Biogeography and Diversification of Neotropical Costaceae

Federal University of Rio de Janeiro

Biodiversity and Evolutionary Biology Graduate Program

Biogeography and Diversification of

Neotropical Costaceae

Thiago José de Carvalho André

A thesis submitted to the Biodiversity

and Evolutionary Biology Graduate

Program of the Federal University of

Rio de Janeiro in partial fulfillment of

the requirements for the degree of

Doctor in Biodiversity and

Evolutionary Biology.

Advisor: Dr Tânia Wendt

Co-Advisor: Dr Clarisse Palma-Silva

Co-Advisor: Dr Chelsea Specht

Rio de Janeiro

February 2015

Federal University of Rio de Janeiro

Biodiversity and Evolutionary Biology Graduate Program

Biogeography and Diversification of Neotropical Costaceae

Thiago José de Carvalho André

A thesis submitted to the Biodiversity and Evolutionary Biology Graduate Program of the

Federal University of Rio de Janeiro in partial fulfillment of the requirements for the

degree of Doctor in Biodiversity and Evolutionary Biology.

Approved by:

Dr Carlos Eduardo Guerra Schrago

Universidade Federal do Rio de Janeiro (UFRJ)

Dr Catarina da Fonseca Lira de Medeiros

Jardim Botânico do Rio de Janeiro (JBRJ)

Dr Evandro Marsola de Moraes

Universidade Federal de São Carlos (UFSCAR)

Dr Fábio Pinheiro

Universidade Estadual Paulista, Campus de Rio Claro (UNESP)

Dr José Ricardo Miras Mermudes

Universidade Federal do Rio de Janeiro (UFRJ)

Rio de Janeiro (RJ), February 2015!

! iii!

!

André, Thiago José de Carvalho

Biogeography and Diversification of Neotropical

Costaceae / Thiago José de Carvalho André – Rio de

Janeiro: UFRJ, 2015.

168 p.: il.; 29,7 cm.

Advisor: Dr Tânia Wendt; Co-Advisor: Dr Clarisse

Palma-Silva; Co-Advisor: Dr Chelsea Specht

Doctoral Thesis – UFRJ / Biodiversity and Evolutionary

Biology Graduate Program, 2015.

1. Evolutionary Biology. 2. Biogeography. 3. Plant

Speciation.

4. Phylogenetic Systematics. 5. Costaceae.

I. Title. II. Thesis

!

! iv!

para Amanda Mortati

! v!

“Claridade quente da manhã vaidosa.

O sol deve ter posto lente nova,

e areou todas as manchas,

para esperdiçar luz.

Dez esquadrilhas de periquitos verdes

receberam ordem de partida,

deixando para as araras cor de fogo,

o pequizeiro morto.

E a árvore, esgalhada e seca, se faz verde,

vermelha e castanha, entre os

mochoqueiros,

braúnas, jatobás e imbaúbas do morro,

na paisagem que um pintor daltônico

pincelou no dorso de um camaleão.

E o lombo da serra é tão bonito e claro,

que até uma coruja,

tonta e míope na luz,

com grandes óculos redondos,

fica trepada no cupim, o dia inteiro,

imóvel e encolhida, admirando as cores,

fatigada, talvez, de tanta erudição…”

João Guimarães Rosa

In Magma (1936)

! vi!

Agradecimentos / Acknowledgements

Este trabalho é fruto de muitas colaborações que só foram possíveis pela imensa

confiaça e maravilhosa orientação que recebi de Tânia Wendt, Clarisse Palma-Silva e

Chelsea Specht. Muito obrigado!

This is the result of much collaborative work only possible through the huge trust and

awesome guidance I received from Tânia Wendt, Clarisse Palma-Silva and Chelsea Specht.

Thank you very much!

Agradeço:

à CAPES pelas bolsas, de doutorado e de estágio de doutorado sanduíche no

exterior;

à banca de defesa de tese e à banca de exame de qualificação por terem avaliado

este trabalho e pelas valiosas contribuições;

aos professores e técnico do Programa de Pós Graduação em Biodiversidade e

Biologia Evolutiva da UFRJ, em especial à coordenadora Michelle Klatau;

à tantas pessoas e instituições que me auxiliaram nas expedições de campo, no

Amazonas, Pará, Mato Grosso, Amapá, Acre, Rondônia, Goiás, Tocantins, Minas Gerais,

Bahia e Espírito Santo;

aos curadores e equipe técnica das coleções botânicas: do Jardim Botânico do Rio

de Janeiro, do Museu Nacional do Rio de Janeiro, do Instituto de Pesquisas Científicas e

Tecnológicas do Amapá, Jardim Botânico de Brasília, Empresa Brasileira de Pesquisa

Agropecuária, EMPRAPA Amazônia Oriental de Belém, Instituto Brasileiro de Geografia e

Estatística em Brasília, Instituto Nacional de Pesquisas da Amazônia, Museu Paraense

Emílio Goeldi, Universidade de Brasília, Universidade Federal do Acre em Rio Branco,

Universidade Federal de Goiás em Goiânia, e Universidade Federal de Uberlândia;

! vii!

aos pesquisadores, alunos e técnicos do Instituto de Botânica de São Paulo, em

especial ao Rafael, Cecília e Rodrigo;

aos colegas do Laboratório de Sistemática Vegetal da UFRJ, Thiago Coser e

Cristiano Lira;

à minha família - os Carvalho, os Mortati e a Berenice também, claro! - que sempre

me apoiou e incansavelmente torceu com muito amor pelo meu sucesso, em especial aos

meus pais, Rita de Cássia de Carvalho André e Washington Luiz André (in memoriam);

à minha amada e linda esposa, Amanda Mortati: por tudo tudo tudo!

Many thanks to:

the Specht lab: Shayla Salzman, Roxana Yochteng, Chodon Sass, Stephen Yee,

Ana Almeida, Grady Pierroz, Jared Nathanson, Maria Eduarda Maldaner, Kelsie Morioka,

Gracie Benson-Martin, Colin Hill, Susan Hepp, Alma Pineyro Nelson, Riva Bruenn, and of

course, Chelsea Specht and her beautiful view from the bay. I learned so much from each

and every one of you… we really are the champions;

the Evolutionary Genetics Lab of the Museum of Vertebrate Zoology, especially the

amazing Lydia Smith;

staff and curators of US herbaria: the Jepson and University Herbaria at UC

Berekely, the Missouri Botanical Garden, the New York Botanical Garden, and Harvard

Herbaria;

Paul Maas and Dave Skinner, for sharing amazing knowledge on Costaceae;

David Bloom, Rob Guralnick and Michele Koo for help with potential distribution

models; Nick Matzke and Dan Rabosky for helping and teaching me how to run their

amazing analytical softwares; Rasmus Nielsen, Monty Slatkin and Patrick O’Grady for

allowing me to learn from their perfect classes at UC Berkeley, and for unvaluable advices;

staff from the University of California at Berkeley, especially from the Plant and

Microbial Biology Department and the International Office.

! viii!

Summary

Abstract…………………………………………………………………………………...…………xvi

Resumo…………………………………………………………………………………...………...xvii

Chapter 1. Introduction: Speciation in Neotropical Costaceae ………..………………………..1

Chapter 2 – Spatiotemporal dynamics of speciation in Neotropical Costaceae…….……….15

2.1. Introduction…………………………………………………………………….………17

2.2. Methods………………………………………………………………………...…...…21

2.3. Results……………………………………………………………………….…………25

2.4. Discussion……………………………………………………………………...………34

2.5. Acknowledgements…………………………………………………………………...38

2.6. References…………………………………………………………………..………...38

Chapter 3 – Spiraling into history: A molecular phylogeny and investigation of biogeographic

origins and floral evolution for the genus Costus...……………………………………………...51

3.1. Introduction………………………………………………………………….…………53

3.2. Methods………………………………………………………………………...……...58

3.3. Results………………………………………………………………………….………65

3.4. Discussion……………………………………………………………………...………80

3.5. Acknowledgments……………………………………………………………..……...84

3.6. References……………………………………………………………………..……...84

Chapter 4 – Evolution of species diversity in the genus Chamaecostus (Costaceae):

molecular phylogenetics and morphometric approaches.......................................................90

4.1. Introduction……………………………………………………………….……………92

4.2. Methods……………………………………………………………………...………...96

4.3. Results……………………………………………………………………………….…99

4.4. Discussion……………………………………………………………………..…..…105

4.5. Taxonomy……………….…………………………………………………...…….…109

4.6. Acknowledgments…………………………………………………………..….……114

4.7. References……………………………………………………………………...……114

! ix!

Chapter 5 – Speciation in the South American dry understory: lessons from Chamaecostus

(Costaceae, Zingiberales)………………………………………………………………………...120

5.1. Introduction…………………………………………………………………….…..…122

5.2. Methods………………………………………………………………………...….…125

5.3. Results………………………………………………………………………….….…134

5.4. Discussion……………………………………………………………………...….…143

5.5. Acknowledgments……………………………………………………………..….…147

5.6. References……………………………………………………………………..….…147

Chapter 6 – Final remarks……………………………………………………………………..…160

Appendix 1………………………………………………………………………………..………..163

Appendix 2…………………………………………………………………………………..……..166

! x!

List of Figures

Chapter 1. Speciation in Neotropical Costaceae: where are we?

Figure 1. From upper left to bottom right: Monocostus uniflorus K.Schum.; Costus L.

stem showing spiromonistichous ligulate leaves with a closed sheath; Chamaecostus

lanceolatus (Petersen) C.D.Specht & D.W.Stev.; Dimerocostus strobilaceus Kuntze;

Costus ligularis Baker; Tapeinochilos ananassae K. Schum.; Costus L. inflorescence

and fruits…………………......……………..….……………………………………………….4

Figure 2. From upper left to bottom right: Costus spiralis (Jacq.) Roscoe; Costus

arabicus L.; Chamaecostus lanceolatus (Petersen) C.D.Specht & D.W.Stev. flower,

calyx, bract, bracteole and petals; Costus pulverulentus C. Presl; Costus arabicus L.

petaloid stamen and pistil; Chamaecostus subsessilis (Nees & Mart.) C.D.Specht &

D.W.Stev. flower……………………………………………………………………………….5

Chapter 2 – Spatiotemporal dynamics of speciation in Neotropical Costaceae

Figure 1. Maximum clade credibility tree estimated from 1,000 trees. Numbers above

branches refer to posterior probabilities, and numbers at right of nodes are age

estimates in million years. Blue bars denote node height probability density at 95%.

Lower scale in million years before present.………...……………………….……………26

Figure 2. Top: (A) Biogeographical analysis of Costaceae using BioGeoBEARS. (B)

Phylorate plots for speciation rate using BAMM; colors at each point in time along

branches denote instantaneous speciation rates, with warmer colors referring to faster

rates; two distinct shift configurations account for most of the posterior probability of

the data with the dark dot indicating the node of the single shift in configuration.

Bottom: Geographic areas included in biogeographic analysis: ! Central America and

the Caribbean, ! Amazonian, ! Andean, ! Central Brazilian Plateau and Atlantic

Rainforests, ! Africa, ! Asia and Oceania. Outgroups are not shown. Central images

show representative Costaceae (top-down): Costus scaber Ruiz & Pav., Costus

arabicus L., Costus ligularis Baker, Chamaecostus lanceolatus (Petersen) C.D. Specht

& D.W. Stev.; C. ligularis photograph by C.Specht, all others by T.André……………..28

! xi!

Figure 3. Speciation rates comparison within Costaceae phylogeny. Color bar and

squares indicate proportion of similarity between rates. Colors at each point in time

along branches of the phylogeny denote instantaneous speciation rates. Outgroup rate

is underestimated due to limited sampling of outgroup clades………………………….30

Figure 4. Top: Degree of sympatry by node age for (A) the South American clade and

(B) the Neotropical Costus clade. Bottom: Maps showing examples of recovered

potential current distributions superimposed for one selected clade from

Chamaecostus (A) and Costus (B)………………………………………………..……32-33


origins and floral evolution for the genus Costus

Figure 1. Representative photos of Neotropical Costus species. (A) Costus

guanaiensis Rusby var. guanaiensis showing the melittophilous morphology. (B)

Costus comosus and (C) Costus scaber showing the ornithophilous morphology.

Photos by C. D. Specht....………………………………………………………………...…55

Figure 2. Maximum likelihood hypothesis for Neotropical Costus. ML cladogram (with

support values) showing character state reconstruction data and phylogram (with

branch lengths) for the concatenated alignment of ITS, ETS, CaM, rps16, and trnL-F

using TIM3 with gamma distribution and log likelihood score of -23,944. Support values

at nodes are Bayesian posterior probabilities, ML, and MP bootstrap proportions.

Nodes supported by a PAUP* strict consensus of 707,000 most parsimonious trees are

shown in bold with 50% consensus in bold grey. Distributions were obtained from

herbaria records housed in the Global Biodiversity Information Facility. Ecoregions are

based on the World Wildlife Fund’s ecoregion designations (Olsen et al. 2001).

Ancestral distribution for the Neotropical clade indicated (Pacific Coast of Mexico and

Central America) based on S-DIVA and MCMC algorithms implemented in RASP. ML

reconstructions performed in Mesquite and BayesTraits are indicated: shifts to

ornithophilous morphology are denoted with hummingbirds and shifts to melittophilous

morphology denoted with bees. Extant pollination syndromes are denoted with

hummingbirds or bees at tips……………………………………………………………..…67

Figure 3. Reconstruction of ancestral character states for morphology-based pollination

syndrome on ML tree. Generalist, melittophilous and ornithophilous morphologies are

! xii!

indicated with likelihood support values from Mesquite and BayesTraits and posterior

probabilities from BayesTraits………………………………………………………………72


syndrome on Bayesian tree. Generalist, melittophilous, and ornithophilous

morphologies are indicated with likelihood support values from Mesquite and

BayesTraits and posterior probabilities from BayesTraits………………………………..73


molecular phylogenetics and morphometric approaches

Figure 1. Schematic representation of measured morphometric variables; LL – Leaf

Length, LW – Leaf Maximum Width, AA – Apex Angle, BA – Base

Angle.………………………………………………...........................................................98

Figure 2. A – Phylogenetic relationships within Chamaecostus (Costaceae). Support

values above branches are Bayesian Posterior Probabilities, while Most Likely

Bootstrap proportions from 1,000 bootstrap replicates are found below branches; B –

Geographic ranges of the Chamaecostus cuspidatus (blue), and Chamaecostus

subsessilis complex: subsessilis clade (orange) and acaulis clade (green). Dashed

lines denote tentative range limits, and the continuous grey line identifies the Araguaia

River……………………………………………………………………………………….…100

Figure 3. Box and whisker plots of three significantly different morphometric variables

between Chamaecostus cuspidatus (n=14), Chamaecostus subsessilis s.str. (n=51)

and Chamaecostus acaulis comb. nov. (n=83), showing means, quartiles and ranges.

A – Leaf length (cm); B – Leaf Maximum Width (cm); C – Leaf Area (cm2)……….…104

Figure 4. Chamaecostus acaulis comb. nov. and Chamaecostus subsessilis s.str.. (B)

photo by W.W.Thomas. (D) photo by D.Skinner……………………………………...…111

! xiii!


(Costaceae, Zingiberales)

Figure 1. Locations of the seventeen sampled sites of Chamaecostus subsessilis (MG,

JA, DF, TG, TO, PK) and C. acaulis (UB, PA, GO, MT, NX, AF, TS, CS, RO, PM, AC).

Topographic variation is shown in the background, with green lowlands.……………126

Figure 2. Relationships of individuals based on the nuclear DNA concatenated

alignment of Chamaecostus subsessilis and C. acaulis.. Maximum Clade Credibility

phylogeny is shown at top, with posterior probabilities of each node shown. The bottom

of the figure shows Bayesian admixture proportions (Q) of individual estimated by

STRUCTURE, assuming K = 7……………………………………………………………137

Figure 3. Relationships of individuals based on the plastid DNA concatenated

alignment. Maximum Clade Credibility phylogeny is shown at top, with posterior

probabilities of each node shown. The bottom of the figure shows Bayesian admixture

proportions (Q) of individual estimated by STRUCTURE, assuming K = 8…………..138

! xiv!

List of Tables


origins and floral evolution for the genus Costus

Table 1. Primer pairs and annealing temperatures used in this study. Primers were

used for both amplification and sequencing....…………………………….………………60

Table 2. Character values and models of evolution for each marker and the

concatenated alignment…………………………………………………………………..…62

Table 3. BayesTraits ML probabilities, Bayesian posterior probabilities, and their

support values at each node tested for ancestral pollination reconstruction using the

ML tree. Nodes supported as having ancestral hummingbird pollination morphology are

underlined……………………………………………………………………………………..75

Table 4. BayesTraits ML probabilities, Bayesian posterior probabilities, and their

support values at each node tested for ancestral state reconstruction of pollination

syndrome using the Bayesian topology. Nodes supported as having ancestral

hummingbird pollination syndrome morphology are underlined…………………………77



Table 1. Morphometric variables tested for diagnose between Chamaecostus

cuspidatus and the Chamaecostus subsessilis complex. Means ± standard deviations;

different letters are indicative of statistical significance (p<0.05; t-test), and F-values

and probabilities of ANOVA are given. Bold values refer to variables that are

significantly different between all three species.…………………………………..........103


(Costaceae, Zingiberales)

Table 1. Name and location of sampled Chamaecostus subsessilis and C. acaulis

populations and respective number of sequences obtained for plastid and nuclear

markers.……………………………………………………………………………..............127

! xv!

Table 2. Outgroup taxa and respective approximate time (Ma) to the most recent

common ancestor with Chamaecostus subsessilis (Nees & Mart.) C.D.Specht &

D.W.Stev……………………………………………………………………………………..128

Table 3. Character values, number of haplotypes, and models of evolution for each

plastid (cp) and nuclear (nr) DNA marker and the concatenated alignments of

Chamaecostus subsessilis and C. acaulis……………………………………….…130-131

Table 4. Parameters for selecting optimal partioning structure. Calculated for plastid

(cp) and nuclear (nr) DNA concatenated alignments of Chamaecostus subsessilis and

C. acaulis. K: Number of population clusters. Likelihood scores for each value of K

genetic clusters from STRUCTURE (Pritchard et al. 2000), Delta K scores for each

value of K genetic clusters following Evanno et al. (2005)……………………………..136

Table 5. Pairwise FST between populations*. Plastid (cp, below diagonal) and nuclear

(nr, above diagonal) DNA concatenated alignments considered individually, of

Chamaecostus subsessilis and C. acaulis**…………………………………………..…139

Table 6. Population molecular diversity. Plastid (cp) and nuclear (nr) DNA

concatenated alignments considered independently……………………………………141

! xvi!

Abstract

The pantropical family of spiral gingers (Costaceae Nakai, Zingiberales) has six genera

comprising about 125 species, distributed primarily in Tropical America, Western Africa,

Southeast Asia and Papuasia-Australia. Here, main aim is to examine spatiotemporal

patterns and processes of speciation in Neotropical Costaceae. A robust dated molecular

phylogeny with most current species reveals a significant variation in speciation rate among

clades. Sympatric processes were likely running Costus radiation in the Neotropics, which

currently comprises nearly half of total familial species richness. Neotropical Costus exhibit

an evolutionary toggle in pollination morphologies, demonstrated by both the multiple

independent evolutions of ornithophily as well as reversals to melittophily. In the Neotropics,

besides Costus, another Costaceae lineage can be found: a small, early divergent clade

endemic to South America (consisting of Monocostus, Dimerocostus and Chamaecostus).

Phylogenetic results show that while Chamaecostus is strongly monophyletic,

Chamaecostus cuspidatus is found to be sister to a clade of some but not all samples of

Chamaecostus subsessilis, making it necessary to acknowledge more than one species in

the Chamaecostus subsessilis complex. Within these lineages there is a deep

phylogeographic structure, recovered at both the population and species level, as inferred

from fixation indices, structure analysis and phylogenetics considered after targeted

hightroughput sequencing by capture using PCR-generated probes. Spatiotemporal patterns

of diversification suggested a divergence between Eastern and Western lineages in the

Eocene. Results uphold Costaceae as an emergening strategic family to understand

speciation patterns and processes in Tropical landscapes.

! xvii!

Resumo

A família botânica das canas-do-brejo ou canas-de-macaco (Costaceae Nakai, Zingiberales)

apresenta distribuição pantropical de seis gêneros compreendendo cerca de 125 espécies,

distribuídas primordialmente na América Tropical, no Oeste da África, no Sudoeste da Ásia

e na Papuásia-Austrália. Aqui, o objetivo prinicipal é examinar os padrões e processos

espaciais e temporais da especiação em Costaceae Neotropicais. Uma filogenia molecular

robusta e datada com ampla maioria das espécies atuais suporta uma varição significantiva

na taxa de especiação entre clados. Processos simpátricos provavelmente levaram à

radiação de Costus na região Neotropical, que compreende praticamente à metade da

riqueza de espécies da família. O clado dos Costus Neotropicais exibe uma alternância

evolutiva na morfologia relacionada à polinização, evidenciado tanto pelas múltiplas origens

independentes da ornitofilia como por reversões à melitofilia. Nos Neotrópicos, além de

Costus, outra linhagem de Costaceae pode ser encontrada: um pequeno clado de

divergência inicial, amplamente endêmico à América do Sul (consistindo em Monocostus,

Dimerocostus e Chamaecostus). Resultados filogenéticos demonstram que, enquanto

Chamaecostus é fortemente monofilético, Chamaecostus cuspidatus é determinado como

clado irmão a algumas mas não todas amostras de Chamaecostus subsessilis, fazendo

necessário o reconhecimento de mais de uma espécie no complexo. Entre essas linhagens

há uma profunda estrutura filogeográfica, revelada tanto na escala populacional quanto

específica, e inferida a partir de índices de fixação, análise de estrutura genética e análise

filogenética, a partir de sequenciamento hightroughput direcionado pelo método de captura

utilizando iscas geradas por PCR. Padrões espaciais e temporais de diversificação sugerem

divergência entre uma linhagem Leste e uma Oeste, durante o Eoceno. Resultados

suportam Costaceae como uma família estratégica emergente para o entendimento de

padrões e processos relacionados à especiação em paisagens tropicais.

! xviii!

!!

! 1!

Chapter 1 – Introduction: Speciation in Neotropical Costaceae

!!

! 2!

The pantropical family of spiral gingers (Costaceae Nakai, Zingiberales) has six genera

comprising about 125 species, distributed primarily in Tropical America, Western Africa,

Southeast Asia and Papuasia-Australia. Spiral gingers can be easily recognised vegetatively

due to their spiromonistichous phyllotaxy and ligulate leaves with a sheath that fully encloses

the stem (Figure 1). Typically, spiral shoots terminate in a condensed, bracteate

inflorescence, from which typically a single fertile flower emerges each day (Stiles 1978;

Schemske 1983; Kay & Schemske 2003). Inflorescences are typically dense, spicate-

capitate, bearing large bracts that are often brightly colored, with or without appendages.

Inflorescence bracts often have extrafloral nectaries (callus) that are visited by ants. In most

species, each inflorescence bract subtends a single axillary flower. Flowers are

monosymmetric with a large labellum comprised from the fusion of five infertile, laminar

staminodes. A single fertile stamen is also petaloid and often blocks entry to the floral tube

and access of pollinators to basal nectaries. The style is very thin and thread-like and is

supported between the two halves (thecae) of the large anther (Figure 2). The stigma in

most species has a protrusion below the stigmatic surface that attaches the stigma to the

fertile thecae and positions the receptive surface above the pollen (Figure 2). Flowers are

unscented and diurnal, producing copious nectar during the hours in which they are open

(Stiles 1978, 1981). There are two to four rows of ovules, and the endosperm is oily

(Panchaksharappa 1963, Tomlinson 1969, Grootjen & Bouman 1981, Kirchoff 1988,

Newman & Kirchoff 1992, Larsen et al. 1998).

All flowers of Costaceae have a tubular labellum comprised of five fused petaloid

staminodes that can be white, yellow, red or orange in color. The labellum may be either

short and broad with a distinct limb bearing red, purple, or yellow stripes and a UV

absorptive ‘landing platform’; or it can form a narrow tube that is largely contained within the

petals (no limb) and lacking in stripes or patterning (Figures 1 and 2). These two

morphologies have been reported to associate with different pollination syndromes,

attracting bees and hummingbirds, respectively. Specht et al. (2001) additionally indicates a

!!

! 3!

generalist pollination syndrome, referencing taxa with a broad broad labellum such as

Chamaecostus subsessilis (Nees & Mart.) C.D.Specht & D.W.Stev., Monocostus uniflorus

(Poepp. ex Petersen) Maas and Dimerocostus strobilaceus Kuntze (Figure 1).

!!

! 4!

Figure 1. From upper left to bottom right: Monocostus uniflorus K.Schum.; Costus L. stem

showing spiromonistichous ligulate leaves with a closed sheath; Chamaecostus lanceolatus

(Petersen) C.D.Specht & D.W.Stev.; Dimerocostus strobilaceus Kuntze; Costus ligularis

Baker; Tapeinochilos ananassae K. Schum.; Costus L. inflorescence and fruits.

!!

! 5!

Figure 2. From upper left to bottom right: Costus spiralis (Jacq.) Roscoe; Costus arabicus L.;

Chamaecostus lanceolatus (Petersen) C.D.Specht & D.W.Stev. flower, calyx, bract,

bracteole and petals; Costus pulverulentus C. Presl; Costus arabicus L. petaloid stamen and

pistil; Chamaecostus subsessilis (Nees & Mart.) C.D.Specht & D.W.Stev flower.

!!

! 6!

In a study on the pollination guilds and variation in floral characteristics of Bornean

gingers (Zingiberaceae and Costaceae), Sakai et al. (1999) verified a limited number of

pollination guilds and low pollinator species diversity, implying that those species of plants

may have evolved without segregating pollinators in each pollination guild. Kay & Schemske

(2003) systematically investigated pollinator relationships in Neotropical Costus L. by

studying 11 species from different sites in Bolivia, Costa Rica and Panama. They

documented visitation rates and pollinator assemblages among a variety of habitats, and

uncovered minimal spatial and temporal variation in visitation rates and pollinator identities.

Pollinator specificity was found to contribute strongly to reproductive isolation for sympatric

species differing in pollination syndrome. Araújo & Oliveira (2007) studied the floral biology

of Costus spiralis (Jacq.) Roscoe and its mechanisms to avoid self-pollination in a gallery

forest of Uberlândia (MG), Brazil. They found that this species is self-compatible but does

not present spontaneous self-pollination, exhibiting movement herkogamy. Verified

pollinators were Phaethornithinae hummingbirds. They found no difference between

germination rates of seeds from self-pollination and cross-pollination, but the seeds

produced from natural fruit-set presented significantly higher germination rates than those

from hand pollination treatments, endorsing the effectiveness and importance of

hummingbirds as pollen vectors for C. spiralis.

On interspecific reproductive barriers, Kathleen Kay’s research group has repetedly

demonstrated (Kay 2006, Kay & Schemske 2008, Yost & Kay 2009, Surget-Groba & Kay

2013) effective reproductive isolation between two closely related hummingbird-pollinated

Costus species: C. pulverulentus C.Presl and C. scaber Ruiz & Pav. Kay (2006) found

evidence of significant prezygotic and evidence of postzygotic isolation, suggesting that

ecological factors, including habitat use and plant-pollinator interactions, may contribute to

speciation in these plants. Kay & Schemske (2008) verified a lower seed set due to pollen-

pistil incompatibility between species pairs that co-occur and experience pollen transfer in

nature compared to species pairs that are otherwise isolated, regardless of genetic distance.

Yost & Kay (2009) review empirical studies of postpollination reproductive isolation in

!!

! 7!

Costus, summarizing isolation due to parental style length differences. They also show that

reduced pollen adhesion, germination, and pollen tube growth contribute to reproductive

isolation between closely related sympatric species, with further geographic separation of

species correlating with greater variation in the strength of these pre-zygotic barriers. More

recently, Surget-Groba & Kay (2013) genotyped microsatellite markers from locations along

the geographical range of both C. scaber and C. pulverulentus, including sympatric sites,

and found high levels of genetic isolation among populations within each species and low

but detectable levels of introgression between species at sympatric sites. These studies

imply that crossing barriers prevent potential hybridization and effectively reinforce the

speciation process.

Specht et al. (2001) presented the first detailed investigation into intrageneric and

interspecific evolutionary relationships within the family Costaceae, based on molecular

markers. Some novel evolutionary trends with respect to floral morphology were proposed

based on that phylogenetic analysis. Phylogenetic results clearly demonstrate that the two

pollination syndromes characterized for Neotropical Costus, ornithophilus and melittophilus

(Maas 1972), display a significant amount of homoplasy, suggesting that transitions between

the two floral forms has been a common event throughout the evolutionary history of this

lineage, and potentially characterize responses to ecological or environmental factors such

as pollinator availability or efficacy. Additionally, the genus Costus was found to be

polyphyletic with respect to genera Tapeinochilos Miq., Dimerocostus and Monocostus.

Later, Specht (2006a) provided a more detailed phylogeny of the family including more

taxonomic units and additional molecular markers. Kay et al. (2005) also estimated

phylogenetic relationships within this family, with sampling focused on reconstructing

biogeographic and pollination history of Costus subgenus Costus Loes (see Maas 1972).

Specht & Stevenson (2006) proposed a generic revision based on these phylogenetic

results. They divided Costus into four genera which largely reflected previous subgeneric

divisions of Costus; Costus L. sensu stricto, Cheilocostus C.D.Specht, Chamaecostus

C.D.Specht & D.W.Stev., and Paracostus C.D.Specht. Cheilocostus was recently shown to

!!

! 8!

be a replacement name for Banksea J.Koenig, which in turn had already been replaced in

1791 by Retzius as Hellenia Retzius (Govaerts 2013). Consequently, species placed in

Cheilocostus are now included in Hellenia. Hellenia is sister to Tapeinochilos Miq., forming

together an Asian-Australian clade. The South American Chamaecostus genus comprises

species previously placed in Costus subgen. Cadalvena (Fenzl) Schumann. These species

are a group of low or diminutive plants, occasionally acaulescent rosettes that do not usually

reach 1m in height. This clade was shown to be sister to the Neotropical Monocostus

K.Schum. + Dimerocostus Kuntze clade in Specht (2006a).

Specht (2006b) tested the hypothesis of an ancient Gondwanan distribution followed

by vicariance via continental drift as the leading cause of the current pantropical distribution

of Costaceae, using a strict clock molecular dating technique of cladogenic events combined

with phylogeny-based dispersal-vicariance biogeographic analyses. Results dated initial

Costaceae diversification to be at least ca. 65 million years, long after the final break up of

the Gondwanan supercontinent, implying Neotropical Costus to have arrived from a long

distance dispersal event of an African ancestral taxon. Kay et al. (2005) calibrated an ITS-

based molecular clock for Neotropical Costus and suggested that diversification of this clade

was recent and rapid, presenting a hypothetical scenario of rapid floral adaptation in

geographic isolation with range shifts in response to environmental changes.

Over 40 species of fossils have been identified as pertaining to the Zingiberales,

comprising a record that extends from the Cretaceous to Pliocene (Berry 1921a, 1921b,

1925, Reid & Chandler 1926, Hickey & Peterson 1978, Friis 1988, Boyd 1992, Manchester &

Kress 1993). However, lack of detailed data on extant and fossil morphology inhibits

confidently determining familial affiliations of many taxa (Smith et al. 2013). Moreover, pollen

- an important source of historical information - lacks an exine in Zingiberales (Kress et al.

1978), impeding efficient fossilization. Nevertheless, there are at least two potential

Costaceae fossils available (J.Benedict pers. comm.): a leaf fossil from the Eocene of Great

Britain (Reid & Chandler 1926) and a stem cast/mold from the Miocene of Trinidad (Berry

1925). The use of these fossils for calibrating phylogenies could help advance our

!!

! 9!

understanding of diversification patterns within Costaceae, although with appropriate caution

given uncertainty in taxonomic affinity (Drummond et al. 2006, Sanders & Lee 2007).

Spiral gingers can be found in many habitats, from Cerrado to Várzeas, and they are

an important element of the understory in Neotropical forests (Drucker et al. 2008,

Figueiredo 2008). Several important evolutionary processes in the family were previously

acknowledged by aformentioned research, such as the importance of prezygotic isolation in

sympatric speciation and the young age of Neotropical taxa indicating rapid diversification.

However, a comprehensive analysis of their biogeographic history using appropriate dating

techniques that take advantage of available fossil calibration within a relaxed molecular clock

approach will help address the spatiotemporal patterns of diversification, both between and

within species, during Costaceae evolution, which are accomplished throughout this thesis.

Main aims of this thesis are: to examine the geographic context of speciation in

Neotropical Costaceae lineages by using diversification analyses and biogeographical

reconstructions through a robust dated molecular phylogeny (CHAPTER 2); to investigate

shifts in pollination syndrome within the Costus clade, with the intention of addressing

potential mechanisms leading to the rapid radiation (CHAPTER 3); to explore the

phylogenetic relationships of species in the Chamaecostus lineage and interpret

evolutionary trends across the entire genus based on a molecular character-based

phylogenetic hypothesis that includes all currently described species of Chamaecostus

(CHAPTER 4); and to examine speciation of Chamaecostus populations from South

American seasonally dry forests of Cerrado and South Amazonia, by estimating

phylogeographic pattern and timing of diversification (CHAPTER 5). A final remarks text

further support major findings (CHAPTER 6). The thesis is presented in the form of scientific

manuscripts. Systematic Botany recently accepted the third chapter for publication, and the

version presented here already features suggestions from anonymous reviewers, and the

fourth chapter was already submitted to Phytoaxa.

!!

! 10!

References

Araújo FP, Oliveira PE. 2007. Biologia floral de Costus spiralis (Jacq.) Roscoe (Costaceae)

e mecanismos para evitar a autopolinização. Revista Brasileira de Botânica 30: 61-70.

Berry EW. 1921a. Tertiary plants from Venezuela. Proceedings of the US National Museum

59: 553-579.

Berry EW. 1921b. Tertiary plants from Costa Rica. Proceedings of the US National Museum

59: 169-185.

Berry EW. 1925. A banana in the Tertiary of Colombia. American Journal of Science. 10:

530-537.

Boyd A. 1992. Musopsis n. gen.: a banana-like leaf genus from the early Tertiary of eastern

North Greenland. American Journal of Botany 79: 1359-1367.

Drucker DP, Costa FRC, Magnusson WE. 2008. How wide is the riparian zone of small

streams in tropical forests? A test with terrestrial herbs. Journal of Tropical Ecology 24:

65-74.

Drummond AJ, Ho SY, Phillips MJ, Rambaut A. 2006. Relaxed phylogenetics and dating

with confidence. PLoS Biol 4:e88.

Figueiredo FOG. 2008. Variação florística e diversidade de Zingiberales em florestas da

Amazônia central e setentrional. Masters Dissertation. Programa de Pós-Graduação em

Biologia Tropical e Recursos Naturais da Amazônia, Manaus, Brazil.

Friis EM. 1988. Spirematospermum chandlerae sp. nov., an extinct species of Zingiberaceae

from the North American Cretaceous. Tertiary Research 9: 7-12.

Govaerts R. 2013. Hellenia Retz., the correct name for Cheilocostus C.D.Specht

(Costaceae). Phytotaxa 151: 63-64.

!!

! 11!

Grootjen CJ, Bouman F. 1981. Development of the ovule and seed in Costus cuspidatus

(Nees & Mart.) Maas (Zingiberaceae), with special reference to the formation of the

operculum. Botanical Journal of the Linnean Society 83: 27-39.

Hickey LJ & Peterson RK.1978. Zingiberopsis, a fossil genus of the ginger family from Late

Cretaceous to early Eocene sediments of western interior North America. Canadian

Journal of Botany 56: 1136-1152.

Kay KM, Schemske DW. 2003. Pollinator Assemblages and Visitation Rates for 11 Species

of Neotropical Costus (Costaceae). Biotropica 35: 198-207.

Kay KM, Schemske DW. 2008. Natural selection reinforces speciation in a radiation of

Neotropical rainforest plants. Evolution 62: 2628-2642.

Kay KM. 2006. Reproductive isolation between two closely related hummingbird-pollinated

Neotropical gingers. Evolution 60: 538-552.

Kay KM, Reeves PA, Olmstead RG, Schemske DW. 2005. Rapid speciation and the

evolution of hummingbird pollination in Neotropical Costus subgenus Costus

(Costaceae): evidence from nrDNA ITS and ETS sequences. American Journal of

Botany 92: 1899-1910.

Kirchoff BK. 1988. Inflorescence and flower development in Costus scaber (Costaceae).

Canadian Journal of Botany 66: 339-345.

Kress WJ, Stone DE, Sellers SC. 1978. Ultrastructure of exine-less pollen: Heliconia

(Heliconiaceae). American Journal of Botany 65: 1064-1076.

Larsen K, Lock JM, Maas H, Maas PJM. 1998. Zingiberaceae. 474-495 pp, In Kubitzki K

(Ed.), The Families and Genera of Vascular Plants. IV. Flowering Plants:

Monocotyledons. Alismatanae and Commelinanae (except Gramineae). Springer, Berlin.

!!

! 12!

Manchester SR & Kress WJ. 1993. Fossil Bananas (Musaceae): Ensete oregonense sp.

nov. from the Eocene of western North America and its phytogeographic significance.

American Journal of Botany 80: 1264-1272.

Newman SWH, Kirchoff BK. 1992. Ovary structure in the Costaceae (Zingiberales).

Internatational Journal of Plant Sciences 153: 471-487.

Panchaksharappa MG. 1963. Embryological studies in the family Zingiberaceae I. Costus

speciosus Smith. Phytomorphology 12: 418-430.

Reid EM & Chandler MEJ. 1926. Catalogue of Cainozoic plants vol. 1: the Bembridge Flora.

British Museum (Natural History): London.

Sakai S, Kato M, Inoue T. 1999. Three pollination guilds and variation in floral characteristics

of Bornean gingers (Zingiberaceae and Costaceae). American Journal of Botany 86(5):

646–658. 1999.

Sanders K, Lee MSY. 2007. Evaluating molecular clock calibrations using Bayesian

analyses with soft and hard bounds. Biol. Lett. 3: 275-279.

Schemske DW. 1983. Breeding system and habitat effects on fitness components in three

Neotropical Costus (Zingiberaceae). Evolution 37: 523-539.

Smith S, Benedict J, Specht C, Collinson M, Leong-Äkornicková J, Kvacek J, Xiao X, Fife J,

Marone F. 2013. Reevaluation of the oldest fossils in Zingiberales and implications for

inferring the evolutionary history of gingers, bananas, and relatives. pp. 43-44. In Botany

2013. Celebrating Diversity! July 27-31 - New Orleans. Abstracts.

Specht CD. 2006a. Systematics and evolution of the tropical monocot family Costaceae

(Zingiberales): A multiple dataset approach. Systematic Botany 31: 89-106.

Specht CD. 2006b. Gondwanan vicariance or dispersal in the tropics? The biogeographic

history of the tropical monocot family Costaceae (Zingiberales). Aliso 22: 633-644.

!!

! 13!

Specht CD, Stevenson DW. 2006. A new phylogeny-based generic classification of

Costaceae (Zingiberales). Taxon 55: 153-163.

Specht CD, Kress WJ, Stevenson DW, DeSalle R. 2001. A molecular phylogeny of

Costaceae (Zingiberales). Molecular Phylogenetics and Evolution 21: 333-345.

Stiles FG. 1978. Temporal Organization of Flowering Among the Hummingbird Foodplants of

a Tropical Wet Forest. Biotropica 10: 194-210.

Stiles FG. 1981. Geographical aspects of bird-flower coevolution, with particular reference to

Central America. Annals of the Missouri Botanical Garden 68: 323-351.

Surget-Groba Y, Kay KM. 2013. Restricted gene flow within and between rapidly diverging

Neotropical plant species. Molecular Ecology 22: 4931-4942.

Tomlinson PB. 1969. In Metcalfe CR (Ed.), Anatomy of Monocotyledons. III. Commelinales-

Zingiberales. Clarendon Press, Oxford.

Yost JM, Kay KM. 2009. The evolution of postpollination reproductive isolation in Costus.

Sex Plant Reprod. 22: 247-255.

!!

! 14!

!!

! 15!

Chapter 2 – Spatiotemporal dynamics of speciation in Neotropical Costaceae

(Zingiberales)

Thiago Andréa,b; Shayla Salzmanc,1; Tânia Wendta; Chelsea Spechtc

a - Departamento de Botânica, Universidade Federal do Rio de Janeiro, Rio de Janeiro (RJ),

Brazil.

b - [email protected]

c - Departments of Plant and Microbial Biology and Integrative Biology, University of

California at Berkeley, Berkeley (CA), USA.

1 - present address: Department of Organismic and Evolutionary Biology, Harvard

University, Cambridge (MA), USA.

!!

! 16!

Abstract

Allopatric forces as well as ecological and reproductive character displacement in sympatry

can lead to new species formation. In the Neotropics, two Costaceae lineages can be found;

a small, early-divergent clade endemic to South America (consisting of Monocostus,

Dimerocostus and Chamaecostus); and the Neotropical Costus clade (ca. 50 species), a

diverse assemblage of understory herbs, comprising nearly half of total familial species

richness. We use a robust dated molecular phylogeny containing ca. 90% of current species

to inform diversification analyses and biogeographical reconstructions, enabling us to

examine the geographic context of speciation in Neotropical Costaceae lineages. Analyses

of speciation rate revealed a significant rate variation among clades, with a rate shift at the

most recent common ancestor of the Neotropical Costus clade. There is an overall

predominance of allopatric speciation in the older-diverging lineages of the South American

clade, as most extant species display little overlap in geographic ranges. In contrast,

sympatry is much higher within the younger Neotropical Costus clade, independent of node

age. Our results show that speciation dynamics during the history of Costaceae is strongly

heterogeneous, and we propose that the Costus radiation in the Neotropics arose mainly by

sympatric processes.

!!

! 17!

2.1. Introduction

Global biodiversity flourishes in the Neotropical region as the result of an extensive

and complex history of evolutionary trends, mediated by ecological processes (Ricklefs

2004, Rull 2011). Despite the extent and importance of the high species diversity found in

the Neotropics, we are still in the early stages of our understanding of its origins (Antonelli &

Sanmartín 2011, Turchetto-Zolet et al. 2013). A critical component of this task is the

examination of the environment-dependent evolutionary processes by which new biological

species arise, i.e. speciation. Although the geography of speciation has long been debated

(Barraclough & Nee 2001, Coyne & Orr 2004, Fitzpatrick et al. 2009), the phylogenetic

approaches of the last few decades have greatly advanced the study of speciation (Nee et

al. 1994, 1998, Ricklefs 2007, Silvestro et al. 2011, Morlon et al. 2011, Paradis 2011,

Rabosky 2014) and its relationship with geographic diversification (Barraclough & Vogler

2000, Graham et al. 2004, Goldberg et al. 2011). Moreover, speciation leaves discernible

signatures in molecular data that can be used to build phylogenetic trees of extant taxa, and

allows for the assessment of speciation rates from phylogenetic data (Rabosky 2009, 2010,

Cusimano & Renner 2010).

Neotropical ecosystems have a strong historical dynamic (Hoorn et al. 2010), with

intense fluctuation both in overall magnitude and in plant diversity (Jaramillo et al. 2010,

Antonelli & Sanmartin 2011). Numerous mechanisms of speciation have been proposed to

explain high species diversity in tropical systems (Mortiz et al. 2000, Antonelli & Sanmartin

2011), most of which assume allopatric speciation as the predominant geographic mode of

speciation. Undeniably, Oligocene and early Miocene were key periods for the development

of modern Neotropical diversity (Antonelli et al. 2009, Hoorn et al. 2011) due to geographic

alterations, especially Andean uplift and the consequent alterations of Amazonian River

drainage patterns (Räsänen et al. 1990, 1992, Hoorn et al. 1995, Gregory-Wodzicki 2000);

ocean level rise and incursions into the South American continent (Flynn & Wyss 1998,

Wesselingh & Salo 2006); and Central America geomorphological dynamics, such as

!!

! 18!

volcanic activity that resulted in oceanic ground shallowing and the closing of the Isthmus of

Panama (Montes et al. 2012, Bacon et al. 2013). In particular, ecological opportunities

provided by the different environments and microclimates available after the uplift of the

Andes are believed to have resulted in the current high South American diversity (Gentry

1982, Luteyn 2002, Rauscher 2002, Bell & Donoghue 2005, Ortiz-Jaureguizar & Cladera

2006, Hughes & Eastwood 2006, Antonelli et al. 2009, Wagner et al. 2013). Compared to

this intense geographic and topographic dynamic during the Tertiary, significant

environmental changes during the Quaternary have been more related to glacial cycles and

their impact on species range continuousness, which could potentially lead to allopatric

speciation (Haffer 2008). Nonetheless, estimates show that nearly half of the dated extant

Neotropical species originated before the Pleistocene (Rull 2008). To some extent, a

proportion of the Neotropical rainforests have also been shown to be ecologically stable over

long periods of time (Carnaval et al. 2009, Couvreur et al. 2011).

Speciation often results from the evolution of reproductive isolation mechanisms, as

once incipient species may not remain distinct entities in their absence (Coyne & Orr 2004).

An alternative to allopatric forces, ecological and reproductive character displacement in

sympatry can also lead to speciation (Schluter 2000, Wendt et al. 2002, Lovette et al. 2002,

Pfennig & Servedio 2013, Rabosky et al. 2014), and several phenotypic novelties have been

associated with species radiations in plants (e.g. Toon et al. 2014, Werner et al. 2014,

Bouchenak-Khelladi et al. 2014).

Adaptive divergence as a response to ecological factors, such as pollinators and

habitat, commonly drives the evolution of prezygotic barriers in plants (Rieseberg & Willis

2007). In particular, the interactions between plants and pollinators can drive adaptive

divergence in floral traits and contribute to the maintenance of reproductive isolation among

closely related sympatric species (Grant 1981, Kay 2006, Wendt et al. 2008). Pollinator

preference can be a strong selection factor for reinforcement between sympatric closely

related species (Hopkins & Rausher 2012). Furthermore, speciation may also occur by

lineage reticulation, instead of divergence (Stebbins 1959, Arnold 1992). Hybridization has

!!

! 19!

been considered a key phenomenon in plant evolution because it results in large amounts of

genetic recombination and may enable the establishment of new traits and evolutionary

lineages (Grant 1981, Ehrlich & Wilson 1991, Soltis & Soltis 1999). Hybridization and

introgression in isolated populations could actually stimulate speciation by providing standing

genetic variation that can contribute to ecological speciation (Seehausen 2004, Palma-Silva

et al. 2011). Moreover, flowering plants maintain genomes with considerable gene

redundancy, much of which is likely the result of allopolyploidy or whole genome duplication

(Soltis et al. 2003).

Spiral Gingers (Costaceae Nakai) comprises ca. 125 pantropically distributed

species, with its center of species diversity located in South and Central America (ca. 95

neotropical species, ca. 15 African, ca. 23 Southeast Asian). Few species have a broad

geographic range, with most species locally restricted geographically and by habitat

heterogeneity (Maas 1972). Two Costaceae lineages inhabit the Neotropics: a small, early

divergent clade primarily endemic to South America consisting of Monocostus K.Schum. (1

species), Dimerocostus Kuntze (3 species) and Chamaecostus C.D.Specht & D.W.Stev. (7

species); and the Neotropical Costus L. clade (ca. 50 species), which encompasses nearly

half of the family’s total species richness. Phylogenetic studies indicate that Neotropical

Costus diverged from African Costus following a long distance dispersal event that occurred

long after the final break up of the Gondwanan supercontinent (Specht 2006a). Hence, the

current distribution of the genus Costus and perhaps the pantropical distribution of the entire

family is better explained by a series of more recent local vicariance and dispersal events

not related to major vicariant events.

Hummingbird pollination (ornithophily) has evolved several times in Neotropical

Costus from bee pollinated (melittophily) ancestors (Kay et al. 2005, Salzman et al. in press),

and flower morphology reflects pollination type mainly by diversity of coloration patterns

marking the showy labellum (an organ comprised of five fused staminodes, Specht et al.

2012), and in overall flower aperture size (Kay et al. 2005, Specht 2006a). This pollination

shift has been suggested to be particularly important in diversification of Neotropical Costus

!!

! 20!

(Kay & Schemske 2003, Kay et al. 2005, Specht 2006b, Salzman et al. in press) but has not

been investigated for the South American clade, i.e. Dimerocostus, Monocostus and

Chamaecostus. Kay et al. (2005) proposed a scenario for the diversification of Costus, in

which range shifts in response to environmental changes and rapid floral adaptation in

geographic isolation could have promoted reproductive isolation among closely related

species. Interestingly, Costaceae’s South American clade is species-poor as compared to

the Neotropical Costus clade, regardless of being older and having potentially experienced

the same environmental changes. Comprehensive and current chronological and

biogeographic approaches for the analysis of Costaceae diversification have not yet been

fully considered, making the Costaceae an appropriate system for studying spatiotemporal

diversification dynamics in the megadiverse Neotropical region.

Here, we examine the spatiotemporal context of speciation through a phylogenetic

framework in two Neotropical Costaceae lineages (Costus and the South American clade)

with significantly different extant species richness.

!!

! 21!

2.2. Materials and Methods

2.2.1. Taxonomic sampling, DNA extraction, amplification and sequencing

A total of 84 ingroup taxa were sampled, including nearly all known species from the

South American clade, most of Neotropical and African Costus species, and a few

representatives of Asian Costaceae. Alpinia zerumbet (Pers.) B.L. Burtt & R.M. Sm. and

Zingiber officinale Roscoe (Zingiberaceae) were designated as outgroups. Available

sequences were downloaded from GenBank (accession numbers in Appendix 1). For the

remaining taxa, total genomic DNA was isolated from silica-gel dried leaf tissue using CTAB

extraction protocol (Doyle & Doyle 1990). Regions of the chloroplast (trnL-trnLF, rps16-trnk)

and nuclear (rpb2, ETS and ITS) genomes were amplified and sequenced using published

primers (Taberlet et al. 1991, Shaw et al. 2007, Specht et al. 2001, Kay et al. 2005, White et

al. 1990, respectively). Novel primers for the calmoduline (CAM) 23rd intron were designed

following a long PCR protocol using Zingiberales primers cam33F and cam328R (Johansen

2005, Salzman et al. in press). PCR fragments were generated and sequenced on an

Applied Biosystems® 3730 DNA Analyzer automated DNA sequencer, at UC Berkeley’s

Museum of Vertebrate Zoology’s Evolutionary Genetics Laboratory. Nucleotide sequences of

the 2 plastid and 4 nuclear genetic markers were concatenated and analyzed under a

Bayesian phylogenetic framework, using BEAST 1.7.4 (Drummond et al. 2012).

Chromatogram files were examined for biases and possible errors using Geneious version

5.6.3 (Biomatters Ltd.). Alignments were made for each marker using MUSCLE algorithm

(Edgar 2004) implemented in Geneious, and alignments were subsequently checked

manually. Alignment regions that could not be unambiguously interpreted were excluded

from analysis. New sequences were deposited in GenBank (accession numbers in Appendix

1).

2.2.2. Phylogenetic inferences and speciation dynamics

!!

! 22!

Nucleotide sequences of the 2 plastid and 4 nuclear genetic markers were

concatenated and analyzed under a Bayesian phylogenetic framework, using BEAST 1.7.4

(Drummond & Rambaut 2007). Sequence data was partitioned to allow different models of

sequence evolution for each region, based on likelihood analyses ran on jModelTest version

2.0 (Darriba et al. 2012) and selected with Bayesian information criterion. A relaxed clock

with an uncorrelated lognormal model of rate variation was used and a Yule speciation

process for branching rates was selected. Two fossil-based time to the most recent common

ancestor (tmrca) calibrations were used; 45 ± 5 Ma for the Costaceae root (Costus incertis!;

Berry 1925) and 85 ± 5 Ma for Alpinia + Zingiber clade (Zingiberopsis!; Hickey & Peterson

1978), and a CTMC rate prior was selected (Ferreira & Suchard 2008). No monophyletic

prior assignment was made. Markov chain Monte Carlo simulations were run twice

independently for 5x107 generations and sampled every 5x103. These analyses were

performed on the CIPRES Science Gateway (Miller et al. 2010). We assessed convergence

of model parameters across the independent runs by analyzing plots of the marginal

posterior distributions in Tracer version 1.5 (Rambaut & Drummond 2009), and by ensuring

high effective sample size values (ESS ≥ 200). Tracer was also used to assess burn-in

levels. A maximum clade-credibility tree was obtained from the posterior sample of trees

using TreeAnnotator version 1.7.4 (Drummond et al. 2012), and visualized on FigTree

(http://tree.bio.ed.ac.uk/software/figtree/).

The Bayesian Analysis of Macroevolutionary Mixtures (BAMM; Rabosky 2014) was

used for the analysis of speciation rates on the ultrametric phylogeny, accounting for non-

random taxon sampling. This model assumes that changes in evolutionary regimes occur

across the branches of phylogenetic trees under a compound Poisson process model of rate

variation and explicitly allows rates to vary both through time and among lineages. Four

MCMC chains were run for 5x107 generations and sampled every 5x103, using the BAMM

default parameters.

2.2.3. Geography of Speciation

!!

! 23!

Locality data for Neotropical Costaceae was downloaded from the Global Biodiversity

Information Facility Data Portal (www.gbif.org, March 2014), and from examination of

herbaria material (A; HAMAB; HEPH; IAN; IBGE; INPA; MG; MO; NY; R; RB; UB; UC;

UFACPZ; UFG, acronyms follows Thiers 2014). Reconstruction of the geographical

distribution of species at the time of speciation is obviously problematic, as the current

distribution of a species is not necessarily a direct indicator of species’ historical

geographical range (Losos & Glor 2003). This issue was addressed by inferring the potential

geographic distribution of species, estimated by ecological niche modeling (Austin 1985,

Graham et al. 2004, Phillips et al. 2006), instead of using geographic range maps based on

locality data. Values were extracted based on specimens’ localities from 43 environmental

variables describing climate (Hijmans et al. 2005, Kriticos et al. 2012), soil (Global Soil Data

Task Group 2000) and net primary productivity (Imhoff & Bounoua 2006). After evaluating

the correlation between each of the variables in R (R Core Team 2014), we removed

variables that had a coefficient -0.75 ≤ r ≥ 0.75, which resulted in 14 correlated variables

describing environmental variation. Extant species potential geographic distributions were

modeled by maximum entropy in MAXENT version 3.3.3k (Phillips et al. 2006), with 25% of

the data used for training. Thresholds for geographic ranges were estimated from median

presence of 50 bootstrap replicates, in QGIS (Quantum GIS Development Team,

www.qgis.org). We then calculated the degree of sympatry between sister clades (Chesser

& Zink 1994: area of overlap in geographic range / range size of clade with smaller range)

for each node of the South American Clade and the Neotropical Costus clade.

Additionally, evolution of the geographic range of Costaceae was interpreted using

historic biogeographic reconstruction. Ancestral ranges were inferred using the

BioGeoBEARS package implemented in R (Matzke 2013), which allows for both probabilistic

inferences of models of range expansion and founder-event speciation (J). An unconstrained

dispersal-extinction-cladogenesis analysis (DEC and DEC+J; Ree et al. 2005, Ree & Smith

2008, Ree & Sanmartín 2009) and the Bayesian analysis for large number of areas

(BAYAREA and BAYAREA+J; Landis et al. 2013), were run using the maximum clade

!!

! 24!

credibility time-calibrated species tree. The Likelihood Ratio Test (LRT) and the Akaike

Information Criterion (AIC; Burnham & Anderson 2002) were used to objectively compare

models (Table 1). Geographic distributions of the terminal taxa were coded based on

environmental discontinuity of Neotropical regions and on extant Costaceae species

distributions: (1) Central America and the Caribbean, (2) Amazonian, (3) Andean, (4) Central

Brazilian Plateau and Atlantic Rainforests, (5) Africa, (6) Asia and Oceania. We set the

maximum number of areas equal to four.

!!

! 25!

2.3. Results

We present a multilocus phylogenetic estimate for Costaceae (Figure 1), along with

corresponding branch age estimates, based on a rich taxon and loci sampling. The family

Costaceae first diversified around 50 million years ago, in the mid-Paleogene period at the

Eocene, and the large Costus genus originated ca. 30 million years ago. The age of the

most-recent common ancestor of Neotropical Costus is estimated to have occurred only ca.

10 million years ago. Furthermore, phylogenetic relations inferred here reaffirm the

monophyly of major lineages within Costaceae: i.e. the South American clade, Costus, and

the Neotropical Costus clade.

The comprehensive phylogeny of Figure 1 clearly shows the unbalanced distribution

of extant species diversity, with most species belonging to the Neotropical Costus clade. The

Bayesian Analysis of Macroevolutionary Mixtures (ESS LogLik = 816.295, ESS N Shifts =

901) favors a configuration with a single shift (posterior distribution = 0.6400).

!!

! 26!

Figure 1. Maximum clade credibility tree estimated from 1,000 trees. Numbers above

branches refer to posterior probabilities, and numbers at right of nodes are age estimates in

million years. Blue bars denote node height probability density at 95%. Lower scale in million

years before present.

!!

! 27!

Biogeographical models including founder-speciation (+J) are favored and

BAYAREALIKE had greater likelihood compared to DEC models (BAYAREALIKE LnL = -

158.034 with +J, and -180.826 without +J; DEC LnL = −165.89 with +J, and −174.02 without

+J). Uncertainty in estimated ancestral range scenarios are indicated with pie charts, but we

particularly discuss Costaceae’s biogeography based on the single scenario with highest

likelihood (Figure 2). Ancestral geographic state reconstructions in the context of our

comprehensive phylogeny indicates that the South American clade originated and diversified

in the Amazon and the Andes (Figure 2A). Neotropical Costus started to diversify (significant

species accumulation indicated by the grey circle, Figure 2B) in Central America from a long

distance dispersal event from Africa. A comparison of speciation rates (Figure 3) suggests

four major rate patterns within Costaceae phylogeny; the South American clade rate, the

Neotropical Costus clade rate, and two separate rate patterns within African Costus.

!!

! 28!

(A) (B)

Figure 2. Top: (A) Biogeographical analysis of Costaceae using BioGeoBEARS. (B)

Phylorate plots for speciation rate using BAMM; colors at each point in time along branches

denote instantaneous speciation rates, with warmer colors referring to faster rates; two

distinct shift configurations account for most of the posterior probability of the data with the

dark dot indicating the node of the single shift in configuration. Bottom: Geographic areas

included in biogeographic analysis: ! Central America and the Caribbean, ! Amazonian, !

Andean, ! Central Brazilian Plateau and Atlantic Rainforests, ! Africa, ! Asia and Oceania.

Outgroups are not shown. Central images show representative Costaceae (top-down):

Costus scaber Ruiz & Pav., Costus arabicus L., Costus ligularis Baker, Chamaecostus

!!

! 29!

lanceolatus (Petersen) C.D. Specht & D.W. Stev.; C. ligularis photograph by C.Specht, all

others by T.André.

!!

! 30!

Figure 3. Speciation rates comparison within Costaceae phylogeny. Color bar and squares

indicate proportion of similarity between rates. Colors at each point in time along branches of

the phylogeny denote instantaneous speciation rates. Outgroup rate is underestimated due

to limited sampling of outgroup clades.

0

0.5

1

!!

! 31!

Degree of sympatry plots (Figure 4, A and B) point to the overall predominance of

allopatric speciation in the South American clade (Figure 4A), where most recently diverged

species display little overlap in geographic ranges. In contrast, sympatry is much higher

within the Neotropical Costus clade, independent of node age (Figure 4B).

!!

! 32!

(A)

!!

! 33!

(B)

Figure 4. Top: Degree of sympatry by node age for (A) the South American clade and (B) the

Neotropical Costus clade. Bottom: Maps showing examples of recovered potential current

distributions superimposed for one selected clade from Chamaecostus (A) and Costus (B).

!!

! 34!

2.4. Discussion

As previously inferred (Specht et al. 2001, Specht 2006b), an early lineage

divergence event splits Costaceae into two major clades: a South American clade

[(Monocostus+Dimerocostus)+Chamaecostus], and a large clade containing the remaining

species of Costaceae. There are important differences between species of the early-

diverging South American clade and the Neotropical Costus clade regarding their

geographical distribution. Nearly half of the species richness of Neotropical Costus is

endemic to Central America, with several predominantly South American species having

large ranges that also include Central America. On the other hand, the distributions of

Chamaecostus, Dimerocostus and Monocostus are almost entirely exclusive to South

America with only a single species of Dimerocostus having a large range expanding into

Central America.

The genus Costus maintains an amphi-Atlantic distribution, with ca. 29 known

species in Africa and ca. 51 in the Neotropics. This rare biogeographic condition (Renner

2004) is shared, for example, with Renealmia L.f. (Zingiberaceae), which has ca. 15 species

in Africa and 61 in the Neotropics (Maas 1977). Särkinen et al. (2007) suggested that

Renealmia acquired this distribution by an oceanic long-distance dispersal event from Africa

to South America during the Miocene or Pliocene, and that speciation in the Neotropics

might have been affected by the orogeny of the Andes. Notably, Specht (2006a) and

Salzman et al. (in press) previously suggested an African origin for Neotropical Costus.

Here, we also demonstrate a long distance dispersal from Africa to Central America ca. 10

million years ago. We also infer subsequent successive migration events from Central to

South America, as well as some apparently recent range expansions by some species, such

as Costus scaber Ruiz & Pav., that still maintain most of their distribution in South America.

Long-distance dispersal of plants, such as transatlantic ones, are challenging to research

because they involves rare events determined by complex and highly stochastic processes.

In fact, extreme climatic events and generalized long distance dispersal vectors are more

!!

! 35!

likely to explain drastic deviations from a mean trend in dispersal, and therefore are likely to

turn exceptionally rare dispersal events into reality (Nathan 2006), a suggestion first made

my Darwin in his Origin of Species.

Disjunct transatlantic patterns have occasionally been ascribed to vicariance

involving the break up of the Gondwanan supercontinent or continental rafting (Raven &

Axelrod 1974), but many tropical plant groups originated more recently (Givnish & Renner

2004, Renner 2004). Previous biogeographic and dating analysis of Costaceae used a strict

molecular clock approach and the calibration points for converting relative to absolute ages

were either based on the divergence time between Costaceae and Zingiberaceae as

estimated in an ordinal analysis of Zingiberales (Specht 2006a, Kress & Specht 2006), or on

an approximated date of the closing of the isthmus of Panama (assumed to be 3.5 Ma and

coincident to Costus arrival in the Neotropics, by Kay et al. 2005). The Specht (2006a)

analysis suggested an initial diversification within Costaceae of approximately 65 million

years ago, long after the final break up of the Gondwanan supercontinent (McLoughlin 2001,

Metcalf 1991) or the existence of a boreotropical dispersal route (ca. 50-40 My ago; Lavin &

Luckow 1993, Morley 2003). Using the closing of the Panama isthmus as a calibration point,

Kay et al. (2005) found Neotropical Costus to be 4.6 Ma old. Additionally, using ITS

substitution rates from across herbaceous taxa, Kay et al. (2005) estimated a date range for

the divergence of subgenus Costus from the rest of the genus of ca. 1.5-7.1 Ma. By using a

relaxed molecular clock we inferred a younger age for the initial Costaceae diversification as

proposed by Specht (2006a), and an older age of Neotropical Costus origin as compared to

Kay et al. (2005). Dates recovered in our analysis point to a recent oceanic long-distance

dispersal (Lavin et al. 2004, Renner 2004) as explicatory of Neotropical Costus origin.

Ancestral biogeographic state reconstruction strongly supports Central America as

ancestral within the Neotropical Costus clade, suggesting that early diverging lineages

evolved mostly in a dynamic and fragmented area. This clade has a remarkably young age

given its large extant species diversity. Its origin is coincident with an abrupt decrease in

global temperatures following the growth in Antarctic ice (Zachos et al. 2001, Katz et al.

!!

! 36!

2008, Jaramillo & Cardenas 2013), and with tectonic activities in West Africa (Diester-Haass

& Chamley 1980, Pearson et al. 2008), as well as the intense geomorphological dynamics in

Central and South America (Hoorn et al. 1995, Montes et al. 2012).

Despite this biogeographic background, our results indicate that ecological and

reproductive evolution in sympatry likely played more important roles in the speciation

dynamics of this lineage. In fact, pollinator dissimilarity is generally sufficient to maintain pre-

zygotic isolation between sympatric species (Sakai et al. 1999, Kay & Schemske 2003),

such that pollination shifts may be sufficient to successfully act as reproductive barriers

between incipient species. Salzman et al. (in press.) showed a relative increase in

diversification for lineages within Costus that display specialized pollination morphologies as

compared to the preseumed generalist pollination morphology. Nevertheless, well-developed

pollination differentiation suggested by the floral morphology, and overall reproductive

biology, remains largely unverified for most species in the family (although see Kay &

Schemske 2003). Moreover, polyploidy may result in the evolution of speciose lineages,

since polyploids often exhibit ecological differentiation, high fecundity, perennial life history,

and self-fertilization or asexual reproduction (Rieseberg & Willis 2007).

Timing of the significant increase in diversification rate, coincident with the origin of

the Neotropical Costus clade, further uncovers a scenario of in situ divergence, including

multiple independent invasions to South America presumably occurring overland by the

Panamanian land bridge, which indeed was likely already formed (Montes et al. 2012, Bacon

et al. 2013). The Late Paleocene of South America represents an important time for plant

evolution as it probably witnessed the first paleoflora attributable to modern day tropical

rainforests (Jaramillo et al. 2006, Wing 2009). However, further ecological studies

investigating tolerance limits to relevant environmental gradients across Costus, such as

edaphic, climatic and topographic variables, will help better define evolution of particular

ecological characters involved in spatial occurrence.

Sympatric speciation significantly contributed to diversification in Costaceae, in

addition to orogenic-driven allopatric processes. Based on the high importance of sympatric

!!

! 37!

speciation for the diversification in the Neotropical Costus clade, its subclades could display

some intrinsic characteristic that facilitates in situ diversification, such as susceptibility to the

evolution of intrinsic barriers to reproduction, genes and genome duplications, or unusually

high amounts of structured genetic diversity. Nonetheless, the founder effect of colonization

of this key geographic region certainly presented new ecological niches for the Costus

lineage, and interactions with a different assembly of pollinators could have been an

important trigger of speciation rates.

!!

! 38!

2.5. Acknowledgements

T.A. received a scholarship grant from Coordenação de Aperfeiçoamento de Pessoal de

Nível Superior. T.W. received a productivity grant from Conselho Nacional de

Desenvolvimento Científico e Tecnológico. C.S. received funding from the National

Geographic Society #8994-11, NSF DEB0866601, and UC Berkeley CNR-BSP, SPUR, and

URAP. We are thankful to Heather Driscoll, Tanya Renner, Irene Liao, Stacey Shen and

Roxana Yockteng for help in selecting molecular markers and generating sequences.

2.6. References

Antonelli A, Sanmartín I. 2011. Why are there so many plant species in the Neotropics?

Taxon 60: 403-414.

Antonelli A, Nylander JAA, Persson C, Sanmartín I. 2009. Tracing the impact of the Andean

uplift on Neotropical plant evolution. Proceeding of the National Academy of Sciences

USA 106: 9749-9754.

Arnold SJ. 1992. Constrains on phenotypic evolution. American Naturalist 140: S85-S107.

Austin MP. 1985. Continuum concept, ordination methods and niche theory. Annual Review

of Ecology and Systematics 16: 39-61.

Bacon CD, Mora A, Wagner WL, Jaramillo CA. 2013. Testing geological models of evolution

of the Isthmus of Panama in a phylogenetic framework. Botanical Journal of the Linnean

Society 171: 287-300.

Barraclough TG, Nee S. 2001. Phylogenetics and speciation. Trends in Ecology and

Evolution 16: 391-399.

Barraclough TG, Vogler AP. 2000. Detecting the geographical pattern of speciation from

species-level phylogenies. American Naturalist 155: 419-434.

Bell CD, Donoghue MJ. 2005. Phylogeny and biogeography of the Valerianaceae

(Dipsacales) with special reference to the South American valerians. Organisms,

Evolution, and Diversity 5: 147-159.

!!

! 39!


530-537.

Bouchenak-Khelladi Y, Muasya AM, Linder HP. 2014. A revised evolutionary history of

Poales: origins and diversification. Botanical Journal of the Linnean Society 175: 4-16.

Burnham KP, Anderson DR. 2002. Model Selection and Multimodel Inference: A Practical

Information-Theoretic Approach. 2nd ed. Springer. 488 p.

Carnaval AC, Hickerson MJ, Haddad CFB, Rodrigues MT, Moritz C. 2009. Stability Predicts

Genetic Diversity in the Brazilian Atlantic Forest Hotspot. Science 323: 785-789.

Chesser RT, Zink RM. 1994. Modes of speciation in birds: An application of Lynch's (1989)

method. Evolution 48: 490-497.

Couvreur TLP, Forest F, Baker WJ. 2011. Origin and global diversification patterns of

tropical rain forests: inferences from a complete genus-level phylogeny of palms. BMC

Biology 9: 44.

Coyne JA, Orr HA. 2004. Speciation. Sinauer Associates Inc. 545 p.

Cusimano N, Renner SS. 2010. Slowdowns in diversification rates from real phylogenies

may not be real. Systematic Biology 59: 458-464.

Darriba D., G. L. Taboada, R. Doallo, and D. Posada. 2012. jModelTest 2: More models,

new heuristics and parallel computing. Nature Methods 9: 772.

Diester-Haass L, Chamley H. 1980. Oligocene climatic, tectonic and eustatic history of NW

Africa (DSDP Leg 41, Site 369). Oceanologica Acta 3: 115-126.

Doyle JJ, Doyle JL. 1990. A rapid total DNA preparation procedure for fresh plant tissue.

Focus 12:13-15.

Drummond AJ, Rambaut A. 2007. BEAST: Bayesian evolutionary analysis by sampling

trees. BMC Evolutionary Biology 7: 214.

Drummond AJ, Suchard MA, Xie D, Rambaut A. 20120. Bayesian phylogenetics with BEAUti

and the BEAST 1.7. Molecular Biology and Evolution 29: 1969-1973.

!!

! 40!

Edgar RC. 2004. MUSCLE: multiple sequence alignment with high accuracy and high

throughput. Nucleic Acids Research 32: 1792-1797.

Ehrlich PR, Wilson EO. 1991. Biodiversity studies: science and policy. Science 253: 758-

762.

Ferreira MAR, Suchard MA. 2008. Bayesian analysis of elapsed times in continuous-time

Markov chains. Canadian Journal of Statistics 26: 355-368.

Fitzpatrick BM, Fordyce JA, Gavrilets S. 2009. Pattern, process and geographic modes of

speciation. Journal of Evolutionary Biology 22: 2342-2347.

Flynn JJ, Wyss AR. 1998. Recent advances in South American mammalian paleontology.

Trends in Ecology and Evolution 13: 449-454.

Gentry AH. 1982. Neotropical floristic diversity: phytogeographical connections between

Central and South America, Pleistocene climatic fluctuations, or an accident of the

Andean orogeny? Annals of the Missouri Botanical Garden 69: 557–593.

Givnish TJ, Renner SS. 2004. Tropical intercontinental disjunctions: Gondwana breakup,

immigration from the Boreotropics, and transoceanic dispersal. International Journal of

Plant Sciences 165: S1-S6.

Global Soil Data Task Group. 2000. Global Gridded Surfaces of Selected Soil

Characteristics (IGBP-DIS). [Global Gridded Surfaces of Selected Soil Characteristics

(International Geosphere-Biosphere Programme - Data and Information System)]. Data

set. Available on-line [http://www.daac.ornl.gov] from Oak Ridge National Laboratory

Distributed Active Archive Center, Oak Ridge, Tennessee, U.S.A.

Goldberg EE, Lancaster LT, Ree RH. 2011. Phylogenetic Inference of Reciprocal Effects

between Geographic Range Evolution and Diversification. Systematic Biology 60: 451-

465.

Graham CH, Ron SR, Santos JC, Schneider CJ, Moritz C. Integrating phylogenetics and

environmental niche models to explore speciation mechanisms in dendrobatid frogs.

Evolution 58: 1781-1793.

Grant V. 1981. Plant speciation. 2nd Ed. Columbia University Press. 432 p.

!!

! 41!

Gregory-Wodzicki KM. 2000. Uplift history of the Central and Northern Andes: A review.

GSA Bulletin 112: 1091-1105.

Haffer J. 2008. Hypotheses to explain the origin of species in Amazonia. Brazilian Journal of

Biology 68: 917-947.

Hickey LJ, Peterson RK. 1978. Zingiberopsis, a fossil genus of the ginger family from Late

Cretaceous to early Eocene sediments of western interior North America. Canadian


Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A. 2005. Very high resolution

interpolated climate surfaces for global land areas. International Journal of Climatology

25: 1965-1978.

Hoorn C, Guerrero J, Sarmiento GA, Lorente MA. 1995. Andean tectonics as a cause for

changing drainage patterns in Miocene northern South America. Geology 23: 237-240.

Hoorn C, Wesselingh FP, ter Steege H, Bermudez MA, Mora A, Sevink J, Sanmartín I,

Sanchez-Meseguer A, Anderson CL, Figueiredo JP, Jaramillo C, Riff D, Negri FR,

Hooghiemstra H, Lundberg J, Stadler T, Sarkinen T, Antonelli A. 2010. Amazonia

through time: Andean uplift, climate change, landscape evolution, and biodiversity.

Science 330: 927-931.

Hoorn C, Wesselingh F (Ed). 2011. Amazonia, landscape and species evolution: a look into

the past. Wiley-Blackwell. 464 p.

Hopkins R, Rausher MD. 2012. Pollinator-mediated selection on flower color allele drives

reinforcement. Science 335: 1090-1092.

Hughes C, Eastwood R. 2006. Island radiation on a continental scale: Exceptional rates of

plant diversification after uplift of the Andes. Proceedings of the National Academy of

Sciences 103: 10334-10339.

Imhoff ML, Bounoua L. 2006. Exploring global patterns of net primary production carbon

supply and demand using satellite observations and statistical data. Journal of

Geophysical Research 111: D22S12.

!!

! 42!

Jaramillo C, Cárdenas A. 2013. Global Warming and Neotropical Rainforests: A Historical

Perspective. Annual Review of Earth and Planetary Sciences 41: 28.1-28.26.

Jaramillo C, Rueda M, Mora G. 2006. Cenozoic plant diversity in the Neotropics. Science

311: 1893–1896.

Jaramillo C, Ochoa D, Contreras L, Pagani M, Carvajal-Ortiz H, Pratt LM, Krishnan S,

Cardona A, Romero M, Quiroz L, Rodriguez G, Rueda MJ, Parra F, Morón S, Green W,

Bayona G, Montes C, Quintero O, Ramirez R, Mora G, Schouten S, Bermudez H,

Navarrete R, Parra F, Alvarán M, Osorno J, Crowley JL, Valencia V, Vervoort J. 2010.

Effects of rapid global warming at the Paleocene-Eocene boundary on Neotropical

vegetation. Science 330: 957–961.

Johansen LB. 2005. Phylogeny of Orchidantha (Lowiaceae) and the Zingiberales based on

six DNA regions. Systematic Botany 30: 106-117.

Katz ME, Miller KG, Wright JD, Wade BS, Browning JV, Cramer BS, Rosenthal Y. 2008.

Stepwise transition from the Eocene greenhouse to the Oligocene icehouse. Nature

Geoscience 1: 329-334.

Kay KM. 2006. Reproductive isolation between two closely related hummingbird-pollinated


Kay KM, Schemske DW. 2003. Pollinator assemblages and visitation rates for 11 species of

Neotropical Costus (Costaceae). Biotropica 35: 198-207.



(Costaceae): Evidence from nrDNA ITS and ETS sequences. American Journal of

Botany 92: 1899-1910.

Kress WJ, Specht CD. 2006. The evolutionary and biogeographic origin and diversification of

the tropical monocot oder Zingiberales. Aliso 22: 619–630.

Kriticos DJ, Webber BL, Leriche A, Ota N, Macadam I, Bathols J, Scott JK. 2012. CliMond:

global high-resolution historical and future scenario climate surfaces for bioclimatic

modeling. Methods in Ecology and Evolution 3: 53-64.

!!

! 43!

Landis MJ, Matzke NJ, Moore BR, Huelsenbeck JP. 2013. Bayesian analysis of

biogeography when the number of areas is large. Systematic Biology 62: 789-804.

Lavin M, Luckow M. 1993. Origins and relationships of tropical North America in the context

of the Boreotropics hypothesis. American Journal of Botany 80: 1-14.

Lavin M, Schrire B, Lewis G, Pennington T, Delgado-Salinas A, Thulin M, Hughes C, Beyra

Matos A, Wojciechoswki MF. 2004. Metacommunity process rather than continental

tectonic history better explains geographically structured phylogenies in legumes.

Philosophical Transactions of the Royal Society London (Biological Sciences) 359: 1509-

1522.

Losos JB, Glor RE. 2003. Phylogenetic comparative methods and the geography of

speciation. Trends in Ecology and Evolution 18: 220-227.

Lovette IJ, Bermingham E, Ricklefs RE. 2002. Clade-specific morphological diversification

and adaptive radiation in Hawaiian songbirds. Proceedings of the Royal Society B

(Biological Sciences) 269: 37-42.

Luteyn JL. 2002. Diversity, adaptation, and endemism in neotropical Ericaceae:

biogeographical patterns in the Vaccinieae. Botanical Review 68: 55-87.

Maas PJM. 1972. Costoideae (Zingiberaceae). Flora Neotropica monograph no 8. Hafner,

New York, New York, USA.

Maas PJM. 1977. Renealmia (Zingiberaceae-Zingiberoideae) Costoideae (Additions)

(Zingiberaceae). Flora Neotropica monograph no 18. Hafner, New York, New York, USA.

Matzke NJ. 2013. Probabilistic Historical Biogeography: New Models for Founder-Event

Speciation, Imperfect Detection, and Fossils Allow Improved Accuracy and Model-

Testing. Frontiers of Biogeography 5: 242-248.

McLoughlin S. 2001. The breakup history of Gondwana and its impact on pre-Cenozoic

floristic provincialism. Australian Journal of Botany 49: 271-300.

Metcalfe I. 1991. Late Palaeozoic and Mesozoic palaeogeography of Southeast Asia.

Palaeogeography, Palaeoclimatology, Palaeoecology 87: 211-221.

!!

! 44!

Miller MA, Pfeiffer W, Schwartz T. 2010. Creating the CIPRES Science Gateway for

Inference of Large Phylogenetic Trees. In Proceedings of the Gateway Computing

Environments Workshop (GCE), New Orleans, LA: 1-8.

Montes C, Cardona A, McFadden R, Morón SE, Silva CA, Restrepo-Moreno S, Ramírez DA,

Hoyos N, Wilson J, Farris D, Bayona GA, Jaramillo CA, Valencia V, Bryan J, Flores JA.

2012. Evidence for middle Eocene and younger land emergence in central Panama:

Implications for Isthmus closure. GSA Bulletin 124: 780-799.

Morley RJ. 2003. Interplate dispersal paths for megathermal angiosperms. Perspectives in

Plant Ecology, Evolution and Systematics 6: 5-20.

Morlon H, Parsons TL, Plotkin JB. 2011. Reconciling molecular phylogenies with the fossil

record. Proceedings of the National Academy of Sciences USA 108: 16327-16332.

Moritz C, Patton JL, Schneider CJ, Smith TB. 2000. Diversification of rainforest faunas: An

integrated molecular approach. Annual Review of Ecology and Systematics 31: 533-563.

Nathan R. 2006. Long-Distance Dispersal of Plants. Science 313: 786-788.

Nee S, May RM, Harvey PH. 1994. The reconstructed evolutionary process. Philosophical

Transactions of the Royal Society of London B (Biological Sciences) 344: 305-311.

Ortiz-Jaureguizar E, Cladera G. 2006. Paleoenvironmental evolution of southern South

America during the Cenozoic. Journal of Arid Environments 66: 489-532.

Palma-Silva C, Wendt T, Pinheiro F, Barbará T, Fay MF, Cozzolino S, Lexer C. 2011.

Sympatric bromeliad species (Pitcairnia spp.) facilitate tests of mechanisms involved in

species cohesion and reproductive isolation in Neotropical inselbergs. Molecular Ecology

20: 3185-3201.

Paradis E. 2011. Time-dependent speciation and extinction from phylogenies: a least

squares approach. Evolution 65: 661-67.

Pearson PN, McMillan IK, Wade BS, Jones TD, Coxall HK, Bown PR, Lear CH. 2008.

Extinction and environmental change across the Eocene-Oligocene boundary in

Tanzania. Geology 36: 179-182.

!!

! 45!

Pfennig DW, Servedio MR. 2013. The role of transgenerational epigenetic inheritance in

diversification and speciation. Non-genetic Inheritance 1: 17-26.

Phillips SJ, Anderson RP, Schapire RE. 2006. Maximum entropy modeling of species

geographic distributions. Ecological Modelling 190: 231-259.

R Core Team. 2014. R: A Language and Environment for Statistical Computing. R

Foundation for Statistical Computing. Vienna, Austria.

Rabosky DL. 2009. Heritability of Extinction Rates Links Diversification Patterns in Molecular

Phylogenies and Fossils. Systematic Biology 58: 629-640.

Rabosky DL. 2010. Extinction rates should not be estimated from molecular phylogenies.

Evolution 64: 1816-1824.

Rabosky DL. 2014. Automatic Detection of Key Innovations, Rate Shifts, and Diversity-

Dependence on Phylogenetic Trees. PLoS ONE 9: e89543.

Rabosky DL, Donnellan SC, Grundler M, Lovette IJ. 2014. Analysis and Visualization of

Complex Macroevolutionary Dynamics: An Example from Australian Scincid Lizards.

Systematic Biology 63: 610-627.

Rambaut A, Drummond AJ. 2007. Tracer v1.5, Available from

http://tree.bio.ed.ac.uk/software/tracer/

Räsänen ME, Salo JS, Jungner H, Pittman LR. 1990. Evolution of the western Amazonian

lowland relief: impact of Andean foreland dynamics. Terra Nova 2: 320-332.

Räsänen M, Neller R, Salo J, Jungner H. 1992. Recent and ancient fluvial deposition

systems in the Amazonian foreland basin, Peru. Geological Magazine 129: 293-306.

Rauscher JT. 2002. Molecular phylogenetics of the Espeletia complex (Asteraceae):

evidence from nrDNA ITS sequences on the closest relatives of an Andean adaptive

radiation. American Journal of Botany 89: 1074-1084.

Raven PH, Axelrod DI. 1974. Angiosperm biogeography and past continental movements.

Annals of the Missouri Botanical Garden 61: 539–673.

!!

! 46!

Ree RH, Sanmartín I. 2009. Prospects and challenges for parametric models in historical

biogeographical inference. Journal of Biogeography 36: 1211-1220.

Ree RH, Smith SA. 2008. Maximum likelihood inference of geographic range evolution by

dispersal, local extinction, and cladogenesis. Systematic Biology 57: 4-14.

Ree RH, Moore BR, Webb CO, Donoghue MJ. 2005. A likelihood framework for inferring the

evolution of geographic range on phylogenetic trees. Evolution 59: 2299-2311.

Renner SS. 2004. Tropical trans-Atlantic disjunctions, sea surface currents, and wind

patterns. Interantional Journal of Plant Sciences 165: S23-S33.

Ricklefs RE. 2004. A comprehensive framework for global patterns in biodiversity. Ecology

Letters 7: 1-15.

Ricklefs RE. 2007. Estimating diversification rates from phylogenetic information. Trends in

Ecology and Evolution 22: 601-610.

Rieseberg LH, Willis JH. 2007. Plant Speciation. Science 317: 910-914.

Rull V. 2008. Speciation timing and neotropical biodiversity: the Tertiary-Quaternary debate

in the light of molecular phylogenetic evidence. Molecular Ecology 17: 2722-2729.

Rull V. 2011. Neotropical biodiversity: timing and potential drivers. Trends in Ecology and


Sakai S, Kato M, Inoue T. 1999. Three pollination guilds and variation in floral characteristics

of bornean gingers (Zingiberaceae And Costaceae). American Journal of Botany 86:

646-658.

Salzman S, Driscoll HE, Renner T, André T, Shen S, Specht CD. Spiraling into history: A

molecular phylogeny and investigation of biogeographic origins and floral evolution for

the genus Costus L. Systematic Botany in press.

Särkinen TE, Newman MF, Maas PJM, Maas H, Poulsen AD, Harris DJ, Richardson JE,

Clark A, Hollingsworth M, Pennington RT. 2007. Recent oceanic long-distance dispersal

and divergence in the amphi-Atlantic rain forest genus Renealmia L.f. (Zingiberaceae).

Molecular Phylogenetics and Evolution 44: 968-980.

!!

! 47!

Schluter D. 2000. Ecological character displacement in adaptive radiation. American

Naturalist 156: S4-S16.

Seehausen O. 2004. Hybridization and adaptive radiation. Trends in Ecology and Evolution

19: 198-207.

Shaw J, Lickey EB, Schilling EE, Small RL. 2007. Comparison of whole chloroplast genome

sequences to choose noncoding regions for phylogenetic studies in angiosperms: the

tortoise and the hare III. American Journal of Botany 94: 275-288.

Silvestro D, Schnitzler J, Zizka G. 2011. A Bayesian framework to estimate diversification

rates and their variation through time and space. BMC Evolutionary Biology 11: 311.

Soltis DE, Soltis PS. 1999. Polyploidy: Recurrent formation and genome evolution. Trends in

Ecology and Evolution 14: 348–352.

Soltis DE, Soltis PS, Tate JA. 2003. Advances in the study of polyploidy since ‘Plant

Speciation’. New Phytologist 161: 173-191.

Specht CD. 2006a. Gondwanan vicariance or dispersal in the tropics? The biogeographic


Specht CD. 2006b. Systematics and evolution of the tropical monocot family Costaceae




Specht CD, Yockteng R, Almeida AMR, Kirchoff BK, Kress JW. 2012. Homoplasy,

Pollination, and Emerging Complexity During the Evolution of Floral Development in the

Tropical Gingers (Zingiberales). Botanical Review. 78:440-462.

Stebbins Jr GL. 1959. The role of hybridization in evolution. Proceedings of the American

Philosophical Society 103: 231-251.

Taberlet P, Gielly L, Pantou G, Bouvet J. 1991. Universal primers for amplification of three

non-coding regions of chloroplast DNA. Plant Molecular Biology 17: 1105-1109.

!!

! 48!

Thiers B. [continuously updated]. Index Herbariorum: A global directory of public herbaria

and associated staff. New York Botanical Garden's Virtual Herbarium.

http://sweetgum.nybg.org/ih/

Toon A, Cook LG, Crisp MD. 2014. Evolutionary consequences of shifts to bird-pollination in

the Australian pea-flowered legumes (Mirbelieae and Bossiaeeae). BMC Evolutionary

Biology 14: 43

Turchetto-Zolet AC, Pinheiro F, Salgueiro F, Palma-Silva C. 2013. Phylogeographical

patterns shed light on evolutionary process in South America. Molecular Ecology 22:

1193-1213.

Wagner N, Silvestro D, Brie D, Ibisch PL, Zizka G, Weising K, Schulte K. 2013. Spatio-

temporal evolution of Fosterella (Bromeliaceae) in the Central Andean biodiversity

hotspot. Journal of Biogeography 40: 869-880.

Wendt T, Canela MBF, Klein DE, Rios RI. 2002. Selfing facilitates reproductive isolation

among three sympatric species of Pitcairnia (Bromeliaceae). Plant Systematics and

Evolution 232: 201-212.

Wendt T, Coser TS, Matallana G, Guilherme FAG. 2008. An apparent lack of prezygotic

reproductive isolation among 42 sympatric species of Bromeliaceae in southeastern

Brazil. Plant Systematics and Evolution 275: 31-41.

Werner GDA, Cornwell WK, Sprent JI, Kattge J, Kiers ET. 2014. A single evolutionary

innovation drives the deep evolution of symbiotic N2-fixation in angiosperms. Nature

Communications 5: 4087.

Wesselingh FP, Salo J. 2006. A Miocene perspective on the evolution of Amazonian biota.

Scripta Geologica, 133: 439-458.

White TJ, Bruns T, Lee S, Taylor J. 1990. Amplification and direct sequencing of fungal

ribosomal RNA genes for phylogenetics. Pp. 315-322 in PCR protocols: A guide to

methods and applications. Innis M, Gelfand D, Sninsky J, White T (Eds). San Diego:

Academic Press.

!!

! 49!

Wing SL, Herrera F, Jaramillo CA, Gómez-Navarro C, Wilf P, Labandeira CC. 2009. Late

Paleocene fossils from the Cerrejón Formation, Colombia, are the earliest record of

Neotropical rainforest. Proceedings of the National Academy of Sciences USA 106:

18627-18632.

Zachos J, Pagani M, Sloan L, Thomas E, Billups K. 2001. Trends, rhythms, and aberrations

in global climate 65 Ma to present. Science 292: 686-693.

!!

! 50!

!!

! 51!

Chapter 3 – Spiraling into history: A molecular phylogeny and investigation of

biogeographic origins and floral evolution for the genus Costus L.

Shayla Salzman,1, 2 Heather E. Driscoll,1,3 Tanya Renner,1,4 Thiago André,1,5 Stacy Shen,1

and Chelsea D. Specht1,6

1 – Departments of Plant and Microbial Biology and Integrative Biology and The University

and Jepson Herbaria, University of California at Berkeley, Berkeley CA, U.S.A.

2 – Current address: Department of Organismic and Evolutionary Biology, Harvard

University, Cambridge MA, U.S.A.

3 – Current address: Vermont Genetics Network, Department of Biology and Physical

Education, Norwich University, Northfield VT, U.S.A.

4 – Current address: Center for Insect Science and the Department of Entomology,

University of Arizona, Tucson AZ, U.S.A.

5 – Current address: Laboratório Integrado de Sistemática Vegetal, Departamento de

Botânica, Universidade Federal do Rio de Janeiro, Rio de Janeiro RJ, Brazil

6 – Author for Correspondence ([email protected])

*manuscript accepted in Systematic Botany

(ISSN, Print: 0363-6445; Online: 1548-2324)

!!

! 52!

Abstract

Rapid radiations are notoriously difficult to resolve, yet understanding phylogenetic patterns

in such lineages can be useful for investigating evolutionary processes associated with

bursts of speciation and morphological diversification. Here we present an expansive

molecular phylogeny of Costus L. (Costaceae Nakai) with a focus on the Neotropical species

within the clade, sampling 47 of the known 51 Neotropical species and including five

molecular markers for phylogenetic analysis (ITS, ETS, rps16, trnL-F, and CaM). We use the

phylogenetic results to investigate shifts in pollination syndrome, with the intention of

addressing potential mechanisms leading to the rapid radiation documented for this clade.

Our ancestral reconstruction of pollination syndrome presents the first evidence in this genus

of an evolutionary toggle in pollination morphologies, demonstrating both the multiple

independent evolutions of ornithophily (bird pollination) as well as reversals to melittophily

(bee pollination). We show that the ornithophilous morphology has evolved at least eight

times independently with four potential reversals to melittophilous morphology, and confirm

prior work showing that neither pollination syndrome defines a monophyletic lineage. Based

on the current distribution for the Neotropical and African species, we reconstruct the

ancestral distribution of the Neotropical clade to the Pacific Coast of Mexico and Central

America. Our results indicate an historic dispersal of a bee-pollinated taxon from Africa to

the Pacific Coast of Mexico/Central America, with subsequent diversification leading to the

evolution of a bird-pollinated floral morphology in multiple derived lineages.

!!

! 53!

3.1. Introduction

Costus L. is a pantropical genus in the monocot family Costaceae. Costus is the most

diverse genus within the Costaceae, with approximately 80 species in total. It comprises two

main biogeographic groups; a paraphyletic assemblage of low-diversity clades native to

tropical Africa and a large derived clade containing approximately 51 species and distributed

exclusively in the Neotropics. The derived Neotropical clade appears to have arisen from a

single long distance dispersal event from Africa occurring approximately 34 ma (Specht

2006b). Costus species can be recognized by their characteristic monistichous spiral

phyllotaxy, tubular sheathing leaf bases each with a pronounced ligule, and terminal (mostly)

inflorescences with imbricate bracts arranged in several series of parastichies. While most

species are terrestrial rhizomatous herbs, a few African species are epiphytic (C. talbotii, C.

lateriflorus) and these have axillary rather than terminal inflorescences. Costus range in

vegetative height from less that one meter to over 3 m tall. They tend to grow most

abundantly in moist low-lands, wet-thickets, clearings or stream beds at relatively low

elevations (< 800 m) but some species have been collected at 2,000 m above sea level.

The evolution of two specific pollination syndromes sets Costus apart from other

genera in the family (Kay et al. 2006). Ornithophilous (hummingbird attracting) Costus have

inflorescences constructed of mostly red, orange or yellow bracts and flowers with narrow,

tubular openings. Melittophilous (bee attracting) Costus have mostly green-bracted

inflorescences and flowers with a wider floral opening and a broad labellum that is white or

yellow with distinct red or purple stripes forming a ‘landing platform’ (Fig. 1). These

morphologies have been shown to be consistent with hummingbird and bee pollination

respectively (Kay and Schemske 2003) and such morphology-based signaling appears to be

more important than reward for defining pollination type, as both types of flowers produce

copious nectar (see Thomson and Wilson 2008). Work done in Bornean gingers

(Zingerberaceae and Costaceae) found no significant difference in sugar concentration

between hummingbird and bee pollinated flowers but a highly elevated daily sugar

!!

! 54!

production in hummingbird plants (Sakai et al. 1999). However, this has not been

investigated within Costus alone. Previous work shows that bee pollination originally evolved

in Africa and is ancestral to the Neotropical clade (Specht et al. 2001; Kay et al. 2005;

Specht 2006b), and that hummingbird pollination is derived within the Neotropical clade and

has evolved at least seven times independently (Specht 2006a).

!!

! 55!

Figure 1. Representative photos of Neotropical Costus species. (A) Costus guanaiensis

Rusby var. guanaiensis showing the melittophilous morphology. (B) Costus comosus and

(C) Costus scaber showing the ornithophilous morphology. Photos by C. D. Specht.

!!

! 56!

It has been suggested that shifts in pollinator-specific morphologies may account for

the rapid radiation seen in the Neotropical clade (Kay et al. 2005). Indeed, diversification

rates with in the New World Costus clade have been shown to be the second highest in the

family, just behind the Asian genus Tapeinochilos Miq. (Specht 2005). Within Costaceae,

Tapeinochilos is the only other genus with species displaying a distinct bird-pollination

syndrome, associated with pollination by native sunbirds (O. Gideon pers. comm.).

Fossil calibrated molecular dating analyses using chloroplast markers (trnL-F and

trnK) puts the diversification of Costus at around 40 million years ago (ma) with early

diverging lineages occurring exclusively in Africa and maintaining a plesiomorphic floral

morphology that is neither specifically associated with bee nor hummingbird pollination

(Specht 2006b). Melittophilous Costus are suggested to have evolved around 34 ma. One of

these melittophilous Costus dispersed to the New World, and the fossil calibrated dating

places the New World radiation at around 22 ma (Specht 2006b) with both floral forms

present by 20 ma. An ITS molecular clock analysis suggests that the Neotropical

diversification occurred much more rapidly, with the ca. 50 species diversifying within the last

four million years (Kay et al. 2005). Parsimony ancestral state reconstruction using broad

geographic species ranges placed the dispersal from Africa to Central America (Kay et al.

2005).

Here we present a phylogenetic hypothesis for the genus Costus with expanded

taxon sampling, including 47 of approximately 51 Neotropical taxa, and increased character

sampling including chloroplast (rps16, trnL-F), nuclear (CaM), and nuclear ribosomal (ITS,

ETS) molecular data. We analyze ancestral pollination syndromes using Bayesian,

maximum parsimony, and maximum likelihood approaches and investigate biogeographic

patterns of dispersal and vicariance using narrowly defined geographic ranges for ancestral

area reconstruction. We confirm the multiple independent origins of the melittophilous and

!!

! 57!

ornithophilous morphologies, and show the first evidence for reversal from ornithophilous to

melittophilous morphology. We hypothesize that the origin and evolution of these specific

pollination-driven morphologies are important contributors to the rapid radiation seen in this

clade.

!!

! 58!

3.2. Material and Methods

3.2.1. Taxon Sampling

Seventy-four Costus individuals were selected in an attempt to sample broadly within

the genus. To provide a strong and robust outgroup, seven individuals representing all other

genera in Costaceae were included. Leaf tissue collected from the field (collectors P. J. M.

Maas, D. Skinner, and C. D. Specht) was supplemented by sampling live tissue from living

collections at the New York Botanical Garden, the UC Botanical Garden at Berkeley, Lyon

Arboretum, and the Smithsonian Greenhouse’s living collection. All newly acquired

sequences are deposited in GenBank with vouchered collection information (Appendix 1).

3.2.2. DNA Sequence Data and Analysis

DNA was extracted from silica dried leaf material using Edwards et al.’s 1991

method. The PCR fragments were generated for two chloroplast markers (rps16 partial

coding sequence with partial trnk-rps16 intergenic spacer and trnL-F intergenic spacer with

partial tRNA-Leu and tRNA-Phe genes), one nuclear gene (partial first intron and partial

coding sequence of calmodulin (CaM)), and two transcribed spacer regions of the ribosomal

DNA array (ITS and ETS) using Phire hot start II DNA polymerase (Thermo Fisher Scientific,

Pittsburgh, Pennsylvania) with a 3 min initial denaturing step at 98 °C, 45 cycles of 5 sec at

98 °C, 15 sec at gene-specific annealing temperatures, and 20 sec at 72 °C, and a final 1

min 72 °C extension. For ETS, 45 cycles included 5 sec at 98 °C followed by 20 sec at 72

°C for a combined anneal and extension. Table 1 lists primer pairs and specific annealing

temperatures. The CaM primers were designed for this study following cloning and

sequencing of PCR products amplified with the Zingiberales-specific primers cam33F and

the reverse complement of cam328R (Johansen 2005). Cycle sequencing was performed

using BigDye v3.1 (Applied Biosystems, Foster City, California) following manufacturer’s

protocol. Sequencing was done at the UC Berkeley Museum of Vertebrate Zoology’s

!!

! 59!

Evolutionary Genetics Laboratory (EGL) on an Applied Biosystems 3730xl DNA analyzer.

Reads were assembled and edited and multiple sequence alignments generated in

Geneious (v5.6.3) using the Geneious alignment algorithm. Manual alignment editing was

done in Geneious (v5.6.3) and Mesquite (Maddison and Maddison 2011).

! 60!

Table 1. Primer pairs and annealing temperatures used in this study. Primers were used for both amplification and

sequencing.

Primer name Primer sequence (5’-3’) Annealing

(oC)

Reference

ITS Leu F GTCCACTGAACCTTATCATTTAG 59.3 Baum et al. 1998

ITS 4 TCCTCCGCTTATTGATATGC White et al. 1990

ETS Costus F CTTTGTTGTGCTCGGCGGAGTTC 72 Kay et al. 2005

18S-IGS GAGACAAGCATATGACTACTGGCAGGATCAACCAG Baldwin & Markos 1998

CaM Costus F TGCTTCTCTCGAACGCTAGAT 66 This study

CaM Costus R GAAACTCGGAATGCCTCCTT This study

rps16x2F2 AAAGTGGGTTTTTATGATCC 57.5 Shaw 2007

trnKx1 TTAAAAGCCGAGTACTCTACC Shaw 2007

trnLc CGAAATCGGTAGACGCTACG 67 Taberlet et al. 1991

trnFf ATTTGAACTGGTGACACGAG Taberlet et al. 1991

!!

! 61!

3.2.3. Phylogenetic Analysis

A concatenated alignment with a total of 4,654 characters containing ITS, ETS, CaM,

rps16, and trnL-F sequences was used to generate phylogenetic hypotheses under

maximum likelihood (ML), maximum parsimony (MP), and Bayesian inference methods.

Gaps and the ends of shorter sequences were treated as missing data. Models of evolution

were determined for each marker alignment as well as the full concatenated alignment by

jModelTest v0.1.1 (Darriba et al. 2012) using Bayesian information criterion. Character

statistics and models of evolution for all alignments are listed in Table 2. Individual marker

trees were run using the appropriate models in PhyML as implemented in Geneious (v5.6.3)

and MrBayes (v3.2; Ronquist and Huelsenbeck 2003) using the CIPRES Scientific Gateway

(Miller et al. 2010). The MCMC parameters for each analysis include 100,000,000

generations using four chains, a cold chain temperature of 0.2, tree sampling every 1,000,

and variable site-specific rate models of: two substitution types and gamma distributed rate

variation with a proportion of invariable sites for ITS, six substitution types with proportion of

sites invariable for CaM, and six substitution types with gamma distributed rate variation for

ETS, rps16, and trnL-F. Runs were stopped early if a convergence diagnostic of 0.01 was

met. A consensus tree for each single marker analysis was generated using the sumtrees

command with a minimum clade frequency of 50% and after a burnin of 10% as determined

by visualizing posterior distributions of the parameter values in Tracer (v1.5; Rambaut and

Drummond 2007). Five separate ML analyses were performed on the full concatenated

alignment in PhyML (Guindon et al. 2010) using TIM3 + G with random starting trees,

random sequence addition, and SPR topology searches. Additionally, to assess clade

support 1,000 ML bootstrap replicates were run using TIM3 + G in PhyML as implemented in

Geneious (v5.6.3). The MP analysis was conducted in PAUP*(v4.0b10) using 1,000

replicates of random sequence addition, TBR topology searches, and holding up to 1,000

trees per replicate. Parsimony bootstrap replications (1,000) with ten replicates each of

random sequence addition and TBR searching were also performed in PAUP*. A partitioned

!!

! 62!

Bayesian analysis was performed on the concatenated alignment using the same models for

each partition as described for individual marker analyses and the same MCMC parameters,

with the exception of the temperature parameter (here set to 0.1). A consensus tree was

generated using the sumtrees command with a minimum clade frequency of 50% and after a

burnin of 10% as determined by visualizing posterior distributions of the parameter values in

Tracer (v1.5; Rambaut and Drummond 2007). Phylogenetic analyses and alignments are

accessible on TreeBASE (study number TB2:S16238).

Table 2. Character values and models of evolution for each marker and the concatenated

alignment.

Marker Number of Characters

Proportion of Missing

Data

Parsimony Informative Characters

Constant Characters

Model of Evolution

ITS 841 0.0958 219 393 K80+I+G ETS 651 0.1228 291 210 TPM2uf+G CaM 818 0.1207 247 381 TPM3+I rps16 1164 0.1270 45 924 TPM1uf+G trnL-F 1180 0.1230 59 949 TrN+G Concatenated 4654 0.1191 861 2857 TIM3+G

3.2.4. Ancestral State Reconstruction Analysis

The topology recovered in the ML analysis (Fig. 2) was used for reconstructing

ancestral character states for pollination syndrome and to reconstruct ancestral areas based

on current distributions. The Bayesian phylogeny showed a slightly different topology

(discussed below) and was also used for ancestral state reconstruction; in this case,

polytomies were resolved to have branch lengths of 0 prior to analyses using the Mesquite

command ‘resolve polytomies.’ Pollination type was coded as generalist, bee, or bird based

on the pollination syndrome morphologies tested by Kay and Schemske (2003; Fig. 1).

Models of pollination rate shift variation were tested in BayesTraits (Pagel and Meade

2004a). All rate shifts were found to be independent. Shifts between generalist and bee or

!!

! 63!

between generalist and bird were below 10 and shifts between bee and bird were above

100. Ancestral state reconstruction for pollination syndrome was tested using a ML

approach. Rates were permitted to be independent and vary in both Mesquite using

StochChar model Mk1 with a decision threshold of 2.0 (Maddison and Maddison 2006) and

in BayesTraits using BayesMultiState (Pagel et al. 2004). Bayesian reconstruction was also

tested in BayesTraits using BayesMultiState with 5,050,000 generations with sampling every

300. A prior of uniform distribution of 0-10 was designated for rates between generalist

pollination and either bird or bee, while a prior of uniform distribution of 50-150 was set for

rates between bee and bird. The rate deviation was set to 15 so as to allow for an

acceptance rate of around 20 percent for new models with each generation. Bayesian

results were tested for statistical significance using the Bayes factor test of twice the

difference between the harmonic means of runs, where the node in question was forced as

bee and then as bird pollination. Any positive value favors the model; a value greater than

(>) two is positive evidence, > five is strong evidence, and > 10 is very strong evidence for

the model. Similarly, the significance of likelihood values given in BayesTraits analyses were

tested using the likelihood ratio test. A significant result is defined as the difference of at

least two log likelihood scores between runs when the node in question was forced as either

bee or bird pollination (Pagel 1999).

Current distributions for all Costus species sampled were assembled as point data

using locality data from the Global Biodiversity Information Facility (GBIF), and point data

were coded as pertaining to any of 14 individual geographic ecoregions (Olsen et al. 2001;

see map, Fig. 2). Data points from GBIF were culled from inclusion (assumed erroneous)

when only one collection for a given species was found in any particular ecoregion, unless

the collection was positively identified and expertly curated. Ancestral distribution was tested

with statistical dispersal-vicariance analysis (S-DIVA) (Yu et al. 2010) and Bayesian

analyses in reconstruct ancestral state in phylogenies (RASP) (Yu et al. 2012). The S-DIVA

reconstruction was set to a maximum of four possible ancestral distribution areas and the

!!

! 64!

bound and hold values were allowed to be the maximum. RASP’s Bayesian binary method

includes two models, the fixed character state JC model and the estimated character state

F81 model. As distributions gathered from GBIF are extensive yet may not cover the breadth

of distribution, both were tested and with and without gamma distribution. All runs used a

wide root distribution at 50,000 generations, sampling every 100, using 10 chains, with a

cold chain temperature of 0.1 and a maximum of four possible ancestral distribution areas.

Additionally, elevations were acquired for the GBIF data points using ArcGIS 9.3 elevation

and imagery data and tested for correlation with pollination syndrome using BayesTraits

(Pagel et al. 2004). A MCMC of 1,010,000 generations was run twice using a continuous

random walk model with autotune rate deviation and a uniform prior of 0 to 1,500 for alpha-

elevation and a uniform prior of 0 to 2 for alpha-pollination. The two runs were compared

after a burnin of 10,000 and the significance of positive or negative correlation was tested

with the Bayes factor test.

3.2.5. Diversification Analysis

The topology recovered in the ML analyses was tested for shifts in diversification

rates using Bayesian analysis of macroevolutionary mixtures (BAMM) version 2.0.0

(Rabosky 2014) and analyzed using BAMMtools version 1.0.1 (Rabosky et al. 2014). The

ML tree was first trimmed of outgroup taxa and made ultrametric using the R package ape

version 3.1-2 (Paradis et al. 2004). Bayesian priors for BAMM were determined with

setBAMMpriors in BAMMtools. 1,000,000,000 generations were tested for convergence with

the coda package (Plummer et al. 2006) in R after a burnin of ten percent.

!!

! 65!

3.4. Results

3.4.1. Molecular Data and Phylogenetic Inference

Single-marker Bayesian and ML trees show similar assemblages of African Costus

taxa and a monophyletic Neotropical clade (data available in TreeBASE TB2:S16238).

However, relationships among African clades and within Neotropical Costus are unresolved

and/or poorly supported in most single marker topologies. Despite some discordance at the

single-marker level, well supported nodes were largely congruent and the cumulative signal

from the concatenated data yielded well-supported trees in Bayesian and ML analyses. As a

result, a multi-marker approach was favored to maximize phylogenetic information (Smith

2000). The following results and discussion are based on analyses using the concatenated

alignment.

The ML tree (one of five topologically identical random starts maximum likelihood

trees of a concatenated alignment of ITS, ETS, CaM, trnL-F, and rps16 using TIM3 with

gamma distribution) has a log-likelihood of -23,899.35952 (Fig. 2). The 1,000 replication ML

bootstrap consensus tree, using TIM3 with gamma distribution, gave largely the same

topology with low bootstrap support scattered across the Neotropical Costus clade.

The MP analysis of the full data set resulted in 707,000 most parsimonious trees with

a score L = 3,027. A strict consensus of these trees supports a number of clades (Fig. 2,

bold) with further 50% consensus support (Fig. 2, bold grey). The parsimony bootstrap

obtained low support values scattered across the tree, however overall the topology does

not disagree with the ML tree with the exception of a few relationships discussed below.

The combined Bayesian analysis shows an average standard deviation of split

frequencies between the two runs of 0.0047 at 100,000,000 generations. After a burnin of

10%, the remaining 180,020,000 trees have a mean log likelihood of -22,178.236 with an

effective sample size of 17,281. All other statistics between runs have mean values with

!!

! 66!

effective sample sizes greater than 16,500. The topology reconstructed by Bayesian

analysis largely supports the ML tree. The exceptions are discussed further below.

!!

! 67!

Figure 2. Maximum likelihood hypothesis for Neotropical Costus. ML cladogram (with

support values) showing character state reconstruction data and phylogram (with branch

lengths) for the concatenated alignment of ITS, ETS, CaM, rps16, and trnL-F using TIM3

!!

! 68!

with gamma distribution and log likelihood score of -23,944. Support values at nodes are

Bayesian posterior probabilities, ML, and MP bootstrap proportions. Nodes supported by a

PAUP* strict consensus of 707,000 most parsimonious trees are shown in bold with 50%

consensus in bold grey. Distributions were obtained from herbaria records housed in the

Global Biodiversity Information Facility. Ecoregions are based on the World Wildlife Fund’s

ecoregion designations (Olsen et al. 2001). Ancestral distribution for the Neotropical clade

indicated (Pacific Coast of Mexico and Central America) based on S-DIVA and MCMC

algorithms implemented in RASP. ML reconstructions performed in Mesquite and

BayesTraits are indicated: shifts to ornithophilous morphology are denoted with

hummingbirds and shifts to melittophilous morphology denoted with bees. Extant pollination

syndromes are denoted with hummingbirds or bees at tips.

!!

! 69!

3.4.2. Phylogenetic Relationships

Costus is reconstructed as monophyletic, with African species forming numerous

small clades as early diverging lineages (Fig. 2). The New World clade is monophyletic and

sister to a clade of African species with a bee-pollinated morphology. Small differences in

relationships suggested by each of the phylogenetic methods make clear that further

nucleotide sampling is necessary to completely resolve species relationships within the

Neotropical lineage. However, questions regarding ancestral distribution and pollination

shifts in the Neotropical clade can be addressed given the resolution acquired. Taxa that

show different placement or form polytomies when using different inference methods are

discussed below.

In the Old World, C. dubius is strongly supported as sister to the C. albiflos/C.

maboumiensis clade in the Bayesian tree (1 posterior probability) and the MP strict

consensus tree. This relationship is weakly supported in the ML tree (BS 49%), yet remains

sister to C. maculatus (a proposed synonym for C. dubius) in all ML runs.

Within the Neotropical lineage, the clade formed by C. vinosus and C. beckii is

weakly supported as sister to the C. dirzoi/C. malortieanus clade in parsimony (BS 21%) and

Bayesian (PP 0.62). Interestingly, however, these two taxa are placed as sister to the

remainder of the New World clade in the ML tree. While C. longebracteolatus clearly

belongs in the C. glaucous/C. malortieanus clade as supported by all methods, its exact

placement is unresolved. Similarly, the placement of the C. asplundii/C. arabicus clade is

unresolved, as Bayesian and ML bootstrap methods very weakly support it as sister to the C.

claviger/C. woodsonii clade (0.22 posterior probability and 16% respectively), while

parsimony methods maintain a polytomy among these taxa. Finally, it is unclear if the C.

laevis/C. wilsonii clade is indeed sister (ML) to the C. lasius/C. allenii clade, as other

inference methods very weakly support it as sister to the larger C. lasius/C. varzearum clade

(0.35 PP, MLBS 35%, MPBS 19%).

!!

! 70!

We feel most confident with the topology represented in the ML tree as, in the four

cases discussed, only the C. vinosus/C. beckii clade shows a different placement altogether.

The other three cases resolved in the ML tree are merely unresolved with other methods,

which could be expected as rapid radiations may not have sufficient data to inform the priors

in Bayesian analysis. However, because these differences were seen between the ML tree

and the Bayesian tree, ancestral pollination reconstruction was run on both, specifically to

investigate if the placement of the C. vinosus/C. beckii clade would affect the number of

hypothesized ancestral character state shifts in pollination syndrome.

3.4.3. Ancestral Character State Reconstruction: Pollination Syndromes

A presumed generalist pollination morphology is recovered as ancestral to the genus,

with a monophyletic shift to a melittophilous morphology within the Old World grade (Fig. 3;

Fig. 4). We recognize that generalist pollination used as the plesiomorphic state may reflect

a lack of knowledge of the pollination biology of the early diverging lineages within the

Costaceae (ex. Dimerocostus, Monocostus and Chamaecostus) which have been

historically treated as generalists because pollination syndromes based on flower

morphology are not conclusive for these lineages, as with the melittophilous and

ornithophilous pollination syndromes of Costus tested by Kay and Schemske (2003). The

monophyletic New World clade was reconstructed as being ancestrally melittophilous,

despite the resolution of ornithophilous taxon C. stenophyllus as the sister to all remaining

Neotropical Costus: this result indicates that bird pollination evolved in C. stenophyllus after

it diverged from the remaining taxa. Based on our recovered topology, there have been up to

nine shifts to hummingbird pollination morphology from the bee-pollinated ancestral form,

with four reversals to a bee pollination syndrome (Fig. 2). Pollination syndromes were weakly

correlated with elevations. The MCMC run showed an average correlation coefficient, R, of

!!

! 71!

0.2 with a standard deviation of 0.01. This correlation is positively supported with a Bayes

factor of 2.78.

!!

! 72!


syndrome on ML tree. Generalist, melittophilous and ornithophilous morphologies are


probabilities from BayesTraits.

!!

! 73!


syndrome on Bayesian tree. Generalist, melittophilous, and ornithophilous morphologies are


probabilities from BayesTraits.

!!

! 74!

3.4.4. Pollination Syndromes Reconstructed on ML Phylogeny

Overall, reconstruction of ancestral character states on the ML tree recover much

stronger statistical significance for nodes tested than are recovered on the Bayesian tree.

The ML pollination syndrome reconstruction tree (Fig. 3) has a negative log likelihood of

69.57 and a rate of 21.479 as determined by the Mk1 model in Mesquite. Bayesian runs had

an average acceptance rate of new models for each generation of 19% showing ideal mixing

of chains. Reconstruction with Mesquite supports nodes along the backbone as statistically

significant for both ornithophilous and melittophilous morphologies even with moderately

high likelihoods pointing to bee pollination. Reconstructed ancestral morphlogies of the

more derived clades, except for C. chartaceus/C. woodsonii (discussed below), are

statistically supported as follows: hummingbird pollination syndrome for the ancestor of the

clade containing C. dirzoi/C. lima; hummingbird for C. dirzoi/C. erythrothyrsus; hummingbird

for C. comosus var. bakeri/C. lima; and bee pollination syndrome morphology for the

ancestor of C. laevis/C. varzearum (Fig. 3). The Bayes factor positively supports all tested

nodes with Bayes factors greater than 2 except for the ancestor of C. chartaceus and C.

woodsonii. However, this ancestor’s posterior probability of 0.95 for the hummingbird

morphology is highly supported (Bayes factor =1.88) and all extant taxa in this clade display

a hummingbird pollination morphology (Fig. 3; Table 3). The ancestor of C. dirzoi and C. lima

shows very strong evidence for having a hummingbird pollination morphology (Bayes factor

= 18.92; posterior probability = 0.77; Mesquite likelihood reconstruction = 0.95), but

interestingly, has only moderate support from BayesTraits likelihood reconstruction (delta

log-likelihood = 0.88; likelihood = 0.72). In fact, the three nodes tested in this C. dirzoi/C.

lima clade are supported as ornithophilous with Mesquite’s likelihood reconstruction, Bayes

factor test and BayesTraits (Fig. 3; Table 3). All three of these methods tested on the ML

tree support a total of nine shifts to the ornithophilous pollination syndrome with four

reversals to a melittophilous pollination syndrome in the C. dirzoi/C. lima clade.

! 75!

Table 3. BayesTraits ML probabilities, Bayesian posterior probabilities, and their support values at each

node tested for ancestral pollination reconstruction using the ML tree. Nodes supported as having

ancestral hummingbird pollination morphology are underlined.

Most Recent Common Ancestor of:

ML Probability

(Bee)

ML Probability

(Bird) Δ log-

likelihood

Posterior Probability

(Bee)


(Bird) Bayes Factor

stenophyllus/woodsonii 0.77 0.23 1.76 0.66 0.34 3.11

barbatus/ woodsonii 0.61 0.39 1.89 0.50 0.50 3.22

beckii/ woodsonii 0.80 0.20 1.90 0.73 0.27 3.32

dirzoi/ woodsonii 0.93 0.07 2.33 0.88 0.12 4.11

dirzoi/lima 0.28 0.72 0.88 0.23 0.77 18.92

dirzoi/erythrothyrsus 0.52 0.48 0.78 0.48 0.52 2.12

comosus v. bakeri/lima 0.23 0.76 1.07 0.19 0.81 2.99

laevis/ woodsonii 0.97 0.03 2.59 0.93 0.07 4.78

laevis/varzearum 0.99 0.01 3.10 0.98 0.02 6.85

asplundii/ woodsonii 0.59 0.41 2.61 0.50 0.50 4.72

chartaceus/ woodsonii 0.07 0.93 1.25 0.05 0.95 1.88

!!

! 76!

3.4.5. Pollination Syndromes Reconstructed on Bayesian Phylogeny

The Mesquite ML reconstruction on the Bayesian tree (Fig. 4) has a log likelihood of 69.02

and a rate of 20.96 as determined by the Mk1 model. Bayesian runs had an average

acceptance rate of new models for each generation of 21% again showing ideal mixing of

chains. Mesquite counts all nodes values as statistically significant for both hummingbird and

bee morphologies, supporting neither as exclusive, except for the three ancestors tested

within the C. dirzoi/C. lima clade, which are all statically significantly for hummingbird

pollination morphology. ML methods in BayesTraits only show strong statistical support for

the ancestor of C. laevis and C. varzearum as having bee pollination morphology, with

moderate to high support for bee pollination morphology in the ancestors of C.

stenophyllus/C. woodsonii, C. pictus/C. woodsonii, C. beckii/C. woodsonii, and C. laevis/C.

woodsonii (Fig. 4; Table 4). Bayesian posterior probabilities at nodes tested in BayesTraits,

however, are all favorably supported with Bayes factors greater than 2, except where there

is positive support for hummingbird pollination morphology in the ancestor of C. beckii and

C. lima (Bayes factor =1.40, posterior probability = 0.70) and an equal probability of

hummingbird or bee pollination syndrome morphologies in the ancestor of C. dirzoi and C.

erythrothrysus (Table 4). The differing topology seen in the Bayesian tree regarding the

placement of C. beckii (hummingbird) and C. vinosus (bee) as sister to the C. dircoi/C. lima

clade, strongly reconstructed as having a hummingbird pollination syndrome as the ancestral

morphology on the ML tree, reduced the number of shifts to ornithophilous morphology to

eight and increased the number of reversals to five with low support (Fig. 4; Table 4).

! 77!

Table 4. BayesTraits ML probabilities, Bayesian posterior probabilities, and their support values at each node tested

for ancestral state reconstruction of pollination syndrome using the Bayesian topology. Nodes supported as having

ancestral hummingbird pollination syndrome morphology are underlined.

Most Recent Common Ancestor of:

ML Probability

(Bee)

ML Probability

(Bird) Δ log-

likelihood


(Bee)


(Bird) Bayes Factor

stenophyllus/ woodsonii 0.72 0.28 1.71 0.57 0.43 2.67

pictus/ woodsonii 0.86 0.14 1.75 0.71 0.29 2.73

beckii/ woodsonii 0.80 0.20 1.56 0.65 0.36 2.36

beckii/lima 0.33 0.67 1.07 0.30 0.70 1.40

dirzoi/lima 0.32 0.68 0.62 0.26 0.74 2.23

dirzoi/erythrothrysus 0.56 0.44 0.56 0.50 0.50 1.83

comosus v. bakeri/lima 0.23 0.77 0.70 0.20 0.80 2.92

laevis/ woodsonii 0.93 0.07 1.72 0.82 0.18 2.72

laevis/varzearum 0.97 0.03 2.27 0.95 0.05 4.26

!!

! 78!

3.4.6. Reconstruction of Ancestral Distribution

S-DIVA as implemented in RASP with a maximum of four possible areas gives a relative

probability = 1.00 for the Pacific Coast of Mexico and Central America as the combined

ancestral area for Neotropical Costus (light yellow area on map, Fig. 2). The Bayesian binary

method was run four times, once with each model, all of which point to a Mexican/Central

American origin in the New World, with the strongest probability of being located along the

Central American Pacific Coast (CAPC; Fig. 2 light yellow), and much smaller probabilities

associated with distributions along the Central American Atlantic Coast (CAAC; Fig.2 pink),

or the Central American Interior (CAI; Fig. 2 red). After 50,000 generations, the F81 model

had a distance between runs of 0.0049 and gave the probabilities of the ancestral

distribution at the Neotropical node (indicated Fig. 2) as CAPC = 69.48%, CAPC plus CAAC

= 5.12%, Old World = 5.17%, and CAPC plus Old World = 8.69%. F81 with gamma

distribution had a distance between runs of 0.0056 and probabilities of CAPC = 36.58%,

CAPC plus CAI = 12.90%, CAPC plus CAAC = 10.64%, CAI = 9.87%, and CAAC = 8.14%.

The JC model had a distance between runs of 0.0060 and probabilities of CAPC = 66.39%,

CAPC plus Old World = 11.05%, Old World = 8.46%. The JC model with gamma distribution

had a distance of 0.0073 and probabilities of CAPC = 48.23%, CAPC plus CAI = 13.67%,

CAI = 8.20%, and CAPC plus CAAC = 7.98%. These methods all support a Mexican/Central

American origin with the Pacific Coast as the most likely area of ancestral distribution (Fig. 2;

light yellow).

3.4.7. Diversification Analysis

Increased rates of speciation seen in Costus appear to be correlated with specialized

pollination morphology. The mean lambda for the clade of all ornithophilous and

melittophilous species (C. afer/C. varzearum) is 78.3732 with a 90% highest density

probability (HDP) of 59.23909 – 100.95127. This is a large increase over the relative rate for

!!

! 79!

species with generalist pollination morphology: 7.680111 (HPD 3.87069 – 13.64808). The

Neotropical clade is further increased with a mean lambda of 205.4129 (HDP 156.0841 –

261.9265). Sub-clades within the Neotropical clade do not show relative rate differences

regardless of pollination morphologies present in the clade. The hummingbird-pollinated

clade of C. chartaceus/C. woodsonii has a mean lambda of 434.3955 (HDP 333.0779 –

558.5928) where the mostly bee-pollinated clade of C. laevis/C. varzearum has 434.322

(HDP 332.7832 – 558.0226). The C. dirzoi/C. lima clade with reversals to melittophilous

morphology has a similar mean lambda of 434.2816 (HDP 332.6214 – 557.7095).

!!

! 80!

3.5. Discussion

Here we have presented a phylogeny for the genus Costus that shows the first

evidence of an evolutionary toggle between hummingbird and bee pollination morphologies

among species in the Neotropical clade. This expanded taxon and molecular marker

phylogeny supports prior work showing that there have been multiple shifts in pollination

syndromes in the genus. The increased sampling and resolution provided in this study

suggests a greater number of origins of hummingbird pollination and finds for the first time

the reversal from ornithophilous to melittophilous morphology, laying the ground work for

further discussion of the factors related to the rapid radiation observed in this clade.

Phylogenetic inference is notoriously difficult for groups that have undergone rapid

radiations, and even the use of rapidly evolving ribosomal spacers and a nuclear intron did

not provide sufficient phylogenetic signal to resolve all evolutionary relationships within this

lineage with strong support. ML analyses, however, resulted in a single topology that was

largely reproduced with a 50% majority rule parsimony tree, excepting only a few close

relationships that were unresolved in parsimony. Bayesian methods supported the same

topology except, again, for a few unresolved polytomies and the placement of the C.

vinosus/C. beckii clade, which was resolved with low support (0.54 posterior probability).

Despite these phylogenetic uncertainties, hypotheses regarding shifts in pollination

morphology can still be tested as can the biogeographic origins of the lineage.

Our results indicate that there have confidently been at least eight independent shifts

to hummingbird pollination from a bee-pollinated ancestor, with the strongest evidence

pointing to nine total shifts. All analyses strongly support Costus stenophyllus, with an

ornithophilous morphology, as sister to the remaining Neotropical taxa suggesting that this

early diverging lineage evolved hummingbird pollination independently from other

hummingbird pollinated species. Further studies investigating the developmental genetics

and morphometrics of floral traits associated with bird pollination across Costus will help

!!

! 81!

define the particular characters involved in the hummingbird pollination trait, and define how

these characters might evolve to produce the complex morphologies associated with the bird

pollination syndrome.

Several unresolved areas of the phylogeny impeded our ability to conclusively state

the total number of character state transitions for the complex pollination syndrome

character, however the support for nodes at which transitions occurred is significant and

indicates that we have arrived at a conservative number of potential shifts for the genus. In

analyses performed on the ML tree, all shifts to hummingbird pollination shown are

statistically significant and nodes are well supported save the C. zingiberoides/C. woodsonii

clade (Fig. 2), which was found in 51 percent of the 707,000 most parsimonious trees; all

extant species of this clade, however, are ornithophilous and are the result of a single

evolution of the bird pollination syndrome. The relationship between the C. asplundii/C.

arabicus and C. quasi-appendiculatus/C. woodsonii clade is still uncertain, as is the

relationship between C. laevis/C. wilsonii and C. lasius/C. allenii, yet the clades themselves

are well supported and indicate the occurrence of multiple, independent shifts to the

ornithophilous morphology from a bee pollinated ancestral form regardless of their

placement. The placement of the C. vinosus/C. beckii clade is still unclear: attempts to infer

the effect of this phylogenetic uncertainty on the hypothesized number of shifts in pollination

morphology were also inconclusive due to low support for a reconstruction of bee pollination

for the most recent common ancestor of C. beckii and C. woodsonii (Table 4) in addition to

the polytomy present in the C. laevis/C. woodsonii clade. Ancestral state reconstruction for

pollination syndrome performed on the Bayesian tree had lower support all around,

suggesting that the phylogenetic uncertainty found in the polytomies confounded ancestral

state reconstruction attempts.

This is the first study to demonstrate there has been a toggle in pollination

morphology, with reversals back to the ancestral bee pollination morphology occurring in the

C. dirzoi/C. lima clade. The monophyly of the pollination-syndrome-toggling C. dirzoi/C. lima

!!

! 82!

clade is very strongly supported with all methods. The ancestral state is strongly supported

as hummingbird pollination morphology with the Bayesian and the Mesquite likelihood

reconstructions (and moderately supported with BayesTraits likelihood reconstruction), and

there is strong support for the ancestral character state reconstructions for nodes within this

clade (Tables 3; Table 4). With only the phylogenetic placement of C. longibracteolatus

remaining to be clarified, this clade clearly shows reversals to melittophilous morphologies

after an initial evolution of ornithophilous morphology. While there appear to have been

eight or nine independent origins of an ornithophilous floral form, there are fewer reversals to

a melittophilous morphology, all of which occur in the C. dirzoi/C. lima clade.

Ancestral area reconstruction analyses all strongly support a Mexican/Central

American point of dispersal to the New World with the highest probability of arrival occurring

along the Pacific Coast. Further dating analyses are required to discuss the biogeographic

history of this group and the effect of distribution on speciation, as temporal evidence is

needed to discuss past biogeographic scenarios (Crisp et al. 2011). We can, however,

discuss the results here in the light of prior work. If the previous fossil calibrated dating is

assumed and the Neotropical radiation occurred around 22 ma, then biogeographic

upheavals such as the Andean uplift (~10 ma) and the near closure of the Isthmus of

Panama [recently shown to be earlier than thought at about 15 ma (Montes et al. 2012) or

perhaps even 31 -16 ma (Bacon et. Al. 2012)], would certainly have played an important role

in speciation through vicariance. Additionally, at that time there were few Costaceae taxa as

part of the flora of the Neotropics, suggesting that intra-lineage competition may not have

had a strong effect on diversification (Specht 2005). This could support the hypothesis that

changing geography pushed populations into higher elevations better populated by

hummingbirds, leading to floral forms that resulted in specific pollination syndromes leading

to reproductive isolation, further supported by the positive support for a weak correlation

between elevation and pollination syndromes shown here with current distributions. If,

however, the ITS molecular clock dating is assumed, with a Neotropical radiation at around

!!

! 83!

four ma (Kay et al. 2005), the factors influencing diversification may be quite the opposite,

with less geographic and climatic upheaval and more competition among Zingiberales taxa

or other bird/bee pollinated plants driving extremely rapid rates of diversification.

The sister clade to Costus, a South American clade consisting of Chamecostus

(seven species), Monocostus (one species), and Dimerocostus (three to five species), would

have experienced a similar biogeographic history, yet failed to diversify either taxonomically

or morphologically to the extent of the diversity exemplified by Costus. These three genera

comprise species that lack the extreme pollination-specific morphologies associated with the

Neotropical Costus clade. Such data support our hypothesis that the rapid radiation

observed in Neotropical Costus is correlated with the specialized pollination morphologies

characteristic of this lineage. An expanded Costaceae phylogeny with outgroup fossil dating

is needed to revisit these questions, and may provide greater insight into the biogeography

of this group and the impact of pollinator-mediated vs. vicariance-related speciation through

means of reproductive isolation and their combined effect on diversification rates.

!!

! 84!

3.6. Acknowledgements

The authors thank Dr. Paul M.J. Maas and Hiltje Maas for their shared expertise in

Costaceae systematics; Y. Wang, I. Liao, H. Cooper, L. Lagomarsino, K. Yu, B. Wong, and

A. Almeida for generating sequence data; Members of the Specht Lab, especially A. Almeida

and R.Yockteng for helpful comments on data generation, analysis and manuscript

preparation. Funding for this project was provided by the National Geographic Society

#8994-11 and UC Berkeley CNR-BSP, SPUR, and URAP.

3.7. References

Bacon, C. D., A. Mora, W. L. Wagner, and C. A. Jaramillo. 2013. Testing geological models

of evolution of the Isthmus of Panama in a phylogenetic framework. Botanical Journal of

the Linnean Society 171: 287-300.

Baldwin, B. G. and S. Markos. 1998. Phylogenetic utility of the external transcribed spacer

(ETS) of 18S–26S rDNA: Congruence of ETS and ITS trees of Calycadenia

(Compositae). Molecular Phylogenetics and Evolution 10: 449-463.

Baum, D., R. Small, and J. Wendel. 1998. Biogeography and floral evolution of baobabs

(Adansonia Bombacaceae) as inferred from multiple data sets. Systematic Biology 47:

181-207.

Crisp, M. D., S. A. Trewick, and L. G. Cook. 2011. Hypothesis testing in biogeography.

Trends in Ecology and Evolution 26: 66-72



Edwards, K., C. Johnstone, and C. Thompson. 1991. A simple and rapid method for the

preparation of plant genomic DNA for PCR analysis. Nucleic Acids Research 19: 1349.

!!

! 85!

Guindon, S., J. F. Dufayard, V. Lefort, M. Anisimova, W. Hordijk, and O. Gascuel. 2010. New

algorithms and methods to estimate maximum-likelihood phylogenies: Assessing the

performance of PhyML 3.0. Systematic Biology 59: 307-331.

Huelsenbeck, J. P. and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogeny.

Bioinformatics 17: 754-755.

Johansen, L. B. 2005. Phylogeny of Orchidantha (Lowiaceae) and the Zingiberales based on

six DNA regions. Systematic Botany 30: 106-117.

Kay, K. M. 2006. Reproductive isolation between two closely related hummingbird-pollinated


Kay, K. M. and D. W. Schemske. 2003. Pollinator assemblages and visitation rates for 11

species of Neotropical Costus (Costaceae). Biotropica 35: 198-207.

Kay, K. M., P. A. Reeves, R. G. Olmstead, and D. W. Schemske. 2005. Rapid speciation

and the evolution of hummingbird pollination in Neotropical Costus subgenus Costus


Botany 92: 1899-1910.

Maddison, W. P. and D. R. Maddison. 2006. StochChar: A package of Mesquite modules for

stochastic models of character evolution. Version 1.1 Available from

http://mesquiteproject.org

Maddison, W. P. and D. R. Maddison. 2011. Mesquite: A modular system for evolutionary

analysis. Version 2.75 Available from http://mesquiteproject.org.

Miller, M. A., W. Pfeiffer, and T. Schwartz. 2010. Creating the CIPRES science gateway for

inference of large phylogenetic trees. New Orleans, Louisiana: Proceedings of the

Gateway Computing Environments Workshop (GCE).

Montes, C., A. Cardona, R. McFadden, S. E. Morón, C. A. Silva, S. Restrepa-Moreno, D. A.

Ramírez, N. Hoyos, J. Wilson, D. Farris, G. A. Bayona, C. A. Jaramillo, V. Valencia, J.

Bryan, and J. A. Flores. 2012. Evidence for middle Eocene and younger land emergence

in central Panama: Implications for Isthmus closure. Geological Society of America

Bulletin 124: 780-799.

!!

! 86!

Olson, D. M., E. Dinerstein, E. D. Wikamanayake, N. D. Burgess, G. V. N. Powell, E.C.

Underwood, J. A. D’Amico, I. Itoua, H.E. Strand, J. C. Morrison, C. J. Loucks, T. F.

Allnutt, T. H. Ricketts, Y. Kura, J. F. Lamoreux, W. W. Wettengel, P. Hedao, and K. R.

Kassem. 2001. Terrestrial ecoregions of the world: A new map of life on Earth.

Bioscience 51: 933-938.

Paradis, E., J. Claude, and K. Strimmer. 2004. APE: Analyses of phylogenetics and

evolution in R language. Bioinformatics 20: 289-290.

Pagel, M. 1999. The maximum likelihood approach to reconstructing ancestral character

states of discrete characters on phylogenies. Systematic Biology 48:612-622

Pagel, M. and A. Meade. 2004a. BayesTraits. Available from http://www.evolution.rdg.ac.uk

Pagel, M., A. Meade, and D. Barker. 2004. Bayesian estimation of ancestral character states

on phylogenies. Systematic Biology 53: 673-684.

Plummer, M., N. Best, K. Cowles, and K. Vines. 2006. CODA: Convergence diagnosis and

output analysis for MCMC. R News 6: 7-11.

Rabosky, D. L. 2014. Automatic detection of key innovations, rate shifts, and diversity-

dependence on phylogenetic trees. PLoS ONE 9: e89543.

Doi:10.1371/journal.pone.0089543

Rabosky, D. L., M. Grundler, C. Anderson, P. Title, J. J. Shi, J. W. Brown, H. Huang, and J.

G. Larson. 2014. BAMMtools: An R package for the analysis of evolutionary dynamics on

phylogenetic trees. Methods in Ecology and Evolution. Doi: 10.1111/2041-210X.12199

Rambaut, A. and A. J. Drummond, 2007. Tracer v1.5, Available from


Ronquist, F. and J. P. Huelsenbeck. 2003. MRBAYES 3: Bayesian phylogenetic inference

under mixed models. Bioinformatics 19: 1572-1574.

Sakai, S., M. Kato, and T. Inoue. 1999. Three pollination guilds and variation in floral

characteristics of Bornean gingers (Zingiberaceae and Costaceae). American Journal of

Botany 86: 646-658.

!!

! 87!

Shaw, J., E. B. Lickey, E. E. Schilling, and R. Small. 2007. Comparison of whole chloroplast

genomes sequences to choose noncoding regions for phylogenetic studies in

angiosperms: The tortoise and the hare III. American Journal of Botany 94: 275-288.

Smith, J. F. 2000. Phylogenetic signal common to three data sets: Combining data which

initially appear heterogeneous. Plant Systematics and Evolution 221: 179–198.

Specht, C. D. 2005. Phylogenetics, floral evolution, and rapid radiation in the tropical

Monocotyledon family Costaceae (Zingiberales). Plant Genome Biodiversity and

Evolution 1B: 29-60.

Specht, C. D. 2006a. Systematics and evolution of the tropical monocot family Costaceae


Specht, C. D. 2006b. Gondwanan vicariance or dispersal in the tropics? The biogeographic

history of the tropical monocot family Costaceae (Zingiberales). Pp. 631-642 in

Monocots: Comparative biology and evolution. eds. J. T. Columbus, E. A. Friar, C. W.

Hamilton, J. M. Porter, L. M. Prince, and M. G. Simpson. Claremont, California: Rancho

Santa Ana Botanic Garden.

Specht, C. D., W. J. Kress, D. W. Stevenson, and R. DeSalle. 2001. A molecular phylogeny

of Costaceae (Zingiberales). Molecular Phylogenetics and Evolution 21: 333-345.

Swofford, D. L. 2003. PAUP*. Phylogenetic analysis using parsimony (*and other methods).

Version 4. Sunderland: Sinauer Associates.

Taberlet, P., L. Gielly, G. Pantou, and J. Bouvet. 1991. Universal primers for amplification of

three non-coding regions of chloroplast DNA. Plant Molecular Biology 17: 1105-1109.

Thomson, J. D., and P. Wilson. 2008. Explaining evolutionary shifts between bee and

hummingbird pollination: Convergence, divergence and directionality. International

Journal of Plant Sciences 169: 23–38.

White, T. J., T. Bruns, S. Lee, and J. Taylor. 1990. Amplification and direct sequencing of

fungal ribosomal RNA genes for phylogenetics. Pp. 315-322 in PCR protocols: A guide

to methods and applications. eds. M. Innis, D. Gelfand, J. Sninsky, and T. White. San

Diego: Academic Press.

!!

! 88!

Yu, Y., A. J. Harris, and X. J. He. 2010. S-DIVA (statistical dispersal-vicariance analysis): A

tool for inferring biogeographic histories. Molecular Phylogenetics and Evolution 56: 848-

850.

Yu, Y., A. J. Harris, and X. J. He. 2012. RASP (reconstruct ancestral state in phylogenies), v.

2.1b. Available at http://mnh.scu.edu.cn/soft/blog/RASP

!!

! 89!

!!

! 90!



Thiago André1,2, Chelsea Specht3, Shayla Salzman4, Clarisse Palma-Silva5, Tânia Wendt2

1 – [email protected]; corresponding author.

2 – Universidade Federal do Rio de Janeiro, Rio de Janeiro (RJ), Brazil.

3 – University of California, Berkeley (CA), USA.

4 – Harvard University, Cambridge (MA), USA.

5 – Universidade Estadual Paulista, Rio Claro (SP), Brazil.

*manuscript submitted to Phytotaxa

(ISSN, Print: 1179-3155; Online 1179-3163)

!!

! 91!

Abstract

While most species within the genus Chamaecostus (Costaceae) are well defined, the broad

geographic range and long list of synonyms associated with Chamaecostus subsessilis led

us to believe there may be some cryptic species within the complex. We thus investigate the

phylogenetic relationships of species in the Chamaecostus lineage and specifically test the

monophyly and diversity of the Chamaecostus subsessilis species complex from a

population perspective by analyzing molecular sequence data and leaf morphometrics. We

interpret evolutionary trends across the entire genus based on a molecular character-based

phylogenetic hypothesis that includes all currently described species of Chamaecostus. Our

results show that while Chamaecostus is strongly monophyletic, C. cuspidatus is found to be

sister to a clade of some but not all samples of C. subsessilis, making it necessary to

acknowledge more than one species in the C. subsessilis complex. Herbarium specimens of

the C. subsessilis complex could be assigned based on geographic proximity to one of the

major three clades recovered in the phylogenetic analysis. Leaf morphometric

measurements were performed on each of these lineages and traits were tested to detect

differences among phylogenetic lineages. We conclude by proposing the recognition of a

new combination, Chamaecostus acaulis, which we describe.

!!

! 92!

4.1. Introduction

Most species are identified on the basis of incongruent patterns and discontinuity of

trait variation, typically using morphological characters. Individuals of a species are

phenomena manifesting itself in rather continuous variation across a population and can

only be understood by sampling through the range of this variation (Frost & Kluge 1994),

which then commonly provide the basis for ascertaining trends. However, perceived

morphological discontinuity it is not always a product of lineage splitting and speciation:

isolated subpopulations bearing genetic structure can lack phenotypic differences between

them (DeSalle et al. 2005, Padial & De La Riva 2009, Padial et al. 2010, Florio et al. 2012).

Such cryptic speciation is characterized by two or more morphologically indistinguishable

groups of organisms that are found to belong to different evolutionary lineages (Sáez &

Lozano 2005). Perceived cryptic speciation can also derive from our inability to distinguish

important, and not always prominent, morphological differences (Shaffer & Thomson 2007).

Using a phylogenetic lineage approach, species-level phylogenies and networks are

able to provide a consistent and predictive evolutionary understanding of species limits

(Funk & Omland 2003, Goldstein & DeSalle 2000). The phylogenetic approach to species

delimitation is particularly promising because distinct species are interpreted as being on

separate evolutionary trajectories (Hey & Pinho 2012), which, in some cases, are expected

to continue to diverge even in the absence of reproductive barriers (Rieseberg et al. 2004).

Here, we investigate the Chamaecostus subsessilis (Nees & Mart.) Specht &

Stevenson (2006: 158) species complex from a population perspective. The broad

geographic range of the complex and the long list of synonyms associated with taxa in this

group indicate that there may be more than one species present, given a phylogenetic

species concept (Nixon & Wheeler 1990). We use DNA sequences to test for monophyly of

species and lineages within Chamaecostus, and study leaf morphometrics to test for

morphological integrity of the lineages examined. Finally, we present a comprehensive

!!

! 93!

molecular phylogeny of Chamaecostus Specht & Stevenson (2006: 157), on which we

interpret evolutionary trends within this genus.

With the sole exception of Chamaecostus subsessilis, species of Chamaecostus are

fairly distinguishable from one another. Taxonomic disputes regarding C. subsessilis are

largely due to its widespread distribution and various subtle morphological differences that

occur across its geographic range. C. subsessilis is a nearly acaulescent herb that inhabits

the seasonally dry forests of Central South America. Individuals placed within this group

comprise the largest geographic range for any species of Chamaecostus, ultimately forming

a species complex that encompasses eleven historically described species. Maas (1972,

1977) emphasized in his monographs that these eleven species could not be separated from

one another based on floral characters alone, which are rather constant throughout the

geographic distribution. Previous taxonomic descriptions referred mainly to variation in

vegetative characters like plant height, leaf shape, and leaf hairiness which were

hypothesized by Maas (1972) to be driven by environmental factors as no clear geographic

isolation separating any of the described forms was observed. Maas combined all eleven

taxa under one species name (Maas 1972, 1976, 1977); initially Maas (1972) used Costus

warmingii Petersen as described in Martius (1890: 57), but later (Maas 1976) revised the

name based on the previously described type of Globba subsessilis, described by Nees and

Martius (1823: 29) but not cited in Martius’ Flora Brasilensis (1890). Maas identified Globba

subsessilis as an earlier synonym of Costus warmingii and replaced the species name with

Costus subsessilis (Nees & Mart.) Maas (1976: 469). Subsequently, Specht and Stevenson

(2006), when describing the new genus Chamaecostus, proposed a new combination

resulting in Chamaecostus subsessilis, a lineage that has henceforth been treated as the

Chamaecostus subsessilis complex.

The C. subsessilis species complex has been considered (Specht 2006, Maas 1972,

Schumann in Engler 1904) to be closely related to Chamaecostus cuspidatus (Nees & Mart.)

Specht & Stevenson (2006: 158), as these taxa share significant morphological similarities

!!

! 94!

such as appendaged green bracts and sheaths that grow beyond the stem node, commonly

covering internodes entirely. Additionally, C. subsessilis and C. cuspidatus have adjacent

geographic distributions, with limited but existing range overlap in the Atlantic Rain Forest

and in the Cerrado transition, Eastern Brazil.

Maas (1972, 1977) comprehensively reviewed all Neotropical Costaceae, in which all

current members of the genus Chamaecostus were included as members of Costus

subgenus Cadalvena (Fenzl) K.Schum. in Engler (1904: 381). The first investigation of the

phylogenetic relationships within Costaceae was presented by Specht et al. (2001), using a

dataset of both morphological and molecular characters. Phylogenetic relationships revealed

members of Maas’ new world Costus subgenus Cadalvena group to be monophyletic, but

more closely related to Dimerocostus and Monocostus than to other Costus lineages,

rendering Costus paraphyletic (Specht et al. 2001; Specht 2006). This analysis also

supported the position of the Cadalvena type species, Costus spectabilis (Fenzl) Schumann

(1892: 422), with other lineages of African taxa within a clade of Costus. Therefore, Specht &

Stevenson (2006) formally placed Maas’ Cadalvena members in the genus Chamaecostus,

with the etymology (chamae-) being indicative of their small stature (≤1 m) relative to plants

remaining in the genus Costus. Together with Monocostus Schumann (1904: 427) and

Dimerocostus Kuntze (1891: 687), these three genera form an early-diverging clade of

approximately 17 species with a distribution encompassing Central and South America

(“South American Clade”, Specht 2006). Morphologically, Chamaecostus, Monocostus and

Dimerocostus share cup-shaped stigmas; tubular and bicarinate bracteoles; presence of

unicellular hairs; and a general flower morphology with long and narrow labellum (fused

petaloid staminodes) opening into a wide and distinct limb. Additionally, Monocostus and

Dimerocostus share a bilocular ovary, while the Chamaecostus ovary is trilocular similar to

that found in Costus L.

As their name implies, Chamaecostus are low plants, even occasionally emerging as

acaulescent rosettes, typically not exceeding 1 m in height and with stems commonly less

!!

! 95!

than 1 cm in diameter. Specht and Stevenson (2006) cite the following synapomorphies for

identification of Chamaecostus: small stature; cup-shaped stigma; open labellum; ovary and

tube of labellum red-brown punctate. Also noteworthy are their very fragile shoots and nodes

that are commonly purplish and lightly geniculate. Additionally, intermittency of aerial shoots

during dry season and presence of subterraneous reserve organs are also very common.

The staminodial labellum is large, ovate at the apex, yellow, orange, red, or white.

Distribution of Chamaecostus is restricted to South America, from the Guyana Shield to the

Amazonian lowlands of Bolivia and Brazil, the western edge of the Brazilian shield and the

Brazilian Atlantic Rainforest. In addition, Chamaecostus species seem to show an

aggregated distribution, with high local abundance but naturally rare occurrence along the

landscape. The genus currently consists of seven known species, all endemic to South

America. Geographic distribution is varied and includes seasonally dry forests of Southwest

Amazonia and Cerrado forest ecosystems of Central Brazil (Chamaecostus subsessilis);

Central Atlantic Forest (Chamaecostus cuspidatus); Amazonian (Chamaecostus fusiformis

(Maas) Specht & Stevenson (2006: 158), Chamaecostus fragilis (Maas) Specht & Stevenson

(2006: 158), Chamaecostus lanceolatus (Petersen) Specht & Stevenson (2006: 158)); and

endemic to the Guyana shield (Chamaecostus congestiflorus (Rich. ex L. F. Gagnep.)

Specht & Stevenson (2006: 158), Chamaecostus curcumoides (Maas) Specht & Stevenson

(2006: 158)).

!!

! 96!

4.2. Material and Methods

4.2.1. Phylogenetic relationships

We analyzed the phylogenetic relationships of species within Chamaecostus, using

Maximum Likelihood (ML; PhyML, Guidon & Gascuel 2003) and Bayesian (MrBayes,

Huelsenbeck & Ronquist 2001) approaches based on a combined dataset of nuclear (ETS,

Kay et al. 2005; ITS, White et al. 1990; rpb2, Specht et al. 2001; CaM, Salzman et al. in

press) and plastid (rps16-trnK, Shaw et al. 2007; petG-trnP, Hwang et al. 2000; tnrL-trnLF,

Taberlet et al. 1991) sequences. All sequences were deposited in GenBank (accession

numbers in appendix 1).

We included individuals from all known species within the genus, and designated

samples of Monocostus uniflorus, two species of Dimerocostus and two species of Costus

as outgroups. For the Chamaecostus subsessilis complex, we analyzed samples from 12

populations collected across its distributional range. Total genomic DNA was isolated from

silica-gel dried leaf tissue using CTAB extraction protocol (Doyle & Doyle 1990). PCR

fragments of the molecular markers above were generated using Phire Hot Start II DNA

Polymerase (Thermo Scientific) with a 3 min. initial denaturing step at 98 °C, 45 cycles of 5

sec. at 98 °C, 15 sec. at gene-specific annealing temperatures, and 20 sec. at 72 °C, with a

final 1 min. 72 °C extension. Cycle sequencing was performed using BigDye v3.1 (Applied

Biosystems) following manufacture’s protocols. Cycle sequencing products were sequenced

on an Applied Biosystems 3730 DNA Analyzer automated DNA sequencer, at UC Berkeley’s

Evolutionary Genetics Laboratory.

We aligned each marker using the MUSCLE algorithm (Edgar 2004) implemented in

Geneious version 6.1.7 (www.geneious.com), and subsequently checked the multiple

sequence alignments manually. Sequence data was partitioned to allow different models of

sequence evolution for each region, with the best model for sequence evolution determined

with jModelTest version 2 (Darriba et al. 2012).

!!

! 97!

Maximum Likelihood (ML) analyses were run with a total of 1,000 bootstrap replicates

to assess statistical clade support. Bayesian analyses were run twice for 50x106

generations, sampling every 10,000 generations. Convergence was assessed via a low

(<0.01) average standard deviation in split frequencies after the first 25% of the sampled

data were discarded as burn-in.

4.2.2. Morphometrics of the Chamaecostus subsessilis complex

We measured leaves of specimens distributed across the range from the

Chamaeostus subsessilis complex (n=134) and from the closely related species

Chamaecostus cuspidatus (n=14) deposited in herbarium collections (A; HUFU; IAN; IBGE;

INPA; MG; MO; NY; R; RB; TANG; UB; UC; UFG - acronyms follow Thiers 2014) for the

assessment of the following morphometric variables: leaf length [LL], leaf maximum width

[LW], apex angle [AA], base angle [BA], leaf elliptical area [LA=�*((LL/2)*LW], leaf area

eccentricity [LE=LA/(LL/2)], leaf length-width proportion [LL/LW], and leaf apex-base

symmetry [LS=AA/BA] (Figure 1). We focused on leaf quantitative variation since vegetative

characters are meaningfully variable within this group, while floral traits are reasonably

constant. Only specimens with well-developed and properly pressed leaves were considered

in the analysis. We also analyzed type specimens of Chamaecostus subsessilis synonyms

(BM, K, MO, P) to review species circumscriptions.

Herbarium specimens of the Chamaecostus subsessilis complex were assigned to

resolved phylogenetic clades based on their recorded geographic location: each herbarium

specimen was assigned to the clade that contained individuals with the closest geographic

proximity. Analysis of variance (ANOVA) and two-sample t-test were performed for each trait

to detect differences between phylogenetic lineages given assignment to clades. Statistical

analyses were computed in R framework (R Development Core Team 2014).

!!

! 98!

Figure 1. Schematic representation of measured morphometric variables; LL – Leaf Length,

LW – Leaf Maximum Width, AA – Apex Angle, BA – Base Angle.

A A

B A

L W

L L

!!

! 99!

4.3. Results

4.3.1. Chamaecostus systematics and phylogenetic relationships

The concatenated multiple sequence alignment is 3,480 base pairs long and the

resulting topology (Figure 2A) is recovered in both Maximum Likelihood (ML) and Bayesian

Inference analyses. The phylogeny is well supported overall, as shown by high node

confidence provided by both Bayesian posterior probabilities and ML bootstrap replicates,

with the exception of the position of Chamaecostus fragilis. In this analysis, Chamaecostus

subsessilis was found to be paraphyletic. Two well-supported major clades are recovered

(Figure 2A; subsessilis and acaulis), one of which is sister to Chamaecostus cuspidatus.

These two clades correspond to previous species definitions and delimitations (see

discussion below), bearing high morphological variability and overlapping characters

between them.

!!

! 100!

(A)

(B)

Figure 2. A – Phylogenetic relationships within Chamaecostus (Costaceae). Support values

above branches are Bayesian Posterior Probabilities, while Most Likely Bootstrap

proportions from 1,000 bootstrap replicates are found below branches; B – Geographic

ranges of the Chamaecostus cuspidatus (blue), and Chamaecostus subsessilis complex:

!!

! 101!

subsessilis clade (orange) and acaulis clade (green). Dashed lines denote tentative range

limits, and the continuous grey line identifies the Araguaia River.

!!

! 102!


Analysis of variance and two-sample t-test detected significant morphometric

differences in three leaf traits between specimens assigned to either acaulis (n=83) or

subsessilis (n=51) clades: leaf length, leaf maximum width, and leaf area (Figure 3). With the

exception of apex-base symmetry, all other morphometric variables significantly separate

Chamaecostus cuspidatus from the other two clades (Table 1). In our molecular phylogeny,

the subsessilis clade is sister to C. cuspidatus, but subssessilis and acaulis clades are highly

similar in morphology, underscoring the significance of these quantitative differences.

!!

! 103!

Table 1. Morphometric variables tested for diagnose between Chamaecostus cuspidatus and

the Chamaecostus subsessilis complex. Means ± standard deviations; different letters are

indicative of statistical significance (p<0.05; t-test), and F-values and probabilities of ANOVA

are given. Bold values refer to variables that are significantly different between all three

species.

Leaf variables

Chamaecostus cuspidatus

Chamaecostus subsessilis

s.str.

Chamaecostus acaulis

ANOVA

(n=14) (n=51) (n=83) F p

Length (cm) 14.9 ± 4.5a 18.6 ± 6.2 b 22.3 ± 5.6 c 13.2 0.000

Maximum Width (cm) 4.5 ± 1.5 a 7.4 ± 2.6 b 8.6 ± 2.1 c 21.2 0.000

Length-Width Proportion

3.4 ± 0.6 a 2.6 ± 0.7 b 2.6 ± 0.4 b 14.5 0.000

Area (cm2) 113.4 ± 87.5 a 235.7 ± 152.9 b 316.0 ± 149.8 c 13.5 0.000

Eccentricity 0.6 ± 0.1 a 0.8 ± 0.2 b 0.8 ± 0.1 b 12.1 0.000

Apex Angle 84.4 ± 37.0 a 44.5 ± 17.4 b 78.6 ± 25.8 b 10.2 0.000

Base Angle 30.4 ± 6.5 a 53.4 ± 18.6 b 52.0 ± 14.3 b 13.1 0.000

Apex-Base Symmetry 1.5 ± 0.6 a 1.7 ± 0.9 a 1.6 ± 0.5 a 0.7 0.508

!!

! 104!

Figure 3. Box and whisker plots of three significantly different morphometric variables

between Chamaecostus cuspidatus (n=14), Chamaecostus subsessilis s.str. (n=51) and

Chamaecostus acaulis comb. nov. (n=83), showing means, quartiles and ranges. A – Leaf

length (cm); B – Leaf Maximum Width (cm); C – Leaf Area (cm2).

!!

! 105!

4.4. Discussion

4.4.1. Chamaecostus systematics and phylogenetic relationships

Here we present a robust molecular phylogeny of Chamaecostus (Figure 2A), from

which novel evolutionary inferences can be inferred. Within Chamaecostus, major clades

generally reflect relationships previously suggested from taxonomic studies (Schumann in

Engler 1904, Maas 1972). A close relationship between Chamaecostus curcumoides and

Chamaecostus fusiformis was suggested by Maas (1972) when he described both species.

Indeed, the well-supported clade formed by these two species is supported morphologically

by various synapomorphies of these species, such as a more complex capitate

inflorescence, ovate-triangular yellow bracts, and a strongly tubular labellum. Likewise,

green, appendaged bracts and a full-grown sheath are synapomorphies of Chamaecostus

cuspidatus and Chamaecostus subsessilis, a species pair whose close taxonomic

relationship had been previously hypothesized (Nees & Martius 1823, Petersen in Martius

1890, Schumann in Engler 1904, Maas 1972). Perhaps the most remarkable implication of

this Chamaecostus phylogeny (Figure 2A) is the paraphyly of Chamaecostus subsessilis,

indicating the need for a revision of the genus.

Our phylogeny suggests some new interpretations and indicates alternative

evolutionary scenarios with respect to biogeography and morphology. Amazonian

distribution is most likely ancestral for the genus, with the appearance in Southern and

Eastern portions of South America being more derived; only C. cuspidatus has a distribution

lying completely outside of the Amazonian domain. Also noteworthy is the position of C.

congestiflorus as sister to the remaining species in the genus; this species has white flowers,

with a conspicuous fimbriate labellum, while all other species have either a yellow, orange,

red or pink glabrous labellum (at least when fully developed and opened), suggesting that a

shift from white flowers and a reduction of the labellum margin complexity evolved only once

in the genus leading to a radiation of colorful-flowered species with a glabrous labellum.

!!

! 106!


Redefining species limits in the Chamaecostus subsessilis complex can be rather

problematic, particularly because of the limited extent of morphological diagnostic

characters, especially when considering morphological differences from every other species

in the genus, which can be exceptionally divergent. Indeed, distinct species criteria describe

different stages in the divergence of lineages, and differences among the many species

concepts are at least partly attributable to the complex and temporally extended nature of

speciation (de Queiroz 2007), and therefore no single definition will be appropriate for all

organisms. Morphological similarity has the disadvantage of using arbitrary determinations of

the threshold of differences (de Queiroz 1998) and is especially challenging to apply when

quantitative character continua are significant. Such is indicated to be the case for the C.

subsessilis complex, where character variation among and between populations was thought

to be solely due to environment (Maas 1972,1977). Nevertheless, determining these two

distinct C. subsessilis lineages to be a single species in face of the strong population genetic

structure would be unfitting and would not reflect the evolutionary differentiation among

sampled populations condition.

Results do show significant morphometric size differences between the two lineages

(Table 1, Figure 3). Additionally, a posteriori interpretation of the descriptions of

Chamaecostus subsessilis synonyms from the historic record also reveals traits that are

useful to separate these lineages. Petersen (in Martius’ Flora Brasiliensis 1890) describes a

Costus warmingii as being over 1 m tall with elliptic to obovate-elliptic leaves, variable in

size, abaxially densely hirsute and adaxially glabrous. Later, Schumann (in Engler 1904)

mentioned other individuals under the synonyms Costus gagnepainii K.Schum. in Engler

(1904: 420), Costus latifolius Gagnepain (1902: 100), Costus paucifolius Gagnepain (1902:

100), Costus pumilus Petersen in Martius (1890: 58), and Costus rosulifer Gagnepain (1902:

!!

! 107!

101), as having either both or at least one leaf side glabrous. All of the types for the

synonyms occur East of the Araguaia River valley. Moore (1895: 480) described Costus

acaulis (=Chamaecostus subsessilis) as having oblong-obovate puberulous leaves based on

a type from Mato Grosso, in the upper Paraguay River basin, Brazil. This specimen greatly

reassembles Loesener’s (1929) descriptions and types for Costus steinbachii Loesener

(1929: 714) and Costus kaempferoides Loesener (1929: 714), from Bolivia and Peru,

respectively. Interestingly, Specht (2006) noted that although indument characters in general

tend to be homoplasious in Costaceae, certain aspects of the indument do help to define

some lineages.

Hence, external morphology, anatomy, ecology, life history and reproductive biology

should be further investigated in a finer and more detailed fashion to help accumulate

diagnostic features for Chamaecostus species, with a specific focus on detecting

synapomorphies between C. cuspidatus and C. subsessilis s.str. (subsessilis clade in Figure

2A), since compilation of relevant information often leads to an improved comprehension of

boundaries between species (e.g. Wendt et al. 2011, Faria et al. 2010). One or only a few

individuals may not be representative of the species as a whole, especially for taxa with

widespread distributions (Goldstein et al. 2000; Walsh 2000). An integrative approach,

combining population genetics, historical biogeography, and environmental data, could be of

great help to elucidate the speciation scenario and demographic history involved. Our

inspection of several individuals from multiple localities across the range of the species

provides discernible differences between clades, highlighting the particular efficacy of

population sampling to fully determine species integrity. Thus, we encourage broad

geographic and genetic sampling to investigate the possibility of cryptic evolutionary

lineages in species complexes.

Species polyphyly can result as an artifact of phylogenetic reconstruction from weak

phylogenetic signal or incomplete lineage sorting, or from taxonomically underestimating or

overestimating genetic exchange among individuals and populations (Funk & Omland 2003).

!!

! 108!

Furthermore, the use of morphological or genetic markers alone could mislead our

understanding of evolutionary history; phylogeographic breaks can form within a

continuously distributed species even when there are no barriers to gene flow if the average

individual dispersal distance and local population size are small (Irwin 2002).

Correspondingly, morphometric data commonly convey ontogenetically or ecologically

governed plasticity and could obscure genetically governed morphological variation relevant

to taxonomic decisions (Tetsana et al. 2014). Our comprehensive approach, combining

phylogenetic analysis of multiple molecular markers and leaf morphometrics, objectively

reveals the recognition of at least two species within the Chamaecostus subsessilis complex

as necessary to appropriately reflect evolutionary relationships, and we formally

acknowledge distinct names below.

!!

! 109!

4.5. Taxonomy

Chamaecostus acaulis (S.Moore) T.André & C.D.Specht comb. nov. = Costus acaulis

S.Moore. Transactions of the Linnean Society of London 4: 480, pl. 33, f. 1–5. 1895; Type:

—BRAZIL. Mato Grosso: Santa Cruz (i.e. Barra do Bugres), November 1891, Spencer

Moore 679 (holotype, BM!).

= Costus steinbachii Leoesener, Notizbl. Bot. Gart. Berl. 10: 714. 1929; Loesener in Engler

& Prantl, Nat. Pflanzenfam. ed. 2. 15A: 634. 1930. Type: — BOLIVIA. Sara: Santa Cruz, 31

December 1925, Steinbach 7386 (lectotype, F; isolectotypes, K!, MO!). Lectotype assigned

by Maas (1972) since the holotype was destroyed at Berlin in 1943.

= Costus kaempferoides Leoesener, Notizbl. Bot. Gart. Berl. 10: 714. 1929. Type: — PERU.

Madre de Dios: Seringal São Francisco, September 1911, Ule 9197 (lectotype, K!).

Lectotype assigned by Maas (1972) since the holotype was destroyed at Berlin in 1943.

Acaulescent or very low plants with stems up to 30 cm long and 1-10 mm wide; Internode

1.0—8.5 (3.5 ± 1.6) cm long; Roots fleshy, with tubers; root tubers fusiform to ellipsoid;

Sheaths membranaceous, 1.5—3.0 (1.6 ± 0.6) cm long, 1.5—8.5 (4.5 ± 1.7) cm wide, obtuse

at the apex, puberulous. Ligule 1 mm long. Leaves (4—6), rosulate, elongate, oblong-

obovate, 22.3 ± 5.6 cm long, 8.6 ± 2.1 cm wide, densely strigose or densely to sparsely

puberulous on both sides, cuneate at the base, shortly acuminate at the apex, apex up to

2.0 cm (0.6 ± 0.4), margins densely ciliate; Inflorescence compact, terminal and short; bracts

herbaceous, green to 30—70 (50 ± 30) mm long, 5—15 (10 ± 5) mm wide, densely

puberulous; appendages foliaceous, green, narrowly triangular to deltate, mucronate at the

apex, densely puberulous. Bracteole membranaceous, tubular, 2—3 cm long. Calyx tubular,

membranaceous to herbaceous, 10—40 mm long, lobes narrowly triangular, mucronate, 1—

15 mm long. Corolla white, 50—70 mm long, tube 25—30 mm long, lobes narrowly elliptic,

mucronate, 30—40 mm long, 6—12 mm wide. Labellum yellow, with white, yellow or orange

nectar guides at the middle, broadly obovate, 60—70 mm long, 70—95 mm wide, margins

!!

! 110!

undulate, lightly fimbriate or glabrous. Stamen yellow, oblong-oblanceolate, Stamen 20—50

mm long, up to 20 mm wide, apex obtuse, irregularly lobed or acuminate, anther attached at

the base, caudate at the apex. Style filiform, glabrous; Stigma cup-shaped; Fruits and seeds

not analyzed.

Chamaecostus acaulis strongly resembles Chamaecostus subsessilis sensu stricto,

but differs by developing shorter habit, bigger oblong-obovate leaves (22.3 ± 5.6 cm long x

8.6 ± 2.1 cm wide), and by possessing puberulous leaves. Chamaecostus subsessilis sensu

stricto have adaxially strigose to glabrous leaves, more elliptical and smaller leaves (18.6 ±

6.2 cm long x 7.4 ± 2.6 cm wide), and variable height, from 0.3 to over 1.0 m.

Since there is substantial overlap in characters between the two (Figure 3, Table 1),

we strongly suggest that location should be taken into account when identifying both

species, in particular because of the strong geographic structure resolved between the

populations analyzed here. Chamaecostus acaulis occurs West of the Araguaia River valley,

through Peruvian, Bolivian and Brazilian South Amazonia, and in Western and Southern

portions of the Central Brazilian Shield, while Chamaecostus subsessilis occurs East from

the Araguaia River valley, and within most of the Central Brazilian Shield and in transition

forests between Cerrado and Central Atlantic Rain Forest (Figure 2B). However, potentially

sympatric populations may occur within the Araguaia River valley, and further detailed

analyses of the populations in this transition zone are necessary.

!!

! 111!

Figure 4. Chamaecostus acaulis comb. nov. and Chamaecostus subsessilis s.str.. (B) photo

by W.W.Thomas. (D) photo by D.Skinner.

!!

! 112!

Measured specimens

Chamaecostus acaulis. BRAZIL. ACRE. Porto Acre: Lowrie 526 (INPA, MG, NY); Rio

Branco: Albuquerque 1389 (NY), Cid 2806 (INPA, NY), Ehrich 8 (NY), Lowrie 445 (INPA, R);

Santa Rosa: Daly 11103 (NY); Sena Madureira: Daly 7855 (NY), Daly 7860 (NY), Prance

7623 (MO), Ramos 682 (INPA); Tarauacá: Prance 7416 (INPA, NY); Xapuri: Alves 2365

(NY), Daly 7195 (MO, NY). GOIÁS. Cabeceiras: Irwin 10358 (MO); Caipônia: Prance 59644

(A, NY); Cidade de Goiás: Kirkbridge 3399 (UB); Santa Rita do Araguaia: Rocha sn (UB).

MATO GROSSO. Alta Floresta: André 804 (RB), Richter 39 (RB); Aripuanã: Berg 18519

(INPA, NY); Cuiabá: Andersson 1623 (A, UB); Barra do Garças: Philcox 4000 (UB);

Palmeiras: Lindman 2485 (A); Poconé: Maciel 144 (INPA); Santa Teresinha: Oliveira 3094

(RB), Thomas 4384 (INPA, MG). Tangará da Serra: Silva 498 (TANG). MINAS GERAIS.

Ituitaba: Macedo 1316 (RB), Macedo 1993 (MO); Uberlândia: Arantes 1124 (HUFU),

Barbosa 281 (HUFU). PARÁ. Altamira: Balée 1996 (NY), Dias 1109 (MG), Ferreira 1068

(NY), Lima 6020 (RB), Nascimento 1177 (NY), Souza 1068 (MG), Souza 1177 (MG);

Conceição do Araguaia: Plowman 8448 (A, INPA, MG, MO, NY); Nova Canaã dos Carajás:

Lobato 2596 (MG); Redenção: Cordeiro 2851 (IAN); Serra do Cachimbo: Prance 25218

(MG). RONDÔNIA. Abunã: Prance 8338 (INPA, MG, NY); Ariquemes: Zarucchi 2677 (A,

INPA, MG, MO, NY, R, RB), Mota 440 (NY), Vieira 440 (MO, R), Vieira 443 (INPA);

Mutumparaná: Prance 8975 (A, INPA, MG, NY, R). BOLIVIA. Bela Vista: Steinbach 7386

(MO); Beni: Maas 8660 (MO, NY, RB), Surubi 313 (NY), Rusby 1399 (A, NY), Ledezma 893

(MO); Ñuflo de Chaves: Ortiz 39 (NY); Santa Cruz: Arroyo sn (MO), Carrión 506 (MO),

Castro 60 (MO), Garvizu 513 (MO), Guillén 3035 (MO), Guillén 3623 (MO), Killeen 7168

(MO), Mamani 1091 (MO), Quevedo 2473 (MO), Rodriguez 572 (MO).

Chamaecostus subsessilis s.str.. BRAZIL. BAHIA. Itamaraju: Mori 10753 (NY, RB); Jussari:

Belém 2274 (UB), Thomas 11937 (MO), Thomas 13401 (MO, NY). DISTRITO FEDERAL.

Brasilia: Barroso 639 (RB), Heringer 10751 (IAN, UB), Irwin 19440 (NY, RB, UB), Pereira

2251 (IBGE, RB), Pires 51 (RB), Pires 57147 (UB). GOIÁS. Alto Paraíso: Felfili 379 (IBGE),

!!

! 113!

Mendonça 2898 (IBGE); Alvorada do Norte: Hatschbach 39011 (NY, UC); Caldas Novas:

Vieira 1652 (RB); Catalão Hatschbach 55820 (MO); Cocalzinho: Mendonça 2206 (IBGE);

Corumbá: Maguire 57147 (MG, MO); Formosa: Irwin 9066 (UB); Goiânia: Rizzo 2566 (UFG),

Rizzo 12305 (UFG); Inhumas: Rizzo 2779 (UFG); Luziânia: Coradin 7397 (RB); Monte

Alegre: Mendonça 4512 (RB); Mossâmedes: Forzza 2500 (RB); Niquelândia: Cordovil 106

(RB), Fonseca 1245 (IBGE, UFG); Nova Roma: Forzza 2541 (RB); São Domingo: Oliveira

1117 (IBGE), Santos 2367 (RB); Trindade: Rizzo 3113 (UFG). MARANHÃO. Tuntum:

Santos 707 (MG, NY). MINAS GERAIS. Abre Caminho: Pereira 59 (RB); Jacinto: Leitman 51

(RB); Januária: Filgueiras 1950 (IBGE), Ratter 2630 (IAN, NY, UB, RB), Ratter 6410 (IBGE);

Lagoa Santa: Hoehne 6206 (R); Minas: Duarte sn (RB); Serra do Cipó: Heringer 7343 (UB);

Unaí: Brina sn (RB); Várzea da Palma: Duarte 7547 (RB). TOCANTINS. Aurora do Norte:

Pereira 2009 (IBGE); Lajeado: Árbocz 6293 (IBGE); Presidente Kennedy: Plowman sn

(INPA).

Chamaecostus cuspidatus. BRAZIL. BAHIA. Belmonte: Mattos 368 (NY), Mattos 1804 (NY),

Santos 828 (RB); Eunápolis: Santos 893 (NY, RB), Mello Filho 2980 (R); Gandú: Santos

1157 (NY, UB); Porto Seguro: Duarte 5668 (RB), Pinheiro 1747 (RB); Wanceslau

Guimarães: Thomas 9329 (MO). ESPÍRITO SANTO. Colatina: Kuhlmann 6660 (RB); Santa

Teresa: Boone 984 (MO).

!!

! 114!

4.6. Acknowledgments

We are thankful to visited herbaria curators and staff. This paper is part of the D.Sc.

requirements of TA at the Biodiversity and Evolutionary Biology Graduate Program of the

Federal University of Rio de Janeiro. TA received a scholarship grant from Coordenação de

Aperfeiçoamento de Pessoal de Nível Superior. TW received a productivity grant from

Conselho Nacional de Desenvolvimento Científico e Tecnológico. CS received support from

the US National Geographic Society (CRE Grant #8994-11) that helped support this

research.

4.7. References

Darriba D, Taboada GL, Doallo R, Posada D. 2012. jModelTest 2: more models, new

heuristics and parallel computing. Nature Methods 9: 772.

de Queiroz K. 1998. The general lineage concept of species, species criteria, and the

process of speciation: A conceptual unification and terminological recommendations.

Pages 57–75 in Endless forms: Species and speciation (D. J. Howard, and S. H.

Berlocher, eds.). Oxford University Press, New York.

de Queiroz K. 2007. Species concepts and species delimitation. Systematic Biology 56: 879-

886.

DeSalle R, Egan M, Sidall M. 2005. The unholy trinity: taxonomy, species delimitation and

DNA barcoding. Philos. Trans. R. Soc. B 360: 1905-1916.


Focus 12:13-15.

!!

! 115!


throughput. Nucleic Acids Research 32:1792–1797.

Engler HGA. (Ed). 1904. Pflanzenreich IV. 46 (Heft 20).

Faria APG, Wendt T, Brown GK. 2010. A revision of Aechmea subgenus Macrochordion

(Bromeliaceae) based on phenetic analyses of morphological variation. Botanical Journal

of the Linnean Society 162: 1-27.

Florio AM, Ingram CM, Rakotondravony HA, Louis Jr EE, Raxworthy CJ. 2012. Detecting

cryptic speciation in the widespread and morphologically conservative carpet chameleon

(Furcifer lateralis) of Madagascar. J. Evol. Biol. 25: 1399-1414.

Frost DR, Kluge AG. 1994. A consideration of epistemology in systematic biology, with

special reference to species. Cladistics 10: 259-294.

Funk DJ, Omland KE. 2003. Species-level paraphyly and polyphyly: frequency, causes, and

consequences, with insights from animal mitochondrial DNA. Annu. Rev. Ecol. Evol.

Syst. 34: 397-423.

Gagnepain F. 1902. Bulletin de la Société Botanique de France. Series 4. Volume 2: 102,

103.

Goldstein PZ, DeSalle R. 2000. Phylogenetic species, nested hierarchies, and character

fixation. Cladistics 16: 364-384.

Goldstein PZ, DeSalle R, Amato G, Vogler AP. 2000. Conservation genetics and the species

boundary. Conserv. Biol. 14: 120-131.

Guindon S, Gascuel O. 2003. A simple, fast, and accurate algorithm to estimate large

phylogenies by maximum likelihood. Systematic Biology 52:696-704.

Hey J, Pinho C. 2012. Population genetics and objectivity in species diagnosis. Evolution 66:

1413-1429.

!!

! 116!

Huelsenbeck JP, Ronquist F. 2001. MRBAYES: Bayesian inference of phylogenetic trees.

Bioinformatics 17:754-755.

Hwang LH, Hwang SY, Lin TP. 2000. Low chloroplast DNA variation and population

differentiation of Chamaecyparis formosensis and Chamaecyparis taiwanensis. Taiwan

Journal of Forest Science 15:229-236.

Irwin DE. 2002. Phylogeographic breaks without geographic barriers to gene flow. Evolution

56: 2383-2394.


evolution of hummingbird pollination in neotropical Costus subgenus Costus

(Costaceae): evidence from nrDNA ITS and ETS sequences. American Journal of

Botany 92:1899–1910.

Kuntze CEO. 1891. Revisio Generum Plantarum. Volume 2: 687.

Linnaeus C. 1753. Species Plantarum. Volume 1: 1-560.

Loesener LET.1929. Notizbl. Bot. Gart. Berl. 10: 714.



Maas PJM. 1976. Notes on New World Zingiberaceae. Acta Botanica Neerlandica 24: 469.



Martius CFP. (Ed.). 1890. Flora Brasiliensis. Volume III. Pars III: 1-128.

Moore S. 1895. Trans. Linn. Soc. series 2. 4: 480.

Nees CGD, Martius CFP. 1823. Nova Acta Phys.-Med. Acad. Caes. Leop.-Carol. Nat. Cur.

Volume

!!

! 117!

Nixon KC, Wheeler QD. 1990. An amplification of the phylogenetic species concept.

Cladistics 6: 211-223.

Padial J, De La Riva I. 2009. Integrative taxonomy reveals cryptic Amazonian species of

Pristimantis (Anura: Strabomantidae). Zool. J. Linn. Soc. 155: 97-122.

Padial J, Miralles A, De la Riva I, Vences M. 2010. The integrative future of taxonomy. Front.

Zool. 7: 16.

R Development Core Team 2014. R: A Language and Environment for Statistical

Computing. R Foundation for Statistical Computing. http://www.R-project.org.

Rieseberg LH, Church SA, Morjan CL. 2004. Integration of populations and differentiation of

species. New Phytol. 161: 59-69.

Sáez AG, Lozano E. 2005. Body doubles. Nature 433:111.

Salzman S, Driscoll HE, Renner T, André T, Shen S, Specht CD. in press. Spiraling into

history: A molecular phylogeny and investigation of biogeographic origins and floral

evolution for the genus Costus L. Systematic Botany.

Schumann K. 1892. Botanische Jahrbücher für Systematik, Pflanzengeschichte und

Pflanzengeographie 15: 422.

Shaffer HB, Thomson RC. 2007. Delimiting Species in Recent Radiations. Syst. Biol. 56:

896–906.


sequences to choose noncoding regions for phylogenetic studies in Angiosperms: the

tortoise and the hare III. American Journal of Botany 94:275–288.

Specht CD, Kress WJ, Stevenson DW, DeSalle R. 2001. A Molecular Phylogeny of

Costaceae (Zingiberales). Molecular Phylogenetics and Evolution 21:333-345.

!!

! 118!

Specht CD, Stevenson DW. 2006. A new phylogeny-based generic classification of

Costaceae (Zingiberales). Taxon 55:153-163.

Specht CD. 2006. Systematics and Evolution of the Tropical Monocot Family Costaceae

(Zingiberales): A Multiple Dataset Approach. Systematic Botany 31: 89-106.

Taberlet P, Gielly L, Pautou G, Bouvet J. 1991. Universal primers for amplification of three

non-coding regions of chloroplast DNA. Plant Molecular Biology 17:1105–1109.

Tetsana N, Pedersen HÆ, Sridith K. 2014. Character intercorrelation and the potential role

of phenotypic plasticity in orchids: a case study of the epiphyte Liparis resupinata. Plant

Syst Evol 300:517–526

Thiers B. 2014. Index Herbariorum: A global directory of public herbaria and associated staff.

New York Botanical Garden's Virtual Herbarium. http://sweetgum.nybg.org/ih/

Walsh P. 2000 Sample size for the diagnosis of conservationunits. Conserv. Biol. 14: 1533-

1535.

Wendt, T, da Cruz DD, Demuner VG, Guilherme FAG, Boudet-Fernandes, H. 2011. An

evaluation of the species boundaries of two putative taxonomic entities of Euterpe

(Arecaceae) based on reproductive and morphological features. Flora 206: 144-150.

White TJ, Bruns T, Lee S, Taylor J. 1990. Amplification and direct sequencing of fungal

ribosomal RNA genes for phylogenetics. In “PCR Protocols: A Guide to Methods and

Applications,” Academic Press, San Diego.

!!

! 119!

!!

! 120!

Chapter 5 – Speciation in the South American dry understory: lessons from

Chamaecostus (Costaceae, Zingiberales)

Thiago André1,2, Chodon Sass3, Roxana Yockteng3, Tânia Wendt2, Chelsea Specht3,

Clarisse Palma-Silva4

1 – [email protected]; corresponding author.

2 – Universidade Federal do Rio de Janeiro, Rio de Janeiro (RJ), Brazil.

3 – University of California, Berkeley (CA), USA.

4 – Universidade Estadual Paulista, Rio Claro (SP), Brazil.

!!

! 121!

Abstract

Central South America is largely comprised of seasonally dry ecoregions, where a sharp

rainy season largely defines growth patterns. Here we examine speciation of the understory

herbs Chamaecostus subsessilis and C. acaulis, inhabitants of South American seasonally

dry forests of Cerrado and South Amazonia. We studied 20 kb of sequence data from

targeted hightroughput sequencing by capture using PCR-generated probes so as to

estimate the phylogeographic pattern and timing of diversification in this Chamaecostus

lineage. We found deep phylogeographic structure at both the population and species level,

as inferred from fixation indices, structure analysis and phylogenetics. Spatiotemporal

patterns of diversification implies that divergence between Eastern and Western lineages

unlikely predates the formation of the Araguaia-Tocantins River system but it may relate to

the uplift of the central Brazilian plateau and the intense shifts in hidrography that likely

occurred during the Miocene due to ocean level variation and Andean orogeny. We also

found a strong supported phylogenetic topological incongruence between plastid and

nuclear genomes. New high-throughput sequencing technologies based on range-wide

sampling of populations allowed us to uncover speciation outlines in a Central South

American lineage adapted to dry ecoregions.

!!

! 122!

5.1. Introduction

South America encompasses exceptional biological diversification (Turchetto-Zolet et

al. 2013) and a complex geomorphological history (Hoorn et al. 2010). Central South

America is largely comprised of seasonally dry ecoregions, mostly by Cerrado savannas and

dry forest patches (Pennington et al. 2009). This biome displays high environmental

heterogeneity due to extremely complex geological history and heterogeneous edaphic and

climatic conditions (Silva et al. 2011). Cerrado landscapes are dominated by vast plateaus

that uplifted from Late Tertiary to Early Quaternary, followed by recent Quaternary cycles of

erosion (Cole 1986, Silva & Bates 2002). Additionally, climatic oscillations from late Tertiary

throughout the Pleistocene lead to alternate contractions and expansions of forests and

savannas on this landscape (Behling & Hooghiemstra 2001). Both Tertiary geological events

and Pleistocene climatic fluctuations were discovered to have promoted divergences among

organism lineages in this region (Turchetto-Zolet et al. 2013). Phylogeography of Cerrado

tree species (Caryocar brasiliense by Collevatti et al. 2003, and Hymenaea stigonocarpa by

Ramos et al. 2007) recovered three genetically differentiated groups essentially congruent

with the Cerrado’s main plateaus. Gallery forests are believed to have maintained the

connectivity between the central Brazil Plateau and adjacent forest formations, i.e. Amazonia

and Atlantic Rainforest, even during the dry and cold climatic periods, offering refuge to a

stable forest biota (Brown & Gifford 2002, Costa 2003, Werneck 2011). Consequently,

savanna formations are scattered with forests and xerophytic communities (Felfili et al. 2001,

Silva et al. 2006). A rainy season, during summer, largely defines growth patterns and

survival for plant species, thus making the Cerrado domain a major barrier to gene flow

between South American rainforest taxa (Prado & Gibbs 1993, Turchetto-Zolet et al. 2012),

and Cerrado itself is not a barrier-free habitat (Silva & Bates 2002, Werneck et al. 2011,

Prado et al. 2012).

!!

! 123!

Chamaecostus subsessilis (Nees & Mart.) C.D.Specht & D.W.Stev. and C. acaulis

(S.Moore) T.André & C.D.Specht are perennial understory herb species from South

American seasonally dry forests. These species represents an ancient lineage of Costaceae

Nakai, a typical component of moist understory of Tropical Rainforests. Taxonomic status of

Chamaecostus subsesilis populations occurring from Bolivia to Atlantic rainforest, including

Southern Amazonia is controversial. Some authors have considered these populations as a

single species, or a species complex, namely C. subsessilis (Maas 1972, 1977). However,

C. acaulis was recently recovered from a synonym of C. subsessilis (André et al.

unpublished), with very limited morphological variation recognized between this two species,

despite a strongly supported paraphyly identified by a phylogenetic approach, involving the

relationship to the Atlantic Rainforest species C. cuspidatus (Nees & Mart.) C.D.Specht &

D.W.Stev. In this novel interpretation, C. subsessilis refers to populations found mainly East

from the Tocantins-Araguaia River Basin, while C. acaulis refers to those found West and

South of it. Populations of both species are commonly locally dense, while these dense

clusters are rare along the landscape. We know very little about reproductive biology of

these plants, but seeds in the family are thought to be ant-dispersed, mainly because of the

presence of an aril (Maas 1972). These species are the only Costaceae species clearly

adapted to climatic seasonality (Maas 1972, Lage et al. 2012).

Studies on population genetic structure are essential to comprehend speciation

processes and lineage divergence. Weak genetic structure is indicative of high rates of gene

flow and species cohesion, while marked structure points to low levels of gene flow and

independent evolutionary trajectories and a possible emerging speciation. In such context,

phylogeography particularly focus on spatial and temporal dimensions of genealogies, within

and among closely related species (Avise 2009) will greatly help to understand speciation

processes. Besides, phylogeographic evidence can distinguish between secondary contacts

of previously isolated lineages and the establishment of phenotypic or genetic clines within a

continuous population across a sharp environmental gradient (Endler 1982, Swenson &

!!

! 124!

Howard 2005, Whinnett et al. 2005, Moritz et al. 2009). Most processes involved in

speciation are not driven by any intrinsic mechanism that operates independently of

environment, and for a lineage to diverge genetic differentiation is necessary, but not

exclusive (Brown 1995). Ecological or geographic variations are widely important, and the

investigation of spatial arrangements among genetic lineages can reveal much of the

historical nature of intraspecific evolution (Avise et al. 1987). To address most questions in

phylogenetics and phylogeography we need a high number of orthologous loci from multiple

individuals of nonmodel species, representing between them a gradient of divergence

(McCormack et al. 2013, Lemmon & Lemmon 2013, Mamanova et al. 2010). The sequence

capture using PCR-generated probes (SCPP) protocol (Peñalba et al. 2014) generates

massive multilocus datasets of homologous loci from the genomic DNA of a nonmodel

organism by its reduction to a targeted genomic subset of interest, removing the need for a

reference genome. Targeting is achieved by hybridizing DNA libraries to short nucleic acid

probes, discarding fragments that did not hybridize, and sequencing the fragments that were

successfully captured. This approach allows to affordably sequencing multiple loci from for

several individuals using individual barcoding.

Here, we take advantage of cutting-edge next-generation sequencing methods to

undertake a range-wide analysis of speciation within this putative cryptic lineage of South

American dry understory plants. We present and examine patterns of nuclear and plastid

DNA diversity in C. subsessilis and C. acaulis. Specifically, our aims are: (1) to determine

the extent of genetic differentiation among populations of the current distribution range,

within and between species; and (2) to test if lineages diverged during Pleistocene. We

further discuss genetic structure in view of available information on the biogeographic history

of South American seasonally dry ecosystems.

!!

! 125!

5.2. Materials and Methods

5.2.1. Sampling and DNA extraction

We sampled leaf tissue of 8-10 individuals from 15 populations (summing 125

individuals) along Chamaecostus acaulis (9 populations) and Chamaecostus subsessilis (6

populations) geographic range (Figure 1, Table 1) and 10 outgroup taxa (Table 2). Genomic

DNA was extracted following the CTAB protocol (Doyle & Doyle 1990) at the Molecular

Biology Lab of the Botanical Institute of São Paulo.

!!

! 126!

Figure 1. Locations of the seventeen sampled sites of Chamaecostus subsessilis (MG, JA,

DF, TG, TO, PK) and C. acaulis (UB, PA, GO, MT, NX, AF, TS, CS, RO, PM, AC).

Topographic variation is shown in the background, with green lowlands.

! 127!

Table 1. Name and location of sampled Chamaecostus subsessilis and C. acaulis populations and respective number

of sequences obtained for plastid and nuclear markers.

Population name and location Species Number of sequences Latitude Longitude cp nr

PK - Presidente Kennedy, Tocantins, Brazil C. subsessilis 9 9 -8.4857 -48.5762 JA - Januária, Minas Gerais, Brazil C. subsessilis 8 8 -15.4320 -44.4162 MG - Rodeador, Minas Gerais, Brazil C. subsessilis 9 9 -18.2941 -44.0363 TG - Teresina de Goiás, Goiás, Brazil C. subsessilis 9 9 -13.6832 -47.2247 DF - Brasília, Distrito Federal, Brazil C. subsessilis 5 5 -15.7413 -47.8854 TO - Lajeado, Tocantins, Brazil C. subsessilis 8 8 -9.9542 -48.3426 UB - Uberlândia, Minas Gerais, Brazil C. acaulis 9 9 -19.1705 -48.3926 GO - Cidade de Goiás, Goiás, Brazil C. acaulis 7 6 -16.0229 -50.0756 NX - Nova Xavantina, Mato Grosso, Brazil C. acaulis 9 9 -14.7665 -52.4688 PA - Canaã dos Carajás, Pará, Brazil C. acaulis 8 8 -6.4523 -50.0735 MT - Confresa, Mato Grosso, Brazil C. acaulis 5 4 -10.5378 -51.4402 AF - Alta Floresta, Mato Grosso, Brazil C. acaulis 9 9 -9.8001 -55.9235 TS - Tangará da Serra, Mato Grosso, Brazil C. acaulis 9 9 -14.6498 -57.4206 CS - Lago Caiman, JM de Velasco, Bolivia C. acaulis 1 1 -13.6100 -60.9112 RO - Ariquemes, Rondônia, Brazil C. acaulis 8 8 -9.9490 -63.0537 AC - Porto Acre, Acre, Brazil C. acaulis 7 7 -9.7539 -67.6755 PM - Pando, Bolivia C. acaulis 1 1 -11.1415 -67.5468

Outgroup C. cuspidatus 1 1 - -

! 128!

Table 2. Outgroup taxa and respective approximate time (Ma) to the most recent common ancestor with Chamaecostus subsessilis (Nees &

Mart.) C.D.Specht & D.W.Stev.

Outgroup taxa Approximate time (Ma) to the

most recent common ancestor with Chamaecostus subsessilis

Reference

Chamaecostus lanceolatus (Petersen) C.D.Specht & D.W.Stev. 44 Specht 2006 Costus spicatus (Jacq.) Sw. 66 Specht 2006 Hellenia globosa (Blume) S.R. Dutta 66 Specht 2006 Alpinia zerumbet (Pers.) B.L. Burtt & R.M. Sm. (Zingiberaceae) 105 Kress & Specht 2006 Canna L. sp (Cannaceae) 106 Kress & Specht 2006 Strelitzia reginae Aiton (Strelitziaceae) 109 Kress & Specht 2006 Musella lasiocarpa (Franch.) C.Y. Wu (Musaceae) 110 Kress & Specht 2006 Allium diabolense (Ownbey & Aase ex Traub) McNeal (Amaryllidaceae) 125 Zeng et al. 2014 Allium denticulatum Kit. (Amaryllidaceae) 125 Zeng et al. 2014

!!

! 129!

5.2.2. High-throughput sequencing

Individual genomic DNA was sheared to a mean 200 base pairs size using a

Bioruptor®, and resulting libraries were double-indexed (Kircher et al. 2012). Probes were

selected from an independent and diverse array of loci: PCR products were generated from

plant genome regions previously acknowledged as informative (Table 3), such as several

plastid (Shaw et al. 2007) and nuclear markers (Specht et al. 2001, Kay et al. 2005, Salzman

et al. in press). Additional nuclear regions were obtained from Costus spicatus flower

transcriptome assembled reads (C. Specht, unpublished) aligned to the banana genome

(D’Hont et al. 2012), allowing the design of primers for the amplification of Chamaecostus

genomic DNA (Appendix 2, Table 3) at the ‘Specht Lab’ at UC Berkeley. Amplicon fragments

of the molecular markers listed in Table 1 were generated using ChoiceTaq® DNA

Polymerase with a 2 min. initial denaturing step at 94 °C, 45 cycles of 45 sec. at 94°C, 15

sec. at gene-specific annealing temperatures (Appendix 2), and 30 sec. at 72 °C, with a final

2 min. 72 °C extension. We largerly prevented paralogous loci because amplicon baits are

predominantly composed of introns. Plastid bait loci were pooled together with nuclear baits

in 1:20 concentrations based on quantification with a Qubit fluorometer (Life Technologies)

to a total of 1.3 µg. Three aliquots of baits were prepared to allow for enrichment

hybridizations.

Genomic libraries were pooled together in three aliquots of 50 individual samples

each at equimolar concentrations based on quantification with NanoDrop spectrophotometer

(ThermoScientific) to a total of 2 µg. Each aliquot was then enriched by hybridization to baits

linked to magnetic beads (Maricic et al. 2010, Peñalba et al. 2014) at the ‘Evolutionary

Genetics Lab’ at UC Berkeley, and then sequenced (100-bp paired-end reads) in one lane of

a Illumina® HiSeq® 2000 platform at ‘Vincent J. Coates Genomics Sequencing Laboratory’

at UC Berkeley.

! 130!

Table 3. Character values, number of haplotypes, and models of evolution for each plastid (cp) and nuclear (nr) DNA marker and the

concatenated alignments of Chamaecostus subsessilis and C. acaulis.

Marker Number of Characters

(bp)

Number of Reads per Sample (Mean ±

Standard Deviation)

Parsimony Informative Characters

Singletons Proportion of

Constant Characters (%)

Number of Haplotypes

Model of Evolution

cpDNA atpF 970 6642.2 ± 3176.2 7 2 99 10 F81+I petG-trnP 626 4675.8 ± 2347.1 5 0 99 6 F81 rbcL 1297 8405.6 ± 4153.8 2 3 99.6 5 F81 rps4-trnS 1021 8535.9 ± 4088.0 7 3 99 9 HKY rps16-trnK 1250 3849.0 ± 1694.5 23 6 96.3 18 TPM1uf trnC-ycf6 1283 4327.4 ± 2453.1 20 3 98.1 10 F81+G trnH-psbA 1631 5444.8 ± 2167.8 22 3 97.5 13 F81+I trnL-rpl32 1233 4165.7 ± 2048.9 40 6 94.2 19 TPM1uf+I trnL-trnLF 1135 9046.9 ± 4599.5 26 0 97.7 15 F81+I trnQ-rps16 1789 3433.0 ± 1238.4 58 10 96.2 34 F81+I concatenated 12235 - 210 36 97.5 50 -

! 131!

Table 3. Continued... nrDNA

BZR 304 156.3 ± 56.8 12 2 95.4 11 F81 CAM 1558 555.3 ± 223.0 26 9 97.6 17 F81 E2 1828 852.0 ± 246.1 60 56 92.2 29 HKY ETS 629 46307.4 ± 31734.8 83 3 85.8 21 K80+G HS 576 331.5 ± 93.2 5 23 95 11 F81 IGPS 931 463.4 ± 135.9 31 3 96.2 14 HKY PSY 338 109.5 ± 37.9 11 2 96.2 10 F81 RAB18 251 85.8 ± 29.4 6 3 96.4 6 F81 RPB2 611 381.4 ± 124.9 16 6 96.2 22 HKY WXY 1078 754.7 ± 295.6 29 14 95.4 13 HKY concatenated 8104 - 279 121 94.5 100 -

!!

! 132!

5.2.3. Bioinformatics

Reads were assembled with ABySS (Simpson et al. 2009) with k-mers of sizes

between 21 and 81, and also varying e/c to e2c2, e10c10 and e20c20. All the assemblies

were merged to get the final assemblies, from which we did reciprocal blast to reference

sequences of the enriched regions. Cleaned reads were then aligned to contigs using

NovoAlign (NovoCraft), with maxscore set at 90. SNPs were called and filtered with GATK

(McKenna et al. 2010). We avoided sequence erros mistaken to rare variants by a

conservative cut-off of coverage values (≥10).

5.2.4. Molecular Evolution and Phylogenetics

We analysed two datasets: (1) with Chamaecostus acaulis, Chamaecostus

subsessilis and Chamaecostus cuspidatus samples (N=120), with sequences fully trimmed

at beginning and at the end, and completely lacking missing information; and (2) dataset (1)

with sequences not fully trimmed and accommodating low level of missing data for a few

sites or samples, plus Costus spicatus and Hellenia globosa sequences. We aligned each

dataset for each marker using the MUSCLE algorithm (Edgar 2004) implemented in

GENEIOUS version 6.1.7 (www.geneious.com), and subsequently checked the multiple

sequence alignments manually. Models of evolution were determined for each marker

alignment in jModelTest version 2 (Darriba et al. 2012) using the Bayesian information

criterion.

5.2.5. Population Structure and Molecular Diversity

Individual and population assignments were conducted using the program

STRUCTURE version 2.3.4 (Pritchard et al. 2000) for dataset (1), which lacks missing data.

!!

! 133!

We explored an admixture model to determine the number of population clusters (k) from 1

through 15 using a burn-in of 1x105 and 2x106 MCMC replications after the burn-in. Five

iterations were run for each k value. The average and standard deviation of the likelihood of

each model were used to calculate Δk (Evanno et al. 2005) using STRUCTURE

HARVESTER web version 0.6.94 (Earl & von Holdt 2011), and the partitioning scheme with

the highest Δk was selected as the model with the most support. To cluster the five iterations

of the optimal partitioning scheme, we used CLUMPP 1.1.2 (Jakobsson & Rosenberg 2007).

Genetic structure was further examined by population pairwise FST comparisons (p < 0.001),

where populations were grouped according to optimal STRUCTURE partitioning scheme,

using ARLEQUIN 3.5 (Excoffier & Lischer 2010). Additionally, molecular diversity indices

were also calculated considering STRUCTURE’s output groups, in ARLEQUIN 3.5.

5.2.6. Divergence time

Nucleotide sequences of the plastid and nuclear genetic markers of the dataset (2)

were independently concatenated and analyzed under a Bayesian phylogenetic framework,

using BEAST 1.7.4 (Drummond & Rambaut 2007). Sequence data was partitioned to allow

different models of sequence evolution for each region, based on likelihood analyses ran on

jModelTest version 2.0 (Darriba et al. 2012). A relaxed clock with an uncorrelated lognormal

model of rate variation was used and a Yule speciation process for branching rates was

selected. One fossil-based time to the most recent common ancestor (tmrca) calibrations

was used: 45 ± 5 Ma for the Costaceae root (Costus incertis!; Berry 1925), and a CTMC rate

prior was selected (Ferreira & Suchard 2008). Markov chain Monte Carlo simulations were

run twice independently for 5x107 generations and sampled every 5x103. These analyses

were performed on the CIPRES Science Gateway (Miller et al. 2010). We assessed

convergence of model parameters across the independent runs by analyzing plots of the

marginal posterior distributions in Tracer version 1.5 (Rambaut & Drummond 2007), and by

!!

! 134!

ensuring high effective sample size values (ESS ≥ 200). Tracer was also used to assess

burn-in levels. A maximum clade-credibility tree was obtained from the posterior sample of

trees using TreeAnnotator version 1.7.4 (Drummond et al. 2012), and visualized on FigTree

(http://tree.bio.ed.ac.uk/software/figtree/).

5.3. Results

5.3.1. High-throughput sequencing

Overall, 11,616.88 ± 8,750.14 reads per sample did not match our targeted

sequences. Two molecular markers (Ih5GT and FT) and three samples, besides non-

Costaceae out-groups, were removed from analysis. Aligned filtered reads per populations

and outgroup taxa for each molecular marker are presented in Supplementary Figures 1 and

2 of Appendix 2. Character statistics, mean number of reads per sample, total number of

haplotypes, and models of evolution for alignments in dataset (1) are listed in Table 3.

5.3.2. Population Structure and Molecular Diversity

The highest Δk value derived from STRUCTURE replicates was found at k = 8 for

plastid DNA and k = 7 for nuclear DNA (Table 4). Resulting clusters denotes: populations

from Western Amazonia (hereafter, WAZ), populations from Southern Brazilian Amazonia

(SBA) and populations from Cerrados West of the Araguaia River (WCR), which comprises

Chamaecostus acaulis; and populations from Cerrados East and South of the Araguaia

River (ECR), the sampled population from upper São Francisco River (SFR), and the

sampled population from the Araguaia-Tocantins interfluve (ATI), which refers to

Chamaecostus subsessilis. Few Chamaecostus subsessilis individuals had a significant

portion (Q > 20) of their nuclear genome assigned to C. cuspidatus (blue in Figure 2).

!!

! 135!

Moreover, haplotype-based phylogenetic tree provides evidence of a strong

phylogeographic structure (Figures 2 and 3) between C. acaulis populations from Amazonia

(WAZ and SBA) and Cerrado (WCR). FST values also indicate geographically heterogeneous

populations (Table 5).

! 136!

Table 4. Parameters for selecting optimal partioning structure. Calculated for plastid (cp) and nuclear (nr) DNA

concatenated alignments of Chamaecostus subsessilis and C. acaulis. K: Number of population clusters. Likelihood

scores for each value of K genetic clusters from STRUCTURE (Pritchard et al. 2000), Delta K scores for each value of

K genetic clusters following Evanno et al. (2005).

K Mean LnP(K) cp / nr

Stdev LnP(K) cp / nr

Ln'(K) cp / nr

|Ln''(K)| cp / nr

Delta K cp / nr

1 -2647.6/-2485.44 0.2828/0.313 - - -

2 -1955.1/-1796.24 110.801/30.9448 692.5/689.2 307.32/297.16 2.773621/9.602894

3 -1569.92/-1404.2 46.9366/20.5926 385.18/392.04 140.12/193.76 2.985302/9.409207

4 -1324.86/-1205.92 53.5869/95.8363 245.06/198.28 69.28/103.68 1.292853/1.081845

5 -1149.08/-1111.32 31.8822/197.0423 175.78/94.6 211.22/62.08 6.625006/0.315059

6 -1184.52/-1078.8 162.1479/105.1456 -35.44/32.52 83.52/120.44 0.515085/1.145459

7 -1303.48/-925.84 494.4949/9.7048 -118.96/152.96 440.5/488.78 0.890808/50.364806

8 -981.94/-1261.66 122.1021/342.4732 321.54/-335.82 989.54/356.74 8.104204/1.041658

9 -1649.94/-1240.74 1280.7859/368.4694 -668/20.92 772.54/54.38 0.603177/0.147583

10 -1545.4/-1274.2 578.0299/411.7445 104.54/-33.46 7.68/71.9 0.013287/0.174623

11 -1433.18/-1379.56 705.1315/821.3339 112.22/-105.36 348.06/463.04 0.49361/0.563766

12 -1669.02/-1021.88 2242.5251/60.3688 -235.84/357.68 432.12/326.58 0.192693/5.409752

13 -1472.74/-990.78 1117.1607/64.9089 196.28/31.1 106.12/32.84 0.094991/0.50594

14 -1170.34/-992.52 571.2659/65.8149 302.4/-1.74 146.1/22.2 0.255748/0.33731

15 -1014.04/-972.06 278.6757/21.7308 156.3/20.46 - -

! 137!

Figure 2. Relationships of individuals based on the nuclear DNA concatenated alignment of Chamaecostus subsessilis and

C. acaulis.. Maximum Clade Credibility phylogeny is shown at top, with posterior probabilities of each node shown. The

bottom of the figure shows Bayesian admixture proportions (Q) of individual estimated by STRUCTURE, assuming K = 7.

! 138!

Figure 3. Relationships of individuals based on the plastid DNA concatenated alignment. Maximum Clade Credibility phylogeny

is shown at top, with posterior probabilities of each node shown. The bottom of the figure shows Bayesian admixture proportions

(Q) of individual estimated by STRUCTURE, assuming K = 8.

! 139!

Table 5. Pairwise FST between populations*. Plastid (cp, below diagonal) and nuclear (nr, above

diagonal) DNA concatenated alignments considered individually, of Chamaecostus subsessilis and

C. acaulis**.

Pairwise FST WCR SBA WAZ ATI SFR ECR

WCR 0.84516 0.87396 0.95247 0.94480 0.91097

nr SBA 0.63175 0.32231 0.90629 0.89137 0.88017 WAZ 0.65019 0.29127 0.90812 0.89007 0.87573 ATI 0.80102 0.84026 0.90871 0.87687 0.87770 SFR 0.81881 0.86619 0.96967 0.87541 0.61835 ECR 0.84652 0.86381 0.90535 0.86559 0.76868

cp *All FST values are significant (p=0.000).

** Chamaecostus acaulis populations = WAZ: Western Amazonia; SBA: Southern Brazilian

Amazonia; WCR: Cerrados West of the Araguaia River.

** Chamaecostus subsessilis populations = ECR: Cerrados East and South of the Araguaia River;

SFR: Upper São Francisco River; ATI: Araguaia-Tocantins Interfluve.

!!

! 140!

Genetic diversity (Nei 1978) ranged from 0.57 to 1.00, and the nucleotide diversity

from 0.0004 to 0.006, between all populations (Table 6). Highest diversity in C. acaulis was

observed in populations from the Cerrados West of the Araguaia River (WCR), and in

populations of Cerrados East of the Araguaia River in C. subsessilis. Overall, total haplotype

and nucleotide diversities were 0.98 and 0.014, and 1.00 and 0.018, respectively for plastid

and nuclear DNA.

! 141!

Table 6. Population molecular diversity. Plastid (cp) and nuclear (nr) DNA concatenated alignments considered

independently.

n Number of

haplotypes Gene diversity Mean number of

pairwise differences

Nucleotide diversity*

C. acaulis

WAZ cp 17 7 0.78 ± 0.07 6.91 ± 3.42 0.001 ± 0.000

nr 17 17 1.00 ± 0.02 16.99 ± 7.95 0.002 ± 0.001

SBA cp 31 8 0.84 ± 0.03 37.94 ± 16.95 0.003 ± 0.002

nr 29 17 0.95 ± 0.02 19.20 ± 8.75 0.002 ± 0.001

WCR cp 25 16 0.96 ± 0.02 70.92 ± 31.62 0.006 ± 0.003

nr 24 19 0.98 ± 0.02 4.74 ± 2.40 0.001 ± 0.000

C. subsessilis

ECR cp 31 11 0.89 ± 0.03 28.77 ± 12.93 0.002 ± 0.001

nr 31 29 1.00 ± 0.01 30.98 ± 13.90 0.004 ± 0.002

SFR cp 8 2 0.57 ± 0.10 4.57 ± 2.51 0.0004 ± 0.0002

nr 8 7 0.96 ± 0.08 35.14 ± 17.17 0.004 ± 0.002

ATI cp 9 5 0.81 ± 0.12 47.89 ± 22.96 0.004 ± 0.002

nr 9 9 1.00 ± 0.05 27.36 ± 13.26 0.003 ± 0.002

*Average over sites. ** Chamaecostus acaulis populations = WAZ: Western Amazonia; SBA: Southern Brazilian

Amazonia; WCR: Cerrados West of the Araguaia River. Chamaecostus subsessilis populations = ECR: Cerrados East

and South of the Araguaia River; SFR: Upper São Francisco River; ATI: Araguaia-Tocantins Interfluve.

!!

! 142!

5.3.3. Divergence Time

Bayesian phylogenetic analysis suggests that each species is monophyletic, with

very high support (posterior probability = 1.0). Moreover, population phylogenetic divergence

reflected STRUCTURE results. Direct observation of population relationships clearly points

to topological incongruence between cpDNA and nrDNA trees: C. subsessilis is sister to C.

cuspidatus in the cpDNA phylogeny (Figure 2), while it is sister to C. acaulis in the nrDNA

tree (Figure 3).

!!

! 143!

5.4. Discussion

New high-throughput sequencing technologies coupled with operational advances in

species-tree phylogeny estimation allow us to concomitantly estimate population divergence

and structure to a level previously unmanageable. Results presented here reveal a deep

phylogenetic structure based on range-wide sampling of populations of a Central South

American lineage adapted to dry ecoregions. We found deep phylogeographic structure at

both the population and species level, in plastid and nuclear genomes, as inferred from

fixation indices, structure analysis and phylogenetics.

The Amazon Basin had high environmental heterogeneity during the last 250,000

years, especially in precipitation amounts and distribution between Western (wetter) and

Eastern (drier) Amazonia (Cheng et al. 2013). Modern Amazonia is a product of terrain

development within an erosional regime since ∼2.5 Ma (Campbell Jr. et al. 2006). The main

lineage split is spatially coincident to the Brasiliano thermo-tectonic event that built the

Araguaia orogenic belt during the Neoproterozoic, between 550 and 1000 Ma (Moura &

Gaudette 1999, Chaves et al. 2008), which could have built an environmental filter to

occurrence and dispersal, isolating surrounding populations. Moreover, the divergence

between Eastern and Western lineages unlikely predates the formation of the Araguaia-

Tocantins River system but it may relate to the intense shifts in level and course that likely

occurred during the Miocene due to ocean level variation and Andean orogeny, which

greatly affected Amazonian hydrography (Campbell Jr. et al. 2006, Hoorn et al. 2010), and

the uplift of the central Brazilian plateau (Saadi et al. 2005).

Population structure of rainforest understory organisms is particularly affected by

integrity and continuity of forest cover (Smith et al. 2014), and in such dynamic scenario,

isolated populations at highlands almost certainly were formed. Morphological similarity

between these close-related populations could have been sustained by stabilizing selection

(Morjan & Rieseberg 2004), thus hiding a strong genetic structure. Hence, the Araguaia-

!!

! 144!

Tocantins River represents a secondary contact zone for the Chamaecostus lineages

considered, once it is a region where long-isolated lineages establish secondary contact

(Swenson & Howard 2005), it is spatially clustered at the approximate mid-point between

taxa extant distributions and divergence times between lineages are similar (Moritz et al.

2009). Indeed, this South American suture zone (Anderson 1949) was also found to have

affected other organism’s evolutionary history, such as lizards (Werneck et al. 2012),

primates (Couette 2007), and Lepidoptera (Brown & Gifford 2002).

Species of Chamaecostus investigated here display a strong supported phylogenetic

topological incongruence between plastid and nuclear genomes. Conflict between plastid

and nuclear datasets commonly indicates reticulating histories due to hybridization (Kellogg

et al. 1996, Rieseberg et al. 1996; Rieseberg 1997, Sang et al. 1997, Seelanan et al. 1997).

Particularly at genus level and below, phylogenetic relationships can be biased by

chloroplast capture resulting from hybridization (Rieseberg & Soltis 1991, Soltis & Kuzoff

1995). Nonetheless, genetic structure depends on species ecological attributes and life

history traits distinctively at nuclear and plastid markers once they represent different gene

dispersal mechanisms in plants, i.e. pollen and seed, respectively (Petit et al. 2005). Once

phylogenetic conservatism and covariation among traits are taken into account, plant genetic

structure is related with mating system for nuclear markers and seed dispersal mode or

geographic range size for organelle markers (Duminil et al. 2007), supporting that topology

inconcruency does relate to distinctive evolutionary processes (Bull et al. 1993, Soltis &

Kuzoff 1995, Huelsenbeck et al. 1996).

In a phylogenetic species concept perspective (Cracraft 1989), species can

frequently exchange individuals and genes with neighbouring species (Isaac et al. 2004). In

C. subsessilis, introgression from C. cuspidatus is further supported by admixture seen in

nrDNA structure. Shared haplotypes in recently diverging taxa can be ascribed to shared

ancestral polymorphism, since descendant lineages are expected to share genetic

polymorphism for some time (Lexer et al. 2006, Nosil et al. 2009). Moreover, divergence

!!

! 145!

between the taxa under consideration here is several million years old, so an ancestral

polymorphism hypotesis is a more likely explanation. Indeed, several Costus species are

interfertile in greenhouse crosses (Kay & Schemske 2008). Straightforward evidence for

hybridization among closely related species might also be achieved from crossing

experiments (Wendt et al. 2001, 2002). Hence, a full range sampling of C. cuspidatus,

including sympatric populations of both, would shed light into a hypothesis of introgression.

Surget-Groba & Kay (2013) found very restrictive gene flow between the sister

species Costus pulverulentus C.Presl and Costus scaber Ruiz & Pav, recovered from

microsatellite loci, and proposed that it could have been a causative factor to the rapid

radiation of Neotropical Costus. They even suggested that to formally test this hypothesis, it

would be necessary to expand taxa sampling, and compare results with lineages that did not

go through such a radiation. Chamaecostus are considerably more ancient than Costus

(Specht 2006, André et al. unpubl.), but it encompasses much less extant species, and we

still found significant strong phylogeographic structure and limited gene flow. Considerable

progress has been made on discerning the importance of different factors and traits

generating reproductive isolation (Schluter 2001, 2009). Gene flow holds species together

much by facilitating the spread of beneficial mutants and associated hitchhiking events

(Morjan & Rieseberg 2004), and additionally by homogenizing neutral loci (Wright 1931).

Reproductive barriers are important mainly to preserve adaptations, but they rarely protect

the entire genome from gene flow in recently diverged species (Rieseberg & Wendel 1993,

Arnold 1997, Rieseberg et al. 2004, Wu 2011). In this sense, populations within a species

evolve collectively due a combination of gene flow and selective sweeps at different spatial

scales (Morjan & Rieseberg 2004, Kane & Rieseberg 2007, Lexer & Widmer 2008, Arnold &

Martin 2010).

Although the clustered and isolated aspects of Chamaecostus populations could

indicate clonal reproduction, the high number of haplotypes in concatenated alignments

sustains that most individuals are genetically distinctive. Rarity along landscape plus the

!!

! 146!

elevated local genetic structure might categorize interglacial refugia (Stewart et al. 2010,

Bonatelli et al. 2014), which will consecutively be verified through a historic demography

approach (DeChaine & Martin 2006). Future works will also focus on comparing coding and

non-coding sequences to understand selection signatures across geographic space and its

relation with environmental tolerance and gene flow.

!!

! 147!

5.5. Acknowledgments

TA received a scholarship grant from Coordenação de Aperfeiçoamento de Pessoal de Nível

Superior. TW received a productivity grant from Conselho Nacional de Desenvolvimento

Científico e Tecnológico. CS received support from the US National Geographic Society

(CRE Grant #8994-11) and CPS received support from BIOTA-FAPESP (2009/52725-3),

which helped support this research. We are thankful to Lydia Smith for assistance with

targeted enrichment, and Grady Pierroz for helping to select and amplify probes.

5.6. References

Almeida FC, Bonvicino CR, Cordeiro-Estrela P. 2007. Phylogeny and temporal

diversification of Calomys (Rodentia, Sigmodontinae): implications for the biogeography

of an endemic genus of the open/dry biomes of South America. Molecular Phylogenetics

and Evolution 42: 449-466.

Anderson E. 1949. Introgressive hybridization. Wiley. 109 p.

Arnold ML, Martin NH. 2010. Hybrid fitness across time and habitats. Trends in Ecology and


Arnold ML. 1997. Natural Hybridization and Evolution. Oxford Series in Ecology and

Evolution. Oxford University Press USA. 232 p.

Avise JC, Arnold J, Ball RM, Bermingham E, Lamb T, Neigel JE, Reeb CA, Saunders NC.

1987. Intraspecific phylogeography: the mitochondrial DNA bridge between population

genetics and systematics. Annual Review of Ecology and Systematics 18: 489-522.

Avise JC. 2009. Phylogeography: retrospect and prospect. Journal of Biogeography 36: 3-

15.

!!

! 148!

Behling H, Hooghiemstra H. 2001. Neotropical savanna environments in space and time:

Late Quaternary interhemispheric comparisons. In Markgraf V. (Ed.). Interhemispheric

Climate Linkages. Academic Press. 307-323 p.


530-537.

Bonatelli IAS, Perez MF, Peterson AT, Taylor NP, Zappi DC, Machado MC, Koch I, Pires

AHC, Moraes EM. 2014. Interglacial microrefugia and diversification of a cactus species

complex: phylogeography and palaeodistributional reconstructions for Pilosocereus

aurisetus and allies. Molecular Ecology 23: 3044-3063.

Brown JH. 1995. Macroecology. University of Chicago Press. 284 p.

Brown KS, Gifford DR. 2002. Lepidoptera in the Cerrado landscape and the conservation of

vegetation, soil, and topographical mosaics. In Oliveira PS, Marquis RJ. (Eds.). The

Cerrados of Brazil: Ecology and Natural History of a Neotropical Savanna. Columbia

University Press. 201-222 p.

Bull JJ, Huelsenbeck JP, Cunningham CW, Swofford DL, Waddell PJ. 1993. Partitioning and

combining data in phylogenetic analysis. Systematic Biology 42: 384-397.

Campbell Jr. KE, Frailey CD, Romero-Pittman L. 2006. The Pan-Amazonian Ucayali

Peneplain, late Neogene sedimentation in Amazonia, and the birth of the modern

Amazon River system. Palaeogeography, Palaeoclimatology, Palaeoecology 239: 166-

219.

Chaves CL, Gorayeb PSS, Moura CAV. 2008. Geologia, geocronologia e litoquímica dos

granitos paleoproterozóicos do tipo A no setor setentrional da Província Tocantins.

Revista Brasileira de Geociências 38: 366-378.

Cheng H, Sinha A, Cruz FW, Wang X, Edwards RL, d’Horta FM, Ribas CC, Vuille M, Scott

!!

! 149!

LD, Auler AS. 2013. Climate change patterns in Amazonia and biodiversity. Nature

Communications 4: 1411.

Cole M. 1986. The Savannas: Biogeography and Geobotany. Academic Press. 438&p.

Collevatti RG, Grattapaglia D, Hay JD. 2003. Evidences for multiple maternal lineages of

Caryocar brasiliense populations in the Brazilian Cerrado based on the analysis of

chloroplast DNA sequences and microsatellite haplotype variation. Molecular Ecology

12: 105-115.

Costa LP. 2003. The historical bridge between the Amazon and the Atlantic Forest of Brazil:

a study of molecular phylogeography with small mammals. Journal of Biogeography 30:

71-86.

Couette S. 2007. Différenciation morphologique et génétique des populations de

douroucoulis (Aotus infulatus, Primates, Platyrhiniens, Cebidae) provenant des rives

droite et gauche du rio Tocantins (Brésil). Comptes Rendus Biologies 330: 148-158.

Cracraft J. 1989. Speciation and its ontology: the empirical consequences of alternative

species concepts for understanding patterns and processes of differentiation. In Otte D,

Endler JA (Eds). Speciation and its Consequences. Sinauer Associates. pp. 28-59.

D’Hont A, Denoeud F, Aury JM, Baurens FC, Carreel F, Garsmeur O, Noel B, Bocs S, Droc

G, Rouard M, Silva C, Jabbari K, Cardi C, Poulain J, Souquet M, Labadie K, Jourda C,

Lengellé J, Rodier-Goud M, Alberti A, Bernard M, Correa M, Ayyampalayam S, Mckain

MR, Leebens-Mack J, Burgess D, Freeling M, Mbéguié-A-Mbéguié D, Chabannes M,

Wicker T, Panaud O, Barbosa J, Hribova E, Heslop-Harrison P, Habas R, Rivallan R,

Francois P, Poiron C, Kilian A, Burthia D, Jenny C, Bakry F, Brown S, Guignon V, Kema

G, Dita M, Waalwijk C, Joseph S, Dievart A, Jaillon O, Leclercq J, Argout X, Lyons E,

Almeida A, Jeridi M, Dolezel J, Roux N, Risterucci AM, Weissenbach J, Ruiz M,

Glaszmann JC, Quétier F, Yahiaoui N, Wincker P. 2012. The banana (Musa acuminata)

!!

! 150!

genome and the evolution of monocotyledonous plants. Nature 488: 213-217.



DeChaine EG, Martin AP. 2006. Using coalescent simulations to test the impact of

Quaternary climate cycles on divergence in an alpine plant-insect association. Evolution

60: 1004-1013.


Focus 12:13-15.

Drummond AJ, Rambaut A. 2007. BEAST: Bayesian evolutionary analysis by sampling

trees. BMC Evolutionary Biology 7: 214.

Drummond AJ, Suchard MA, Xie D, Rambaut A. 2012. Bayesian phylogenetics with BEAUti

and the BEAST 1.7. Molecular Biology and Evolution 29: 1969-1973.

Duminil J, Fineschi S, Hampe A, Jordano P, Salvini D, Vendramin GG, Petit RJ. 2007. Can

Population Genetic Structure Be Predicted from Life‐History Traits? American Naturalist

169: 662-672.

Earl DA, vonHoldt BM. 2011. STRUCTURE HARVESTER: a website and program for

visualizing STRUCTURE output and implementing the Evanno method. Conservation

Genetics Resources DOI: 10.1007/s12686-011-9548-7


throughput. Nucleic Acids Research 32: 1792-1797.

Endler JA. 1982. Problems in distinguishing historical from ecological factors in

biogeography. American Zoologist 22: 441-452.

Evanno G, Regnaut S, Goudet J. 2005. Detecting the number of clusters of individuals using

!!

! 151!

the software STRUCTURE: a simulation study. Molecular Ecology 14: 2611-2620.

Excoffier L, Lischer HEL. 2010. Arlequin suite ver 3.5: A new series of programs to perform

population genetics analyses under Linux and Windows. Molecular Ecology Resources.

10: 564-567.

Felfili JM, Mendonça RC, Walter BMT, Silva Júnior MC, Fagg CW, Nóbrega MGG, Sevilha

AC, Silva MA. 2001. Flora fanerogâmica das matas de galeria e ciliares do Brasil

Central. In Ribeiro JF, Fonseca CEL, Sousa-Silva JC (Eds) Cerrado: caracterização e

recuperação de Matas de Galeria. Embrapa. 195-263 p.

Ferreira MAR, Suchard MA. 2008. Bayesian analysis of elapsed times in continuous-time

Markov chains. Canadian Journal of Statistics 26: 355-368.

Ho SYW, Jermiin LS. 2004. Tracing the decay of the historical signal in biological sequence

data. Systematic Biology 53: 628-637.

Hoorn C, Wesselingh FP, ter Steege H, Bermudez MA, Mora A, Sevink J, Sanmartín I,

Sanchez-Meseguer A, Anderson CL, Figueiredo JP, Jaramillo C, Riff D, Negri FR,

Hooghiemstra H, Lundberg J, Stadler T, Sarkinen T, Antonelli A. 2010. Amazonia

through time: Andean uplift, climate change, landscape evolution, and biodiversity.

Science 330: 927-931.

Huelsenbeck JP, Bull JJ, Cunningham CW. 1996. Combining data in phylogenetic analysis.

Trends in Ecology and Evolution 11: 152-157.

Isaac NJB, Purvis A. 2004. The ‘species problem’ and testing macroevolutionary

hypotheses. Divers. Distrib. 10, 275–281

Jakobsson M, Rosenberg NA. 2007. CLUMPP: a cluster matching and permutation program

for dealing with label switching and multimodality in analysis of population structure.

Bioinformatics 23: 1801-1806.

!!

! 152!

Kane NC, Rieseberg LH. 2007. Selective sweeps reveal candidate genes for adaptation to

drought and salt tolerance in common sunflower, Helianthus annuus. Genetics

175:1803-1812.

Kay KM, Schemske DW. 2008. Natural selection reinforces speciation in a radiation of

Neotropical rainforest plants. Evolution 62: 2628-2642.




Botany 92: 1899-1910.

Kellogg EA, Appels R, Mason-Gamer RJ. 1996. When genes tell different stories: the diploid

genera of Triticeae (Gramineae). Systematic Botany 21: 321-347.

Kircher M, Sawyer S, Meyer M. 2012. Double indexing overcomes inaccuracies in multiplex

sequencing on the Illumina platform. Nucleic Acids Research 40: e3.

Lage LA, Silva IV, André T. 2012. Root tuber morphoanatomy of Chamaecostus subsessilis

(Costaceae), a drought-tolerant ground herb from seasonal forests of central South

America. Bulletin of the Heliconia Society International 18: 10-11.

Lemmon EM, Lemmon AR. 2013. High-Throughput Genomic Data in Systematics and

Phylogenetics. Annual Review of Ecology, Evolution and Systematics 44: 99-121.

Lexer C, Widmer A. 2008.The genic view of plant speciation: recent progress and emerging

questions. Philosophical Transactions of the Royal Society B 363: 3023-3036.

Lexer C, Kremer A, Petit R. 2006. Shared alleles in sympatric oaks: recurrent gene flow is a

more parsimonious explanation than ancestral polymorphism. Comment. Molecular

Ecology 15: 2007-2012.

Librado P, Rozas J. 2009. DnaSP v5: A software for comprehensive analysis of DNA

!!

! 153!

polymorphism data. Bioinformatics 25: 1451-1452.

Lomolino MV, Riddle BR, Whittaker RJ, Brown JH. 2010. Biogeography. 4th Ed. Sinauer

Associates 560 p.





Mamanova L, Coffey AJ, Scott CE, Kozarewa I, Turner EH, Kumar A, Howard E, Shendure

J, Turner DJ. 2010. Target-enrichment strategies for next-generation sequencing. Nature

Methods 7: 111-118.

Maricic T, Whitten M, Pääbo S. 2010. Multiplexed DNA sequence capture of mitochondrial

genomes using PCR products. PLoS ONE 5: e14004.

McCormack JE, Hird SM, Zellmer AJ, Carstens BC, Brumfield RT. 2013. Applications of

next-generation sequencing to phylogeography and phylogenetics. Molecular

Phylogenetics and Evolution 66: 526-538.

McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, Garimella K,

Altshuler D, Gabriel S, Daly M, DePristo MA. 2010. The Genome Analysis Toolkit: a

MapReduce framework for analyzing next-generation DNA sequencing data. Genome

Research 20: 1297-1303.

Miller MA, Pfeiffer W, Schwartz T. 2010. Creating the CIPRES Science Gateway for

Inference of Large Phylogenetic Trees. In Proceedings of the Gateway Computing

Environments Workshop (GCE), New Orleans, LA: 1-8.

Moritz C, Hoskin CJ, MacKenzie JB, Phillips BL, Tonione M, Silva N, VanDerWal J, Williams

SE, Graham CH. 2009. Identification and dynamics of a cryptic suture zone in tropical

!!

! 154!

rainforest. Proceedings of the Royal Society B 276: 1235-1244.

Morjan CL, Rieseberg LH. 2004. How species evolve collectively: implications of gene flow

and selection for the spread of advantageous alleles. Molecular Ecology 13: 1341-1356.

Moura CAV, Gaudette HE. 1999. Zircon ages of basement orthogneisses from the northern

of the Araguaia Belt, Brazil. Basement Tectonics 13: 155-178.

Nosil P, Funk DJ, Ortiz-Barrientos D. 2009. Divergent selection and heterogeneous genomic

divergence. Molecular Ecology 18: 375-402.

Peñalba JV, Smith LL, Tonione MA, Sass C, Hykin SM, Skipwith PL, McGuire JA, Bowie

RCK, Moritz C. 2014.Sequence capture using PCR-generated probes: a cost-effective

method of targeted high-throughput sequencing for nonmodel organisms. Molecular

Ecology Resources 14: 1000-1010.

Pennington RT, Lavin M, Oliveira-Filho A. 2009. Woody plant diversity, evolution, and

ecology in the tropics: perspectives from seasonally dry tropical forests. Annual Review

of Ecology, Evolution and Systematics 40: 437-457.

Petit RJ, Duminil J, Fineschi S, Hampe A, Salvini D, Vendramin GG. 2005. Comparative

organisation of chloroplast, mitochondrial and nuclear diversity in plant populations.

Molecular Ecology 14: 689-711.

Prado DE, Gibbs PE. 1993. Patterns of species distributions in the dry seasonal forests of

South America. Annals of the Missouri Botanical Garden 80: 902-927.

Prado CPA, Haddad CFB, Zamudio KR. 2012. Cryptic lineages and Pleistocene population

expansion in a Brazilian Cerrado frog. Molecular Ecology 21: 921-941.

Pritchard JK, Stephens M, Donnelly P. 2000. Inference of population structure using

multilocus genotype data. Genetics 155: 945-959.

!!

! 155!

Rambaut A, Drummond AJ. 2007. Tracer v1.5, Available from


Ramos ACS, Lemos-Filho JP, Ribeiro RA, Santos FR, Lovato MB. 2007. Phylogeography of

the tree Hymenaea stigonocarpa (Fabaceae: Caesalpinioideae) and the influence of

Quaternary climate changes in the Brazilian cerrado. Annals of Botany 100: 1219-1228.

Reiseberg LH, Soltis DE. 1991. Phylogenetic consequences of cytoplasmic gene flow in

plants. Evolutionary Trends in Plants 5: 65-84.

Rieseberg LH, Wendel J. 1993. Introgression and its consequences in plants. In Harrison R.

(Ed). Hybrid Zones and the Evolutionary Process. Oxford University Press USA. 70-114

p.

Rieseberg LH, Church S, Morjan CL. 2004. Integration of populations and differentiation of

species. New Phytologist 161: 59-69.

Rieseberg LH, Whitton J, Linder CR. 1996. Molecular marker incongruence in plant hybrid

zones and phylogenetic trees. Acta Botanica Neerlandica 45: 243-262.

Rieseberg LH. 1997. Hybrid origins of plant species. Annual Review of Ecology and

Systematics 28: 359–389.

Saadi A, Bezerra FHR, Costa RD, Igreja HLS, Franzinelli E. 2005. Neotectônica da

plataforma brasileira. In Souza CRG, Suguio K, Oliveira AMS, Oliveira PE (Eds).

Quaternário do Brasil. Holos Editora. 211-234 pp.

Salzman S, Driscoll HE, Renner T, André T, Shen S, Specht CD. Spiraling into history: A

molecular phylogeny and investigation of biogeographic origins and floral evolution for

the genus Costus L. Systematic Botany in press.

Sang T, Crawford DJ, Stuessy TF. 1997. Chloroplast DNA phylogeny, reticulate evolution,

and biogeography of Paeonia (Paeoniaceae). American Journal of Botany 84: 1120–

!!

! 156!

1136.

Schluter D. 2001. Ecology and the origin of species. Trends in Ecology and Evolution 16:

372-380.

Schluter D. 2009. Evidence for ecological speciation and its alternative. Science 323: 737-

741.

Seelanan T, Schnabel A, Wendel JF. 1997. Congruence and consensus in the cotton tribe

(Malvaceae). Systematic Botany 22: 259-290.


sequences to choose noncoding regions for phylogenetic studies in angiosperms: the

tortoise and the hare III. American Journal of Botany 94: 275-288.

Silva IA, Silva DM, Carvalho GH, Batalha MA. 2011. Reproductive phenology of Brazilian

savannas and riparian forests: environmental and phylogenetic issues. Annals of Forest

Science 68: 1207-1215.

Silva JF, Farinas MR, Felfili JM, Klink CA. 2006. Spatial heterogeneity, land use and

conservation in the Cerrado region of Brazil. Journal of Biogeography 33: 536-548.

Silva JMC, Bates JM. 2002. Biogeographic patterns and conservation in the South American

Cerrado: a tropical savanna hotspot. BioScience 52: 225-234.

Simpson JT, Wong K, Jackman SD, Schein JE, Jones SJM, Birol İ. 2009. ABySS: A parallel

assembler for short read sequence data. Genome Research 19: 1117-1123.

Smith BT, McCormack JE, Cuervo AM, Hickerson MJ, Aleixo A, Cadena CD, Pérez-Emán J,

Burney CW, Xie X, Harvey MG, Faircloth BC, Glenn TC, Derryberry EP, Prejean J,

Fields S, Brumfield RT. 2014. The drivers of tropical speciation. Nature 515: 406-409.

Soltis DE, Kuzoff RK. 1995. Discordance between nuclear and chloroplast phylogenies in

!!

! 157!

the Heuchera group (Saxifragaceae). Evolution 49: 727-742.

Specht CD. 2006. Gondwanan vicariance or dispersal in the tropics? The biogeographic




Stewart JR, Lister AM, Barnes I, Dalén L. 2010. Refugia revisited: individualistic responses

of species in space and time. Proceedings of the Royal Society of London Series B 277:

661-671.

Surget-Groba Y, Kay KM. 2013. Restricted gene flow within and between rapidly diverging

Neotropical plant species. Molecular Ecology 22: 4931-4942.

Swenson N G, Howard DJ. 2005. Clustering of contact zones, hybrid zones, and

phylogeographic breaks in North America. American Naturalist 166: 581-591.

Turchetto-Zolet AC, Pinheiro F, Salgueiro F, Palma-Silva C. 2013. Phylogeographical

patterns shed light on evolutionary process in South America. Molecular Ecology 22:

1193-1213.

Turchetto-Zolet AC, Cruz F, Vendramin GG, Simon MF, Salgueiro F, Margis-Pinheiro M,

Margis R. 2012. Large-scale phylogeography of the disjunct Neotropical tree species

Schizolobium parahyba (Fabaceae-Caesalpinioideae). Molecular Phylogenetics and


Wendt T, Canela MBF, Faria APG, Rios RI. 2001. Reproductive biology and natural

hybridization between two endemic species of Pitcairnia (Bromeliaceae). American


Wendt T, Canela MBF, Klein DE, Rios RI. 2002. Selfing facilitates reproductive isolation

among three sympatric species of Pitcairnia (Bromeliaceae). Plant Systematics and

Evolution 232: 201-212.

!!

! 158!

Werneck FP. 2011. The diversification of eastern South American open vegetation biomes:

historical biogeography and perspectives. Quaternary Science Reviews, 30: 1630-1648.

Werneck FP, Costa GC, Colli GR, Prado DE, Sites Jr JW. 2011. Revisiting the historical

distribution of seasonally dry tropical forests: new insights based on palaeodistribution

modelling and palynological evidence. Global Ecology and Biogeography 20: 272-288.

Werneck FP, Gamble T, Colli GR, Rodrigues MT, Sites Jr JW. 2012. Deep diversification

and long-term persistence in the South American ‘dry diagonal’: integrating continent-

wide phylogeography and distribution modeling of Geckos. Evolution 66: 3014-3034.

Whinnett A, Zimmermann M, Willmott KR, Herrera N, Mallarino R, Simpson F, Joron M,

Lamas G, Mallet J. 2005. Strikingly variable divergence times inferred across an

Amazonian butterfly ‘suture zone’. Proceedings of the Royal Society B 272: 2525-2533.

Wright S. 1931. Evolution in Mendelian populations. Genetics 16: 97-169.

Wu CI. 2001. The genic view of the process of speciation. Journal of Evolutionary Biology

14: 851-865.

Zeng L, Zhang Q, Sun R, Kong H, Zhang N, Ma H. 2014. Resolution of deep angiosperm

phylogeny using conserved nuclear genes and estimates of early divergence times.

Nature Communications 5: 4956.

!!

! 159!

!!

! 160!

Chapter 6 – Final remarks

!!

! 161!

Biological diversity varies through space, time, and among different kinds of

organisms. The conspicuous disparities in both species richness and morphological diversity

that exist between close-related groups of organisms comprise a common pattern in the

organization of biological diversity, and such cases appear in nearly any taxonomic group.

The genus Costus is by far the most species-rich in the spiral gingers family (Costaceae

Nakai), with species occurring natively in tropical Africa as well as Central and South

America. The majority of the diversity in Costus species number and morphology occurs in

the Neotropics. However, an ancient lineage of Costaceae also inhabits the New World

tropics (consisting of Monocostus, Dimerocostus and Chamaecostus) and is likely to have

played a role in Neotropical forest ecosystems long before the arrival of Costus from Africa.

The causes of phylogenetic, temporal, and geographic variation are at least partially related

to the evolutionary process of diversification. Results presented in this thesis erect

propositions related to speciation and the evolution of species diversity in the spiral gingers.

Costaceae display a significant variation in speciation rate among clades, with a

significant rate shift at the most recent common ancestor of the Neotropical Costus clade.

The geographic context of speciation in Neotropical Costaceae lineages uncovers an overall

predominance of allopatric speciation in the older divergent South American clade, where

most extant species display little overlap in geographic ranges. In contrast, sympatry is much

higher within the Neotropical Costus clade, independent of node age.

Rapid radiations are fundamental to examining evolutionary processes associated

with bursts of speciation and morphological diversification. Ancestral reconstruction of

pollination syndrome within Costus presents an evolutionary toggle in pollination

morphologies, demonstrating both the multiple independent evolutions of ornithophily (bird

pollination) as well as reversals to melittophily (bee pollination). Based on the current

distribution for the Neotropical and African species, an historic dispersal of a bee-pollinated

taxon from Africa occurred to the Pacific Coast of Mexico and Central America, subsequently

diversificating and leading to the evolution of a bird-pollinated floral morphology in multiple

!!

! 162!

derived lineages.

Most species within the genus Chamaecostus (Costaceae) are well defined, but there

is a broad geographic range and long list of synonyms associated with Chamaecostus

subsessilis, which could indicate the presence of some cryptic species within this species

complex. Indeed, while Chamaecostus is strongly monophyletic, C. cuspidatus is found to be

sister to a clade of some but not all samples of C. subsessilis, making it necessary to

acknowledge more than one species in the C. subsessilis complex. Leaf morphometric

measurements of herbarium specimens reveal limited distinction among phylogenetic

lineages, demonstrating a cryptic speciation scenario.

A further examination of speciation of the understory herbs from the Chamaecostus

subsessilis complex along South American seasonally dry forests of Cerrado and South

Amazonia support a deep phylogeographic structure at both the population and species

level, as inferred from fixation indices, structure analysis and phylogenetics. These lineages

exhibit a strong supported phylogenetic topological incongruence between plastid and

nuclear genomes and admixture analysis.

Biogeography and trait evolution in Costaceae is still insufficiently investigated, and

to advance in the study of the evolutionary scenarios of diversification, other approaches

would be largely valuable, such as the analysis of differences in niche dimensions and of

niche conservation over evolutionary time, the correlation of species’ phenotypic variability

and geographic range attributes, and the assessment of selection signature, and of the

timing and geographical pattern of demographic evolution. Their charismatic spiral foliage

and normally bold-colored inflorescences make Costaceae species prominent features of

most lowland Neotropical forests, and as for its manageable species richness, variable

morphological traits, and easy of cultivation, they are emergening as a strategic family to

understand speciation patterns and processes in Tropical landscapes.

!!

! 163!

Appendix 1. Voucher Information for species sampled and GenBank accession numbers for

DNA sequences. Information is as follows: species name, voucher information, GenBank

accession numbers for five loci: ITS, ETS, CaM, rsp16-trnK, trnL-trnLF. Genbank accessions

beginning KJ were created in this study. Previously generated sequences for a species that

used a separate vouchered individual have the voucher information indicated in

parentheses. Dashes denote missing data. Collector abbreviations for herbarium vouchers:

BS = B. Stahl; BTM = B.T. Meise; CS = Chelsea Specht; M = P.J.M. Maas; Mood = John

Mood; KMN = Lyon Arboretum Herbarium (HLA) collector Ken Nagata; R = David Skinner;

TvA = Tinde van Andel; US = Uwe Scharf; WJK = W. John Kress. Abbreviations for living

collections: NMNH = Smithsonian Greenhouse Collections; L = Lyon Arbortetum; BB =

Burgers Bush; NY = New York Botanical Garden Living Collection; RB = Rio de Janeiro

Botanical Garden herbarium collections.

Chamaecostus lanceolatus (Petersen) var. pulchriflorus (Ducke) C. D. Specht & D. W. Stev., RB550388, KJ011475, KJ011224, KJ011299, KJ011419, KJ011347; Chamaecostus cuspidatus (Nees & Mart.) C. D. Specht & D. W. Stev., RB584184, AY994739 (WJK94-3681), KJ011223, KJ011298, KJ011418, KJ011346; Chamaecostus subsessilis (Nees & Mart.) C. D. Specht & D. W. Stev., RB584183, KJ011476, KJ011225, KJ011300, KJ011420, KJ011348; Hellenia globosa (Blume) S.R. Dutta, CS02-351 as Cheilocostus globosus, KJ011479, KJ011228, -, KJ011423, AY994592 (Mood 1713); Costus afer Ker Gawl., M10205, KJ011425, KJ011151, KJ011230, KJ011351, KJ011304; Costus albiflos Maas & H. Maas , M9968, KJ011426, KJ011152, KJ011231, KJ011352, KJ011305; Costus allenii Maas, M9563, AY994743 (NY347/95A), KJ011153, -, KJ011353, AY994587 (NY347/95A); Costus amazonicus J. F. Macbr., M9036 , AY994742 (CS02-327), KJ011154, KJ011232, KJ011354, AY994586 (CS02-327); Costus arabicus L., CS98-193, AY041034, KJ011155, KJ011233, KJ011355, KJ011306; Costus asplundii (Maas) Maas, R2976, KJ011427, KJ011156, KJ011234, KJ011356, KJ011307; Costus barbatus Suess., CS01-256, AY994741 (NY1413/91B), KJ011157, KJ011235, KJ011357, KJ011308; Costus beckii Maas & H. Maas, CS99-232 , KJ011428, KJ011158, KJ011236, KJ011358, KJ011309; Costus bracteatus Rowlee, M9409, KJ011429, KJ011159, KJ011237, KJ011359, KJ011310; Costus chartaceus Maas, WJK90-3124, AY994719 (WJK99-6356), KJ011160, KJ011238, KJ011360, AY994559 (WJK 99-6356); Costus claviger Benoist, M9306, AY994740 (KMN2361), KJ011161, KJ011239, KJ011361, AY994584 (KMN 2361); Costus comosus var. bakeri (K. Schum.) Maas, BB1999-0126011, KJ011430, KJ011162, KJ011240, KJ011362, KJ011311; Costus curvibracteatus Maas, M9381, KJ011431, KJ011163, KJ011241, KJ011363, AY994583 (NY356-95A); Costus deistelii K. Schum., M9298, AY994752, KJ011165, KJ011242, KJ011365, AY994599; Costus aff. dubius, L92.0048, KJ011434, KJ011167, KJ011244, KJ011367, AY994596 (M3549); Costus dinklagei K. Schum., TvA3549, KJ011433, KJ011166, KJ011243, KJ011366, KJ011312; Costus dirzoi García-Mend. & Ibarra-Manr., 1996-0228002, KJ011435, KJ011168, KJ011245, KJ011368, KJ011313; Costus dubius K. Schum., M10206, KJ011436, KJ011169, KJ011246, KJ011369, KJ011314; Costus erythrocoryne, K. Schum., AY994738 (CS02-326), AY972950 (KK98.73), -, -, AY994579 (CS02-326); Costus erythrophyllus Loes., R2847, KJ011437, KJ011170,

!!

! 164!

KJ011247, -, KJ011315; Costus erythrothyrsus Loes., M9317, KJ011438, KJ011171, KJ011248, KJ011370, KJ011316; Costus gabonensis Koechlin, M10291, KJ011439, KJ011172, KJ011249, KJ011371, KJ011317; Costus geothyrsus K. Schum., BS6782, KJ011440, KJ011173, KJ011250, KJ011372, KJ011318; Costus glaucus Maas, Delft 46-325, KJ011441, KJ011174, KJ011251, KJ011373, KJ011319; Costus guanaiensis var. guanaiensis Rusby, R2666, AY994737 (L94.0306), KJ011175, KJ011252, KJ011374, AY994577 (L94.0306); Costus guanaiensis var. macrostrobilus (K. Schum.) Maas, CS00-253, KJ011442, KJ011176, KJ011253, KJ011375, KJ011320; Costus guanaiensis var. tarmicus (Loes.) Maas, KMN2811, AY994751, KJ011177, KJ011254, KJ011376, AY994597; Costus laevis Ruiz & Pav., CS01-257, KJ011443, KJ011178, KJ011255, KJ011377, AY994576 (NY351/95A); Costus lasius Loes., M9155, AY994735 (NMNH 94-670), KJ011179, -, KJ011378, AY994575 (NMNH 94-670); Costus lateriflorus Gagnep., M10331, KJ011444, KJ011180, KJ011256, KJ011379, KJ011321; Costus letestui Pellegr., CS02-331, AY994733, KJ011181, KJ011257, KJ011380, AY994573; Costus leucanthus Maas, BB1996-1105001, KJ011445, KJ011182, KJ011258, KJ011381, KJ011322; Costus aff. ligularis, BB1998-0923003, KJ011446, KJ011183, KJ011259, KJ011382, KJ011323; Costus aff. ligularis , M10226, KJ011448, KJ011185, KJ011261, KJ011384, -; Costus aff. ligularis , US237, KJ011447, KJ011184, KJ011260, KJ011383, KJ011324; Costus sp. nov., M9995, KJ011449, KJ011186, KJ011262, KJ011385, KJ011325; Costus lima K. Schum., R3014, KJ011450, KJ011187, KJ011263, KJ011386, -; Costus lima var. scabremarginatus Maas, BTM75-0400, KJ011451, KJ011188, KJ011264, KJ011387, KJ011326; Costus loangensis H. Maas & Maas, M10184, KJ011452, KJ011189, KJ011265, KJ011388, KJ011327; Costus longibracteolatus Maas, R3003, KJ011453, KJ011190, KJ011266, KJ011389, KJ011328; Costus lucanusianus J. Braun & K. Schum., M10000, KJ011455, KJ011192, KJ011268, KJ011391, -; Costus maboumiensis Pellegr., M10227, KJ011456, KJ011193, KJ011269, KJ011392, KJ011330; Costus aff. afer, WJK94-3683, AY994731, KJ011194, KJ011270, KJ011393, AY994571; Costus malortieanus H. Wendl., M9791, KJ011457, KJ011195, KJ011271, KJ011394, KJ011331; Costus montanus Maas, R2972, AY994729 (KKsn), KJ011196, -, -, AY994569 (KKsn); Costus nitidus Maas, M9524, KJ011458, KJ011197, KJ011272, -, KJ011332; Costus osae Maas & H. Maas, M9501, KJ011459, KJ011198, KJ011273, KJ011395, KJ011333; Costus phaeotrichus Loes., CS02-323, AY994721, KJ011199, KJ011274, KJ011396, AY994561; Costus aff. phyllocephalus, BB870057, KJ011460, KJ011200, KJ011275, KJ011397, KJ011334; Costus aff. phyllocephalus, M10389, KJ011461, KJ011201, KJ011276, KJ011398, KJ011335; Costus pictus D.Don ex Lindl., WJK94-3691, AY994726, KJ011202, KJ011277, KJ011399, AY994566; Costus plicatus Maas, WJK94-5376, AY994725, KJ011203, KJ011278, KJ011400, AY994565; Costus productus Gleason ex Maas, R2693, KJ011462, KJ011204, KJ011279, KJ011401, KJ011336; Costus pulverulentus C. Presl, WJK94-3680, AY994723, KJ011205, KJ011280, KJ011402, AY994563; Costus quasi-appendiculatus Woodson ex Maas, CS99-233, KJ011463, KJ011206, KJ011281, KJ011403, KJ011337; Costus ricus Maas & H. Maas, R2970, KJ011464, KJ011207, KJ011282, KJ011404, KJ011338; Costus scaber Ruiz & Pav., R2253, KJ011465, KJ011208, KJ011283, KJ011405, KJ011339; Costus spectabilis K. Schum., WJK97-6118, AY994718, KJ011209, -, KJ011406, AY994556; Costus spicatus Sw., WJK02-7143, KJ011432, KJ011164, KJ011284, KJ011364, KJ011340; Costus spiralis Roscoe, M9335, KJ011466, KJ011210, KJ011285, KJ011407, -; Costus stenophyllus Standl. & L. O. Williams, L75.0305, KJ011467, KJ011211, KJ011286, KJ011408, AY994560 (CS02-313); Costus talbotii Ridl., 2003-0109009, KJ011471, KJ011216, KJ011291, KJ011412, KJ011343; Costus aff. tappenbeckianus, WJK94-3697, AY994715, KJ011215, KJ011290, -, AY994553; Costus aff. tappenbeckianus, TvA3675, KJ011468, KJ011212, KJ011287, KJ011409, -; Costus aff. tappenbeckianus , US235, KJ011469, KJ011213, KJ011288, KJ011410, KJ011341; Costus aff. tappenbeckianus , US236, KJ011470, KJ011214, KJ011289, KJ011411, KJ011342; Costus varzearum Maas, WJK94-5379, AY994722, KJ011217, KJ011292, KJ011413, AY994551 (CS01-277); Costus villosissimus Jacg., KMN632, AY994713, KJ011218, KJ011293, KJ011414, AY994550; Costus vinosus Maas, M9568, KJ011472 (WJK94-3727), KJ011219, KJ011294, -, KJ011344; Costus

!!

! 165!

wilsonii Maas, M9507, KJ011473, KJ011220, KJ011295, KJ011415, -; Costus woodsonii Maas, CS01-264, AY994712, KJ011221, KJ011296, KJ011416, AY994549; Costus aff. lucanusianus (yellow flower form), Breteler5297, KJ011454, KJ011191, KJ011267, KJ011390, KJ011329; Costus zingiberoides J. F. Macbr., BTM86-00-01, KJ011474, KJ011222, KJ011297, KJ011417, KJ011345; Dimerocostus strobilaceus Kuntze, CS01-272, KJ011478, KJ011227, KJ011302, KJ011422, KJ011350; Monocostus uniflorus (Peterson) Maas, WJK75-0065, KJ011477, KJ011226, KJ011301, KJ011421, KJ011349; Tapeinochilos queenslandiae K. Schum., DavisB68.168, KJ011480, KJ011229, KJ011303, KJ011424, AY994542 (Hay7052)

! 166!

Appendix 2

Supplementary Table 1. Primer pairs developed in this study and respective annealing temperatures.

Molecular Marker Direction Sequence (5' --> 3') Annealing

temperature (oC)

HS Heat-Shock Protein forward TAG TTG ATC CTA TGG ATG AGA TTG

56 reverse CCA TGT TGG CAG ACC ATC C

BZR Brassinazole-Resistant Protein forward CTT GGC GGA GTT GAA TTT GGT AGG

57 reverse TCT GCG ACG AGG CCG GAT G

RAB18 Phosphatase 2c forward GGG AYA CAT TCA CCA ATC TTT CTG

56 reverse GCA TTG CTC CAC ATT TAC ACG AG

IGPS Imidazole Glycerol-phosphate Synthase forward TCA AGC TTG GAA GTG GCA TCA G

56 reverse ACC TCC TTC CGA TGG AAA ATT CC

WXY Granule-bound Starch Synthase forward GGA ATA GCT GCA GCG AGA ATA TCT

56 reverse GAT CAT CCT WTG TTT CTK GAG AAG G

PSY Phytoene Synthase forward GCT CAT CTG TCC TTC TGC ACC A

57 reverse TCC GCT GGT GTC GAA CCT GCT

E2 Ubiquitin-conjugating Enzyme forward CGA AAC TTC AGA TTG CTC GAA G

54 reverse AAG GTG CCC TCA GGA GG

FT Flowering Locus T forward GAA ATT GTT TGC TAT GAG AGT CCA CG 56 reverse GCA CCT TCT TCC ACC ACA ACC

Ih5GT Anthocyanin 5-glucosyltransferase forward CAT GGT TTG TTC ACG CTG CC 54 reverse GGT GGA ACT CGA CGC TGG A

!!

! 167!

Supplementary Figure 1. Nuclear baits capture specificity. Bars depict the total number of

the filtered reads that map to the targets.

!!

! 168!

Supplementary Figure 2. Plastid baits capture specificity. Bars depict the total number of the

filtered reads that map to the targets.

Biogeography and Diversification of Neotropical Costaceae

Documents

Transcript of Biogeography and Diversification of Neotropical Costaceae