<%BANNER%>

Phylogeny and Character Evolution of Kielmeyeroideae (Clusiaceae) based on Molecular and Morphological Data

Permanent Link: http://ufdc.ufl.edu/UFE0005660/00001

Material Information

Title: Phylogeny and Character Evolution of Kielmeyeroideae (Clusiaceae) based on Molecular and Morphological Data
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0005660:00001

Permanent Link: http://ufdc.ufl.edu/UFE0005660/00001

Material Information

Title: Phylogeny and Character Evolution of Kielmeyeroideae (Clusiaceae) based on Molecular and Morphological Data
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0005660:00001


This item has the following downloads:


Full Text

PAGE 1

PHYLOGENY AND CHARACTER EVOL UTION OF KIELMEYEROIDEAE (CLUSIACEAE) BASED ON MOLECULAR AND MORPHOLOGICAL DATA By CHRISTINE NOTIS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Christine Notis

PAGE 3

ACKNOWLEDGMENTS I would like to express my gratitude to my committee members (Walter Judd, Doug Soltis, and Pam Soltis) for all of their help and commitment. I thank the Soltis lab for providing the financial support that allowed me to complete this project. I thank Peter Stevens for providing me with many of his unpublished manuscripts on Clusiaceae, and for his continued help throughout my project. My thanks also extend to Kent Perkins for his help in obtaining herbarium specimen loans; and to Matt Gitzendanner, Norris Williams, Mark Whitten, and my fellow graduate students for their help with running analyses and troubleshooting problems in the laboratory. I thank several people for collecting plants in the field, or specimens from herbaria: Mario Blanco, Fred Damon, Utkarsh Ghate, Cesar Grandez, Gretchen Ionta, Louis Santiago, and Katia Silvera. I thank the following herbaria for their loan of specimens: University of Florida, New York Botanical Garden, Harvard University, Missouri Botanical Garden, Fairchild Tropical Botanic Garden, and the Smithsonian. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................vi LIST OF FIGURES..........................................................................................................vii ABSTRACT.........................................................................................................................x CHAPTER 1 INTRODUCTION........................................................................................................1 2 MATERIALS AND METHODS.................................................................................5 Taxon Sampling for Molecular Data Sets....................................................................5 DNA Amplification, Sequencing, and Alignment........................................................5 Morphological Characters............................................................................................6 Phylogenetic Analyses................................................................................................17 3 RESULTS...................................................................................................................37 Analysis 1: rbcL of Clusiaceae...................................................................................37 Analysis 2: rbcL + matK of Kielmeyeroideae............................................................38 Analysis 3: ITS of Kielmeyeroideae...........................................................................39 Analyses 4 and 5: rbcL + matK + ITS of Kielmeyeroideae.......................................40 Analysis 6: Morphology of Kielmeyeroideae.............................................................41 Analyses 7 and 8: DNA + Morphology of Kielmeyeroideae.....................................41 Analyses 9 and 10: DNA + Morphology of Kielmeryoideae, Including Neotatea....42 Character Evolution....................................................................................................43 4 DISCUSSION.............................................................................................................70 Analytical Issues.........................................................................................................70 The Value of Morphology..........................................................................................71 Taxonomic History.....................................................................................................71 Character Evolution....................................................................................................78 Biogeography..............................................................................................................83 iv

PAGE 5

5 KEY TO THE GENERA OF KIELMEYEROIDEAE...............................................99 APPENDIX A TAXA USED FOR DNA ANALYSIS.....................................................................101 B SPECIMENS EXAMINED......................................................................................104 LIST OF REFERENCES.................................................................................................111 BIOGRAPHICAL SKETCH...........................................................................................116 v

PAGE 6

LIST OF TABLES Table page 2-1 Morphological characters used in a cladistic analysis of Kielmeyeroideae (Clusiaceae)..............................................................................................................13 2-2 Morphological character-state coding used in phylogenetic analysis of Kielmeyeroideae (Clusiaceae)..................................................................................22 3-1 Bootstrap support values by parsimony analysis.....................................................68 A-1 Taxa used for DNA analysis..................................................................................102 vi

PAGE 7

LIST OF FIGURES Figure page 2-1 Stipuliform structures...............................................................................................27 2-2 Overview of leaf of Endodesmia calophylloides showing mesophyll with latex/ resin cavities and canals...........................................................................................28 2-3 Inflorescence architectures of Kielmeyeroideae......................................................29 2-4 Androecial features of Kielmeyeroideae..................................................................30 2-5 Androecium of Clusiella axillaris............................................................................31 2-6 Gynoecial features of Kielmeyeroideae...................................................................32 2-7 Diagrammatic longitudinal sections of ovaries showing ovule position at anthesis.....................................................................................................................33 2-8 Three types of winged seeds found in Kielmeyeroideae..........................................34 2-9 Leaf blade cross sections illustrating anatomical characters used in phylogenetic analyses....................................................................................................................35 2-10 Leaf blade cross sections illustrating anatomical characters used in phylogenetic analyses....................................................................................................................36 3-1 Strict consensus of 1530 most parsimonious trees of length 845 from Analysis 1 (rbcL alone)..............................................................................................................50 3-2 Majority-rule consensus tree based on rbcL data for Clusiaceae (Analysis 1)........51 3-3 Maximum likelihood tree of rbcL sequences across Clusiaceae (Analysis 1).........52 3-4 Strict consensus of 510 most parsimonious trees of length 361 from Analysis 2 (rbcL + matK)...........................................................................................................53 3-5 Maximum likelihood tree of rbcL + matK data set (Analysis 2).............................54 3-6 Strict consensus of 10 most parsimonious trees of length 987 from Analysis 3 (ITS alone)................................................................................................................55 vii

PAGE 8

3-7 Maximum likelihood tree of ITS sequences (Analysis 3)........................................56 3-8 Strict consensus of 2 most parsimonious trees of length 1326 from Analysis 4 (rbcL + matK + ITS, unpruned)...............................................................................57 3-9 Majority-rule consensus tree based on the unpruned rbcL + matK + ITS data set (Analysis 4)..............................................................................................................58 3-10 Majority-rule consensus tree based on the pruned rbcL + matK + ITS data set (Analysis 5)..............................................................................................................59 3-11 Strict consensus of 61 most parsimonious trees of length 227 from Analysis 6 (morphology alone)..................................................................................................60 3-12 Strict consensus of 4 most parsimonious trees of length 1990 from Analysis 7 (DNA + morphology, unpruned)..............................................................................61 3-13 Strict consensus of 4 most parsimonious trees of length 1297 from Analysis 8 (DNA + morphology, pruned)..................................................................................62 3-14 Strict consensus of 28 most parsimonious trees of length 2009 from Analysis 9 (DNA + morphology, unpruned, including Neotatea).............................................63 3-15 Strict consensus of 12 most parsimonious trees of length 1315 from Analysis 10 (DNA + morphology, pruned, including Neotatea).................................................64 3-16 Morphological character-state transformations mapped onto one of the most parsimonious trees (tree #2) from analysis 7 (DNA + morphology, unpruned)......65 3-17 Morphological character-state transformations mapped onto one of the most parsimonious trees (tree #1) from analysis 9 (DNA + morphology with Neotatea, pruned) showing the clade containing Neotatea......................................67 4-1 Character-state distribution for resin/latex canals in leaf mesophyll within Kielmeyeroideae.......................................................................................................84 4-2 Character state distribution for resin/latex cavities in leaf mesophyll within Kielmeyeroideae.......................................................................................................85 4-3 Character-state distribution for leaf arrangement within Kielmeyeroideae.............86 4-4 Character-state distribution for lignification of the leaf margin within Kielmeyeroideae.......................................................................................................87 4-5 Character-state distribution for transcurrent lateral bundles in the leaf blade within Kielmeyeroideae...........................................................................................88 4-6 Character-state distribution for petiole bundle architecture within Kielmeyeroideae.......................................................................................................89 viii

PAGE 9

4-7 Character-state distribution for anther glands within Kielmeyeroideae...................90 4-8 Character-state distribution for carpel number within Kielmeyeroideae.................91 4-9 Character-state distribution for fruit type within Kielmeyeroideae.........................92 4-10 Character-state distribution for seed form within Kielmeyeroideae........................93 4-11 Character-state distribution for exotegmen presence within Kielmeyeroideae........94 4-12 Character-state distribution for endosperm in mature seeds within Kielmeyeroideae.......................................................................................................95 4-13 Character-state distribution for embryo length within Kielmeyeroideae.................96 4-14 Character-state distribution for cordate cotyledons within Kielmeyeroideae..........97 4-15 Biogeographical patterns within Kielmeyeroideae..................................................98 ix

PAGE 10

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science PHYLOGENY AND CHARACTER EVOLUTION OF KIELMEYEROIDEAE (CLUSIACEAE) BASED ON MOLECULAR AND MORPHOLOGICAL DATA By Christine Notis August 2004 Chair: Walter S. Judd Cochair: Douglas E. Soltis Major Department: Botany Clusiaceae are a family of approximately 1000 species with a pantropical distribution. Based on a morphology-based cladistic analysis, the family has been divided into three subfamilies: Hypericoideae, Clusioideae, and Kielmeyeroideae. Kielmeyeroideae are divided into two tribes: Calophylleae, a large, pantropical tribe (ca. 450 spp), and Endodesmieae, a small, tropical African group (comprising the two monotypic genera Endodesmia and Lebrunia). Subfamilial assignment of Endodesmieae has been uncertain. Based on fruit characters, Endodesmieae are similar to Calophylleae; however, vegetatively, they are similar to Clusioideae. A previous family-level study based on rbcL sequences confirmed the monophyly of the three traditional subfamilies except that Clusiella, traditionally a member of Clusioideae, was placed in Kielmeyeroideae. Internal support for relationships within Kielmeyeroidae was weak, leaving it unclear as to which genera Clusiella is most closely related. Species of Endodesmieae were not included in the rbcL study. The present study determined the x

PAGE 11

generic relationships within Kielmeyeroideae based on phylogenetic analysis of rbcL, matK, and ITS sequence data and morphological characters. Parsimony analyses were conducted on each molecular and morphological data set separately and combined. Bayesian and maximum likelihood analyses were performed on separate and combined molecular data sets. Sampling included several species of most genera of Calophylleae, as well as Clusiella and the enigmatic Endodesmia. The rbcL and matK data sets support the sister-group relationship of Endodesmia to the remaining Kielmeyeroideae and the sister-group relationship of Mammea to the rest of Calophylleae. Within the core Calophylleae (all Calophylleae except Mammea), the strictly New World genera (Kielmeyera, Caraipa, Haploclathra, Clusiella, Mahurea, Neotatea, and Marila) likely form a clade, and the primarily Old World genera (Kayea, Poeciloneuron, Mesua, and Calophyllum) constitute a second clade. All morphological characters are mapped onto a total evidence tree in order to infer their evolutionary pattern. Character state reconstructions of several noteworthy morphological characters, such as the presence/absence of latex cavities and canals, leaf arrangement, presence/absence of anther glands, carpel number, fruit type, and seed form are discussed in more detail. Putative morphological synapomorphies for each genus were determined, and a key to the genera of Kielmeyeroideae is provided. xi

PAGE 12

CHAPTER 1 INTRODUCTION Clusiaceae are a large, pantropical family of approximately 1000 species in the Malpighiales (Stevens, 2001). Within Malpighiales, Clusiaceae are probably most closely related to Bonnetiaceae and Podostemaceae (Gustafsson et al., 2002; Weitzman and Stevens, 1997), but relationships to other families in this order remain uncertain (Soltis et al., 2000). Stevens (1980; in press, and unpubl.) based on morphological and anatomical research, divided Clusiaceae into three subfamilies: Clusioideae, Hypericoideae, and Kielmeyeroideae. Clusioideae are usually glabrous, dioecious plants with latex/resin canals, fasciculate or nonfasciculate stamens, and short or obsolete styles. Members of Hypericoideae are herbs to shrubs of temperate or high-elevation tropical areas with resin cavities, fasciculate stamens, and free styles. Most Kielmeyeroideae have an indumentum of unior multicellular hairs, latex/resin cavities and/or canals, nonfasciculate stamens, and fused styles. Gustafsson et al. (2002) reconstructed the phylogeny of Clusiaceae using the chloroplast gene rbcL. The analysis provided support for the monophyly of three clades: Kielmeyeroideae, Clusioideae, and Hypericoideae + Podostemaceae, except that Clusiella, traditionally placed in Clusioideae (but questioned by Hammel, 1999), appeared in Kielmeyeroideae. Resolution of the relationships within Kielmeyeroideae was poor, and several genera of Kielmeyeroideae (Haploclathra, Neotatea, Poeciloneuron, and Endodesmia) were not included. 1

PAGE 13

2 Stevens (in press) recognized two tribes within Kielmeyeroideae: Calophylleae, a large, pantropical tribe (11 genera, ca. 450 spp); and Endodesmieae, a small, tropical African group (2 monotypic genera). The placement of Endodesmieae within Kielmeyeroideae was considered tentative because of its similarities to both Clusioideae and Kielmeyeroideae. Many members of Kielmeyeroideae are economically important. For example, Mammea americana is a popular fruit (called the mammey apple) in the Caribbean. Several species of Calophyllum, Kayea, and Haploclathra are used for timber in construction (Stevens, 1980, unpubl.; Vasquez, 1993). Mesua ferrea and some species of Calophyllum are planted as ornamentals. The fruits of Calophyllum yield an oil that is sometimes used medicinally or in lamps (Stevens, 1980). Calophyllum lanigerum is being investigated as a possible use for the treatment of AIDS because it contains a nonnucleoside inhibitor of HIV 1 reverse transcriptase (Greer, 2001). Kielmeyeroideae offer the opportunity to examine the evolution of several noteworthy morphological characters, such as latex system, anther glands, androecial form, and fruit type. The latex system may consist of cavities (more or less spherical latexor resin-containing structures) and/or canals (elongated latexor resin-containing structures). In Calophyllum and Neotatea, canals largely replace intersecondary veins. The mesophyll in the leaves of Clusiella and Endodesmia contains both cavities and canals; other genera have either cavities or canals. In this study, I assessed the phylogenetic significance of these structures, and attempted to determine the ancestral condition for the subfamily.

PAGE 14

3 Many members of Kielmeyeroideae have apical glands on their anthers, which contain a resinous or oily substance that may serve to attract pollinators (Stevens, unpubl.). The glands range in shape from spherical to elongate or bowl-shaped (crateriform). In addition to having anther glands, the flowers of Clusiella also have resin-secreting staminodes, similar to the large genus Clusia (Clusioideae: Clusieae), to which Clusiella was thought to be related based on its epiphytic habit, sessile, cupuliform stigmas, and non-ascendent ovules (Planchon and Triana, 1860; Engler, 1925). The phylogenetic significance of these characters is assessed in connection with a reconsideration of the phylogenetic placement of Clusiella. The taxonomic value of anther glands is considered within a phylogenetic context. Most members of Kielmeyeroideae produce capsular fruits (septicidal or septifragal); however, several genera produce indehiscent fruits. Mammea, Calophyllum, and Endodesmia produce one-seeded berries; Clusiella has a several-seeded berry. A phylogeny of Kielmeyeroideae has allowed me to determine that from a capsular ancestor, fleshy, indehiscent fruits likely have evolved several times within the subfamily. The fruit form within Clusiaceae is quite homoplasious, as in many other families (Kron et al., 2002; Wilson et al., 2001; Judd et al., 2002). Kielmeyeroideae are distributed throughout the tropics. Within Calophylleae, Neotatea, Marila, Mahurea, Clusiella, Kielmeyera, Caraipa, and Haploclathra are found in the New World tropics; Poeciloneuron, Mesua, and Kayea are found only in the Old World tropics; and Mammea and Calophyllum each have a few species in the New World and a majority of their species in Madagascar to the Pacific (Stevens, 1980, unpubl.). The two genera of Endodesmieae (Endodesmia and Lebrunia) are found in tropical

PAGE 15

4 Africa. Estimation of a phylogeny for the group allowed analysis of the evolution of taxa in relation to the geographic region in which they occur. Kielmeyeroideae have apparently diversified within both the Paleoand Neotropics. The primary goal of this research project was to resolve the generic relationships within Kielmeyeroideae, using molecular and morphological data. Although a previous phylogenetic analysis of Clusiaceae using rbcL (Gustafsson et al., 2002) included many members of Kielmeyeroideae and showed support for the monophyly of this subfamily (including Clusiella), generic relationships within Kielmeyeroideae were not resolved, and several genera were not included in that analysis (i.e., Haploclathra, Neotatea, Poeciloneuron, and Endodesmia). In the present study, rbcL and matK sequences were used to determine the phylogenetic position of Endodesmia (which had never before been included in a molecular cladistic analysis); and rbcL, matK, ITS sequences and morphology were used to infer generic relationships and circumscription within Kielmeyeroideae. The value of morphology in cladistic analyses was examined by comparing resolution and levels of support in trees resulting from analyses with and without morphology. In addition, all 74 morphological characters were mapped onto the total evidence topology to gain a better understanding of patterns of evolution.

PAGE 16

CHAPTER 2 MATERIALS AND METHODS Taxon Sampling for Molecular Data Sets Taxa selected and gene regions sequenced are summarized in Appendix A, along with voucher information and GenBank accession numbers. Generally, several species of each genus of Kielmeyeroideae were included in molecular analyses. Neotatea, found in the tepuis of northern South America; and Lebrunia, an African member of Endodesmieae, are the only genera not represented in the molecular data sets. Several taxa of Kielmeyeroideae not included in the family-level rbcL study of Gustafsson et al. (2002) were sequenced and added to the previously published sequences. DNA Amplification, Sequencing, and Alignment Total DNA was isolated from 34 species of Kielmeyeroideae according to the methods of Soltis et al. (1991) modified from Doyle and Doyle (1987), scaled down to 1.0-mL extraction volumes. Fifteen to 20 mg of silica-dried leaves or leaf tissue removed from herbarium specimens were ground using ceramic mortars and pestles with liquid nitrogen and a pinch of sand. DNA was incubated overnight at 60C. After the chloroform extraction, the DNA was precipitated overnight at C with 0.08 volume of 7.5 M ammonium acetate and 0.54 volume of isopropanol. DNA was centrifuged, washed in 70% ethanol, washed again in 95% ethanol, dried, and resuspended in 50 L of Tris-EDTA buffer and stored at C. Three different gene regions were amplified to resolve phylogenetic relationships within Kielmeyeroideae: rbcL, matK, and ITS. Amplification of DNA was performed using 25-L reactions containing 5.0 mmol/L KCl, 5

PAGE 17

6 2.0 mmol/L Tris pH 8.3, 1.5 mmol/L MgCl2, 0.001% Tween, 0.05% DMSO, 0.2 mmol/L dNTPs, 1.25 units of Taq polymerase, and 0.5 mol/L primer. The primers used for amplification of rbcL were Z-1F, Z-895R, Z-674F, and 3 rbcL, designed by G. Zurawski (DNAX Research Institute, Palo Alto, CA, USA). The PCR thermal cycling profile consisted of 35 cycles with an annealing temperature of 55C. A ca. 500-basepair (bp) region of matK was amplified with the primers trnK-710F and matK-1168R (Johnson and Soltis, 1995). The PCR thermal cycling profile consisted of ten cycles with an annealing temperature of 49C and 30 cycles with an annealing temperature of 39C. ITS was amplified using the forward primer N-nc18S10 and reverse primer C26A (Wen and Zimmer 1996). A touchdown thermal cycling program was used; the initial annealing temperature was 53C, decreasing 1C per cycle for six cycles, followed by 36 cycles at 48C. All PCR products were sequenced in both directions with the PCR primers used as sequencing primers using Beckman-Coulter Dye Terminator Cycle Sequencing Quick-Start kits and a Beckman-Coulter CEQ 2000 or CEQ 8000 automated sequencers following the manufacturers protocols (Beckman Coulter, Inc., Fullerton, CA, USA), except that half reactions were used. Sequences in each data set were aligned by eye using SeAl 2.0 (Rambaut, 1996). The alignments were straightforward except for the ITS sequences, where alignment among genera was sometimes difficult. Several alignments were tried, and all alignments resulted in the same topology. Morphological Characters Taxa used in the morphological data set were chosen to match those taxa available for the molecular analyses. The morphological data set included all genera of

PAGE 18

7 Kielmeyeroideae except Lebrunia, for which herbarium specimens were not available. Clusia lanceolata and Hypericum tetrapetalum, representatives of Clusioideae and Hypericoideae, respectively, were selected as outgroups based on previous morphological (Stevens, unpubl.) and molecular (Gustafsson et al., 2002) cladistic analyses. When herbarium specimens in reproductive condition were not available for a given species used in the molecular data set, a congeneric species was used as a placeholder. (In combined molecular and morphological analyses, the morphological data of Mammea subsessifolia were combined with the DNA data of Mammea sessiliflora, the morphological data of Kielmeyera speciosa were combined with the DNA data of Kielmeyera rosea, the morphological data of Kielmeyera coriacea were combined with the DNA data of Kielmeyera petiolaris, the morphological data of Caraipa sp. were combined with the DNA data of Caraipa utilis, the morphological data of Haploclathra leleanii were combined with the DNA data of Haploclathra cordata, the morphological data of Kayea borneensis were combined with the DNA data of Kayea stylosa, and the morphological data of Calophyllum fibrosum were combined with the DNA data of Calophyllum soulattri.) In the characters examined for this analysis, infrageneric variation is minimal, and the analysis focused on intergeneric (not intrageneric) relationships. Therefore, a congeneric placeholder seems justified following the approach of Kron and Judd (1997). Appendix B lists herbarium specimens examined. Morphological and anatomical characters were examined using herbarium material supplemented by alcohol-preserved flowers collected from cultivated plants at Fairchild Tropical Botanic Garden, in Miami, FL. Dried material was rehydrated overnight in a solution of tap water and a drop of liquid hand soap (Howard, 1974; Judd et al., 2002).

PAGE 19

8 Sectioning for anatomical observations was done by hand with a razor blade, and the sections were stained with 2-3 drops of phloroglucinol and 1 drop of HCl. Temporary slides were created by suspending the sections in glycerol and applying clear nail polish to seal the cover slip. A data matrix of 74 characters was ultimately constructed (Tables 2-1 and 2-2). Most of the characters examined were easily divisible into discrete states, therefore avoiding arbitrary decisions regarding state delimitation (Stevens, 1991). Several characters, such as bud length to width ratio, bracteole position, and petiole base excavation, were initially observed but excluded from the final matrix because they could not be delimited into clear states. Characters 54-58 in the matrix were taken from Stevens (unpubl.) morphological analysis of the family. These characters are testa vasculature pattern, exotegmen presence or absence, endosperm presence or absence, embryo length, and cotyledon shape. The interpretation of stipuliform structures (character 8) was sometimes difficult. When present, most stipuliform structures are attached on the stem adjacent to the region where the basal part of the petiole attaches to the stem (Figure 2-1a); those of Mahurea exstipulata are attached towards the upper part of where the petiole attaches to the stem, adjacent to the axillary bud (Figure 2-1b). It is possible that the stipuliform structures of Mahurea exstipulata are actually modified colleters (Stevens, unpubl.), but in this analysis they were treated the same as other paired stipuliform structures. The nodes of Clusiella have a distinctive interpetiolar stipuliform structure (Figure 2-1c), which was coded as a different state from the petiole-adjacent structures of other taxa.

PAGE 20

9 In most taxa, the distinction between cavities (more or less spherical latexor resin-containing structures surrounded by veins; character 12) and canals (elongated latexor resin-containing structures, independent of the course of veins; character 13) is clear. However, the mesophyll of both Endodesmia and Clusiella contains cavities with a continuous range of sizes from spherical to oblong or elongate cavities that may span almost the entire distance from midvein to margin (Figure 2-2). These taxa were scored as having both cavities and canals. Inflorescence types (character 17) are defined following Judd et al. (2002). Some inflorescences did not correspond clearly to a specific category. For example, several different types of cymose inflorescences were observed. The Neotatea-type cyme can be described as being a congested raceme-like structure with spiraled units associated with large bracts (Figure 2-3a). The Clusiella-type cyme has primary and secondary axes that produce several internodes before producing the terminal flower (Figure 2-3b). The other types of inflorescences detected in this study are diagramed in Figure 2-3c-i. Calyx and corolla aestivation (characters 22 and 26, respectively) were observed at the point of insertion of the sepals or petals, respectively, and were defined according to Endress (1994). A flower with quincuncial aestivation of the calyx or corolla is defined as having two sepals or two petals entirely outside the others; imbricate aestivation refers to having one sepal or one petal entirely outside the others. A contorted corolla is one in which each petal has one side covering the adjacent petal and the other side under the other adjacent petal; no petals are entirely outside the others. When only two sepals are present, aestivation is opposite; aestivation is decussate when there are more than two

PAGE 21

10 sepals present and they are opposite and decussate. Opposite and decussate aestivation patterns were treated as the same state. A fasciculate androecium (character 30) is a striking feature of many members of Clusioideae and Hypericoideae, but has also been reported in a few members of Kielmeyeroideae. Endodesmia has been described as having a fasciculate androecium (Engler, 1925; Stevens, unpubl.), but is scored here as lacking fascicles. Observations of several flowers from different specimens did not indicate that the monadelphous tube of stamens (Figure 2-4a,b) represents a fasciculate androecium. Several species of Calophyllum have been described by some authors (Smith, 1920; Robson, 1976) as having fascicles. Stevens (1980) stated that the appearance of fascicles is probably caused by separation of the androecium along lines of weakness due to the removal of the petals during dissection. Following most previous interpretations, C. inophyllum and C. brasiliense were coded as having fascicles. Developmental studies could clarify the nature of the apparent fasciculation in Calophyllum. Stamen dimorphism (character 31) was observed on perfect and staminate flowers. Endodesmia calophylloides has some anthers that produce pollen and others in which pollen is evidently not produced (Figure 2-4c). Both types of anthers produce resin: fertile anthers produce resin in an expanded, elongate connective, while nonfertile anthers secrete resin from most of the anther surface. Clusiella also has a dimorphic androecium, but this was coded as a separate state from that of Endodesmia. The staminate flowers of Clusiella contain both resin-secreting staminodes (at the base of the staminal tube) and fertile stamens, the latter also apparently secrete resin or oil through the connective (Figure 2-5).

PAGE 22

11 In most taxa, the presence or absence of an anther gland (character 34) is an obvious feature (e.g., Kayea borneensis, Marila laxiflora, and Mahurea exstipulata; Figure 2-4d-f). Some taxa have an expanded connective and it is difficult to determine whether or not this structure is secreting resin. In these cases, the presence of a shiny or oily substance on the structure was considered sufficient evidence to characterize it as an anther gland (e.g., Endodesmia calophylloides, Figure 2-4c; Clusiella axillaris, Figure 2-5c). Placentation type (character 38) was generally clear; however, the distinction between axile and intruded parietal placentation is subtle, and it seems probable that one is derived from the other (Figure 2-6a,b). The ovaries of some taxa (e.g., Mahurea exstipulata) have both axile and intruded parietal placentation, depending on where the cross section of the ovary is made. My character-state delimitations reflect this by treating these two conditions as the same state (state 0: placentation axile or intruded-parietal). A nonseptate ovary with parietal placentation (e.g., Hypericum tetrapetalum, Figure 2-6c), however, was easily distinguishable and considered a distinct state from the axile or deeply intruded parietal (and thus falsely septate) conditions characteristic of most of the other taxa. The ovary of Clusiella is distinctive: the placentation is laminar, with the ovules scattered on the partitions (Figure 2-6d). The mature ovaries of a few taxa have a noteworthy feature of the septa (character 40). In both Mahurea and Neotatea, the intruded placentae are bordered by the carpel walls that have curled in (Figure 2-6e). In all other taxa with intruded placentae, the carpel walls do not curl in and border the placentae.

PAGE 23

12 Ovule position at anthesis (character 42) may be more or less median, basal, or apical. In a median position, the ovules may be attached throughout the length of the axis or placental region, or may be clumped somewhere in the middle (Figure 2-7a). In a basal position (Figure 2-7b), the ovule(s) are attached only at the base of the ovary or at the base of the axis. An apical position (Figure 2-7c) indicates that the ovule(s) are attached only at the apex of the ovary. Several taxa of Kielmeyeroideae have winged seeds (character 52). A careful examination of wing morphology and anatomy was needed to hypothesize whether or not wings of these taxa are homologous. A wing may be either two cell layers thick (i.e., Kielmeyera) or more than two cell layers thick (i.e., Caraipa, Haploclathra, Mahurea, and Neotatea), and each type may extend completely around the seed (i.e., Kielmeyera, Caraipa, and Haploclathra) or only partly around the seed (i.e., Mahurea and Neotatea). Some of the wings contain vascular tissue (i.e., Mahurea and Neotatea), while others do not (Kielmeyera, Caraipa, and Haploclathra; Stevens, unpubl.). In this analysis, four different states were delimited: (1) having no wing, (2) having a wing more than two cell layers thick that completely surrounds the seed (Figure 2-8a), (3) having a wing two cell layers thick that completely surrounds the seed (Figure 2-8b), and (4) having a wing more than two cell layers thick that does not completely surround the seed (Figure 2-8c). Sixteen characters included in the morphological data set were taken from leaf and petiole anatomy (characters 59-74). Most of these characters are diagramed in Figures 2-9 and 2-10.

PAGE 24

Table 2-1. Morphological characters used in a cladistic analysis of Kielmeyeroideae (Clusiaceae) 1. Terminal bud present (0); terminal bud not present (1) 2. Terminal bud without scales (0); terminal bud with scales (1) 3. Axillary bud visible but flush with stem (0); axillary bud visible, small to prominent, but not flush with stem (1); axillary bud immersed in stem (2) 4. Indumentum of unbranched unicellular hairs absent (0); indumentum of unbranched unicellular hairs present (1) 5. Indumentum of multicellular hairs absent (0); indumentum of multicellular hairs present (1) 6. Leaf arrangement opposite (0); leaf arrangement alternate (1) 7. Colleters present (0); colleters absent (1) 8. Stipuliform structures absent (0); stipuliform structures present and paired, adjacent to point of attachment of petiole (1); stipuliform structures present and interpetiolar (2) 9. Petiole or leaf bases enclosing terminal bud (0); petiole or leaf bases not enclosing terminal bud (1) 10. Secondary venation eucamptodromous (0); secondary venation brochidodromous (1) 11. Tertiary venation not evident (0); tertiary venation reticulate (1); tertiary venation percurrent (2) 12. Free latex/resin glands not in mesophyll (0); free latex/resin glands present in mesophyll (1) 13. Canals in mesophyll present (0); canals in mesophyll absent (1) 14. Intersecondary veins not replaced by canals (0); intersecondary veins largely replaced by canals (1) 15. Inflorescence/flower position terminalor terminal and axillary (0); inflorescence/flower position axillary (1) 16. Shoot growth monopodial (0); shoot growth sympodial (1) 17. Inflorescence a cyme (0); infloresence a panicle-like cyme (1); inflorescence a Clusiella-type cyme (2); inflorescence a Neotatea-type cyme (3); inflorescence a raceme (4); inflorescence fasciculate (5) 18. Terminal flower present (0); terminal flower absent (1) 19. Bracteoles present (0); bracteoles absent (1) 20. Bracteoles without abaxial gland (0); bracteoles with abaxial gland (1) 21. Plant dioecious or androdioecious (0); plant monoecious (1) 13

PAGE 25

Table 2-1. Continued 22. Calyx aestivation decussate and/or opposite (0); calyx aestivation quincuncial (1); calyx aestivation imbricate (2) 23. Sepals eight (0); sepals five (1); sepals four (2); sepals two (3) 24. Sepals enclosing petals before anthesis (0); sepals not enclosing petals before anthesis (1) 25. Sepals free (0); sepals fully or partly connate, at least part way (1); outer whorl of sepals connate and inner whorl of sepals free (2) 26. Corolla aestivation imbricate (0); corolla aestivation contorted (1); corolla aestivation decussate (2) 27. Petals five (0); petals four (1); petals two (2) 28. Each petal bilaterally symmetric (0); each petal asymmetric (1) 29. Stamens many (0); stamens fifteen (1) 30. Stamens not in fascicles (0); stamens in fascicles (1) 31. Stamens monomorphic (0); stamens with Clusiella-type dimorphism (1); stamens with Endodesmia-type dimorphism (2) 32. Filaments free (0); filaments connate only at base (1); filaments monadelphous, Clusiella-type (2); filaments monadelphous, Endodesmia-type (3) 33. Anthers less than or equal to three mm long (0); anthers four to five mm long (1); anthers greater than or equal to six mm long (2) 34. Anther glands absent (0); anther glands present (1) 35. Anther glands spherical to elongate (0); anther glands bowl-shaped (1) 36. Anther dehiscence by slits (0); anther dehiscence by terminal pores (1) 37. Carpels four to eight (0); carpels one (1); carpels two (2); carpels three (3); carpels four (4); carpels ten to thirty (5) 38. Placentation axile or intruded-parietal (0); placentation non-intruded-parietal (1); placentation laminar, with ovules scattered on the partitions (2); placentation basal (3); placentation apical (4) 39. U-shaped structure absent in ovary (0); U-shaped structure emerging from base of ovary forming a pseudopartition (1) 40. Placentation not intruded-axile bordered by mesocarp tissue that has curled in (0); placentation intruded-axile bordered by mesocarp tissue that has curled in (1) 41. Ovules one to eight per carpel (0); ovules fifteen or more per carpel (1); ovule one per gynoecium (2) 42. Ovule position at anthesis +/median (0); ovule position basal at anthesis (1); ovule position apical at anthsis (2) 14

PAGE 26

Table 2-1. Continued 43. Style length to ovary length ratio less than or equal to four (0); style length to ovary length ratio greater than or equal to six (1) 44. Styles +/free (0); styles +/fused (1) 45. Stigma shape expanded (0); stigma shape narrow (1) 46. Calyx persistent (0); calyx not persistent (1) 47. Calyx not accrescent (0); calyx accrescent (1) 48. Pedicel not swollen in fruit (0); pedicel swollen in fruit (1) 49. Endocarp not bony (0); endocarp bony (1) 50. Fruit dehiscence septicidal (0); fruit dehiscence septifragal (1); fruit indehiscent (2) 51. Seeds not winged (0); seeds with wing several cells thick, going completely around seed, with no vascular tissue (1); seeds with wing two cells thick, going completely around seed, with no vascular tissue (2); seeds with elongate wing, several cells thick, not going completely around seed, with a peripheral vascular bundle (3) 52. Seeds not plumose (0); seeds plumose (1) 53. Seeds not arillate (0); seeds arillate (1) 54. Testa with unbranched or braided raphal bundle (0); testa with bundles throughout (1) 55. Exotegmen present (0); extotegmen absent (1) 56. Endosperm present in ripe seeds (0); endosperm absent in ripe seeds (1) 57. Embryo less than four mm long (0); embryo more than four mm long (1) 58. Cotyledon not cordate in shape (0); cotyledon cordate in shape (1) 59. Abaxial phloem not partitioned by fibers/lignified cells (0); abaxial phloem partitioned by fibers/lignified cells (1); abaxial phloem doubly partitioned by fibers/lignified cells (2) 60. Group of fibers adaxial to midrib bundle absent (0); patch of fibers adaxial to midrib bundle present (1) 61. Fibers adjacent to abaxial epidermis absent (0); fibers adjacent to abaxial epidermins present, scattered only (1); fibers adjacent to abaxial epidermis present, in continuous band (2) 62. Fiber strand not associated with adaxial side of abaxial-most xylem (0); fiber strand associated with adaxial side of abaxial-most xylem (1) 15

PAGE 27

Table 2-1. Continued 63. Lateral bundles not transcurrent (0); lateral bundles transcurrent (1) 64. Canals not immediately adaxially and abaxially positioned in relation to secondary veins (0); canals immediately adaxially and abaxially positioned in relation to secondary veins (1) 65. Marginal lignification absent (0); marginal lignification present (1) 66. Marginal canal present (0); marginal canal absent (1) 67. One layer of xylem in midrib (0); two layers of xylem in midrib (1); three to six layers of xylem in midrib (2); midrib xylem in bundles (3) 68. Lignified spongy mesophyll absent (0); lignified spongy mesophyll present (1) 69. Leaf blade without abaxial palisade mesophyll (0); leaf blade with abaxial palisade mesophyll 70. Hypodermis absent (0); hypodermis present (1) 71. Epidermis not lignified (0); epidermis lignified (1) 72. Abaxial epidermis not papillose (0); abaxial epidermis papillose (1) 73. Druse crystals in petiole absent (0); druse crystals in petiole present (1) 74. Petiole bundle arched to circular, solid or fragmented (0); petiole bundle with three layers (1); petiole bundle an arch with dorsal groupings of xylem and phloem (2) 16 Characters 54-58 are taken from Stevens (unpubl.). State zero (0) is used to represent the state found in the taxa, Clusia lanceolata and Hypericum tetrapetalum. When outgroups differed, the state was arbitrarily assigned to one of the outgroup genera.

PAGE 28

17 Phylogenetic Analyses Maximum parsimony, Bayesian inference, and maximum likelihood were used to estimate phylogeny. Maximum parsimony analyses were conducted using PAUP* 4.0 b10 (Swofford, 2000), with all characters weighted equally and gaps treated as missing data. Analyses were conducted using heuristic searches with 100 random addition replicates, TBR (tree-bisection-reconnection) branch swapping, and the MulTrees option in effect. Support for clades was estimated using 500 bootstrap replicates with 10 random addition replicates, and TBR branch swapping; 1000 trees were saved per replicate (with MulTrees option in effect). MrBayes 3.0 (Huelsenbeck, 2000) was used for Bayesian inference of phylogeny. The GTR+I+ substitution model was employed in each data set (rbcL, matK, ITS, rbcL + matK + ITS unpruned, and rbcL + matK + ITS pruned) based on the results of a likelihood ratio test done for each data set using Modeltest 3.06 (Posada and Crandall, 1998). The GTR+I+ model consists of separate, time-reversible parameters for all of the possible base substitutions (GTR; Yang, 1994), and two parameters to account for substitution rate heterogeneity across sites, a fixed proportion of invariant sites (I; Hasegawa et al., 1985), with the sites free to vary fitted to a discrete approximation (four categories) of a gamma distribution (; Yang, 1994). Each Bayesian analysis was run with four chains of two million generations. Results were printed to the screen every 100 generations and one tree was saved to the file every 10 generations. Log likelihood values were graphed in Microsoft Excel and burn-in was determined when likelihoods became stationary. Posterior probabilities of clades were calculated from the frequency at which each clade appeared among the trees visited by creating a majority-rule

PAGE 29

18 consensus (excluding trees produced during burn-in phase) using PAUP* 4.0 b10 (Swofford, 2000) Maximum likelihood analyses were completed using PAUP* 4.0 b10 (Swofford, 2000). For each molecular data set, base frequencies were calculated and a model of substitution was determined by Modeltest 3.06 (Posada and Crandall, 1998). For each data set (rbcL, matK, ITS, rbcL + matK + ITS unpruned, and rbcL + matK + ITS pruned), it was determined that a GTR+I+ model of substitution best fit the data. Analyses were conducted using heuristic searches with 100 random addition replicates, TBR (tree-bisection-reconnection) branch swapping, and the MulTrees option in effect. Support for clades was estimated using 100 bootstrap replicates, with starting trees obtained by neighbor joining. Bootstrap analyses were performed on all molecular separate and combined data sets except for rbcL of Clusiaceae and because of the extremely long computational time to complete this analysis. Preliminary analyses of each separate data set (using maximum parsimony for molecular and morphological data sets, and Bayesian inference and maximum likelihood for the molecular data sets) showed they differed only in their amount of resolution. Several analyses were performed, each with its own objectives. Analysis 1 included rbcL sequences from throughout Clusiaceae taken from a previous study (Gustafsson et al., 2002), and also included several new sequences from Kielmeyeroideae (i.e., Calophyllum goniocarpum, Calophyllum inophyllum, Calophyllum leleanii, Calophyllum soulattri, Calophyllum vexans, Caraipa densifolia, Caraipa savannarum, Caraipa utilis, Caraipa valioli, Clusiella isthmensis, Endodesmia calophylloides, Haploclathra cordata, Kielmeyera petiolaris, Kielmeyera rosea, Mammea sessiliflora, Mammea siamensis,

PAGE 30

19 Mammea usambarensis, Marila laxiflora, Marila tomentosa, Mesua ferrea, and Poeciloneuron indicum). Bonnetia roraimae (Bonnetiaceae), Viola sororia (Violaceae), and Chrysobalanus icaco (Chrysobalanaceae), representing three different families of Malpighiales, were used as outgroups in this analysis based on the multi-gene angiosperm topology by Soltis et al. (2000) and the rbcL phylogeny of Gustafsson et al. (2002). The rbcL family-level analysis was done to determine the phylogenetic placement of Endodesmia, a genus not included in the study of Gustafsson et al. (2002). Ascertaining the placement of Endodesmia (see Results) then allowed its use as a functional outgroup (Watrous and Wheeler, 1981) in subsequent analyses. The rbcL analysis was also done to test the possible sister-group relationship of Mammea and the rest of Calophylleae. This relationship appeared in the strict consensus of the analysis of Gustafsson et al. (2002), but did not have bootstrap support. Analysis 2 combined rbcL and matK sequences, which increased the number of variable sites. This analysis included members of Kielmeyeroideae and was performed in an effort to resolve major clades within this subfamily. In Analysis 2, two members of Clusioideae, Clusia lanceolata and Garcinia spicata, served as outgroups. (DNA was isolated from five different species of Hypericum and one species of Triadenum in an attempt to include an outgroup from Hypericoideae, but none of the nine different primer combinations tried resulted in the successful amplification of the matK region.) Analysis 3 used only ITS sequence data. Due to the difficulty in aligning some portions of ITS among phylogenetically distant genera, the ITS analysis was performed primarily to resolve relationships between phylogenetically close genera. Endodesmia, based on its position in analyses 1 and 2 (see Results), was used as a functional outgroup

PAGE 31

20 in this analysis because of alignment problems with sequences from taxa outside of Kielmeyeroideae. Analysis 4 combined all three gene regions to maximize character evidence. This combined molecular data matrix included taxa that lacked sequences from one or more of the three gene regions employed because of their inability to amplify. Of 37 taxa included in this multigene analysis, 16 were missing sequence data from one or more of the gene regions (ca. 22% missing data). Clusia lanceolata was used as the outgroup in this analysis. To minimize the effect of missing data, Analysis 5 combined ITS, matK, and rbcL sequences, but included only those taxa for which all three sequences were available. This pruning resulted in a data set of 21 taxa. Endodesmia was used as a functional outgroup. (Clusia lanceolata was not used as an outgroup in this pruned analysis because it was missing sequence data for ITS as a result of the problematic alignment of sequences from outside Kielmeyeroideae.) Analysis 6 included only morphological characters for genera of Kielmeyeroideae. Representatives of the other two subfamilies, Clusia lanceolata and Hypericum tetrapetalum, were used as outgroups. Analysis 7 combined the 74 morphological characters with all three molecular data sets (Analysis 4) and included all taxa except Neotatea, for which molecular data were lacking. Clusia lanceolata was used as the outgroup in this analysis. Analysis 8 combined the morphological characters with the molecular characters, but only included those taxa common to all four data sets. Endodesmia was used as a functional outgroup based on its position in analyses 1 and 2 (see results).

PAGE 32

21 Additional analyses were performed that included Neotatea, for which only morphological characters were used. In Analysis 9, Neotatea was added to the unpruned total evidence data set (DNA + morphology), and Analysis 10 added Neotatea to the pruned total evidence data set. Using MacClade 4.05 (Maddison and Maddison, 2001), morphological character-state transformations were mapped onto one of the most parsimonious trees from the unpruned total evidence analysis (Analysis 7), following the approach of Kron et al. (2002). In the case of equivocal tracing of characters, MacClade was set to show all most parsimonious states at each node. In addition, morphological characters were mapped onto a most parsimonious tree from the pruned total evidence analysis that included Neotatea (Analysis 10). In the unpruned total evidence analysis (Analysis 9), the position of Neotatea was unresolved.

PAGE 33

22 Table 2-2. Morphological character-state coding used in phylogenetic analysis of Kielmeyeroideae (Clusiaceae) 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 Clusia lanceolata 0 0 0 0 0 0 0 0 0 A 0 0 0 0 0 0 0 0 Hypericum tetrapetalum 0 0 1 0 0 0 1 0 1 ? 0 1 1 0 0 0 0 0 Endodesmia calophylloides 1 ? 1 0 0 0 0 1 1 0 1 1 0 0 0 0 0 0 Mammea subsessifolia 0 1 1 0 0 0 0 0 1 A 1 1 1 0 1 0 5 1 Mammea siamensis 0 1 1 0 0 0 0 0 1 A 1 1 1 0 1 0 5 1 Mammea americana 0 1 1 0 0 0 0 0 1 A 1 1 1 0 1 0 5 1 Neotatea duidae 0 0 1 1 0 1 1 0 1 1 0 0 0 1 0 0 3 0 Neotatea longifolia 0 0 1 1 0 1 1 0 1 1 0 0 0 1 0 0 3 0 Neotatea neblinae 0 0 1 1 0 1 1 0 1 1 0 0 0 1 0 0 3 0 Marila racemosa 0 0 1 0 1 0 0 0 1 1 2 1 1 0 1 0 4 1 Marila laxiflora 0 0 1 1 0 0 1 0 1 1 2 1 1 0 1 0 4 1 Marila plumbaginea 0 0 1 1 0 0 0 0 1 1 2 1 1 0 1 0 4 1 Mahurea exstipulata 0 1 1 1 0 1 0 1 1 0 2 1 1 0 0 0 1 0 Kielmeyera speciosa 0 1 1 0 1 1 0 0 1 1 1 1 1 0 0 0 1 0 Kielmeyera coriacea 0 1 1 1 1 1 0 0 1 1 1 1 1 0 0 0 1 0 Caraipa densifolia 0 0 1 0 1 1 0 0 1 A B 1 1 0 0 0 1 0 Caraipa sp 0 0 1 0 1 1 0 0 1 A B 1 1 0 0 0 1 0 Caraipa savanarum 0 0 1 0 1 1 0 0 1 1 1 1 1 0 0 0 1 0 Haploclathra paniculata 0 0 1 0 1 0 0 0 1 A B 1 1 0 0 0 1 0 Haploclathra leiantha 0 0 1 0 1 0 0 0 1 A 1 1 1 0 0 0 1 0 Clusiella isthmensis 0 1 1 0 0 0 0 2 1 1 1 1 0 0 0 1 2 0 Clusiella axillaris 0 1 1 0 0 0 0 2 1 0 1 1 0 0 0 1 2 0 Kayea kunstleri 0 1 1 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 Kayea elmeri 0 1 1 0 0 0 0 0 1 A 1 1 1 0 0 0 0 0 Kayea borneensis 0 1 1 0 0 0 0 0 1 A 1 1 1 0 0 0 1 0 Mesua ferrea 1 ? 2 0 0 0 1 0 1 1 1 1 1 0 1 0 0 0 Poeciloneuron indicum 0 1 1 1 0 0 0 1 1 1 1 1 1 0 0 0 1 0 Calophyllum inophyllum 0 0 1 0 1 0 1 0 1 0 0 0 0 1 1 0 1 0 Calophyllum brasiliense 0 0 1 0 1 0 1 0 1 0 0 0 0 1 1 0 1 0 Calophyllum fibrosum 0 0 1 0 1 0 1 0 1 0 0 0 0 1 1 0 1 0

PAGE 34

23 Table 2-2. Continued 1 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 Clusia lanceolata 0 0 0 0 0 0 0 0 1 0 0 0 0 0 ? 0 ? 0 Hypericum tetrapetalum 1 0 1 0 2 0 0 1 1 1 0 ? 0 0 0 0 ? 0 Endodesmia calophylloides 0 0 1 1 1 1 0 1 0 1 0 0 2 3 0 1 0 0 Mammea subsessifolia 0 0 0 0 3 0 1 2 1 0 0 0 0 1 0 0 ? 0 Mammea siamensis 0 0 0 0 3 0 1 2 1 0 0 0 0 1 0 1 0 0 Mammea americana 0 0 0 0 3 0 1 2 1 0 0 0 0 0 0 1 0 0 Neotatea duidae 1 0 1 2 1 0 0 1 0 0 0 0 0 0 ? 1 0 0 Neotatea longifolia 1 0 1 2 1 0 0 1 0 0 0 0 0 0 2 1 0 0 Neotatea neblinae 1 0 1 2 1 0 0 1 0 0 0 0 0 0 2 1 0 0 Marila racemosa 0 0 1 1 1 0 1 A 0 0 0 0 0 1 0 1 1 0 Marila laxiflora 0 0 1 1 1 0 1 0 0 0 0 0 0 1 0 1 0 0 Marila plumbaginea 0 0 1 1 1 0 1 0 0 0 0 0 0 1 0 1 0 0 Mahurea exstipulata 0 0 1 1 1 A 0 1 0 0 0 0 0 0 0 1 1 0 Kielmeyera speciosa 0 0 1 1 1 1 0 1 0 0 0 0 0 0 0 0 ? 0 Kielmeyera coriacea 0 0 1 1 1 1 0 1 0 0 0 0 0 0 0 1 0 0 Caraipa densifolia 0 0 1 1 1 1 1 1 0 0 0 0 0 0 0 1 1 0 Caraipa sp 0 0 1 1 1 1 0 1 0 0 0 0 0 1 0 1 1 0 Caraipa savanarum 0 0 1 1 1 1 1 1 0 0 0 0 0 1 0 1 1 0 Haploclathra paniculata 0 0 1 1 1 1 0 1 0 0 0 0 0 1 1 0 ? 0 Haploclathra leiantha 0 0 1 1 1 1 0 1 0 0 0 0 0 1 0 0 ? 0 Clusiella isthmensis 0 1 0 2 1 1 0 1 0 0 0 0 1 2 0 1 0 0 Clusiella axillaris 0 1 0 2 1 1 0 1 0 0 0 0 1 2 0 1 0 0 Kayea kunstleri 0 0 1 0 2 0 2 0 1 0 0 0 0 1 0 0 ? 0 Kayea elmeri 0 0 1 0 2 0 2 0 1 0 0 0 0 1 0 1 0 0 Kayea borneensis 0 0 1 0 2 0 2 0 1 0 0 0 0 0 0 1 0 0 Mesua ferrea 1 0 1 0 2 A 0 2 1 0 0 0 0 1 0 0 ? 0 Poeciloneuron indicum 0 0 1 1 1 1 0 1 0 0 1 0 0 0 0 0 ? 1 Calophyllum inophyllum 1 0 1 0 3 0 0 1 1 0 0 1 0 1 0 0 ? 0 Calophyllum brasiliense 1 0 1 0 3 1 ? 2 2 0 0 1 0 1 0 0 ? 0 Calophyllum fibrosum 1 0 1 0 2 A 0 2 1 0 0 0 0 1 0 0 ? 0

PAGE 35

24 Table 2-2. Continued 3 3 3 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 Clusia lanceolata 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 Hypericum tetrapetalum 3 1 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 Endodesmia calophylloides 1 4 0 0 0 2 0 ? 1 0 0 1 0 2 0 0 0 1 Mammea subsessifolia 2 3 0 0 0 1 0 1 0 1 0 0 0 2 0 0 0 1 Mammea siamensis 2 3 0 0 0 1 0 1 0 1 0 0 0 2 0 0 0 1 Mammea americana 2 3 0 0 0 1 0 1 0 1 0 0 0 2 0 0 0 1 Neotatea duidae 3 0 0 1 1 0 ? 1 0 0 0 0 0 0 3 0 0 0 Neotatea longifolia 3 0 0 1 1 0 0 1 0 0 0 0 0 0 3 0 0 0 Neotatea neblinae 3 0 0 1 1 0 0 1 0 0 0 0 0 0 3 0 0 0 Marila racemosa C 0 0 0 1 0 0 1 0 1 0 0 0 0 0 1 0 0 Marila laxiflora 3 0 0 0 1 0 0 1 0 0 0 0 0 0 0 1 0 0 Marila plumbaginea 3 0 0 0 1 0 0 1 0 0 0 0 0 0 0 1 0 0 Mahurea exstipulata 3 0 0 1 1 0 0 1 0 0 0 0 0 0 3 0 0 0 Kielmeyera speciosa 3 0 0 0 1 0 0 1 0 1 0 0 0 0 2 0 0 0 Kielmeyera coriacea 3 0 0 0 1 0 0 1 0 1 0 0 0 0 2 0 0 0 Caraipa densifolia 3 0 0 0 0 0 0 1 0 0 0 0 0 1 1 0 0 1 Caraipa sp 3 0 0 0 0 0 0 1 0 0 0 0 0 1 1 0 0 1 Caraipa savanarum 3 0 0 0 0 0 0 1 0 0 0 0 0 1 1 0 0 1 Haploclathra paniculata 3 0 0 0 0 1 0 1 0 0 0 0 0 1 1 0 0 1 Haploclathra leiantha 3 0 0 0 0 1 0 1 0 0 0 0 0 1 1 0 0 1 Clusiella isthmensis 5 2 0 0 1 0 0 ? 0 0 0 0 0 2 0 0 0 0 Clusiella axillaris 5 2 0 0 1 0 0 ? 0 0 0 0 0 2 0 0 0 0 Kayea kunstleri 4 3 0 0 0 1 1 1 1 0 1 0 0 0 0 0 0 1 Kayea elmeri 4 3 0 0 0 1 0 1 1 0 0 0 0 2 0 0 0 1 Kayea borneensis 4 3 0 0 0 1 1 1 1 0 1 0 0 ? 0 0 0 1 Mesua ferrea 2 3 1 0 0 1 0 1 0 0 0 0 0 1 0 0 0 1 Poeciloneuron indicum 2 3 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1 Calophyllum inophyllum 3 3 0 0 2 1 0 1 0 1 0 0 1 2 0 0 0 1 Calophyllum brasiliense 2 3 0 0 2 1 0 1 0 1 0 0 1 2 0 0 0 1 Calophyllum fibrosum 2 3 0 0 2 1 0 1 0 1 0 0 1 2 0 0 0 1

PAGE 36

25 Table 2-2. Continued 5 5 5 5 5 6 6 6 6 6 6 6 6 6 6 7 7 7 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 Clusia lanceolata 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 1 0 0 Hypericum tetrapetalum 0 0 1 0 0 0 0 ? 0 0 0 1 0 0 1 0 0 0 Endodesmia calophylloides 1 1 0 ? 0 0 0 ? 0 0 0 0 0 0 0 0 0 0 Mammea subsessifolia 1 1 0 0 0 0 1 ? 1 0 1 0 0 0 0 0 0 0 Mammea siamensis 1 1 0 0 0 0 0 ? 1 0 1 0 0 0 0 0 0 0 Mammea americana 1 1 0 0 0 0 0 ? 1 0 1 0 0 0 0 0 0 0 Neotatea duidae 0 0 1 0 2 0 2 ? A 1 1 0 0 0 0 1 0 1 Neotatea longifolia 0 0 1 0 2 0 0 ? 0 1 1 0 0 0 0 1 0 0 Neotatea neblinae 0 0 1 0 2 0 0 ? 0 1 1 0 0 0 0 1 0 1 Marila racemosa 0 0 1 0 0 0 1 0 1 0 0 0 1 0 0 0 0 0 Marila laxiflora 0 0 1 0 0 0 1 0 1 0 1 0 2 0 0 0 1 0 Marila plumbaginea 0 0 1 0 0 0 0 0 1 0 1 0 2 0 0 0 0 0 Mahurea exstipulata 0 0 1 0 0 0 0 0 1 0 1 0 2 0 0 0 0 0 Kielmeyera speciosa 1 1 0 1 0 0 0 0 1 0 0 0 1 0 0 0 0 0 Kielmeyera coriacea 1 1 0 1 0 0 1 0 1 1 1 0 1 0 0 0 0 0 Caraipa densifolia 1 1 0 1 0 1 0 0 1 0 1 0 1 0 0 0 0 0 Caraipa sp 1 1 0 1 0 1 0 0 1 0 1 0 1 0 0 0 0 1 Caraipa savanarum 1 1 0 1 0 1 0 0 1 0 1 0 2 0 0 0 0 0 Haploclathra paniculata 1 1 0 1 1 1 2 0 1 1 1 0 B 0 0 0 0 1 Haploclathra leiantha 1 1 0 1 1 1 2 0 0 0 1 0 1 0 0 0 0 1 Clusiella isthmensis 0 0 1 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 Clusiella axillaris 0 0 1 0 0 0 0 1 0 0 0 1 2 0 0 0 0 0 Kayea kunstleri 1 1 0 ? 1 1 2 1 1 0 1 0 1 0 0 0 0 0 Kayea elmeri 1 1 0 ? 1 1 2 1 1 0 0 0 1 0 0 0 0 0 Kayea borneensis 1 1 0 ? 0 1 2 1 1 0 1 0 2 0 0 0 0 0 Mesua ferrea 1 1 0 ? 0 0 2 0 1 0 1 0 1 0 0 0 0 0 Poeciloneuron indicum 1 1 0 ? 1 1 2 0 1 0 1 0 1 0 0 0 0 0 Calophyllum inophyllum 1 1 0 0 0 1 0 ? 1 0 1 0 0 0 0 0 0 0 Calophyllum brasiliense 1 1 0 0 1 1 0 ? 1 0 1 0 0 0 0 0 0 0 Calophyllum fibrosum 1 1 0 0 1 1 0 ? 1 0 1 0 0 1 0 0 0 0

PAGE 37

26 Table 2-2. Continued 7 7 3 4 Clusia lanceolata 0 0 Hypericum tetrapetalum ? ? Endodesmia calophylloides 0 0 Mammea subsessifolia 1 0 Mammea siamensis 1 0 Mammea americana 1 0 Neotatea duidae 0 ? Neotatea longifolia 0 0 Neotatea neblinae 0 0 Marila racemosa 1 1 Marila laxiflora 1 1 Marila plumbaginea 1 1 Mahurea exstipulata 0 1 Kielmeyera speciosa 0 0 Kielmeyera coriacea 0 0 Caraipa densifolia 1 0 Caraipa sp 0 0 Caraipa savanarum 0 0 Haploclathra paniculata 0 2 Haploclathra leiantha 0 2 Clusiella isthmensis 0 0 Clusiella axillaris 1 0 Kayea kunstleri 1 0 Kayea elmeri 1 0 Kayea borneensis 1 0 Mesua ferrea 1 0 Poeciloneuron indicum 0 0 Calophyllum inophyllum 1 0 Calophyllum brasiliense 1 0 Calophyllum fibrosum 1 0 State is used to represent the condition found in outgroup taxa, Clusia lanceolata and Hypericum tetrapetalum. When outgroups differed, the state was arbitrarily assigned to one of the outgroup genera. A = 0&1; B = 1&2; C = 3&4.

PAGE 38

27 Figure 2-1. Stipuliform structures. A) Endodesmia calophylloides. Paired stipuliform structures attached on the stem adjacent to the region where the basal part of the petiole attaches to the stem. B) Mahurea exstipulata. Stipuliform structure attached on stem just above point of petiole attachment, adjacent to axillary bud. C) Clusiella isthmensis. Stipuliform structure interpetiolar.

PAGE 39

28 Figure 2-2. Overview of leaf of Endodesmia calophylloides showing mesophyll with latex/resin cavities and canals. A similar pattern is found in Clusiella spp.

PAGE 40

29 Figure 2-3. Inflorescence architectures of Kielmeyeroideae. A) Neotatea-type cyme. B) Clusiella-type cyme. C) Cyme, as in Endodesmia calophylloides and Kayea elmeri. D) Reducded cyme, as in Kayea kunstleri. E) Single-flowered cyme, as in Mesua ferrea. F-G) Panicle-like cymes, as in Calophyllum spp., Poeciloneuron indicum, Kayea borneensis, Kielmeyera spp., Haploclathra spp., and Caraipa spp. H) Fascicle, as in Mammea spp. I) Raceme, as in Marila spp.

PAGE 41

30 Figure 2-4. Androecial features of Kielmeyeroideae. A-C) Endodesmia calophylloides. A) Flower with petals removed. B) Longitudinal section of stamen tube, showing inside surface covered by anthers. C) Two types of anthers, with lines representing resin like substance; nonfertile anther on left and fertile anther on right. D) Stamen of Kayea borneensis, showing spherical anther gland at apex of C-shaped theca. E) Stamen of Marila laxiflora, showing elongate anther gland at apex. F) Stamen of Mahurea exstipulata, illustrating bowl-shaped (crateriform) anther gland.

PAGE 42

31 Figure 2-5. Androecium of Clusiella axillaris. A) Staminate flower bud with sepals and petals removed. Resin-secreting staminodes surround a monadelphous tube of functional stamens. B) Diagram of longitudinal section through A. C) Close up of functional anther showing glandular connective and theca.

PAGE 43

32 Figure 2-6. Gynoecial features of Kielmeyeroideae. A-D) Placentation types. A) Ovary cross section of Mahurea exstipulata, showing axile placentation. B) Ovary cross section of Kielmeyera speciosa (hairs on ovary not shown), showing intruded parietal placentation (and falsely septate ovary). C) Ovary cross section of Hypericum tetrapetalum, showing parietal placentation and nonseptate ovary. D) Ovary cross section of Clusiella axillaris, showing laminar placentation. E) Diagram of young fruit cross section of Mahurea exstipulata and Neotatea spp. The intruded placentae are bordered by the carpel walls (septae) that have curled in.

PAGE 44

33 Figure 2-7. Diagrammatic longitudinal sections of ovaries showing ovule position at anthesis. A) Median ovule position. B) Basal ovule position. C) Apical ovule position.

PAGE 45

34 Figure 2-8. Three types of winged seeds found in Kielmeyeroideae. A) Caraipa densifolia. The wing is more than two cell layers thick, completely surrounds the seed, and contains no vascular tissue. B) Kielmeyera coriacea. The wing is two cell layers thick, completely surrounds the seed, and contains no vascular tissue. C) Mahurea exstipulata. The wing is more than two cell layers thick, does not completely surround the seed, and contains vascular tissue.

PAGE 46

35 Figure 2-9. Leaf blade cross sections illustrating anatomical characters used in phylogenetic analyses. Xylem is represented by white with closely spaced, narrow, black lines. Phloem is represented by even stipling. Fibers are shown in black. Latex/resin canals are shown as circles, while latex/resin cavities are shown as circles with uneven stipling. A) Neotatea neblinae. Canals are present immediately adaxially and abaxially positioned in relation to secondary veins (chr. 64:1). Lateral vascular bundles are included (chr. 63:0). One layer of midrib xylem is present (chr. 67:0). Phloem is doubly partitioned by fibers (chr. 59:2). Hypodermis is present (chr. 70:1). Abaxial epidermis is papillose (chr. 72:1). B) Haploclathra paniculata. Cavities are present in mesophyll (chr. 12:1). Lateral vascular bundles are transcurrent (chr. 63:1). Group of fibers adaxial to midrib bundle is present (chr. 60:1). Two layers of midrib xylem are present (chr. 67:1). Phloem is partitioned by fibers (chr. 59:1). Fibers are adjacent to abaxial epidermis, in a continuous band (chr. 61:2). Abaxial epidermis is papillose (chr. 72:1).

PAGE 47

36 Figure 2-10. Leaf blade cross sections illustrating anatomical characters used in phylogenetic analyses. Xylem is represented by white with closely spaced, narrow, black lines. Phloem is represented by even stipling. Fibers are shown in black. Dark stipling represents less strongly lignified fibers. Latex/resin canals are shown as circles, while latex/resin cavities are shown as circles with uneven stipling. A) Marila laxiflora. Cavities are present in mesophyll (chr. 12:1). Lateral vascular bundles are transcurrent (chr. 63:1). Six layers of midrib xylem are present (chr. 67:2). Phloem is not partitioned by fibers (chr. 59:0). Fibers are adjacent to abaxial epidermis, in a continuous band (chr. 61:2). Adaxial epidermis is lignified (chr. 71:1). B) Kayea borneensis. Cavities are present in mesophyll (chr. 12:1). Lateral vascular bundles are transcurrent (chr. 63:1). Group of fibers adaxial to midrib bundle is present (chr. 60:1). Three layers of midrib xylem are present (chr. 67:2). Fiber strand is associated with adaxial side of abaxial-most xylem (chr.62:1). Fibers are adjacent to abaxial epidermis, in a continuous band (chr. 61:2).

PAGE 48

CHAPTER 3 RESULTS Analysis 1: rbcL of Clusiaceae Of the 1408 base pairs sequenced in rbcL, 410 are variable, and 240 varied in a parsimony-informative manner. The analysis resulted in 1530 most parsimonious trees of length 845 (CI = 0.612, RI = 0.755, RC = 0.462). The rbcL data support the monophyly of three major clades: Clusioideae, Hypericoideae + Podostemaceae, and Kielmeyeroideae (Figure 3-1). Kielmeyeroideae, including Endodesmia calophylloides, are supported with a bootstrap value of 75%. Endodesmia calophylloides, of Endodesmieae, is placed as the sister group to the rest of the subfamily, the Calophylleae, which are strongly supported as a clade (bootstrap support = 95%). In the strict consensus, Mammea appears as the sister group to the rest of Calophylleae, but this relationship does not receive bootstrap support >50%. Appearing in the strict consensus tree, but without bootstrap support, is a clade containing Calophyllum + Mesua, which is a sister to the Kayea + Poeciloneuron clade. The Bayesian analysis of rbcL (Figure 3-2) also shows support for three major clades within Clusiaceae: Hypericoideae + Podostemaceae, Clusioideae, and Kielmeyeroideae, with posterior probabilities of 100, 100, and 98, respectively. Within Kielmeyeroideae, Endodesmia appears sister to Calophylleae with a posterior probability of 100 and Mammea appears sister to the remaining Calophylleae with a posterior probability of 71. 37

PAGE 49

38 The maximum likelihood analysis of rbcL also shows the same three major clades within Clusiaceae: Hypericoideae + Podostemaceae, Clusioideae, and Kielmeyeroideae. Two trees were obtained, with likelihood scores of 6792.82386 and 6792.82649. The topologies are identical to that of the Bayesian analysis except in their placement of Haploclathra. In one of the maximum likelihood trees, Haploclathra appears sister to a clade of Mahurea + Clusiella + Marila, and in the other tree (Figure 3-3), Haploclathra appears in a clade with Kielmeyera and Caraipa. Analysis 2: rbcL + matK of Kielmeyeroideae Combining rbcL and matK resulted in a total of 1885 characters; 120 of the 279 variable characters were parsimony-informative. The analysis recovered 510 most parsimonious trees of length 361 (CI = 0.831, RI = 0.762, RC = 0.633). Endodesmia appears as sister to Calophylleae with 100% bootstrap support, and Mammea is placed as sister to the rest of Calophylleae with 78% bootstrap support (Figure 3-4). As in the analysis of rbcL alone, a clade containing Calophyllum + Mesua is sister to Kayea + Poeciloneuron in all shortest trees, but these relationships again do not receive bootstrap support >50%. Mesua appears sister to Calophyllum with 86% bootstrap support. Caraipa, Kielmeyera, and Haploclathra form a monophyletic group with 71% bootstrap support. Bayesian inference resulted in a topology identical to that of parsimony, and therefore the tree is not shown here. Kielmeyeroideae and Calophylleae appear monophyletic, each with posterior probabilities of 100. Mammea is placed sister to the rest of Calophylleae with a posterior probability of 92. Caraipa, Kielmeyera, and Haploclathra form a clade with a posterior probability of 100.

PAGE 50

39 The maximum likelihood analysis resulted in a tree with a score of 4912.90425 and a similar topology to that of the parsimony and Bayesian analyses; however, in the maximum likelihood topology, Marila appears sister to Kayea + Mahurea, although without bootstrap support (Figure 3-5). Analysis 3: ITS of Kielmeyeroideae The ITS region provided 809 characters, 383 of which are variable; 232 were parsimony-informative. Ten most parsimonious trees were recovered of length 987 (CI = 0.584, RI = 0.658, RC = 0.384). Unlike analyses 1 and 2 (rbcL and rbcL + matK, respectively), ITS shows variation at lower taxonomic levels; however, the ITS analysis shows no bootstrap support for branches along the spine of the tree. All genera appear monophyletic, and most are well supported (Figure 3-6). Caraipa + Haploclathra + Kielmeyera form a clade with 86% bootstrap support. Poeciloneuron and Kayea are supported only weakly (bootstrap support = 60%). The topology resulting from Bayesian analysis of ITS sequences (results not shown) is similar to that of parsimony, except that in the Bayesian topology, Mammea appears in a weakly supported clade (posterior probability of 64) with Kayea and Poeciloneuron. All genera are supported by posterior probabilities of 100. Caraipa, Haploclathra, and Kielmeyera are supported as a clade with a posterior probability of 100, within which Caraipa and Haploclathra are sisters with a posterior probability of 93. Maximum likelihood analysis of the ITS data set resulted in a tree with a score of 5734.93593. The topology is similar to that of the Bayesian analysis: Mammea appears sister to a clade of Kayea + Poeciloneuron, although without bootstrap support (Figure 3-7).

PAGE 51

40 Analyses 4 and 5: rbcL + matK + ITS of Kielmeyeroideae Combining all three gene regions resulted in 2696 characters; 311 of 645 variable characters were parsimony-informative in the unpruned analysis, while 251 of 534 variable characters were parsimony-informative in the pruned analysis. The combined analysis that included all taxa (Analysis 4, rbcL + matK + ITS, unpruned; Figure 3-5) produced two most parsimonious trees of length 1327 (CI = 0.643, RI = 0.665, RC = 0.427). All genera are monophyletic and are supported by higher bootstrap values than in the previous analyses (Table 3-1). Calophylleae are monophyletic with 100% bootstrap support. Caraipa, Haploclathra, and Kielmeyera form a clade with 95% bootstrap support, Mesua is sister to Calophyllum with 84% support, and Kayea and Poeciloneuron are sisters with 72% bootstrap support. While the strict consensus shows resolution along the spine of the tree, these relationships do not receive bootstrap support >50%. The strict consensus places Mammea sister to the rest of Calophylleae, within which a clade containing Mahurea, Clusiella, and Marila is sister to a clade of Caraipa, Haploclathra, and Kielmeyera + Kayea, Poeciloneuron, Mesua, and Calophyllum (Figure 3-8). The topology and levels of support in the pruned combined analysis (Analysis 5) are nearly identical to those of the unpruned combined analysis (Analysis 4), and results are therefore not shown here. The only topological difference between analyses 4 and 5 is that in the unpruned combined analysis (Analysis 4; Figure 3-8), the relationships among Caraipa, Haploclathra, and Kielmeyera are unresolved, whereas in the pruned combined analysis (Analysis 5), Caraipa and Kielmeyera are sister groups with 57% bootstrap support.

PAGE 52

41 Bayesian analyses of the unpruned and pruned combined DNA data sets resulted in similar topologies and levels of support compared with the parsimony analyses of these data sets (Figures 3-9 and 3-10). The maximum likelihood analysis of the unpruned and pruned combined DNA data sets resulted in nearly identical topologies to that of parsimony and therefore results are not shown here. The only difference in the results is in the relationships among Kielmeyera, Caraipa, and Haploclathra. In strict consensus of the parsimony analysis (Figure 3-8), the relationships among these three genera are unresolved; in the maximum likelihood tree, Kielmeyera is sister to a clade of Caraipa + Haploclathra. The likelihood scores of the unpruned and pruned combined DNA analyses were 10,988.21023 and 9321.34755, respectively. Bootstrap support for clades was equivalent to that of the bootstrap of the parsimony analysis. Analysis 6: Morphology of Kielmeyeroideae The morphological analysis was based on 74 variable characters, 64 of which were parsimony-informative. Twenty most parsimonious trees of length 224 were recovered (CI = 0.482, RI = 0.740, RC = 0.357). Kielmeyeroideae are supported as a clade with a bootstrap value of 52% (Figure 3-11). In the strict consensus, all genera except Caraipa appear monophyletic with bootstrap support >50%. Caraipa and Haploclathra form a clade with 58% bootstrap support. Mesua, Calophyllum, Kayea, and Mammea form a clade with 72% bootstrap support. Analyses 7 and 8: DNA + Morphology of Kielmeyeroideae Combining all three gene regions with the morphological data matrix results in a total of 2770 characters. In the unpruned analysis (Analysis 7), 436 of the 962 variable characters were parsimony-informative; in the pruned analysis (Analysis 8), 306 of the

PAGE 53

42 712 variable characters were parsimony-informative. The unpruned combined DNA + morphology analysis (Analysis 7; Figure 3-12) resulted in four most parsimonious trees of length 1991 (CI = 0.671, RI = 0.648, RC = 0.435). The topology is almost identical to that of the combined DNA analyses without morphology (analyses 4 and 5); in some clades (i.e., Kielmeyera, and Kielmeyera + Caraipa + Haploclathra), bootstrap support is slightly lower in the unpruned DNA + morphology combined analysis (Analysis 7), whereas in other clades (i.e., Marila, Marila + Mahurea + Clusiella and Kayea + Poeciloneuron), the bootstrap support is slightly higher when morphology is added (Table 3-1). The only significant difference in support is evident in the Mesua + Calophyllum clade: this clade received 84% bootstrap support in the unpruned and pruned combined DNA analyses, and received 97% bootstrap support in the unpruned DNA + morphology analysis. The pruned combined DNA + morphology analysis (Analysis 8; Figure 3-13) resulted in four most parsimonious trees of length 1297 (CI = 0.705, RI = 0.548, RC = 0.386). The topology of the strict consensus is nearly identical to that of the unpruned analysis (Analysis 7), with only two differences: relationships among Caraipa, Haploclathra, and Kielmeyera are unresolved in the pruned analysis (Analysis 8) whereas Caraipa and Haploclathra are sisters with 56% bootstrap support in the unpruned analysis (Analysis 7). Furthermore, in the strict consensus of the pruned analysis, Kayea and Poeciloneuron do not form a clade; however, in the bootstrap analysis, Kayea and Poeciloneuron are supported as sisters with weak support (bootstrap support = 61%). Analyses 9 and 10: DNA + Morphology of Kielmeryoideae, Including Neotatea Combining all three gene regions with the morphological data matrix, including Neotatea, results in a total of 2770 characters. In the unpruned analysis (Analysis 9), 439

PAGE 54

43 of the 962 variable characters were parsimony-informative; in the pruned analysis (Analysis 10), 310 of the 714 variable characters were parsimony-informative. Including Neotatea in the unpruned combined DNA + morphology analysis (Analysis 9) resulted in 28 most parsimonious trees of length 2009 (CI = 0.666, RI = 0.654, RC = 0.436). The relationships of Neotatea to other genera within Kielmeyeroideae are unresolved in the unpruned analysis (Analysis 9; Figure 3-14). Adding Neotatea to the pruned combined DNA + morphology analysis (Analysis 10; Figure 3-15) resulted in twelve most parsimonious trees of length 1315 (CI = 0.698, RI = 0.569, RC = 0.398). Neotatea appears in a clade with Clusiella, Marila, and Mahurea with 69% bootstrap support. Neotatea is sister to Mahurea, but with only weak support (bootstrap support = 52%). Character Evolution Morphological character-state transformations were mapped onto one of the most parsimonious trees (tree #2) from Analysis 7 (DNA + morphology, unpruned; Figure 3-16). Kielmeyeroideae are distinguished from Clusia lanceolata by 15 characters. An informal survey of Clusioideae and Hypericoideae provides evidence that none of these characters is likely synapomorphic for Kielmeyeroideae. However, having a nonfasciculate androecium might be a synapomorphy for Kielmeyeroideae; although most members of Clusieae do not have fasciculate androecia, most of the remaining members of Clusioideae (Garcinieae and Symphonieae), as well as most Hypericoideae, do have fasciculate androecia. Following is a nonexhaustive list of putative morphological synapomorphies for clades within Kielmeyeroideae. Equivocal, very homoplasious character state transformations (i.e., 0/1 0/1) are not mentioned except where this condition is only an

PAGE 55

44 artifact of missing morphological data from taxa that were included in the molecular data set, but not the morphological data set. Character states changing from equivocal to unequivocal (i.e., 0/1 1) are mentioned. Unequivocal, but homoplasious, character-state transformations (i.e., 0 1) are designated by an asterisk. Unequivocal, unique, character-state transformations (i.e., 0 1) are designated by two asterisks. Refer to Table 2-1 for a description of character states. Endodesmia has the following potential morphological synapomorphies: a nonfunctional terminal bud (chr. 1:1*), presence of stipuliform structures (chr. 8:1*), eucamptodromous secondary venation (chr. 10:0), quincuncial calyx aestivation (chr. 22:1*), sepals not enclosing petals (chr. 24:1), five petals (chr. 27:0*), asymmetrical petals (chr. 28:1**), Endodesmia-type stamen dimorphism (chr. 31:2**), Endodesmia-type monadelphous stamens (chr. 32:3**), one carpel (chr. 37:1), apical placentation (chrs. 38:4** and 42:2**), a narrow stigma (chr. 45:1*), a swollen pedicel (chr. 48:1**), and an indehiscent fruit (chr. 50:2). Potential synapomorphies for Calophylleae include having a terminal bud without scales (chr. 2:1), absence of latex/resin canals in mesophyll (chr. 13:1*), fused styles (chr. 44:1), presence of transcurrent lateral bundles in leaf blade (chr. 63:1*), and marginal lignification of leaf blade (chr. 65:1*). Mammea may be diagnosed by several putative morphological synapomorphies including an axillary inflorescence position (chr. 15:1*), fasciculate inflorescence (chr. 17:5**), absence of terminal flower (chr. 18:1*), being androdioecious (21:0), having two sepals (chr. 23:3*), sepals that enclose the petals before anthesis (chr. 24:0), fused sepals (chr. 25:1*), decussate corolla aestivation (chr. 26:2*), basal placentation (chrs.

PAGE 56

45 38:3 and 42:1), the absence of a persistent calyx (chr. 46:1*), indehiscent fruit (chr. 50:2), and the presence of druse crystals in the petiole (chr. 73:1). Putative synapomorphies for the remaining members of Calophylleae (all genera except Mammea) are: having brochidodromous secondary venation (chr. 10:1), a panicle-like cyme inflorescence (chr. 17:1), being monoecious (chr. 21:1), and two layers of leaf midrib xylem (chr. 67:1*). The clade containing Kayea, Poeciloneuron, Mesua, and Calophyllum has the following potential synapomorphies: basal placentation (chrs. 38:3 and 42:1) and the presence of a group of fibers adaxial to the leaf midrib bundle (60:1). A likely synapomorphy for Kayea and Poeciloneuron is narrow stigmas (chr. 45:1*). Potential synapomorphies for Kayea include sepals that enclose the petals before anthesis (chr. 24:0), the outer whorl of sepals fused and the inner whorl free (chr. 25:2**), imbricate corolla aestivation (chr. 26:0*), four carpels (chr. 37:4**), a style length to ovary length ratio greater than or equal to six (chr. 43:1*), an accrescent calyx (chr. 47:1*), a fiber strand associated with adaxial side of abaxial-most xylem of leaf midrib (chr. 62:1*), and the presence of druse crystals in the petiole (chr. 73:1). Poeciloneuron may be diagnosed by the following putative morphological synapomorphies: an indumentum of unbranched, unicellular hairs (chr. 4:1*), the presence of stipuliform structures (chr. 8:1*), quincuncial calyx aestivation (chr. 22:1), five sepals (chr. 23:1), sepals not enclosing the petals before anthesis (chr. 24:1), five petals (chr. 27:0), 15 stamens (chr. 29:1**), the absence of anther glands (chr. 34:0), anthers with terminal pores (chr. 36:1**), free styles (chr. 44:0*), septicidal capsule (chr. 50:0), abaxial phloem in leaf midrib partitioned by fibers (chr. 59:1), and the absence of druse crystals in the petiole (chr. 73:0). Calophyllum and

PAGE 57

46 Mesua have several potential morphological synapomorphies, including the absence of colleters (chr. 7:1*), axillary inflorescence position (chr. 15:1*), the absence of bracteoles (chr. 19:1*), opposite or decussate corolla aestivation (chr. 26:2*), filaments fused at base (chr. 32:1), and the absence of anther glands (chr. 34:0). Mesua is recognized by the following putative synapomorphies: a nonfunctional terminal bud (chr. 1:1*), immersed axillary bud (chr. 3:2**), cyme inflorescence (chr. 17:0*), the presence of a U-shaped partition in the ovary (chr. 39:1**), septifragal capsule (chr. 50:1), the absence of a group of fibers adaxial to the midrib bundle (chr. 60:0), and the presence of a continuous band of fibers adjacent to the abaxial epidermins (chr. 61:2). Calophyllum can be recognized by many potential synapomorphies, including a terminal bud without scales (chr. 2:0), an indumentum of multicellular hairs (chr. 5:1*), eucamptodromous secondary venation (chr. 10:0*), tertiary venation that is not evident (chr. 11:0*), absence of latex/resin cavities in mesophyll (chr. 12:0*), presence of latex/resin canals in mesophyll (chr. 13:0*), intersecondary veins that are replaced by canals (chr. 14:1*), one ovule per gynoecium (chr. 41:2**), absence of persistent calyx (chr. 46:1*), bony endocarp (chr. 49:1**), indehiscent fruit (chr. 50:2), absence of fibers adjacent to abaxial epidermis (chr. 61:0), and a single layer of leaf midrib xylem (chr. 67:0*). The clade containing Caraipa, Haploclathra, Kielmeyera, Clusiella, Mahurea, and Marila has a few potential morphological synapomorphies, including quincuncial calyx aestivation (chr. 22:1*), five petals (chr. 27:0*), three carpels (chr. 37:3*), axile or intruded-parietal placentation (chr. 38:0), and median ovule position (chr. 42:0). Potential synapomorphies for the clade containing Kielmeyera, Caraipa, and Haploclathra include an indumentum of multicellular hairs (chr. 5:1*), alternate leaves

PAGE 58

47 (chr. 6:1), sepals that enclose the petals before anthesis (chr. 24:1), cordate cotyledons (58:1**), and the absence of druse crystals (chr. 73:0). Putative synapomorphies for Kielmeyera include more than 15 ovules per carpel (chr. 41:1), absence of a persistent calyx (chr. 46:1), and a seed with a wing two cell layers thick, going completely around the seed, with no vascular tissue (chr. 51:2**). Caraipa and Haploclathra are united by a terminal bud without scales (chr. 2:0*), a septifragal capsule (chr. 50:1*), a seed with a wing several cell layers thick, going completely around the seed, with no vascular tissue (chr. 51:1**), and the presence of a group of fibers adaxial to the leaf midrib bundle (60:1*). One possible synapomorphy for Caraipa is bowl-shaped anther glands (chr. 35:1). Haploclathra may be diagnosed by the following putative synapomorphies: opposite leaves (chr. 6:0), slightly connate filaments (chr. 32:1), absence of anther glands (chr. 34:0*), basal ovule position (chr. 42:1*), abaxial phloem in leaf midrib partitioned by fibers (chr. 59:1), the presence of a continuous band of fibers adjacent to the abaxial epidermis (chr. 61:2*), a papillose epidermis (chr. 72:1*), and petiole bundle an arch with dorsal groupings of xylem and phloem (chr. 74:2**). The clade containing Clusiella, Mahurea, and Marila has the following potential synapomorphies: more than 15 ovules per carpel (chr. 41:1), a testa vasculature of unbranched or braided raphal bundle (chr. 54:0), presence of an exotegmen (chr. 55:0*), presence of endosperm in the mature seeds (chr. 56:0*), and an embryo less than four mm long (chr. 57:0*). Clusiella has many potential synapomorphies, including the presence of an interpetiolar stipuliform structure (chr. 8:2**), presence of latex/resin canals in the mesophyll (chr. 13:0*), sympodial shoot growth (chr. 16:1**), Clusiella-type cyme (chr. 17:2**), bracteoles with an abaxial gland (chr. 20:1**), dioecy (chr.

PAGE 59

48 21:0*), imbricate calyx aestivation (chr. 22:2*), Clusiella-type stamen dimorphism (chr. 31:1**), Clusiella-type monadelphous stamens (chr. 32:2**), ten to 30 carpels (chr. 37:5**), laminar placentation with ovules scattered on the partitions (chr. 38:2**), indehiscent fruit (chr. 50:2), absence of transcurrent lateral bundles (chr. 63:0*), absence of marginal lignification (chr. 65:0*), and absence of a marginal canal in leaf blade (chr. 66:1**). Mahurea and Marila have the following potential synapomorphies: percurrent tertiary venation (chr. 11:2**), septicidal capsule (chr. 50:0), and a petiole bundle with three layers of xylem and phloem (chr. 74:1*). Potential synapomorphies of Mahurea include an indumentum of unbranched, unicellular hairs (chr. 4:1), alternate leaves (chr. 6:1*), presence of stipuliform structures (chr. 8:1*), eucamptodromous secondary venation (chr. 10:0*), free filaments (chr. 32:0), bowl-shaped anther glands (chr. 35:1*), intruded placentae bordered by in-curled carpel walls (chr. 40:1**), a seed with an elongate wing several cell layers thick, not going completely around the seed, with a peripheral vascular bundle (chr. 51:3**), three to six layers of xylem in the leaf midrib (chr. 67:2) and the absence of druse crystals in the petiole (chr. 73:0). Marila can be diagnosed by the following likely synapomorphies: an axillary inflorescence (chr. 15:1*), a raceme (chr. 17:4**), absence of a terminal flower (chr. 18:1*), fused sepals (chr. 25:1), imbricate corolla aestivation (chr. 26:0), plumose seeds (chr. 52:1**), and the presence of druse crystals in the petiole (chr 73:1). Morphological character-state transformations mapped onto a representative most parsimonious tree from Analysis 9 (DNA + morphology with Neotatea, pruned) are shown for the clade containing Neotatea (Figure 3-17). The clade containing Clusiella, Marila, Mahurea, and Neotatea has the same synapomorphies as does the clade of

PAGE 60

49 Clusiella, Mahurea, and Marila in Analysis 7 (Figure 3-16). Putative synapomorphies for Marila, Mahurea, and Neotatea are the presence of unicellular hairs (chr. 4:1*), percurrent tertiary venation (chr. 11:2*), sepals that enclose the petals before anthesis (24:0*), septicidal capsule (chr. 50:0*), and a petiole bundle with three layers of xylem and phloem (chr. 74:1*). Mahurea and Neotatea are united by the following potential synapomorphies: alternate leaves (chr. 6:1*), free filaments (chr. 32:0*), intruded placentae bordered by in-curled carpel walls (chr. 40:1**), and a seed with an elongate wing several cell layers thick, not completely surrounding the seed, with a peripheral vascular bundle (chr. 51:3**). Potential synapomorphies of Neotatea are the absence of colleters (chr. 7:1), tertiary venation that is not evident (chr. 11:0*), absence of latex/resin cavities in the mesophyll (chr. 12:0*), presence of latex/resin canals (chr. 13:0*), intersecondary veins that are replaced by canals (chr. 14:1*), Neotatea-type inflorescence (chr. 17:3**), absence of bracteoles (chr. 19:1*), imbricate calyx aestivation (chr. 22:2*), anthers longer than six mm (chr. 33:2**), phloem of leaf midrib doubly partitioned by fibers (chr. 59:2**), lateral bundles of leaf midrib not transcurrent (chr. 63:0), presence of canals immediately adaxially and abaxially positioned in relation to secondary veins (chr. 64:1**), a single layer of leaf midrib xylem (chr. 67:0*), a lamina with a hypodermis (chr. 70:0*), and a petiole bundle with vascular tissue arched to circular (chr. 74:0).

PAGE 61

50 Figure 3-1. Strict consensus of 1530 most parsimonious trees of length 845 from Analysis 1 (rbcL alone). CI = 0.612, RI = 0.755, RC = 0.462, HI = 0.388; of 1408 aligned positions, 240 are parsimony-informative. Numbers above branches represent bootstrap values. Kayea sp. was originally labeled as Mesua sp. in the Gustafsson et al. (2002) analysis, but is probably a species of Kayea (see Discussion).

PAGE 62

51 Figure 3-2. Majority-rule consensus tree based on rbcL data for Clusiaceae (Analysis 1). Bayesian analysis was run using the GTR+I+ substitution model. The analysis employed four chains of two million generations. Trees produced during the burn-in phase were not used to produce the consensus tree. Numbers above branches are the percent of the time that the clade occurs among the sampled trees (i.e., the posterior probability).

PAGE 63

52 Figure 3-3. Maximum likelihood tree of rbcL sequences across Clusiaceae (Analysis 1). Likelihood score = 6792.82649. Maximum likelihood analysis employed the GTR+I+ substitution model.

PAGE 64

53 Figure 3-4. Strict consensus of 510 most parsimonious trees of length 361 from Analysis 2 (rbcL + matK). CI = 0.831, RI = 0.762, RC = 0.633, HI = 0.169; of 1885 aligned positions, 120 are parsimony-informative. Numbers above branches represent bootstrap values. Kayea sp. was originally labeled as Mesua sp. in the Gustafsson et al. (2002) analysis, but is probably a species of Kayea (see Discussion).

PAGE 65

54 Figure 3-5. Maximum likelihood tree of rbcL + matK data set (Analysis 2). Likelihood score = 4912.90425. Maximum likelihood analysis employed the GTR+I+ substitution model. Numbers above branches represent bootstrap values.

PAGE 66

55 Figure 3-6. Strict consensus of 10 most parsimonious trees of length 987 from Analysis 3 (ITS alone). CI = 0.584, RI = 0.658, RC = 0.384, HI = 0.416; of 809 aligned positions, 232 are parsimony-informative. Numbers above branches represent bootstrap values.

PAGE 67

56 Figure 3-7. Maximum likelihood tree of ITS sequences (Analysis 3). Likelihood score = 5734.93593. Maximum likelihood analysis employed the GTR+I+ substitution model. Numbers above branches represent bootstrap values.

PAGE 68

57 Figure 3-8. Strict consensus of 2 most parsimonious trees of length 1326 from Analysis 4 (rbcL + matK + ITS, unpruned). CI = 0.643, RI = 0.665, RC = 0.427, HI = 0.357; of 2696 aligned positions, 311 are parsimony-informative. Numbers above branches represent bootstrap values.

PAGE 69

58 Figure 3-9. Majority-rule consensus tree based on the unpruned rbcL + matK + ITS data set (Analysis 4). Bayesian analysis was run using the GTR+I+ substitution model. The analysis employed four chains of two million generations. Trees produced during the burn-in phase were not used to produce the consensus tree. Numbers above branches are the percent of the time that the clade occurs among the sampled trees (i.e., the posterior probability).

PAGE 70

59 Figure 3-10. Majority-rule consensus tree based on the pruned rbcL + matK + ITS data set (Analysis 5). Bayesian analysis was run using the GTR+I+ substitution model. The analysis employed four chains of two million generations. Trees produced during the burn-in phase were not used to produce the consensus tree. Numbers above branches are the percent of the time that the clade occurs among the sampled trees (i.e., the posterior probability).

PAGE 71

60 Figure 3-11. Strict consensus of 61 most parsimonious trees of length 227 from Analysis 6 (morphology alone). CI = 0.482, RI = 0.740, RC = 0.357, HI = 0.518; of 74 characters, 64 are parsimony-informative. Numbers above branches represent bootstrap values.

PAGE 72

61 Figure 3-12. Strict consensus of 4 most parsimonious trees of length 1990 from Analysis 7 (DNA + morphology, unpruned). CI = 0.671, RI = 0.648, RC = 0.435, HI = 0.329; of 2770 characters, 436 are parsimony-informative. Numbers above branches represent bootstrap values.

PAGE 73

62 Figure 3-13. Strict consensus of 4 most parsimonious trees of length 1297 from Analysis 8 (DNA + morphology, pruned). CI = 0.705, RI = 0.548, RC = 0.386, HI = 0.295; of 2770 characters, 306 are parsimony-informative. Numbers above branches represent bootstrap values.

PAGE 74

63 Figure 3-14. Strict consensus of 28 most parsimonious trees of length 2009 from Analysis 9 (DNA + morphology, unpruned, including Neotatea). CI = 0.666, RI = 0.654, RC = 0.436, HI = 0.334; of 2770 characters, 439 are parsimony-informative. Numbers above branches represent bootstrap values.

PAGE 75

64 Figure 3-15. Strict consensus of 12 most parsimonious trees of length 1315 from Analysis 10 (DNA + morphology, pruned, including Neotatea). CI = 0.698, RI = 0.569, RC = 0.398, HI = 0.302; of 2790 characters, 310 are parsimony-informative. Numbers above branches represent bootstrap values.

PAGE 76

Figure 3-16. Morphological character-state transformations mapped onto one of the most parsimonious trees (tree #2) from Analysis 7 (DNA + morphology, unpruned). Equivocal, very homoplasious character-state transformations (i.e., 0/1 0/1) are not included in this list. See Table 2-1 for a description of characters and character states. A: 2(0/11), 13(01), 44(0/11), 63(01), 65(01); B: 1(01), 8(01), 10(0/11), 22(01), 24(0/11), 27(10); 28(01), 31(02), 32(03), 37(0/1/21), 38(0/3/44), 42(0/1/22), 45(01), 48(01), 50(0/22); C: 15(01), 17(0/1/55), 18(01), 21(0/10), 23(13), 24(0/10), 25(01), 26(12), 38(0/33), 42(0/11, 46(01), 50(0/22), 73(0/11); D: 10(0/11), 17(0/1/51), 21(0/11), 67(01); E: 38(0/33), 42(0/11), 60(01); F: 45(01); G: 23(1/22), 24(0/10), 25(02), 26(10), 37(24), 43(01), 47(01), 62(01), 73(0/11); H: 4(01), 8(01), 22(21), 23(1/21), 24(0/11), 27(0/10), 29(01), 32(0/10), 34(0/10), 36(01), 44(10), 50(0/20), 59(0/11), 73(0/10); I: 7(01), 15(01), 19(01), 26(12), 32(0/11), 34(0/10)73(0/11); J: 1(01), 3(12), 17(10), 39(01), 50(0/1/21), 60(10), 61(0/22); K: 2(0/10), 5(01), 10(10), 11(10), 12(10), 13(10), 14(01), 41(02), 46(01), 49(01), 50(0/1/22), 61(0/20), 67(10); L: 22(01), 27(10), 37(23), 38(0/30), 42(0/10); M: 5(01), 6(01), 24(0/11), 58(01), 73(0/10); N: 41(0/11), 46(0/11), 51(0/1/22); O: 2(10), 50(0/21), 51(0/1/21), 54(0/11), 60(01); P: 35(0/11); Q: 6(10), 32(0/11), 34(10), 42(01), 59(01), 61(02), 72(01), 74(02); R: 41(0/11), 54(0/10), 55(10), 56(10), 57(10); S: 8(02), 13(10), 16(01), 17(12), 20(01), 21(10), 22(12), 31(01), 32(0/12), 37(35), 38(02), 50(0/22), 63(10), 65(10), 66(01); T: 11(12), 50(0/20), 74(01); U: 4(0/11), 6(01), 8(01), 10(10), 32(0/10), 35(01), 40(01), 51(03), 67(1/22), 73(0/10; V: 15(01), 17(14), 18(01), 25(01), 26(10), 52(01), 73(0/11). 65

PAGE 77

66

PAGE 78

67 Figure 3-17. Morphological character-state transformations mapped onto one of the most parsimonious trees (tree #1) from Analysis 9 (DNA + morphology with Neotatea, pruned) showing the clade containing Neotatea. Equivocal, very homoplasious character-state transformations (i.e., 0/1 0/1) are not included in this list. See Table 2-1 for a description of characters and character states. A: 41(0/11), 42(0/10), 54(0/10), 55(10), 56(10), 57(10); B: 4(01), 11(12), 24(10), 50(20), 67(12); C: 6(01), 32(10), 40(01), 51(03); D: 7(0/11), 11(20), 12(10), 13(10), 14(01), 17(13), 19(01), 22(12), 33(02), 59(02), 63(0/10), 64(01), 67(20), 70(01), 74(0/10).

PAGE 79

Table 3-1. Bootstrap support values by parsimony analysis 68 BootstrapSupport Clade rbcL matK matK + rbcL ITS ITS + matK + rbcL, unpruned ITS + matK + rbcL, pruned Kielmeyeroideae 75 100 100 NA NA Calophylleae 95 98 100 100 Mammea + rest of Calophylleae 57 52 78 Kayea 91 NA 93 100 99 NA Kayea + Poeciloneuron 60 72 70 Mammea 88 99 100 99 100 100 Marila 85 88 97 Marila + Mahurea + Clusiella Neotatea + Mahurea + Marila + Clusiella NA NA NA NA NA NA Calophyllum 99 100 99 Mesua + Calophyllum 64 86 84 84 Haploclathra NA NA NA 100 100 NA Kielmeyera 53 97 96 69 98 98 Caraipa 52 74 95 98 95 Caraipa + Haploclathra Haploclathra + Kielmeyera + Caraipa 57 71 86 100 92

PAGE 80

Table 3-1. Continued. Bootstrap Support Clade morpho. DNA + morpho w/o Neotatea, unpruned DNA + morpho w/o Neotatea, pruned DNA + morpho w/ Neotatea, unpruned DNA + morpho w/ Neotatea, pruned Kielmeyeroideae 52 NA NA Calophylleae 98 66 Mammea + rest of Calophylleae Kayea 87 100 NA 100 NA Kayea + Poeciloneuron 75 61 78 Mammea 95 100 100 100 100 Marila 96 93 NA 92 NA Marila + Mahurea + Clusiella 51 67 Neotatea + Mahurea + Marila + Clusiella NA NA 69 Calophyllum 100 100 100 100 100 Mesua + Calophyllum 97 97 96 96 Haploclathra 94 100 NA 100 NA Kielmeyera 75 92 100 92 100 Caraipa 96 98 97 95 Caraipa + Haploclathra 58 56 55 Haploclathra + Kielmeyera + Caraipa 91 94 92 93 69 Bootstrap support generally increases with combination of data sets. A dash () signifies that the clade is not supported by a bootstrap value >50%.

PAGE 81

CHAPTER 4 DISCUSSION Analytical Issues The topologies from each data set analyzed separately did not generally conflict with each other. The only conflict involves the position of Mammea as supported by morphology compared to rbcL and matK nucleotide sequences. Morphology alone (Figure 3-8) places Mammea in a clade with Calophyllum, Mesua, and Kayea, while the analyses using rbcL alone (Figures 3-1 through 3-3) and matK alone (results not shown) place Mammea as sister to the rest of Calophylleae (i.e., all Kielmeyeroideae except Endodesmia). In each case, the position of Mammea in the separate analyses is not well-supported. Combining the three gene regions into one data set increased internal support (Table 3-1) and decreased the resulting number of most parsimonious trees. For example, Kayea and Poeciloneuron are sisters with 72% bootstrap support in the parsimony unpruned combined DNA analysis (Figure 3-5), whereas this clade does not receive more than 60% bootstrap support in any of the separate analyses (Figures 3-1 and 3-4). Marila does not receive bootstrap support >50% in the rbcL and matK separate analyses and receives 85% support in the ITS analysis; in the unpruned combined DNA analysis, Marila is supported with 89% bootstrap support. Furthermore, the clade of Kielmeyera + Caraipa + Haploclathra is supported by 95% bootstrap support in the unpruned combined DNA analysis, whereas this clade does not receive support greater than 86% in the separate analyses. While the sister-group relationship of Mammea to the 70

PAGE 82

71 rest of Calophylleae is not supported in the combined DNA analyses, it is supported by 78% bootstrap support in the rbcL + matK analysis (Figure 3-3). The total evidence approach (i.e., combining all data sets into a single matrix for analysis) was also shown to have positive results in the large data set analyses of Soltis et al. (1998) and Chase et al. (1998). The Value of Morphology The value of morphology in phylogenetic studies has recently been questioned by Scotland et al. (2003). These authors argue that one of the major constraints in using morphology involves the limited number of unambiguous characters provided by this time-consuming approach. It is interesting to note that in this study, morphology alone provides 64 parsimony-informative characters, while the matK region used (488 bp) provides only 55 parsimony-informative characters and rbcL (1408 bp, of Kielmeyeroideae) provides 66 parsimony-informative characters. The number of informative characters provided by including 809 bp of ITS is much greater (232) than for the other data sets, but the homoplasy index is also higher for ITS (0.416, compared to 0.205 for rbcL and 0.116 for matK). Of the four separate data sets (rbcL, matK, ITS, and morphology), the morphological characters had the highest homoplasy index (0.518). Despite the greater amount of homoplasy in the morphological characters, combining morphology with the three molecular data sets (Figures 3-9 through 3-12) results in values of support equivalent to those in the combined DNA analyses (Table 3-1). In some cases, bootstrap support is higher when morphology is added. For example, a clade containing Clusiella, Mahurea, and Marila, present in the strict consensus of the combined molecular data set (but without bootstrap support >50%), receives bootstrap support of 51% (in the unpruned analysis) and 67% (in the pruned analysis) with the

PAGE 83

72 addition of morphology to the combined molecular data set (Figures 3-9 and 3-10). The sister-group relationship of Calophyllum and Mesua also receives higher support (97% vs. 84%) with the addition of morphology to the combined DNA analyses (Figures 3-9 and 3-10). In the analysis that included only morphological characters, every genus except for Caraipa (which was placed in a clade with Haploclathra) appeared monophyletic. Scotland et al. (2003) stated that much of the useful morphological diversity has already been scrutinized. This statement, which discourages morphological phylogenetics, is misleading. While it may be true that many of the morphological attributes of an organism have been observed at some earlier point in the taxonomic history of the organism, these characters may not have been applied in a phylogenetic context. Characters that have been stressed in historical classifications do not always turn out to be the appropriate (i.e., accurate indicators of phylogenetic relationships). The historical classification of Clusiella illustrates this point. Characters that had already been scrutinized and stressed in Englers (1925) classification of Clusiella (resiniferous staminodes, cupuliform stigmas, epiphytic habit) were characters that incorrectly placed the genus near Clusia. Only when numerous morphological characters were examined in a phylogenetic context did the taxonomic placement of Clusiella within Kielmeyeroideae become clear. For example, the presence of stipuliform structures, resin cavities, anther glands, seeds without an aril, and an embryo with large cotyledons support this placement. Scotland et al. (2003) did support the systematic use of anatomical characters, after the necessary rigorous and critical studies. The present investigation does show that

PAGE 84

73 anatomical characters can be phylogenetically useful (i.e., presence/absence of transcurrent lateral bundles, position of resin canals relative to vascular bundles, presence/absence of fibers adaxial to the midrib bundle, and presence/absence of fibers partitioning the phloem). Through observations of just leaf and petiole anatomy, 16 characters were discovered, and 13 of these are phylogenetically informative. But why only study the anatomy of few morphological characters, as Scotland et al. suggest? If the anatomy of other parts of the plant (i.e., stems, roots, flowers, fruits, etc.) were observed, it seems likely that many additional, phylogenetically useful, anatomical characters could have been found. In a response to the Scotland et al. (2003), Jenner (2004) stated that, in the case of well-studied seed plants, the assertion that much of the morphological diversity has already been scrutinized may be true. The present study provides evidence that, not even in the case of seed plants, is this statement valid. Several morphological characters, which had never before been used phylogenetically in a published work, were shown to be phylogenetically informative. The use of morphology in the phylogenetic analysis of Kielmeyeroideae allowed the inclusion of Neotatea, a genus endemic to the tepuis of northern South America. DNA could not be obtained from Neotatea because the herbarium specimens were originally collected in alcohol. Because only morphological characters could be used to infer its phylogenetic relationships, including Neotatea in the analysis is similar to including a fossil taxon (Gandolfo, 2002; Hermsen et al., 2003). Although its position does not receive strong support (Figure 3-12), the placement of Neotatea as the sister group to Mahurea, in a clade with Clusiella and Marila, is logical based on seed and fruit

PAGE 85

74 characters (Figure 3-14) and is a good hypothesis of relationships until molecular data for Neotatea can be acquired. The inclusion of morphological data also allowed clear comparison of phylogenetic results with traditional classifications of Kielmeyeroideae and determination of putative morphological synapomorphies for the genera of Kielmeyeroideae (see below and Results). Taxonomic History The taxonomic placement of several genera of Kielmeyeroideae has been a source of confusion throughout the taxonomic history of this group. The genera that are primarily South American in distribution have been especially problematic. While Englers (1925) treatment of Clusiaceae included Kielmeyera, Mahurea, Marila, Caraipa, and Haploclathra (in Kielmeyeroideae), other workers had different placements for these genera. Hutchinson (1959, 1969) placed Kielmeyera, Marila, and Caraipa in Bonnetiaceae, while Maguire (1972) also included Haploclathra, Mahurea, and Neotatea, but not Marila, in Bonnetiaceae. Kubitzki (1978) included Caraipa and Mahurea in Bonnetiaceae and Field (1978) placed Haploclathra in Theaceae, which he included as part of Bonnetiaceae. Cronquist (1981) tentatively placed Kielmeyera and Neotatea within Bonnetioideae (Theaceae), stating that although these genera have alternate leaves like Bonnetioideae and other Theaceae, they have the resin ducts and anatomy of Clusiaceae. In the morphological cladistic analysis of Clusiaceae-Bonnetiaceae-Elatinaceae of Stevens (unpubl.), Marila, Caraipa, Kielmeyera, Haploclathra, Neotatea, and Mahurea appear as basal branches within Kielmeyeroideae (Clusiaceae). The rbcL-based phylogeny of Clusiaceae by Gustafsson et al. (2002) confirms the placement of Marila, Mahurea, Kielmeyera, and Caraipa within

PAGE 86

75 Kielmeyeroideae, but does not support a basal position of these genera. The results presented here, based on rbcL sequences across Clusiaceae, including a larger sampling of Kielmeyeroideae (Figures 3-1 through 3-3), confirm the placement of Marila, Mahurea, Neotatea, Kielmeyera, Caraipa, and Haploclathra within Kielmeyeroideae. These New World genera, in addition to Clusiella, likely form one of three main clades within Calophylleae (Figures 3-2, 3-3, 3-8, 3-9, 3-12, 3-13, and 3-15). The taxonomic placement of Clusiella has also been a cause of debate. When first described, Planchon and Triana (1860) placed Clusiella in their tribe Clusieae based on its sessile, cupuliform stigmas and non-ascendent ovules. Engler (1925) also believed that Clusiella belonged in Clusioideae, and placed the genus next to Clusia. Besides stigma shape and ovule orientation, Clusiella and Clusia are similar in having an epiphytic habit, resiniferous androecium, and dioecious habit. Hammel (1999) noted that Clusiella has many characteristics (i.e., stipuliform structures, bud scales, resin cavities in the leaf, psilate pollen exine, baccate fruits, small foveolate seeds without an aril, and an embryo with large cotyledons) that are not common in clusioid genera, and thus he questioned its placement in Clusioideae. Stevens (in press) placed Clusiella in its own tribe, Clusielleae, sister to the rest of Clusioideae. In the rbcL-based phylogeny of Gustafsson et al. (2002), Clusiella appears within a well-supported Kielmeyeroideae, but its relationship to other genera of the subfamily is not supported. In the separate molecular parsimony analyses of the present study (Figures 3-1 and 3-6), Clusiellas position within Kielmeyeroideae is unresolved. In the combined molecular DNA analyses, Clusiella appears in a clade with Mahurea and Marila, but without bootstrap support >50% (Figure 3-8). The Bayesian analysis of rbcL data across Clusiaceae

PAGE 87

76 (Figure 3-2) weakly supports a clade of Mahurea, Clusiella, and Marila with a posterior probability of 72. In the maximum likelihood analysis of rbcL (Figure 3-3) and rbcL + matK + ITS (results not shown), Clusiella appears in a clade with Marila and Mahurea. When the morphological characters are added to the combined molecular data set (analyses 7-10), Clusiella appears with weak to moderate bootstrap support in a clade with Marila and Mahurea (Figures 3-12 and 3-13), or with Marila, Mahurea, and Neotatea, in the pruned analysis that includes Neotatea (Figure 3-15). This clade is diagnosed by several seed and fruit characters, such as an unbranched or braided raphal bundle in the testa, the absence of an exotegmen, presence of endosperm in ripe seeds, a small embryo, and gynoecia having more than 15 ovules per carpel. Based on its unisexual flowers, drupaceous fruit, and large embryo, Robson (1978) included Mammea in Clusioideae, tribe Garcinieae. The large embryos of Garcinieae and Mammea are only superficially similar; the embryo of Garcinia is composed almost entirely of a swollen hypocotyl, while that of Mammea is mainly two immense cotyledons (Stevens, 1980). Mammea was properly placed in Kielmeyeroideae by Takhtajan (1997) and Stevens (in press). In the morphological cladistic analysis of Stevens (unpubl.), Mammea appears in a clade with other primarily Old World genera (e.g., Calophyllum, Mesua, Kayea, and Poeciloneuron). Mammea, Calophyllum, Mesua, Kayea, and Poeciloneuron all have basal placentation and usually two or four carpels. In my analysis based solely on morphological characters (Figure 3-8), Mammea appears in a clade with Calophyllum, Mesua, and Kayea. Mammea appears distinct from these genera (and sister to all other Calophylleae) in the rbcL phylogeny of Gustafsson et al. (2002), although without bootstrap support. With an increased taxon sampling of Mammea and

PAGE 88

77 other Kielmeyeroideae, this relationship still does not receive support greater than 50% in the parsimony analysis of rbcL sequences (Figure 3-1). The Bayesian analysis of rbcL data (Figure 3-2) places Mammea sister to the remaining Calophylleae with a posterior probability of 71. By adding an additional gene region, the sister-group relationship of Mammea + remaining Calophylleae receives stronger bootstrap support (Analysis 2, rbcL + matK; Figures 3-4 and 3-5). The Bayesian analysis of the rbcL + matK data set supports the sister group relationship of Mammea + remaining Calophylleae with a posterior probability of 92 (results not shown). In the combined DNA + morphology analyses (Figures 3-9 through 3-12), the position of Mammea is again unresolved. The position of Endodesmieae (which includes two monotypic genera, Endodesmia and Lebrunia) within Clusiaceae has been questioned due to the morphological similarities of the genera of Endodesmieae to both Clusioideae and Kielmeyeroideae. Vegetatively, Endodesmieae are similar to Clusioideae in that they are glabrous, possess resin canals, and have lateral vascular bundles in the leaf blade that are not transcurrent; however, the presence of large seeds with large cotyledons allies Endodesmieae with Kielmeyeroideae. Stevens (in press) tentatively placed Endodesmieae sister to Calophylleae within Kielmeyeroideae. For the first time, Endodesmia is included here in a molecular phylogenetic analysis, and its placement is confirmed. In both Analysis 1 (rbcL; Figures 3-1 through 3-3) and Analysis 2 (rbcL + matK; Figures 3-4 and 3-5), Endodesmia appears sister to the remaining Kielmeyeroideae, the Calophylleae, with high support. In the rbcL phylogeny of Gustafsson et al. (2002), Kayea stylosa and Mesua sp. appear as sister taxa; this is likely an artifact of a misidentification of their Mesua sp.,

PAGE 89

78 which is probably a species of Kayea. Kayea has been included in Mesua in the past (Kostermans, 1969), but was properly separated from Mesua by Bentham (1863) and Anderson (1874) based on stigma shape and ovary type. Kayea has four narrow stigma lobes while Mesua has two broad stigma lobes. The ovary of Kayea is composed of a single locule (but is four-carpellate) and contains four or more ovules. Mesua has a two-loculate (two-carpellate) ovary with four ovules. Despite its history of being treated as closely related to Mesua, Kayea is actually more closely related to the Western Ghats endemic Poeciloneuron, while Mesua is the sister to Calophyllum (Figures 3-1 through 3-10 and 3-12 through 3-15). Kayea and Poeciloneuron both have narrow stigmas, and Mesua and Calophyllum have broad stigma lobes. Character Evolution One of the probable synapomorphies for the Clusiaceae is the presence of resin or other exudates in secretory canals or cavities (Stevens, 2001; Judd et al., 2002). Determining which of these (i.e., canals or cavities) is the ancestral condition cannot be answered until we have an understanding of subfamilial relationships within Clusiaceae. In the strict consensus of the rbcL family-level parsimony analysis (Figure 3-1), Kielmeyeroideae are sister to a Clusioideae + Hypericoideae clade, but this relationship receives no bootstrap support. The Bayesian analysis also shows this topology, but the probability is low (Figure 3-2). Within Kielmeyeroideae, if it is assumed that the presence of canals is the ancestral state, it is most parsimonious to assume that canals were lost in the ancestor of Calophylleae and re-evolved in three different lineages: Neotatea, Clusiella, and Calophyllum (Figure 4-1). It is important to note that while the four species of Mammea included in my analysis do not have canals, some species (i.e., M. vatoensis, M. nervosa, and M. mirabilis) are reported to have canals (Stevens,

PAGE 90

79 unpubl.), thus possibly making the loss of canals, as assumed by parsimony, occur one node up in the cladogram. Assuming that the absence of resin cavities is the ancestral state, then resin cavities evolved in the ancestor of Kielmeyeroideae and have been lost two separate times, in Neotatea and Calophyllum (Figure 4-2). It is interesting that Endodesmia, which is sister to the rest of Kielmeyeroideae, possesses both resin cavities and canals. Having opposite leaves is one character that is often given as a diagnostic feature of Clusiaceae, but within Kielmeyeroideae, alternate leaves evolved two or three times. Alternate leaves evolved once in the Mahurea-Neotatea clade and once or twice in the Kielmeyera-Caraipa-Haploclathra clade (Figure 4-3). The change to alternate leaves could have occurred in the ancestor of Kielmeyera-Caraipa-Haploclathra with a reversal to opposite leaves in Haploclathra or alternate leaves could have evolved in parallel in Kielmeyera and Caraipa (Figure 4-3). Some of the anatomical characters used in this analysis are highly homoplasious (i.e., presence of druse crystals in the petiole); however, despite their homoplasy, many of the anatomical characters varied in a phylogenetically informative manner. The presence of a lignified leaf margin and transcurrent lateral bundles in the leaf blade are two features that likely evolved in the ancestor of Calophylleae and were lost only sporadically within this clade (Figures 4-4 and 4-5). Petiole bundle architecture is another anatomical character that is phylogenetically useful. A petiole bundle with three layers of xylem and phloem is probably a synapomorphy for Marila, and a petiole bundle comprised of an arch with dorsal groupings of xylem and phloem is unique to Haploclathra (Figure 4-6).

PAGE 91

80 A noteworthy feature of many members of Kielmeyeroideae is the presence of an apical gland on the anthers. The shape of the anther gland may be spherical to elongate or bowl-shaped (crateriform). The anther glands were described by Stevens (in press) as containing latex, resin, or other material; other reports (i.e., Gustafsson et al., 2002) describe the glands as containing oily fluids. I observed that the anther glands of Endodesmia secrete a shiny, resin-like substance; in the anther glands of other taxa, the secretion is not evidently resin-like and may indeed be oils of some kind. Work needs to be done to characterize the substances secreted by these anther glands. The presence of anther glands may be a synapomorphy for Kielmeyeroideae; however, many members of Hypericoideae also have anther glands (Stevens, in press), but their homology with those of Kielmeyeroideae is uncertain. Although anther glands are present in Symphonia (Stevens, unpubl.), most Clusioideae lack anther glands. If we assume the absence of anther glands is ancestral, then anther glands evolved once in the ancestor of Kielmeyeroideae and were lost several times within this subfamily (Figure 4-7). The absence/presence of anther glands may vary intragenerically (i.e., Mammea, Kielmeyera, and Kayea), be consistently absent within a genus (i.e., Haploclathra, Poeciloneuron, Mesua, and Calophyllum), or be consistently present within a genus (i.e., Marila, Mahurea, Clusiella, and Caraipa). Little is known about the pollination biology of members of Kielmeyeroideae. Taxa that secrete resin via anther glands (i.e., Endodesmia and possibly others) or staminodes (i.e., Clusiella) are likely pollinated by resin-collecting bees. The bees use the resin for nest construction (Armbruster, 1984; Oliveira et al., 1996). The resin polymerizes slowly and provides the bees with a waterproof protection and probably

PAGE 92

81 antiviral and antimicrobial activity for their larvae (Oliveira et al., 1996). Anther glands that may contain oils might attract pollinators by fragrance (Ribeiro et al., 1999) or be used in nest construction. Carpel number varies greatly within Kielmeyeroideae, but the variation is phylogenetically informative (Figure 4-8). Having two carpels is likely the ancestral state in the Calophylleae, and the two-carpellate taxa do not form one clade. Kayea, whose closest relatives have two carpels, apparently experienced a doubling of its carpel number to four. The presence of three carpels is a likely synapomorphy for the clade containing Kielmeyera, Caraipa, Haploclathra, Clusiella, Mahurea, Neotatea, and Marila. Within this clade, there is a secondary increase in carpel number in Clusiella. Fruit type within Clusiaceae is homoplasious, as it is in many other angiosperm families (Kron et al., 2003; Judd et al., 2002; Wilson et al., 2001). The number of times a fleshy, indehiscent fruit has evolved within Kielmeyeroideae is unclear. Using a DELTRAN optimization and assuming that capsular fruits are ancestral, a fleshy, indehiscent fruit may have evolved five different times in Kielmeyeroideae (Figure 4-9). The fibrous, drupaceous fruits of Calophyllum are dispersed by birds, bats, or sometimes water in the case of strand species (Stevens, in press). Most of the fleshy fruits of Mammea are dispersed by mammals (Stevens, in press). The seeds of Clusiella isthmensis have been found in a fecal sample of a thrush-like manakin (Schiffornis turdinus) in Costa Rica (A. Boyle, pers. comm.). Most species of Kayea have capsular fruits, but in some species (i.e., K. elmeri), the calyx becomes highly accrescent and surrounds the fruit, making it indehiscent. Septifragal capsules evolved twice within Kielmeyeroideae: in Mesua and in the Caraipa-Haploclathra clade.

PAGE 93

82 Winged seeds have probably evolved three separate times within Kielmeyeroideae (Figure 4-10). By examining their morphology and anatomy, it is clear that all winged seeds in the group are not homologous. The wing on the seeds of Mahurea and Neotatea does not completely surround the seed, is several cell layers thick, and contains a peripheral vascular bundle. The wing on the seeds of Kielmeyera completely surrounds the seed, is two cell layers thick, and does not have any vascular tissue. Caraipa and Haploclathra have a seed wing that completely surrounds the seed, is several cell layers thick, and does not contain vascular tissue. The small, dry seeds of Marila and the winged seeds of Mahurea, Neotatea, Kielmeyera are wind-dispersed (Stevens, in press). The narrowly winged seeds of Caraipa and Haploclathra are wind dispersed or possibly water-dispersed. Many species of Caraipa and Haploclathra live in periodically flooded habitats of the Amazonian region (Kubitzki, 1989; Vasquez, 1993), in which the production of a water-dispersed seed would be advantageous. Seed and embryo characters are shown to vary in a phylogenetically informative manner. The loss of an exotegmen occurred in the ancestor of Kielmeyeroideae and a reversal to the presence of an exotegmen occurred once within the subfamily, in the Clusiella-Mahurea-Neotatea-Marila clade (Figure 4-11). The presence of endosperm in mature seeds shows the same evolutionary pattern: a loss of endosperm occurred in the presumed common ancestor of Kielmeyeroideae and a reversal to presence of endosperm occurred in the Clusiella-Mahurea-Neotatea-Marila clade (Figure 4-12). Having an embryo less than four mm long is another synapomorphy for the Clusiella-Mahurea-Neotatea-Marila clade; all other members of Kielmeyeroideae have an embryo greater

PAGE 94

83 than four mm in length (Figure 4-13). Cotyledon shape is also a useful character: cordate cotyledons are unique to the clade containing Kielmeyera, Caraipa, and Haploclathra (Figure 4-14). Biogeography The ancestral distribution of Kielmeyeroideae is unclear based on my reconstructions (Figure 4-15). Endodesmia (along with the other genus of Endodesmieae, Lebrunia) is restricted to tropical Africa. Mammea, whose center of diversity is in Madagascar (Stevens, unpubl.), also has representatives in Africa and in the Neotropics. According to Stevens (unpubl.), phenetically primitive species of Mammea exist in both Central America and Africa. Taxon sampling for Mammea (a genus of about 70 species) must be increased before its biogeographical pattern becomes apparent. Kielmeyera, Caraipa, Haploclathra, Clusiella, Mahurea, Neotatea, and Marila make up an entirely New World clade within Kielmeyeroideae. The other major clade within Kielmeyeroideae, composed of Kayea, Poeciloneuron, Mesua, and Calophyllum, is entirely Old World except for some species of Calophyllum. Calophyllum inophyllum is distributed in both the Neotropics and Paleotropics, while Calophyllum brasiliense, along with about ten other species not included in this study, are restricted to the Neotropics. It is clear from Figure 4-15 that Calophyllum originated in the Paleotropics, and subsequently spread to the Neotropics.

PAGE 95

84 Figure 4-1. Character-state distribution for resin/latex canals in leaf mesophyll within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has canals in the leaf mesophyll. MacClade was set to show all most parsimonious states at each node.

PAGE 96

85 Figure 4-2. Character state distribution for resin/latex cavities in leaf mesophyll within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, does not contain cavities in the mesophyll. MacClade was set to show all most parsimonious states at each node.

PAGE 97

86 Figure 4-3. Character-state distribution for leaf arrangement within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has alternate leaves. MacClade was set to show all most parsimonious states at each node.

PAGE 98

87 Figure 4-4. Character-state distribution for lignification of the leaf margin within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has lignified margins. MacClade was set to show all most parsimonious states at each node.

PAGE 99

88 Figure 4-5. Character-state distribution for transcurrent lateral bundles in the leaf blade within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has lateral bundles that are not transcurrent. MacClade was set to show all most parsimonious states at each node.

PAGE 100

89 Figure 4-6. Character-state distribution for petiole bundle architecture within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has an arched petiole bundle architecture. MacClade was set to show all most parsimonious states at each node.

PAGE 101

90 Figure 4-7. Character-state distribution for anther glands within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has anther glands. MacClade was set to show all most parsimonious states at each node.

PAGE 102

91 Figure 4-8. Character-state distribution for carpel number within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has three carpels. MacClade was set to show all most parsimonious states at each node.

PAGE 103

92 Figure 4-9. Character-state distribution for fruit type within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has a septicidal capsule. MacClade was set to show all most parsimonious states at each node.

PAGE 104

93 Figure 4-10. Character-state distribution for seed form within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has seeds with a wing like that of Mahurea. MacClade was set to show all most parsimonious states at each node.

PAGE 105

94 Figure 4-11. Character-state distribution for exotegmen presence within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has an exotegmen. MacClade was set to show all most parsimonious states at each node.

PAGE 106

95 Figure 4-12. Character-state distribution for endosperm in mature seeds within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has endosperm in the ripe seeds. MacClade was set to show all most parsimonious states at each node.

PAGE 107

96 Figure 4-13. Character-state distribution for embryo length within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has an embryo of less than 4 mm long. MacClade was set to show all most parsimonious states at each node.

PAGE 108

97 Figure 4-14. Character-state distribution for cordate cotyledons within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, has non-cordate cotyledons. MacClade was set to show all most parsimonious states at each node.

PAGE 109

98 Figure 4-15. Biogeographical patterns within Kielmeyeroideae. Tree topology is one tree (Tree #2) from the four most parsimonious trees obtained in the unpruned combined molecular and morphological analysis (Analysis 7). Neotatea, which appears sister to Mahurea in Analysis 10, but was not included in this analysis, occurs in the Neotropics. Although Clusia lanceolata occurs in the Neotropics, Clusioideae are a pantropical group and therefore C. lanceolata is coded as equivocal. MacClade was set to show all most parsimonious states at each node.

PAGE 110

CHAPTER 5 KEY TO THE GENERA OF KIELMEYEROIDEAE In this key, all genera of Kielmeyeroideae are included except Lebrunia (Endodesmieae), which was omitted due to the lack of knowledge regarding its morphological variation. 1. Leaves alternate............................................................................................................2 1. Leaves opposite............................................................................................................5 2. Ovules 1-3 per carpel; seeds with narrow wing completely surrounding seed; stipuliform structures absent; colleters present; latex/resin cavities present in leaf blade..................................................................................................................Caraipa 2. Ovules more than 15 per carpel; winged seeds absent or present; stipuliform structures present or absent; colleters present or absent; latex/resin cavities or canals present in leaf blade......................................................................................................3 3. Anther glands deeply crateriform; stipuliform structures present; colleters present; tertiary venation percurrent to reticulate; latex/resin cavities present in leaf blade....... ........................................................................................................................Mahurea 3. Anther glands spherical, subcrateriform, or absent; stipuliform structures absent; colleters present or absent; tertiary venation reticulate or not evident; latex/resin cavities or canals present in leaf blade.........................................................................4 4. Anther glands spherical; colleters absent; seeds with elongate wing; tertiary venation not evident; latex/resin canals present in leaf blade........................................Neotatea 4. Anther glands spherical to subcrateriform or absent; colleters present; seeds with broad wing; tertiary venation reticulate; latex/resin cavities present in leaf blade......... .....................................................................................................................Kielmeyera 5. Plant an epiphytic shrub or liana; dioecious; interpetiolar stipuliform structures present; staminate flowers with filaments fused into a tube for most of their length and resiniferous staminodes at the base of tube; carpellate flowers with an ovary with 5-20 expanded, sessile stigmas, the ovary surrounded at its base by resiniferous staminodes; fruit a many-seeded berry............................................................Clusiella 5. Plant a tree; monoecious or (andro)dioecious; stipuliform structures present or absent, but if present, then not interpetiolar; flowers without resiniferous staminodes; ovary with fewer than 5 stigma lobes; style(s) present; fruit a capsule or berry, but if berry, then with 4 or fewer seeds..................................................................................6 99

PAGE 111

100 6. Fruit indehiscent...........................................................................................................7 6. Fruit dehiscent..............................................................................................................9 7. Inflorescence a fascicle; plant (andro)dioecious; bracteoles present; anther glands spherical but sometimes absent; stipuliform structures absent.......................Mammea 7. Inflorescence a cyme or racemeto panicle-like cyme; flowers bisexual; bracteoles present or absent; anther glands elongate or absent; stipuliform structures present or absent............................................................................................................................8 8. Inflorescence a cyme; filaments fused into tube, with anthers arising from the apex and covering the entire inside surface of tube; anther glands elongate; stigma narrow (same thickness as style) and unlobed; placentation apical; leaf blade with latex/resin cavities and canals that cross secondary veins..........................................Endodesmia 8. Inflorescence a racemeor panicle-like cyme; filaments fused at base or appearing fasciculate, but not fused into a tube; anther glands absent; stigma expanded, with 2 or 3 poorly-developed lobes; placentation basal; leaf blade with latex/resin canals that alternate with secondary veins...........................................................Calophyllum 9. Inflorescence an axillary raceme (without terminal flower); ovules more than 15 per carpel; seeds plumose or not; placentation axile to intruded-parietal.................Marila 9. Inflorescence a cyme, panicle-like cyme, or cyme of 1-3 flowers (and terminal flower present); ovules 1-4 per carpel; seeds not plumose; placentation basal or axile to intruded-parietal......................................................................................................10 10. Inflorescence a panicle-like cyme; seeds with a narrow wing; placentation axile to intruded-parietal; ovary 3-loculate; stigmas 3, expanded.........................Haploclathra 10. Inflorescence a cyme or 1-3-flowered reduced cyme; seeds unwinged; placentation basal; ovary 1or 2-loculate; stigmas 2 or 4, expanded or narrow.............................11 11. Stigmas 2, expanded; inflorescence a reduced cyme of 1 or 2 flowers..............Mesua 11. Stigmas 2 or 4, narrow (same thickness as style).......................................................12 12. Style 4-parted; stamens more than 15; anthers open by longitudinal slits; calyx usually accrescent................................................................................................Kayea 12. Styles 2; stamens 15; anthers open by terminal pores; calyx not accrescent.................. ...............................................................................................................Poeciloneuron

PAGE 112

APPENDIX A TAXA USED FOR DNA ANALYSIS

PAGE 113

Table A-1. Taxa used for DNA analysis. 102 Taxon Collector and # Herbarium with voucher ITS matK rbcL Calophyllum brasiliense Camp. C. Notis 387 FLAS AY625643 Calophyllum goniocarpum P.F. Stevens F. Damon 318 MO, HUH AY625638 Calophyllum inophyllum L. C. Notis 391 FLAS AY625640 AY625043 AY625020 Calophyllum leleanii P.F. Stevens F. Damon 317 MO, HUH AY625641 AY625045 AY625022 Calophyllum soulattri Burm. f. F. Damon 320 MO, HUH AY625639 AY625044 AY625021 Calophyllum sp. nov. P.F. Stevens F. Damon 323 MO, HUH AY625642 Calophyllum vexans P.F. Stevens F. Damon 321 MO, HUH AY625637 Caraipa densifolia Mart. C. Grandez 16239 FLAS AY625626 AY625035 AY625012 Caraipa savannarum Kubitzki G. Aymard sn PORT AY625628 AY625034 Caraipa tereticaulis Tul. Vormisto 578 AAU, AMAZ AY625627 Caraipa utilis R. Vasquez C. Grandez 16240 FLAS AY625625 AY625036 AY625013 Caraipa valioli Paula C. Grandez 16243 FLAS AY625624 Clusia lanceolata Cambess. C. Notis 389 FLAS AY625054 Clusiella isthmensis Hammel M. Whitten 2657 FLAS AY625631 AY625042 AY625019 Endodesmia calophylloides Benth. H. Ndoma & J. Ntui 899 MO AY625610 AY625053 AY625030 Garcinia spicata Hook. f. C. Notis 388 FLAS AY625055 Haploclathra cordata R.Vsquez C. Grandez 16237 FLAS AY625630 AY625040 AY625017 Haploclathra paniculata Benth. C. Grandez 16246 FLAS AY625629 Hypericum tetrapetalum Lam. C. Notis 444 FLAS Kayea elmeri Merrill A. Kalat sn HUH AY625633 Kayea kunstleri King K. Larsen et al. 42186 MO AY625632 Kayea stylosa Thw. Kostermans 11106 HUH AY625634 AY625048 AY625025 Kielmeyera lathrophyton Saddi F. Feres sn UEC AY625623 AY625038 AY625015

PAGE 114

Table A-1. Continued 103 Kielmeyera petiolaris Mart. F. Feres 75 UEC AY625039 AY625016 Kielmeyera rosea Mart. Kubitzki et al. 97-5 HBG AY625622 AY625037 AY625014 Mahurea exstipulata Benth. Kubitzki et al. 97-27 HBG AY625621 AY625041 AY625018 Mammea americana L. C. Notis 392 FLAS AY625613 AY625052 AY625029 Mammea sessiliflora Planch. & Triana McPherson 18377 MO AY625611 AY625050 AY625027 Mammea siamensis (Miq.) T. Anders. P. Sweeney 1039 MO AY625614 AY625051 AY625028 Mammea usambarensis B.Verdcourt MO AY625612 AY625049 AY625026 Marila laxiflora Rusby M. Samaniego 124 STRI AY625618 AY625031 AY625009 Marila laxiflora Rusby van der Werff et al. 16246 MO AY625619 AY625033 Marila plumbaginea P.F. Stevens M. Nee 48390 MO AY625617 Marila racemosa Sw. B. Gunn 84 MO AY625615 AY625008 Marila tomentosa Poepp. & Endl. van der Werff et al. 16215 MO AY625620 AY625032 AY625010 Marila sp. van der Werff et al. 14476 MO AY625616 Mesua ferrea L. C. Notis 390 FLAS AY625635 AY625047 AY625024 Poeciloneuron indicum Bedd. U. Ghate sn FLAS AY625636 AY625046 AY625023

PAGE 115

APPENDIX B SPECIMENS EXAMINED Calophyllum brasiliense Camp.: Puerto Rico, Municipio de Loiza, Barrio Torrecilla Baja, 2.2-2.6 km due WSW of Punta Vacia Talega, elev. near sea level, G.R. Proctor 50003 (FTG); USA, Florida, Dade Co., Miami, Fairchild Tropical Garden, C. Notis 387 (FLAS). Calophyllum fibrosum P.F. Stevens: Madagascar, Toamasina, Nosy Mangabe, 5 km S of Maroantsetra, 1530S, 49E, elev. 0-330 m, G.E. Schatz & E. Carlson 2852 (MO); Madagascar, Toamasina, Masoala Peninsula, South trail, S of Androka River, 1538S, 49E, elev. 150-450 m, G.E. Schatz & G. Modeste 3049 (MO). Calophyllum inophyllum L.: USA, Florida, Dade Co., Miami, Fairchild Tropical Garden, W.T. Gillis 11103 (FTG); USA, Florida, Dade Co., Miami, Fairchild Tropical Garden, C. Notis 391 (FLAS). Caraipa densifolia Mart.: Peru, Dpto. Loreto, Prov. Maynas, Puerto Almendras, 7325W, 03S, elev. 122 m, R. Vasquez & N. Jaramillo 7659 (MO); Peru, Dpto. Loreto, Prov. Maynas, Quebrada Sucusari, afluente iquierdo del Rio Napo, 0320S, 72W, elev. 130 m, R. Vasquez & N. Jaramillo 11750 (MO). Caraipa savannarum Kubitzki: Venezuela, Amazonas, Rio Sipapo, entre Barranco Tonina y Cano Veneno, 0454-05N, 67-67W, A. Castillo 6980 (MO). Caraipa sp.: Venezuela, Amazonas, Dpto. Atures, Serrania Paru, planicie central, SW sector, 0425N, 65W, elev. 1200-1250 m, O. Huber & J. Rosales 4993 (MO). 104

PAGE 116

105 Clusia lanceolata Cambess.: Brazil, Guanabara, Restinga de Jacarepagua, Recreio dos Bandeirantes, B. Maguire, C.K. Maguire, & J. Murca Pires 44586; USA, Florida, Dade Co., Miami, Fairchild Tropical Garden, C. Notis 389 (FLAS). Clusiella axillaris (Engl.) Cuatr.: Venezuela, Amazonas, 3 km NE of San Carlos de Rio Negro, ca. 20 km S of confluence of Rio Negro and Brazo Casiquiare, 0156N, 67W, elev. 120 m, R.L. Liesner 6777 (MO); Brasil, Amazonas, Reserva Ducke, 0253S, 59W, M.T.V. Campos 548 (MO); Brasil, Amazonas, Sao Paulo de Olivenca, Ducke 1626 (HUH); Colombia, Amazonas, Vaupes, Rio Apaporis, Soratama (above mouth of Rio Kananari) and vicinity, 05N, 70W, elev. 900 ft, R.E. Schultes & I. Cabrera 16089 (HUH). Clusiella isthmensis B. Hammel: Costa Rica, Limon, Canton de Talamanca, Bratsi, Amubri, Alto Lari, Kivut., alfluente innominado del Rio Lari, margen izquierda, 0922N, 83W, elev. 1500m, G. Herrera 5422 (MO); Panama, Comarca de San Blas, along El Llano-Carti Rd., W of road 13.8 km to 15.8 km from Interamerican Hwy, 0919N, 78W, elev. 325 m, G. de Nevers, C. de Leon, & M. Marcus 3747 (MO); Panama, Prov. Panama, along El Llano-Carti Rd., 0915N, 79W, elev. 400 m, G. McPherson 7585 (MO); Panama, Comarca de San Blas, El Llano-Carti Rd., 19.1 km from Interamerican Hwy, 0919N, 78W, elev. 350 m, G. de Nevers 4300 (MO). Endodesmia calophylloides Benth.: Gabon, Estuaire, S of Estuaire du Gabon along Remboue River, British gas site, 0012S, 10E, elev. 10 m, G. McPherson 15100 (HUH); Gabon, 12-15 km NE of Asok, elev. 660 m, F.J. Breteler & J.J.F.E. de Wilde 123 (MO); Cameroun, 16 km on the road from Ebolowa to Minkok and 2 km to the left

PAGE 117

106 along a forest exploitation track, 0258N, 11E, elev. 650 m, J.J.F.E. de Wilde 7926 (MO). Garcinia spicata Hook. f.: USA, Florida, Dade Co., Miami, Fairchild Tropical Garden, C. Notis 388 (FLAS). Haploclathra leiantha (Benth.) Benth.: Brasil, Igapo, Foz do rio Marie, afluente do rio Negro, 0025S, 66W, M.L. Kawasaki 360 (MO). Haploclathra paniculata (Mart.) Benth.: Brasil, Amazonas, Reserva Ducke, Igarape do Tinga, Floersta de Baixio, Arvore, 0253S, 59W, A. Vicentini 1144 (MO); Brasil, Amazonas, Mun. de Maues, basin of Rio Apoquitaua, lower Rio Pacoval, 0348S, 58W, J.L. Zarucchi et al. 3181 (MO); Brasil, Secus Rio Negro, inter Barcellos et San Isabel, R. Spruce sn (HUH). Hypericum tetrapetalum Lam.: USA, Florida, Levy Co., FL 24, 1/8 mile west of Bronson city limit, C. Notis 444 (FLAS). Kayea borneensis P.F. Stevens: Sarawak, Sabal Kruim, Sabal F.R., 74th Mile Kuching/Simanggang Rd., 1st Division, Y.P. Ching 41133 (MO). Kayea elmeri Merrill: Kalimantan, 7 km along side road at km 57 from Sangai (S. Mentaya), Kab. Kotawaringin Timur, 0129S, 112E, P. Wilkie sn (HUH). Kayea kunstleri (King) Kosterm.: Malaya, Gunung Jerai, Kedah, road side, elev. 2100 ft, K.M. Kochummen sn (HUH); Malaya, Sungkop VJR Kedah, elev. 300 ft, Y.C. Chan sn (HUH); Thailand, Prov. Songkla, Koh Hong, 10020E, 07N, elev. 350-450 m, K. Larson et al. 42652 (MO); Thailand, Prov. Songkla, Khao Mot Daeng, Ko Hong, Hat Yai, 10020E, 07N, elev. 100 m, K. Larson et al. 42186 (MO).

PAGE 118

107 Kielmeyera coriacea Mart.: Bolivia, Dpto. Santa Cruz, Prov. Velasco, Estancia Flor de Oro, 7 km S of buildings, 1336S, 61W, elev. 190 m, M. Nee 41338 (MO); Brasil, Planalto do Brasil, Distrito Federal, ca. 10 km S of Brasilia, on road to Belo Horizonte, elev. 1200 m, H.S. Irwin, R. Souza, R. Reis dos Santos 8576 (HUH); Brasil, Planalto do Brasil, Serra do Rio Preto, Estado de Goias, 16S, 47W, elev. 1000 m, H.S. Irwin, R. Souza, R. Reis dos Santos 10395 (MO). Kielmeyera speciosa St. Hil.: Brasil, Minas Gerais, Municipio de Formoso, Parque Nacional Grande Serto Veredas, ca. 23 km da Sede da FUNATURA, 1519S, 46W, elev. 820 m, R.C. Mendona, M.L. Fonseca, F.C.A. Oliveira, & F. das Chagas e Silva 3473 (MO); Brasil, Gois, Municipio de Niquelndia, Estrada para Colinas do Sul, ca. 41 km de Niquelndia, 1423S, 48W, elev. 485 m, M.L. Fonseca, M.A. da Silva, D. Alvarenga, & A.J.V. Santos 1783 (MO); Brasil, Distrito Federal, proximo ao porto principal da Reserva Biologica de Aguas Emendadas, ca. 40 km NE of Brasilia, A.P. Silva 06 (MO). Mahurea exstipulata Benth.: Colombia, Caqueta, Araracuara, trocha a Yari, parcelas experimentales de regeneracion, 0025S, 72W, elev. 200 m, A. Gentry & M. Sanchez 64919 (MO); Venezuela, Estd. Bolivar, along road from Icabaru to Santa Elena, 16-17 km NE of Icabaru, ca. 0428N, 61W, elev. 735-790 m, T.B. Croat 54167 (MO); Venezuela, Municipio Gran Sabana, entre km 114 y 113 de la Carretera Santa Elena, Piedra de La Virgen, elev. 950 m, C. Benitez & W.G. DArcy 5274 (MO). Mammea americana L.: USA, Florida, Dade Co., Miami, Fairchild Tropical Garden, C. Notis 392 (FLAS).

PAGE 119

108 Mammea siamensis (Miq) T. Anders.: Thailand, Chi Avg Mai, 76/1 Sei 5, Sutep Rd., Chisne Mai, base of Doi Sutep, east side, elev. 350 m, J.F. Maxwell 92-70 (HUH); Thailand, Songkla, Maa Bee Yah Village, foothills of Klong Rhang Hill, elev. 50 m, J.F. Maxwell 86-70 (HUH); Thailand, Chiang Mai, Biology Dept., Faculty of Science, Chiang Mai Univeristy, elev. 350 m, Franco Paleo 30 (HUH). Mammea subsessifolia (Hochr.) Kosterm.: Toamasina, Masoala Peninsula, ca. 11 km S of Ambanizana, 1543S, 49-58E, elev. 0-5 m, G.E. Schatz & G. Modeste 3099 (HUH); Country Unknown, Est (Nord), Fort dAnalamateza au Sud dAntsirabe-Nord, R. Capuron sn (HUH). Marila laxiflora Rusby: Panama, Prov. of Panama, in high ridges of the Serrania de Maji, S of the Choco village of Ipeti, ca. 5 hours walk from the village, 0845N, 78W, elev. 500-650 m, S. Knapp, R. foster, J. Mallet, & M. Huft 4476; Panama, Prov. of Panama, Cerro Azul, R.A. Lopez 38; Panama, San Blas-Panama border, on Kuna divide trail west of Llano-Carti Rd, 0920N, 79W, elev. 250 m, G. McPerson 11882; Panama, Canal Zone, Barro Colorado Island, upper Allee creek, R. Foster 2820. Marila plumbaginea P.F. Stevens: Bolivia, Dpto. La Paz, Prov. Larecaja, Copacabana, ca. 10 km S of Mapiri, elev. 850-950 m, B.A. Krukoff 11057; Bolivia, Dpto. Santa Cruz, Prov. Ichilo, ca. 2 km W of Villa San German on highway from Buena Vista to Rio Ichilo, 1721S, 64W, elev. 275 m, M. Nee 48390; Ecuador, Prov. Napo, Canton Tena, Estacion Biologica Jatun Sacha, Rio Napo, 8 km al E de Misahualli, 0105S, 77W, elev. 450 m, W. Palacios 3273. Marila racemosa Sw.: Dominica, on road bordering Imperial Rd., Sylvania to Mahaut River, elev. 549 m, W.H. Hodge 532; Saint Lucia, Quarter of Castries, vicinity of

PAGE 120

109 Piton Flore, elev. ca. 1500 ft, G.L. Webster, J.R. Ellis, & K. Miller 9302; Saint Lucia, Piton Flore, elev. 1200 ft, V. Slane, L. JnPierre, C. Jimber 489. Mesua ferrea L.: Malaya, Kepong F.R.I. area, Selangor, elev. 350 ft., M.B. Sider 13223; Malaya, Selangor, University of Malaya Campus, road to 5th Res. College, elev. 100 ft., B.C. Stone 11039; Puerto Rico, Villa leon, Bayamon, Toa Alta Rd., L.R. Holdridge 244; USA, Florida, Dade Co., Miami, Fairchild Tropical Garden, Plot 42, W.B. Zomlefer 748. Neotatea duidae (Kobuski & Steyermark) A.L. Weitzman & P.F. Stevens: Venezuela, Terr. Federal de Amazonas, Dpto. Atabapo cumber del Cerro Duida, 0322N, 65W, elev. 1640 m, O. Huber 13444. Neotatea longifolia (Gleason) Maguire: Venezuela, Cerro Duida, Rio Cunucunuma, Amazonas, along North Escarpment above Culebra, elev. 1400 m, B. Maguire, R.S. Cowan, & J.J. Wurdack 29585; Venezuela, Dpto. Atures, Serrania Paru, planicie central, SW sector, 0425N, 65W, elev. 1200-1250 m, P.E. Berry, O. Huber, & J. Rosales 4976. Neotatea neblinae Maguire: Venezuela, Cerro de la Neblina, Rio Yatua, south slope of Cumbre Camp Cao toward Cao Grande, elev. 1500 m, B. Maguire, J.J. Wurdack, & G.S. Bunting 37321; Venezuela, Territorio Federal Amazonas, Dpto. Rio Negro Cerro de Neblina, Puerto Chimo Camp on Rio Mawarinuma and up north slope of canyon, 5 km E of La Neblina Base Camp by air, 0050N, 66W elev. 150-1800 m, R. Liesner & C. Brewer 15874.

PAGE 121

110 Poeciloneuron indicum Bedd.: India, Palghat District, Kerala State, Vattapparai Shola, Sirurani Slopes, E. Vajrarelu 77764; India, Shimoga District, Kashataka State, 13.3N, 75.7E, elev. 650 m, U. Ghate sn.

PAGE 122

LIST OF REFERENCES Anderson, T. 1874. Guttiferae. In J.D. Hooker (ed.), Flora of British India 1: 209464. London: Reeve. Armbruster, W.S. 1984. The role of resin in angiosperm pollination: ecological and chemical considerations. American Journal of Botany 71: 1149-1160. Bentham, G. 1863. Flora Australiensis Vol. 1. London: Lovell Reeve. Chase, M.W. and A.V. Cox. 1998. Gene sequences, collaboration, and analysis of large data sets. Australian Systematic Botany 11: 215-229. Cronquist, A. 1981. An integrated system of classification of the flowering plants. New York: Columbia University Press. Doyle, J.J. and J.L. Doyle. 1987. A rapid DNA isolation procedure for small amounts of fresh leaf tissue. Phytochemical Bulletin 19: 11-15. Endress, P.K. 1994. Diversity and evolutionary biology of tropical flowers. Cambridge: Cambridge University Press. Engler, A. 1925. Guttiferae. In A. Engler and K. Prantl, Die natrlichen Pflanzenfamilien 21: 154-237. Field, B.S. 1978. Theaceae. Pp. 82-83. In: V.H. Heywood (ed.). Flowering plants of the world. New York: Mayflower Books. Gandolfo, M.A., K.C. Nixon, and W.L. Crepet. 2002. Triuridaceae fossil flowers from the Upper Cretaceous of New Jersey. American Journal of Botany 89: 19401957. Greer, M. 2001. Single calanolide A doses safe for humans (HIV inhibitor being developed from Calophyllum lanigerum, Malaysian rain forest tree). June 18. AIDS Weekly. Gustafson, M.H.G., V. Bittrich, and P.F. Stevens. 2002. Phylogeny of Clusiaceae based on rbcL sequences. International Journal of Plant Sciences, 163: 1045-1054. Hammel, B.E. 1999. Two new species of Clusiella (Clusiaceae) with a synopsis of the genus. Novon 9: 349-359. 111

PAGE 123

112 Hasegawa, M., K. Kishino, and T. Yano. 1985. Dating the human-ape splitting by a molecular clock of mitochondrial DNA. Journal of Molecular Evolution 32: 443445. Hermsen, E.J., M.A. Gandolfo, K.C. Nixon, and W.L. Crepet. 2003. Divisestylus gen. nov. (aff. Iteaceae), a fossil saxifrage from the Late Cretaceous of New Jersey, USA. American Journal of Botany 90: 1373-1388. Howard, R.A. 1974. The stem-node-leaf continuum of the Dicotyledoneae. Journal of the Arnold Arboretum of Harvard University 55: 125-181. Huelsenbeck, J.P. 2000. MrBayes: Bayesian inference of phylogeny. Distributed by the author. Department of Biology, University of Rochester. Hutchinson, J. 1959. The families of flowering plants (2nd ed.) Vol. 1: Dicotyledons arranged according to a new system based on their probable phylogeny. England: Oxford at the Clarendon Press. Hutchinson, J. 1969. Evolution and phylogeny of flowering plants; Dicotyledons: facts and theory. London and New York: Academic Press. Jenner, R.A. 2004. Accepting partnership by submission? Morphological phylogenetics in a molecular millennium. Systematic Biology 53: 333-342. Johnson, L.A. and D.E. Soltis. 1995. Phylogenetic inference in Saxifragaceae sensu stricto and Gilia (Polemoniaceae) using matK sequences. Annals of the Missouri Botanical Garden 82: 149-175. Judd, W.S., C.S. Campbell, E.A. Kellogg, P.F. Stevens, and M.J. Donoghue. 2002. Plant Systematics: A Phylogenetic Approach (2nd ed.). Sunderland, MA: Sinauer Associates, Inc. Kostermans, A.J.G.H. 1969. Kayea Wall. and Mesua L. (Guttiferae). Reinwardtia 7: 425-431. Kron, K.A. and W.S. Judd. 1997. Systematics of the Lyonia group (Andromedeae, Ericaceae) and the use of species as terminals in higher-level cladistic analyses. Systematic Botany 22: 479-492. Kron, K.A., W.S. Judd, P.F. Stevens, D.M. Crayn, A.A. Anderberg, P.A. Gadek, C.J. Quinn, and J.L. Luteyn. 2002. Phylogenetic classification of Ericaceae: molecular and morphological evidence. The Botanical Review 68: 335-423. Kubitzki, K. 1978. The botany of the Guyana Highland. X. Caraipa and Mahurea (Bonnetiaceae). Memiors of the New York Botanical Garden 29:82-128.

PAGE 124

113 Kubitzki, K. 1989. The ecogeographical differentiation of Amazonian inundation forests. Plant Systematics and Evolution 162: 285-304. Maddison, D.R. and W.P. Maddison. 2001. MacClade 4.05. Sunderland, MA: Sinauer Associates. Maguire, B. 1972. The botany of the Guyana Highland. IX. Bonnetiaceae. Memiors of the New York Botanical Garden 23:131-165. Oliveira, C.M.A. de, A.M. Porto, V. Bittrich, I. Vencato, and A.J. Marsaioli. 1996. Floral resins of Clusia spp.: chemical composition and biological function. Tetrahedron Letters 37: 6427-6430. Planchon, J. and J. Triana. 1860. Mmoire sur la famille des guttifres. Annales des Sciences Naturelles Botanique sr. 4, 13: 306-376; 14: 226-367; 15: 240-319; 16: 263-308. Posada, D. and K. A. Crandall. 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14 (9): 817-818. Rambaut, A. 1996. SeAl: Sequence Alignment Editor. http://evolve.zoo.ox.ac.uk/. Ribeiro, J.E.L. da S., M.J.G. Hopkins, A. Vicentini, C.A. Sothers, M.A. da S. Costa, J.M. de Brito, MAD de Souza, L.H.P. Martins, L.G. Lohmann, P.A.C.L. Assuno, E. da C. Pereira, C.F. da Silva, M.R. Mesquita, and L.C. Procpio. 1999. Flora da reserva Ducke: guia de identicao das plantas vasculares de uma floresta de terra-firme na Amaznia central. Manaus: INPA. Robson, N.K.B. 1976. Guttiferae. In: Flora of Taiwan Editorial Committee (ed.), Flora of Taiwan Vol. 2. Taiwan: Epoch Publishing Co. Robson, N.K.B. 1978. Guttiferae. Pp. 85-87. In: V.H. Heywood (ed.), Flowering plants of the world. New York: Mayflower Books. Scotland, R.W., R.G. Olmstead, and J.R. Bennett. 2003. Phylogeny reconstruction: the role of morphology. Systematic Biology 52: 539-548. Smith, J.J. 1920. Plantae novae vel criticae ex herbario et horto Bogoriensis 1. Bulletin. du Jardin botanique de Buitenzorg 3: 390-410. Soltis, D.E., P.S. Soltis, M.W. Chase, M.E. Mort, D.C. Albach, M. Zanis, V. Savolainen, W.H. Hahn, S.B. Hoot, M.F. Fay, M. Axtell, S.M. Swensen, L.M. Prince, W.J. Kress, D.C. Nixon, and J.S. Farris. 2000. Angiosperm phylogeny inferred from 18S rDNA, rbcL, and atpB sequences. Botanical Journal of the Linnean Society 133: 381-461.

PAGE 125

114 Soltis, D.E., P.S. Soltis, T.G. Collier, and M.L. Edgerton. 1991. Chloroplast DNA variation within and among genera of the Heuchera group (Saxifragaceae): evidence for chloroplast transfer and paraphyly. American Journal of Botany 78: 1091-1112. Soltis, D.E., P.S. Soltis, M.E. Mort, M.W. Chase, V. Savolainen, S.B. Hoot, and C.M. Morton. 1998. Inferring complex phylogenies using parsimony: an empirical approach using three large DNA data sets for angiosperms. Systematic Biology 47: 32-42. Stevens, P.F. 1980. A revision of the Old World species of Calophyllum (Guttiferae). Journal of the Arnold Arboretum 61: 117-690. Stevens, P.F. 1991. Character states, morphological variation, and phylogenetic analysis: a review. Systematic Botany 16: 553-583. Stevens, P.F. 2001. Angiosperm Phylogeny Website. Version 4, May 2003 [and more or less continuously updated since]. http://www.mobot.org/ MOBOT/ research/ APweb/. Stevens, P.F. In press. Clusiaceae In: K. Kubitzki (ed.). The families and genera of vascular plants. Vol. 3. Flowering plants, dicotyledons, dilleniid families. Heidelberg: Springer. Swofford, D.L. 2000. PAUP*. Phylogenetic Analysis Using Parsimony (*And Other Methods). Version 4. Sunderland, Massachusetts: Sinauer Associates. Takhtajan, A. 1997. Diversity and classification of flowering plants. New York: Colombia University Press. Vasquez, R. 1993. Una Nueva Haploclathra (Clusiaceae) de la Amazona Peruana. Novon 3: 499-501. Watrous, L.E. and Q.D. Wheeler. 1981. The out-group comparison method of character analysis. Systematic Zoology 30: 1-11. Weitzman, A.L. and P.F. Stevens. 1997. Notes on the circumscription of Bonnetiaceae and Clusiaceae, with taxa and new combinations. BioLlania, Edicin Especial 6: 551-564. Wen, J. F. and E. A. Zimmer. 1996. Phylogeny and biogeography of Panax L. (the ginseng genus, Araliaceae): inferences from ITS sequences of the nuclear ribosomal DNA. Molecular Phylogenetics and Evolution 6: 167-177. Wilson, P.G., M.M. OBrien, P.A. Gadek, and C.J. Quinn. 2001. Myrtaceae revisited: a reassessment of infrafamilial groups. American Journal of Botany 88: 2013-2025.

PAGE 126

115 Yang, Z. 1994. Maximum likelihood phylogenetic estimation from DNA sequences with variable rates over sites: approximate methods. Journal of Molecular Evolution 39: 306-314.

PAGE 127

BIOGRAPHICAL SKETCH Christine H. Notis was born on June 13, 1979, in Des Moines, Iowa. Christine graduated from Johnston High School in 1997, after which she attended Iowa State University in Ames, Iowa. At Iowa State, Christine was interested in biology, and became fascinated with plants while taking an introductory biology class from Lynn G. Clark. She gained her first research experience in the lab of Jonathan F. Wendel, where she harvested ovules from Gossypium spp. as part of a larger project on the seed trichomes of these plants. Christine became interested in fern gametophyte ecology, and decided to complete her honors project on the longevity and swimming distance of fern sperm (under the direction of Donald R. Farrar). She received her Bachelor of Science degree (with honors) in botany from Iowa State in the spring of 2001. In the fall of 2001, Christine came to the University of Florida to start a masters degree under the direction of Walter S. Judd and Douglas E. Soltis. Her project focused on a phylogeny of Kielmeyeroideae (Clusiaceae) using molecular and morphological data. Christine had first become intrigued by Clusiaceae when she left Iowa for a summer to take Walter Judds Tropical Botany class at the National Tropical Botanical Garden-Kampong and Fairchild Tropical Botanic Garden in Miami. While a masters student at the University of Florida, Christine was a teaching assistant in several courses, including introductory biology, introductory botany, plant diversity, and plant taxonomy. Christine received her Master of Science degree in August, 2004. 116