mesosoma of each specimen. All continuous data were divided by Weber’s length (except for 280 Weber’s length and colour brightness) to generate correction for individual body size, because 281 these data (continuous traits measured) were not normally distributed (Arnan et al., 2018; 282 Fichaux et al., 2019). 283 115 Body measurements were made using a digital capture micrometre (accurate to 0.01 mm) 284 provide in the LC Micro 2.2 OLYMPUS® software. We carried out image acquisition using 285 Microscope Digital Camera LC30 OLYMPUS® mounted on a stereomicroscope SZ61 286 OLYMPUS®. We randomly selected individuals of every species recorded in the dataset to take 287 the measures. At least six individuals were measured per species or all individuals for species 288 with fewer individuals (e.g. only one measure per trait for singletons). We took measures of 289 available individuals (only minor workers were used; N = 1386 images from 693 individuals 290 measured; average = 4.1 individuals per species). We used genera/species information available 291 at AntWeb (https://www.antweb.org) and AntWiki website (https://www.antwiki.org) (Guénard 292 et al., 2017) (and by our own observations) to attribute categorical and ordinal morphological 293 traits to each species (polymorphism and integument sculpture). 294 295 Macrohabitats variables 296 We used WorldClim data (version 2, 1970-2000) (https:// www.worldclim.org) extracted 297 for each collection site geographical coordinates (N=24: 12 upper and 12 lower sites) as climatic 298 explanatory variables (mean and variance) that may shape environmental site patterns, such as 299 temperature, wind speed, precipitation, vapor pressure deficit, and solar radiation. Additionally, 300 we used as geographical variables the elevation data for each sampling site (m a.s.l.), and its 301 respective latitudinal and longitudinal position (decimal degrees) (Appendix 1 – A.2). 302 Functional diversity metrics 303 Functional structure (FS) of Functional Diversity (FD) was measured using a 304 multidimensional framework proposed by Mouillot et al. (2013). With this approach, we can 305 access a particularly useful way to characterize functional spatial-temporal changes in 306 116 communities. We first computed the functional distance between each pair of species using the 307 Gower distance since functional traits were not all continuous, which allows considering different 308 types of traits (see de Bello et al., 2011; Leitão et al., 2016; Nunes et al., 2016 for more details) 309 (Appendix 1 – A.3). A principal coordinate analysis (PCoA) on each functional distance matrix 310 was built to a multidimensional functional space and then the different functional indexes of 311 assemblage structure was estimated (Mouillot et al. 2013). The number of PCoA axes (i.e. 312 dimensions) was chosen based on the quality of the functional space, with “quality functional 313 space” function (Maire et al., 2015) of the R software (R Core Team, 2019). This function shows 314 the extent to which it accurately represents the initial functional distances between species pairs. 315 Thus, we kept the first six PCoA axes with the minimum number of axes that yield a high-quality 316 functional space (Maire et al., 2015). 317 Then, we calculated the functional indices using the “multidimFD” function in the “FD” 318 R package (Laliberté & Legendre, 2010). After that, we selected four functional indices: FRic 319 (Functional Richness) – volume of the functional space filled by all species within the 320 community, which shows the range of trait combination; FEve (Functional Evenness) – the 321 regularity of the distribution and relative abundance of species in functional space for a given 322 community, which applies only to the distribution of abundance in given niche space; FDiv 323 (Functional divergence) – the proportion of total abundance supported by species with the most 324 extreme trait values within a community and indicates the degree of niche differentiation; and 325 FOri (Functional originality) – the isolation of a species in the functional space occupied by a 326 given community and reflects the degree of uniqueness of species traits. The raw values of each 327 index were standardized between 0 and 1 by dividing them, respectively, by the maximum 328 distance to the barycentre and by the maximum nearest-neighbour distance observed overall 329 species present in dataset (Mouillot et al., 2013; Leitão et al., 2016). 330 117 Data analysis 331 Species richness at each mountain (lower and upper sites) were estimated using the 332 sample completeness of our samples based on a sample coverage value (Chao & Jost, 2012; Chao 333 et al., 2013). Analyses were done using iNEXT (Hsieh et al., 2016), available at 334 https://chao.shinyapps.io/iNEXTOnline/. 335 To verify the relationship between functional dissimilarity and environmental distance 336 we calculated functional β-diversity dissimilarity among sites (N=24) based on the Jaccard 337 dissimilarity coefficient. Thus, using the “functional.beta.pair” function of betapart (Baselga & 338 Orme, 2012) to produce a dissimilarity matrix based on each site with Jaccard dissimilarity 339 coefficient. Then we modelled the variation in biological dissimilarities using Generalized 340 Dissimilarity Modelling (GDM: Ferrier et al., 2007). With GDM we could modelling spatial 341 variation in functional community composition between pairs of geographical locations, and it 342 can be based on any dissimilarity matrix as response. In our case, the response matrix was the 343 pairwise Jaccard dissimilarity matrix (functional β-diversity) for each site. GDM is based on 344 matrix regression and is specified based on a link function defining the relationship between the 345 response (i.e., dissimilarity matrix) and predictor environments variables (Ferrier et al., 2007), 346 to uncover the effects of ecological distance on functional diversity. For all variables, the 347 multicollinearity was tested via Pearson correlation to define the variables to be used in the 348 models as explanatory predictors (i.e., all of which with correlation values lower than ± 0.6; 349 Appendix 1 – A.2). Among these variables we kept eleven variables used in the analysis 350 (Appendix 1 – A.2). Finally, we assessed the impacts of the predictor variables on this matrix 351 using the functions “gdm” and “gdm.varImp” available in the “gdm” R package (Manion et al., 352 2018). The GDM with Jaccard dissimilarity matrix is an important functional diversity approach, 353 which allows us to elucidate the differences in functional β-diversity and, then, the extent of the 354 118 functional dissimilarity between communities, their responses to environmental changes, and so, 355 determine which mechanisms will act as environmental filters on communities. 356 Differences in functional diversity indices (FRic, FEve, FDiv, and FOri) among 357 mountains (latitudinal gradient) and between upper and lower sites (elevational gradient) were 358 evaluated using generalized linear models (GLM – binomial error family, corrected for 359 overdispersion) (Crawley, 2013). We performed four models (Model 1 to Model 4), considering 360 each functional index as response variables using the related variables to the three melanism 361 hypothesis (i.e., mean temperature, vapor deficit pressure [VPD] and mean solar radiation [UVB-362 B]; Appendix 1 – A.2) and the interaction between elevation and latitude as explanatory 363 variables. 364 We also used assembly weighted averages (AWM) of lightness (the V variable of HSV, 365 Table 1), as proposed in Bishop et al. (2016) and Law et al. (2019) and body size (Weber’s 366 Length – WL, Table 1), calculated for each community (N=24): 367 where S is the number of species in each community site, pi is the proportional abundance of 368 each species and xi is the trait value (lightness – AWM V - or body size – AWM WL) of each 369 species (Appendix 1 – A.4). For ant’s proportional abundance, we used the proportion of 370 incidence of each species per site (Castro et al., 2012; Banschbach & Ogilvy, 2014), e.g., there 371 were 20 pitfall traps per site then if a species has fallen on all pitfall traps, the frequency equals 372 1; if a species has fallen in six traps, the frequency equals 0.3. 373 Generalized linear models (GLM) were also used to assess how much variation in 374 assemblage‐weighted lightness (AWM V) and assemblage‐weighted body size (AWM WL) 375 119 could be explained by the elevational and latitudinal gradients. We tested six models, 376 contemplating in the first three models the AWM V as response variable (Model 5 to Model 7, 377 with binomial corrected for overdispersion). In every model, we test the effect of each melanism 378 hypothesis, mean temperature (related to TMH), mean VPD (related to MDH) and mean UVB-379 B (related to PPH), respectively, plus elevation (m a.s.l.) and latitude (DD) values for each site 380 as explanatory variables, considering the interaction between these two geographical variables. 381 For the other models (Model 8 to Model 10, with Gaussian distribution), we used the AWM WL 382 as response variable and the same explanatory variables described above for AWM V. 383 All models were corrected for overdispersion for complete models to the minimal 384 adequate models. The significance of the tests in models was determined by computing an 385 analysis of deviance using the F test (Crawley, 2013). All analyses were performed in R software 386 (R Core Team, 2019). 387 Results 388 Taxonomic structure 389 We collected 2,357 specimens distributed among 169 morphospecies and seven ants’ 390 subfamilies (Appendix 1 – A.5). The most representative subfamilies in numbers of records 391 were Myrmicinae (47.0%), Formicinae (26.6%), Dolichoderinae (12.2%), and Ectatominae 392 (6.8%). The richer subfamilies were Myrmicinae (61.5%), Formicinae (14.2%), 393 Dolichoderinae (6.5%), Ponerinae (6.5%), and Ectatominae (5.9%). Most subfamilies 394 occurred on all mountains and elevations, except for Dorylinae, which did not occur on any 395 upper sites. We found a high number of singletons (49 spp. - 28.9%) and doubletons (15 spp. 396 - 8.8%). Only two morphospecies were sampled on all mountains, Ectatomma edentatum 397 (Ectatomminae) and Pheidole oxyops (Myrmicinae). Additionally, we found 29 (17.2%) 398 120 morphospecies sampled exclusively on upper sites, 73 spp. (43.2%) exclusively on lower 399 sites, and 68 spp. (40.2%) on both sets of sites. 400 Rarefaction-extrapolation accumulation curves suggested high sample coverage for 401 all sites (Appendix 1 – A.6.a; A.6.b; A.7), where the lower sites presented the higher values 402 of species richness. The sample completeness found for the upper sites was high (sample 403 coverage ranging from 0.741 to 0.969). Also, we found high sample completeness for the 404 lower sites (sample coverage ranging from 0.832 to 0.938). 405 406 Functional structure (FS) 407 Functional β-diversity of ant communities did not change along the latitudinal 408 gradient, nor on the elevation gradient as a function of geo-climatic gradient changes. There 409 were no differences in the explained deviance among the pairwise functional dissimilarity 410 of ants according to GDM analysis. The environmental distance explained only 3.64% of the 411 pairwise functional diversity dissimilarity model (Table 2). 412 413 Table 2. Summary of the Generalized Dissimilarity Modelling (GDM) of Jaccard dissimilarity 414 (β-diversity) and the predictor variable impacts. 415 Dissimilarity GDM deviance 17.71 Explained (%) 3.64 p-value 0.69 Variables impacts Latitude 0.97 Elevation 0.36 UVB-B (mean) 0.42 Temperature Max (mean) 0.98 Temperature (mean) 0.97 UVB-B (variance) 0.42 Precipitation (variance) 0.92 121 Nonetheless, we found a significant decrease in functional richness (FRic) due to an 416 increase in elevation (Model 1, Table 3; Figure 3), independently of the latitudinal variation 417 (Model 1, Table 3; Figure 3b). We do not find any effects of spatial variables (elevation and 418 latitude) on other functional metrics FEve, FDiv and FOri (Table 3; Figure 3). 419 420 Table 3. Results of generalized linear models with functional metrics (FRic = Functional 421 Richness; FEve = Functional Evenness; FDiv = Functional Divergence; FOri = Functional 422 Originality) as responses variables across the elevational and latitudinal gradients. The response 423 variables denote a model with binomial error distribution corrected for overdispersion 424 (quasibinomial). d.f = Degrees of Freedom; F = Fisher value’s; Pr(>F) = P-values (significance 425 codes: *** ≤ 0.001; **≤ 0.01; * ≤ 0.05); VPD = Vapor Pressure Deficit; UV-B = Ultra-violet 426 radiation band B. 427 Response variables Explanatory variables F Pr(>F) Model 1 FRic a Mean Temperature 4.8458 0.04182 * d.f. = 23/ d.f. residuals = 17 Mean VPD 0.1011 0.75439 Mean UV-B 0.3264 0.57527 Elevation 9.3781 0.00705 ** Latitude 4.4694 0.04959 * Elevation: Latitude 0.6687 0.42482 FRicb Elevation 8.7094 0.007383 ** d.f. = 23/ d.f. residuals = 22 122 Response variables Explanatory variables F Pr(>F) Model 2 FEvea Mean Temperature 15.9627 0.0009366 *** d.f. = 23/ d.f. residuals = 17 Mean VPD 0.2728 0.6082310 Mean UV-B 0.9555 0.3420281 Elevation 1.8275 0.1941359 Latitude 0.4635 0.5051834 Elevation: Latitude 3.4674 0.0799683 FEvea Mean Temperature 14.699 0.0009036 *** d.f. = 23/ d.f. residuals = 22 Model 3 FDiva Mean Temperature 0.0093 0.92437 d.f. = 23/ d.f. residuals = 22 Mean VPD 0.6529 0.43024 Mean UV-B 2.4890 0.13307 Elevation 2.3770 0.14154 Latitude 0.0646 0.80235 Elevation: Latitude 3.3630 0.08424 FDivb Mean Temperature 0.0089 0.9259 d.f. = 23/ d.f. residuals = 18 Mean VPD 0.6249 0.4395 Mean UV-B 2.3822 0.1401 123 Response variables Explanatory variables F Pr(>F) Elevation 2.2750 0.1488 Latitude 0.0619 0.8064 Model 4 FOria Mean Temperature 0.0119 0.91454 d.f. = 23/ d.f. residuals = 22 Mean VPD 0.0251 0.87591 Mean UV-B 0.6414 0.43427 Elevation 0.0111 0.91751 Latitude 0.4924 0.49236 Elevation: Latitude 4.1998 0.05619 FOrib Mean Temperature 0.0103 0.9202 d.f. = 23/ d.f. residuals = 18 Mean VPD 0.0219 0.8840 Mean UV-B 0.5587 0.4644 Elevation 0.0096 0.9229 Latitude 0.4289 0.5208 a Complete model b Minimal adequate model 428 124 429 Fig. 3. Results of generalized linear models between functional diversity metrics and geographical 430 variables of elevation and latitude.Legend: FRic – Functional richness; FEve – Functional evenness; 431 FDiv – Functional divergence; FOri – Functional originality. Elevation: meters above sea level (m 432 125 a.s.l.); Latitude: decimal degrees. Colour and size dots: black tiny dots – upper sites; white big dots 433 – lower sites. 434 Also, no effects of environmental variables on functional metrics were observed 435 (Table 3), with exception of functional evenness (FEve). We found a negative effect on FEve 436 due to increased mean temperature (Model 2, Table 3; Figure 4). 437 438 Fig. 4. Results of generalized linear models between functional evenness and mean temperature. 439 Colour and size dots: black tiny dots – upper sites; white big dots – lower sites. 440 441 126 Effects of environmental filters on ant’s traits - Testing three hypotheses of cuticle 442 colour and body size on elevational and latitudinal gradients 443 We found a positive effect of assemblage‐weighted lightness (AWM V) and assemblage‐444 weighted body size (AWM WL), being body size a predictor of lightness on ants (Anova: F 445 (1,22) = 52.573, p= 2.902e-07, R 2= 0.69; Figure 5). 446 447 Fig. 5. Results of correlation between assemblage‐weighted lightness (AWM V) and assemblage‐448 weighted body size (AWM WL). Colour and size dots: black tiny dots – upper sites; white big dots – 449 lower sites. 450 451 Increasing the elevation, regardless of the latitude, has a negative effect on the 452 assemblage lightness (Table 4; Figure 6) and body size (Table 4; Figure 7a). 453 454 127 Table 4. Results of generalized linear models with AWM index of lightness (with Mean of colour 455 brightness - AWMV) and AWM index of body size proxy (with Mean Weber’s Length – 456 AWMWL) variables as responses variables across the elevational and latitudinal gradients. Each 457 model was performed to test the three melanism hypothesis: TMH (Thermal Melanism 458 Hypothesis), MDH (Melanism Dissection Hypothesis) and PPH (Photo-Protection Hypothesis). 459 The response variables denote a model with quasibinomial error distribution for the AWM V 460 index and Gaussian error distribution for the AWM WL index. All models corrected for 461 overdispersion for complete models and for minimal adequate models. d.f = Degrees of Freedom; 462 F = Fisher value’s; Pr(>F) = P-values (significance codes: *** ≤ 0.001; **≤ 0.01; * ≤ 0.05); VPD 463 = Vapor Pressure Deficit; UV-B = Ultra-violet radiation band B. 464 Response variables Explanatory variables F Pr(>F) Model 5 – TMH AWM V a Mean Temperature 9.9125 0.005291 ** d.f. = 23/ d.f. residuals = 19 Elevation 6.2709 0.021549 * Latitude 1.3825 0.254197 Elevation: Latitude 0.7877 0.385902 AWM V b Mean Temperature 9.7985 0.005058 ** d.f. = 23/ d.f. residuals = 21 Elevation 6.1988 0.021244 * Model 6 – MDH AWM V a Mean VPD 9.9304 0.005258 ** 128 Response variables Explanatory variables F Pr(>F) d.f. = 23/ d.f. residuals = 19 Elevation 8.3119 0.009527 ** Latitude 0.8504 0.367989 Elevation: Latitude 0.5088 0.484344 AWM V b Mean VPD 10.3195 0.004181 ** d.f. = 23/ d.f. residuals = 21 Elevation 8.6376 0.007840 ** Model 7 – PPH AWM V a Mean UV-B 3.3542 0.082758 d.f. = 23/ d.f. residuals = 19 Elevation 14.4169 0.001218 ** Latitude 1.8328 0.191680 Elevation: Latitude 1.3626 0.257531 AWM V b Elevation 16.296 0.0005513 *** d.f. = 23/ d.f. residuals = 22 Model 8 – TMH AWM WLa Mean Temperature 4.3732 0.05018 d.f. = 23/ d.f. residuals = 19 Elevation 1.1678 0.29339 Latitude 1.3243 0.26410 Elevation: Latitude 0.4198 0.52478 129 Response variables Explanatory variables F Pr(>F) AWM WLb Mean Temperature 4.2080 0.05292 d.f. = 23/ d.f. residuals = 21 Latitude 0.0839 0.77487 Model 9 – MDH AWM WL a Mean VPD 3.8867 0.06341 d.f. = 23/ d.f. residuals = 19 Elevation 1.7827 0.19760 Latitude 1.2692 0.27394 Elevation: Latitude 0.3292 0.57287 AWM WLb Mean VPD 4.0216 0.05864 d.f. = 23/ d.f. residuals = 21 Elevation 1.8445 0.18954 Latitude 1.3133 0.26533 Model 10 -PPH Mean UV-B 2.1102 0.16264 AWM WL a Elevation 4.4788 0.04774 * d.f. = 23/ d.f. residuals = 19 Latitude 0.0005 0.98313 Elevation: Latitude 0.6211 0.44036 AWM WLb Elevation 5.8302 0.02451 * d.f. = 23/ d.f. residuals = 22 130 Response variables Explanatory variables F Pr(>F) a Complete model b Minimal adequate model 465 131 466 Fig. 6. Results of generalized linear models between cuticle lightness and both spatial and environmental 467 variables testing the hypotheses of TMH, MDH, and PPH for patterns in cuticle colour Legend: TMH – 468 132 Thermal Melanism Hypothesis; MDH - -Melanism Desiccation Hypothesis; PPH - -Photo-Protection 469 Hypothesis; AWM V – assembly weight means of cuticle lightness; VPD – Vapor Pressure Deficit (kPa 470 - kilo Paschal); Elevation: meters above sea level (m a.s.l.); Latitude: decimal degrees. Colour and size 471 dots: black tiny dots – upper sites; white big dots – lower sites. 472 133 473 Fig. 7. Results of generalized linear models between body size and both spatial and environmental 474 variables testing the hypotheses of TMH, MDH, and PPH for patterns in cuticle colour. Legend: TMH – 475 134 Thermal Melanism Hypothesis; MDH - -Melanism Desiccation Hypothesis; PPH - -Photo-Protection 476 Hypothesis; AWM WL – assembly weight means of body size; VPD – Vapor Pressure Deficit (kPa - kilo 477 Paschal); Elevation: meters above sea level (m a.s.l.); Latitude: decimal degrees. Colour and size dots: 478 black tiny dots – upper sites; white big dots – lower sites. 479 For the GLM models in which we tested the TMH hypothesis, we observed an 480 increase in cuticle lightness due to average temperature increase, together with the negative 481 effect of the elevation increase revealing a darker ant community in the upper sites than 482 lower ones (Model 5, Table 4; Figure 6). We also observed an increase in cuticle lightness 483 as a function of vapor pressure deficit (VPD) when we tested the MDH hypothesis and, 484 similarly, we found the same pattern of the negative effect of increased elevation indicating 485 the occurrence of a darker community in environments with most moisture or with lower 486 VPD values (Model 6, Table 4; Figure 6). When we tested the PPH hypothesis, no effects of 487 UV-B radiation on cuticle colour were observed. 488 We found no effects of the environmental variables in the models generated to test 489 the relationship of body size to the three melanisation hypotheses (Table 4; Figure 7). 490 However, we observed that ants are tiny in upper sites than lower sites (Model 10, Table 4). 491 492 Discussion 493 In the latitudinal gradient where the Espinhaço mountain range occurs, ant species tend 494 to be tinier and darker on mountaintops than ants of lower sites in the campo rupestre regardless 495 of the latitude. Despite the redundant pattern of functional diversity found in the latitudinal 496 gradient (as found by Silva & Brandão, 2014), there is a strong and positive relationship between 497 cuticle colour brightness and body size for snow-free mountaintops ants’ communities in the 498 elevational gradient. Beyond the elevational effects, the main drivers of functional ant diversity 499 135 that act as environmental filters were the decrease in mean temperature and vapor pressure deficit 500 (VPD) values. Generally, ants are considered thermophilic organism, being this variable 501 important to ants’ community structuring (Kaspari et al., 2015; Bishop et al., 2017), including 502 campo rupestre ants’ communities (Costa et al., 2018). Nevertheless, we presented here the first 503 study relating changes in distribution and diversity patterns of ant communities to colour attribute 504 and to body size variation (i.e., shifts on functional diversity by also considering other traits) in 505 a tropical elevational and latitudinal geo-climatic gradient. 506 Contrary to our expectations, we find a redundant pattern of functional diversity across 507 the latitudinal gradient, the same pattern was found for ants in the Atlantic Forest by Silva & 508 Brandão (2014). As we expected, we found ants’ communities with lower functional richness 509 (FRic) in upper sites than lower ones of campo rupestre, as a direct effect of the decrease in 510 taxonomic α-diversity due to elevation increasing. The species richness decrease due to the 511 increasing elevation is a recurrent pattern on campo rupestre’s ant communities (Castro et al. in 512 prep.; Perillo et al., in prep.). Ants’ functional diversity in campo rupestre shows redundant 513 functional patterns in the elevational gradient (Castro et al. in prep), even with the high species 514 turnover found between the elevation bands. However, unlike this study, the highest elevation in 515 Castro et al. (in prep) did not exceed 1400 m, which makes it difficult to report these functional 516 diversity patterns directly to our results, especially for upper site communities’ patterns, since 517 we have sites up to 2000 m a.s.l. across the Espinhaço range. 518 In addition, according to our expectations, regardless of the elevation or latitude range, 519 the functional evenness (FEve) index was high in the whole system and negative related to mean 520 temperature (as seen by Silva & Brandão, 2014). Since the FEve is dependent of abundance 521 distribution (in our case based on species incidence) (Mason et al., 2005; Mouillot et al., 2013), 522 we found high FEve in upper sites than lower sites, as seen as by Silva & Brandão (2014) for 523 136 ants in Atlantic Forest. In all mountains surveyed on the Espinhaço mountain range, the mean 524 temperature decreases with elevation increase (Fernandes et al., 2016; Ferrari et al., 2016). This 525 FEve pattern evidenced higher evenness in incidence of mountaintops’ ants (upper sites) in the 526 functional space, especially on mountains with greater geographical distance between upper and 527 lower sites (such as Caraça, Pico do Breu, Pico das Almas e Barbado), whereas the mean 528 temperature and vapor pressure deficit varies more. Additionally, no effects were observed on 529 functional divergence (FDiv) and functional originality (FOri). Since functional divergence and 530 functional originality are low and not driven by variables related to elevation and latitude 531 (Mouillot et al., 2013). These results evidencing the functional redundancy pattern on these ants’ 532 communities across the Espinhaço mountain range. The high number of epigeic omnivores ants 533 in the whole system (like Pheidole spp., Linepithema spp and Ectatomma spp.) and the recurrent 534 decrease in occurrence of predators’ species due to the increased elevation may have contributed 535 to these results. In all mountains, there was a decreased occurrence in mountaintop of Ponerinae, 536 a taxon with predominant predators’ species, and the absence of Dorylinae, exclusively formed 537 by predators’ species, in all upper sites surveyed. At high elevations Ponerinae’s species richness, 538 and their incidence as well, tend to be lower than in lowland communities in a tropical mountains 539 range in South America and Meso America (Tiede et al., 2017; Longino & Branstetter, 2019). In 540 the same way, most Dorylinae species were related to the forest habitats and only the 541 Neivamyrmex spp. were related to the montane habitat, while the others species of this subfamily 542 were not sampled beyond 1400 m.a.s.l (Tiede et al., 2017; Longino & Branstetter, 2019) 543 including the campo rupestre (Anjos et al., 2015; Costa et al., 2015, 2018; Castro et al. in prep). 544 As seen by Tiede et al. (2017), we also found that the species richness and incidence of 545 ants were important drivers of ant functional diversity. Different from these authors, we found a 546 negative effect of elevation and mean temperature on ants’ incidence and species richness, 547 137 consequently in functional evenness and functional richness. As a consequence, a species poor 548 community dominated by tinier and darker ants was selected due to their adaptation to harsh 549 environmental conditions at high elevations (Machac et al., 2011; Smith, 2015; Bishop et al., 550 2016; Castro et al., in prep) under effects of environmental filters such as the increase in elevation 551 and hence decrease in mean temperature and vapor pressure deficit. 552 Our findings confirm, only across the elevational gradient independently of latitude, the 553 thermal melanism hypothesis (TMH) in campo rupestre’s ant communities and refute the 554 melanism desiccation hypothesis (MDH). This last pattern was found in temperate mountains 555 with snow-mountaintops, whereas ants were darker and bigger as the mean temperature 556 decreased with elevation increase (Bishop et al., 2016). Also, in forest canopy, which presents 557 darker ants in high strata as a function of temperature, vapor pressure deficit and UV-B radiation 558 rates arisen (Law et al., 2019). We did not find any relationship between UV-B radiation and 559 colour brightness and body size with the photo-protection hypothesis (PPH). However, due to 560 the constantly lower temperature in temperate mountains along the day, Bishop et al.(2016) 561 found a negative relationship between brightness and body size. On the other hand, Law et al. 562 (2019), despite encountering darker ants in the hotter forest strata, did not find any relationship 563 between colour brightness and body size. In our study, we found a positively correlation between 564 colour brightness and body size, an important trade-off found in ectothermic species (Clusella 565 Trullas et al., 2007; Clusella-Trullas et al., 2008; Schweiger & Beierkuhnlein, 2016), where 566 large-sized ants are brighter than small-sized ones. In upper sites we found smaller ant 567 communities predominantly tinier and darker. Habitat filtering species unable to tolerate abiotic 568 conditions of a given habitat, which limiting the establishment, and resulting in co-occurring 569 species with similar ecological attributes (Keddy, 1992). These findings were driven for example, 570 by the distribution of Ectatomma spp. found in the campo rupestre. We sampled five Ectatomma 571 138 spp. and, one of them was one of two species that occurred on all sampled mountains and the 572 only species of the genus sampled in upper sites: E. edentatum, the tinier and darker Ectatomma 573 surveyed in this study. The other species found only in lower sites, like E. tuberculatum and E. 574 opaciventre, were bigger ants with brighter colour. 575 In this way, we can directly relate the decrease in temperature, and the vapor pressure 576 deficit (moisture rise) as important environmental filters due to the increase in elevation, 577 revealing a trade-off pattern in the relationship among body size and colour brightness in ants’ 578 communities of campo rupestre inhabiting the Espinhaço mountain range. In general, small 579 ectothermic organism could take heat faster and have a greater convective heat-transfer capacity 580 with the environment compared to large ones. These characteristics improve the fitness of small 581 ectothermic species under low temperatures (e.g. small lizards) and they are more melanic to 582 catch more heat from the environment since they can dissipate heat faster in windy conditions 583 (Clusella-Trullas et al., 2008), like a “bio-radiator”, adjusting more efficiently body temperature 584 than lighter and bigger individuals. During the cool early-morning periods typical of campo 585 rupestre (Fernandes et al., 2016; Ferrari et al., 2016) tiny darker ants could benefit from heating 586 faster than big lighter ones, because with their small size, rapid heat loss occurs during warm 587 periods of the day as seen for tiny and black beetles in Namibia desert (Turner & Lombard, 1990). 588 In conclusion, we found a redundant pattern of functional diversity across the 589 latitudinal gradient and a decrease in functional richness in the elevational gradient. Thus, 590 we show evidences for the vertical stratification of cuticle lightness and body size in the 591 tropical ant communities of snow-free mountains as a response to the increase in elevation, 592 following by a decrease in mean temperature and vapor pressure deficit. This can be 593 interpreted by the thermal melanism hypothesis (TMH), which occurs in old mountains of 594 an OCBIL but not in new mountains in landscapes characterised by fertile soils (young, often 595 139 disturbed, fertile landscapes - YODFEL). These findings highlight the importance of the 596 climatic variables on ants’ community structuring and enhanced the importance of 597 understanding and preserve them, reaffirming the importance to monitoring ant communities 598 in a global warming crisis. Changes to climate on mountains, like temperature enhanced, 599 was predicted by IPPC World Climatic panel (Minx et al., 2017). Shifts in the relative 600 abundance or species incidence, and consequently in the functional structure of mountain 601 communities, could be a result of changes in environmental conditions that act as filters for 602 species distributions. 603 604 Acknowledgments 605 We thank all friends involved in data collection during fieldwork and processing material in lab, 606 specially LEI Team (Lab. Ecology of Insects - years 2013 to these days). We also thanks Nubia 607 Campos, Rayana Melo, Daniela Melo, Luiza Azevedo, Jéssica Martins, Matteus Carvalho, André 608 Araújo, Heron Hilário and Ivan Monteiro for help us in the field and laboratory work as well. 609 We also thank Flávio Camarota and Scott Powel (Cephalotes) and Rodrigo Feitosa for ant 610 identification. We also thank all the Espinhaço’s people, like “ Seo” José Camilo (Alitota from 611 Monte Azul - MG), Sinval, Zé do Pilão and Dona Maria (Rio de Contas – BA), for kindly 612 received us in their home and their accepting we maked our “madness” to look up the ants, wasps 613 and bees at the highlands! We thank the Fundação de Amparo à Pesquisa do Estado de Minas 614 Gerais (FAPEMIG - CRA - APQ-00311-15) for financial support. FSN received a Research 615 Productivity Fellowship from CNPq. RRCS is supported by Pesquisa & Desenvolvimento of 616 Agência Nacional de Energia Elétrica and Companhia Energética de Minas Gerais (P&D 617 ANEEL/CEMIG, PROECOS project GT-599), FAPEMIG (APQ-00288-17) and CNPq 618 (428298/2018-4) research project grants. This study was financed in part by the Coordenação de 619 140 Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001. FSN, 620 PGdS, and FSC thank the CAPES for grants. We are also grateful to Instituto Estadual de 621 Florestas (IEF-MG), Instituto Chico Mendes para Biodiversidade (ICMBio-MG and ICMBio-622 BA) and GSG for logistical support. The author(s) declare(s) that there is no conflict of interest 623 regarding the publication of this article. 624 625 References 626 Andersen AN. 1995. A Classification of Australian Ant Communities, Based on Functional 627 Groups Which Parallel Plant Life-Forms in Relation to Stress and Disturbance. Journal of 628 Biogeography 22: 15–29. 629 Anjos D V., Campos RBF & Ribeiro SP. 2015. Temporal turnover of species maintains ant 630 diversity but transforms species assemblage recovering from fire disturbance. Sociobiology. 631 Arnan X, Andersen AN, Gibb H, Parr CL, Sanders NJ, Dunn RR, Angulo E, Baccaro FB, 632 Bishop TR, Boulay R, Castracani C, Cerdá X, Del Toro I, Delsinne T, Donoso DA, Elten 633 EK, Fayle TM, Fitzpatrick MC, Gómez C, Grasso DA, Grossman BF, Guénard B, 634 Gunawardene N, Heterick B, Hoffmann BD, Janda M, Jenkins CN, Klimes P, Lach L, 635 Laeger T, Leponce M, Lucky A, Majer J, Menke S, Mezger D, Mori A, Moses J, Munyai 636 TC, Paknia O, Pfeiffer M, Philpott SM, Souza JLP, Tista M, Vasconcelos HL & Retana J. 637 2018. Dominance-diversity relationships in ant communities differ with invasion. 638 Baccaro FB, Feitosa RM, Fernandez F, Fernandes IO, Izzo TJ, Souza JLP & Solar R. 2015. 639 Guia para os gêneros de formigas do Brasil. Manaus, AM, Brazil: INPA. 640 Banschbach VS & Ogilvy E. 2014. Long-Term Impacts of Controlled Burns on the Ant 641 Community (Hymenoptera: Formicidae) of a Sandplain Forest in Vermont. Northeastern 642 141 Naturalist 21: NENHC-1-NENHC-12. 643 Barden P. 2017. Fossil ants (Hymenoptera: Formicidae): ancient diversity and the rise of modern 644 lineages. Myrmecological News 24: 1–30. 645 Baselga A. 2010. Partitioning the turnover and nestedness components of beta diversity. Global 646 Ecology and Biogeography 19: 134–143. 647 Baselga A & Orme CDL. 2012. Betapart: An R package for the study of beta diversity. Methods 648 in Ecology and Evolution 3: 808–812. 649 Basset Y, Novotny V, Miller SE & Springate ND. 1998. Assessing the impact of forest 650 disturbance on tropical invertebrates: some comments. Journal of Applied Ecology 35: 461–466. 651 Bastide H, Yassin A, Johanning EJ & Pool JE. 2014. Pigmentation in Drosophila melanogaster 652 reaches its maximum in Ethiopia and correlates most strongly with ultra-violet radiation in sub-653 Saharan Africa. BMC Evolutionary Biology 14: 1–14. 654 Begon M, Harper JL & Townsend CR. 2006. Ecology: Individuals, Populations and 655 Communities. 656 de Bello F, Lavorel S, Albert CH, Thuiller W, Grigulis K, Dolezal J, Janeček Š & Lepš J. 657 2011. Quantifying the relevance of intraspecific trait variability for functional diversity. Methods 658 in Ecology and Evolution 2: 163–174. 659 Bishop TR, Robertson MP, van Rensburg BJ & Parr CL. 2014. Elevation-diversity patterns 660 through space and time: Ant communities of the Maloti-Drakensberg Mountains of southern 661 Africa. Journal of Biogeography. 662 Bishop TR, Robertson MP, Gibb H, van Rensburg BJ, Braschler B, Chown SL, Foord SH, 663 Munyai TC, Okey I, Tshivhandekano PG, Werenkraut V & Parr CL. 2016. Ant assemblages 664 142 have darker and larger members in cold environments. Global Ecology and Biogeography 25: 665 1489–1499. 666 Bishop TR, Robertson MP, Van Rensburg BJ & Parr CL. 2017. Coping with the cold: 667 minimum temperatures and thermal tolerances dominate the ecology of mountain ants. 668 Ecological Entomology 42: 105–114. 669 Bolton B, Alpert G, Ward PS & Naskrecki P. 2005. Bolton ’ S Catalogue of Ants of the World. 670 Environmental Health: 9145. 671 Castro FS De, Gontijo AB, Castro PDTA & Ribeiro SP. 2012. Annual and seasonal changes 672 in the structure of litter-dwelling ant assemblages (Hymenoptera: Formicidae) in Atlantic 673 semideciduous forests. Psyche 2012,: 12 pages. 674 Castro FS, da Silva PG, Solar RRC, Fernandes GW & Neves FS. (in prep). Environmental 675 drivers of taxonomic and functional diversity of ant communities in a tropical mountain. 676 Manuscript submitted for publication. 677 Chao A & Jost L. 2012. Coverage-based rarefaction and extrapolation: Standardizing samples 678 by completeness rather than size. Ecology 93: 2533–2547. 679 Chao A, Wang YT & Jost L. 2013. Entropy and the species accumulation curve: A novel 680 entropy estimator via discovery rates of new species. Methods in Ecology and Evolution 4: 1091–681 1100. 682 Clusella-Trullas S, Terblanche JS, Blackburn TM & Chown SL. 2008. Testing the thermal 683 melanism hypothesis: A macrophysiological approach. Functional Ecology 22: 232–238. 684 Clusella Trullas S, van Wyk JH & Spotila JR. 2007. Thermal melanism in ectotherms. Journal 685 of Thermal Biology 32: 235–245. 686 143 Costa F V., Mello R, Lana TC & Neves FS. 2015. Ant fauna in megadiverse mountains: A 687 checklist for the rocky grasslands. Sociobiology 62: 228–245. 688 Costa F V., Blüthgen N, Viana-Junior AB, Guerra TJ, Di Spirito L & Neves FS. 2018. 689 Resilience to fire and climate seasonality drive the temporal dynamics of ant-plant interactions 690 in a fire-prone ecosystem. Ecological Indicators 93: 247–255. 691 Crawley MJ. 2013. The R Book-Second Edition. 692 Dolédec S, Chessel D, ter Braak CJF & Champely S. 1996. Matching species traits to 693 environmental variables: a new three-table ordination method. Environmental and Ecological 694 Statistics 3: 143–166. 695 Dunn RR, Agosti D, Andersen AN, Arnan X, Bruhl CA, Cerdá X, Ellison AM, Fisher BL, 696 Fitzpatrick MC, Gibb H, Gotelli NJ, Gove AD, Guenard B, Janda M, Kaspari M, Laurent 697 EJ, Lessard JP, Longino JT, Majer JD, Menke SB, McGlynn TP, Parr CL, Philpott SM, 698 Pfeiffer M, Retana J, Suarez A V., Vasconcelos HL, Weiser MD & Sanders NJ. 2009. 699 Climatic drivers of hemispheric asymmetry in global patterns of ant species richness. Ecology 700 Letters 12: 324–333. 701 Eliot C. 2007. Method and metaphysics in Clements’s and Gleason’s ecological explanations. 702 Studies in History and Philosophy of Science Part C :Studies in History and Philosophy of 703 Biological and Biomedical Sciences 38: 85–109. 704 Feener DH, Lighton JRB & Bartholomew GA. 1988. Curvilinear Allometry, Energetics and 705 Foraging Ecology: A Comparison of Leaf-Cutting Ants and Army Ants. Functional Ecology 2: 706 509–520. 707 Fernandes GW. 2016. The Megadiverse Rupestrian Grassland. In: Fernandes GW, ed. Ecology 708 144 and Conservation of Mountaintop grasslands in Brazil. Springer, Cham, 3–14. 709 Fernandes GW, Almeida HA, Nunes CA, Xavier JHA, Cobb NS, Carneiro MAA, 710 Cornelissen T, Neves FS, Ribeiro SP, Nunes YRF, Pires AC V. & Beirão M V. 2016. Cerrado 711 to Rupestrian Grasslands: Patterns of Species Distribution and the Forces Shaping Them Along 712 an Altitudinal Gradient. In: Fernandes GW, ed. Ecology and Conservation of Mountaintop 713 grasslands in Brazil. Springer, Cham, 345–377. 714 Ferrari LT, Schaefer CEGR, Fernandes RBA, Mendonça BAF, Gjorup DF, Corrêa GR & 715 E.O. S. 2016. Thermic and Hydric Dynamics of Ironstone (Canga) and Quartzite Rupestrian 716 Grasslands in the Quadrilátero Ferrífero: The Ecological Importance of Water. In: Fernandes 717 GW, ed. Ecology and Conservation of Mountaintop Grasslands in Brazil., 72–85. 718 Ferrier S, Manion G, Elith J & Richardson K. 2007. Using generalized dissimilarity 719 modelling to analyse and predict patterns of beta diversity in regional biodiversity assessment. 720 Diversity and Distributions 13: 252–264. 721 Fichaux M, Béchade B, Donald J, Weyna A, Delabie JHC, Murienne J, Baraloto C & Orivel 722 J. 2019. Habitats shape taxonomic and functional composition of Neotropical ant assemblages. 723 Oecologia 189: 501–513. 724 Gaston KJ. 2000. Global patterns in biodiversity. Nature 405: 220–7. 725 Gaston KJ. 2009. Geographic range limits: Achieving synthesis. Proceedings of the Royal 726 Society B: Biological Sciences 276: 1395–1406. 727 Gibb H & Parr CL. 2013. Does Structural Complexity Determine the Morphology of 728 Assemblages? An Experimental Test on Three Continents. PLoS ONE 8: 1–7. 729 Giulietti AM, Pirani JR & Harley RM. 1997. Centres of plant diversity: a guide and strategy 730 145 for their conservation. Oxford: Information Press: 397–4004. 731 Gleason H. 1926. The Individualistic Concept of the Plant Association Author ( s ): H . A . 732 Gleason Source : Bulletin of the Torrey Botanical Club , Vol . 53 , No . 1 , ( Jan ., 1926 ), pp . 7-733 26 Published by : Torrey Botanical Society Stable URL : http://www.jstor.org/st. Society 53: 7–734 26. 735 Gras P, Tscharntke T, Maas B, Tjoa A, Hafsah A & Clough Y. 2016. How ants, birds and 736 bats affect crop yield along shade gradients in tropical cacao agroforestry. Journal of Applied 737 Ecology 53: 953–963. 738 Guénard B, Weiser MD, Gómez K, Narula N & Economo EP. 2017. The Global Ant 739 Biodiversity Informatics (GABI) database: Synthesizing data on the geographic distribution of 740 ant species (Hymenoptera: Formicidae). Myrmecological News 24: 83–89. 741 Heino J, Alahuhta J, Fattorini S & Schmera D. 2019. Predicting beta diversity of terrestrial 742 and aquatic beetles using ecogeographical variables: insights from the replacement and richness 743 difference components. Journal of Biogeography 46: 304–315. 744 Hopper SD. 2009. OCBIL theory: Towards an integrated understanding of the evolution, 745 ecology and conservation of biodiversity on old, climatically buffered, infertile landscapes. Plant 746 and Soil 322: 49–86. 747 Hsieh TC, Ma KH & Chao A. 2016. Introduction to iNEXT ( R package ): Excerpt from iNEXT 748 User ’ s Guide. : 1–8. 749 Johnson RA. 2015. A taxonomic revision of South American species of the seed-harvester ant 750 genus Pogonomyrmex (Hymenoptera: Formicidae). Part I. Zootaxa 4029: 001–142. 751 Kalmus H. 1941. The resistance to desiccation of Drosophila mutants affecting body colour. 752 146 Proceedings of the Royal Society of London. Series B - Biological Sciences 130: 185–201. 753 Kaspari M, Clay NA, Lucas J, Yanoviak SP & Kay A. 2015. Thermal adaptation generates a 754 diversity of thermal limits in a rainforest ant community. Global Change Biology 21: 1092–1102. 755 Kaspari M & Weiser MD. 1999. The size–grain hypothesis and interspecific scaling in ants. 756 Functional Ecology 13: 530–538. 757 Keddy PA. 1992. Assembly and response rules: two goals for predictive community ecology. 758 Journal of Vegetation Science 3: 157–164. 759 Laliberté E & Legendre P. 2010. A distance-based framework for measuring functional 760 diversity from multiple traits. Ecology 91: 299–305. 761 Lamanna C, Blonder B, Violle C, Kraft NJB, Sandel B, Šímová I, Donoghue JC, Svenning 762 JC, McGill BJ, Boyle B, Buzzard V, Dolins S, Jørgensen PM, Marcuse-Kubitza A, 763 Morueta-Holme N, Peet RK, Piel WH, Regetz J, Schildhauer M, Spencer N, Thiers B, Wiser 764 SK & Enquist BJ. 2014. Functional trait space and the latitudinal diversity gradient. 765 Proceedings of the National Academy of Sciences of the United States of America 111: 13745–766 13750. 767 Law SJ, Bishop TR, Eggleton P, Griffiths H, Ashton L & Parr C. 2019. Darker ants dominate 768 the canopy: Testing macroecological hypotheses for patterns in colour along a microclimatic 769 gradient. Journal of Animal Ecology: 1–13. 770 Leal IR, Filgueiras BKC, Gomes JP, Iannuzzi L & Andersen AN. 2012. Effects of habitat 771 fragmentation on ant richness and functional composition in Brazilian Atlantic forest. 772 Biodiversity and Conservation 21: 1687–1701. 773 Leitão RP, Zuanon J, Villéger S, Williams SE, Baraloto C, Fortunel C, Mendonça FP & 774 147 Mouillot D. 2016. Rare species contribute disproportionately to the functional structure of 775 species assemblages. Proceedings of the Royal Society B: Biological Sciences 283: 20160084. 776 Leitão RP, Zuanon J, Mouillot D, Leal CG, Hughes RM, Kaufmann PR, Villéger S, Pompeu 777 PS, Kasper D, de Paula FR, Ferraz SFB & Gardner TA. 2018. Disentangling the pathways 778 of land use impacts on the functional structure of fish assemblages in Amazon streams. 779 Ecography 41: 219–232. 780 Longino JT & Branstetter MG. 2019. The truncated bell: an enigmatic but pervasive 781 elevational diversity pattern in Middle American ants. Ecography 42: 272–283. 782 Losey JE & Vaughan M. 2006. The Economic Value of Ecological Services Provided by 783 Insects. BioScience 56: 311. 784 Machac A, Janda M, Dunn RR & Sanders NJ. 2011. Elevational gradients in phylogenetic 785 structure of ant communities reveal the interplay of biotic and abiotic constraints on diversity. 786 Ecography 34: 364–371. 787 Maire E, Grenouillet G, Brosse S & Villéger S. 2015. How many dimensions are needed to 788 accurately assess functional diversity? A pragmatic approach for assessing the quality of 789 functional spaces. Global Ecology and Biogeography 24: 728–740. 790 Manion G, Lisk M, Ferrier S, Nieto-Lugilde D, Mokany K & Fitzpatrick MC. 2018. Package 791 ‘gdm’: Generalized Dissimilarity Modeling Version. 792 Måren IE, Kapfer J, Aarrestad PA, Grytnes JA & Vandvik V. 2018. Changing contributions 793 of stochastic and deterministic processes in community assembly over a successional gradient. 794 Ecology 99: 148–157. 795 Mason NWH, Mouillot D, Lee WG & Wilson JB. 2005. Functional richness, functional 796 148 evenness and functional divergence: The primary components of functional diversity. Oikos 111: 797 112–118. 798 de Melo E. 2000. Polygonaceae da Cadeia do Espinhaço, Brasil. Acta Botanica Brasilica 14: 799 273–300. 800 Minx JC, Callaghan M, Lamb WF, Garard J & Edenhofer O. 2017. Learning about climate 801 change solutions in the IPCC and beyond. Environmental Science and Policy 77: 252–259. 802 Mouillot D, Graham NAJ, Villéger S, Mason NWH & Bellwood DR. 2013. A functional 803 approach reveals community responses to disturbances. Trends in Ecology & Evolution 28: 167–804 177. 805 Moura MR, Costa HC, Argôlo AJS & Jetz W. 2017. Environmental constraints on the 806 compositional and phylogenetic beta-diversity of tropical forest snake assemblages. Journal of 807 Animal Ecology 86: 1192–1204. 808 Nation JL. 2008. Integument: Structure and Function. In: Capinera JL, ed. Encyclopedia of 809 Entomology. Dordrecht: Springer Netherlands, 2015–2019. 810 Nunes CA, Braga RF, Figueira JEC, Neves F de S & Fernandes GW. 2016. Dung Beetles 811 along a Tropical Altitudinal Gradient: Environmental Filtering on Taxonomic and Functional 812 Diversity (O Ovadia, Ed.). PLOS ONE 11: e0157442. 813 Nunes CA, Quintino A V., Constantino R, Negreiros D, Reis Júnior R & Fernandes GW. 814 2017. Patterns of taxonomic and functional diversity of termites along a tropical elevational 815 gradient. Biotropica 49: 186–194. 816 Paolucci LN, Maia MLB, Solar RRC, Campos RI, Schoereder JH & Andersen AN. 2016. 817 Fire in the Amazon: impact of experimental fuel addition on responses of ants and their 818 149 interactions with myrmecochorous seeds. Oecologia 182: 335–346. 819 Parr CL, Dunn RR, Sanders NJ, Weiser MD, Photakis M, Bishop TR, Fitzpatrick MC, 820 Arnan X, Baccaro F, Brand�o CRF, Chick L, Donoso DA, Fayle TM, G�mez C, 821 Grossman B, Munyai TC, Pacheco R, Retana J, Robinson A, Sagata K, Silva RRR, Tista 822 M, Vasconcelos H, Yates M, Gibb H, Brandão CRF, Chick L, Donoso DA, Fayle TM, 823 Gómez C, Grossman B, Munyai TC, Pacheco R, Retana J, Robinson A, Sagata K, Silva 824 RRR, Tista M, Vasconcelos H, Yates M, Gibb H, Brand�o CRF, Chick L, Donoso DA, 825 Fayle TM, G�mez C, Grossman B, Munyai TC, Pacheco R, Retana J, Robinson A, Sagata 826 K, Silva RRR, Tista M, Vasconcelos H, Yates M & Gibb H. 2017. GlobalAnts: a new database 827 on the geography of ant traits (Hymenoptera: Formicidae). Insect Conservation and Diversity 10: 828 5–20. 829 Perillo LN, Castro FS, Solar RRC & Neves FS. (in prep.). Can we disentangle the importance 830 of latitudinal and altitudinal gradients of variation for ants, bees and wasps? Manuscript in 831 preparation. 832 Petchey OL & Gaston KJ. 2006. Functional diversity: Back to basics and looking forward. 833 Ecology Letters 9: 741–758. 834 Peters MK, Hemp A, Appelhans T, Behler C, Classen A, Detsch F, Ensslin A, Ferger SW, 835 Frederiksen SB, Gebert F, Haas M, Helbig-Bonitz M, Hemp C, Kindeketa WJ, Mwangomo 836 E, Ngereza C, Otte I, Röder J, Rutten G, Schellenberger Costa D, Tardanico J, Zancolli G, 837 Deckert J, Eardley CD, Peters RS, Rödel MO, Schleuning M, Ssymank A, Kakengi V, 838 Zhang J, Böhning-Gaese K, Brandl R, Kalko EKV, Kleyer M, Nauss T, Tschapka M, 839 Fischer M & Steffan-Dewenter I. 2016. Predictors of elevational biodiversity gradients change 840 from single taxa to the multi-taxa community level. Nature Communications 7. 841 150 R Core Team. 2019. R: A language and environment for statistical computing. R Foundation 842 for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/. 843 Ricotta C. 2017. Of beta diversity, variance, evenness, and dissimilarity. Ecology and Evolution 844 7: 4835–4843. 845 Rosa AHB, Perillo LN, Neves FS, Ribeiro DB & Freitas AVL. 2019. Butterflies collected 846 using malaise traps as useful bycatches for ecology and conservation. Journal of Threatened 847 Taxa 11: 14235–14237. 848 Schluter D & Pennell MW. 2017. Speciation gradients and the distribution of biodiversity. 849 Nature 546: 48–55. 850 Schweiger AH & Beierkuhnlein C. 2016. Size dependency in colour patterns of Western 851 Palearctic carabids. Ecography 39: 846–857. 852 Silva PG, Lobo JM, Hensen MC & Hernández FZV de mello MIM. 2018. Turnover and 853 nestedness in subtropical dung beetle assemblages along an elevational gradient. : 1–14. 854 Silva RR & Brandão CRF. 2014. Ecosystem-wide morphological structure of leaf-litter ant 855 communities along a tropical latitudinal gradient. PLoS ONE 9. 856 Silveira FAO, Negreiros D, Barbosa NPU, Buisson E, Carmo FF, Carstensen DW, 857 Conceição AA, Cornelissen TG, Echternacht L, Fernandes GW, Garcia QS, Guerra TJ, 858 Jacobi CM, Lemos-Filho JP, Le Stradic S, Morellato LPC, Neves FS, Oliveira RS, Schaefer 859 CE, Viana PL & Lambers H. 2016. Ecology and evolution of plant diversity in the endangered 860 campo rupestre: a neglected conservation priority. Plant Soil 403: 129–152. 861 Smith AR. 1978. Color gamut transform pairs. ACM SIGGRAPH Computer Graphics 12: 12–862 19. 863 151 Smith MA. 2015. Ants, elevation, phylogenetic diversity and community structure. Ecosphere 864 6: art221. 865 Soares NC & Morellato LPC. 2018. Crepuscular pollination and reproductive ecology of 866 Trembleya laniflora (Melastomataceae), an endemic species in mountain rupestrian grasslands. 867 Flora: Morphology, Distribution, Functional Ecology of Plants 238: 138–147. 868 Southwood TRE. 1977. Habitat as the template for ecological strategies? The Journal of animal 869 ecology 46: 337–365. 870 Stein A, Gerstner K & Kreft H. 2014. Environmental heterogeneity as a universal driver of 871 species richness across taxa, biomes and spatial scales. Ecology Letters. 872 Stevens RD, Cox SB, Strauss RE & Willig MR. 2003. Patterns of functional diversity across 873 an extensive environmental gradient: Vertebrate consumers, hidden treatments and latitudinal 874 trends. Ecology Letters 6: 1099–1108. 875 Swenson NG, Anglada-Cordero P & Barone JA. 2011. Deterministic tropical tree community 876 turnover: Evidence from patterns of functional beta diversity along an elevational gradient. 877 Proceedings of the Royal Society B: Biological Sciences 278: 877–884. 878 Terblanche JS. 2012. Thermal relations. In: Chapman RF, Simpson SJ, Douglas AE, eds. The 879 Insects: Structure and Function. New York: Cambridge University Press, 588–621. 880 Tiede Y, Schlautmann J, Donoso DA, Wallis CIB, Bendix J, Brandl R & Farwig N. 2017. 881 Ants as indicators of environmental change and ecosystem processes. Ecological Indicators 83: 882 527–537. 883 Tuomisto H. 2010. A diversity of beta diversities: Straightening up a concept gone awry. Part 2. 884 Quantifying beta diversity and related phenomena. Ecography 33: 23–45. 885 152 Turner JS & Lombard AT. 1990. Body color and body temperature in white and black Namib 886 desert beetles. Journal of Arid Environments 19: 303–315. 887 Vandewalle M, de Bello F, Berg MP, Bolger T, Dolédec S, Dubs F, Feld CK, Harrington R, 888 Harrison PA, Lavorel S, da Silva PM, Moretti M, Niemelä J, Santos P, Sattler T, Sousa JP, 889 Sykes MT, Vanbergen AJ & Woodcock BA. 2010. Functional traits as indicators of 890 biodiversity response to land use changes across ecosystems and organisms. Biodiversity and 891 Conservation 19: 2921–2947. 892 Vasconcelos HL, Maravalhas JB, Feitosa RM, Pacheco R, Neves KC & Andersen AN. 2018. 893 Neotropical savanna ants show a reversed latitudinal gradient of species richness, with climatic 894 drivers reflecting the forest origin of the fauna. Journal of Biogeography 45: 248–258. 895 Vellend M. 2016. The Theory of Ecological Communities (MPB-57). Princeton University Press. 896 Villéger S, Grenouillet G & Brosse S. 2013. Decomposing functional β-diversity reveals that 897 low functional β-diversity is driven by low functional turnover in European fish assemblages. 898 Global Ecology and Biogeography 22: 671–681. 899 Villéger S, Mason NWH & Mouillot D. 2008. New Multidimensional Functional Diversity 900 Indices for a Multifaceted Framework in Functional Ecology Published by : Ecological Society 901 of America content in a trusted digital archive. Ecology 89: 2290–2301. 902 Villéger S, Novack-Gottshall PM & Mouillot D. 2011. The multidimensionality of the niche 903 reveals functional diversity changes in benthic marine biotas across geological time. Ecology 904 Letters 14: 561–568. 905 Weber NA. 1938. The biology of the fungus-growing ants. Part IV. Additional new forms. Part 906 V. The Attini of Bolivia. Biología de las hormigas cultivadoras de hongos. Parte IV. Nuevas 907 153 formas adicionales. Part V. Las Attini de Bolivia. Revista de Entomologia 7: 154–206. 908 Wills BD, Powell S, Rivera MD & Suarez A V. 2017. Correlates and Consequences of Worker 909 Polymorphism in Ants. Annual Review of Entomology 63: 575–598. 910 Yates ML, Andrew NR, Binns M & Gibb H. 2014. Morphological traits: predictable responses 911 to macrohabitats across a 300 km scale. PeerJ 2: e271. 912 913 154 Appendix 1 - Supplementary Information 1 A.1. Information of the 12 campo rupestre mountains across Espinhaço Mountain Range, southeast Brazil. 2 Locality (Abbreviation) Lower elevation sites Upper elevation sites Latitude Longitude Elevation Latitude Longitude Elevation P.E. Serra do Ouro Branco (OB) 20o 30' 20.10'' S 43o 37' 23.21 '' W 1150 20o 29' 12.37'' S 43o 42' 43.96'' W 1569 P.E. Pico do Itacolomi (OP) 20o 28' 53.04'' S 43o 27' 48.14'' W 1137 20o 25' 40.44' S 43o 28' 50.96'' W 1583 RPPN Santuário do Caraça (CA) 20o 04' 37.99'' S 43o 29' 37.90'' W 1207 20o 08' 07.19'' S 43o 27' 08.87'' W 2066 Pico do Breu - Serra do Cipó (PB) 19o 05' 46.00'' S 43o 41' 14.89'' W 1108 19o 05' 44.02'' S 43o 39' 56.42'' W 1583 P.E. Pico do Itambé (IT) 18o 24' 00.45'' S 43o 18' 05.42'' W 1140 18o 23' 52.59'' S 43o 20' 11.85'' W 1798 P.E. do Rio Preto (RP) 18o 09' 35.89'' S 43o 19' 14.18'' W 1018 18o 13' 05.21'' S 43o 18' 56.35'' W 1601 Botumirim (BO) 16o 53' 15.06'' S 43o 01' 50.16'' W 1122 16o 50' 27.23'' S 43o 04' 15.56'' W 1420 P.E. Serra Nova (SN) 15o 43' 11.67'' S 42o 50' 10.32'' W 1043 15o 43' 33.77'' S 42o 49' 42.76'' W 1276 Pico da Formosa - Monte Azul (PF) 15o 12' 25.39'' S 42o 48' 04.10'' W 1151 15 o 13' 54.52'' S 42o 48' 54.52'' W 1416 Pico das Almas (AL) 13o 30' 33.86'' S 41o 53' 29.94'' W 1171 13o 31' 01.16'' S 41o 57' 29.63'' W 1564 Pico do Barbado (BA) 13o 15' 43.80'' S 41o 52' 37.72'' W 1165 13o 17' 47.70'' S 41o 54' 13.08'' W 1943 P.N. Chapada Diamantina – Guiné (GU) 12o 45' 48.28'' S 41o 30' 40.05'' W 1114 12o 45' 10.47'' S 41o 30' 20.59'' W 1396 3 155 A.2. Environmental variables extracted from WorldClim data (version 2, 1970-2000). Units: Wind (Mean and variance): m s-1; VPD (vapor pressure deficit) (Mean and variance): kPa; UV-B (Mean and variance): kJ m-2 day-1; Precipitation (Mean and variance): mm; TMax (Mean and variance): ºC; Temperature (Mean and variance): ºC; Elevation: m a.s.l. (above sea level); Longitude and Latitude: decimal degrees (DD). Bold columns represent the variables selected for GDM and GLM's (with *) models after Pearson’s correlation multicollinearity test. Site Wind (Mean) *VPD (Mean) *UV-B (Mean) Precipitation (Mean) TMax (Mean) *Temperature (Mean) Wind (variance) VPD (variance) UV-B (variance) Precipitation (variance) TMax (variance) Temperature (variance) *Elevation Longitude *Latitude OB Lower 1.63 1.63 14884.93 123.97 23.44 17.94 0.23 0.43 2623531.80 10775.53 5.26 4.23 1149.61 -43.6231147 -20.5055833 OB Upper 1.86 1.46 14905.33 132.83 21.92 16.56 0.24 0.36 2227368.24 12114.70 4.81 3.91 1569.47 -43.7122106 -20.4867697 OP Lower 1.57 1.69 14614.72 122.20 24.02 18.22 0.23 0.45 3907401.20 9867.01 6.17 4.66 1137.12 -43.4633711 -20.4814008 OP Upper 1.91 1.43 14250.02 137.45 21.13 15.91 0.22 0.35 3760208.77 11471.25 4.96 4.03 1582.92 -43.4808217 -20.4279011 CA Lower 1.58 1.62 14510.32 132.45 23.12 17.65 0.20 0.40 4009626.69 11461.68 5.06 3.92 1207.00 -43.4938572 -20.0772203 CA Upper 2.11 1.25 14682.25 143.58 18.79 14.00 0.22 0.27 2509313.84 12874.27 4.35 3.76 2066.00 -43.4524625 -20.1353306 PB Lower 1.37 1.65 15587.92 122.58 24.23 18.56 0.14 0.37 2801926.45 12010.81 5.36 4.08 1108.00 -43.6874689 -19.0961119 PB Upper 1.54 1.50 14613.45 135.03 22.28 16.98 0.16 0.31 3056712.32 12400.66 4.44 3.53 1582.60 -43.6656711 -19.0955617 IT Lower 1.29 1.66 14801.87 123.58 23.41 18.22 0.16 0.30 3878510.71 10760.49 4.61 3.76 1139.76 -43.3015058 -18.4001244 IT Upper 1.62 1.39 14883.00 129.50 20.39 15.65 0.17 0.23 2856372.91 11553.91 3.35 2.85 1797.78 -43.3366247 -18.3979431 RP Lower 1.20 1.76 15239.07 110.48 24.96 19.41 0.16 0.33 4149096.59 9517.74 5.04 3.60 1017.91 -43.3206047 -18.1599686 RP Upper 1.51 1.47 15199.88 121.97 21.88 16.53 0.17 0.24 3908683.17 10880.80 3.71 2.90 1600.95 -43.3156525 -18.2181147 BO Lower 1.35 1.70 15491.68 87.60 25.67 19.81 0.16 0.22 4172134.97 6657.52 3.33 2.74 1122.46 -43.0306000 -16.8875178 BO Upper 1.53 1.55 15625.25 89.77 23.84 18.26 0.18 0.19 3620895.62 7109.12 2.91 2.49 1420.22 -43.0709894 -16.8408969 SN Lower 1.58 1.69 16965.08 74.67 25.70 19.93 0.19 0.18 2785448.08 5599.33 2.87 2.73 1042.62 -42.8361989 -15.7199089 SN Upper 1.72 1.57 16663.75 75.25 24.32 18.75 0.23 0.16 3288265.07 5553.69 2.72 2.60 1276.03 -42.8285450 -15.7260461 PF Lower 1.79 1.61 16770.15 75.00 24.96 19.49 0.24 0.19 3418514.43 5193.71 2.04 1.95 1150.70 -42.8011381 -15.2070519 PF Upper 1.92 1.51 16680.88 76.72 23.55 18.27 0.28 0.17 3338734.19 5257.14 2.02 1.93 1416.27 -42.8151450 -15.2318111 AL Lower 2.18 1.67 16743.42 92.42 25.23 19.41 0.21 0.09 3732885.17 4040.45 1.48 1.41 1170.76 -41.8916511 -13.5094061 AL Upper 2.54 1.35 16935.75 92.08 20.87 15.73 0.25 0.08 3073133.48 3081.72 1.41 1.36 1563.67 -41.9582300 -13.5169889 BA Lower 2.17 1.68 16845.58 78.68 25.37 19.65 0.21 0.10 3632592.00 3474.26 1.67 1.61 1165.18 -41.8771439 -13.2621669 BA Upper 2.39 1.48 16955.82 88.00 22.76 17.54 0.26 0.09 3153682.19 3775.27 1.58 1.51 1942.59 -41.9036339 -13.2965831 GU Lower 2.12 1.74 16594.23 70.02 25.34 19.63 0.14 0.08 3559622.99 1902.33 1.65 1.60 1113.67 -41.5111250 -12.7634100 GU Upper 2.26 1.62 16571.00 78.65 23.88 18.33 0.18 0.08 4152143.79 1565.68 1.66 1.57 1396.29 -41.5057219 -12.7529081 4 156 A.3. Species of ants captured and traits used to calculate functional diversity (FD): (1) morphological 5 traits (a) WL: Weber’s length (μm); (b) HFL: Hind femur length (μm); (c) V: measure of colour 6 brightness (%); (d) Workers polymorphism: categorical (monomorphic, dimorphic and polymorphic); 7 (e) integument sculpture (ordinal data (smooth/often shine, intermediate and texturized); and (2) 8 ecological traits or life history traits (a): functional groups based on global-scale responses of ants to 9 environmental stress and disturbance AA = Army Ants; AD = Arboreal Dominant; AP = Arboreal 10 Predator; AS = Arboreal Subordinate; CO = Cryptic Omnivores; CP = Cryptic Predators; DD = 11 Dominant Dolichoderinae; EO = Epigeic Omnivores; EP = Epigeic Predators; Hatt = High Attini; Latt 12 = Low Attini; OPP = Opportunist; SC = Subordinate Camponotini SH = Seed Harvester; UT = 13 Underground trophobiont. 14 Morphospecies Polymorphism Sculpturing FG WL HFL V Acromyrmex cf. niger polymorphic intermediate Hatt 2337.03 1.1170 0.5702 Acromyrmex RPsp.1 polymorphic intermediate Hatt 1208.43 0.9377 0.8235 Acromyrmex sp.2 polymorphic intermediate Hatt 2325.11 1.0800 0.6856 Acromyrmex sp.4 polymorphic intermediate Hatt 2186.92 1.1318 0.6307 Acromyrmex subterraneus polymorphic intermediate Hatt 2000.10 1.2960 0.6327 Acropyga goeldii monomorphic smooth UT 523.92 0.8404 0.9216 Anochetus sp.1 monomorphic intermediate EP 2620.73 0.8879 0.7882 Apterostigma gp. pilosum sp.1 monomorphic intermediate Latt 1381.88 0.9575 0.7660 Apterostigma gp. pilosum sp.2 monomorphic intermediate Latt 1359.25 1.0274 0.7046 Atta bisphaerica polymorphic intermediate Hatt 3187.84 1.2435 0.7680 Atta sexdens polymorphic intermediate Hatt 2999.99 1.5110 0.6895 Azteca sp.1 polymorphic smooth AD 1180.16 1.0282 0.6333 Brachymyrmex cf. pictus monomorphic smooth OPP 347.35 0.8303 0.6725 Brachymyrmex cordemoyi monomorphic smooth OPP 491.83 0.8734 0.6797 Camponotus (Tanaemyrmex) sp.3 dimorphic intermediate SC 1848.11 0.8847 0.7078 Camponotus atriceps polymorphic intermediate SC 2657.95 1.0252 0.6359 Camponotus bidens dimorphic intermediate SC 1197.02 0.7747 0.8510 Camponotus blandus dimorphic intermediate SC 2180.12 0.9366 0.6765 Camponotus cingulatus dimorphic intermediate SC 2648.08 0.9730 0.6255 Camponotus crassus dimorphic intermediate SC 1741.61 0.8071 0.7621 Camponotus fastigatus dimorphic intermediate SC 1505.01 0.8217 0.8490 Camponotus lespesii dimorphic intermediate SC 4171.45 1.0810 0.6078 Camponotus leydigi dimorphic intermediate SC 2993.01 0.8530 0.6928 Camponotus melanoticus polymorphic intermediate SC 2976.27 1.0032 0.7882 Camponotus novogranadensis dimorphic intermediate SC 1802.05 0.7962 0.7405 Camponotus renggeri polymorphic intermediate SC 3430.93 0.9354 0.7755 Camponotus rufipes polymorphic intermediate SC 2998.64 0.9685 0.6634 Camponotus sp.6 dimorphic intermediate SC 1944.56 0.8435 0.8523 Camponotus textor dimorphic intermediate SC 2096.84 0.9408 0.7314 Camponotus vitatus dimorphic intermediate SC 3200.65 0.8942 0.6431 Camponotus westermanni dimorphic intermediate SC 1781.13 0.9258 0.7732 Cephalotes depressus dimorphic intermediate AS 1250.75 0.6224 0.9712 Cephalotes maculatus dimorphic intermediate AS 974.54 0.6177 0.8902 Cephalotes minutus dimorphic intermediate AS 1112.30 0.7101 0.7673 157 Morphospecies Polymorphism Sculpturing FG WL HFL V Cephalotes pinelli dimorphic intermediate AS 1011.15 0.6333 0.9412 Cephalotes pusillus dimorphic intermediate AS 1249.10 0.6933 0.8157 Crematogaster acuta polymorphic intermediate AD 1238.96 0.8884 0.9373 Crematogaster BAsp.1 polymorphic intermediate AD 843.72 0.8336 0.8010 Crematogaster BOsp.1 polymorphic intermediate AD 751.35 0.8796 0.6444 Crematogaster brasiliensis polymorphic intermediate AD 709.58 0.7649 0.8301 Crematogaster cf. erecta polymorphic intermediate AD 912.38 0.9013 0.8510 Crematogaster obscurata polymorphic intermediate AD 631.46 0.8671 0.8745 Crematogaster OBsp.1 polymorphic intermediate AD 731.79 0.7231 0.6020 Crematogaster PFsp.1 polymorphic intermediate AD 840.56 0.8665 0.7349 Crematogaster SNsp.1 polymorphic intermediate AD 719.89 0.8525 0.6745 Crematogaster sp.1 polymorphic intermediate AD 653.64 0.7660 0.6314 Crematogaster sp.10 polymorphic intermediate AD 557.10 0.7844 0.7948 Crematogaster sp.7 polymorphic intermediate AD 678.52 0.7722 0.9059 Crematogaster sp.8 polymorphic intermediate AD 566.21 0.7182 0.5804 Crematogaster sp.9 polymorphic intermediate AD 631.88 1.0087 0.6451 Cyphomyrmex sp.1 monomorphic intermediate Latt 896.27 0.8682 0.6569 Cyphomyrmex sp.2 monomorphic intermediate Latt 944.99 0.8775 0.7771 Cyphomyrmex sp.6 monomorphic intermediate Latt 894.34 0.9035 0.6643 Dorymyrmex brunneus monomorphic smooth OPP 1181.72 1.0537 0.6935 Dorymyrmex goeldii monomorphic smooth OPP 1316.34 0.9711 0.6170 Dorymyrmex piramicus monomorphic smooth OPP 1019.26 1.0432 0.8516 Dorymyrmex sp.5 monomorphic smooth OPP 1103.42 0.9810 0.7340 Dorymyrmex sp.6 monomorphic smooth OPP 1046.68 0.9757 0.6373 Eciton vagans polymorphic intermediate AA 2167.79 0.9079 0.9059 Ectatomma brunneum monomorphic textured EP 3758.86 0.8571 0.7203 Ectatomma edentatum monomorphic textured EP 2850.07 0.9294 0.7732 Ectatomma opaciventre monomorphic textured EP 4508.32 0.9763 0.7739 Ectatomma permagnum monomorphic textured EP 3887.23 0.7976 0.7418 Ectatomma tuberculatum monomorphic textured EP 3876.18 0.8324 0.9882 Eurhopalothrix bruchi monomorphic textured CP 473.06 0.6214 0.6412 Forelius maranhaoensis monomorphic smooth DD 778.43 1.0943 0.6928 Gnamptogenys cf. menozzii monomorphic textured EP 1660.06 0.7318 0.4706 Gnamptogenys gp. striatula monomorphic textured EP 1526.58 0.8254 0.7791 Gnamptogenys sp.2 monomorphic textured EP 1592.71 0.7780 0.8235 Gnamptogenys sp.3 monomorphic textured EP 1266.40 0.7647 0.9503 Gnamptogenys sulcata monomorphic textured EP 1700.09 0.7670 0.5229 Hypoponera distinguenda monomorphic smooth CP 1055.38 0.7148 0.7549 Kalathomyrmex emeryi monomorphic intermediate Latt 901.01 0.7900 0.5882 Kalathomyrmex sp.1 monomorphic intermediate Latt 889.14 0.8910 0.9098 Labidus praedator polymorphic smooth AA 1783.46 0.9842 0.6176 Linepithema cerradensis monomorphic smooth EO 741.50 0.7646 0.7484 Linepithema iniqum monomorphic smooth EO 863.79 0.7920 0.6980 Linepithema micans monomorphic smooth EO 864.63 0.7855 0.8353 Linepithema neotropicum monomorphic smooth EO 756.42 0.7582 0.7627 158 Morphospecies Polymorphism Sculpturing FG WL HFL V Mycetophylax sp.5 monomorphic textured Latt 916.62 0.8726 0.7245 Mycetophylax sp.7 monomorphic textured Latt 1035.32 0.8986 0.8137 Mycetophylax sp.8 monomorphic textured Latt 675.74 0.8108 0.7592 Mycocepurus goeldii monomorphic textured Latt 1008.20 0.8545 0.6993 Mycocepurus smithii monomorphic textured Latt 721.08 0.7189 0.8235 Myrmelachista ruszkii monomorphic intermediate AS 560.56 0.4227 0.2824 Myrmelachista sp.1 monomorphic intermediate AS 599.48 0.5816 0.8608 Myrmicocrypta sp.1 monomorphic intermediate Latt 949.19 0.9161 1.0000 Myrmicocrypta sp.2 monomorphic intermediate Latt 910.19 0.8981 0.8196 Neivamyrmex pseudops polymorphic intermediate AA 1423.86 0.9358 0.5459 Neivamyrmex sp.2 polymorphic intermediate AA 916.45 0.6682 0.3765 Neoponera verenae monomorphic intermediate EP 3666.04 0.8217 0.9000 Nesomyrmex sp.2 monomorphic textured AS NA NA NA Nomamyrmex hartigii polymorphic intermediate AA 2065.72 0.8517 0.7176 Nylanderia sp.1 monomorphic smooth OPP 768.83 0.9711 0.6612 Nylanderia steinheili monomorphic smooth OPP 807.71 0.9246 0.7105 Ochetomyrmex semipolitus monomorphic intermediate EO 509.50 0.6875 0.7229 Odontomachus bauri monomorphic intermediate EP 3521.84 0.7690 0.8853 Odontomachus meinerti monomorphic intermediate EP 2505.14 0.8407 0.6000 Oxyepoecus sp.2 monomorphic intermediate EP 664.17 0.6242 0.7778 Oxyepoecus sp.5 monomorphic intermediate EP 673.24 0.6470 0.7059 Pachycondyla striata monomorphic intermediate EP 4989.69 0.7926 0.8464 Pachycondyla harpax monomorphic intermediate EP 3189.35 0.6839 0.7755 Pheidole capillata dimorphic textured EO 858.38 0.9582 0.8595 Pheidole diligens dimorphic intermediate EO 652.69 0.9257 0.6908 Pheidole diligens sp.2 dimorphic intermediate EO 713.50 0.9993 0.7359 Pheidole gp. flavens sp.1 dimorphic intermediate EO 456.76 0.8902 0.6359 Pheidole gp. flavens sp.2 dimorphic intermediate EO 462.98 0.8950 0.6712 Pheidole gp. flavens sp.3 dimorphic intermediate EO 625.07 0.9062 0.8837 Pheidole oxyops dimorphic intermediate EO 826.96 1.0488 0.7268 Pheidole radoszkowskii dimorphic intermediate EO 768.17 1.0651 0.7203 Pheidole sp.1 dimorphic intermediate EO 1214.31 1.0727 0.8699 Pheidole sp.10 dimorphic intermediate EO 799.29 1.0343 0.8569 Pheidole sp.11 dimorphic intermediate EO 849.37 0.7436 0.7314 Pheidole sp.12 dimorphic intermediate EO 712.90 0.8446 0.8190 Pheidole sp.13 dimorphic intermediate EO 848.68 0.9264 0.2855 Pheidole sp.14 dimorphic intermediate EO 769.03 1.0553 0.6000 Pheidole sp.17 dimorphic intermediate EO 1229.00 1.2434 0.7908 Pheidole sp.18 dimorphic intermediate EO 705.65 0.9362 0.7451 Pheidole sp.19 dimorphic intermediate EO 669.13 1.0206 0.6369 Pheidole sp.20 dimorphic intermediate EO 806.47 1.0412 0.7209 Pheidole sp.21 dimorphic intermediate EO 705.09 0.8181 0.4549 Pheidole sp.22 dimorphic intermediate EO 1014.65 1.2990 0.7843 Pheidole sp.23 dimorphic intermediate EO 1102.34 1.0660 0.7085 Pheidole sp.24 dimorphic intermediate EO 1338.25 1.2973 0.6569 159 Morphospecies Polymorphism Sculpturing FG WL HFL V Pheidole sp.25 dimorphic intermediate EO 980.23 1.1163 0.8980 Pheidole sp.26 dimorphic intermediate EO 654.23 0.8736 0.6059 Pheidole sp.27 dimorphic intermediate EO 1371.20 1.0651 0.6275 Pheidole sp.5 dimorphic intermediate EO 858.84 0.8645 0.7928 Pheidole sp.6 dimorphic smooth EO 976.26 0.9256 0.6752 Pheidole sp.7 dimorphic intermediate EO 945.03 1.0890 0.7843 Pheidole sp.9 dimorphic intermediate EO 802.47 1.0872 0.6908 Pheidole subarmata dimorphic smooth EO 580.50 0.8749 0.8634 Pheidole vafra dimorphic intermediate EO 845.15 1.0654 0.6451 Platythyrea cf. angusta monomorphic textured AP 2777.42 0.7959 0.9804 Pogonomyrmex naegelii monomorphic textured SH 1581.88 0.8488 0.7856 Pseudomyrmex gp. pallidus sp.1 monomorphic intermediate AS 1042.05 0.5026 0.6346 Pseudomyrmex sp.1 monomorphic intermediate AS 2329.55 0.5927 0.8784 Pseudomyrmex sp.2 monomorphic intermediate AS 1444.89 0.5267 0.7412 Pseudomyrmex termitarius monomorphic intermediate AS 1795.57 0.6547 0.6706 Pseudoponera sp.1 monomorphic intermediate EP 2287.51 0.6513 0.7739 Pseudoponera sp.2 monomorphic intermediate EP 2415.93 0.6065 0.7098 Rasopone sp.1 monomorphic intermediate EP 736.78 0.5505 0.6627 Rogeria besucheti monomorphic textured CO 757.25 0.7424 0.5255 Sericomyrmex amabilis monomorphic intermediate Hatt 1514.82 0.8861 0.7392 Sericomyrmex sp.1 monomorphic intermediate Hatt 1341.57 0.9268 0.1569 Solenopsis globularia sp.1 polymorphic smooth CO 648.44 0.6436 0.7876 Solenopsis globularia sp.2 polymorphic smooth CO 677.46 0.6618 0.6948 Solenopsis saevissima polymorphic smooth EO 906.83 0.7177 0.8425 Solenopsis sp.1 monomorphic smooth CO 416.84 0.5984 0.7281 Solenopsis sp.2 polymorphic smooth CO 562.92 0.6815 0.8516 Solenopsis sp.3 polymorphic smooth EO 577.03 0.6985 0.7569 Solenopsis sp.7 polymorphic smooth EO 491.50 0.6557 0.7993 Solenopsis substituta polymorphic smooth EO 914.39 0.8837 0.8085 Strumigenys cf. grytava monomorphic textured CP 500.71 0.5342 0.8235 Strumigenys louisianae monomorphic textured CP 541.94 0.7431 0.7588 Strumigenys subedentata monomorphic textured CP 716.89 0.6806 0.8549 Trachymyrmex ALsp.2 monomorphic intermediate Latt 1992.15 1.0158 0.8275 Trachymyrmex ALsp.3 monomorphic intermediate Latt 1666.33 0.9331 0.8353 Trachymyrmex RPsp.1 monomorphic intermediate Latt 1474.82 0.9021 0.6654 Trachymyrmex sp.1 monomorphic intermediate Latt 1486.19 0.8954 0.7405 Trachymyrmex sp.2 monomorphic intermediate Latt 1705.35 0.8988 0.7255 Trachymyrmex sp.3 monomorphic intermediate Latt 1851.45 0.9483 0.7373 Trachymyrmex sp.4 monomorphic intermediate Latt 1051.19 0.8351 0.7634 Trachymyrmex sp.5 monomorphic intermediate Latt 1858.54 0.9407 0.7147 Trachymyrmex sp.8 monomorphic intermediate Latt 2023.80 0.8932 0.8353 Tranopelta gilva monomorphic smooth CO 644.49 0.6121 0.8627 Wasmannia affinis monomorphic textured EO 583.92 0.7648 0.8752 Wasmannia auropunctata monomorphic textured EO 495.41 0.8352 0.6641 Wasmannia lutzi monomorphic textured EO 549.84 0.7439 0.9049 15 160 A.4. Assembly weighted averages of lightness (AWM V) and body size (AWM WL), calculated for 16 each community (N=24). 17 Site AWM_V AWM_WL Site AWM_V AWM_WL OB Lower 0.59 7.65 BO Lower 0.39 6.91 OB Upper 0.41 6.67 BO Upper 0.43 6.01 OP Lower 0.63 6.50 SN Lower 1.00 14.99 OP Upper 0.38 4.96 SN Upper 0.39 3.44 CA Lower 0.46 8.47 PF Lower 0.63 8.15 CA Upper 0.20 1.60 PF Upper 0.48 6.37 PB Lower 0.90 9.74 AL Lower 0.83 13.90 PB Upper 0.52 10.03 AL Upper 0.60 8.78 IT Lower 0.57 6.43 BA Lower 0.74 8.71 IT Upper 0.38 6.42 BA Upper 0.19 3.73 RP Lower 0.43 5.33 GU Lower 0.64 8.97 RP Upper 0.56 11.31 GU Upper 0.36 7.21 18 161 A.5. Records of Formicidae species sampling in campo rupestre localities across Espinhaço Mountain Range, southeast Brazil. Legend: 1) Ouro Branco; 19 2) Ouro Preto; 3) Caraça; 4) Pico do Breu; 5) Pico do Itambé; 6) Rio Preto; 7) Botumirim; 8) Serra Nova; 9) Monte Azul; 10) Pico das Almas; 11) Pico 20 Barbados; 12) P.N. Chapada Diamantina – Guiné; Sobs= Taxonomic richness; Records = % Frequency records of specimens on the whole system. 21 Lower elevation sites Upper elevation sites Subfamily/ Morphospecies 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 Genera Dolichoderinae (Sobs=11; Records= 12.2%) Azteca Azteca sp.1 X Dorymyrmex Dorymyrmex brunneus X X X X X X X Dorymyrmex goeldii X X X X X X X X X X Dorymyrmex piramicus X X X Dorymyrmex sp.5 X X Dorymyrmex sp.6 X Forelius Forelius maranhaoensis X X X Linepithema Linepithema cerradensis X X X X X X X Linepithema iniqum X Linepithema micans X X X X X X X X X X X X X X X X X X Linepithema neotropicum X Dorylinae (Sobs=5; Records= 0.4%) Eciton Eciton vagans X Labidus Labidus praedator X Neivamyrmex Neivamyrmex pseudops X X Neivamyrmex sp.2 X Nomamyrmex Nomamyrmex hartigii X Ectatomminae (Sobs=10; Records= 6.8%) Ectatomma Ectatomma brunneum X X X X X 162 Lower elevation sites Upper elevation sites Subfamily/ Morphospecies 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 Genera Ectatomma edentatum X X X X X X X X X X X X X X X X X Ectatomma opaciventre X X Ectatomma permagnum X Ectatomma tuberculatum X Gnamptogenys Gnamptogenys cf. menozzii X Gnamptogenys gp. striatula X X X X X X Gnamptogenys sp.2 X Gnamptogenys sp.3 X X Gnamptogenys sulcata X X X Formicinae (Sobs=24; Records= 26.6%) Acropyga Acropyga goeldii X Brachymyrmex Brachymyrmex cf. pictus X X X X X X X X X X X X X X X X X Brachymyrmex cordemoyi X X X X X X X X X X X X X X X Camponotus Camponotus atriceps X X X X X Camponotus bidens X Camponotus blandus X X X X Camponotus cingulatus X X X X X X X X X X X X X Camponotus crassus X X X X X X X X X X X X X X X X X Camponotus fastigatus X X X X X X X X X X Camponotus lespesii X Camponotus leydigi X X X Camponotus melanoticus X X X X X X X X X X X X X X X X X X X Camponotus novogranadensis X X X X Camponotus renggeri X X Camponotus rufipes X X X X X X X X X X X X X X X 163 Lower elevation sites Upper elevation sites Subfamily/ Morphospecies 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 Genera Camponotus sp.6 X X X X X Camponotus Tanaemyrmex sp.3 X Camponotus textor X X Camponotus vitatus X X X Camponotus westermanni X X X X X X X X X X X X X X X X X X X Myrmelachista Myrmelachista ruszkii X Myrmelachista sp.1 X X X X X Nylanderia Nylanderia sp.1 X X Nylanderia steinheili X X X X X X X X X X Myrmicinae (Sobs=104; Records= 47.0%) Acromyrmex Acromyrmex cf. niger X Acromyrmex RPsp.1 X Acromyrmex sp.2 X X Acromyrmex sp.4 X X X X X X X Acromyrmex subterraneus X X X X X Apterostigma Apterostigma gp. pilosum sp.1 X Apterostigma gp. pilosum sp.2 X Atta Atta bisphaerica X X Atta sexdens X Cephalotes Cephalotes depressus X Cephalotes maculatus X Cephalotes minutus X X X Cephalotes pinelli X Cephalotes pusillus X X X X X X X Crematogaster Crematogaster acuta X 164 Lower elevation sites Upper elevation sites Subfamily/ Morphospecies 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 Genera Crematogaster BAsp.1 X Crematogaster BOsp.1 X Crematogaster brasiliensis X X X Crematogaster cf. erecta X Crematogaster obscurata X Crematogaster OBsp.1 X X Crematogaster PFsp.1 X Crematogaster SNsp.1 X Crematogaster sp.1 X X Crematogaster sp.10 X Crematogaster sp.7 X Crematogaster sp.8 X Crematogaster sp.9 X Cyphomyrmex Cyphomyrmex sp.1 X X X Cyphomyrmex sp.2 X X X X X X Cyphomyrmex sp.6 X Eurhopalothrix Eurhopalothrix bruchi X Kalathomyrmex Kalathomyrmex emeryi X Kalathomyrmex sp.1 X Mycetophylax Mycetophylax sp.5 X X Mycetophylax sp.7 X Mycetophylax sp.8 X Mycocepurus Mycocepurus goeldii X X X Mycocepurus smithii X X Myrmicocrypta Myrmicocrypta sp.1 X 165 Lower elevation sites Upper elevation sites Subfamily/ Morphospecies 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 Genera Myrmicocrypta sp.2 X Nesomyrmex Nesomyrmex sp.2 X Ochetomyrmex Ochetomyrmex semipolitus X X X X X Oxyepoecus Oxyepoecus sp.2 X X X X Oxyepoecus sp.5 X Pheidole Pheidole capillata X X X X X X Pheidole diligens X X X X X X X X X X X Pheidole diligens sp.2 X Pheidole gp. flavens sp.1 X X X X Pheidole gp. flavens sp.2 X X X X X X X X X X Pheidole gp. flavens sp.3 X X Pheidole oxyops X X X X X X X X X X X X X X X X X X X Pheidole radoszkowskii X X X X X X X X X X X X X X X X X Pheidole sp.1 X X X X X X Pheidole sp.10 X X X X X X X Pheidole sp.11 X Pheidole sp.12 X X X X X Pheidole sp.13 X X Pheidole sp.14 X X Pheidole sp.17 X X X Pheidole sp.18 X Pheidole sp.19 X Pheidole sp.20 X X Pheidole sp.21 X Pheidole sp.22 X 166 Lower elevation sites Upper elevation sites Subfamily/ Morphospecies 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 Genera Pheidole sp.23 X X Pheidole sp.24 X Pheidole sp.25 X X Pheidole sp.26 X Pheidole sp.27 X Pheidole sp.5 X X Pheidole sp.6 X X X X X X X Pheidole sp.7 X Pheidole sp.9 X X Pheidole subarmata X X X X Pheidole vafra X X X X X X Pogonomyrmex Pogonomyrmex naegelii X X X X X X X X X X X X X X X Rogeria Rogeria besucheti X Sericomyrmex Sericomyrmex sp.1 X Sericomyrmex sp.2 X X X X Solenopsis Solenopsis globularia sp.1 X X X X X X X Solenopsis globularia sp.2 X X Solenopsis saevissima X X X X X X X X X X X X X Solenopsis sp.1 X X X X X X X X X X X X X X X X X Solenopsis sp.2 X X X X X X X Solenopsis sp.3 X Solenopsis sp.7 X X X X X X X X X X X X Solenopsis substituta X X X X Strumigenys Strumigenys cf. grytava X Strumigenys louisianae X X 167 Lower elevation sites Upper elevation sites Subfamily/ Morphospecies 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 Genera Strumigenys subedentata X Trachymyrmex Trachymyrmex ALsp.2 X Trachymyrmex ALsp.3 X Trachymyrmex RPsp.1 X Trachymyrmex sp.1 X Trachymyrmex sp.2 X X X X X X X X X X X X Trachymyrmex sp.3 X Trachymyrmex sp.4 X X X Trachymyrmex sp.5 X X X Trachymyrmex sp.8 X Tranopelta Tranopelta gilva X X Wasmannia Wasmannia affinis X X X X X Wasmannia auropunctata X X X X X X X X X X X X X X X X X Wasmannia lutzi X X X Ponerinae (Sobs=11; Hits= 4.7%) Anochetus Anochetus sp.1 X Hypoponera Hypoponera distinguenda X X Neoponera Neoponera verenae X X Odontomachus Odontomachus bauri X X X Odontomachus meinerti X Pachycondyla Pachycondyla striata X X X X X X X X X X X X Pachycondyla harpax X X Platythyrea Platythyrea cf. angusta X Pseudoponera Pseudoponera sp.1 X X Pseudoponera sp.2 X 168 Lower elevation sites Upper elevation sites Subfamily/ Morphospecies 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 Genera Rasopone Rasopone sp.1 X Pseudomyrmicinae (Sobs=4; Records= 2.3%) Pseudomyrmex Pseudomyrmex gp. pallidus sp.1 X X X X X X X X Pseudomyrmex sp.1 X Pseudomyrmex sp.2 X 22 169 23 (a) (b) (c) (d) (e) (f) A.6.a. Rarefaction-extrapolation species accumulation curves of ant richness for each site in the 12 mountains survey: A:(OB) Ouro Branco; B: (OP) Pico do Itacolomi, Ouro Preto; C: (CA) RPPN Santuário do Caraça; D: (PB) Pico do Breu, Serra do Cipó; E: (IT) Parque Estadual Pico do Itambé; F: (RP) Parque Estadual do Rio Preto. Colour code: dark orange = Lower Sites; dark blue = Upper Sites. 24 170 (g) (h) (i) (j) (k) (l) A.6.b. Rarefaction-extrapolation species accumulation curves of ant richness for each site in the 12 mountains survey: G: (BO) Botumirim; H: (SN) Parque Estadual Serra Nova; I: (PF) Pico da Formosa, Monte Azul; J: (AL) Pico das Almas; K: (BA) Pico do Barbado; L:. (GU) Parque Nacional Chapada Diamantina, Guiné. Colour code: dark orange = Lower Sites; dark blue = Upper Sites. 25 171 A.7. Table with the results of rarefaction-extrapolation accumulation curves of the 24 sites of the 12 mountains in the Espinhaço Mountain Range. Sample coverage is expressed by C.ha index Site S.(obs) Sample Coverage Site S (obs) Sample Coverage OB Lower 29 0.9127 BO Lower 24 0.9144 OB Upper 32 0.7969 BO Upper 23 0.8322 OP Lower 31 0.9119 SN Lower 48 0.9384 OP Upper 25 0.8677 SN Upper 27 0.7816 CA Lower 28 0.9251 PF Lower 37 0.8787 CA Upper 10 0.8086 PF Upper 26 0.8851 PB Lower 44 0.8934 AL Lower 44 0.9202 PB Upper 25 0.9243 AL Upper 26 0.9503 IT Lower 26 0.9016 BA Lower 44 0.8301 IT Upper 20 0.9145 BA Upper 18 0.741 RP Lower 25 0.8445 GU Lower 24 0.9238 RP Upper 23 0.9689 GU Upper 18 0.8942 26 172 Conclusão Geral 27 Apesar do elevado turnover de espécies de formigas em todas as dimensões 28 espaço-temporais ao longo do gradiente de elevação e de latitude, as comunidades de 29 formigas são funcionalmente redundantes, ou seja, as espécies mudam entre habitats, 30 elevações ou latitudes, mas as principais características funcionais e funções ecológicas 31 permanecem basicamente inalteradas. 32 Na escala da montanha, diferenças taxonômicas na composição da comunidade 33 como consequência da elevada diversidade β taxonômica ao longo do gradiente em 34 distâncias geográficas curtas enfatizam a importância de conservar toda a montanha, pois 35 a perda de qualquer parte da comunidade pode acarretar perda de diversidade taxonômica 36 e funcional. Como as espécies são restritas, caso as comunidades de topo de montanha 37 não sejam preservadas, haverá uma possível perda das principais funções ecológicas 38 avaliadas neste estudo. As comparações dos padrões taxonômicos e funcionais 39 demonstraram a importância e relevância do uso de diferentes facetas de β diversidade 40 (taxonômico e funcional) em escalas distintas (espaço-tempo). Com essas informações 41 sobre dos padrões das múltiplas facetas da diversidade, podemos relacionar diretamente 42 nossas observações e descobertas sobre biodiversidade à ecologia aplicada. Por exemplo, 43 ao elaborar políticas públicas ou identificar áreas prioritárias para conservação da 44 biodiversidade. 45 No gradiente latitudinal, as espécies de formigas de campos rupestre tendem a ser 46 menores e mais escuras no topo das montanhas do que comunidades em locais mais 47 baixos, independentemente da latitude. Apesar do padrão redundante de diversidade 48 funcional encontrada no gradiente latitudinal existe uma forte e positiva relação entre o 49 brilho da cor da cutícula e o tamanho do corpo para as comunidades de formigas em 50 173 montanhas antigas. Além dos efeitos de elevação, a diminuição da temperatura média e 51 déficit de pressão de vapor em função do aumento da elevação atuam como filtros 52 ambientais, sendo os principais fatores estruturadores da diversidade taxonômica e 53 funcional das formigas em gradientes latitudinais. Por serem consideradas organismos 54 termofílicos, essas variáveis são de extrema importância para a estruturação da 55 comunidade das formigas nos ambientes montanos de campo rupestre. 56 Investigar a relação entre os padrões de diversidade taxonômica e funcional e a 57 distribuição de formigas em ambientes montanos tropicais nos mostrou a grande 58 capacidade adaptativa das comunidades de formigas frente aos diferentes filtros 59 encontrados nesses ambientes, evidenciando a importância do entendimento de como 60 esses organismos tão antigos podem viver e interagir no e com o ambiente natural, bem 61 como as mudanças nesses ambientes podem afetar a estruturação dessas comunidades. 62 63