A Systematic Revision of Passiflora Section Dysosmia (Passifloraceae)

Home , Arida (plant), Capanemia, Los Cabos Municipality, Passiflora, Passifloraceae, Passiflora foetida

A dissertation presented to

the faculty of

the College of Arts and Sciences of Ohio University

In partial fulfillment

of the requirements for the degree of

Doctor of Philosophy

Harlan T. Svoboda

August 2018

This dissertation titled

A Systematic Revision of Passiflora Section Dysosmia (Passifloraceae)

HARLAN T. SVOBODA

has been approved for

the Department of Environmental and Plant Biology

and the College of Arts and Sciences by

Harvey E. Ballard, Jr.

Professor of Environmental and Plant Biology

Joseph Shields

Interim Dean, College of Arts and Sciences 3

ABSTRACT

SVOBODA, HARLAN T., Ph.D., August 2018, Environmental and Plant Biology

A Systematic Revision of Passiflora Section Dysosmia (Passifloraceae)

Director of Dissertation: Harvey E. Ballard, Jr.

One of the most taxonomically difficult groups in the passionflower genus

(Passiflora, Passifloraceae) is section Dysosmia. For centuries, the taxa belonging to this group have proven difficult to classify or even name. Various authors over the years have attempted to tackle the complex task of describing the diversity in this section and have done so with varying degrees of success; often times shuffling names between species or recognizing taxa at differing ranks.

Ellsworth Killip, in his 1938 monograph of the family, introduced a litany of varietal names for one of the most notoriously complicated species, Passiflora foetida, in an attempt to make sense of the staggering degree of morphological variation. His work remains a foundation for the systematics of Dysosmia as well as that of the rest of the genus to this day. The section was most recently revised by John Vanderplank in 2013, who took a more conservative approach by synonymizing many names and renaming others to more accurately reflect our changing taxonomic hypotheses. At the outset of this project 21 species and nine varieties were recognized and accepted in the section.

The primary goal of this research was to use modern techniques and provide new evidence to aid in the delimitation of the taxa within section Dysosmia. An exhaustive nomenclatural study was conducted in order to ascertain the extent to which the sectional 4 diversity has already been named. Dozens of names were subsequently typified and corrected in order to bring them into compliance with the current Code of Nomenclature.

Intensive morphological studies were performed to investigate the use of macro- and micromorphological features—some of which never before assessed—for the delimitation of taxa. Robust statistical analyses revealed at least three species assemblages or lineages within section Dysosmia that were then the focus of subsequent work, particularly a conspicuous red-fruited complex. Geometric morphometric descriptors were also pursued for the purposes of classification and show promise for objective classification of an otherwise unruly group. Molecular approaches were explored to test the hypotheses of monophyly and relationships within the section but were not successful.

Finally, climatic and soil data were employed to investigate a complex of taxa and delimit distinct entities at the rank of species. These data were also informative in describing habitat characteristics of the species. This is the first time that evidence of this kind has been used for taxonomic delimitation in Passiflora.

The culmination of this research was a comprehensive synopsis of a portion of the section, comprising all of the red-fruited taxa, in which several new species and combinations were proposed. The current understanding of section Dysosmia brings the number of recognized taxa to 31 species and five putative varieties.

DEDICATION

This dissertation is dedicated to my partner, Ricky, for his unwavering support and love

throughout the adventure of grad school. I also dedicate this to my family who have

believed in my every endeavor from the very start.

ACKNOWLEDGMENTS

First and foremost, I thank my advisor and mentor, Dr. Harvey Ballard, Jr., for his steadfast support throughout every aspect of my development as a researcher, scholar, and educator. I would also like to thank my current committee members, Drs. Morgan

Vis, Rebecca Snell, Alycia Stigall, and John MacDougal as well as previous committee members, Drs. Elizabeth Hermsen and Shawn Kuchta. I also thank several Passiflora specialists who have guided and influenced my research, including Drs. Christian

Feuillet, Peter Jørgensen, Shawn Krosnick, John MacDougal, and John Vanderplank. I express my gratitude to my current and former labmates: Ben Gahagen, Jen Hastings,

Anne Sternberger, Bina Swasta Sitepu, and Bethany Zumwalde for their support. I also appreciate the work of Harold Blazier for growing plants for my research, and to Zach

Parinello and Todd Roth for assistance with digitizing and databasing herbarium specimens. I extend my thanks to the Ohio University Graduate Student Senate for funding my research with two Original Work Grants and several Travel Awards to attend scientific meetings, the Department of Environmental and Plant Biology for travel funding, the Ohio University Graduate College for a Student Enhancement Award, and the Ohio Center for Ecology and Evolutionary Studies for a research fellowship. I am indebted to the following herbaria for access to their collections: A, AMD, ASU, B, BAR,

BHCB, BM, BOLO, BRIT, C, CAS, CHIC, DES, DUKE, E, ECON, F, FLAS, FTG, G,

GENT, GH, GOET, HA, HCIB, HUA, HUEFS, IND, INPA, K, L, LINN, LISC, LSU,

MA, MBM, MEXU, MICH, MIN, MISSA, MPU, MO, NY, P, PH, R, RB, RSA, S, SD,

SP, TEX-LL, U, UC, UCR, UPCB, UPRRP, US, USF, UTC, VT, W, WAG, and YU. 7

TABLE OF CONTENTS

Page

Abstract ...... 3 Dedication ...... 5 Acknowledgments ...... 6 List of Tables ...... 10 List of Figures ...... 11 Chapter 1: Introduction ...... 14 Morphology ...... 16 Current Systematics ...... 20 Phylogenetic Relationships ...... 25 Geographic Distribution...... 26 Chapter 2: Typifications and Nomenclatural Notes in Passiflora Section Dysosmia (Passifloraceae) ...... 28 Abstract ...... 28 Introduction ...... 29 Nomenclatural Treatment ...... 30 Typifications ...... 30 Nomenclatural Clarification ...... 43 Ambiguous Names ...... 51 Invalid and Illegitimate Names ...... 55 Acknowledgements ...... 56 Chapter 3: Phenetic and Cladistic Studies Help Clarify Species Assemblages in Passiflora Section Dysosmia (Passifloraceae) ...... 63 Abstract ...... 63 Introduction ...... 64 Methods...... 66 Results and Discussion ...... 68 Conclusions ...... 70 Acknowledgements ...... 71 Chapter 4: Leaf Geometric Morphometrics ...... 85 Introduction ...... 85 8

Comprehensive Methods for Leaf Geometric Morphometric Analyses ...... 86 Abstract ...... 87 Background ...... 87 Materials and Reagents ...... 89 Equipment ...... 89 Software ...... 89 Procedure ...... 90 Data Analysis...... 117 Notes ...... 117 Acknowledgments ...... 118 Systematic Explorations in Section Dysosmia Using Leaf Geometric Morphometric Shape Descriptors ...... 120 Methodology...... 120 Results and Discussion...... 122 Acknowledgements ...... 129 Chapter 5: Contributions Toward Understanding the Biodiversity of Passiflora in North America: Updates and a New Combination From the Baja California Peninsula, Mexico and Vicinity ...... 130 Abstract ...... 131 Introduction ...... 131 Material and Methods ...... 134 Taxon Sampling and Georeferencing ...... 134 Environmental Data ...... 135 Geometric Morphometrics ...... 137 Results ...... 137 Environmental Analyses ...... 137 Leaf Shape Analysis ...... 139 Peninsular Endemism ...... 139 Discussion ...... 140 Reevaluation of the Passiflora arida Complex ...... 140 Habitat Characterization and Conservation...... 142 Ecology, Distribution, and Taxonomic Treatment ...... 143 Key to the Baja California Peninsula Species of Passiflora ...... 151 Acknowledgements ...... 151 9

Chapter 6: Molecular Phylogenetics ...... 162 Introduction ...... 162 Methods...... 162 Taxon Sampling and Outgroup Selection ...... 163 DNA Extraction and Purification ...... 163 Marker Selection and Amplification ...... 165 Sequencing ...... 167 Results ...... 167 Discussion and Future Directions ...... 168 Acknowledgements ...... 169 Chapter 7: Synopsis of the Red-Fruited Members of Passiflora section Dysosmia ...... 170 Introduction ...... 170 Distribution and Endemism ...... 171 Taxonomic History, Nomenclature, and Classification ...... 172 Key to the Red-Fruited Dysosmia Taxa ...... 173 Taxonomic Treatment ...... 175 Accepted and New Taxa ...... 176 Excluded Names ...... 211 Discussion ...... 211 Chapter 8: Overall Conclusions ...... 213 References ...... 215 Appendix A: Permissions to Reproduce Published Articles ...... 234 Appendix B: Herbarium Specimens and iNaturalist Observations Used in the Study from Chapter 5 ...... 237 Appendix C: Samples Used in the Molecular Study ...... 247 Appendix D: Conditions Used in the PCR Trials ...... 250

LIST OF TABLES

Page

Table 1-1. The historical and current treatments of the Dysosmia group...... 21 Table 3-1. List of specimens used in this study and their codes used in the phenetic analysis ...... 81 Table 3-2. The characters and their states used in the phenetic analysis ...... 84 Table 5-1. Climatic and soil variables obtained from ArcMap used in this study ...... 153 Table 5-2. Results of linear discriminant analyses (LDA) of environmental data based on different combinations and groupings of taxa ...... 154 Table 5-3. Comparison of features used to distinguish Passiflora arida from P. pentaschista ...... 155 Table 6-1. Primers used in the molecular study ...... 165

LIST OF FIGURES

Page

Figure 1-1. The shapes of some representative Passiflora leaves ...... 15 Figure 1-2. Characteristic morphological features of section Dysosmia ...... 17 Figure 1-3. A sample of the floral diversity found in section Dysosmia ...... 18 Figure 1-4. A comparison of the two major classifications of the Dysosmia group ...... 24 Figure 2-1. Designated lectotype of Passiflora baraquiniana...... 58 Figure 2-2. Designated neotype of Passiflora ciliata...... 59 Figure 2-3. Designated lectotype of Passiflora gossypiifolia ...... 60 Figure 2-4. Designated lectotype of Passiflora hispida ...... 61 Figure 2-5. Lectotype of Passiflora foetida ...... 62 Figure 3-1. A sampling of the variation in leaf shape and lobing found in section Dysosmia ...... 73 Figure 3-2. Geographical distribution of the OTUs used in this study ...... 74 Figure 3-3. Representative structures seen in Dysosmia taxa ...... 75 Figure 3-4. A representation of the leaf measurements used in this study ...... 76 Figure 3-5. Ordinations resulting from a principal coordinates analysis (PCoA) ...... 77 Figure 3-6. Ordination resulting from a non-metric dimensional scaling (NMDS) analysis ...... 78 Figure 3-7. Phenogram showing the clustering of two major groups ...... 79 Figure 3-8. Fifty percent majority rule consensus cladogram from a parsimony analysis in PAUP ...... 80 Figure 4-1. The ImageJ menu bar with the ‘Point selection’ tool selected ...... 91 Figure 4-2. Placement of landmarks on some representative leaves ...... 91 Figure 4-3. Example of the ‘Results’ window produced by ImageJ ...... 92 Figure 4-4. Example format of the ‘master spreadsheet’ ...... 93 Figure 4-5. The reformatted ‘master spreadsheet’ (‘reformatted.txt’) as seen in Excel ... 95 Figure 4-6. Landmarks are checked for accuracy using the R package ggplot ...... 96 Figure 4-7. PCA ordination resulting from a Generalized Procrustes Analysis using an example landmark dataset ...... 98 Figure 4-8. The ImageJ menu bar with the ‘Wand (tracing)’ tool selected ...... 99 Figure 4-9. The ‘Tools’ menu in Photoshop ...... 101 12

Figure 4-10. An example of areas edited out using the ‘Brush Tool’ ...... 103 Figure 4-11. Selection of .bmp files for chain coding in the software SHAPE ...... 105 Figure 4-12. The ‘Config’ tab of the ChainCoder program ...... 106 Figure 4-13. The CHC2NEF window ...... 108 Figure 4-14. Orienting leaf image chain code to normalized elliptical Fourier descriptors ...... 109 Figure 4-15. The ‘NEF File Information’ window ...... 110 Figure 4-16. The PrinComp program toolbar ...... 111 Figure 4-17. The ‘Information of Principal Component Analysis’ window ...... 112 Figure 4-18. Select report information for printing...... 113 Figure 4-19. The ‘Reconstruct Contours’ window...... 114 Figure 4-20. Eigenleaves visualized in the ‘PrinPrint’ program ...... 115 Figure 4-21. Linear regression of Aspect Ratio (AR) and Circularity (Circ.) data ...... 116 Figure 4-22. Ordination representing a generalized Procrustes analysis (GPA) of landmark data for all of the taxa in section Dysosmia ...... 123 Figure 4-23. Ordination of outline (elliptical Fourier descriptors; EFD) data for section Dysosmia ...... 124 Figure 4-24. Landmark overlays and outline overlays of leaves ...... 125 Figure 4-25. Leaf overlays of the Passiflora foetida complex ...... 126 Figure 4-26. Ordinations representing the leaf shape variation of Passiflora bahamensis ...... 127 Figure 4-27. Ordination of the Passiflora ciliata complex based on landmark data ...... 128 Figure 5-1. Map of the Baja California Peninsula and vicinity ...... 156 Figure 5-2. Some of the Passiflora taxa found on the Baja California Peninsula ...... 157 Figure 5-3. Six homologous landmarks representing the primary venation and overall leaf shape ...... 157 Figure 5-4. The resulting ordination (Axes 1 and 2) from a linear discriminant analysis (LDA) of the environmental data ...... 158 Figure 5-5. Representation of a linear discriminant analysis (LDA) of the Passiflora arida complex based on environmental data ...... 159 Figure 5-6. Distributions of the Passiflora species native to the Baja California Peninsula ...... 160 Figure 5-7. Illustration of Passiflora arida and P. pentaschista ...... 161 13

Figure 7-1. Leaves from the proposed isotype of Passiflora aurea, sp. nov...... 177 Figure 7-2. Leaf from the proposed holotype of Passiflora subauriculata, sp. nov...... 202 Figure 7-3. Comparison of the major treatments of the red-fruited members in section Dysosmia ...... 212

CHAPTER 1: INTRODUCTION

The family Passifloraceae Juss. ex Roussel belongs to the order Malpighiales

Juss. ex Bercht. & J.Presl along with closely related families the Salicaceae Mirb.

(willow family) and Violaceae Batsch (violet family) (APG IV, 2016). The

Passifloraceae is comprised of approximately 750 species divided into 17 genera that are distributed throughout the tropics in both the Old and New World (Feuillet &

MacDougal, 2006). The most species rich of the Passifloraceae genera, Passiflora L., includes over 570 species of passionflowers with a distribution primarily in Central and

South America (Ocampo Pérez & Coppens d’Eeckenbrugge, 2017). Members of

Passiflora show an array of unusual morphological characters, such as extrafloral (often petiolar) nectaries, an elevated androgynophore, and a corona of one to many series of filaments, which together uniquely define the genus. With few exceptions, species also have a 5-merous corolla and calyx, three fused carpels, and five stamens (Ulmer &

MacDougal, 2004). Members of the genus also express substantial variation in leaf form, leading Killip (1938) to state that few, if any, other plant groups encompass as much diversity in leaf morphology as Passiflora (see Figure 1-1).

Figure 1-1. The shapes of some representative Passiflora leaves (Figure 2 from Chitwood

& Otoni, GigaScience, 6, 2017).

Three major monographs have treated the genus in great detail, the earliest being that of Masters (1871, 1872), followed more than 50 years later by Harms (1925), and 16 finally the most complete work by Killip (1938). The last is the “reigning bible of passionflowers,” according to Ulmer and MacDougal (2004), and provides the foundation for our modern understanding of Passiflora systematics. The infrageneric classification has since been modified and updated (see Feuillet & MacDougal, 2003; Ulmer &

MacDougal, 2004) to reflect newly discovered species and accommodate molecular phylogenetic evidence that has become available in the decades since Killip’s revision.

Whereas Killip segregated 22 subgenera with two series, Feuillet and MacDougal (2003) separated the genus into four subgenera (Astrophea (DC.) Mast., Decaloba (DC.) Rchb.,

Deidamioides (Harms) Killip, and Passiflora), each with several supersections, sections, and series within them. Subgenus Passiflora is further divided into six supersections, one of which is Stipulata Feuillet & J.M.MacDougal containing the section of interest,

Dysosmia DC. This section, originally introduced by A. P. de Candolle (1822), was one of seven sections he proposed for Passiflora.

Morphology

Dysosmia as a group has been distinguished from other Passiflora groups in having: 1) deeply cleft, often glandular, stipules; 2) highly filiform, pinnatisect, and often glandular involucral bracts; 3) two to four series of filaments in the corona; 4) a non- plicate operculum with denticulate margins; and 5) leaves often bearing glandular trichomes on the abaxial and/or adaxial surfaces (Figure 1-2). Although often scarce on herbarium sheets, flower morphology has been used in several cases of taxonomic circumscription—mostly from fresh specimens—and is clearly variable among the taxa of Dysosmia (see Figure 1-3). 17

Figure 1-2. Characteristic morphological features of section Dysosmia. A. Pyriform long- stipitate glands of Passiflora arizonica. B. Lepidote glands of P. chrysophylla. C. Highly dissected stipules of P. arizonica bearing glands. D. Stipules of P. vellozoi. E. Highly pinnatisect involucral bracts of P. clathrata. F. Involucral bracts of P. lepidota. Images taken from herbarium specimens by H. Svoboda. 18

Figure 1-3. A sample of the floral diversity found in section Dysosmia. A. Passiflora vellozoi (photo: passiflorae.fr). B. P. bahamensis (photo: D. Goldman). C. P. pectinata (photo: J. MacDougal). D. P. foetida (s.l.) (photo: R. Boender). E. P. foetida var. lanuginosa (photo: J. MacDougal). F. P. foetida (s.l.). G. P. sublanceolata (photo: R. Boender). H. P. arida (photo: J. Ochoa). I. P. urbaniana (photo: R. Boender).

The present concepts of species relationships in Dysosmia come from incomplete morphological observations based on limited herbarium or field material. Gross morphology has played an important role in delimiting species and infraspecific taxa (see

Masters, 1871; Killip, 1938), but very little, if any, emphasis has been put on 19 micromorphological features such as type and structure of glands or trichomes. Most of the historical descriptions (older protologues) of taxa provided no information on finer morphological features such that they are of little help in separating taxa at this level. A relatively thorough analysis using the same gross morphological traits examined by others was conducted by Killip (1938), but he, too, mostly ignored micromorphological details as a means of delimiting species or indicating relationships. An examination of fine-scale details of both vegetative and floral morphology will reveal new important taxonomic characters to distinguish taxa and better understand phylogenetic relationships in the section.

Glandular trichomes, a hallmark feature of Dysosmia, have thus far been overlooked for potential taxonomic characters. Although abundant and conspicuous, many taxonomists have ignored these structures as potential systematically informative features. Several studies attempting to document coarse glandular trichome structural diversity (Durkee, 1982, 1983; Durkee, Baird, & Cohen, 1984; Jáuregui, García, & Pérez,

2002; Lemos, da Costa Silva, & Flavia de Albuquerque Melo-de-Pinna, 2017) as well as exudate composition (Baker, Opler, & Baker, 1978; Radhamani, Sudarshana, &

Krishnan, 1995) using a handful of Passiflora species have been conducted, but no study to date has looked at more than one to a few species within Dysosmia. Similar to the investigations by Nogueira, Ottra, Guimarães, Machado, and Lohmann (2013) into glandular trichome variation in Bignoniaceae, these fine-scale structures will likely provide new evidence to inform trait evolution and phylogenetic relationships in the group. 20

Pubescence is generally regarded as a typical trait of Dysosmia, but several species and varieties are in fact consistently glabrous. Interestingly, every taxon that is glabrous also bears red fruit, a condition that is exceptionally rare within Passiflora (J.

MacDougal, personal communication). Although a small number of pubescent taxa also have red fruit, the combination of glabrous features and red fruit likely constitutes a unique group within Dysosmia, which warrants further systematic investigation.

Current Systematics

Since its circumscription nearly 200 years ago, the group known as Dysosmia has undergone many changes and taxonomic expansions. When first described (de Candolle,

1822), this section included three species P. foetida L., P. ciliata Aiton, and P. hibiscifolia Lam. that were known at the time. Since then taxonomists have recognized

Dysosmia in at various ranks, as is evident in Table 1-1, at one point even raising its status to a separate genus (Roemer, 1846). The most recent evaluation of Dysosmia by

Vanderplank (2013), currently at the rank of section within supersect. Stipulata, subgen.

Passiflora, expanded the group to formally include five species from a segregate subgenus that Killip (1938) called Dysosmioides Killip. He also greatly reduced the number of infraspecific taxa from that of Killip (1938) by synonymizing dozens of names. Gross morphology played a role in this revision, but traits such as pubescence and fruit color were also used to reassess the taxonomy within sect. Dysosmia.

Table 1-1

The historical and current treatments of the Dysosmia group. Treatment of Dysosmia Author(s) Date Publication A section in genus Passiflora A.P. de Candolle 1822 Mém. Soc. Phys. Genève 1: 436. A subgenus in genus Passiflora L. Reichenbach 1828 Consp. Regn. Veg. 132. Tripsilina, a genus in C. Rafinesque 1838 Fl. Tellur. 4: 103. Passifloraceae Dysosmia, a genus in M. Roemer 1846 Fam. Nat. Syn. Monogr. 2: Passifloraceae 131, 149. A section in genus Passiflora G. Bentham & 1862 Gen. Pl. 1: 810. J. Hooker A section in subgenus M. Masters 1871 Trans. Linn. Soc. London Plectostemma 27: 631. A section in genus Passiflora H. Harms 1925 Nat. Pflanzenfam. 2(21): 502. A subgenus in Passiflora E.P. Killip 1938 Publ. Field Mus. Nat. Hist., Bot. Ser. 19: 487. A section in subgenus Passiflora, C. Feuillet & J.M. 2003 J. News. Passiflora Soc. Int. supersection Stipulata MacDougal 13(2): 34. A section in subgenus Passiflora, J. Vanderplank 2013 Curtis’s Bot. Mag. 30(4): supersection Stipulata 318-387.

The current taxonomic classification of Dysosmia (mainly prior to this dissertation) includes 22 recognized species and eight varieties (mostly following

Vanderplank, 2013). Currently recognized taxa are as follows: Passiflora arida (Mast. &

Rose) Killip, P. arizonica (Killip) D.H.Goldman, P. bahamensis Britton, P. boticarioana

Cervi, P. campanulata Mast., P. chrysophylla Chodat, the P. ciliata Aiton complex (with three varieties, P. ciliata var. hibiscifolia (Lam.) Vanderpl., P. ciliata var. orinocensis

(Killip) Vanderpl. and P. ciliata var. santiagana (Killip) Vanderpl.), P. clathrata Mast., the P. foetida L. complex (with four varieties, P. foetida var. acapulcensis Killip, P. 22 foetida var. ellisonii Vanderpl., P. foetida var. nigelliflora (Hook.) Mast., and P. foetida var. oaxacana Killip), P. fruticosa Killip, P. hypoglauca Harms, P. lepidota Mast., P. palmeri Rose, P. pectinata Griseb., P. pentaschista (Killip) H.T.Svoboda, P. setulosa

Killip, P. sublanceolata (Killip) J.M.MacDougal, P. urbaniana Killip, P. vellozoi

Gardner, P. vesicaria L. (with one variety, P. vesicaria var. galapagensis (Killip)

Vanderpl.), P. vestita Killip, and P. villosa Vell. This list includes several species that have been raised from varietal status of P. foetida by various authors as well those previously classified to subgenus Dysosmioides: P. campanulata, P. hypoglauca, P. setulosa, P. vellozoi, and P. villosa.

The infraspecific taxa have often been “shuffled” between species by various authors, with some like Killip (1938) treating the varieties originally ascribed to the P. ciliata complex as varieties of the P. foetida complex, and vice versa. A comparison between Killip’s (1938) and Vanderplank’s (2013) treatment of the taxa is shown in

Figure 1-4. In this way, it is perhaps more suitable to think of the taxa in this section as belonging to one of several potential sublineages or complexes, most of which may not have been identified and characterized yet. The groups hypothesized by current specialists are, as yet, poorly circumscribed themselves and more investigation is needed to assess species assignments and distinctions of these species assemblages relative to other species in sect. Dysosmia. For the purposes of this dissertation research, four complexes or groups will be informally recognized: 1) the P. foetida complex characterized as having green fruits and pubescent vegetation, 2) the P. ciliata complex characterized as having red fruit and glabrous vegetation, 3) the red-fruited pubescent (or 23

“RFP”) group characterized as having red fruit, but pubescent vegetation, and 4) the

Dysosmioides group characterized as generally lacking glands and having significantly leafier stipules and involucral bracts.

The largest of the sublineages in Dysosmia, the Passiflora foetida group (a large species complex), has been the repository of approximately 50 varieties, subspecies and forms (International Plant Names Index [IPNI], 2018) over the years. The patterns of variation found in the section are bewildering, making the status of many taxa poorly understood. Killip (1936) states that because of the high variability of P. foetida,

Dysosmia is “taxonomically one of the most difficult” groups in the genus, echoing a similar remark in an 1894 letter to John Hart from Hermann Harms who also expressed frustration in dealing with the complexities of trying to untangle the taxonomic mess in the P. foetida complex. This complexity is also seen within the other sublineages, bringing into question the evolutionary history of taxa in the section and the distinct possibility of convergent evolution. A number of factors likely contribute to the overwhelming variation in the group, including phenotypic plasticity, underrepresented population diversity, geographic variation, cryptic species, hybridization, and inconsistently defined species delimitations. 24

Figure 1-4. A comparison of the two major classifications of the Dysosmia group. Dashed lines indicate an unchanged taxon, solid lines indicate a different classification, asterisks (*) indicate a new name or taxonomic combination. 25

Another sublineage, the P. ciliata complex, has been difficult to circumscribe due to its highly variable nature and overall similarity in leaf form to the many other taxa in

Dysosmia, especially P. foetida. Initial studies of specimens of P. ciliata and other similar taxa appear to have red fruits and glabrous foliage, but other taxa with red fruits in Dysosmia vary in pubescence. The latter perhaps belong to a third sublineage, the RFP group, having a combination of both red fruit and pubescent vegetative features. The shared morphological traits and variable leaf shape of this potential sublineage make assessing its uniqueness difficult at best. The final sublineage, the Dysosmioides group, is thought to be the most basally diverging group in Dysosmia—or perhaps a closely related lineage outside of Dysosmia (sensu stricto). Leaf shape and weakly divided stipules and bracts appear to ally this group with the section of interest, but the lack of glands and unusual flower morphology call into question its true affinity with sect. Dysosmia.

Phylogenetic Relationships

Recent studies (Clifford, 2017; Hansen et al., 2006; Krosnick, Porter-Utley,

MacDougal, Jørgensen, & McDade, 2013; Muschner et al., 2003; Muschner, Zamberlan,

Bonatto, & Freitas, 2012; Souza-Chies, Yockteng, & Nadot, 2005; Yockteng & Nadot,

2004) provide a phylogenetic foundation for the relationships of the subgenera and sections in Passiflora, but no molecular studies to date address the phylogenetic relationships of the taxa in Dysosmia with sufficient sampling. The markers used in these studies (e.g., ITS, ncpGS, trnT-trnF, and ndhF [Clifford, 2017]; ITS, ncpGS, trnL-trnF, and ndhF [Krosnick et al., 2013]; matK, trnL-trnF, ncpGS, and ITS [Souza-Chies et al.,

2012], rpoC1 and trnL-trnT [Hansen et al., 2006]; ITS, trnL-trnF, and rps4 [Muschner et 26 al., 2004]; and ncpGS [Yockteng & Nadot, 2004]) showed varying degrees of utility in resolving relationships among closely related taxa.

Each of these studies did, however, provide evidence for the position of section

Dysosmia as sister to the rest of subgenus Passiflora. Investigations that have used taxa belonging to both Dysosmia (sensu stricto) and the Dysosmioides group (see Clifford,

2017; Hansen et al., 2006; Muschner et al., 2003) group have also supported this relationship, but further show that the two groups or lineages may not, in fact, be closely related. While taxa belonging to Dysosmia (sensu stricto) are consistently resolved basally to subgen. Passiflora, the Dysosmioides species are placed more “nested” within the subgenus, usually affiliating with members of section Granadillastrum Triana &

Planch.

Geographic Distribution

Dysosmia has a broad distribution throughout the Neotropics, with Mexico harboring a particularly diverse subset of the section. Hotspots can also be found in

Brazil (particularly the Dysosmioides group) and throughout the West Indies (namely the

P. ciliata complex). The P. foetida complex (sensu lato), however, is also known from

Paleotropical regions around the world including every continent except for Antarctica.

The circumstances of this taxon’s (or complex’s) presence into the Old World are thus far unclear but may have been due to a single introduction as a cultivated plant in Europe in the 17th or 18th century.

The current understanding of the distributions of Dysosmia species come from intensive herbarium studies, supplemented by personal or citizen science observations 27

(particularly those from iNaturalist (iNaturalist Network, 2018)). Floristic treatments— both historical and contemporary—are also often consulted for species occurrences in particular geographic areas, though the identification of specimens may be suspect when voucher specimens are not cited. Differences in taxonomic opinion, particularly those of non-specialists, also lead to discrepancies in the literature.

While some species (or species complexes) are quite widespread, others have a limited range or are endemic to a particular country or region. For example, Passiflora boticarioana, P. campanulata, P. clathrata, P. hypoglauca, P. lepidota, P. setulosa, P. vellozoi, and P. villosa are all endemic to Brazil; most of which belong to the

Dysosmioides group. Endemics are also present among the other major groups: Passiflora fruticosa (a member of the P. foetida complex; endemic to the state of Baja California

Sur in Mexico), P. bahamensis (a member of the P. ciliata complex; endemic to the

Bahamas), and P. sublanceolata (a member of the RFP group; endemic to the Yucatan

Peninsula). Due to our poor understanding of the taxon boundaries within the section, it is possible that some of the apparently broadly distributed taxa may in fact represent multiple narrowly-distributed and narrowly-circumscribed taxa.

CHAPTER 2: TYPIFICATIONS AND NOMENCLATURAL NOTES IN PASSIFLORA

SECTION DYSOSMIA (PASSIFLORACEAE)

This article was published in the journal Phytotaxa (Svoboda, MacDougal, &

2016, Magnolia Press; see Appendix A for more information). All text is retained from the publication, with the citations reformatted to comply with the formatting of this dissertation.

This work formally typifies all names within section Dysosmia and further clarifies the nomenclature for others that require nomenclatural attention.

Harlan T. Svoboda1, John M. MacDougal2,3 & Harvey E. Ballard, Jr.1

1Department of Environmental and Plant Biology, Ohio University, 315 Porter Hall,

Athens, Ohio 45701, U.S.A.

2Department of Mathematics and Natural Sciences, Harris-Stowe State University, 3026

Laclede Ave., St. Louis, Missouri 63103, U.S.A.

3Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166, U.S.A.

Received 11 April 2016; Accepted 23 November 2016; Published 14 December 2016.

Abstract

Names within Passiflora section Dysosmia (Passifloraceae) were evaluated for proper typification and nomenclatural clarity. Because several names lacked type material, 13 lectotypes, two neotypes, and four epitypes are here designated to maintain 29 nomenclatural stability. Many more names suffered from other nomenclatural discrepancies and thus four orthographic errors, six author citations, and 12 type designations are here corrected to comply with the International Code of Nomenclature

(ICN). Seven names in Dysosmia are treated as ambiguous and 12 names are identified as either invalid or illegitimate.

Introduction

One of the most enigmatic and taxonomically difficult groups in Passiflora

Linnaeus (1753), section Dysosmia De Candolle (1822), has troubled botanists for almost

200 years. An ongoing revision of the section by the first author has revealed several names as never properly typified or lacking a type specimen. Further, some taxa needed re-typification as previous types had been destroyed or lost. Other discrepancies in the nomenclature such as orthographic errors, improper author citations, and erroneous herbarium citations were also discovered.

In this article, typifications are made where necessary. In some instances, epitypes are proposed to aid in the interpretation of the name. All names that require nomenclatural clarification in one way or the other are also treated to ensure that every name associated with Dysosmia is in compliance with the current International Code of

Nomenclature for algae, fungi, and plants (ICN; McNeill et al., 2012). All herbarium acronyms follow Index Herbariorum (Thiers, 2014). Where applicable, only homotypic synonyms are listed. The taxa are divided into four distinct groups for the purposes of this article: A) names requiring typification; B) names needing nomenclatural clarification; C) ambiguous names; D) invalid and illegitimate names. 30

Nomenclatural Treatment

Typifications

The following names are here lectotypified, neotypified, or epitypified to comply with Chapter II, Section 2 of the ICN (McNeill et al., 2012).

1. Dysosmia fluminensis M.Roemer (1846, p. 150)

≡ Passiflora foetida [sensu Vellozo] (1831, tab. 86; 1881, 379).

≡ Passiflora foetida Linnaeus var. fluminensis (M.Roemer) Killip (1938, p. 499).

Lectotype (here designated):—Illustration from Vellozo, Fl. Flumin. 9, tab. 86. 1831.

Epitype (here designated):—BRAZIL. Rio de Janeiro: Nova Friburgo, 23 November

1890, A. Glaziou 20334 (P04882507!).

Notes:—Vellozo’s Flora Fluminensis (1831; 1881), although an important work for newly described taxa from Brazil, has proven difficult to interpret for a number of reasons. It has been documented that Vellozo sent specimens to the herbarium at the

Royal Museum of Lisbon (Borgmeier, 1961), but all attempts to find his original collections at LISU, or other herbaria, have been unsuccessful (Pastore, 2013; also see

Carauta, 1973).

Originally published as Passiflora foetida, Roemer (1846) considered it different from Linnaeus’ species and thus named it Dysosmia fluminensis, which was subsequently changed to P. foetida var. fluminensis by Killip (1938). In an intensive study of the species of Passifloraceae described in the Flora Fluminensis, Cervi and Rodrígues (2010) agreed with Killip’s taxonomic assignment but did not typify the name. Thus, Vellozo’s

(1831) plate number 86 is here selected as the lectotype of the name as it is the only 31 known original material. The name P. foetida [sensu Vellozo] was validly published under Art. 38.8 and 38.9 of the ICN (McNeill et al., 2012) because the plate shows dissection of the flower, bracts, and fruit. In selecting a herbarium specimen to aid in the identification of this taxon, A. Glaziou 20334 from the type locality of Rio de Janeiro,

Brazil is here chosen as the epitype.

2. Passiflora baraquiniana Lemaire (1861, plate 276)

≡ Passiflora foetida Linnaeus var. hirsuta Masters (1871, p. 631).

Lectotype (here designated):—Illustration from Lemaire, Ill. Hort. 8, plate 276. 1861

(Figure 2-1).

Notes:—The original illustration of this species (Figure 2-1) shows a plant that is clearly in the Dysosmia group, but is unique in its leaf morphology. Lemaire (1861) indicated that this plate was based on a specimen collected in the “forêts du territoire des

Amazones,” by M. Baraquin, for whom he named the species. Such a specimen, however, has not been found in any of the major herbaria worldwide. Because Lemaire worked mainly with living plants, the location of his collections and types, if any, are unknown (Stafleu & Cowan, 1979). Lemaire’s (1861) plate number 276 is here chosen as the lectotype as it is the only known original material. Although an epitype was sought to help with the interpretation of this name, no suitable herbarium specimen has yet been found.

The currently accepted name of this taxon, P. foetida var. hirsuta, has had a tumultuous nomenclatural past and often times not even been associated with P. baraquiniana. Masters (1871, p. 631), however, made a direct reference to P. 32 baraquiniana as the only synonym. In the same publication, he also published the name

P. suberosa Linnaeus var. hirsuta (Linnaeus) Masters (1871, p. 630) with P. hirsuta

Linnaeus listed as one of several synonyms and apparent basionym for his new name.

The following year, in the Flora Brasiliensis, Masters (1872, p. 583) also listed P. hirsuta as a synonym of P. foetida var. hirsuta, while still retaining it as the basionym of P. suberosa var. hirsuta (Masters, 1872, p. 579; Porter-Utley, 2014, p. 93).

Linnaeus’ (1753) description of P. hirsuta is a mix of polynomials and cited illustrations that depict passionflowers from two very different sections of the genus: the apetalous section Cieca Medikus (P. suberosa) and section Dysosmia (P. foetida). Thus,

Masters cited P. hirsuta Linnaeus [pro parte] under both species and sections in 1872.

While it is clear in the 1871 publication that Masters intended for P. hirsuta to be the basionym of the new combination P. suberosa var. hirsuta, his Passiflora foetida var. hirsuta is validly published there by citing a previously validly published description (P. baraquiniana). Because of this confusion, the author of this variety of P. foetida is often seen as “(Linnaeus) Masters”, but is to be cited correctly as simply “Masters.” The confusion surrounding P. hirsuta Linnaeus has also been reduced by the recent lectotypification of the name on an illustration clearly representing a species from section

Cieca in the P. suberosa complex (Porter-Utley, 2014, p. 86).

3. Passiflora ciliata Aiton (1789, p. 310)

≡ Dysosmia ciliata (Aiton) M.Roemer (1846, p. 149).

≡ Passiflora foetida Linnaeus var. ciliata (Aiton) Masters (1871, p. 631). 33

Neotype (here designated):—JAMAICA. Cultivated in “Hon. Mrs. Barrington’s stove at

Mongewell,” [Crowmarsh, England, UK], July 1791, collector and collection unknown

(LINN-HS1418.43!) (Figure 2-2).

Notes:—For authorship purposes, the names published in Hortus Kewensis should be attributed to William Aiton alone as explained by Reveal (1985) and Art. 46, Ex. 37 of the ICN (McNeill et al., 2012).

The only documented, original material that Aiton (1789) used in circumscribing this species was listed as “Nat. of Jamaica; Cult. 1783 by Mrs. Norman.” Killip (1938, p.

509) made a note that this type specimen was at the Natural History Museum, London

(BM) then. Intense search for this sheet at both BM and the herbarium of the Royal

Botanic Gardens, Kew (K), however, failed to trace that specimen (J. Gregson and S.

Edwards, personal communication).

A specimen from the herbarium of Sir James E. Smith, held at the Linnean

Society of London (LINN-HS1418.43; Figure 2-2), is annotated as being P. ciliata and matches well with the protologue’s description and accompanying illustration. It is probable that Mrs. Barrington’s and Mrs. Norman’s plants were from the same seed introduction and were circulated around England in the late 1700s. This specimen is here designated as the neotype of the name.

4. Passiflora clathrata Masters (1872, p. 580)

Type:—Sello 2335 (lectotype B, destroyed).

Lectotype (here designated):—BRAZIL. Minas Gerais: Without locality, 1816–1821, A. de Saint-Hilaire 2157 (P00464294!). 34

Syntypes:—BRAZIL. Goiás: Gardner 3192 (K); Minas Gerais: Lindberg 361a (BR!),

Lindberg 623 (BR!), Claussen 377 (P!, W), Saint-Hilaire 465 (P!), Warming s.n. (P!,

US!).

Notes:—Masters (1872) listed a total of eight syntypes for P. clathrata. Killip

(1938, p. 462) later lectotypified the name with one of these syntypes, Sello 2335 at the herbarium of the Botanisches Museum Berlin-Dahlem (B). This type was undoubtedly lost in the 1943 fire that destroyed parts of the herbarium and museum (Hiepko 1987).

Because no duplicates of this specimen have been found, a new lectotype is here chosen from the remaining syntypes. From the images that were available online, we chose a specimen (A. de Saint-Hilaire 2157) that we believe best matches the description while also showing the range in variability of leaf morphology.

For authorship purposes, the correct citation should be simply “Masters” following Art. 46.6 of the ICN (McNeill et al., 2012).

5. Passiflora foetida Linnaeus var. vesicaria Persoon (1807, p. 222)

Lectotype (here designated):—Illustration from Plukenet, Phytographia, tab. 104, fig. 1.

1691.

Notes:—Persoon (1807) indicated that his new variety was based on a copper engraving from Plukenet’s (1691) Phytographia, originally drawn by J. Collins. This figure was used as part of the circumscription of P. vesicaria Linnaeus, but differs morphologically from the designated lectotype of the species (see name 15, below).

Because this was the only original material used by Persoon, Plukenet’s tab. 104, fig. 1 is here designated as the lectotype of the name. 35

6. Passiflora foetida Linnaeus var. vitacea Masters (1872, p. 583)

Lectotype ([second-step] here designated):—URUGUAY. “Banda oriental,” 1816–1821,

A. de Saint-Hilaire 2529 (P00605791!; isolectotypes F977125! [lectotype fragment],

P00605792!, P00605793!).

Notes:—Masters (1872) chose A. de Saint-Hilaire 2529 to represent the type of this variety. Killip (1938, p. 491) listed the type (= first-step lectotype) as being in the collections at the Muséum National d’Histoire Naturelle, Paris (P). Since there are three extant duplicates of Saint-Hilaire 2529 at P and none of them were annotated by Killip as the type, we are selecting the best representative specimen (P00605791) as a second-step lectotype.

7. Passiflora gossypiifolia Hamilton (1825, p. 48), as ‘gossypifolia’

≡ Dysosmia gossypiifolia (Hamilton) M.Roemer (1846, p. 149).

≡ Passiflora foetida Linnaeus var. gossypiifolia (Hamilton) Masters (1871, p. 631).

Lectotype (here designated):—Without locality, no date, collector and collection unknown (P00605797!) (Figure 2-3).

Notes:—William Hamilton (1825) indicated in the protologue that the specimen used to describe P. gossypiifolia was from N.A. Desvaux’s herbarium, by the notation

“Herb. Prof. Desv.”, but did not indicate a collector, collection number, or locality.

Specimens examined for the Prodromus can be easily distinguished by their labels which bear a scientific name, Desvaux’s name, and reference to its corresponding page in the

Prodromus, all in Desvaux’s handwriting (Howard, Clausen, & Gillis, 1981). 36

Only one specimen originally from the Desvaux herbarium, now in the general collection at P (P00605797), has all of these properties and matches well with the description. This sheet bears the notation “Passiflora gossypifolia Desv. prod. p. 48.” but also has an additional slip of paper that reads “Pérou [Peru]”, which is perhaps in conflict with the title of the work. However, because the slip is separate from and has different handwriting than Desvaux’s label, it is unclear if it actually belongs with the specimen.

Another Desvaux herbarium specimen (P00605798) likely also bears Desvaux’s handwriting, but differs in the typical label information (i.e., two scientific names are listed, there is no mention of the Prodromus, etc.).

Killip (1938, p. 468) listed the type as a P specimen collected by “Poiteau in

1802” from the Dominican Republic, despite the fact that nearly all of the sheets in

Desvaux’s herbarium were from various correspondents and lacked any collection information. Such a Poiteau specimen could not be located at P. Because of these facts, the most suitable specimen (P00605797; Figure 2-3) is here chosen as the lectotype in place of the erroneous type proposed by Killip.

The name “gossypiifolia” has suffered from a multitude of misspellings and incorrect author citations since its publication. Although it is clear that N.A. Desvaux participated in writing the Prodromus, Howard et al. (1981) assert that “all of the taxa described in the Prodromus should be attributed to Hamilton alone.” Another common mistake is the misspelling of the epithet. Originally published as “gossypifolia,” the correct spelling of this epithet should in fact be “gossypiifolia” in compliance with Art.

60.8 (see also Recommendation 60G.1) of the ICN (McNeill et al., 2012). 37

8. Passiflora hermannii De Candolle (1828, p. 332), as ‘Hermanni’

Lectotype (here designated):—Illustration from Hermann, Parad. Bat., plate 176. 1698.

[The plate based on cultivated material grown in Leiden, Netherlands, originating from

Curaçao.]

Notes:—The very brief protologue of this species describes a plant from the island of Curaçao with “foliis velutinus trilobatis involucro minimo foliolis 3 integris…”.

Notable is the mention of a small involucre of three entire bracts, an unusual morphological state for taxa in Dysosmia. The protologue matches closely with, and is undoubtedly derived from, one of Hermann’s (1698, p. 176) Passiflora species, “Flos

Passionis albus folio Ibisci sericeo trilobato…” including an illustration later in his work

(plate 176).

De Candolle (1828) indicated that his new species was not well known, using the notation “species non satis notae.” Probably lacking a specimen at hand, he based the name on Hermann’s work alone. Furthermore, the epithet is to be corrected to

‘hermannii’ following Art. 60.12 and Recommendation 60C.1 of the ICN (McNeill et al.,

2012).

9. Passiflora hispida De Candolle ex Triana & Planchon (1873, p. 172).

≡ Passiflora foetida Linnaeus var. hispida (De Candolle ex Triana & Planchon) Killip in

Gleason, Bulletin of the Torrey Botanical Club, 58, 408. 1931.

Lectotype (here designated):—MARTINIQUE. [Saint-Pierre]: Saint-Pierre, 4 May 1839,

A. Steinheil 30 (P04881879!) (Figure 2-4). 38

Syntypes:—BRAZIL. Collector and collection unknown (G [specimen not located]).

COLOMBIA. Nariño: Triana 2944 (COL!). GUADELOUPE. Perrottet s.n. (G!).

MARTINIQUE. Plée s.n. (P [specimen not located]). SENEGAL. (“Herb. Cambess.”,

[specimen not located]). SURINAME. Kappler 1518, Hostmann 652 (G!). VIRGIN

ISLANDS. Saint Thomas: Wydler 100 (G!). VENEZUELA. Zulia: Plée s.n. (specimen not located).

Notes:—The protologue lists ten specimens (= syntypes) that Triana and

Planchon (1873) associated with this new name. Killip (1938, p. 496) erroneously chose another specimen from Brazil at P as the neotype. Because several syntypes are extant, these take precedence over Killip’s erroneous neotypification. One such specimen showing diagnostic characters, Steinheil 30 (Figure 2-4)—listed as an additional specimen examined in Killip’s (1938) monograph—is here chosen as the lectotype of the name.

The basionym, P. hispida, should be cited as “De Candolle ex Triana &

Planchon” due to the citation “DC., mss., in herb. Mus. Paris” by Triana and Planchon

(1873). At the level of variety, Killip was identified as the combining author by a footnote in Gleason’s (1931, p. 58) publication.

10. Passiflora hypoglauca Harms (1922, p. 296)

Type:—E. Ule 2569 (holotype B, destroyed [image F!]).

Lectotype (here designated):—BRAZIL. [Minas Gerais]: Ouro Preto, “Serra de Ouro

Preto,” 1892, E. Ule 2569 (R000090093!). 39

Notes:—The protologue of P. hypoglauca indicates that that the holotype, E. Ule

2569, was at the Berlin-Dahlem herbarium (B). However, this specimen was destroyed in the fire of 1943 (Hiepko, 1987). A photograph of the holotype was deposited in the herbarium of the Field Museum (F) but does not show appropriate detail to serve as the lectotype. A duplicate specimen deposited at the Museu Nacional, Rio de Janeiro (R) is here designated as the lectotype.

11. Passiflora lepidota Masters (1872, p. 581)

Type:—Sello s.n. (holotype B, destroyed).

Neotype (here designated):—BRAZIL. [Paraná]: “Itararé opp., Morungava praedium, in campo,” 740 m, 30 January 1915, P. Dusén 16569 (MO905635!; isoneotype S08-

10927!).

Notes:—The protologue of P. lepidota (Masters, 1872) lists a single unnumbered specimen collected by “Sello” as the holotype and indicates that it was located at the collections at B (“in herb. Reg. Berol. absque numero”). This specimen was undoubtedly destroyed in the fire of 1943 (Hiepko, 1987) and no duplicates have been traced elsewhere. Killip (1938) observed six additional specimens of this species, five of which were at the herbarium of the Naturhistoriska Riksmuseet (S). In selecting a neotype for this name, we chose one of them, P. Dusén 16569, with well-preserved vegetative and fertile parts.

12. Passiflora muralis Barbosa Rodrigues (1891a, p. 29)

≡ Passiflora foetida Linnaeus var. muralis (Barbosa Rodrigues) Killip (1938, p. 497). 40

Lectotype (here designated):—Illustration from Barbosa Rodrigues, Vellosia 2(3), est.

13b. 1891b.

Epitype (here designated):—BRAZIL. Bahia: Vicinity of “Toca de Onca,” 27–29 June

1915, J.N. Rose & P.G. Russell 20080 (NY00484180!; isoepitype US762369!).

Notes:—If Barbosa Rodrigues actually used any prepared vouchers for his descriptions, they were likely destroyed in a flood (Buzatto, Singer, Romero-González, &

Van Den Berg, 2011). Being an avid artist, he illustrated most of his newly described species, though a large portion of these original illustrations were also lost or destroyed around the same time (Brito, 2013). The reproduced illustration of this species in Vellosia

(Barbosa Rodrigues, 1891b) is here chosen to serve as the lectotype of the name.

In selecting an epitype we chose a specimen (J.N. Rose & P.G. Russell 20080) that, following Killip (1938, p. 498) and our own opinion, best depicts the features presented by Barbosa Rodrigues (1891a; 1891b).

13. Passiflora nigelliflora Hooker (1839, plate 3635)

≡ Dysosmia nigelliflora (Hooker) M.Roemer (1846, p. 151).

≡ Passiflora foetida Linnaeus var. nigelliflora (Hooker) Masters (1871, p. 631).

Type:—ARGENTINA. [Santiago del Estero]: “St. Jago de Estero” on the Rio Dulce,

1835, J. Tweedie 1171 (holotype K000323473!).

Epitype (here designated):—Illustration from Hooker, Curtis’s Bot. Mag. 12 [Bot. Mag.

65], plate 3635. 1839. 41

Notes:—In the protologue of this species, William Hooker (1839) cited a specimen, “Tweedie Pl. Exsicc. (n. 1170),” presumably at K where J. Tweedie deposited his material and where Hooker worked (Stafleu & Cowan, 1986). Such a specimen was found in the collections at K, but it is a type specimen for Passiflora palmatisecta

Masters and bears no morphological similarity to that of P. nigelliflora. Another specimen, Tweedie 1171, however, matches perfectly with Hooker’s description and accompanying illustration. The small mistake in numbering was likely a typographical error when published and Tweedie 1171 at K should therefore be accepted as the holotype of this species.

Because the holotype specimen has no flowers, Hooker’s original illustration of the species that was published alongside the protologue (which does include flowers), plate 3635, is further selected as an epitype to aid in the interpretation of the name.

14. Passiflora palmeri Rose (1892, p. 131)

Lectotype (here designated):—MEXICO. [Baja California Sur]: Carmen Island, 1–7

November 1890, E. Palmer 868 (US46866!, isolectotypes A00112633!, F118680!

[holotype fragment], F265984!, G00441014!, GH0068012!, K323266!, NY110410!, S-

G-4585!, UC108345!, US46867!, US46868!, US1360073!).

Notes:—Rose (1892) indicated that the type specimen of this name was to be E.

Palmer 868, presumably at the United States National Herbarium (US) where Palmer deposited the majority of his type material. Killip (1938, p. 465) indicated that this type was indeed at US, but because four duplicates of this collection exist at US, a single type specimen needed to be selected from among them. Neither Rose nor Killip made any 42 indication of which of the four specimens should be considered the “holotype,” even on the sheets, and thus US46866 is here designated as the lectotype from the original material.

15. Passiflora vesicaria Linnaeus (1759, p. 20)

Lectotype (here designated):—JAMAICA. Without locality, no date, P. Browne s.n.

(S08-4074!).

Notes:—The history of P. vesicaria, as well as its relationship with P. foetida, is a complicated one. When Linnaeus first described P. foetida in his Species Plantarum

(1753, p. 959) he referenced five polynomials published in seven separate works, one being Passiflora florum involucris triphyllis multifido-capillaribus from his own Hortus

Cliffortianus (Linnaeus 1737, p. 431). Three years later in the Civil and Natural History of Jamaica, Patrick Browne (1756, p. 327) used this polynomial as the basis for a new taxon that he distinguished separate from P. foetida, calling it “Passiflora 1. Vesicaria; florum involucris triphyllis multifido-capillaribus.” However, because Browne did not consistently use binomials in his work, it was not a validly published name and therefore has no nomenclatural standing.

It was another three years before Linnaeus (1759, p. 20) conceded with Browne that some of the polynomials originally used to describe P. foetida were in fact representing a different taxon. Linnaeus thus named the new species P. vesicaria referencing “327,” the page on which Browne’s (1756) new taxon was identified, thereby creating a valid binomial name. 43

There exists only one specimen that can be considered original material for this name: a pressed plant that Browne himself collected in Jamaica (S08-4074) which bears the species name in Linnaeus’ hand. This specimen was first cited as original material in

Jarvis (2007, p. 727), and later confirmed by Vanderplank (2013, p. 349), who cited it as a holotype thus not effecting typification under Art. 9.23 and 7.10 of the ICN (McNeill et al., 2012). The herbarium specimen S08-4074 is here formally designated as the lectotype of the name.

16. Passiflora villosa Vellozo (1831, tab. 87; 1881, p. 380)

Lectotype (here designated):—Illustration from Vellozo, Fl. Flumin. 9, tab. 87. 1831.

Epitype (here designated):—BRAZIL. São Paulo: “In a wood,” 20 November 1941, B.

Pickel 5512 (US1803974!).

Notes:—Vellozo’s (1831) plate 87 represents an illustration with analysis according to Art. 38.8 and 38.9 of the ICN (McNeill et al., 2012) and is here chosen as the lectotype for P. villosa as it is the only known original material of the name.

Furthermore, B. Pickel 5512 is here selected as the epitype to aid in the interpretation of morphological features diagnostic for this species.

Nomenclatural Clarification

Orthographic errors, inaccurate type information, and improper author citations are here treated in order to maintain transparency in the nomenclature.

1. Passiflora bolstadii Dusén (1903, p. 50)

Type:—BRAZIL. “Serra do Itatiaia, ad marginem silvulae,” 2100 m, May 1902, P.

Dusén s.n. (holotype S04-594!). 44

Notes:—Although the Dusén s.n. specimen from 1902 bears an ‘isotypus’ label, it is the only specimen at S (Per Dusén’s home institution) annotated by the author and should be considered the holotype. Another specimen at S, Dusén 14078 (S04-593), mistakenly bears a ‘typus’ label but was collected nine years after publication and is not original material.

2. Passiflora campanulata Masters (1872, pp. 568, 615)

Lectotype (designated by Killip 1938, p. 517):—BRAZIL. Rio de Janeiro: Organ

Mountains, no date, W. Lobb s.n. (K000323327!).

Syntype:—BRAZIL. Minas Gerais: Weddell 1333 (specimen not located).

3. Passiflora ciliata Aiton var. polyadena Grisebach (1866, p. 285)

≡ Passiflora foetida Linnaeus var. polyadena (Grisebach) Killip (1938, p. 512).

Type:—CUBA. Without locality, 1865, C. Wright s.n. (holotype GOET009389!; isotype

US44904!).

Notes:—Grisebach worked at the Herbarium Göttingen (GOET) and deposited most of his types there (Stafleu & Cowan, 1976). Searching for this name on the

[JSTOR] Global Plants database (Ithaka, 2014) turned up an image annotated as the holotype by M.E. Duharte and H. Manitz in 1987 housed at GOET.

This holotype specimen bears what, at first sight, appears to be a collector number

(“368”), but is not in the collector’s handwriting and was likely added many years after the author’s publication of this name. Because this specimen is the only sheet that matches the protologue and specimen citation (Wright s.n.) in the GOET herbarium (M.

Appelhans, personal communication), it can safely be assumed that this is the holotype. 45

4. Passiflora ciliata Aiton var. quinqueloba Grisebach (1866, p. 113)

≡ Passiflora foetida Linnaeus subvar. quinqueloba (Grisebach) Masters (1871, p. 631).

≡ Passiflora foetida Linnaeus var. quinqueloba (Grisebach) Killip (1938, p. 511).

Type:—CUBA. “Cuba Orientali, prope,” 1867, C. Wright 2601 (holotype GOET009390!; isotypes G00441004!, G00441005!, GH00068022!, K000323250!, MA607468!,

MO2063293!, NY01444432!, P00605794!, US448904!, YU244604!).

Notes:—Because the collector and collection number match with the protologue, and it was verified that there are no duplicates at GOET (M. Appelhans, personal communication), it is safe to assume that this specimen is indeed the holotype.

An additional specimen in the herbarium of the New York Botanical Garden

(NY00084319) bears a ‘Wright 2601’ label, but differs morphologically from all other type specimens of this variety. Furthermore, the collection number on the label is written in a hand contrary to Wright’s and should, therefore, not be considered a type.

5. Passiflora ciliata Aiton var. riparia C.Wright ex Grisebach (1866, p. 113)

≡ Passiflora foetida Linnaeus var. riparia (C.Wright ex Grisebach) Killip (1938, p. 510).

Type:—CUBA. Without locality, 1860–1864, C. Wright 2602 (holotype GOET009391!; isotypes BM000798356!, G00441002!, G00441002!, GH00068023!, K000323251!,

MA607467!, MO2063294!, NY01444433!, P00604269!, US943566!, YU244603!).

6. Passiflora foetida Linnaeus (1753, p. 959)

≡ Granadilla foetida (Linnaeus) Gaertner (1788, p. 289).

≡ Tripsilina foetida (Linnaeus) Rafinesque (1836, p. 103), as ‘fetida’.

≡ Dysosmia foetida (Linnaeus) M.Roemer (1846, p. 149). 46

Lectotype (designated by Killip 1938, p. 483):—COUNTRY UNKNOWN. Without locality, no date, collector and collection unknown (LINN1070.24!) (Figure 2-5).

Notes:—Killip (1938) designated a specimen in the LINN herbarium

(LINN1070.24; Figure 2-5A) as the lectotype of the name, presumably because

Linnaeus’ own handwriting appears at the bottom of the sheet (Figure 2-5B). Although this specimen is not the most typical form for P. foetida, it is still well within the range of variation for the species as circumscribed by Linnaeus. Irrespective of Vanderplank’s

(2013, p. 370) comments, Killip’s lectotypification is not superseded as it is not in conflict with the protologue (McNeill et al., 2012, Art. 9.19).

7. Passiflora foetida Linnaeus var. ellisonii Vanderplank (2013, p. 380)

Type:—AUSTRALIA. New South Wales: Cultivated at the National Collection of

Passiflora U.K., Somerset, London, England, from wild seed collected by Don Ellison,

January 2010, R.J.R. Vanderplank NCP 2005/12 (holotype K).

Paratypes:—CULTIVATED. Vanderplank NCP 2006/12 (K!), Vanderplank NCP

2007/12 (K!), Vanderplank NCP 2008/12 (K). NEW GUINEA. Wells 3797 (K).

Notes:—When publishing this name, Vanderplank (2013, p. 380) designated a single holotype but erroneously designated four other collections as ‘isotypes.’ Since one specimen (Wells 3797) is of different origin than the holotype and the other three are apparently preparations from the same living plant at different times, none can be considered as isotypes but paratypes instead (see McNeill et al., 2012, Art. 8.2, Art. 9.6).

8. Passiflora foetida Linnaeus var. maxonii Killip (1836, p. 326), as ‘Maxoni’ 47

Type:—NICARAGUA. [Managua]: Vicinity of Managua, mostly along the shore of Lake

Managua, 24 June 1923, W.R. Maxon, A.D. Harvey & A.T. Valentine 7219 (holotype

US1180248!; isotypes US1180249!, US1180250!).

Notes:—The epithet is to be corrected to ‘maxonii’ following Art. 60.12 and

Recommendation 60C.1 of the ICN (McNeill et al., 2012).

9. Passiflora foetida Linnaeus var. orinocensis (Killip) Feuillet (2007, p. 144)

≡ Passiflora foetida Linnaeus var. orinocensis Killip (1938, p. 510), nom. inval.

≡ Passiflora foetida Linnaeus subsp. orinocensis Killip (Bailey 1930, p. 205).

Type:—VENEZUELA. [Bolívar]: Ciudad Bolívar and vicinity, “Isla Degrero,” on the

Orinoco [River], 200 ft, lat. 8:10˚N, 6 March 1921, L.H. Bailey & E.Z. Bailey 1773

(holotype US1059996!; isotype NY00110440!).

Notes:—Killip (in Bailey, 1930) published the name of this taxon as “P. foetida subsp. orinocensis,” but in his 1938 monograph of the family listed the name as “P. foetida var. orinocensis” without indicating a change in rank. This name thus remained not validly published until Feuillet (2007).

10. Passiflora foetida Linnaeus var. strigosa S.Moore (1895, p. 365)

Lectotype (designated by Killip 1938, p. 499):—BRAZIL. Matto Grosso do Sul: “In ripâ fl. Paraguay inter Santa Cruz et Villa Maria, itaque juxta Corumbá,” 1891–1892, S.

Moore 915 (BM00056156!).

Syntype:—BRAZIL. Matto Grosso do Sul: Moore 820 (BM!).

Notes:—Moore (1895) listed two syntypes in his original description of this variety, S. Moore 820 and 915. Both collections were mounted on a single sheet at BM, 48 but collection number 820 is deteriorated beyond recognition and is of little use for taxonomic purposes. P. Jørgensen annotated collection 915 as being the ‘holotype,’ but since Moore explicitly chose no single holotype, Killip’s (1938) selection is to be considered a lectotype.

11. Passiflora hastata Bertoloni (1840, p. 427)

≡ Dysosmia hastata (Bertol.) M.Roemer (1846, p. 149).

≡ Passiflora foetida Linnaeus var. hastata (Bertol.) Masters (1871, p. 631).

Type:—GUATEMALA. Without locality, no date, J. Velásquez s.n. (holotype

BOLO508029!).

Notes:—During a visit to Università di Bologna’s herbarium (BOLO) in 1979,

Duncan (1983) was unable to locate 57 of the total 61 type specimens for names published in Antonio Bertoloni’s Florula Guatimalensis (1840). These “missing” specimens have since been located during an ongoing restoration of the BOLO herbarium

(Cristofolini, Conte, & Mossetti, 1987). Not Bertoloni himself, but Captain J. Velásquez was the collector and later presented it to Bertoloni for his work (G. Cristofolini, personal communication).

12. Passiflora moritziana Planchon in Triana & Planchon, [Annales des Sciences

Naturelles, Botanique, 5(17), 175. 1873].

≡ Passiflora foetida Linnaeus var. moritziana (Planchon) Killip in Pulle, Flora of

Suriname [(Vol. 3), 318. 1937].

Type:—VENEZUELA. [Aragua]: Colonia Tovar, no date, J. Moritz 437 (holotype

P00605763!). 49

Notes:—The name P. moritziana was published in a section of a journal authored by J. Triana and J.E. Planchon (1873), but was listed as an unpublished name attributed to Planchon (“Planch., mss.”). The correct author citation should, therefore, be written as

“Planchon” in accordance with Art. 46.2 of the ICN (McNeill et al., 2012). A new combination using this basionym, P. foetida var. moritziana, was published by E.P. Killip in Pulle (1937), who validly published the name (McNeill et al., 2012, Art. 41.3) but incorrectly attributed the basionym to De Candolle. However, we have been unable to find any evidence that De Candolle ever published a name using moritziana. Killip corrected the author citation for this name in his 1938 monograph.

Killip (1938) listed duplicates of the type material to be at BM and K, with additional material found at FI by the authors, but those labels give different collection localities, namely “Caracas” (K000323448!, BM000566157!, FI004378!) and “La Ceiba”

(K000323449!). Though they have the same collection number (“437”), they are not of the same gathering according to Art. 8.2 of the ICN (McNeill et al., 2012) and thus should not be considered isotypes.

13. Passiflora pectinata Grisebach (1860, p. 294)

Type:—TURKS AND CAICOS ISLANDS. Turks Islands: Without locality, no date,

Hjalmarson s.n. (holotype GOET009308!; isotype K000323253!).

Notes:—Grisebach’s protologue of this species lists an unnumbered Hjalmarson specimen as the type of the name. Such a specimen exists at GOET where Grisebach worked and deposited his type specimens (Stafleu & Cowan, 1976), and therefore should be considered the holotype. 50

14. Passiflora pseudociliata Britton (1917, p. 19)

Type:—CUBA. Camagüey: Savannas near Camagüey, 2–7 April 1912, N.L. Britton, E.G.

Britton & J.F. Cowell 13155 (holotype NY00991121!; isotypes F459317!,

NY00084323!, P00605795!, US718224!).

Notes:—Two type specimens of P. pseudociliata were deposited at NY. Britton wrote “Type” on the label of one of them (NY00991121), indicating which specimen should be considered the holotype.

15. Passiflora vellozoi Gardner (1845, p. 103), as ‘Vellozii’

≡ Cieca vellozoi (Gardner) M.Roemer (1846, p. 142), as ‘Vellozii’.

Neotype (designated by Killip 1938, p. 514):—BRAZIL. Rio de Janeiro: Organ

Mountains, 3,000 ft, July 1837, Gardner 427 (BM566167!; isoneotype K00323328!).

Notes:—The spelling of the epithet is to be corrected to “vellozoi” to comply with

Art. 60.12 and Recommendation 60C.1 of the ICN (McNeill et al., 2012).

Gardner’s (1845) protologue of P. vellozoi included Vellozo’s (1831) plate 86,

“Passiflora foetida” [sensu Vellozo], as a synonym for his new name. Max Roemer

(1846, p. 142) later asserted that Vellozo’s (1831) plate 86 actually referred to Dysosmia fluminensis, which now serves as the lectotype of that name (see name 1, above). Killip

(1938, p. 515) agreed with Roemer’s decision stating that, “Gardner described a wholly different plant [from Vellozo’s illustration] under the name P. Vellozii,” and chose

Gardner 427 to serve as the neotype for the name.

Ambiguous Names

The descriptions for several validly published names are too ambiguous to ascribe to any one taxon and are therefore considered questionable or dubious names.

1. Passiflora foetida Linnaeus var. angustifolia Schlechtendal (1853, p. 632)

Type:—VENEZUELA. Vargas: “In arenosis ad Cabo blanco,” 1000 ft, no date, P.

Wagener 142 (specimen not located).

Notes:—Schlechtendal (1853) published this name with uncertainty as indicated by a question mark in the protologue (“Passiflora foetida L. ? var. angustifolia v. nova species”). Although validly published under Art. 36.1 of the ICN (McNeill et al., 2012), the short, unhelpful description makes the name too ambiguous to apply to one particular taxon despite listing P. Wagener 142 in the protologue.

The location of Wagener’s collections is so far unknown (Stafleu & Cowan,

1988) and therefore cannot be consulted for study. The Wagener collection was sought at the Martin-Luther-Universität herbarium (HAL) since Schlechtendal worked mostly at this institution, but no such specimen exists there (U. Braun, personal communication).

2. Passiflora foetida Linnaeus f. longifolia Kuntze (1891, p. 254)

Type:—VENEZUELA. La Guaira and Caracas, no date, collector and collection unknown (specimen not located).

Notes:—Kuntze (1891) published this name with a brief description (“folia longius acuminata longiora [long, acuminate leaves]”) which does not provide enough information to distinguish it from other taxa in Dysosmia. No Kuntze specimen(s) matching this description were found at NY (T. Zanoni, personal communication), nor 52 any major herbarium worldwide, which further impedes the positive application of this name.

3. Passiflora foetida Linnaeus var. variegata G.Meyer (1818, p. 226)

Type:—Type unknown.

Notes:—Meyer’s (1818) description of this variety under P. foetida included a brief diagnosis, “β variegata, corona ex albo coeruleo et purpureo variegata,” a discussion of how it differed from the typical species, and reference to Plukenet’s (1696, p. 282) “Passiflora Vesicaria Hederacea.” There is also reference to an illustration in

Plukenet’s Phytographia (1691, tab. 104, fig. 1), which is now the lectotype of P. foetida var. vesicaria (see name 5, above).

Meyer was of the opinion that “this variety is probably a species on its own [a good species]” (1818, p. 227, translated). He noted that it was widespread throughout the area treated by his flora, and that it bloomed in January. We have not located a type specimen, and are hesitant to typify this name without further searching. The taxonomic identity of this variety is thus far unclear.

4. Passiflora hibiscifolia Lamarck var. velutina J.Jacquin ex Fenzl (1844, p. 5)

Type:—“Insulis Caribaeis,” no date, collector and collection unknown (specimen not located).

Notes:—Joseph Jacquin published volume one of Eclogae Plantarum Rariorum between the years of 1811 and 1816, and was preparing to publish volume two before his death in 1839 (see introduction to Jacquin, 1844). A colleague, Eduard Fenzl, decided to finish the second volume and revised some of the nomenclature in the work. 53

Fenzl noticed that an illustration, tab. 123, labeled by Jacquin as “P. hirsuta”

J.Jacquin (non Linnaeus, 1753; see name 9, below) was a homonym of the name and attempted to apply a new name to the plant represented in the plate. However, the illustration and the vague description provided little evidence to distinguish it from other taxa already described in Dysosmia. Furthermore, none of the references for the listed synonyms match with the illustration. A single specimen from Jacquin’s herbarium

(W0073465), now at the Naturhistorisches Museum Wien (W), is similar in morphology to the illustration but does not match closely enough to consider typification.

5. Passiflora pectinifera W.Thompson (1876, p. 80)

Type:—Type unknown.

Notes:—The lack of a detailed description, illustration, or herbarium specimen makes it impossible to distinguish it from other Dysosmia taxa (e.g., P. foetida).

6. Passiflora polyadena Vellozo (1881, p. 381; 1831, tab. 92), as ‘polyaden’

≡ Dysosmia polyadena (Vellozo) M.Roemer (1846, p. 150).

Type:—BRAZIL. Rio de Janeiro: “Collibus maritimis prope urbem,” no date, collector and collection unknown (specimen not located).

Notes:—Vellozo’s (1831) illustration of this species has no diagnosis (i.e., cross section or detail) and thus the name was not validly published until 1881 (Carauta, 1973).

It is nearly indistinguishable from P. foetida Vellozo (non Linnaeus, 1753) published just six plates earlier. Although the description was published 50 years later, it is of little use in differentiating it from many other Dysosmia taxa. Roemer’s (1846) added description accompanying Dysosmia polyadena also does not help in this way. 54

Because none of Vellozo’s specimens have yet been found (see taxon 1, above), it is impossible to ascribe this name accurately with the iconograph and description alone.

7. Passiflora variegata Miller (1768, Passiflora No. 8)

Type:—Type unknown.

Notes:—Miller’s (1768) introduction of this new name included brief Latin and

English diagnoses as well as a quotation of a polynomial and reference to an entry in

Tournefort (1700, p. 241). The protologue also includes detailed observations in English from living, cultivated material. Miller compared it to P. foetida but said “there can be no doubt of its being a different species.”

The diagnosis (“foliis hastatis pilosis amplioribus, involucris multifido capilaribus [largest halbert-pointed hairy leaves, and empalements composed of many- pointed hairs]”) is general and could potentially apply to many of the P. foetida-type taxa.

Tournefort’s polynomial is also overly general, though he refers (1700, p. 241) to an icon and polynomial from Hermann’s earlier Paradisus Batavus (“Flos passionis hirsutofolio, flore purpureo, variegata”) corresponding to page and plate 176 (Hermann 1698, p. 176), a species or form with notably small and non-multifid bracts.

Miller’s description from cultivated plants is more helpful, including bracts

“netted” but not long and enveloping the flower, corona “light blue,” and fruit maturing to a “deep yellow colour.” The taxonomic identity of this species is unclear, though the color of the fruit indicates an affinity with P. vesicaria. We have not located a type specimen in any major herbarium, and are hesitant to typify this name without further searching. 55

Invalid and Illegitimate Names

Names that are invalid, illegitimate, or otherwise do not meet the requirements of a validly published name are listed and discussed for completeness.

1. Dysosmia hircina (Sweet ex Loudon) M.Roemer (1846, p. 150), pro syn.

≡ Passiflora hircina Sweet ex Loudon (1830, p. 270), nom. nud.

Notes:—Because there is no description or illustration for this species nor for its basionym, both names are invalid.

2. Passiflora chrysophylla Chodat var. typica Chodat (1902, p. 745), nom. inval.

Notes:—Invalid according to Art. 24.3 of the ICN (McNeill et al., 2012).

3. Passiflora foetida Linnaeus var. baraquiniana (Lemaire) Vanderplank (2013, p. 383), nom. superfl.

Notes:—This combination is superfluous and therefore invalid according to Art.

52 of the ICN (McNeill et al., 2012), due to the existence of the homotypic P. foetida var. hirsuta (see taxon 2, above).

4. Passiflora foetida Linnaeus var. glabrifolia Miquel ex Triana & Planchon (1873, p.

172), pro syn.

5. Passiflora foetida Linnaeus f. latifolia Kuntze (1891, p. 254), nom. nud.

6. Passiflora foetida Linnaeus var. pectinata Holtz (1892, p. 2), nom. nud.

7. Passiflora foetida Linnaeus f. quinqueloba Masters ex Killip (1938, p. 511), pro syn.

8. Passiflora hibiscifolia Lamarck var. glabrata J.Jacquin (1844, p. 5), as ‘glabratum’, nom. nud.

9. Passiflora hirsuta J.Jacquin (non Linnaeus, 1753) (1844, tab. 123), nom. illeg. 56

Notes:—Joseph Jacquin’s (1844) “P. hirsuta” is a later homonym of P. hirsuta

Linnaeus and represents an ambiguous taxon in Dysosmia named P. hibiscifolia var. velutina by Eduard Fenzl (see name 4, above).

10. Passiflora marigouja Perrottet ex Triana & Planchon (1873, p. 172), pro syn.

11. Passiflora villosa Dombey ex Triana & Planchon (non Vellozo, 1831) (1873, p. 154), pro syn.

12. Passiflora villosa Macfadyen (non Vellozo, 1831) (1850, p. 151), nom. inval.

Notes:—Macfadyen’s Flora of Jamaica (1850) was not effectively published, and therefore any names published therein are invalid.

Acknowledgements

We are grateful to the directors of the MO, NY, and US herbaria for allowing the study of their collections and for loans of specimens. We also thank Marc Appelhans

(GOET), Uwe Braun (HAL), Lynda Brooks (LINN), Giovanni Cristofolini (BOLO), Åsa

Dalsätt (S), Sara Edwards (K), Robin Everly (US), Colleen Funkhouser (US), Laurent

Gautier (G), Jonathan Gregson (BM), Esther Jackson (NY), Ingrid Lin (US), Armin

Löckher (W), Vera Lúcia C. Martins (R), Carlos Parra (COL), Henrik Pedersen (C), Nina

Rønsted (C), Olof Ryding (C), Piet Stoffelen (BR), Ernst Vitek (W), Bruno Walnöfer

(W), Thomas Zanoni (NY), and Sue Zmarzty (K) for taking the time to search for specimens and/or providing images from their respective herbaria.

We further acknowledge John McNeill (E), Kanchi Gandhi (GH) and the late

James Reveal (MARY) for their nomenclatural expertise, Charlie Jarvis (BM) for his help and advice with Linnaean names and material, and Roy Gereau for Latin translation. 57

We also extend our thanks to an anonymous reviewer and Christian Bräuchler who helped improve this manuscript.

Figure 2-1. Designated lectotype of Passiflora baraquiniana. Image provided by the Peter H. Raven Library, Missouri Botanical Garden.

Figure 2-2. Designated neotype of Passiflora ciliata. Used by permission of the Linnean Society of London. 60

Figure 2-3. Designated lectotype of Passiflora gossypiifolia. Image provided by the Muséum national d’Histoire naturelle, Paris. 61

Figure 2-4. Designated lectotype of Passiflora hispida. Image provided by the Muséum national d’Histoire naturelle, Paris. 62

Figure 2-5. Lectotype of Passiflora foetida. A. The specimen in its entirety. B. An enlarged portion of the specimen showing Linnaeus’ handwriting. Used by permission of the Linnean Society of London.

CHAPTER 3: PHENETIC AND CLADISTIC STUDIES HELP CLARIFY SPECIES

ASSEMBLAGES IN PASSIFLORA SECTION DYSOSMIA (PASSIFLORACEAE)

This article has already been published and is reprinted by permission from

Springer Nature (Phenetic and cladistic studies help clarify species assemblages in

Passiflora section Dysosmia (Passifloraceae), H. T. Svoboda and H. E. Ballard, Jr.,

Brittonia 70(1), Copyright © 2017, The New York Botanical Garden, Springer Nature; see Appendix A for more information). All text is retained from the publication, with minor adjustments to some citations to comply with the formatting of this document.

In the article, we test whether section Dysosmia could be divided into more manageable groups or lineages based on morphology to aid in future studies of the group.

Notably, the results are the first to provide evidence for groupings based on vegetative pubescence and fruit color.

Harlan T. Svoboda and Harvey E. Ballard, Jr.

Department of Environmental and Plant Biology, Ohio University, 315 Porter Hall,

Athens, OH 45701, USA

Received 24 January 2017; Accepted 22 August 2017; Published 22 September 2017

(online), 30 March 2018 (print).

Abstract

One of the most enigmatic and unusual groups in the passionflower genus,

Passiflora section Dysosmia (Passifloraceae), stands out as a group that is notoriously taxonomically complicated. Phenetic and cladistic analyses of Dysosmia were carried out with representatives from all 21 species currently recognized in the section, in order to 64 delineate morphological groups above the species level. The study was based mainly on vegetative morphological characters and included principal coordinates, non-metric multidimensional scaling, cluster, and cladistic analyses. Results from each analysis reveals that three major species assemblages exist within Dysosmia, corresponding largely to vegetative pubescence and fruit color. The results presented here set the stage for future systematic and phylogenetic studies in Passiflora sect. Dysosmia.

Introduction

The family Passifloraceae Juss. ex Roussel is comprised of approximately

750 species divided into 17 genera that are distributed throughout the tropics in both the

Old and New World (Feuillet & MacDougal, 2006). The most species-rich of the genera,

Passiflora L., includes over 560 species of passionflowers with a distribution primarily in

Central and South America (Krosnick et al., 2013). Members of Passiflora show an array of unique floral characters, and also express substantial phenotypic variation. Our current understanding of Passiflora classification comes from Feuillet and MacDougal (2003) in which the genus was divided into four subgenera: Astrophea (DC.) Mast., Decaloba

(DC.) Rchb., Deidamioides (Harms) Killip, and Passiflora.

A notoriously difficult lineage, Dysosmia DC. (currently circumscribed at the level of section in Passiflora subgenus Passiflora), has perplexed botanists for nearly 200 years, due in part to the section’s vast diversity in leaf morphology (see Figure 3-1). This diversity was a reason for Killip (1938) to hypothesize that few, if any, other plant groups encompass as much diversity in leaf morphology as the passionflowers. Based on the classification proposed by Feuillet and MacDougal (2003), sect. Dysosmia belongs to 65 supersection Stipulata Feuillet & J. M. MacDougal, subgen. Passiflora. The section was revised recently (Vanderplank, 2013) and currently contains 21 species and a number of infraspecific varieties—the second largest section in supersect. Stipulata (the largest being sect. Granadillastrum Triana & Planch. with approximately 68 species). Taxa in

Dysosmia are distributed from southern North America, throughout Central America, south to southern South America, and east among the Caribbean islands (Figure 3-2), with some recent introductions into the Old World tropics (Vanderplank, 2013).

As a section, Dysosmia can be been distinguished from other Passiflora groups in having 1) deeply cleft, often glandular, stipules (Figure 3-3A, B), 2) highly filiform, pinnatisect, and often glandular involucral bracts (Figure 3-3C, D), and 3) leaves that bear epidermal glands or glandular trichomes on the abaxial surface (Figure 3-3E, F).

Although taxa from distantly related groups in the genus may also have one or more of these traits (e.g., the filiform involucral bracts of Passiflora trisecta Mast.), this suite of characters is definitively unique to sect. Dysosmia.

Dysosmia, at its many ranks, has been treated in a number of monographs and synopses (namely Masters, 1871; Killip, 1938; Vanderplank, 2013) but to this day a comprehensive, comparative morphological study of each taxon in the section has not been conducted. Although gross morphology has historically played an important role in delimiting species and infraspecific taxa, the suite of traits (including micromorphological characters) that define any given specimen in Dysosmia has largely been ignored for taxonomic purposes. 66

Svoboda, MacDougal, and Ballard (2016) provided typification of every valid name in Dysosmia, thereby reducing the confusion between nomenclature and biology, but because of the difficulty in distinguishing between the taxa within Dysosmia, phenetic and cladistic analyses were carried out to assess the utility of morphological characters in taxonomic circumscription. The study presented here is based on a typological approach in order to encompass the breadth of morphological diversity within the section, above the species level, without the necessity of assessing the entirety of species-level variation.

Methods

A total of 55 specimens, representing all 21 species and the limits of gross morphological variation in section Dysosmia, were consulted for this study (Table 3-1).

Also included in the material are multiple accessions of the two largest species complexes, P. ciliata Aiton and P. foetida L. Because this study focuses on broader scale

(i.e., above the level of species) differences, one to two specimens per species were used, with 26 type specimens acting as “anchors” of known morphological diversity within

Dysosmia. When nomenclatural types were not available for a particular species, a single specimen was selected to represent the most typical form of that taxon in terms of characters measured in this study (i.e., pubescence, gland shape, bract dividedness, etc.).

Herbarium specimens from MO, NY and US were studied in person, supplemented with additional scanned images from BHCB, CAS, GH, K, and P found online (acronyms follow the Index Herbariorum, http://sweetgum.nybg.org/science/ih/). 67

A total of 24 qualitative and quantitative characters (Table 3-2) were evaluated in this study. These characters include some that have been traditionally used by taxonomists (i.e., leaf lobing, bract divisions, stem pubescence) as well as some that that have so far been overlooked for systematic purposes (i.e., gland structure, leaf basal angle). Quantitative characters taken from the leaves, as depicted in Figure 3-4, were measured from herbarium specimens using digital calipers or digitally from scanned images using ImageJ developed by W. Rasband (version 1.49; Schneider, Rasband, &

Eliceiri, 2012).

The data were compiled into a matrix of 55 rows of specimens, or operational taxonomic units (OTUs), and 24 columns of characters [see Suppl. Material 1 of Svoboda

& Ballard (2018)]. Both of the phenetic analyses and the cluster analyses were performed using the Paleontological Statistics (PAST; version 3.11) software developed by

Hammer, Harper, and Ryan (2001). Each column was coded to be either “nominal” or

“binary” depending on the type of data collected for that character.

Principal coordinates analysis (PCoA; Figure 3-5) and non-metric multidimensional scaling (NMDS; Figure 3-6) were performed on the morphological data using Gower’s similarity index (Gower, 1966). For the cluster analysis, a Gower similarity index was calculated and a phenogram produced using the unweighted pair group with the arithmetic means (UPGMA) method (Figure 3-7).

The 18 discrete variables were subjected to a cladistic analysis using parsimony with the software PAUP* (version 4.0b10; Swofford, 2002). The characters were treated as ‘unordered’ and of equal weight. A heuristic search with 1000 replicates was used, 68 with P. boticarioana Cervi (“bo_465”), P. campanulata Mast. (“cm_327”), P. hypoglauca Harms (“hy_131”), and P. vellozoi Gardner (“vz_631”) serving as the outgroup taxa because of their morphological affinities with other taxa outside of

Dysosmia (not included in this study). Trees were constructed with the tree-bisection- reconnection (tbr) algorithm and evaluated using a stepwise-addition search limited to

100 most parsimonious trees. Although the strict consensus tree was unresolved, a well- resolved 50% majority-rule consensus tree was selected (Figure 3-8) and visualized and formatted in FigTree version 1.4.2 (Rambaut, 2014).

Results and Discussion

The ordination resulting from the PCoA (Figure 3-5) shows that there are at least three morphologically distinct groups within section Dysosmia. One of these groups, designated as the Passiflora ciliata complex and depicted with red squares on the left of the ordination, includes all OTUs that are completely glabrous. Interestingly this group also bears vivid red fruit—a character that was not recorded in the study, and a condition that is exceptionally rare across Passiflora (J. MacDougal, personal communication).

The second group occupies the upper right of the ordination, depicted with purple triangles, and represents pubescent species that were previously placed in a segregate subgenus, Dysosmioides Killip (Killip, 1938) with the addition of the species P. boticarioana which likely would have been placed in this section had it been discovered at the time. These taxa, designated as the Dysosmioides group, share many of the unusual morphological features of Dysosmia but lack glandular trichomes on the plant body. 69

A third group, depicted by green circles in the right portion of the ordination and designated as the P. foetida complex, includes OTUs that are pubescent and bear green

(or rarely yellow) fruit. A hypothesized fourth group, here informally designated as the red-fruited pubescent (RFP) group and depicted with blue crosses in the lower right portion of the ordination, contains OTUs that are also pubescent but bear red fruit.

However, this putative group does not separate out well from the former group and appears not to be morphologically distinct from other pubescent OTUs in our analyses.

The first and second coordinates of the PCoA represent 27.28% and 14.96% of the variation, respectively, followed by coordinate three representing only 7.02% of the variation.

Similarly, the ordination resulting from the NMDS analysis (Figure 3-6) shows separation of the P. ciliata group (red squares in the right portion of the ordination) as well as the Dysosmioides group (purple triangles in the upper left of the ordination) from the remaining pubescent taxa. Once again the P. foetida (green circles) and RFP (blue crosses) groups, in the lower left portion of the ordination, overlap considerably. The

Shepard plot [see Suppl. Material 2 of Svoboda & Ballard (2018)] R2 values for axes one and two were 0.6060 and 0.2287, respectively, with a stress value of 0.2010, indicating a fair representation of the input data.

The results from the cluster analysis (Figure 3-7) are consistent with the results from the PCoA and NMDS analyses. The glabrous, red-fruited OTUs (red squares) form a distinct cluster occupying the right of the phenogram. Most of the Dysosmioides group

(purple triangles) appears as a cluster at the far left of the phenogram, with two OTUs 70 found in the center along with the rest of the pubescent taxa. The phenogram also shows little evidence that the RFP (blue crosses) and pubescent, green-fruited (green circles)

OTUs are morphologically distinct from each other, as these taxa appear mixed in with one another.

The cladogram resulting from the cladistic analysis (Figure 3-8) further supports that three distinct groups exist in Dysosmia. The seven most basal OTUs near the root of the tree (depicted by purple triangles) are from the Dysosmioides group, though they do not form a monophyletic clade. The RFP taxa (blue crosses) show no order, as they are scattered throughout the pubescent, green-fruited OTUs (green circles) in the upper portion of the tree. Once again the glabrous, red-fruited OTUs (red squares) form a separate group near the middle of the cladogram.

Conclusions

The consistent separation of three morphological clusters in each of the analyses indicates that there are at least three major phenotypically (and potentially evolutionarily) distinct lineages within section Dysosmia, with a unique suite of traits defining each of them.

Although fruit color was not measured as part of the original dataset, it has been shown in in each of our analyses to be a defining feature of the Passiflora ciliata group.

The combination of glabrous vegetative features and red fruit are, therefore, likely a synapomorphy for the group. Researchers of the genus have long suspected the existence of a “red-fruited clade” in Dysosmia (Feuillet, 2007; J. MacDougal, personal communication), but it has, until now, never been shown with substantial data. Fruit color 71 has also proven to be a phylogenetically informative character in a number of other studies (see Givnish, Sytsma, Smith, & Hahn, 1995; Miller, Mione, Phan, & Olmstead,

2011), despite complicated vegetative morphologies. In each case, fruit color appears to match with molecular phylogenies and provide a framework for species-level relationships.

The putative fourth group—the red-fruited pubescent (RFP) assemblage— although exhibiting traits found in both the P. ciliata and P. foetida groups, did not show any separation from the two other pubescent groups in any of the analyses. This hinders our understanding of the evolutionary history of the RFP group, but may indicate that red fruit has evolved more than once in Dysosmia or that reticulate evolution and hybridization is occurring.

Based on the foundational phenetic data presented here, future studies using geometric morphometrics of leaf shape will aim to resolve species boundaries and discovery of novel phenotypes in Dysosmia. Likewise, the data present a basis for comprehensive molecular studies that will be used to test hypotheses of monophyly within the section and whether such clades do correspond to fruit color.

Acknowledgements

This paper is part of the doctoral dissertation of the first author. The project was supported in part by an Ohio University Graduate Student Senate Original Works Grant awarded to HTS. We are grateful to the directors of the MO, NY, and US herbaria for kindly providing access to their collections and for loans of specimens, as well as three 72 anonymous reviewers who helped improve this manuscript. We further extend our thanks to John MacDougal for creative insights and guidance throughout the project.

Figure 3-1. A sampling of the variation in leaf shape and lobing found in section Dysosmia. A. Leaf of P. campanulata Mast. B. Leaf of P. ciliata Aiton var. quinqueloba Griseb. (isotype). C. Leaf of P. ciliata [variety undesignated]. D. Leaf of P. hypoglauca Harms. E. Leaf of P. foetida L. [variety undesignated]. F. Leaf of P. lepidota Mast. G. Leaf of P. setulosa Killip. H. Leaf of P. pectinata Griseb. I. Leaf of P. hibiscifolia Lam. (holotype).

Figure 3-2. Geographical distribution of the OTUs used in this study. Created using SimpleMappr (http://simplemappr.net).

Figure 3-3. Representative structures seen in Dysosmia taxa. A. Highly dissected stipule of P. foetida var. longipedunculata Killip (holotype). B. Leafier stipule of P. vellozoi Gardner. C. Highly pinnatisect involucral bract of P. vestita Killip. D. Leafier involucral bract of P. setulosa. E. Bulbose glands on the abaxial surface of a P. lepidota leaf. F. Capitate glands on the leaf of P. palmeri Rose.

Figure 3-4. A representation of the leaf measurements used in this study. Apical and basal angles reported in degrees, while all other measurements were reported in centimeters.

Figure 3-5. Ordinations resulting from a principal coordinates analysis (PCoA). A. Coordinates one and two of the PCoA. B. Coordinates one and three of the PCoA. In each ordination, the P. ciliata group (red squares) shows separation from the other clusters. Blue crosses represent the red-fruited pubescent (RFP) group; purple triangles represent the Dysosmioides group; green circles represent the P. foetida group. 78

Figure 3-6. Ordination resulting from a non-metric dimensional scaling (NMDS) analysis. At least three clusters appear in the ordination, with the P. ciliata (red squares) and Dysosmioides (purple triangles) groups showing clear separation. See Figure 3-5 for explanation of other colors and symbols.

Figure 3-7. Phenogram showing the clustering of two major groups: one with glabrous features and the other with pubescence. See Figure 5 for explanation of colors and symbols. 80

Figure 3-8. Fifty percent majority rule consensus cladogram from a parsimony analysis in PAUP. The cladogram is based on 18 discrete morphological characters. See Figure 3-5 for explanation of colors and symbols.

Table 3-1

List of specimens used in this study and their codes used in the phenetic analysis. Herbarium acronyms follow the Index Herbariorum (http://sweetgum.nybg.org/science/ih/). Code Taxon name (current determination) Country, State Collection data Herbarium ar_405 Passiflora arida Mexico, Sonora J. N. Rose 1206b NY az_002 P. arizonica United States, Arizona G. J. Harrison 4774a US ba_318 P. bahamensis Bahamas, New Providence N. L. Britton & L. J. K. Brace 392a NY bo_465 P. boticarioana Brazil, Minas Gerais J. A. Lombardi et al. 5741a BHCB cm_327 P. campanulata Brazil, Rio de Janeiro W. Lobb s.n.a K ch_805 P. chrysophylla var. hastata Canindeyú, Paraguay E. Hassler 5934b P ci_672 P. ciliata [complex] Nicaragua, Atlántico Norte Stevens 7809 MO ci_696 ——— Nicaragua, Atlántico Norte Stevens 7758 MO ci_911 ——— Honduras, Gracias a Dios Abraham P. 121 MO ci_736 ——— Mexico, Oaxaca Pringle 4847 MO ci_818 ——— Mexico, Quintana Roo Carnevali et al. 4841 MO ci_631 ——— Mexico, Oaxaca Spellman 2109 MO ci_423 ——— Mexico, Tabasco Menéndez L. et al. 276 MO ci_713 ——— Mexico, Quintana Roo Cabrera C. & de Cabrera 3528 MO ci_995 ——— Mexico, Yucatán Steere 1089 US ci_098 ——— Cuba, Holguin Howard 6133 US ci_426 ——— Belize, Toledo Balick et al. 3589 MO ci_170 ——— Honduras, Atlantida MacDougal et al. 3447 US 82

Table 3-1 continued

ci_293 P. ciliata var. quinqueloba Cuba C. Wright 2601b MO ci_294 P. ciliata var. riparia Cuba C. Wright 2602b MO cl_093 P. clathrata Brazil, Goiás Heringer & Eiten 14195 US fo_955 P. foetida [complex] Belize, Cayo Dwyer et al. 328 MO fo_671 ——— Mexico, Yucatán Cabrera C. & de Cabrera 8792 MO fo_301 ——— Guatemala, Petén Contreras 715 MO fo_655 ——— Mexico, Campeche Cabrera C. 2277 MO fo_825 P. foetida var. acapulcensis Mexico, Guerrero E. Palmer 306a US fo_964 P. foetida var. eliasii Colombia, Atlántico Bro. Elias 467a US fo_556 P. foetida var. hibiscifolia Mexico, Oaxaca Rose et al. 10056 US fo_900 P. foetida var. hirsutissima Guatemala, Alta Verapaz M. W. Owen 9a US fo_247 P. foetida var. isthmia Panama, Colón W. R. Maxon 7013a US fo_914 P. foetida var. lanuginosa Mexico, Veracruz F. M. Liebmann 4096 [53]a US fo_641 P. foetida var. longipedunculata Mexico, Tamaulipas H. H. Bartlett 10987a US fo_248 P. foetida var. maxonii Nicaragua, Managua W. R. Maxon 7219a US fo_061 P. foetida var. mayarum Belize, Belize P. H. Gentle 6a US fo_235 P. foetida var. oaxacana Mexico, Oaxaca E. W. Nelson 2762a US fo_996 P. foetida var. orinocensis Venezuela, Bolívar L. H. Bailey & E. Z. Bailey 1773a US fo_833 P. foetida var. parvifolia Mexico, Guerrero E. Palmer 315a US fo_761 P. foetida var. salvadorensis El Salvador, Sonsonate P. C. Standley 22006a US fo_044 P. foetida var. sanctae-martae Colombia, Magdalena E. P. Killip & A. C. Smith 21103a US 83

Table 3-1 continued

fo_007 P. foetida var. santiagana Cuba, Santiago de Cuba Morton et al. 3319 US fo_078 P. foetida var. subintegra Belize, Stann Creek W. A. Schipp 648b US fo_891 P. foetida var. tepicana Mexico, Nayarit J. N. Rose 1998a US fr_013 P. fruticosa Mexico, Baja California Sur León de la Luz & Goldman 6108 US hy_131 P. hypoglauca Brazil, Minas Gerais Barreto 8969 US le_940 P. lepidota Brazil, Paraná Cordeiro & Hatschbach 499 US pl_012 P. palmeri Mexico, Baja California Sur E. Palmer 868b GH pe_451 P. pectinata Bahamas, Exuma Correll 40773 MO st_360 P. setulosa Brazil, Paraná Kummrow 2090 MO su_638 P. sublanceolata Guatemala, Petén H. H. Bartlett 12788a US ur_166 P. urbaniana Belize, Belize Arvigo et al. 354 NY vz_631 P. vellozoi Brazil, Paraná Hatschbach & Ramos 57920 MO ve_099 P. vesicaria var. galapagensis Ecuador, Galápagos J. T. Howell 8833a CAS vs_904 P. vestita Ecuador, Napo Brandbyge et al. 32693 MO vi_675 P. villosa Brazil, Minas Gerais Macedo 3088 US vi_825 ——— Brazil, Goiás Irwin et al. 33065 NY a Holotype specimen. b Isotype specimen.

Table 3-2

The characters and their states used in the phenetic analysis. Character State Leaf midrib length (cm) Leaf widest point (cm) Leaf proximal basal vein length (cm) Leaf distal basal vein length (cm) Leaf apical angle (degrees) Leaf basal angle (degrees) Leaf lobing 0 = unlobed, 1 = shallowly three-lobed, 2 = deeply three-lobed, 3 = shallowly five-lobed, 4 = deeply five-lobed Leaf pubescence 0 = absent, 1 = present Abaxial leaf surface glands 0 = absent, 1 = present Adaxial leaf surface glands 0 = absent, 1 = present Leaf gland structure 0 = absent, 1 = capitate (long), 2 = capitate (short), 3 = bulbous Leaf marginal glands 0 = absent, 1 = present Leaf pubescence 0 = absent, 1 = present Petiole pubescence 0 = absent, 1 = present Petiole glands 0 = absent, 1 = present Stipule pubescence 0 = absent, 1 = present Stipule glands 0 = absent, 1 = present Stipule structure 0 = thick or leafy, 1 = crown-like, 2 = entirely divided Stipules pinnate 0 = no, 1 = yes Bract pubescence 0 = absent, 1 = present Bract glands 0 = absent, 1 = present Bract divisions 0 = bipinnatisect, 1 = tripinnatisect, 2 = not pinnate, 3 = once pinnatisect Bract axis 0 = thick, 1 = medium, 2 = thin Ovary pubescence 0 = absent, 1 = present

CHAPTER 4: LEAF GEOMETRIC MORPHOMETRICS

Introduction

Morphology, particularly that of leaf shape, has long played a role in plant systematic and alpha taxonomic research. Before the advent of modern genetic techniques, researchers were often confined to using only morphology to describe new species and distinguish taxa from one another. Now, with rapid progression in molecular phylogenetics, genomics, next generation sequencing, etc., reliance on gross morphology has somewhat lessened in many fields with respect to species discovery and taxonomy; this being particular true when morphology alone has not consistently convinced researchers of taxonomic (or nomenclatural) decision.

Still, even after the “DNA revolution” much of systematic research relies heavily on morphology. A quick glance at articles published in plant systematic journals (e.g.,

Systematic Botany, Taxon, etc.) reveals that gross morphology still plays an integral role in a wide array of taxonomic work, including the description of new species which requires a description or diagnosis that separates said new taxon from all other known taxa in accordance with the Code of Nomenclature (McNeill et al., 2012). For tasks such as identifying herbarium specimens, it could be argued that much of the sorting, at least initially, is based on gross morphology—with subsequent, finer-scale observations leading to proper identification. The same can be said for identifying plants in the field, in which gross morphology can be used, at least coarsely, to separate groups of plants (at various taxonomic levels) and identify those that are of interest. 86

Much of the morphological descriptors or decision-making based on leaf morphology in the past, even today, has been arguably subjective or qualitative in nature.

An evaluation of “how different” one leaf is from another leaf may differ greatly between taxonomists. Geometric morphometrics is one answer to this disparity that aims to quantify exactly “how different” one leaf is from another in an objective manner. Various methodologies exist, with the two main techniques being generalized Procrustes analysis

(GPA; landmarks) and elliptical Fourier descriptors (EFD; outlines). These methods, separately or together, provide an objective way of describing morphology that can be used in a variety of investigations—taxonomic delimitation being the focus of this article and research presented in this chapter.

Comprehensive Methods for Leaf Geometric Morphometric Analyses

This article was published in the online journal Bio-protocol (Klein & Svoboda,

2017) and is reprinted here by permission of the authors. All text is retained from the publication, with minor adjustments to the citations to comply with the formatting of this document.

Here we present a novel workflow for using geometric morphometrics for leaf shape quantification that can be useful in a number of fields.

Laura L. Klein1 and Harlan T. Svoboda2

1Department of Biology, Saint Louis University, St. Louis, Missouri, USA

2Department of Environmental and Plant Biology, Ohio University, Athens, Ohio, USA

Received 26 December 2016; Accepted 27 March 2017; Published 5 May 2017.

Abstract

Leaf morphometrics are used frequently by several disciplines, including taxonomists, systematists, developmental biologists, morphologists, agronomists, and plant breeders to name just a few. Leaf shape is highly variable and can be used for identifying species or genotypes, developmental patterning within and among individuals, assessing plant health, and measuring environmental impacts on plant phenotype. Traditional leaf morphometrics require hand tools and access to specimens, but modern efforts to digitize botanical collections make digital morphometrics a readily accessible and scientifically rigorous option. Here we provide detailed instructions for performing some of the most informative digital geometric morphometric analyses available: generalized Procrustes analysis, elliptical Fourier analysis, and shape features.

This comprehensive procedure for leaf shape analysis is comprised of six main sections:

A) scanning of material, B) acquiring landmarks, C) analysis of landmark data, D) isolating leaf outlines, E) analysis of leaf outlines, and F) shape features. This protocol provides a detailed reference for applying landmark and outline analysis to leaf shape as well as describing leaf shape features, thus empowering researchers to perform high throughput phenotyping for diverse applications.

Background

There are a variety of approaches to digital leaf shape morphometrics, including outline or Fourier analysis, contour signatures, landmark analysis, shape features, fractal dimensions, and texture analysis (Cope, Corney, Clark, Remagnino, & Wilkin, 2012).

Among these analyses, landmark and Fourier analysis together perform exceptionally 88 well at distinguishing between groups among leaf shapes (McLellan & Endler, 1998;

Hearn, 2009). Landmark analysis is ideal for capturing aspects of shape that are consistent among all leaves within a given dataset. The selection of landmarks should include points that are biologically homologous and adequately represent the morphology of the leaf (see more pointers in Bookstein, 1991; Zelditch, Swiderski, Sheets, & Fink,

2004). If leaves in the dataset do not have evolutionarily conserved shape features,

‘pseudo-landmarks’ can instead be placed (Chitwood & Sinha, 2016); that is, landmarks can be placed at equidistant points along the leaf outline relative to homologous points that act as anchors. Landmarks can then be analyzed using Generalized Procrustes

Analysis (GPA), which normalizes shape data (annotated by landmarks) at equal scale, allowing for an accurate comparison of shapes regardless of their size. Outline analysis offers a more broadly applicable phenotyping method in that Elliptical Fourier

Descriptors (EFDs) are used to build shape descriptors of the leaf outline (Kuhl &

Giardina, 1982; Iwata & Ukai, 2002). While sensitive to noise, EFDs are ideal for large leaf datasets that have subtle differences between shapes. Shape features are an additional, simple method of outline analysis that can include the perimeter to area ratio, aspect ratio, and circularity measurements (Cope et al., 2012). In this protocol, we focus on aspect ratio and circularity, as they detect signatures of lobing and serration. Aspect ratio is the ratio of the major axis to the minor axis of a fitted ellipse, in which case values close to ‘1’ are more circular in shape regardless of lobing. Circularity is the ratio of the leaf area to perimeter outline. This measurement is useful for discriminating leaves with lobing and serration, with low circularity values indicating significant lobing and 89 serration. This protocol is designed such that researchers can choose between all three methods (GPA, EFD, and shape features) based on which analyses best fit their data.

Materials and Reagents

1. Herbarium specimens and/or fresh leaf material

Equipment

1. Computer that can run Microsoft® Windows® XP (or later) and/or Mac® OS X®

10.4 (or later)

2. Flatbed photo scanner (Epson Expression, model: 10000XL)

Software

1. Adobe® Photoshop® CS4 (or later)

2. Epson® Scan Utility v3.4.9.6 (https://support.epson.com/)

3. JavaTM (https://java.com/)

4. ImageJ (https://imagej.nih.gov/ij/)

5. Microsoft® Excel® 2011 (or later)

6. SHAPE v1.3 (http://lbm.ab.a.u-tokyo.ac.jp/~iwata/shape/)

6a. SHAPE is built for Windows but if using a Mac, Wine and

Winebottler (http://winebottler.kronenberg.org/) are required

7. R (https://r-project.org/)

7a. Packages: ‘shapes,’ ‘ggplot2,’ ‘devtools,’ ‘ellipse,’ and ‘roxygen2’

8. RStudio (https://rstudio.com/products/rstudio/)

8a. RStudio is an optional user interface for R

Procedure

Note: Examples of R scripts and ImageJ macros referenced throughout the protocol can be freely downloaded from GitHub (link: https://github.com/htsvoboda/LeafGeometricMorphometrics.git; Note 1).

Scanning Fresh Leaves or Herbarium Specimens

1. For fresh leaves: Place leaves, with petioles removed, flat on the scanner bed.

Multiple leaves can be placed on the scanning bed at once, so long as they do not overlap.

If the background of the scanner is not already white, place a white piece of paper on top of the leaves; subsequent analyses work best with a solid white background (Note 2).

2. For herbarium specimens: carefully place the entire sheet face-down on the scanner bed. Any loose material should be placed in a fragment packet on the sheet.

3. Scans should be saved as .jpg or .tif files and named with a unique identifier

(e.g., accession number). Note: It is best to scan at 400 dpi (or a higher resolution) as this improves the quality of the image.

Acquiring and Exporting Landmarks

1. Create a spreadsheet in Excel with the following column names: ‘order,’

‘label,’ ‘x,’ and ‘y.’ This file will be referred to as the ‘master spreadsheet.’

2. Open ImageJ. For the first use: Select ‘Analyze > Set Measurements…’ and check only the ‘Display label’ checkbox. ‘Redirect’ and decimal place parameters can be left at their defaults.

3. From the ImageJ menu bar, select the ‘Point selection’ tool (Figure 4-1; Note

3). 91

Figure 4-1. The ImageJ menu bar. Here, the ‘Point selection’ tool button is selected.

4. Open the first image to be landmarked. In the menu bar, select ‘File > Open...’

5. Using the ‘Point selection’ tool in multi-point mode, begin placing landmarks on your predetermined landmark points (see Figure 4-2; Notes 4 and 5).

Figure 4-2. Placement of landmarks on some representative leaves. A. Seventeen landmarks placed on two fresh leaves of Vitis riparia; B. Six landmarks placed on a leaf of Passiflora campanulata from a herbarium specimen. 92

6. Once all landmarks have been applied to a leaf (or to multiple leaves per image), view the landmark data. In the menu bar, select ‘Analyze > Measure.’ A new window, called ‘Results’ (Figure 4-3), will appear with the x, y coordinates for each landmark.

Figure 4-3. Example of the ‘Results’ window produced by ImageJ. This is viewed by selecting ‘Analyze > Measure’ in the menu bar. 93

7. Copy (CTRL + C) and paste (CTRL + V) these data from the ‘Results’ window directly into the ‘master spreadsheet’ (see Figure 4-4).

Figure 4-4. Example format of the ‘master spreadsheet.’

8. Close the ‘Results’ and image windows and repeat this process with each image file until all leaves have been landmarked. ‘File > Open Next’ can be used if all of the images are in the same folder. 94

9. Landmark coordinates for every leaf will be pasted as-is into the ‘master spreadsheet’. The column ‘order’ may need to be adjusted at the end to reflect a continuous numerical set.

10. Before analyzing the landmark data, the ‘master spreadsheet’ will need to be converted from Excel format (.xlsx) to a Tab Delimited Text format (.txt).

Click ‘File > Save As…’ and select the .txt option.

11. Data within the .txt file will need to be reformatted such that the rows are single leaves and the adjacent columns represent the landmark data (x, y coordinates) in sequential order. This can be done in R using the code from GitHub (file:

‘Protocol_stepB-11.R’; R Core Team, 2017).

12. Import the reformatted file (referred to as ‘reformatted.txt’) into Excel to verify the data was properly written.

12a. Click ‘File > Import...’ and choose ‘space’ and/or ‘tab’ as the delimiter (see

Note 6).

13. Supplemental columns and information (i.e., species, genotype, sex, etc.) can be added at this point to help with downstream analyses (Figure 4-5).

Figure 4-5. The reformatted ‘master spreadsheet’ (‘reformatted.txt’) as seen in Excel. The spreadsheet now displays each row as a leaf with its landmarks distributed across the columns. Additional information can be added to the reformatted spreadsheet that may be useful in downstream analyses (e.g., columns B-F).

14. Check landmarks for accuracy. This can be done by plotting the landmark coordinates in R with the package ‘ggplot2’ (Figure 4-6; Wickham, 2009). The corresponding R script can be downloaded from GitHub (file: ‘Protocol_stepB-14.R’).

Figure 4-6. Landmarks are checked for accuracy using the R package ggplot. A. Example of specimen scan; B. The companion plotted landmark coordinates to the scan (A); C. Incorrectly placed landmarks will be immediately apparent visually. Note that B) and C) are inverted to that of A), as pixel coordinates map inversely in a linear regression.

15. If there appears to be an obvious error in the position of the landmarks for any image (as in Figure 4-6C), redo landmark placement (step B5) for the affected leaf/file and paste the updated coordinates into the ‘master spreadsheet.’ Repeat steps B11-B12 to reformat and check landmark accuracy for the new spreadsheet.

Analysis of Landmarks: Generalized Procrustes Analysis (GPA)

1. To perform GPA using the R package ‘shapes,’ (Dryden, 2017) the input file must consist of only landmark data.

1a. Import (‘File > Import...’) the ‘reformatted.txt’ file into Excel.

1b. Remove all column headers and any other column information.

1c. In the menu bar, select ‘File > Save As…’ and rename the file to distinguish that it contains only x, y coordinates (e.g., ‘reformatted_coords.txt’).

2. The data can now be processed in R using the ‘shapes’ package. 97

2a. The analysis produces Procrustes principal component scores and percent variance explained, Eigenleaves, Eigenvalues, among other informative outputs. The R script includes code that will write files containing PC scores and percentages for further analysis. Example R script can be found on GitHub (file: ‘Protocol_stepC-2-a.R’).

2b. Because leaf order is preserved in the output files, any additional information

(e.g., individual leaf identity, species, genotype, etc.) can be re-added to these output files by adding additional columns to the file, then pasting the additional information for each leaf from the ‘reformatted.txt’ file into the PC score file.

2c. The leaves can be visualized in morphospace, using the packages ‘ggplot2,’

‘devtools,’ (Wickham and Chang, 2015) ‘ellipse,’ (Murdoch, Chow, & Celayeta, 2007) and ‘roxygen2’ (Wickham, Danenberg, & Eugster, 2015). Example code can be found on

GitHub (file: ‘Protocol_stepC-2-c.R’). An example ordination can be seen in Figure 4-7. 98

Figure 4-7. PCA ordination resulting from a Generalized Procrustes Analysis using an example landmark dataset. In this example, the leaves were labeled and color-coded according to the species identity.

Isolating Leaves from Scans

1. For scans of fresh leaves

1a. Open ImageJ.

1b. Create a macro to isolate leaf images as binary images.

1b.i. In the menu bar, select ‘Plugins > New > Macro.’

1b.ii. A new window will appear titled ‘Macro.txt.’ Input the text from the script found on GitHub (file: ‘Protocol_stepD-1-b-ii.txt’) and leave this window open. We will refer to this macro as the ‘multiple open’ macro (Note 7).

1c. Next, create a second macro that will select individual leaves from a scan. 99

1c.i. With the ImageJ toolbar as the active window, navigate to the menu bar and select ‘Plugins > New > Macro.’

1c.ii. A second window will appear titled ‘Macro.txt.’ Input the text from the script found on GitHub (file: ‘Protocol_stepD-1-c-ii.txt’). Leave this macro window open. We will refer to this macro as the ‘select’ macro.

1d. Open the first image to be analyzed.

1d.i. With the ImageJ toolbar as the active window, navigate to the menu bar and select ‘File > Open.’

1e. Execute the ‘multiple open’ macro by highlighting all code line and clicking either ‘⌘ + R’ (for a Mac) or ‘CTRL + R’ (for a PC) (also see Note 8).

1f. The leaves within the image should now be thresholded in black and white.

1g. Isolate single leaves to extract and save as separate files by selecting the

‘Wand (tracing)’ tool from the ImageJ menu toolbar (Figure 4-8).

Figure 4-8. The ImageJ menu toolbar. Here, the ‘Wand (tracing)’ tool is selected.

1g.i. Select an individual leaf by clicking on it and confirm its outline is properly highlighted.

1g.ii. Highlight all code lines and execute the ‘select’ macro. 100

1g.iii. A prompt will appear to save a binary image of a single leaf. Name the file appropriately and save in a different folder with other binary .jpeg files generated during this process for later use.

1h. Repeat steps D1a-D1h until all leaves in the image have been isolated.

1i. When each leaf in the open file has been successfully converted to binary images, add the text “run("Open Next");” as the first line of script in the ‘multiple open’ macro.

1j. Highlight all code lines, including the new line, then execute. The addition of the ‘Open Next’ command will now open the next image file in the folder.

1k. A prompt will appear to save the changes made to the current open, binary image. Do not save changes, as this will alter the original image scan.

1l. The next scan will open. Repeat the steps in D1 until all files have been converted to binary images of one leaf per image.

2. For scans of herbarium specimens

2a. Open Photoshop CS4.

2b. If the ‘Tools’ menu bar (Figure 4-9) is not already on the main screen, manually open it by selecting ‘Window > Tools.’

101

Figure 4-9. The ‘Tools’ menu in Photoshop. A. Quick Selection Tool; B. Eyedropper Tool; C. Brush Tool; D. Crop Tool; E. Clone Stamp Tool; F. Default Foreground/Background Color Tool.

2c. From the menu, select ‘File > Open…’ and select the first scan to be processed.

2d. Identify an appropriate leaf (i.e., one that is in good condition, flat, and does not have many [or any] structures intersecting it). 102

2e. If the leaf is not in an upright position, rotate it using ‘Image > Image

Rotation’ until the apex of the leaf is pointing upward and the base or petiole is pointing downward.

2f. Use the ‘Crop Tool’ (Figure 4-9D) to isolate this leaf from most of its surroundings, but leaving a buffer around each side of the leaf.

2g. If any structures (i.e., stems, tendrils, flowers, other leaves, etc.) intersect the leaf of interest, these must be removed for further analyses.

2g.i. Click on the ‘Default Foreground/Background Color Tool’ button (Figure 4-

9F), making sure that ‘white’ is indicated in the foreground box (as seen in Figure 4-9).

2g.ii. Select the ‘Brush Tool’ (Figure 4-9C) and ensure that under the Options

Menu (‘Window > Options’) the ‘Mode’ is set to ‘Normal’ and the ‘Opacity’ is set to

‘100%’.

2g.iii. Using the ‘Brush Tool’, select an appropriate brush diameter and then paint over any structures or tissues that intersect with the leaf of interest (Figure 4-10).

2h. Click on the ‘Quick Selection Tool’ (Figure 4-9A). Left-click on the leaf and drag along the blade until the entire lamina, and only the lamina, is outlined.

2i. Right-click in the selected area and then click ‘Select Inverse’ from the new menu. 103

Figure 4-10. An example of areas edited out using the ‘Brush Tool.’ By using this tool in Photoshop, intersecting structures have been painted over to remove them.

2j. Switch back to the ‘Brush Tool’ and increase the brush diameter. Paint over the background (with white) until only the leaf remains.

2k. Right-click the image and select ‘Select Inverse’ again.

2l. Any holes, tears, or anomalies on the leaf surface need to be filled in so as to match the color of the lamina. This can be done using either the ‘Clone Stamp Tool’

(Figure 4-9E) or the ‘Eyedropper Tool’ (Figure 4-9B).

2l.i. To use the ‘Clone Stamp Tool,’ hold the Alt/Option key and click an intact area on the leaf. Now use the tool to fill in any damaged spots. 104

2.l.ii. To use the ‘Eyedropper Tool,’ left-click on an intact area of the leaf near the damaged area to extract the color. Switch to the ‘Brush Tool’ and paint over the damaged area(s) to match the color of the rest of the leaf.

2m. In some cases, especially when using .tif files, it may be necessary to ‘flatten’ the layers of the image by clicking ‘Layers > Flatten Image’ before saving. This will merge all of the edits into a single, savable image.

2n. Save each leaf (‘File > Save as…’) with a slightly different filename, but still keeping track of the original scan that it came from (e.g., ‘MO1624745_leaf1,’

‘MO1624745_leaf2,’ etc.).

2o. Repeat steps D2a-D2n for multiple leaves per scan, and for multiple scans.

2p. To Convert isolated leaf outlines to black and white images for SHAPE analysis (step D3), follow steps D1a-D1f.

3. The program SHAPE (Iwata & Ukai, 2002) uses binary leaf outline image files in BMP format. Leaf images can be converted in ImageJ by creating a batch macro.

3a. Open ImageJ.

3b. Create a macro to easily convert many images at once.

3b.i. From the ImageJ menu click ‘Process > Batch > Macro...’

3b.ii. Select the appropriate ‘Input’ folder containing the images to be converted.

3b.iii. Select an ‘Output’ folder to contain the new binary images.

3b.iv. Choose ‘BMP’ from the ‘Output Format’ drop-down menu.

3b.v. Input the text from the script found on GitHub (file: ‘Protocol_stepD-3-b- v.txt’) into the large blank space provided. 105

3c. View the output folder to check that the .bmp files were properly converted

(Note 9).

Analysis of Outlines: Elliptical Fourier Descriptors (EFDs)

1. Use the SHAPE software to convert image outlines to chain code. We encourage users become familiar with the SHAPE User Manual in order to better explore parameter choice.

1a. Open the ChainCoder program within SHAPE.

1a.i. Select ‘Files > Select Image File(s).’

1a.ii. In the following window (Figure 4-11), select the folder of BMP files. The

BMP files will then appear in the field ‘File(s).’

Figure 4-11. Selection of .bmp files for chain coding in the software SHAPE. Use the double arrow button to move all files to the ‘Selected File(s)’ field.

106

1a.iii. Select all images and select the double arrow button to transfer the desired files into the ‘Selected File(s)’ field. Press ‘OK.’

1b. Before beginning the analyses, select the ‘Config’ tab (Figure 4-12).

Figure 4-12. The ‘Config’ tab of the ChainCoder program. This allows for configuration of parameters before beginning the chain coding process.

1b.i. Set ‘Object Color’ to ‘Dark (Black),’ and ‘Scale Included’ to ‘No.’ Leave the other fields at their defaults.

1c. Select the ‘Processing’ tab to begin processing photos.

1c.i. Select ‘Load Image’ (Note 10). 107

1c.ii. Deselect the ‘Select Area’ box.

1c.iii. Select ‘Gray Scale.’

1c.iv. Select ‘Make Histogram.’

1c.v. Select ‘Binarize Image.’

1c.vi. Check the ‘Ero Dil Filter’ and ‘Dil Ero Filter’ boxes and select to what degree to ‘erode’ and ‘dilate’ the outline; this option adds and subtracts the amount of pixels from the image to give a smoother border.

1c.vii. Select the ‘Labeling Object’ button. A new window named ‘Chain Code

Data’ will appear, allowing the user to indicate which objects, over a certain number of pixels, should be isolated for analysis.

1c.viii. Select ‘Chain Coding.’ This will add the chain code to the user selection.

1c.ix. Select ‘Save to File.’ For the first image, you will have to name the file.

Subsequently, chain codes for further images will be saved as processed (Note 11).

1d. Repeat steps E1a-E1c for all further images. If many images need to be processed, hold down the ‘enter’ key (or put a weight on it until the image processing is finished).

2. Convert chain code to normalized EFDs.

2a. Open the CHC2NEF program within SHAPE. This program converts the chain code created above to normalized EFDs.

2b. A new window will appear (Figure 4-13). Select the chain code file produced in step E1c.ix (‘CHC File Name’), then create a name for the new NEF file that will be generated in the following steps (‘NEF File Name’). 108

2c. Set the ‘Max Harmonic No.’ Higher harmonic numbers lead to better shape approximations, but usually 20 is sufficient to recapitulate leaf shape accurately.

Figure 4-13. The CHC2NEF window. Select the chain code input file (.chc) and name the resulting normalized elliptical Fourier descriptors file (.nef).

2d. Select the Normalization Method to be ‘based on the longest radius.’ This is the way the image will initially be oriented, and this option allows better subsequent manipulation to align properly.

2e. Click ‘OK.’

2f. A new window will appear. Click ‘Start !!’

2g. Orient the image so that all images are similarly aligned. This is an arbitrary designation, but needs to be consistent among all images.

2g.i. There are a number of buttons to assist in this process (see Figure 4-14). The image can be turned by any number of degrees in either direction, and arrow buttons can 109 be adjusted by rotation degree. Alternatively, images can rotate by 180°. Utilize the three turning buttons at the right side of the screen to orient the leaf image appropriately. NEF code will reflect these changes (bottom of window).

Figure 4-14. Orienting leaf image chain code to normalized elliptical Fourier descriptors. Utilize the three turning buttons at the right side of the screen to orient the leaf image appropriately. NEF code will reflect these changes (bottom of window).

2g.ii. Once aligned, click ‘Save/Next Obj.’ and repeat until all images have been normalized. 110

3. Visualize Elliptical Fourier Descriptors (EFDs) using Principal Components

Analysis (PCA). These analyses can be done within the SHAPE program itself (see below; Note 12).

3a. Open the ‘PrinComp’ program within SHAPE.

3b. A new window will appear. In the menu bar, select ‘Files > Open Nef File’ and select the desired .nef file.

3c. An additional window will appear (Figure 4-15) with parameters to determine for the PCA.

Figure 4-15. The ‘NEF File Information’ window. Desired PCA parameters can be set for the normalized Elliptical Fourier Descriptors. 111

3c.i. ‘Number of Header Lines’ is automatically set according to the NEF file.

3c.ii. Select an appropriate number of harmonics on which to perform the PCA

(the default is 20).

3c.iii. Select the desired coefficients to keep constant.

3c.iii.1. To analyze both symmetric and asymmetric variance: select ‘a-d.’

3c.iii.2. To analyze symmetric variance: select ‘a’ and ‘d.’

3c.iii.3. To analyze asymmetric variance: select ‘b’ and ‘c.’

3d. Perform the PCA by clicking the ‘Principal Components Analysis’ button (see

Figure 4-16).

Figure 4-16. The PrinComp program toolbar. A. ‘Principal Component Analysis’ button performs the PCA; B. ‘Calculate Principal Component Scores’ button generates PC scores for the dataset; C. ‘Reconstruct Principle Component Contours’ button generates Eigenleaves for visualization.

3e. Verify that parameters are correct in a new window, and click ‘OK.’

3f. Once the PCA results have been generated, a new window will appear to save the results file (.pcr file) in the desired folder.

3g. A new window will appear with information from the PCA (Figure 4-17).

There are a number of tabs with useful information about the analysis. 112

Figure 4-17. The ‘Information of Principal Component Analysis’ window. This window will appear after the program has completed the PCA run, providing useful information about the results of the analysis in various tabs.

3g.i. At the bottom of the window, select the ‘Make Report’ button.

3g.ii. A new window titled ‘Report Option Dialog’ (Figure 4-18) will appear to select analysis information that will be written into a report (.txt file). Confirm the

‘Eigenvalues & Eigenvalue Proportions’ box is checked, as this contains the percent variance explained by each principal component (PC). 113

Figure 4-18. Select report information for printing. Boxes checked in this window will include relevant information in a report that will be written as a .txt file.

3g.iii. Click ‘OK.’ A new window will appear with the text file containing a PCA report.

3h. To retrieve the PC values for further analysis, click the ‘Calculate Principal

Components Scores’ button in the PrinComp toolbar (Figure 4-16B) to create a PC score file.

3h.i. A new window will appear to name and select the output file. Click ‘OK.’

3i. To visualize the ‘Eigenleaves’ and what each PC represents, click the

‘Reconstruct Principal Component Contours’ button in the PrinComp toolbar (Figure 4-

16C).

3i.i. A dialogue box will appear to select the number of components to reconstruct, options being ‘Reconstruct Effective Components Only’ or ‘Select

Manually...’ (Figure 4-19). This will be the number of components that will be visualized. 114

Figure 4-19. The ‘Reconstruct Contours’ window. In order to visualize the Eigenleaves, select the desired number of components to be reconstructed.

3j. Save the resulting ‘PC contours’ file. New windows will appear (opens automatically in SHAPE’s PrinPrint program; Figure 4-20), one with a graphic showing

Eigenleaves, another to select draw options.

4. The PrinPrint program can be used to view the ‘PC contours’ file at a later time.

115

Figure 4-20. Eigenleaves visualized in the ‘PrinPrint’ program.

Shape Features: Aspect Ratio and Circularity

1. Open ImageJ.

2. Navigate to the menu bar and select ‘Analyze > Set Measurements...’ 116

3. Check the ‘Area,’ ‘Shape Descriptors,’ and ‘Display label’ boxes.

4. Navigate to the menu bar and select ‘Process > Batch > Macro...’

4a. A new window will appear titled ‘Batch Process.’ Select the appropriate input

(i.e., binary leaf image files) and output folders (Note 13).

4b. In the macro field, input the ImageJ code found on GitHub (file:

‘Protocol_stepF-4-b.txt’).

4c. Select ‘Process.’

4d. A new window will appear titled ‘Results.’ Save this report.

5. The resulting data can be visualized using linear regression (Figure 4-21).

Figure 4-21. Linear regression of Aspect Ratio (AR) and Circularity (Circ.) data. In this example, multiple leaves from several genotypes of Vitis riparia (purple) and V. rupestris (green) have been measured for AR and Circ., with the resulting data visualized in this linear regression. Low AR and Circ. values capture the deeper lobing and significant serration of V. riparia leaves compared to that of V. rupestris leaves. 117

Data Analysis

Data analysis techniques are fundamental to the purpose of this protocol and are integrated within the procedure (i.e., C, E3, F5), as they are often challenging for new users to develop de novo. However, our methods represent a sampling of available methods for digital morphometric analysis. We encourage users to explore the literature and available programs to develop a method that is best suited for their material and the particular scientific inquiry.

Notes

1. In the scripts available from GitHub, any text preceded with ‘#’ should be considered user notes—these will not be read by the computer if the whole script is copied and pasted into the R or macro consoles.

2. Because Generalized Procrustes Analysis (GPA) and outline analysis do not require scaling, it is not necessary to include a ruler in scans for morphometric analyses; however, we recommend this as good practice.

3. If the point tool is in the single ‘Point’ setting, simply right-click on the button to change it to the ‘Multi-point’ setting.

4. Landmarks must be placed in the same order on all leaves.

5. If an error was made during placement, points can be deleted by holding the

CTRL key and clicking on the point. It is also possible to move points by clicking on the imprecise point, then moving it to the desired location.

6. It may be necessary to adjust column headers or other information that might be erroneously askew from the conversion process. 118

7. If there is only one fresh leaf per scan, refer to the directions in step D1e-D1f.

Repeat until all images have been converted to black and white.

8. For the first image file of the dataset, do not use the first line of code (e.g., run(“Open Next”)).

9. BMPs are large files, and it may be convenient to use an external hard drive or cloud storage client to store and use them from this point forward.

10. Viewing the first image in the viewing field may require using the ‘Zoom

Out’ button in the top right corner to visualize properly.

11. This produces one file for all of the images’ chain code. If chain-coding cannot be completed in one sitting, the file can be updated to include the remaining files at a later date.

12. Alternatively, using the R package ‘Momocs’ (Bonhomme, Picq, Gaucherel,

& Claude, 2014) converts NEFs to objects for a variety of graphical visualizations, including PCA.

13. The desired output will not be a folder of files, rather the measurement report that can be saved as a single file. Therefore, the output folder can be the same as the input folder. To minimize file size, select a small file format (we suggest .jpeg) in the ‘Output

Format’ drop down menu.

Acknowledgments

This protocol was developed in part for the publications of Chitwood et al. (2016) and Klein et al. (2017). The authors are grateful to Dr. Dan Chitwood for his comments, guidance, and expertise in using and developing these methods. We would also like to 119 thank Dr. Allison Miller lab undergraduate members for their comments and suggestions on the protocol, as well as three reviewers who helped improve this manuscript. A Saint

Louis University Graduate Research Assistantship to LLK and an Ohio University

Original Work Grant to HTS supported this work.

120

Systematic Explorations in Section Dysosmia Using Leaf Geometric Morphometric

Shape Descriptors

As discussed above, geometric morphometrics provides an objective and quantitative method for recognizing differences in morphology. For plants specifically, leaf shape continues to play an important role in taxonomic delimitation; especially when genetic or other data are not available. The taxa in section Dysosmia are notoriously difficult to identify because of their apparently highly variable leaf shape; likewise, trying to sort herbarium specimens of Dysosmia collections can often become a futile task for this very reason. Leaf shape has historically been a major factor in describing new species and inferring taxonomic boundaries in the section. Some species have unique morphologies and are not easily confused with other taxa, but others share an overall leaf morphology that seems to represent a grade of shape and lobing. Certainly, one expects that leaves will tend to look different between two taxa, but “how different” are these leaves? How can researchers move beyond subjective descriptors of shape and objectively test for differences that might represent novel taxa? Geometric morphometrics may provide the answer to these questions.

Methodology

The methodologies presented here are based largely on the protocol of Klein and

Svoboda (2017, above), with minor changes to better visualize the results (e.g., using the

R package Momocs).

121

Taxonomic Sampling and Digitization

Each of the 21 accepted species in section Dysosmia, including many of the varieties currently in synonymy, were used in this study. Three additional groups or complexes (the red-fruited pubescent [RFP], the orange-fruited pubescent [OFP], and the yellow-fruited pubescent [YFP]) were designated for this study to determine if fruit color might be a taxonomically informative character. A total of 514 leaves from 351 unique herbarium sheets, including 48 type specimens, were consulted in order to give the greatest range in morphological diversity. Multiple leaves were often used from a single sheet so as to provide more leaves for the analysis; likewise, collections with multiple sheets (usually from different herbaria) were also consulted for the same reason.

Physical specimens were either sent on loan and scanned at Ohio University using an Epson Expression 10000XL flatbed scanner (Seiko Epson, Suwa, Japan) or received as a digital loan of scans, from the following herbaria (herbarium codes follow the Index

Herbariorum; Thiers, 2018): A, ASU, BHCB, BOLO, BM, C, CAS, E, ECON, F, FTG,

G, GH, GOET, HUEFS, K, LINN, MEXU, MICH, MO, NY, P, RB, S, SD, TEX/LL,

UC, UCR, UPRRP, US, and W.

Image Processing and Morphometric Analysis

The digital images were processed using ImageJ (Schneider, Rasband, & Eliceiri,

2012) using the protocol developed by Klein and Svoboda (2017) in order to remove overlapping structures and produce a “clean” leaf image for downstream analyses. Due to the incredible diversity in leaf lobing and shape, only six homologous landmarks were identified and placed on each leaf (as seen in Figure 5-3; Svoboda & Harris, 2018), 122 mapping the primary venation of the lamina. Images of individual leaves were converted into black and white binary files (see Klein & Svoboda, 2017) for outline analysis.

Following Klein and Svoboda (2017), the landmark dataset was transformed and subjected to generalized Procrustes analysis (GPA) using a suite of packages developed in the software R (R Core Team, 2017; RStudio Team, 2018). While Klein and Svoboda

(2017) used the software SHAPE (Iwata & Ukai, 2002) for elliptical Fourier descriptor

(EFD) analysis, I have chosen to use the R package Momocs (Bonhomme et al., 2014) in this study because of its ability to easily handle both the landmark and outline datasets without additional conversion or reformatting. Principal Components Analysis (PCA) and

Linear Discriminant Analysis (LDA) were conducted in order to visualize the results of the analyses. The ability of LDA to classify the accessions to the correct taxon was tested using a leave-one-out cross-validation approach.

The objectives of this study were to 1) test whether landmark and/or outline descriptors are appropriate for describing varying leaf shapes, 2) determine if one descriptor is more successful than the other, and 3) assess whether these methods can be used for taxonomic discovery in section Dysosmia.

Results and Discussion

Ordinations resulting from PCA showed near-identical patterns to those of the

LDA, and so only the former are shown here. When all of the taxa were analyzed simultaneously, the amount of overlap in overall leaf shape was apparent in both the landmark (Figure 4-22) and outline analyses (Figure 4-23). The large cloud of points in each indicates that many disparate taxa, regardless of assumed affinity or fruit color, 123 share a great deal of morphological diversity in regard to their leaf shape. As expected, the three main species complexes (Passiflora ciliata, P. foetida, and the RFP) occupy nearly the entire morphospace; indicating that each represents an enormous amount of shape variation as currently grouped.

Figure 4-22. Ordination representing a generalized Procrustes analysis (GPA) of landmark data for all of the taxa in section Dysosmia.

124

Figure 4-23. Ordination of outline (elliptical Fourier descriptors; EFD) data for section Dysosmia.

Analyses were also performed on individual species (or complexes) to better understand the variation within a “known” taxonomic context. Some of the variation within well-defined species is shown in Figure 4-24. When considering a complex, for example the P. foetida complex, the leaves are more variable than one might expect for a given species (Figure 4-25). Ordinations of a single species, Passiflora bahamensis, reveal that the landmark and outline analyses (Figure 4-26) are showing largely the same pattern across their respective morphospaces. 125

Figure 4-24. Landmark overlays (A–C) and outline overlays (D–F) of leaves. A, D. Passiflora bahamensis. B, E. P. sublanceolata. C, F. P. urbaniana.

The classification ability of LDA varied by species, but generally improved using landmark data compared to the outline data; a finding consistent with Chitwood and

Otoni (2017b). For example, P. bahamensis was correctly classified 86.6% of the time based on landmarks, compared to 85.0% with outlines whereas P. urbaniana was correctly classified with 38.1% accuracy based on landmarks but only 29.5% of the time using outlines. The results of the landmark and outline analyses are widely consistent with one another, indicating that each type of descriptor—even though they rely on different kinds of data—is fairly accurately describing the shapes of the leaves. 126

Figure 4-25. Leaf overlays of the Passiflora foetida complex. A. Outline overlays. B. Landmark overlays.

Viewing the distribution of accessions in morphospace has allowed me to identify patterns that I otherwise probably would not have noticed. Clusters of points, representing accessions with very similar leaf shapes, can be observed in the PCA ordinations of landmarks of the “P. ciliata complex” (Figure 4-27) and might be indicative of taxonomic identity. Those accessions circled in blue in Figure 4-27 all represent a taxon called Passiflora foetida var. subintegra Killip that I am considering a distinct species based on these data and micromorphological novelties (see Chapter 7).

Scrutiny of the ordinations of this and other complexes has likewise revealed more clusters of unique morphologies that may correspond to novel taxa. 127

Figure 4-26. Ordinations representing the leaf shape variation of Passiflora bahamensis. A. Landmark data. B. Outline data.

While the results of both the landmark and outline descriptors show promise for taxonomic delimitation in section Dysosmia, the overall highly plastic nature of the 128 leaves in the section complicate the utility of these methods. Geometric morphometric descriptors alone are not adequate for taxonomic identification but can easily be integrated with other datasets (e.g., micromorphology, ecology, genetics, etc.) for robust systematic investigations. Refining these methods and realizing the full potential of geometric morphometrics in Passiflora taxonomy will be the focus of future studies.

Figure 4-27. Ordination of the Passiflora ciliata complex based on landmark data. Blue circle represents a taxon called P. foetida var. subintegra Killip.

129

Acknowledgements

I would like to thank the curators and staff of the herbaria mentioned in the

Methodology section for loans of specimens or sending digital scans of their collections. I am also indebted to Vincent Bonhomme (Université de Montpellier), Dan Chitwood

(Michigan State University), and Laura Klein (St. Louis University) who have helped me throughout this project with discussions of morphometric theory, carrying out analyses, and interpretation of the results. 130

CHAPTER 5: CONTRIBUTIONS TOWARD UNDERSTANDING THE

BIODIVERSITY OF PASSIFLORA IN NORTH AMERICA: UPDATES AND A NEW

COMBINATION FROM THE BAJA CALIFORNIA PENINSULA, MEXICO AND

VICINITY

This article has already been published and is reprinted by permission

(Contributions toward understanding the biodiversity of Passiflora in North America:

Updates and a new combination from the Baja California Peninsula, Mexico and vicinity,

H. T. Svoboda and AJ Harris, Journal of Systematics and Evolution [online], Copyright

Appendix A for more information). All text is retained from the publication, with minor adjustments to some citations to comply with the formatting of this document.

Here, we use environmental measurements and intensive morphological studies to describe the Passiflora taxa found on the Baja California Peninsula, Mexico, and raise a taxon from the rank of variety to species based on these data.

Harlan T. Svoboda1 and AJ Harris2,†

1Department of Environmental and Plant Biology, Ohio University, 315 Porter Hall,

Athens, Ohio 45701, USA

2Department of Botany, National Museum of Natural History, Smithsonian Institution,

MRC 166, Washington, D.C. 20013, USA

†Present address: Department of Biology, Oberlin College, Science Center K123, 119

Woodland Street, Oberlin, Ohio 44074, USA

Received 1 December 2017; Accepted 17 May 2018; Published 24 June 2018 (online). 131

Abstract

The Baja California Peninsula and surrounding landmasses harbor an abundant flora in an otherwise harsh and arid environment. Of the many plant groups native to this peninsular and insular region, passionflowers (Passiflora, Passifloraceae) are represented by several conspicuous taxa that all belong to a single lineage, section Dysosmia. Basic questions remain regarding this group, particularly the taxonomic status among the

Passiflora arida complex. Therefore, we investigated the claims of endemism, habitat characteristics, and taxon boundaries with section Dysosmia in the Baja California region using extensive sampling of herbarium specimens and iNaturalist observations. We confirmed that only one of the native Passiflora taxa (P. fruticosa) was endemic to the

Baja California Peninsula, with an additional taxon (P. palmeri) considered near- endemic. Environmental data revealed significant distinctions between the habitats of many of the native taxa as well as within the P. arida complex, especially with respect to precipitation and temperature tolerances. Geometric morphometric analyses of leaf shape were largely not successful at separating taxa, indicating leaf shape may not be a good indicator of taxon identity in this particular group. Based on ecological differences and discrete macro- and micromorphological features, a varietal name is here synonymized and a new combination is proposed: Passiflora pentaschista.

Introduction

The flora of North America comprises a rich vegetation over a vast, heterogeneous physiographic region. The flora exhibits a classical latitudinal diversity gradient such that species richness is greater at lower latitudes (Thorne, 1993), though 132 finer-scale patterns also exist (Gentry & Dodson, 1987; Harris, Walker, Dee, & Palmer,

2016). Mexico harbors more than 20 000 species within 1.9 million km2 of land area compared to approximately 16 000 species over 21.5 million km2 for the rest of North

America (Qian, 1999; Villaseñor, 2016). Disturbances over at least the last 100 million years, including aridification (Graham, 1999), have led to a fairly recent assembly of floristic compositions, especially the warm desert floras of northwestern Mexico and the southwestern United States.

In modern times, these desert floras cover roughly 3% of the land area of North

America, and are particularly expansive in western Mexico (MacMahon, 2000). Deserts tend to exhibit lower diversity than in adjacent, wetter areas, probably due to both their recent origin and the biotic constraints on potential colonizers (Preston, 1960; Thorne,

1993; Smith, Monson, & Anderson, 2012). Nevertheless, the deserts are striking for their high levels of endemism and for several clear patterns of geographic distribution (Clark-

Tapia & Molina-Freaner, 2004; Garrick, Nason, Meadows, & Dyer, 2009) that probably arose relatively recently. Endemism in the Baja California Peninsula of Mexico is particularly well documented in a multitude of plant groups (see Riemann & Ezcurra,

2005; Rebman, Gibson, & Rich, 2016), with 897 taxa (nearly 26% of the flora) being endemic or near-endemic (Rebman et al., 2016).

The Baja California Peninsula of Mexico (Figure 5-1) and its associated islands are home to at least five of the 30 passionflower taxa currently recognized (fide

Vanderplank, 2013) in section Dysosmia DC. (Passiflora L., Passifloraceae Juss. ex

Roussel). Broadly, passionflowers are distributed throughout the Neotropics with more 133 than 570 species (Ocampo Pérez & Coppens d’Eeckenbrugge, 2017), of which 17 occur in the United States (Goldman & MacDougal, 2015) and more than 80 species, including undescribed ones, occur in Mexico (J. M. MacDougal, unpublished data). Passionflowers comprise an array of beautiful and striking features, such as the filamentous corona and elevated androgynophore, which make this genus easily recognizable. Section Dysosmia was delimited within supersection Stipulata Feuillet & J.M.MacDougal according to the subgeneric classification of Feuillet and MacDougal (2003) and is one of the most widespread and unusual taxonomic groups. The group possesses morphological novelties that include ciliate involucral bracts and ubiquitous glands or glandular trichomes on the foliar plant structures.

The five currently recognized taxa of sect. Dysosmia (fide Vanderplank, 2013) native to the Baja California Peninsula are Passiflora foetida L. (sensu lato), P. arida

(Mast. & Rose) Killip, P. arida var. pentaschista Killip (Figure 5-2A), P. fruticosa Killip

(Figure 5-2B), and P. palmeri Rose (Figure 5-2C), of which the latter three have historically been considered endemic to the Peninsula. In addition, Wiggins (1980) recognized a sixth taxon, Passiflora arida var. cerralbensis Killip, and indicated that it was also endemic to the peninsula. More recent studies (e.g., Goldman, 2003;

Vanderplank, 2013), however, have reported an expanded range of variety cerralbensis into mainland Mexico where typical P. arida is also known to occur (the type of the species being collected in Guaymas). Endemism and distribution is difficult to assess in this group, however, because “P. arida” as currently circumscribed comprises a complex of phenotypes with questionable morphological distinctions among them. 134

Another taxonomically difficult Dysosmia taxon within the study area is the P. foetida complex. This “species” has been regarded as problematic since the time of Linnaeus, leading Killip (1938) to recognize 38 varieties within it and many more names in synonymy. Passiflora foetida from the Peninsula is represented by specimens that exhibit an extreme gradation of morphology that is unsurpassed in the rest of sect. Dysosmia, with many of the phenotypes being found throughout and across the extensive range of the species. Ongoing research by the first author (HTS) is attempting to resolve relationships and taxonomic boundaries in the complex (see Svoboda & Ballard, 2018), but it remains a problem that will require more in-depth investigation beyond the scientific and geographic scope of this study.

In prior work to delimit and better understand the Passiflora species of the Baja

California Peninsula, quantitative habitat characteristics have not been used, and reports on the ecological affinities of the species have been limited to qualitative data ascertained from herbarium labels and field observations.

In this study, our objectives are to (i) evaluate the habitat characteristics of the taxa of section Dysosmia native to the Baja California Peninsula, (ii) delimit the taxonomic boundaries of the Passiflora arida complex, and (iii) reassess the peninsular endemic status for each native taxon.

Material and Methods

Taxon Sampling and Georeferencing

We examined a total of 204 herbarium specimens and 18 iNaturalist observations

(iNaturalist Network, 2018) representing the majority of the known collections for each 135 taxon native to the Baja California Peninsula (Appendix B, as Doc. S1 of Svoboda &

Harris, 2018). Peninsular members of the Passiflora foetida complex were considered “P. foetida (sensu lato)” for this study without designation of variety, as the variation observed is too great to properly assess for the scope of this paper. Three additional species have also been reported from the Peninsula: Passiflora arizonica (Killip)

D.H.Goldman, of sect. Dysosmia, and P. bryonioides Kunth, and P. subpeltata Ortega, of other sections. However, the identification of these occurrences is either dubious, cannot be confirmed, or well beyond their known range and ecological habitat, and thus are likely not native to the peninsula. While peninsular specimens of the latter two taxa probably represent simple misidentifications, the several collections previously identified as P. arizonica from the peninsula are inconsistent with the circumscription of that species and, therefore, we do not confirm its presence on the Baja California Peninsula at this time. The specimens appear to be unusual phenotypes of P. arida or possible hybrids with P. palmeri and warrant additional investigation beyond the scope of this study.

Consequently, these three taxa are not included in the current study.

Using herbarium and observational records of the native species, we gathered geolocation data either verbatim or estimated according to GEOLocate version 3.22 (Rios

& Bart, 2010) with curation in SimpleMappr (Shorthouse, 2010) to ensure geographic accuracy.

Environmental Data

We obtained climatic and soil variables for each of the six taxa of Passiflora native to the Baja California Peninsula and the Mexican state of Sonora based on the 136 georeferenced occurrences records (discussed above) and using Global Information

Systems (GIS) layers. We extracted 11 soil and 20 climatic variables (Table 5-1) from

GIS layers derived from the Harmonized World Soil Index

(FAO/IIASA/ISRIC/ISSCAS/JRC, 2012) and WorldClim version 2.0 (Fick & Hijmans,

2017), respectively, for 222 spatially unique specimen records from herbarium specimens and iNaturalist observations. The extractions were performed in ArcMap version 10.4.1

(Environmental Systems Research Institute [ESRI], 2016) at 30 arc second resolution.

Subsoil and/or topsoil data was not available for 36 occurrences, which therefore had

“missing data” for those variables. For 15 occurrences, particularly those on small islands in the Gulf of California, no climatic data was available and the extraction returned a “No

Data” result. Thus, we omitted these occurrences from subsequent analyses, giving a total of 207 usable accessions (see Data S1 of Svoboda & Harris, 2018).

The extracted, curated environmental data from ArcMap were transformed with z- scores (Fisher, 1921) using the ‘scale’ function in RStudio version 1.1.423 (RStudio

Team, 2018; R Core Team, 2017) to ensure homogenization of the covariance matrices.

We then carried out linear discriminant analysis (LDA) and multivariate analysis of variance (MANOVA) using the program PAST version 3.16 (Hammer et al., 2001). In order to understand the relative correlations between habitat characteristics and taxonomy, we tested three taxonomic hypotheses, with (i) the Passiflora arida complex as three distinct entities (fide Killip, 1938), (ii) the P. arida complex as only two entities, with vars. arida and cerralbensis combined (proposed by Vanderplank, 2013), and (iii) the P. arida complex as only two entities, with vars. cerralbensis and pentaschista 137 combined. The ability of LDA to predict the correct taxon based on environmental data was tested using a leave-one-out cross-validation approach in PAST.

Geometric Morphometrics

We placed six homologous landmarks (Figure 5-3) on 265 leaves from scanned herbarium specimens (Appendix B, as Doc. S1 of Svoboda & Harris, 2018) in ImageJ

(Abràmoff, Magalhães, & Ram, 2004; Schneider et al., 2012) following procedures outlined in Klein and Svoboda (2017). The landmarks, which map the primary venation, serve as a proxy for the general shape of the leaf and tend to outperform elliptical Fourier descriptors (outlines) when used for taxon identification (Chitwood & Otoni, 2017a,

2017b).

We performed Procrustes superimposition of the x,y landmark coordinates using the R package Momocs (Bonhomme et al., 2014) for shape analysis. This method standardizes each leaf so that differences in size and rotation do not affect downstream analyses. Principal component analysis (PCA) and LDA were used to assess the utility of leaf shape for taxonomic discrimination (see Data S1 of Svoboda & Harris, 2018). To test the ability of LDA to predict the correct taxon based on leaf shape, we implemented a leave-one-out cross-validation approach in R.

Results

Environmental Analyses

The results of the various linear discriminant analyses (LDA) based on environmental data are reported in Table 5-2 for each of the iterations performed. Note that because the highly variable Passiflora foetida complex was found to occupy nearly 138 the entire ordination space, these accessions were removed from the analyses so as to allow better resolution among the rest of the taxa. In general, the separation between groups increased when the “Passiflora arida complex” was treated as two entities with vars. cerralbensis and pentaschista combined into one group and P. arida (sensu stricto) treated separately. Statistically significant (p = 0.0084) differences were observed between groups when all the taxa were analyzed together (Figure 5-4) and when just the

P. arida complex was analyzed separately (Figure 5-5A; p = 1.711-9). We did not detect significant differences between groups when P. arida vars. arida and cerralbensis were combined into one group and variety pentaschista was considered another (Figure 5-5B; p = 0.2671). Passiflora fruticosa is represented by a distinct cluster of accessions that are completely separate from all of the other taxa (Figure 5-4).

Likewise, the ability of LDA to correctly classify accessions based on taxonomic identity increased with the same grouping scheme; notably a 100% correct classification rate within the P. arida complex when vars. cerralbensis and pentaschista were treated as one group (Figure 5-5A), compared to 89.2% when vars. arida and cerralbensis were grouped together (Figure 5-5B).

When all of the remaining taxa were analyzed together, and P. arida vars. cerralbensis and pentaschista combined into one group (Figure 5-4), the largest contributors to the separation along Axis 1 of the LDA were several temperature variables, especially minimum temperature (BIO6), temperature annual range (BIO7), and temperature seasonality (BIO4). Contributing greatly to the separation along Axis 2 139 of the LDA were precipitation variables, particularly precipitation of the wettest month and quarter (BIO13 and BIO16, respectively), and mean annual precipitation (BIO12).

Leaf Shape Analysis

Procrustes analysis of the six foliar landmarks (Figure 5-3) revealed little to no separation between the groups of taxa. When all of the taxa were analyzed together, the ordinations resulting from PCA and LDA (see Figure S1 of Svoboda & Harris, 2018) showed a significant amount of overlap between each group, with the LDA correctly classifying the accessions to the correct taxon only 52.4% of the time. Likewise, when just the three members of the “Passiflora arida complex” were analyzed, PCA and LDA ordinations (see Figure S2 of Svoboda & Harris, 2018) exhibited similar levels of overlap between groups. The ability of LDA to correctly classify the P. arida taxa was slightly improved, with a 63.7% success rate.

Peninsular Endemism

Wiggins (1980) reported that Passiflora arida var. cerralbensis and var. pentaschista, P. fruticosa, and P. palmeri were all endemic to the Peninsula. Based our expanded sampling from multiple herbaria and confirmed iNaturalist observations, only

P. fruticosa was confirmed to be endemic to the Peninsula (Figure 5-6A); specifically, to the state of Baja California Sur. Several collections of P. palmeri were confirmed from the islands of Isla Tiburón, Isla Turners [Isla Dátil], and Isla Turón [Isla San Esteban], which are politically part of the Mexican state of Sonora. Thus, P. palmeri is considered a near-endemic of the Peninsula (Figure 5-6B). 140

The three varieties of the P. arida complex, when considered as a single species, are found not only on the Baja California Peninsula but also in the neighboring Mexican states of Sonora and Sinaloa. Several collections of var. pentaschista also exist from the states of Arizona and California (United States), but these are likely recent introductions that date back to at least the late 1980s (see MacDougal, 2001; Goldman, 2003) and do not constitute native, but perhaps naturalized, populations.

Discussion

Reevaluation of the Passiflora arida Complex

The distinctions among the three varieties in the Passiflora arida complex have been dubious since their circumscription. In fact, the author of P. arida var. pentaschista proclaimed that although the leaves “differ greatly from those of typical P. arida... no important differences are discernible,” (Killip, 1938, p. 471). Vanderplank (2013) retained P. arida var. pentaschista, and synonymized var. cerralbensis with typical P. arida but did not explain his reasoning for this.

Goldman (2003, p. 247) states that morphologically intermediate specimens proliferate and that “one variety seems to flow into another,” when they are considered separate entities. But upon closer examination of confirmed specimens of varieties cerralbensis and pentaschista, it became clear that they actually share a great deal of morphological affinities despite the apparent disparity between the leaf shapes of the type specimens. In fact, individual plants grown in cultivation by the first author (H. T.

Svoboda, personal observation) bear leaves that are the typical shape of both “varieties” on the same individual. The confusion in identifying specimens correctly is eliminated, or 141 at least largely reduced, when vars. cerralbensis and pentaschista are considered one taxon. It is interesting to note that leaf shape descriptors did not accurately classify the taxa, thus indicating that identifications based solely on leaf shape should be suspect.

Considering varieties cerralbensis and pentaschista as one taxon is further supported by the results of the environmental analyses. We observed statistically significant differences between groups when these two taxa were combined (Figures 5-4,

5-5A), whereas we did not detect significant differences when we treated “P. arida” as three entities. We have, consequently, chosen to synonymize P. arida var. cerralbensis with P. arida var. pentaschista based on these lines of evidence. Because both names were published at the same time and, therefore, have equal priority according to the Code of Nomenclature (McNeill et al., 2012), the correct name for the combined taxon is not readily apparent. We have chosen to retain the name P. arida var. pentaschista over P. arida var. cerralbensis on the basis that Vanderplank (2013, p. 363) previously reduced the latter to synonymy (see Art. 11, Ex. 23 of McNeill et al., 2012).

In examining hundreds of specimens of the taxa within the complex, we further detected a suite of morphological characters (Table 3) that were considerably and consistently different between P. arida (sensu stricto) and the newly proposed combined taxon. We found leaf vestiture, in particular, to be a major and consistent difference.

Although we did find the color pattern of the corona filaments to be reliably different between the taxa, we did not find any significant difference in flower size as suggested by

Vanderplank (2013, p. 365). Significant differences were also consistently observed between these two taxa based on environmental and habitat data (see Figures 5-4, 5-5A). 142

For these reasons, we propose a new combination for the recognition of this distinct taxon at the rank of species: Passiflora pentaschista (Killip) H.T.Svoboda, comb. nov.

Habitat Characterization and Conservation

Much work has been done in the past to describe the various vegetation regions and plant communities of the Baja California Peninsula (see Wiggins, 1980; Garcillán,

González-Abraham, & Ezcurra, 2012). While none of the taxa in this study show affinities strictly consistent with the vegetation regions or communities, our data do reveal broad scale differences in multiple temperature and precipitation tolerances that distinguish the taxa from each other (Figure 5-4). For example, Passiflora pentaschista appears to be affected by increased precipitation, driving its separation from the rest of the taxa, while P. arida (sensu stricto) and P. palmeri tend to separate based on the influence of temperature seasonality and temperature annual range. Soil texture also plays a role in the distinction between taxa, such that P. fruticosa appears to have a strong preference for soils comprised of a large percentage of sand (upwards of 67%) and gravel

(17%), whereas Passiflora palmeri shows preference for soils with high clay content (up to 40%), contributing to its separation from most of the other taxa (Figure 5-4). These soil textures are related to drainage, with larger textures (e.g., gravel and sand) having faster rates of drainage compared to finer textures (e.g., silt and clay) (Brouwer, Goffeau,

& Heibloem, 1985, Chapter 2).

Notably, many of these Passiflora species occur along the coasts of the peninsular and insular Baja California region, and their habitats will likely be at risk from coastal upwells (Bakun, 1990), changes in precipitation (Trenberth, 2011; Cavazos & Arriaga- 143

Ramírez, 2012), and increased temperatures (Intergovernmental Panel on Climate

Change [IPCC], 2014) due to the effects of global climate change. Therefore, a conservation assessment for the IUCN Red List is currently underway by the first author

(HTS) to determine the level of threat and potential conservation priorities for these species.

Passiflora pentaschista is known to display an inclination for invasiveness or weediness (Goldman, 2003, as P. arida var. pentaschista). The species exhibits drought tolerance, rapid growth rate, self-fertility, large seed set, and extended seed dormancy, all of which suggest the “potential to be a weed if introduced outside its native habitat,”

(Ulmer & MacDougal, 2004, as P. arida). Monitoring of this species, particularly in

Arizona, should be carried out to track its progression, ability to invade new habitats, and effects on new areas that it colonizes.

Ecology, Distribution, and Taxonomic Treatment

The data below come from herbarium specimens and iNaturalist observations that represent the vast majority of the known collections or verified observations of the species under study. Values reported in the habitat descriptions, and those in parentheses, are means derived from the GIS layers described in Materials and Methods and Table 5-

1, unless otherwise stated. Minimum and maximum temperatures reported below represent the average highs for hottest month and average lows for coldest month.

Passiflora arida (Mast. & Rose) Killip, J. Wash. Acad. Sci. 12: 256. 1922. Passiflora foetida L. var. arida Mast. & Rose, Contr. U.S. Natl. Herb. 5: 182. 1899. Type: Mexico, 144

Sonora, Guaymas, at Guaymas, 5–11 June 1897, J. N. Rose 1206 (lectotype, designated by Killip (1938, p. 470), US!; isolectotypes, NY!, GH!).

Description: Habit a low, vining shrub; stems densely lanate with soft, white hairs; stipules deeply divided, usually inconspicuous, densely pubescent, lacking resin glands (or very sparingly present) (Figure 5-7A); petioles 6.5–16.5 mm long, lanate, lacking resin glands and petiolar nectaries; leaves densely lanate-tomentose with golden- white hairs, 3-lobed (or rarely sub 5-lobed) with rounded, subequal lobes, lacking resin glands (or very sparingly present), margins entire, base cordate to subtruncate (Figure 5-

7C); peduncles 18–41 mm long, lanate; involucral bracts bi- to tripinnatisect, densely white pilose-lanate, lacking resin glands (or very sparingly present) (Figure 5-7B); flowers 2.5–3.5 cm wide; sepals 1.5 cm long, obtuse, hirsute without only, greenish- white, awned; petals white, 1 cm long, oblong, glabrous within and without; corona filaments in several series, outermost blue (basal fifth), apical third blue-violet, with white between; androgynophore glabrous, with purple spots; ovary subglobose, sericeous; fruit globose, 3 cm in diameter, variously villose, green at maturity; seeds 3- toothed apically, reticulate, base cordate, 5 mm long, 3 mm wide.

Distribution: Found in Mexico, state of Baja California, municipalities of

Ensenada and Mexicali (including Isla San Lorenzo); state of Baja California Sur, municipalities of La Paz (including Isla San Francisquito), Loreto (including Isla

Monserrate), Mulegé (including Isla Tortuga), and Santa Rosalia; state of Sonora, municipalities of Empalme, Guaymas, and Hermosillo (including Isla Tiburón and Isla

Turón [Isla San Esteban]); see Figure 5-6D. 145

Habitat: Sub-/topsoil 13.3/13.3% clay, 40.1/28.5% silt, 46.6/58.2% sand,

9.5/16.3% gravel; sub-/topsoil pH 7.5/7.0; minimum/maximum temperature 5.5/38.1

(22.8 mean)˚C; annual precipitation 14.6 mm; elevation 0–891 (118 mean) m.

Phenology: Budding February–April, flowering April–June, fruiting April–

December.

Passiflora foetida L., Sp. Pl. 2: 959. 1753. [Only homotypic synonyms are listed here.]

Granadilla foetida (L.) Gaertn., Fruct. Sem. Pl. 1: 289. 1788. Tripsilina foetida (L.) Raf.,

Fl. Tellur. 4: 103. 1836. Dysosmia foetida (L.) M.Roem., Fam. Nat. Syn. Monogr. 2: 149.

1846. Type: Locality and collection information unknown (lectotype, designated by

Killip (1938, p. 483), LINN!).

Description [for P. foetida of the Baja California Peninsula]: Habit a climbing, vining shrub; stems sparingly to densely hirsute; stipules hirsute, deeply divided, gland- tipped; petioles hirsute, bearing resin glands, but lacking petiolar nectaries; leaves 3- to sub-5-lobed, lobes rounded or acute, sparingly or densely pilose-hirsute with brown hairs, bearing numerous resin glands, margins entire to coarsely dentate, base cordate to truncate; peduncles densely hirsute, length highly variable; involucral bracts bi- to tripinnatisect, densely pilose-hirsute, gland-tipped; flowers 2–4 cm wide; sepals whitish- green, oblong, hirsute without only, awned; petals white or pale pink, oblong, glabrous within and without; corona with several series, the outermost usually pink-violet and white; androgynophore glabrous, reddish to purple spots; ovary subglobose, densely hirsute; fruit globose, pilose-hirsute, green at maturity; seeds oblong, variously 3-toothed, reticulate, base obtuse or truncate. 146

Distribution [for P. foetida of the Baja California Peninsula]: State of Baja

California, municipality of Ensenada; state of Baja California Sur, municipalities of

Comondú, La Paz, Los Cabos, and Mulegé; see Figure 5-6C.

Habitat: Sub-/topsoil 10.6/11.3% clay, 27.0/25.1% silt, 62.4/63.6% sand,

15.6/15.9% gravel; sub-/topsoil pH 6.8/6.5; minimum/maximum temperature 3.4/37.0

(20.1)˚C; annual precipitation 27.1 mm; elevation 20–1740 (704) m.

Phenology: Budding October–May, flowering October–May, fruiting October–

June.

Notes: The varietal names most frequently given to the peninsular collections of this species complex include Passiflora foetida var. gossypiifolia (Ham.) Mast. and P. foetida var. longipedunculata Killip. Few, if any, specimens bear any resemblance to the lectotype of the former (see Svoboda et al., 2016) and likely just represent misapplications of the name. The latter, although cited from the Peninsula by Killip

(1938, p. 487), apparently has a vastly disjunct distribution with the nearest mainland specimens (and type) being in the state of Tamaulipas, Mexico. It is thus far dubious whether any peninsular collections are actually attributable to this poorly understood taxon.

Passiflora fruticosa Killip, J. Wash. Acad. Sci. 12: 256. 1922. Type: Mexico, Baja

California Sur, [Comondú], Santa Margarita Island, 19 March 1911, J. N. Rose 16285

(holotype, US638347!; isotypes, US1480377!, NY!).

Description: Habit a low, spreading shrub, 20–40 cm high; stems densely lanate with short, soft, wooly hairs, shoots appearing short and compacted; stipules lanulate, 147 highly divided, tips covered with resin glands; petioles 5–10 mm long, densely lanate with golden or white hairs, bearing numerous resin glands, but lacking petiolar nectaries; leaves densely lanate, many stalked glandular trichomes, 3-lobed with rounded, subequal lobes, margins entire with ample resin glands, base cordate; peduncles 1–3 cm long, lanate (Figure 5-2B); involucral bracts hirsute, bi- to tripinnatisect, segments copiously gland-tipped; flowers 2.5–3 cm wide; sepals white, pubescent without only, 10 mm long, awned; petals white, glabrous within and without, 5–7 mm long; corona filaments in several series; androgynophore glabrous, reddish-purple spots; ovary pubescent; fruit subglobose, pubescent, 2.5–3.5 cm in diameter, green at maturity; seeds oblong, reticulate, 3-toothed apically, nearly truncate at base, 5 mm long, 2.5 mm wide.

Distribution: Endemic to peninsular and insular Baja California Sur, Mexico, municipalities of Comondú (including Isla Magdalena, Isla San Diego, Isla Santa Cruz, and Isla Santa Margarita), La Paz (including Isla Espíritu Santo, Isla Partida, Isla San

Francisquito, and Isla San José), Loreto (including Isla Santa Catalan), Los Cabos, and

Mulegé; see Figure 5-6A.

Habitat: Sub-/topsoil 9.8/11.2% clay, 23.0/23/3% silt, 67.2/65.5% sand,

16.3/17.2% gravel; sub-/topsoil pH 6.9/6.5; minimum/maximum precipitation 8.7/36.5

(22.5)˚C; annual precipitation 15.3 mm; elevation 0–415 (80) m.

Phenology: Budding September–June, flowering October–June, fruiting year- round.

Notes: The label on the holotype specimen (US638347) states it was collected at

“Santa Maria Bay” (Los Cabos municipality, Baja California Sur), whereas the isotypes 148 cite the locality as “Santa Margarita Island” (Comondú, BCS). Killip (1938, p. 466) apparently received notes from the collector to correctly determine the type collection was from Isla Santa Margarita, as the isotypes state.

Passiflora palmeri Rose, Contr. U.S. Natl. Herb. 1: 131. 1892. Type: Mexico, [Baja

California Sur, Loreto], Isla Carmen, 1–7 November 1890, E. Palmer 868 (lectotype, designated by Svoboda et al. (2016, p. 110), US46866!; isolectotypes, A!, F118680!,

F265984!, G!, GH!, K!, NY!, S-G!, UC!, US46867!, US46868!, US1360073!).

Description: Habit a low, trailing shrub, 40–50 cm high; stems densely pilose with straight, yellow-white hairs; stipules pilose, gland-tipped; petioles pilose, 1–1.5 cm long, bearing numerous resin glands, but lacking petiolar nectaries; leaves deeply 3-lobed with rounded or weakly acute lobes, densely pilose-tomentose, bearing many resin glands, margins serrate with copious resin glands, base subcordate; peduncles 4–5 cm long, pilose; involucral bracts 20–24 mm long, 8–12 mm wide, pilose, deeply tripinnatisect, segments gland-tipped; flowers 5–7 cm wide, perianth reflexed (Figure 5-2C); sepals pure white, linear, 3.5 cm long, pilose without only, slender-awned; petals pure white, 2–

3 cm long, glabrous within and without; corona filaments in several series, outer 2 series erect, 2.5 mm tall, blue-purple; androgynophore extended (up to 2 cm tall), glabrous, not spotted; ovary subglobose, pilose; fruit globose, pilose, 2.5–3 cm in diameter, green at maturity; seeds oblong-cuneate, 1–3-toothed apically, the middle apical tooth up to 1 mm long, reticulate, 5–6 mm long, 2–2.25 mm wide.

Distribution: Near-endemic to the Baja California Peninsula, Mexico, state of

Baja California, municipalities of Ensenada, Mexicali (including Isla Ángel de la Guarda 149 and Isla San Lorenzo Sur); state of Baja California Sur, municipalities of Comondú, La

Paz (including Isla San Francisquito), Loreto (including Isla Carmen, Isla Coronados, and

Isla Danzante), Mulegé (including Isla San Marcos), and Santa Rosalia; state of Sonora, municipality of Hermosillo (Isla Tiburón, Isla Turners [Isla Dátil], and Isla Turón [Isla

San Esteban]); see Figure 5-6B.

Habitat: Sub-/topsoil 39.6/33.2% clay, 25.7/23.7% silt, 33.1/43.1% sand,

7.7/12.8% gravel; sub-/topsoil pH 7.7/7.3; minimum/maximum temperature 4.3/38.2

(22.3)˚C; annual precipitation 13.2 mm; elevation 0–574 (96) m.

Phenology: Budding, flowering, and fruiting year-round.

Passiflora pentaschista (Killip) H.T.Svoboda, comb. nov. Passiflora arida (Mast. &

Rose) Killip var. pentaschista Killip, Publ. Field Mus. Nat. Hist., Bot. Ser. 19: 470. 1938.

Type: Mexico, [Baja California Sur, Los Cabos], San José del Cabo, March–June 1897,

A. W. Anthony 333 (holotype, US!; isotypes, GH!, F!, MO!). =Passiflora arida (Mast. &

Rose) Killip var. cerralbensis Killip, Publ. Field Mus. Nat. Hist., Bot. Ser. 19: 470. 1938.

Type: Mexico, Baja California Sur, La Paz, Ruffo’s Ranch, “Cerralbo Island” [Isla

Cerralvo/Isla Jacques Cousteau], 3 June 1921, I. M. Johnston 4043 (holotype, US!; isotypes, A! [left specimen], CAS!, GH!, K!, NY!, UC!).

Description: Habit an upright vine; stems densely lanuginose with golden-white hairs; stipules deeply divided, conspicuous, pubescent, lacking resin glands (or very sparingly present) (Figure 5-7D); petioles 10–29 mm long, lanuginose, lacking resin glands and petiolar nectaries; leaves deeply 3- to 5-lobed (occasionally 7), apices rounded, the middle lobe spathulate, densely sericeous-lanuginose with straight, golden- 150 brown hairs (rarely velutinous), lacking resin glands (or very sparingly present), margins entire, base cordate (Figure 5-7F, G, H); peduncles 21–54 mm long, villosulous- pubescent; involucral bracts bi- to tripinnatisect, densely lanuginose with straight, golden hairs, sometimes reddish, lacking resin glands (or inconspicuously present) (Figure 5-

7E); flowers 2.5–3.5 cm wide (Figure 5-2A); sepals obtuse, hirsute without only, greenish-white, awned; petals white, oblong, glabrous within and without; corona filaments in several series, the outermost magenta (basal third), apical half pink-violet, with white between; androgynophore glabrous, with reddish spots; ovary subglobose, densely hirsute; fruit globose, villose, green at maturity; seeds 3-toothed apically, oblong, reticulate, base rounded to subtruncate, 3.5 mm long, 2 mm wide.

Distribution: Found in Mexico, state of Baja California, municipality of Ensenada

(including Isla Todos Santos); state of Baja California Sur, municipalities of Comondú,

La Paz (including Isla Jacques Cousteau [Isla Cerralvo]), Loreto (including Isla Santa

Catalina [Isla Catalan]), Los Cabos, and Mulegé; state of Sinaloa, municipality of

Mazatlán; state of Sonora, municipalities of Bacadéhuachi, Benito Juárez, Guaymas,

Hermosillo, and Navojoa; see Figure 5-6E; recent introductions in the states of Arizona and California, United States.

Habitat: Sub-/topsoil 23.3/25.3% clay, 26.1/25.2% silt, 50.6/49.4% sand,

11.1/11.7% gravel; sub-/topsoil pH 7.5/7.3; minimum/maximum temperature 2.0/38.5

(22.7)˚C; annual precipitation 21.6 mm; elevation 0–1020 (140) m.

Phenology: Budding September–March, flowering and fruiting October–May. 151

Notes: Some specimens, particularly those collected in the United States, exhibit a velutinous-lanulose leaf pubescence that is much shorter than what is typical for the species. Cultivated plants also tend to show this type of vestiture across the plant body.

Key to the Baja California Peninsula Species of Passiflora

1a. Plants bearing conspicuous glands ...... 2

1b. Plants lacking glands (or very sparingly present) ...... 4

2a. Corona filaments erect, less than 5 mm long; leaf margins finely serrate ......

...... P. palmeri

2b. Corona filaments spreading, greater than 5 mm long; leaves entire to coarsely dentate

...... 3

3a. Leaves variously 3-lobed, sparingly to densely pilose-hirsute...... P. foetida

3b. Leaves equally 3-lobed, densely lanate ...... P. fruticosa

4a. Leaves unequally 3- to 5-lobed, sericeous-lanuginose ...... P. pentaschista

4b. Leaves equally 3-lobed, densely lanate-tomentose ...... P. arida

Acknowledgements

This work represents part of the dissertation research of the first author (HTS).

The authors are grateful to three reviewers who helped improve this paper; to Harvey

Ballard, Jr. (Ohio University), Max Fancourt (IUCN Red List Unit), Bob Holzinger, John

MacDougal (Harris-Stowe State University), and Jon Rebman (San Diego Natural

History Museum) for insightful conversations; to Vincent Bonhomme (Université de

Montpellier) for assistance with the morphometric analyses; to Mattias Lanas for the illustrations; and to Jon Rebman and W. Erik Voss for supplying photographs. We also 152 thank the curators and staff of the A, ARIZ, ASU, B, BRIT, CAS, COL, DES, DUKE, E,

GH, HCIB, MEXU, MO, NY, P, RSA, SD, TEX-LL, UC, UCR, US, and VT herbaria for access to collections, loans of specimens, and/or scanned material. This work was supported in part by an Ohio University Student Enhancement Award and a fellowship through the Ohio Center for Ecology and Evolutionary Studies awarded to HTS.

153

Table 5-1

Climatic and soil variables obtained from ArcMap used in this study. Source Variable Units Harmonized World Soil Annual potential evapotranspiration mm Index (FAO/IIASA/ISRIC/ Elevation m ISSCAS/ JRC, 2012) Subsoil clay content % weight

Subsoil gravel content % volume

Subsoil sand content % weight

Subsoil silt content % weight

Subsoil pH pH

Topsoil clay content % weight

Topsoil gravel content % volume

Topsoil sand content % weight

Topsoil silt content % weight

Topsoil pH pH WorldClim version 2.0 Annual mean temperature (BIO1) ˚C (Fick & Hijmans, 2017) Mean diurnal range (BIO2) ˚C

Isothermality (BIO3) %

Temperature seasonality (BIO4) ˚C

Maximum temperature of warmest month (BIO5) ˚C

Minimum temperature of coldest month (BIO6) ˚C

Temperature annual range (BIO7) ˚C

Mean temperature of wettest quarter (BIO8) ˚C

Mean temperature of driest quarter (BIO9) ˚C

Mean temperature of warmest quarter (BIO10) ˚C

Mean temperature of coldest quarter (BIO11) ˚C

Annual precipitation (BIO12) mm

Precipitation of wettest month (BIO13) mm

Precipitation of driest month (BIO14) mm

Precipitation seasonality (BIO15) mm

Precipitation of wettest quarter (BIO16) mm

Precipitation of driest quarter (BIO17) mm

Precipitation of warmest quarter (BIO18) mm

Precipitation of coldest quarter (BIO19) mm

154

Table 5-2

Results of linear discriminant analyses (LDA) of environmental data based on different combinations and groupings of taxa. Combination or group(s) of taxa LD1 (% LD2 (% LD3 (% Classifier variation) variation) variation) (% correct) All taxa together (including P. foetida) 48.5 24.7 12.6 72.4 All taxa, with P. arida as 3 groups 54.9 27.9 09.3 80.9 All taxa, with P. arida as 2 groups 56.5 32.6 10.9 81.6 (vars. arida + cerralbensis) All taxa, with P. arida as 2 groups 66.4 23.8 09.8 85.6 (vars. cerralbensis + pentaschista) P. arida complex, as 3 groups 88.1 11.9 — 86.8 P. arida complex, as 2 groups (vars. 100 — — 89.2 arida + cerralbensis) P. arida complex, as 2 groups (vars. 100 — — 100 cerralbensis + pentaschista) LD1, linear discriminant axis 1; LD2, linear discriminant axis 2; LD3, linear discriminant axis 3.

155

Table 5-3

Comparison of features used to distinguish Passiflora arida from P. pentaschista. Values in parentheses are means. Character Passiflora arida Passiflora pentaschista Stems Densely lanate Lanuginose Stipules Often small, inconspicuous Conspicuous, deeply divided Petiole length 6–16 (9.5) mm 9–29 (15.3) mm Leaf lobing 3 rounded, subequal lobes 3–7 elongate, rounded lobes, the middle lobe spathulate Leaf vestiture Yellow-gray, lanate to tomentose Golden brown, sericeous- lanuginose Peduncle length 18–41 (28.6) mm 21–55 (32.4) mm Involucral bract vestiture Densely white, pilose Golden, lanuginose Color pattern of outer corona Basal fifth blue, apical third Basal third magenta, apical half filaments blue-violet, white between pink-violet, white between Seeds 3 prominent teeth apically, base Weakly 3-toothed apically, base cordate rounded to subtruncate

156

Figure 5-1. Map of the Baja California Peninsula and vicinity. Base map rendered with SimpleMappr (www.simplemappr.net).

157

Figure 5-2. Some of the Passiflora taxa found on the Baja California Peninsula. A. Flower and leaf of Passiflora pentaschista (Killip) H.T.Svoboda, comb. nov. (º P. arida (Mast. & Rose) Killip var. pentaschista Killip). B. Leaves and involucral bracts of P. fruticosa Killip. C. Flower, leaves, and buds of P. palmeri Rose. Photo credits: A, W. Erik Voss; B, C, Jon Rebman.

Figure 5-3. Six homologous landmarks (red circles) representing the primary venation (blue lines) and overall leaf shape. Leaf isolated from an isotype specimen of Passiflora arida var. cerralbensis (GH00067991). 158

Figure 5-4. The resulting ordination (Axes 1 and 2) from a linear discriminant analysis (LDA) of the environmental data. Passiflora arida (sensu stricto), red triangles; P. arida vars. cerralbensis and pentaschista (combined), blue and green squares, respectively; P. fruticosa, orange triangles; P. palmeri, purple circles.

159

Figure 5-5. Representation of a linear discriminant analysis (LDA) of the Passiflora arida complex based on environmental data. A. Significant separation (p = 1.711-9) between P. arida vars. cerralbensis and pentaschista as one group (red) and variety arida as another (blue). B. Non-significant separation (p = 0.2671) between P. arida vars. arida and cerralbensis as one group (blue) and variety pentaschista as another (red).

160

Figure 5-6. Distributions of the Passiflora species native to the Baja California Peninsula. A. Passiflora fruticosa (orange triangles). B. P. palmeri (purple circles). C. P. foetida, peninsular specimens only (blue triangles). D. P. arida (red triangles). E. P. pentaschista, comb. nov. (combined collections of P. arida vars. pentaschista and cerralbensis; green squares). Base maps rendered with SimpleMappr (www.simplemappr.net).

161

Figure 5-7. Passiflora arida (A–C) and P. pentaschista, comb. nov. (D–H). A. Stipule. B. Involucral bract. C. Leaf; inset showing detail of vestiture. D. Stipule. E. Involucral bract. F–H. Representative leaves; inset showing detail of vestiture. Drawn by Mattias Lanas from herbarium material.

162

CHAPTER 6: MOLECULAR PHYLOGENETICS

Introduction

The relationships among species of Passiflora, much less those of section

Dysosmia, have been a major research focus since the first phylogenetic study was published by Muschner et al. (2003). Several other studies have sought to test various evolutionary hypotheses in the genus (Clifford, 2017; Hansen et al., 2006; Krosnick et al.,

2013; Muschner et al., 2012; Souza-Chies et al., 2005; Yockteng & Nadot, 2004) with varying degrees of success at different taxonomic levels. These studies have provided researchers fair-to-good resolution of the relationships of the proposed subgenera and, to a lesser degree, supersections.

These studies have paved the way for expanding molecular investigations throughout the genus but had a limited sampling for selecting appropriate markers to use in phylogenetic reconstruction in other parts of Passiflora. Section Dysosmia, in particular, is often neglected in such studies and usually represented by one or two species of the section. This lack of Dysosmia accessions has hindered our understanding of the section’s placement in subgenus Passiflora and the relationships of the species within it. Here, I set out to provide the first-ever phylogenetic investigation focusing entirely on section Dysosmia.

Methods

The methodologies described here were carried out in the laboratory of Harvey

Ballard at Ohio University following protocols and procedures used regularly in the lab.

Modifications were made where necessary to optimize the procedures. 163

Taxon Sampling and Outgroup Selection

Forty-one specimens representing a thorough sampling the morphological diversity of section Dysosmia (sensu lato), including members from each of the four informal groups, were selected for this molecular study (see Appendix C). Additionally, six outgroup species were chosen to test the monophyly of Dysosmia based on the results of previous phylogenetic studies (namely Hansen et al., 2006; Krosnick et al., 2013;

Muschner et al., 2003, 2012; Souza-Chies et al., 2005; Yockteng & Nadot, 2004) and expert opinion (J. MacDougal, personal communication): Passiflora menispermifolia

Kunth and P. nephrodes Mast. (both subgenus Passiflora, supersection Stipulata Feuillet

& J.M.MacDougal, section Granadillastrum Triana & Planch.) as hypothesized sister taxa to Dysosmia (sensu stricto), and P. ampullacea (Mast.) Harms (subgenus Passiflora, supersection Tacsonia (Juss.) Feuillet & J.M.MacDougal, section Colombiana

L.K.Escobar, series Leptomischae L.K.Escobar), P. holosericea L. (subgenus Decaloba

(DC.) Rchb., supersection Auriculata J.M.MacDougal & Feuillet), P. maliformis L.

(subgenus Passiflora, supersection Laurifolia (Cervi) Feuillet & J.M.MacDougal, series

Tiliifoliae Feuillet & J.M.MacDougal), and P. trisecta Mast. (subgenus Passiflora, supersection Tacsonia, section Manicata (Harms) Feuillet & J.M.MacDougal) as more distantly related taxa that may place the Dysosmioides group apart from Dysosmia (sensu stricto).

DNA Extraction and Purification

Total genomic DNA was extracted primarily from herbarium tissue, supplemented by silica dried material where available, using a CTAB method used by 164

Krosnick et al. (2013) that was modified from Doyle and Doyle (1987). Approximately

50 mg of leaf tissue was flash frozen with liquid nitrogen in a 1.5 mL microcentrifuge tube and ground using a minipestle to obtain a fine powder. To each tube, the following was added: 500 µL of 1´ CTAB solution (prepared with 2 g cetyl trimethylammonium bromide [CTAB], 8.18 g 5M sodium chloride [NaCl], 4 mL 0.5M ethylenediaminetetraacetic acid [EDTA; 8.0 pH], 10 mL 1M Tris-Cl [8.5 pH], 1 g polyvinylpyrrolidone [PVP], 2 µL β-mercaptoethanol [per sample], and 97 mL distilled water). Samples were incubated in a water bath first at 50˚C for 30 minutes then at 65˚C for 15 minutes. After incubation, 400 µL of 24:1 chloroform:isoamyl alcohol was added to each tube. The tubes were centrifuged at 12,500 rpm for eight minutes in an Eppendorf

5424 microcentrifuge (Eppendorf, Hamburg, Germany) to separate the phases, the supernatant transferred to a new 1.5 mL microcentrifuge tube, and then 30 µL of 3M sodium acetate (5.2 pH) and 510 µL of cold isopropanol added to each new tube. Samples were centrifuged at 12,500 rpm for 10 minutes following precipitation in a -20˚C freezer for two weeks. The resulting pellet was washed with 500 µL 70% ethanol, then allowed to dry in a fume hood for approximately one hour. Pellets were rehydrated with 100 µL of

1´ TE buffer (8.5 pH; prepared with 1 mL 1M Tris-Cl, 200 µL 0.5M EDTA, and 98.8 mL distilled water) and incubated in a 65˚C water bath for five minutes.

Because of the ubiquitous resin glands found on the leaves of Dysosmia taxa— which may have hindered downstream applications—samples were preemptively purified immediately following precipitation with a GeneCleanÒ Turbo Kit (MP Biomedicals,

Santa Ana, CA) using the provided protocol. DNA concentration (ng/µL) and purity 165

(A260:A280) were quantified using a NanoDropÔ 1000 spectrophotometer (Thermo Fisher

Scientific, Waltham, MA).

Marker Selection and Amplification

Among the molecular markers with the potential to provide adequate phylogenetic resolution are the nuclear ITS (internal transcribed spacer) marker and the nuclear- encoded, chloroplast-expressed ncpGS (or as GScp; glutamine synthetase) gene (both previously used in Passiflora, see Krosnick et al., 2013; Yockteng & Nadot, 2004). An additional two markers, the chloroplast trnS-trnfM and psbZ-trnG, were selected as they showed promise in Passiflora from preliminary screenings (S. Krosnick, personal communication). The primers used in this study were developed by other sources (Shaw et al., 2005; Sun, Skinner, Liang, & Hulbert, 1994; White, Bruns, Lee, & Taylor, 1990) and are shown in Table 6-1.

Table 6-1

Primers used in the molecular study. Primer Name Sequence 17SE, forward (for amplification) 5’ ACGAATTCATGGTCCGGTGAAGTGTTCG 3’ 26SE, reverse (for amplification) 5’ TAGAATTCCCCGGTTCGCTCGCCGTTA 3’ ITS5, forward (for sequencing) 5’ GGAAGTAAAAGTCGTAACAAGG 3’ ITS4, reverse (for sequencing) 5’ TCCTCCGCTTATTGATATGC 3’ GScp839, forward 5’ CACCAATGGGGAGGTTATGC 3’ GScp1056, reverse 5’ CATCTTCCCTCATGCTCTTTGT 3’ trnS(GCU), forward 5’ AGATAGGGATTCGAACCCTCGGT 3’ trnfM(CAU), reverse 5’ GCGGAGTAGAGCAGTTTGGT 3’ psbZ, forward 5’ GGTACMTCATTATGGATTGG 3’ 5’trnG2S, reverse 5’ TTTTACCACTAAACTATACCCGC 3’ 166

All genetic markers were amplified from genomic DNA samples via polymerase chain reaction (PCR) using the KAPA3G Plant PCR Kit (Kapa Biosystems, Wilmington,

MA). The constituents for each 50 µL reaction included: 25 µL of 2´ KAPA Plant PCR

Buffer (containing 1.5 mM MgCl2 [1´] and 0.2 mM of each dNTP [1´]), 20.6 µL PCR- grade water, 1.5 µL of each primer (10 mM), 1 µL template DNA (5 ng/µL), and 0.4 µL of KAPA3G Plant DNA Polymerase (2.5 U/µL). PCR was performed using a 2720

Thermal Cycler (Applied Biosystems, Foster City, CA) starting with an initial denaturing step of four to five minutes at 95˚C and ending with a final extension step of 5 minutes at

72˚C. The cycle conditions and number varied depending on the marker being amplified

(see Appendix D). Namely, the annealing temperature was adjusted between trials to test for optimized amplification: for ITS, annealing temperatures between 50˚C and 65˚C (at

1˚ C increments) were used; for ncpGS, annealing temperatures between 50˚C and 57˚C

(at 1˚ C increments) were used; for trnS-trnfM, temperatures of 58˚C and 60˚C were used; for psbZ-trnG, temperatures between 53˚C and 57˚C (at 2˚C increments) were used.

The products from PCR were electrophoresed on a 1.3% agarose gel (0.39 g agarose, 30 mL 1´ TBE [Tris-Borate-EDTA] buffer, and 1.5 µL ethidium bromide) at 96

V for approximately one hour using a FisherBiotech FB-SB-710 Mini-Horizontal Unit

(Thermo Fisher Scientific) and visualized with a Gel DocÔ EZ Imager (Bio-Rad,

Hercules, CA) to determine the success or failure of the PCR amplification. Following the manufacturers’ protocols, successful amplifications were cleaned to remove leftover

PCR constituents either with the WizardÒ SV and Gel Clean-Up System (Promega,

Madison, WI) in the case of samples with only a single target band, or the MonarchÒ 167

DNA Gel Extraction Kit (New England BioLabs, Ipswich, MA) in the case of samples having one or more extraneous bands in addition to the target band. DNA concentration and purity were again checked after cleanup using the NanoDrop spectrophotometer. For samples that could not be measured accurately with the NanoDrop (e.g., those whose concentrations did not register), a QubitÒ 3.0 fluorometer (Invitrogen, Carlsbad, CA) was used to measure DNA quantity.

Sequencing

Direct Sanger cycle sequencing reactions and sequence analysis were carried out at the Ohio University Genomics Facility (Athens, OH) using a 3130xl Genetic Analyzer

(Applied Biosystems) and the BigDye™ Terminator v3.1 Cycle Sequencing Kit (Applied

Biosystems). Each reaction contained 0.5 µL of BigDye™ Reaction Mix, 1 µL of 3.2 µM primer, 4 µL of PCR-grade water, and 1 µL of 5-10 ng/µL template DNA. An initial incubation step of 96˚C for 1 minute was followed by 30 cycles of 96˚C for 10 seconds

(denature), 53˚C for 5 seconds (anneal), and 60˚C for 2.5 minutes (extension), with a final incubation at 4˚C until clean-up could occur. Sequencing products were then cleaned using the BigDye XTerminator™ Purification Kit (Applied Biosystems).

Additional sequencing trials was performed at Molecular Cloning Laboratories (South

San Francisco, CA) using a 3730xl DNA Analyzer (Applied Biosystems).

Results

Conditions for successful amplification of the four markers varied throughout the optimization process. For two of the markers (ITS and trnS-trnfM), amplification was initially high but often produced multiple extraneous (non-target) bands at lower 168 annealing temperatures. The optimal annealing temperature for ITS and trnS-trnfM was found to be 63.5˚C and 60˚C, respectively. The ncpGS marker had an overall lower initial success rate, but subsequent trials with varying annealing temperatures (50–57˚C) improved amplification differentially. Conditions optimized by Yockteng & Nadot

(2004) provided the best amplification results with an annealing temperature of 55˚C.

The marker psbZ-trnG did not amplify during any of the PCR trials despite numerous changes to the reaction conditions.

Of the 37 direct sequencing reactions attempted across the three markers that amplified, only three samples (using the trnS forward primer) were successful: Passiflora vestita, P. foetida var. maxonii, and P. fruticosa. Percent similarity was calculated between these sequences and showed a 98% similarity between the former two taxa; these, in turn, having a ~54% similarity with the latter taxon. However, all direct sequencing trials of ITS and ncpGS failed outright.

Discussion and Future Directions

The inability to sequence the samples remains puzzling. Although high quality

DNA was successfully extracted from herbarium material and the markers properly amplified, sequencing has not been equally as successful. Previous phylogenetic studies involving members of section Dysosmia, and using many of the same markers, have apparently had no difficulties directly sequencing samples. One putative, albeit not compelling, explanation for failure might be the ubiquitous foliar resin glands interfering with the Sanger sequencing reactions. Another unexplored possibility is the existence of secondary structures within the DNA molecule, which may be the cause of aborted 169 sequencing reactions. Primer solutions, DNA cleaning protocols, and target sequence concentrations have all been checked multiple times to ensure correct parameters, and each time the assessment is that all of these variables are adequate for sequencing. The reasons for the continued failure are thus far unclear.

Based on the results of this work, future phylogenetic studies in section Dysosmia may benefit from using genomic approaches rather than the individual marker method used here. Among several potential approaches, anchored phylogenomics appears to be a viable option (S. Krosnick, unpublished data) as well as hybrid enrichment-sequence capture (S. Achá, unpublished data), as they are both currently being used for robust phylogenetic studies in subgenus Decaloba. Future molecular work in Dysosmia may require use of these methods to produce a phylogenetic reconstruction.

Acknowledgements

I would like to thank Bill Broach and Rachel Yoho (Ohio University Genomics

Facility), Danny Wolf (Ohio University), and Serena Achá (University of Missouri, St.

Louis) for helpful discussions and advice for the molecular work and sequencing. I am especially grateful to Shawn Krosnick (Tennessee Tech University) for her unyielding and continuing collaboration and assistance in developing this project. I also appreciate the support, use of lab equipment, and guidance provided by Harvey Ballard throughout.

170

CHAPTER 7: SYNOPSIS OF THE RED-FRUITED MEMBERS OF PASSIFLORA

SECTION DYSOSMIA

This chapter represents a major portion an eventual monograph of section

Dysosmia intended for submission to the journal Systematic Botany Monographs; the style guidelines of this journal have been generally followed for the taxonomic section of this chapter. New species or new combinations described in this dissertation are not intended for effective or valid online publication here; one or more separate manuscripts will be submitted to scientific journals to publish these taxonomic novelties at a later time.

Introduction

Section Dysosmia DC. is considered one of the most taxonomically complicated in the genus Passiflora L. (Passifloraceae Juss. ex Roussel). Understanding of this group is confounded by the apparently highly variable morphology of the taxa within the section, which seems to be made up of two or more broad species complexes. Botanists have created over one hundred infrageneric names within this relatively small section and repeatedly shuffled and reshuffled taxa between these names.

One of the species complexes revealed by Svoboda and Ballard (2018), the

Passiflora ciliata Aiton complex, is distinguished by its fully glabrous vegetative features and red fruit. A second complex, informally recognized as the red-fruited pubescent

(RFP) group, contains species that also bear red fruit but have pubescent vegetative parts.

Svoboda and Ballard (2018) found that although both of these complexes have red fruit, a rare trait in the genus, only the P. ciliata complex consistently clustered together in each 171 of the morphological analyses. The RFP complex did not form a discrete group in any these same analyses. One possible explanation for this result is that red fruit has evolved multiple times in Dysosmia and did not arise from a single red-fruited ancestor, or perhaps that vegetative pubescence is labile and represents a reversion in the RFP group.

Future DNA sequence studies will be needed to test these hypotheses.

While the monophyly of a single red-fruited clade—or that of the two complexes—could not be confirmed (see Chapter 6), all of the red-fruited taxa are here treated together in this synopsis for convenience. Fruit color, it seems, provides a potentially natural grouping in other plant groups (Givnish et al., 1995; Miller et al.,

2011) and may similarly be supported by phylogenetic analyses of section Dysosmia in the future.

Distribution and Endemism

The red-fruited members of section Dysosmia have a wide distribution nearly comparable to that of the pubescent, green-fruited members of the section. The glabrous taxa (the “P. ciliata complex”) are concentrated in Mexico and the countries of Central

America, with additional collections from the West Indies including Anguilla, the

Bahamas, Bermuda, the British Virgin Islands, Cuba, the Dominican Republic, Haiti,

Jamaica (the type locality of P. ciliata [sensu stricto]), Martinique, St. Kitts and Nevis, and the Turks and Caicos. Collections are also known from four South American countries: Brazil, Colombia, French Guiana, and Venezuela. I have also seen a handful of glabrous collections from Florida and Texas (United States of America), but these likely represent introduced populations (see Goldman & MacDougal, 2015). Red-fruited 172 pubescent (RFP) taxa also tend to be concentrated in Mexico and Central America

(collections known from Belize, Costa Rica, El Salvador, Guatemala, Nicaragua). Two island nations in the West Indies (Cuba and Jamaica) are also known to harbor RFP taxa.

Many of the red-fruited taxa, whether in the RFP or P. ciliata complex, appear to be endemic to specific countries, islands, or regions. Passiflora bahamensis, for example, is endemic to the islands of the Bahamas, while P. ciliata var. santiagana is known only from two provinces in Cuba. Other endemics, at different political levels, include P. lepidota (endemic to Brazil), P. urbaniana and P. foetida var. subintegra (Belize), P. sublanceolata (Yucatan Peninsula), and P. ciliata var. hibiscifolia (southern Mexico).

Taxonomic History, Nomenclature, and Classification

In an intensive nomenclatural investigation of section Dysosmia (see Chapter 2;

Svoboda et al., 2016), all of the names associated with the section were reviewed to ensure proper typification and the extent to which published names represent the biological diversity of the group. I have determined at least 11 names representing strictly glabrous taxa (the P. cilata complex) and an additional seven representing the RFP complex. Generally, these names correspond well with the known diversity, though undescribed taxa certainly still exist as more collections are examined in finer detail. In- depth investigations of micromorphology, for example, are providing evidence for the recognition of taxa that have been previously lumped into one (or both, by different researchers) of the species complexes mentioned.

Many of the Dysosmia names that were recognized by Killip (1938), and variously treated by Vanderplank (2013), have been intensely studied and confirmed to 173 represent distinct species with suites of traits that distinguish them from all other taxa.

Other names that were previously treated as varieties or as synonyms of other taxa were likewise investigated, and I believe that some of them also represent distinct species. In this work I have adopted a classification scheme that primarily agrees with previous authors (e.g., Killip, 1938; Vanderplank, 2013), but have decided to recognize some of these infraspecific taxa at the rank of species based on their geographic ranges and distinct suites of morphological traits.

All members of section Dysosmia share a number of morphological characters, many of which are also general characters of the genus: leaves alternate; involucral bracts

3; flowers bisexual, axillary, solitary; sepals 5, awned; petals 5, glabrous; androgynophore erect and glabrous; stamens 5, glabrous; styles 3; fruit a berry.

Key to the Red-Fruited Dysosmia Taxa

1. Plants with pubescent vegetative features ...... 2

1’. Plants with glabrous vegetative features ...... 7

2. Leaves bearing glands on the abaxial and/or adaxial surface ...... 3

2’. Leaves without glands on either surface; Tabasco, Mexico ...... P. aurea

3. Leaves deeply 3-lobed to deeply 5-lobed ...... 4

3’. Leaves nearly unlobed to weakly 3-lobed ...... 6

4. Leaves yellow lanuginose, deeply 3- to 5-lobed; Cuba ...... P. santiagana

4’. Leaves variously pubescent, deeply 3- or 5-lobed; Central America and Mexico ...... 5

174

5. Leaves pilosulous, sub 5-lobed to deeply 5-lobed; Central America ...... P. maxonii

5’. Leaves variously pubescent, deeply 3-lobed; Central America and Mexico......

...... RFP complex (see discussion under P. maxonii)

6. Leaves densely tomentose; flowers with an elongated androgynophore, perianth bright pink ...... P. sublanceolata

6’. Leaves velvety-scabrous, coriaceous; flowers with a typical-length androgynophore, perianth blue-purple ...... P. urbaniana

7. Leaves bearing ubiquitous lepidote glands abaxially ...... 8

7’. Leaves bearing capitate or pyriform long-stipitate glands on the laminar surface and/or the margins ...... 10

8. Leaves narrowly lanceolate or deeply 3- to 5-lobed ...... 9

8’. Leaves 3-lobed, triangular in shape, glands drying yellow; Brazil ...... P. lepidota

9. Leaves narrowly lanceolate or obscurely 3-lobed near base, coriaceous; Belize ......

...... P. subintegra

9’. Leaves deeply 3-lobed to deeply 5-lobed; southern Mexico ...... P. hibiscifolia

10. Leaves lacking glands on both surfaces, but often present on the margins ...... 11

10’. Leaves with glands on one or both surfaces ...... 12

11. Leaves shallowly 3-lobed, narrowly pandurate, stipules inconspicuous; Bahamas ......

...... P. bahamensis

11’. Leaves deeply 3- to sub 5-lobed, stipules conspicuous; West Indies and Central

America...... P. ciliata (s.s.)

175

12. Leaves bearing glands on both surfaces; Cuba ...... P. wrightiana

12’. Leaves bearing glands only on the abaxial surface ...... 13

13. Leaves cordate-deltoid, involucral bract segments not deeply pinnatisect; flowers with elongated androgynophore; West Indies and Bermuda ...... P. pectinata

13’. Leaves variously 3- or 5-lobed, bracts deeply pinnatisect, flowers with typical length androgynophore; Central America, South America, and West Indies ...... 14

14. Leaves subhastate, weakly 3-lobed; South America ...... P. orinocensis

14’. Leaves deeply 3-lobed or 5-lobed; Central America or West Indies ...... 15

15. Leaves linear-subauriculate, 3 unequal lobes, middle lobe apex acuminate; Central

America...... P. subauriculata

15’. Leaves subcordate in general shape, deeply and equally 3- or 5-lobed, lobe apices subacute to rounded; Cuba ...... P. pseudociliata

Taxonomic Treatment

The names that appear below are formatted for different taxonomic statuses: numbered names in small caps are considered accepted in this work; those in regular italics are synonyms (homotypic or heterotypic); those in bold italics are provisional new names or combinations (which are not intended for effective publication here). All herbarium codes follow the Index Herbariorum (Thiers, 2018). Collections marked with an asterisk (*) indicate specimens that have an affinity with the taxon they are listed under but are potentially in conflict with the protologue or type collection; those marked with a dagger (†) are paratypes of the name; types marked with an exclamation point (!) were personally seen and confirmed as types of the name. 176

Accepted and New Taxa

1. Passiflora aurea H.T.Svoboda, sp. nov. [ined.] TYPE: MEXICO. Tabasco:

Huimanguillo municipality, km. 10.4 de la desviación de Huimanguillo-Fco.

Ruedo, 27 Jun 1981, C. Cowan 3325 (holotype: MO2921420!; isotype:

MEXU365664!). (See Figure 7-1.)

Description—Stems slender, terete, sparingly golden tomentulose-villose. Leaves ensiform to hastate, golden tomentulose, lacking glands save a single, long apical gland; narrowly 3-lobed, the middle lobe up to 5´ longer than the basal lobes; petioles golden villose, glandless; stipules pubescent, long filiform, divided, with capitate glands.

Flowers morphology unknown; peduncles villose; involucral bracts pubescent, bipinnatisect, the ultimate segments gland-tipped. Fruits sparingly hirsute, globose, red at maturity; seed morphology unknown.

Distribution—Currently only known from the type locality in Tabasco, Mexico.

Habitat—Known to grow in flooded pastures; elevation 30 m.

Phenology—Type specimen fruiting in June.

Etymology—The species is named for the yellowish vestiture and overall golden appearance of the leaves when dried.

Taxonomic Notes—Currently this new species is known only from a single collection, hindering our complete understanding of the taxon. However, the extreme leaf shape, golden tomentulose vestiture, and lack of foliar glands set this entity apart. Other species with a similar leaf shape include Passiflora subauriculata, sp. nov. and P. subintegra, comb. nov. (below). 177

Figure 7-1. Leaves from the proposed isotype of Passiflora aurea, sp. nov.

2. PASSIFLORA BAHAMENSIS Britton, Bull. New York Bot. Gard. 5: 315. 1907. TYPE:

BAHAMAS. New Providence Island, South Pike Road, 31 Aug 1904, N. L. Britton

& L. J. K. Brace 392 (holotype: NY00084318!; isotypes: F171822!, F171823!,

K000323252!).

Description—Stems subterete, striate, glabrous, often purplish-red. Leaves glabrous, lustrous above, coriaceous, narrowly pandurate, margins denticulate with glands; 3-lobed, the middle lobe elongated, the tip acuminate, lateral lobes smaller, 178 rounded to subangular; petioles glabrous, with few glands; stipules small, deeply pinnatisect almost to the stem, glabrous, gland-tipped. Flowers white to pink-violet; peduncles glabrous; involucral bracts glabrous, deeply bi- or tripinnatisect, the ultimate segments gland-tipped; sepals white to pink-violet, ovate-lanceolate, glabrous; petals white to pink-violet, ovate-oblong; corona filaments in 5 series, the inner three capillary, the outer two filiform, spreading, the proximal portion banded with violet and white, the distal third to half purple to magenta; androgynophore cream-colored; stigmas glabrous; ovary ovoid, glabrous. Fruits glabrous, globose, lustrous, deep red at maturity; seeds obovoid, punctate, brown-black at maturity.

Distribution—Endemic to the Bahamas, known from Andros, Cat, and New

Providence islands.

Habitat—Rocky pine barrens, disturbed limestone shrublands; elevation 0–18 m

(8 m average).

Phenology—Budding, flowering, and fruiting year-round.

Taxonomic Notes—This species is easily distinguishable by its unique, pandurate leaf shape and the relatively limited geographic distribution in the Bahamas.

Additional Specimens Examined—BAHAMAS. Andros Island: Central Andros,

Fehling 31 (FTG), Freid & Szczecinski 06-724 (FTG), Sauleda 1214 (FTG), Sauleda

1214A (FTG), Sauleda 1214F (FTG), Sauleda 1242 (FTG), Sauleda 1660 (FTG), Small

& Carter 8754 (NY, US); Mangrove Cay, Popenoe 211 (FTG), Small & Carter 8507

(NY); North Andros, Black & Black 737 (FTG), Brace 6861† (NY), Brace 7101† (NY),

Correll & Evans 43934 (FTG, MO, NY), Correll & Godfrey 41299 (FTG), Correll & 179

Proctor 47785 (NY), Correll & Proctor 47862 (FTG), Dawson 26911 (US), Goldman &

Berry 3222 (FTG, MO), Goldman & Rice 3221 (US), Harvey 7802 (TEX), Hill 3458

(FTG), Jestrow et al. 2012-001 (FTG), Jestrow et al. 2012-049 (FTG), Jestrow et al.

2012-060 (FTG), Kjellmark 106 (DUKE), Kjellmark 148 (DUKE), Northrop & Northrop

391a (NY), Popenoe 171 (FTG), Vincent et al. 11559 (UPRRP); South Andros, Brace

5022† (NY), Brace 5191† (NY), Hill 3113 (FTG, NY), Small & Carter 8571 (NY),

Vincent et al. 14209 (FTG); Unknown district, Correll et al. 49657 (NY), Lippincott s.n.

(FTG), Nickrent 2445 (NY). Cat Island: Britton & Millspaugh 5833† (NY), Richey &

Freid 00-893 (NY), Wilson 7189 (NY). New Providence Island: Brace 223 (NY), Britton

& Millspaugh 2102† (NY), Britton 55† (NY), Coker 70† (NY), Correll & Correll 48290

(FTG, NY), Correll 48452 (FTG, NY, US), Correll 51145 (FTG), Curtiss 209† (MO,

NY, US), Degener 18963 (NY), Gillis 8382 (FTG), Gillis 9200 (FTG), Ledin 333 (FTG),

Webster et al. 10460 (DUKE, US).

3. PASSIFLORA CILIATA Aiton, Hort. Kew. 3: 310. 1789. Dysosmia ciliata (Aiton)

M.Roem., Fam. Nat. Syn. Monogr. 2: 149. 1846. Passiflora foetida L. var. ciliata

(Aiton) Mast., Trans. Linn. Soc. London 27: 631. 1871. TYPE: UNITED KINGDOM.

England: [Oxfordshire county], cultivated in “Hon. Mrs. Barrington’s stove at

Mongewell,” [Crowmarsh parish], original material from Jamaica, Jul 1791,

collector and collection unknown (neotype, designated by Svoboda et al. (2016:

102): LINN-HS1418.43!). (See Figure 2-2.)

Passiflora foetida var. nicaraguensis Killip, Publ. Carnegie Inst. Wash. 461: 328. 1936.

Passiflora hastata Bertol. var. nicaraguensis Killip ex Standl., ined., Publ. Field 180

Mus. Nat. Hist., Bot. Ser. 10: 293. 1931. TYPE: HONDURAS. Atlántida: La Fragua,

7 Dec 1927, P. C. Standley 52665 (holotype: US1406059!; isotype: F582783!).

Description—Stems glabrous, terete. Leaves glabrous, bearing pyriform long- stipitate glands on the margins only; deeply 3-lobed to sub 5-lobed; petioles glabrous, bearing few stalked glands; stipules glabrous, long, filiform, with pyriform long-stipitate glands. Flowers pinkish; peduncles glabrous; involucral bracts glabrous, bipinnatisect, the ultimate segments gland-tipped with pyriform glands; sepals glabrous, lanceolate- oblong; petals lanceolate-oblong; corona filaments radiate, banded with violet; stigmas glabrous; ovary subglobose, glabrous. Fruits glabrous, globose, red at maturity; seeds obovate, punctate, brown-black at maturity.

Distribution—Known from Mexico (the states of Campeche, Chiapas, Quintana

Roo, Tabasco, Veracruz, and Yucatán), throughout Central America (Belize, Costa Rica,

Guatemala, Honduras, and Nicaragua), and Cuba and Jamaica in the West Indies.

Habitat—Coastal dunes, riparian thickets; elevation 0–163 m (45 m average).

Phenology—Flowering and fruiting year-round.

Taxonomic Notes—This species has so far been a dumping ground for various glabrous, red-fruited taxa (hence the “P. ciliata complex” designation in some works).

Upon typifying the name (Svoboda et al., 2016) and examining thousands of specimens,

Passiflora ciliata (sensu stricto) appeared to be more consistent than previously thought.

The combination of general leaf shape (deeply 3- to sub 5-lobed) and micromorphological features (i.e., no glands on the abaxial or adaxial leaf surfaces) easily separate this species from all other known taxa in section Dysosmia. 181

In addition to its distribution in Mexico, Central America, and the West Indies, collections of this species (or complex) have been reported from the states of Florida and

Texas, USA. Only one of these collections, Jones s.n. from Florida in 1920, could be confirmed as Passiflora ciliata (sensu stricto) for this treatment. Regardless of their ultimate taxonomic assignment, these glabrous, red-fruited individuals likely represent introductions into the United States (see Goldman & MacDougal, 2015).

Representative Specimens Examined—COSTA RICA. Alajuela: Upala, Herrera

1827 (DUKE, MO). CUBA. Matanzas: Jovellanos, León 16679 (GH). GUATEMALA. Petén:

San Andres, Steyermark 46022 (NY, US); Sayaxché, Contreras 9774 (LL, MO, NY),

Harmon & Fuentes 5806 (MO, NY), Lundell 17892 (LL), Lundell 18036 (DUKE, LL,

MO). HONDURAS. Atlántida: Tela, MacDougal et al. 3447 (MEXU, MO, US), Yuncker

4668 (NY). Colón: Trujillo, Saunders 317 (MO, NY), Saunders 318 (MEXU, MO, NY,

TEX). JAMAICA. Cornwall: Westmoreland, Britton 2880 (NY), Harris 11816 (MO, NY,

US). Surrey: Kingston, Maxon & Killip 1423 (NY, US). MEXICO. Campeche: Carmen,

Lloyd & Mueller 3690 (GH); Champotón, Cabrera C. & de Cabrera 14072 (MO);

Hopelchén, Álvarez 8532 (MEXU, NY); Solidaridad, Bequaert 54 (A). Chiapas:

Ocosingo, Martínez S. 13476 (MO), Martínez S. 16234 (MO). Quintana Roo: Adolfo de la Huerta, Álvarez et al. 10515 (NY). Tabasco: Cunduacan, Ventura A. 20093 (NY);

Frontera, Gliessman et al. 1702 (NY); Nacajuca, Tenorio L. et al. 19594 (TEX); Paraiso,

Conrad & Conrad 2933 (MO, US), Magaña & Cowan 2110 (NY); Unknown municipality, Hernández M. 270 (LL, MEXU), Rovirosa 560 (NY, US). Veracruz:

Coatzacoalcos, Calzada 370 (GH, MEXU, MO); Minatitlán, Juárez G. & Leyva 2618 182

(MEXU, TEX); San Andrés Tuxtla, Gilbert 78-25 (TEX); Tlacotalpan, Nee & Taylor

26725 (F, MEXU, MO, NY, XAL); Veracruz, Gonzalez G. 146 (GH, MEXU, US);

Unknown municipality, Hernández M. 194 (LL, MEXU, NY). Yucatán: Celestún, Durán

1933 (NY), Durán 2043 (NY), Ferrer et al. 157 (TEX); Dzemul, Prinzie et al. 188

(MEXU, MO); Progreso, Gaumer et al. 23355 (ECON, GH, MO, NY, US), Lundell &

Lundell 8152 (US); Santa Elena, Cabrera C. & de Cabrera 10818 (TEX), Degener &

Degener 26768 (NY, US); Telchac Puerto, Cabrera C. & de Cabrera 9703 (MO),

Cabrera C. & de Cabrera 10249 (NY), Cabrera C. & de Cabrera 11307 (MEXU, TEX),

Carnevali et al. 4335 (NY); Unknown municipality, Gaumer 630 (MO, NY, P), Gaumer

1888 (F, MO, US). NICARAGUA. Granada: Granada, Sandino 692 (MO). UNITED STATES

OF AMERICA. Florida: Lake, Jones s.n. (US).

4. PASSIFLORA HIBISCIFOLIA Lam., Encycl. 3(1): 39. 1798. Dysosmia hibiscifolia (Lam.)

M.Roem., Fam. Nat. Syn. Monogr. 2: 149. 1846. Passiflora foetida var.

hibiscifolia (Lam.) Killip, Publ. Field Mus. Nat. Hist., Bot. Ser. 19: 507. 1938.

Passiflora ciliata var. hibiscifolia (Lam.) Vanderpl., Bot. Mag. 30(4): 358. 2013

[2014]. TYPE: COUNTRY UNKNOWN. Locality information unknown, though likely

in Mexico, 13 Sep 1779, J.-B. Lamarck s.n. (holotype: P00307573!).

Passiflora liebmannii Mast. (as ‘Liebamanni’), Fl. Bras. 13(1): 547. 1872. TYPE:

MEXICO. Puebla: [Coxcatlán municipality], Venta Salada, May 1842, F. M.

Liebmann 4078 [Passiflorea 41] (holotype: C10016407!; isotype: US1084033!).

Description—Stems glabrous, terete. Leaves glabrous, margins entire to sparingly crenate, bearing short-stalked obconical glands, abaxial surface with spheroidal, 183 lepidote glands; deeply 3-lobed when immature, developing 5 (up to 7) lobes at maturity; petioles glabrous, bearing few short-stalked glands; stipules glabrous, deeply divided with pyriform long-stipitate glands. Flowers white to lavender; peduncles glabrous; involucral bracts glabrous, deeply bi- to tripinnatisect, the ultimate segments gland- tipped; sepals glabrous, ovate-lanceolate; petals oblong-lanceolate; corona filaments magenta (inner fourth) and purple (outer third) with white in between; androgynophore mostly red; stigmas glabrous; ovary ovoid, glabrous. Fruits glabrous, globose, deep red at maturity; seeds oblong, punctate brown-black at maturity.

Distribution—Found throughout southern Mexico, with a high concentration in the state of Oaxaca.

Habitat—Most commonly in coastal dunes, secondary forests, sandy roadsides, and stream beds; elevation 0–1069 m (273 m average).

Phenology—Budding, flowering, and fruiting year-round.

Taxonomic Notes—This taxon has been recognized as a species as well as a variety of each of the two main species complexes (P. foetida and P. ciliata).

Morphological investigations have revealed a suite of traits, including ubiquitous lepidote glands on the abaxial leaf surface and generally deeply 5-lobed leaves, that are unique among Dysosmia taxa. One specimen studied thus far, Spellman 2102 (marked with an asterisk below), matches with all of the morphological traits of this species, including lepidote glands, but is covered in a soft pubescence on the vegetative features. This individual may represent a hybrid of P. hibiscifolia and a pubescent taxon (e.g., P. foetida). 184

Additional Specimens Examined—MEXICO. Chiapas: Chiapilla, Gould &

Williams 138 (TEX), Martínez S. & Reyes G. 20194 (MEXU, RSA, TEX); Cintalapa,

Alexander 314 (NY); Ocozocoautla de Espinosa, Téllez V. et al. 8163 (MEXU); Tonalá,

Breedlove 25574 (MEXU, MO), Breedlove & Thorne 20852 (MEXU, MO, NY), Lent 1

(MO). Guerrero: Acapulco de Juárez, Barkley 14171 (BRIT, MEXU, MO, NY, RSA,

TEX), Forment 763 (MEXU), Forment 767 (MEXU), Forment 1140 (MEXU), García S. s.n. (MEXU), Ramírez C. s.n. (MEXU), Soto N. & Solórzano G. 12856 (MEXU);

Arcelia, Mexia 8911 (B, US); Copala, Seigler et al. 13919 (MEXU); Coyuca de Benítez,

Espinosa 89 (MEXU), Lozada P. 2 (MEXU), Perez 2 (MEXU), Perez 194 (MEXU),

Perez 515 (MEXU); Cruz Grande, Martínez S. & Villaseñor 475 (MEXU, MO); San

Jeronimo, Lozada P. 331 (MEXU); Unknown municipality, Ramírez C. s.n. (MEXU).

Michoacán: La Huacana, Martínez S. & Téllez 110. Oaxaca: Asunción Ixtaltepec,

Castillo 2288 (MEXU), Cedillo T. & Torres C. 1067 (LL, MEXU, MO); Cosolapa,

Acosta C. 788 (MEXU); Cuicatlan, Conzatti 3911 (MEXU); Heroica Ciudad de Juchitán de Zaragoza, Cedillo T. et al. 1362 (LL, MEXU, MO), Greene 8027 (IND), Nava Z. et al.

1911 (MEXU); Juchitán de Zaragoza, Alvarado C. et al. 63 (MEXU); Magdalena

Tlacotepec, Cabrera C. & de Cabrera 7363 (MEXU); Matías Romero, Sousa et al. 9194

(MEXU); San Antonio Nanahuatipam, Salinas T. & Ramos F-3880 (MEXU, RSA, US),

Tenorio L. et al. 8861 (MEXU, US); San Francisco del Mar, Franco et al. 21 (MEXU),

Martínez et al. 51 (MEXU); San Juan Bautista Cuicatlán, Aguirre B. 12 (NY), Cruz-

Espinosa & San Pedro 659 (MEXU), Cruz-Espinosa & San Pedro 1511 (MEXU),

Miranda-González 979 (MEXU), Miranda-González 4621 (MEXU), Nelson 1632 (US), 185

Rose et al. 10056 (US), Salinas T. & Ramos F-3978B (RSA), Salinas T. et al. 4160

(MEXU, MO); San Juan de los Cués, Calzada & Paredes F. 22981 (MEXU); San Juan

Quiotepec, Cruz-Espinosa 2185 (MEXU); San Martín Toxpalan, Soule & Brunner 2249

(MEXU, MO, TEX); San Mateo del Mar, Zizumbo & Colunga 67 (MEXU, MO); San

Miguel Chimalapa, Torres C. & Cedillo T. 108 (DUKE, MEXU, MO, NY); San Miguel del Puerto, Elorsa C. 2772 (MEXU); San Pedro Huamelula, Elorsa C. 1157 (MEXU),

Elorsa C. 1567 (MEXU), Elorsa C. 2647 (MEXU, MO), Perret et al. 310 (MEXU),

Rivera H. et al. 382 (MEXU); San Pedro Tapanatepec, Spellman 2102* (MO); Spellman

2103 (MO), Spellman 2104 (MO), Spellman 2106 (MEXU, MO), Spellman 2108

(MEXU, MO), Spellman 2109 (MO); San Pedro Totolapa, Webster et al. 12996 (GH,

MEXU, MO); Santa Catarina Juquila, Rusby 48 (NY, US); Santa María Guienagati,

Torres C. et al. 2561 (MEXU, MO); Santiago Astata, Elorsa C. 187 (MEXU), Elorsa C.

1530 (MEXU), Elorsa C. 1627 (MEXU), Elorsa C. 1661 (MEXU), Elorsa C. 1926

(MEXU), Elorsa C. 1962 (MEXU), Elorsa C.1982 (MEXU), Elorsa C. 2707 (MEXU),

Salas M. 1910 (MEXU); Santiago Niltepec, Pennel et al. 206 (MEXU, MO); Santo

Domingo Tehuantepec, Shapiro 3 (MEXU); Tehuantepec, Cedillo T. 562 (MEXU, NY),

González O. 334 (MEXU), Gould & J. Williams 137 (TEX), Hess et al. 1265 (ASU,

MEXU, MO, NY), King 1686 (NY, TEX, US), MacDougal 314 (DUKE, US),

MacDougall 25.S (NY), MacDougall s.n. (NY), Nelson 2589 (US), Orcutt 5272 (US),

Spellman 2110 (MEXU), Spellman 2111 (MO), Spellman 2112 (MO), Spellman 2116

(MO), Spellman 2117 (MO); Valerio Trujano, Juárez G. & F. Marini 608 (MEXU),

Salinas T. et al. 4143 (MEXU); Unknown municipality, Conzatti & Conzatti 2111 186

(MEXU), Gentry 4847 (RSA), Pringle 4847 (E, MEXU, MO, NY, P, US), Seler 1665

(NY), Smith 245 (MO, RSA, US). Puebla: Coxcatlán, MacDougal 285 (DUKE, US),

Rzedowski 25576 (MEXU), Tenorio L. & Martínez C. 17380 (MEXU); Tehuacán,

Liebmann 4079 (US). Tabasco: Centla, Novelo R. & Ramos V. 2702 (MO); Emiliano

Zapata, Cabrera C. et al. 2660 (DUKE, MEXU, MO). Yucatán: Santa Elena, Schwabe s.n. (MEXU).

Other Observations Examined—MEXICO. Guerrero: Copala, H. V. Barcenas

(“horacio_barcenas”; iNaturalist ID: 10360363), D. Fuentes (“dianafr”; iNaturalist ID:

10393452). Jalisco: Puerto Vallarta, M. C. Álvarez R. (“cecilia_mayanjays”; iNaturalist

ID: 5874307), E. Canales (“erendiracanales”; iNaturalist ID: 9524959), M. Guel A.

(“maribel_guel20”; iNaturalist ID: 9872887); K. E. Peña J. (“karenjoya”; iNaturalist ID:

9529551). Michoacán: Lázaro Cárdenas, W. Ostria (“wero-ostria”; iNaturalist ID:

8291767), H. Suzuki (“hakikuchi”; iNaturalist ID: 6096896). Morelos: Tlaquiltenango,

“annemirdl” (iNaturalist ID: 325005). Nayarit: Bahía de Banderas, A. López H. (“alexiz”; iNaturalist ID: 5115933). Oaxaca: Asunción Ixtaltepec, “tereso30” (iNaturalist ID:

7605012); Heroica Ciudad de Juchitán de Zaragoza, “tereso30” (iNaturalist ID:

7605025), A. Jossafat (“abeljossafat”; iNaturalist ID: 6168756), J. C. Pérez M. (“charro”; iNaturalist ID: 2663734); San Pedro de Huamelula, E. N. Tapia B. (“momoto-erick”; iNaturalist ID: 12537177); Santa María Huatulco, C. Domínguez-Rodríguez (“blakesito”; inaturalist IDs: 14115527, 14114385), J. C. Garcia M. (“juancarlosgarciamorales1”; iNaturalist ID: 1232027); Santo Domingo Tehuantepec, J. C. Garcia M.

(“juancarlosgarciamorales1”; iNaturalist IDs: 1166121, 1418031); Villa de Tututepec de 187

Melchor Ocampo, H. V. Barcenas (“horacio_barcenas”; iNaturalist IDs: 10710804,

10728801, 12695916), D. Fuentes (“dianafr”; iNaturalist IDs: 10311329, 10311694,

10408089).

5. PASSIFLORA LEPIDOTA Mast., Fl. Bras. 13(1): 547, 581. 1872. TYPE: BRAZIL. Paraná:

“Itararé opp., Morungava praedium,” 30 Jan 1915, P. Dusén 16569 (holotype, F.

Sellow s.n.: B, destroyed; neotype, designated by Svoboda et al. (2016: 109):

MO905635!; isoneotype S08-10927!).

Description—Stems glabrous, slender, terete. Leaves glabrous, bearing lepidote glands abaxially (turning yellow when dried); triangularly 3-lobed; petioles glabrous, bearing few short-stalked glands; stipules glabrous, small, divided, with lepidote glands.

Flowers white; peduncles glabrous; involucral bracts glabrous, bipinnatisect, the ultimate segments gland-tipped with lepidote glands; sepals glabrous, oblong-lanceolate, white; petals white, oblong-lanceolate; corona filaments in four series, radiate; stigmas glabrous; ovary subglobose, glabrous. Fruits glabrous, globose, reddish at maturity; seeds oblong, punctate, brown-black at maturity.

Distribution—Endemic to Brazil, known from the states of Mato Grosso do Sol,

Paraná, and São Paulo.

Habitat—Dry, open fields; elevation 318–1104 m (802 m average).

Phenology—Flowering November–February, fruiting December–February.

Taxonomic Notes—This species is unique in several respects. The bright yellow, lepidote foliar glands (when dried), triangular leaf shape, and its isolated distribution in southeastern Brazil truly set this entity apart. Three collections seen thus far, marked with 188 an asterisk below, bear the distinctive yellow, lepidote glands on the leaves but are also pubescent. These individuals may represent hybrids with a pubescent taxon.

Additional Specimens Examined—BRAZIL. Mato Grosso do Sul: Amambai,

Hatschbach 65519 (MBM), Hatschbach et al. 51532 (MBM); Rio Brilhante, Hatschbach

25034 (MBM). Paraná: Arapoti, Caxambu 1894 (HCF), Hatschbach 20435 (MBM);

Campo Bonito, Dusén 16906 (S, US); Campo Mourão, Hatschbach et al. 15909 (MBM);

Curitiba, Motta & Silva 4586* (RB); Guarapuava, Hatschbach & Ravenna 23088*

(FLOR, MBM, MEXU, NY); Jaguariaíva, Cervi 2999 (MBM, UPCB), Hatschbach

43412 (MBM); Lapa, Hatschbach 1095 (MBM); Palmas, Hatschbach 15459 (MBM);

Palmeira, Cervi 8562 (UPCB), Cordeiro & Hatschbach 499 (MBM, US), Freitas 192

(MBM); Ponta Grossa, Dombrowski 6736 (MBM), Dusén 7263 (S), Gonçalves 55

(UPCB), Hatschbach 2903 (MBM, US), Hatschbach 9616 (L, MBM), Jönson 1071a (S),

Schwartsburd 502 (UPCB); Porto Amazonas, Gurgel 46 (HUEFS, RB), Gurgel 63*

(RB), Gurgel 64 (RB), Gurgel 66 (HBR, RB), Gurgel s.n. (RB); Tibagi, Takeda s.n.

(UPCB); Unknown municipality, Dusén s.n. (S). São Paulo: Bom Sucesso de Itararé,

Souza et al. 4677 (IAC); Iperó, Hoehne & Gehrt 36728 (B); Itapetininga, Lima s.n. (RB),

Mattos 10972 (UPCB); São Paulo, Handro s.n. (IAC).

6. Passiflora maxonii (Killip) H.T.Svoboda, comb. nov. [ined.] Passiflora foetida var.

maxonii Killip (as ‘Maxoni’), Publ. Carnegie Inst. Wash. 461: 326. 1936. TYPE:

NICARAGUA. Managua: [Managua municipality], vicinity of Managua, mostly

along the shores of Lake Managua, 24 Jun 1923, W. R. Maxon, A. D. Harvey & A.

T. Valentine 7219 (holotype: US1180248!; isotypes: US1180249!, US1180250!). 189

Description—Stems softly pilosulous, terete. Leaves subhastate, pilosulous, bearing few, scattered capitate long-stipitate glands abaxially; sub 5-lobed to deeply 5- lobed; petioles pilosulous, bearing several capitate long-stipitate glands; stipules pubescent, filiform segments tipped with capitate glands. Flowers purple; peduncles pilosulous; involucral bracts glabrous, bipinnatisect, the ultimate segments gland-tipped with capitate glands; sepals glabrous, oblong-lanceolate; petals oblong-lanceolate; corona filaments radiate; stigmas glabrous; ovary subglobose, glabrous. Fruits glabrous, globose, red at maturity; seeds oblong, punctate, brown-black at maturity.

Distribution—Found in Central America, in the countries of Costa Rica, El

Salvador, Guatemala, Honduras, and Nicaragua.

Habitat—Dry tropical forests and gallery forests; elevation 4–680 m (100 m average).

Phenology—Flowering and fruiting year-round.

Taxonomic Notes—No other entity in section Dysosmia possesses the suite of traits seen in this taxon (e.g., pubescent, red-fruited, deeply 5-lobed). Because of these unique traits, I have decided to recognize this taxon at the rank of species.

The red-fruited pubescent (RFP) complex in Mexico and Central America, much like the P. ciliata and P. foetida complexes, is comprised of potentially several unique taxa. Passiflora maxonii represents the 5-lobed, hispidulous members of the complex, but it is so far unclear what the correct name for taxa with variously 3-lobed leaves and various vestiture from this region should be. Potential names may include Passiflora foetida var. hastata (Bertol.) Killip (º Passiflora hastata Bertol.), P. foetida var. 190 lanuginosa Killip, P. foetida var. mayarum Killip, P. foetida var. salvadorensis Killip.

While only varieties lanuginosa and mayarum are reported to have red fruit by the publishing author, vegetatively all of the names could be applied to specimens that are observed to have red fruit, though these tend to introgress into one another and may represent more than one RFP taxon. More studies are needed to determine if any of these names represent unique evolutionary species within the RFP complex.

Additional Specimens Examined—COSTA RICA. Guanacaste: Carrillo, Gómez et al. 23063 (DUKE), Gómez P. et al. 23063 (MO); La Cruz, Liesner 2353 (MO). EL

SALVADOR. Ahuachapán: El Refugio, Chinchilla s.n. (MO). San Miguel: Chirilagua,

Tucker 870 (LL, US). San Salvador: San Salvador, Romero s.n. (MO), Standley 22763

(GH). GUATEMALA. Peten: Villa Canales, Contreras 2296 (LL). HONDURAS. Choluteca:

Marcovia, Molina & Molina 31974 (US). NICARAGUA. Carazo: Jinotepe, Estrada 105

(MO). Chinandega: Corinto, Robleto T. 1805 (MO). Chontales: Juigalpa, Stevens &

Montiel 17428 (MO). Granada: Granada, Guzman et al. 530 (MO); Nandaime, Grijalva

P. 2470 (MO). Leon: La Paz Centro, Araquistain 2904 (MO). León: La Paz Centro,

Araquistain & Moreno 1097 (MO); León, Moreno 6561 (MO). Managua: Ciudad

Sandino, Moreno & Henrich 9018 (MO); Managua, Coe 998 (MO), Croat 39008 (MO),

Moreno 990 (MO), Seymour 6281 (MO), Stevens 13172 (MO), Stevens 20468 (MO);

Mateare, Moreno 8575 (MO), Moreno 8589 (MO), Moreno 9083 (MO), Moreno &

Guzmán 615 (MO), Sandino 634 (MO); San Rafael del Sur, Moreno 21493 (MO);

Tipitapa, Moreno & Sandino 7285 (MO), Robbins 6093 (GH, MEXU, MO). Rivas: 191

Belén, Grijalva P. & Medina 2583 (MO); San Juan del Sur, Stevens & Montiel 3773

(MO).

7. Passiflora orinocensis (Killip) H.T.Svoboda, comb. nov. [ined.] Passiflora foetida

subsp. orinocensis Killip in L.H.Bailey, Gentes Herbarum 2: 205. 1930.

Passiflora foetida var. orinocensis (Killip) Feuillet, J. Bot. Res. Inst. Texas 1(1):

144. 2007. Passiflora ciliata var. orinocensis (Killip) Vanderpl., Bot. Mag. 30(4):

356. 2013. TYPE: VENEZUELA. Bolívar: Isla Degrero, on the Orinoco [River], 6

Mar 1921, L. H. Bailey & E. Z. Bailey 1773 (holotype: US1059996!; isotype:

NY00110440!).

Description—Stems terete, glabrous. Leaves glabrous, subhastate, bearing many pyriform long-stipitate glands abaxially; 3-lobed, the middle lobe lanceolate, lateral lobes suborbicular; petioles glabrous, with many glands; stipules small, deeply pinnatisect almost to the stem, glabrous, gland-tipped. Flowers color unknown; peduncles glabrous; involucral bracts glabrous, deeply bi- or tripinnatisect, the ultimate segments gland- tipped; sepals ovate-lanceolate, glabrous; petals ovate-oblong; corona filaments radiate; stigmas glabrous; ovary ovoid, glabrous. Fruits glabrous, globose, red at maturity; seeds obovoid, reticulate, black at maturity.

Distribution—South America, known currently from Bolivia, Guyana, and

Venezuela.

Habitat—River banks and gallery forests; elevation 1–124 m (60 m average).

Phenology—Flowering January–May, fruiting January–May, and August. 192

Taxonomic Notes—This name represents the known glabrous specimens from

South America (except for P. lepidota which is morphologically and geographically unique) and appears to be distinct from those of Central America and the West Indies.

Hence, I am recognizing the taxon at the rank of species.

Additional Specimens Examined—BOLIVIA. Beni: Guayaramerín, Anderson

11900 (NY). GUYANA. Barima-Waini: Hollowell 356 (MO, NY). VENEZUELA. Apure:

Muñoz, Zuloaga et al. 4317 (MO); Pedro Camejo, Castillo et al. 3213 (MO); San

Fernando, Davidse & González 12147A (MO, NY), Davidse & González 13238 (MO).

Barinas: Antonio José de Sucre, Stergios et al. 5769 (MO).

8. PASSIFLORA PECTINATA Griseb., Fl. Brit. W. Ind. 294. 1860. TYPE: TURKS AND

CAICOS. Turks Islands, no date, Hjalmarson s.n. (holotype: GOET009398!;

isotype: K000323253!).

Description—Stems subangular, striate, glabrous. Leaves glabrous, cordate- deltoid, apex obtuse, base deeply cordate, coriaceous, lustrous, bearing few short-stalked glands abaxially, having a mottled appearance when dried; petioles glabrous, with few short-stalked glands; stipules deeply pinnatisect, robust, glabrous, gland-tipped. Flowers pure white; peduncles glabrous; involucral bracts glabrous, once pinnatisect or rarely bipinnatisect, the ultimate segments gland-tipped; sepals linear-lanceolate, glabrous, long-awned; petals linear; corona filaments in several series, radiate; androgynophore elongated; stigmas glabrous; ovary ovoid, glabrous. Fruits glabrous, subglobose, spotted, pink-red at maturity; seeds obovoid, coarsely reticulate, black at maturity. 193

Distribution—Found in the Greater Antilles (Haiti), the Lucayan Archipelago

(Bahamas and Turks and Caicos), and Bermuda.

Habitat—Coastal sand dunes, low scrub or coppices; elevation 0–29 m (7 m average).

Phenology—Flowering and fruiting year-round.

Taxonomic Notes—The lustrous, coriaceous leaves, pectinate involucral bracts, and flowers with an elongated androgynophore, make this species very recognizable and distinct.

Additional Specimens Examined—BAHAMAS. Andros District: North Andros

Island, Svensson & Bremer 713 (S); South Andros Island, Coker 242 (NY). Cat Island

District: Cat Island, Britton & Millspaugh 5926 (NY, US), Byrne 272 (A, FTG, NY),

Hitchcock s.n. (MO); Little San Salvador Island, Britton & Millspaugh 5665 (NY).

Central Abaco District: Great Guana Cay, Britton & Millspaugh 2917 (NY). Crooked

Island District: Crooked Island, Brace 4609 (NY), Correll 44381 (FTG, NY), Gillis

10656 (A, FTG, MO); Long Cay, Brace 4090 (NY), Brace 455 (NY), Correll & Proctor

48811 (FTG, NY), Eggers 3802 (E, L, P), Hitchcock s.n. (MO). Exuma District: Great

Exuma Island, Correll & Correll 42349 (FTG), Correll & Correll 42432 (FTG, NY),

Correll & Correll 42453 (FTG, NY), Correll 40773 (FTG, LL, MO), Eldridge s.n. (MO),

Nickerson & Kessler 2787 (FTG, MO), Nickerson & Semple 2967 (A, FTG, MO), Perry

1591 (DUKE); Hummingbird Cay, Dunn s.n. (GH), Gillis 9361 (A), Morin 4088 (GH);

Little Exuma Island, Correll 40802 (FTG, NY). Inagua District: Great Inagua Island,

Correll 41603 (FTG, MO, NY), Correll 45804 (FTG, NY), Dunbar 220 (A), Gillis & 194

Proctor 11695 (A, MO), Hill 515 (FTG), Proctor & Gillis 33276 (A); Little Inagua

Island, Correll & Haxby 46001 (FTG, MO, NY), Correll 47376 (FTG, NY), Nash &

Taylor 1231 (NY); Unknown island, Austin & Conroy 4723 (FTG), Nash & Taylor 985

(NY), Nash & Taylor 1134 (NY). Long Island District: Cash’s Cay, Cerbin C166 (FTG,

GH, NY); Long Island, Correll 44961 (FTG, NY), Hackett 122 (DUKE, LL), Hill 515b

(A, FTG). Mayaguana District: Mayaguana Island, Freid & Lefaivre 06-064 (FTG).

North Abaco District: Cave Cay, Britton & Millspaugh 2825 (NY). Ragged Island

District: Great Ragged Island, Gillis & Proctor 11966 (A), Wilson 7838 (NY); Ragged

Island, Correll 48381 (FTG, NY). Rum Cay District: Conception Island, Britton &

Millspaugh 5988 (NY), Fairchild 2569 (MO, P); Rum Cay, Correll & Wasshausen

46755 (FTG). San Salvador District: San Salvador Island, Correll 43903 (FTG, NY),

Darlington & Hanley 6 (FTG), Gillis 5264 (DUKE), Gillis 8838 (FTG), Smith s.n.

(FTG), Vincent et al. 9230 (FTG), Wilson 7278 (NY). South Eleuthera District: South

Eleuthera Island, Sauleda & Sauleda 5698 (FTG, MO). BERMUDA. Hamilton: Brown &

Britton 899 (NY), Brown et al. 1134 (NY, US); Hall’s Island, Brown & Britton 873 (MO,

NY). St. George’s: Brown & Britton 974 (NY). HAITI. Nord-Ouest: Île de la Tortue,

Ekman 4158 (A, LL, MO, NY, US), Jestrow et al. 2016-H-016 (FTG), Leonard &

Leonard 15335 (NY), Leonard & Leonard 15336 (MO), Leonard & Leonard 15382 (GH,

US). TURKS AND CAICOS. Caicos Islands: Big Ambergris Cay, Manco 307 (FTG, K,

TCI), Vincent et al. 16334 (FTG); East Caicos Island, Pollard et al. 1343 (FTG); Middle

Caicos Island, Gillis & Proctor 12308 (A), Sadle et al. 312 (FTG); North Caicos Island,

Corcoran et al. 34 (FTG, K, TCI), Correll 43431 (FTG), Gillis & Proctor 12272 (A); 195

Providenciales Island, Correll 43163 (FTG, NY), Correll 46365 (FTG, NY), Gillis 11827

(A, MO), Raven 28190 (MO); South Caicos Island, Burch 4245 (MO, NY), Millspaugh

& Millspaugh 9318 (NY), Proctor 8820 (A), Wilson 7661 (MO, NY); West Caicos

Island, Gillis 12427 (A). Turks Islands: Grand Turk Island, Correll 46589 (FTG), Gillis

11790 (A), Millspaugh & Millspaugh 9033 (NY), Nash & Taylor 3804 (NY); Salt Cay,

Buden 27 (FTG).

Other Observations Examined—BAHAMAS. Cat Island District: Little San

Salvador Island, N. Frade (“noaboa”; iNaturalist ID: 3883047). Exuma District: Great

Exuma Island, N. Frade (“noaboa”; iNaturalist ID: 14301481); Stocking Island, N. Frade

(“noaboa”; iNaturalist ID: 14178861). San Salvador District: San Salvador Island, H.

Svoboda (personal observation, January 2016).

9. PASSIFLORA PSEUDOCILIATA Britton, Bull. Torrey Bot. Club 44: 19. 1917. TYPE: CUBA.

Camagüey: [Camagüey municipality], savannas near Camagüey, 2–7 Apr 1912,

N. L. Britton, E. G. Britton & J. F. Cowell 13155 (holotype: NY00991121!;

isotypes: F459317!, NY00084323!, P00605795!, US718224!).

Passiflora ciliata var. polyadena Griseb., Cat. Pl. Cub.: 285. 1866. Passiflora foetida var.

polyadena (Griseb.) Killip, Publ. Field Mus. Nat. Hist., Bot. Ser. 19: 512. 1938.

TYPE: CUBA. Locality information unknown, 1865, C. Wright s.n. (holotype:

GOET009389!; isotypes: NY00084319!, US44904!).

Passiflora ciliata var. quinqueloba Griseb., Cat. Pl. Cub.: 113. 1866. Passiflora foetida

subvar. quinqueloba (Griseb.) Mast., Trans. Linn. Soc. London 27: 631. 1871.

Passiflora foetida var. quinqueloba (Griseb.) Killip, Publ. Field Mus. Nat. Hist., 196

Bot. Ser. 19: 512. 1938. TYPE: CUBA. “Cuba Orientali,” 1861, C. Wright 2601

(holotype: GOET009390!; isotypes: G00441004!, G00441005!, GH00068022!,

K000323250!, MA607468!, MO2063293!, NY01444432! [left specimen on

sheet], P00605794!, US44905!, YU244604!).

Description—Stems terete, glabrous. Leaves glabrous, bearing some pyriform long-stipitate glands abaxially and marginally; 3- to deeply 5-lobed, the lobe apices subacute to rounded; petioles glabrous, with few to no glands; stipules small, deeply pinnatisect and filiform, glabrous, gland-tipped. Flowers white-violet; peduncles glabrous; involucral bracts glabrous, deeply bi- to tripinnatisect, the ultimate segments thin and gland-tipped; sepals ovate-lanceolate, glabrous; petals ovate-oblong; corona filaments radiate; stigmas glabrous; ovary subglobose, glabrous. Fruits glabrous, globose, scarlet at maturity; seeds oblong, reticulate, black at maturity.

Distribution—Apparently endemic to Cube, found throughout the island.

Habitat—Savannas and palm barrens; elevation 4–748 m (157 m average).

Phenology—Flowering and fruiting year-round.

Taxonomic Notes—This name represents glabrous plants found in Cuba with deeply 3- to 5-lobed leaves, which also consistently have glands only on the abaxial leaf surface. Vegetatively, this species exhibits a wide range of leaf lobing that may represent different developmental stages, as this pattern is also seen in Passiflora pentaschista

(Killip) H.T.Svoboda (see Svoboda & Harris, 2018).

A putative isotype specimen of Passiflora ciliata var. quinqueloba (NY00084319) was found to be inconsistent with the type collection Wright 2601 (see Svoboda et al., 197

2016), and upon further investigation actually matches perfectly with the type collection of P. ciliata var. polyadena. Killip (1938) stated that a specimen of the latter, in fact, existed in the herbarium of the New York Botanical Garden (NY), but has since been

“lost” after his monograph. It appears that the collection number ‘2601’ was added to the label later (sometime between 1938 and present) by an individual who mistakenly thought it represented variety quinqueloba but should actually have remained blank as on the type of variety polyadena (which is an unnumbered [s.n.] collection). Because both varieties are now synonyms of the newly circumscribed Passiflora pseudociliata presented here, the specimen in question unequivocally represents the species despite its convoluted history.

Representative Specimens Examined—CUBA. Camagüey: Camagüey, Britton

2370 (NY); Nuevitas, Shafer 684 (NY), Shafer 850 (NY), Shafer 1101 (NY, US). Ciego de Ávila: Morón, Shafer 2690 (MO, NY), Shafer 2714 (NY, P, US). Cienfuegos: Abreus,

Combs 44 (MO, NY, P); Cienfuegos, Alain 2844 (MO), Britton & Wilson 5698 (NY),

Howard 6234 (GH, NY, P), Jack 5335 (GH, NY, US), Jack 5864 (A), Jack 7144 (A,

NY); Cumanayagua, Britton et al. 5886 (NY), Howard 5609 (GH), Howard 6346 (GH),

Jack 8027 (A, NY); Rodas, Singleton 522 (ECON); Unknown municipality, Howard

5157 (GH), Singleton 402 (ECON). Havana: Habana del Este, León 4125 (NY). Holguín:

Mayarí, Carabia 3786 (NY), Howard 6133 (GH, NY, P, US), Morton & Acuña 2909

(NY), Shafer 3081 (NY), Shafer 3618 (NY). Las Tunas: Jesus Menendez, Kay et al. 228

(MO); Las Tunas, León 15763 (GH); Manatí, Acuña 17103 (NY). Matanzas: Jovellanos,

Alain 1649 (GH); Matanzas, Britton & Wilson 427 (NY); Unknown municipality, León 198

13135 (NY). Mayabeque: Madruga, Britton et al. 675 (NY), León 3340 (NY); Santa Cruz del Norte, Gentry 50912 (MO). Pinar del Rio: Los Palacios, Shafer 11858 (NY). Sancti

Spíritus: Banao, León & Roca 7941 (NY), Luna & León 370 (NY); Sancti Spíritus, Alain

1531 (GH), León 5361 (NY), León 5369 (NY), Shafer 12173 (MO, NY); Trinidad,

Britton et al. 5511 (NY). Villa Clara: Santa Clara, Howard 5028 (GH, NY), Howard et al. 417 (A), León 9296 (NY), Smith et al. 3146 (US), Webster et al. 19 (A); Santo

Domingo, Ekman 13873 (NY). Unknown province: Unknown municipality, León & Roca

8172 (NY).

10. PASSIFLORA SANTIAGANA (Killip) Borhidi, Bot. Közlem. 58: 176. 1971. Passiflora

foetida var. santiagana Killip, Publ. Field Mus. Nat. Hist., Bot. Ser. 19: 491.

Passiflora ciliata var. santiagana (Killip) Vanderpl., Bot. Mag. 30(4): 358. 2013.

TYPE: CUBA. Santiago [de Cuba]: Vicinity of “Santiago City” [Santiago de Cuba],

14 Feb 1902, C. L. Pollard, E. Palmer & W. Palmer 279 (holotype: US403071!;

isotypes: F125781!, GH00112844!, MIN1001947!, PH507792!, NY!,

MO2063283!).

Description—Stems terete, hirsute. Leaves densely yellow lanuginose-hirsute or pannose, bearing some pyriform long-stipitate glands abaxially; deeply 3- to deeply 5- lobed, the lobes oblong; petioles hirsute, with few glands; stipules small, deeply pinnatisect almost to the stem, pubescent, gland-tipped. Flowers white tinged with light violet; peduncles glabrous; involucral bracts hirsute, deeply tripinnatisect, the ultimate segments gland-tipped; sepals oblong-lanceolate, hirsute; petals oblong-lanceolate; corona filaments radiate, magenta (proximal fourth) and pale violet (distal half) with 199 white in between; androgynophore spotted with red; stigmas sparingly hirsute; ovary ovoid, white-pilose. Fruits hirsute-pilose, globose, red at maturity; seeds oblong, reticulate, brown-black at maturity.

Distribution—Endemic to southern Cuba, known from the provinces of

Guantánamo and Santiago de Cuba.

Habitat—Coastal thickets, dry hillsides, and limestone terraces; elevation 0–64 m

(18 m average).

Phenology—Flowering and fruiting July–March.

Taxonomic Notes—Glandular, lanuginose-hirsute, 3- to 5-lobed leaves, and red fruit are an unusual combination of traits. This, and its restricted distribution in just two

Cuban provinces, make the taxon here a unique species. Borhidi (1971) originally changed this taxon to the rank of species (from that of variety by Killip, 1938), but gave no reason for the change in rank.

Additional Specimens Examined—CUBA. Guantánamo: Caimanera, Britton

1891† (NY), Morton & Alain 8843 (DUKE, US); Guantánamo, Britton 1972† (NY, US),

Earle 83 (F, NY); Imías, Michelangeli et al. 1449 (NY), Morton & Alain 8926 (DUKE,

US), Schultes et al. 534 (GH); San Antonio del Sur, Axelrod et al. 10234 (UPRRP),

Morton & Alain 9224 (US); Unknown municipality, Hioram 1887 (NY). Santiago de

Cuba: Santiago de Cuba, Britton & Cowell 12546† (NY), Britton & Cowell 12662† (MO,

NY, US), Britton et al. 12852† (NY), León 3732† (NY, US), León 3934† (NY), León et al. 10555† (NY), Lopez F. 1151 (US), Morton 3319 (US); Unknown municipality,

Clemente 2629 (US), Clemente 6504 (US), Hamilton 151† (NY), Havard 34 (NY), 200

Havard 36 (NY), Linden 1703† (BM, BR, G, GENT, K, L, NY, P). Unknown province:

Unknown municipality, Hioram 1936 (NY).

Other Observations Examined—CUBA. Guantánamo: Caimanera, K. A.

Bakkegard (“kbakkegard”; iNaturalist ID: 9355422), W. Fidler (“wayne_fidler”; iNaturalist IDs: 12584567, 12135487, 12357864).

11. Passiflora subauriculata H.T.Svoboda, sp. nov. [ined.] TYPE: NICARAGUA. North

Caribbean Coast Autonomous Region [formerly part of Zelaya]: Puerto Cabezas

municipality, along road to Panua (entrance ca. 7.6 km NW of Santa Marta), ca.

1.6 km NW of Panua and 5.7 km W from main road, 20 Apr 1978, W. D. Stevens

7758 (holotype: MO2923696!). (See Figure 7-2.)

Description—Stems terete, glabrous. Leaves lamina narrowly lanceolate, glabrous, bearing pyriform long-stipitate glands abaxially, margins ciliate with glands; 3- lobed, the middle lobe extremely long, lanceolate, and acuminate, the lateral lobes much reduced, usually rounded; petioles glabrous, bearing few glands; stipules deeply pinnatisect almost to the stem, glabrous, gland-tipped. Flowers morphology unknown; peduncles glabrous; involucral bracts glabrous, delicate, deeply bipinnatisect, the ultimate segments gland-tipped. Fruits glabrous, globose, red at maturity; seed morphology unknown.

Distribution—Currently known only from northeastern Nicaragua and eastern

Honduras.

Habitat—Habitat unknown; elevation known to be 20–25 m.

Phenology—Known to fruit in February, April, and June. 201

Etymology—The epithet refers to the subauriculate leaf bases that are characteristic of this new species. Other species with a similar leaf shape include

Passiflora aurea, sp. nov. (above) and P. subintegra, comb. nov. (below).

Taxonomic Notes—This species is not well known, as there are only three specimens currently available for study. The leaf shape is quite remarkable in that the middle lobe is extremely long and slender, while the two basal lobes are small and nearly auriculate. These basal lobes appear to sometimes spread out distally, perhaps as the plant ages.

Additional Specimens Examined—HONDURAS. Gracias a Dios: Puerto Lempira,

Barkley & Hernández 40646 (GH). NICARAGUA. North Caribbean Coast Autonomous

Region: Siuna, Stevens 12747 (MEXU).

202

Figure 7-2. Leaf from the proposed holotype of Passiflora subauriculata, sp. nov.

12. Passiflora subintegra (Killip) H.T.Svoboda, comb. nov. [ined.] Passiflora foetida

var. subintegra Killip, Publ. Carnegie Inst. Wash. 461: 328. 1936. TYPE: [BELIZE].

Stann Creek: All Pines, 12 Sep 1930, W. A. Schipp 648 (holotype:

MICH1115896!; isotypes: A!, BM000543035!, F658701!, G0016989!,

MO990078!, NY00073845!, S04-605!).

Description—Stems subterete, wiry, dark, glabrous. Leaves narrowly lanceolate, glabrous, coriaceous, glaucous beneath, margins revolute, bearing ubiquitous white lepidote glands abaxially; often unlobed, sometimes with two small, rounded lateral 203 lobes; petioles glabrous, with few white, lepidote glands; stipules small, inconspicuous, deeply pinnatisect, glabrous, gland-tipped. Flowers pinkish-mauve; peduncles glabrous; involucral bracts glabrous, sparsely bipinnatisect, the ultimate segments gland-tipped; perianth and corona morphology unknown; stigmas glabrous; ovary ovoid, glabrous.

Fruits glabrous, globose, crimson at maturity; seed morphology unknown.

Distribution—Endemic to Belize, occurring in the departments of Stann Creek and Toledo.

Habitat—Savannas, and edges of mangrove swamps; elevation 0–29 m (13 m average).

Phenology—Flowering and fruiting September–March.

Taxonomic Notes—The unusually narrow shape, coriaceous nature, and ubiquitous lepidote abaxial glands of the leaves are absolutely unique among Dysosmia taxa. For this reason, I have chosen to recognize this taxon at the rank of species. Other species with a similar leaf shape include Passiflora aurea, sp. nov. and P. subauriculata, sp. nov. (above).

Two collections seen thus far, marked with an asterisk below, have a scabrous pubescence on the abaxial leaf surface. The type collection and the protologue indicate that the taxon is unambiguously glabrous throughout, suggesting that these unusual specimens may be hybrids with pubescent taxa.

Additional Specimens Examined—BELIZE. Stann Creek: Gentle 8150 (LL, NY),

Gentle 8451 (LL), Proctor 36528 (MO). Toledo: Gentle 3924* (MO, NY), Goodwin &

Lopez 1391* (MO), Goodwin et al. 1318 (MO). 204

13. PASSIFLORA SUBLANCEOLATA (Killip) J.M.MacDougal, Novon 14: 459. 2004.

Passiflora palmeri var. sublanceolata Killip, Publ. Carnegie Inst. Wash. 461: 322.

1936. TYPE: GUATEMALA. Petén: Uaxactun to San Clemente, 30 Apr 1931, H. H.

Bartlett 12788 (holotype: US1492638!; isotypes: CAS48980!, MICH1115898!,

NY02331748!).

Description—Stems terete, tomentose. Leaves sublanceolate, densely tomentose- villose, bearing ubiquitous capitate long-stipitate glands abaxially and marginally; weakly

3-lobed; petioles tomentose, with abundant capitate long-stipitate glands; stipules conspicuous, deeply pinnatisect, tomentose, tips bearing many glands. Flowers bright pink; peduncles tomentose; involucral bracts hirsute, deeply bipinnatisect, the ultimate segments heavily gland-tipped; sepals oblong-lanceolate, reflexed, hirsute; petals oblong- lanceolate, reflexed; corona filaments very short, erect, the outermost series white; androgynophore elongated, spotted with red; stigmas sparingly hirsute; ovary ovoid, white-tomentose. Fruits hirsute-pilose (sometimes very sparingly so), ovoid, scarlet at maturity; seeds oblong, reticulate, black at maturity.

Distribution—Endemic to the Yucatan Peninsula, occurring throughout the states of Campeche, Tabasco, Quintana Roo, and Yucatán (Mexico), the department of Petén in

Guatemala, and the district of Corozal in Belize.

Habitat—Low, often inundated, tropical forests; elevation 8–294 m (131 m average).

Phenology—Flowering and fruiting year-round. 205

Taxonomic Notes—Although the overall shape and glandular nature of the leaves is not entirely unique, the bright pink, reflexed flowers with a short, erect corona set this species apart from all others in section Dysosmia.

Additional Specimens Examined—BELIZE. Corozal: Dwyer 14513 (MO).

GUATEMALA. Petén: Flores, Contreras 1507 (LL, MEXU), Contreras 1729 (MO),

Contreras 478 (LL), Contreras 8394 (LL), Contreras 8497 (LL), Ortíz 391 (US).

MEXICO. Campeche: Calakmul, Bye et al. 12010 (MEXU), Cabrera C. & Canul 5048

(MEXU, MO), Cabrera C. & de Cabrera 11859 (MO), Cabrera C. & de Cabrera 13524

(COL, MEXU), Cabrera C. & de Cabrera 2451 (MEXU, MO), Cabrera C. et al. 8439

(MEXU), Calónico S. et al. 21581 (MEXU, MO), Calónico S. et al. 22709 (MEXU,

MO), Gutiérrez B. 9095 (MEXU), Madrid N. & Arandia 976 (MEXU, MO), Madrid N.

& Gamboa 460 (MEXU), Madrid N. et al. 113 (MEXU, TEX), Madrid N. et al. 433

(MEXU, MO), Madrid N. et al. 790 (MEXU, MO, TEX), Madrid N. et al. 832 (MEXU,

MO), Martínez S. & Álvarez 30853-A (MEXU, TEX), Martínez S. et al. 27244 (COL,

MEXU), Martínez S. et al. 27606 (MEXU), Martínez S. et al. 27801 (MEXU), Martínez

S. et al. 27891-A (MEXU), Martínez S. et al. 29261 (MEXU), Martínez S. et al. 29855

(MEXU), Martínez S. et al. 30501-A (MEXU), Martínez S. et al. 30670 (MEXU),

Martínez S. et al. 35087-B (MEXU), Martínez S. et al. 35088 (MEXU), Martínez S. et al.

35320 (MEXU, MO); Campeche, Gutiérrez B. 5884 (MEXU), Gutiérrez B. 5889

(MEXU), Gutiérrez B. 5922 (MEXU), Gutiérrez B. 6933 (MEXU), Martín 468 (MEXU),

Martín 817 (MEXU), Ramirez R. 161 (MEXU), Webster & Lynch 17521 (DUKE, GH,

MEXU, MO, TEX), Zamora C. & Méndez D. 5090 (MEXU, TEX); Carmen, Saucedo C. 206

45 (MEXU); Champotón, Cabrera C. & de Cabrera 14163 (MEXU), Cabrera C. & de

Cabrera 7080 (MEXU), Carnevali et al. 4474 (NY), Flores & Chan 8706 (MEXU),

Pascual A. et al. 539 (MEXU), Zamora C. & Hernández T. 4340 (TEX), Zamora C. &

Hernández T. 4376 (MEXU, TEX), Zamora C. et al. 4651 (TEX); Escárcega, Bravo H.

1323 (MEXU), Cabrera C. & de Cabrera 10937 (MEXU, MO), Cabrera C. & de

Cabrera 2187 (DUKE, MEXU), Harriman 16007 (US), Hernández X. et al. ES-282

(MEXU), Téllez V. et al. 5608 (MEXU, MO), Webster & Lynch 17713 (MEXU, MO);

Hopelchén, Álvarez & Jiménez 7913 (MEXU), Álvarez & Ramírez 10174 (NY), Álvarez

8646 (MEXU, NY), Álvarez et al. 10321 (MEXU), Álvarez et al. 8852 (MEXU), Álvarez et al. 8910 (NY), Álvarez et al. 9036 (MEXU, NY), Leitner s.n. (MO), Martínez S. &

Álvarez 27411 (MEXU); Unknown municipality, Cabrera C. & de Cabrera 11775

(MEXU, MO). Oaxaca: Heroica Ciudad de Huajuapan de León, Torres C. et al. 784*

(NY); Huautla de Jiménez, Torres C. & Cedillo T. 1498* (TEX). Quintana Roo: Adolfo de la Huerta, Álvarez & Ramírez 8463 (MEXU, NY), Álvarez et al. 9624 (MEXU, NY);

Felipe Carrillo Puerto, Téllez V. & Cabrera C. 1502 (MEXU); José María Morelos,

Gaumer et al. 23671† (NY, US); Othón P. Blanco, Cabrera C. & de Cabrera 1280

(MEXU), Cabrera C. & Torres 1004 (MEXU), Carnevali et al. 4505 (MO, NY), Cowan

3028 (MEXU, MO, NY), Ramírez et al. 1209 (BRIT), Téllez V. & Cabrera C. 2654

(MEXU, MO). Tabasco: Balancán, Menendez et al. 275 (MEXU, MO). Yucatán: Muna,

Tolosa 6 (MEXU); Tekax, Enriquez 341 (MEXU). Unknown state: Unknown municipality, Linden 1367 (GENT). 207

Other Observations Examined—MEXICO. Campeche: Calakmul, M. Álvaro M.

(“miguel-a-m”; iNaturalist ID: 566700); Champotón, A. Dorantes E. (“adorantes”; iNaturalist ID: 1298977).

14. PASSIFLORA URBANIANA Killip, J. Wash. Acad. Sci. 17: 426. 1927. TYPE: CUBA.

Habana: Cultivated, seeds originally from Belize, May 1906, C. F. Baker 2588

(holotype: US529493!; isotypes: B [destroyed], GH [not located], HAC [not

located], NY [not located], P00605815!, PH597801!, UC [not located]).

Description—Stems terete, short villosulous. Leaves lanceolate-oblong, scabrous above, lanulose beneath, bearing few lepidote glands abaxially, subcoriaceous, margins entire, apex rounded, base cordate; weakly 3-lobed; petioles tomentose, lacking glands; stipules very short, inconspicuous, pubescent, the tips bearing glands. Flowers blue- purple; peduncles villosulous; involucral bracts hirsutulous, deeply bipinnatisect, the ultimate segments gland-tipped; sepals oblong, hirsutulous; petals oblong-linear; corona filaments in 5 series, radiate, the outermost series violet proximally; stigmas hirsutulous; ovary subglobose, densely white-tomentose. Fruits hirsutulous, globose, red at maturity; seeds oblong, reticulate, black at maturity.

Distribution—Found throughout Belize, also occurs in the department of Petén,

Guatemala.

Habitat—Pine savannas, edges of mangrove swamps; elevation 0–550 m (72 m average).

Phenology—Flowering March–October, fruiting March–November. 208

Taxonomic Notes—The short, velvety vestiture of the leaves and blue-purple flowers (in addition to the red fruits) represent an uncommon combination of features in section Dysosmia.

Additional Specimens Examined—BELIZE. Belize: Arvigo et al. 354 (LL), Balick et al. 2084 (MO), Croat 23258 (BRIT, MEXU, MO, US), Croat 24057 (MO), Croat

24779 (US), Davidse & Brant 32870 (MO), Davidse & Brant 32942 (MEXU, MO),

Davidse & Brant 33161 (MEXU, MO), Dwyer & Dieckman 9063 (MO), Dwyer 10484

(MO), Dwyer 12428 (GH, MO), Dwyer 12471A (LL, MEXU, MO), Dwyer 14651

(MEXU, MO), Dwyer 14989 (MEXU, MO), Gentle 9407 (LL), Goodwin & Bridgewater

1032 (E), Holmes 4625 (LL), Kay 20 (MO), Lazor & Tyson 2128 (MO), McDaniel 13100

(FSU), Ramamoorthy et al. 3631 (MEXU), Spellman 1534 (MO), Whitefoord 2364

(MO), Whitefoord 2692 (MO), Wiley 87 (MO, US), Worthington 21212 (MO),

Worthington 23831 (BRIT). Cayo: Cowan et al. 5196 (MEXU), Davidse & Holland

36334 (MO), Gentry 7639 (DUKE, MEXU, MO), Nee et al. 46799 (LL, MEXU, MO),

Sutton et al. 171 (MEXU, MO). Orange Walk: Davidse & Brant 32807 (DUKE, MEXU,

MO), Davidse & Brant 32822 (MEXU, MO), Durán et al. 3261 (MEXU), Grether et al.

2228 (MO). Stann Creek: Balick et al. 3049 (MO), Goodwin & Bridgewater 1049 (E),

Goodwin & Lopez 1415 (E, MO). CUBA. Habana: Asima s.n. [cultivated] (P).

GUATEMALA. Petén: Flores, Contreras 7 (LL), Contreras 1261 (LL).

Other Observations Examined—BELIZE. Belize: A. Lindqvist (“annikaml”; iNaturalist ID: 10183754), K. Sasan (“kimberlietx”; iNaturalist ID: 10359728). 209

15. Passiflora wrightiana H.T.Svoboda, sp. nov. [ined.] Passiflora ciliata var. riparia

C.Wright ex Griseb., Cat. Pl. Cub.: 113. 1866. Passiflora foetida var. riparia

(C.Wright ex Griseb.) Killip, Publ. Field Mus. Nat. Hist., Bot. Ser. 19: 510. 1938.

TYPE: CUBA. Locality information unknown, 1860–1864, C. Wright 2602

(holotype: GOET009391!; isotypes: BM000798356!, G00441002!, G00441003!,

GH00068023!, K000323251!, MA607467!, MO2063294!, NY01444433! [right

specimen on sheet], P00604269!, US943566!, YU244603!).

Description—Stems terete, robust, glabrous. Leaves lamina cordate to orbicular, glabrous, bearing numerous capitate long-stipitate glands on both surfaces, margins minutely ciliate with clands, apex acuminate, base cordate; 3-lobed (sometimes weakly so); petioles glabrous, bearing few to many glands; stipules deeply once to bipinnatifid, conspicuous, half leafy, glabrous, the tips bearing glands. Flowers whitish-purple; peduncles glabrous; involucral bracts glabrous, deeply bi- to tripinnatisect, the ultimate segments gland-tipped; sepals oblong, hirsutulous; petals oblong-linear; corona filaments purple proximally, white distally; stigmas glabrous; ovary subglobose, glabrous. Fruits glabrous, globose, red at maturity; seeds obovoid, reticulate, black at maturity.

Distribution—Native to the West Indies, known from Anguilla, the British Virgin

Islands, Cuba, the Dominican Republic, and Haiti.

Habitat—Found on roadside and in lowland thickets; elevation 0–200 m (45 m average).

Phenology—Flowering and fruiting year-round. 210

Taxonomic Notes—Glands on both the abaxial and adaxial surfaces of glabrous, cordate-orbicular leaves represent a one-of-a-kind morphology, and is here recognized as a species. The epithet of the basionym, ‘riparia,’ cannot be used for the name of this taxon at the rank of species because it is not available (it is blocked by the name

Passiflora riparia Mart. ex Mast. of supersection Laurifolia). Thus, I have decided to name this species in honor of the prominent botanist and collector of the type specimen,

Charles Wright (1811–1885). Killip (1938, p. 511) indicated a single collection of this taxon (as P. foetida var. riparia) from the state of Florida, USA, but the specimen in question clearly represents Passiflora ciliata (sensu stricto).

Additional Specimens Examined—ANGUILLA. East End: Howard & Kellogg

19056 (A, NY). BRITISH VIRGIN ISLANDS. Guana Island: Proctor 43454 (NY). Tortola

Island: Fishlock 7 (NY). Virgin Gorda Island: Smith 10598 (A, NY). CUBA. Santiago de

Cuba: San Luis, Pollard & Palmer 351 (MO, NY). DOMINICAN REPUBLIC. Duarte: Villa

Riva, Abbott 548 (NY). El Seibo: Miches, Zanoni et al. 15863 (MO, NY); Unknown municipality, Ekman 12096 (NY). La Altagracia: San Rafael del Yuma, Zanoni et al.

10566 (MO, NY); Unknown municipality, Alain & Liogier 25407 (NY). La Romana:

Cumayasa, Mejía & Zanoni 9365A (FTG, MO, NY); La Romana, Mejía & Zanoni 7693

(FTG, MO, NY), Zanoni & Mejía 17110 (FTG, MO, NY). Puerto Plata: Luperón,

Augusto 640 (NY). Samaná: Samaná, Liogier 14374 (GH, NY, P), Zanoni et al. 29327

(MO, NY); Sanchez, Zanoni et al. 34137 (MO, NY). HAITI. Nord: Plaisance, Nash 598

(NY). Nord-Ouest: Port-de-Paix, Ekman H4289 (US), Leonard & Leonard 11233 (GH,

MO), Leonard & Leonard 14018 (NY). Ouest: Port-au-Prince, Leonard 2772 (NY), 211

Leonard 3569 (NY). SAINT KITTS AND NEVIS. Saint Kitts Island: St. George, Britton &

Cowell 252 (NY), Meagher 4033 (A, MO), Meagher 02/21/09/94 (A).

Excluded Names

16. Passiflora foetida var. parvifolia Killip, Publ. Field Mus. Nat. Hist., Bot. Ser. 19:

501. 1938. TYPE: MEXICO. Guerrero: [Acapulco de Juárez municipality],

Acapulco and vicinity, Oct 1894–Mar 1895, E. Palmer 315 (holotype:

US252833!; isotypes: GH135264!, GH135265!).

Distribution—Confirmed from the state of Guerrero, Mexico.

Taxonomic Notes—The protologue described the fruit of this variety as “reddish”

(Killip, 1938, p. 501). So far there has been no evidence from specimens (even the type), collectors’ notes, or other observations that the plant actually bears red fruit. Based on leaf morphology and its geographic location, this name probably represents young lateral branches (hence the small leaves) of the taxon Passiflora foetida var. acapulcensis Killip, which is very abundant in that area and known to have green fruits. The name is, therefore, excluded from the treatment of red-fruited taxa presented here.

Discussion

The red-fruited members of section Dysosmia have been revised here following the intent of the Unified Species Concept in defining evolutionary species as “separately evolving metapopulation lineages” (de Queiroz, 2007). Species recognized here each have distinct suites of traits that separate them from other known taxa in the section, as discussed within the treatments above. The two complexes treated in-depth in this revision (the P. ciliata and RFP complexes) now have a combined total of 15 recognized 212 species (see Figure 7-3). Expanded studies of each complex are needed in order to more fully understand the diversity of these groups, especially in light of the handful of additional names that do or may represent red-fruited taxa that were not treated here.

Based on the taxonomic revisions presented here, the number of species in the section overall has increased to 31 with five varieties (of Passiflora foetida and P. vesicaria, not treated here).

Figure 7-3. Comparison of the major treatments of the red-fruited members in section Dysosmia. Dashed lines indicate an unchanged taxon, solid lines indicate a different classification, asterisks (*) indicate a new name or taxonomic combination.

213

CHAPTER 8: OVERALL CONCLUSIONS

Section Dysosmia is a complicated conglomerate of species and species complexes that continues to perplex researchers. The sheer amount of morphological diversity in a relatively small group provides an abundance of taxonomic confusion, but also a wealth of discovery.

Throughout my research I have started to chip away at the complexities of the section, as others have done before me, in order to better understand the group. I am making headway in answering old questions by using new techniques, providing clarity where there has been only obscurity, and reevaluating the previously-held assumptions about Dysosmia.

In Chapter 2, I have definitively dealt with the confusing nomenclature in section

Dysosmia by typifying dozens of names and clarifying orthographic errors. This was the first step to understanding to what degree the names in the literature represent the known diversity in the section. The study from Chapter 3 relied on these names as anchors to create a matrix of macro- and micromorphological characters to examine complexes or sublineages in section Dysosmia, taking into account much the taxonomic diversity. The analyses revealed at least two major complexes in the section, which have been the basis for subsequent studies and the recognition of new species.

Chapter 4 provided a new workflow for objectively discriminating taxa using leaf shape. This method led to the discovery of unique phenotypes that, upon closer inspection, represent distinct species based not only on shape descriptors but also micromorphological features. In Chapter 5, I delineated and described species based on 214 environmental variables—a set of tools that have not yet been used for taxonomic purposes in Passiflora.

The utility of micromorphological traits for taxonomic delimitation will, undoubtedly, lead to new discoveries and supply myself and future researchers with additional characters for classifying taxa. Suites of these traits have already been implemented in describing the new species proposed here. Other types of data, including environmental variables, are likely to also elucidate taxonomic boundaries or understand the ecological requirements of different taxa. Likewise, opening the door to geometric morphometrics will allow for greater objectivity and more robust evidence as we attempt to improve our classification of passionflowers. I hope that the stance I have taken here in regard to recognizing unique taxa at the rank of species will be implemented, where appropriate, by other researchers who are struggling with infraspecific names.

A major missing piece of the puzzle remains the lack of molecular phylogenetic information for section Dysosmia. This group may require more advanced techniques

(e.g., phylogenomics, shotgun sequencing, or hybrid probe capture approaches) in order to derive usable markers for this group and eventually fully understand their apparently complex relationships. Questions still remain about section Dysosmia that can really only be answered in the context of a resolved phylogeny; namely the monophyly of the section, the inclusion of the “Dysosmioides” group within its boundaries, and the evolution of red fruit. Future work in Dysosmia, I think, will focus on these questions and likely give rise to more interesting inquiries that will propel a renewed interest in this section. 215

REFERENCES

Abràmoff, M. D., Magalhães, P. J., & Ram, S. J. (2004). Image processing with ImageJ.

Biophotonics International, 11, 36-42.

Aiton, W. (1789) Hortus Kewensis: A catalogue of the plants cultivated in the Royal

Botanic Garden at Kew (Vol. 1). London, England: G. Nicol.

APG IV. (2016). An update of the Angiosperm Phylogeny Group classification for the

orders and families of flowering plants: APG IV. Botanical Journal of the

Linnean Society, 181, 1-20. https://doi.org/10.1111/boj.12385

Bailey, L. H. (1930). South American novelties. Gentes Herbarum, 2, 200-206.

Baker, H. G., Opler, P. A., & Baker, I. (1978). A comparison of the amino acid

complements of floral and extrafloral nectars. Botanical Gazette, 139, 322-332.

Bakun A. (1990). Global climate change and intensification of coastal ocean upwelling.

Science, 247, 198-201.

Barbosa Rodrigues, J. (1891a). Eclogae plantarum novarum. Vellosia, 2(1), 1-88.

Barbosa Rodrigues, J. (1891b). Vellosia, 2(3). Rio de Janeiro, Brazil: Imprensa Nacional.

Bentham, G., & Hooker, J. D. (1862). Genera plantarum: Ad exemplaria imprimis in

Herberiis Kewensibus servata definite (Vol. 1). London, England: A. Black.

Bertoloni, A. (1840). Florula Guatimalensis. Novi Commentarii Academiae Scientiarum

Instituti Bononiensis, 4, 403-443.

Bonhomme, V., Picq, S., Gaucherel, C., & Claude, J. (2014). Momocs: Outline analysis

using R. Journal of Statistical Software, 56, 1-24. 216

Bookstein, F. L. (1991). Morphometric tools for landmark data: Geometry and biology.

Cambridge, England: Cambridge University Press.

Borgmeier, T. (1961). A história da “Flora Fluminensis”: Documentos. Publicações do

Arquivo Nacional, 48, 3-21.

Britton, N. L. (1917). Studies of West Indian plants-IX. Bulletin of the Torrey Botanical

Club, 44, 1-37.

Brouwer, C., Goffeau, A., & Heibloem, M. (1985). Irrigation water management:

Training Manual No. 1 - Introduction to Irrigation. Food and Agriculture

Organization of the United Nations. Retrieved from

http://www.fao.org/docrep/R4082E/R4082E00.htm

Browne, P. (1756). The Civil and Natural History of Jamaica (Vol. 1). London, England:

Osborne & Shipton.

Buzatto, C. R., Singer, R. B., Romero-González, G. A., & Van Den Berg, C. (2011).

Typifications and new synonymies in Capanemia (Orchidaceae, Oncidiinae).

Novon, 21, 28-33. https://doi.org/10.3417/2009058

Carauta, J. P. P. (1973). The text of Vellozo’s Flora Fluminensis and its effective date of

publication. Taxon, 22(2/3), 281-284. https://doi.org/10.2307/1218138

Cavazos, T., & Arriaga-Ramírez, S. (2012). Downscaled climate change scenarios for

Baja California and the North American monsoon during the twenty-first century.

Journal of Climate, 25, 5904-5915.

Cervi, A. C., & Rodrígues, W.A. (2010). Nomenclatural and taxonomic review of

Passifloraceae species illustrated and described by Vellozo in Flora Fluminensis. 217

Acta Botanica Brasilica, 24, 1109-1111. https://doi.org/10.1590/S0102-

33062010000400029

Chitwood, D. H., & Sinha, N. R. (2016). Evolutionary and environmental forces

sculpting leaf development. Current Biology, 26(7), R297-306.

Chitwood, D. H., & Otoni, W. C. (2017a). Morphometric analysis of Passiflora leaves:

The relationship between landmarks of the vasculature and elliptical Fourier

descriptors of the blade. GigaScience, 6, 1-13.

Chitwood, D. H., & Otoni, W. C. (2017b). Divergent leaf shapes among Passiflora

species arise from a shared juvenile morphology. Plant Direct, 1(5), e00028.

Chitwood, D. H., Klein, L. L., O’Hanlon, R., Chacko, S., Greg, M., Kitchen, C., ...

Londo, J. P. (2016). Latent developmental and evolutionary shapes embedded

within the grapevine leaf. New Phytologist, 210, 343-355.

Chodat, R. H. (1902). Plantae Hasslerianae. Bulletin de l’Herbier Boissier, sér. 2(2), 733-

747.

Clark-Tapia, R., & Molina-Freaner, F. (2004). The genetic structure of a columnar cactus

with a disjunct distribution: Stenocereus gummosus in the Sonoran Desert.

Heredity, 90, 443-450.

Clifford, M. R. (2017.) Scents and sense ability: The evolution and role of chemical cues

in the pollination and herbivory of Passiflora (Unpublished doctoral dissertation).

University of Washington, St. Louis, MO.

218

Cope, J. S., Corney, D., Clark, J. Y., Remagnino, P., & Wilkin, P. (2012). Plant species

identification using digital morphometrics: A review. Expert Systems with

Applications, 39, 7562-7573.

Cristofolini, G., Conte, L., & Mossetti, U. (1987). Antonio Bertoloni’s herbarium

(Bologna): A rectification. Taxon, 36, 405–407. https://doi.org/10.2307/1221432 de Brito, A. L. V. T. (2013). Lectotypification of Aristolochia alpestris

(Aristolochiaceae) and Monostychosepalum monanthum (Burmanniaceae): Two

obscure Brazilian species described by João Barbosa Rodrigues (1842-1909).

Harvard Papers in Botany, 18, 237-239. https://doi.org/10.3100/025.018.0216 de Candolle, A. P. (1822). Sur les plantes rares ou nouvelles qui ont fleuri dans le Jardin

de Botanique de Genève pendant les anées 1819, 1820 et 1821. Mémoires de la

Société de Physique et d’Histoire Naturelle de Genève, 1, 431-463. de Candolle, A. P. (1828). Passifloreae. Prodromus Systematis Naturalis Regni

Vegetabilis, 3, 321-338. de Queiroz, K. (2007). Species concepts and species delimitation. Systematic Biology, 56,

879-886.

Doyle, J. J. & Doyle, J. L. (1987). A rapid DNA isolation procedure for small quantities

of fresh leaf tissue. Phytochemical Bulletin, 19, 11-15.

Dryden, I. L. (2017). Shapes: Statistical shape analysis (Version 1.1-8) [R software

package].

Duncan, T. (1983). Antonio Bertoloni’s herbarium and the Florula Guatimalensis types.

Taxon, 32, 299-301. https://doi.org/10.2307/1221990 219

Durkee, L. T. (1982). The floral and extrafloral nectaries of Passiflora. II. The extra-

floral nectary. American Journal of Botany, 69, 1420-1428.

Durkee, L. T. (1983). The ultrastructure of floral and extrafloral nectaries. In B. Bentley

& T. Elias (Eds.), The biology of nectaries (pp. 1-29). New York, NY: Columbia

University Press.

Durkee, L. T., Baird, C. W., & Cohen P. F. (1984). Light and electron microscopy of the

resin glands of Passiflora foetida (Passifloraceae). American Journal of Botany,

71, 596-602.

Dusén, P. (1903). Sur la flore de la Serra do Itatiaya. Arquivos do Museu Nacional do Rio

de Janeiro, 13, 1-119.

Ellison, A. M., Davis, C. C., Calie, P. J., & Naczi, R. F. C. (2014). Pitcher plants

(Sarracenia) provide a 21st-century perspective on infraspecific ranks and

interspecific hybrids: A modest proposal for appropriate recognition and usage.

Systematic Botany, 39, 939-949.

Environmental Systems Research Institute. (2016). ArcMap (Version 10.4.1) [computer

software]. Redlands, CA: Environmental Systems Research Institute, Inc.

FAO/IIASA/ISRIC/ISSCAS/JRC. (2012). Harmonized world soil database (Version 1.2)

[Data file]. Retrieved from http://www.fao.org/soils-portal/soil-survey/soil-maps-

and-databases/harmonized-world-soil-database-v12/en/

Feuillet, C. (2007). Validation of two taxa from northern South America. Journal of the

Botanical Research Institute of Texas, 1, 143-144. 220

Feuillet, C., & MacDougal, J. M. (2003 [2004]). A new infrageneric classification of

Passiflora L. (Passifloraceae). Passiflora: The Journal & Newsletter of Passiflora

Society International, 13(2), 34-38.

Feuillet, C., & MacDougal, J. M. (2006). Passifloraceae. In K. Kubiitzki (Ed.), The

families and genera of vascular plants, 9 (pp. 270-281). New York, NY: Springer.

Fick, S. E., & Hijmans, R. J. (2017). WorldClim 2: New 1-km spatial resolution climate

surfaces for global land areas. International Journal of Climatology, 37, 4302-

4315. https://doi.org/10.1002/joc.5086

Fisher, R. A. (1921). On the probable error of a coefficient of correlation deduced from a

small sample. Metron, 1, 3-32.

Gaertner, J. (1788). De fructibus et seminibus plantarum (Vol. 1). Stuttgart, Germany:

Author.

Garcillán, P. P., González-Abraham, C., & Ezcurra, E. (2012). Phytogeography,

vegetation, and ecological regions. In: J. P. Rebman & N. C. Roberts (Eds.), Baja

California plant field guide (3rd ed.) (pp. 23-24). San Diego, CA: Sunbelt

Publications.

Gardner, G. (1845). Contributions towards a Flora of Brazil. London Journal of Botany,

4, 97-136.

Garrick, R.C., Nason, J. D., Meadows, C. A., & Dyer, R. J. (2009). Not just vicariance:

Phylogeography of a Sonoran Desert euphorb indicates a major role of range

expansion along the Baja peninsula. Molecular Ecology, 18, 1916-1931. 221

Gentry, A. H., & Dodson, C. (1987). Diversity and biogeography of Neotropical vascular

epiphytes. Annals of the Missouri Botanical Garden, 74, 205-233.

Givnish, T. J., Sytsma, K. J., Smith, J. F., & Hahn, W. J. (1995). Molecular evolution,

adaptive radiation, and geographic speciation in Cyanea (Campanulaceae,

Lobelioideae). In W. L. Wagner & V. A. Funk (Eds.), Hawaiian biogeography:

Evolution on a hot spot archipelago (pp. 288-337). Washington, D.C.:

Smithsonian Institution Press.

Gleason, H. A. (1931). Botanical results of the Tyler-Duida expedition. Bulletin of the

Torrey Botanical Club, 58, 405-464.

Goldman, D. H., & MacDougal, J. M. (2015). Passifloraceae. In Flora of North America

Editorial Committee (Eds.), Flora of North America north of Mexico (Vol. 6) (pp.

170-182). Oxford, England: Oxford University Press.

Goldman, D. H. (2003 [2004]). Two species of Passiflora (Passifloraceae) in the Sonoran

Desert and vicinity: A new taxonomic combination and an introduced species in

Arizona. Madroño, 50, 243-264.

Gower, J. C. (1966). Some distance properties of latent root and vector methods used in

multivariate analysis. Biometrika, 53, 325-338.

Graham, A. (1999). Late Cretaceous and Cenozoic history of North American vegetation:

North of Mexico. Oxford, England: Oxford University Press.

Grisebach, A. H. R. (1860). Flora of the British West Indian islands. London, England:

L. Reeve. 222

Grisebach, A. H. R. (1866). Catalogus plantarum Cubensium. Leipzig, Germany: G.

Englemann.

Hamilton, W. (1825). Prodromus plantarum Indiae Occidentalis. London, England:

Treuttel & Würtz.

Hammer, Ø., Harper, D. A. T., & Ryan, P. D. (2001). PAST: Paleontological statistics

software package for education and data analysis. Palaeontologia Electronica

4(1), n.p.

Hansen, A. K., Gilbert, L. E., Simpson, B. B., Downie, S. R., Cervi, A. C., & Jansen, R.

K. (2006). Phylogenetic relationships and chromosome number evolution in

Passiflora. Systematic Botany, 31, 138-150.

Harms, H. A. T. (1922). Neue arten der gattung Passiflora L. Repertorium Specierum

Novarum Regni Vegetabilis, 18, 294-299.

Harms, H. A. T. (1925). Passifloraceae. In A. Engler (Ed.), Die Natürlichen

Pflanzenfamilien (Vol. 2) (pp. 470-507). Leipzig, Germany: W. Engelmann.

Harris, AJ, Walker, C., Dee, J. R., & Palmer, M. W. (2016). Latitudinal trends in genus

richness of vascular plants in the Eocene and Oligocene of North America. Plant

Diversity, 38, 1-9.

Hearn, D. J. (2009). Shape analysis for the automated identification of plants from

images of leaves. Taxon, 58, 934-954.

Hermann, P. (1698). Paradisus batavus. Leiden, Natherlands: Petrum van der Aa.

Hiepko, P. (1987). The collections of the Botanical Museum Berlin-Dahlem (B) and their

history. Englera, 7, 219-252. 223

Holtz, M. (1892). Introduced plants in the Northern Territory. Transactions and

Proceedings of the Royal Society of South Australia, 15, 1-4.

Hooker, W.J. (1839). Curtis’s botanical magazine (Vol. 12) [Botanical magazine (Vol.

65)]. London, England: E. Couchman.

Howard, R.A., Clausen, K.S. & Gillis, W.T., Jr. (1981). William Hamilton (1783–1856)

and the Prodromus Plantarum Indiae Occidentalis (1825). Journal of the Arnold

Arboretum, 62, 211-242. https://doi.org/10.5962/bhl.part.11247 iNaturalist Network. (2018). iNaturalist: A community for naturalists [Website].

Available from http://www.inaturalist.org/

International Plant Names Index. (2018). The International Plant Names Index [Website].

Available from http://www.ipni.org/

Intergovernmental Panel on Climate Change. (2014). Climate change 2014: Synthesis

report. Contribution of Working Groups I, II and III to the fifth Assessment

Report of the Intergovernmental Panel on Climate Change. Geneva, Switzerland:

IPCC. Retrieved from http://www.ipcc.ch/report/ar5/syr/

Ithaka. (2014). JSTOR Global Plants database [Website]. Available from:

http://plants.jstor.org/

Iwata, H., & Ukai, Y. (2002). SHAPE: A computer program package for quantitative

evaluation of biological shapes based on elliptic Fourier descriptors. Journal of

Heredity, 93, 384-385.

Jacquin, J. F. (1844). Eclogae plantarum rariorum aut minus cognitarum (Vol. 2).

Vienna, Austria: Author. 224

Jarvis, C. E. (2007). Order out of chaos: Linnaean plant names and their types. London,

England: Linnean Society of London.

Jáuregui, D., García, M., & Pérez, D. (2002). Morfoanatomía de las glándulas en cuatro

especies de Passiflora L. (Passifloraceae) de Venezuela. Caldasia, 24, 33-40.

Killip, E. P. (1936). Passifloraceae of the Mayan region. Publications of the Carnegie

Institution of Washington, 461, 301-328.

Killip, E. P. (1938). The American species of Passifloraceae. Publications of the Field

Museum of Natural History, botanical series, 19, 1-613.

Klein, L. L, & Svoboda, H. T. (2017). Comprehensive methods for leaf geometric

morphometric analyses. Bio-protocol, 7(9), e2269.

https://doi.org/10.21769/BioProtoc.2269.

Klein, L. L., Caito, M., Chapnick, C., Kitchen, C., O’Hanlon Regan, Chitwood, D. H., &

Miller, A. J. (2017). Digital morphometrics of two North American grapevines

(Vitis: Vitaceae) quantifies leaf variation between species, within species, and

among individuals. Frontiers in Plant Science, 8, 373.

https://doi.org/10.3389/fpls.2017.00373

Krosnick, S. E., Porter-Utley, K. E., MacDougal, J. M., Jørgensen, P. M., & McDade, L.

A. (2013). New insights into the evolution of Passiflora subgenus Decaloba

(Passifloraceae): Phylogenetic relationships and morphological synapomorphies.

Systematic Botany, 38, 692–713.

Kuhl, F. P., & Giardina, C. R. (1982). Elliptic features of a closed contour. Computer

Vision Graphics, 18, 236-258. 225

Kuntze, O. (1891). Revisio generum plantarum (Vol. 1). Leipzig, Germany: A. Felix.

Lamarck, J.-B. (1789). Encyclopédie méthodique, botanique (Vol. 3). Paris, France:

Chez. Panckoucke.

Lemaire, C. (1861). L’illustration horticole (Vol. 8). Ghent, France: Gyselynk &

Gyselynk.

Lemos, R.C.C. de, da Costa Silva, D., & Flavia de Albuquerque Melo-de-Pinna, G.

(2017). A structural review of foliar glands in Passiflora L. (Passifloraceae).

PLoS ONE, 12(11), e0187905. https://doi.org/10.1371/journal.pone.0187905

Linnaeus, C. (1737). Hortus Cliffortianus. Amsterdam, Netherlands: Unknown.

Linnaeus, C. (1753). Species plantarum (Vol. 2). Stockholm, Sweden: L. Salvius.

Linnaeus, C. (1759). Flora Jamaicensis. Uppsala, Sweden: Unknown.

Loddiges, C. (1818). Botanical cabinet (Vol. 2). London, England: Arch & Arch.

Loudon, J.C. (1830). Loudon’s hortus Britannicus: A catalogue of all the plants

indigenous, cultivated in, or introduced to Britain. London, England: Longman,

Rees, Orme, Brown, Green, & Longman.

MacDougal, J. M. (2001). Vascular plants of Arizona: Passifloraceae. Journal of the

Arizona-Nevada Academy of Science, 33, 41-45.

Macfadyen, J. (1850). The Flora of Jamaica: a description of the plants of that island,

arranged according to the natural orders (Vol. 2). London, England: Longman,

Orme, Brown, Green, & Longman. 226

MacMahon, J. A. (2000). Warm deserts. In: M. G. Barbour & W. D. Billings (Eds.),

North American terrestrial vegetation (2nd ed.) (pp. 285-321). Cambridge,

England: University of Cambridge Press.

Masters, M. T. (1871). Contributions to the natural history of the Passifloraceae. The

Transactions of the Linnaean Society of London, 27, 593-645.

Masters, M. T. (1872). Passifloraceae. In C. F. P. de Martius & A. G. Eichler (Eds.),

Flora Brasiliensis, enumeratio plantarum in Brasilia hactenus detectarum (Vol.

13) (pp. 531-628). Leipzig, Germany: F. Fleischer.

McLellan, T., & Endler, J. A. (1998). The relative success of some methods for

measuring and describing the shape of complex objects. Systematic Biology, 47,

264-281.

McNeill, J., Barrie, F. R., Buck, W. R., Demoulin, V., Greuter, D. L., Hawksworth, D. L.,

... Turland, N. J. (2012). International code of nomenclature for algae, fungi and

plants (Melbourne Code): Adopted by the eighteenth International Botanical

Congress, Melbourne, Australia, July 2011. Regnum Vegetabile, 154, 1-274.

Meyer, G. F. W. (1818). Primitiae florae essequeboensis adjectis descriptionibus centum

circiter stirpium novarum, observationibusque criticis. Göttingen, Germany: H.

Dieterich.

Miller, P. (1768). The gardener’s dictionary (Vol. 8). London, England: J. Rivington et

al. 227

Miller, R. J., Mione, T., Phan, H.-L., & Olmstead, R. G. (2011). Color by numbers:

Nuclear gene phylogeny of Jaltomata (Solanaceae), sister genus to Solanum,

supports three clades differing in fruit color. Systematic Botany, 36, 153-162.

Moore, S. (1895). The phanerogamic botany of the Matto Grosso expedition, 1891-92.

Transactions of the Linnean Society of London, Botany, 4, 265-516.

Murdoch, D., Chow, E. D., & Celayeta, J. F. (2007). ellipse: Functions for drawing

ellipses and ellipse-like confidence regions (Version 0.3-5) [R software package].

Muschner, V. C., Lorenz, A. P., Cervi, A. C., Bonatto, S. L., Souza-Chies, T. T., Salzano,

F. M., ... Freitas, L. B. (2003). A first molecular phylogenetic analysis of

Passiflora (Passifloraceae). American Journal of Botany, 90, 1229-1238.

https://doi.org/10.3732/ajb.90.8.1229

Muschner, V. C., Zamberlan, P. M., Bonatto, S. L., & Freitas, L. B. (2012). Phylogeny,

biogeography and divergence times in Passiflora (Passifloraceae). Genetics and

Molecular Biology, 35, 1036-1043.

Nogueria, A., Ottra, J. H. L. el, Guimarães, E., Machado, S. R., & Lohmann, L. G.

(2013). Trichome structure and evolution in Neotropical lianas. Annals of Botany,

112, 1331-1350.

Ocampo Pérez, J., & Coppens d’Eeckenbrugge, G. (2017). Morphological

characterization in the genus Passiflora L.: An approach to understanding its

complex variability. Plant Systematics and Evolution, 303, 531-558. doi:

10.1007/s00606-017-1390-2 228

Pastore, J. F. B. (2013). A review of Vellozo’s name for Polygalaceae in his Flora

Fluminensis. Phytotaxa, 108, 41-48. https://doi.org/10.11646/phytotaxa.108.1.2

Persoon, C. H. (1807). Synopsis Plantarum (Vol. 2). Paris, France: J. G. Cottam.

Plukenet, L. (1691). Phytographia (Vol. 2). London, England: Author.

Plukenet, L. (1696). Almagestum botanicum. London, England: Author.

Porter-Utley, K. (2014). Revision of Passiflora subgenus Decaloba supersection Cieca

(Passifloraceae). Phytokeys, 43, 1-224. https://doi.org/10.3897/phytokeys.43.7804

Preston, F. W. (1960). Time and space and the variation of species. Ecology, 41, 611-627.

Pulle, A. A. (1937). Flora of Suriname (Vol. 3). Amsterdam, Netherlands: Instituut voor

de Tropen.

Qian, H. (1999). Spatial pattern of vascular plant diversity in North America north of

Mexico and its floristic relationship with Eurasia. Annals of Botany, 83, 271-283.

R Core Team. (2017). R: A language and environment for statistical computing (Version

3.4.3) [Software]. Available from https://www.r-project.org/

Radhamani, T. R., Sudarshana, L., & Krishnan R. (1995). Defense and carnivory: Dual

role of bracts in Passiflora foetida. Journal of Biosciences, 20, 657-664.

Rafinesque, C. S. (1836 [1838]). Flora Telluriana (Vol. 4). Philadelphia, PA: H.

Probasco.

Rambaut, A. (2014). FigTree (Version 1.4.3) [computer software]. Available from:

http://tree.bio.ed.ac.uk/software/figtree/ 229

Rebman, J. P, Gibson, J., & Rich, K. (2016). Annotated checklist of the vascular plants of

Baja California, Mexico. Proceedings of the San Diego Society of Natural

History, 45, 1-352.

Reginato, M. (2016). monographaR: An R package to facilitate the production of plant

taxonomic monographs. Brittonia, 68, 212-216.

Reichenbach, H. G. (1828). Conspectus regni vegetabilis per gradus naturales evoluti.

Leipzig, Germany: C. Cnobloch.

Reveal, J. L. (1985). Colonial Maryland plants in D.C. Solander’s “Descriptions of Plants

from Various Parts of the World”-an unpublished 1767 manuscript. Bartonia, 51,

80-92.

Riemann, H., & Ezcurra, E. (2005). Plant endemism and natural protected areas in the

peninsula of Baja California, Mexico. Biological Conservation, 122, 141-150.

Rios, N. E, & Bart, H. L. (2010). GEOLocate (Version 3.22) [Website]. Available from

http://www.museum.tulane.edu/geolocate/

Roemer, M. J. (1846). Familiarum naturalium regni Vvgetabilis synopses

monographicae (Vol. 2). Weimar, Germany: Landes-Industrie-Comptoir.

Rose, J. N. (1892). List of plants collected by Dr. Edward Palmer in 1890 on Carmen

Island. Contributions from the United States National Herbarium, 1, 129-134.

RStudio Team. (2016). RStudio: Integrated development for R (Version 1.0.143)

[Software]. https://www.rstudio.com/products/rstudio/download/

RStudio Team. (2018). RStudio: Integrated development for R (Version 1.1.423)

[Software]. https://www.rstudio.com/products/rstudio/download/ 230

Schlechtendal, D. F. L. von (1853). Plantae Wagenerianae Columbicae. Linnaea, 26,

631-674.

Schneider, C. A., Rasband, W. S., & Eliceiri, K. W. (2012). NIH Image to ImageJ: 25

years of image analysis. Nature Methods, 9, 671-675.

Shaw, J., Lickey, E. B., Beck, J. T., Farmer, S. B., Liu, W., Miller, J., ... Small, R. L.

(2005). The tortoise and the hare II: Relative utility of 21 noncoding chloroplast

DNA sequences for phylogenetic analysis. American Journal of Botany, 92, 142-

166.

Shorthouse, D. P. (2010). SimpleMappr: An online tool to produce publication-quality

point maps [Website]. Available from www.simplemappr.net/

Smith, S. D., Monson, R., & Anderson, J. E. (2012). Physiological ecology of North

American desert plants. New York, NY: Springer.

Souza-Chies, T. T., Yockteng, R., & Nadot, S. (2005). Systématique moléculaire

compare en France et au Brésil du genre Passiflora L. (Passifloraceae). Cahiers

Agricultures, 14, 209-215.

Stafleu, F. A., & Cowan, R. S. (1976). Taxonomic literature: A selective guide to

botanical publications and collections with dates, commentaries and types (Vol.

1) (2nd ed.). Utrecht, Netherlands: Bohn, Scheltema, & Holkema.

Stafleu, F. A., & Cowan, R. S. (1979). Taxonomic literature: A selective guide to

botanical publications and collections with dates, commentaries and types (Vol.

2) (2nd ed.). Utrecht, Netherlands: Bohn, Scheltema, & Holkema. 231

Stafleu, F. A., & Cowan, R. S. (1986). Taxonomic literature: A selective guide to

botanical publications and collections with dates, commentaries and types (Vol.

6) (2nd ed.). Utrecht, Netherlands: Bohn, Scheltema, & Holkema.

Stafleu, F. A., & Cowan, R. S. (1988). Taxonomic literature: A selective guide to

botanical publications and collections with dates, commentaries and types (Vol.

7) (2nd ed.). Utrecht, Netherlands: Bohn, Scheltema, & Holkema.

Sun, Y., Skinner, D. Z., Liang, G. H., & Hulbert, S. H. (1994). Phylogenetic analysis of

Sorghum and related taxa using internal transcribed spacers of nuclear ribosomal

DNA. Theoretical and Applied Genetics, 89, 26-32.

Svoboda, H. T., & Harris, AJ. (2018). Contributions toward understanding the

biodiversity of Passiflora in North America: Updates and a new combination

from the Baja California Peninsula, Mexico and vicinity. Journal of Systematics

and Evolution. https://doi.org/10.1111/jse.12434

Svoboda, H. T. &, Ballard, H. E., Jr. (2018). Phenetic and cladistic studies help clarify

species assemblages in Passiflora section Dysosmia (Passifloraceae). Brittonia,

70, 15-24.

Svoboda, H. T., MacDougal, J. M., & Ballard, H. E., Jr. (2016). Typifications and

nomenclatural notes in Passiflora section Dysosmia (Passifloraceae). Phytotaxa,

288, 101-119.

Swofford, D. L. (2002). PAUP: Phylogenetic analysis using parsimony (Version 4.0)

[Software]. Available from: http://paup.sc.fsu.edu/ 232

Thiers, B. (Continuously updated). Index Herbariorum: A global directory of public

herbaria and associated staff [Website]. Available from

http://www.sweetgum.nybg.org/science/ih/

Thompson, W. (1876). The Garden: An illustrated weekly journal of gardening in all its

branches (Vol. 9). London, England: Unknown.

Thorne, R. F. (1993). Vegetation. In Flora of North America Editorial Committee (Eds.),

Flora of North America north of Mexico (Vol. 1) (pp. 132-153). Oxford, England:

Oxford University Press.

Tournefort, J. P. de (1700). Institutiones rei herbariae (Vol. 1). Paris, France: Unknown.

Trenberth, K. E. (2011). Changes in precipitation with climate change. Climate Research,

47, 123-138.

Triana, J., & Planchon, J.E. (1873). Prodromus florae Novo-Granatensis. Annales des

Sciences Naturelles, Botanique, 5(17), 111-194.

Ulmer, T., & MacDougal, J. M. (2004). Passiflora: Passionflowers of the world.

Portland, OR: Timber Press.

Vanderplank, J. (2013 [2014]). A revision of Passiflora section Dysosmia. Curtis’s

Botanical Magazine, 30, 317-387. https://doi.org/10.1111/curt.12050

Vellozo, J. M. C. (1831 [1827]). Flora fluminensis icones (Vol. 9). Paris, France:

Senefelder.

Vellozo, J. M. C. (1881). Flora fluminensis. Archivos do Museu Nacional do Rio de

Janeiro, 5, 1-467. 233

Villaseñor, J. L. (2016). Checklist of the native vascular plants of Mexico. Revista

Mexicana de Biodiversidad, 87, 559-902.

White, T. J., Bruns, T., Lee, S., & Taylor, J. (1990). Amplification and direct sequencing

of fungal ribosomal RNA genes for phylogenetics. In M. A. Innis, D. H. Gelfand,

J. J. Sninsky & T. J. White (Eds.), PCR protocols: A guide to methods and

applications (pp. 315-322). San Diego, CA: Academic Press.

Wickham, H. (2009). ggplot2: Elegant graphics for data analysis. New York, NY:

Springer.

Wickham, H., & Chang, W. (2015). devtools: Tools to make developing R packages

easier (Version 1.0) [R software package].

Wickham, H., Danenberg, P., & Eugster, M. (2015). roxygen2: In-source documentation

for R (Version 5.1) [R software package].

Wiggins, I. L. (1980). Flora of Baja California. Stanford, CA: Stanford University Press.

Willdenow, C. L. (1800). Caroli a Linné species plantarum (Vol. 3). Berlin, Germany: G.

C. Nauk

Yockteng, R., & Nadot, S. (2004). Phylogenetic relationships among Passiflora species

based on the glutamine synthetase nuclear gene expressed in chloroplast (ncpGS).

Molecular Phylogenetics and Evolution, 31, 379-396.

Zelditch, M. L., Swiderski, D. L., Sheets, H. D., & Fink, W. L. (2004). Geometric

morphometrics for biologists: A primer (1st ed.). Amsterdam, Netherlands:

Elsevier Academic Press.

234

APPENDIX A: PERMISSIONS TO REPRODUCE PUBLISHED ARTICLES

Permission to reproduce the text and figures from the already-published portions of this dissertation was granted, where necessary, by the appropriate copyright holders.

Permission for Chapter 2

Permission to reproduce this work was granted by the Editor-in-Chief of

Phytotaxa, Zhi-Qiang Zhang, via email.

Permission for Chapter 3

This Agreement between Mr. Harlan Svoboda (“You”) and Springer Nature

(“Springer Nature”) consists of your license details and the terms and conditions provided by Springer Nature and Copyright Clearance Center.

License Number 4362740041619

License date Jun 05, 2018

Licensed Content Publisher Springer Nature

Licensed Content Publication Brittonia

Licensed Content Title Phenetic and cladistic studies help clarify species

assemblages in Passiflora section Dysosmia

(Passifloraceae)

Licensed Content Author Harlan T. Svoboda, Harvey E. Ballard

Licensed Content Date Jan 01, 2017

Licensed Content Volume 70

Licensed Content Issue 1

Type of Use Thesis/Dissertation 235

Requestor type Author of this Springer article

Portion Full article

Will you be translating? No

Title A Systematic Revision of Passiflora Section

Dysosmia (Passifloraceae)

Author Harlan T. Svoboda

Institution Ohio University

Permission for Chapter 4

Author retains the following non-exclusive rights to use the Article without further permission after the publication of the Article by Bio-protocol, provided that

Author will reference (and, where applicable, require any third party to reference) the appropriate citation of the publication by Bio-protocol of the Article when exercising any of the rights retained by the Author: (2.b.ii.) Reprint the Article in print format for inclusion in a thesis or dissertation that the Author writes for a degree granted by an educational institution.

Bio-protocol articles deriving from eLife papers are distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use and redistribution provided that the original author and source are credited.

Permission for Chapter 5

This Agreement between Mr. Harlan Svoboda (“You”) and John Wiley and Sons

(“John Wiley and Sons”) consists of your license details and the terms and conditions provided by John Wiley and Sons and Copyright Clearance Center. 236

License Number 4367821442177

License date Jun 14, 2018

Licensed Content Publisher John Wiley and Sons

Licensed Content Publication Journal of Systematics and Evolution

Licensed Content Title Contributions toward understanding the biodiversity

of Passiflora in North America: Updates and a new

combination from the Baja California Peninsula,

Mexico and vicinity

Licensed Content Author Harlan T. Svoboda, AJ Harris

Licensed Content Date Jun 06, 2018

Type of Use Thesis/Dissertation

Requestor type Requestor of this Wiley article

Portion Full article

Will you be translating? No

Title A Systematic Revision of Passiflora Section

Dysosmia (Passifloraceae)

Author Harlan T. Svoboda

Institution Ohio University

237

APPENDIX B: HERBARIUM SPECIMENS AND INATURALIST OBSERVATIONS

USED IN THE STUDY FROM CHAPTER 5

This list of specimens and iNaturalist observations was originally published as supplementary material (Doc. S1 of Svoboda & Harris, 2018) and the formatting has been retained. Herbarium codes follow the Index Herbariorum (Thiers, 2018; http://sweetgum.nybg.org/science/ih/). Specimen lists generated using the R package monographaR (Reginato, 2016).

Accessions Used for the Environmental and Morphometrics Portion of the Study

Samples in this section are organized by the taxonomy used prior to the revision proposed in this study. Specimens marked with a dagger (†) indicate that it was used in the morphometrics portion of the study. Online usernames, in addition to surnames, where possible, are provided in quotation marks for observations gathered from iNaturalist (https://inaturalist.org/).

Passiflora arida (Mast. & Rose) Killip. MEXICO. Baja California: Ensenada,

Harbison 27235 (CAS328002†, SD27235†, CAS31726†), Hill s.n. (SD193123). Baja

California Sur: Loreto, Moran 9286 (SD66357†); Mulegé, Ferris 8632 (CAS0575462†),

Hammerly 99 (CAS295001), Hastings 71-98 (SD78717†), Johnston 3598 (CAS48981†),

Johnston 4200 (CAS48979†), Porter 11065 (HCIB13000, RSA610927†), Rempel 190

(RSA408953†), Sanders 6382 (UCR43153), Wiggins 7948 (CAS0575435), Wiggins

17353 (CAS0575445†); Santa Rosalia, Tenorio L. & Romero T. 12860 (MO5578529).

Sonora: Empalme, Felger & Russell 14662 (ASU0044092†); Guaymas, Daniel 2365

(ASU0044089), Felger et al. 10034 (MEXU799292†), Gomez C. (“usvaldo”; iNaturalist 238

3673467), Gomez C. (“usvaldo”; iNaturalist 5235069), Plagens (“mjplagens”; iNaturalist

7957695), Siegler & Richardson 11706 (MEXU747111, MEXU796053); Hermosillo,

Felger 7039 (MEXU257199), Felger et al. 07-126 (SD199676), Lott & Atkinson 2456

(RSA501797†, MEXU693977†), Moran 28172 (SD105169†), Reina G. & Van Devender

96-633 (MEXU763667, RSA624806†), Wiggins 17198 (MEXU106450, CAS0575430),

Wilder & Turner 07-40 (SD183341†). Passiflora arida var. cerralbensis Killip. MEXIO.

Baja California Sur: Comondú, Carter et al. 2483 (MEXU55092, UC916026†,

CAS0575464†), Domínguez L. 3163 (HCIB17618†, SD157736†), Domínguez L. 3570

(SD153610†, HCIB19329), Gentry 4453 (UC709206†), León de la Luz 559 (SD257902,

HCIB24698), Lung s.n. (UC211934), Thomas 8405 (MEXU707677, CAS0575452,

SD60298†), Wiggins 15218 (CAS0575440†, UC1210636); La Paz, Agúndez 142

(MEXU547114†), Atwood, D. et al. 18979 (MO5563315), “heidip” (iNaturalist

8236297), León de la Luz 3746 (MEXU1336825, HCIB17983), León de la Luz 9587

(HCIB012445), Moran 18980 (MEXU217303, UC1437968, SD100366†), Moran 6894

(SD50539), Moran 9484 (CAS502208†, SD66359), Pérez N. 787 (HCIB009047), Pio-

León (“pioleon”; iNaturalist 2730403), Pio-León (“pioleon”; iNaturalist 4366361), Pio-

León (“pioleon”; iNaturalist 4908100), Pio-León (“pioleon”; iNaturalist 4965172), Pio-

León (“pioleon”; iNaturalist 5145291), Rebman (“jrebman”; iNaturalist 9157694),

Rebman 6994 (HCIB19861, RSA685303, UCR149731†, SD155768†), Rebman et al.

9680 (SD153890†, HCIB20457), Wiggins 14387 (CAS575438†), Wiggins 15619

(CAS0575442†, CAS0575443†), Wiggins 15662 (MEXU107314, CAS0575441,

CAS0575444, MEXU173140†), Wiggins et al. 458 (UC1224942†, MEXU106775†, 239

CAS0575433); Loreto, Carter 4857 (UC1437974, MO2684651, MEXU217289), Carter

5885 (UC1437969†), Jones 27460 (RSA0027099); Los Cabos, Agúndez 327

(MEXU802425†), Knight (“knightericm”; iNaturalist 1748398), Mudie 960 (SD93754),

Wiggins & Ernst 558 (UC1224940†); Mulegé, Daniel & Butterwick 6776

(CAS928258†), Domínguez L. 2457 (SD146829†, HCIB11259), Gentry 3681

(MEXU858619, MO1157423, UC602744), Rebman & Johnson 1056 (RSA530424),

Spjut & Marin 6046 (MEXU336135, MEXU751470), Valov 2004-089

(MEXU1301901†). Sinaloa: Mazatlán, Jones 22345 (RSA0027129†). Sonora:

Bacadéhuachi, Felger 3627 (MEXU848110); Bentio Juárez, Friedman & Davis 007-94

(ASU0044083, MO04649306); Guaymas, Felger et al. 85-1408 (MEXU807845†),

Gallagher et al. 259 (ASU0044087†), Palmer 91 (E00567945†, US0094452†), Sanders et al. 2495 (RSA301457†, ASU0044088†); Hermosillo, Van Devender & Penalba 95-

538 (MO04649275), Wilkinson 159 (MEXU777633†); Unknown municipality, Pringle s.n. (VT224333†). Passiflora arida var. pentaschista Killip. MEXICO. Baja California:

Ensenada, Jones 24966 (RSA0027132†). Baja California Sur: Comondú, Rebman &

Davis 1693 (ASU0044086); La Paz, Breedlove & Axelrod 43116 (CAS513523†), Daniel

2525 (ASU0044090†), Domínguez C. 449 (MEXU536355, HCIB14276, SD126794),

León de la Luz 07-020 (HCIB23058, SD189301†), Moran 7002 (CAS0575453†,

CAS0575454†), Roberts & Roberts s.n. (SD182770†), Sanders et al. 3307

(UCR28363†), Whitehead 866 (CAS0575459), Wiggins 14506-A (CAS0575439); Los

Cabos, Fishbein et al. 3069 (MEXU955082†), Harbison s.n. (SD46335), León de la Luz

11704 (HCIB28171, SD259562), Peters 35 (UC2031055†), Pio-León (“pioleon”; 240 iNaturalist 4023011), Rebman (“jrebman”; iNaturalist 5690417), Rempel 752-37

(RSA408959†). Sonora: Navojoa, Sanders et al. 8981 (RSA507312). Unknown state:

Unknown municipality, Jones 24293 (RSA0027131†). Passiflora foetida L. MEXICO.

Baja California Sur: La Paz, Domínguez L. 92 (HCIB002103), Rebman et al. 5787

(UCR110502, HCIB013899, SD143112), Rebman et al. 27350 (SD237087), Rebman et al. 29082 (SD242034); Los Cabos, Domínguez C. 968 (HCIB063530), Peters 163

(SD176795); Mulegé, Boyd & Ross 5844 (RSA576496), Domínguez L. 2255

(HCIB010233). Passiflora fruticosa Killip. MEXICO. Baja California Sur: Comondú,

Beauchamp 2125 (RSA222027†, SD79277), Flores F. 455 (TEX455494,

MEXU459618), León de la Luz 6126 (HCIB4471, RSA594241†, SD139891), Mason

1919 (CAS146827†), Medel N. (“kardoncito”; iNaturalist 9081152), Medel N. 11-368

(SD227085†, HCIB027451), Medel N. et al. 872 (HCIB028329), Moran 3828

(UC1001524†, CAS0575417), Moran 9563 (CAS502207†), Moran 9581

(MEXU164385†, SD66353†), Rebman (“jrebman”; iNaturalist 5236092), Rebman &

Pérez N. 3492 (UCR112675†, SD140384†), Rebman & Roberts 4803 (UCR112196,

SD144728†), Rebman et al. 1721 (SD137079, ASU0044119), Rebman et al. 2795

(RSA589286, ASU0044120), Wiggins 17781 (CAS0575408†), Wiggins 17793

(CAS0575409†), Wiggins 17810 (CAS0575410†); La Paz, Domínguez C. 1579

(MEXU1363956), Fuentes-Sorario et al. 112 (MO6398007), Johnston 3951

(CAS48065†), Johnston 3978 (CAS48064†), Jones 24766 (RSA27127†, CAS173076†),

León de la Luz 7062 (HCIB023169), León de la Luz 9648 (MEXU1353132), López G.

(“elg56”; iNaturalist 4795607), Maskarinec s.n. (DES00029394), Moran 3712 241

(UC1001571†, CAS0575416†), Moran 9438 (RSA408939, UCR44814†, SD66356†),

Moran 9625 (SD66354†), Rebman & Medel N. 26947 (HCIB000739†), Sousa P. 196

(RSA506330, MEXU838277†), Taylor & Taylor 15731 (MO2242343, US2915014†),

Tenorio L. et al. 11810 (RSA511142, MEXU819034), Wiggins 15227 (CAS0575404†),

Wiggins 16147 (CAS0575406†, MEXU107820†), Wiggins 17660 (MEXU106672†,

CAS0575407), Wiggins 17682 (MEXU106673†), Wiggins 17829 (US3582941,

CAS0575411†), Wiggins 19107 (CAS0575413†), Wiggins et al. 380 (CAS0575401†),

Wiggins et al. 384 (UC1224939†, CAS0575402†, MEXU106774†); Loreto, Domínguez

L. 2155 (HCIB9875, RSA607371†), Moran 9349 (CAS502210, SD66358†). Passiflora palmeri Rose. MEXICO. Baja California: Ensenada, Carter & Kellogg 2958

(UC1095101, GH, CAS0575389, MEXU21259, SD48066), Fuentes-Sorario et al. 145

(MO6397169), Gentry & McGill 23295 (DES00008486, ASU0044144), Harvey 646

(RSA60675†), Henrickson 1132 (RSA182223, GH), Johnston 3500 (A†), Moran 10262

(CAS0575379, MEXU110219, SD54601), Moran 12363 (SD65144), Moran 12882

(RSA193840†, CAS0575377, SD65145), Porter & Machen 11450 (RSA743892,

SD208201), Rebman et al. 3126 (DES00041145, ASU0044137, HCIB8206, RSA609668,

SD139523†), Vanderplank et al. 5128 (SD246121), Vinton s.n. (SD222701); Mexicali,

Hastings 71-24 (SD78646†), Johnston 3406 (CAS48022†, GH†), Johnston 3536

(UC251985†, GH†), Moran 7208 (SD49645), Moran 8589 (SD61552), Moran 10438

(SD54133†), Rebman & Hirales L. 2088 (ARIZ328033, ASU0044168), Rebman &

Shepard-Espinoza 5896 (UCR112251, SD146438), Rebman et al. 8836 (ASU0044140,

DES00042988, UCR115739†, SD145007, SD202947†), Roos s.n. (MO3663335, 242

UCR24217†, RSA450213, RSA512551), Tenorio L. & Romero T. 10953 (LL00455496,

RSA494898, MEXU831241), Tenorio L. & Romero T. 9477 (MEXU493872), Wiggins

16998 (CAS0575360), Wiggins 17282 (MEXU106452†, CAS0575361†), Wilder &

Bowen 10-349 (UCR236091, ARIZ; USON), Wiggins & Wiggins 15772 (TEX00455500,

GH, CAS0575368†, MEXU106557†). Baja California Sur: La Paz, Domínguez C. 4337

(HCIB030024†), Johnston s.n. (A†), León de la Luz 10485 (MEXU1377509,

SD189302†), León de la Luz 7821 (SD148768†), Vinton s.n. (SD222702†), Wiggins

13024 (UC1009301, GH, RSA97831, CAS0575367†, SD47451†), Wiggins 16856

(CAS0575359†); Loreto, Carter & Ferris 3745 (UC1437970, MEXU217270†, GH†),

Carter & Kellogg 3154 (LL00455498, UC1095100†, CAS0575390†), Carter &

Sharsmith 4939 (UC1437973, GH, MEXU217262, SD100365†), Carter 4327

(TEX00455501, UC1437971†, MO2684650), Carter 5977 (UC1593900), Collins et al.

234 (US1531357, A), Dawson 6303 (RSA408967), Domínguez C. 1140 (BRIT),

Domínguez L. 3411 (SD160060, SD151801†), Fuerte O. 165 (SD176794†), Fuerte O.

188 (UC1593985), Harder & Appleby 1060 (MO4338463, RSA553003), Hastings 71-

129 (SD78686†), Johnston 3759 (CAS48029†), Johnston 3823 (CAS48030†), Johnston

3882 (US1209259†, CAS48032†), León de la Luz 10192 (MEXU1340878†,

HCIB018768, SD151744†), León de la Luz 9127 (HCIB18032), Moran 9085

(SD66387†), Moran 9182 (SD66388†), Moran 9223 (SD66389†, CAS502206), “reilhag”

(iNaturalist 7238531), Rebman (“jrebman”; iNaturalist 6538188), Rebman et al. 5704

(UCR110480, SD143111†), Roberts & Roberts 4490 (SD164435), Tenorio L. et al.

11845 (MEXU831098, RSA510795), Tenorio L. et al. 11867 (MO3719902†, 243

RSA511144, MEXU830824), Wiggins 17462 (CAS0575362); Mulegé, Baker & Johnson

8720 (ASU0044105†, SD154162), Boyd & Ross 5963 (MEXU624567, RSA573583,

UCR84426, CAS792247), Chambers 772 (CAS0575421†, MEXU21260, SD46845),

Daniel 205 (ASU0044139†, CAS681950), Douglas 2230 (DUKE397636), Ferris 8632A

(CAS0575385†), Gershenzon et al. s.n. (TEX00455502), Hammerly 424 (CAS295003†,

GH; CAS0575395), Hastings 71-93 (SD78782), Holler s.n. (DES00023944), Johnston

3604 (CAS48023†), Johnston 3659 (CAS48027†), Johnston 3721 (UC251983†), León de la Luz et al. 6141 (MO6280375, GH01147926†), León de la Luz 9316 (SD160061),

Leuenberger et al. 2999 (B100272021), Lynch 1111 (BRIT, NY), Merello & Brunner

321 (TEX00457234, COL521510, MO5578492†, MEXU1051508†), Miller et al. 7327

(MO5578493†), Moran 3990 (CAS0575378), Moran 8938 (RSA207425, SD66386†),

Porter & Machen 11067 (RSA610929), Rebman 4213 (SD142386†), Rebman & Davis

1722 (ASU0044142†, SD137078), Rempel 195 (RSA408968†), Seigler & Richardson

11731 (MEXU747061), Tenorio L. et al. 11875 (RSA511143, MEXU831161†), Thomas

7967 (CAS0575382, MEXU107163), Thomas 7969 (MEXU107678†), Van Devender et al. 91-455 (MO5161099†), Webster & Lynch 19608 (DUKE281148, MO2878372, GH†,

MEXU287390†, SD109080†), Wiggins 17310 (MEXU106461†), Wiggins & Wiggins

17977 (CAS0575370†), Wiggins & Wiggins 18242 (CAS443512†, CAS0575372†),

Wiggins & Wiggins 18273 (CAS444279†), Zippin 45 (SD132133†); Santa Rosalia,

Tenorio L. & Romero T. 12856 (MO5578527†). Sonora: Hermosillo, Felger & Cooper

15315 (MEXU296402†), Felger & Cooper 15485 (MEXU257196), Johnston 3167

(CAS48033†, GH), Moran 8745A (MEXU164386, SD66390), Tenorio L. et al. 9535 244

(RSA501086†, MEXU819073†), Wiggins 17135 (CAS0575363, MEXU106453), Wilder

& Bertelsen 06-86 (ARIZ379495†, ASU0044081†, SD191685), Wilder et al. 06-17

(SD191686), Wilder et al. 07-450a (SD199677), Wilder et al. 08-147 (BRIT).

Additional Specimens Examined

In addition to the specimens cited above, these specimens were examined for morphological and/or phenological data. Specimens in this section are organized following the taxonomy accepted in this paper.

Passiflora arida. MEXICO. Baja California: Mexicali, Lindsay s.n. (CAS). Baja

California Sur: La Paz, Johnston 3544 (CAS, UC); Mulegé, Tenorio L. et al. 11881

(MEXU, RSA). Sonora: Guaymas, J. N. Rose 1206 (GH, NY, US; type of P. foetida L. var. arida Mast. & Rose), Jones 22961 (RSA); Hermosillo, Johnston 3200 (CAS),

Johnston 4402 (CAS). Unknown state: Unknown municipality, Harbison 27247 (SD).

Passiflora foetida. MEXICO. Baja California Sur: Los Cabos, Bailey 225 (US),

Brandegee s.n. (SD); Mulegé, Johnston 3660 (CAS, US), Nelson & Goldman 7194 (US).

Passiflora fruticosa. MEXICO. Baja California Sur: Comondú, Brandegee s.n. (UC), J.

N. Rose 16285 (NY, US; type of P. fruticosa); La Paz, Brandegee s.n. (GH), Bryant s.n.

(UC), Moran 9445 (RSA, SD), Wiggins 19086 (CAS). Passiflora palmeri. MEXICO.

Baja California: Ensenada, Burgess & Van Devender 7633 (MO), Marin 71-43 (SD);

Mexicali, Copp 110 (CAS), Johnston 3397 (CAS, GH, UC), Lott & Atkinson 2495 (CAS,

MEXU), Moran 4098 (CAS, UC), Valiente B. & Chiang C. 593 (MEXU, RSA);

Unknown municipality, Hallé 7826 (P). Baja California Sur: Comondú, Simpson & Neff

03-16-80-5 (TEX); Loreto, Carter & Reese 4547 (UC), E. Palmer 868 (A, F, GH, K, NY, 245

S, UC, US; type of P. palmeri), Johnston 3848 (CAS), León de la Luz 7573 (SD),

Stephenson s.n. (VT), Tenorio L. et al. 11866 (MEXU, RSA), Unknown 11 (MEXU),

Unknown 33 (MEXU); Mulegé, Berry 107 (CAS), Rempel 175 (RSA), Rose 16691 (NY),

Sousa P. 281 (MEXU, RSA), Wiggins 5451 (CAS, GH, NY, RSA, UC), Wiggins &

Wiggins 18113 (CAS, MEXU); Santa Rosalia, Spjut & Marin 6047 (CAS, MEXU, US).

Sonora: Hermosillo, Lott & Atkinson 2520 (MEXU), Moran 13029 (SD), Wilder &

Turner 07-111 (SD). Unknown state: Unknown municipality, Diguet s.n. (P), Whitehead

765 (CAS). Passiflora pentaschista (Killip) H.T.Svoboda, comb. nov. MEXICO. Baja

California: Ensenada, Jones 24294 (RSA). Baja California Sur: La Paz, I. M. Johnston

4043 (A, CAS, GH, K, NY, UC, US; type of P. arida var. cerralbensis), Johnston 3069

(CAS), Jones s.n. (RSA), Porter 125 (CAS, MEXU, TEX, UC), Rebman et al. 3649 (SD,

UCR); Loreto, Fuerte O. 127 (SD, UC), Fuerte O. 185 (UC), Wiggins 17625 (CAS,

MEXU); Los Cabos, A. W. Anthony 333 (F, GH, MO, US; type of P. arida var. pentaschista), Agúndez 607 (MEXU), Gander 9738 (SD), Grabendörfer s.n. (UC),

Harbison s.n. (SD), Howell 10609 (CAS), Peters s.n. (UC), Unknown 44999 (MEXU),

Webster & Murphey 24352 (MEXU, SD); Unknown municipality, Peters s.n. (UC).

Sinaloa: Mazatlán, Bravo H. 203-5372 (MEXU), Gonzalez O. 7230 (MEXU), Jones s.n.

(RSA). Sonora: Hermosillo, Felger & Bezy 14012 (MO). Unknown state: Unknown municipality, Diguet s.n. (P). UNITED STATES OF AMERICA. Arizona: Pima,

Goldman 2092 (MO, US), Spellenberg & Zucker 12959 (MO, NMC). Incertae sedis specimens. MEXICO. Baja California: Ensenada, Faulkner s.n. (SD), Rebman 17339

(SD). Baja California Sur: Comondú, Rebman 31169 (SD), Vanderplank et al. 5037 246

(SD); La Paz, Rebman & Medel N. 26947 (SD); Mulegé, De Groot et al. 4887 (CAS,

RSA, SD), Moran 21375 (SD), Rebman 5131 (HCIB, SD, UCR), Valov 1030 (SD), Valov

2010-039 (SD).

247

APPENDIX C: SAMPLES USED IN THE MOLECULAR STUDY

The samples used in the molecular portion of this study are cited here, with geographic information and herbarium accession listed.

Taxon Collection Country State/Province Specimen Accession ampullacea Jørgensen 61434 Ecuador Azuay MO04645924 arida Tenorio L. & Romero T. 12860 Mexico Baja California Sur MO5578529 arizonica Goldman 2120 USA Arizona MO5749728 bahamensis Goldman & Rice 3221 Bahamas Andros Island US3518016 boticarioana Pena s.n. Brazil Minas Gerais BHCB156369 chrysophylla Vanni et al. 2251 Paraguay Boquerón MO4004577 ciliata (complex) Rosales 974 El Salvador Ahuachapán MEXU1177270 ciliata (complex) Abraham P. 121 Honduras Garcias a Dios MO3449911 ciliata (complex) Hanan A. et al. 940 Mexico Veracruz TEX00261763 ciliata (complex) Balick et al. 3589 Belize Toledo GH ciliata (complex) Alcorn 3385 Mexico San Luis Potosí TEX00157712 ciliata (complex) Reveal et al. 7335 Venezuela Zelaya NY ciliata (complex) Cabrera & Cabrera 3528 Mexico Quintana Roo MEXU348693 ciliata (complex) Stevens 7809 Nicaragua Atlántico Norte MO2923672 ciliata var. santiagana Michelangeli et al. 1449 Cuba Guantánamo NY01163956 clathrata Esteves et al. 15496 Brazil Minas Gerais NY00955169 248

Appendix C continued

foetida (complex) Fuentes P. 37 Mexico Veracruz MEXU827000 foetida (complex) Reyes G. 542 Mexico Chiapas MEXU665709 foetida var. isthmia Espina 2485 Colombia Choco MO3861180 foetida var. maxonii Grijalva 2470 Nicaragua Granada MO3215398 foetida var. moritziana Cremers & Bastos 12900 French Guiana Saint-Laurent-du- US3357132 Maroni foetida var. nigelliflora Krapovickas & Cristóbal 46385 Argentina Formosa GH foetida var. oaxacana Miller; Tenorio L. 497 Mexico Oaxaca MO3446908 foetida var. parvifolia Lott et al. 4154 Mexico Jalisco TEX00157745 foetida var. subintegra Proctor 36528 Belize Stann Creek MO2604244 hibiscifolia Gould & Williams 137 Mexico Oaxaca TEX00157716 holosericea Benítez de Rojas et al. 5166 Venezuela Carabobo MO5621816 hypoglauca Mota 2922 Brazil Minas Gerais BHCB008754 lepidota Cordeiro & Hatschbach 499 Brazil Paraná US3232940 maliformis Ramon 240 [11706] Vanuatu Shefa P menispermifolia Galdames 6398 Panama Bocas del Toro MO6421738 nephrodes Delanoy 263 Bolivia La Paz MO6147647 palmeri León de la Luz et al. 6141 Mexico Baja California Sur GH01147926 pectinata Svoboda 2 Bahamas San Salvador Island n/a pentaschista Goldman 2092 USA Arizona US3405669

249

Appendix C continued

RFP Carr 30042 USA Texas TEX00464207 RFP Castillo C. & Zamora C. 6715 Mexico Guerrero MEXU732162 RFP Cowan 3325 Mexico Tabasco MO2921420 setulosa Cervi & Dunaiski 3251 Brazil Paraná NY01550460 sublanceolata Madrid N. et al. 433 Mexico Campeche MO5678901 trisecta Galiano et al. 4556 Peru Anta MO5900098 urbaniana Goodwin & Lopez 1415 Belize Belize MO6313278 vellozoi Milward-de-Azevedo et al. 132 Brazil Rio de Janeiro RB475520 vesicaria var. galapagensis Bentley 3 Ecuador Galápagos US3021369 vestita Rojas & Ortiz 6087 Peru Pasco MO6241033 villosa Ratter et al. 3607 Brazil Federal District MO4372191 YFP Vigo & Graham 16525 Peru Padre Abad MO6147266 Abbreviations: RFP, red-fruited pubescent; YFP, yellow-fruited pubescent.

250

APPENDIX D: CONDITIONS USED IN THE PCR TRIALS

Conditions used for the optimization of PCR for the samples of Dysosmia taxa used in this study.

Optimiza- Primer Set Denaturing Annealing Extension Number tion Trial Step Step Step of Cycles 1 ITS17SE- 1 m at 94˚ C 1 m at 50˚ C 2 m at 72˚ C 35 26SE 2 ITS17SE- 1 m at 94˚ C 1 m at 53˚ C 2 m at 72˚ C 35 26SE 3 ITS17SE- 1 m at 94˚ C 30 s at 53˚ C 2 m at 72˚ C 35 26SE 4 ITS17SE- 1 m at 94˚ C 30 s at 55˚ C 2 m at 72˚ C 35 26SE 5 ITS17SE- 1 m at 94˚ C 30 s at 60˚ C 2 m at 72˚ C 25 26SE 6 ITS17SE- 1 m at 94˚ C 30 s at 58˚ C 2 m at 72˚ C 35 26SE 7 ITS17SE- 30 s at 94˚ C 30 s at 59˚ C 30 s at 72˚ C 35 26SE 8 ITS17SE- 30 s at 94˚ C 30 s at 60˚ C 30 s at 72˚ C 35 26SE 8 trnS-trnfM 30 s at 94˚ C 30 s at 60˚ C 30 s at 72˚ C 35 9 ITS17SE- 30 s at 94˚ C 30 s at 60˚ C 30 s at 72˚ C 30 26SE 10 ITS17SE- 30 s at 94˚ C 30 s at 60˚ C 30 s at 72˚ C 27 26SE 10 ITS17SE- 30 s at 94˚ C 30 s at 55˚ C 30 s at 72˚ C 35 26SE 11 ITS17SE- 30 s at 94˚ C 30 s at 60˚ C 30 s at 72˚ C 26 26SE 12 ITS17SE- 30 s at 94˚ C 30 s at 60˚ C 30 s at 72˚ C 26 26SE 13 ITS17SE- 20 s at 95˚ C 15 s at 63˚ C 30 s at 72˚ C 35 26SE 14 ncpGS839- 20 s at 95˚ C 15 s at 57˚ C 30 s at 72˚ C 35 1056 14 psbZ- 20 s at 95˚ C 15 s at 57˚ C 30 s at 72˚ C 35 5'trnG2S 15 ITS17SE- 20 s at 95˚ C 15 s at 65˚ C 30 s at 72˚ C 35 26SE 16 ncpGS839- 1 m at 94˚ C 30 s at 55˚ C 1.5 m at 72˚ C 35 1056 17 ITS17SE- 20 s at 95˚ C 15 s at 63˚ C 30 s at 72˚ C 30 26SE 251

Appendix D continued

18 ITS17SE- 20 s at 95˚ C 15 s at 64˚ C 30 s at 72˚ C 31 26SE 19 ITS17SE- 20 s at 95˚ C 15 s at 64˚ C 30 s at 72˚ C 33 26SE 20 ncpGS839- 20 s at 95˚ C 15 s at 53˚ C 30 s at 72˚ C 35 1056 21 ITS17SE- 20 s at 95˚ C 15 s at 56˚ C 30 s at 72˚ C 35 26SE 21 ncpGS839- 20 s at 95˚ C 15 s at 56˚ C 30 s at 72˚ C 35 1056 22 psbZ- 20 s at 95˚ C 15 s at 55˚ C 30 s at 72˚ C 35 5'trnG2S 24 ncpGS839- 30 s at 94˚ C 30 s at 55˚ C 30 s at 72˚ C 35 1056 25 ncpGS839- 30 s at 94˚ C 45 s at 50˚ C 30 s at 72˚ C 35 1056 26 trnS-trnfM 30 s at 94˚ C 30 s at 60˚ C 30 s at 72˚ C 35 27 ncpGS839- 1 m at 94˚ C 45 s at 53˚ C 1.5 m at 72˚ C 35 1056 27 psbZ- 1 m at 94˚ C 45 s at 53˚ C 1.5 m at 72˚ C 35 5'trnG2S 29 ncpGS839- 1 m at 94˚ C 45 s at 55˚ C 1.5 m at 72˚ C 35 1056 30 ncpGS839- 1 m at 94˚ C 45 s at 55˚ C 1.5 m at 72˚ C 40 1056 32 ncpGS839- 1 m at 94˚ C 45 s at 55˚ C 1.5 m at 72˚ C 35 1056 35 ncpGS839- 1 m at 94˚ C 45 s at 55˚ C 1.5 m at 72˚ C 43 1056 36 ncpGS839- 1 m at 94˚ C 45 s at 53˚ C 1.5 m at 72˚ C 43 1056 37 ITS17SE- 20 s at 95˚ C 15 s at 64˚ C 30 s at 72˚ C 31 26SE 38 ITS17SE- 20 s at 95˚ C 15 s at 60˚ C 30 s at 72˚ C 32 26SE 39 ncpGS839- 1 m at 94˚ C 45 s at 51˚ C 1.5 m at 72˚ C 40 1056 40 trnS-trnfM 45 s at 94˚C 45 s at 60˚ C 1.5 m at 72˚ C 35 41 ITS17SE- 20 s at 95˚ C 15 s at 63˚ C 30 s at 72˚ C 32 26SE 42 ITS17SE- 20 s at 95˚ C 15 s at 63˚ C 30 s at 72˚ C 33 26SE Abbreviations: s, seconds; m, minutes. ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

Thesis and Dissertation Services ! !