<%BANNER%>

Systematics of Tribe Sobralieae (Orchidaceae)

Permanent Link: http://ufdc.ufl.edu/UFE0044898/00001

Material Information

Title: Systematics of Tribe Sobralieae (Orchidaceae) Phylogenetics, Pollination, Anatomy, and Biogeography of a Group of Neotropical Orchids
Physical Description: 1 online resource (296 p.)
Language: english
Creator: Neubig, Kurt Maximilli
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: anatomy -- biogeography -- elleanthus -- phylogeny -- pollination -- sobralia
Biology -- Dissertations, Academic -- UF
Genre: Botany thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: PollinationBiogeographyElleanthusSobrIncluding Elleanthus,Epilyna, Sertifera, and Sobralia,tribe Sobralieae has over 200 species and is a major constituent of theNeotropical flora.  However, little isknown of the phylogenetic relationships within the group.  Therefore, one nuclear (nrITS) and 5 plastidloci (matK, rpl16, trnL-F, trnS-G, and ycf1) were sequenced to elucidate the phylogeny of this group.  Phylogenetic analysis of these datasetsrevels that Elleanthus, Epilyna, and Sertifera are monophyletic, but that Sobralia is polyphyletic. This polyphyly is centered on Sobraliasect. Sobralia, which itself is notmonophyletic.  Sobralieae have an exclusively Neotropicaldistribution.  However, according to thephylogenetic data, the plesiomorphic condition is South American.  This indicates a South American origin of thetribe, with crown groups that are more widespread.  These data were analyzed in an ultrametricanalysis to determine temporal correlation of these two groups and to estimatethe actual date of these phylogenetic events. These data support these two crown clades diversified and expanded theirrange at about the same time, and possibly during or somewhat before theclosing of the Isthmus of Panama. It has long been recognized that there are bee- andhummingbird-pollinated taxa within Sobralieae. However, it is not been clear how these conditions have evolved, norwhether the mechanisms are by deceit or reward (including composition ofrewards).  To understand the evolution ofpollination within Sobralieae, nectar volume and composition was measured, aswell as nectary structure and composition.  Also, we mapped the known and inferredpollinators of this group.  The resultshows that there are at least four independent evolutionary events that haveled to hummingbird-pollination and that bee-pollination is probablyplesiomorphic within the group.  Also,nectar is primarily produced in bird-pollinated taxa (especially Elleanthus), whereas deceit is theprimary mechanism in bee-pollinated taxa (Sobralia,in part).The vegetative anatomy has been poorly examinedwithin this tribe.  In order to betterunderstand vegetative anatomical variation, leaves, stems and roots weresectioned and compared using light microscopy. In combination, these data lend a better understanding of this richgroup of orchids.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kurt Maximilli Neubig.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Williams, Norris H.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044898:00001

Permanent Link: http://ufdc.ufl.edu/UFE0044898/00001

Material Information

Title: Systematics of Tribe Sobralieae (Orchidaceae) Phylogenetics, Pollination, Anatomy, and Biogeography of a Group of Neotropical Orchids
Physical Description: 1 online resource (296 p.)
Language: english
Creator: Neubig, Kurt Maximilli
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: anatomy -- biogeography -- elleanthus -- phylogeny -- pollination -- sobralia
Biology -- Dissertations, Academic -- UF
Genre: Botany thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: PollinationBiogeographyElleanthusSobrIncluding Elleanthus,Epilyna, Sertifera, and Sobralia,tribe Sobralieae has over 200 species and is a major constituent of theNeotropical flora.  However, little isknown of the phylogenetic relationships within the group.  Therefore, one nuclear (nrITS) and 5 plastidloci (matK, rpl16, trnL-F, trnS-G, and ycf1) were sequenced to elucidate the phylogeny of this group.  Phylogenetic analysis of these datasetsrevels that Elleanthus, Epilyna, and Sertifera are monophyletic, but that Sobralia is polyphyletic. This polyphyly is centered on Sobraliasect. Sobralia, which itself is notmonophyletic.  Sobralieae have an exclusively Neotropicaldistribution.  However, according to thephylogenetic data, the plesiomorphic condition is South American.  This indicates a South American origin of thetribe, with crown groups that are more widespread.  These data were analyzed in an ultrametricanalysis to determine temporal correlation of these two groups and to estimatethe actual date of these phylogenetic events. These data support these two crown clades diversified and expanded theirrange at about the same time, and possibly during or somewhat before theclosing of the Isthmus of Panama. It has long been recognized that there are bee- andhummingbird-pollinated taxa within Sobralieae. However, it is not been clear how these conditions have evolved, norwhether the mechanisms are by deceit or reward (including composition ofrewards).  To understand the evolution ofpollination within Sobralieae, nectar volume and composition was measured, aswell as nectary structure and composition.  Also, we mapped the known and inferredpollinators of this group.  The resultshows that there are at least four independent evolutionary events that haveled to hummingbird-pollination and that bee-pollination is probablyplesiomorphic within the group.  Also,nectar is primarily produced in bird-pollinated taxa (especially Elleanthus), whereas deceit is theprimary mechanism in bee-pollinated taxa (Sobralia,in part).The vegetative anatomy has been poorly examinedwithin this tribe.  In order to betterunderstand vegetative anatomical variation, leaves, stems and roots weresectioned and compared using light microscopy. In combination, these data lend a better understanding of this richgroup of orchids.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Kurt Maximilli Neubig.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Williams, Norris H.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044898:00001


This item has the following downloads:


Full Text

PAGE 1

1 SYSTEMATICS OF TRIBE SOBRALIEAE (ORCHIDACEAE): PHYLOGENETICS, POLLINATION, ANATOMY, AND BIOGEOGRAPHY OF A GROUP OF NEOTROPICAL ORCHIDS By KURT MAXIMILLIAN NEUBIG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVE RSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Kurt Maximillian Neubig

PAGE 3

3 To all the magnificent people of this plane t without whom I would be alone

PAGE 4

4 ACK NOWLEDGMENTS I would like to thank my wife, Julie Kay, and my son, Henry George, for their unending efforts to prolong this dissertation. Henry and Linda Neubig my parents, have also been an unending source of emotional and monetary support and deserve the mos t credit for my putative intellect and exceptional modesty I also thank Norris Williams, my advisor, and the other members of my graduate committee: Pam Soltis, Rob Ferle, and Walter Judd. Unfortunately, although W. Mar k Whitten and Barbara Carls ward contributed an enormous amount towards the intellectual and methodological components of this dissertation, they do not get the recognition they deserve as members of my graduate committee. Also, Robert Dressler, the foremost authority on the taxono my of the orchids discussed in this dissertation, has provided endless help in identification and taxonomy. In the course of this dissertation, he published over a dozen papers on the taxonomy of Sobralia and Elleanthus Therefore, I thank him for keepin g me on my toes and for I would like to thank my fellow graduate students for the many inspirational conversations, highly intellectual interactions, and professional attitud es They are Lorena Endara, Mario Blanco, Iwan Molgo, Paul Corogin, J. Richard Abbott, Gretchen Ionta, Charlotte Germain Aubrey, and Lucas Majure among many others This dissertation could not have been possible without the collections made by many res earchers including, but not limited to Mark Whitten, Mario Blanco, Samantha Koehler, Cassio van den Berg, Delsy Trujillo, Franco Pupulin, Andres Maduro, Robert Dressler, and Tom Mirenda.

PAGE 5

5 In the course of this dissertation, I have been distracted by multipl e side projects and for which I am eternally grateful Stuart McDaniel has been the driving force in duping me into working on non vascular plants. Sharon Talley (USDA) influenced me into working on Mikania (Asteraceae). Norris Williams and Mark Whitten urged me into working on Oncidiinae (Orchidaceae). Fabiany Herrera used hi s shiny head to distract me while he tempted me with an Ulmaceous agenda. J. Richard Abbott, perhaps the largest offender of mass distraction, has diverted me in work on Asimina Polygalaceae, and on a total Florida floristic DNA project we have deemed Barcoding the Flora of Florida (a.k.a., BarFF). Skip Blanchard has been highly successful in tricking me into working on various groups of Malvaceae. Emily Schwartz, then undergrad uate, has commiserated tirelessly on that Malvaceae work also. This work could not have been possible without the support of the FLAS herbarium. In that regard Kent Perkins and Trudy Lindler have been inspirational and supportive terriers nipping at my h eels. I would also like to thank Savita Shanker and Patrick Thimote of the Interdisciplinary Center for Biotechnology Research for their tireless efforts in the sequencing core that have facilitated this dissertation and other projects in producing a large quantity of data Computation time was provided by the FLMNH Phyloinformatics Cluster for High Performance Computing in the Life Sciences awarded to Pam and Doug Soltis with technical assistance provided by Matt Gitzendanner. Many different individuals, especially Charlotte Germain Aubrey, Julie Allen, J. Gordon Burleigh, and Lorena Endara, provided analytical assistance.

PAGE 6

6 This dissertation was supported by the many teaching assistantships through the extinct Botany department, and extant Biology departme nt, at UF. Additional support was provided by NSF through research assistantships (though grants IOB 0543659 and DEB 0234064). I received funding through the Vaughn Jordan Fellowship in orchid biology in the form of fellowships, from the American Orchid Sobralia the th World Orchid Conference Fellowship for research funds. Additional funding was provided by USDA APHIS in a coopera tive agreement for Mikania micrantha foundation and the Global Plants Initiative.

PAGE 7

7 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ .......... 10 LIST OF FIGURES ................................ ................................ ................................ ........ 11 LIST OF ABBREVIATIONS ................................ ................................ ........................... 17 ABSTRACT ................................ ................................ ................................ ................... 18 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 20 Background ................................ ................................ ................................ ............. 20 Phylogenetics and Classification of Sobralieae ................................ ...................... 21 Geography ................................ ................................ ................................ .............. 23 General Vegetative Morphology and Anatomy ................................ ....................... 24 General Floral Morphology and Pollination ................................ ............................. 25 Why Study Sobralieae? ................................ ................................ .......................... 27 2 PHYLOGENETICS AND TAXONOMY OF SOBRALIEAE ................................ ...... 40 Background ................................ ................................ ................................ ............. 40 Materials and Methods ................................ ................................ ............................ 41 Taxon Sampling ................................ ................................ ............................... 41 Types and Nomenclature ................................ ................................ ................. 41 Extractions, Amplification, and Sequencing ................................ ...................... 42 Data Analysis ................................ ................................ ................................ ... 45 Results ................................ ................................ ................................ .................... 46 Phylogenetic Information ................................ ................................ .................. 46 Sequence Divergence Heterogeneity ................................ ............................... 48 Discussion ................................ ................................ ................................ .............. 49 Phylogenetic Relationships ................................ ................................ .............. 49 Broad Taxonomic Implications ................................ ................................ ......... 50 Taxonomy of Elleanthus ................................ ................................ ................... 51 Elleanthu s sect. Elleanthus and sect. Chloidelyna ................................ ..... 52 Elleanthus sect. Cephalelyna ................................ ................................ ..... 53 Elleanthus sect. Stachydelyna ................................ ................................ ... 54 Elleanthus sect. Hymenophora and sect. Elongatae ................................ .. 56 Elleanthus sect. Calelyna and sect. Strobilifera ................................ ......... 57 Elleanthus sect. Otiophora ................................ ................................ ......... 58 Elleanthus sect. Laterales ................................ ................................ .......... 58 Taxonomy of Epilyna ................................ ................................ ........................ 59

PAGE 8

8 Taxonomy of Sertifera ................................ ................................ ...................... 59 Taxonomy of Sobralia ................................ ................................ ...................... 61 Sobralia sect. Sobralia ................................ ................................ ............... 63 Core Sobralia (including sect. Racemosae ) ................................ ............... 67 Inflorescence Position, Flower Size, and Why Sobralieae Taxonomy Is Bad? ................................ ................................ ................................ .............. 75 3 BIOGEOGRAPHY OF SOBRALIEAE: ORIGINS IN SOUTH AMERICA AND DIVERSIFICATION IN CROSSING THE ISTHMUS OF PANAMA ....................... 112 Background ................................ ................................ ................................ ........... 112 Materials and Methods ................................ ................................ .......................... 116 Distribution Maps ................................ ................................ ............................ 116 Mesquite and R8s Analyses ................................ ................................ ........... 117 Results ................................ ................................ ................................ .................. 117 Discussion ................................ ................................ ................................ ............ 118 Limiting Factors of Distribution? ................................ ................................ ..... 118 Geographic Origin ................................ ................................ .......................... 122 Dating in Sobralieae ................................ ................................ ....................... 123 4 REWARD AND DECEIT P OLLINATION IN SOBRALIA AND ELLEANTHUS : A COMPARISON OF NECTAR PRODUCTION AND ADAPTATION THROUGH FLORAL ANATOMY ................................ ................................ ............................. 141 Background ................................ ................................ ................................ ........... 141 Materials and Methods ................................ ................................ .......................... 146 Nectar Volume and Quantity ................................ ................................ .......... 146 Floral Anatomy ................................ ................................ ............................... 146 Floral Fragrances ................................ ................................ ........................... 147 Elevational Variation by Pollination Syndrome ................................ ............... 147 Reconstruction of Pollination Related Char acters ................................ .......... 147 Results ................................ ................................ ................................ .................. 148 Overall Floral Syndrome ................................ ................................ ................. 148 Callus Struct ure ................................ ................................ .............................. 148 Starch ................................ ................................ ................................ ............. 149 Double Cuniculus ................................ ................................ ........................... 149 Nectary and Nectar ................................ ................................ ........................ 150 Fragrance Production ................................ ................................ ..................... 150 Elevational Maps ................................ ................................ ............................ 151 Discussion ................................ ................................ ................................ ............ 151 The Callus ................................ ................................ ................................ ...... 152 The Double Cuniculus ................................ ................................ .................... 156 Nectar Quantification ................................ ................................ ...................... 157 Fragrances ................................ ................................ ................................ ..... 160 Floral Structure as It Relates to Pollination ................................ ..................... 162 M imicry and Deceit ................................ ................................ ......................... 168 Sphingophily ................................ ................................ ................................ ... 172

PAGE 9

9 Elevation and Pollinators ................................ ................................ ................ 173 Evolution of Pollination Syndromes ................................ ................................ 174 Future Directions ................................ ................................ ............................ 178 5 DESCRIPTIVE VEGETATIVE ANATOMY IN SOBRALIEAE ................................ 210 Background ................................ ................................ ................................ ........... 210 Materials and Methods ................................ ................................ .......................... 212 Results ................................ ................................ ................................ .................. 213 Sobralieae ................................ ................................ ................................ ...... 214 Sobralia Sensu Stricto (Core Sobralia ) ................................ ........................... 215 Sobralia ciliata ................................ ................................ ................................ 215 Sobralia dichotoma Clade ( S. caloglossa + S. dichotoma + S. mandonii ) ...... 216 Sobralia portillae ................................ ................................ ............................. 216 Sertifera ................................ ................................ ................................ .......... 216 Epilyna ................................ ................................ ................................ ............ 217 Elleanthus ................................ ................................ ................................ ....... 217 Discussion ................................ ................................ ................................ ............ 218 Character Evolution ................................ ................................ ........................ 218 Anatomy and Morphology as They Relate to Epiphytism ............................... 223 6 CONCLUSIONS ................................ ................................ ................................ ... 268 Discussion ................................ ................................ ................................ ............ 268 Key to the Proposed Genera of Sobralieae: ................................ ......................... 270 LIST OF REFERENCES ................................ ................................ ............................. 274 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 296

PAGE 10

10 LIST OF TABLES Table page 2 1 Species names, DNA number voucher information, herbarium of de position, and country of origin for all material sequenced in this study.. ......................... 103 2 2 Subgene ric/sectional classification schemes for Elleanthus ............................ 110 2 3 Subgeneric/sectional classification schemes for all taxa currently referred to as Sobralia in this text ................................ ................................ .................... 111 3 1 Clade age estimates in millions of years using r8s based on the same 6 DNA region maximum likelihood tree.. ................................ ............................. 140 4 1 Previously published accounts of floral visitors and pollinators in Sobralieae. 205 4 2 Observations of n ectar secretion in this study. ................................ ................. 207 4 3 Floral characters as they relate to pollination syndromes in groups within Sobralieae.. ................................ ................................ ................................ ...... 209 5 1 List of species used in this study, voucher numbers, and parts sectioned. ...... 263 5 2 List of vegetative anatomical characters and character states used in phylogenetic reconstruction. ................................ ................................ ............. 265 5 3 Character states for taxa used in ancestral character re construction analyses of Sobralieae. ................................ ................................ ................................ ... 266

PAGE 11

11 LIST OF FIGURES Figure page 1 1 Phylogenetic consensus tree of broad relationships in Orchidaceae based on previously published phylogenetic papers. ................................ ......................... 29 1 2 Epiphytic habit ................................ ................................ ................................ .... 30 1 3 Vegetative features found in Sobralieae ................................ ............................. 31 1 4 Inflorescence structure and position ................................ ................................ ... 32 1 5 Inflorescence structure and position ................................ ................................ ... 33 1 6 General floral features ................................ ................................ ........................ 34 1 7 Lip morphology demonstrating the various ornamentations ............................... 35 1 8 Floral columns, ge neral morphology ................................ ................................ ... 36 1 9 Floral columns within the short lived group of Sobralia showing variation in shape and color as a utility f or species level identification ................................ 37 1 10 Pollinia showing different sizes, textures, and colors ................................ .......... 38 1 11 Fruit and seed structures ................................ ................................ .................... 39 2 1 Photos of plants used in this study ................................ ................................ ..... 78 2 2 Additional phot os of plants used in this study ................................ ..................... 79 2 3 Additional phot os o f plants used in this study ................................ ..................... 80 2 4 Additional photos of plants used in this study ................................ ..................... 81 2 5 Additional photos of plants used in thi s study ................................ ..................... 82 2 6 Additional phot os of plants used in this study ................................ ..................... 83 2 7 Additional phot os of plants used in this study ................................ ..................... 84 2 8 Additional phot os of plants used in this study ................................ ..................... 85 2 9 Phylogram of one of >2000 most parsimonious trees generated from the analysis o f ITS data set ................................ ................................ ...................... 86 2 10 Bootstrap consensus tree from the parsimony analysis of ITS data set. ............ 87

PAGE 12

12 2 11 Phylogram of the single mos t likely tree generated from the analysis of ITS data set. ................................ ................................ ................................ .............. 88 2 12 Bootstrap consensus tree from the likelihood analysis of ITS data set. .............. 89 2 13 Phylogram of one of >2000 most parsimonious trees generated from the analysis of comb ined plastid (5 locus) data set ................................ .................. 9 0 2 14 Bootstrap consensus tree from the parsimony analy sis of combined plastid (5 locus) data set. ................................ ................................ ............................... 91 2 15 Phylogram of the single most likely tree generated from the analysis of combined plastid (5 locus) data set. ................................ ................................ ... 92 2 16 Bootstrap consensus tree from the likelihood analysis of combined plastid (5 locus) data set. ................................ ................................ ................................ ... 93 2 17 Phylogram of one of >2000 most parsimonious tre es generated from the analysis of total combined (6 locus) data set ................................ ...................... 94 2 18 Bootstrap consensus tree from the parsimony analysis of total combined (6 locus) data set. ................................ ................................ ................................ ... 95 2 19 Phylogram of the single most likely tree generated from the analysis of total combined (6 locus) data set. ................................ ................................ .............. 96 2 20 Bootstrap consensus tree fr om the likelihood analysis of total combined (6 locus) data set. ................................ ................................ ................................ ... 97 2 21 Simplified bootstrap consensus cladogram (bootstrap values not shown) from the likelihood analysis of total combined ( 6 locus) data set, showing taxonomy at various levels as discussed in this chapter ................................ .... 98 2 22 Phylograms from maximum likelihood analyses of A) nrITS and B) combined plastid data sets ................................ ................................ ................................ .. 99 2 23 Simplified bootstrap consensus cladogram (bootstrap values not shown) from the likelihood analysis of total combined (6 locus) data set, showing Illustrations of major inflorescence types in So bralieae ................................ .... 100 2 24 Ancestral character state reconstruction of axillary versus terminal inflorescence position ................................ ................................ ....................... 101 2 25 Ancestral character state reconstruction of flower size ................................ .... 102 3 1 Distributional map of Elleanthus referenced in red. Histogram represents minimum elevation recorded for same specimens used to create ma p. ........... 128 3 2 Distributional map of Epilyna ................................ ................................ ........... 129

PAGE 13

13 3 3 Distributional map of Sertifera ................................ ................................ .......... 130 3 4 Distributional map of Sobralia ................................ ................................ ........... 131 3 5 Distributional map of Sobralia ................................ ................................ ........... 132 3 6 Distributional map of core Sobralia (excluding sect. Racemosae ) .................... 133 3 7 All data presented in figures 3 1 through 3 6 are represented in these maps .. 134 3 8 Ancestral characters state (ML) reconstruction of Sobralieae based on geographic distribution using the total evidence tree of the 6 DNA region maximum likelihood analysis ................................ ................................ ............ 135 3 9 Ancestral characters state (MP) reconstruction of Sobralieae based on geographic distribution using the total evidence tree of the 6 DNA region maximum likelihood analysis ................................ ................................ ............ 136 3 10 Chronogra m of the 6 DNA region maximum likelihood tree based on a root age of 51 Ma ................................ ................................ ................................ ..... 137 3 11 Chronogram of the 6 DNA region maximum likelihood tree based on a root age of 61 Ma ................................ ................................ ................................ ..... 138 3 12 Theoretical phylogenetic patterns for biogeographic interactions with a South American origin ................................ ................................ ................................ 139 4 1 Sobralia warscewiczii ( Blanco 2676 ) showi ng the various morphological aspects of flowers associ ated with large bee pollination ................................ .. 180 4 2 Elleanthus sp. ( Neubig 203 ) showing the various morphological aspects of flowers associat ed with h ummingbird pollination ................................ .............. 181 4 3 Sobralia decora ( Whitten 3280 ) flower; a bee pollinated flower with no nectar reward ................................ ................................ ................................ .............. 182 4 4 Sobr alia warscewiczii ( Blanco 2677 ) flower, a bee pollinated flower with no nectar reward ................................ ................................ ................................ .... 183 4 5 Sobralia chrysostoma ( Neubig 213 ) flower, a bee pollinated flower with no nectar reward. ................................ ................................ ................................ ... 184 4 6 Sobralia macrophylla ( Blanco 3022 ) a bee pollinated flower with a nectar reward ................................ ................................ ................................ .............. 185 4 7 Flowers of Sobralia rosea a bee pollinat ed flo wer that produces nectar reward ................................ ................................ ................................ .............. 186

PAGE 14

14 4 8 Flower of Sobralia bouchei ( Blanco 3009 ), a bee pollinated flower that produces nectar rewards ................................ ................................ .................. 187 4 9 Flower of Sobralia callosa ( Blanco 3021 ), a hummingbird pollinated flower that produces nectar rewards ................................ ................................ ........... 188 4 10 Floral morphology of Sobralia amabilis a hummingbird poll inated species that probably produces nectar ................................ ................................ .......... 189 4 11 Flower of Sobralia crocea ( Neubig 206 ), a nectar deceit species with putatively hummingbird pollination ................................ ................................ ... 190 4 12 Floral morphology of S. rarae avis ( Blanco 870 ) ................................ .............. 191 4 13 Flowers of Elleanthus caravata ( Neubig 202 ), a hummingbird pollinated flower that produces necta r rewards ................................ ................................ 192 4 14 Flowers of Elleanthus sodiroi ( Neubig 246 ) a hummingbird pollinated flower that produces nectar rewards ................................ ................................ ........... 193 4 15 Flowers of Sobralia ciliata ( Whitten 3529 ), a flower that produces nectar rewards and putatively hummingbird pollinated ................................ ............... 194 4 16 Flowers of Sobralia caloglossa ( Whitten 3530 ), a flower tha t produces no nectar reward and is bee pollinated ................................ ................................ .. 195 4 17 Flowers of Sobralia mandonii ( Whitten 3531 ), a flower that produces no nectar rewards and is bee pollinated ................................ ................................ 196 4 18 concentration (in brix) for four species of Sobralia ................................ ............ 197 4 19 sucrose concentration (in brix) for four species of Elleanthus ................................ ....... 198 4 20 concentration (in brix) comparing putatively hummi ngbird pollinated (blue diamonds) and bee pollinated (red squares) nectar observed data in this study. ................................ ................................ ................................ ................ 199 4 21 Lips of Sobralia showing putative osmophore ................................ .................. 200 4 22 Sobralia visitation by euglossines ................................ ................................ ..... 201 4 23 Elevational differences of among groups within Sobralieae, separated by assessed pollinator ................................ ................................ ........................... 202 4 24 Phylogenetic tree based on 6 region maximum likelihood analysis (from chapter two) modified to show characteristics of nectary structure, nectar

PAGE 15

15 presence, pollinia color, flower longevity, double cunicu lus, and pollination syndrome. ................................ ................................ ................................ ......... 203 4 25 Phylogenetic tree based on 6 region maximum likelihood analysis (from chapter two) with the character of pollination syndrome, as outlined in this study ................................ ................................ ................................ ................. 204 5 1 TS of leaf sheath and developing leaf blade of two adjacent nodes ( Sobralia lacerate, Neubig 209 ) ................................ ................................ ....................... 226 5 2 TS of leaf sh eaths ................................ ................................ ............................. 227 5 3 TS of leaf blades at the margins ................................ ................................ ....... 228 5 4 TS of leaf blade demonstrating isodiametr ic cells within the mesophyll ........... 229 5 5 TS of leaf blade demonstrating pali sade cells within the mesophyll ................. 230 5 6 Cuticular peels of abaxial epidermis ................................ ................................ 231 5 7 TS of leaf mesophyll and abaxial leaf surface, with focus on stomata .............. 232 5 8 TS of leaves showing bulliform cells of the epidermis ................................ ...... 233 5 9 TS of leaves showing details of primary veins ................................ .................. 234 5 10 TS of leaf mesophyll showing the dense presence of plastids ......................... 235 5 11 Raphide crystals ................................ ................................ ............................... 236 5 12 Lysogenic, irregularly furcated trichomes in Sobralieae ................................ ... 237 5 13 TS of whole mature stems ................................ ................................ ................ 238 5 14 TS of partial mature stems ................................ ................................ ............... 239 5 15 TS of mature stem at the epidermis. ................................ ................................ 240 5 16 TS of stem showing detail of individual vascular bundles ................................ 241 5 18 TS of roots showing variation in cortex ................................ ............................. 243 5 19 TS of roots showing details of velamen ................................ ............................ 244 5 20 TS of roots showing the details of the tilo somes ................................ ............... 245 5 21 Pressed specimen of Epilyna hirtzii ( Whitten 2938 ) showing some synapomorphies for the genus ................................ ................................ ......... 246

PAGE 16

16 5 22 Ancestral c haracter state reconstruction of root tilosome anatomy (character 1). ................................ ................................ ................................ ..................... 247 5 23 Ancestral character state reconstruction of root sclerified idioblasts (character 2). ................................ ................................ ................................ .... 248 5 24 Ancestral character state reconstruction of root pith cell wall thickness (character 3). ................................ ................................ ................................ .... 249 5 25 Ancestral character state reconstruction of stem shape (character 4). ............. 250 5 26 Ancestral character state reconstruction of leaf sheath tubercles (character 5). ................................ ................................ ................................ .... 251 5 27 Ancest ral character state reconstruction of leaf sheath fiber bundle presence (character 6). ................................ ................................ ................................ .... 252 5 28 Ancestral character state reconstruction of leaf sheath air space presence (character 7). ................................ ................................ ................................ .... 253 5 29 Ancestral character state reconstruction of leaf idioblasts with raphides (character 8). ................................ ................................ ................................ .... 254 5 30 Ancestral character state r econstruction of mesophyll composition (character 9). ................................ ................................ ................................ .... 255 5 31 Ancestral character state reconstruction of leaf vernation (character 10). ........ 256 5 32 Ancestral character state reconstruction of leaf margin shape (character 11). 257 5 33 Ancestral character state reconstruction of midrib exsertion (character 12). .... 258 5 34 Ancestral character state reconstruction of leaf bundle caps on midribs (character 13). ................................ ................................ ................................ .. 259 5 35 Ancestral character state reco nstruction of leaf subepidermal parenchyma (character 14). ................................ ................................ ................................ .. 260 5 36 Ancestral character state reconstruction of leaf abscission layer presence between sheath and blade (character 15). ................................ ....................... 261 5 37 Ancestral character (character 16). ................................ ................................ ................................ .. 262 6 1 Phylogenetic tree from 6 locus ML a nalysis of DNA data, outlining generic circumscriptions of current genera (in black boxes) and new genera to be delimited (black text, to right). ................................ ................................ ........... 273

PAGE 17

17 LIST OF ABBREVIATION S MA Million years ago; refers to geologic time befo re present. MYR Million years; refers to an interval of time, irrelevant of present. S L Sensu lato, i n the broad sense; refers to a taxonomic group that is more widely circumscribed or a phylogenetic entity that is larger. S S Sensu stricto, i n the stri ct sense; refers to a taxonomic group that is more narrowly circumscribed or a phylogenetic entity that is smaller.

PAGE 18

18 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requi rements for the Degree of Doctor of Philosophy SYSTEMATICS OF TRIBE SOBRALIEAE (ORCHIDACEAE): PHYLOGENETICS, POLLINATION, ANATOMY, AND BIOGEOGRAPHY OF A GROUP OF NEOTROPICAL ORCHIDS By Kurt Maximillian Neubig December 2012 Chair: Norris H. Williams Majo r: Botany Including Elleanthus Epilyna Sertifera and Sobralia Sobralieae (Orchidaceae) have over 200 species and are a major constituent of the Neotropical flora. However, little is known of the phylogenetic relationships within the group. Therefore one nuclear (nrITS) and five plastid loci ( matK rpl16 trnL F trnS G and ycf1 ) were sequenced to elucidate the phylogeny of this group. Phylogenetic analysis of these data set s reve a ls that Elleanthus Epilyna and Sertifera are monophyletic, but tha t Sobralia is polyphyletic. This polyphyly is centered on Sobralia sect. Sobralia which itself is not monophyletic. Sobralieae have an exclusively Neotropical distribution. However, according to the phylogenetic data, the plesiomorphic geographic regi on is South America indicating a South American origin of the tribe, with more widespread crown groups T hese data were analyzed in u ltramet r ic analyses to determi ne temporal correlation of those two crown groups ( Sobralia and Elleanthus ) and to estimate the actual date of these phylogenetic events. These data support that these two crown clades diversified and

PAGE 19

19 expanded their range s at about the same time and possibly during or somewhat before the closing of the Isthmus of Panama. I t has long been recog nized that there are bee and hummingbird pollinated taxa within So bralieae. However, it is not clear how these conditions have evolved, or whether the mechanisms are by deceit or reward (inc luding composition of rewards). To understand the evolution of pollination within Sobralieae, nectar volume and composition were measured, as well as nectary structure and composition. Also, I mapped the known and inferred pollinators of this group onto cladograms At least four independent evolutionary even ts have led to hummingbird pollination and bee pollination is probably plesiomorphic within the clade Also, necta r is primarily produced in bird pollinated taxa (especially Elleanthus ), whereas deceit is the primary mechanism in bee pollinated taxa ( Sobralia in part). T he vegetative anato my has been poorly examined within this tribe. In order to better understand vegetative anatomical variation, leaves, stems and roots were sectioned and compared using light microscopy. In combination, these data interpreted phylogenetically, lend a better understanding of this diverse group of orchids.

PAGE 20

20 CHAPTER 1 INTRODUCTION Background With over 20,000 species ( Dressler, 1993a ) Orchidaceae represent ~5% of the total di versity of all seed plant species ( Govaerts, 2001 ; Scotland and Wortley, 2003 ) Orchids form the second largest family of flowering plants, exceeded by, or perhaps equal to, the sunflower family (Asteraceae). Orchids have filled many ecological niches, from temperate areas to the tropics, in deserts to rainfore sts, and from sea level to high elev ation pramo The family is most diverse in the tropics, where epiphytism is the dominant habit of growth. Additi onally, orchids are an over 100 million dollar horticultural industry because of their appeal to hobbyist growers and because of the cut flowe r industry ( Griesbach, 2002 ) Furthermore orchids are the subjects of multiple popular magazines because of their horticultural appeal. V irtually all kinds of biological investigations require the understanding of relationships in order to bett er understand evolut ion. B ecause there are so many species of orchids, understanding relationshi ps is even more crucial because the increased number of species increases the complexity of possible relationships. For these reasons, study of the species rich orchid famil y is important, especially in a species level evolutionary context This dissertation focuses on one tribe of orchids, the Sobralieae. Orchids, and in particular Sobralieae, are a poorly understood group of plants. The goals of this study are 1) to produce a robust phylogenetic hypothesis of relationships within Sobralieae; 2) to clarify taxonomic problems; 3) to elucidate patterns of geographic distributions; 4) to describe the pattern s of vegetative and floral anatomy

PAGE 21

21 in a phylogenetic context ; and 5) to cla rify evolutionary patterns of pollination related traits. Phylogenetics and C lassification of Sobralieae The large number of species of Orchidaceae has hindered our understanding of phylogenetic relationships within the family. Studies that included fami ly wide sampling of taxa ( Cameron, 2001 2004 ; Gorniak, Paun, and Chase, 2010 ) have clarified relationships among subfamilies and tribes. Phylogenetic studies reveal a large paraphyletic grade of many tribes within Epidendroideae that share plesiomorphic features (relative to other tribes within the subfamil y ) such as a generally terrestrial habit, a lack of pseudobulbs, and floral f eatures such as the lack of a true viscidium and the lack of a stipe in the pollinarium. These tribes include, but are not restricted to, the Neottieae, Nervilieae, Tropidieae, S obralieae, Arethuseae, and Epidendreae. Unfortunately, relationships among the se tribes are still relatively poorly known (Fig. 1 1). In the course of the past several hundred years, general orchid classification has changed drastically. Very few tropica l orchids were known to science until the mid 1800s when botanists such as Lindley, Reichenbach and others started to describe a very large number of genera and species. This greater knowledge of diversity fostered higher level classifications within Orc hidaceae. Most rece ntly, taxonomists have followed a classification of orchids that includes five subfamilies (i.e., Apostasioideae, Vanilloideae, Cypripedioideae, Orchidoideae, and Epidendroideae). As defined by this c lassification, Epidendroideae are t h e most species rich and include Sobralieae. In general, members of Sobralieae have groups of subfamily Epidendroideae based on the above mentioned plesiomorphic characters.

PAGE 22

22 More recent and objective phylogenetic analy ses using DNA data have demonstrated that Sobralieae are closely related to genera such as Tropidia and Corymborkis ( Cameron et al., 1999 ; Cameron, 2001 2004 ; Neubig et al., 2009b ) A phylogenetic ( Rothacker, 2007 ) The classification of the clade that includes Sobralia and related genera has varied considerably during the last 150 years; differences in taxono mic treatment depended upon which morpholog ical characters were emphasized Sobralieae, d escribed by Pfitzer in 1887, have long been recognized as a natural group, at least in part. For much of its nomenclatural history it has been known as subtribe Sob raliinae (although placed in several different tribes). Schlechter ( 1915 ) placed Elleanthus Sertifera and Sobralia in subtribe Sobraliinae, which he in turn placed in a very large tribe Polychondreae, an artificial group now including various Orchidoideae, Vanilloideae, and Epidendroideae, based on soft pollinia. Dressler ( 1981 ) placed his subtribe Sobraliinae in tribe Arethuseae based on symplesiomorphies such as presence of corms, plicate leaves, and eight soft pollinia (although he also included ab errant genera such as Arpophyllum and Xerorchis ). Dressler ( 1993a ) later placed subtribe Sobraliinae in tribe Epidendreae based on the distinctive velamen and seed morphology. All these previous classifications recognized a natural Sobralieae, but its relationship to other groups was not clear. Because this group is not closely related to other taxa in tribes Epidendreae and Arethuseae, currently accepted classifications elevate the former subtribe Sobraliinae to the r ank o f tribe ( Pridgeon et al., 2005 ) Consisting of ca. 200 species in four genera, Sobralieae are a poorly known group of plants in terms of generic limits and species level taxonomy Sobralieae as

PAGE 23

23 circumscribed in this treatise consi st of only four genera of unequal species richness. Two genera, Elleanthus C. Presl. and Sobralia Ruiz & Pav., each consist of about 100 species. The other two genera, Epilyna Schltr. and Sertifera L indl. & Rchb.f., each have fewer than 10 species. Howe ver, some recent works have subsumed Epilyna into Elleanthus because of the vegetative similarity to Elleanthus sect. Chloidelyna ( Pridgeon et al., 2005 ) Some other alternative genera also have been described for these same taxa. Fo r example, some workers place Sobralia callosa in Lindsayella (as L. amabilis ) and Sobralia amabilis in Fregea (as Fregea amabilis ) ( Brieger, 1983 ) Taxonomic study of Sobralieae has been hindered by the poor quality o f most herbarium specimens ( Dressler, 2009 2011 ) These problems are related to the ephemeral nature of the flowers. Many species of Sobralia produce very short lived, delicate flowers that deliquesce usually at the end of a single d ay. The flowers require special effort to press, dry, and mount to prevent them from either deliquescing in the press or becoming stuck to the pressing paper. Consequently, the herbarium record mostly consists of flowers that have been very poorly preserve d over time. In combination with the large flower size, the net result is a herbarium specimen that has very brittle and thin flowers whose three dimensional structure and internal color patterns are obscured. Geography The tribe is widely distributed in tropical America and is a common element in Neotropical communities in mid eleva tion s of the Andes (1000 20 00 m), but some members c an deviate greatly from that range Many species proliferate in highly disturbed habitats and are therefore commonly encoun tered on roadsides. Sertifera is restricted to relatively high elevations in the northern Andes. Epilyna is found in

PAGE 24

24 southern Central America and northern South America. Elleanthus is distributed throughout tropical America, and Sobralia is similar in d istribution except for the notable absence in the West Indies. The tribe has an interesting biogeographic pattern with some groups being restricted to South America, whereas other groups are distributed throughout the Neotropics. A closer inspection of th e distribution shows that widespread clades are correlated with high species richness in Elleanthus and Sobralia However, no study has ever incorporated phylogenetics and distributional data (in a study of historical biogeography ) Phylogenetic data are essential to address the patterns of distribution among taxa in the tribe. This can be accomplished by various methods, including ancestral character state reconstruction and ultrametric analysis of the phylogenetic data, to correlate patterns of timing and geographic distribution. General Vegetative Morphology and A natomy Sobralieae are mostly terrestrial ( Dunsterville, 1979 ; Pridgeon et al., 2005 ) but can be epiphytic (Fig. 1 2) with many species equally suited to both habits The vegetative morphology has been well characterized in various floristic treatments ( Dressler, 1993b ; Dressler, 2003a ) The tribe is easily recognizable with a typically bamboo like growth architecture most commonly with multiple thin, vertical stems emerging from rhizomes (or sometimes corms) with no pseudobulbs (Fig. 1 3 ). Although many species in the tribe are an average height of two meters, some species can be much smaller (10 cm) or very tall (up to 13.4 m). In fact, one species, Sobralia altissima is the tallest orchid known (13.4 m) ( Bennett and Christenson, 1999 ; Collantes and Leon, 1999 ) There are also some closely related species of So bralia ( as well as Elleanthus ) that are almost as tall ( Dunsterville and Dunsterville, 1982 )

PAGE 25

25 The leaves are usually very thin and the mode of photosynthetic metabolism is always C3 whereas many more advanced e pidendroid taxa have developed thickened and succulent leaves that are associated with CAM metabolism ( Silvera, Santiago, and Winter, 2005 ; Silvera et al., 2010a ) The leaves can be plicate to rar ely conduplicate. Roots can be very thick, sometimes over 1 cm in diameter, with a very thick velamen (outer cell layers), but with few discernable and unique features. Little is known of vegetative anatomical variation in Sobralieae, even though anatomy can provide useful (though cryptic) characters for phylogenetic analysis and identification. Some species of Sobralia and Elleanthus have been examined within broad orchid comparisons ( Solereder and Meyer, 1930 ; Benzing, Ott, and Friedman, 1982 ; Pridgeon, Stern, and Benzing, 1983 ; Moreira and Isaias, 2008 ) but Sobralieae have not been intensely sampled and these features have never been considered in a phylogenetic context. General Floral Morphology and P ollination Floral presentation varies due to the p osition and form of the inflorescence Inflorescences are racemes, panicles, or capitate and may be terminal or axillary or very rarely a short scape like inflorescence is produced from the base of the plant (Fig. 1 4 & 1 5 ). The flowers are spirally ar ranged or distichous. Flowers of Sobralieae have man y of the typical features of orchid s (Fig. 1 6 ). Sobralia for example, bears a striking resemblance to other orchids such as Vanilla Mill. Epistephium Kunth and Cattleya Lindl They are zygomorphic, both structurally and in color pattern. The flowers may be resupinate or non resupinate and range drastically in size (less than 1 cm across to more than 20 cm). The color of the flowers

PAGE 26

26 range s from white, yellow, orange, magenta, to purple, all in a wi de variety of combinations. The three sepals are similar to each other, with two lateral s and one dorsal. The three petals are differentiated, with two laterals and one ventral, the latter always differentiated into a labellum (lip). As with many other o rchids, the lip can have calli, which consist of raised tissue on the adaxial surface (such structures are not homologous across orchids). In Sobralieae, the calli are p aired structure s that have differentiated into various shapes, sizes and cellular com positions. The calli are imp or tant in pollination because their function can be in nectar secretion. Lips can be additionally ornamented with keels the ridged structures running longitudina lly along the middle of the lip (Fig. 1 7 ). As in all orchids, t he reproductive parts (andr oecium and gynoecium) are fused, forming a compound structure called the column (Figs. 1 8 & 1 9 ). The column in Sobralieae is similar to that of many other Epidendroideae in having only a single fertile anther that is dorsally disposed with two wings that are commonly described as staminodes. The stigma is ventral to the anther in orientation within the column, sometimes with an infrastigmatic ligule. In Sobralia there is often a rostellum consisting of an elastic flap, which is distal to the stigma. In Elleanthus the stigma is forward facing with no rostellar flap. The column form can also be very useful in species identification, especially within Sobralia (Fig. 1 9). The pollinia range from approx imately 1 10 mm in widt h (Fig. 1 10 ). The ir texture and size can be either relatively hard and small, or mealy and large. Also, the color can be bluish purple or whitish yellow; the darker colors occur in hummingbird pollinated

PAGE 27

27 taxa and are hypothesized to reduce loss of polli nia from birds grooming their beaks ( Dressler, 1971 ) Although pollination in Sobralia and Elleanthus has not received significant attention some aspects of pollination are known. T he general morphology of the f lowers is very well known; because of their showy flowers, many species have been very well illustrated. Flowers of Sobralia are typically bee pollinated and can have red, pink, orange, yellow, purple or white perianths that have distinct nectar guides o n the lip that are contrasting in color, with very few species producing nectar. Elleanthus and Sertifera are typically hummingbird pollinated and all have relatively reduced flowers that are often brightly colored (magenta, yellow, orange or white), wit h a thick callus that produces copious nectar. The pollination of Epilyna is still unknown. Whereas many angiosperm flowers produce nectar via nectaries supplied by phloem, most Sobralieae that secrete nectar do so via starch filled calli on the lip that rapidly convert starch to sugars. This research attempts to elucidate the patterns of nectar sec re tion, nectar composition, floral anatomy, and the evolution of pollination systems. Fruits range greatly in size (Fig. 1 11) from 1 cm (especially in Elleant hus Epilyna and Sertifera ) to over 15 cm (only in Sobralia ) in length The fruits are often strongly ribbed and dehisce along six valves. Seeds are usually fusiform especially in Sobralia and are Bletia species of Elleanthus hav Elleanthus seeds that are more similar to Laeliinae ( Dressler, 1993a ; Prutsch, Schardt, and Schill, 2000 ) Why S tudy Sobralieae? Orchids, and in particular Sobra lieae, are still a poorly understood group of plants. This study aims to provide phylogenetic patterns, tackle taxonomic problems, elucidate

PAGE 28

28 patt erns of geographic distribution describe veget ative and floral anatomy, and clarify patterns in the evolution of pollination syndromes

PAGE 29

29 Figure 1 1. Phylogenetic consensus tree of broa d relationships in Orchidaceae based on previously published phylogenetic papers ( Chase et al., 2003 ; Neubig et al., 2009b ; Gorniak, Paun, and Chase, 2010 ) Note that, although the relationships among subfamilies are well resolved, t he relationships among the many tribes of subfamily Epidendroideae are still largely unresolved. The study focus, Sobralieae is bolded; asterisks indicate tribes sampled in this study.

PAGE 30

30 Figure 1 2. Epiphytic habit. A) Sobralia crocea growing on a rece ntly cleared road side. Though not explicitly epiphytic, many epiphytes like this sp ecies are found on cleared road sides. B) Elleanthus graminifolius on the side of a tree. C) Elleanthus fractiflexus growing within a matrix of mosses and other epiphytes on a lateral branch of a tree. Photos by K. M. Neubig.

PAGE 31

31 Figure 1 3 Vegetative features found in Sobralieae. A) Sobralia powellii showing mottling common on sheaths; B) Sobralia atropubescens showing simple pubescence; C) Elleanthus caravatus show ing simple pubescence; D) Sobralia yauaperyensis showing purple leaf undersides; E) Sobralia callosa showing large brown irregularly stellate hairs and a plicate leaf. Photos by W. M. Whitten and K. M. Neubig.

PAGE 32

32 Figure 1 4 Inflorescence structure an d position. A) Elleanthus fractiflexus with a distichous, terminal, and fractiflex raceme; B) E. lancifolius with a distichous, terminal, and fractiflex raceme that grows perpendicular to the axis of the stem; C D) E aff. lateralis showing whole plant and the short scape like inflorescence that is borne at the base of the plant; E) Mark Whitten holding E. aurantiacus with a characteristic branching habit, with multiple terminal inflorescences and spirally arranged flowers; F G) E. capitatus with capitate i nflorescences and spirally arranged flowers, from the side and top, respectively (note the mucilage secreted within the inflorescence). Photos by K. M. Neubig.

PAGE 33

33 Figure 1 5. Inflorescence structure and position. A) Elleanthus oliganthus with terminal spiral racemes; B) Epilyna hirtzii with terminal distichous racemes; C) Sertifera sp., with axillary secund racemes; D) Lorena Endara photographing Sobralia dorbignyana with terminal panicles; E) Sobralia ciliata with axillary racemes; F) J. Richard Abb ot holding Sobralia caloglossa (left), with axillary racemes, and S. mandonii (right), with axillary panicles; G J) terminal and Sobralia recta ; H) Sobralia macrantha ; I) Sobralia chrysostom a ; J) Sobralia lancea Photos by R. Arevalo, C. H. Dodson, W. M. Whitten, and K. M. Neubig.

PAGE 34

34 Figure 1 6 General floral features A) Elleanthus cynarocephalus longitudinal section; B) Sobrali a macrophylla S. mucronata and S. citrea from left to r ight, showing a portion of the range of flower sizes; C) Sobralia warszewiczii longitudinal section; D) Sobralia macrophylla cross section at base of perianth. Scale bars = 1 cm. Photos by W. M. Whitten and K. M. Neubig.

PAGE 35

35 Figure 1 7 Lip morphology demonstrating the various ornamentations. A) Elleanthus sp., with a single bulbous and fused callus; B) Elleanthus caravatus with two bulbous scrotiform calli; C) Elleanthus longibracteatus with two bulbous scrotiform calli; D) Elleanthus cynarocephalu s with two bulbous scrotiform calli; E) Sobralia mandonii with two bulbous scrotiform calli; F) Sobralia macrantha with two raised ridge like calli; G) Sobralia warszewiczii with two raised ridge like calli; H) Sobralia sp., with two raised ridge like calli; I) Sobralia caloglossa with irregular finger like projections; J) Sobralia mucronata with raised and undulating keels; K) Sobralia callosa with a single raised undulating keel; L) Sobralia bouchei with the lip in side view showing highly undulat e margin and recurved apex. Photos by W. M. Whitten and K. M. Neubig.

PAGE 36

36 Figure 1 8. Floral columns, general morphology A, B) Elleanthus sp., side and ventral surfaces, respectively; C) Elleanthus sodiroi ; D) Sobralia ciliata ; E) Sobralia caloglossa ; F ) Sobralia mandonii ; G) Sobralia rosea ; H) Sobralia sp., with anther intact; I) Sobralia sp., with anther removed; J) Sobralia sp., side view. Photos by W. M. Whitten and K. M. Neubig.

PAGE 37

37 Figure 1 9 Floral columns within the short lived group of Sobrali a showing variation in shape and color as a utility for s pecies level identification. A, B) S. andreae ; C) S. bouchei ; D) S. callosa ; E) S. chrysostoma ; F) S. citrea ; G) S. crocea ; H) S. decora ; I) S. ecuadorana ; J) S. exigua ; K) S. klotzscheana ; L) S. kr uskayae ; M) S. lacerata ; N) S. lindleyana ; O) S. macrantha ; P) S. macrophylla ; Q) S. mucronata ; R) S. powellii ; S) S. rarae avis ; T) S. recta ; U) S. theobromina ; V) S. violacea ; W X) S. warszewiczii ; and Y) S sp. Photos by W. M. Whitten and K. M. Neubig.

PAGE 38

38 Figure 1 10 Pollinia showing different sizes, textures, and colors. A B) note the relatively small size, hardened texture and purple color o f hummingbird pollinated taxa: A) Elleanthus caravatus ; B) Sobralia callosa ; C G) Note the relatively large m ealy, and yellowish white color: C) Sobralia bouchei ; D, E) Sobralia warszewiczii dor s al view and side view, respectively; F) Sobralia sp.; G) Sobralia caloglossa Scale bars = 1 mm. Photos by W. M. Whitten and K. M. Neubig.

PAGE 39

39 Figure 1 11 Fruit an d seed structures. A C) Infructescences with every flower setting fruit. A) Elleanthus sp.; B) E. lancifolius ; C) E. capitatus ; D F) Sobralia bouchei ; D) Fruit showing dehiscence scale bare = 1 cm ; E) close up of placental surface with seeds attached; F) f usiform seeds scale bar = 1 mm G) Seeds of S. citrea with the seed coat removed, showing the embryo, SEM, scale bar = 0.5 mm; H) Seeds of S. gloriana SEM, scale bar = 0.5 mm. Photos by K. M. Neubig.

PAGE 40

40 CHAPTER 2 PHYLOGENETICS AND TAXONOMY OF SOBRALIE AE Background The purposes of this chapter are to utilize a molecular based phylogenetic hypothesis to resolve taxonomic problems, to analyze patterns of morphological diversification, shifts in pollination syndrome, changes in geographic ranges over time, and to demonstrate the evolution of this rich and variable clade Although this study is aimed at creating a phylogenetic framework of the tribe, a secondary goal is to examine taxonom ic circumscriptions at the generic, sectional, and specific level. Th erefore, taxonomy at these levels is discussed within the context of phylogenetic relationships. Some morphological characters, such as inflorescence position and architecture are analyzed through character state reconstruction and are discussed because of taxonomic relevance. Relationships within the tribe have been poorly assessed to date. Szlachetko et al. (2009 ) used nrITS to reconstruct the phylogeny of the tribe. They sampled Elleanthus Sertif era and Sobralia but did not sample Epilyna However, most of their samples were unidentified species, demonstrating generic relationships, but not interspecific relationships. Szlachetko et al. (2009 ) also showed the polyphyly of Sobralia at least partially. Neubig ( 2011 ) demonstrated a phylogenetic hypothesis of the tribe with complete generic sampling, but only us ed three loci (nrITS and two plastid loci, ycf1 and trnS G ). That phylogeny was also limited by the number of species sampled, with only 42 species of Sobralieae. They demonstrated the placement of Epilyna as sister to Elleanthus and also showed the poly phyly of Sobralia

PAGE 41

41 Despite these previous efforts to create a phylogenetic framework of Sobralieae, much could be done to improve the resolution of relationships and the taxon sampling. Therefore, the objective of this study is to reconstruct the phylogen y of the tribe using additional DNA data and to outline species sampled and unsampled per clade especially for the polyphyletic genus Sobralia Materials and Methods T axon S ampling Specimens were obtained from wild collected and cultivated plants (Table 1 ). Sampling of Elleanthus Epilyna Sertifera and Sobralia included 150 accessions in 84 uncertainty). Outgroups (five spp.) included three other genera of basal Epidendro id tribes: Neottieae ( Palmorchis and Epipactis ), Arethuseae ( Bletilla ), and Tropidieae ( Tropidia ). Outgroups were chosen based on phylogenetic placement of Sobralia and Elleanthus in previous work s ( Cameron et al., 1999 ; Whitten, Williams, and Chase, 2000 ; Chase et al., 2003 ; Cameron, 2004 ) Photographs of representative vouchers are presented in Figs. 2 1 through 2 8. Ty pes a nd N omenclature Given that most plant ty pe specimens are too old to provide DNA sequences, it is rare for type specimens to have publi shed molecular data. For some species, I was lucky enough to receive tissue directly from type specimens by way of the collector or the describing author, such as Sobralia quinanta ( Pupulin et al. 3644 ), Sobralia maduroi ( Maduro and Olmos 206 ), Sobralia n utans ( Maduro and Olmos 236 ), Sobralia sanfelicis ( Maduro and Olmos 269 ), Sobralia citrea ( Dressler 6338 ), and Sobralia sororcula ( Dressler 6415 ).

PAGE 42

42 The advantage of sequencing type specimens is obvio us, in that the sequence data are unambiguously linked to the species, except as it relates to synonymy. The second less obvious advantage is that, by making sequences of type specimens available through major public DNA data resources such as GenBank, DNA sequences can be used as a tool for identification of un known or sterile specimens. Many species of Sobralieae described by R. Dressler in the past 15 years have come from greenhouse material cultivated at the Florida Museum of Natural History in Gainesville. For many of these taxa, although I did not sequence the actual holotype, my vouchers are often from cultivated specimens from which the holotype was made (i.e., clonotypes). Examples include: Elleanthus capitatellus Sobralia chrysostoma Sobralia crispissima Sobralia exigua Sobralia gloriana Sobralia kruskayae Sobralia recta and Sobralia theobromina Because of the taxonomic nature of this study, I also include many excerpts from diagnoses or descriptions at various taxonomic le vels. These items are in quotation marks to denote them. These are most ly in English for the ease of the reader, though sometimes translated, especially from Latin. Extractions, A mplification a nd S equencing All freshly collected material was preserved in silica gel ( Chase and Hills, 1991 ) Ge nomic DNA was extracted using a modified cetyl trimethylammonium bromide (CTAB) technique ( Doyle and Doyle, 1987 ) scaled to a 1 mL volume r eaction. Approximately 10 mg o f dried tissue were ground in 1 mL of CTAB 2X buffer and either K. Some total DNAs were then cleaned with Qiagen QIAquick PCR purification columns to remove any inhibitory secon dary compounds. Amplifications were performed using a Biometra Tgradient or an

PAGE 43

43 Eppendorf Mastercycler EP Gradient S thermocycler and Sigma brand reagents in 25 (~10 100 ng), 2 (25 mM), Taq For the plastid regions the following reaction components were used: 0.5 (~10 100 ng), 16 17 2 (25 Taq Sequenced regions included one nuclear (nrITS) and five plastid regions. nrITS (ITS 1 + 5.8S rDNA+ ITS 2) This region was ampli fied with a touchdown protocol using the parameters 94C, 2 min; 15X (94C, 1 min; 76C, 1 min, reducing 1C per cycle; 72C, 1 min); 21X (94C, 1 min; 59C, 1 min; 72C, 1 min); 72C, 3 min with the primers 17SE (ACG AAT TCA TGG TCC GGT GAA GTG TTC G) and 26SE (TAG AAT TCC CCG GTT CGC TCG CCG TTA C) from Sun et al. (1994 ) matK trnK This region includes the matK end of matK trnK the parameters 94C, 3 min; 8X (94C, 30 sec; 60 51C, 1 min; 72C, 3 min); 30X (94C, 30 sec; 50C, 1 min; 72C, 3 min); 72C, 3 min, with the primers 19F ( CGT TCT GAC CAT ATT GCA CTA TG ) from Molvray, Kores, and Chase (2000 ) and trnK2R ( ACC TAG TCG GAT GGA GTA G ) from Johnson and Soltis (1994 ) Additional primers intF ( TGA GCG AAC ACA TTT CTA TGG ) and intR ( ATA AGG TTG AAA CCA AAA GTG ) from Neubig et al. (2012 ) were used for sequencing to attain adequate coverage for overlap of data.

PAGE 44

44 rpl16 This region includes an intron in the rpl16 gene and was amplified using a protocol with the p arameters 94C, 3 min; 33X (94C, 30 sec; 52C, 30 sec; 72C, 2 min); 72C, 3 min, with the primers 71F ( GCT ATG CTT AGT GTG TGA CTC GTT G ) and 1661R ( CGT ACC CAT ATT TTT CCA CCA CGA C ) from Jordan, Courtney, and Neigel (1996 ) trnL F This region includes the trnL intron as well as the trnL F intergenic spacer parameters 94C, 3 min; 30X (94C, 30 sec; 56C, 30 sec; 72C, 1 min); 72C, 3 min, with the primers C ( CGA AAT CGG TAG ACG CTA CG ) and F ( ATT TGA ACT GGT GAC ACG AG ) from Taberlet et al. (1991 ) trnS GCU trnG UCC This intergenic spacer region w as amplified with the parameters 94C, 3 min; 33X (94C, 30 sec; 50C, 30 sec; 72C, 2 min); 72C, 3 min, with the primers trnS GCU (AGA TAG GGA TTC GAA CCC TCG GT) and 3'trnG UUC (GTA GCG GGA ATC GAA CCC GCA TC) from Shaw et al. (2005 ) ycf1 An approximately 1500 base gene was sequenced ( Neubig et al., 2009b ) It was amplified with the same program as matK with primers 3720F (TAC GTA TGT AAT GAA CGA ATG G) and 5500R (GCT GTT ATT GGC ATC AAA CCA ATA GCG). Additional internal primers intF (GAT CTG GAC CAA TGC ACA TAT T) and intR (TTT GAT TGG GAT GAT CCA AGG) were also required for sequencing. co, HCl (pH 8.5) and stored at 4C. Purified PCR products were then cycle sequenced using

PAGE 45

45 the parameters 96C, 10 sec; 25X (96C, 10 sec; 50C, 5 sec; 60C, 4 min), with a mix protocols. Purified cycle sequencing products were directly sequenced o n an ABI 377, 3100 Biosystems, Foster City, CA, USA). Electropherograms were edited and assembled dep osited in GenBank (Table 2 1). Data A nalysis Sequence data were manually aligned using Se Al v2.0a11 ( Rambaut, 1996 ) Indels (insertions/deletions) were not coded as characters. A nalyses were performed using PAUP*4.0b10 ( Swofford 1999 ) Fitch parsimony [unordered characters with equal weights; ( Fitch, 1971 ) ] analyses used a heuristic search strategy which consisted of branch swapping by tree bisect ion reconnection (TBR), Deltran character optimization, stepwise addition with 1000 random addition replicates holding 5 trees at each step, and saving multiple trees (MulTrees). Levels of support were assessed using the bootstrap ( Felsenstein, 1985 ) Bootstrap percentages under parsimony were estimated with 1000 bootstrap replicates, using TBR swapping for 50 random addition replicates per bootstrap replicate. For maximum likelihood (ML), Modeltest ( Posada and Crandall, 1998 ) w as used to determine the appropriate model for analysis using all combined data under the Akaike Information Criterion. M aximum likelihood analyses were combined plastid data gene data set.

PAGE 46

46 Bootstrap percentages under ML were estimated with 100 bootstrap replicates, using TBR swapping for one random addition replicate per bootstrap replicate. All data were included in these analyse s, except in trnL F where bps 345 862 were deleted because of indels that were difficult to align with any certainty. All analyses were performed for data sets including ITS only, plastid only, and all data combined. Data congruence was tested using the partition homogeneity test (ILD ) in PAUP*4.0b10 ( Swofford, 1999 ) as described by Johnson and Soltis ( 1998 ) Heuristic searches for the ILD tests were performed using 100 replicates and TBR branch swapping. Probability values lower than 0.05 were used to identify data sets that were significantly different from one another. Ancestr al character state reconstructions were implemented in Mesquite ( Maddison and Maddison, 2005 ) using likelihood character reconstruction for inflorescence characters on the single most likely tree from the combined 6 locus analysis ( see Fig. 2 19). Results Phylogenetic I nformation The aligned length of the ITS data set was 902 bp. Of these, 323 were parsimony informative (35.8%). Fitch parsimony analysis of the ITS region found 100 equally parsimonious trees (Figs. 2 9 and 2 10) of 1202 steps (consistency index (CI) = 0.525, retention index (RI) = 0.878). Maximum likelihood analysis of ITS only ( lnL = 8156.9; Figs. 2 11 and 2 12) yielded trees similar in topology to parsimony analyses. Bootstrap support for all nodes was similar t o that from parsimony. The aligned length of the combined plastid data set ( matK rpl16 trnL F trnS G and ycf1 ) data set was 7738 bp Of these, 926 were parsimony informative (8.4%). Fitch parsimony analysis of the combined plastid data set found 100 equally

PAGE 47

47 parsimonious trees (Figs. 2 13 and 2 14) of 3331 steps (CI = 0.678, RI = 0.878). Maximum likelihood analysis of plastid data only ( lnL = 34151.4; Figs. 2 15 and 2 16) yielded trees similar in topology to parsimony analyses. Bootstrap support for all nodes was similar to that from parsimony. The aligned length of the total combined (six DNA regions) data set was 8640 bp. Of these, 1249 were parsimony informative (14.5%). Parsimony analysis of all six DNA regions found 100 equally parsimonious tr ees (Figs. 2 17 and 2 18) of 4603 steps (CI = 0.629, RI = 0.873). Maximum likelihood analysis of all six regions ( lnL = 43555.8; Figs. 2 19 and 2 20) yielded trees similar in topology to parsimony analyses. Bootstrap support for all nodes was similar to that from parsimony. Partition homogeneity (ILD) tests showed mixed results for congruence among the different partitions of these data. Because all plastid loci are inherited together on a single molecule, they were treated as a single locus for the pur pose of congruence tests [those examined were not significantly different ( Neubig et al., 2011 ) ]. However, the test comparing ITS and the combined pla stid data showed significant incongruence compared with random partitions of the same size ( P comparison of phylograms and bootstrap percentages between the different trees indicates that there are only a few examples of strong incongruence. For example, Sobralia ciliata Sobralia according to ITS but is sister to the rest of the tribe in the plastid data set. Other incongruences can be found in the relative positions of S. dorbignyana S. portillae S. mandonii S. dichotoma and Sertifera colombiana All data were combined because the pa rtition homogeneity test has been demonstrated to be overly sensitive ( Graham et al., 1998 ; Reeves et al.,

PAGE 48

48 200 1 ) and because a total evidence approach yields highly resolved and relatively strongly supported topology. Additionally, many of those relationships in ITS that differ from plastid relationships are not well supported. With limited outgroup taxon sampl ing, relationships among the basal Epidendroideae tribes Neottieae ( Palmorchis ), Tropidieae ( Tropidia ), Arethuseae ( Bletilla ), and Sobralieae remain unclear. However, Sobralieae are monophyletic in all data sets. Within Sobralieae, there are many consiste nt features among different data sets. Sobralia (see Fig. 2 21), Elleanthus Epilyna and Sertifera are all consistently monophyletic. Incongruent features of phylogenetic topology are centered on Sobralia species within section Sobra lia: S. dichotoma S. ciliata S. dorbignyana S. mandonii and S. portillae the trees; however, their position relative to each other varies among different data sets. Sequence Divergence H eterogeneity These d ata consistently show heterogeneity in sequence divergence among taxa (Fig. 2 22). Although this phenomenon is more pronounced in the ITS data set it is also evident within the plastid data set This inconsistency lies between the two main clades of the tribe : Sobralia Sobralia sect. Sobralia Sertifera Epilyna and Elleanthus The latter clade shows a much higher level of sequence divergence than the former with the net result that the relations Sobralia this is the root cause of the poor ability to differentiate m aterial at the species level within this clade on the basis of sequence data alone (although there is extensive mor phological differentiation) Additional ly, the effect of the non phylogenetic pattern on the normalization of this tree i n an ultrametric analysis will be

PAGE 49

49 discussed in more detail within the context of biogeography and geologic timing (see ch apter 3). Discussion Phylogenetic R elationships Molecular data provide strong support for the monophyly of Sobralieae. However, the outgroups and/or the DNA loci sampled in this study were not sufficient to identify the closest relative of Sobralieae amon particular taxonomic groups. For a more detailed evaluation on the evolution of morphology and anatomy, both floral and vegetat ive, see chapters 4 and 5. Within Sobralieae, the monophyly of all genera is demonstrated, except for Sobralia which has nomenclatural problems that hinder its disintegration. The artificiality of Sobralia has long been recognized ( Reichenbach, 1853 ) and each of the sub clades Sobralia is r eadily diagnosable but Sobralia s.l. has never been spl it taxonomically. Although the eventual goal is to describe ne w genera for the small groups within Sobralia do so now because of the limited taxon sampling in this phylogenetic investigatio n and because of the delay in a nomenclatural vote at the next Internat ional Botanical Conference (IBC, in 2017 ) This relates to the proposal to conserve the name Sobralia with a conserved type ( Dressler et al., 2011 ) which e ssentially asks permission to redesignate the type species for Sobralia so as to preserve the largest monophyletic group as Sobralia Sobralia a generic reclassification is not presented in this disser tation (see chapter 6) but generic and subgeneric relationships are presented as they are currently circumscribed.

PAGE 50

50 Unfortunately, both Sobralia and Elleanthus have non monophyletic sections which also require recircumscription but these changes also wi ll be delayed until further phylogenetic study can be made. Broad Taxonomic I mplications Species of Sobralieae have been described over a long period Although many species have been described from throughout its Neotropical range, Robert Dressler has de scribed the most species in recent years especially from Central America. Therefore, the relatively high number of species from Central America (especially in the genus Sobralia ) represents the focus of study, rather than an accurate representation of ov erall diversity. In light of this Central American diversity, and because the general trend in orchids is to have the greatest diversity in the Andes, many species probably remain to be discovered in South America, where t here has been considerably less c ollecting in the last 30 years. A group of Polish botanists ( Dudek and Tukallo, 2007 ; Dudek and Szlachetko, 2010 ) has published treatments splitting Sobralieae into two subtribes: Elleanth inae and Sobraliinae. Howeve r, both subtribes are non monophyletic according to my analyses Sobraliinae ( Dudek and Tukallo, 2007 ) consists of only one genus, Sobralia which is demonstrated to be polyphyletic according to the analyse s here presented and previous publications ( Szlachetko et al., 2009 ; Neubig et al., 2011 ) Elleanthinae ( Dudek and Szlache tko, 2010 ) consists of 5 genera based on the occurrence of relatively small flowers: Elleanthus Epilyna Evelyna Adeneleuterophora and Sertifera Unfortunately, the recognition of Evelyna (= Elleanthus sect. Cephalelyna ) and Adeneleuterophora (= Ell eanthus sect. Chloidelyna ) makes Elleanthus polyphyletic. Because of the non

PAGE 51

51 monophyletic circumscriptions, these tre atments are not here recognized (i.e., if Evelyna is synonymized with Elleanthus Elleanthus is monophyletic). Taxonomy o f Elleanthus Many species of this group did not start their taxonomic journey in Elleanthus Although Elleanthus was described in 1827, and Evelyna in 1835, many species of what is now recognized as Elleanthus were described as species of Evelyna until 1862, when Reichenba ch (1862) published his treatment of the genus Elleanthus including Evelyna Following Reichenbach, subsequent floristic and taxonomic treatments included Evelyna within Elleanthus For a synopsis of subgeneric classification scheme s of Elleanthus see T able 2 2. The first subgeneric classification was outlined in Reichenbach (1862 ) He described two subgenera Calelyna (with a lip depressed at the base) and Euevelyna (with a paired callus at the base of the lip), the latter with two sections, Cephalelyna (with capitate inflorescences) and Stachydelyna (with spicate inflorescences, mentum small or none, and stems simple or branched). Section Stachydelyna was then further subdivided into sub section Chloidelyna (with m entum narrow and leaves linear) and a series of other undiagnosed, informal groups that appear to be artificial in light of recent phylogenetic data: Kermesinae Furfuraceae Oliganthae and Ensatae Sectional classification wit hin Elleanthus s.l. was later put forth by Garay (1978) and soon after by Brieger (1983) (see Table 2 2 for details). Unfortunately, Brieger does not seem to have noticed the sections that Garay described five years earlier. Although it is clear that Bri eger was conver ging on the same general idea in relation to some of the groups that Garay described, he chose different type species. In so doing, Brieger created a classification scheme that more finely dissects the genus into sections. Therefore, my tr eatment (below) follows a 10 section

PAGE 52

52 classification scheme of Elleanthus that mostly follows Garay (1978a ) but also incorporates the work of Brieger (1983), for the sake of completeness. Elleanthus sect. Elleanthus and sect. Chloidelyna Garay ( 1969) originally described sect. Virgatae for what is now recognized as sect. Elleanthus (lectotypified with E. lancifolius ). Section Elleanthus is similar to sect. Chloidelyna in having relatively small plants with fractiflex and distichous inflorescence s. However, sect. Elleanthus differs in having plicate leaves, larger bracts with mottled coloration, and although some species such as E. lancifolius and E. condorensis have white flowers, other species have brightly colored flowers. Additional species in this section include E. virgatus E. ampliflorus E. confusus and E. formosus all with perianths of various shades of red and similar vegetative morphologies. T he species level taxonomy of these red flowered species is likely dubious. In particular, E. virgatus was originally described as Sertifera virgata but it has terminal inflorescences (differing from the typical axillary inflorescences seen in Sertifera ; see later discussion); it was transferred by Schweinfurth (1938 ) to Elleanthus b ut the muriculate leaf sheaths typical of Sertifera makes the placement of this species uncertain Another species, E. bifarius has brightly colored, distichous and tightly imbricate bracts that place it in sect. Elleanthus but it is morphologically distinct from all other species in the section. This study sampled E. lancifolius and E. ampliflorus only, but strongly supports the monophyly of sect. Elleanthus Reichenbach described sect. Chloidelyna with E. graminifolia as the type. This group is diagnosed by having conduplicate and narrow leaves (usually less than 5 mm wide), small white flowers, and distichous inflorescences that can be either loosely fractiflex or tightly imbricate. The flowers are always inconspicuous and white, with

PAGE 53

53 yellow pollinia. In addition to th e type species, this section includes E. fractiflexus E. ligularis E. linifolius E. muscicola E. poiformis E. stolonifer and E. tillandsioides This study includes E. fractiflexus E. graminifolius E. poifor mis and E. stolonifer This section is perhaps the best studied within Elleanthus with regard to species d elimitation, probably due to its relative abundance (both in herbaria and in the field, pers. obs.) compared to other groups of Elleanthus (Fig. 2 2 1 ) ( Ljtnant, 1976 ; Barringer, 1985 ; Dressler and Bo garin, 2007b ) These results weakly support the sister relationship of sect. Chloidelyna and sect. Elleanthus especially according to plastid data, but they are also consistent with inflorescence architecture. Elleanthus sect. Cephalelyna Garay describ ed sect. Cephalelyna ( with E. casapensis as the type ) for the group of species with strongly capitate inflorescences Although other species groups in Elleanthus can have inflorescences almost as s trongly capitate this is the only section with infloresce nces that secrete copious mucilage (jelly). The jelly is secreted by floral bracts and encases the entire inflorescence, and the flowers protrude through the jelly at anthesis. Similar jelly encased inflorescences are found in some bromeliads (also in ex tremely wet forests) and are hypothesized to protect the buds and developing fruits from herbivory. Other species in this section include E. brasiliensis E. capitatus E. cephalophora E. cephalotus E. cynarocephalus E. glomera E. hooker i anus E. killi pii E. sodiroi E. sphaerocephalus and E. zamorensis ( Garay, 1978a ) But some of these species are dubiously recognized. For example, in Ecuador, E. capitatus E. sodiroi E. sphaerocephalus are accepted names, but E. casapensis E. cephalophora and E. z amorensis are treated as synonyms of E. capitatus ( Dodson and Luer, 2010 ) Nir

PAGE 54

54 (2000 ) also treated E. cephalotus as a synonym of E. capitatus Alternatively, E. capitatus is replaced entirely by E. cynarocephalus in Central America ( Dressler, 2003a ) Clearly, much taxonomic work at the species level is needed for this group. This study included E. cephalotus E. cynarocephalus and E. sodiroi with only on e sample per s pecies, so specie s delimitation is not assessed in these data. However, the relative ly stro ng sequence divergence among the se samples indicate s that further phylogenetic work would support some level of species distinctiveness among members of this sectio n. Nonetheless, these data support a monophyletic sect. Cephalelyna Elleanthus sect. Stachydelyna Garay described sect. Stachydelyna with E. lindenii as the type (however, E. lindenii is frequently recognized as a synonym of E. aurantiacus ). This sectio n, as defined by Garay (1978a ) is polymorphic and not surprisingly, polyphyletic (Fig. 2 21). These phylogenetic results indicate that there are at least two groups that correspond to latives. One group consists of species most like the type, with short inflorescences and highly branching stems that are recumbent. Garay included the following species fitting that description: E. phorcophyllus E. hallii E. petrogeiton and E. auranti acus However, the widespread and variable E. aurantiacus has been frequently divided into numerous species, e.g., E. lindenii E. cajamarcae E. bractescens E. gallipanensis E. hallii E. hoppii E. pallidiflorus and E. pallidiflavus ( Dodson and Luer, 2010 ) E lleanthus phorcophyllus also has a branching habit (Fig. 1 4e), branching inflorescences (Fig. 2 3a, b, & c), and orange flowers identical to those of E. aurantiacus and differs only in its slightly narrower leaves; E. phorcophyllus grows on rock faces, so its narrow leaves may be an adaptation to its x eric habitat. Elleanthus flavescens also shares the branching habit

PAGE 55

55 and inflorescence structure of E. aurantiacus but has a yel lowish flower color. This study includes four samples of what is called E. aurantiacus two of which had yellow flowers and two with orange fl owers, and all of which fit the broad vegetative morphological spectrum discussed above These phylogenetic anal yses do not consistently group color forms. And, because color form is the only discernible and continuum o f size and shape), I opt for a broadly defined E. aurantia cus which includes all of the previously mentioned species, except for E. petrogeiton Elleanthus petrogeiton is reported to have greenish white flowers, but is sympatric with E. aurantiacus (type: Pichincha, Ecuador). It might be based on an inaccurate description of a plant of E. aurantiacus with immature flowers, so its status is unclear. The other group that Garay included within sect. Stachydelyna has relatively erect stems that often branch near the apex ( Dodson and Luer, 2010 ) and inflorescences that bend downwards, regardless of stem position (Fig. 2 2 c, d, e and f). He included the following species that fit this description: E. asplundii, E. maculatus E. lupulinus E. arista tus E. xanthocomos E. amethystinoides E. tovarensis E. rhodolepis E. gracilis E. gastrochilus E. magnicallosus E. scopula E. ventricosus and E. reichenbachianus Other species likely fall wi thin this group, but because inflorescence recurvature has not been emphasized in species descriptions and often is not preserved in herbarium specimens, th ose species cannot be attributed at this time. This study included E. amethystinoides E. ensatus E. gracilis E. longibracteatus and E. maculatus Spe cies identifications in this group are very challenging because of the number of species described, and the ir app arent morphological plasticity so the names attributed are

PAGE 56

56 provisional at this point, except for E. longibracteatus which differs in its crea m colored flowers. Elleanthus steyermarkii is aberrant within this group because of its very narrow conduplicate leaves which more resemble those of sect. Chloidelyna ( Barringer, 1987 ) Its floral morphology is not similar to either of these previously mentioned groups Elleanthus sect. Hymenophora and sect. Elongatae Garay (1978a ) described sect. Hymenophora ( with E hymenophorus as the type ) This group is diagnosed by having racemose inflorescences with spirally arranged flowers. Garay also included E vitellinus E ecuadorensis E curtii E discolor E ruizi i and E oliganthus Brieger (1983) separated sect. Elongatae (with E oliganthus as the type) from Hymenophora He differentiated this section from others by its having elongate inflorescences. This section was treated as monotypic by Bri eger, but al classification ( Garay, 1978a ) it is likely that he was unaware of it It is diagnosed by having relatively long internodes and small filiform bracts. From these two sections, this study sampled E oliganthus E hymenophorus and E. discolor Because E oliganthus is sister to sect. Hymenophora sect. Elongatae does not make any other section nonmonophyletic; therefore, this group is probably best treated as a single section because they share many features and are poorly differentiated from each other. In addition to the relatively elongate racemes with spirally arranged flowers, these two sections also tend to have erect inflorescences that are particularly distinct because the internodes are rel atively long and the rachis relatively gracile (as compared to sect. Calelyna which has the most similar

PAGE 57

57 inflorescences). If these two sections were merged, the name sect. Hymenophora should be applied because of priority. Elleanthus sect. Calelyna and s ect. Strobilifera Reichenbach described sect. Calelyn a ( with the type as E. myrosmatis ) Garay (1978a ) also included E. robustus E. aureus E. arpophyllostachys E. vernicosus E. conifer E. weberbaurianus E. smithii and E. caravata Howe ver, E. caravata does not morphologically fit within this group, nor do these phylogenetic data (Fig. 2 20) support its inclusion. In the strict sense (i.e., excluding E. caravata ), this section is monophyletic and is readily identified by the larger size of the plants, long spiral racemes, and lips with a large spot that appears rusty (probably due to a concentration of carotenoids filling papillose epidermal cells). This group is remarkably variable in morphology, and has flowers that frequently change color as they age, making species level identification problematic because flower color is emphasized in species identification. Other species that morphologically fit within this group are E. bradeorum E. columnaris E. glaucophyllus E. norae E. strob ilifer and E. wageneri ( Dunsterville and Garay, 1979 ; Dressler, 2003a ; Dodson and Luer, 2010 ) When Brieger (1983 ) described sect. Strobilifera he designated the type as E. caravata and also included E. crinipes a Brazilian taxon. Garay (1978 a ) considered this species to be in sect. Calelyna Although E. caravata and E. crinipes are sister, my data (Fig. 2 20) support the separation of sect. Strobilifera because those two species are not most closely related to the rest of sect. Calelyna (s ensu Garay). Because of this and despite the overlap in the original sense of sect. Calelyna both sections s hould be recognized. Section Strobilifera is easily identifiable by the deeply lacerate lips, dense hairs that cover the plant, small stature (<1 m), and floral bracts that are relatively

PAGE 58

58 cartilaginous. Only E. caravatus and E. crinipes are unquestionably placed within the section. But Barringer (1987 ) described E. hirsutus citing material that Dodson and Luer (2010 ) called E. caravata This seems to be a misidentification of that partic ular specimen by Barringer, and E. hirsutus is a distinct species, but with unclear phylogenetic affinities. Elleanthus sect. Otiophora Garay (1978a ) described sect. Otiophora (with E. carolii as the type) He also included E. auriculatus and E. blatteus ( Garay, 1 978b ) The original description of this section is somewhat vague, with plants having terminal and spicate inflorescences, flowers on all sides, lateral sepals more or less connate, infrastigmatic ligules angled, and the clinandrium biauriculate. These are not particularly meaningful characters within Elleanthus I attribute the group that includes E. caricoides E. tonduzii and E. tricallosus to this section based on the similarity to previously mentioned species, but this is only preliminary. Howev er, another species that is distantly related, E. capitatellus is morphologically similar in having subcapitate inflorescences and broad and somewhat lacerate lips. At this moment, the placement of E. capitatellus is uncertain. Elleanthus sect. Laterales Garay (1978) described sect. Laterales with E. lateralis as the type, and also includ ed E. rhizomatosus Barringer (1987) described an additional species from Peru, E. hirsutus (see previous discussion of sect. Stachydelyna ), which conforms to the unusua l inflorescence position of this section. In the description of this section Garay mentioned the diagnostic feature of lateral inflorescences. His description is cryptic and unclear, and does not refer to axillary inflorescences, but rather scapose infl orescences

PAGE 59

59 that are closely appressed to the ground (See Fig. 1 4 c and d). However, Barringer (1987) dismantled the small sect. Laterales because of the lack of consistency in E. hirsutus and E. lateralis share lateral inflorescence s they also have the floral morphology of sect. Otiophora (a mesochile with a single transverse ridge and an unwinged column with a slight infrastigmatic ligule). The other species, E. rhizomatosus share s more floral similarities (a mesochile with a pair of transverse ridges and a winged column with no infrastigmatic ligule) to the polymorphic sect. Stachydelyna (Brieger, 1987). Unfortunately, I did not sample any members of this section, so I cannot assess ei t her its monophyly or phylogenetic affinity. Taxonomy o f Epilyna Epilyna consists of only three described species. Epilyna is sometimes treated as a member of the genus Elleanthus because they share many features such as reduced flowers and conduplicate leaves, as in Elleanthus sect. Chloidelyna ( Dressler, 2003a ) However, as Epilyna is sister to Elleanthus and be cause the vegetative features used to unite these two genera are homoplasious, it is kept separate in this treatment, but either treatment is justified on the basis of the ir reciprocal monophyly. Only two species were sampled in this study, E. jimenezii from Central America, and E. hirtzii from Ecuador. The only other species described is E. embreei from Ecuador, which is distinguished by a much larger and less branched habit and an erose rather than lacerate lip margin. Taxonomy o f Sertifera Sertifera includes only six species. Although this genus has small flowers that are similar to Elleanthus this similarity is obviously due to convergence (see Chapter 4

PAGE 60

60 regarding floral convergence). When Reichenbach and Lindley described the genus Sertifera ( Reichenbach, 1877 ) they described two new species, S. purpurea and S. virgata the former with lateral corymbose inflorescences and the latter with terminal racemes. Schweinfurth (1938 ) transferred S. virgata to Elleanthus (based on the termi nal racemes; see previous discussion of E. virgatus in sect. Elleanthus ). Sertifera is diagnosed by having leaf sheaths with tubercles, lateral and secund inflorescences (often with flowers clustered on a long, flattened peduncle), flowers comparable in s ize to Elleanthus magenta and white tepals (mostly), and a saccate lip (Fig. 1 5c). The species recognized are : S. aurantiaca S. colombiana S. grandifolia S. major S. parviflora and S. purpurea One taxonomically aberrant species, Sertifera lehmanni ana (Kraenzl.) Garay, was originally described in the genus Diothonea Lindl., but then later placed in Epidendrum [ E. lehmannianum (Kraenzl.) Hgsater & Dodson ( ) ]. Nonetheless, the Sertifera co mbination for this taxon has been applied in Ecuadorian treatments. In reviewing type specimens of Sobralieae and identification of material used in this study, one aberrant species, Sobralia rigidissima seems to fit better with Sertifera based on morphol ogy, although it has never been attributed to that genus. It was in combination are unique to Sertifera Garay (1978) adds that this species has secund flowers with a compressed peduncle, further supporting the placement of this species within Sertifera However, the type of Sobralia rigidissima bear s a strong resemblance to Sobralia c iliata because they share small reddish flowers (but larger and with longer

PAGE 61

61 s epals and petals than the typical Sertifera ) and axillary inflorescences. Unfortunately, this species was not sampled in this study, so phylogenetic affinities based on DNA dat a are not known. Finally, a lthough it was possible to identify two samples used in this study as Sertifera by vegetative characters, it was not possible to make positive species level identification; therefore, th e two Ecuadorian samples are denoted as S ertifera Taxonomy o f Sobralia Sobralia is a species rich genus with a long and tumultuous history, some species having been shuffled around in many different genera. Although the first species were accurately attributed to what is now known as Sobral ia some unrelated taxa were originally described as Sobralia based on floral convergence (e.g., Prosthechea citrina (La Llave & Lex.) W.E. Higgins). Conversely, Sobralia proper has been described in other genera. The genus Cyathoglottis Poepp. & Endl. w as described in 1836 for just two species, S. candida and S. crocea based on their smaller flowers. Often, species or groups of species with different pollination syndromes have been traditionally placed in different genera because of the ir strikingly di fferent morphology. This historical tendency to recognize genera because of variation in gross floral morphology has been demonstrated to conflict with phylogenetic relationships due to homoplasy in pollination related floral characters. The emphasis on gross floral traits also tends to result in the creation of paraphyletic groups that express the plesiomorphic floral condition. This bias is particularly apparent within Sobralia in the case of S. callosa which has been segregated as Lindsayella Ames & C.Schweinf. because of its distinctive hummingbird floral syndrome, as opposed to the typical bee floral syndrome that is characteristic of most species of Sobralia However, the recognition of Lindsayella would exacerbate the

PAGE 62

62 polyphyly in Sobralia The floral morphology is misleading in this example because be found in the case of Sobralia amabilis This hummingbird pollinated species is sometimes segregated as Frege a Rchb.f. because of the distinct, diminutive magenta flowers. See chapter 4 for a more detailed discussion on floral morphology as it relates to pollination biology. Subgeneric classifications within Sobralia have changed over time (Table 2 3). Reichenbach (1853 ) first proposed an unranked infra generic classification which he called Euso bralia and Brasolia the former referring to the group here designated as Sobralia as sect. Sobralia ). He placed S. dichotoma as the only species of Brasolia Lindley ( 1854) proposed some very informal infrageneric groups, in which he agreed with Reichenbach recognizing Brasolia which was the species S. dichotoma and S. stenophylla But then Lindley listed a series of groups denoted by numbers, letters and asterisks, which are difficult to convey. The first group, other than Brasolia ncluded S. rosea S. chrysantha and S. paradisiaca This refers to what would later be named as sect. Racemosae by Brieger (1983). The last group included on presence of hairs and venation of the lip. These further subdivisions are artificial, so are not enumerated in detail here. However, in this overarching group, he included many species, e.g., S. bletiae S. candida S. chlorantha S. crocea S. decora S.

PAGE 63

63 fenzli ana S. fimbriata S. fragrans S. galeottiana S. klotzscheana S. labiata S. lindleyana S. liliastrum S. macrantha S. macrophylla S. sessilis S. setigera S. violacea and S. warszewiczii All these species that were sampled form a clade in our an alyses, with the exception of S. liliastrum which differs in having an elongate raceme The next to define sections within Sobralia was Brieger (1983 ) He recognized sect. So bralia (for Brasolia ) with lateral and elongate inflorescences, predicated on the typification of Sobralia by Angely (197 3 ) Next, he described sect. Racemosae for species with terminal and racemose inflorescences. He also described sect. Intermediae for species having an elongate basal internode of the inflorescence, condensed internodes above and with small flowers ; s ect. Abbreviatae for species with very short and condensed inflorescences and ephemeral flowers ; and l astly, sect. Globosae in synonymy with Cyathoglottis to include S. candida and S. crocea but his diagnosis does not differentiate this section from sect s. Abbreviatae or Intermediae Dressler (2002 ) in his assessment of the sections of Sobralia largely agreed with B rieger. Because these two treatments of infrageneric classification are the most recent, and because they divide the genus most finely, these will be the focus of the following discussion. Sobralia sect. Sobralia The most taxonomically complex of these se ctions is the paraphyletic sect. Sobralia diagnosed by axil lary and/or elongate inflorescences and with relatively coriaceous, long lived flowers (vs. deliquescent flowers lasting a single day). The genus was typified by Angely (1973 ) by S. dichotoma which by default is also the type species for sect. Sobralia When Ruiz and Pavon first described the genus in 1794,

PAGE 64

64 the y did not designate a type species, nor did they even describe any species at all. It was not until four years later that they described three species, the first of which was S. dichotoma which is presumably why Angely (1973) chose that species. Most sp ecies of Sobralia are not members of sect. Sobralia and members of sect. Sobralia are not commonly collected. Because one goal of taxonomic research is to create a classification system that reflects monophyly yet maintains stability, collaborators and I submitted a proposal to conserve Sobralia using S. biflora a member of the core Sobralia, as type of the genus ( Dressler et al., 2011 ) Because of nomenclatural rules under the International Code of Nomenclature of Algae, Fun gi and Plants it will be some years before a stable and phylogenetically based nomenclatural system can be Sobralia the more extensively discussed proposed nomenclatural changes in chapter 6). However, our propos ed nomenclatural conservation is recommended based on the Nomenclature Committee for Vascular Plants ( Applequist, 2012 ) ; voting will occur in 2017. My sampling of sect. Sobralia included S. caloglossa S. ciliata S. dichotoma S. dorbignyana S. mandonii S. portillae and S. roezlii. Species in sect. Sobralia not sampled include S. altissima S. boliviensis S. cattleya S. hirtzii S. kermesina S. rigidissima (see previous discussion of Sertifera ), S. scopulorum S. sobralioides S. speciosa S. stenophylla S. uribei and S. weberbauriana However, the exact placement of these unsampled species is a major unanswered question in this study, as their relationships are relatively important for redefining generic limits in the tribe. Sobralia cil iata like most other sect. Sobralia has axillary inflorescences. However, it is the only member of that section that has magenta flowers that are

PAGE 65

65 relatively small (perianth less than 2 cm long) with a thickened pad like callus on the lip with numerous r idges on the lip distal to that callus. It is the only putatively hummingbird pollinated member of sect. Sobralia (see chapter 4 for details of pollination biology). This taxon is the source of the most distinct incongruence between data set s ITS seque nce data place it sister to core Sobralia, whereas plastid data place it one node back, sister to the rest of the tribe. The combined analysis (Fig. 2 20) agrees with the plastid data (Fig. 2 16) probably because that data set has more characters th us o utweigh ing the ITS data (Fig. 2 12) Regardless of the position, its unique morphology and phylogenetic isolation suggest that it deserves generic segregation and could be monotypic. However, S. alstromerioides S. boliviensis S. kermesina S. rigidissi ma and S. scopulorum may be closely related to S. ciliata because of floral similarity in size and color, but this has proven to be a homoplasious feature and so may very well be misleading ( Christenson and Moretz, 2003 ) Most species of sect. Sobralia belong to a morphologically cohesive group that resembles Sobralia dichotoma but the species do not form a clade Two clades are paraphyletic relative to Sertifera Sobralia dorbignya na Elleanthus and Epilyna One clade includes at least S. portillae and S. roezlii and the other includes at least S. dichotoma S. caloglossa and S. mandonii Although S. portillae and S. roezlii are strongly supported as sister taxa, they differ gr eatly in inflorescence architecture, making the separation of this paraphyletic part of sect. Sobralia even more complex. Sobralia portillae has terminal racemes, such as the distantly related S. liliastrum ( Christenson, 2003 ) Sobralia roezlii only has axill ary inflorescences (and they are even paniculate) which makes it morphologically more akin to the clade including S.

PAGE 66

66 dichotoma S. caloglossa and S. mandonii ( Christenson, 2002 ) However, the floral morphology of S. portillae is remarkabl y similar to that of S. mandon ii ( Chr istenson, 2003 ) Thus, these two groups of sect. Sobralia are diffic ult to delimit. Finding discre t e morphological characters that clearly disti nguish these groups and specifying the group to which the remaining unpl aced species should be placed are di fficult to determine It is likely that the unplaced species such as S. altissima S. cattleya S. hirtzii S. speciosa S. uribei and S. weberbauriana would belong to one or the other of these two clades. Sobralia dorbignyana is sister to Epilyna and El leanthus Because of its phylogenetic isola tion and unique morphology it is readily recognizable and probably deserves generic segregation. Garay (1978) transferred Chloraea sobralioides to Sobralia and i n so doing, he did not recognize the similarity i n his treatment of S. sobralioides to S. dorbignyana Schlechter described Sobralia kalbeyeri almost concurrently with Kraenzlin, when the latter described Chloraea sobralioides. Puzzlingly, Schlechter noted the perfect similarity of S. dorbignyana to S. kalbeyeri yet still decided to describe the new species. Because of this, S. sobralioides and S. kalbeyeri seem to be clear synonyms of S. dorbignyana; they differ only in subtle differences in flower color. All these share a terminal paniculate inflor escence structure, which is not found in any other known species within the tribe. Sobralia stenophylla endemic to Venezuela and Guyana is the most enigmatic member of sect. Sobralia because its floral morphology is Pleione like (i.e., flower texture del icate, with numerous fimbriae on the central portion of lip) and it has narrow, coriaceous leaves that are unlike those of any other member of Sobralieae. Its axillary

PAGE 67

67 inflorescences are a symplesio morph y that indicates its placement in sect. Sobralia I t was not sampled in this study, but it is a crucial taxon that, given the rampant polyphyly of other members of sect. Sobralia is likely to be an isolated lineage and might deserve generic segregation. Withou t molecular data, I regard it as uncertainly placed Core Sobralia (including sect. Racemosae ) This clade of Sobralia has terminal inflorescences (racemose or with a condensed cone) and usually relatively large flowers. Sobralia liliastrum is sister to the remainder of core Sobralia It has rather simple, usually unbranched terminal racemes, which would probably place it within the general concept of sect. Racemosae but it lacks the large foliaceous bracts so common to that group of species. As mentioned previously, Brieger placed this species wit hin sect. Abbreviatae which is highly discordant with its inflorescence architecture. Sobralia liliastrum shares a similar terminal raceme and floral morphology (large white flowers with yellow keels on the lip) with the putatively closely related S. eli zabethae although I did not sample the latter species If sectional classifications are to be maintained, this species, along with S. elizabethae should be placed in a new separate section. Sobralia sect. Racemosae is a group of species with fractiflex elongate inflorescences and relatively large, boat shaped bracts. Based upon the four species sampled ( Sobralia lindenii, S. luerorum S. pulcherrima, S. rosea), the section is monophyletic. Other species in this section (not sampled) include S. glorios a S. tamboana S. ruckeri and S. paradisiaca Molecular data show that S. pulcherrima and S. lindenii are phylogenetically embedded within S. rosea ; therefore, I lump these two species into a more widely circumscribed S. rosea The floral morphology of S. pulcherrima S. lindenii and S. rosea grade into each other and differ primarily in flower

PAGE 68

68 color (ranging from white to very pink ) and in overall vegetative size (from <1 m to >4 m); see chapter 4 for further detail Sobralia ruckeri has long been sy nonymized with S. rosea However, S. ruckeri has floral morphology that does not match the continuum previously discussed for S. rosea Sobralia ruckeri and S. gloriosa were not sampled in this study, but they are more similar to each other in overall ve getative morphology and in floral coloration patterns to each other (mostly white or pinkish perianth with a dark maroon distal margin of the lip and a yellow throat) than they are to S. rosea Whatever the case, they likely belong within the clade that i ncludes S. rosea and S. luerorum The other sections of core Sobralia are sect s Abbreviatae Intermediae and Globosae Collectively, they form a monophyletic group defined by cone like inflorescences (highly reduced terminal racemes that produce only on e or two flowers at a time). However each of these morphologically diagnosed sections has various degrees problems regarding non monophyly Unfortunately, these sections were poorly defined by Brieger, with overlapping characters and each had very few species directly attributed to them when described. For these sections, he mostly listed only the type species and vague diagnoses, such that it is very difficult to understand his intentions without detailed lists of species for each section: sect. Abbre viatae ( S. fimbriata as type), sect. Intermediae ( S. fragrans as type), and sect. Globosae ( S. candida as type and S. crocea as a species additionally included ). Section Intermediae is particularly hard to understand and because Dressler (2002 ) has taken a broad definition of sect. Abbreviatae including virtually everything in the cone like inflorescence group of taxa (except for sect. Globosae circumscription. It should be mentioned

PAGE 69

69 that recognizing either sect. Globosae or sect. Intermediae makes sect. Abbreviatae non monophyletic. Section Globosae has been defined d ifferently over time. Brieger (1983 ) included only two species when he described this section: S. candida and S. crocea He 2 sect. Globosae does appear to be monophyletic, including S. candida S. lancea S. nutans and S. quinata sample d in this study mo no phyletic when only including the extremely minu te bract group (see Fig. 1 5j), although he also included S. pardalina not sampled in this study ( Dressler, 2002 2003b ) However, because Brieger (1983 ) included S. crocea in this section the section is non monophyletic. Section Intermediae was described with onl y a single species, S. fragrans It is condensed internodes above, and few sect. Intermediae other than the type b ecause of the unique inflorescence features, so this study will take a conservative approach and only assign S. fragrans (sensu Brieger) to this group by adding the species S. bl etiae S. mariannae S. mucronata and S. crocea ( Dressler, 2002 2003b ) This phylogeny sampled only S. mu cronata and S. crocea It is clear that the circumscriptions of these two sections are problematic because of their close ly intertwined evolutionary history. They share relatively small flowers that are usually whitish ( S. crocea is the exception with or ange flowers). At the heart of these sectional

PAGE 70

70 discrepancies is S. crocea a widespread and common species found throughout the Andes. Garay described S. persimilis which I consider to be a synonym of S. crocea It differs only in subtle differences of lip size and shape, but shares the short stature and bright orange, tubular flowers, probably pollinated by hummingbirds (see chapter 4). Section Abbreviatae was defined by Brieger (1983 ) as ortest inflorescences, with condensed internodes, imbricate bracts, the leaf sheaths inflated above, and with 1 Dressler (2002 ) treats most species of the core Sobralia as members of section Abbreviatae It is clear that all of those sectional delimitations are variations on a theme and they have probably been over ly split regarding inflorescence structure Therefore, for the sake of simplicity and because of the consistent cone like inflorescence type throughout this group, I suggest a very broadly defined group sect. Abbreviatae sensu latissimo, with sects. Globosae and Intermediae as synonyms. A broad discussion of this morphologically variable group is most meaningful in the context of phylogenetic groupings, so some distinct morphological and phylogenetic informal groups are discussed below Recta complex Plants in this clade are usually large (1 2 m), flowers are usually nodding and somewhat campanulate, with a white perianth and dark wine red lips. The only two species known to this group are S. recta and S. helleri Although this study only has a single accession of S. helleri the three accessi ons of S. recta are monophyletic. Macrantha complex. The flowers of this clade as the name suggests, are relatively large, but are particularly remarkable in the size of the lip, which can be more than 10 cm long. The two species in this group that wer e sampled, S. macrantha and

PAGE 71

71 S. sanderiana are sister taxa. One species that fits morphologically with these is S. labiata but is not closely related. A s pecies that is likely to be in this complex S. rogersiana differs from S. macrantha only by a sub tle difference in the coloration pattern on the lip. Undatocarinata complex This complex is diagnosed by having an extremely undulate lip margin, with multiple parallel crests running longitudinally within the center of the lip, which are usually strongly raised and also undulating, and broad columns with wings that are splayed out and much longer than other Sobralia groups. This study included S. maduroi S. sanfelicis and S. undatocarinata Dressler (2004b ) described this complex in detail and described a species S. dissimilis in this group as well as a number of other undescribed morphological entities. These analyses (Fig. 2 20) indicate that the complex is monophyletic. Leucoxantha complex Historically, many w hite flowered species of Sobralia were attributed to S. leucoxantha but are not closely related ( Dressler, 2001 ) The clade s swept back lateral sepals and small trumpet shaped lips. Other species in this complex include S. kruskayae S. sotoana S. blancoi S. pendula and S. aspera ( Dressler, 2005 ; Dressler a nd Pupulin, 2008 ; Dressler and Bogarin, 2010 ) This study included only S. kruskayae and S. leucoxantha and with this sampling, the group is monophyletic (Fig. 2 20) Lindleyana complex. Dressler (2007 ) described a group of Sobralia species that have inflated, tubular bracts. These species include S. crispissima S. infundibuligera S. lindleyana S. macrophylla S. madisonii S. rarae avis and S. sororcula ( Dressler,

PAGE 72

72 2007 ) This grouping as a whole is polyphyletic, according to these DNA data. However, a s ubset of these species, S. crispissima S. lindleyana and S. sororcula do form a clade which Lindleyana diagnosed by being relatively small plants (usually less than 40 50 cm tall), with open, cup shaped, l eafy inflorescence bracts. This bract character is obviously homoplasious as it is extremely similar to S. macrophylla ( Dressler, 2007 ) Bouchei complex. This clade is diagnosed by having a magenta perianth, lip with fimbriate margins and a thickened pad like callus, forward facing, slit like stigmas, and large globose shiny anther caps. Also in this group are S. triandra and S. sanderae although these are likely s ynonyms of S. bouchei ( Dressler, 2003a ) Dressler (2003a ) has attributed S. wilsoniana to this group ; however that species does not show any close morphologic al affinity to these species In the course of this stu dy, material belonging to this group was obtained from Colombia (Orquideas del Valle, vendor), labeled as S. rosea This material is obviously not closely related to S. rosea and it is likely an undescribed species of Sobralia sister to the Central Amer ican material of S. bouchei (see S cf. bouchei in trees), but it is not formally described here because of a paucity of herbarium material. This clade is so named because it is a species rich group wit h little sequence diverge nce among the members, yet there is huge variation in plant size, flower size, flower color, and flower shape. Because of the lack of resolution, little can be said of the relationships. However, despite this lack of resolution, some informative aspects can be discussed. Sobralia andreae is unusual in having bright magenta flowers, but with a tubular gullet

PAGE 73

73 ( Dressler, 20 06 ) These data do not support its monophyly, even though the source of all specimen material in Colombia (Orquideas del Valle, vendor) was the same. Another two species, S. setigera and S. virginalis form a clade, although with poor support ; both are found in South America and h ave white flowers (but differ in lip morphology). Sobralia gloriana when originally described, was hy p othesized as a possible natural hybrid between S atropubescens and some unknown species ( Dressler, 2002 ) ; these data somewhat support that hypothesis, because none of the elatives ( McDade, 1995 ) Several complexes are present within the BOGUS group and these make some morphological sense. Warszewiczii complex This group of species is taxonomica lly problematic, as emphasized by Dressler (2011 ) Within this complex, there are some species tha t have purple flowers such as S. warszewiczii S. bradeorum S. violacea and S. amparoae These are possibly distinct, despite overall similarity and general flower color. This is especially true of S. violacea which is not a part of this complex, des pite having extremely similar floral morphology (but note the differences in column structure detailed in Fig. 1 9). Within this mix there are white flowered entities that differ subtly, including a white form of S. warszewiczii S. chrysostoma and S. theobromina Although these latter two species are very similar morphologically, they are phylogenetically distinct from each other. All these taxa have similar column structure and shape, which unites them (note hourglass shape in Fig. 1 9e, u, w, and x). Decora complex This complex is widespread from Mexico to Brazil and species limits within it are undoubtedly poorly defined. There are many included species, e.g.,

PAGE 74

74 S. decora S. sessilis S. yauaperyensis S. galleotiana S. macdougallii and S. fenzliana ( Dressler, 2012 ) Dressler treated some species such as S. panamensis S. neglecta and S. fe nzliana as synonyms of S. decora ( Dressler, 2003a ) but separated S. fenzliana again based on column wing length a nd perianth color ( Dressler, 2012 ) This study sampled S. decora S. fenzliana S. sessilis and S. yauaperyensis, which collectively form a c lade, but species monophyly is not known. The flower color seems to be a subtle gradation across the geographic range, ranging from white to dark purple means that in the region of the inflorescence, the stem will flop over and a new vegetative growth will emerge from a leaf axil, effectively functioning as a stolon. The flowers range in color from whitish to light pink to dark purple. Many of these species also h ave a relatively strong floral fragrance that is likely coumarin. Remaining species. Some additional species of core Sobralia have been treated as different genera. Sobralia callosa has been placed alternatively as Lindsayella amabilis and S. amabilis a s Fregea amabilis Both of these examples have converg ent morphologies to hummingbird pollination syndromes with reduced flower size, bright magenta flowers, column reduction, stigma re orientation, and pollinia that are reduced, condensed and purple. A ll these features are homoplasious and are mechanistically related to pollination by and attraction of hummingbirds. Despite the similarity of these two species, i.e., S. callosa and S. amabilis closest relatives However, this homoplasy illustrate s the remarkable forces that selection plays on the morphology of flowers.

PAGE 75

75 Inflorescence Position, Flower Size, and Why Sobralieae Taxonomy Is Bad? I used ancestral character state reconstruction to demonstrate the evolutionary pattern s of inflorescence position and structure. Inflorescences in Sobralieae may be axillary or terminal. Terminal inflorescences are formed at the apex of a shoot and axillary inflorescences are borne from axillary buds, basal to the shoot terminus. The di stinction between these two positions can be blurred in some plant groups, but in Sobralieae, the difference is usually clear (see Fig. 2 23 for variation in inflorescence structure). However, in a few species (e.g., Sobralia dorbignyana ), both terminal a nd axillary inflorescences are produced, because the inflorescence is a compound panicle with multiple leafy bracts intermingled. Inflorescences also have bracts (leaf derived structures), and these can vary in size and shape. Furthermore, the axis of an inflorescence (i.e., the rachis) may be highly condensed (capitate in some species of Elleanthus ) or elongate, branched or unbranched, erect or (less commonly) nodding, and may have either spiral or distichous phyllotaxy. In a few species of Elleanthus sp ecialized short shoots with reduced leaves bear the (terminal) inflorescences (i.e a scape), whereas the taller, leafy shoots do not produce inflorescences at all. In Sobralieae, most of these inflorescence structural variants exist in some combination. These differences are presented in simplified illustrations (Fig. 2 23). As delimited in Figure 2 23, the core Sobralia is a group distinguished by two main types of inflorescence morphology. Both types are terminal, but in species such as S. rosea and S. luerorum (sect. Racemosae ), the floral displays are strongly distichous and the rachis Sobralia liliastrum also has an elongate inflorescence morphology, but it is not frac tiflex and the

PAGE 76

76 bracts are much reduced. In the remainder of core Sobralia the inflorescence rachis is highly condensed, such that the internodes of the rachis are extremely short (often 1 2 mm). The resulting morphology appears acaulescent, usually with relatively large bracts (except in sect. Globosae sensu Dressler). This condensed inflorescence is present in many Sobralia species with ephemeral flowers. In the combined analysis (Figs. 2 17 through 2 20), Sobralia ciliata is sister to all other Sobra lieae, whereas other members of sect. Sobralia ( S. caloglossa S. dichotoma S. dorbignyana S. mandonii, S. portillae, and S. roezlii ) form a paraphyletic grade relative to the remainder of the tribe. In addition to the genus Sertifera all these species have axillary inflorescences that may or may not branch to form panicles and have relatively small inflorescence bracts except for S. dorbignyana and S. portillae which have termin al inflorescences. Terminal inflorescences are shared with virtually all species of Epilyna and Elleanthus Elleanthus has the most variable inflorescences in the whole tribe, being distichous or spirally arranged, capitate to loosely racemose, and can be oriented downwards, upwards or even horizontally (parallel to the grou nd). Epilyna although similar to Elleanthus has the unique combination of conditions: conduplicate leaves (shared with Elleanthus sect. Chloidelyna ) and non fractiflex, terminal racemes. Sertifera is small in stature (almost always <1 m) and has secund axillary inflorescences. The evolutionary trends within Sobralieae demonstrate the plesiomorphic condition of axillary inflorescences and relatively large flowers (Fig. 2 24). This apparently symplesiomorphic grade across both major clades is represente d by some taxa of sect s Sobralia and Sertifera There has been independent convergence to terminal

PAGE 77

77 inflorescences across both large clades in Sobralieae. Terminal inflorescences are even more homoplasious with the inclusion of S. portillae There has a lso been independent reduction of internodes within the inflorescence to capitate inflorescences although with very different floral presentation strategies. Core Sobralia uses these reduced inflorescences to present a single flower (or two), whereas Ell eanthus presents many flowers at a time. The relative reduction of flower size also seems to have played a role in the evolution of Sobralieae. Because Sobralia is a paraphyletic group relative to Epilyna Elleanthus and Sertifera its morphology is rela tively plesiomorphic (Fig 2 25). T he relatively large flower of Sobralia a plesiomorphic condition, as compared to the much reduced flower s of the other three genera stands in stark contrast. It is this large flower size that has misled taxonomists ca using them to recognize Sobralia despite its obvious artificial and polymorphic nature. Thus Sobralia delimited on the basis of its large flower, in combination with variable inflorescence architecture, is not surprisingly polyphyletic, even though its large flowers are visually striking.

PAGE 78

78 Figure 2 1. Photos of plants used in this study: A) Tropidia polystachya an outgroup, Whitten 2830 ; B) Bletilla striata an outgroup, Neubig 192 ; C) Sobralia ciliata Whitten 2529 ; D) Norris Williams with S. dor bignyana Blanco 3129 ; E) S. dorbignyana Blanco 3129 ; F) S. roezlii no voucher; G) S. portillae Whitten 2433 ; H) S. aff. dichotoma Trujillo 330 ; I) S. mandonii Whitten 3530 ; J) S. caloglossa Whitten 3531 Photos by M. Blanco, D. Trujillo, W. M. Whi tten, and K. Neubig.

PAGE 79

79 Figure 2 2. Additional photos of plants used in this study: A) Elleanthus tricallosus Blanco 2961 ; B) E. sodiroi Neubig 246 ; C) E. longibracteatus Whitten 99205 ; D) E amethystinoides Trujillo 321 ; E) E sp., Trujillo 328 ; F) E. maculatus Trujillo 391 ; G) E. lancifolius Whitten 1575 ; H) E. ampliflorus Blanco 2949 ; I) E. caravata Neubig 202 ; J) E. oliganthus Whitten 2861 Photos by M. Blanco, D. Trujillo, W. M. Whitten, and K. Neubig.

PAGE 80

80 Figure 2 3. Additional photos of plants used in this study: A) Elleanthus aurantiacus (yellow form), Trujillo 308 ; B) E. aurantiacus (yellow form), Whitten 1610 ; C) E. aurantiacus (orange form), Whitten 1611 ; D) E sp., Neubig 203 ; E) E conifer Whitten 1740 ; F) E conifer Whitten 289 9 ; G) E robustus Trujillo 331 ; H) E robustus Trujillo 341 ; I) E robustus Trujillo 367 ; J) E robustus Trujillo 377 ; K) E robustus Whitten 3546 Photos by M. Blanco, D. Trujillo, W. M. Whitten, and K. Neubig.

PAGE 81

81 Figure 2 4. Additional photos of plants used in this study: A) Sobralia liliastrum ; B) S. rosea Whitten 531 ; C) S. rosea (= S. pulcherrima ), Neubig 215 ; D) S. callosa Neubig 224 ; E) S. callosa Blanco 3021 ; F) S cf. bouchei Neubig 208 ; G) S. bouchei Blanco 3009 ; H) S. bouchei Blanc o 3000 ; I) S. amabilis Blanco 2960 ; J) S. exigua Whitten 3003 Photos by M. Blanco, D. Trujillo, W. M. Whitten, and K. Neubig.

PAGE 82

82 Figure 2 5. Additional photos of plants used in this study: A) Sobralia leucoxantha Blanco 2914 ; B) S. macrantha Whitt en 3533 ; C) S. recta Neubig 207 ; D) S. recta Whitten 2851 ; E) S. recta Blanco 2840 ; F) S. macrophylla Blanco 3022 ; G) S cf. mucronata Neubig 228 ; H) S mucronata Neubig 210 ; I) S mucronata Blanco 2971 ; J) S. crocea Neubig 206 ; K) S. quinata Whit ten 2869 Photos by M. Blanco, D. Trujillo, W. M. Whitten, and K. Neubig.

PAGE 83

83 Figure 2 6. Additional photos of plants used in this study: A) Sobralia warscewiczii (white form), Blanco 2689 ; B) S. warscewiczii (purple form), Neubig 211 ; C) S. chrysostoma Neubig 213 ; D) S. warscewiczii (white form), Blanco 2676 ; E) S. warscewiczii (purple form), Whitten 2831 ; F) S. theobromina Blanco 2679 ; G) S. theobromina Whitten 2989 ; H) S. violacea Neubig 218 ; I) S. citrea Whitten 2977 ; J) S. gloriana Blanco 2678 Photos by M. Blanco, D. Trujillo, W. M. Whitten, and K. Neubig.

PAGE 84

84 Figure 2 7. Additional photos of plants used in this study: A) Sobralia yauaperyensis Blanco 3023 ; B) S. decora Whitten 2862 ; C) S. virginalis Whitten 2853 ; D) S. virginalis Whitt en 1589 ; E) S. andreae Neubig 221 ; F) S kerryae Blanco 2943 ; G) S. ecuadorana Whitten 2850 ; H) S gentryi Whitten 2852 Photos by M. Blanco, D. Trujillo, W. M. Whitten, and K. Neubig.

PAGE 85

85 Figure 2 8. Additional photos of plants used in this study: A) Sobralia atropubescens Blanco 3031 ; B) S. sp., Whitten 2863 ; C) S aff. chrysostoma Blanco 2926 ; D) S lacerata Neubig 209 ; E) S. powellii Whitten 3257 ; F) S sp., Neubig 205 ; G) S aff. andreae Neubig 219 ; H) S. klotzscheana Blanco 3011 Photos by M. Blanco, D. Trujillo, W. M. Whitten, and K. Neubig.

PAGE 86

86 Figure 2 9. Phylogram of one of >20 00 most parsimonious trees generated from the analysis of ITS data set Nodes that collapse in the strict consensus are indicated by red arrows.

PAGE 87

87 Figure 2 1 0. Bootstrap consensus tree from the parsimony analysis of ITS data set

PAGE 88

88 Figure 2 11. Phylogram of the single most likely tree generated from the analysis of ITS data set

PAGE 89

89 Figure 2 12. Bootstrap consensus tree from the likelihood analysis of ITS d ata set

PAGE 90

90 Figure 2 13. Phylogram of one of >20 00 most parsimonious trees generated from the analysis of combined plastid (5 locus) data set Nodes that collapse in the strict consensus are indicated by red arrows.

PAGE 91

91 Figure 2 14. Bootstrap consensus t ree from the parsimony analysis of combined plastid (5 locus) data set

PAGE 92

92 Figure 2 15. Phylogram of the single most likely tree generated from the analysis of combined plastid (5 locus) data set

PAGE 93

93 Figure 2 16. Bootstrap consensus tree from the likelih ood analysis of combined plastid (5 locus) data set

PAGE 94

94 Figure 2 17. Phylogram of one of >20 00 most parsimonious trees generated from the analysis of total combined (6 locus) data set Nodes that collapse in the strict consensus are indicated by red arro ws.

PAGE 95

95 Figure 2 18. Bootstrap consensus tree from the parsimony analysis of total combined (6 locus) data set

PAGE 96

96 Figure 2 19. Phylogram of the single most likely tree generated from the analysis of total combined (6 locus) data set

PAGE 97

97 Figure 2 20. Boo tstrap consensus tree from the likelihood analysis of total combined (6 locus) data set

PAGE 98

98 Figure 2 21. Simplified bootstrap consensus cladogram (bootstrap values not shown) from the likelihood analysis of total combined (6 locus) data set showing taxon omy at various levels as discussed in this chapter. Asterisks indicate non monophyletic taxa.

PAGE 99

99 Figure 2 22. Phylograms from maximum likelihood analyses of A) nrITS and B) combined plastid data sets. Note the heterogeneity of sequence divergence among the various clades within the tribe.

PAGE 100

100 Figure 2 23. Simplified bootstrap consensus cladogram (bootstrap values not shown) from the likelihood analysis of total combined (6 locus) data set showing Illustrations of major inflorescence types in Sobralieae although not all types are shown. A) Axillary racemes or panicles (sect. Sobralia ), B) terminal, spiral and erect raceme (various Elleanthus ), C) terminal panicle ( Sobralia dorbignyana ), D E) terminal, fractiflex and/or tightly imbricate raceme ( Elleanth us sect. Elleanthus and sect. Chloidelyna ), F) terminal, fractiflex large bracted raceme ( Sobralia sect. Racemosae Sobralia ).

PAGE 101

101 Figure 2 24. Ancestral character state reconstruction of axillary ver sus terminal inflorescence position, based on the 6 region maximum likelihood tree in Fig. 2 19. A) All portions of the tree, except for core Sobralia B) All of core Sobralia

PAGE 102

102 Figure 2 25. Ancestral character state reconstruction of flower size, ba sed on the 6 region maximum likelihood tree in Fig. 2 the tree, except for core Sobralia B) All of core Sobralia

PAGE 103

103 Table 2 1 Species names, DNA number (usually an abbreviated form of the voucher number used in these trees, Figs. 2 9 through 2 21) voucher information, herbarium of deposition, and country of origin (where known) for all material sequenced in this study. Some sequences we re deposited in G enBank for previous studies; otherwise, SPECIES DNA NUMBER VOUCHER Herbarium Origin ITS trnS G rpl16 3'ycf1 trnL F matK Bletilla striata (Thunb.) Rchb. f. K N192 Neubig 1 2006 FLAS unknown Y Y Y EU490723 Y EU490679 Elleanthus aff. gracilis (Rchb. f.) Rchb. f. W3552 Whitten 3552 FLAS Peru Y Y Y Y Y Y Elleanthus amethystinoides Garay DT321 Trujillo 321 HURP Peru Y Y Y Y N Y Elleanthus ampliflorus Schltr. B294 9 Blanco 2949 FLAS Panama EU490663 Y Y EU490732 Y EU490682 Elleanthus aurantiacus (Lindl.) Rchb. f. W1610 Whitten 1610 FLAS Ecuador Y Y Y Y Y Y Elleanthus aurantiacus (Lindl.) Rchb. f. W1611 Whitten 1611 FLAS Ecuador EU490664 Y Y EU490733 Y EU490683 Ell eanthus aurantiacus (Lindl.) Rchb. f. W2355 Whitten 2355 FLAS Ecuador Y Y Y Y Y Y Elleanthus aurantiacus (Lindl.) Rchb. f. DT308 Trujillo 308 HURP Peru Y Y Y Y N Y Elleanthus capitatellus Dressler KN201 Neubig 201 FLAS Panama Y Y Y Y Y Y Elleanthus cara vata (Aubl.) Rchb. f. KN202 Neubig 202 FLAS unknown Y Y Y Y Y Y Elleanthus caricoides Nash B3055 Blanco 3055 FLAS Panama Y Y Y Y Y Y Elleanthus caricoides Nash B3106 Blanco 3106 FLAS Panama EU490665 Y Y EU490734 Y EU490684 Elleanthus cephalotus Garay & H.R.Sweet N282 Ackerman 3292 NA unknown Y Y Y Y Y Y Elleanthus cf sodiroi Schltr. KN246 Neubig 246 FLAS Colombia Y N Y N N Y Elleanthus conifer (Rchb. f. & Warsz.) Rchb. f. B2527 Blanco 2527 FLAS Ecuador EU490666 Y Y EU490735 Y EU490685 Elleanthus coni fer (Rchb. f. & Warsz.) Rchb. f. W1740 Whitten 1740 FLAS Ecuador Y Y Y Y Y Y Elleanthus conifer (Rchb. f. & Warsz.) Rchb. f. W2899 Whitten 2899 FLAS Ecuador Y Y Y Y Y Y Elleanthus crinipes Rchb. f. SB36 ESA Brazil Y Y Y Y Y Y Elleanthus cynarocephalus Rchb.f. B3105 Blanco 3105 FLAS Panama Y Y Y EU490736 Y EU490686 Elleanthus discolor (Rchb. f. & Warsz.) Rchb. f. KN231 Neubig 231 FLAS unknown Y Y Y Y Y Y

PAGE 104

104 Table 2 1. Continued. SPECIES DNA NUMBER VOUCHER Herbarium Origin ITS trnS G rpl16 3'ycf1 trnL F matK Elleanthus discolor (Rchb. f. & Warsz.) Rchb. f. KN232 Neubig 232 FLAS Colombia Y Y Y Y Y Y Elleanthus ensatus (Lindl.) Rchb. f. W3555 Whitten 3555 FLAS Ecuador Y Y Y Y N Y Elleanthus ensatus (Lindl.) Rchb. f. DT423 Trujillo 423 HURP Peru Y Y Y Y Y Y Elleanthus fractiflexus Schltr. W1517 Whitten 1517 FLAS Ecuador Y Y Y Y Y Y Elleanthus fractiflexus Schltr. W1594 Whitten 1594 FLAS Ecuador Y Y Y Y Y Y Elleanthus graminifolius (Barb. Rodr.) Ljtnant W1542 Whitten 1542 FLAS Ecuador Y Y Y Y N Y Ell eanthus hymenophorus (Rchb. f.) Rchb. f. N563 Maduro & Olmos 245 NA Panama Y Y Y Y Y Y Elleanthus lancifolius C. Presl B2918 Blanco 2918 FLAS Panama Y Y Y Y Y Y Elleanthus lancifolius C. Presl W1575 Whitten 1575 FLAS Ecuador EU490667 Y Y EU490737 Y EU49 0687 Elleanthus longibracteatus (Lindl. ex Griseb.) Fawc. N399 Whitten 99205 FLAS Jamaica? Y Y Y Y Y Y Elleanthus maculatus (Lindl.) Rchb. f. DT328 Trujillo 328 HURP Peru Y Y Y Y Y Y Elleanthus oliganthus (Poepp. & Endl.) Rchb. f. B2528 Blanco 2528 FLAS Ecuador Y Y Y Y Y Y Elleanthus oliganthus (Poepp. & Endl.) Rchb. f. W1502 Whitten 1502 FLAS Ecuador EU490668 Y Y EU490738 Y EU490688 Elleanthus oliganthus (Poepp. & Endl.) Rchb. f. W2861 Whitten 2861 FLAS Ecuador Y Y Y Y Y Y Elleanthus poiformis Schltr B3075 Blanco 3075 FLAS Panama EU490669 Y Y EU490739 Y EU490689 Elleanthus robustus (Rchb. f.) Rchb. f. W3539 Whitten 3539 FLAS Ecuador Y Y Y Y Y Y Elleanthus robustus (Rchb. f.) Rchb. f. W3546 Whitten 3546 FLAS Ecuador Y Y Y Y Y Y Elleanthus robustus (Rchb. f.) Rchb. f. DT331 Trujillo 331 HURP Peru Y Y Y Y Y Y Elleanthus robustus (Rchb. f.) Rchb. f. DT341 Trujillo 341 HURP Peru Y Y Y Y Y Y Elleanthus robustus (Rchb. f.) Rchb. f. DT367 Trujillo 367 HURP Peru Y N Y Y Y Y Elleanthus robustus (Rchb. f.) Rchb. f. DT377 Trujillo 377 HURP Peru Y Y Y Y Y Y Elleanthus sp. KN203 Neubig 203 FLAS unknown Y Y Y Y Y Y Elleanthus sp. W3538 Whitten 3538 FLAS Ecuador Y Y Y Y Y Y

PAGE 105

105 Table 2 1. Continued. SPECIES DNA NUMBER VOUCHER Herbarium Origin ITS trnS G rpl16 3' ycf1 trnL F matK Elleanthus sp. DT391 Trujillo 391 HURP Peru Y Y Y Y Y Y Elleanthus stolonifer Barringer B2934 Blanco 2934 FLAS Panama Y Y Y Y Y Y Elleanthus tonduzii Schltr. N564 Maduro & Olmos 246 NA Panama Y Y Y Y Y Y Elleanthus tricallosus Ames & C. Schweinf. B2961 Blanco 2961 FLAS Panama EU490670 Y Y EU490740 Y EU490690 Elleanthus tricallosus Ames & C. Schweinf. N565 Maduro & Olmos 247 NA Panama Y Y Y Y Y Y Epilyna hirtzii Dodson N526 Hirtz 1741 NA Ecuador Y Y Y Y Y Y Epilyna hirtzii Dodson W 2938 Whitten 2938 FLAS Ecuador Y Y Y Y Y Y Epilyna jimenezii Schltr. B0869 Blanco 869 USJ Costa Rica Y Y Y Y Y Y Epilyna jimenezii Schltr. B2997 Blanco 2997 FLAS Panama Y Y Y Y Y Y Epipactis helleborine (L.) Crantz W3326 Whiiten 3326 FLAS USA Y Y Y EU 490742 Y EU490692 Palmorchis powellii (Ames) C. Schweinf. & Correll V2115 Vargas 2115 NA unknown Y Y Y EU490757 Y EU490697 Palmorchis trilobulata L.O. Williams W3628 Whitten 3628 FLAS Colombia Y N Y Y Y Y Sertifera sp. N546 Hirtz 7160 NA Ecuador Y Y Y Y Y Y Sertifera sp. W2937 Whitten 2937 FLAS Ecuador Y Y Y Y Y Y Sobralia aff. andreae Dressler KN219 Neubig 219 FLAS Colombia Y Y Y Y Y Y Sobralia aff. bouchei Ames & C. Schweinf. KN208 Neubig 208 FLAS Colombia Y Y Y Y Y Y Sobralia aff. chrysostoma Dressler B2926 Blanco 2926 FLAS Panama Y Y Y Y Y Y Sobralia aff. dichotoma Ruiz & Pav. DT330 Trujillo 330 HURP Peru Y Y Y Y Y Y Sobralia aff. maduroi Dressler M&O601 Maduro & Olmos 601 NA Panama Y N Y Y Y Y Sobralia allenii L.O. Williams W2834 Whitten 2834, Whitten 3281 FLAS Panama Y Y Y Y Y Y Sobralia amabilis (Rchb. f.) L.O. Williams B2960 Blanco 2960 FLAS Panama Y Y Y Y Y Y Sobralia andreae Dressler KN221 Neubig 221 FLAS Colombia Y Y Y Y Y Y Sobralia atropubescens Ames & C. Schweinf. B3031 Blanco 3031 FLAS Panama Y Y Y Y Y Y Sobralia bouchei Ames & C. Schweinf. B3000 Blanco 3000 FLAS Panama EU490671 Y Y EU490781 Y EU490708

PAGE 106

106 Table 2 1. Continued. SPECIES DNA NUMBER VOUCHER Herbarium Origin ITS trnS G rpl16 3'ycf1 trnL F matK Sobralia bouchei Ames & C. Schweinf. B3009 Blanco 3009 FLAS Panama Y Y Y Y Y Y Sobralia callosa L.O. Williams B3021 Blanco 3021, Whitten 3275 FLAS Panama Y Y Y Y Y Y Sobralia callosa L.O. Williams KN224 Neubig 224 FLAS unknown Y Y Y Y Y Y Sobralia caloglossa Schltr. W3248 Whitten 3248 FLAS Peru Y Y Y Y Y Y Sobralia caloglossa Schltr. W3531 Whitten 3531 FLAS Peru Y Y Y Y Y Y Sobralia candida (Poepp. & Endl.) Rchb. f. RLD13555 Dressler 13555 NA unknown Y N Y Y Y Y Sobralia cf. mucronata Ames & C. Schweinf. KN228 Neubig 2 28 FLAS Colombia Y Y Y Y Y Y Sobralia cf. withneri D.E. Benn. & Christenson W3249 no voucher FLAS Peru Y Y Y Y Y Y Sobralia chrysostoma Dressler KN213 Neubig 213 FLAS Costa Rica Y Y Y Y Y Y Sobralia ciliata (C. Presl) C. Schweinf. & Foldats W3529 Whitte n 3529 FLAS Venezuela Y Y Y Y Y Y Sobralia citrea Dressler B3030 Blanco 3030 FLAS Panama Y Y Y Y Y Y Sobralia citrea Dressler W2977 Whitten 2977 FLAS Panama Y Y Y Y Y Y Sobralia citrea Dressler RLD6338 Dressler 6338 (type) NA Panama Y N Y Y Y N Sobr alia crispissima Dressler KN196 Neubig 5 2006 FLAS Panama Y Y Y Y Y Y Sobralia crocea (Poepp. & Endl.) Rchb. f. KN206 Neubig 206 FLAS unknown Y Y Y Y Y Y Sobralia crocea (Poepp. & Endl.) Rchb. f. W0533 no voucher NA Ecuador Y Y Y Y Y Y Sobralia crocea (Poepp. & Endl.) Rchb. f. W1578 Whitten 1578 FLAS Ecuador EU490672 Y Y EU490782 Y EU490709 Sobralia decora Bateman W2862 Whitten 2862 FLAS Panama Y Y Y Y Y Y Sobralia dichotoma Ruiz & Pav. W3532 Whitten 3532 FLAS Peru Y Y Y Y N Y Sobralia dorbignyana R chb. f. B3129 Blanco 3129 FLAS Ecuador Y Y Y Y Y Y Sobralia dorbignyana Rchb. f. DT276 Trujillo 276 HURP Peru Y Y Y Y Y Y Sobralia doremiliae Dressler B2973 Blanco 2973 FLAS Panama Y Y Y Y Y Y Sobralia ecuadorana Dodson W2850 Whitten 2850 FLAS Ecuador Y Y Y Y Y Y Sobralia exigua Dressler W3003 Whitten 3003 FLAS Panama Y Y Y Y Y Y Sobralia fenzliana Rchb. f. KN212 Neubig 212 FLAS Panama Y Y Y Y Y Y

PAGE 107

107 Table 2 1. Continued. SPECIES DNA NUMBER VOUCHER Herbarium Origin ITS trnS G rpl16 3'ycf1 trnL F matK Sobralia gentryi Dodson W2852 Whitten 2852 FLAS Ecuador Y Y Y Y Y Y Sobralia gloriana Dressler B2678 Blanco 2678, Whitten 3307 FLAS Panama Y Y Y Y Y Y Sobralia gloriana Dressler B3029 Blanco 3029 FLAS Panama Y Y Y Y Y Y Sobralia gloriana Dressler KN200 Neubig 200 FLAS Panama Y Y Y Y Y Y Sobralia helleri A.D. Hawkes B2840 Blanco 2840 FLAS Panama Y Y Y Y Y Y Sobralia kerryae Dressler B2943 Blanco 2943 FLAS Panama Y Y Y Y Y Y Sobralia klotzscheana Rchb. f. B3011 Blanco 3011 FLAS Ecuador Y Y Y Y Y Y S obralia kruskayae Dressler B2675 Blanco 2675 FLAS Costa Rica Y Y Y Y Y Y Sobralia kruskayae Dressler B3020 Blanco 3020 FLAS Panama Y Y Y Y Y Y Sobralia labiata Warsz. & Rchb. f. W2832 Whitten 2832 FLAS Panama Y Y Y Y Y Y Sobralia lacerata Dressler & Pupulin KN209 Neubig 209, Whitten 3570 FLAS Colombia Y Y Y Y Y Y Sobralia lancea Garay W2869 Whitten 2869 FLAS Ecuador Y Y Y Y Y Y Sobralia leucoxantha Rchb. f. B2681 Blanco 2681 FLAS Panama Y Y Y Y Y Y Sobralia leucoxantha Rchb. f. B2914 Blanco 2914 FLAS Panama Y Y Y Y Y Y Sobralia liliastrum Lindl. FS01418 FS01418 NA Brazil Y Y Y Y Y Y Sobralia liliastrum Lindl. SB10 Koehler 34146 ESA Brazil Y Y Y Y Y Y Sobralia lindleyana Rchb. f. N573 Maduro & Olmos 223 NA Panama Y Y Y Y Y Y Sobralia lindley ana Rchb. f. W2973 Whitten 2973 FLAS Panama Y Y Y Y Y Y Sobralia luerorum Dodson W2729 Whitten 2729 FLAS Ecuador Y Y Y Y Y Y Sobralia macrantha Lindl. W3533 Whitten 3533 FLAS Mexico Y Y Y Y Y Y Sobralia macrophylla Rchb. f. B3022 Blanco 3022, Whitten 3266 FLAS Panama Y Y Y Y Y Y Sobralia macrophylla Rchb. f. SB07 ESA Brazil Y N Y Y Y Y Sobralia madisonii Dodson W2854 Whitten 2854 FLAS Ecuador Y Y Y Y Y Y Sobralia maduroi Dressler M&O206 Maduro & Olmos 206 (type) NA Panama Y Y Y Y N Y Sobralia ma ndonii Rchb. f. W3247 Whitten 3247 FLAS Peru Y Y Y Y Y Y Sobralia mandonii Rchb. f. W3530 Whitten 3530 FLAS Peru Y Y Y Y Y Y Sobralia mucronata Ames & C. Schweinf. B2971 Blanco 2971 FLAS Panama Y Y Y Y Y Y Sobralia mucronata Ames & C. Schweinf. KN210 Ne ubig 210 FLAS unknown Y Y Y Y Y Y

PAGE 108

108 Table 2 1. Continued SPECIES DNA NUMBER VOUCHER Herbarium Origin ITS trnS G rpl16 3'ycf1 trnL F matK Sobralia nutans Dressler N575 Maduro & Olmos 236 (type) NA Panama Y Y Y Y Y Y Sobralia portillae Christenson W2433 Whitten 2433 FLAS Ecuador Y Y Y Y Y Y Sobralia powellii Schltr. W3257 Whitten 3257, Neubig 227 FLAS Ecuador Y Y Y Y Y Y Sobralia quinata Dressler FP3644 Pupulin 3644 (type) USJ Costa Rica Y Y Y Y Y Y Sobralia recta Dressler B3010 Blanco 3010 FLAS Pana ma Y Y Y Y Y Y Sobralia recta Dressler KN207 Neubig 207 FLAS Panama Y Y Y Y Y Y Sobralia recta Dressler W2851 Whitten 2851 FLAS Panama Y Y Y Y Y Y Sobralia roezlii Rchb. f. KN226 no voucher FLAS Colombia Y Y Y Y Y Y Sobralia rosea Poepp. & Endl. KN215 Neubig 215 FLAS unknown Y Y Y Y Y Y Sobralia rosea Poepp. & Endl. W0531 no voucher NA Ecuador Y Y Y Y Y Y Sobralia rosea Poepp. & Endl. W1507 Whitten 1507 FLAS Ecuador Y Y Y Y Y Y Sobralia rosea Poepp. & Endl. DT280 Trujillo 280 HURP Peru Y Y Y Y Y Y Sobralia rosea Poepp. & Endl. DT314 Trujillo 314 HURP Peru Y Y Y Y Y Y Sobralia sanderiana Rolfe W3258 Whitten 3258 FLAS unknown Y Y Y Y Y Y Sobralia sanfelicis Dressler M&O269 Maduro & Olmos 269 (type) NA Panama Y N Y Y Y Y Sobralia sessilis Lindl. SB05 ESA Brazil Y Y Y Y Y Y Sobralia setigera Poepp. & Endl. DT296 Trujillo 296 HURP Peru Y Y Y Y Y Y Sobralia sororcula Dressler RLD6415 Dressler 6415 (type) NA Panama Y Y Y Y Y Y Sobralia sp. KN205 Neubig 205 FLAS Colombia Y Y Y Y Y Y Sobralia sp. W2863 Whitten 2863 FLAS unknown Y Y Y Y Y Y Sobralia theobromina Dressler B2679 Blanco 2679 FLAS Panama Y Y Y Y Y Y Sobralia theobromina Dressler W2989 Whitten 2989 FLAS Panama Y Y Y Y Y Y Sobralia turkeliae Christenson KN225 no voucher FLAS unknown Y Y Y Y Y Y Sobralia undatocarinata C. Schweinf. M&O227 Maduro & Olmos 227 NA unknown Y Y Y Y N Y Sobralia violacea Linden ex Lindl. KN218 Neubig 218 FLAS unknown Y Y Y Y Y Y

PAGE 109

109 Table 2 1. Continued SPECIES DNA NUMBER VOUCHER Herbarium Origin ITS trnS G rpl16 3'ycf1 trnL F matK Sobralia virginalis F. Peeters & Cogn. W1589 Whitten 1589 FLAS Ecuador Y Y Y Y Y Y Sobralia virginalis F. Peeters & Cogn. W2853 Whitten 2853 FLAS Ecuador Y Y Y Y Y Y Sobralia warszewiczii Rchb. f. B2676 Blanco 2676, Blanco 268 6 FLAS Costa Rica EU490673 Y Y EU490783 Y EU490710 Sobralia warszewiczii Rchb. f. W2831 Blanco 2677, Whitten 2831 FLAS Costa Rica Y Y Y Y Y Y Sobralia warszewiczii Rchb. f. B2689 Blanco 2689, Dressler 3198 FLAS Costa Rica Y Y Y Y Y Y Sobralia warszewicz ii Rchb. f. KN211 Neubig 211 FLAS Costa Rica Y Y Y Y Y Y Sobralia warszewiczii Rchb. f. N596 Dressler 6287 NA unknown Y Y Y Y Y Y Sobralia warszewiczii Rchb. f. N597 Dressler 6181 NA unknown Y Y Y Y Y Y Sobralia yauaperyensis Barb. Rodr. B3023 Blanco 3023 FLAS Ecuador Y Y Y Y Y Y Tropidia polystachya (Sw.) Ames W2830 Whitten 2830 FLAS Ecuador EU490674 Y Y EU490793 Y EU490714

PAGE 110

110 Table 2 2 Subgeneric/sectional classification schemes for Elleanthus Reichenbach (1862) Garay (1978) Brieger (1983) Su bgen. Calelyna Subgen. Euevelyna Sect. Cephalelyna Sect. Stachydelyna Subsect. Chloidelyna Kermesinae Furfuraceae Oliganthae Ensatae Sect. Calelyna Sect. Cephalelyna Sect. Stachydelyna Sect. Ellea nthus Sect. Chloidelyna Sect. Laterales Sect. Hymenophora Sect. Otiophora Gen. Evelyna Gen. Pseudelleanthus Gen. Elleanthus Sect. Elleanthus Sect. Elongatae Sect. Strobilfera

PAGE 111

111 Table 2 3 Subgeneric/sectional classification schemes for all taxa c urrently referred to as Sobralia in this text Note that Lindley (1854) did not give formal names for groupings, but rather used letters, numbers and asterisks for delimitation, so details of key morphological characters for each group or key taxa includ ed are shown for clarity. Reichenbach ( 1853 ) Lindley (1854) Brieger (1983) Brasolia Eusobralia Brasolia ) S. rosea etc.) Sobralia ) without crests) Gen. Sobralia Sect. Sobralia Sect. Racemosae Sect. Abbreviatae Sect Intermediae Sect. Globosae Gen. Lindsayella Gen. Fregea

PAGE 112

112 CHAPTER 3 BIOGEOGRAPHY OF SOBR ALIEAE: ORIGINS IN S OUTH AMERICA AND DIVERSIFICATION IN C ROSSING THE ISTHMUS OF PANAMA Background Using phylogenetic relationships within Sobralieae presented in Chapter 2, this chapter address es biogeographic patterns, including geographic distributions of clades, geologic history of the Neotropics, and phylogenetic structure over time within the tribe. The current distributions of plant taxa reflect a complex interplay between their evolutionary history and the geologic history of the landmasses they inhabit. Present day distributions a rise by various processes including vicariance and dispersal. Vicariance is the process by which populations are isolated from one another, often by means of a geological event, which may lead to genetic divergence over time and possible speciation. Disp ersal is a different mechanism by which a lineage can be split and where populations are separated by range expansion or crossing of barriers to colonize new habitats. Both vicariance and dispersal may result in the genetic isolation of lineages and the p ossibility of speciation. Both processes likely influence the distribution of orchid species, but the relative contribution of each may be difficult to determine One clear pattern of biogeography is by isolation over time. Isolation has the effect of al lowing a lineage to diverge phylogenetically, potentially allowing for niche specialization. Theoretically, if an island, or some other effectively isolated area, drifted geologically and came in contact with another landmass, that lineage of plants could further explore new niches and more area in the novel landmass. Such is the geological history of South America, home to many groups of endemic plants, relative to Central America.

PAGE 113

113 As all the major landmasses of Gondwana split, Africa and South America remained together until ~130 Ma ( Fitzgerald, 2002 ; Macdonald et al., 2003 ) South America was isolated from other continents since it split from Gondwana, although it may have temporarily come in contact with various Caribbean landmasses over that time ( Briggs, 19 94 ) The geologic history of the Caribbean landmasses is complex ( Hedges, 1982 ; Hedges, Hass, and Maxson, 1 992 ; Hedges, 1996 ; Graham, 2003 ) and although important in many plant groups, seems to be irrelevant in Sobralieae (see later discussion). In South America the orogeny of the Andes has been a long and complex geological process starting ~90 Ma ( Cardona et al., 2010 ) and is still occurring today ( Husson, Conrad, and Faccenna, 2008 ) The process of subduction and crustal thickening has repeated over this period ( Haschke et al., 2002a ; Haschke et al., 2002b ) creating topological variation. T hrough their orogeny, the Andes have changed drainage patterns in South America and have likely played a large role in the generation of biodiversity ( Hoorn et al., 1995 ; Antonelli et al., 2009 ; Hoorn et al., 2010 ) The dynamic nature of these events may have led to historically significant periods of disjunct landmasses and variation in elevation in montane areas. There is geological evidence that Central and South America have not always been connected as they are today. The narrow Isthmus of Panama is supposed to have closed the oceanic gap connecting the P acific Ocean and the Caribbean Sea around 3 5 Ma through a very complicated series of events ( Briggs, 1994 ) This closure is supported by various marine biological lines of evidence ( Lessios, 1981 ; Briggs, 1984 ; Coates et al., 1992 ; Jackson et al., 1993 ) and marine salinity ( Burton, Ling, and

PAGE 114

114 O'Nions, 1997 ) However, there is both biological evidence ( Knowlton et al., 1993 ; Knowlton and Weigt, 1998 ; Dick, Herrera Cubilla, and Jackson, 2003 ) and geological evidence ( Montes et al., 2012 ) that the closing of isthmus may have occurred earlier. in which there is fossil evidence that many animals suddenly began to move from North America and Central America to South America and vice versa ( Marshall et al., 1979 ; Webb, 1991 ; Dacosta and Klicka, 2008 ) In the same way that animals began to move freely between these two large continental masses, some pl ants did the same. Undoubtedly this complex geological process o f the isolation o f the South American continent (after its breakup from Gondwana) explains at least some endemic groups to the New World. Because South America is no longer disjunct from other landmasses, one would expect those taxa that evolved there to have migrated into the Caribbean, Central America, and even North America. If one were to suppose that a group diversified over time on the South American continent, the phylogenetic signal might represent a paraphyletic grade of taxa that are still only known from South America and crown clades that have moved into Central America either exclusively (thereby forming a clade that is endemic to Central America) or mixed (thereby dispersing back and forth among Central and South America). Examples that clea rly demonstrate this South American origin hypothesis based on phylogenetic data are Opuntia in Cactaceae ( Majure et al., 2012 ) Bromeliaceae ( Givnish et al., 2007 ; Sass and Specht, 2010 ) Siparuna in Sip arunaceae ( Renner and Won, 2001 ) and various orchid groups such as those of tribe Cymbidieae, including Zygopetalinae ( Whitten et al., 2005 ; Neubig et al., 2009a )

PAGE 115

115 Maxillariinae ( Whitten et al., 2007 ; Kirby, 2011 ) among many others ( Lia et al., 2001 ; Kron and Luteyn, 2005 ; Levin and Miller, 2005 ; Simpson, Tate, and Weeks, 2005 ; Erkens et al., 2007 ; Saslis Lagoudakis et al., 2008 ; Yuan and Olmstead 2008 ; Endara, Williams, and Whitten, 2011 ; Symmank et al., 2011 ; Gardner et al., 2012 ) Ma ny other potential New World endemic examples have been enumerated ( Gentry, 1982 ; Burnham and Graham, 1999 ) but many have no conclusive phylogenetic data to address biogeographic origins. Conversely, the invasion of North American/Central American plant taxa such as in Lupinus (Fabaceae) ( Hughes and Eastwood, 2006 ) and Sideroxylon (Sapotaceae) ( Smedmark and Anderberg, 2007 ) into South America has occurred B ut when did such migration events take place ? Estimating the age of Orchidaceae is necessary to corroborate geolo gic events with phylogenetic events. Unfortunately, orchids have a very incomplete fossil record. The most convincing orchid fossils include some fossil leaves putatively belonging to the genera Dendrobium and Earina ( Conran, Bannister, and Lee, 2009 ) and pollinia attributed to subtribe Goodyerinae ( Ramirez et al., 2007 ) Phylogenies incorporating these fossils yield an estimate for the crown clade of Epidendroideae diverging around 51 Ma, whic h is directly re levant to early diverging groups of that subfamily, including Sobralieae ( Ramirez et al., 2007 ) Sobralieae are an exclusively Neotropical group of plants, although the most closely related tribes are typically present in both the New and Old W orld (e.g., Arethuseae, Neottieae, Triphoreae, Tropidieae, etc.). The distributions of Elleanthus and one subclade of core Sobralia are bo th widespread throughout the Neotropics. The two other minor genera are Epilyna which is found only in northern South America and

PAGE 116

116 southern Central America, and Sertifera which is only found in the northern Andes of South America. The distribution of So bralia is complicated by its non monophyly. Sobralia sect. Sobralia is exclusively found in South America. A paraphyletic grade of taxa within core Sobralia (see chapter 2 for details) is exclusively South American, including Sobralia liliastrum S. luer orum and S. rosea (among other species in sect. Racemosae ). No species of Epilyna Sertifera or Sobralia occur in the Caribbean. Only Elleanthus is present in the Caribbean, with only five species ( Nir, 2000 ) three of which are widespread in Central and South America ( E. cephalotus E. caravata and E. longibracteatus ) and the other two of which are endemic ( E. cordidactylus and E. du ssii ). These latter two species belong to sects. Chloidelyna and Calelyna respectively, and if they are in fact distinct species, they are very recently diverged phylogenetically. This assemblage of species in the Caribbean indicates that the presence o f Elleanthus in the Caribbean is a relatively recently derived condition. Because the Caribbean islands do not represent any significant cladogenesis within Sobralieae and because of the low level of endemism, this study will not focus on the Caribbean, e xcept in the context of geologic history. Materials and Methods Distribution M aps To understand the distribution of Sobralieae, specimen data were downloaded from the Global Biodiversity Information Facility (GBIF), including digitized locality information from 8,492 collecti ons (from 77 herbaria) in Sobralieae. The 5,601 specimens with coordinate data were used to map distributions of taxonomic groups in Sobralieae using ArcGIS TM Taxonomy associated with the accessions was modified for

PAGE 117

117 consistency with generic and species level circumscriptions outlined in chapter 2 of this dissertation. Mesquite and R 8s A nalyses The total evidence tree using six DNA regions generated from a RAxML ( Stamatakis, Ludwig, and Meier, 2005 ) maximum likelihood analysis from chapter 2 was used to reconstruct ancestral areas through Mesquite ( Maddison and Maddison, 2005 ) Geographic character state s were code d as either Eurasia (outgroups only), Central America, South America or both. This 6 region likelihood tree was also used to create an ultrametric tree of Sobralieae using r8s ver. 1.7.1 ( Sand erson, 2003 ) based on penalized likelihood ( Sanderson, 2002 ) with a TN algorithm. Estimates of the crown age s of Epidendroideae from Ramirez et al. (2007 ) was used to designate the root age of the tree (i.e., 51 Ma and 61 Ma, along with ranges in error) Cross validation determined an optimal smoothing parameter o f 3.2 with penalized likelihood analyses ( Sanderson, 2003 ) Results Maps of all groups had distributions that were exclusively found in South America, except for Elleanthus Epilyna and a portion of c ore Sobralia (Figs. 3 1 through 3 6). Elevation for these same specimens can be seen in histograms (Figs. 3 1 through 3 6). To understand the influence of sample density as it relates to species richness, number of specimens and number of species (richne ss) were plotted independently in polygons across the range for all specimen data plotted in figures 3 1 through 3 6 (Fig. 3 7). When reconstructed on a phylogenetic tree, a South American distribution is plesiomorphic and a Central American distribution i s derived (Fig. 3 8). The reconstruction shows relatively frequent interchange between South America and

PAGE 118

118 Central America. The cladogenetic origins of Elleanthus and the subclade of core Sobralia coincide relatively close in time around 10 Ma for a root fixed at 61 Ma and around 8 Ma for a root fixed at 51 Ma (Figs. 3 9 & 3 10 ; Table 3 1 ), but the subsequent introduction of Epilyna jimenezii into Central America occurred much later around 2 Ma (regardless of root age) However, the range in dates for th ese nodes differs drastically because of the range in estimated dates for the origin of the crown clade of Epidendroideae as previously defined ( Ramirez et al., 2007 ) The age of the crown clade of Sobralieae is estimated at 15.7 29.7 Ma. The crown ages of the Elleanthus clade and the subclade of core Sobralia are similar: 6.6 11.5 Ma and 6.7 14.6 Ma, respectively. The crown ag e of Epilyna is much younger, at 1.6 2.6 Ma. Discussion Limiting Factors of D istribution? Geographic distributions can be biologically and phylogenetically informative, but aspects that do not reflect evolutionary history may bias current distributions, an d th us confound phylogenetic interpretations A variety of abiotic, biotic and anthropogenic forces determine species distributio ns. Thus geographic specimen collection bias could be the limiting factor in our understanding of distribution. These spec imen data show very high species richness for areas within Costa Rica. But does this represent an area of higher species richness, or an artifact of more extensive collection here than elsewhere ? This phenomenon might best be explained by high collecting activity and many recent species descriptions focused on Sobralia especially within Costa Rica and Panama ( Dressler, 1995 2001 2002 2003b a 200 4b a 2005 2006 2007 ; Dressler and Bogarin, 2007a ; Dressler and Pupulin, 2008 ; Dressler, 2009 ; Dressler and Bogarin, 2010 ; Dressler and Pupulin, 2010 ; Dressler, 2011 2012 ) In addition areas with larger

PAGE 119

1 19 number s of collections generally have more species represented (Fig. 3 7), perhaps indicating a sampling bias. This phenomenon of apparent higher species richness has been noted in Ecuador, also because of higher collec tion activity ( Jrgensen and L en Ynez, 1999 ) Nonetheless, these data indicate that Costa Rica and Ecuador have the highest species richness (Fig. 3 7b), reflecting high species numbers in floristic treatments of those areas compared to the remainder of the Neotropics. The signif icance of the collection bias is unimportant for overall estimations of geographic origin and movement across geographic barriers over time. However, this taxonomic bias for the relatively well studied countries of Costa Rica and Ecuador makes it such tha ( Rabosky and Lovette, 2008 ) is probably premature given the taxon sampling of this phylogenetic component, e specially considering the taxon sampli ng within this phylogenetic study. Therefore, in the context of the discussion of diversification, it is meant here as a relative increase in species richness. Recently, the importance of mycorrhiz al associations with plants has been examined as a major limiting factor of orchid distributions ( Rasmussen, 2002 ; Dearnaley, 2007 ) Th e importance of mycorrhizae in the growth and function of orchids and other plants has long been acknowledged. However, this specificity of this plant fungus relationship can differ among species, among seedlings versus adults ( ) and among terrestrial versus epiphytic habits ( Martos et al., in press ) Edaphic qualities may also limit fungal presence and therefore the presence of the associated plant ( McCormick et al., 2012 ) Furthermore ecological specialization and mycorrhizae rarity seem to be the major limiting factor in the distribution of some

PAGE 120

120 orchids ( Swarts et al., 2010 ) Sobralieae mycorrhizal associations are completely unknown. Nonetheless, because Sobralieae grow in a vari ety of areas, both terrestrial and epiphytic, which would have very different soil types, it is likely that the fungal associates are varied and ubiquitous. Although mycorrhizae may determine local establishment of Sobralieae, there is no reason to suspec t mycorrhizal presence/absence has affected continent level movements. Sobralieae species are most diverse in the montane Neotropics (e.g., the cordilleras of Central America, the Andes, the Guyanan shield), and the major clades display differences in thei r elevational ranges. Elevational differences among genera or groups within Sobralieae might be indicative of a bias that has led to present day distributions. Epilyna (1250 m), Sobralia sect. Racemosae (950 m), and the subclade of core Sobralia (950 m), all have modest average elevations. Some Elleanthus live at high elevations, but most are found at modest elevations (1650 m). Perhaps most remarkable is Sertifera which grows at the highest elevations on average (2450 m). Furthermore, Sobralia sect. Sobralia grows at high elevations (1950 m). Non geological forces may largely influence some distributional aspects such as elevation. Distributions of organisms that may play intimate role s in the life history of Sobralieae (i.e., pollinators) can great ly affect their geography by limiting their reproductive capacity. For example, tax a such as Elleanthus and Sertifera live on average at higher elevations than Epilyna or most groups of Sobralia Elleanthus and Sertifera are both primarily pollinated by hummingbirds, whereas Sobralia is primarily pollinated by bees and the pollinators of Epilyna are unknown (see chapter 4 for a more detailed discussion of pollination and the elevational differences likely driven by

PAGE 121

121 pollinator specificity; Fig. 4 23). Hum mingbird diversity and density matches the elevational preferences of Elleanthus and Sertifera from sea level to ~3000 m ( Feinsinger et al., 1979 ; Bleiweiss, 1998b ; Stiles, 2004 ) ; likewise, the diversity and density of bees (especially Euglossini) matches that of Sobralia of sea level to 1600 m ( Dressler, 1982 ) This association of plant s overlapping the distribution of pollinator s probably also at least partially explains the absence of Sobralia in the Caribbean; that is, there are also no euglossine bees in the Greater a nd Lesser Antilles, except for one species in Jamaica. However, elevational preference by pollinators for groups within Sobralieae does not adequately explain current overall distribution throughout the Neotropics, nor the patterns of diversification withi n Elleanthus and core Sobralia Indeed, both Elleanthus and core Sobralia take advantage of a wide range of elevations (Figs. 3 1 & 3 6). The subclade of core Sobralia (Fig. 3 6) and sect. Racemosae (Fig. 3 5) both have an average elevation of 950 m, yet the former is widespread and the latter is restricted to South America. Sertifera with an average of 2450 m in elevation, represents the highest average for any group in the tribe. Likewise, sect. Sobralia occurs at an average of 1950 m, which is relat ively high for the tribe and sect. Sobralia is restricted to South America. Any quantitative differences in highest elevation amon g Central and South America do not explain the distribution bias in South America, because Sobralieae does not utilize the h ighest elevations available in either area Therefore, neither high nor low elevation niches could be the limiting factor to explain why certain phylogenetic groups are present in South America, but not in Central America. This

PAGE 122

122 lack of limitation is als o true for pollinator availability, because both bees and hummingbirds are present in Central and South America. Diversification within the core group of Sobralia may be associated w ith the exploitation of a lower elevation niche. The subclade of core So bralia that has repeatedly crossed the Isthmus of Panama can be found at much lower elevations (often <500 m) across its distribution (Fig. 3 6). Overall, elevation has likely played a role in the evolution of Sobralieae because of elevational preferences of pollinating bees and hummingbirds, but has not been a restricting factor in its migration across the Isthmus of Panama. The phylogenetic significance of this is in homoplasy, because pollinator class has shifted multiple times within the tribe (see ch apter 4 for a detailed discussion of pollination within a phylogenetic context). Geographic O rigin The phylogenetic and distributional data are most consistent with a South American origin for Sobralieae. To estimate at which point within the phylogeny d ispersal to Central America occurred, ancestral character state reconstruction was implemented of geographic character state s [ either Eurasia, South America or Central America ; ML and MP, in Mesquite ( Maddison and Maddison, 2005 ) ]. This was done on range of a species was coded for an individual. These strategies outline the bias of sampling per geographic range in the former and emphasized the ambiguity per species in the latter. Nonetheless, the latter is emphasized because it errs on the side of conservative estimates overall. The selection of three clades ( Elleanthu s Epilyna and

PAGE 123

123 the subclade of core Sobralia ) demonstrates the most inclusive groups where widespread distributions are likely to have occurred evolutionarily. Phylogenetic patterns as they relate to geographic structure have been dissected in various way s. However, they do not tell us much of general phylogenetic/geographic correlations with major geological events. By comparing published phylogenetic hypotheses as they relate to geographic patterns across the Isthmus of Panama, we might better understa nd the complex manifestations of migration. In Fig. 3 12, I outline some discrete phylogenetic situations that represent a South American origin for groups that have diversified into Central and North America. These patterns can be simple, or if found in combination, are likely variable because of the many biotic (e.g., pollination, fruit dispersal, and mycorrhizal associations) and abiotic (e.g., geologic events and climatic fluctuations) reasons for current patterns of distribution. Nonetheless, the anc estral state within th e Sobralieae is South America (Figs. 3 8 and 3 9), so its diversification into Central America must be derived. There are two clades that include the majority of species (i.e., Elleanthus and a subclade of core Sobralia ). However, t hat pattern of radiation into Central America is somewhat stochastic; that is, there have been many different migrations between South and Central America based on character state reconstructions. This pattern of structured randomness fits one of the mode ls of evolution proposed here of short distance dispersal, with frequent interchange between ancestral and derived distributional states. Dating in Sobralieae The age of Orchidaceae has been difficult to estimate due to scarcity of fossils. Recent publica tions include convincing fossil leaves putatively belonging to the genera Dendrobium and Earina ( Conran, Bannister, and Lee, 2009 ) An orchid pollinium

PAGE 124

124 attached to the back of a bee was recently described from amber ( Ramirez et al., 2007 ) That fossil is definitively placed as a member of subtribe Goodyerinae in subfamily Orchidoideae. The dates of the Dendrobium and Earina fossils are congruent with the pollinium fossil ( Conran, Bannister, and Lee, 2009 ) so therefore support the hypothesis of the dates proposed by Ramirez et al. ( 2007 ) These fossils were incorporated within a combined ultrametric analysis ( Guo et al., 2012 ) Unfortunately, they did not explicitly describe the taxon sampling within Epidendroideae for their analysis, so relative dates as they apply to various taxa are unclear and therefore not usable in this study. The impact of using various calibrations on dating a phylogeny has been explored extensively ( Sauquet et al., 2012 ) The a ge of Sobralieae and clades within were estimated using a range of crown age estimates of Epidendroideae from Ramir ez et al. (2007 ) in order to provide an estimate of error (Table 3 1). Two primary hypotheses of ages of the crown clade of Epidendroideae were presented in that paper: 51 and 61 Ma (Figs. 3 10 & 3 11). H owever, because they gave a more conservative es timate using penalized likelihood with a root age of 51 Ma for Epidendroideae, this discussion will focus on that age despite the associated range in error. These analyses indicate that the crown group of Sobralieae began to diverge around 23.4 Ma. Also, the species rich clades of Elleanthus and the subclade of core Sobralia diverged at about the same time, 9.3 (6.6 11.5) and 10.7 (6.7 14.6) Ma respectively These ultrametric analyses indicate that Sobralieae did not phylogenetically diversify for more t han 20 myr since its split from its closest relatives, including Tropidieae and Arethuseae, at approximately 40 Ma (Fig. 3 10). The isolation of

PAGE 125

125 Sobralieae in South America relatively long ago explains the deep divergence among this tribe and any of its c losest relatives, although its origin by dispersal or vicariance is also are not clear, probably due to a rapid radiation, so accurately dating their divergence dates is problematic. Nonet heless, based on the previous study of fossil dating for the whole family ( Ramirez et al., 2007 ; Guo et al., 2012 ) it is highly unlikely that Epidendroideae orchids began to diversify before the K/T boundary 65 million years ago. The influence of geological processes in South America might explain patterns of phylogeneti c divergence. Species richness in many Neotropical plant groups, including orchids, is relatively high in the Andes ( Young et al., 2002 ) The uplift of the Ande s is hypothesized to have played a major role in the diversification of these orchids. The majority of species of Sobralieae in South America are found in the Andes and other montane area s of the Neotropics but only the members of Sobralia sect. Sobralia sect. Racemosae and Sertifera are restricted to South America The formation of the Andes has been a complex process, with multiple periods of uplift. The orogeny of the northern Andes (especially with high elevations) likely predates the phylogenetic origin of groups in Sobralieae that are now endemic to that area, making the estimates of the age of these Andean endemic taxa (e.g., Sobralia sect. Sobralia and Sertifera ) highly plausible. That is, endemic taxa may have originated in the current area o f endemism, but this is most likely if the geological proce ss pre dates or is concurrent with the origin of the organism. Species from the Guyanan shield or in southeastern Brazil (e.g., Sobralia ciliata and S liliastrum) diverged ~18 and 11 Ma, respectiv ely (Fig. 3 10) and so may

PAGE 126

126 represent an eastern, relictual, pre Andean distribution. This hypothesis is consistent with the presence of the Pebas system of waterways that created a disjunction between eastern and western South America between 10 and 23 Ma ( Hoorn et al., 2010 ) Therefore, the Pebas system may also have been a barrier to dispersal from eastern to western South America. This hypothesis is also consistent with the diversification of Andean taxa such as Sobralia sect. Sobrali a and Sertifera starting at ~13 Ma and Sobralia sect. Racemosae at ~9 Ma (Fig. 3 10) with the dissipation of the Pebas waterways ( Hoorn et al., 2010 ) I hypothesize that three geographically widespread clades ( Elleanthus Epilyna and the s ubclade of core Sobralia ) diverged in response to the closing of the Isthmus of Panama. The dates of these three clades are of particular interest. Their relative ages and absolute ages were inferred using a fossil pollinarium for calibration of ML trees All estimates, including the range of errors are older than the classic estimate of the closing of the isthmus at 3.5 Ma. However, various studies have shown that the estimation of ~3.5 Ma for the closing of the isthmus may be too recent and that this uncertainty may range into 3 9 Ma, or even 18 Ma supported by both biological evidence ( Knowlton et al., 1993 ; Knowlton and Weigt, 1998 ; Dick, Herrera Cubilla, and Jackson, 2003 ) and geological evidence ( Montes et al., 2012 ) that the closing of isthmus may have occurred earlier than 3 Ma. which is congruent with the estimates of this study. Moreover, many previous studies concerning the biogeography around the isthmus have centered on land fauna, which probably have rather stringent requirements of land connection for dispersal. Nonetheless, despite the evidence of the

PAGE 127

127 G reat American I nterchange, some mammals such as peccaries may have migrated into Panama as much as 19 Ma ( MacFadden et al., 2010 ) Sobralieae have small, wind dispersed seeds (Fig. 1 11) and might have dispersed to Central America across an oceanic gap prior to closure of the isthmus. Long distance dispersal e vents from Africa to the New World (a much greater distance) are hypothesized for other orchid taxa (e.g., Bulbophyllum Campylocentrum Dendrophylax Polystachya Tropidia ), so this mechanism of dispersal over large bodies of water, although rare, might e xplain the relatively old dates for dispersal from South America into Central America. The ability of Sobralieae to traverse bodies of water is also evidenced by the presence of Elleanthus throughout the Caribbean. A more densely sampled and better resol ved phylogeny and a fine scale understanding of speciation patterns as they relate to geography may lead to a better picture of dating events. These estimated ages of Sobralieae are consistent with estimated ages of their pollinators. Ramirez et al. (2010 ) used a molecular phylogeny of Euglossini and mo lecular clock methods to estimate that the bees evolved ca. 27 42 Ma. The fossil record of hummingbirds is poor, but molecular phylogenetic work and fossil calibrations of hummingbirds show that they diversified at ~18 Ma, but that their stem lineage is ~ 58 Ma ( Bleiweiss, 1998a ) All of these dates sufficiently pre date the diversification of Sobralieae and so do not contradict the dates estimated within Sobralieae.

PAGE 128

128 Figure 3 1 Distributional map of Elleanthus base d on 3 214 specimen data points. Arrows indicate countries where Elleanthus is present, but no coordinate specimen data were available. Clade in tree to which map refers is referenced in red. Histogram represents minimum elevation recorded for same spec imens used to create map.

PAGE 129

129 Figure 3 2 Distributional map of Epilyna based on 61 specimen data points. Clade in tree to which map refers is referenced in red. Histogram represents minimum elevation recorded for same specimens used to create map.

PAGE 130

130 Figure 3 3. Distributional map of Sertifera based on 33 specimen data points. Clade in tree to which map refers is referenced in red. Histogram represents minimum elevation recorded for same specimens used to create map.

PAGE 131

131 Figure 3 4 Distribut ional map of Sobralia sect. Sobralia based on 55 specimen data points. Grade in tree to which map refers is referenced in red. Histogram represents minimum elevation recorded for same specimens used to create map.

PAGE 132

132 Figure 3 5 Distributional map of Sobralia sect. Racemosae based on 120 specimen data points. Grade in tree to which map refers is referenced in red. Histogram represents minimum elevation recorded for same specimens used to create map.

PAGE 133

133 Figure 3 6 Distributional map of core Sobr alia (excluding sect. Racemosae ), based on 2 112 specimen data points. Clade in tree to which map refers is referenced in red. Histogram represents minimum elevation recorded for same specimens used to create map.

PAGE 134

134 Figure 3 7 All data presented in figures 3 1 through 3 6 are represented in these maps. A) An overlay of number of specimen collections per polygon. B) An overlay of number of species per polygon.

PAGE 135

135 Figure 3 8 Ancestral characters state (ML) reconstruction of Sobralieae based on ge ographic distribution using the total evidence tree of the 6 DNA region maximum likelihood analysis. This tree emphasizes the importance of node selection to represent portions of the tree that represent the expansion of taxa from South America to Central America.

PAGE 136

136 Figure 3 9 Ancestral characters state (MP) reconstruction of Sobralieae based on geographic distribution using the total evidence tree of the 6 DNA region maximum likelihood analysis. This tree emphasizes the importance of node selection to represent portions of the tree that represent the expansion of taxa from South America to Central America as indicated by arrows

PAGE 137

137 Figure 3 10 C hronogram of the 6 DNA region maximum likelihood tree based on a root age of 51 Ma. Scale bar represe nts millions of years. Arrows indicate the node s that represent the expansion of taxa from South America to Central America as indicated by arrows

PAGE 138

138 Figure 3 11 Chronogram of the 6 DNA region maximum likelihood tree based on a root age of 61 Ma. Sc ale bar represents millions of years. Arrows indicate the node s that represent the expansion of taxa from South America to Central America as indicated by arrows

PAGE 139

139 Figure 3 12 Theoretical phylogenetic patterns for biogeographic interactions with a So uth American origin. A ) A stochastic model of distribution with no clear origin. B ) A short distance dispersal model of taxa originating in South America (new locality is close enough that it freely interchanges between plesiomorphic and derived localiti es). C ) A long distance model of taxa originating in South America (new locality is far enough that it diversifies with little to no interchange).

PAGE 140

140 Table 3 1 Clade age estimates in millions of years using r8s based on the same 6 DNA region maximum lik elihood tree. The root of the tree was fixed at various ages to estimate the range in error for the age of the crown Epidendroideae (Ramirez et al, 2007). A smoothing parameter of 3.2 with a TN algorithm under penalized likelihood was used for all analys es. The range for two dates (plus their maximum/minimum error ranges) is given that represents A) the oldest ( 618 Ma ) and B) youngest ( 517 Ma ) scenarios for Epidendroideae (see Ramirez et al., 2007 for details). Ages are rounded to nearest decimal. A oldest scenario Estimated root age from fossil 53 .0 61 .0 73 .0 Ramirez et al. (2007) Sobralieae (crown age) 19.6 23.4 29.7 this study Epilyna + Elleanthus (crown age) 9.6 11.2 13.9 this study Epilyna (crown age) 1.9 2.2 2.6 this study Elleanthus (cro wn age) 8 .0 9.3 11.5 this study "core" Sobralia subclade (crown age) 8.7 10.7 14.6 this study B youngest scenario Estimated root age from fossil 44 .0 51 .0 58 .0 Ramirez et al. (2007) Sobralieae (crown age) 15.7 18.7 21.9 this study Epilyna + Elleant hus (crown age) 7.9 9.2 10.6 this study Epilyna (crown age) 1.6 1.8 2 .0 this study Elleanthus (crown age) 6.6 7.7 8.8 this study "core" Sobralia subclade (crown age) 6.7 8.2 9.9 this study

PAGE 141

141 CHAPTER 4 REWARD AND DECEIT POLLINATION IN S OBRALIA AND E LLE ANTHUS : A COMPARISON OF NECTAR PRODUCTION AND ADAPTATION THROUGH FLORAL ANATOMY Background The objectives of this study are to document pollination related traits including flower size, color, morphology, nectary structure, nectar production, and floral ph enology, in Sobralia and Elleanthus and to analyze these traits in a phylogenetic context. Plant pollinator interactions are dynamic, based on many different interactions and stimuli specific to pollinators. Evolutionary adaptations to various pollen ve ctors can lead to the evolution of novel pollin ation systems within groups. While foraging bees are attracted to many different colors, fragrances, nectar guides, and nectar rewards, birds are attracted to bright, contrasting colors in flowers and/or brac ts with nectar, but no fragrances. At the core of the attraction is the promise of reward. Floral rewards in Orchidaceae are extremely diverse ( van der Pijl and Dodson, 1966 ; van der Cingel, 2001 ) and include nectar, oils, fragrances and resin or wax; in contrast to many angiosperms, pollen is never offered as a legitimate reward. Some orchids (subtribe Oncidiinae) offer oil as a reward to Cent ris bees, but have close relatives that deceive oil reward ( Silvera, 2002 ; Alisconi et al., 2009 ; Neubig et al., 2012 ) Another exotic mode of deceit is pseudocopulation, in which the male pollinator is deceived into copulating with the flower which is disguised as a female. Pseudocopulation has evolved ma ny times in orchids, on different continents and in different subfamilies (e.g., Ophrys Tolumnia henekenii Telipogon Lepanthes and Chiloglottis ) ( van de r Cingel, 2001 ; Blanco and Barboza, 2005 ; Devey et al., 2008 ; Neubig et al., 2012 ) Many orchids

PAGE 142

142 produce no nectar (e.g., Cattleya Sobralia and Cochleanthes Raf. ). It is estimated that one third of all species of orchids are deceitful ( Cozzolino and Widmer, 2005 ) This high percentage gives direct evidence that pollination by deceit is a successful strategy. But deceit is a fo rm of mimicry, at least as defin ed by some There are two forms of mimicry: Mllerian, wh ere two or more species benefit, and Batesian, wh ere one or more species benefit from a model ( Johnson et al., 2003 ; Jerskov, Johnson, and Ki ndlmann, 2006 ) and morphologies, and thus be deceitful but without any specific model to mimic. Despite the frequency of deceit in orchids many orchids do reward their pollinators. So me orchids give food bodies such as hairs on the labellum ( Whitten et al., 2007 ; Davies and Stpiczynska, 2008b ) and many Oncidiinae secrete oil that is col lected by female bees and used to provision nests. About 650 species of Neotropical orchids (e.g.,Zygopetalinae, Stanhopeinae, Catasetinae) are pollinated by male euglossine bees (Apidae: Euglossini) that collect fragrances from the orchids. The fragranc es are hypothesized to be involved in courtship behavior and mate attraction ( Roubik and Hanson, 2004 ) But by far, the most common reward is nectar, which is presented by varying morphologies to many different pollinators: Diptera (flies), Hymenoptera (bees and wasps ), Lepidoptera (moths and butterflies) Trochilidae (hummingbirds) in the New World ( van der Pijl and Dodson, 1966 ; van der Cingel, 2001 ) and s u nbirds in the Old W orld ( Johnson, Linder, and Steiner, 1998 ) In Sobralieae, pollination vectors include various bees and hummingbirds. Some species produce nectar rewards, whereas others produce no apparent reward (deceit flowers). Although the pollination biology of Sobralia is poorly studied, most species are

PAGE 143

143 reported to be pollinated by a variety of large solitary bees, especially by euglossine bees ( van der Pijl and Dodson, 1966 ) Sobralia species pollinated by bees have a typical gullet type morphology, of ten with yellow streaks (nectar guides) in the throat and comparatively l arge pollinia (Fig. 4 1). The cor e Sobralia ( primarily bee pollinated) exhibit gregarious flowering, a phenology that is rare in Orchidaceae. In species with gregarious flowering, n early all individuals within a population flower simultaneously, resulting in hundreds or thousands of flowers available (open and receptive) to pollinators for a single day. After this one day of floral maturity, all flower parts undergo rapid deliquescen ce in which the perianth shrivels rapidly; unpollinated flowers are abscised from the plant within another day or two. Each flowering stem of an individual often produces flowers simultaneously, but not continuously and constantly throughout the months of flowering for the plant (i.e., gregarious flowering) This is accomplished by having many flower buds within an inflorescence develop to near maturity and then pause; the trigger for synchronous flowering is unclear, but might be an environmental cue, su ch as a change in temperature, often associated with rainfall. Synchronous flowering has not been studied in detail in Sobralia but experiments with the synchronous flowering Dendrobium crumenatum Sw. have demonstrated that a drop in temperature triggers flowering ( Seifriz, 1923 ) Further synchrony has been documented in Psilochilus Barb. Rodr. ( Pansarin and Amaral, 2008 ) It may be a relatively strong pollen efficiency for plants rare in the landscape ( Jerskov, Johnson, and Kindlmann, 2006 )

PAGE 144

144 A few species such as S. macrantha and S. turkeliae have reverted to having longer lived flowers, but only for several days ( Rogers, 2007 ) Other species of Sobralia such as those of Sobralia sect. Racemosae (e.g., S. rosea ) and sect. Sobralia (e.g., S. caloglossa S. dichotoma and S. mandonii ) have long l ived flowers (of more than several days). These flowers are presented sequentially and continuously for many week s or months and are all bee pollinated. Hu mmingbird pollination occurs in Elleanthus Sertifera and a few Sobralia species ( e.g., S. amabilis and S. callosa ) ( Dressler, 2002 ) Elleanthus flowers are much smaller than those of Sobralia with a typical gestalt for hummingbird type pollination with white, pink, purple, red, or orange flowers and bracts (Fig. 4 2). The brightly colored species of Elleanthus produce copious nect ar, and like Sobralia callosa and S. amabilis have cryptic pollinia. Dressler (1971) noted that various hummingbird pollinated orchids often have pollinia that are dark bluish gray, and pollinia are often smaller than in related bee pollinated taxa. Both traits (dark color and reduced size) are hypothesized to make the pollinia less visible to the hummingbird vector, and therefore Nectar concentration and volume are two traits that are thought to be linked to the class of pollinator ( Baker and Baker, 1983 ) Hummingbird pollinated flowers are hypothesized to produce relatively larger volum es of dilute nectar, whereas bee pollinated taxa produce comparatively smaller volumes of more concentrated nectar ( Bolten and Feinsinger, 1978 ) The characteristics of the nectar and nectaries in Sobralia flowers have never been reported. The first to describe the nectary structure of Elleanthus was Darwin ( 1862 ) who described them a large Because nectar

PAGE 145

145 concentration and volume have been a used as indicators of putative pollinator classes, I measured these traits in all available Sobralieae species. Also, nec tary structure was examined across all Sobralia and Elleanthus Flower size varies extensively in Sobralieae. Species of Elleanthus Epilyna, and Sertifera have relatively small flowers compared to the flowers of Sobralia Variation in floral size is likely a consequence of shifts in pollination mode. The large flowers of Sobralia are mostly pollinated by large bees (e.g. Eulaema ). The small flowers of Elleanthus and Sertifera are usually pollinated by hummingbirds. However, pollinators of Epilyna and those of smaller, white flowered species of Elleanthus are unknown. Floral anatomy may also play a role in pollination. Veyret (1981 ) reported that Sobralia sessilis flowers have two channels that run alo ng the lateral sepals where the perianth is ad is a single tube that runs within the perianth, and often into the ovary region, providing a nectary and is found in subtribe Laeliinae ( Dressler, 1993a ) The double cuniculus probably serves as a pair of tubular false nectaries for long tongued bees [especially including Apidae ( Danforth et al., 2006 ) ]. I surveyed the presence of these structures across species of Sobralia to understand the distribution of this morphological character in a phylogenetic context and to ev aluate the association of cuniculi with other pollination related traits. Overall, I used anatomical and morp hological data to investigate the mechanics of pollination, the propensity of deceit and reward, and the physical floral properties of secretion w ithin the context of known pollinators to better understand the evolution of pollination in a phylogenetic context.

PAGE 146

146 Materials and Methods Nectar Volume and Q uantity Flowers (mostly cultivated; some in the field) were examined for nectar presence/absence. If nectar was found, measurements were made of both volume and sucrose concentra tion. Sucrose concentration was measured with a 0 53 brix Atago refractometer at various times of the day ( Corbet, 2003 ) Concentrations are presented in brix units (g sucrose p er 100 g water) rather than weight per volume, because it is a common unit used in nectar and food science ( Bolten et al., 1979 ; Corbet, 2003 ) Nectar was pipetted and measured with a 0.5 20 Rainin micropipetter. Minor nectar constituents, such as amino acids ( Gottsberger, Schrauwen, and Linskens, 19 84 ) were not examined in this study. Floral A natomy Flowers of selected species were fixed in FAA ( 70% ethanol: glacial acetic acid: commercial formalin, 9.0:0.5:0.5 ) for several days, and then stored in 70% ethanol until prepared for sectioning. Flow ers were dehydrated in series until in 100% ethanol, and then transferred to miner al oil, then embedded of paraffin. Embedded tissues were then sectio Sections were ir on alum hematoxylin and co unter stained with safranin Observations and photographs were taken with a PixeraPro 150es digital camera attached to a Zeiss Axioskop 40 microscope. Additional hand sections were made of various species of Sobralia and Elleanthus to demonstrate variati on and presence of the double cuniculus. 2 KI: iodine potassium iodide) to test for presence of starch. Labella were also hand sectioned in the morning, at noon and evening for mature flowers and also stained with I 2 KI.

PAGE 147

147 Floral F ragrances Gas chromatography mass spectrometry (GC/MS) analyses were outside the scope of this study. However, because of the importance of fragrances in some pollination syndromes, it is important to note the presence and provisional identification of fragrances The experienced noses of Mark Whitten and Norris Williams (noted authorities of fragrances involved in euglossine pollination) were used to help identify major notes within fragrant flowers. These i dentifications were partially confirmed by smell through a reference library of fragrances. Elevational V ariation by P ollination S yndrome To understand empirical distribution of Sobralieae by pollination syndrome, the s pecimen data from chapte r 3 were used including digitized locality information from 8,492 collections from GBIF Of those, only 5,601 had coordinate data associated. And of those, only 5,256 specimens had species level identifications with which pollination syndrome could be associated. T hese data w ere used to create histograms to show trends in elevational variation among pollination syndromes R econstruction of Pollination Related C haracters To reconstruct the evolution of pollination systems, the total 6 region combined DNA maximum like lihood tree from chapter 2 was used to map this character pollination using Mesquite, ver. 2.75 ( Maddison and Maddison, 2005 ) Three character state s were used: one for bee pollination, a second for hummingbird pollination, and a th ird for the unknown pollinator system of the small white flowered species of Elleanthus and Epilyna

PAGE 148

148 Results Overall F loral S yndrome Many species of core Sobralia show a gullet shape floral shape for bee pollination. Taxon sampling was based on broad phyl ogenetic sampling across the tribe and included some species that give no reward, but do have a false nectary: S. decora (Fig. 4 3), S. warszewiczii (Fig. 4 4), and S. chrysostoma (Fig. 4 5). Some species are bee pollinated but offer a nectar reward: S. m acrophylla (Fig. 4 6), S. rosea (Fig. 4 7), and S. bouchei (Fig. 4 8). Two species of core Sobralia have a hummingbird pollination syndrome and provide nectar: S. callosa (Fig. 4 9) and S. amabilis (Fig. 4 10). O ne species of Sobralia is putatively hummi ngbird pollinated without nectar: S. crocea (Fig. 4 11). And one species of Sobralia has a putative sphingophilous syndrome: S. rarae avis (Fig. 4 12). Elleanthus has a predominant ly hummingbird pollination syndrome: E. caravata (Fig. 4 13) and E. sodiro i (Fig. 4 14). Section Sobralia was represented by two clades, one including the putatively hummingbird pollinated S. ciliata (Fig. 4 15) and the other including the bee pollinated S. caloglossa (Fig. 4 16) and S. mandonii (Fig. 4 17). Callus S tructure T he callus of a labellum is probably not homologous across Orchidaceae. However, there is little doubt that the callus in the Sobralieae is homologous. In Elleanthus the callus consists of two relatively large globose balls that sit at the base of the li p. Most, if not all, species of Elleanthus produce nectar from these large stores of starch in the callus. The typical callus of Sobralia also cons ists of two distinct structures; however, they are not the globose structures found in Elleanthus Sobrali a calli are raised ridges that have migrated a small distance up to the sides of the labellum, such that they are borne opposite each other along the sides. When seen at

PAGE 149

149 the distal end of the labellum, the space between the calli forms a small tube (Fig. 1 3C & D), which may guide the tongue of the visiting bee, channeling it to the double cuniculus. The calli of S. bouchei and S. callosa differ from either of the previously mentioned types. They are fused and expanded to form a pad at the median portion of the base of the labellum. Starch All species stained with I 2 KI produced at least some starch. However, the density of starch and thickness of the tissue co ntaining the starch varied. In all of the nectar secreting species, a rich store of starch can be found in the pad like or globose callus, depending on the genus. In species that produced no nectar (i.e., most species of Sobralia ), the starch was less extensive, often restricted to the epidermis of the callus and any hairs t hat are present. S ectio ns of the callus were made with a rotary microtome. These sections show starch grains that fluoresce as a typical cross shape under polarized light (see Fig 4 14F, starch grains/amyloplasts). Double C uniculus This unique structure of having a paired tubu lar false nectary was surveyed and found in S. chrysostoma ( Neubig 213 ), S. decora ( Whitten 3280 ), S. gloriana ( Blanco 2678, Whitten 3307 ), S. macrophylla ( Blanco 3022, Whitten 3266 ), S. helleri ( Neubig 207 ), S. klotzscheana ( Blanco 3011 ), S. powellii ( Wh itten 3257 ), S. warszewiczii ( Blanco 2676, 2677, 2686, 2689, Whitten 2831 ), and S. sp. ( Neubig 205 ). Species lacking a double cuniculus include S. bouchei ( Blanco 3009 ), S. callosa ( Blanco 3021, Whitten 3275 ), S. crocea ( Neubig 206 ), and S. rosea ( Neubig 214 ). No species of Elleanthus Epilyna or Sertifera observed has this double cuniculus.

PAGE 150

150 Nectary and N ectar N ectar concentration (brix) and volume ( ) were measured in four species of Sobralia ( S. bouchei S. callosa S. macrophylla and S. rosea ; Table 4 2; Fig. 4 18) and f our species of Elleanthus ( E. aurantiacus E. caravata E. cynarocephalus and E. sodiroi ; Table 4 2; Fig. 4 19). The following species were observed to produce nectar, but the amounts were too small to be measured: S. ciliata E. lancifolius E. graminifolius E. fractiflexus and E. robustus The following species have been observed that do not produce nectar: S. andreae S. at ropubescens S. caloglossa S. chrysostoma (Fig. 4 5), S. citrea S. crispissima S. crocea (Fig. 4 11), S. decora (Fig. 4 3), S. dichotoma S. doremi liae S. exigua S. gloriana S. helleri S. kerryae S. lacerata S. leucoxantha S. lindleyana S. macra ntha S. mandonii S. mucronata S. quinata S. recta S. theobromina S. violacea S. warscewiczii (Fig. 4 4), and S. yauaperyensis All these species were examined in cultivation, except for S. rosea and E. aurantiacus where plants were measured from m ultiple populations, in situ, across Ecuador in February, 2009. The only other species of Sobralia reported to produce nectar, S. amabilis was not observed in this study. Its flower s are similar in size and shape to those of Sobralia callosa so it coul d also be similar in other aspects of floral anatomy. Fragrance P roduction Fragrance is absent in all species of Elleanthus and Sertifera and some species of Sobralia Fragrant species often produced starch in the callus of the lip, even though they d id not produce nectar (Fig. 4 24 ). However, many species of Sobralia are subtly to highly fragrant. I observed a very strong fragrance of coumarin in multiple membe rs of the Sobralia decora complex ( S. decora S. fenzliana and S. yauaperynesis ). Multiple populations of Sobralia rosea were observed to have vanillin (same population

PAGE 151

151 sampling in Ecuador measured for nectar). Sobralia bouchei has an unidentifiable fragrance that can only be described as acrid. Sobralia rarae avis has a str ong fragrance of su nscreen that is present at night only described in detail by Kaiser (1993 ) ] No floral fragrances have ever been published for Sobralieae. One species has been analyzed using GCMS for floral fragrances: Sobralia virginalis has benzyl al cohol (17%) and phenyl ethyl alc ohol (83%), both strong attract ants of male e Whitten, unpublished data). Elevational M aps When separated by putative pollinator, specimen data of the different groups of Sobralieae show clear affinities to the elevational prefe rence of pollinators (Fig. 4 23) Discussion E volutionary forces on flower morphology and functionality are, at least in part, a function of selection through pollination. Because effective pollination is dependent upon a complex suite of features and in teractions, it is necessary to study multiple qualities of both plant and pollinator In orchids, an integral part of the pollination process is the effective transfer of pollinia to the pollinator and the reciprocal transfer of pollinia to the receptive stigma. Unfortunately, during the course of this dissertation, it was not possible to observe a diverse sampling of species for precise placement of pollinia and the mechanisms of such events. Nonetheless, observations were made of the flowers directly with implications for pollinator attraction and interaction, if not on the efficiency of pollinia transfer. Such traits are often referred to constituting floral syndrome. The usage of

PAGE 152

152 different pollinators. However, this faithfulness of pollinators towards floral syndromes ( Fenster et al., 2004 ) Because nectaries and false nectaries play an integral role in reward and deceit, they are discussed here i n the context of adapt at ion to pollination systems. Also, the role of nectar volume and sugar concentration is examined in light of putative pollinators. T he total pollination syndrome, using all of these morphological and anatomical data, along with pu b lished accounts of pollinators of these plants is discussed in a phylogenetic context. The C allus 2 KI) revealed the distribution of starch within floral tissues (Figs. 4 8D & E; 4 9C, D & G; 4 13C) and with polarized light under a light microscope (Fig 4 14F). In all species in the tribe, there are two calli at the base of the lip, but with various shapes, sizes, degrees of fusion, and quantity of starch produced. Although the flowers are relative ly small compared to Sobralia the calli of Elleanthus are large relative to the flower (approximately 2 3 mm long, each) and packed with starch prior to anthesis (Figs. 4 13 & 4 14). However, Sobralia is more variable. Some Sobralia s pecies that produce nectar (e.g. S. bouchei S. callosa S. macrophylla and S. rosea ) have the two c alli fused into a thickened pad, which has relatively densely packed starch at an early stage of anthesis (Figs. 4 6, 4 7, 4 8, & 4 9). This starch is no longer present lat e in the day (Fig. 4 9C & D), presumably lost through metabolism in the production of nectar. But most species of Sobralia differ from this model by deceiving t he pollinator in several ways: 1) t here is no nectar produced; and 2) the callus is ridged and contorts to form a tube that serves as a funnel

PAGE 153

153 ovary (Figs. 4 3, 4 4, & 4 5; see later discussion). Presence of starch in the callus is a consistent feature among all flowers in the Sobralieae. The density and quantity of the starch is variable however. Even species of Sobralia that do not secrete nectar nonetheless accumulate a minimal amount of starch (Fig 4 21), although this likely serves in another function in p ollination, such as fragrance production (see later discussion of osmophores). Those species with the thickest accumulation of starch relative to flower size do produce nectar. I conclude that the callus is the nectariferous tissue of Sobralieae because o f the following three factors : 1) It is on the callus and only the callus that, if observed early in floral maturity, there is nectar deposition. Early in the life of the flower, droplets of nectar can be seen to form directly on the surface of the callus (Fig. 4 7E). 2) All calli observed for taxa with nectar secretion have starch early in the life of the flower, and by the time the flower senesces, the starch is largely exhausted. 3) Although no stomata or trichomes that might be associated with the ep idermis of som e flowers are present in any discerni ble quantity, th ese calli have a dense cellular structure that is consistent In some sections of floral tissue, the epidermis was slightly rupt ured, which might indicate secretion by lysogeny. However, this is believed to be an artifact of the lab processes in combination with the extremely delicate nature of the flowers themselves. A nectariferous callus has been reported in other orchids, but the frequency and distribution throughout the family is poorly understood. A callus that secretes nectar has been demonstrated in Maxillaria anceps Ames & C. Schweinf. ( Davies,

PAGE 154

154 Stpiczynska, and Gregg, 2005 ) Stenorrhynchos Rich. ex Spr eng. ( Galetto, Bernardello, and Rivera, 1997 ) an d in many other orchid groups ( Davies and Stpiczynska, 2008b ) In orchids, the callus is simply a term given to any unusual, raised, or ornamented portion at the base of the lip. And although the callus is not likely homologous in orchids, the ability to produce thickened tissue (i.e., a callus) may be an exaptation for r, oils or fragrances. In the case of Sobralia and Elleanthus it is nectar. Orchids are known to accumulate starch for various pollinat ion secretion activities. S tarch accumulation, followed by reduction associated with nectar secretion also has been found in some other orchids such as Hexisea Lindl. ( Stpiczynska, Davies, and Gregg, 2005 ) Acianthera Scheidw. ( de M elo, Borba, and Sousa Paiva, 2010 ) Limodorum L. ( Figueiredo and Pais, 1992 ) Epipactis Zinn ( Pais and Fi gueiredo, 1994 ) in multiple species in subfamily Orchidoideae ( Galetto, Bernardello, and Rivera, 1997 ; Stpiczynska et al., 2005 ) and among other plant families ( Durkee, 1983 ) From this widespread occurrence the advantage of having a fixed reservoir of carbohydrates, such as starch, is evident. It allows for rapid dissemination of secretions, and in the case of Sobralia and Elleanthus nectar, especially in species with short lived flowers. The ultrastructure of nectaries in flowers i s, in general, fairly well understood ( Fahn, 1979 ; Vassilyev, 2010 ) as well as the general manner of conveyance and secretion ( Pacini and Nepi, 2007 ) In orchids, floral reward anatomy (includin g nectaries, osmophores, res in secreting structures, and elaiophores) have been studied in a handful of species throughout the orchids ( Stpiczynska, 2003 ; Stpiczynska, Davies, and Gregg, 2003 ; Davies, Stpiczynska, and Gregg, 2005 ; Stpiczynska, Davies, and

PAGE 155

155 Gregg, 2005 ; Davies and Stpiczynska, 2008b ; Davies and Stpiczynska, 2008a ; Stpiczynska, Davies, and Kaminska, 2010 ) With these works in mind, floral secretions transported from cell to cell in a nectary via the symplast (the interconnected membrane system allow ing for transport of solutes from cell to cell via plasmodesmata), allowing for sugars to migrate freely through a nectary ( Fahn, 1979 ) However, the secretion process in particular can be achieved in three main ways. The first is to have stomata on the surface of the nect ary. The second is to have glandular hairs through which the nectar is secreted. Lastly, nectar can be secreted through no discerni ble structures, but simply through the cell walls of the epidermis. It is the last that seems apparent from SEM of Sobrali eae, because there are virtually no stomata or hairs on the nectar secreting tissue of the callus (Figs. 4 6C & D, 4 8C, and 4 9E & F). Also, except for S. bouchei the epidermis of the o ther species examined is highly papillose (with no intercellular spa ces), so as to provide additional surface area through which s ecretion can occur. In S. bouchei the callus surface is relatively smooth with brick shaped cells and large intercellular gaps as seen from above (Fig. 4 8C). Some orchids ( Aerangis Rchb.f. a nd Platanthera Rich.) and non orchids ( Brassica napus L.) have the ability to reabsorb the sugars secreted in nectar ( Burquez and Corbet, 1991 ; Koopowitz and Marchant, 1998 ; Stpiczynska, 2003 ) There is no evidence whether or not Sobralieae reabsorb nectar. However, because the nectary structure in So bralieae is by starch accumulation rather than direct vascularization (transfer from the phloem) and because the flowers are so short lived, it is unlikely that they reabsorb nectar.

PAGE 156

156 The D ouble C uniculus Orchids have a plethora of structures for conveying nectar to pollinators, especially long tongued insects. Som e members of tribe Vandeae (especially Angraecum Bory ) have long tubular spurs (formed from an inundation of the lip) that are associated with hawk moth pollination ( van der Cingel, 2001 ) In some orchids, a cuniculus is formed through the fusion of a hypanthium like structure, as in the Laeliinae (e.g., Brassavola R.Br. ), and forms a single tube serving much the same function as in Angraecum ( Stpiczynska, Davies, and Kaminska, 2010 ) Although all of these various structures are o bviously not homologous due to being distantly related taxa, they all converge upon a similar function to facilitate pollination, either by deceit or through legitimate nectar reward. The double cuniculus is unusual among orch ids and is probably unique to a portion of the core group of Sobralia It is formed from an open air space between the lateral sepals and the ovary and can extend for up to several centimeters through the ovary region (Figs. 4 3, 4 4, & 4 5). Perhaps the most significant aspect of the double cuniculus is that it runs deeply into the ovary. All flowers examined in this study with this structure offer no nectar reward, neither in the typical callus position of other Sobralieae nor within this cunicular reg ion deep within the ovary. Because Sobralia usually has a typical gullet shaped flower (zygomorphic, with a tubular lip and nectar as being a paired false nectary, and probably adds to the effectiveness of deceit in bee pollinated species of Sobralia This interpretation is supported by the fact that the width of the double cunicular tubes exceeds the width of the tongues of bees (e.g. euglossines). I hypothesize tha t the function of the double cuniculus is as a false

PAGE 157

157 nectary and to bring the bee further into the throat of the flower to give a higher likelihood of effective pollination ( Nilsson, 1988 ) Because the long tongued bees of tribe Euglossini are the most common observed pollinators of S obralia (Table 4 1, Fig. 4 22), this length mediated deceit is likely a significant adaptation. Unfortunately, the sampling of the double cuniculus is not complete relative to the species sampled in the phylogenetic analysis. This is largely due to the di fficulty of obtaining material of these ephemeral flowers. However, there is a general trend of small flowers lacking the cuniculus, and larger flowers having the cuniculus. Moreover, the double cuniculus is prominent within the ephemeral flowered clade of core Sobralia (Fig. 4 24 ). Nectar Q uantification Sugar composition in nectar has been studied extensively in various flowering groups, but not in Sobralieae. Many studies have demonstrated that there are differences between various nectars for differen t pollinators. For example, the range of sucrose concentrations for solitary bee nectar is 16 50% whereas the range for hummingbirds is 13 30% ( Baker, 1975 ; Baker and Baker, 1983 ) This trend has a large overlap and the immedi ate difference is only in the upper range of concentrations for bees. The largest difference among pollinators is the relative proportion of sucrose to glucose and fructose, but this also overlaps extensively. Hummingbird pollinated flowers high proportions of glucose and fructose (i.e., relatively high sucrose concentrations) in the nectar ( Baker and Baker, 1983 ) N ectar in h ummingb ird pollinated flowers is frequently reported as more dilute ( Hainsworth and Wolf, 1972 1976 ; Bolten and Feinsinger, 1978 ; Pyke and Waser, 1981 ) However, a relatively high proportion of sucrose was found in a broad sampling

PAGE 158

158 of hummingbird pollinated plan ts in Costa Rica ( Stiles and Freeman, 1993 ) Therefore, there is a large overlap between the nectar concentrations in flowers assessed of these different pollinators. To address this difference in c oncentration, it has been questioned whether hummingbirds select for flowers in this range. As hummingbird pollination is generally a derived condition, usually within insect pollinated groups ( Beardsley, Yen, and Olmstead, 2003 ; Kay et al., 2005 ) it is reasonable to assume that hummingbirds select for a specific type of nectar. However, this hypothesis of nect ar preferences in hummin gbirds has unraveled since the early works ( Johnson and Nicolson, 2007 ) Likewise, there are similar trends of sugar ratios as they relate to p ollinators of Ipomoea ( Galetto and Bernardello, 2004 ) and in other plant groups ( Galetto, Bernardello, and Sosa, 1998 ; Burke et al., 2000 ) in which no significant differences occur among nectars of plants served by different pollinators. Other surveys across many unrelated plant taxa have shown variable nectar concentrations for h ummingbird pollinated taxa ( McDade and Weeks, 2004 ) Indeed, nectar concentration in Sobralieae is inconsistent and is not particularly correlated with pollination syndrome (Table 4 2 & Fig. 4 20). Unfortunately, I was not able to quantify the different sugars (sucrose, glucose, and fructose) among these species. Nonetheless, there are some features prevalent within these data. Many of the humming bird pollinated species produced small er volumes (perhaps owing to the smaller f per flower (except in E. sodiroi which ranged he bee pollinated flowers such as S. bouchei and S. rosea which produced an average crop of 8.4 14.1 seem to have higher volumes. The aberrantly low volume of S. macrophylla can probably be explained by its inconsistency.

PAGE 159

159 It is a very poor nectar producer, and although approximately 50 flowers were examined over this study from several different plants, only r arely was nectar observed (n=6). The nectar of some e uglossine bee pollinated non orchids has been studied in ( Borrell, 2005 2006 ) However, the bee pollinated ta xa that produce rewards have relatively lo w viscosity nectar (i.e., S. bouchei S. macrophylla and S. rosea ) and the deceitful species of Sobralia have no nectar; thus, a large sampling of species that are bee pollinated and produce nectar within Sobrali a is impossible. The energetic cost of nectar as a reward has been explored. In some plants, nectar production can represent 37% of the total available energy ( Pyke, 1991 ) Removal of nectar can cause plant s to produce net higher nectar secretion, with even greater costs to the plant ( Pyke, 1991 ) As noted nectar is secreted by the callus, which accumulates starch in Sobralieae. This starch appears to be the total sum of carbohydrates allocated for nectar per flower because, after the nectar is removed, no further nectar is secreted. Extra floral nectar of Sobralia has been described ( Baskin and Bliss, 1969 ; Jeffrey, Arditti, and Koopowitz, 1970 ) However, in the course of this study, many hundreds of plants across many species sam pled in all genera were observed, and n o extra flor al nectaries were ever found t hus raising questions about the accuracy of these accounts Unfortunately, there are some potential sources of error in the measurements of sugars. The first, although minor is in the refractometer used. Because a sucrose refractometer was used three potential problems may have occurred; the first is a deviation from 20C, although all measurements were made as close as possible to that

PAGE 160

160 temperature. The second potential s ource of error is relative humidity ( Bertsch, 1983 ; Corbet, 2003 ) although measurements w ere made as quickly as possible, s o that equilibration with the ambient environment would not occur. And third although most plants were cultivated and so not available to insects or other potential pollinators to remove nectar, all plants of Sobralia rosea and Elleanthus aurantiacus were sampled in the wild. These plants could have been visited by pollinators that removed nectar (modifying volume), or been rained upon (modifying volume and concentration). However, except for v isitation by pollinators, the se factors were controlled for by careful inspection. Fragrances Fragrances are a common feature of bee pollinated taxa. In some orchids, such as those pollinated by male euglossine bees, the fragrance is a legitimate reward in itself ( Dodson et al., 1969 ; Roubik and Hanson, 2004 ) In other plant groups, fragrances can simply be attractants for many different bees. Many species of Sobralia produce fragrances, but male euglossine brushin g behavior (androeuglossophily ) has never bee n observed. Therefore, the visitation of euglossines on Sobralia is likely nectar foraging behavior and the fragrances might simply add to the attractive quality of the flowers. However, there is still the question of where the fragrances are produced. As with many other orchids, fragrance production is also associated with the accumulation of starch at the site of secretion (i.e., the osmophore). Starch accumulation for fragrance secretion in osmophores is known from Stanhopea Frost ex Hook. and relati ves ( Curry et al., 1991 ) Scaphosepalum ( Pridgeon and Stern, 1985 ) and Acianthera ( de Melo, Borba, and Sousa Paiva, 2010 ) Many species of Sobralia produ ce fragrances and

PAGE 161

161 accu mulate starch in the epidermis (and sometimes a few cell layers below, but never forming a thick pad such as in the nectary) of the callus. Therefore, the callus is likely a multifunctional organ and probably serves as an osmophore a s well as a nectary The structure of this putative osmophore differs from the nectary by usually having a hairier texture (Fig. 4 3B, 4 21). This argument is further supported by the dissection of flowers and smelling their individual parts ; the lip is the most fragrant (pers. obs.). However, species can both produce nectar and be fragrant (e.g., S. bouchei and S. rosea ), so the callus might have a dual function. In fact, for both S. bouchei and S. rosea the basal part of the callus is nectariferous, and the more distal end is still rich in starch, but does not have nectar (Figs 4 7E, 4 8D). This distal callus region may be involved in fragrance production, but further testing is required to support this hypothesis. Although I was not able to directl y quantify or qualify the fragrances of other species with GCMS, some fragrances were identified by nose. Many unidentifiable and faint odors were produced by the var ious species of core Sobralia b ut one of the more distinct fragrances identified was van illin in Sobralia rosea This is a common attractant to euglossines ( Williams and Whitten, 1983 ) S pecies in the S. decora complex (see chapter 2 for details) commonly produce coumarin. Although when baited with coumarin, euglossines are not attracted ( Bennett, 1972 ) it may still play a role in attracti ng other groups of bees. S obralia recta produces a sharp, resinous (terpinoid?) odor. The inference is that although an androeuglossophilous pollination syndrome is not out of the question, it seems unlikely. Both male and female bees have been observed to visit Sobralia flowers, but likely for nectar ( Dodson, 1966 ) Other orchid

PAGE 162

162 groups such as Cattleya ( Kaiser, 1993 ) and Cochleanthes ( Ackerman, 1983 ) are known to have deceptive nectar fora ging morphology, pollinated by e uglossine bees, and produce androeuglossophilous fragrances. Only one species ( S. rarae avis ) was observed to produce fragrance at night. Flowers of most species senesce and deliquesce in the late evening, and therefore could not be fragrant, much less pollinated at night Floral S tructure as It Relates to P ollination pidendroideae are soft and mealy ( Harder and Johnson, 2008 ) The presence of this type of pollinia in many Sobralia is a symplesiomorphy. The derived condition of having sligh tly hardened, but especially non mealy textures, such as in Elleanthus Sobralia callosa and S. amabilis is derived. However, because it is common among hummingbird pollinated orchids, it is likely a n adaptation to that syndrome. Additionally, pollinia of hummingbird pollinated orchids are darkened, an adaptation that has been known for many groups of Neotropical orchids ( Dressler, 1971 ) These derived, hummingbird associated features of pollinia have a morpho logy that seems to be somewhat labile, considering the number of orchid lineages that have converged upon the same suite of characteristics (see later discussion). In most species of Sobralia there is a typical gullet type, resupinate flower with yellow o r red nectar guides (Fig s. 4 1, 4 3 through 4 8). Flowers of this type are visited and effectively pollinated by bees of all types, but especially by euglossine bees ( van der Pijl and Dodson, 1966 ) Bees are drawn in with the probable interest of nectar foraging based on these featur es. Because the lip is presented on the lower surface and the column on the upper, the likelihood of pollinia deposition on the dorsal surface

PAGE 163

163 of the bee is extremely high. Species of Sobralia associated with the aforementioned floral morphology also hav e a downward facing stigma and an elastic rostellar flap, secreting a sticky fluid that binds to the pollinia. This rostellar flap can only function properly as the bee is exiting a flower, walking in reverse (posteriorly) because there is insufficient ro om within the narrow tubular confines of the labellar entry to turn around. The pollinia are deposited on the scutellum (a posterior and dorsal thoracic segment) of the bee because it is the first segment of the bee of sufficient size and shape to effecti vely contort the rostellar flap (Fig. 4 22). That is, because the rostellar flap faces inwards within the flower, its function of spreading a sticky secretion onto the pollinator can only occur as the insect exits. Pollination would therefore occur in a different flower as the bee makes the sa me series of movements; the rostellar flap serves its second function of scraping the pollinia off and depositing them on the stigmatic surface, which is adjacent to the rostellum. It is extremely plausible that a s econd pollinial mass could be deposited on the same bee simultaneously with this pollination event. Finally, the pollinial mass in bee pollinated flowers seems to be relatively large (~1 10 mm wide), mealy in texture, and white to yellow in color. A larg e pollen mass maximizes the number of pollen grains delivered and might be adaptive for maximizing the number of seeds produced per capsule. Anecdotal evidence of a positive correlation between pollinial size and fruit size support this argument (see late r discussion). If the function of the double cuniculus is a false nectary, it is likely to serve the function of fooling long tongued bees into deeply probing the flowers. Conversely, when a bee probes the depth of a flower, never finding nectar, and has not gone deep enough to remove the pollinia, then po llination will not occur. It is therefore likely that

PAGE 164

164 species that produce nectar (in the callus rather than the double cuniculus, which is always a deceitful mechanism) must have adaptations of length t o effectively fool the bees. For example, Sobralia rosea has very large flowers with a particularly long tubular throat (~5 cm), at the base of which is the true nectary. Another nectar producing species, S. bouchei is much smaller than S. rosea and has a different conformation of the column. In S. bouchei the stigma is slit l ike, forward facing, and lacks an elastic rostellum (Fig. 4 8) such that a pollinator would deposit the pollinia into the stig ma upon entry of a flower. W ith that change, the posi tioning of the pollinia would have to be on the front of the head (frons) or the front of the thorax (pronotum), depending on the size and shape of the bee. It is not clear why this elaborate structural change has occurred in S. bouchei when the plesiomor phic condition of column structure previously discussed seems well suited to bee pollination. Pollinia transfer by hummingbirds appears to be a very different process from that in a typical bee pollinated flower, with multiple floral morphological changes (Fig. 2). In Sobralia callosa S. amabilis and Elleanthus the lip is modified to form a small cup like structure in which the nectar collects in a reservoir. The cup that results is enclosed by a notch in the lip, which is constricted, and by the colum n where the lower edge of the stigmatic rim comes in contact with the lip The stigma is forward facing and the l arge rostellar flap is absent Specifically for Sobralia callosa the two lateral lobes on the stigma would center the bill of the hummingbir d so that deposition of the pollinia would be on the upper surface. In some species of Elleanthus the presentation of the flowers is likely non resupinate, so that the pollinia would be deposited on the lower bill. The pollinia in hummingbird pollinated species share the common features of being

PAGE 165

165 relatively reduced in size, relatively hard in texture, and usually bluish gray in color. This syndrome differs in the pollination event, because the forward facing stigmas would require that the pollinia be dep osited on the stigma as the bill is being pushed into a flower. In this sense, it is the lower lobes of the stigma (as in Sobralia callosa ) or the infrastigmatic ligule (as in Elleanthus ) that provide a rigid structure to force the pollinia already on the bill of the bird into the stigma. The pollinia of the same flower could then be removed and deposited on the bill of the hummingbird as it is removed from the flower. This functional difference of pollination, as manifested in the stigmatic orientation is probably important to pollination success. One species of Sobralia S. crocea appears to be hummingbird pollinated (L. Endara, pers. comm.). It has small tubular flowers that are orange to orange red and are pendant (Fig. 4 11 ) It is common in the Andes in wet montane forests where it is sympatric with various hummingbird pollinated taxa with similar floral size, shape, color, and pendant orientation. However, although the pollinia are highly reduced (~2 mm in width), they are white. Furthermore, t he flowers produce no nectar, and if pollinated by hummingbirds, this is the only known species to be deceitful. Because S. crocea lies within the core group of Sobralia where virtually all of the species give no nectar reward, it is simply a retained sy mplesiomorphy for this species to use deceit. The column structure and pollinia deposition mechanism of S. crocea however, are more similar to those of bee pollinated flowers. There is some evidence that Elleanthus flowers undergo protandrous activity, a t least in Elleanthus brasiliensis ( Singer, 2003 ) This evidence is apparently based on the conformation of the column within the flower. As the flower first opens, the anther is

PAGE 166

166 presented at a lower position relative to the lip, perhaps making the likelihood of pollinia remov al higher. Later in the development of the flower, the anther rotates towards a more dorsal position within the flower, exposing the stigma in a more forward facing orientation. This subtle difference in column orientation may have a significant effect o n receptivity. However, during the course of this study, some observations of other species of Elleanthus (e.g., E. sodiroi E. caravatus and E. cynarocephalus ) in a greenhouse setting showed that this putative protandry might be relatively common in the genus. Protandry has never been observed in Sobralia nor is it likely because of the ephemeral nature of the flowers (i.e., protandry in Elleanthus is a process th at occurs over several or more days). In virtually all species of Elleanthus the flowers are non resupinate with respect to the axis of the inflorescence. However, inflorescences are presented in a wide variety of orientations Some groups have inflores cences that point down (either by direct growth downwards, such as in the E. longibracteatus clade of sect. Stachydelyna or by being so heavy that the inflorescence can not be supported by the stem, such as in E. capitatus of sect. Cephalelyna ) and theref ore have a resupinate orientation as it relates to gravity. Other groups of Elleanthus have vertically held inflorescences with flowers in the non resupinate orientation (such as sect. Calelyna sect. Hymenophora s.l., and E. aurantiacus ). The significan ce of this is the likely consequence of positioning of pollinia as they are deposited on the hummingbird s A resupinate flower would likely place the pollinia on the upper surface of the bill, whereas the non resupinate flower would likely place the polli nia on the lower surface. The advantage of color size, texture, and

PAGE 167

167 position of pollinia was proposed decades ago ( Dressler, 1971 ) based on the preening behavior by birds, but it has never been tested. The obvi ous conclusion is that a pollinium placed on the lower surface of th e bill would have a greater fit ness because the bird could not see it and remove it. This aspect of positioning was not observed in this study, so cannot be tested. However, the point ma y largely be moot in comparing taxa that are resupinate versus non resupinate (either quantified by species richness or by fruit set); fruit set is extremely high among many Elleanthus species, regardless of floral orientation (pers. obs., and see Fig. 1 1 1). And although exact species numbers as they apply to these floral orientations cannot be estimated easily, the resupinate position seems to be much more common. Further more, the resupinate position in other groups of hummingbird pollinated orchids is much more common (e.g., Ada Lindl., Broughtonia R. Br., Camaridium Lindl., Coccineorchis Schltr., Cochlioda Lindl., Cyrtochilum Kunth, Domingoa Schltr., Fernandezia Ruiz & Pav., Guarianthe Dressler & W.E. Higgins, Prosthechea (Lindl.) W.E. Higgins, Hexisea Lindl., Isabelia Barb. Rodr., Isochilus R. Br., Odontoglossum Kunth, Sacoila Raf., Ornithidium Salisb. ex R. Br., Sertifera Sobralia Sophronitis Lindl., Stenorrhynchos Rich. ex Spreng., among others) versus non resupinate (e.g., Arpophyllum LaLlave & Lexa ra and Elleanthus in part). F or an illustrative description of Sobralieae pollinated by hummingbirds see Siegel ( 2011 ) The apparent conclusion is that pollinia color, commonly darkened in hummingbird pollinated orchids, especially in the taxa above ( Dre ssler, 1971 ) is more important than its positioning on the bill. Whatever the case, pollinia size reduction is obviously adaptive to fitting the bill and is a consistent feature am ong hummingbird pollinated taxa in Sobralieae and across Orchidaceae

PAGE 168

168 Ac cording to the phylogenetic relationships within the tribe, long lived flowers (2 7 days) are relatively plesiomorphic. Most species of Sobralia as indicated by the clade Sobralia l flowered species of Sobralia tend to flower gregariously, having populations that simultaneously present a single flower per inflorescence at a time, and then wither, and after a varying amount of time, they will repeat the process for a season. Because the flowers of Sobralia are ephemeral, we infer that it is physiologically important to localize starch in early flower development for rapid nectar or fragrance production. Mimicry and D eceit Mimicry is the process by which species evolve shared perceiv ed similarities that confer some evolutionary advantage. Mimicry is divided into two broad categories: Batesian and Mllerian. In pollination biology, Mllerian mimicry is the presence of two or more rewarding species tha t share a common morphology to ac cess a pollination niche, mutually benefitting by providing resources to pollinators. Conversely, in Batesian mimicry, one or more species is deceitful, and with no reward, it is effectively taking advantage of another species, without having to produce a n investment in energy (i.e., nectar). This latter case is the only observable interaction known from Sobralieae, mimicry is a complex topic that involves a wide arra y of biological processes and interactions. To keep this discussion concise, I will address food mimicry, as that is likely the only process at work. The supposition of mimicry is based on the presence of deceit within Sobralia Although there may exist examples of specific interactions of Sobralia species taking advantage of rewarding species in their local environs, Sobralia like many food deceitful orchids, probably take s advantage of a general gestalt that is

PAGE 169

169 attractive to a wide variety of pollinat ors. This is termed generalized food deception ( Jerskov, Johnson, and Kindlmann, 2006 ) and the mechanism is apparently frequent and sometimes referred to as naivet ( Ackerman, 1986 ) Sobralia can have many bee pollinators (Table 1), which accords with them being generalized food decei ving. Although many species of pollinators are not necessary for this deceit to be successful it is reasonable to assume that these generalized features are attractive to many different species of bee s and that many species lend a higher likelihood of pollination success, in general. Furthermore, generalized food deception based on generalized foraging behavior has been demonstrated in many orchids; see Jerskov Johnson, and Kindlmann (2006) for a detailed list of such groups Some take a very strict definition of mimicry (i.e., a one to one sense) and discount the generalized forms of ( Roy and Widmer, 1999 ) However, strong one to one pollinator to plant relationships can exist ( Schiestl and Schluter, 2009 ) although it is not known in any pollination system in Sobralieae. The lack of floral rewards might benefit the deceptive Sobralia by promoting outcrossing, as the lack of nectar might drive the pollinators to visit more plants ( Jerskov and Johnson, 2006 ) Of course, the co within the landscape of deceitful ones enhances their likelihood of being pollinated ( Feinsinger, 1978 ; Johnson et al., 2003 ) One common measure of success of pollination is fruit set. Nearly all Sobralieae are self compatible and outcrossing; autogamy is frequent in only one species, Sobralia madisonii reely, the stipe of the pollinia recurves to deposit the pollinia into the stigma of the same flower or there is some environmental mediator,

PAGE 170

170 such as rain ( Catling, 1980 ) This structural constraint makes it virtually impossible in Sobralieae to self pollinate because the stipe does not exist and the pollinia sit naked within the clinandrium, making it very difficult to be repositioned into the stigma without external forces. However autogamy is known in several species: Sobralia exigua which is cleistogamous ( Dressler, 2005 ) and S. madisonii in which the mechanism is unknown (Whitten, pers. comm.). Ne ctar rewards are associated with higher pollination success, and thus higher fruit set, in various orchids ( Neiland and Wilcock, 1998 ) This also seems to be true in Elleanthus versus Sobralia Although the evidence is largely anecdotal, rather than quantitative, Elleanthus observed in the wild exhibit high fruit set, with nearly every flower within an inflorescence setting fruit (Fig. 1 11A, B, & C); plants in cultivation (without pollinators) r arely set fruit. This prodigious fruit set is contrasted with many species of Sobralia which exhibit lower fruit set (usually less than 20%). The net seed produced is likely offset, because these different rates of fruiting are inversely proportional to fruit size. Elleanthus has relatively small fruits, whereas Sobralia has relatively large (usually >10 cm in Sobralia and usually <2 cm in Elleanthus ). I was not able to quantify the differences in number of seeds produced per fruit for practical reason s, but there is probably a strong difference in the number of seeds within each fruit. Likewise, the pollinia of Sobralia are larger than Elleanthus and may convey a larger quantity of pollen that can effectively fertilize more ovules making more viable seeds. The cost/benefit of fruit production versus nectar reward may be evident in fruit size. If pollination of deceitful species (i.e., most of Sobralia ) is indeed more rare than those of rewarding species (i.e., Elleanthus ), then it seems reasonable t hat fruit size is an adaptive response to this shortcoming.

PAGE 171

171 The architecture of inflorescences can have a large effect on the presentation to pollinators. As previously discussed in chapter 2, many members of Sobralia sect. Sobralia have axillary inflores cences that are displayed for a season, and then die. The stems are long lived and continue to grow for multiple seasons. However, most species in the tribe have terminal inflorescences. More importantly, the number of flowers produced and the structure of presentation differ greatly. Elleanthus inflorescences typically produce many flowers simultaneously, often with 5 50 open at the same time and lasting at least several days. But these inflorescences undergo anthesis sequentially ( i.e., they are inde terminate) so that they may produce flowers for many weeks or even months. Because of these factors, a high rate of geitonogamy (fertilization from another flower on the same plant) in rewarding species such as in Elleanthus is probable. This argument is supported in the case of artificial nectar additions in deceptive species of Orchis L. where higher rates of geitonogamy were observed ( Johnson and Nilsson, 1999 ) However, Smi thson ( 2006 ) found no difference in inbreeding depression between a group of rewarding and deceitful species in Barlia Parl. and Anacamptis Rich This result of course, must be examined on a case by case basis. After all deceptive specie s of Sobralia generally produce only a single flower at any given time per inflorescence, but there can be multiple inflorescences per plant. I f ephemeral flowers are adapted for outcrossing in the context of deceit, then this explains much of the diversi ty and success of Sobralia because this combination of traits is predominant within the genus. Synchrony may enhance the likelihood of pollination success in Myrmecophila ( Parra Tabla and Vargas, 2007 ) but those inflorescences have multiple flowers, so the dynamics of plant pollinator interactions is

PAGE 172

172 likely different. Deceptive species may be more variable than rewarding species ( Ackerman, Cuevas, and Hof, 2011 ) b ut it is not clear that higher variation translates into a function of floral recognition (or lack thereof) and thus higher rates of p ollination. However, this concept of floral density is complex. Floral density affects the gene flow by a function of distance ( Knight, 2003 ) R are flowering species may be at a disadvantage ( Kunin and Iwasa, 1996 ) S ynchronous flowering probably increases effective population size and probably increas es gene flow by reducing distance [i.e., ( Feinsinger, 1978 ) ]. And, if individuals of Sobralia are naturally rare in the landscape, but flower synchronously, the large dista nces between plants may be compensated by the ability of euglossines to travel very long distances regularly ( Janzen, 1971 ) Because of the different kinds of pollinators and floral diversity, Sobralieae presents a great model system for studying the effect of inflorescen ce architecture and floral presentation on pollination. Sphingophily One species has such unusual features that it should be mentioned, even if there is no firm evidence regarding its pollinators Sobralia rarae avis smell emi tted at night, with a white and green floral presentation, and a long tubular portion of the perianth, similar to the hawk moth pollinated (sphingophilous) flowers of Angraecum ( Kaiser, 1993 ) or Brassavola ( Roebuck and Steinhart, 1978 ; Williams, 1981 ) This species has an atypical column structure where the anther is presented more towards the dorsal surface, making room for a very b road and rigid clinandrium and broad, downward c urving column wings that create an inverted U shaped opening between the column and the lip (Fig. 4 12). The result is a small tubular opening that is extreme ly difficult to open and therefore most likely to be probed

PAGE 173

173 by a long tongued insect. Because of this long tubular stru cture in combination with nocturnal fragrance, this taxon is likely pollinated by moths (especially Sphingidae) with pollinia deposition on or near the tongue ( Maad and Nilsson, 2004 ) ; other possibilities include nocturnal beetles, or ptiloglossine bees that are active after dusk and before dawn. Nectar was not observed because of the paucity of floral mate rial (only a single flower was observed), a problem that plagued the describer of this species ( Dressler, 2007 ) Also, even though this plant was carefully watched d uring its flowering time, it was not suspected of having anthesis at night until several flowers had already expired. Field observations of pollination are needed to determine the pollinator for this unusual species. Elevation and P ollinators In montane a reas, such as those of Central and South America, elevation can play an integral role in plant and pollinator distributions (see chapter 2 for a detailed discussion of elevation and biogeography). In a reciprocal sense, elevation can differentiate the pol linator guilds among plant groups and can be used as an additional line of evidence to support plant pollinator interactions But where are the pollinators of Sobralieae found? Hummingbird diversity ranges from sea level to well over 4 000 m but the highe st abundance is between 2500 and 3500 m ( Feinsinger et al., 1979 ; Bleiweiss, 1998b ; Stiles, 2004 ) Li kewise, the diversity of bees (especially Euglossini) primarily ranges from sea level to 1600 m ( Dressler, 1982 ) but mostly from sea level to 900 m ( Ramirez et al., 2010 ) Other insects such as moths, although diverse at high elevati ons, have the highest diversities at less than 2000 m ( Beck, Brehm, and Fiedler, 2011 ) Although these metrics are not censuses of

PAGE 174

174 the exact pollinators of Sobralieae, they do represent general tren ds of where the major groups of known pollinators are found Elevations within the groups of Sobralieae differ greatly, depending on the assessed pollination syndrome (Fig. 4 23 ). For some Elleanthus and Epilyna species that are putatively pollinated by an unknown insect have a relatively low average elevation (950 m), compared to those Elleanthus species pollinated by hummingbirds (1800 m). Within core Sobralia species that are primarily pollinated by bees have a lower average elevation (850 m) compar ed to those pollinated by hummingbirds (1800 m). And in Sobralia sect. Sobralia along with Sertifera species pollinated by bees have a relatively low elevation on average (1850 m) compared to those pollinated by hummingbirds (2300 m). With these ranges and averages in elevations in mind, the elevational differences among the different groups of Sob ralieae are consistent with those of the ir putative pollinators. One of the greatest inconsistencies within these data is the relatively high elevational pref erence of Sobralia sect. Sobralia even though it is bee pollinated. This is most likely explained by the group using a d ifferent group of bees than euglossine bees which prefer relatively low elevations However, little is known of the pollination in t his group, so this is merely speculation. Evolution of Pollination S yndromes Within a phylogenetic context, many hypotheses relating to the evolution of pollination can be addressed, either through a study of morphological change in the plants (or even the pollinators) or of the syndrome itself. Hummingbird pollination has evolved many times from groups that use different pollinators, usually various kinds of bees. Among these many taxa, Aquilegia L. shows bee, moth and hummingbird

PAGE 175

175 adaptations ( Whittall and Hodges, 2007 ) A larger review paper ( Thomson and Wilson, 2008 ) also discussed the evolutionary histories of Costus L. Ipomoea L. Mimulus L. Penstemon Schmidel and Silene L The overarching pollinator shifts are correlated with change in color, corolla shape, and floral orientation. More importantly, evolutionary shifts in poll inator class are directional, usually transitioning from bee pollination to hummingbird pollin ation and not vice versa. The phylogenetic and pollination data of Sobralieae indicate that there have been six independent origins of hummingbird pollination in the tribe It has been suggested that hummingbirds drive speciation ( Schmidt Lebuhn, Kessler, and Hensen, 2007 ) and in some cases (e.g., Tillandsia L. ) with sister clades of bee and hummingbird pollinated species the clade with hummingbird pollination has significantly highe r species richness This is indeed consistent with Sobralieae in the case of Elleanthus which is sister to the sing le species, Sobralia dorbignyana but that is not the case for Sobralia callosa S. amabilis S. crocea S. ciliata or Sertifera The latter examples are independent origins of hummingbi rd pollination (Fig s 4 24 & 4 25 ). However, S. amabilis S. callos a and S. crocea are all species with recently derived hummingbird pollinated systems that have not had as much time to diversify as has Elleanthus (see Chapter 3 for temporally calibrated cladograms, Fig s 3 9 & 3 10 ). This may explain the relative differ ence in species richness among the independe nt origins of hummingbird polli n a ted taxa. However, sister clade comparisons of species richness in the case of Sobralieae is probably not fair because of the number of independent origins of hummingbird pollina tion and the subsequent number of sister clades that include heterogeneous mixtures of pollination syndromes.

PAGE 176

176 Molecular phylogenetic studies of the Sobralieae (Fig. 4 25 ) show that Sobralia callosa S. ciliata and S. amabilis are not very closely related, despite having very similar flowers (Figs. 4 9, 4 10, & 4 15). According to these data, S. callosa is more closely related to S. bouchei while it is unclear which species is the closest relative of S. amabilis More enigmatic is S. ciliata which has u nclear relationships (see chapter 2 for ambiguity between nuclear and plastid data sets). The classic visual cues of flowers include color and nectar guides ( Jerskov, Johnson, and Kindlmann, 2006 ) Although it is widely held that hummingbirds prefer red as a visual cue, this is an oversimplification ( Healy and Hurl y, 2001 ) Despite the fact that some hummingbirds avoid yellow feeders ( Lyerly, Riess, and Ross, 1950 ) some hummingbird pollinated species of Elleanthus have yellow flowers (e.g., E. caravata ). Yellow, magenta, orange, purple, and white are among the colo rs found in Elleanthus and Sobralia that are attractive to hummingbirds. Because these colors are used in combination, it could be the stark contrast that is the primary attractant of these birds. Furthermore, the ability of hummingbirds to learn color a nd reward association may be the best advantage to the plants ( Healy and Hurly, 2001 ) and might explain the labi le nature of shifts to hummingbird pol linators Color as we recognize it may not be as clear cut to pollinators. It is now widely known that vision in bees takes advantage of the ultraviolet wavelengths ( Peitsch et al., 1992 ) However, vision in the ultraviolet spectrum is also present in hummingbirds, but to a lesser degree ( Goldsmith, 1980 ) The importance of ult raviolet vision may be over emphasized relative to normal color vision ( Kevan, Chittka, and Dyer, 2001 ) Indeed, similar col ors in the visual range of humans can have a highly divergent UV

PAGE 177

177 reflectance and so attract bees and hummingbirds differently ( Lunau et al., 2011 ) Red may be a poor attractant to bees, but it is nonetheless still found in putatively bee pollinated flowers (e.g., S. helleri and S. recta ). Regardless of color vision, the most plausible explanation for the changes of vi sual cues among different pollinators may be their cognitive preferences and learned behaviors ( Nilsson, 1992 ) According to Endress ( 1994 ) flowers are solitary or loosely clustered in a pendant position; Elleanthus and Sobralia do not fit this model. However, the character suites associated with this syndrome in orchids are somewhat different from those of other angiosperm groups. They inc lude diurnal anthesis, weak zygomorphy, tubular flowers, vivid colors, no fragrances, abundant nectar, no nectar guide, horizontal or pendant orientation, and coriaceous flowers ( van der Pijl and Dodson, 1966 ) H aving some, but not all of these characters is obviously sufficient to a ttract hummingbirds successfully Among the six independent evolutionary transitions of hummingbird pollination is Sertifera (Fig. 4 25 ), although there are no definitive observations of hummingbird pollination in this genus Interestingly, these plants g row at relatively high elevations (2200 3150 m) in the Andes and are frequently sympatric with other hummingbird pollinated families such as Ericaceae (e.g., Cavendishia Lindl. which often has the same contrasting colors of pink or purple and white). Ser tifera may mimic (i.e., by Mllerian mimicry ) other species that attract birds It is not clear if Sertifera has cryptic pollinia or if the flowers produce nectar. It was not possible to get floral material of this genus during the course of this study. However, from descriptions, the lips do produce

PAGE 178

178 a cupule structure which is similar to Elleanthus and Sobralia callosa in which nectar might be secreted and stored. The largest unknown in the evolution of pollination systems in Sobralieae is that of the small white flowered species in Elleanthus and Epilyna (Fig. 4 25 labeled as E. lancifolius (sect. Elleanthus ), all of sect. Chloidelyna (e.g., E. fractiflexus E. graminifolius E. linifolius E. p oiformis and E. stolonifer ), E. caricoides and all of Epilyna Although in combination with the rest of Elleanthus they all form a group with reduced flowers, these flowers are even smaller than the typical hummingbird pollinated species and have no bri ght colors. It is highly unlikely that these species could be pollinated effectively by hummingbirds because of their size, nor is it likely that hummingbirds would be attracted to them. Moreover, these species have yellow pollinia and very small quantit ies of nectar (<1 ). Small nectar seeking moths (e.g, Noctuidae) have been hypothesized as pollinators (C. Dodson, pers. comm.). Future D irections When Darwin wrote his seminal book on orchid pollination, he outlined many remarkable behavioral and morphological features in orchids that are still significant. His only observation of polli nation in Sobralieae was second hand account of a British bumblebee on Sobralia macrantha in a greenhouse ( Darwin, 1862 ) Upon exiting the flower, the bee was described as being clumsy and confused. This observation of in toxication was reflected a century later in another landmark orchid pollination book, but on Sobralia rosea and S. violacea in Ecuador ( van der Pijl and Dodson, 1966 ) However, as discussed elsewhere, of those species, only S. rosea actually produces havior is significant and mediated by the plants, then the

PAGE 179

179 cause is not likely to be nectar, as the presence of nectar is inconsistent among the species of Sobralia Although these floral anatomical, morphological and pollinator observations are indicativ e of clear pollination shifts, more data are needed. The most glaring gaps in knowledge are with species level plant pollinator relationships, especially for the white flowered species of Elleanthus sect. Elleanthus sect. Chloidelyna and Epilyna Verif ication of hummingbird pollination in other taxa, such as Sertifera Sobralia ciliata S. callosa and S. crocea is also critical for accurate interpretations of the number of changes to this derived pollination syndrome. And perhaps even more intriguing is S. rarae avis (and the putatively closely related S. madisonii and S. infundibuligera neither of which was examined morphologically in this study). If these species are sphingoph ilous, then it would be evidence for a more complex array of floral evol ution within Sobralieae than previously thought. It may never be practical to observe the ephemeral flowering species of Sobralia and thus to extensively study many different species. However the advent of relatively cheap digital video cameras might pr ove to be a useful tool for documenting pollinators.

PAGE 180

180 Figure 4 1 Sobralia warscewiczii ( Blanco 2676 ) showing the various morphological aspects of flowers associated with large bee pollination. A) Front view of relatively large flower, showing the en try point of the pollinator in yellow, in between the lip and the column, scale bar = 5 cm. B) Flower from the side view, emphasizing the depth of gullet morphology, scale bar = 5 cm. C) Ventral side of the column (which faces the lip) with distal end to the right, scale bar = 1 cm. Note the stigma, rostellar flap (rostellum) and anther. D) Pollinia (mass of pollen) showing the relatively large size, the pale yellowish color, and the mealy texture, scale bar = 1 mm.

PAGE 181

181 Figure 4 2 Elleanthus sp. ( Neu big 203 ) showing the various morphological aspects of flowers associated with hummingbird pollination. A) Front view of relatively small flowers grouped into a spiral inflorescence, showing the visual target as a brown spot, in between the lip and the col umn, scale bar = 1 cm. B) A longitudinal section of the flower from the side view, emphasizing the saccate labellar morphology, and the median ridge of the column forming a puncture of lip showing the relatively large globose calli that secrete nectar, scale bar = 0.5 cm. D) Pollinia (mass of pollen) showing the relatively small size, the purple color, and the hardened texture, scale bar = 1 mm.

PAGE 182

182 Figure 4 3 Sobralia decora ( Whi tten 3280 ) flower; a bee pollinated flower with no nectar reward. All blue surfaces are stained with methylene blue. A) General floral morphology, scale bar = 1 cm. B) A longitudinal section of the flower, scale bar = 1 cm. Note the long ridged callus. C H) Serial transverse sections of the pedicel, ovary, and perianth, scale bar = 1 mm. Note the two vacant spaces (paired cuniculi) present between the sepals and the column fused to the lip; these form a pair of false nectar spurs.

PAGE 183

183 Figure 4 4 Sob ralia warscewiczii ( Blanco 2677 ) flower, a bee pollinated flower with no nectar reward. All blue surfaces are stained with methylene blue. A) A longitudinal section of the flower, scale bar = 1 cm. Note the long ridged callus and that the pollinaria hav e been deposited within the stigmatic cavity. B G) Serial transverse sections of the pedicel, ovary, and perianth, scale bar = 1 mm. Note the two vacant spaces (cuniculi) present between the sepals and the column fused to the lip.

PAGE 184

184 Figure 4 5 Sobra lia chrysostoma ( Neubig 213 ) flower, a bee pollinated flower with no nectar reward. All blue surfaces are stained with methylene blue. A) General floral morphology, scale bar = 1 cm. B G) Serial transverse sections of the pedicel, ovary, and perianth, s cale bar = 1 mm. Note the two vacant spaces (cuniculi) present between the sepals and the column fused to the lip.

PAGE 185

185 Figure 4 6 Sobralia macrophylla ( Blanco 3022 ) a bee pollinated flower with a nectar reward. A) General floral morphology, scale bar = 1 cm. B) Light microscopy of callus showing dense cellular arrangement a t surface, the secretory area of nectar, scale bar = 0.5 cm. C) SEM of callus surface, scale bar = 1 mm. D)

PAGE 186

186 Figure 4 7 Flowers of Sobralia rosea a bee pollinated flower that produces nectar rewards. A D ) Flowers across a population in Ecuador, showing the range of variation in color, scale bars = 1 cm. This color difference is the primary morphological difference S. rosea and S. pulcherrima supporting the taxonomic circumscription of a larger, more bro adly defined species (i.e., S. rosea s.l.). E) The callus of the lip of one of these flowers, showing the local secretion of nectar upon the callus, scale bar = 1 cm.

PAGE 187

187 Figure 4 8 Flower of Sobralia bouchei ( Blanco 3009 ), a bee pollinated flower that produces nectar rewards. A) Frontal view of this flower, scale bar = 1 cm. B) Ventral surface of the column, scale bar = 1 cm. Note the highly differentiated stigmatic orientation of front facing surface which would require very different mechanical dep osition during the pollination process; the pollinia would be scraped off during entry of the flower, and thus deposited in the stigma. C) SEM of the surface of the callus, showing very different cellular surface texture compared to other Sobralia species scale bar = 1 mm. Note the pores (intercellular spaces), which probably serve to increase surface area for nectar secretion. D) Basal portion of young lip, stained with potassium iodide to indicate starch, precisely outlining the callus, scale bar = 1 cm. This starch is the putative carbohydrate source for the nectar. E) Transverse section of lip, showing the same callus staining the starch black, scale bar = 0.5 cm. This thick pad represents the fusion of the two distinct calli seen in most other me mbers of the tribe.

PAGE 188

188 Figure 4 9 Flower of Sobralia callosa ( Blanco 3021 ), a hummingbird pollinated flower that produces nectar rewards. A) Frontal view of this flower, scale bar = 1 cm. B) Oblique view of ventral surface of the column, scale bar = 0.5 cm. Note the highly differentiated stigmatic orientation of front facing surface which would require very different mechanical deposition during the pollination process; the pollinia would be scraped off during entry of the flower, and thus deposited in the stigma. C D) Transverse sectio ns of the lip and stained with L tection in a young flower (morning ) and an old flower (evening), respectively, showing the gradual reduction in starch over time, sca le bar = 1 mm. E) SEM o f the wh ole callus, scale bar = 1 mm. F) Surface of the callus, showing very different cellular surface texture from S. bouchei (Fig. 4 8) extremely papillose surface texture which probably serve s to increase surface area for nectar secretion. G) Basal portion of young l ip, cleared, then stained with L allus, scale bar = 1 mm. H) Transverse section of lip showing detail of cellular structure,

PAGE 189

189 Figure 4 10 Floral morphology of Sobralia amabilis a hummingbird pollinated species that probably produces nectar. Photo by R. Hoyer.

PAGE 190

190 Figure 4 11 Flower of Sobralia crocea ( Neubig 206 ) a nectar deceit spec ies with putatively hummingbird pollination. A) Plant in situ, scale bar = 5 cm. B) Flower from side, showing tubular, orange perianths, scale bar = 1 cm. C) Flower from fron t, scale bar = 1 cm. D) Distal end of lip showing dark red striations on the low raised keels, scale bar = 1 cm. E) Basal end of lip, showing two ridged callus type (with no nectar), characteristic of the core Sobralia, scale bar = 1 cm. F) Ventral surf ace of column, scale bar = 1 cm. G) Oblique view of column with anther moved to show the cream color, scale bar = 1 cm.

PAGE 191

191 Figure 4 12 Floral morphology of S. rarae avis ( Blanco 870 ) A) Front view of flower. B) Side view of flower. C) Close up of throat of flower, not e the expanded clinandrium and pair of column wings that forms a rigid structure that fills a large portion of the throat, except for a narrow tubular opening. D) Top view of lip. E) Side view of lip. F) Side view of column. G) Ven tral view of column. Scale bars = 1 cm.

PAGE 192

192 Figure 4 13 Flowers of Elleanthus caravata ( Neubig 202 ), a hummingbird pollinated flower that produces nectar rewards. A) Inflorescence showing the bright color contrast of bract and flower, typical of bird pollination, scale bar = 1 cm. B) Flower showing saccate base where nectar is secreted and stored, scale bar = 1 cm. C) The callus of the lip in a young flower, stained with potassium iodide indicating the presence of starch, scale bar = 1 mm. D) Pollin ia showing relatively small size, dark color, and hard texture, scale bar = 1 mm. E) Longitudinal section of callus, scale bar = 1 mm. F) Transverse section of callus, scale bar = 1mm.

PAGE 193

193 Figure 4 14 Flowers of Elleanthus sodiroi ( Neubig 246 ) a hummin gbird pollinated flower that produces nectar rewards. A) Dense capitate inflorescence, scale bar = 1 cm. B) Frontal view of flower showing the entrance point for the pollinator, scale bar = 0.5 cm. C) Oblique view of ventral surface of the column, scale bar = 0.5 cm. Note the median ridge of the column which ower showing the two ball like calli at the base, scale bar = 1 cm. E) Transverse section of one callus, scale bar = 1 mm F) Transverse secti on of callus under polarized light scale Note the birefringent granules within each cell indicating the presence of starch.

PAGE 194

194 Figure 4 15 Flowers of Sobralia ciliata ( Whitten 3529 ), a flower that produces nectar rewards and putatively hummingbird pollinated. A) Frontal view of flower, scale bar = 1 cm. B) Oblique view of ventral surface of the column, scale bar = 0.5 cm. Note the similar structure of the elastic rostellar flap with a similar mechanical interaction of pollination as the core group of Sobralia C D) Li p of flower from side and from the top, respectively, scale bar = 1 cm. Note the nectar secreted on the surface of the thickened pad like callus at the base as indicated by arrow

PAGE 195

195 Figure 4 1 6 Flowers of Sobralia caloglossa ( Whitten 3530 ), a flower that produces no nectar reward and is bee pollinated. A) Whole flower, scale bar = 1 cm. B) Ventral side of column, showing rostellar flap, scale bar = 1 cm. C) Pollinia, scale bar = 2 mm. D) Vaguely scrotiform callus, scale bar = 0.5 cm. E) Cross sec tion of callus, scale bar = 1 mm.

PAGE 196

196 Figure 4 17 Flowers of Sobralia mandonii ( Whitten 3531 ), a flower that produces no nectar rewards and is bee pollinated. A) Whole flower, scale bar = 1 cm. B) Ventral side of column, showing rostellar flap, scale bar = 1 cm. C) Pollinia, scale bar = 1 mm. D) Vaguely scrotiform callus, scale bar = 1 cm. E) Cross section of callus, scale bar = 1 mm. F) Cross section of callus, scale bar = 0.5 mm.

PAGE 197

197 Figure 4 18 Graphs with lines of regression of nectar volum e (in ) and sucrose concentration (in brix) for four species of Sobralia Each data point is a measurement from a single flower per inflorescence per day, on first day of anthesis.

PAGE 198

198 Figure 4 19 Graphs with lines of regression of nectar volume (in ) and sucrose concentration (in brix) for four species of Elleanthus

PAGE 199

199 Figure 4 20 Graphs with lines of regression of nectar volume (in ) and sucrose concentration (in brix) comparing putatively hummingbird pollinated (blue diamonds) and bee pollinated (red squares) nectar observed data in this study.

PAGE 200

200 Figure 4 2 1 Lips of Sobralia showing putative osmophore. A) Sobralia powellii ( Whitten 3257 ); lip is cut longitudinally down center and stained with I 2 KI to show presence of starch, scale bar = 1 cm Note mealy (papillose) and hairy texture. B C ) Sobralia decora ( Whitten 3280 ); B) lip in cross section stained with bromophenol blue (to show general cellular structure and vasculature) and I 2 KI, note that only starch accumulation is shallow within the tissue, scale bar = 1 cm ; C) cross section of callus stained with safranin and hematoxylin showing mealy texture of hairs indicated by arrow, scale bar = 1 mm D) Sobralia recta ( Blanco 3010 ); cross section of callus stained with safranin and hematoxylin showing mealy texture of hairs indicated by arrow, scale bar = 1 mm.

PAGE 201

201 Figure 4 22 Sobralia visitation by euglossines. A) Euglossa sp. at a member of the S. decora complex. B F) Sobralia pollinia on variou s euglossine bee pollinators; n ote that the pos i tioning is always on the scutellum of the thorax, despite size of bee and pollinia. A) Eulaema bombiformis male B) Eulaema nigrita male C) Eulaema sororia male D) Eufriesea surinamensis female E) Euglossa hansoni male Scale bars = 1 cm Photos by Norris H. Williams, W. Mark Whitten and K. Neubig.

PAGE 202

202 Figure 4 23 Elevational differences of among groups within Sobralieae, separated by assessed pollinator. A B) Elleanthus and Epilyna with those species pollinated by an unknown insect vs. those species pollinated by hummingbirds, respectively. C D) Core Sobralia with those species pollinated by bees vs. those pollinated by hummingbirds, respectively. E F) Sertifera and Sobralia sect. Sobralia with those species pollinated by bees v s. those species pollinated by hummingbirds, respectively. Note the difference of elevational preferences among pollination guilds.

PAGE 203

203 Figure 4 24 Phylogenetic tree based on 6 region maximum likelihood analysis (from chapter two) modified to show char acteristics of nectary structure, nec tar presence, pollinia color, flower longevity double cuniculus, and pollination syndrome.

PAGE 204

20 4 Figure 4 25 Phylogenetic tree based on 6 region maximum likelihood analysis (from chapter two) with the character of pol lination syndrome, as outlined in this study, mapped on the tree using likelihood reconstruction as implemented in Mesquite. Note that S. rarae avis was not sampled in the phylogenetic study, so the putative sphingophyly is not is not mapped here. Some t axa are, such as Elleanthus sects. Elleanthus and Chloidelyna have an unknown

PAGE 205

205 Table 4 1. Previously published accounts of floral v isitors and pollinators in Sobralieae; compiled based on the followin g resources ( Ducke, 1902 ; Dodson, 1962 1965 ; van der Pijl and Dodson, 1966 ; Dressler, 1971 1976 ; Braga, 1977 ; Roubik and Ackerman, 1987 ; Roubik, 2000 ; Dziedzioch, Stevens, and Gottsberger, 2003 ; Singer, 2003 ; Fogden and Fogden, 2006 ) Orchid Pollinator/visitor Higher taxon Literature cited Sobralia amabilis Razisea spicata Trochilidae (bird) Fo gden & Fogden (2006) Pantrope insignis Trochilidae (bird) Dodson (1965) Sobralia decora Euglossa viridissi m a Apidae, Euglossini (bee) Van der Pijl & Dodson (1966) Sobralia leucoxantha Eulaema speciosa Apidae, Euglossini (bee) Dodson (1965) Sobralia macrophylla Euglossa cf. ignita Apidae, Euglossini (bee) Braga (1977) Sobralia rosea Bombus morio Apidae, Bombini (bee) Dodson (1965) Eulaema polyzona Apidae, Euglossini (bee) Dodson (1965) Eufriesea ornata Apidae, Euglossini (bee) Dodson (1965) Sobr alia sessilis Euglossa cordata Apidae, Euglossini (bee) Ducke (1902) Sobralia violacea Bombus morio Apidae, Bombini (bee) Dodson (1965) Bombus hortulans var. robusta Apidae, Bombini (bee) Van der Pijl & Dodson (1966) Xylocopa frontalis Apidae, Xylocop ini (bee) Dodson (1962) Xylocopa cf. transitoria Apidae, Xylocopini (bee) Dodson (1965) Eufriesea surinamensis Apidae, Euglossini (bee) Dodson (1962) Eulaema cingulata Apidae, Euglossini (bee) Dodson (1962) Eulaema polychroma Apidae, Euglossini (be e) Dodson (1962) Eulaema speciosa Apidae, Euglossini (bee) Dodson (1965) Epicharis rustica Apidae, Centridini (bee) Dodson (1965) Sobralia aff. weberbaueriana Eulaema polychroma Apidae, Euglossini (bee) Van der Pijl & Dodson (1966) Sobralia sp. (unid entified) Euglossa championi Apidae, Euglossini (bee) Roubik & Ackerman (1987) Euglossa cybelia Apidae, Euglossini (bee) Roubik & Ackerman (1987) Euglossa deceptrix Apidae, Euglossini (bee) Roubik & Ackerman (1987) Euglossa despecta Apidae, Euglossin i (bee) Roubik & Ackerman (1987) Euglossa dissimula Apidae, Euglossini (bee) Roubik & Ackerman (1987) Euglossa dressleri Apidae, Euglossini (bee) Roubik & Ackerman (1987) Euglossa variabilis Apidae, Euglossini (bee) Roubik & Ackerman (1987) Eulaema bombiformis Apidae, Euglossini (bee) Roubik & Ackerman (1987) Eulaema meriana Apidae, Euglossini (bee) Roubik & Ackerman (1987) Eulaema nigrita Apidae, Euglossini (bee) Roubik & Ackerman (1987), Dressler (1976) Exaerete frontalis Apidae, Euglossini (bee) Roubik & Ackerman (1987) Plebeia minima Apidae, Meliponini (bee) Roubik (2000) Elleanthus amethystinoides Doryfera ludovicae Trochilidae (bird) Dziedzioch et al., 2003 Colibris thalassinus Trochilidae (bird) Dziedzioch et al., 2003 Adelomyia melanogenys Trochilidae (bird) Dziedzioch et al., 2003 Coeligena coeligena Trochilidae (bird) Dziedzioch et al., 2003 Heliangelus amethysticollis Trochilidae (bird) Dziedzioch et al., 2003 Ocreatus underwoodii peruanus Trochilidae (bird) Dziedzioch e t al., 2003 El leanthus arpophyllostachys Ocr e atus underwoodii Trochilidae (bird) Dodson (1965) Elleanthus aureus unknown hummingbird Trochilidae (bird) Dodson (1962)

PAGE 206

206 Table 4 1. Continued. Orchid Pollinator/visitor Higher taxon Literature cited Ellean thus aurantiacus unknown hummingbird Trochilidae (bird) Dodson (1962) Elleanthus aurantiacus (as E. hallii ) unknown hummingbird Trochilidae (bird) Dodson (1965) Elleanth u s bifarius Ocreatus underwoodii peruanus Trochilidae (bird) Dziedzioch et al., 2003 Elleanthus capitatus unknown hummingbird Trochilidae (bird) Dressler (1971) unknown hummingbird Trochilidae (bird) Dodson (1962) Elleanthus brasiliensis Phaethornis petrei Trochilidae (bird) Singer (2003) Elleanthus glaucophyllus Colibri thalassinus T rochilidae (bird) Fogden & Fogden (2006) Colibri delphinae Trochilidae (bird) Fogden & Fogden (2006) Selasphorus scintilla Trochilidae (bird) Fogden & Fogden (2006) Elleanthus hymenophorus Lampornis calolaema Trochilidae (bird) Fogden & Fogden (2006) Amaz a lia tzacatl Trochilidae (bird) Dodson (1965) Elleanthus lentii Colibri thalassinus Trochilidae (bird) Fogden & Fogden (2006) Elleanthus maculatus Doryfera ludovicae Trochilidae (bird) Dziedzioch et al., 2003 Colibris thalassinus Trochilidae (bi rd) Dziedzioch et al., 2003 Coeligena coeligena Trochilidae (bird) Dziedzioch et al., 2003 Coeligena torquata Trochilidae (bird) Dziedzioch et al., 2003 Heliangelus amethysticollis Trochilidae (bird) Dziedzioch et al., 2003 Haplophaedia aureliae Tr ochilidae (bird) Dziedzioch et al., 2003 Ocreatus underwoodii peruanus Trochilidae (bird) Dziedzioch et al., 2003 Elleanthus rosea unknown hummingbird Trochilidae (bird) Dodson (1965) Elleanthus sp. (unidentified) Heliangelus amethysticollis Trochilida e (bird) Dziedzioch et al., 2003 Haplophaedia aureliae Trochilidae (bird) Dziedzioch et al., 2003 Ocreatus underwoodii peruanus Trochilidae (bird) Dziedzioch et al., 2003 unknown hummingbird Trochilidae (bird) Dziedzioch et al., 2003

PAGE 207

207 Table 4 2 Observations of nectar secretion in this study. Although some species were observed and confirmed to have nectar, not all had measurable amounts of nectar. Species sample size, flowers (n) mean (avg) standard deviation range Vouchers Syndrome Elleanthus aurantiacus 5 none (population sampling) hummingbird volume ( ) 4.4 1.8 2 to 7 concentration (brix) 22.8 1.9 21 to 26 E. caravata 52 Neubig 202 hummingbird vol ume ( ) 5.7 2.4 2 to 10.1 concentration (brix) 24.3 6.9 12 to 40 E. cynarocephalus 5 Neubig 247 hummingbird volume ( ) 6 3.1 2 to 10 concentration (brix) 9.2 6.6 5 to 21 E. sodiroi 46 Neubig 246 hummingbird volume ( ) 13.6 6.5 4 to 31.5 concentration (brix) 15.9 6.4 7 to 25 Sobralia bouchei 52 Blanco 3009, Neubig 208 bee volume ( ) 14.1 7.9 2 to 43 concentration (brix) 21.2 3.3 12 to 28 S. callosa 27 Blanc o 3021, Neubig 224 hummingbird volume ( ) 6.3 2.4 1.5 to 12 concentration (brix) 16.3 2 12 to 19.5

PAGE 208

208 Table 4 2. Continued. Species sample size, flowers (n) mean (avg) standard deviation range Vouchers Syndrome S. macrophylla 6 Blanco 3022 bee volume ( ) 4.9 2.4 1 to 8 concentration (brix) 20.6 1.4 18 to 22 S. rosea 46 none (population sampling) bee volume ( ) 8.4 8.6 3 to 35 concentration (brix) 13.8 3.1 5 to 19.5

PAGE 209

209 Table 4 3 Floral characters as they relate to pollination syndromes in groups wi thin Sobralieae. Data are mostly from personal observations, but also from literature. diurnal anthesis strong or weak zygomorphy tubular perianth nectar "pocket" double cuniculus nectar nectar guide horizontal or pendant color fragrant gullet resupinate? syndrome Sobralia sect. Sobralia Sobralia ciliata yes strong no no no yes yes horizontal vivid no no resupinate hummingbird? all other species y es strong no no no no yes horizontal vivid or dull unknown yes resupinate bee core Sobralia Sobralia amabilis yes strong no unknown no probable yes horizontal vivid no no resupinate hummingbird Sobralia callosa yes stron g no yes no yes yes horizontal vivid no no resupinate hummingbird? Sobralia crocea yes weak yes no no no yes pendant vivid no no resupinate hummingbird? Sobralia bouchei yes strong no no yes yes yes horizontal vivid or dull mixed yes resupinate bee Sobralia macrophylla yes strong no no yes yes yes horizontal vivid or dull mixed yes resupinate bee Sobralia rosea yes strong no no yes yes yes horizontal vivid or dull mixed yes resupinate bee Sobralia rarae avis no strong yes* no no unknown no horizontal whitish yes no resupinate hawk moth? all other species yes strong mixed no yes no yes horizontal vivid or dull mixed yes resupinate bee Sertifera yes strong yes yes no unknown no horizontal vivid no no resupinate hummingbird? Epilyna y es weak yes yes no unknown no horizontal whitish no no non resupinate unknown Elleanthus Elleanthus lancifolius yes weak yes yes no yes no horizontal whitish no no mixed unknown sect. Chloidelyna yes weak yes yes no yes n o horizontal whitish no no mixed unknown all other species yes weak yes yes no yes mixed mixed vivid no no mixed hummingbird

PAGE 210

210 CHAPTER 5 DESCRIPTIVE VEGETATIVE ANATOMY IN SOBRALIEAE Background This chapter describes the variation in vegetative anato my for exemplar taxa of Sobralieae T he purpose of this study is to provide additional structural characters that help diagnose clades of Sobralieae and to improve our understand ing of anatomical evolution within Sobralieae. Although general vegetative m orphology has been discussed as it relates to taxonomy ( C hapter 2) and floral morphology and anatomy have been discussed as they relate to pollination ( C hapter 4), little is known of the variation in veget ative anatomy across Sobralieae, as only a few anat omical studies have included any representatives of Sobralieae ( Solereder and Meyer, 1930 ; Benzing, Ott, and Friedman, 1982 ; Arevalo, Figueroa, and Madrinan, 2011 ) Characters from vegetative anatomy have been used systematically for over 150 years ( Judd et al., 2008 ) and whereas most structural characters show varying levels of homoplasy, anatomical characters are some of the most phylogenetically informative structural characters in higher order phy logenetics of angiosperms ( Carlquist, 1961 ; Barthlott et al., 1998 ) as well as orchids ( Freudenstein and Rasmussen, 1999 ; Stern and Judd, 1999 ; Stern and Whitten, 1999 ; Stern and Judd, 2000 ; Stern and Carlsward, 2006 2008 ) Vegetative anatomical adaptation s make many orchids uniquely suited for ecological adaptations such as epiphytism. Leaf succulence thickened roots velamen, and pseudobulbs are common features in orchids that help in water conservation and foster evolution of epiphytism. Leaf succulence in combi nation with growth at low elevations is correlated in orchids with crassulacean acid metabolism, a photosynthetic

PAGE 211

211 pathway that conserves water through stomate closure during the day ( Silvera, Santiago, and Winter, 2005 ; Silvera et al., 2010a ) Although frequently epiphytic, Sobralieae never have pseudobulbs or succulent leaves, which is consistent with the C3 photosynthetic pathway found in this group ( Silvera, Santiago, and Winter, 2005 ; Silvera et al., 2010a ) Sobralieae are part of a paraphyletic grade of taxa that includes Epidendroideae (e.g., Neott i eae, Nervilieae, and Triphoreae). D espite the propensity of the epiphyt ic habit within the family, many major lineages are plesiomorphically terrestrial. The transitions between terrestrial and epiphytic habits within Sobralieae have not been studied extensively because of the ambiguity in character state s and because small plant size is more likely to be a predictor of epiphytism than is phylogenetic relationship ( Benzing and Ott, 1981 ) Although some vegetative traits are useful for identifying species or monophyletic groups of species within Sobralieae, there is homoplasy in vegetative anatomical and morpholog ical characters, and thus distantly related taxa are often similar due to reversals or parallelisms in these characters Sobr alieae are largely uniform in their vegetative gestalt because of their large (often >1 m), pseudobulbless stems with plicate, wide (often >5 cm) leaves, and thick roots (often >0.5 cm). Some notable exceptions are species of Epilyna and Elleanthus sect. Chloidelyna which have highly reduce d morphology and conduplicate, narrow (often <0.5 cm) leaves. G eneric delimitation has been based on relatively few gross floral characters, especially size ( C hapters 2 and 4). Because characters derived from vegetati ve anatomy have often been found to be more predictive of phylogenetic relationships than morphology

PAGE 212

212 ( Freudenstein and Rasmussen, 1999 ) I sampled the tribe with an emphasis on major clades as revealed by molecular data Materials and Methods Fresh material for this study was obtained primarily from greenhouse cultivated plants, although some wild collected specimens were also used All vouchers were deposi ted at the University of Florida herbarium ( FLAS, Table 5 1) and in nearly all cases represent the same vouchered specimens used in the molecular phylogenetic study ( C hapter 2) Plant part s available for study were preserved in FAA (9 part s 70% ethanol, 0 .5 part 37% commercial formalin, and 0.5 part glacial acetic acid) for at least 48 hours, and then stored in 70% ethanol. Transverse sections (TS) of leaves stems, and roots were made with a sliding microtome at a thickness of 60 S alum hematoxylin and s afranin ( Carlsward et al., 1997 ) Differentiation and dehyd ration of stained sections were carried out in a graduated ethanol series followed by clearing in limonene (Citrisolv, Fisher Scientific Company). Sections were then mounted on microscope slides with Canada balsam and photographed with a Zeiss Axioscope 4 0 compound microscope attached to a Pixera Pro 150ES digital camera. Cuticular peels to view epidermal surfaces were made using clear nail polish, which was painted on, allowed to dry, then removed using clear tape. To study foliar trichomes, hand sectio ns of fresh leaf tissue were stained with toluidine blue. D escriptions of the anatomical features are organized phylogenetically, based on results in Chapter 2. space sizes relative to the siz e of the surrounding cells of a plant organ. Only consistent anatomical features were used for phylogenetic analyses.

PAGE 213

213 Sixteen vegetative features of anatomy and morphology (Table 5 2) were used to construct a character matrix (Table 5 3) The larger phy logenetic sampling of six DNA loci from c hapter 2 was reduced to include only the accessions sampled for anatomy Because the maximum likelihood (ML) analysis with reduced taxon sampling yield ed a tree that was incongruent with the ML analysis of greater taxon sampling (see chapter 2 regarding incongruent placement of Sobralia ciliata ) a maximum parsimony (MP) tree showing a similar topology to the large ML topology was used to reconstruct character state s C haracters were mapped onto the MP phylogeny us ing a likelihood based analysis of ancestral character state reconstruction as implemented in Mesquite ( Maddison and Maddison, 2005 ) Results The following f eatures of vegetative anatomy in Sobralieae are shown: leaf development (F ig. 5 1), leaf sheath (Fig. 5 2), leaf margins (Fig. 5 3), leaf mesophyll composition (Figs. 5 4 and 5 5), leaf epidermis and stomata (Figs. 5 6 and 5 7), leaf bulliform cells (Fig. 5 8), leaf veins (Fig. 5 9), leaf plastids (Fig. 5 10), raphides (Fig. 5 1 1), lysogenic trichomes (Fig. 5 12), stem shape (Fig. 5 13), stem general anatomy (Fig. 5 14), stem epidermis (Fig. 5 15), stem vascular bundle (Fig. 5 16), root vascular cylinder (Fig. 5 17), root cortex (Fig. 5 18), root velamina (Fig. 5 19), and tilosom es (Fig. 5 20). In addition to those anatomical characters, I used two morphological characters : abscission layer between sheath and the blade (Fig. 5 21). Character state reconstructions of all characters (Tables 5 2 and 5 3) are shown for tilosome shape (Fig. 5 22), presence /absence of sclerified idioblasts in root cortex (Fig. 5 23), pith cell wall variation (Fig. 5 24), stem shape (Fig. 5 25), presence /absence of sheath tubercles (Fig. 5 26), presence /absence of sheath fiber bundles (Fig. 5 27),

PAGE 214

214 presence /absence of sheath air spaces (Fig. 5 28), presence /absence of idioblasts with raphides (Fig. 5 29), mesophyll composition (Fig. 5 30), leaf vernation (Fig. 5 31), leaf margin shape (Fig. 5 32), midrib exsertion (Fig. 5 33), sclerifica tion around bundle caps (Fig. 5 34), presence of subepidermal parenchyma around midrib (Fig. 5 35), presence /absence of abscission layer between leaf and sheath (Fig. 5 36), and presence /absence 37). A natomical descrip tions of the tribe, as well as each of the major lineages within Sobralieae, are presented here. Uniform features throughout the tribe are included in the description of Sobralieae and omitted from each generic description. Sobralieae Leaf sheath Abscis sion layer between sheath and blade present. Abaxial glandular tubercles absent. MESOPHYLL homogeneous. FIBER BUNDLES sometimes present between collateral vascular bundles in the mesophyll Air spaces present in larger leaves or absent in smaller leave s. Leaf blade STOMATA abaxial and tetracytic. Outer ledges thin, inner ledges thick and short. Substomatal chambers small to large, irregularly shaped, especially when present with palisade. EPIDERMAL CELLS rectangular, periclinally oriented to isodia metric in TS. Bulliform cells frequent on both abaxial and adaxial surfaces, rarely restricted to the adaxial surface above the midrib. FIBER BUNDLES absent. MESOPHYLL homogeneous. VASCULAR BUNDLES collateral, in one row. Sclerenchyma associated with both xylem and phloem poles. Bundle sheath distinct. Stem EPIDERMIS papillose, covered by a thick cuticle. CORTEX 2 10 cells thick. CAULINE SCLERENCHYMATOUS BAND composed of thick walled cells

PAGE 215

215 sometimes extending into ground tissue between vascular bundles. GROUND TISSUE of thick walled and /or thin walled parenchyma. Root VELAMEN 2 6 cells thick. TILOSOMES present, scrotiform. EXODERMAL CELLS thickened, walls tapering in thickness near the inner tangential surface CORTEX homogeneous. ENDODE RMAL CELLS o thickened PERICYCLE cells thin walled opposite xylem and thick walled opposite phloem. VASCULAR CYLINDER 6 to 40 arch. PITH heterogeneous or homogeneous, sclerenchymatous and/or parenchymatous towards center. Sobralia Sensu Stricto (C ore Sobralia ) Leaf sheath FIBER BUNDLES present in S. cri s pissima S. rosea S. theobromina and S. warszewiczii Air spaces present. Leaf blade MESOPHYLL heterogeneous in S. bouchei S. callosa S. cri s pissima and S. lindleyana Stem CORTEX 2 4 cells thick. GROUND TISSUE of thick walled or thin walled parenchyma. Root VELAMEN 2 4 (typically 3) cells thick. CORTEX homogeneous; heterogeneous with sclerified idioblasts scattered in S. gloriana VASCULAR CYLINDER 14 to 24 arch. PITH heterogeneous, sclerenchymatous outside grading into a distinct parenchymatous center. Sobralia ciliata Leaf sheath FIBER BUNDLES present, occurring in groups of two or three between vascular bundles. Leaf blade MESOPHYLL heterogeneous.

PAGE 216

216 Stem CORTEX 3 cells thick. CAULINE SCLERENCHYMATOUS BAND not extending into ground tissue. GROUND TISSUE of thin walled parenchyma. Root VELAMEN 4 6 cells thick. CORTEX heterogeneous, with sclerified idioblasts scattered. VASCULAR CYLINDER approximately 40 arch. PITH homogen eous and parenchymatous. Sobralia dichotoma C lade ( S. caloglossa + S. dichotoma + S. mandonii ) Leaf sheath FIBER BUNDLES present, alternating with vascular bundles. Leaf blade. EPIDERMAL CELLS with bulliform cells frequent and on both abaxial and adaxia l surfaces. MESOPHYLL heterogeneous ; homogeneous in S. dichotoma Stem CORTEX 3 4 cells thick. CAULINE SCLERENCHYMATOUS BAND not extending into ground tissue. GROUND TISSUE of thin walled parenchyma. Root VELAMEN 3 5 cells thick. VASCULAR CYLINDER 20 to 30 arch. PITH heterogeneous ; sclerenchymatous on outside, gradually becoming more parenchymatous towards center. Sobralia portillae Leaf sheath N ot sampled. Stem Not sampled. Root VELAMEN 3 cells thick. TILOSOMES scrotiform, short. VASCULAR CYLINDER 8 arch. PITH homogeneous and sclerenchymatous. Sertifera Leaf sheath Abaxial glandular tubercles present. FIBER BUNDLES present, alternating with vascular bundles. Leaf blade MESOPHYLL heterogeneous.

PAGE 217

217 Stem CORTEX 3 cells thick. CAULINE SC LERENCHYMATOUS BAND not extending into ground tissue. GROUND TISSUE of thin walled parenchyma. Root Not sampled. Epilyna Leaf sheath A bscission layer absent. FIBER BUNDLES present, alternating with vascular bundles. Leaf blade EPIDERMAL CELLS bulli form only above the midrib. Stem CORTEX 2 10 cells thick. CAULINE SCLERENCHYMATOUS BAND not extending into ground tissue. GROUND TISSUE of thick walled parenchyma. Root VELAMEN 2 cells thick. TILOSOMES flattened. CORTEX heterogeneous, with scattere d sclerified idioblasts. VASCULAR CYLINDER 6 arch. PITH homogeneous and sclerenchymatous. Elleanthus Leaf sheath FIBER BUNDLES minute, alternating between vascular bundles. Leaf blade EPIDERM IS with bulliform cells frequent on both abaxial and adaxial surfaces, except in E. stolonifer where those cells are only above the midrib. MESOPHYLL homogeneous ; heterogeneous in E. ensatus and E. stolonifer Stem CORTEX 2 4 cells thick. Root VELAMEN 2 3 (rarely 4) cells thick. TILOSOMES scrotiform ; flattene d in E sp. ( Neubig 203 and Whitten 3538 ) and E. caravata CORTEX usually heterogeneous, with sclerified idioblasts scattered ; homogeneous in E. gracilis E. ensatus and E. longibracteatus VASCULAR CYLINDER 8 to 30 arch.

PAGE 218

218 Discussion Character E volutio n Some anatomical characters that have been of utility in previous orchid anatomical phylogenetic analyses ( Stern and Judd, 1999 ; Carlsward, 2004 ) were examined in this study but were either invariable or the variation was incompatible with phylogenetic analysis o f Sobralieae Among those characters, root endodermal cell wall thickenings were always o thickened. Similarly, the root exodermal cell wall thickenings were thickened and tapering [not as distinctly thickened as in Vandeae ( Carlsward, 2004 ) ]. Root pith cell s were homogeneous (i.e., with no distinct banding) or heterogeneous, but the distinction between these two conditions was difficult to code in some taxa. Leaf s heath fiber bundle orientatio n (whether closer to the adaxial or abaxial surfaces) was difficult to code because of the paucity of fibers within each bundle Indeed, fiber bundles in the sheaths were generally located closer to the abaxial surface. B ulliform cells in the foliar epid ermis were always present in the taxa sampled but their size and number differ ed drastically and showed a continuum of variation with respect to distribution Likewise, leaf epidermal cell shape differ ed in size, but not discretely, and was more or less rectangular shape throughout Silica bodies ( Carlsward, 2004 ; Prychid, Rudall, and Gregory, 2004 ) were not seen in tribe Sobralieae, but were present and conical in the outgroup, Bletilla Overall, very few anatomical characters are synapomorphic and therefore useful for generic delimitation. In particular, the disparate clades that make up polyphyletic sect. Sobralia are not strongly distinguished by unique anatomical characters although they vary in combination of some characters. Further sampling of species and more individuals from each species, epecially the

PAGE 219

219 phylogenetically interesting S. dorbignyana and S. stenophylla is needed before definitive statements of character e volution can be asserted. Although several anatomical characters were phylogenetically uninformative several showed distinct and phylogenetically informative variation within Sobralieae. All Sobralieae have leaves that appear to develop in a duplicate fa shion, with the leaf margins in direct contact during development ( Dressler, 1993a ) However, Sobralieae vary in the folding pattern of the blade Most Sobralieae are plicate (folded upon multiple planes) while some taxa are conduplicate (folded only once at the midrib). Conduplication has evolved independently in Epilyna and Elleanthus sect. Chloidelyna and is a synapomorphy for each of those clades (Fig. 5 31). In both cases, this character state is a ssociated with extreme vegetative reduction. Leaf sheath tubercles are an aut apomorphy for Sertifera (Figs. 5 2 & 5 26). These tubercles are glandular (i.e., raised portions of the epidermis that are topped by a gland ular hair ). I n addition to Sertifera Sobralia rigidissima also has tubercles on the leaf sheath (pers. obs.) Additional anatomical material would be needed to determine the homology of this character across taxa of Sobralieae M any Sobralieae also have furcate secretory hairs on the abax ial blade and sheath surface, which likely contain mucilage that lubricates leaves during elongation [ ( Pridgeon, 1981 ) Fig. 5 12 ] Frequently these furcate hairs are ephemeral, lys ogenic, and no longer present at m aturity ( except in Sobralia atropubescens S. callosa and Elleanthus caravata ). The basal cell, which appears glandular because of its rounded outline, is usually the only part of the hair left at maturity (stained light blue in Fig. 5 12 A & B). Althou gh these

PAGE 220

220 remnant basal cells are common their presence is homoplasious and phylogenetically scattered throughout Sobralieae Leaf sheath fiber bundles were commonly present between the vascular bundles in the mesophyll Their presence was only coded when consistently found between most of the vascular bundles within the sheath. Th e presence of fiber bundles was homoplasious (Fig. 5 27) and seem ed to be more prevalent in species with medium to large plants. The multiple evolutionary occurrences of plant size reduction in Sobralia and Elleanthus likely explains the homoplasious nature of fiber bundle presence Leaf sheath air spaces were common throughout the Sobralieae and the outgroup Bletilla Their loss is unique to Epilyna and Elleanthus sect. Chloi delyna and is likely an evolutionary consequence of extreme vegetative reduction (Fig. 5 28). Leaf idioblasts with raphid e s were common in Sobralieae and were absent in two species of distantly related Sobralia ( i.e., S. kruskayae and S. portillae Fig 5 29 ). The lack of raphid e s is probably an artifact of inadequate sampling, as these structures are common in orchids. Heterogeneous foliar mesophyll refers to the differentiation of palisade and spongy mesophyll while a homogeneous mesophyll (the common c ondition in many monocots) is composed of uniform isodiametric cell s The effect of light on leaf anatomy has been well studied ( Yano and Terashima, 2001 ) Frequently, high light levels induce production of a thicker mesophyll with a thicker palisade ( Lichtenthaler et al., 1981 ; Eschrich, Burchardt, and Essiamah, 1989 ; Yano and Terashima, 2001 ) Although high light intensity induces more layers of palisade, the differentiation of palisade and spongy parenchyma does not seem to be affected by light intensity. Most of the plants used in

PAGE 221

221 this study were grown under uniform greenhouse conditions, so the anatomical variation can not be correlated with different light levels Mesophyll differentiation is homoplasious (Fig. 5 30), so the basis for variation in anatomy does not seem to be under tight evolutionary constraint. Leaf margin shape refers to the appearance of the margin in TS, with respect to the length of the abaxial and adaxial surfaces This trait is also homoplasious (Fig. 5 32) and may be strongly influenced by leaf folding during development because of compressio n within the sheath of the subtending leaf and the number of plications. Sclerified cells extending from the bundle caps may completely encircle the midrib, or the bundle caps may be distinct and only at the xylem and phloem poles This character is homop lasious (Fig. 5 34) and t hese sclerified bundles do not seem to be associated with any other anatomical trait. The propensity of equivocal states within the dee per nodes of the reconstruction illustrate s the extreme homoplasy in this character. Leaf meso phyll parenchyma between the midrib vascular bundle and epidermis can be found at either the abaxial or adaxial surface, but is most common on the abaxial surface in Sobralieae. The exsertion of the midrib from the abaxial surface is an anatomical synapom orphy for Epilyna (Fig. 5 33). morphological synapomorph y of Epilyna (Figs. 5 21 & 5 37). Although the ligule is unique within the Sobralieae, similar structures at the same position within the leaf occur in Pityphyllum a member of the distantly related Cymbidieae ( Blanco et al., 2007 ) However, this structure is clearly not homologous to the ligule found in other monocot groups, such as grasses or gingers Also synapomorphic for the species of Epilyna is the absence of an abscission layer between

PAGE 222

222 the blade and leaf sheath (Figs. 5 21 & 5 36). This is a unique feature of Epilyna within Sobralieae, but it can be found in other groups of orchids, including species of Dichaea ( Neubig et al., 200 9a ) The function of this abscission layer loss is not known for any group. Stem shape in TS is elliptic and is largely uniform within the tribe. Epilyna has a synapomorphy of flattened stem s in TS (Fig. 5 25). This flattening is caused by a prolifera tion of parenchymatous cells between the epidermis and the sclerified ring surrounding the vascular bundles (Fig. 5 13E). The epidermis may be smooth or crenate in TS but this variation is not uniform and is difficult to diagnose (Fig. 5 13). Sclerified idioblasts can be found in the root cortex, surrounded by large areas of parenchyma (Fig. 5 18). The presence of sclerified idioblasts is homoplasious but such cells are common in Elleanthus Epilyna Sobralia gloriana and S. ciliata (Fig. 5 23). Root p ith cell walls can vary distinctly in thickness (Fig. 5 17). Thickened cell wall s (at least in the center most portion of the pith) were relatively rare in Sobralieae (restricted to some species of Elleanthus Epilyna and Sobralia sect. Sobralia ) but the ir presence is homoplasious (Fig. 5 24). The v elamen is a multilayered epidermis of many orchid root s ( Engard, 1944 ; Porembski and Barthlott, 1988 ) Tilosomes, which are also common in orchids, are distinctively branched or coralloid masses of cellulosic/ lignic material of the innermost velamen layer occurring over the passage cells of the exodermis They are hypothesized to increase surface area for transport of materials across the passage cells. In Sobralia e of their webbed texture ( Pridgeon, Stern, and Benzing, 1983 ) The variation I have observed within

PAGE 223

223 Sobralieae indicates that two shapes of tilosomes exist: large globose bodies (scrotiform) and flattened bodies. The presence of tilosomes is a synapomorphy of Sobralieae, and m ost members have scrotiform shaped tilosomes. Flattened tilosomes are less common and might be a synapomorphy for Epilyna and Elleanthus Within Elleanthus ( sect. Stachydelyna with pendant inflorescence s), r eversals to the scrotiform type (such as in members of Sobralia ) have occurred (Fig. 5 22) Anatomy and Morphology as They R elate to E piphytism Epiphytism has long been believed to be a driving force in the evolution of vegetative morphology and anatomy of orchids ( Benzing, 1990 ) Atwood (1986 ) stated that all species of Sobralia and Elleanthus are epiphytes, but that Sertifera is terrestrial. However, this is an oversimplification. Many species of Sobra lia and Elleanthus are obligate epiphytes. Therefore, a great deal of intermediacy exists between the states of obligate terrestrials and obligate epiphytes. The plasticity of epiphytism and the loose constraints by which many orchids can move between ter restrial and epiphytic habits mean that discussing these as discre t e character state s in a phylogenetic context is an overly simplistic approach to understanding the evolution of epiphytism However given that some Sobralieae are strongly terrestrial or strongly epiphytic, some qualitative differences can be discussed in order to ascertain consistent features at the ends of this spectrum of variation. Epiphytic species are generally reduced in size, which is a common trend in orchids ( Benzing and Ott, 1981 ) For example, members of Elleanthus sect. Chloidelyna Epilyna Sobralia callosa and S. amabilis are all small in stature (usually <30cm). The vegetative reduction is even more extreme in Elleant hus sect. Chloidelyna and in Epilyna with narrow leaves (i.e., usually less than 5 mm). On the other hand, typical

PAGE 224

224 terrestrial species tend to be large r plants These larger plants can rarely become epiphytic and when they are epiphytic, they rarely att ain the same stature as their terrestrial conspecifics. Orchids that are strongly adapted to the epiphytic habit usually show succulen ce in the stem ( e.g., pseudobulb s ) and/or leaf ( Arevalo, Figueroa, and Madrinan, 2011 ) which is associated with CAM photosynthesis T he metabolic requirements of being a successful epiphyte in the hottest and driest communities often also require the use of CAM to conserve water ( Silvera et al., 2010b ) Although there are notable exceptions to these structural adaptations t hey do represent consistent t rends in most Epidendroideae Because n o member of Sobralieae has pseudobulbs or leaf succulence ( only Epilyna comes close to having succulent lea ves), it is not surprising that C3 is the only photosynthetic pathway reported f or Sobralieae ( Silvera et al., 20 10a ) Their lack of succulence and CAM photosynthesis is consistent with the fact that Sobralieae are successful epiphytes mostly in wet montane forests. Pridgeon (1981 ) described foliar hairs in p leu rothallid orchids that are similar in morphology to those of Sobralieae. These hairs are multicellular and rupture at maturity leaving a brown residue. This lyso gen ic activity is consistent with observations in Sobralieae and is not likely to represent water absorption such as is found in the multicellular scale like trichomes of Tillandsia ( Pridgeon, 1981 ) The mucilage secretion activity of these lyso gen ic hairs is evidenced in all taxa as brown mucilage and is p articularly exaggerated in Elleanthus sect. Cephalelyna where mucilage is copious within the inflorescence However, this mucilage secretion does not appear to be related to epiphytism and these lysogenic cells do not seem to be water absorbing.

PAGE 225

225 The v ela men is also frequently well developed and multiseriate in epiphytes. Given that the velamen is prevalent in all orchids, but epiphytism is mostly present in That exaptat ion is probably why epiphytism has evolved so many times in orchids ( Atwood, 1986 ) Studies based on water conductivity indicate that the velamen may play a minimal physiological role in water relations, while cortical thickness might influence water conduc tion more effectively ( Rieger and Litvin, 1999 ) Comparatively, epiphytes can often have thinner roots compared to larger terrestrial s ( Moreira and Isaias, 2008 ) presumably to decrease water loss by decreasing the surface area to volume ratio That finding is consistent in Sobralieae, where the largest roots are found in the terrestrial members o f sect. Sobralia The thickest velamina of Sobralieae are also found in sect. Sobralia so a thicker velamen is not an identifiable aspect of epiphytism. Clearly, there is a complex suite of traits that function in combin ation to yield epiphytism in orchids. Many of those epiphytic traits, like CAM photosynthesis, leaf succulence and pseudobulb presence, are absent in the prima rily terrestrial Sobralieae. Through reduction of the overall body size (i.e., shorter plants and narrower leaves), some species have transitioned to a strictly epiphytic habit. This reduction in plant size likely also reduces the hydrological severity of periodic drought, a perennial problem in epiphytes.

PAGE 226

226 Figure 5 1 TS of leaf sheath and d eveloping leaf blade of two adjacent nodes ( Sobralia lacerate, Neubig 209 ) Note that the developing leaf blade is plicate (folded) and that the sheath supports the growth of this young leaf. Scale bars = 100

PAGE 227

227 Figure 5 2 TS of leaf sheaths. A) Epilyna jimenezii Blanco 2997 B) Sobralia callosa Blanco 3021 C) Elleanthus ensatus Whitten 3555 D) Sertifera sp., Whitten 2937 (arrow indicates glandular tubercle) Scale bars = 100

PAGE 228

228 Figure 5 3 TS of leaf blades at the margins. A) Epilyna hirtzii Whitten 2938 B) Sobralia mucronata Neubig 210 C) Sobralia portillae Whitten 2433 D) Elleanthus aff. gracilis Whitten 3552 E) Sobralia bouchei Blanco 3009 F) Sobralia recta Blanco 3010 G) Sobralia fenzliana Neubig 212 H) Elleanthus ensatus Whitten 3555

PAGE 229

229 Figure 5 4 TS of leaf blade demonstrating isodiametric cells within the mesophyll. A) Sobralia lancea Whitten 2869 B) Sobralia rosea Whitt en 531 C) Sobralia recta Blanco 3010 D) Sobralia portillae Whitten 2433 E) Sobralia fenzliana Neubig 212

PAGE 230

230 Figure 5 5 TS of leaf blade demonstrating palisade cells within the mesophyll. A) Elleanthus aff. gracilis Whitten 3552 B) Elleanthus ensatus Whitten 3555 C) Sertifera colombiana Whitten 2937 D) Sobralia callosa Blanco 3021

PAGE 231

231 Figure 5 6 Cuticular peels of abaxial epidermis. A) Elleanthus sodiroi Neubig 246 B) Sobralia macrophylla Bla nco 3022 C) Sobralia bouchei Blanco 3009 D) Sobralia recta Neubig 207 E) Elleanthus longibracteatus Whitten 99205 F) Sobralia roezlii no voucher.

PAGE 232

232 Figure 5 7 TS of leaf mesophyll and abaxial leaf surface, with focus on sto mata. Note the open spaces of the spongy mesophyll proximal to the stomata, allowing for gas exchange. A) Sobralia fenzliana Neubig 212 B) Sobralia rosea Whitten 531 C) Sobralia bouchei Blanco 3009 (SC = substomatal cavity; arrow indicates a singl e guard cell of a stomate)

PAGE 233

233 Figure 5 8 TS of leaves showing bulliform cells of the epidermis. A) Sobralia dichotoma Whitten 3532 B) Sobralia lancea Whitten 2869. C) Elleanthus caravata Neubig 202 Arrows indicate clusters of bu lliform cells.

PAGE 234

234 Figure 5 9 TS of leaves showing details of primary veins. A) Sobralia bouchei Blanco 3009. B) Sobralia mandonii Whitten 3530 C) Abaxially exserted midrib in Epilyna jimenezii Blanco 2997 D) Sobralia callosa Blanco 3021 E) Sobral ia warscewiczii Blanco 2677

PAGE 235

235 Figure 5 10 TS of leaf mesophyll showing the dense presence of plastids, Epilyna hirtzii Whitten 2938 Arrow indicates a single plastid among many.

PAGE 236

236 Figure 5 11 Raphide crystals. A E) TS of leaves sh owing the presence of raphides and/or the characteristic circular idioblast with raphides. A) Sobralia macrophylla Blanco 3022 B) Sobralia rosea Whitten 531 C) Epilyna hirtzii Whitten 2938 D) Sobralia macrophylla Blanco 3022 (IR = idioblast that contained raphides that washed away during slide preparation) E) Elleanthus caravata Neubig 202 F) TS of root showing raphide crystal, Sobralia klotzscheana Blanco 3011 Arrows indicate raphide crystals.

PAGE 237

237 Figure 5 12 Lys ogenic irregularly furc ated trichomes in Sobralieae. A, B) Sobralia warscewiczii leaf abaxial surface, Blanco 2676 C) Sobralia callosa Blanco 3021 leaf abaxial surface. D) Elleanthus caravata Neubig 202 abaxial sheath surface

PAGE 238

238 Figure 5 13 TS of whole mature stems. A) Elleanthus longibracteatus Whitten 99205 B) Elleanthus caravata Neubig 202 C) Sobralia mucronata Neubig 210 D) Sobralia callosa Blanco 3021 E) Epilyna jimenezii Blanco 2997 Scale bars

PAGE 239

239 Figure 5 14 TS of partial mature stems A) Elleanthus ensatus Whitten 3555 B) Elleanthus capitatellus Neubig 201

PAGE 240

240 Figure 5 15 TS of mature stem at the epidermis. A) Sobralia ciliata Whitten 3529 B) Sobralia mandonii Whitten 3530

PAGE 241

241 Figure 5 16 TS of stem showing detail of individual vascular bundles. A) Sobralia ciliata Whitten 3529 B) Sobralia rosea Whitten 531

PAGE 242

242

PAGE 243

243 Figure 5 18 TS of roots showing variation in cortex. Note that A C have homogeneous cortices and that D E have sclerified idioblasts intermixed within parenchyma. A) Sobralia callosa Blanco 3021 B) Sobralia recta Blanco 3010 C) Sobralia mandonii Whitten 3547 D) Elleanthus sp., Whitten 3538 E) Elleanthus capitatellus Neubig 201 Arrows indicate sclerified cells within cortex.

PAGE 244

244 Figure 5 19 TS of roots showing details of velamen. A) Epilyna hirtzii Whitten 2938 B) Sobralia portillae Whitten 2433 C) Sobralia ciliata Whitten 3529 D) Sobralia fenzliana Neubig 212 E) Elleanthus ensatus Whitten 3555. Scale

PAGE 245

245 Figure 5 20 TS of roots showing the details of th e tilosomes. A) Sobralia bouchei Blanco 3009 B) Elleanthus aff. gracilis Whitten 3552. C) Elleanthus caravata Neubig 202 Arrows indicate tilosomes.

PAGE 246

246 Figure 5 21 Pressed specimen of Epilyna hirtzii ( Whitten 2938 ) showing s ome synapomorphies for the genus, including apiculate leaf apices, a lack of abscission layer in the leaves between the sheath and t he blade, and a at the apex of the sheath, opposite the blade.

PAGE 247

247 Figure 5 22 Ancestral character state reconstru ction of root tilosome anatomy (character 1).

PAGE 248

248 Figure 5 23 Ancestral character state reconstruction of root sclerified idioblasts (character 2).

PAGE 249

249 Figure 5 24 Ancestral character state reconstruction of root pith cell wall thickness (character 3).

PAGE 250

250 Figure 5 25 Ancestral character state reconstruction of stem shape (character 4).

PAGE 251

251 Figure 5 26 Ancestral character state reconstruction of leaf sheath tubercles (character 5).

PAGE 252

252 Figure 5 27 Ancestral character state reconstruction of leaf sheat h fiber bundle presence (character 6).

PAGE 253

253 Figure 5 28 Ancestral character state reconstruction of leaf sheath air space presence (character 7).

PAGE 254

254 Figure 5 29 Ancestral character state reconstruction of leaf idioblasts with raphid e s (character 8).

PAGE 255

255 F igure 5 30 Ancestral character state reconstruction of mesophyll composition (character 9).

PAGE 256

256 Figure 5 31 Ancestral character state reconstruction of leaf vernation (character 10).

PAGE 257

257 Figure 5 32 Ancestral character state reconstruction of leaf mar gin shape (character 11).

PAGE 258

258 Figure 5 33 Ancestral character state reconstruction of midrib exsertion (character 12).

PAGE 259

259 Figure 5 34 Ancestral character state reconstruction of leaf bundle caps on midribs (character 13).

PAGE 260

260 Figure 5 35 Ancestral cha racter state reconstruction of leaf subepidermal parenchyma (character 14).

PAGE 261

261 Figure 5 36 Ancestral character state reconstruction of leaf abscission layer presence between sheath and blade (character 15).

PAGE 262

262 Figure 5 37 Ancestral character state reco (character 16).

PAGE 263

263 Table 5 1. List of species used in this study, voucher numbers, and parts sectioned. All specimens were deposited in the University of Florida herbarium (FLAS) Asterisks indicate individuals that wer e not sequenced in this study; similar taxa were chosen as phylogenetic placeholders. L = leaf; Sh = leaf sheath; St = stem; R = root. Voucher Taxon Parts collected Neubig 1 2006 Bletilla striata (Thunb.) Rchb. f. L, Sh, St, R Whitten 3552 Elleanthus a ff. gracilis (Rchb. f.) Rchb. f. L, Sh, St, R Neubig 201 Elleanthus capitatellus Dressler L, Sh, St, R Neubig 202 Elleanthus caravata (Aubl.) Rchb. f. L, Sh, St, R Whitten 3555 Elleanthus ensatus (Lindl.) Rchb. f. L, Sh, St, R Whitten 99205 Elleanthus longibracteatus (Lindl. ex Griesb.) Fawc. L, Sh, St, R Whitten 3546 Elleanthus robustus (Rchb. f.) Rchb. f. L, Sh, St Neubig 203 Elleanthus sp. L, Sh, St, R Whitten 3538 Elleanthus sp. L, Sh, St, R Blanco 2934 Elleanthus stolonifer Barringer L, Sh, St Whitten 2938 Epilyna hirtzii Dodson L, Sh, St, R Blanco 2997 Epilyna jimenezii Schltr. L, Sh, St, R Whitten 2937 Sertifera sp. L, Sh, St Blanco 3009 Sobralia bouchei Ames & C. Schweinf. L, Sh, St, R Blanco 3021, Whitten 3275 Sobralia callosa L.O. Wil liams L, Sh, St, R Whitten 3531 Sobralia caloglossa Schltr. L, Sh, St, R Whitten 3529 Sobralia ciliata (C. Presl) C. Schweinf. & Foldats L, Sh, St, R Neubig 5 2006 Sobralia crispissima Dressler L, Sh, St Whitten 3532 Sobralia dichotoma Ruiz & Pav. L, R Neubig 212 Sobralia fenzliana Rchb. f. L, Sh, St, R Blanco 2678, Whitten 3307 Sobralia gloriana Dressler L, Sh, St, R Blanco 3011 Sobralia klotzscheana Rchb. f. L, St, R

PAGE 264

264 Table 5 1. Continued. Voucher Taxon Parts collected Blanco 3020 Sobralia krus kayae Dressler L, St Whitten 2941* Sobralia kruskayae Dressler L, Sh, St, R Neubig 209 Sobralia lacerata Dressler & Pupulin L, Sh, St Whitten 2869 Sobralia lancea Garay L, Sh, St Blanco 2681 Sobralia leucoxantha Rchb. f. L, R Whitten 3547, Whitten 932 11* Sobralia lindleyana Rchb. f. L, Sh, St, R Blanco 3022, Whitten 3266 Sobralia macrophylla Rchb. f. L, Sh, St, R Whitten 3530 Sobralia mandonii Rchb. f. L, Sh, St, R Neubig 210 Sobralia mucronata Ames & C. Schweinf. L, St Whitten 2433 Sobralia portil lae Christenson L, R Blanco 3010 Sobralia recta Dressler L, St, R Whitten 0531 Sobralia rosea Poepp. & Endl. L, Sh, St Blanco 2679 Sobralia theobromina Dressler L, Sh, St, R Blanco 2677, Whitten 2831 Sobralia warszewiczii Rchb. f. L, Sh, St, R

PAGE 265

265 Ta ble 5 2 List of vegetative anatomical characters and character state s used in phylogenetic reconstruction. Character state s are ordered so that the Character number Anatomical or morphological character/ character state 1 RO OT, TILOSOME: absent (0), flattened hat shape (1), scrotiform (2) 2 ROOT, SCLERIFIED IDIOBLASTS IN CORTEX: absent (0), present (1) 3 ROOT, PITH CELL WALLS: thin (0), thick (1) 4 STEM, SHAPE: rounded (0), dorsiventrally flattened and winged (1) 5 SHEATH TUBERCLES: absent (0), present (1) 6 SHEATH, FIBER BUNDLES: absent (0), present (1) 7 SHEATH, AIR SPACES: present (0), absent (1) 8 LEAF, IDIOBLASTS WITH RAPHIDES: absent (0), present (1) 9 LEAF, MESOPHYLL: homogeneous (0), heterogeneous (1) 10 LEAF VERNATION: plicate (0), conduplicate (1) 11 LEAF, MARGIN SHAPE: abaxially longer (0), isodiametric (1), adaxially longer (2) 12 LEAF, MIDRIB: within mesophyll (0), abaxially exserted (1) 13 LEAF, BUNDLE CAPS OF MIDRIB: extensive thick walled cells ass ociated with midrib (0), thick walled cells restricted to phloem and xylem poles (1) 14 LEAF, SUBEPIDERMAL PARENCHYMA BETWEEN MIDRIB VASCULAR BUNDLE AND EPIDERMIS: present (0), absent (1) 15 LEAF, ABSCISSION LAYER BETWEEN LEAF AND SHEATH: present (0), ab sent (1) 16

PAGE 266

266 Table 5 3 Character state s for taxa used in ancestral character reconstruction analyses of Sobralieae. Taxon 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Bletilla striata 0 0 0 0 0 0 0 0 0 0 2 0 1 0 0 0 Elleanthus aff. gracilis 2 0 0 0 0 1 0 1 1 0 0 0 0 0 0 0 Elleanthus capitatellus 1 1 1 0 0 0 0 1 0 0 2 0 1 1 0 0 Elleanthus caravata 1 1 0 0 0 0 0 1 0 0 1 0 1 1 0 0 Elleanthus ensatus 2 0 0 0 0 1 0 1 1 0 2 0 1 0 0 0 Elleanthus longi bracteatus 2 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 Elleanthus robustus ? ? ? 0 0 1 0 1 0 0 ? 0 0 0 0 0 Elleanthus sp. (Neubig 203) 1 1 1 0 0 1 0 1 0 0 0 0 1 0 0 0 Elleanthus sp. (Whitten 3538) 1 1 0 0 0 1 0 1 0 0 ? 0 1 0 0 0 Elleanthus stolonifer ? ? ? 0 0 1 1 1 1 1 1 0 0 0 0 0 Epilyna hirtzii 1 0 1 1 0 1 1 1 0 1 0 1 0 0 1 1 Epilyna jimenezii 1 1 0 1 0 0 1 1 0 1 0 1 0 0 1 1 Sertifera sp. ? ? ? 0 1 0 0 1 1 0 0 0 1 0 0 0 Sobrali a bouchei 2 0 0 0 0 0 0 1 1 0 ? 0 1 0 0 0 Sobralia callosa 2 0 0 0 0 0 0 1 1 0 2 0 1 0 0 0 Sobralia caloglossa 2 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 Sobralia ciliata 2 1 0 0 0 1 0 1 1 0 2 0 0 1 0 0 Sobralia crispissima ? ? ? 0 0 1 0 1 1 0 0 0 0 0 0 0 Sobralia dichotoma 2 0 0 ? ? ? ? 1 0 0 0 0 0 0 0 0 Sobralia fenzliana 2 0 0 0 0 0 0 1 0 0 2 0 0 0 0 0 Sobralia gloriana 2 1 0 0 0 0 0 1 0 0 2 0 0 0 0 0 Sobralia klotzscheana 2 0 0 0 ? ? ? 1 0 0 2 0 0 1 0 0 Sobralia kruskayae (Blanco 3020) ? ? ? 0 ? ? ? 0 0 0 ? 0 1 0 0 0 Sobralia kruskayae (Whitten 2941) 2 0 0 0 0 0 0 1 0 0 ? 0 1 0 0 0 S obralia lacerata ? ? ? 0 0 0 0 1 0 0 ? 0 1 0 0 0 Sobralia lancea ? ? ? 0 0 0 0 1 0 0 2 0 0 0 0 0 Sobralia leucoxantha 2 0 0 ? ? ? ? 1 0 0 2 0 1 0 0 0 Sobralia lindleyana 2 0 0 0 0 0 0 1 1 0 0 0 0 1 0 0 Sobralia macrophylla 2 0 0 0 0 0 0 1 0 0 ? 0 0 0 0 0 Sobralia mandonii 2 0 0 0 0 1 0 1 1 0 0 0 0 0 0 0

PAGE 267

267 Table 5 3. Continued Taxon 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Sobralia mucronata ? ? ? 0 ? ? ? 1 0 0 ? 0 0 1 0 0 Sobralia portillae 2 0 1 ? ? ? ? 0 0 0 2 0 1 0 0 0 Sobralia recta 2 0 0 0 ? ? ? 1 0 0 0 0 1 0 0 0 Sobralia rosea ? ? ? 0 0 1 0 1 0 0 1 0 1 0 0 0 Sobralia theobromina 2 0 0 0 0 1 0 1 0 0 2 0 1 0 0 0 Sobralia warszewiczii 2 0 0 0 0 1 0 1 0 0 2 0 1 0 0 0

PAGE 268

268 CHAPTER 6 CONCLUSIONS Discussion The objectives of my dissertation were to understand pollination syndromes and vegetative features of Sobralieae within a phylogenetic framework. I inferred phylogenetic relationships using DNA data from nrITS and various plastid loci. I then documented the floral anatomy, nectar volume and con centration, and the anatomical structures associated with nectar secretion. These data were combined with published accounts of pollinators to hypothesize the evolution of pollination syndromes. These data support the conclusion of multiple origins of hu mmingbird pol lination within the tribe. Bee pollination is likely the plesiomorphic condition for this group, but the pollinator(s) of the small flowered species of Elleanthus and Epilyna is still a mystery. These anatomical data reveal the presence of f alse nectaries and deceit pollination within many species of Sobralia The species of this tribe that do secrete nectar do so by rapidly converting starch stored in the expanded callus within the lip into sugars. I also examined evolutionary patterns in v egetative anatomy and morphology across the DNA based phylogeny of Sobralieae. Most vegetative anatomical characters were homoplasious and many features of the stem were invariant, but a few vegetative anatomical features were informative. Epilyna is de fined by conduplicate leaves with ligules and a lack of leaf abscission layer, and Sertifera has tubercles on the leaf sheaths. Inflorescence architecture within the Sobralieae is plesiomorphically axillary, with independently derived terminal inflorescen ces in Elleanthus plus Epilyna and in the core Sobralia Many sections of Elleanthus are readily identified by inflorescence structure (i.e., distichous versus spirally arranged, capitate versus racemose, etc.).

PAGE 269

269 Unfortunately, although Sobralia sect. Sob ralia is readily distinguished from the core Sobralia based on axillary inflorescences, sect. Sobralia is not monophyletic, and inflorescence variation within that section does not separate the group more finely. Species distribution data were combined wit h phylogenetic data to infer biogeographic, temporal, and diversification patterns. These data show a South American origin for the tribe, with two major lineages (i.e., Elleanthus and a subclade of core Sobralia ) that have diversified as they radiated in to Central America. Analysis reveals that these two movements into Central America were contemporaneous and resulted in increases in the rate of diversification. A final objective is to reconcile the generic classification of the tribe with phylogenetic d ata, resulting in monophyletic genera; the data show that Sobralia is not monophyletic. One nomenclatural solution is to lump all members of the tribe into a single genus, for which Sobralia would have priority. However, inflorescence architecture and fl ower size are largely correlated with the generic concepts Elleanthus Epilyna and Sertifera; nomenclatural stability favors retention of these generic concepts. Retention of these genera will necessitate the division of Sobralia into several monophyleti c groups However, the vast majority of species of Sobralia are not found in the same clade as the type species, S. dichotoma Because the goal of taxonomy is to provide a usage of names that is as consistent as possible with phylogeny collaborators and I have proposed a change of type species for Sobralia ( Dressler et al., 2011 ) that would minimize nomenclatural changes as the smaller groups of Sobralia sect. Sobralia are recognized at the generic level Any proposed generic recircumscriptions will be delayed until this proposed change of type is ruled

PAGE 270

270 upon at the next International Botanical Congress in 2017 Taxon sampling for molecular analyses needs to be expanded to clarify what species belong to each clade before gener ic recircumscriptions are finalized. Using current taxon sampling an informal, new generic classification based upon this work is proposed here (Fig. 6 1). The broader implications of these phylogenetic, biogeographic, morphological and anatomical data are the increase of basic knowledge of this yet poorly understood group of orchids. With this work, we move towards the goal of better und erstanding in species diversity. A more stable classification system provides a better context by which other scient ific endeavors can be applied Because of this, a key to the proposed classification system is presented. Note that two new genera (i.e., new genera 4 and 5) although phylogenetically distinct according to the DNA data, cannot be morphologically separat ed at this time. Key to the Proposed G enera of Sobralieae: Sertifera 2. Inflorescences terminal, sometimes appearing axillary beca use of highly branching 3. Leaves plicate, plants often relatively large (stems usually >20 cm), flowers spirally arranged, flowers and/or bracts usually with vivid Elleanthus in part. 3. Leaves conduplicate, usually small plants (stems usually <20 cm), flowers distichously arranged, flowers white, bracts with green, white or brown

PAGE 271

271 4. Inflorescence held erect (and vaguely fractiflex), bracts much shorter than ................ Epilyna 4. Inflorescence usually held horizontally, projected perpendicular to the axis of the stem (and strongly fractiflex), bracts barely shorter than the f lower and Elleanthus sect. Cloidelyna 1. Flowers more than 1 cm wide, and more than 3 cm long (usually much Sobralia sensu lato). 5. Inflorescences terminal, produced once, and o nly in one flowering 6. Inflorescences simple, fractiflex, or condensed racemes (with densely overlapping bracts and very short internodes) with white, purple, magenta, red, Sobralia sensu stricto (type = S. biflora ) new genus 1 ( Sobralia dorbignyana ). 5. Inflorescences axillary, produced in succession with stems often lastin g for several or more years, and with older inflorescences remaining attached below the newer 7. Leaves conduplicate, ~5 mm wide, with no prominent veins, except for the new genus 2 ( Sobral ia stenophylla ). 7. Leaves plicate, usually more than 5 cm wide, with many prominent 8. Flowers magenta, with a bright orange spot, perianth usually less than 2 cm .. new genus 3 ( Sobra lia ciliata ).

PAGE 272

272 8. Flowers variously colored, perianth usually more than 4 cm in length new genera 4 & 5 ( Sobralia portillae group & Sobrali a dichotoma group)

PAGE 273

273 Figure 6 1. Phylogenetic tree from 6 locus ML analysis o f DNA data, outlining generic circumscriptions of current genera (in black boxes) and new genera to be delimited (black text, to right).

PAGE 274

274 LIST OF REFERENCES A CKERMAN J. 1983. Euglossine bee pollination of the orchid, Cochleanthes lipscombiae : a food source mimic. American Journal of Botany 70: 830 834. ______. 1986. Mechanisms and evolution of food deceptive polliantion systems in orchids. Lindleyana 1: 1 08 113. A CKERMAN J., A. A. C UEVAS AND D. H OF 2011. Are deception pollinated species more variable than those offering a reward? Plant Systematics and Evolution 293: 91 99. A LISCONI S. S., J. P. T ORRETTA M. E. B ELLO AND B. G. G ALATI 2009. Elaiophores in Gomesa bifolia (Sims) M. W. Chase & N. H. Williams (Oncidiinae: Cymbidieae: Orchidaceae): structure and oil secretion. Annals of Botany 104: 1141 1149. A NGELY Paulo. A NT ONELLI A., J. A. A. N YLANDER C. P ERSSON AND I. S ANMARTIN 2009. Tracing the impact of the Andean uplift on Neotropical plant evolution. Proceedings of the National Academy of Sciences 106: 9749 9754. A PPLEQUIST W. L. 2012. Report of the Nomenclature Co mmittee for Vascular Plants: 64. Taxon 61: 1108 1117. A REVALO R., J. F IGUEROA AND S. M ADRINAN 2011. Anatomia foliar de ocho especies de orquideas epifitas. Lankesteriana 11: 39 54. A TWOOD J. T. 1986. The size of the Orchidaceae and the systematic distr ibution of epiphytic orchids. Selbyana 9: 171 186. B AKER H. G. 1975. Concentrations in nectars form hummingbird flowers. Biotropica 7: 37 41. B AKER H. G., AND I. B AKER 1983. A brief historical review of the chemistry of floral nectar. In B. Bentley and T. S. Elias [eds.], The biology of nectaries. Columbia University Press, New York, New York. B ARRINGER K. 1985. Three new species of Elleanthus (Orchidaceae) from Central America. Brittonia 37: 286 290. ______. 1987. Two new species of Elleanthus (Orchida ceae) from Ecuador and Peru. Systematic Botany 12: 162 166. B ARTHLOTT W., C. N EINHUIS D. C UTLER F. D ITSCH I. M EUSEL I. T HEISEN AND H. W ILHELMI 1998. Classification and terminology of plant epicuticular waxes. Botanical Journal of the Linnean Society 126: 237 260.

PAGE 275

275 B ASKIN S. I., AND C. A. B LISS 1969. Sugar occurring in the extrafloral exudates of the Orchidaceae. Phytochemistry 8: 1139 1145. B EARDSLEY P. M., A. Y EN AND R. G. O LMSTEAD 2003. AFLP phylogeny of Mimulus section Erythranthe and the evol ution of hummingbird pollination Evolution 57: 1397 1410. B ECK J., G. B REHM AND K. F IEDLER 2011. Links between the environment, abundance and diversity of Andean moths. Biotropica 43: 208 217. B ENNETT D. E., AND E. A. C HRISTENSON 1999. Sobralia altiss ima Orchids 68: 1112 1113. B ENNETT F. D. 1972. Baited McPhail fruitfly traps to collect euglossine bees. Journal of the New York Entomological Society 80: 137 145. B ENZING D. H. 1990. Vascular epiphytes: general biology and related biota. Cambridge Univ ersity Press, New York, New York. B ENZING D. H., AND D. W. O TT 1981. Vegetative reduction in epiphytic Bromeliaceae and Orchidaceae: Its origin and significance. Biotropica 13: 131 140. B ENZING D. H., D. W. O TT AND W. E. F RIEDMAN 1982. Roots of Sobral ia macrantha (Orchidaceae): structure and function of the velamen exodermis complex. American Journal of Botany 69: 608 614. B ERTSCH A. 1983. Nectar production of Epilobium angustifolium L. at different air humidities; nectar sugar in individual flowers a nd the optimal foraging theory. Oecologia 59: 40 48. B LANCO M. A., AND G. B ARBOZA 2005. Pseudocopulatory pollination in Lepanthes (Orchidaceae: Pleurothallidinae) bu fungus gnats. Annals of Botany 95: 763 772. B LANCO M. A., G. C ARNEVALI W. M. W HITTEN R. B. S INGER S. K OEHLER N. H. W ILLIAMS I. O JEDA K. M. N EUBIG AND L. E NDARA 2007. Generic realignments in Maxillariinae (Orchidaceae). Lankesteriana 7: 515 537. B LEIWEISS R. 1998a. Tempo and mode of hummingbird evolution. Biological Journal of the Li nnean Society 65: 63 76. ______. 1998b. Slow rate of molecular evolution in high elevation hummingbirds. Proceedings of National Academy of Sciences 95: 612 616. B OLTEN A. B., AND P. F EINSINGER 1978. Why do hummingbird flowers secrete dilute nectar? Biot ropica 10: 307 309. B OLTEN A. B., P. F EINSINGER H. G. B AKER AND I. B AKER 1979. On the calculation of sugar concentration in flower nectar. Oecologia 41: 301 304.

PAGE 276

276 B ORRELL B. J. 2005. Long tongues and loose niches: evolution of euglossine bees and their nectar flowers. Biotropica 37: 664 669. ______. 2006. Mechanics of nectar feeding in the orchid be Euglossa imperialis : pressure, viscosity and flow. The Journal of Experimental Biology 209: 4901 4907. B RAGA P. 1977. Aspectos biologicos das Orchidaceae d a Amaznica Central. Acta Amazonica, Manaus 7: 1 89. B RIEGER F. G. 1983. Sobraliinae. In F. G. Brieger, R. Maatsch, and K. Senghas [eds.], Die Orchideen. Verlag Paul Parey, Berlin. B RIGGS J. C. 1984. Freshwater fishes and biogeography of Central America and the Antilles. Systematic Zoology 33: 428 435. ______. 1994. The genesis of Central America: biology versus geophysics. Global Ecology and Biogeography Letters 4: 169 172. B URKE J. M., R. W YATT C. W. DE P AMPHILIS AND M. L. A RNOLD 2000. Nectar charact eristics of interspecific hybrids and their parents in Aesculus (Hippocastanaceae) and Iris (Iridaceae). Journal of the Torrey Botanical Society 127: 200 206. B URNHAM R. J., AND A. G RAHAM 1999. The history of Neotropical vegetation: new developments and status. Annals of the Missouri Botanical Garden 86: 546 589. B URQUEZ A., AND S. A. C ORBET 1991. Do flowers reabsorb nectar? Functional Ecology 5: 369 379. B URTON K. W., H. F. L ING AND R. K. O N IONS 1997. Closure of the Central American Isthmus and its effect on deep water formation in the North Atlantic. Nature 386: 382 385. C AMERON K. M. 2001. An expanded phylogenetic analysis of Orchidaceae using three plastid genes: rbcL atpB and psaB American Journal of Botany 88: 104. ______. 2004. Utility of plastid psaB gene sequences for investigating intrafamilial relationships within Orchidaceae. Molecular Phylogenetics and Evolution 31: 1157 1180. C AMERON K. M., M. W. C HASE W. M. W HITTEN P. J. K ORES D. C. J ARRELL V. A. A LBERT T. Y UKAWA H. G. H ILLS AND D. H. G OLDMAN 1999. A phylogenetic analysis of the Orchidaceae: evidence from rbcL nucleotide sequences. American Journal of Botany 86: 208 224.

PAGE 277

277 C ARDONA A., V. A. V ALENCIA G. B AYONA J. D UQUE M. D UCEA G. G EHRELS C. J ARAMILLO C. M ONTES G. O JED A AND J. R UIZ 2010. Early subduction related orogeny in the northern Andes: Turonian to Eocene magmatic and provenance record in the Santa Marta Massif and Rancheria Basin, northern Colombia. Terra Nova 23: 26 34. C ARLQUIST S. 1961. Comparative plant an atomy: a guide to taxonomic and evolutionary applications of anatomical data in angiosperms. Holt, Reinhart & Winston, New York, New York. C ARLSWARD B. S. 2004. Molecular systematics and anatomy of Vandeae (Orchidaceae): the evolution of monopodial leafle ssness, University of Florida, Gainesville. C ARLSWARD B. S., W. L. S TERN W. S. J UDD AND T. W. L UCANSKY 1997. Comparative leaf anatomy and systematics in Dendrobium sections Aporum and Rhizobium (Orchidaceae). Internation Journal of Plant Sciences 158: 332 342. C ATLING P. M. 1980. Rain assisted autogamy in Liparis loeselii (L.) L. C. Rich. (Orchidaceae). Bulletin of the Torrey Botanical Club 107: 525 529. C HASE M. W., AND H. G. H ILLS 1991. Silica gel: an ideal material for field preservation of leaf samples for DNA studies. Taxon 40: 215 220. C HASE M. W., J. V. F REUDENSTEIN K. M. C AMERON AND R. L. B ARRETT 2003. DNA data and Orchidaceae systematics: a new phylogenetic classification. In K. W. Dixon, S. P. Kell, R. L. Barrett, and P. J. Cribb [eds.] Orchid Conservation, 69 89. Natural History Publications, Kota Kinabalu, Malaysia. C HRISTENSON E. A. 2002. Three Sobralias new to Peru. Orchids 71: 994 1001. ______. 2003. Three showy new orchids from Ecuador. The Orchid Review 111: 243 246. C HRISTENSON E. A., AND C. M ORETZ 2003. Sobralia boliviensis The Orchid Review 111: 374 376. C OATES A. G., J. B. J ACKSON L. S. C OLLINS T. M. C RONN H. J. D OWSETT L. M. B YBELL P. J UNG AND J. A. O BANDO 1992. Closure of the Isthmus of Panama: the near shore mar ine record of Costa Rica and western Panama. Geological Society of America Bulletin 104: 814 828. C OLLANTES B., AND M. L EON 1999. Inquil Symbol of carnival in Huachocolpa, Peru. Orchids 68: 1102 1111. C ONRAN J. G., J. M. B ANNISTER AND D. E. L EE 2009 Earliest orchid macrofossils: early Miocene Dendrobium and Earina (Orchidaceae: Epidendroideae) from New Zealand. American Journal of Botany 96: 466 474.

PAGE 278

278 C ORBET S. A. 2003. Nectar sugar content: estimating standing crop and secretion rate in the field. Apidologie 34: 1 10. C OZZOLINO S., AND A. W IDMER 2005. Orchid diversity: an evolutionary consequence of deception? TRENDS in Ecology and Evolution 20: 487 494. C URRY K. J., L. M. M C D OWELL W. S. J UDD AND W. L. S TERN 1991. Osmophores, floral features, and systematics of Stanhopea (Orchidaceae). American Journal of Botany 78: 610 623. D ACOSTA J. M., AND J. K LICKA 2008. The Great American Interchange in birds: a phylogenetic perspective with the genus Trogon Molecular Ecology 17: 1328 1343. D ANFORTH B N., S. S IPES J. F ANG AND S. G. B RADY 2006. The history of early bee diversification based on five genes plus morphology. Proceedings of National Academy of Sciences 103: 15118 15123. D ARWIN C. 1862. On the various contrivances by which British and fo reign orchids are fertilised by insects, and on the good effects of intercrossing. John Murray, London. D AVIES K. L., AND M. S TPICZYNSKA 2008a. Labellar micromorphology of two euglossine pollinated orchid genera; Scuticaria Lindl. and Dichaea Lindl. Anna ls of Botany 102: 805 824. D AVIES K. L., AND M. S TPICZYNSKA 2008b. The anatomical basis of floral, food reward production in Orchidaceae. In J. Teixeira da Silva [ed.], Floriculture, ornamental and biotechnology: advances and topical issues, V, 392 406. Global Science Books Isleworth, Middlesex. D AVIES K. L., M. S TPICZYNSKA AND A. G REGG 2005. Nectar secreting floral stomata in Maxillaria anceps Ames & C. Schweinf. (Orchidaceae). Annals of Botany 96: 217 227. DE M ELO M. C., E. L. B ORBA AND E. A. S OUSA P AIVA 2010. Morphological and histological characterization of the osmophores and nectaries of four species of Acianthera (Orchidaceae: Pleurothallidinae). Plant Systematics and Evolution 286: 141 151. D EARNALEY J. D. W. 2007. Further advances in orchid mycorrhizal research. Mycorrhiza 17: 475 486. D EVEY S. S., R. M. B ATEMAN M. F. F AY AND J. A. H AWKINS 2008. Friends or relatives? Phylogenetics and species delimitation in the controversial European orchid genus Ophrys Annals of Botany 101: 385 402.

PAGE 279

279 D ICK M. H., A. H ERRERA C UBILLA AND J. B. C. J ACKSON 2003. Molecular phylogeny and phylogeography of free living Bryozoa (Cupuladriidae) from both sides of the Isthmus of Panama. Molecular Phylogenetics and Evolution 27: 355 371. D ODSON C. H. 1962. The i mportance of pollination in the evolution of the orchids of tropical America. American Orchid Society Bulletin 31: 525 534, 641 649, 731 735. ______. 1965. Agentes de polinizacin y su influencia sobre la evolucin en la familia orquidacea. Universidad Nac ional de la Amazona Peruana, Iquitos, Peru. ______. 1966. Ethology of some bees of the tribe Euglossini (Hymenoptera: Apidae). Journal of the Kansas Entomological Society 39: 607 609. D ODSON C. H., AND C. A. L UER 2010. Orchidaceae: Genera Cyrtochiloides Epibator Elanders Sverige AB, Mlnlycke, Sweden. D ODSON C. H., R. L. D RESSLER H. G. H ILLS R. M. A DAMS AND N. H. W ILLIAMS 1969. Biologically active compounds in orchid fragrances. Science 164: 1243 1249. D OYLE J. J., AND J. L. D OYLE 1987. A rapid D NA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11 15. D RESSLER R. L. 1971. Dark pollinia in hummingbird pollinated orchids or do hummingbirds suffer from strabismus? The American Naturalist 105: 80 83. ______. 1976. How to study orchid pollination without any orchids. In K. Senghas [ed.], Proceedings of the 8th World Orchid Conference, 534 537. German Orchid Society Inc., Frankfurt. ______. 1981. The orchids: natural history and classification. Harvard Universi ty Press, Cambridge. ______. 1982. Biology of the orchid bees (Euglossini). Annual Review of Ecology and Systematics 13: 373 394. D RESSLER R. L. 1993a. Phylogeny and classification of the orchid family. Dioscorides Press, Portland, Oregon. ______. 1993b. Field guide to the orchids of Costa Rica and Panama. Comstock Publishing Associates, Ithaca, New York. D RESSLER R. L. 1995. New orchid species from Costa Rica. Novon 5: 140 145. ______. 2001. Two Sobralias: Sobralia chrysostoma is described and the identi ty of Sobralia leucoxantha is resolved. Orchids 70: 750 751.

PAGE 280

280 ______. 2002. The major sections or groups within Sobralia with four new species from Panama and Costa Rica, S. crispissima S. gloriana S. mariannae and S. nutans Lankesteriana 5: 9 15. _____ _. 2003a. Orchidaceae. In B. E. Hammel, M. H. Grayum, C. Herrera, and N. Zamora [eds.], Manual de Plantas de Costa Rica, volumen III, 1 595. Missouri Botanical Garden Press, St. Louis, MO. ______. 2003b. Sobralia quinata a new species in section Globosae Lankesteriana 6: 27 28. ______. 2004a. Two new species of Elleanthus from Panama Journal fur den Orchideenfreund 11: 142 148. ______. 2004b. The Sobralia undatocarinata complex: where have they been hiding? Orchids 73: 774 778. ______. 2005. Six new speci es of Sobralia from Panama. Lindleyana [in Orchids] 74: 937 941. ______. 2006. Sobralia andreae a new and attractive species, and easy to grow! Orquideloga 26: 130 132. ______. 2007. Sobralia macrophylla and other sobralias with tubular bracts, with two new species, S. rarae avis and S. sororcula Orquideloga 25: 43 49. ______. 2009. Can sobralias be classified? Orchids 78: 658 663. ______. 2011. Las Sobralias se pueden clasificar? El complejo de Sobralia warszewiczii Lankesteriana 11: 239 243. _____ _. 2012. Sobrali decora : the species and its cousins from Mexico and Central America. Orchids 81: 308 310. D RESSLER R. L., AND D. B OGARIN 2007a. Two attractive new species of Sobralia from Panama. Lindleyana [in Orchids] 76: 696 701. ______. 2007b. Ellea nthus ligularis a name for a relatively common "new" species of Elleanthus sect. Chloidelyna Lankesteriana 7: 539 542. D RESSLER R. L., AND F. P UPULIN 2008. The identity of Sobralia leucoxantha with three new species, two closely allied and one more di stant. Orquideloga 25: 152 158. ______. 2010. Sobralia lacerata una especie Colombiana nueva, atractiva y con clasificacion confusa por largo tiempo. Orquideloga 27: 35 42. D RESSLER R. L., AND D. B OGARIN 2010. Some new Sobralieae from Costa Rica and P anama. Lankesteriana 9: 475 485.

PAGE 281

281 D RESSLER R. L., M. A. B LANCO F. P UPULIN AND K. M. N EUBIG 2011. (2019) Proposal to conserve the name Sobralia (Orchidaceae) with a conserved type. Taxon 60: 907 908. D UCKE A. 1902. As especies Paraenses do genero Euglos sa Latr. Boletim Museu Paraense 3: 1 19. D UDEK M., AND P. T UKALLO 2007. Sobraliinae Schlecht. & Elleanthinae Szlach. classificatory differences. Genus (Supplement) 14: 51 52. D UDEK M., AND D. L. S ZLACHETKO 2010. New combinations in subtribe Elleanthi nae (Orchidaceae, Epidendroideae). Richardiana 11: 1 12. D UNSTERVILLE G. C. K. 1979. Some Venezuelan Elleanthus species. American Orchid Society Bulletin 48: 665 671. D UNSTERVILLE G. C. K., AND L. A. G ARAY 1979. Orchids of Venezuela: an illustrated fiel d guide, A G. Harvard University Printing Office, Allston, MA. D UNSTERVILLE G. K. C., AND E. D UNSTERVILLE 1982. Pinelia alticola and Elleanthus norae : a very short and a very tall story. The Orchid Review 90: 322 324. D URKEE L. T. 1983. The ultrastructu re of floral and extrafloral nectaries. In B. Bentley and T. S. Elias [eds.], The biology of nectaries. Columbia University Press, New York, New York. D ZIEDZIOCH C., A. D. S TEVENS AND G. G OTTSBERGER 2003. The hummingbird plant community of tropical mont ane rain forest in southern Ecuador Plant Biology 5: 331 337. E NDARA C. L., N. H. W ILLIAMS AND M. W. W HITTEN 2011. Filogenia molecular preliminar de Scaphosepalum (Orchidaceae: Pleurothallidinae). Lankesteriana 11: 245 252. E NDRESS P. K. 1994. Diversit y and evolutionary biology of tropical flowers. Cambridge University Press, New York, New York. E NGARD C. J. 1944. Morphological identity of the velamen and exodermis in orchids. Botanical Gazette 105: 457 462. E RKENS R. H. J., L. W. C HATROU J. W. M AAS T. VAN DER N IET AND V. S AVOLAINEN 2007. A rapid diversification of rainforest trees ( Guatteria ; Annonaceae) following dispersal from Central into South America. Molecular Phylogenetics and Evolution 44: 399 411. E SCHRICH W., R. B URCHARDT AND S. E SSIAM AH 1989. The induction of sun and shade leaves of the European beech ( Fagus sylvatica L.): anatomical studies. Trees 3: 1 10.

PAGE 282

282 F AHN A. 1979. Ultrastructure of nectaries in relation to nectar secretion. American Journal of Botany 66: 977 985. F EINSINGER P 1978. Ecological interactions between plants and hummingbirds in a successional tropical community. Ecological Monographs 48: 269 287. F EINSINGER P., R. K. C OLWELL J. T ERBORGH AND S. B. C HAPLIN 1979. Elevation and morphology, flight energetics, and f oraging ecology of tropical hummingbirds. American Naturalist 113: 481 497. F ELSENSTEIN J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783 791. F ENSTER C. B., W. S. A RMBRUSTER P. W ILSON M. R. D UDASH AND J. D. T HOMSON 2004. Pollination syndromes and floral specialization. Annual Review of Ecology, Evolution, and Systematics 35: 375 403. F IGUEIREDO A. C. S., AND M. S. P AIS 1992. Ultrastructural aspects of the nectary spur of Limodorum abortivum (L.) Sw. (Orch idaceae). Annals of Botany 70: 325 331. F ITCH W. M. 1971. Toward defining the course of evolution: minimum change for a specific tree topology. Systematic Zoology 20: 406 416. F ITZGERALD P. 2002. Tectonics and landscape evolution of the Antarctic plate s ince the breakup of Gondwana, with an emphasis on the West Antarctic Rift System and the Transantarctic Mountains. Royal Society of New Zealand Bulletin 35: 453 469. F OGDEN M., AND P. F OGDEN 2006. Hummingbirds of Costa Rica. Firefly Books, Buffalo, New Y ork. F REUDENSTEIN J. V., AND F. N. R ASMUSSEN 1999. What does morphology tell us about orchid relationships? a cladistic analysis. American Journal of Botany 86: 225 248. G ALETTO L., AND G. B ERNARDELLO 2004. Floral nectaries, nectar production dynamic s and chemical composition in six Ipomoea species (Convolvulaceae) in relation to pollinators. Annals of Botany 94: 269 280. G ALETTO L., G. B ERNARDELLO AND G. L. R IVERA 1997. Nectar, nectaries, flower visitors, and breeding system in five terrestrial Or chidaceae from central Argentina. Journal of Plant Research 110: 393 403. G ALETTO L., G. B ERNARDELLO AND C. A. S OSA 1998. The relationship between floral nectar composition and visitors in Lycium (Solanaceae) from Argentina and Chile: what does it refle ct? Flora 193: 303 314. G ARAY L. A. 1978a. Studies in American orchids, X. Botanical Museum Leaflets 26: 1 38.

PAGE 283

283 ______. 1978b. Elleanthus In G. Harling and B. Sparre [eds.], Flora of Ecuador, 57 110. Department of Systematic Botany, University of Goetebor g and Section for Botany, Riksmuseum, Stockholm. G ARDNER A. G., M. V AIO M. G UERRA AND E. E MSHWILLER 2012. Diversification of the American bulb bearing Oxalis (Oxalidaceae): dispersal to North America and modification of the tristylous breeding system. American Journal of Botany 99: 152 164. G ENTRY A. H. 1982. Neotropical floristic diversity: phytogeographical connections between Central and South America, Pleistocene climatic fluctuation, or an accident of the Andean orogeny? Annals of the Missouri Bot anical Garden 69: 557 593. G IVNISH T. J., K. C. M ILLIAM P. E. B ERRY AND K. J. S YTSMA 2007. Phylogeny, adaptive radiation, and historical biogeography of Bromeliaceae inferred from ndhF sequence data. Aliso 23: 3 26. G OLDSMITH T. H. 1980. Hummingbirds see near ultraviolet light. Science 207: 786 788. G ORNIAK M., O. P AUN AND M. W. C HASE 2010. Phylogenetic relationships within Orchidaceae based on a low copy nuclear coding gene, Xdh : Congruence with organellar and nuclear ribosomal DNA results. Molecul ar Phylogenetics and Evolution 56: 784 795. G OTTSBERGER G., J. S CHRAUWEN AND H. F. L INSKENS 1984. Amino acids and sugars in nectar, and their putative evolutionary significance. Plant Systematics and Evolution 145: 55 77. G OVAERTS R. 2001. How many spe cies of seed plants are there? Taxon 50: 1085 1090. G RAHAM A. 2003. Historical phytogeography of the Greater Antilles. Brittonia 55: 357 383. G RAHAM S. W., J. R. K OHN B. R. M ORTON J. E. E CKENWALDER AND S. C. H. B ARRETT 1998. Phylogenetic congruence a nd discordance among one morphological and three molecular data sets from Pontederiaceae. Systematic Biology 47: 545 567. G RIESBACH R. J. 2002. Development of Phalaenopsis orchids for the mass market. In J. Janick and A. Whipkey [eds.], Trends in new crop s and new uses. ASHS Press, Alexandria, VA. G UO Y. Y., Y. B. L UO Z. J. L IU AND X. Q. W ANG 2012. Evolution and biogeography of the slipper orchids: Eocene vicariance of the conduplicate genera in the Old and New World tropics. PLoS One 7: e38788. H TER E. 1992. New combinations in Epidendrum 12: 296.

PAGE 284

284 H AINSWORTH F. R., AND L. L. W OLF 1972. Crop volume, nectar concentration and hummingbird energetics. Comparative Biochemistry and Physiology Part A: Physiology 42: 359 366. ______. 1976. Ne ctar characteristics and food selection by hummingbirds. Oecologia 25: 101 113. H ARDER L. D., AND S. D. J OHNSON 2008. Function and evolution of aggregated pollen in angiosperms. International Journal of Plant Sciences 169: 59 78. H ASCHKE M. R., E. S CHEU BER A. G UNTHER AND K. J. R EUTTER 2002a. Evolutionary cycles during the Andean orogeny: repeated slab breakoff and flat subduction? Terra Nova 14: 49 55. H ASCHKE M. R., W. S IEBEL A. G UNTHER AND E. S CHEUBER 2002b. Repeated crustal thickening and recyc ling during the Andean orogeny in north Chile (21 26). Journal of Geophysical Research 107: 1 18. H EALY S. D., AND T. A. H URLY 2001. Foraging and spatial learning in hummingbirds. In L. Chittka and J. D. Thomson [eds.], Cognitive Ecology of Pollination Cambridge University Press, New York. H EDGES S. B. 1982. Caribbean biogeography: implications of recent plate tectonic studies. Systematic Zoology 31: 518 522. ______. 1996. Vicariance and dispersal in Caribbean biogeography. Herpetologica 52: 466 473. H EDGES S. B., C. A. H ASS AND L. R. M AXSON 1992. Caribbean biogeography: molecular evidence for dispersal in West Indian terrestrial vertebrates. Proceedings of the National Academy of Sciences 89: 1909 1913. H OORN C., J. G UERRERO G. A. S ARMIENTO AND M. A. L ORENTE 1995. Andean tectonics as a cause for changing drainage patterns in Miocene northern South America. Geology 23: 237 240. H OORN C., F. P. W ESSELINGH H. TER S TEEGE M. A. B ERMUDEZ A. M ORA J. S EVINK I. S A. S ANCHEZ M ESEGUER C. L. A NDERSON J. P. F IGUEIREDO C. J ARAMILLO D. R IFF F. R. N EGRI H. H OOGHIEMSTRA J. L UNDBERG T. S TADLER T. S A RKINEN AND A. A NTONELLI 2010. Amazonia through time: Andean uplift, climate change, landscape evolution, and biodiversity. Science 330: 927 931. H UGHES C. E., AND R. E ASTWOOD 2006. Island radiation on a continental scale: Exceptional rates of plant diversification after uplift of the Andes. Proceedings of National Academy of Sciences 103: 10334 10339.

PAGE 285

285 H USS ON L., C. P. C ONRAD AND C. F ACCENNA 2008. Tethyan closure, Andean orogeny, and westward drift of the Pacific Basin. Earth and Planetary Science Letters 271: 303 310. J ACKSON J. B. C., P. J UNG A. G. C OATES AND L. S. C OLLINS 1993. Diversity and extinc tion of tropical American mollusks and emergence of the Isthmus of Panama. Science 260: 1624 1626. J ANZEN D. H. 1971. Euglossine bees as long distance pollinators of tropical plants. Science 171: 203 205. J EFFREY D. C., J. A RDITTI AND H. K OOPOWITZ 1970 Sugar content in floral and extrafloral exudates of orchids: pollination, myrmecology and chemotaxonomy implication. The New Phytologist 69: 187 195. J ERSKOV J., AND S. D. J OHNSON 2006. Lack of floral nectar reduces self pollination in a fly pollinat ed orchid. Oecologia 147: 60 68. J ERSKOV J., S. D. J OHNSON AND P. K INDLMANN 2006. Mechanisms and evolution of deceptive pollination in orchids. Biological Review 81: 219 235. J OHNSON L. A., AND D. E. S OLTIS 1994. matK DNA sequences and phylogenetic reconstruction in Saxifragaceae s.st. Systematic Botany 19: 143 156. ______. 1998. Assessing congruence: empirical examples from molecular data. In D. E. Soltis, P. S. Soltis, and J. J. Doyle [eds.], Molecular systematics of plants II: DNA sequencing, 297 348. Kluwer Academic Publishers, Boston, Massachusetts. J OHNSON S. D., AND L. A. N ILSSON 1999. Pollen carryover, geitonogamy, and the evolution of deceptive polination systems in orchids. Ecology 80: 2607 2619. J OHNSON S. D., AND S. W. N ICOLSON 2007. E volutionary associations between nectar properties and specificity in bird systems. Biology Letters 4: 49 52. J OHNSON S. D., H. P. L INDER AND K. E. S TEINER 1998. Phylogeny and radiation o pollination systems in Disa (Orchidaceae). American Journal of Bo tany 85: 402 411. J OHNSON S. D., C. I. P ETER L. A. N ILSSON AND A. J 2003. Pollination success in a deceptive orchid is enhanced by co occuring rewarding magnet plants. Ecology 84: 2919 2927. J ORDAN W. C., M. W. C OURTNEY AND J. E. N EIGEL 1996. Low le vels of intraspecific genetic variation at a rapidly evolving chloroplast DNA locus in North American duckweeds (Lemnaceae). American Journal of Botany 83: 430 439. J RGENSEN P. M., AND S. L E"N Y NEZ 1999. Catalogue of the vascular plants of Ecuador. Mis souri Botanical Garden Press, St. Louis, MO.

PAGE 286

286 J UDD W. S., C. S. C AMPBELL E. A. K ELLOGG P. F. S TEVENS AND C. J. D ONOGHUE 2008. Plant Systematics: A Phylogenetic Approach, 3rd ed. Sinauer Associates, Inc., Sunderland, Massachusetts. K AISER R. 1993. The scent of orchids: olfactory and chemical investigations. Elsevier Science Publishers, New York. K AY K. M., P. A. R EEVES R. G. O LMSTEAD AND D. W. S CHEMSKE 2005. Rapid speciation and the evolution of hummingbird pollination in neotropical Costus subgenu s Costus (Costaceae): evidence from the nrDNA ITS and ETS sequences. American Journal of Botany 92: 1899 1910. K EVAN P. G., L. C HITTKA AND A. G. D YER 2001. Limits to the salience of ultraviolet: lessons from colour vision in bees and birds. The Journal of Experimental Biology 204: 2571 2580. K IRBY S. H. 2011. Active mountain building and the distribution of "core" Maxillariinae species in tropical Mexico and Central America. Lankesteriana 11: 275 291. K NIGHT T. M. 2003. Floral density, pollen limitatio n, and reproductive success in Trillium grandiflorum Oecologia 442: 557 563. K NOWLTON N., AND L. A. W EIGT 1998. New dates and new rates for divergence across the Isthmus of Panama. Proceedings of the Royal Society of London B 265: 2257 2263. K NOWLTON N ., L. A. W EIGT L. A. S OLORZANO D. K. M ILLS AND E. B ERMINGHAM 1993. Divergence in proteins, mitochondrial DNA, and reproductive compatibility across the Isthmus of Panama. Science 260: 1629 1632. K OOPOWITZ H., AND T. A. M ARCHANT 1998. Postpollintion n ectar reabsorption in the African epiphyte Aerangis verdickii (Orchidaceae). American Journal of Botany 85: 508 512. K RON K., AND J. L. L UTEYN 2005. Origins and biogeographic patterns in Ericaceae: New insights from recent phylogenetic analysis. Biologis ke Skrifter 55: 479 500. K UNIN W., AND Y. I WASA 1996. Pollinator foraging strategies in mixed floral arrays: density effects and floral constancy. Theoretical Population Biology 49: 232 263. L ESSIOS H. A. 1981. Divergence in allopatry: molecular and mor phological differentiation between sea urchins seperated by the Isthmus of Panama. Evolution 35: 618 634. L EVIN R. A., AND J. S. M ILLER 2005. Relationships within tribe Lycieae (Solanaceae): paraphyly of Lycium and multiple origins of gender dimorphism. American Journal of Botany 92: 2044 2053.

PAGE 287

287 L IA V. V., V. A. C ONFALONIERI C. I. C OMAS AND J. H. H UNZINKER 2001. Molecular phylogeny of Larrea and its allies (Zygophyllaceae): reticulate evolution and the probable time of creosote bush arrival to North Am erica. Molecular Phylogenetics and Evolution 21: 309 320. L ICHTENTHALER H. K., C. B USCHMANN M. D LL H. J. F IETZ T. B ACH U. K OZEL D. M EIER AND U. R AHMSDORF 1981. Photosynthetic activity, chloroplast ultrastructure, and leaf characteristics of high l ight and low light plants and of sun and shade leaves. Photosynthesis Research 2: 115 141. L JTNANT B. 1976. Obervations on the Elleanthus linifolius alliance (Orchidaceae) in S. America. Botaniska Notiser 129: 445 454. L UNAU K., S. P APIOREK T. E LTZ AN D M. S AZIMA 2011. Avoidance of achromatic colours by bees provides a private niche for hummingbirds. The Journal of Experimental Biology 214: 1607 1612. L YERLY S. B., B. F. R IESS AND S. R OSS 1950. Color preference in the Mexican violet eared hummingbir d, Calibri t. thalassinus (Swainson). Behaviour 2: 237 248. M AAD J., AND L. A. N ILSSON 2004. On the mechanism of floral shifts in speciation: gained pollination efficiency from tongue to eye attachment of pollinia in Platanthera (Orchidaceae). Biologica l Journal of the Linnean Society 83: 481 495. M ACDONALD D., I. G OMEZ P EREZ J. F RANZESE L. S PALLETTI L. L AWVER L. G AHAGAN I. D ALZIEL C. T HOMAS N. T REWIN M. H OLE AND D. P ATON 2003. Mesozoic break up of SW Gondwana: implications for regional hydroc arbon potential of the southern South Atlantic. Marine and Petroleum Geology 20: 287 308. M AC F ADDEN B. J., M. X. K IRBY A. R INCON C. M ONTES S. E. M ORON N. S TRONG AND C. J ARAMILLO 2010. Extinct peccary "Cynorca" occidentale (Tayassuidae, Taya ssuinae) from the Miocene of Panama and correlations to North America. Journal of Paleontology 84: 288 298. M ADDISON W., AND D. M ADDISON 2005. Mesquite: A modular system for evolutionary analysis. Evolution 62: 1103 1118. M AJURE L., W. S. J UDD P. S. S O LTIS AND D. E. S OLTIS 2012. Phylogeny of Opuntia s.s. (Cactaceae): clade delineation, geographic origin, and reticulate evolution. American Journal of Botany 99: 847 864. M ARSHALL L. G., R. F. B UTLER R. E. D RAKE G. H. C URTISS AND R. H. T EDFORD 1979. Calibration of the great American interchange. Science 204: 272 279. M ARTOS F., F. M UNOZ T. P AILLER I. K OTTKE C. G ONNEAU AND M. A. S ELOSSE in press. The role of epiphytism in architecture and evolutionary constraint within mycorrhizal networks of tr opical orchids. Molecular Ecology

PAGE 288

288 M C C ORMICK M. K., D. L. T AYLOR K. J UHASZOVA R. K. B URNETT J R D. F. W HIGHAM AND J. P. O N EILL 2012. Limitations on orchid recruitment: not a simple picture. Molecular Ecology 21: 1511 1523. M C D ADE L. A. 1995. Hybri dization and phylogenetics. In P. C. Hoch and A. G. Stephenson [eds.], Experimental and Molecular Approaches to Plant Biosystematics, 305 331. Missouri Botanical Garden Press, St. Louis, MO. M C D ADE L. A., AND J. A. W EEKS 2004. Nectar in hummingbird polli nated Neotropical plants I: patterns of production and variability in 12 species. Biotropica 36: 196 215. M OLVRAY M., P. J. K ORES AND M. W. C HASE 2000. Polyphyly of mycoheterotrophic orchids and functional influences of floral and molecular characters. In K. L. Wilson and D. A. Morrison [eds.], Monocots: systematics and evolution, 441 448. CSIRO Publishing, Collingswood, Victoria, Australia. M ONTES C., A. C ARDONA R. M C F ADDEN S. E. M ORON C. A. S ILVA S. R ESTREPO M ORENO D. A. R N. H OYOS J. W I LSON D. F ARRIS G. A. B AYONA C. A. J ARAMILLO V. V ALENCIA J. B RYAN AND J. A. F LORES 2012. Evidence for middle Eocene and younger land emergence in central Panama: Implications for Isthmus closure. Geological Society of America Bulletin 124: 780 799. M OREIRA A. S. F. P., AND R. M. D S. I SAIAS 2008. Comparative anatomy of the absorption roots of terrestrial and epiphytic orchids. Brazilian Archives of Biology and Technology 51: 83 93. N EILAND M. R. M., AND C. C. W ILCOCK 1998. Fruit set, nectar rewar d, and rarity in the Orchidaceae. American Journal of Botany 85: 1657 1671. N EUBIG K. M., N. H. W ILLIAMS W. M. W HITTEN AND F. P UPULIN 2009a. Molecular phylogenetics and the evolution of fruit and leaf morphology of Dichaea (Orchidaceae: Zygopetalinae). Annals of Botany 104: 457 467. N EUBIG K. M., M. W. W HITTEN M. A. B LANCO C. L. E NDARA N. H. W ILLIAMS AND S. K OEHLER 2011. Preliminary molecular phylogenetics of Sobralia and relatives (Orchidaceae: Sobralieae). Lankesteriana 11: 307 317. N EUBIG K. M ., M. W. W HITTEN B. S. C ARLSWARD M. A. B LANCO C. L. E NDARA N. H. W ILLIAMS AND M. M OORE 2009b. Phylogenetic utility of ycf1 in orchids: a plastid gene more variable than matK Plant Systematics and Evolution 277: 75 84. N EUBIG K. M., M. W. W HITTEN N H. W ILLIAMS M. A. B LANCO L. E NDARA J. G. B URLEIGH K. S ILVERA J. C. C USHMAN AND M. W. C HASE 2012. Generic recircumscriptions of Oncidiinae (Orchidaceae: Cymbidiee) based on maximum likelihood analysis of combined DNA datasets. Botanical Journal of the Linnaean Society 168: 117 146.

PAGE 289

289 N ILSSON L. A. 1988. The evolution of flowers with deep corolla tubes. Nature 334: 147 149. ______. 1992. Orchid pollination biology. TREE 7: 255 259. N IR M. 2000. Orchidaceae Antillanae. DAG Media Publishing, New York, New York. P ACINI E., AND M. N EPI 2007. Nectar production and presentation. In S. W. Nicolson, M. Nepi, and E. Pacini [eds.], Nectaries and nectar. Springer Press, Dordrecht, Holland. P AIS M. S., AND A. C. S. F IGUEIREDO 1994. Floral nectaries from Limod orum abortivum (L.) Sw. and Epipactis atropurpurea Rafin. (Orchidaceae); ultrastructural changes in plastids during the secretory process. Apidologie 25: 615 626. P ANSARIN E. R., AND M. C. E. A MARAL 2008. Pollen and nectar as a reward in the basal epiden droid Psilochilus modestus (Orchidaceae: Triphoreae): A study of floral morphology, reproductive biology and pollination strategy. Flora 203: 474 483. P ARRA T ABLA V., AND C. F. V ARGAS 2007. Flowering synchrony and floral display size affect pollination s uccess in a deceit pollinated tropical orchid. Acta Oecologica 32: 26 35. P EITSCH D., A. F IETZ H. H ERTEL J. DE S OUZA D. F. V ENTURA AND R. M ENZEL 1992. The spectral input systems of hymenopteran insects and their receptor based colour vision. Journal of Comparative Physiology A 170: 23 40. P OREMBSKI S., AND W. B ARTHLOTT 1988. Velamen radicum micromorphology and classification of Orchidaceae. Nordic Journal of Botany 8: 117 137. P OSADA D., AND K. A. C RANDALL 1998. Modeltest: tesing the model of DNA substitution. Bioinformatics 14: 817 818. P RIDGEON A. M. 1981. Absorbing trichomes in the Pleurothallidinae (Orchidaceae). American Journal of Botany 68: 64 71. P RIDGEON A. M., AND W. L. S TERN 1985. Osmophores of Scaphosepalum (Orchidaceae). Botanical G azette 146: 115 123. P RIDGEON A. M., W. L. S TERN AND D. H. B ENZING 1983. Tilosomes in roots of Orchidaceae: morphology and systematic occurence. American Journal of Botany 70: 1365 1377. P RIDGEON A. M., P. J. C RIBB M. W. C HASE AND F. N. R ASMUSSEN 20 05. Genera Orchidacearum, Vol. 4, Epidendroideae (Part One). Oxford University Press, New York, New York.

PAGE 290

290 P RUTSCH J., A. S CHARDT AND R. S CHILL 2000. Adaptations of an orchid seed to water uptake and storage. Plant Systematics and Evolution 220: 69 75. P RYCHID C. J., P. J. R UDALL AND M. G REGORY 2004. Systematics and biology of silica bodies in monocotyledons. The Botanical Review 69: 377 440. P YKE G. H. 1991. What does it cost a plant to produce floral nectar? Nature 350: 58 59. P YKE G. H., AND N. M. W ASER 1981. The production of dilute nectars by hummingbird and honeyeater flowers. Biotropica 13: 260 270. R ABOSKY D. L., AND I. J. L OVETTE 2008. Explosive evolutionary radiations: decreasing speciation or increasing extinction through time? Evolution 62: 1866 1875. R AMBAUT A. 1996. Se Al: Sequence Alignment editor, v2.0a11. University of Oxford, Oxford, UK. R AMIREZ S. R., D. W. R OUBIK C. S KOV AND N. E. P IERCE 2010. Phylogeny, diversification patterns and historical biogeography of euglossine orc hid bees (Hymenoptera: Apidae). Biological Journal of the Linnean Society 100: 552 572. R AMIREZ S. R., B. G RAVENDEEL R. B. S INGER C. R. M ARSHALL AND N. E. P IERCE 2007. Dating the origin of the Orchidaceae from a fossil orchid with its pollinator. Natu re 448: 1042 1045. R ASMUSSEN H. N. 2002. Recent developments in the study of orchid mycorrhiza. Plant and Soil 244: 149 163. R EEVES G., M. W. C HASE P. G OLDBLATT P. R UDALL M. F. F AY A. V. C OX B. L EJEUNE AND T. S OUZA C HIES 2001. Molecular systematic s of Iridaceae: evidence from four plastid regions. American Journal of Botany 88: 2074 2087. R EICHENBACH H. G. 1853. Sobralia Flore des Serres et des Jardins de l'Europe 8: 245 247. ______. 1862. Orchides. Annales Botanices Systematicae 6: 167 933. ___ ___. 1877. Orchideae Roezliana seu criticae descriptae auctore. Linnaea 41: 1 98. R ENNER S. S., AND H. W ON 2001. Repeated evolution of dioecy from monoecy in Siparunaceae (Laurales). Systematic Biology 50: 700 712. R IEGER M., AND P. L ITVIN 1999. Root s ystem hydraulic conductivity in species with contrasting root anatomy. Journal of Experimental Botany 50: 201 209. R OEBUCK K. I., AND W. L. S TEINHART 1978. Pollination ecology and the scent response in the genus Brassavola American Orchid Society Bullet in 47: 507 511. R OGERS B. 2007. Growing Sobralias. Orchids 76: 912 919.

PAGE 291

291 R OTHACKER E. P. 2007. The primitive Epidendroideae (Orchidaceae): phylogeny, character evolution and the systematics of Psilochilus (Triphoreae), Ohio State University, Columbus, Ohi o. R OUBIK D. W. 2000. Deceptive orchids with Meliponini as pollinators. Plant Systematics and Evolution 222: 271 279. R OUBIK D. W., AND J. A CKERMAN 1987. Long term ecology of euglossine orchid bees (Apidae: Euglossini) in Panama. Oecologia 73: 321 333. R OUBIK D. W., AND P. E. H ANSON 2004. Orchid bees of tropical America: biology and field guide. Instituto Nacional de Biodiversidad, Santo Domingo de Heredia, Costa Rica. R OY B., AND A. W IDMER 1999. Floral mimicry: a fascinating yet poorly understood ph enomenon. TRENDS in Plant Science 4: 325 330. S ANDERSON M. J. 2002. Estimating absolute rates of molecular evolution and divergence times in the absence of a molecular clock. Bioinformatics 19: 101 109. ______. 2003. r8s: inferring absolute rates of molec ular evolution and divergence times in the absence of a molecular clock. Bioinformatics 19: 301 302. S ASLIS L AGOUDAKIS C., M. W. C HASE D. N. R OBINSON S. J. R USSELL AND B. B. K LITGAARD 2008. Phylogenetics of neotropical Platymiscium (Leguminosae: Dalbe rgieae): systematics, divergence times, and biogeography inferred from nuclear ribosomal and plastid DNA sequence data. American Journal of Botany 95: 1270 1286. S ASS C., AND C. D. S PECHT 2010. Phylogenetic estimation of the core Bromelioids with an emph asis on the genus Aechmea (Bromeliaceae) Molecular Phylogenetics and Evolution 55: 559 571. S AUQUET H., S. Y. W. H O M. A. G ANDOLFO G. J. J ORDAN P. W ILF D. J. C ANTRILL M. J. B AYLY L. B ROMHAM G. K. B ROWN R. J. C ARPENTER D. M. L EE D. J. M URPHY J. M. K. S NIDERMAN AND F. U DOVICIC 2012. Testing the impact of calibration on molecular divergence times using a fossil rich group: the case of Nothofagus (Fagales). Systematic Biology 61: 289 313. S CHIESTL F. P., AND P. M. S CHLUTER 2009. Floral isolation specialized pollination, and pollinator behavior in orchids. Annual Review of Entomology 54: 425 446. S CHLECHTER R. 1915. Die Orchideen. Verlagsbuchhandlung Paul Parey, Berlin. S CHMIDT L EBUHN A. N., M. K ESSLER AND I. H ENSEN 2007. Hummingbirds as driv ers of plant speciation? TRENDS in Plant Science 12: 329 331. S CHWEINFURTH C. 1938. Nomenclatorial notes. VII. Botanical Museum Leaflets 6: 112.

PAGE 292

292 S COTLAND R. W., AND A. H. W ORTLEY 2003. How many species of seed plants are there? Taxon 52: 101 104. S EIFRI Z W. 1923. The gregarious flowering of the orchid Dendrobium crumenatum American Journal of Botany 10: 32 37. S HAW J., E. B. L ICKEY J. T. B ECK S. B. F ARMER W. L IU J. M ILLER K. C. S IRIPUN C. T. W INDER E. E. S CHILLING AND R. L. S MALL 2005. The to rtoise and the hare II: relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. American Journal of Botany 92: 142 166. S IEGEL C. 2011. Orchids and hummingbirds: sex in the fast lane. Orchid Digest 75. S ILVERA K. 2002. Adapt ive radiation of oil reward compounds among Neotropical orchid species (Oncidiinae), University of Florida, Gainesville, Florida. S ILVERA K., L. S. S ANTIAGO AND K. W INTER 2005. Distribution of crassulacean acid metabolism in orchids of Panama: evidence of selection for weak and strong modes. Functional Plant Biology 32: 397 407. S ILVERA K., L. S. S ANTIAGO J. C. C USHMAN AND K. W INTER 2010a. The incidence of crassulacean acid metabolism in Orchidaceae derived from carbon isotope ratios: a checklist of the flora of Panama and Costa Rica. Botanical Journal of the Linnaean Society 163: 194 222. S ILVERA K., K. M. N EUBIG W. M. W HITTEN N. H. W ILLIAMS K. W INTER AND J. C. C USHMAN 2010b. Evolution along the crassulacean acid metabolism continuum. Functiona l Plant Biology 37: 995 1010. S IMPSON B. B., J. A. T ATE AND A. W EEKS 2005. The biogeography of Hoffmannseggia (Leguminosae, Caesalpinioideae, Caesalpinieae): a tale of many travels. Journal of Biogeography 32: 15 27. S INGER R. B. 2003. Orchid pollinati on: recent developments from Brazil. Lankesteriana 7: 111 114. S MEDMARK J. E. E., AND A. A. A NDERBERG 2007. Boreotropical migration explains hybridization between geographically distant lineages in the pantropical clade Sideroxyleae (Sapotaceae). America n Journal of Botany 94: 1491 1505. S MITHSON A. 2006. Pollinator limitation and inbreeding depression in orchid species with and without nectar rewards. New Phytologist 169: 419 430. S OLEREDER H., AND F. M EYER 1930. Systematische Anatomie der Monokotyled onen, vol. 6. Gebruder Borntraeger, Berlin.

PAGE 293

293 S TAMATAKIS A., T. L UDWIG AND H. M EIER 2005. RAxML III: a fast program for maximum likelihood based inference of large phylogenetic trees. Bioinformatics 21: 456 463. S TERN W. L., AND W. M. W HITTEN 1999. Comp arative vegetative anatomy of Stanhopeinae (Orchidaceae). Botanical Journal of the Linnean Society 129: 87 103. S TERN W. L., AND W. S. J UDD 1999. Comparative vegetative anatomy and systematics of Vanilla (Orchidaceae). Botanical Journal of the Linnean So ciety 131: 353 382. ______. 2000. Comparative anatomy and systematics of the orchid tribe Vanilleae excluding Vanilla Botanical Journal of the Linnean Society 134: 179 202. S TERN W. L., AND B. S. C ARLSWARD 2006. Comparative vegetative anatomy and system atics of the Oncidiinae (Maxillarieae, Orchidaceae). Botanical Journal of the Linnean Society 152: 91 107. ______. 2008. Vegetative anatomy of Calypsoeae (Orchidaceae). Lankesteriana 8: 105 112. S TILES F. G. 2004. Phylogenetic constraints upon morphologic al and ecological adaptation in hummingbirds (Trochilidae): why are there no hermits in the paramo. Ornitologia Neotropical 15: 191 198. S TILES F. G., AND C. E. F REEMAN 1993. Patterns in floral nectar characteristics of some bird visited plant species fr om Costa Rica. Biotropica 25: 191 205. S TPICZYNSKA M. 2003. Nectar resorption in the spur of Platanthera chlorantha Custer (Rchb.) Orchidaceae structural and microautoradiographic study. Plant Systematics and Evolution 238: 119 126. S TPICZYNSKA M., K. L. D AVIES AND A. G REGG 2003. Nectary structure and nectar secretion in Maxillaria coccinea (Jacq.) L.O. Williams ex Hodge (Orchidaceae). Annals of Botany 93: 87 95. ______. 2005. Comparative account of nectary structure in Hexisea imbricata (Lindl.) Rchb .f. (Orchidaceae). Annals of Botany 95: 749 756. S TPICZYNSKA M., K. L. D AVIES AND M. K AMINSKA 2010. Structure of the cuniculus nectary in Brassavola flagellaris Barb. Rodr. (Laeliinae Benth., Orchidaceae). Acta Agrobotanica 63: 3 10. S TPICZYNSKA M., C. M ILANESI C. F ALERI AND M. C RESTI 2005. Ultrastructure of the nectar spur of Platanthera chlorantha (Custer) Rchb. (Orchidaceae) during successive stages of nectar secretion. Acta Biologica Cracoviensia 47: 111 119.

PAGE 294

294 S UN Y., D. Z. S KINNER G. H. L IANG AND S. H. H ULBERT 1994. Phylogenetic analysis of Sorghum and related taxa using internal transcribed spacers of nuclear ribosomal DNA. Theoretical and Applied Genetics 89: 26 32. S WARTS N. D., E. A. S INCLAIR A. F RANCIS AND K. W. D IXON 2010. Ecological specialization in mycorrhizal symbiosis leads to rarity in an endangered orchid. Molecular Ecology 19: 3226 3242. S WOFFORD D. L. 1999. PAUP*: phylogenetic analysis using parsimony (*and other methods), version 4.0b10. Sinauer Associates, Sunderland, Mas sachusettts. S YMMANK L., M. S. S AMAIN J. F. S MITH G. P INO A. S TOLL P. G OETGHEBEUR C. N EINHUIS AND S. W ANKE 2011. The extraordinary journey of Peperomia subgenus Tildenia (Piperaceae): insights into diversification and colonization patterns from its cradle in Peru to the Trans Mexican Volcanic Belt. Journal of Biogeography 38: 2337 2349. S ZLACHETKO D. L., M. D UDEK M. G ORNIAK J. M YTNIK E JSMONT AND R. F ERNANDEZ 2009. Sobralia abadorum (Orchidaceae) a new orchid species from Peru. Biodiversity: R esearch and Conservation 15: 3 8. T ABERLET P., L. G IELLY G. P AUTOU AND J. B OUVET 1991. Universal primers for amplification of three non coding regions of chloroplast DNA. Plant Molecular Biology 17: 1105 1109. T T., J. T J. J ERSKOV G. HOV AND M. A. S ELOSSE 2012. Symbiotic germination capability of four Epipactis species (Orchidaceae) is broader than expected from adult ecology. American Journal of Botany 99: 1020 1032. T HOMSON J. D., AND P. W ILSON 2008. Explaining evolutionary shifts between bee and hummingbird pollination: convergence, divergence, and directionality. International Journal of Plant Sciences 169: 23 38. VAN DER C INGEL N. A. 2001. An atlas of orchid pollination: America, Africa, Asia and Australia. A. A. Balkema Rotterdam, Netherlands. VAN DER P IJL L., AND C. H. D ODSON 1966. Orchid flowers: their pollination and evolution. University of Miami Press, Coral Gables, Florida. V ASSILYEV A. E. 2010. On the mechanisms of nectar secretion: revisited. Annals of Botany 105: 349 354. V EYRET Y. 1981. Quelques aspects du pistil et de son devenir chez quelques Sobraliinae (Orchidaceae) de Guyane. Adansonia 3: 75 83. W EBB S. D. 1991. Ecogeography and the Great American Interchange. Paleobiology 17: 266 280.

PAGE 295

295 W HITTALL J. B. AND S. A. H ODGES 2007. Pollinator shifts drive increasingly long nectar spurs in columbine flowers. Nature 447: 706 709. W HITTEN M. W., N. H. W ILLIAMS AND M. W. C HASE 2000. Subtribal and generic relationships of Maxillarieae (Orchidaceae) with emphas is on Stanhopeinae: combined molecular evidence. American Journal of Botany 87: 1842 1856. W HITTEN W. M., N. H. W ILLIAMS R. L. D RESSLER G. G ERLACH AND F. P UPULIN 2005. Generic relationships of Zygopetalinae (Orchidaceae: Cymbidieae): combined molecula r evidence. Lankesteriana 5: 87 107. W HITTEN W. M., M. A. B LANCO N. H. W ILLIAMS S. K OEHLER G. C ARNEVALI R. B. S INGER C. L. E NDARA AND K. M. N EUBIG 2007. Molecular phylogenetics of Maxillaria and related genera (Orchidaceae: Cymbidieae) based on com bined molecular data sets. American Journal of Botany 94: 1860 1889. W ILLIAMS N. H. 1981. Floral fragrance components of Brassavola Selbyana 5: 279 285. W ILLIAMS N. H., AND W. M. W HITTEN 1983. Orchid floral fragrances and male euglossine bees: methods and advances in the last sesquidecade. Biological Bulletin 164: 355 395. Y ANO S., AND I. T ERASHIMA 2001. Separate localization of light signal perception for sun and shade type chloroplast and palisade tissue fifferentiation in Chenopodium album Plant C ell Physiology 42: 1303 1310. Y OUNG K. R., C. U LLOA U LLOA J. L. L UTEYN AND S. K NAPP 2002. Plant evolution and endemism in Andean South America: an introduction. Botanical Review 68: 4 21. Y UAN Y. W., AND R. G. O LMSTEAD 2008. A species level phylogene tic study of the Verbena complex (Verbenaceae) indicates two independent intergeneric chloroplast transfers. Molecular Phylogenetics and Evolution 48: 23 33.

PAGE 296

296 BIOGRAPHICAL SKETCH Kurt Maximillian Neubig was born in Baton Rouge Louisiana, to Henry Conra d Neubig of Plaquemine and Linda Hutchins Neubig o f Shreveport He graduated from peers (his crowning achievement in life to this point) He attended and graduated from Louisiana State University in Baton Rouge. While at LSU, Kurt met Lowell Urbatsch, a plant systematist of infinite wisdom who helped steer Kurt into the dismally hopeless career of botany. After graduating with an aptly named BS Kurt moved to Gainesville with great haste to start his graduate career at the University of Florida under the guidance of the magnanimous Norris H. Williams. Within a year of moving to UF he married the beautiful and enchanting Julie Kay Eggert, who b ore him a son of great bearing, Henry George Neubig. Through the masterful guidance of Norris, Kurt was able to squeak through with a n MS in 2005, and likewise with a PhD in 2012. career goals include having an office big enough to hold a couch s o that he can finally get some rest, and earning enough money to feed his small, but very hungry family.