The Evolution of Angiosperms

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/311541611

Systematics and Evolution of Menyanthaceae and the Floating-Leaved Genus Nymphoides

Article · January 2010

CITATIONS READS 0 212

1 author:

Nicholas P. Tippery University of Wisconsin - Whitewater

92 PUBLICATIONS 645 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Aquatic plant systematics View project

Molecular taxonomy of Hyrcanian Alnus using nuclear ribosomal ITS and chloroplast trnH-psbA DNA barcode markers View project

All content following this page was uploaded by Nicholas P. Tippery on 11 July 2017.

The user has requested enhancement of the downloaded file. Systematics and Evolution of Menyanthaceae and the Floating-Leaved Genus Nymphoides

Nicholas Peter Tippery, Ph.D.

University of Connecticut, 2010

Menyanthaceae (70 species) are a family of aquatic and wetland plants that occur worldwide. This diverse group contains different growth habits (emergent and floating-leaved), reproductive systems (heterostyly, homostyly, gynodioecy and dioecy), floral and seed morphologies, and inflorescence architectures. In this study, I have evaluated the phylogenetic relationships, taxonomy, biogeography, and morphological character evolution for over half of the approximately 50 Nymphoides species, plus all species in related genera within the family.

In Chapter 1 I investigated generic relationships across Menyanthaceae and found that the contemporary circumscription of Villarsia included three paraphyletic lineages that graded toward a monophyletic Nymphoides. Biogeographical reconstruction supported an Australian origin for the family and also for all of the major clades, with dispersal events corresponding to the boreal sister taxa Menyanthes and Nephrophyllidium, and the South African Villarsia clade.

Chapter 2 is a study of the genus Nymphoides that examined all Australian species of the genus and synthesized for the first time the morphological data for all the species worldwide.

Morphological data analysis indicated several relationships, including the grouping of species with similar inflorescence habits. Molecular phylogenetic analyses supported a similar division, but species resolution was thoroughly incongruent on trees derived from nuclear (nrITS) or chloroplast (matK/trnK) data, suggesting widespread hybridization during the diversification of the genus.

In Chapter 3 I studied the Nymphoides inflorescence architecture, which comprises three types: expanded (pairs of flowers separated by internodes), condensed (single floating leaves each supporting a cluster of flowers), and a morphology unique to N. peltata (pairs of leaves supporting clusters of flowers). I determined that these quite different inflorescence types likely were derived from a common blueprint, from which they deviate only by their relative elongation of internodes or expansion of bracts into foliage leaves.

In Chapter 4 I examined the life history and reproductive potential of N. peltata, a

Eurasian native that is naturalized in North America. Populations of N. peltata are able to produce abundant fruits with highly germinable seeds, but the plants are genetically identical throughout their naturalized range, possibly the result of inbreeding while in cultivation.

Nicholas Peter Tippery – University of Connecticut, 2010

Systematics and Evolution of Menyanthaceae and the Floating-Leaved Genus Nymphoides

Nicholas Peter Tippery

B.A., University of Dallas, 2000

A Dissertation

Submitted in Partial Fulfillment of the

Requirements for the Degree of

Doctor of Philosophy

at the

University of Connecticut

2010

Nicholas Peter Tippery

2010

APPROVAL PAGE

Doctor of Philosophy Dissertation

Systematics and Evolution of Menyanthaceae and the Floating-Leaved Genus Nymphoides

Presented by

Nicholas Peter Tippery, B.A.

Major Advisor ______

Donald H. Les

Associate Advisor ______

Gregory J. Anderson

Associate Advisor ______

Cynthia S. Jones

University of Connecticut

2010

Dedicated to my father, Kevin Tippery, who introduced me to the beauty and perpetual wonder of nature, who taught me to persevere through adversity, who encouraged me to be active and enjoy life, and who always supported me in whatever I did.

iii

ACKNOWLEDGEMENTS

I could not have gotten even close to the point of finishing my Ph.D. without the continuing support of so many people. I am grateful foremost to my advisor Don Les, who was my guide through the graduate school experience, who taught me by example and by gentle correction, who provided me with laboratory resources whenever I could not support myself, and who generously included me on his research to help me become well published. I thank also my advisory commitee, who helped to round out my education with their phenomenal experience in their respective areas of expertise. Greg Anderson, whose friendliness knows no equal, was always encouraging and supportive, and he was a valuable resource for helping me think through pollination biology and reproductive system evolutionary concerns. Cindi Jones, who knows all the inner workings of plant anatomy and morphology, was very generous with her time and resources in allowing me to add the morphological component to my dissertation, and her enthusiasm continues to inspire me.

I had a number of professional acquatintances who helped to further my development as a scientist. I could not thank Surrey Jacobs enough for sharing two months of his life with me and instructing me in everything from Australian customs to good plant pressing technique. Surrey took my poorly conceived plan to collect Nymphoides across Australia and turned it into a workable project that became rousingly successful. Surrey and Betty Jacobs made Australia a second home for me, and I am forever grateful to have met and shared time with them. Surrey‘s friends and colleagues also helped greatly to make our trip more successful and enjoyable. I thank especially John and Marion Clarkson for opening their home to us, and I was privileged also to meet Helen Aston, Janice and Roger Carolin, Geoff Sainty, Lance Smith, and Karen

Wilson while in Australia. Walter Pagels was a steady source of plant material, knowlege and inquiry about Menyanthaceae, and I continue to try to repay him for helping to get me started. I relied heavily on the information gained from herbarium loans, for which I thank the curators of

the CONN herbarium, Andrew Doran and Bob Capers, and the directors and staff of the many herbaria that kindly sent loans for my research.

The graduate students of the UConn EEB Department were my constant support, both emotionally and academically. I enjoyed having made so many kind friends here, and I hope to carry them with me through the rest of my life. I could scarcely name them all, and I deeply regret that I cannot list everyone here, but some people who were particularly helpful include Lori

Benoit, Jessica Budke, Jane Carlson, Karolina Fučíková, Laura Forrest, Geert Goemans, Chris

Martine, Hilary McManus, Michael Moody, Brigid O‘Donnell, Rachel Prunier, Krissa Skogen,

Frank Smith, Kathryn Theiss, Dan Vanderpool, Juan Carlos Villarreal, Amy Weiss, and Norm

Wickett. Without these and so many other inspirational people, I would have abandoned hope long ago.

In addition to fellow EEB graduate students, I had many friends help me through their personal support, many of whom only tangentially knew what sort of research I was pursuing.

These include Dan and Mary Bernier, Russ Bird, Kate and Matt Gilmore, Jeff Kotz, Kate

O‘Sullivan, Nicole and Pete Palumbo, Marissa Prosser, Anna and Scott Russell, Emily Simmons, and Liz Werle. I thank them for keeping me grounded and helping me to have fun in the midst of my most serious times.

My family has always been there to support me, and they helped me to become the kind of person who could successfully finish a Ph.D. My mom Kitty has always been the rock of our family, keeping us on the straight and narrow, motivating us to be better people, and always presenting a positive outlook on life. My sister Janie is always optimistic and supportive, and there is nothing she wouldn‘t do to help out a friend. My brother Chris helped to raise me by his example, and he was the constant companion of my youth whom I could always rely on to laugh at my jokes. In our adult years, he has continued to support me, and I treasure the time I have been able to spend with him, his wife Allison, and their lovely children.

During the majority of my time at UConn, I have been privileged to have the love and companionship of my girlfriend Megan Johnston. She has helped me to see the end goal through the whole long process of graduate school, and she has made my experience vastly more enjoyable. Megan has supported my role as a graduate student far and above what should be expected of any partner, and at times she was an even stronger motivator than I was for my own work. Her patient support was really one of the major reasons why I was able to finish.

TABLE OF CONTENTS

Chapter 1: Systematics of Menyanthaceae

Abstract...... 1

Introduction...... 1

Materials and Methods...... 2

Results...... 4

Discussion...... 6

Literature Cited...... 10

Appendices...... 12

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Abstract...... 16

Introduction...... 17

Materials and Methods...... 21

Results...... 30

Discussion...... 33

Literature Cited...... 40

Tables...... 46

Appendices...... 52

Figure Captions...... 61

Figures...... 63

Chapter 3: Evolution of inflorescence morphology in Nymphoides

Abstract...... 69

Introduction...... 70

Materials and Methods...... 74

Results...... 76

vii

Discussion...... 82

Literature Cited...... 89

Figure Captions...... 90

Figures...... 93

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

Abstract...... 106

Introduction...... 107

Materials and Methods...... 115

Results...... 118

Discussion...... 122

Literature Cited...... 128

Tables...... 132

Appendices...... 136

Figure Captions...... 137

Figures...... 139

viii

TABLE AND APPENDIX INDEX

TABLE 1.01 – Statistics for molecular and morphological data and phylogenetic

analyses of Menyanthaceae species...... 4

APPENDIX 1.01 – List of voucher specimens for Menyanthaceae molecular data...... 12

APPENDIX 1.02 – Non-molecular characters and states scored for Menyanthaceae...... 13

APPENDIX 1.03 – Matrix of non-molecular data for Menyanthaceae...... 14

APPENDIX 1.04 – GenBank accession data for Menyanthaceae outgroup taxa...... 15

TABLE 2.01 – Matrix of non-molecular data for Nymphoides species...... 46

APPENDIX 2.01 – List of voucher specimens for Nymphoides molecular data...... 52

APPENDIX 2.02 – Non-molecular characters and states scored for Nymphoides...... 60

TABLE 4.01 – Populations of N. peltata studied in North America...... 132

TABLE 4.02 – Alleles obtained from microsatellite analysis of N. peltata populations...... 134

APPENDIX 4.01 – Locality and voucher information for N. peltata plants collected

outside of the U.S.A. or from cultivated sources...... 136

FIGURE INDEX

FIG. 1.01 –Phylogeny of Menyanthaceae constructed using non-molecular data...... 4

FIG. 1.02 – Phylogeny of Menyanthaceae constructed using nrITS data...... 5

FIG. 1.03 – Phylogeny of Menyanthaceae constructed using matK/trnK data...... 5

FIG. 1.04 – Phylogeny of Menyanthaceae constructed using rbcL data...... 6

FIG. 1.05 – Phylogeny of Menyanthaceae constructed using combined molecular data...... 7

FIG. 1.06 – Estimated relative and calibrated ages for major Menyanthaceae clades...... 8

FIG. 2.01 – Phylogenetic relationships among Menyanthaceae genera...... 63

FIG. 2.02 –Phylogeny of Nymphoides constructed using non-molecular data...... 64

FIG. 2.03 – Full-data phylogeny of Nymphoides constructed using nrITS data...... 65

FIG. 2.04 – Full-data phylogeny of Nymphoides constructed using matK/trnK data...... 66

FIG. 2.05 – Trimmed-data phylogeny of Nymphoides comparing topology of nrITS and

matK/trnK data trees...... 67

FIG. 2.06 –Ancestral state reconstructions for heterostyly and inflorescence morphology...... 68

FIG 3.01 – Schematic depiction of the overall Nymphoides inflorescence morphology

for the three types of inflorescence architecture...... 93

FIG. 3.02 – Illustrated inflorescence cross-sections and diagrammatic interpretations

for N. aurantiaca and N. peltata...... 94

FIG. 3.03 – Cross-section, diagram, and dissected bud of N. aurantiaca inflorescence

(expanded type)...... 95

FIG. 3.04 – Cross-section and diagram of N. crenata inflorescence, and dissected bud

of N. exiliflora inflorescence (both expanded type)...... 96

FIG. 3.05 – Cross-section and diagram of N. indica inflorescence (condensed type)...... 97

FIG. 3.06 – Cross-section, diagram, and dissected bud of N. aquatica inflorescence

(condensed type)...... 98

FIG. 3.07 – Cross-section, diagram, and dissected bud of N. peltata inflorescence

(peltata type)...... 99

FIG. 3.08 – Further dissected inflorescence bud of N. peltata (peltata type)...... 100

FIG. 3.09 – Dissected bud of N. indica inflorescence (conensed type) showing floral and

inflorescence continuation axes...... 101

FIG. 3.10 – Dissected inflorescence continuation axis of N. indica (condensed type)...... 102

FIG. 3.11 – Simplified, generalized models of inflorescence organization in Nymphoides...... 103

FIG. 3.12 – Elaboration of the simplified Nymphoides inflorescence models by expansion

of floral and inflorescence continuation axes...... 104

FIG. 3.13 – Dissected inflorescence bud of Ornduffia umbricola...... 105

FIG. 4.01 – Study populations of N. peltata in North America...... 139

FIG. 4.02 – Photograph of Apis mellifera with N. peltata pollen...... 140

FIG. 4.03 – Chart of hourly pollen accumulation for a naturalized N. peltata population...... 141

FIG. 4.04 – Seed set per fruit in studied N. peltata populations...... 142

FIG. 4.05 – Percent seed germination per fruit in studied N. peltata populations...... 143

FIG. 4.06 – Plot of percent seed germination against seed number per fruit for studied

N. peltata populations...... 144

FIG. 4.07 – Plot of percent seed germination against seed color for studied N. peltata

populations...... 145

FIG. 4.08 – Genetic distance tree for native, cultivated, and naturalized N. peltata...... 146

Chapter 1: Systematics of Menyanthaceae

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Chapter 2

Phylogenetic relationships and morphological evolution in Nymphoides

ABSTRACT

The predominantly floating-leaved aquatic plant genus Nymphoides consists of approximately 50 species that occur worldwide and are considerably variable with respect to their growth habit, inflorescence architecture, and traits related to their vegetative, floral, and seed morphologies. I undertook a phylogenetic study using both morphological and molecular data in order to evaluate relationships among species and to infer the evolution of two characters of interest, heterostyly and inflorescence architecture. Phylogenetic analysis of morphological data resolved several clades of species, but with poor nodal support. Molecular data from nuclear

(nrITS) and chloroplast (matK/trnK) DNA sequences were significantly incongruent regarding the placement of many clades and species. Two major clades were resolved consistently, including the clade sister to N. minima and the clade comprised of these two lineages. The widespread incongruency of Nymphoides species was interpreted to represent hybrid origins of species, some of which retain conflicting data from their parental lineages in the chloroplast

(maternally inherited) and nucleus (biparentally inherited, but able to resemble one parent exclusively by gene conversion). Ancestral character state reconstruction indicated that heterostyly was widespread among ancestral species, from which homostyly has evolved on multiple occassions. The ancestral sexual condition for the dioecious species N. aquatica and N. cordata could not be ascertained with confidence, owing to incomplete taxon sampling, poor resolution of crown clades, and incongruence between nuclear and chloroplast data. Ancestral

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides state reconstruction also indicated that the expanded inflorescence morphology found in several

Australian and tropical Asian Nymphoides represents the ancestral condition, from which a condensed morphology (found in most Nymphoides worldwide) evolved at least twice independently, and from which the unique N. peltata morphology also arose.

INTRODUCTION

Aquatic plants represent a morphologically and phylogenetically diverse category of tracheophytes, in which multiple terrestrial lineages have independently recolonized aquatic habitats (Cook 1999). Plants that grow in water have evolved many adaptations to aquatic environments and exist in a variety of growth forms (e.g., Sculthorpe 1967). One such growth form is the floating-leaved habit, in which plants are rooted in the substrate underwater but have leaves that float on the water surface. This habit conspicuously characterizes the water lily family

(Nymphaeaceae), but it also occurs sporadically in groups of monocots (e.g., Potamogeton) and dicots (e.g., Persicaria; Cook 1996). Nymphoides (literally, ―resembling Nymphaea‖), a genus in the aquatic plant family Menyanthaceae, is one of the most diverse and widespread groups of floating-leaved aquatic plants (Cook 1996; Kadereit 2007). Species of Nymphoides are found worldwide; several have broad distributions (e.g., N. indica), but many are restricted to a single continent or even to a local geographic region (Raynal 1974b; Sivarajan and Joseph 1993; Aston

1973, 1982). Although a number of regional taxonomic treatments exist for Nymphoides, there has been no recent comprehensive treatment of the genus worldwide.

With only two exceptions (N. cambodiana and N. exiliflora), Nymphoides species produce floating leaves that arise from a submersed rhizome anchored in the substrate. Floating leaves in Nymphoides support their entomophilous flowers and help to keep them above water.

Inflorescences are either expanded, having internodes that elongate between pairs of flowers, or condensed, with dense flower clusters supported by a single floating leaf. With few exceptions,

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Nymphoides flowers are pentamerous, with yellow or white petals that may be glabrous or variously ornamented with hairs or laciniate wings. Flowers are borne above water, but fruit development of the irregularly dehiscent capsule occurs underwater. Morphological characters like flower appearance and the form and number of seeds vary greatly among species, but other features like the inflorescence architecture and fruit structure are less variable across taxa.

Sexual condition. One of the more intriguing floral characteristics of Nymphoides is the presence of dimorphic heterostyly, a sexual condition characterized by spatial separation of anthers and stigma that occurs reciprocally in different plants (i.e., the anthers of one morph type are at the height of stigmas in the other morph type). In addition to the conspicuous stamen and pistil dimorphism, heterostylous plants frequently have corresponding differences in pollen and stigma micromorphology, and a diallelic, sporophytic self-incompatibility mechanism that prevents plants from selfing or interbreeding with plants of the same morph type (Ganders 1979).

Heterostyly promotes outcrossing by presenting pollen at a different height from the stigma on the same flower, which also is the same height as the receptive stigma on a compatible flower of the opposite morph type (Barrett and Shore 2008). Heterostylous species occur along with non- heterostylous ones in a number of unrelated lineages, and the phylogenetic distribution of heterostyly supports the inference of multiple, independent origins (Ganders 1979; Barrett 1992;

Barrett and Shore 2008; Cohen 2010). In Menyanthaceae, phylogenetic analyses reconstructed heterostyly as the ancestral condition, with several non-heterostylous lineages having lost the syndrome independently (Chapter 1; Tippery et al. 2008).

Four Nymphoides species (N. aquatica, N. cordata, N. krishnakesara, N. macrosperma) exhibit a further modification of their sexual system with dioecy, i.e., the condition where flowers are unisexual and a single individual bears flowers of only one sex (Ornduff 1966; Sivarajan and

Joseph 1993). Whereas heterostyly promotes outcrossing, dioecy absolutely requires fertilization to be between genetically different individuals. Still another condition is present in one

Nymphoides species (N. cristata) that is gynodioecious, i.e., with individual plants bearing either

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides female or hermaphroditic flowers (Vasudevan Nair 1973). Dioecy and gynodioecy are derived conditions in Nymphoides (Chapter 1; Tippery et al. 2008), and their presence in an ancestrally heterostylous group offers the opportunity to compare evolutionary patterns of the different sexual conditions, and to address the question of why such systems (i.e., dioecy and gynodioecy) might evolve in the presence of other systems (i.e., heterostyly) that already promote outcrossing.

Vegetative (i.e., clonal) reproduction is an additional factor that potentially affects sexual system evolution, particularly in Nymphoides where vegetative clones can grow to occupy entire water bodies or portions thereof (Takagawa et al. 2006; Larson 2007). The various means of clonal propagation in Nymphoides include fragmentation (e.g., N. cristata; Vasudevan Nair

1973), stoloniferous growth (e.g., N. indica; Raynal 1974a), and the dispersal of floating propagules that have pre-formed roots (e.g., N. aquatica; Godfrey and Wooten 1981). Vegetative reproduction further complicates the evolution of sexual condition, because it can increase the likelihood of inbreeding (e.g., matings between two ramets of the same clone) or provide plants with an alternate method of propagation in the absence of suitable mating partners (e.g., if plants are self-incompatible and without appropriate mating options or dioecious and clonal for a single sex). In small or genetically impoverished populations with greater inbreeding depression, clonally-reproducing plants are more likely to evolve systems that promote outcrossing

(Jacquemyn and Honnay 2008). Such mechanisms include self-incompatibility (possibly with associated heterostyly) and conditions like gynodioecy and dioecy.

In Nymphoides, where heterostyly represents the ancestral condition, self-incompatibility in theory should be widespread also, in which case there would be no need to evolve any further mechanism to promote outcrossing. Nonetheless, both dioecious and gynodioecious species exist, and the sexual and ecological conditions of their closest relatives could be greatly informative regarding the pathway by which dioecy evolved in the genus. The evolution of dioecy from heterostyly has been proposed to occur through an increase in ‗femaleness‘ or

‗maleness‘ of the different floral morphs (Lloyd 1979). Ornduff (1966) postulated that dioecy

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides had evolved from heterostyly in Nymphoides, based on the prevalence of the latter condition.

However, there also are numerous non-heterostylous species that have lost the floral dimorphism and presumably the associated self-incompatibility. Characterizing the various aspects of heterostyly requires thorough research (e.g., crossing experiments to determine self- incompatibility). For example, in the homostylous species N. geminata the evolutionary loss of heterostyly also corresponded with an increase in vegetative propagation and self-compatibility relative to the related, heterostylous species N. montana (Haddadchi 2008). Considering the occasional presence of non-heterostylous species in Nymphoides, and the largely untested distribution of self-incompatibility, it is possible that dioecy evolved from a self-compatible (e.g.,

> 90% seed set from selfing relative to outcrossing) or weakly self-incompatible (10-90% seed set) ancestor. Furthermore, even a strongly self-incompatible species (< 10% seed set) can produce seeds by autogamy, albeit at a low frequency (e.g., N. peltata; Wang et al. 2005; see

Chapter 4). Consequently, inbreeding depression in a highly clonal species could be sufficiently adaptive to drive further evolution toward dioecy.

Hypotheses regarding the evolution of sexual condition, inflorescence habit, floral and seed morphology, along with biogeography and taxonomy, can be addressed using phylogenetic methods. In the following study I conducted a molecular phylogenetic analysis of the genus

Nymphoides in order to determine the relationships among species, their biogeographical relationships, and to evaluate various hypotheses of character evolution. Specifically, I addressed the following questions: 1) whether species with the two different inflorescence types (i.e., expanded or condensed) would resolve in separate clades, and if the anomalous N. peltata (which differs from all other Nymphoides by inflorescence, fruit, and seed characters) would resolve as the sister to the rest of the genus, which has been inferred by its taxonomic placement in a separate section, 2) whether characters related to flower and seed morphology would correlate with recognizable phylogenetic groups, 3) whether any pattern would emerge that correlates with the geographic distribution of Nymphoides species worldwide, for example indicating that there

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides were relatively few colonization events in areas like Africa and the Americas, where species endemism is high, 4) whether dioecious Nymphoides species evolved from a heterostylous or non-heterostylous ancestor, and whether gynodioecy could be implicated as an intermediate condition in the evolution of dioecy. By answering these questions, I hoped to gain better insight into evolutionary patterns in Nymphoides and the relative adaptive value of different morphologies both independently and in combination.

MATERIALS AND METHODS

Taxon sampling. Species of Nymphoides were studied using both morphological and molecular data. Morphological data were obtained from literature sources (Aston 1973, 1982,

1984, 1986, 1987, 1997, 2002, 2003; Raynal 1974b; Sivarajan and Joseph 1993) and examination of preserved specimens (in ethanol or dried) or live plants. Two species were omitted that until recently were considered part of Nymphoides: N. exigua (F.Muell.) Kuntze, which Tippery and

Les (2009) transferred to Liparophyllum (as L. exiguum), and N. stygia (J.M.Black) H.Eichler, a rare and possibly extinct species that Aston (2009) declared a nomen ambiguum and recommended be considered a species of Villarsia (with affinity to species of Villarsia now belonging to Ornduffia Tippery & Les; Tippery and Les 2009). Other taxa of questionable status, which some authors have synonymized (e.g., Nymphoides cristata / N. hydrophylla, N. elegans /

N. moratiana, N. humboldtiana / N. indica / N. thunbergiana; Ornduff 1966; Klackenberg 1990), were coded separately and provisionally retained as distinct species. Using molecular data, I analyzed 136 accessions of 31 Nymphoides species, including previously reported nucleotide sequences for ten taxa (Tippery et al. 2008; Tippery and Les 2009; Tippery et al. 2009) that were retrieved from GenBank (Appendix 2.01). Specimens were identified using relevant literature sources (see above).

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Morphological data. Fifty-four species of Nymphoides were scored for 28 vegetative and reproductive features (Appendix 2.02), which commonly have been used to distinguish species. Morphological data were verified against specimens where possible, otherwise they were obtained only from literature sources. Morphological data were compiled for each taxon, and in total 54 taxa were coded (Table 2.01). Seven taxa (N. flaccida, N. hastata, N. lungtanensis, N. microphylla, N. minor, N. siamensis, and N. verrucosa) were excluded from the phylogenetic analysis of morphological data (see below) because insufficient data were available to distinguish them. The following discussion of morphological variation in Nymphoides assesses the relative taxonomic value of each character.

Duration. Menyanthaceae have a base chromosome number of x = 9 (Tippery et al.

2008), and the Nymphoides clade includes diploid, triploid, tetraploid, and hexaploid cytotypes

(character 1, Appendix 2.02). Species have been described as having either an annual or perennial habit (character 2), and the evidence for this distinction often is derived from a simple comparison of the rhizome length. Plants that persist for a number of years generate substantial rhizomes (Raynal 1974a), whereas those of annual plants are much shorter. The value of this trait is diminished somewhat by the probability that duration depends on the habitat, with species treated as ‗annual‘ perhaps being unable to persist vegetatively through the dry season. Species described as annuals occur predominantly in the tropics, in places where water bodies tend to be only seasonally wet and remain inhospitable to aquatic plants during other seasons.

Habit. Nearly all Nymphoides have floating leaves that support their inflorescence

(character 3). This morphology differs from species in related genera that have facultatively (e.g.,

Ornduffia albiflora (F.Muell.) Tippery & Les, O. reniformis (R.Br.) Tippery & Les) or obligately

(O. submersa (Aston) Tippery & Les) floating leaves that never support the inflorescence. The inflorescence remains erect in species with facultatively floating leaves, and in O. submersa the flowers actually rest on the water surface, being borne on a lax inflorescence that originates from the rhizome separately from the floating leaves. Two Nymphoides species (N. cambodiana and

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

N. exiliflora) notably deviate from the floating-leaved habit of the genus, and in these the leaves also possess truncate or acute leaf bases in lieu of the typical cordate shape (Tippery et al. 2009).

Exceptions can be observed in individual plants of various Nymphoides species, which occasionally take on a somewhat erect habit when they become stranded after water levels recede

(Tippery and Les 2009).

Sexual condition. Species of Menyanthaceae are predominantly heterostylous (character

4), although several species exhibit homostyly, dioecy, or gynodioecy (see Introduction).

Dioecious species were coded as ambiguous for heterostyly, but the gynodioecious species

Nymphoides cristata, whose hermaphroditic flowers resemble the short-styled flowers of heterostylous species (Vasudevan Nair 1973), was coded as exhibiting heterostyly. Dioecious and gynodioecious species were coded as having unisexual flowers (character 5).

Leaves. The leaves of Nymphoides vary among species in size, shape, thickness, and their position relative to the inflorescence. Leaf diameter ranges from less than 1 cm in N. furculifolia to over 20 cm in N. indica (character 6), with shapes that are orbicular or elliptical

(length/width greater than 1.15). The rearward-facing lobes of the typically cordate or reniform leaves either overlap or diverge in a sinus at up to a 90° angle; in some species the lobes impart a furculiform (e.g., N. furculifolia) shape to the leaf. Leaf texture may be smooth to deeply roughened (especially on the underside, e.g., N. aquatica), on leaves that are membranous to several millimeters thick. In relation to the inflorescence, leaves that originate directly from the rhizome (i.e., ‗sterile‘ or ‗vegetative‘ leaves; Raynal 1974a) often differ from those that arise from within the inflorescence itself (i.e., ‗fertile‘ leaves). The most conspicuous instance of this dimorphism occurs in species with the condensed inflorescence type (discussed below), in which the fertile leaves each subtend a flower cluster and have only a short petiole (< 2 cm), and the combined leaf and inflorescence are borne on a long, flexuous stem that resembles the petiole.

Some authors have distinguished species that produce both sterile and fertile leaves from those that produce only fertile leaves (e.g., Raynal 1974b); however, as the production of sterile leaves

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides may relate to ontogeny (e.g., Raynal 1974a; van der Velde et al. 1979), this character should be evaluated with caution. A third type of leaf also produced by some Nymphoides species is the strictly submersed leaf, which is constitutive in some taxa. Submersed leaves depart from the cordate morphology of floating leaves, tending to be spathulate in form, also lacking the thickness of the former. The presence of submersed leaves also should be considered relative to the ontogeny of the plant, as the first leaves produced by an underwater seedling often are of this type

(e.g., Vasudevan 1968; Aston 1982, 1984).

Inflorescence. There are two major inflorescence types in Nymphoides (condensed or expanded), which differ primarily by the relative degree of internode elongation between pairs of flowers (character 7). A comparative examination of inflorescence morphology and development is provided in Chapter 3 and is simply summarized here. In all floating-leaved species, flowers are carried above water by the flexion of their pedicels, one or two per day, after which the pedicels relax and the fruits complete their development underwater. The condensed inflorescence type consists of a dense flower cluster that is supported by a single floating leaf and arises at the end of a long, flexuous stem. Subsequent fertile branches may arise from the inflorescence node, and these have been termed ‗water surface stolons‘ (―stolons qui courent à la surface de l’eau‖, Raynal 1974a: p. 234). In the expanded inflorescence type, flowers arise in subopposite pairs subtended by two bracts, one of which may resemble a floating foliage leaf.

Flower pairs are separated by internodes that elongate up to several centimeters, although the internodes of some specimens remain short, producing an inflorescence similar to the condensed type (pers. obs.). A third type of inflorescence exists only in N. peltata, where nodes each have two floating leaves and a cluster of several flowers. Inflorescences of the expanded or N. peltata type also may branch at nodes without flowers, in which case the node is subtended by a single floating leaf. The inflorescences of the two erect habit species (N. cambodiana, N. exiliflora) most closely resemble the expanded inflorescence type because their flowers are borne in subopposite pairs; more distal bracts are small and lanceolate, whereas proximal bracts are

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides expanded to be foliose. The number of flowers per node, then, can approximate the inflorescence type, with condensed inflorescences having 4-10 or more flowers per node and expanded inflorescences having only two (character 8). The unique inflorescence of N. peltata combines elements of both condensed and expanded inflorescence types, having clusters of flowers that are supported by pairs of floating leaves and separated by internodes. For the purpose of this study,

N. peltata was coded as having condensed floral internodes (i.e., internodes of the floral continuation axis, rather than the inflorescence elongation axis; Chapter 3)

Flower. The actinomorphic flowers of Nymphoides are borne on pedicels that bend to elevate them above water, where they are pollinated in the open air by insects. Both the calyx and corolla are syngamous, being divided distally into lobes that typically number five, with a few species regularly having more or fewer elements (character 11). Corolla lobes (i.e., petals) vary in color from yellow/orange to white (character 12), the latter sometimes tinged purple or with a yellow throat (character 13). Two broad categories exist for corolla lobe ornamentation, namely whether the lobes have lateral wings or are covered with ciliate hairs. Intermediate conditions exist that should be mentioned. The lateral wings may be enitre, finely serrate, or laciniate to varying degrees, in the extreme appearing simply as a line of hairs down the margin of the petal (character 14; Raynal 1974a). The adaxial petal surface may be glabrous, have a single median wing, or be densely covered with hairs (character 15). If a median wing exists, it takes the form of the lateral wings, i.e., being entire (e.g., N. cristata), laciniate (e.g., N. crenata), or existing only as a line of hairs (e.g., N. guineensis). Species that appear to have densely hairy petals should be distinguished as to whether the hairs are dense throughout (e.g., N. brevipedicellata) or exist only as a single median line. The corona, which in many cases exists as a single transverse fringe of hairs across the base of the petal at the corolla throat (character 16), further complicates the distinction of petal types. Species that otherwise have glabrous petals nonetheless can have a pubescent corona (e.g., N. cristata).

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Stamens in Nymphoides are adnate to the corolla tube, from which they depart at the corolla throat. In heterostylous species (see above), their length depends on the morph type, as stigma and anther height vary reciprocally between long-styled and short-styled flowers. Anthers are sagittate and vary in color from yellow to black, even to blue in some species (e.g., N. krishnakesara; Sivarajan and Joseph 1993). Alternating with the stamens and located in the corolla tube of most Nymphoides are interstaminal glands (character 17), which consist of a cluster of glandular hairs that may be sessile or pedicellate (character 18). The gynoecium of all

Nymphoides species is unilocular with parietal placentation, superior or fused to varying degrees with the perianth, and consists of two carpels in nearly all species. The ovary may be stout or elongate, and the stigma can range from nearly entire to laciniate or deeply divided (e.g., N. macrosperma; Sivarajan and Joseph 1993). In most species, the base of the ovary is encircled by carpellary glands, which are opposite and equal to the stamens in number (character 19); these are elliptical and tend to be tipped by a row of ciliate hairs, and may be colored green or shades of orange.

Fruit. All Nymphoides have capsular fruits that dehisce irregularly (i.e., not by valves like related genera; Aston 1973). In species where the fruits remain underwater, the capsule may decay away to release seeds that float on the water surface. Species in ephemeral habitats have capsules that dry and break open haphazardly. Seed numbers per capsule (character 21) vary from only one or two (e.g., N. cambodiana, N. disperma; Aston 1986; Tippery et al. 2009) to over

50 (e.g., N. montana; Aston 1982).

Seeds. Seeds are among the most variable characters in Nymphoides, yet they are remarkably consistent within a species and therefore are highly diagnostic (Aston 2003), especially for dried specimens, in which the flowers are notoriously difficult to examine. They range in size from 0.5 mm (e.g., N. parvifolia) to over 5 mm (e.g., N. peltata) in length (character

22), with shapes that are orbicular or elliptical (character 23) and which may be globose or strongly compressed (character 24). The surfaces can be unornamented or consist of single- or

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides multi-cell projections (character 25), often termed ‗tubercles‘ (e.g., Aston 2003), that may arise densely (i.e., one per epidermal cell), in clusters, or sparsely (i.e., with unornamented cells interspersed. The projections of a single cell may have mostly parallel sides or taper gradually, or they may be expanded and bulbous (character 26). Seed epidermal cell junctures also differ in being polygonal or interdigitate (character 27), although both types can be found on the same seed (Aston 2003). Some seeds possess a distinctive caruncle, which is a cluster of expanded and often nutritive cells that surround the hilum (character 28; Chuang and Ornduff 1988; Aston

2003).

Molecular data. Nucleotide sequence data were obtained from specimens that were collected and preserved in the field using liquid CTAB (Rogstad 1992) or from dried herbarium specimens. Genomic DNA was extracted and amplified for select gene regions following Les et al. (2008). Primers used for amplification and sequencing were as follows: nuclear ribosomal internal transcribed spacer (nrITS) – ITS2, ITS3, ITS4, and ITS 5 (Baldwin 1992); matK and trnK introns – trnK-3914 (dicot) and trnK-2R (Johnson and Soltis 1995), 1F (Bremer et al. 2002), and

0445F, 0503R, 1011R, 1556R, 1749F, 1848F, and 1966R (Tippery et al. 2008). Amplified DNA fragments were purified using 0.1 µL ExoSAP-IT® enzyme mixture (Affymetrix, Inc., Santa

Clara, California), 0.9 µL water, and 1.0 µL amplification product in a 1.5 µL reaction.

Sequencing reactions were conducted using 2.0 µL of cleaned amplicon, 1.0 µL of Big Dye®, 2.0

µL of 5x ABI buffer (Applied Biosystems, Foster City, California), and 3.2 pmol of sequencing primer in a 10 µL reaction. Cycle sequencing and cleanup followed Les et al. (2008); sequencing was performed on an ABI PRISM® 3100 genetic analyzer (Applied Biosystems).

Chromatograms were edited using the program 4Peaks ver. 1.7 (Griekspoor and

Groothuis 2005) and assembled into contigs using CodonCode Aligner ver. 3.0.3 (CodonCode

Corporation, Dedham, Massachusetts). Nucleotide sequences were aligned manually using

MacClade ver. 4.06 (Maddison and Maddison 2000). Insertions and deletions (indels) were

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides scored for the aligned nucleotide matrices using simple indel coding (Simmons and Ochoterena

2000) implemented with the program SeqState ver. 1.4.1 (Müller 2005, 2006).

Phylogenetic Analyses. Data were analyzed separately (morphology, nrITS and matK/trnK nucleotide and indel data) using both equally-weighted maximum parsimony and maximum likelihood methods. Prior to combining data, partition-homogeneity / incongruence- length difference (ILD) tests were conducted using PAUP* ver. 4.0b10 (heuristic search, 1,000 replicates, maxtrees=1,000; Farris et al. 1994; Swofford 2002) with a significance threshold of p

< 0.01, in order to evaluate the relative congruency of the nuclear (nrITS DNA and indels) and chloroplast (matK/trnK DNA and indels) data partitions. After initial results indicated that the partitions were significantly incongruent (see Results), I employed taxon jackknifing (Lecointre and Deleporte 2005) to determine the taxa responsible for the incongruence, a procedure that proved effective in a previous study of N. montana (Tippery and Les, submitted). Single taxa or clades of taxa that resolved consistently on both nrITS and matK/trnK phylogenies (see Results) were pruned singly or in combinations. In addition, taxon jackknifing was conducted for two subgroups of taxa, which resolved consistently on both nrITS and matK/trnK phylogenies relative to N. minima (see Results). One of these subgroups resolved as the sister to N. minima, and the other comprised a grade of taxa toward N. minima and the other subgroup. In the subgroups also, taxa or clades of taxa were removed singly or in combinations. After determining that congruence could be achieved only by removing more than two of the ingroup taxa or clades (see

Results), the nuclear and chloroplast data partitions were evaluated independently. Incongruence was evaluated further using the Templeton test in PAUP* (Templeton 1983) to identify nodes that were significantly incongruent between the two data sets.

For each data matrix, the data were analyzed twice, once with all available accessions

(referred to as the ‗full‘ data set) and once with only one representative accession per taxon

(referred to as ‗trimmed‘). In the trimmed data analysis, N. indica comprised two ‗taxa‘, one each from Australia and India, because molecular data resolved accessions from these regions to

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides different positions on the matK/trnK phylogeny (see Results). Heuristic tree searches were performed under parsimony in PAUP* ver. 4.0b10 (Swofford 2002) with 100 replicates of random stepwise addition and branch swapping by tree bisection and reconnection (TBR), using maxtrees=100,000. Multistate taxa in the morphology data were treated as polymorphisms and ambiguous nucleotide states in the molecular data as uncertainties. Support for internal nodes was evaluated using 1,000 bootstrap replicates in PAUP* with the following options: heuristic search, one random stepwise addition per replicate, swapping by TBR, and maxtrees=10,000.

Trees were depicted as one of many most-parsimonious phylograms (full-data trees) or strict- consensus cladograms of all most-parsimonious topologies (trimmed-data trees).

After model selection with Modeltest ver. 3.4 under the AIC criterion (Posada and

Crandall 1998; Posada and Buckley 2004; Posada 2006), likelihood analysis was implemented using GARLI ver. 0.96.r396 (Zwickl 2006), on both full and trimmed accession trees.

Morphology and indel data were evaluated using the Mkv model (Lewis 2001), and models of

DNA evolution were applied to nrITS (GTR+I+G) and matK/trnK (TVM+ G) aligned nucleotide data. Combined data matrices for nrITS and matK/trnK were partitioned between nucleotide and indel data, with their respective models applied. Ten separate likelihood runs were performed using different random starting seeds, and the tree with the maximum likelihood score was compared with the parsimony consensus tree. Bootstrap analysis was conducted in GARLI using

1,000 replicates.

Character evolution. State transitions for two features, heterostyly (character 4,

Appendix 2.02) and inflorescence internodes (character 7), were mapped onto the trimmed taxa nrITS and matK/trnK trees under the likelihood criterion in Mesquite ver. 2.72 (Maddison and

Maddison 2009). States for outgroup taxa were assigned following previously published data

(Chapter 1; Tippery et al. 2008).

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

RESULTS

Morphological data. All of the 28 morphological characters included in the analysis were parsimony-informative (Appendix 2.02), and 10.4% of cells lacked data (Table 2.01).

Molecular data. The full nrITS data matrix (955/169 characters [nucleotide / indel],

339/136 parsimony-informative, 4.3% missing data excluding gaps) comprised 148 accessions, and the full matK/trnK matrix (2,654/45 characters, 217/25 parsimony-informative, 35.6% missing data overall, 5.4% missing in the 5‘ trnK intron region) had 142 accessions. Each trimmed data matrix (with only one accession per taxon) comprised 39 accessions, for which fewer data were missing for both nrITS (2.6% missing) and matK/trnK (15.5% missing overall,

2.9% missing in the 5‘ trnK intron). The ILD test indicated significant incongruence between the nuclear and chloroplast data partitions (p = 0.001). Successive pruning of taxa (cf. Lecointre and

Deleporte 2005) failed to yield a non-significant incongruence (p > 0.01) after removing all possible pairs of taxa or consistently resolved clades. Taxon jackknifing directed at the clade sister to N. minima (i.e., in ILD tests only containing taxa belonging to the clade; see below) achieved non-significant incongruence by removing N. ezannoi (p = 0.12) or the N. indica accession from Australia (p = 0.02), whereas pruning of taxa in the grade toward this clade revealed that removal of both N. montana and N. spinulosperma could recover a non-significant incongruence value (p = 0.06).

Phylogenetic analyses. Phylogenetic analysis of morphological data reached the imposed limit of 100,000 most-parsimonious trees (156 steps, CI = 0.28, RI = 0.64; Fig. 2.02) and returned a maximum likelihood score (lnL) of -566. Morphological data resolved several clades, but none with high bootstrap support. All morphological characters but three (characters 3, 5, and

19; Appendix 2.02) had high homoplasy (homoplasy index > 0.5, determined in PAUP*;

Swofford 2002).

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Analysis of the full nrITS data matrix obtained the set maximum of 100,000 most- parsimonious trees (1,050 steps, CI = 0.68, CIexc = 0.65, RI = 0.94; Fig. 2.03), as did analysis of the full matK/trnK data matrix (414 steps, CI = 0.92, CIexc = 0.88, RI = 0.98; Fig. 2.04).

Parsimony analysis of trimmed data matrices yielded 12 trees for nrITS (879 steps, CI = 0.71,

CIexc = 0.60, RI = 0.82) and the maximum of 100,000 trees for matK/trnK (344 steps, CI = 0.92,

CIexc = 0.82, RI = 0.94; Fig. 2.05). Likelihood analysis resulted in trees with natural log likelihood scores of -7,066 (full nrITS), -6,522 (full matK/trnK), -5,820 (trimmed nrITS), and -

5,854 (trimmed matK/trnK). The topologies of likelihood trees (not shown) were fully congruent with the corresponding strict consensus trees recovered under parsimony.

The majority of taxa for which multiple accessions were analyzed resolved to single, well-supported clades on both the nrITS (Fig. 2.03) and matK/trnK (Fig. 2.04) trees, sometimes with no branch lengths between accessions. Clades with notable sequence variation within a species included the geographically widespread N. indica, for which accessions from Australia resolved variously to clades or grades that were rather distant from the single N. indica accession from India. Nymphoides crenata comprised two subclades that had substantial genetic distance between them, N. aurantiaca comprised a clade with considerable genetic variation, and a single accession of N. exiliflora from the Northern Territory was distinct from four other accessions of that taxon from Queensland.

There was little agreement between the nuclear and chloroplast data about phylogenetic relationships in Nymphoides (Fig. 2.05). Both data matrices resolved the genus as monophyletic relative to Liparophyllum, with strong support. The only other major point of agreement involved the placement of N. minima, which resolved as the sister species of a clade containing about half of the species analyzed. The nodes that resolved the position of N. minima were recovered by both nuclear and chloroplast data, with moderate support. In addition, the sister relationship of N. crenata to all other ingroup taxa was supported marginally by the nrITS data and unresolved by matK/trnK data. Elsewhere on the trees, both data sets resolved clades containing N. aquatica +

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

N. cordata, N. aurantiaca + N. cambodiana, N. brevipedicellata + N. thunbergiana, N. elliptica +

N. triangularis, N. fallax + N. humboldtiana, N. furculifolia + N. parvifolia + N. quadriloba, and

N. planosperma + N. simulans + N. spongiosa.

The nrITS and matK/trnK trees differed with respect to 17 nodes on the nrITS tree and six on the matK/trnK tree. Templeton tests (Templeton 1983) of topological congruency revealed significant incongruence at 12 nodes resolved on the nrITS phylogeny and five nodes on the matK/trnK tree (Fig. 2.05, circles). Nodes that were significantly incongruent tended to have only moderate bootstrap support values, although several well-supported nodes (bootstrap > 90%) also were incongruent.

Character evolution. The evolution of two characters was reconstructed on the nuclear and chloroplast trees constructed from trimmed data (Fig. 2.06). The condition of heterostyly

(character 4, Appendix 2.02) resolved as the most likely ancestral state for most nodes on the nrITS and matK/trnK trees (Fig. 2.06A), whereas several backbone nodes, including the common ancestor of N. minima and its sister clade, and several nodes descended from this ancestor, were reconstructed to be homostylous. The dioecious species N. aquatica and N. cordata were largely unresolved relative to other taxa, although their most recent ancestor was resolved to be homostylous on the nrITS tree and marginally heterostylous on the matK/trnK tree. Inflorescence morphology (character 7) mapped to four transitions on the nrITS tree and at least two transitions on the matK/trnK tree (Fig. 2.06B), depending on the resolution of internal nodes. The nrITS tree unambiguously reconstructed transitions from expanded to condensed morphology for three clades, plus one independent transition to the N. peltata morphology from an ancestor with an expanded inflorescence.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

DISCUSSION

Hybrid origin of species. Phylogenetic study of Nymphoides using two genomic regions, the nuclear ITS region (nrITS) and the chloroplast matK gene and trnK introns

(matK/trnK), failed to recover a congruent tree topology. Templeton tests (Templeton 1983) indicated that the majority of nodes that differed between the nuclear and chloroplast trees were significantly incongruent (p < 0.01), although five nodes on the nrITS and one node on the matK/trnK tree showed non-significant incongruence (Fig. 2.05). In a previous study that analyzed fewer Nymphoides taxa, the source of incongruence was traced to a single taxon, N. montana, which was deemed to be of hybrid origin because its nuclear data were most similar to

N. geminata but its chloroplast data closely matched N. spinulosperma (Tippery and Les, in press). However, in the current study, which included 32 ingroup taxa, the disagreement between nuclear and chloroplast data trees was too great to be attributed to one or even several taxa.

Systematic removal of taxa did not show non-significant ILD p-values in any scenario where two taxa (or consistently resolved clades) were removed, and systematic removal of three or more taxa was computationally intractable (23 taxa or clades arranged into inclusion groups of 20, or n

= 1,771 combinations). Therefore, subgroups of taxa or clades, which resolved consistently relative to N. minima, were analyzed independently for sources of incongruence. In the sister clade of N. minima, N. ezannoi and N. indica from Australia were found to contribute to the incongruence between the nrITS and matK/trnK phylogenies, and in the grade toward this clade

N. montana and N. spinulosperma were major sources of incongruence.

There are several possible explanations for the incongruence between nuclear and chloroplast data. The different signal could be due to incomplete lineage sorting, introgression, or ancestral hybridization, possibly coincident with speciation. Under a hybrid species origin, the different genetic background of one or the other parental taxon, along with any corresponding phylogenetic signal, subsequently could have been lost. Studies in other groups have shown that

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides nuclear data can be preferentially lost from one genome or another (Buggs et al. 2009).

Furthermore, the nrITS region itself is prone to homogenization through concerted evolution

(Álvarez and Wendel 2003), which could lead to the signal from the maternal nuclear genome being lost and evidence of the maternal parent retained only in the chloroplast genome. On the

Nymphoides phylogeny, approximately half of the taxa resolved differently on the nuclear and chloroplast trees, which might appear to be an unreasonable amount of hybridization to postulate.

However, a similar proportion of hybrid species has been demonstrated in the genus Persicaria

(Kim and Donoghue 2008). In that study, the authors were able to correlate hybridization with allopolyploid speciation, and thus determine that polyploids were the result of hybridization from diploid parental lineages.

In Nymphoides (x = 9), chromosome numbers are known for relatively few species, although multiple species with either diploid (2n = 18) and tetraploid (2n = 36) counts and one hexaploid species (2n = 54; N. peltata) have been reported for the genus (Ornduff 1970; Li et al.

2002). Without being able to ascribe ploidy levels to most of the Nymphoides species in this study, it was impossible to conclude which species represented the parental vs. derived (i.e., allopolyploid) lineages. Three species included in this study, N. crenata, N. cristata, and N. indica, have been reported to be diploid, and seven are tetraploid (Table 2.01); N. peltata represents the only species reported to be solely hexaploid (2n=54), although both tetraploid and hexaploid counts have been obtained for N. geminata and N. montana. One anomalous species not included in the molecular analysis, N. lungtanensis, is triploid (2n=27) and also produces no fruit (Li et al. 2002). This species could represent a sterile hybrid between diploid and tetraploid species, but genetic data have not been evaluated yet to address this possibility or to identify putative parental taxa. If N. lungtanensis did arise by hybridization, it would be a contemporary example of a hybrid species origin and a system in which to examine post-hybridization effects on genome organization.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Of the taxa for which molecular data were analyzed, chromosome numbers are known for too few to conclude which species are directly descended and which originated by hybridization.

Nymphoides crenata is diploid and resolved in both nrITS and matK/trnK analyses near the base of the genus. The geographically widespread and phylogenetically nested species N. indica was reported to be diploid, as was the gynodioecious species N. cristata. Given the genetic diversity of N. indica, chromosome numbers should be evaluated for a greater diversity of samples in order to confirm that all lineages of the species are diploid. The morphologically similar neotropical species N. humboldtiana, which Ornduff (1969) recommended be merged with N. indica, has a tetraploid chromosome complement and in my analyses consistently resolved to a different clade than Old World N. indica. In fact, samples of N. humboldtiana resolved with N. fallax, another

New World species and a tetraploid, which differs from the former in having yellow flowers and verrucose seeds. The data here presented thus support the retention of N. humboldtiana as a species separate from N. indica. Moreover, the diversity of genotypes that were recovered from

N. indica specimens in the Old World suggests the presence of several entities, which subsequent morphological analyses might reveal to be proper species.

Several anomalous Nymphoides species are polyploid, and their potential hybrid origin could have been a factor in the evolution of their distinct morphologies. Nymphoides peltata is the most morphologically divergent species in the genus, having a unique inflorescence architecture, large, compressed fruits, and flattened seeds with a marginal ring of stiff projections.

Its distinctness prompted Grisebach (1839) to erect a monotypic section Waldschmidtia (Weber ex F.H.Wigg.) Griseb. (= sect. Nymphoides; Sivarajan and Joseph 1993) However, both nuclear and chloroplast data resolved N. peltata as nested within clades of other Nymphoides, thus potentially indicating a hybrid origin, particularly considering the species‘ hexaploid chromosome number (2n = 54). Another interesting chromosomal pattern occurs in N. aquatica and N. cordata, two tetraploid species native to temperate North America and the only dioecious species in my molecular phylogenetic analyses. The closest relative of these species was not

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides determined conclusively, but notably the gynodioecious species N. cristata, which resolved near the dioecious species on the tree, is diploid.

Heterostyly and dioecy. Evolutionarily, one of the most important distinctions among

Nymphoides species involves their sexual condition. The phylogenetic distribution of heterostyly, homostyly, gynodioecy, and dioecy did not show a clear pattern in this analysis on either the nrITS or the matK/trnK tree (Fig. 2.06A). The homostylous plants that were sampled occupied rather dispersed phylogenetic positions, indicating that for these taxa, each instance represents an independent acquisition of homostyly from a heterostylous progenitor. An exception to this pattern was observed for the homostylous species N. minima and members of its sister clade (e.g.,

N. furculifolia, N. parvifolia, N. rautanenii), whose phylogenetic relationships support a brief maintenance of homostyly in an ancestral taxon, from which heterostyly re-evolved subsequently.

However, the reconstruction of states must be interpreted with caution, because evolutionary theory does not readily support the re-acquisition of complex characters like the heterostylous syndrome (cf. Barrett and Shore 2008). Moreover, several of the taxa that provide signal for the ancestral state reconstruction have rather small flowers (< 1 cm; character 10, Table 2.01), and these could either have lost the style/filament polymorphism as their size reduced or simply be too small for researchers to have observed heterostyly. In one comparative study, flowers of the homostylous and self-compatible species N. geminata were found to be smaller relative to the heterostylous (but otherwise morphologically similar) species N. montana (Haddadchi 2008).

Furthermore, the apparent homostyly of species does not necessarily implicate the loss of self- incompatibility. Data are currently unavailable regarding the presence or effectiveness of self- incompatibility in most homostylous Nymphoides species. Given the uncertainty of relationships and the possible hybrid origins of taxa, an acceptable elucidation of the evolution of homostyly will require further research before conclusions can be drawn regarding the phylogenetic relationships of homostylous taxa and the effectiveness of their self-incompatibility mechanism.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

The ancestral condition of the dioecious species N. aquatica and N. cordata is similarly difficult to assess, because of the poor resolution of their related taxa. The nrITS phylogeny placed the homostylous African species N. ezannoi as their closest relative, whereas the matK/trnK data ambiguously resolved five species or clades as the most closely related (Fig.

2.06A). Consequently, the reconstruction of homostyly or heterostyly in the most immediate ancestor requires more confident determination of phylogenetic relationships, possibly also including increased taxon sampling, in order to be conclusive. The gynodioecious species N. cristata resolved as a potential sister taxon to N. aquatica / N. cordata in the matK/trnK analysis, and both data sets resolved it as a close relative.

Inflorescence architecture. The incongruence of phylogenetic trees that were resolved using nuclear and chloroplast data hinders the reconstruction of morphological character evolution and historical biogeography in Nymphoides. Nonetheless several patterns emerged that were robust to the incongruence between trees. Likelihood reconstruction of ancestral states revealed that there were at least two transitions in Nymphoides from an expanded inflorescence morphology to a condensed type (Fig. 2.06B), with three transitions mapped onto the more thoroughly resolved nrITS tree. The repeated evolution of condensed inflorescence morphology and its prevalence in the genus suggest that it has a strong selective advantage, at least under certain conditions, and thus may have contributed to the diversification of some Nymphoides clades, particularly the clade containing N. minima and its sister taxa, all of which have a condensed inflorescence. Although complete taxon sampling was not attained for Nymphoides, it is fairly likely that the unsampled species with condensed morphology would fall into one of the clades determined in this study, because of their geographic distance from species with expanded morphology (e.g., species in Africa and the Americas) and the broad morphological diversity

(e.g., in floral and seed traits) of species that were analyzed with molecular data. Notably, all extra-Australian species with condensed inflorescences that were included in this study resolve to the same clade sister to N. minima. Furthermore, all Nymphoides species with expanded

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides inflorescence morphology except N. tonkinensis were included in this study. Further attention should be paid to the multiple independent origins of condensed inflorescence morphology and determining, for example, whether the developmental mechanism that produces the morphology is the same in every case.

Biogeography. Geographically, Nymphoides species are found worldwide, and it is noteworthy that all species with the expanded inflorescence morphology grow in Australia and tropical Asia. The phylogenetic distribution of species in Menyanthaceae supports an origin for the genus in Australia (Chapter 1; Tippery et al. 2008), and the analysis of additional Nymphoides species presented here also supports diversification of the genus from an Australian ancestor.

Moreover, the distribution of Australian Nymphoides species on the nuclear and chloroplast phylogenies coincides with the occurrence of expanded inflorescence morphology, which mapped as the ancestral condition for the genus. Superficially, the expanded morphology found in

Nymphoides more closely resembles the inflorescences of related outgroup taxa in the sister genus Liparophyllum, which mostly have expanded internodes and nodes with one to three flowers, although both single-flowered inflorescences (L. gunnii) and inflorescences with clusters of sessile flowers (L. capitatum, L. congestiflorum) also occur (Fig. 2.06B).

In the grade of species leading to N. minima, all species are exclusively Australian except for N. cambodiana, which has a tropical Asian distribution that overlaps with its sister species N. aurantiaca, and the wide-ranging Eurasian species N. peltata (Fig. 2.05). Species in the clade sister to N. minima are predominantly extra-Australian, except the consistently resolved clade of

N. furculifolia + N. parvifolia + N. quadriloba, and N. indica, which has a range that extends throughout the Old World tropics. Resolution within the clade sister to N. minima was relatively poor in this study, however several African and Neotropical species associated closely on the nrITS tree (Fig. 2.03). Further sampling of African and Asian species will be necessary to reconstruct more robustly the biogeographical patterns of Nymphoides species with condensed inflorescences.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Conclusion. The analysis of phylogenetic relationships in Nymphoides clearly requires further work to unravel patterns of ancestral hybridization, which hamper the reconstruction of ancestral states and evolutionary patterns. Moreover, the existence of hybrid species origins must be verified using chromosomal data and potentially additional genetic markers that can provide further evidence. Nonetheless, several patterns have emerged from the current study, most notably that inflorescence architecture transitioned at least two times from an ancestrally expanded morphology to a derived, condensed morphology. Other hypotheses, for example regarding the evolution of dioecy and the disposition of its precursor (e.g., dioecious, homostylous, heterostylous), require additional taxon sampling and better phylogenetic resolution.

ACKNOWLEDGEMENTS. This study depended greatly upon the generosity of several plant collectors, including Lori Benoit, Claudia Bove, Surrey Jacobs, Ken Langeland, Don Les, Chris

Martine, Michael Moody, C. Mogalhães, A. Moreira, Walter Pagels, Hyunchur Shin, C. Sigueira,

Jackie Smith, Amy Weiss, Dave Wilson, and the directors and staff of herbaria (BRI, DNA, L,

MO, NSW, PERTH, TUR, UC). Research was funded in part by the American Society of Plant

Taxonomists, the University of Connecticut Center for Conservation and Biodiversity, the

University of Connecticut Graduate School, and the Ronald L. Bamford endowment. I am grateful to have received Australian hospitality from Janice and Roger Carolin, John and Marion

Clarkson, Lance Smith, and especially Betty and Surrey Jacobs, who welcomed me into their home as they would their own family. I am forever indebted to the National Herbarium of New

South Wales and the late Surrey Jacobs for enabling and facilitating the field collecting trip where most of the Australian Nymphoides were collected.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

LITERATURE CITED

Álvarez, I. and J. F. Wendel. 2003. Ribosomal ITS sequences and plant phylogenetic inference.

Molecular Phylogenetics and Evolution 29: 417-434.

Aston, H. I. 1973. Aquatic plants of Australia. Carlton, Victoria: Melbourne University Press.

Aston, H. I. 1982. New Australian species of Nymphoides Séguier (Menyanthaceae). Muelleria 5:

35-51.

Aston, H.I. 1984. Nymphoides triangularis and N. elliptica (Menyanthaceae): two new Australian

species. Muelleria 5: 265-270.

Aston, H.I. 1986. Nymphoides disperma (Menyanthaceae): a new Australian

species. Muelleria 6(3-4): 197-200.

Aston, H.I. 1987. Nymphoides beaglensis (Menyanthaceae): a new Australian

species. Muelleria 6(5):359-362.

Aston, H. I. 1997. Nymphoides spinulosperma (Menyanthaceae): a new species from south-

eastern Australia. Muelleria 10: 21-25.

Aston, H.I. 2002. Nymphoides simulans (Menyanthaceae): a new species from northern

Australia. Muelleria 16:83-86.

Aston, H. I. 2003. Seed morphology of Australian species of Nymphoides (Menyanthaceae).

Muelleria 18: 33-65.

Baldwin, B. G. 1992. Phylogenetic utility of the internal transcribed spacers of ribosomal DNA in

plants: an example from the Compositae. Molecular Phylogenetics and Evolution 1: 3-16.

Barrett, S. C. H. 1992. Evolution and Function of Heterostyly. Berlin: Springer.

Barrett, S. C. H. and J. S. Shore. 2008. New insights on heterostyly: Comparative biology,

ecology and genetics. Pp. 3-32 in V. E. Franklin-Tong, ed. Self-Incompatibility in

Flowering Plants: Evolution, Diversity, and Mechanisms. Berlin: Springer.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Bremer, B., K. Bremer, N. Heidari, P. Erixon, R. G. Olmstead, A. A. Anderberg, M. Källersjö,

and E. Barkhordarian. 2002. Phylogenetics of asterids based on 3 coding and 3 non-

coding chloroplast DNA markers and the utility of non-coding DNA at higher taxonomic

levels. Molecular Phylogenetics and Evolution 24: 273-300.

Buggs, R. J. A., A. N. Doust, J. A. Tate, J. Koh, K. Soltis, F. A. Feltus, A. H. Paterson, P. S.

Soltis, and D. E. Soltis. 2009. Gene loss and silencing in Tragopogon miscellus

(Asteraceae): comparison of natural and synthetic allotetraploids. Heredity 103: 1-9.

Chuang, T. I. and R. Ornduff. 1992. Seed morphology and systematics of Menyanthaceae.

American Journal of Botany 79: 1396-1406.

Cohen, J. I. 2010. ―A case to which no parallel exists‖: The influence of Darwin‘s Different

Forms of Flowers. American Journal of Botany 97: 701-716.

Cook, C. D. K. 1996. Aquatic Plant Book. The Hague: SPB Academic Publishing.

Cook, C. D. K. 1999. The number and kinds of embryo-bearing plants which have become

aquatic: a survey. Perspectives in Plant Ecology, Evolution and Systematics 2: 79-102.

Farris, J. S., M. Källersjö, A. G. Kluge, and C. Bult. 1994. Constructing a significance test for

incongruence. Systematic Biology 44: 570-572.

Ganders, F. R. 1979. The biology of heterostyly. New Zealand Journal of Botany 17: 607-635.

Godfrey, R. K. and J. W. Wooten. 1981. Aquatic and Wetland Plants of Southeastern United

States: Dicotyledons. Athens: University of Georgia Press.

Griekspoor, A. and T. Groothuis. 2005. 4Peaks ver. 1.7. http://mekentosj.com/4peaks/

Grisebach, A. H. R. 1839. Tribus VII. Menyanthideae. Pp. 336-348 in Genera et species

Gentianearum. Stuttgart: J. G. Cotta.

Haddadchi, A. 2008. Floral variation and breeding system in distylous and homostylous species

of clonal aquatic Nymphoides (Menyanthaceae). PhD. Thesis. Armidale: University of

New England.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Jacquemyn, H. and O. Honnay. 2008. Mating system evolution under strong clonality: towards

self-compatibility or self-incompatibility? Evolutionary Ecology 22: 483-486.

Johnson, L. A. and D. E. Soltis. 1995. Phylogenetic inference in Saxifragaceae sensu stricto and

Gilia (Polemoniaceae) using matK sequences. Annals of the Missouri Botanical Garden

82: 149-175.

Kadereit, G. 2007. Menyanthaceae. Pp. 599-604 in J. W. Kadereit and C. Jeffrey, eds. The

Families and Genera of Vascular Plants, vol. 8. Berlin: Springer.

Kim, S.-T., and M. J. Donoghue. 2008. Incongruence between cpDNA and nrITS trees indicates

extensive hybridization within Eupersicaria (Polygonaceae). American Journal of Botany

95: 1122-1135.

Klackenberg, J. 1990. Famille 168 bis. Menyanthacées. Pp. 169-180 in P. Morat, ed. Flore de

Madagascar et des Comores. Paris: Muséum National d‘Histoire Naturelle.

Larson, D. 2007. Reproduction strategies in introduced Nymphoides peltata populations revealed

by genetic markers. Aquatic Botany 86: 402-406.

Lecointre, G. and P. Deleporte. 2005. Total evidence requires exclusion of phylogenetically

misleading data. Zoologica Scripta 34: 101-117.

Les, D. H., S. W. L. Jacobs, N. P. Tippery, L. Chen, M. L. Moody, and M. Wilstermann-

Hildebrand. 2008. Systematics of Vallisneria (Hydrocharitaceae). Systematic Botany 33:

49-65.

Lewis, P. O. 2001. A likelihood approach to estimating phylogeny from discrete morphological

character data. Systematic Biology 50: 913-925.

Li, S.-P., T.-H. Hsieh, and C.-C. Lin. 2002. The genus Nymphoides Séguier (Menyanthaceae) in

Taiwan. Taiwania 47: 246-258.

Lloyd, D. G. 1979. Evolution towards dioecy in heterostylous populations. Plant Systematics and

Evolution 131: 71-80.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Ornduff, R. 1966. The origin of dioecism from heterostyly in Nymphoides (Menyanthaceae).

Evolution 20: 309-314.

Ornduff, R. 1969. Neotropical Nymphoides (Menyanthaceae): Meso-American and West Indian

species. Brittonia 21: 346-352.

Ornduff, R. 1970. Cytogeography of Nymphoides (Menyanthaceae). Taxon 19: 715-719.

Maddison, D. R. and W. P. Maddison. 2000. MacClade: Analysis of phylogeny and character

evolution, ver. 4.0. Sunderland: Sinauer Associates.

Maddison, W.P. and D.R. Maddison. 2009. Mesquite: A Modular System for Evolutionary

Analysis, Version 2.6. http://mesquiteproject.org.

Müller, K. 2005. SeqState – primer design and sequence statistics for phylogenetic DNA data

sets. Applied Bioinformatics 4: 65–69.

Müller, K. 2006. Incorporating information from length-mutational events into phylogenetic

analysis. Molecular Phylogenetics and Evolution 38: 667–676.

Posada, D. 2006. ModelTest Server: a web-based tool for the statistical selection of models of

nucleotide substitution online. Nucleic Acids Research 34: W700-W703.

Posada, D. and T. R. Buckley. 2004. Model selection and model averaging in phylogenetics:

advantages of the AIC and Bayesian approaches over likelihood ratio tests. Systematic

Biology 53: 793-808.

Posada, D. and K. A. Crandall. 1998. Modeltest: testing the model of DNA substitution.

Bioinformatics 14: 817-818.

Raynal, A. 1974a. Le genre Nymphoides (Menyanthaceae) en Afrique et a Madagascar. 1re partie:

Morphologie. Adansonia ser. 2, 14: 227-270.

Raynal, A. 1974b. Le genre Nymphoides (Menyanthaceae) en Afrique et a Madagascar. 2e partie:

Taxonomie. Adansonia ser. 2, 14: 405-458.

Rogstad, S. H. 1992. Saturated NaCl-CTAB solution as a means of field preservation of leaves

for DNA analyses. Taxon 41: 701-708.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Sculthorpe, C. D. 1967. The Biology of Aquatic Vascular Plants. London: Edward Arnold.

Simmons, M. P. and H. Ochoterena. 2000. Gaps as characters in sequence-based phylogenetic

analyses. Systematic Biology 49: 369-381.

Sivarajan, V. V. and K. T. Joseph. 1993. The genus Nymphoides Séguier (Menyanthaceae) in

India. Aquatic Botany 45: 145-170.

Swofford, D. L. 2002. PAUP*. Phylogenetic analysis using parsimony (*and other methods), ver.

4. Sunderland: Sinauer Associates.

Takagawa, S., I. Washitani, R. Uesugi, and Y. Tsumura. 2006. Influence of inbreeding depression

on a lake population of Nymphoides peltata after restoration from the soil seed bank.

Conservation Genetics 7: 705-716.

Templeton, A. R. 1983. Phylogenetic inference from restriction endonuclease cleavage site maps

with particular reference to the evolution of humans and the apes. Evolution 37: 221–224.

Tippery, N. P. and D. H. Les. 2009. A new genus and new combinations in Australian Villarsia

(Menyanthaceae). Novon 19: 406-413.

Tippery, N. P., D. H. Les, D. J. Padgett, and S. W. L. Jacobs. 2008. Generic circumscription in

Menyanthaceae: a phylogenetic evaluation. Systematic Botany 33: 598-612.

Tippery N. P., D. H. Les, J. C. Regalado Jr., L. V. Averyanov, Vu Ngoc Long, and P. H. Raven

2009. Transfer of Villarsia cambodiana to Nymphoides (Menyanthaceae). Systematic

Botany 34: 818-823.

Uesugi, R., K. Goka, J. Nishihiro, and I. Washitani. 2004. Allozyme polymorphism and

conservation of the Lake Kasumigaura population of Nymphoides peltata. Aquatic

Botany 79: 203-210.

Van der Velde, G., T. G. Giesen, and L. Van der Heijden. 1979. Structure, biomass and seasonal

changes in biomass of Nymphoides peltata (Gmel.) O. Kuntze (Menyanthaceae), a

preliminary study. Aquatic Botany 7: 279-300.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Vasudevan, R. 1968. A new species of Nymphoïdes (Menyanthaceae) from south India. Kew

Bulletin 22: 101-106.

Vasudevan Nair, R. 1973. Heterostyly and breeding mechanism of Nymphoides cristatum (Roxb.)

O. Kuntze. Journal of the Bombay Natural History Society 72: 677-682.

Wang, Y., Q. Wang, Y.-H. Guo, and S. C. H. Barrett. 2005. Reproductive consequences of

interactions between clonal growth and sexual reproduction in Nymphoides peltata: a

distylous aquatic plant. New Phytologist 165: 329-336.

Zwickl D. J. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological

sequence datasets under the maximum likelihood criterion. Ph.D. Dissertation. Austin:

University of Texas

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

TABLE 2.01 – Matrix of morphological data for Nymphoides species, using the characters and states given in Appendix 2.02.

1 2 3 4 5 6 7 8 9 10 N. aquatica (J.F.Gmel.) Kuntze 2 1 1 ? 1 12 0 2 2 1 N. aurantiaca (Dalzell) Kuntze ? 01 1 1 0 1 1 0 1 01 N. beaglensis Aston ? 0 1 1 0 1 1 0 12 12 N. bosseri A.Raynal ? 1 1 1 0 1 0 1 01 1 N. brevipedicellata (Vatke) A.Raynal ? 1 1 0 0 1 0 2 0 1 N. cambodiana (Hance) Tippery ? ? 0 1 0 1 1 0 1 2 N. cordata (Elliott) Fernald 2 ? 1 ? 1 1 0 2 01 0 N. coreana (Léveille) Hara 2 ? 1 0 0 1 0 1 01 01 N. crenata (F.Muell.) Kuntze 0 1 1 1 0 12 1 0 12 2 N. cristata (Roxb.) Kuntze 0 ? 1 1 1 12 0 2 12 0 N. disperma Aston ? 0 1 0 0 01 1 0 12 12 N. elegans A.Raynal ? 0 1 1 0 1 0 1 01 1 N. elliptica Aston ? 0 1 1 0 1 0 1 1 12 N. exiliflora (F.Muell.) Kuntze ? ? 0 1 0 01 1 0 01 ? N. ezannoi Berhaut ? 1 1 0 0 12 0 2 0 12 N. fallax Ornduff 2 1 1 1 0 1 0 12 1 ? N. flaccida L.B.Sm. ? ? 1 ? ? 1 0 ? 2 12 N. forbesiana (Griseb.) Kuntze ? 1 1 1 0 12 0 2 12 1 N. furculifolia Specht ? ? 1 0 0 0 0 1 01 0 N. geminata (R.Br.) Kuntze 23 ? 1 0 0 1 1 0 ? ? N. grayana (Griseb.) Kuntze ? 0 1 1 0 01 0 2 12 1 N. guineensis A.Raynal ? 0 1 0 0 1 0 1 0 0 N. hastata (Dop) Kerr ? ? 1 ? ? 01 ? ? 12 ? N. herzogii A.Galàn-Mera & G.Navarro ? ? 1 ? ? 1 0 2 12 0 N. humboldtiana (Kunth) Kuntze 2 ? 1 1 0 2 0 2 2 1 N. humilis A.Raynal ? ? 1 0 0 01 0 1 01 1 N. hydrophylla (Lour.) Kuntze 0 01 1 0 0 01 0 1 12 0 N. indica (L.) Kuntze subsp. indica 0 01 1 1 0 12 0 2 12 12 N. indica subsp. occidentalis A.Raynal ? 1 1 1 0 2 0 2 12 2 N. krishnakesara K.T.Joseph & Sivar. ? 01 1 ? 1 1 0 2 1 1

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

TABLE 2.01 (continued)

1 2 3 4 5 6 7 8 9 10 N. lungtanensis S.P.Li, T.H.Hsieh & 1 ? 1 ? 0 12 0 1 1 1 C.C.Lin N. macrosperma Vasudevan 2 01 1 ? 1 2 0 2 12 1 N. microphylla (A.St.-Hil.) Kuntze ? ? 1 ? ? 01 0 ? ? ? N. milnei A.Raynal ? 0 1 1 0 1 0 1 1 1 N. minima (F.Muell.) Kuntze ? 0 1 0 0 0 0 1 0 0 N. minor (D.Don ex G.Don) S.Gupta, ? ? 1 ? ? ? 0 ? ? ? A.Mukherjee & M.Mondal N. montana Aston 23 1 1 1 0 12 1 0 12 2 N. moratiana A.Raynal ? 0 1 1 0 1 0 1 01 1 N. parvifolia (Wall.) Kuntze ? 01 1 0 0 01 0 1 01 0 N. peltata (S.G.Gmel.) Kuntze 3 1 1 1 0 12 0 1 12 2 N. planosperma Aston ? 0 1 1 0 0 0 12 0 0 N. quadriloba Aston ? 0 1 1 0 1 0 2 01 1 N. rautaneni (N.E.Br.) A.Raynal ? 0 1 0 0 1 0 2 1 0 N. siamensis (Ostenf.) Kerr ? ? 1 ? ? ? 0 12 1 01 N. simulans Aston ? 0 1 0 0 01 0 1 0 0 N. sivarajanii K.T.Joseph ? 01 1 0 0 1 0 12 1 0 N. spinulosperma Aston ? 1 1 1 0 1 1 0 12 2 N. spongiosa Aston ? 0 1 1 0 1 0 12 01 1 N. subacuta Aston ? 0 1 1 0 1 1 0 12 2 N. tenuissima A.Raynal ? 0 1 0 0 1 0 2 0 1 N. thunbergiana (Griseb.) Kuntze ? 1 1 1 0 12 0 2 12 2 N. tonkinsensis (Dop) P.H.Hô ? ? 1 ? ? 0 1 0 1 01 N. triangularis Aston ? 0 1 1 0 1 0 2 12 12 N. verrucosa (R.E.Fries) A.Galàn-Mera & ? ? 1 ? ? ? ? ? ? ? G.Navarro

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

TABLE 2.01 (continued)

11 12 13 14 15 16 17 18 19 20 N. aquatica (J.F.Gmel.) Kuntze 1 0 1 0 0 0 1 1 1 2 N. aurantiaca (Dalzell) Kuntze 1 1 1 1 0 1 1 0 1 1 N. beaglensis Aston 1 0 0 1 0 3 1 0 1 1 N. bosseri A.Raynal 1 0 1 2 1 0 1 0 1 1 N. brevipedicellata (Vatke) A.Raynal 1 0 1 2 2 2 1 1 1 12 N. cambodiana (Hance) Tippery 1 1 1 1 0 1 1 0 1 0 N. cordata (Elliott) Fernald 1 0 1 0 0 0 1 ? ? 1 N. coreana (Léveille) Hara 01 0 1 1 1 0 1 ? 1 1 N. crenata (F.Muell.) Kuntze 1 1 1 1 1 1 1 0 1 12 N. cristata (Roxb.) Kuntze 1 0 1 0 1 1 1 1 1 1 N. disperma Aston 1 1 1 1 0 3 1 0 ? 1 N. elegans A.Raynal 1 0 ? 2 1 0 1 1 1 1 N. elliptica Aston 1 0 1 1 0 1 1 0 1 1 N. exiliflora (F.Muell.) Kuntze 1 1 1 1 0 1 0 ? 1 01 N. ezannoi Berhaut 1 0 1 0 1 1 0 ? 0 1 N. fallax Ornduff 1 1 1 1 1 2 1 ? ? 1 N. flaccida L.B.Sm. ? 0 ? ? ? ? ? ? ? ? N. forbesiana (Griseb.) Kuntze 1 1 1 2 1 2 1 1 1 1 N. furculifolia Specht 0 0 1 0 0 0 1 0 1 0 N. geminata (R.Br.) Kuntze 1 1 1 1 0 1 1 ? ? 1 N. grayana (Griseb.) Kuntze 1 1 1 1 1 12 ? ? 1 ? N. guineensis A.Raynal 1 1 1 2 1 1 1 0 1 1 N. hastata (Dop) Kerr 1 1 1 ? ? ? ? ? ? 1 N. herzogii A.Galàn-Mera & G.Navarro 1 1 1 1 0 0 ? ? ? 1 N. humboldtiana (Kunth) Kuntze 1 0 1 2 2 2 1 ? ? 1 N. humilis A.Raynal 1 ? ? 2 1 ? 1 1 1 1 N. hydrophylla (Lour.) Kuntze 1 0 0 1 ? ? ? ? ? 01 N. indica (L.) Kuntze subsp. indica 12 0 1 2 12 2 1 01 1 12 N. indica subsp. occidentalis A.Raynal 1 0 1 2 2 2 1 1 1 ? N. krishnakesara K.T.Joseph & Sivar. 1 0 0 1 1 0 1 ? 1 0

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

TABLE 2.01 (continued)

11 12 13 14 15 16 17 18 19 20 N. lungtanensis S.P.Li, T.H.Hsieh & 01 0 1 2 1 0 1 ? 1 ? C.C.Lin N. macrosperma Vasudevan 1 0 ? 1 1 2 1 1 1 12 N. microphylla (A.St.-Hil.) Kuntze ? 1 1 ? ? ? ? ? ? ? N. milnei A.Raynal 1 1 1 2 ? ? 1 0 1 01 N. minima (F.Muell.) Kuntze 1 0 1 0 0 0 0 ? ? 0 N. minor (D.Don ex G.Don) S.Gupta, ? ? ? ? ? ? ? ? ? ? A.Mukherjee & M.Mondal N. montana Aston 1 1 1 1 0 1 1 0 1 12 N. moratiana A.Raynal 1 1 1 2 2 2 1 1 1 1 N. parvifolia (Wall.) Kuntze 1 0 1 1 0 1 1 1 1 01 N. peltata (S.G.Gmel.) Kuntze 1 1 1 1 0 0 1 ? 1 2 N. planosperma Aston 1 0 1 0 0 1 1 1 1 0 N. quadriloba Aston 0 0 1 1 1 1 1 01 1 01 N. rautaneni (N.E.Br.) A.Raynal 1 1 1 2 1 0 1 1 1 1 N. siamensis (Ostenf.) Kerr 1 0 ? ? ? ? ? 1 ? N. simulans Aston 0 0 1 01 0 1 0 ? 0 0 N. sivarajanii K.T.Joseph 1 0 1 1 0 1 1 1 1 1 N. spinulosperma Aston 1 1 1 1 0 1 1 ? ? 2 N. spongiosa Aston 1 0 1 0 0 1 1 1 1 01 N. subacuta Aston 1 1 1 1 0 1 1 0 1 1 N. tenuissima A.Raynal 1 0 ? 2 1 0 1 0 0 1 N. thunbergiana (Griseb.) Kuntze 1 1 1 1 1 3 1 1 1 1 N. tonkinsensis (Dop) P.H.Hô 1 1 1 1 0 ? ? ? 1 1 N. triangularis Aston 1 0 1 1 0 1 1 0 1 1 N. verrucosa (R.E.Fries) A.Galàn-Mera & 1 0 1 ? ? ? ? ? ? ? G.Navarro

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

TABLE 2.01 (continued)

21 22 23 24 25 26 27 28 N. aquatica (J.F.Gmel.) Kuntze 2 1 0 0 1 1 ? 0 N. aurantiaca (Dalzell) Kuntze 1 1 0 0 1 0 0 1 N. beaglensis Aston 2 0 0 0 2 0 0 1 N. bosseri A.Raynal 2 1 0 1 0 ? 1 0 N. brevipedicellata (Vatke) A.Raynal 2 1 0 1 0 ? 1 0 N. cambodiana (Hance) Tippery 0 1 0 0 1 0 ? 0 N. cordata (Elliott) Fernald 2 1 0 0 1 1 0 0 N. coreana (Léveille) Hara 12 1 1 1 0 ? ? 0 N. crenata (F.Muell.) Kuntze 2 01 1 1 01 0 1 0 N. cristata (Roxb.) Kuntze 2 1 0 0 1 1 ? 0 N. disperma Aston 0 1 0 0 2 0 0 0 N. elegans A.Raynal 1 0 0 1 0 ? 0 0 N. elliptica Aston 1 1 0 0 1 2 0 1 N. exiliflora (F.Muell.) Kuntze 2 0 0 0 1 0 0 1 N. ezannoi Berhaut 1 1 0 1 2 0 1 0 N. fallax Ornduff 12 1 1 0 0 ? ? 0 N. flaccida L.B.Sm. ? ? ? ? ? ? ? ? N. forbesiana (Griseb.) Kuntze 12 1 0 1 2 0 0 0 N. furculifolia Specht 1 0 0 0 0 ? 0 0 N. geminata (R.Br.) Kuntze 2 0 0 0 1 0 0 1 N. grayana (Griseb.) Kuntze ? 1 0 1 0 ? ? 0 N. guineensis A.Raynal 0 1 0 1 0 ? 0 0 N. hastata (Dop) Kerr ? 1 ? ? 0 ? ? ? N. herzogii A.Galàn-Mera & G.Navarro 1 1 1 ? 0 ? ? 0 N. humboldtiana (Kunth) Kuntze 12 1 0 0 0 ? 1 0 N. humilis A.Raynal 12 1 0 0 1 0 1 ? N. hydrophylla (Lour.) Kuntze 1 1 0 ? 1 0 ? ? N. indica (L.) Kuntze subsp. indica 12 1 01 0 012 01 1 0 N. indica subsp. occidentalis A.Raynal 2 1 0 ? 1 1 1 0 N. krishnakesara K.T.Joseph & Sivar. 1 1 1 1 1 0 ? ?

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

TABLE 2.01 (continued)

21 22 23 24 25 26 27 28 N. lungtanensis S.P.Li, T.H.Hsieh & ? ? ? ? ? ? ? ? C.C.Lin N. macrosperma Vasudevan 01 2 1 ? 2 0 0 0 N. microphylla (A.St.-Hil.) Kuntze ? ? 0 1 ? ? ? ? N. milnei A.Raynal 0 1 0 1 0 ? 0 0 N. minima (F.Muell.) Kuntze 1 0 0 0 0 2 0 0 N. minor (D.Don ex G.Don) S.Gupta, ? ? ? ? ? ? ? ? A.Mukherjee & M.Mondal N. montana Aston 2 1 0 1 0 ? 0 0 N. moratiana A.Raynal 1 1 0 1 ? ? 1 0 N. parvifolia (Wall.) Kuntze 1 0 0 1 0 ? 0 0 N. peltata (S.G.Gmel.) Kuntze 2 2 1 1 1 0 1 0 N. planosperma Aston 0 1 1 1 1 0 0 1 N. quadriloba Aston 12 0 0 0 1 0 0 0 N. rautaneni (N.E.Br.) A.Raynal 1 1 0 1 2 0 0 0 N. siamensis (Ostenf.) Kerr 01 1 0 1 0 ? ? ? N. simulans Aston 1 0 0 0 1 0 0 1 N. sivarajanii K.T.Joseph 1 0 0 1 1 1 1 0 N. spinulosperma Aston 2 1 1 1 1 0 1 1 N. spongiosa Aston 1 01 0 0 1 0 0 1 N. subacuta Aston 01 1 0 0 1 0 ? 1 N. tenuissima A.Raynal 1 1 1 1 0 ? 1 0 N. thunbergiana (Griseb.) Kuntze 012 1 0 0 0 ? 1 0 N. tonkinsensis (Dop) P.H.Hô 1 1 0 0 0 ? ? 0 N. triangularis Aston 2 0 0 0 1 0 0 1 N. verrucosa (R.E.Fries) A.Galàn-Mera & ? ? 0 ? 1 1 ? ? G.Navarro

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

APPENDIX 2.01 – Voucher specimens for accessions cited in this study. Accession numbers in parentheses correspond to those depicted on phylogenetic trees (Figs. 2-3). Locality and voucher information is followed by GenBank accession numbers for nrITS and trnL, respectively (— = not sequenced). Asterisks (*) indicate previously published sequences that were retrieved from GenBank.

Liparophyllum Hook.f.

L. capitatum (Nees ex Lehm.) Tippery & Les AUSTRALIA: Western Australia.

Kenwick, Perth, Ornduff 9349 (UC; EF173053, EF173086).

L. congestiflorum (F.Muell.) Tippery & Les AUSTRALIA: Western Australia.

Eneabba, Lake Logue, Coughran s.n. 01 Nov 1983 (UC; EU257161, EU257174).

L. exaltatum (Sol. ex Sims) Tippery & Les AUSTRALIA: New South Wales. Coffs

Harbour, (1) Ornduff s.n. 01 Oct 1985 (UC; EU257162, EU257175). CULTIVATED. Mount

Annan Botanic Garden, (2) Jacobs 9381 (NSW; EF173054, EF173087).

L. exiguum (F.Muell.) Tippery & Les AUSTRALIA: Tasmania. Tributary of Dervent

River, sw side of Lake St. Clair Road, ―power lines overhead‖, Balmer s.n. 20 Jan 2000 (CONN;

EF173035, EF173071).

L. gunnii Hook.f. AUSTRALIA: Tasmania. Alkaline pan in far SW Tasmania, Olga

River Valley, (1) Balmer s.n. 11 May 2000 (CONN; EF173023, EU257173); Mount Field

National Park, E edge of Twisted Tarn, (2) Balmer s.n. 16 Jan 2000 (CONN; EF173024,

EF173061).

L. lasiospermum (F.Muell.) Tippery & Les AUSTRALIA: Western Australia.

Albany, (1) Ornduff 9414 (UC; EF173056, EF173089); Marbellup, (2) Cranfield 1160 (A;

EF173055, EF173088).

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

L. latifolium (Benth.) Tippery & Les AUSTRALIA: Western Australia. Augusta, (1)

Ornduff 9374 (UC; EU257163, EU257176); Dombakup River, (2) Ornduff 10000 (UC;

EU257164, EU257177).

L. violifolium (F.Muell.) Tippery & Les AUSTRALIA: Western Australia. Jandakot,

Kunzea, Ornduff 9342 (UC; EU342366, —).

Nymphoides Ség.

N. aquatica — U.S.A.: Florida. Clay County, Keystone Heights, Lake Geneva, (1)

Tippery 232 (CONN); Collier County, Big Cypress National Preserve, (2) Tippery 243 (CONN);

Highlands County, E shore of Lake Grassy, (3) Langeland s.n. 05 May 1999 (CONN); Leon

County, Tallahassee, A.B. Maclay Gardens State Park, Lake Hall, (4) Tippery 253 (CONN); Leon

County, Tallahassee, Lake Jackson, (5) Weiss 127 (CONN); Marion County, Ocala National

Forest, Hopkins Prairie, (6) Tippery 233 (CONN); Osceola County, Lake Tohopekaliga, (7)

Benoit 06-018 (CONN); Union County, Lake Butler, (8) Weiss 134 (CONN). Georgia. Decatur

County, Bainbridge, Open Pond, (9) Tippery 223 (CONN). South Carolina. Berkeley County,

Lake Moultrie, (10) Tippery 205 (CONN); Kershaw County, Camden, N.R. Goodale State Park,

(11) Tippery 204 (CONN). CULTIVATED. PetsMart, Manchester, Connecticut, (12) Tippery s.n. (CONN)

N. aurantiaca — AUSTRALIA: Northern Territory. Darwin, Knuckey Lagoons, (1)

Tippery 122 (CONN); Howard Springs, Girraween Lagoon, (2) Tippery 127 (CONN); Kakadu

National Park, 1 km N of Arnhem Highway on Wildman River Road, (3) Tippery 143 (CONN;

FJ391922, FJ391930); Kakadu National Park, 5 km N of Arnhem Highway on Wildman River

Road, (4) Tippery 147; Kakadu National Park, Creek above Gunlom, (5) Tippery 156 (CONN;

FJ391923, —); Kakadu National Park, Gunlom Falls, slow moving rock pool above falls, (6)

Martine 744 (CONN; FJ391919, FJ391931); Kakadu National Park, Maguk, where footpath crosses Barramundie Creek, (7) Tippery 155 (CONN); Mistake Billabong, c. 28 km from Arnhem

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Highway, (8) Tippery 142 (CONN); pool alongside road to Point Stuart, c. 38 km from Arnhem

Highway, (9) Tippery 140 (CONN); Queensland. c. 1 km E of Musgrave on road to Marina

Plains, (10) Tippery 164 (CONN). Western Australia. Edith Falls / Leliyn, upper pool, (11)

Tippery 119 (CONN). THAILAND: Trang. (12) Chansilpa s.n. (CONN; FJ391921, —).

N. beaglensis — AUSTRALIA: Western Australia. c. 61 km NE of Derby on Gibb

River Road, (1) Tippery 107 (CONN); c. 62 km NE of Derby on Gibb River Road, (2) Tippery

106 (CONN); Taylor‘s Lagoon, (3) Tippery 103 (CONN); Yabbagoody Claypan, c. 14.4 km E of

Derby on Gibb River Road, (4) Tippery 105 (CONN).

N. brevipedicellata — ZAMBIA: Luapula District. Lake Bangweulu, Renvoize 5604

(MO).

N. cambodiana — VIETNAM: Tây Ninh. Lò Gò-Xa Mát National Park, Regalado

1621 (MO; FJ391929, FJ391936).

N. cordata — U.S.A.: Alabama. Escambia County, Atmore, Little River State Forest,

(1) Tippery 256 (CONN). Connecticut. Middlesex County, Messerschmidt Pond, (2) Murray

99-044 (CONN; EF173027, EF173063); New London County, Glasgo, Glasgo Pond, (3) Tippery

80 (CONN); New London County, Voluntown, Hopeville Pond, (4) Tippery 79 (CONN); New

London County, Voluntown, Pachaug Pond, (5) Tippery 73 (CONN); Tolland County, above

Mansfield Hollow Dam, (6) Benoit 06-037 (CONN); Tolland County, Mansfield, Echo Lake, (7)

Tippery 2 (CONN; EF173028, EF173064); Windham County, Woodstock, Black Pond, (8)

Tippery 72 (CONN). Georgia. Clay County, Fort Gaines, G.T. Bagby State Park, (9) Tippery

216 (CONN); Early County, Kolomoki Mounds State Park, (10) Tippery 218 (CONN); Warren

County, Hamburg State Park, (11) Tippery 212 (CONN). New Hampshire. Rockingham

County, Lee, Weelwright Pond, (12) Wells s.n. 26 Sep 1972 (NHA; EF173029, EF173065).

N. coreana — SOUTH KOREA: Gangwon-do. Yeounpo-ri, Sonyang-myeon,

Yangyang-gun, (1) Hyunchur Shin s.n. (CONN). TAIWAN: Pu-hsing, Tao yuan Hsion, (2)

Chuang 3131 (UC).

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

N. crenata — AUSTRALIA: New South Wales. Northwest Plains, Yelarbon, ephemeral watercourse c. 40 km southeast of Yelarbon, on the Texas road, (1) Les 618 & Jacobs

8622 (CONN; EF173030, EF173066). Queensland. c. 1 km S of Yelarbon, (2) Jacobs 9435

(NSW; EF173032, EF173068); c. 127 km S of Lakeland on road to Mareeba, (3) Tippery 179

(CONN); c. 155 km W of Chillagoe on road to Normanton, (4) Tippery 181 (CONN); c. 185 km

W of Chillagoe on road to Normanton, (5) Tippery 180 (CONN). Western Australia. Russ

Creek, (6) Fryxell & Craven 3985 (UC). CULTIVATED. (7) Pagels s.n. (CONN; EF173031,

EF173067); (8) Pagels s.n. (CONN).

N. cristata — INDIA. northern India, (1) Pagels s.n. (CONN; EF173034, EF173070).

SRI LANKA: Eastern Province. Trincomalee District, Muthur, (2) Robyns 7269 (L). U.S.A.:

Florida. Collier County, in canal, (3) Smith s.n. (FSU; EF173033, EF173069).

N. disperma — AUSTRALIA: Western Australia. Anjo Peninsula, c. 44 km NW of

Kalumburu and c. 7 km E of Truscott Airport, Kalumburu Aboriginal Reserve, (1) Broun 4

(PERTH); N Kimberley, unnamed creek running into Pauline Bay, Vansittart Bay, (2) Forbes

2098 (PERTH).

N. elliptica — AUSTRALIA: Queensland. c. 19 km E of Musgrave on road to Marina

Plains, Tippery 171 (CONN).

N. exiliflora — AUSTRALIA: Northern Territory. Central Arnhem road, 10 km W of

Goyder River, (1) Cowie & Dunlop 8333 (DNA; EF173036, EF173072). Queensland. c. 3.5 km

E of Musgrave on road to Marina Plains, (2) Tippery 166 (CONN); c. 19 km E of Musgrave on road to Marina Plains, (3) Tippery 173 (CONN); c. 30 km N of Laura on road to Musgrave, (4)

Tippery 175 (CONN); c. 59 km N of Hann River roadhouse on road to Musgrave, (5) Tippery 177

(CONN); 15 km due E of Mount Molloy, freehold land in State Forest, (6) Halford Q326 (BRI).

N. ezannoi — BURKINA FASO: Oudalan. Some km SW of Gorom Gorom, Madsen

6073 (TUR).

N. fallax — MÉXICO: Querétaro. Amealco, Novelo & Ramos 3801 (MO).

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

N. furculifolia —AUSTRALIA: Northern Territory. Nitmiluk National Park, (1)

Michell 3389 (DNA). Western Australia. Edith Falls / Leliyn, rivulet on right-hand trail to upper pool, (2) Tippery 120 (CONN).

N. geminata — AUSTRALIA: New South Wales. Carroll‘s Creek, Tenterfield, (1)

Constable s.n. 02 Dec 1965 (UC; EF173037, EF173073). Queensland. Mount Moffatt National

Park, NW of Injune, (2) Bean 14273 (NSW); Tarong State Forest, (3) Bean 13183 (NSW;

EF173038, EF173074).

N. humboldtiana —BRAZIL: Rio de Janeiro. Lagoa de Jurubatiba, (1) Moreira,

Mogalhães, & Sigueira 70 (CONN). Roraima. BR-210, cerca de 30 km a oeste da BR-174, município de Boa Vista, (2) Bove et al. 1969 (CONN); BR-174, cerca de 28 km ao sul de

Rorainópolis, (3) Bove et al. 1978 (CONN). ECUADOR: Pichincha. Forest pond, Tinalandia

Hotel, E of Santo Domingo, (4) Pagels s.n. (CONN). MÉXICO: Veracruz. (5) Calzada 9328

(UC).

N. indica — AUSTRALIA: New South Wales. Lismore, (1) Jacobs 9395 (NSW;

EF173042, —). Northern Territory. Darwin, Knuckey Lagoons, (2) Tippery 124 (CONN); floodplain along old Jim Jim road, c. 85 km S of Annaburroo, (3) Tippery 153 (CONN);

Girraween Lagoon, (4) Tippery 128 (CONN); Howard Springs, private residence of D. Wilson,

(5) Tippery 131 (CONN); Kakadu National Park, Four-Mile Hole on Wildman River, (6) Tippery

145 (CONN); Kakadu National Park, Anbangbang Billabong, (7) Tippery 151 (CONN); Kakadu

National Park, Swamp near Alligator River upstream boat launch, (8) Tippery 150 (CONN);

Roadside pool, Girraween Road, (9) Tippery 129 (CONN). Queensland. c. 1 km E of Musgrave on road to Marina Plains, (10) Tippery 162 (CONN); c. 5 km N of Hann River roadhouse on

Laura to Musgrave road, (11) Tippery 159 (CONN); near Cairns, (12) Pagels s.n. (CONN;

EF173040, EF173076); road to town of 1770, Miriamvale, about 5 km N of Rd 24 Jct., (13) Les

541 & Jacobs 8518 (NSW;EF173041, EF173077); Topaz, private residence of J. and M.

Clarkson, (14) Tippery 158 (CONN). Western Australia. Adcock Gorge, 270 km NE of Derby,

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

(15) Tippery 108 (CONN); Adcock Gorge, 270 km NE of Derby, (16) Tippery 110 (CONN);

Kumbidgee Lodge, 10 km E of Katherine on road to Nitmiluk Gorge, (17) Tippery 117 (CONN);

Kununurra, Lily Creek Lagoon, (18) Tippery 100 (CONN); Leach Lagoon, 46 km SE of

Katherine on Stuart Highway, road N from rest/camping area, (19) Tippery 115 (CONN); Marlgu

Billabong, Parry Lagoons, c. 10 km S of Wyndham, (20) Tippery 113 (CONN); Saddler Springs, near Imtintji rest stop, 225 km NE of Derby on Gibb River Road, (21) Tippery 111 (CONN);

Taylor‘s Lagoon, (22) Tippery 101 (CONN). INDIA: Rajasthan. Bharatpur Wetlands, (23)

Pagels s.n. (CONN; EF173039, EF173075).

N. minima — AUSTRALIA: Northern Territory. c. 1 km W of Humpdy Doo on

Arnhem Highway, (1) Tippery 137 (CONN); 4 km W of Humpty Doo, (2) Tippery 138 (CONN);

Arnhem Land, (3) Short & Harwood 5018 (DNA; FJ391924, —); Berry Creek, McHenry Drive, off Stuart Highway, (4) Tippery 136 (CONN); Kakadu National Park, 5 km N of Arnhem

Highway on Wildman River Road, (5) Tippery 149 (CONN); roadside pool, Girraween Road, (6)

Tippery 130 (CONN). Western Australia. Adcock Gorge, 270 km NE of Derby, (7) Tippery

109 (CONN; FJ391932, —); Edith Falls / Leliyn, lower part of right-hand trail to upper pool, (8)

Tippery 121 (CONN; FJ391925, FJ391933); Edith Falls / Leliyn, upper pool, (9) Tippery 118

(CONN). CULTIVATED. (10) Pagels s.n. (CONN); private residence of D. Wilson, (11)

Tippery 132 (CONN).

N. montana — AUSTRALIA: New South Wales. Black Bob‘s Creek, (1) Moody 477

(CONN; EF173043, EF173078); Braidwood to Nerriga road, Black Bobs Creek, about 2.6 km by road SW of the Corang River, equals about 13 km direct line SSW of Nerriga, (2) Aston 1823

(NSW); Cecil Hopkins Reserve, Moss Vale Road, Bowral, (3) Jacobs 9950 (NSW); Paddys

River, Hume Highway, (4) Jacobs 9951 (NSW); SW end of Wingecarribee Swamp, c. 6.5 km due

WNW of Robertson, 2.5 km due NNW of Burrawang, (5) Kodela 208 (NSW). CULTIVATED.

Collected in Victora, (6) Pagels s.n. (CONN; EF173044, EF173079); Lake Nadungamba, Mount

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

Annan Botanic Gardens, (7) Jacobs 9376 (NSW; EF173045, EF173080); probable origin Otways,

(8) Jacobs 9949 (NSW); Hunter Valley, (9) Jacobs 9948 (NSW).

N. parvifolia — AUSTRALIA: Northern Territory. Arnhem Land, c. 44 km SSE of

Maningrida, (1) Cowie 8511 (DNA). Queensland. c. 100 km S of Coen, Musgrave road, (2)

Jacobs 9316 (NSW).

N. peltata — HUNGARY. (1) Wilstermann s.n. (CONN). INDIA: Jammu and

Kashmir. Srinagar, Golmarg, (2) Rodin 8236 (UC; FJ391920, —). SOUTH KOREA:

Changnyeong-gun. Gyeongsangnam-do, Uponeup (Upo) Marsh, (3) Hyunchur Shin s.n.

(CONN). U.S.A.: Connecticut. Tolland County, Columbia, private pond, (4) Tippery 198

(CONN). New York. Saratoga County, Stillwater, Hudson River, (5) Tippery 19 (CONN;

EF173046, EF173081); Washington County, Dresden, private boat launch, (6) Tippery 293

(CONN). CULTIVATED. Paradise Water Gardens, Whitman, Massachusetts, (7) Romanowski s.n. (CONN); Maryland Aquatic Nurseries, Jarrettsville, Maryland, (8) Tippery s.n. (CONN;

EF173047, —).

N. planosperma — AUSTRALIA: Northern Territory. Kakadu National Park, (1)

Craven 6544 (MO); Kakadu National Park, (2) Craven 6544 (NSW).

N. quadriloba — AUSTRALIA: Western Australia. About 0.3 km NE of Kalumburu

Mission in a shallow depression between community and airstrip, (1) Mitchell 3875 (NSW);

Gardner, c. 55 km N of Drysdale River, on the Kalumburu road, (2) Jacobs 8025 (NSW).

N. rautanenii — NAMIBIA. Ovamboland, Oshikango, Rodin 9385 (UC).

N. simulans — AUSTRALIA: Queensland. c. 1 km E of Musgrave on road to Marina

Plains, (1) Tippery 165 (CONN); c. 5 km N of Hann River roadhouse on Laura to Musgrave road,

(2) Tippery 160 (CONN); c. 59 km N of Hann River roadhouse on road to Musgrave, (3) Tippery

176 (CONN).

N. spinulosperma — AUSTRALIA: New South Wales. c. 13 km NNW of Collie, (1)

Aston 2878 (NSW); Oxley Highway, between Gilgandra and Warren, 14-15 km E of Collie, and

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

0.5 km E of the Berida-Innisfail road junction, (2) Aston 2880 (NSW; FJ391928, —).

Queensland. c. 96 km S of Surat on the Saint George road, (3) Les 616 & Jacobs 8605 (NSW;

FJ391926, FJ391934). Victoria. c. 16 km by road (14 km in a straight line) W of Saint Arnaud along the Wimmera Highway, (4) Aston 2869 (NSW). CULTIVATED. Collected in Victora,

(5) Pagels s.n. (CONN; FJ391927, FJ391935); Hunter Valley, (6) Jacobs 9947 (NSW).

N. spongiosa — AUSTRALIA: Northern Territory. 4 km W of Humpty Doo, (1)

Tippery 139 (CONN); Kakadu National Park, Buba Billabong, (2) Smith 3704 (NSW).

N. subacuta — AUSTRALIA: Northern Territory. 12 km NE of Gunn Point Road, on track to Melacca Swamp, Wilson 5076 (NSW).

N. thunbergiana — SOUTH AFRICA: Eastern Cape. Near Port Elizabeth, Pagels s.n.

(CONN; EF173048, EF173082).

N. triangularis — AUSTRALIA: Queensland. c. 3.5 km E of Musgrave on road to

Marina Plains, (1) Tippery 168 (CONN); c. 5 km N of Hann River roadhouse on Laura to

Musgrave road, (2) Tippery 161 (CONN).

Nymphoides sp. 1 — AUSTRALIA: Western Australia. Beverley Springs Station,

Brolga Swamp, Cowie 4390 (PERTH).

Nymphoides sp. 2 — AUSTRALIA: Western Australia. Mitchell Plateau, Barrett &

Barrett 2640 (PERTH).

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

APPENDIX 2.02 – Morphological and other non-molecular characters and states that were scored for species of Nymphoides (see Table 2.01).

1. Chromosome number (2n): 0 = 18, 1 = 27, 2=36, 3=54. 2. Duration: 0 = annual, 1 = perennial. 3. Growth habit: 0 = erect, 1 = floating-leaved. 4. Heterostyly: 0 = absent, 1 = present. 5. Unisexual flowers: 0 = absent, 1 = present. 6. Leaf diameter (longest dimension):

0 = < 2 cm, 1 = 2-10 cm, 2 = > 10 cm. 7. Floral axis internodes: 0 = condensed, 1 = elongated.

8. Flower number per node: 0 = < 4, 1 = 4-10, 2 = > 10. 9. Pedicel length: 0 = < 2 cm, 1 = 2-5 cm, 2 = > 5 cm. 10. Corolla diameter: 0 = < 1 cm, 1 = 1-2 cm, 2 = > 2 cm. 11. Petal number: 0

= < 5, 1 = 5, 2 = > 5. 12. Petal color: 0 = white (or purplish), 1 = yellow (or orange). 13.

Corolla throat color: 0 = white, 1 = yellow. 14. Petal margin: 0 = entire, 1 = laciniate, 2 = line of hairs. 15. Petal surface: 0 = glabrous, 1 = median wing, 2 = hairy throughout. 16. Corona: 0

= absent, 1 = single line of hairs, 2 = dense field of hairs. 17. Interstaminal glands: 0 = absent,

1 = present. 18. Interstaminal gland form: 0 = sessile, 1 = pedicellate. 19. Carpellary glands:

0 = absent, 1 = present. 20. Capsule length: 0 = < 3 mm, 1 = 3-7 mm, 2 = > 7 mm. 21. Seed number: 0 = < 5, 1 = 5-20, 2 = > 20. 22. Seed length: 0 = < 1 mm, 1 = 1-4 mm, 2 = > 4 mm.

23. Seed shape: 0 = orbicular (l/w < 1.15), 1 = ellipsoid (l/w > 1.15). 24. Seed lateral compression: 0 = uncompressed, 1 = compressed. 25. Seed cell projections: 0 = absent, 1 = evenly covering surface (one per epidermal cell), 2 = sparsely distributed, 3 = distributed on margin only. 26. Seed cell projection type: 0 = tapering, 1 = bulbous, 2 = densely packed together. 27. Seed cell margin: 0 = polygonal, 1 = interdigitate. 28. Seed caruncle: 0 = absent,

1 = present.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

FIG. 2.01 – Relationships among Menyanthaceae genera and recent nomenclatural changes (Tippery and Les 2009). Terminal clade identifiers represent recently accepted names, and bracketed genus names at right represent currently proposed generic circumscription.

FIG. 2.02 – Strict consensus (MP) phylogenetic tree of relationships among Nymphoides species, constructed using morphological character data (Table 2.01). The tree was rooted at N. crenata, following results from molecular data analyses.

FIG. 2.03 – Phylogeny of Nymphoides, constructed using nrITS data. The tree represents one of many most-parsimonious trees, with branch lengths indicated. Multiple accessions of the same species (Appendix 2.01) are differentiated by numbers in parentheses. Nodes that were resolved on the strict consensus (MP) tree are indicated with an asterisk (*).

FIG. 2.04 – Phylogeny of Nymphoides, constructed using matK/trnK data. The tree represents one of many most-parsimonious trees, with branch lengths indicated. Multiple accessions of the same species (Appendix 2.01) are differentiated by numbers in parentheses.

Nodes that were resolved on the strict consensus (MP) tree are indicated with an asterisk (*).

FIG. 2.05 – Strict consensus (MP) phylogenetic trees constructed using nrITS (left) or matK/trnK (right) data for Nymphoides. Except for N. indica (two accessions), one accession was analyzed per species (Appendix 2.01), indicated by numbers in parentheses. Central lines connect identical species or clades that were recovered consistently in both analyses. Nodes for which the resolution differed significantly between nrITS and matK/trnK analyses (see text) are circled, whereas nodes that were not significant are indicated with ‗n.s.‘. Nodes without such labels were congruent between nrITS and matK/trnK analyses.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

FIG. 2.06 – Reconstruction of ancestral states using maximum likelihood, onto strict consensus (MP) phylogenetic trees (see Fig. 2.05). A. Heterostyly (character 4, Appendix 2.02) reconstructed on the nrITS (top) and matK/trnK (bottom) trees. B. Inflorescence morphology, coded as the presence or absence of internodes (character 7), reconstructed on the nrITS (top) and matK/trnK (bottom) trees.

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

FIG. 2.01

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

FIG. 2.02

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

FIG. 2.03

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

FIG. 2.04

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

FIG. 2.05

Chapter 2: Phylogenetic relationships and morphological evolution in Nymphoides

FIG. 2.06

Chapter 3: Evolution of inflorescence morphology in Nymphoides

Chapter 3

Evolution of inflorescence morphology in Nymphoides

ABSTRACT

Plants of the genus Nymphoides have a growth habit in which floating leaves support flowers that open above water. Two major inflorescence morphologies exist in Nymphoides, an

‗expanded‘ type in which pairs of flowers occur at nodes with two leaves each, with nodes separated by expanded internodes, and a ‗condensed‘ type that consists of dense flower clusters that each are supported by a single floating leaf. In addition, the anomalous species N. peltata has a unique inflorescence type, in which pairs of floating leaves support clusters of flowers, with expanded internodes between successive nodes. I examined the developmental morphology of several representatives of the three Nymphoides inflorescence architectures, using dissection and cross-sectional data from inflorescence buds. Using these data to reconstruct the order and position of organ initiation, I corroborated prior evidence that inflorescence growth is sympodial, with each node terminating in a flower. Furthermore, I identified a repeating modular structure consisting of a terminal flower preceded by two leaves and two flowers, which has the same arrangement in all Nymphoides inflorescence morphology types. The types, which are considerably different in their gross morphology, were found to differ at the developmental level by only the relative elongation of internodes and/or expansion of leaves. In addition, two kinds of modular growth were identified: a ‗floral continuation axis‘ that reiterates the two-flower and two-leaf module, and an ‗inflorescence continuation axis‘ that develops into the floating leaf plus flower cluster in the condensed type, and the pair of floating leaves plus flower cluster in N.

Chapter 3: Evolution of inflorescence morphology in Nymphoides peltata. Simplified models that depict only the repeating structures of these continuation axes, when repeated diagrammatically, are able to reconstruct the gross morphologies of all three

Nymphoides inflorescence types.

INTRODUCTION

Aquatic plants make up only about two percent of all vascular plant species, yet they occur throughout the world in a wide variety of habitats and exhibit remarkable morphological diversity (Cook 1990). One particularly distinctive growth form of aquatic plants is the floating- leaved habit, in which leaves expand upon the water surface after originating from a submersed rootstock. The leaves tend to be broad, with adaxial stomata that allow them to obtain gases from the atmospheric side of the leaf, in contrast to terrestrial plants which often breathe through the underside of their leaves. Like many aquatic plants, floating-leaved species have flowers that open above water, where they accomplish pollination using biotic or abiotic vectors.

Menyanthaceae are a family of aquatic plants in which the largest and most widespread genus Nymphoides consists almost entirely of floating-leaved species. Species of other genera are variously aquatic, but generally they have an emergent wetland habit with leaves and inflorescences that support themselves above the water. Nymphoides species, however, typically do not grow erect, and their floating leaves are closely associated with inflorescences that either extend along the water surface or consist of dense flower clusters. Phylogenetic study of

Menyanthaceae revealed that Nymphoides species are monophyletic and evolved relatively recently (5-20 Ma) from an ancestor that lacked a floating-leaved habit (Chapter 1; Tippery et al.

2008). The presence of floating leaves in Nymphoides and their absence in related groups allows for phylogenetic comparisons and the reconstruction of steps that were involved in the evolution of the floating-leaved habit. Moreover, Nymphoides species have a specialized morphology

Chapter 3: Evolution of inflorescence morphology in Nymphoides whereby the floating leaves are integrated with the inflorescence and help to keep it afloat, which increases the value of comparing their morphology to related species without floating leaves.

The Nymphoides inflorescence. Two major inflorescence types are found in

Nymphoides, which differ in the form of associated leaves and the relative elongation of internodes (Fig. 3.01). Some species have inflorescences that are expanded (referred to herein as the ‗expanded type‘; Fig. 3.01A), with internodes separating pairs of flowers that are associated with two leaves, which are either two bracts (e.g., N. montana; Aston 1982) or a bract and a broad floating leaf (referred to herein as a ‗foliage leaf‘; e.g., N. aurantiaca; Sivarajan and Joseph

1993). Other species have a condensed inflorescence (the ‗condensed type‘; Fig. 3.01B), in which flowers all emerge from one position and are subtended by a single floating leaf (e.g., N. indica; Sivarajan and Joseph 1993). In these species the junction of stem and floating leaf petiole appears to be continuous and nearly indistinguishable from the petiole of vegetative leaves (i.e., leaves without an associated inflorescence), giving the false impression that flowers are borne on the petiole. A third type of inflorescence occurs only in N. peltata (Fig. 3.01C) and consists of nodes, each with two foliage leaves and a cluster of flowers, that repeat along the flowering stem and are separated by internodes (Sivarajan and Joseph 1993).

Of approximately 50 Nymphoides species (Chapter 2), 11 have an expanded inflorescence

(Table 2.01), and these occur only in Australia and tropical Asia. Expanded inflorescences are associated with flowers that have laciniate lateral petal wings, a glabrous petal surface (exclusive of the median wing), and typically a yellow or orange corolla. The condensed inflorescence characterizes roughly 40 species throughout the cosmopolitan range of the genus, and is associated with yellow or white corollas, a ciliate or glabrous petal surface, and petal wings that may be entire to dissected. Flowers of the N. peltata inflorescence type resemble those of species with expanded inflorescences (i.e., with yellow petals with laciniate margins). Two Nymphoides species deserve additional mention because they grow with an erect or recumbent habit and do not have floating leaves supporting their inflorescences (although floating leaves may be present).

Chapter 3: Evolution of inflorescence morphology in Nymphoides

Despite their distinct growth habit, N. cambodiana and N. exiliflora have inflorescences that most closely resemble the expanded type (i.e., their flowers emerge in pairs subtended by two bracts, which have internodes between them).

Inflorescence habit is not always consistent within a Nymphoides species. Specifically, species that typically have an expanded inflorescence occasionally lack elongated internodes, due to unknown genetic or environmental cues. For example, N. crenata often has clustered flowers, although in many specimens the internodes clearly have expanded to reveal the paired flower arrangement (pers. obs.). For the purpose of this study, species that have the ability to produce elongated internodes are categorized as having an expanded inflorescence, whereas species in the condensed inflorescence category never have been observed with elongated internodes.

Possibly owing to the variability within species or to the preponderance of condensed- inflorescence species worldwide, the different inflorescence types never received any taxonomic designation, except that the anomalous species N. peltata was assigned to a separate section by

Grisebach (1839). The inflorescence categories presented herein were first outlined by Aston

(1982), who noted a correlated pattern of inflorescence and floral morphology in Australian

Nymphoides species. She categorized newly described species under either the ―geminata group‖

(i.e., with expanded inflorescences) or the ―indica group‖ (i.e., with condensed inflorescences).

Species that occur outside of Australia (except N. peltata) also conform to the same basic categories, which thus can be applied to the genus worldwide.

Some Nymphoides species exhibit a further modification of the inflorescence by branching. In N. peltata and species with the expanded inflorescence type, more proximal inflorescence nodes often have branching nodes that are subtended by a single floating leaf. A more complex branching system exists in the condensed inflorescence group, some species of which (e.g., N. indica) can produce fertile leaves (i.e., floating leaf plus flower cluster) from the nodes of existing flower clusters.

Chapter 3: Evolution of inflorescence morphology in Nymphoides

Morphological study. Previous work on elucidating inflorescence morphology in

Nymphoides was conducted by Wagner (1895) on N. peltata, and by Goebel (1891) on several species including N. aurantiaca and N. indica. These species span all three categories of inflorescence morphology and thus provide a useful framework for undertaking an updated study.

Because the early work was illustrated only with line drawings of dissections and cross sections, more rigorous (i.e., photographic) documentation is necessary to verify the associated interpretations. The early studies did convincingly disclose the sympodial nature of the

Nymphoides inflorescence, by demonstrating in every species examined that the main axis terminates in a flower, followed by subsequent growth from the axils of foliage leaves or bracts.

In both N. peltata and N. aurantiaca, each node was observed to contain a terminal flower and at least one other flower on the axis leading to the first. Species of the condensed inflorescence type were not examined in similar cross-section, although the initiation of fertile leaves in N. indica was shown to be sympodial, with successive fertile leaves arising from the axils of bracts on the rhizome (Goebel 1891). In this study, the data of Goebel (1891) and Wagner (1895) were used to provide testable hypotheses regarding the relative positions and initiation of organs in

Nymphoides species.

Phylogenetic relationships have been determined for Nymphoides using molecular data

(Chapter 2), in a study that evaluated all expanded-inflorescence species but one, 20 species with condensed inflorescences, and N. peltata. Both nuclear and chloroplast data resolved a large clade containing only species with the condensed type, and at least one other clade with condensed inflorescences that was interspersed among a grade of mostly expanded-inflorescence species. Nymphoides peltata also resolved within the grade of species with expanded inflorescences, not as sister to the rest of the genus as its morphological uniqueness and prior classification as a monotypic section (Grisebach 1839) would suggest. Ancestral state reconstruction determined that the N. peltata inflorescence type evolved from an ancestor that had

Chapter 3: Evolution of inflorescence morphology in Nymphoides an expanded inflorescence, as did the condensed inflorescence morphology on at least two occasions (Chapter 2).

Several aspects of the inflorescence morphology and phylogenetic relationships in

Nymphoides require further research. Primarily, the multiple evolutionary origins of groups with the condensed inflorescence type indicate that it should not be considered a homologous trait without additional evidence. In addition, the morphological data currently available are insufficient to resolve the positional relationships among organs. Another complication relates to the variation within species, some of which can produce inflorescences that appear to be either expanded or condensed under different conditions. In order to determine the positional homology of inflorescence organs, and to understand the nature of intraspecific variation, it is necessary to undertake a more directed study that examines multiple species and obtains more explicit morphological data.

The aims of this study were (1) to identify morphological points of comparison that could be evaluated among species in the different inflorescence categories, (2) to establish similarities and differences among species with respect to the relative positions and development of organs, and (3) to propose mechanisms for the evolutionary transitions between inflorescence types.

These questions were addressed using data from dissections and cross sections of inflorescence buds of multiple species with condensed or expanded inflorescence types and N. peltata.

MATERIALS AND METHODS

Developing inflorescence buds were collected and preserved in 70% ethanol. Plants of

N. peltata were collected on 18 August 2009 from Rogers Island Wildlife Management Area,

New York (see Chapter 4), and buds of N. aquatica and N. indica were obtained on 15 May 2009 from material growing in the University of Connecticut research greenhouse, originally collected by L. Benoit in Florida and by W. Pagels in Queensland, Australia, respectively (see Appendix

Chapter 3: Evolution of inflorescence morphology in Nymphoides

2.02). Material of other species was collected in Australia in 2008, concurrent with specimens that were used for molecular work (Appendix 2.02). Buds were dissected to a size appropriate for microtomy (5-10 mm in length), then prepared for embedding and sectioning following Ruzin

(1999). Samples were taken through a dehydration series to 100% ethanol in 5% increments, each lasting at least 24 hours at room temperature. Following dehydration, samples were taken to

100% tert-butyl alcohol (TBA) in a series of ethanol dilutions (50% / 75% / 75% / 100% / 100%

TBA), then to 67% TBA and 33% paraffin oil, all at room temperature. Next, samples were transferred to a 60°C incubator where liquid Paraplast® Plus (McCormick Scientific, Richmond,

Illinois) was added, first by replacing one-half of the liquid volume with Paraplast (repeated twice at 24-hour intervals) and then replacing the entire volume with fresh Paraplast (repeated once after 24 hours). The samples were then embedded in a tray made from cardstock paper, where they were oriented and cooled slowly until solid.

Embedded samples were trimmed with a razor blade and mounted to wooden blocks using additional Paraplast. Samples were sectioned on a American Optical / Spencer ―820‖ rotary microtome (American Optical Corporation, Buffalo, New York) at 7 m thickness, with the resulting ribbons floated on dH2O for one minute at 50°C then blotted of water and dried overnight at 30°C to slides pre-coated with Haupt‘s adhesive (Haupt 1930). Samples were de- paraffinized by immersion (2x; 5 min each) in 100% CitriSolv® (Decon Labs Inc., King of

Prussia, Pennsylvania), then 100% ethanol (2x; 4 min each), then taken through a series to dH2O

(95% / 75% / 50% / 30% / 0% ethanol; 4 min each), then stained for 60 minutes in 1% safranin O

(aqueous). Next they were taken rapidly through a series to 95% ethanol (30% / 50% / 75% /

95% ethanol) and stained for 30 seconds in 0.5% fast green (in 95% ethanol). Samples were then rinsed in 100% ethanol (2x), then 1:1 ethanol:xylene (2x), then 100% xylene (2x). Cover slips were adhered to slides using one drop of Permount® adhesive (Fisher Scientific, Pittsburgh,

Pennsylvania) and allowed to dry at 30°C for one week. Slides were viewed on a Leica Wild

M10 microscope (Leica Microsystems Inc., Bannockburn, Illinois). Digital photographs were

Chapter 3: Evolution of inflorescence morphology in Nymphoides taken using the software qCapture Pro ver. 6.0 (QImaging, Surrey, British Columbia) with a

MicroPublisher 3.3 RTV camera (QImaging).

RESULTS

Expanded inflorescence type. Previously published depictions of N. aurantiaca

(Goebel 1891), representative of the expanded inflorescence type, show a repeating, modular pattern whereby each node consists of a lower foliage leaf and an upper bract (Fig. 3.02A).

Diagrammatic interpretation of this pattern (Fig. 3.02C) shows a sympodial arrangement, in which the axis terminates in a flower at each node (F1, orange). A second flower (F2, yellow) is produced in the axil of a upper bract (L2) that is subopposite to the foliage leaf (L1). The flowering axis (blue) continues from the axil of the foliage leaf, repeating the organ number and arrangement of the current node.

Cross-sections of N. aurantiaca obtained during this study corresponded to the depiction by Goebel (1891), with respect to the number and positions of organs (Fig. 3.03A-B). The first terminal flower (TF1) is preceded by a distal leaf (DL1), which subtends an axillary flower (AF1).

Subopposite to DL1 is a proximal leaf (PL1), in the axil of which is a cluster of flowers that represents the floral continuation axis, i.e., subsequent nodes that replicate the organ number and arrangement of the current node. In plants of the expanded inflorescence type, internodes are present between subopposite pairs of flowers. In Fig. 3.03, organs with the same numerical subscript remain at the same node and are separated by an internode from organs with the consecutive number subscript. For example, the main axis terminates again at the next node with a flower (TF2) preceded by two leaves, a proximal leaf (PL2) that subtends the next continuation axis and a distal leaf (DL2) that subtends an axillary flower (AF2). In the bud depicted in Fig.

3.03A, four terminal flowers are visible (TF1-TF4), each associated with the same organs in the same relative positions.

Chapter 3: Evolution of inflorescence morphology in Nymphoides

Sections of N. crenata, which also has an expanded inflorescence, showed the same pattern, (Fig. 3.04A-B), in which subopposite pairs of flowers (TF and AF) and their associated leaves (DL and PL) repeat over successive nodes. The section of N. crenata (Fig. 3.04A) depicts slices through various portions of different flowers. Most often, flowers are seen as cross- sections through the pedicel (e.g., TF1, AF1), in which are visible several vascular traces (darker blue areas) and numerous air space tissues (i.e., lacunae, seen as empty white areas). Other flowers show whorls of floral organs (e.g., AF2, AF3, the former sliced tangentially), including the unilocular gynoecium, five stamens, five valvate corolla lobes, and five imbricate calyx lobes.

In N. crenata, the form of the proximal leaf differs from N. aurantiaca; in the former this leaf is reduced to a bract, and in the latter the proximal leaf expands to be a foliage leaf. These different forms are indistinguishable in cross-section, however, because both types of leaves have sheathing bases. In addition, N. crenata often varies with respect to the elongation of internodes, and quite often specimens or illustrations (e.g., Aston 1973) show no elongation.

In addition to cross-sections, material of species with expanded inflorescences was dissected to verify the interpretation of cross-sectional data. Two species were examined, N. aurantiaca (Fig. 3.03C) and N. exiliflora (Fig. 3.04C), the latter of which has an erect habit but otherwise conforms to the model of the expanded inflorescence type. For both species, the pattern observed in cross-section for the expanded inflorescence type was evident also in dissected material. Specifically, the buds of both species examined had a large, conspicuous terminal flower (TF1), which was preceded by a distal leaf (DL1) subtending an axillary flower

(AF1) and a proximal leaf (PL1) subtending the continuation axis. Some of the organs produced at subsequent nodes are visible within the continuation bud, most conspicuously the next pair of leaves (PL2 and DL2).

Condensed inflorescence type. Cross-sections of N. indica, representative of the condensed inflorescence type, showed a pattern similar to what was observed in species of the expanded inflorescence type (Fig. 3.05A-B), namely that each terminal flower (TF) is preceded

Chapter 3: Evolution of inflorescence morphology in Nymphoides by two leaves (both in the form of bracts), the proximal one (PL) subtending a continuation axis and the distal one (DL) subtending a single flower (AF). In the cross-section depicted, the first terminal flower (TF1) is preceded by a distal leaf (DL1) subtending an axillary flower (AF1); the lower, proximal leaf probably was dissected away, in which case the outermost visible bract (PL0) represents the proximal leaf on the axis of a flower at the preceding node (not visible). Although the proximal leaf of the TF1 axis is missing from the section, the pattern established for the expanded inflorescence type is fully apparent on the TF2 axis, with DL2 subtending AF2 and PL2 subtending the continuation axis. This pattern continues also on the TF3 axis. The use of the term ‗node‘ and repetition of the numbering scheme from the expanded inflorescence type are used as a convenience in reference to the overall similarity of organ positions in the condensed type, although in this inflorescence type none of the internodes on the continuation axis elongate.

Nymphoides aquatica, which also has a condensed inflorescence, showed the same pattern as N. indica (Fig. 3.06A-B), except that in the slide shown, the repeating pattern is preceded first by a terminal flower axis (TF0) that has in the axil of its upper leaf (DL0) not one flower but a cluster that appears to replicate the modular pattern of other nodes. Beginning at the flower labeled TF1, the pattern is more regular and again shows the terminal flower (TF1) preceded by a distal leaf (DL1) subtending an axillary flower (AF1) and a proximal leaf (PL1) subtending the continuation axis, in which the pattern repeats. Bud dissections of N. aquatica

(Fig. 3.06C) also showed two bracts associated with the terminal flower (TF1), the lower one

(PL1) subtending the continuation axis (in which two flowers, AF2 and TF2, are visible) and the upper one (DL1) subtending a single flower (AF1). The inflorescence architecture of species with condensed inflorescences thus showed the same numbers of bracts and flowers, in the same general arrangement, as species with expanded inflorescences, with the only apparent difference being the elongation of internodes between subopposite pairs of flowers.

Nymphoides peltata type. Sections of N. peltata illustrated by Wagner (1895; Fig.

3.02B,D) depict a terminal flower (F1, orange) preceded by two large foliage leaves, the upper of

Chapter 3: Evolution of inflorescence morphology in Nymphoides

which (L2) subtends a cluster of flowers (yellow), and the lower leaf (L1) subtends a continuation axis (blue). Wagner (1895) described the continuation axis, contained in the axil of L1, as the first of three elements axillary to F1, with the second being a flower in the axil of L2 (F2) and the third a flower (F3) subtended by a bract, which itself is the third leaf (L3). The final flower (F4) was depicted as axillary on the axis leading to F2 and subtended by a bract (L4). Under this interpretation, the proximal foliage leaf (L1) subtends a bud that replicates the current node and represents the next node to be produced on the inflorescence, whereas the distal foliage leaf (L2) subtends an axis of only two flowers.

My examination of inflorescence cross-sections of N. peltata (Fig. 3.07A-B) revealed the same structures depicted by Wagner (1895), in approximately the same relative positions.

However, cross-sections obtained for this study show an additional cluster of organs that was not accounted for in earlier illustrations. Several flower buds were observed in the axil of the upper foliage leaf (PL1), but notably the flowers were arranged similarly to those of N. aurantiaca.

Specifically, two leaves exist on the axis of the terminal flower (TF1, ignoring for now the lower foliage leaf BL), one of which (PL1) subtends a cluster of flowers and another (DL1), which has the form of a bract, subtends only one flower (AF1). The cluster of flowers in the axil of the upper foliage leaf (PL1) repeats the pattern of the other inflorescence types, with one terminal flower (TF2) preceded by two leaves (PL2 and DL2), etc. Similarly to the condensed inflorescence type, internodes do not elongate in this portion of the inflorescence, and apart from the upper foliage leaf (PL1), subsequent leaves have the form of bracts. Dissection of the floral continuation bud (Fig. 3.07C) supported the interpretation given, showing the terminal flower

(TF1) preceded immediately by a bract (DL1) and axillary flower (AF1). Lower down on the axis leading to TF1 are flowers in the axil the upper foliage leaf (removed), two of which (TF2 and

AF2) are apparent in the photograph.

The lower foliage leaf (BL, Fig. 3.07A) of N. peltata subtends a continuation axis of a different sort than what was described previously. Other continuation axes, which were

Chapter 3: Evolution of inflorescence morphology in Nymphoides distinguished under the term ‗floral continuation axis‘, consisted of repeating elements with two flowers (TF and AF) and two leaves (DL and PL), with or without an expanded internode between them. However, the continuation axis in the axil of the basal leaf (BL), here termed the

‗inflorescence continuation axis‘ (IA), replicates the entirety of the current node, including all leaves (BL, PL1, DL1) and all organs contained in the repeating element axillary to PL1. The nodes observed in the overall morphology of N. peltata (Fig. 3.01C) thus represent the repetition of complex elements, with each node containing two types of continuation axes.

Dissection of the N. peltata inflorescence bud revealed the same arrangement of structures that was observed in cross-section (Fig. 3.08). Specifically, the terminal flower (TF1) is preceded by two foliage leaves (removed), the lower of which subtends the inflorescence continuation axis (IA1), and the upper subtends the floral continuation axis (AF1, TF2, AF2, etc.).

The floral continuation axis (Fig. 3.08B, seen from the opposite side than Fig. 3.08A) contains a cluster of flowers, in which are visible two pairs of flowers (see also Fig. 3.07C). The inflorescence continuation axis (Fig. 3.08C) contains the next two foliage leaves (BL2 and PL2B), and further dissection (Fig. 3.08D) shows the repetition of the structures of the current node, including another terminal flower (TF2B) and inflorescence continuation axis (IA2)

Additional examination of the total inflorescence architecture of species with condensed and expanded inflorescences revealed structures that were similar, if not homologous to the inflorescence continuation axis of N. peltata.

The inflorescence continuation axis. Another feature that is expressed variously across species of Nymphoides involves a secondary growth axis. In all species examined, some portion of the inflorescence comprises a floral continuation axis, which produces successive repeating units of two flowers and two leaves each. In N. peltata another axis, the inflorescence continuation axis, is required to repeat the basic modular unit of the inflorescence, i.e., two subopposite floating leaves and a cluster of flowers. The description presented thus far of inflorescence buds of the condensed inflorescence type has not accounted for a major factor of

Chapter 3: Evolution of inflorescence morphology in Nymphoides the inflorescence, namely the large floating leaf that supports the flower cluster. This leaf is integral to the inflorescence and must be considered when modeling the developmental morphology of species. In N. indica and other species with condensed inflorescences, the floating leaf subtends a bud, which Goebel (1891: p. 122) called the ‗vegetative accessory shoot‘

(―vegetativer Beispross der Inflorescenz‖). This bud grows out to replicate the larger inflorescence unit of floating leaf plus flower cluster, and thus it can be considered analogous, if not homologous, to the inflorescence continuation axis in N. peltata.

The accessory shoot bud of N. indica (IA, Fig. 3.09) grows, as Goebel (1891) described, in a position axillary to the floating leaf (BL). The floating leaf displaces the inflorescence axis to such a degree that its petiole (BL) appears to be a continuation of the inflorescence stem (BI), but the terminal position actually is occupied by the floral continuation axis (FA). In adaxial view (Fig. 3.09B), the floral (FA) and inflorescence (IA) continuation axes are seen to emerge from within the sheathing base of the floating leaf. With the sheathing leaf base and several associated floral bracts removed (Fig. 3.09C), flower arrangement follows the pattern described above, with flowers occurring in successive pairs that occupy terminal (TF) and axillary (AF) positions. The arrangement of bracts is less obvious due to the dense clustering of flowers, and can be interpreted more accurately using cross-sections (e.g., Fig. 3.05).

Further dissection of the accessory shoot bud (Fig. 3.10) revealed a fully sheathing (i.e., around the entire circumference of the bud) prophyll (PP1) that encloses two elements, a fertile leaf and an accessory bud. The fertile leaf occupies the terminal position and consists of a basal internode (BI1, Fig. 3.10B-C), a floating leaf (BL1), and the organ cluster that contains the floral

(FA1) and inflorescence (IA1) continuation axes. Axillary to the prophyll PP1 is a bud that replicates the structure of the current node, with prophyll (PP2) and fertile leaf (BI2, BL2, FA2,

IA2, Fig. 3.10D). Continuing growth from the accessory shoot bud, therefore, also is sympodial, with the axis terminating in a fertile leaf and subsequent growth from the same node being possible from the axillary accessory shoot bud.

Chapter 3: Evolution of inflorescence morphology in Nymphoides

Goebel (1891) reported that the floating leaf in condensed inflorescence species was actually the second of two leaves associated with the inflorescence, the first being a bract.

Wagner (1895) offered a contrary explanation, that the foliage leaf was the lower of the two. It is uncertain exactly what part of the developing plants each author was discussing, but it is possible that they both were correct. Accessory shoot buds examined in N. indica were found to have an enveloping bract, within which was a concentric series of similar bracts, each with a fertile leaf in the terminal position (Fig. 3.10). The arrangement of these organs corresponds to the description by Goebel (1891). At a different stage in development, the foliage leaf itself precedes a terminal inflorescence bud and contains in its axil the next vegetative accessory shoot bud (Fig. 3.09A), thus showing the foliage leaf to be lower as reported by Wagner (1895).

A structure analogous to the inflorescence continuation axis was not clearly evident in species with the expanded inflorescence type, although the structure may be reflected in branching nodes of the inflorescence. The axillary buds of branching nodes grow out to replicate the appearance of the floral continuation axis, and thus they serve the same function as inflorescence continuation axes in the other inflorescence types (i.e., to duplicate larger portions of the inflorescence). In each case, the inflorescence continuation axis grows out to replicate the entire node at hand, and so differs from the floral continuation axis, which in N. peltata and the condensed inflorescence type produce only a flower cluster (without internodes) and in the expanded inflorescence type produce pairs of flowers with internodes between them.

DISCUSSION

Dissection and microtomy of Nymphoides inflorescence buds clarified several interpretations that had been reported by previous authors (Goebel 1891; Wagner 1895) and allowed for further interpretation regarding the homology of structures across different inflorescence types. Diagrams of the different Nymphoides inflorescence types reveal overall

Chapter 3: Evolution of inflorescence morphology in Nymphoides similarity among species and the relatively minor differences required to relate them to the consensus diagram (Fig. 3.11A). Two kinds of repeating structures were observed in

Nymphoides, which represent the floral (FA) and inflorescence (IA) continuation axes. In the former, modular units of two leaves (PL, DL) and two flowers (AF, TF) repeat, with or without internodes between them. The lower of the two leaves (PL) subtends the further continuation of the floral pattern (FA). The distal internode (DI) remains condensed in all inflorescence types, but elongation of the proximal internode (PI) varies among inflorescence types. The inflorescence continuation axis (IA), contained in the axil of a basal leaf (BL), repeats the organs of the entire inflorescence, including BL and IA. The basal internode (BI) between successive units of the inflorescence axis elongates in all Nymphoides species.

The structures depicted in Fig. 3.11A were observed in all species of the three different inflorescence types, in the same relative positions. Furthermore, the positional relationships of the two continuation axes were consistent across inflorescence types. In the expanded inflorescence type (Fig. 3.11B), only the distal internode (DI) is altered from the consensus diagram, and the condensation of this internode occurs in all inflorescence types. Species with the expanded type differ among themselves, however, in whether their proximal leaf (PL) expands to be foliose (e.g., N. aurantiaca; Sivarajan and Joseph 1993) or takes the form of a bract

(e.g., N. montana; Aston 1982). The condensed inflorescence type (Fig. 3.11C) differs from the consensus model in having a reduced proximal internode (PI), and the basal leaf (BL) comprises the characteristic floating leaf of condensed-inflorescence species. Species of the condensed inflorescence type vary in whether they expand the inflorescence continuation axis (IA) to produce additional fertile leaves. These leaves (which some authors term ‗stolons‘; Chapter 2) may be produced facultatively (e.g., N. indica; Sivarajan and Joseph 1993) or be entirely absent

(e.g., N. thunbergiana; Raynal 1974). The inflorescence of N. peltata (Fig. 3.11D) has a reduced proximal internode (PI) and two types of proximal leaves (PL), the first of which expands to be

Chapter 3: Evolution of inflorescence morphology in Nymphoides foliose (depicted in Fig. 3.11D), whereas subsequent proximal leaves comprise bracts in the flower cluster (see Fig. 3.12F,I).

Inflorescence architecture reconstruction. Repetition of the pattern established for the various Nymphoides inflorescence types allows for reconstruction of the overall morphological relationships within the inflorescence, as illustrated in Fig. 3.12. Simplified diagrams of the three inflorescence types are depicted in the top row (Fig. 3.12A-C). Adding two subsequent units to the floral continuation axis (FA1) results in further pairs of flowers with their associated leaves

(Fig. 3.12D-F). In the expanded inflorescence type (Fig. 3.12D), these pairs are separated by an elongated proximal internode (PI), whereas in the condensed type (Fig. 3.12E) and N. peltata

(Fig. 3.12F) the additional pairs simply augment the cluster of flowers, without having visible internodes between them. Elaboration of the inflorescence continuation axis (IA1) reveals different types of repeating structures (Fig. 3.12G-I). In the expanded inflorescence type, the inflorescence continuation axis comprises another unit of paired flowers along a sympodial axis

(Fig. 3.12G). Growth of the inflorescence axis in condensed-inflorescence species produces an additional fertile leaf, along with its associated flower cluster (Fig. 3.12H). In N. peltata, inflorescence axis growth is required to produce the repeating unit of two subopposite foliage leaves (Fig. 3.12I), in contrast to the other inflorescence types where growth of the inflorescence axis could be considered facultative (in species that are able to elaborate the axis).

In addition to comparing specific organ relationships, a broader general pattern has emerged for species of Nymphoides, in which the inflorescence axis terminates frequently, with growth continuing from axillary buds. The nature of axillary growth varies, and the two most conspicuous examples are presented here, namely the floral and inflorescence continuation axes.

However, at least one other type of organization was uncovered during this project (Fig. 3.06A), in which the terminal flower was immediately preceded by two continuation axes, one having the form of a floral continuation axis and another that had an uncertain floral arrangement. Minor

Chapter 3: Evolution of inflorescence morphology in Nymphoides variants in inflorescence organization were not considered within the scope of this project, but they would be important to evaluate in future work.

Under the generalized interpretation of the Nymphoides inflorescence structure (Fig.

3.11A), the inflorescence would be categorized as cymose, with each axis terminating in a flower

(TF) and continuing to grow from the axil of preceding leaves (Weberling 1989). Flowers and branches that originate from the axis leading to the terminal flower would be considered paraclades of the first order. There are three different forms of paraclades in Nymphoides, which are the axillary flower (AF), the floral continuation axis (FA), and the inflorescence continuation axis (IA). Elaboration of the basic structure for different Nymphoides inflorescence types (Fig.

3.12) produces paraclades of the second and further orders. For example, in the expanded inflorescence diagram (Fig. 3.12G), up to three orders of paraclades are shown on the elaborated continuation axes. The floral continuation axis has two orders of paraclades that each terminate in a flower (TF2 and TF3), and the inflorescence continuation axis has three orders, also with terminal flowers (TF2B, TF3B, TF4B). In these diagrams the subscript number indicates the node to which organs belong, and the paraclade order of any terminal flower corresponds to the subscript number minus one. Each successive order of paraclade reproduces some portion of the main axis, which was simplified in Fig. 3.11A.

Evolution of inflorescence types. The fact that phylogenetic analyses reconstruct multiple origins for the condensed inflorescence type in Nymphoides (Chapter 2) is less surprising in light of the discovery that the inflorescence types differ from each other by relatively minor degrees of internode elongation and leaf elaboration. It is conceivable that selective forces favoring either a condensed or expanded inflorescence could readily prompt the conversion of one type into another with relatively minor alterations to their developmental pathways. Species with the condensed inflorescence type are more numerous and widespread than the expanded- inflorescence species, which grow only in Australia and tropical Asia. Perhaps the relative abundance and geographic range of species with condensed inflorescences reflects a greater

Chapter 3: Evolution of inflorescence morphology in Nymphoides selective advantage for this type of morphology, or it might also indicate a genetic canalization of the pathway leading to condensed inflorescences. As mentioned previously, several species with expanded inflorescences occasionally do not elongate their internodes. However, none of the well-studied species with condensed inflorescences has been observed conversely with elongated internodes.

All Menyanthaceae other than Nymphoides have an erect, expanded inflorescence, except for a small number of anomalous species in Liparophyllum (sensu Tippery and Les 2009), which possess few, solitary, or clustered flowers, and Ornduffia submersa, which produces lax panicles and flowers that lie upon the water surface (Tippery and Les 2009). The prevalence of expanded inflorescences in species of the basal grade of Nymphoides species supports the reconstruction of this type as ancestral for the genus (Chapter 2). Species in related genera are similar in having expanded inflorescences (although sessile flowers are clustered together in L. capitatum and L. congestiflorum), and these are a valuable point of comparison to the Nymphoides inflorescence.

Patterns of organ arrangement in related genera apparently do not exhibit the same organization that occurs in Nymphoides, however. Species of Nymphoides that have the expanded inflorescence type always have flowers arranged in subopposite pairs, but paired flowers are rarely observed in the sister genus Liparophyllum or other Menyanthaceae genera (Tippery and

Les 2009). In addition, the overall architecture of most species outside of Nymphoides appears to be paniculate, in which bracts subtend flowers or branches but overall growth is not sympodial.

The issue of whether inflorescence nodes are sympodial, however, has not been addressed previously in those genera related most closely to Nymphoides (e.g., Liparophyllum,

Ornduffia, Villarsia), but a study of the more distantly-related Menyanthes trifoliata (Sjörs 1988) did reveal its inflorescence initiation to be sympodial, with subsequent rhizome growth continuing from an axillary position. Sympodial initiation of the inflorescence also was reported for N. indica (Goebel 1891) and N. peltata (Wagner 1895). Although no Liparophyllum species were analyzed for this study, dissection of O. umbricola revealed a pattern strikingly similar to

Chapter 3: Evolution of inflorescence morphology in Nymphoides what has been described here for Nymphoides (Fig. 3.13). Specifically, the inflorescence bud contains a terminal flower (TF1) preceded by two bracts; the lower bract (PL1) subtends a continuation bud (FA1) and the upper bract (DL1) subtends a flower (AF1). Insufficient material was available to obtain cross-sections, though, and further study will be required to characterize the inflorescence bud of O. umbricola and other outgroup species properly.

The relative adaptive benefits of expanded or condensed inflorescences should be considered when evaluating the abundance, distribution, and phylogenetic relationships of species with different inflorescence types. The expanded inflorescence offers a greater spatial spread over the water surface and, through the localized activity of pollinators, might increase the number of neighboring individuals with which one ramet can exchange pollen. In addition, spatially separate flowers produce fruits that likewise are separated, allowing for greater local seed dispersal. The condensed inflorescence has the advantage of associating many flowers closely with one floating leaf, which can be particularly large relative to the individual flower size

(e.g., N. indica) and thus confer greater stability to the supported inflorescence. Furthermore, many species with clustered inflorescences can produce roots from their inflorescence nodes.

Consequently, if the combined leaf and flower cluster becomes separated from the rhizome it becomes a propagation unit, possibly bearing seeds (in attached fruits) in addition to having the potential to establish vegetatively. The unique inflorescence of N. peltata combines features of both inflorescence types, having the clustered flowers and large, supportive leaves of the condensed type but also spreading over the water surface like the expanded type.

Clearly a variety of reproductive strategies exists in Nymphoides, and these are facilitated to some degree by the various inflorescence architectures that are present in the genus. The different types of inflorescence architecture have different adaptive advantages, and selective pressure may have driven transitions from one inflorescence type to another in the evolutionary history of Nymphoides. At least two independent origins of the condensed inflorescence type have been reconstructed phylogenetically, perhaps driven by conditions that favored the

Chapter 3: Evolution of inflorescence morphology in Nymphoides condensed-type morphology. The transition from expanded to condensed morphology, although it appears to be a major evolutionary event, was determined in this study to have resulted from relatively minor developmental changes, specifically the suppression of internode elongation and the elaboration of floral-associated leaves as bracts. The overall similarity among Nymphoides species that I uncovered also enables a comparison to species outside the genus, which preliminary data suggest have a similar pattern of organ development to Nymphoides.

Understanding the morphological relationships of floating-leaved Nymphoides species to their nearest relatives with emergent growth forms will allow for the complete reconstruction of habit evolution in this diverse, widespread, and ecologically successful genus.

ACKNOWLEDGEMENTS. I am grateful to Jessica Budke and Cynthia Jones for patiently instructing me in the use of plant morphology techniques, providing helpful feedback on earlier drafts of the morphology project, and allowing me to use their reagents, equipment, and laboratory space. I benefitted also from attending the MORPH minicourse ―Homology:

Conceptual and Historical Integration from the Morphological to the Molecular‖, funded by the

National Science Foundation, and engaging other participants in discussion. Clint Morse and the

University of Connecticut greenhouse staff were a great help in keeping plants alive that were later used for morphology work. This project also relied on the generosity of the National

Herbarium of New South Wales and the late Surrey Jacobs, who enabled me to collect material of the Australian Nymphoides species that was suitable for morphology work.

Chapter 3: Evolution of inflorescence morphology in Nymphoides

LITERATURE CITED

Aston, H. I. 1982. New Australian species of Nymphoides Séguier (Menyanthaceae). Muelleria 5:

35-51.

Cook, C. D. K. 1990. Aquatic Plant Book. The Hague: SPB Academic Publishing.

Goebel, K. 1891. Morphologische und biologische Studien VI: Limnanthemum. Annales du

Jardin Botanique de Buitenzorg 9: 120-126.

Grisebach, A. H. G. 1839. Genera et Species Gentianearum. Stuttgart: Sumtibus J. G. Cottae.

Haupt, A. W. 1930. A gelatin fixative for paraffin sections. Stain Technology 5: 97-98.

Raynal, A. 1974. Le genre Nymphoides (Menyanthaceae) en Afrique et a Madagascar. 2e partie:

Taxonomie. Adansonia ser. 2, 14: 405-458.

Ruzin, S. E. 1999. Plant Microtechnique and Microscopy. New York: Oxford University Press.

Sivarajan, V. V. and K. T. Joseph. 1993. The genus Nymphoides Séguier (Menyanthaceae) in

India. Aquatic Botany 45: 145-170.

Sjörs, H. 1988. Vattenklövern, Menyanthes trifoliata – en minimonografi. [Menyanthes trifoliata,

a short monograph.] Svensk Botanisk Tidskrift 82: 51-64.

Tippery, N. P. and D. H. Les. 2009. A new genus and new combinations in Australian Villarsia

(Menyanthaceae). Novon 19: 406-413.

Tippery, N. P., D. H. Les, D. J. Padgett, and S. W. L. Jacobs. 2008. Generic circumscription in

Menyanthaceae: a phylogenetic evaluation. Systematic Botany 33: 598-612.

Wagner, R. 1895. Die Morphologie des Limnanthemum nymphaeoides. Botanische Zeitung 53:

189-205.

Weberling, F. 1989. Morphology of Flowers and Inflorescences. Translated by R. J. Pankhurst.

New York: Cambridge University Press.

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG 3.01 – Representations of the overall morphology of Nymphoides inflorescences of the expanded type (A; e.g., N. aurantiaca), condensed type (B; e.g., N. indica), and N. peltata

(C). Diagrams are based on Sivarajan and Joseph (1993).

FIG. 3.02 – Illustrated cross-sections of Nymphoides inflorescences for N. aurantiaca (A;

Goebel 1891) and N. peltata (B; Wagner 1895), and corresponding diagrammatic interpretations

(C-D). Flowers (F) and leaves (L) are labeled with subscripts that are discussed in the text.

FIG. 3.03 – Cross-section through the inflorescence bud of Nymphoides aurantiaca (A;

Tippery 122) and corresponding diagram (B). Subscript numbers indicate the node order; each node has a terminal flower (TF), an axillary flower (AF) subtended by a distal leaf (DL), and a proximal leaf (PL) that subtends a continuation axis containing the organs of subsequent nodes.

A dissected inflorescence bud of N. aurantiaca also is shown (C; Tippery 122). Scale bar = 1 mm.

FIG. 3.04 – Cross-section through the inflorescence bud of Nymphoides crenata (A;

Tippery 179) and corresponding diagram (B). Abbreviations are as in Fig. 3.03. A dissected inflorescence bud of N. exiliflora also is shown (C; Tippery 166). Scale bar = 1 mm.

FIG. 3.05 – Cross-section through the inflorescence bud of Nymphoides indica (A;

Tippery 162) and corresponding diagram (B). Abbreviations are as in Fig. 3.03. Leaf PL1, which is not evident on the cross-section, is drawn as an outline in the diagram. Scale bar = 1 mm.

FIG. 3.06 – Cross-section through the inflorescence bud of Nymphoides aquatica (A;

Benoit 06-018) and corresponding diagram (B). Abbreviations are as in Fig. 3.03. Organs shaded red on the diagram are of a more complex pattern and are shaded away for simplicity of

Chapter 3: Evolution of inflorescence morphology in Nymphoides comparison to other diagrams. A dissected inflorescence bud of N. aquatica also is shown (C;

Benoit 06-018). Scale bar = 1 mm.

FIG. 3.07 – Cross-section through the inflorescence bud of Nymphoides peltata (A;

Tippery 83) and corresponding diagram (B). The lower foliage leaf (BL) subtends an inflorescence continuation axis (IA). Other abbreviations as in Fig. 3.03. Organs colored red on the diagram are part of the accessory shoot system (discussed in the text) and are shaded away to clarify organ arrangement on the floral continuation axis. A dissected inflorescence bud of N. peltata also is shown (C; Tippery 83). Scale bar = 1 mm.

FIG. 3.08 – Dissected inflorescence bud of Nymphoides peltata (Tippery 83). Whole bud with subtending foliage leaves removed (A), floral continuation axis subtended by the upper foliage leaf (B), inflorescence continuation axis (C), dissected inflorescence continuation axis

(D). Abbreviations discussed in the text. Scale bar = 1 mm.

FIG. 3.09 – Relative organ arrangement of the floral (FA) and inflorescence continuation axes (IA) at the junction of inflorescence stem (BI) and floating-leaf petiole (BL) of Nymphoides indica (Tippery 162), lateral (A) and adaxial (B) views. Higher-magnification of the adaxial view with bracts removed (C) shows the arrangement of terminal (TF) and axial flowers (AF) of various orders (subscript numbers). Scale bar = 1 mm.

FIG. 3.10 – A. Dissected inflorescence continuation axis of Nymphoides indica (Pagels s.n. 30 Aug 2005), showing prophyll (PP1) and developing foliage leaf (BL1). B. Same bud, with prophyll partially dissected away to reveal floral (FA1) and inflorescence (IA1) continuation axes and subsequent foliage leaf (BL2). C. Prophyll removed, foliage leaf (BL2) and prophyll (PP2) of axillary visible. D. Dissection of axillary bud to reveal floral (FA2) and inflorescence (IA2)

Chapter 3: Evolution of inflorescence morphology in Nymphoides

continuation axes at the junction of inflorescence stem (BI2) and floating leaf (BL2). Scale bar =

1 mm.

FIG. 3.11 – A. Simplified, generalized diagram of inflorescence organization in

Nymphoides. Each modular system contains a terminal (TF) and axillary flowers (AF), basal

(BL), proximal (PL), and distal leaves (DL), basal (BI), proximal (PI), and distal (DI) internodes.

Two kinds of continuation axis exist, the floral (FA) and inflorescence axes (IA). Modification of the generalized model results in the modular system for species with expanded (B) or condensed

FIG. 3.12 – Elaboration of the simplified model for Nymphoides with expanded (A,D,G) or condensed (B,E,H) inflorescence types and N. peltata (C,F,I). The top row (A-C) repeats the simplified model from Fig. 3.11. The middle row (D-F) shows the model with two units added to the floral continuation axis (FA1). The bottom row (G-I) shows the same model with units added to the inflorescence continuation axis (IA1).

FIG. 3.13 – Dissected inflorescence bud of Ornduffia umbricola (Pagels s.n.). Whole apical bud (A), dissected between two bracts (B), detail of lower bract (C), opposite view of upper bract (D). The terminal flower (TF1) is preceded by two leaves, the distal leaf (DL1) subtending an axillary flower (AF1), and the proximal leaf (PL1) subtending the floral continuation axis (FA1). A closer view of the floral continuation axis is shown also (C), in which two leaves of the next node (DL2 and PL2) are visible. Scale bar = 1 mm.

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.01

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.02

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.03

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.04

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.05

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.06

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.07

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.08

100

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.09

101

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.10

102

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.11

103

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.12

104

Chapter 3: Evolution of inflorescence morphology in Nymphoides

FIG. 3.13

105

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

Chapter 4

Population biology and the invasive potential of Nymphoides peltata

ABSTRACT

Nymphoides peltata is a floating-leaved aquatic plant native to Eurasia that comprises several naturalized populations in North America, where it also is cultivated as an ornamental water garden plant. In their native range, populations of the heterostylous N. peltata consist of both long- and short-styled floral morphs, and genetic self-incompatibility prevents (to varying degrees) the production of seeds after self- or intramorph pollination. In addition to sexual reproduction, N. peltata also can reproduce vegetatively (i.e., clonally) through fragmentation or stoloniferous growth. I surveyed populations of N. peltata in North America to determine the proportion of long- and short-styled floral morphs, the amount of pollen transferred by insects, the seed set and seed germination. In addition, using microsatellite markers I assessed the amount of genetic variability within populations and compared this to cultivated plants and to populations in Eurasia. Populations of N. peltata in North America were found to consist entirely of short-styled flowers, which nonetheless were able to produce abundant seeds that germinated under natural conditions. Pollen transfer was observed in one population to be accomplished primarily by Apis mellifera, which was a regular visitor to flowers. Genetic analysis revealed that

North American N. peltata populations were entirely homozygous at eight microsatellite loci, and they were genetically identical to one another and to material in cultivation. In contrast, plants from most native populations were heterozygous at one or more loci, including plants from adventive populations in Sweden. Analysis of genetic distance indicated that N. peltata in North

106

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

America and in cultivation was most similar to plants from Europe. Because the genotypes of

North American N. peltata were homozygous, molecular data were unable to distinguish whether vegetative or sexual reproduction characterizes contemporary naturalized populations. The paucity of genetic variation in North American populations likely reflects a bottleneck event, which probably occurred in cultivation prior to introduction, evidenced by the identical genotypes recovered from naturalized and cultivated plants.

INTRODUCTION

Humans are having an increasingly noticeable effect on the world, with respect to its natural cycles and biological interrelationships. One example of human-mediated habitat modification involves the introduction of non-native species that then go on to persist independently (i.e., without human intervention), some of which ultimately become invasive (i.e., spread aggressively and have a negative impact on native organisms). Some exotic species introductions have been accidental, while others can be traced directly to an intentional event, such as cultivation or even intentional establishment for a perceived ecological benefit (e.g., kudzu, cane toad; Pfeiffer and Voeks 2008).

Invasive plants can spread rapidly through propagules (dispersal units), which are either sexual (e.g., seeds, pollen) or vegetative (e.g., by fragmentation or through specialized structures).

In cases where pollination occurs, invasive plants may be pollinated abiotically (e.g., by wind) or utilize species from either native or invasive pollinator fauna. Dispersal of fruits or seeds may similarly be either abiotic or biotic. In addition to requiring effective pollination and dispersal vectors, invasive exotic plants must overcome various habitat or climate properties that may differ than what they were exposed to in their native territory. Not every exotic species that ends up in a new territory becomes invasive, and it has been proposed that only one in ten introduced species becomes naturalized, of which only one-tenth then becomes invasive (i.e., the so-called

107

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

‗tens rule‘; Williamson and Fitter 1996). However, the period between initial introduction and becoming invasive varies in length, and many currently invasive plants persisted at low levels for

40 years or more before becoming nuisance weeds (Les and Mehrhoff 1999).

Aquatic plants represent a particularly noxious category of invasive species, because they interfere with commercial and recreational activities and tend to spread rapidly over a wide range

(Les and Mehrhoff 1999). Aquatic plants occupy habitats that may be contiguous over large areas (e.g., a river basin) or be visited frequently by vagile biotic vectors (e.g., migratory waterfowl). Aquatic habitats also are remarkably similar over large geographic areas, being less affected by climate and temperature variation than their terrestrial counterparts. For these reasons, invasive aquatic plants can establish rapidly and persist over broad ranges. Aquatic plants also have a number of specialized dispersal methods. Some species can disperse and establish from tiny fragments of leaf or other tissue, others produce abundant winter-buds or tubers, and some reproduce using seeds that are adapted to the aquatic habitat, e.g., by floating or clinging externally to biotic vectors (Sculthorpe 1967).

Nymphoides peltata life history. Nymphoides peltata is an aquatic plant native to temperate Eurasia, whose habit consists of a submersed rhizome and floating leaves, some of which support a lax inflorescence that can trail for several meters along the water surface (van der

Velde and van der Heijden 1981). Its native range extends from Japan to western Europe, with the northern limit approximated by the 16°C July isotherm (van der Voo and Westhoff 1961;

Darbyshire and Francis 2008). In recent years an apparent range expansion has brought N. peltata to Sweden, where it grows in several lakes (Larson 2007). In North America, naturalized populations of N. peltata have persisted after first being introduced in the late 19th century

(Stuckey 1974; Les and Mehrhoff 1999). Initial collections of N. peltata were from ornamental ponds in the eastern United States, after which it began to expand into natural areas (Stuckey

1974). In a study of herbarium specimens, Stuckey (1974) summarized the species‘ contemporary range extent, which included portions of 15 states in the U.S.A. Since then,

108

Chapter 4: Population biology and the invasive potential of Nymphoides peltata additional populations have been located, and records now exist for 29 states (Howard 2010) and four Canadian provinces (Darbyshire and Francis 2008). Although it has spread over considerable geographic area, N. peltata exists in relatively localized populations, and on the whole it is not considered to be of major concern (Bartodziej and Ludlow 1997; Les and

Mehrhoff 1999). Nymphoides peltata currently is listed as a noxious weed in six states, however it continues to be cultivated and sold as an ornamental water garden plant (Howard 2010). In native populations, N. peltata occurs in shallow lake inlets and slow-moving rivers (van der

Velde and van der Heijden 1981; Nishihiro et al. 2009), and these are typical habitats for naturalized populations also (Larson 2007; Darbyshire and Francis 2008).

Morphologically, N. peltata is unique in many ways relative to its congeners. The inflorescence consists of nodes with two subopposite leaves, both of which expand to be foliose, and which support a cluster of flowers with initiation patterns that differ from all other

Nymphoides species (Chapter 3). Flowers of N. peltata are large (average diameter 3.5 cm; van der Velde and van der Heijden 1981) and abundant; a single ramet can have two consecutive nodes flowering simultaneously and produce over 20 fruits in one season (van der Velde and van der Heijden 1981). Van der Velde and van der Heijden (1981) estimated seed production at over

3,000 seeds per square meter in the Netherlands population they studied. In addition to seeds, N. peltata can reproduce vegetatively through a type of stoloniferous growth mediated by its trailing inflorescence, which may root at the nodes and establish new ramets for subsequent seasons.

Inflorescence nodes also can disperse after being removed from the parent plant (Darbyshire and

Francis 2008). Nymphoides peltata overwinters by dying back to the underwater rhizome, and this structure even survives freezing when exposed (Darbyshire and Francis 2008).

Flowers of N. peltata are brought above the water surface gradually by flexion of the pedicel, and one day prior to anthesis the corolla extends beyond the calyx to become visible in bud (van der Velde and van der Heijden 1981). Both stigma and anthers are mature at anthesis, and pollinators have been observed visiting flowers immediately after they open; the flowers

109

Chapter 4: Population biology and the invasive potential of Nymphoides peltata remain open for several hours, depending on ambient conditions, after which they wilt and return underwater (van der Velde and van der Heijden 1981). Flowers are visited by insects, chiefly hymenopterans (e.g., Apis, Bombus, Halictus) and to a lesser extent epihydrid and syrphid flies

(van der Velde and van der Heijden 1981; Wang et al. 2005). In Netherlands populations, hymenopterans were observed seeking nectar at the ovary base. These insects were determined to be effective pollinators, because of their repetitive flower visitation pattern and hairy bodies, and because their tongues were sufficiently long to overcome the 5-6 mm mechanical constraint imposed by the narrow corolla tube and obstructive staminodal hairs (van der Velde and van der

Heijden 1981). Pollination occurs almost exclusively within local or contiguous populations, as there was no evidence of pollen transfer between populations that were separated by 1.2 km

(Wang et al. 2005). Autonomous pollination (i.e., without pollinator activity) was inferred in the

Netherlands populations, but only in the short-styled floral morph (see below) and mostly in conjunction with late-day flower wilting (van der Velde and van der Heijden 1981).

The fruits of N. peltata are the largest in the genus by far, and typically contain 20-50 seeds (van der Velde and van der Heijden 1981; Wang et al. 2005). Seeds are obovate and compressed laterally, 4.0-5.2 mm long, with a marginal ring of stiff hairs (about 1 mm long) that are ornamented distally with micropapillae (Chuang and Ornduff 1992). Fruits may remain attached to the maternal plant or become detached and float away. After their enclosing fruit decays, the seeds float on the water surface, where they are dispersed by water or by clinging to biotic vectors. Cook (1990) found that N. peltata seeds are optimized to be hydrophobic enough to float, yet sufficiently hydrophilic to attach to the surface of animals as they leave the water.

However, they do not remain viable after passing through the digestive tract of waterfowl (Smits et al. 1989). Seeds may float for two months or more, but they often sink when disturbed (van der Velde and van der Heijden 1981). Seed germination depends on a period of cold stratification, followed by exposure to light and oxygenated water (Smits et al. 1990).

110

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

Heterostyly in Nymphoides peltata. Like many species of Nymphoides, N. peltata exhibits dimorphic heterostyly, a complex syndrome most conspicuously identified by reciprocal herkogamy (i.e., spatial separation of anthers and stigma), whereby individual plants produce flowers that are either long-styled (with short filaments) or short-styled (with long filaments).

The reciprocal positions of organs is such that the anthers of one morph are at the same height as the stigma of the opposite morph (van der Velde and van der Heijden 1981). Heterostyly thus promotes outcrossing mechanically in two ways, by separating the anthers and stigma within a flower and by offering pollen to a biotic vector at the same relative height where it would be received by a compatible stigma. The average spatial separation of anthers and stigma was measured in Netherlands population to be 7.7 mm on long-styled flowers and 2.5 mm on short- styled flowers, with the most similar heights observed between short-styled stigmas and long- styled stamens, which both were about 10.5 mm from the flower base (van der Velde and van der

Heijden 1981). In addition to reciprocal herkogamy, the syndrome of heterostyly also includes a genetic self-incompatibility mechanism that prevents fertilization after pollination within a flower

(i.e., self-pollination, either autogamy or geitonogamy) or between plants of the same floral morph type (i.e., intramorph pollination). Self- or intramorph pollinations have been termed

‗illegitimate‘ crosses, and crosses between different floral morphs ‗legitimate‘ (Darwin 1877).

The function of heterostyly to promote outcrossing becomes even more important in light of the prevalent vegetative reproduction that characterizes N. peltata. When plants reproduce vegetatively (i.e., clonally), they produce offspring that are genetically identical to themselves.

Clonal reproduction has the advantage that it requires no pollination or seedling establishment, but the disadvantage that it results in zero genetic variability among offspring (barring somatic mutation), which may be required to adapt to changing or novel environments (Philbrick and Les

1996; Holsinger 2000). In addition, clonal reproduction avoids any negative effects that might be associated with either inbreeding or outcrossing (Fenster and Dudash 1994 and references therein). The potential disadvantages of clonal reproduction include increased susceptibility to

111

Chapter 4: Population biology and the invasive potential of Nymphoides peltata herbivores or pathogens, owing to a lower diversity of resistance traits (Hamilton 1990;

Alexander 1991; Silvertown 2008), as well as a greater potential for inbreeding, because the majority of pollinations within a highly clonal population would be between genetically identical individuals. In the latter scenario, sexual reproduction actually could become detrimental in populations that were derived principally through clonal reproduction, in contrast to sexually reproducing populations that already house genetic variation among individuals. Genetic incompatibility mechanisms (including heterostyly) are able to counteract the negative effects of inbreeding that are associated with clonal populations, because they prevent fertilization between individuals that are genetically identical. In order to assess the population dynamics of a population that has both clonal and sexual modes of reproduction, it is important to understand many aspects, including the amount of each reproductive mode, the genetic diversity within and among individuals, and the degree to which genetic self-incompatibility mechanisms operate.

Clonal vegetative propagation is relatively simple to assess in N. peltata, because genetically identical individuals have the same floral morph type. Strictly clonal reproduction results in populations that consist entirely of one morph, in which all pollinations are illegitimate.

In contrast, populations that reproduce only sexually have offspring with equal proportions of each floral morph (Ganders 1979; Barrett 1992). In heterostylous species, then, the relative frequency of floral morphs serves as a proxy for the amount of sexual versus clonal reproduction, with the former resulting in a 1:1 ratio of morph types and the latter having uniform populations of one or the other type. In China, for example, populations of N. peltata were found to have a range of floral morph ratios, with some consisting entirely of one morph type and others having near equal proportions of both morph types (Wang et al. 2005). Experimental populations with more balanced morph ratios achieved higher fruit set, with a sharp decline in fruit production when the proportion of either floral morph exceeded 80% (Wang et al. 2005). A subpopulation of

N. peltata in Japan that had both long- and short-styled flowers showed a variety of genotypes, whereas all monomorphic subpopulations but one were genetically uniform (Uesugi et al. 2004,

112

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

2007). Genetic examination of adventive populations in Sweden also revealed a high proportion of monomorphic populations, all of which were genetically uniform (Larson 2007).

Variations within the heterostylous syndrome have been observed in native and exotic populations of N. peltata. One notable variation is in the relative strength of the self- incompatibility system. Under strict self-incompatibility, all illegitimate crosses, whether the result of self- or intramorph pollination, result in zero seed set. Experimental crosses between incompatible flowers in China resulted in practically no fruits (Wang et al. 2005), but illegitimate pollinations in other studies have tended to produce fruits with a modest number of seeds (1-10 in the Netherlands, van der Velde and van der Heijden 1981; 10-20 in western Eurasia, Ornduff

1966; mostly fewer than 10 in Japan, Takagawa et al. 2006). Although seeds may be produced by illegitimate pollinations in N. peltata, studies that examined seedling viability found it to be much lower in selfed than outcrossed offspring (Ornduff 1966; Takagawa et al. 2006).

Perhaps the most extreme breakdown of self-incompatibility in N. peltata was found to associate also with a breakdown in floral morphology. In Japan, a rare subpopulation consisting of homostylous flowers (i.e., with a single floral morph in which anthers and stigma are at the same height) was characterized to be highly self-compatible, with self-pollinations of homostylous flowers resulting in 84% seed set relative to legitimate crosses (Takagawa et al.

2006). Breakdown of reciprocal herkogamy concomitant with weakened self-incompatibility has been observed in many heterostylous groups (e.g., Ornduff 1972; Piper et al. 1984; Barrett and

Shore 1987; Schoen et al. 1997; Nakamura et al. 2007; Sakai and Wright 2008), including in

Nymphoides the well-characterized homostylous species N. geminata, which also evolved to have smaller flowers and more clonal reproduction than the related congener N. montana (Haddadchi

2008). The breakdown of heterostyly and the evolution of increased selfing are advantageous for small or colonizing populations that reproduce by seed, where the likelihood of a legitimate pollination is decreased (Baker 1955).

113

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

Invasive potential of Nymphoides peltata. The interrelationships of heterostyly and clonal reproduction in N. peltata are important to consider in light of the species‘ potentially invasive status. Depending on the mechanism of spread (e.g., vegetative fragmentation, seed dispersal), different control methods would be variously effective. Furthermore, if populations are reproducing sexually, they could be generating variation that would increase their adaptability and thus augment their potential to spread and become more aggressive. A study of adventive N. peltata populations in Sweden revealed that clones were genetically variable but were not recombining within most of the populations surveyed (Larson 2007). Moreover, seedlings produced from artificial crosses had novel recombinant genotypes, indicating that populations have the potential to increase their genetic variability. The process whereby a species becomes invasive involves a period of acclimation and adaptation (Williamson and Fitter 1996; Les and

Mehrhoff 1999), and the recombination of genetic material in N. peltata could potentially produce enough adaptive variability to turn it into a truly noxious weed. The production of seeds, however, requires either populations that are diverse enough to have both floral morphs, or a sufficiently weak self-incompatibility system that allows illegitimate crosses to produce seed. In

Sweden, monomorphic populations of N. peltata were not observed to produce fruit (Larson

2007), and the lower viability of illegitimately sired offspring (Ornduff 1966; Takagawa et al.

2006) would make this an unlikely pathway for reproduction. Nonetheless, selection for increased selfing and seedling viability in invasive populations could eventually offset the negative consequences that have been observed in other populations.

In order to formulate a more comprehensive understanding of the biology and invasive potential of exotic N. peltata in North America, I undertook a study to examine the pollination biology, seed production, seedling establishment, and genetic diversity of populations in

Connecticut, New York, Pennsylvania, and Washington. In addition, I compared the genetic composition of these populations to plants in cultivation and material from the species‘ native range in Hungary, India, Japan, and Korea, and the adventive populations in Sweden. Using

114

Chapter 4: Population biology and the invasive potential of Nymphoides peltata observational data, pollination and germination experiments, and population genetic

(microsatellite) analyses, I addressed the following questions: (1) Are plants morphologically heterostylous, and what is the distribution of floral morphs? (2) Do plants exhibit self- incompatibility, and to what extent? (3) Are plants reproducing sexually or vegetatively (i.e., clonally)? (4) Are flowers being pollinated? (5) Are fruits and seeds being produced? (6) Are seeds germinable, and are the resulting seedlings viable?

MATERIALS AND METHODS

Population localities. In Connecticut, N. peltata has been newly reported from two localities in the past five years, after having been absent from the state for a century prior

(Howard 2010). The first is a private pond in Columbia (abbreviated ColCT; Table 4.01), and the second population grows 38 km away in the Pachaug State Forest (PacCT) in Voluntown. The

New York populations examined all were collected in the Hudson River, near the towns of

Dresden (DreNY), Ticonderoga (TicNY), Stillwater (StiNY), and in the Rogers Island Wildlife

Management Area (RogNY) near Hudson/Catskill. Two Pennsylvania populations were surveyed, in the town of Lower Pottsgrove (PotPA) and at the Ockanickon Boy Scout Camp

(OckPA), both in the eastern part of the state. In Washington a population-level sample of multiple individuals was collected from Lake Spokane (SpoWA, also known as Long Lake), and single plants also were acquired from the vicinities of Bellingham (BelWA) and Lynden

(LynWA). In addition to North American populations, material was obtained for single plants from Hungary, India, and Korea, and from populations that had been studied previously in Japan

(Takagawa et al. 2006) and Sweden (Larson 2007; Appendix 4.01).

Where possible, an assessment was made of the structure of each population, including the area of coverage, the ratios of flower morph types, the production of fruits and seeds, and the presence of seedlings. Areas of coverage and areas of bounding water bodies were assessed using

115

Chapter 4: Population biology and the invasive potential of Nymphoides peltata satellite maps in Google Earth (Google Inc., Mountain View, California), on which polygons were drawn to approximate water body boundaries and the limits of N. peltata coverage. Polygon coordinate data were exported, and areas were calculated in R ver. 2.10.0 (R Development Core

Team 2009) using the areapl function of the splancs package (Rowlingson and Diggle 1993), with a conversion of 111.12 km per degree latitude. Floral morph ratios were assessed by counting at least 20 individuals per population. For five populations (ColCT, RogNY, PotPA, OckPA, and

SpoWA), leaves of multiple individuals, separated from each other by at least 2 m, were taken for genetic analysis, and fruits were taken from two populations (ColCT and RogNY) to study germination and the genetic composition of seedlings.

Pollination. In the RogNY population, pollination experiments were conducted on two occasions (20 Sep 2008 and 18 Aug 2009), and pollinator activity was noted. The activity of pollinators was assessed as the frequency of visits and the relative abundance of any one species.

Several individual insects were collected in 70% ethanol (18 Aug 2009) for later identification and to determine the distribution of pollen on their bodies.

Several flowers were bagged prior to the anthesis of any flowers in the population

(approximately 0800 hr), using 7.5 x 10 cm organza bags with drawstring closure. Pollination experiments entailed four categories of pollination: autonomous (flowers bagged all day), selfed

(bagged flowers self-pollinated), outcrossed (bagged flowers hand-pollinated with stamens from plants at least 2 m distant), and open (flowers unmanipulated). Following hand-pollination, flowers were bagged again for the remainder of the day. Each category contained up to eight replicates. Controlled pollinations were conducted by removing stamens from flowers and brushing the anthers four times against the stigma of the target flower. Effective transfer of pollen by hand-pollination was verified by eye (i.e., white stigmas were observed to acquire a yellow hue).

Both bagged and hand-pollinated flowers were collected to verify the effectiveness of these treatments (20 Sep 2008). In addition, samples of open-pollinated flowers were collected at

116

Chapter 4: Population biology and the invasive potential of Nymphoides peltata approximately 1.5-hr intervals during the day. The flowers were preserved in 70% ethanol, and later their stigmas were assayed for the presence and relative density of pollen, which was done by taking digital photographs at 200x magnification. A filled shape was drawn around the stigma area that was in focus, and the number of pollen grains visible in that area was counted; image quantification was accomplished using the histogram function in the GNU Image Manipulation

Program (GIMP) ver. 2.6.6 (GIMP Team 2010) to determine the number of pixels occupied by pollen or stigma.

The inflorescence stems associated with experimentally pollinated flowers were each labeled with a numbered band. Fruits were allowed to mature for four weeks, after which they were collected and tested for seed germination (see below). Open-pollinated fruits of unknown age also were collected from ColCT (22 Sep 2008) and RogNY (20 Sep 2008 and 21 Sep 2009).

In addition, fruits from consecutive nodes on the same plant, replicated over eight plants, were collected from RogNY (20 Oct 2008).

Seed germination. Seeds were germinated following the procedure of Smits et al.

(1990), which first involved four weeks of cold stratification (4°C in tap water, in shaded ambient light). Fruits were separated into individual cups, and seeds were allowed to rot away from their fruit, after which fruit material was removed and seeds were counted. Fruits that did not rot within the cold-stratification period were incised and their seeds were removed. Seeds were then exposed to 15 hr light / 7 hr dark cycles at 20°C, with daylight intensity at approximately 150

Einstein m-2 hr-1. Germination was scored as the emergence of the radicle within three weeks of light treatment. Some seedlings were preserved in CTAB for genetic analysis, and others were potted and grown in the greenhouse. Seed color was quantified for each fruit of the 2009 RogNY group by taking a digital photograph, removing background (i.e., non-seed) portions of the image, and obtaining the mean pixel value in GIMP (GIMP Team 2010).

Microsatellite data. Genetic analyses were performed on material from native and naturalized N. peltata populations and cultivated plants (Table 4.01), using seven of the eight

117

Chapter 4: Population biology and the invasive potential of Nymphoides peltata microsatellite primers that were designed by Uesugi et al. (2005). DNA was extracted from live or CTAB-preserved plants, or from dried herbarium specimens, according to the method described in Chapter 2, and stored in 1x TE buffer. DNA concentration was quantified using a

NanoDrop™ ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, Delaware) and then diluted to 5 ng/L in dH2O. Amplification was performed using 1 L of diluted DNA, 0.8

L of 2.5 mM dNTP, 0.1  of each 10 mM primer, 1 L of buffer, and 0.05 L of Titanium™

Taq polymerase (Clontech Laboratories, Inc., Mountain View, California), in a 10 L reaction.

Thermal cycling protocol followed Shibayama et al. (2006), with a range of annealing temperatures (Np152: 60.4°C, Np207: 52.3°C, Np274: 68.1°C, Np306: 58.1°C , Np641: 64.7°C,

Np694: 66.3°C, Np730: 67.5°C; primer set Np382 was not analyzed), using an Eppendorf®

Mastercycler pro S thermal cycler (Eppendorf North America, Hauppauge, New York).

Fragments were analyzed on an ABI Prism® 3130xl Genetic Analyzer (Applied Biosystems, Inc.,

Foster City, California), and scored using the program GeneMarker® ver. 1.80 (SoftGenetics,

LLC, State College, Pennsylvania). Microsatellite data were scored as the presence or absence of each allele, and the binary data were used to construct a neighbor-joining tree under the distance criterion in PAUP* (Swofford 2002).

RESULTS

Population localities. Eleven North American populations of N. peltata were located for this study (Table 4.01; Fig. 4.01). These included several sites that were reported by Stuckey

(1974), mainly including portions of the Hudson River (e.g., DreNY, StiNY). Other populations were located with the assistance of state or local agents, and these were of relatively recent origin.

Populations occupied small ponds (ColCT, OckPA, PotPA, BelWA, LynWA) or portions of slow-moving rivers (SpoWA and all NY populations). The area of local population coverage

118

Chapter 4: Population biology and the invasive potential of Nymphoides peltata ranged from 0.04 to over 50 ha (Table 4.01). Populations where multiple individuals were counted (ColCT [n = 20], RogNY [n = 87], OckPA [n = 20], SpoWA [n = 28]) always consisted only of short-styled flower morphs. Anther and stigma heights were not quantified, but the spatial separation of these organs conformed roughly to the measurement given for short-styled flowers in the Netherlands (van der Velde and van der Heijden 1981). Fruits were readily observed in these populations as well; fruit production in the RogNY population was approximately 4-7 fruits per inflorescence node, with up to four or more nodes present on any one ramet. Seedlings were not observed in the RogNY population, which existed 10 m distant from the nearest shoreline. Other populations (ColCT, StiNY, TicNY) had plants that were scored as seedlings, because they grew on exposed mud with few leaves and no observable rhizome.

Pollination. In the RogNY population, floral visitors consisted primarily of Apis mellifera, which was observed to make directed and repeated visits to N. peltata flowers. Flower buds of N. peltata are elevated above water for several hours before anthesis; flowering commences after sunrise and continues until early evening. On 18 Aug 2009, buds began to open around 0800hr, when individuals of A. mellifera were already active. Some flower buds were observed to have been robbed of nectar, evidenced by a hole bored through the petal (which extends beyond the calyx). In addition, some individuals of A. mellifera were observed prying flower buds open; after flowers were pried, they continued to open. The time of anthesis varied among flowers, and some flowers that apparently were prepared to open on the observation date

(i.e., were elevated above water and the same size as flowers that did open) had not opened as late as 1600hr. When visiting a N. peltata flower, A. mellifera typically oriented its head toward the ovary base, where nectaries are located, but some individuals also were observed collecting pollen from anthers or petals. Other, less abundant insect species were seen visiting N. peltata flowers also, including other hymenopterans and some dipterans, although the type of activity varied from resting on the petals to apparently collecting nectar.

119

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

Insect floral visitors were observed to have a small amount of N. peltata pollen on their bodies; however, the distribution of pollen on specimens may be an artifact of the collection method (insects collected along with flowers, in 70% ethanol). Specimens of A. mellifera, however, all had dense aggregates of N. peltata pollen on their corbiculae (Fig. 4.02).

Observation of stigmas collected throughout the day showed an overall increase in pollen density

(Fig. 4.03), with a significant increase in pollen deposition after 1200hr (Welch two sample t-test of unmanipulated flowers; p < 0.0001). Stigmas from flowers that were bagged all day or failed to open showed a modest amount of pollen deposition. Flowers that were hand-pollinated received amounts of pollen that were comparable to open-pollinated flowers.

Seed germination. Fruits (of both known and unknown age) that resulted from open pollinations showed remarkable variation in seed number (Fig. 4.04), and the character of seeds recovered also varied considerably. Seed number per capsule ranged over all populations from

11 to 88. Seeds varied most conspicuously in color, with darker seeds also being more firm.

Seeds that were pale green tended to be more fragile and apparently indicated a lesser degree of development, although the range of color from green to dark brown was observed also in the collection of fruits that were all the same age. Seed color and texture were quite similar within a fruit, and thus the variation primarily existed between fruits.

The results of experimental pollinations were largely inconclusive. There was substantial failure to relocate the fruits of bagged flowers, possibly due to spatial variability of a plant resulting from water activity, or to plant fragmentation. Additionally, the study design did not account sufficiently for pollinations that would not result in fruit, and presumably the decay of flowers that did not set fruit would also dissociate the pollination bag. Although the numbered bands were affixed to the inflorescence stem, many of these still were not located. There was circumstantial evidence (i.e., bags that would have contained fruits being located floating independently near the pollinated plant), however, that some pollinations failed to produce fruits.

These included in the 2009 RogNY population one selfed plant for which an empty bag was

120

Chapter 4: Population biology and the invasive potential of Nymphoides peltata recovered, and in addition one selfed, one open-pollinated, and two outcrossed plants for which no bags or fruits were located. The identification tags of fruits recovered from the 2008 pollination experiment at RogNY were effaced, so collected fruits could be identified only to the date of pollination, but not to any particular pollination treatment.

Germination of seeds varied widely, from zero to 100% in open-pollinated fruits of unknown age (35.0 ± 36.5 [mean ± standard deviation]; Fig. 4.05). Thirty percent of fruits had zero germination, and the majority of fruits had less than 25%; percent germination was not correlated with seed number (Fig. 4.06). Seed color in open-pollinated plants of unknown age averaged a shade of light brown (mean color value 81.3 ± 20.0, on a scale of 0 = white and 255 = black, n = 29), and among these plants seed color was not correlated with germination success

(Fig. 4.07). Experimental outcross pollinations from the 2009 RogNY population (n = 3) all resulted in significantly higher seed germination (mean 79.3%; Welch two sample t-test; p <

0.05) and significantly darker seeds (mean color value 174.6 ± 38.9; p < 0.05); a similar pattern was observed also for the one open-pollinated fruit of the same age. Seedlings that were potted following the 2008 germination experiments have persisted to 2010, although seedlings of some fruits entirely died out. Seedlings growing in the greenhouse have not flowered to date.

Microsatellite data. Genetic data were obtained for plants from all populations surveyed

(Table 4.01), and for seven of the eight microsatellite primer pairs that were designed by Uesugi et al. (2005). The alleles recovered for each primer pair were variable across populations, and allele number ranged from two (Np306) to six (Np730; Table 4.02). Two primer pairs (Np274 and Np694) each amplified two loci, which had been reported previously (Uesugi et al. 2005;

Larson 2007).

Plants from within the native range of N. peltata and plants from Sweden were heterozygous at one or more loci, except for the accession from Korea, which was homozygous at all loci. Plants from the United States and plants in cultivation were all identical to each other, and heterozygous at only one locus, Np152. The alleles for Np152, however, were not observed

121

Chapter 4: Population biology and the invasive potential of Nymphoides peltata to recombine in four seedlings from the RogNY population (RogNY 2 seedlings A-D, for which the maternal parent was RogNY 2; Table 4.02). Alleles recovered from North American N. peltata were shared variously with plants from all parts of the species‘ native range, but plants collected from Japan shared the fewest alleles.

The distance tree constructed from microsatellite data indicated that N. peltata plants in

Japan are the most genetically dissimilar to plants from North America (Fig. 4.08). Plants from

India and Korea had intermediate similarity, whereas plants from Hungary and Sweden were the most similar out of native populations sampled. Considerable variation was observed within regions that had multiple samples (i.e., Japan and Sweden), and the minimum distance from

North American samples to their most similar European counterparts (Hungary or Lake Sågsjön,

Sweden with four differences each) was less than the maximum distance observed within either

Sweden (nine differences between Lake Sågsjön and Lake Mälaren) or Japan (11 differences between Oofunatsu and Wadamisaki subpopulations of Lake Kasumigaura).

DISCUSSION

Pollination and seed production. In this study, plants of N. peltata were observed to consist entirely, across all North American populations surveyed, of short-styled plants, in contrast to interbreeding populations of N. peltata and other heterostylous species, in which the ratio of short-styled to long-styled plants approaches 1:1. The presence in North America of exclusively short-styled plants would seem to argue for an entirely clonal existence, provided that the plants are intramorph incompatible, which usually is the case for heterostylous species.

Nymphoides peltata, however, has been observed in other populations to be weakly (Ornduff

1966; van der Velde and van der Heijden 1981) or strongly (Wang et al. 2005) self-incompatible, and in either case there is at least some proportion of ‗illegitimate‘ pollinations that result in fruit and seed set. In this study I found evidence, based on fruit abundance in populations of entirely

122

Chapter 4: Population biology and the invasive potential of Nymphoides peltata short-styled plants, that self- or intramorph pollinations regularly produce fruits and seeds. Open- pollinated plants produced an average of 46.9 ± 13.5 seeds per fruit (Fig. 4.04), or approximately

72% seed set relative to the reported mean ovule number of 64 ± 14 per gynoecium (van der

Velde and van der Heijden 1981). The average seed production observed in North American N. peltata was nearly identical to seeds produced by intermorph crosses in Japan (46.4 ± 1.6;

Takagawa et al. 2006) and represents the highest value reported thus far for intramorph crosses in the species. In addition, seed germination ranged from 0% to 100% under natural conditions

(Fig. 4.05), with the modal value of 0% germination observed in 30.6% of fruits.

Seed production in N. peltata requires the activity of pollinators, and in at least one population this task was accomplished primarily by Apis mellifera, itself a non-native species but a regular feature of the North American fauna. Apis mellifera were regular visitors to flowers of the RogNY N. peltata population, in which they were observed to pry flowers open and actively seek the nectar at the base of the corolla tube. Individuals of A. mellifera were dusted with pollen, which they also packed onto their corbiculae. The steady activity of pollinators also correlated with an increase in pollen deposition on stigmas of N. peltata, in contrast to flowers from which pollinators were excluded, that received negligible amounts of pollen (Fig. 4.03).

Open-pollinated flowers received pollen equivalent to what was achieved by hand-pollination, indicating that flowers generally are not pollen limited. However, seed set and seed viability varied considerably among plants, and further research should be directed toward determining the factors that promote the production of abundant, viable seeds. The relative age and maturity of fruits could be a factor, but age was known for only a small number of fruits collected during this study. The fruits resulting from both open pollination and hand-pollinated outcrosses had significantly higher seed germination and significantly darker seeds than did fruits that were collected haphazardly. Notably, although fruit collection targeted fruits that looked mature (i.e., were relatively large in size), these may not have been fully developed. Fruits of N. peltata dissociate from the maternal plant as part of the maturation process, and the biased collection of

123

Chapter 4: Population biology and the invasive potential of Nymphoides peltata fruits that remained attached might have skewed the results toward immature fruits. Dissociation of mature fruits also could explain the failure to recover many fruits that were expected to result from experimental pollinations. Nevertheless, fruits of some experimentally pollinated flowers remained attached to their maternal plant, and these were visually indistinguishable from attached, open-pollinated fruits that were collected concurrently.

Although seed germination was examined, seedling establishment was not quantified in this study and may represent a substantial hurdle to the sexual propagation of N. peltata. Most of the populations surveyed for this study inhabited small, contained water bodies, in which vegetative spread could easily account for the distribution of individuals. The relative isolation of populations in North America and their apparent inability to disperse widely could indicate the failure of some aspect of seed propagation. Certainly the frequency of individuals within any one water body can be explained by vegetative reproduction alone, without invoking seed dispersal or seedling establishment. The distribution of N. peltata over a long portion the Hudson River, however, seems to indicate dispersal by propagules, whether vegetative or sexual. The putative agent of introduction for many naturalized populations of N. peltata has been the release of privately owned water garden plants, and the evidence presented here cannot dispute the possibility that every population represents an independent introduction, as the species is obtained without difficulty from commercial sources and features rather commonly in the water garden trade.

Molecular data. Genetic analyses using microsatellite markers revealed low heterozygosity in naturalized North American populations of N. peltata, with six out of seven analyzed markers being homozygous. Furthermore, populations sampled from both eastern and western states of the U.S.A. were genetically identical to one another, and also identical to plants in cultivation. The low variability within North American populations of N. peltata compared to plants in the native range supports the existence of a genetic bottleneck somewhere in the history of its introduction. In contrast to the homogeneous North American populations, populations in

124

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

Sweden that resulted from a northward range expansion showed high amounts of genetic diversity (Larson 2007). The extensive homozygosity in North American plants relative to their

European relatives, however, indicates that they probably underwent a period of inbreeding, perhaps concomitant with a long history of cultivation. The original introduction of N. peltata into North America stemmed from cultivated populations (Stuckey 1974), which prior to their arrival in the U.S.A. could have been cultivated for many years in Europe. Genetic data support the introduction of a genetically impoverished strain of plants, rather than a loss of genetic variability in situ, because the homozygous genotype found in naturalized populations is identical to the genotype of cultivated N. peltata plants.

The one microsatellite locus that was heterozygous for North American N. peltata,

Np152, curiously did not produce recombinant genotypes in four artificially generated seedlings.

Sequencing is underway to ascertain whether the ‗alleles‘ of Np152 actually reflect the same locus or are unrelated amplifications (e.g., cross-amplifications or paralogous loci). Alleles of the same length as North American plants had not been reported previously, but same-length alleles were recovered in material from Sweden that was analyzed for this study (Table 4.02). Artificial seedlings were generated in the original microsatellite study of N. peltata in Sweden (Larson

2007), however data from the Np152 locus were not reported.

The widespread homozygosity in North American N. peltata complicates the effort to ascertain whether plants are reproducing vegetatively or sexually. If the alleles of Np152 are paralogous, as hypothesized above, then the species is truly homozygous across all microsatellite loci examined. Although this result strongly supports the existence of a historical bottleneck, it precludes a determination of the reproductive mode, because either vegetative or sexual reproduction would continue to propagate homozygous individuals. Nonetheless, results from other facets of this study inform the likelihood of either sexual or vegetative reproduction. The evidence largely favors a vegetative reproductive mode, for two major reasons. First, although seed production and viability were found to be high, seedlings were rarely observed, and these

125

Chapter 4: Population biology and the invasive potential of Nymphoides peltata were always few in number and isolated on mudflats, at considerable distance from the deep water habitat of mature individuals. Secondly, N. peltata continues to have a narrow distribution in spite of repeated introductions over more than a century. Particularly in the northeastern

United States, where water bodies are abundant, plants that disperse by seed should be able to spread rather effectively. Although vegetative propagules are not known to transport between water bodies, the seeds of N. peltata are superbly adapted to disperse on the exterior surface of aquatic animals (Cook 1990).

The management implications of this study are primarily associated with the discovery that N. peltata has a prolific capacity to produce fruits and viable seeds, despite the existence of only one morph type in a species where other populations are strongly self- and intramorph incompatible. Some aspect of the heterostylous self-incompatibility system clearly has broken down in North American N. peltata, although further research will be required to determine more precisely the nature of the breakdown. Genetic variability, and resulting phenotypic variability, are implicated in a species‘ ability to adapt to environments. Although N. peltata underwent a genetic bottleneck associated with years of cultivation and the release of few individuals into natural habitats, it still retains the ability to recombine genetic material through sexual reproduction. The process whereby a species transitions from naturalized to invasive requires time, which some authors have hypothesized to represent a period of genetic adaptation, after which the non-native species becomes a superior competitor against native species. Nymphoides peltata may yet have the potential to become much more invasive, and this possibility should be considered in decisions regarding the control of its trade and importation.

ACKNOWLEDGEMENTS. I am extremely grateful for the following people for providing me with collected material or information on populations of Nymphoides peltata in the U.S.A.:

Laurel Baldwin (Whatcom County Noxious Weed Control Board, Washington), Ann Bove

(Vermont Department of Environmental Conservation), Les Mehrhoff (Department of Ecology

126

Chapter 4: Population biology and the invasive potential of Nymphoides peltata and Evolutionary Biology, University of Connecticut), Nancy Murray (Connecticut Department of Environmental Protection), Ann Rhoads (Morris Arboretum, University of Pennsylvania),

Carol Williamson (Columbia, Connecticut). Other people generously sent specimens from abroad, including Jun Nishihiro (Graduate School of Agricultural and Life Sciences, University of Tokyo), Hyunchur Shin (Soonchunhyang University, Chungnam, Korea), Maike Wilstermann-

Hildebrand (Remseck am Neckar, Germany), and I also wish to thank the staff of the University of California Herbarium (UC) for granting access to their specimens. I am particularly indebted to Daniel Larson (Swedish University of Agricultural Sciences, Uppsala, Sweden), who sent both material of N. peltata from Sweden and fluorescent-labeled primers for microsatellite analysis.

Research was funded in part through the Ronald L. Bamford endowment, University of

Connecticut.

127

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

LITERATURE CITED

Alexander, H. M. 1991. Plant population heterogeneity and pathogen and herbivore levels: a field

experiment. Oecologia 86: 125-131.

Baker, H. G. 1955. Self-compatibility and establishment after ―long-distance‖ dispersal.

Evolution 9: 347-348.

Barrett, S. C. H. 1992. Genetics of weed invasions. Pp. 91-119 in Jain, S. K. and L. W. Botsford,

eds. Applied Population Biology. Dordrecht: Kluwer Academic Publishers.

Barrett, S. C. H. and J. S. Shore. 1987. Variation and evolution of breeding systems in the

Turnera ulmifolia L. complex (Turneraceae). Evolution 41: 340-354.

Bartodziej, W. and J. Ludlow. 1997. Aquatic vegetation monitoring by natural resources agencies

in the United States. Journal of Lake and Reservoir Management 13: 109-117.

Chuang, T. I. and R. Ornduff. 1992. Seed morphology and systematics of Menyanthaceae.

American Journal of Botany 79: 1396-1406.

Cook, C. D. K. 1990. Seed dispersal of Nymphoides peltata (S.G. Gmelin) O. Kuntze

(Menyanthaceae). Aquatic Botany 37: 325-340.

Darbyshire, S. J. and A. Francis. 2008. The biology of invasive alien plants in Canada. 10.

Nymphoides peltata (S. G. Gmel.) Kuntze. Canadian Journal of Plant Science 88: 811-

829.

Darwin, C. D. 1877. The different forms of flowers on plants of the same species. London: John

Murray.

Fenster, C. B. and M. R. Dudash. 1994. Genetic considerations for plant population restoration

and conservation. Pp. 34-62 in Bowles, M. L. and C. J. Whelan, eds. Restoration of

Endangered Species. Cambridge: Cambridge University Press.

Ganders, F. R. 1979. The biology of heterostyly. New Zealand Journal of Botany 17: 607-635.

128

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

GIMP Team. 2010. GIMP: The GNU Image Manipulation Program, ver. 2.6.6.

http://www.gimp.org/

Haddadchi, A. 2008. Floral variation and breeding system in distylous and homostylous species

of clonal aquatic Nymphoides (Menyanthaceae). PhD. Thesis. Armidale: University of

New England.

Hamilton, W. D. 1980. Sex versus non-sex versus parasite. Oikos 35: 282-290.

Holsinger, K. E. 2000. Reproductive systems and evolution in vascular plants. Proceedings of the

National Academy of Sciences, USA 97: 7037-7042.

Howard, V. 2010. Nymphoides peltata. USGS Nonindigenous Aquatic species Database,

Gainsville, FL. http://nas.er.usgs.gov/queries/FactSheet.aspx?speciesID=243

RevisionDate: 4/4/2007. Accessed 19 April 2010.

Larson, D. 2007. Reproduction strategies in introduced Nymphoides peltata populations revealed

by genetic markers. Aquatic Botany 86: 402-406.

Les, D. H. and L. J. Mehrhoff. 1999. Introduction of nonindigenous aquatic vascular plants in

southern New England: a historical perspective. Biological Invasions 1: 281-300.

Nakamura. K., T. Denda, O. Kameshima, and M. Yokota. 2007. Breakdown of distyly in a

tetraploid variety of Ophiorrhiza japonica (Rubiaceae) and its phylogenetic analysis.

Journal of Plant Research 120: 501-509.

Nishihiro, J., R. Uesugi, S. Takagawa, and I. Washitani. 2009. Toward the restoration of a

sustainable population of a threatened aquatic plant, Nymphoides peltata: integrated

genetic/demographic studies and practices. Biological Conservation 142: 1906-1912.

Ornduff, R. 1966. The origin of dioecism from heterostyly in Nymphoides (Menyanthaceae).

Evolution 20: 309-314.

Ornduff, R. 1972. The breakdown of trimorphic incompatibility in Oxalis Section Corniculatae.

Evolution 26: 52-65.

129

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

Pfeiffer, J. M. and R. A. Voeks. 2008. Biological invasions and biocultural diversity: linking

ecological and cultural systems. Environmental Conservation 35: 281-293.

Philbrick, C. T. and D. H. Les. 1996. Evolution of aquatic angiosperm reproductive systems.

BioScience 46: 813-826.

Piper, J. G., B. Charlesworth, and D. Charlesworth. 1984. A high rate of self-fertilization and

increased seed fertility of homostyle primroses. Nature 310: 50-51.

R Development Core Team. 2009. R: A language and environment for statistical computing.

Vienna, Austria: R Foundation for Statistical Computing. http://www.R-project.org

Rowlingson, B. and P. Diggle. 1993. Splancs: spatial point pattern analysis code in S-Plus.

Computers and Geosciences 19: 627-655.

Sakai, S. and S. J. Wright. 2008. Reproductive ecology of 21 coexisting Psychotria species

(Rubiaceae): When is heterostyly lost? Biological Journal of the Linnean Society 93:

125-134.

Schoen, D. J., M. O. Johnston, A.-M. L‘Heureux, and J. V. Marsolais. 1997. Evolutionary history

of the mating system in Amsinckia (Boraginaceae). Evolution 51: 1090-1099.

Shibayama, Y., R. Uesugi, Y. Tsumura, and I. Washitani. 2006. Conservation of the Lake

Kasumigaura population of Nymphoides indica (L.) Kuntze based on genetic evaluation

using microsatellite markers. Limnology 7: 193-197.

Silvertown, J. 2008. The evolutionary maintenance of sexual reproduction: evidence from the

ecological distribution of asexual reproduction in clonal plants. International Journal of

Plant Sciences 169: 157-168.

Smits, A. J. M., R. van Ruremonde, and G. van der Velde. 1989. Seed dispersal of three

nymphaeid macrophytes. Aquatic Botany 35: 167-180.

Smits, A. J. M., P. H. van Avesaath, and G. van der Velde. 1990. Germination requirements and

seed banks of some nymphaeid macrophytes: Nymphaea alba L., Nuphar lutea (L.) Sm.

and Nymphoides peltata (Gmel.) O.Kuntze. Freshwater Biology 24: 315-326.

130

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

Stuckey, R. L. 1974. The introduction and distribution of Nymphoides peltatum (Menyanthaceae)

in North America. Bartonia 42: 14-23.

Swofford, D. L. 2002. Phylogenetic analysis using parsimony (*and other methods), ver. 4.

Sunderland, Massachusetts: Sinauer Associates.

Takagawa, S., I. Washitani, R. Uesugi, and Y. Tsumura. 2006. Influence of inbreeding depression

on a lake population of Nymphoides peltata after restoration from the soil seed bank.

Conservation Genetics 7: 705-716.

Uesugi, R., K. Goka, J. Nishihiro, and I. Washitani. 2004. Allozyme polymorphism and

conservation of the Lake Kasumigaura population of Nymphoides peltata. Aquatic

Botany 79: 203-210.

Uesugi, R., N. Tani, K. Goka, J. Nishihiro, Y. Tsumura, and I. Washitani. 2005. Isolation and

characterization of highly polymorphic microsatellites in the aquatic plant, Nymphoides

peltata (Menyanthaceae). Molecular Ecology Notes 5: 343-345.

Uesugi, R., J. Nishihiro, Y. Tsumura, and I. Washitani. 2007. Restoration of genetic diversity

from soil seed banks in a threatened aquatic plant, Nymphoides peltata. Conservation

Genetics 8: 111-121.

Van der Velde, G. and L. A. van der Heijden. 1981. The floral biology and seed production of

Nymphoides peltata (Gmel.) O.Kuntze (Menyanthaceae). Aquatic Botany 10: 261-293.

Van der Voo, E. E. and V. Westhoff. 1961. An autecological study of some limnophytes and

helophytes in the area of large rivers. Wentia 5: 163-258.

Wang, Y., Q. Wang, Y.-H. Guo, and S. C. H. Barrett. 2005. Reproductive consequences of

interactions between clonal growth and sexual reproduction in Nymphoides peltata: a

distylous aquatic plant. New Phytologist 165: 329-336.

Williamson, M. and A. Fitter. 1996. The varying success of invaders. Ecology 77: 1661-1666.

131

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

TABLE 4.01 – Populations of Nymphoides peltata studied in North America.

Population ID State County Locality Voucher Tippery 198 ColCT Connecticut Tolland private pond, Columbia (CONN) Pachaug State Forest, Tippery 301 PacCT Connecticut New London Voluntown (CONN) Tippery 293 DreNY New York Washington private boat launch, Dresden (CONN) Rogers Island Wildlife Tippery 83 RogNY New York Greene Management Area (CONN) river mudflat, town of Tippery 19 StiNY New York Saratoga Stillwater (CONN) Tippery 96 TicNY New York Essex near Fort Ticonderoga (CONN) Ockanickon Scout Tippery 97 OckPA Pennsylvania Bucks Reservation (CONN) Ringing Rocks Park, Lower Tippery 98 PotPA Pennsylvania Montgomery Pottsgrove (CONN) unnamed lake, near Baldwin s.n. BelWA Washington Whatcom Bellingham (CONN) Baldwin s.n. LynWA Washington Whatcom unnamed lake, near Lynden (CONN) Lake Spokane Tippery 183 SpoWA Washington Spokane (a.k.a. Long Lake) (CONN)

132

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

TABLE 4.01 (continued)

Population ID Coordinates Water body area (ha) Estimated N. peltata cover (ha) 41d 42m 10s N ColCT 0.04 0.04 72d 15m 29s W 41d 37m 05s N PacCT 36.43 Unknown 71d 48m 58s W 43d 40m 20s N DreNY River Unknown 73d 24m 31s 42d 14m 05s N RogNY River 1.17 73d 50m 04s W 42d 56m 13s N StiNY River Unknown 73d 39m 25s W 43d 50m 21s N TicNY River Unknown 73d 23m 11s W 40d 25m 57s N OckPA 1.62 0.21 75d 05m 15s W 40d 16m 08s N PotPA 0.32 0.24 75d 36m 08s W 48d 43m 10s N BelWA Unknown Unknown 122d 29m 52s W 48d 56m 16s N LynWA 0.42 Unknown 122d 38m 50s W 47d 47m 42s N SpoWA River 56.96 117d 33m 58s W

133

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

TABLE 4.02 – Alleles obtained from microsatellite analysis of Nymphoides peltata

populations. Loci for which only one length variant was recovered are depicted as homozygous.

Np152 Np207 Np274_1 Np274_2 Np306 1 2 1 2 1 2 1 2 1 2 U.S.A.: ColCT 154 160 128 128 165 165 134 134 175 175 U.S.A.: PacCT 154 160 128 128 165 165 134 134 175 175 U.S.A.: DreNY 154 160 128 128 165 165 134 134 175 175 U.S.A.: RogNY 1 154 160 128 128 165 165 134 134 175 175 U.S.A.: RogNY 2 154 160 128 128 165 165 134 134 175 175 U.S.A.: RogNY 2 seedling A 154 160 128 128 165 165 134 134 175 175 U.S.A.: RogNY 2 seedling B 154 160 128 128 165 165 134 134 175 175 U.S.A.: RogNY 2 seedling C 154 160 128 128 165 165 134 134 175 175 U.S.A.: RogNY 2 seedling D 154 160 128 128 165 165 134 134 175 175 U.S.A.: StiNY 154 160 128 128 165 165 134 134 175 175 U.S.A.: TicNY 154 160 128 128 165 165 134 134 175 175 U.S.A.: OckPA 1 154 160 128 128 165 165 134 134 175 175 U.S.A.: OckPA 2 154 160 128 128 165 165 134 134 175 175 U.S.A.: PotPA 154 160 128 128 165 165 134 134 175 175 U.S.A.: BelWA 154 160 128 128 165 165 134 134 175 175 U.S.A.: LynWA ? 160 128 128 165 165 134 134 175 175 U.S.A.: SpoWA 154 160 128 128 165 165 134 134 175 175 Cultivated: MD Aquatic Nurs. 154 160 128 128 165 165 134 134 175 175 Cultivated: Paradise W. G. 154 160 128 128 165 165 134 134 175 175 Hungary 160 160 128 128 165 165 134 134 175 175 India: Srinagar, Golmarg 160 160 130 155 165 165 134 136 175 175 Japan: Kajiyama 155 155 130 159 167 167 134 136 177 177 Japan: Oofunatsu 155 155 130 155 151 167 134 136 177 177 Japan: Tomita 155 165 130 159 167 167 136 138 177 177 Japan: Tsumagi ? ? 155 155 151 167 136 138 177 177 Japan: Wadamisaki 155 165 130 159 167 167 136 138 177 177 Korea: Uponeup Marsh 160 160 128 128 165 165 134 134 177 177 Sweden: Lk Glan 160 160 128 128 165 165 138 138 175 175 Sweden: Lk Mälaren 160 160 128 128 165 165 134 138 175 175 Sweden: Lk Sågsjön 154 160 ? ? 163 165 138 138 175 175 Sweden: Lk Sommen 154 160 128 128 163 165 138 138 175 175

134

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

TABLE 4.02 (continued)

Np641 Np694_1 Np694_2 Np730 1 2 1 2 1 2 1 2 U.S.A.: ColCT 221 221 157 157 131 131 163 163 U.S.A.: PacCT 221 221 157 157 131 131 163 163 U.S.A.: DreNY 221 221 157 157 131 131 163 163 U.S.A.: RogNY 1 221 221 157 157 131 131 163 163 U.S.A.: RogNY 2 221 221 157 157 131 131 163 163 U.S.A.: RogNY 2 seedling A 221 221 157 157 131 131 163 163 U.S.A.: RogNY 2 seedling B 221 221 157 157 131 131 163 163 U.S.A.: RogNY 2 seedling C 221 221 157 157 131 131 163 163 U.S.A.: RogNY 2 seedling D 221 221 157 157 131 131 163 163 U.S.A.: StiNY 221 221 157 157 131 131 163 163 U.S.A.: TicNY 221 221 157 157 131 131 163 163 U.S.A.: OckPA 1 221 221 157 157 131 131 163 163 U.S.A.: OckPA 2 221 221 157 157 131 131 163 163 U.S.A.: PotPA 221 221 157 157 131 131 163 163 U.S.A.: BelWA 221 221 157 157 131 131 163 163 U.S.A.: LynWA 221 221 157 157 131 131 163 163 U.S.A.: SpoWA 221 221 157 157 131 131 163 163 Cultivated: MD Aquatic Nurs. 221 221 157 157 131 131 163 163 Cultivated: Paradise W. G. 221 221 157 157 131 131 163 163 Hungary 221 245 161 161 131 131 ? ? India: Srinagar, Golmarg ? ? 161 161 131 131 ? ? Japan: Kajiyama 219 235 169 169 125 125 144 144 Japan: Oofunatsu 235 235 169 169 125 125 144 169 Japan: Tomita 221 221 161 161 125 125 144 144 Japan: Tsumagi 221 235 169 169 125 125 144 169 Japan: Wadamisaki 219 221 161 161 125 125 144 144 Korea: Uponeup Marsh 221 221 155 155 125 125 137 137 Sweden: Lk Glan 221 245 157 161 131 131 148 148 Sweden: Lk Mälaren 221 245 157 161 131 131 148 150 Sweden: Lk Sågsjön 221 221 157 157 131 131 163 163 Sweden: Lk Sommen 221 221 161 161 131 131 150 150

135

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

APPENDIX 4.01 – Locality and voucher information for Nymphoides peltata plants collected outside of the U.S.A. or from cultivated sources.

Cultivated: Maryland Aquatic Nurseries, Tippery s.n. (CONN); Paradise Water Gardens,

Tippery s.n. (CONN); Hungary: Wilstermann s.n. (CONN); India: Srinagar, Golmarg,

R.J.Rodin 8236 (UC); Japan: Lk Kasumigaura: Kajiyama, J. Nishihiro s.n.; Lake Kasumigaura:

Oofunatsu, J. Nishihiro s.n.; Lake Kasumigaura: Tomita, J. Nishihiro s.n.; Lake Kasumigaura:

Tsumagi, J. Nishihiro s.n.; Lake Kasumigaura: Wadamisaki, J. Nishihiro s.n.; Korea:

Changnyeong-gun, Gyeongsangnam-do: Uponeup (Upo) Marsh, Hyunchur Shin s.n. (CONN);

Sweden: Lake Glan, D.Larson s.n.; Lake Mälaren, D.Larson s.n.; Lake Sågsjön, D.Larson s.n.;

Lake Sommen, D.Larson s.n.

136

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

FIG. 4.01 – Populations of Nymphoides peltata in North America that were examined for this study. Locality abbreviations are given in the text.

FIG. 4.02 – Photograph of Apis mellifera collected at the RogNY population of

Nymphoides peltata on 20 Sep 2008 and preserved in 70% ethanol, focusing on the corbicula, on which are packed numerous grains of N. peltata pollen.

FIG. 4.03 – Chart of pollen accumulation throughout the day on 20 Sep 2008 at the

RogNY population of Nymphoides peltata. Unopened flowers were collected at 1030hr and

1600hr, and hand-pollinated flowers also were collected at 1600hr. Points are randomized slightly to show multiple points that overlap.

FIG. 4.04 – Histogram of seed set per fruit across all populations, onto which are overlaid per-population means and standard deviations for open-pollinated fruits of unknown age.

Individual points are plotted for fruits of known age collected on 20 Oct 2008 (with unknown treatment) and 21 Sep 2009 (either open-pollinated or outcrossed by hand).

FIG. 4.05 – Histogram of percent seed germination per fruit across all populations, onto which are overlaid per-population means and standard errors for open-pollinated fruits of unknown age. Individual points are plotted for fruits of known age collected on 20 Oct 2008

(with unknown treatment) and 21 Sep 2009 (either open-pollinated or outcrossed by hand).

FIG. 4.06 – Plot of percent seed germination against seed number per fruit. Populations of open-pollinated fruits of unknown age have different symbols, and additional symbols are provided for fruits of known age collected on 20 Oct 2008 (with unknown treatment) and 21 Sep

2009 (either open-pollinated or outcrossed by hand).

137

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

FIG. 4.07 – Plot of percent seed germination against seed color, quantified as a number between 0 (white) and 255 (black). Fruits of the RogNY population collected on 21 Sep 2009 are plotted, with separate symbols provided for fruits of known age collected on 20 Oct 2008 (with unknown treatment) and 21 Sep 2009 (either open-pollinated or outcrossed by hand).

FIG. 4.08 – Neighbor-joining tree of genetic distance between populations of Nymphoides peltata sampled, evaluated using a binary matrix of presence or absence for each allele.

Populations from the native range of N. peltata have their country of origin indicated, with populations from Sweden annotated to the genotypes reported previously (Larson 2007). Plants from naturalized North American populations and grown in cultivation are identical and clustered together; their identities and genotypes are given in Table 4.02.

138

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

FIG. 4.01

139

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

FIG. 4.02

140

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

FIG. 4.03

141

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

FIG. 4.04

142

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

FIG. 4.05

143

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

FIG. 4.06

144

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

FIG. 4.07

145

Chapter 4: Population biology and the invasive potential of Nymphoides peltata

FIG. 4.08

146

View publication stats