Investigations of morphological and molecular variation in wild and cultivated violets (Viola; Violaceae)

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By

Daniel William Howard Robarts, B.S.

Graduate Program in Evolution, Ecology and Organismal Biology

The Ohio State University

2013

Dissertation Committee:

Andrea D. Wolfe, Advisor

Pablo Jourdan

Laura Kubatko

Harvey Ballard

Copyright by

Daniel William Howard Robarts

2013

Abstract

The genus Viola is a large and diverse group of flowering plants. The objectives of this study were to develop and explore new morphological and molecular tools with horticultural and systematic applications.

Chapter 1 employed digital image analysis software, Tomato Analyzer, for flower morphology analysis of a 127 accession collection of sect. Melanium violets ("pansy group"). Seventy-seven traits associated with shape, size, and color were scored separately as categorical or continuous variables. The qualitative analysis was favored, capturing more of the variation and receiving higher bootstrap support in cluster analysis dendrograms. Cluster and ordination analyses indicated that the presence of blotch was the primary grouping factor, and secondarily, measures of color and shape (e.g., petal width). Uniformity across accessions of some hybrid morphotypes (e.g., "white with blotch") led to tight clustering across analyses. There were no significant correlations between clustering patterns and accessions' originating country or parent company, as had been previously reported.

Chapter 2 builds on the morphological analysis described in Chapter 1 by utilizing sequence-related amplified polymorphism (SRAP) markers to further characterize the collection of Melanium violets. Here, SRAP fragments indicated no significant differences between the horticultural classes of violets, though more were generated from

ii species types than hybrid types of the same ploidy. Bayesian analysis suggested distinctive structure clusters within the collection, but was obscured by high levels of admixture. Some color forms (white, white with blotch, yellow with blotch, and orange) tended to cluster strongly together. Correlation analysis of morphological and molecular datasets, as well as analysis of a combined dataset, underscored the conclusion that some genetic lines could be generalized by blotch presence and flower color. The relationship between these data may help in optimizing germplasm collection management and development of SRAPs by breeders for marker-assisted selection.

Chapter 3 describes the genetic diversity within Viola pedata, a North American perennial violet with a bicolorous floral syndrome similar to that seen in taxa of the pansy group. Four hundred fifty-eight individuals collected from 42 unique populations and one cultivar were sampled in eastern North America, and compared using microsatellite markers. Results confirm that this species is a high order polyploid (>4x), and underlying structure indicates previous refugial populations in the Driftless Area of the upper Midwest during previous glacial maxima. A genetic discontinuity east and west of the Appalachian Mountain range was depicted by neighbor-joining, principal coordinates, and Bayesian approaches, a pattern seen in many other plant taxa.

Chapter 4 reviews the uses of SRAP markers in the current literature. Since their inception in 2001, over 300 published articles have described the use of markers, with incidence of use increasing every year. Analysis of comparative studies shows SRAP markers generating similar levels of amplicons and polymorphism as AFLP, but with drastically fewer, more simple steps. In presenting relevant case studies, SRAP markers

iii are shown as valuable, but underexplored tools to address hypotheses in topics beyond those in the applied sciences for which they were developed. These research areas include systematics, biogeography, ecology, and conservation.

iv

Dedication

I dedicate this dissertation to my family, with the deepest gratitude and appreciation for their enduring support of my efforts to complete this endeavor. Their consistent patience and faith

in me have not waivered through time or distance, and truly allowed me to follow my

dreams.

I offer special thanks to my partner in life, Claire, who followed me though countless long

nights of dissertation toiling as girlfriend, fiancée, wife, and soon-to-be mother of our twin

boys. Thank you, especially, for acceptance of my phytomania, and embracing my passion

for the natural world.

v

Acknowledgments

While the first page of this document proffers only my name as author, it would not have come into existence without the generous contributions of many.

I humbly thank my advisor, Dr. Andrea Wolfe, for her guidance and support though the entirety of this endeavor. Her patience and mentoring have been a true blessing, and helped me develop as an independent researcher. A debt of gratitude is also owed to the financial support of the Ornamental Plant Germplasm Center (OSU-USDA), and its director,

Dr. Pablo Jourdan. Without their interest and provision, none of this research would have come to fruition. It is my hope the research described herein may serve to underscore the immense value of this institution and its endeavors. Also, I give thanks to my other advisors,

Dr. Harvey Ballard, who was always in my corner and provided expertise in Viola, and to Dr.

Laura Kubatko, for lending her guidance in resolving my statistical quandaries.

Thanks to all who assisted in locating, acquiring, and maintaining plant materials for analysis, including Bryan Connolly (MA NHESP), Tony Recznicek (UM), Ted Cochrane

(UW), Robert Capers (UConn), Steven Carroll (State Arboretum of Virginia), Dennis Bell

(ULM), Peter Stefany (PanAmerican Seed), David Snodgrass (OSU), Peter Zale (OSU), and many other researchers, naturalists, and photographers (thank you Flickr friends!). Special thanks to Jose Diaz, fellow WOLFE PACK comrades Aaron Wenzel, Paul Blischak,

Shannon Kilkenny, and graduate colleagues including Bruce Ackley, Chia-Hua Lin, and

Mike Sovic. Your daily research assistance was critical for my forward progress.

vi

Vita

June 2000 ...... White Mountains Regional High School

May 2004 ...... B.S. Biology, Bates College

2006-2010 ...... Graduate Research/Teaching Associate,

Department of Horticulture and Crop

Science, The Ohio State University

2010 to present ...... Graduate Research/Teaching Associate,

Department of Evolution, Ecology, and

Organismal Biology, The Ohio State

University

Publications

Robarts, D., & Baum, M. J. (2007). Ventromedial hypothalamic nucleus lesions disrupt olfactory mate recognition and receptivity in female ferrets. Hormones and Behavior, 51(1), 104-113.

Batterton, M. N., Robarts, D., Woodley, S. K., & Baum, M. J. (2006). Comparison of odor and mating-induced glomerular activation in the main olfactory bulb of estrous female ferrets. Neuroscience Letters 12, 400(3), 224-229.

McCormick, C. M., Robarts, D., Kopeikina, K., Kelsey, J. E. (2005). Long-lasting, sex- and age-specific effects of social stressors on corticosterone responses to restraint and on locomotor responses to psychostimulants in rats. Hormones and Behavior, 48(1), 64-74.

vii

McCormick, C. M., Robarts, D., Gleason, E., Kelsey, J. E. (2004). Stress during adolescence enhances locomotor sensitization to nicotine in adulthood in female, but not male, rats. Hormones and Behavior, 46(4), 458-466.

Fields of Study

Major Field: Evolution, Ecology and Organizmal Biology

Focus: Ornamental Plant Characterization

viii

Table of Contents

Abstract ...... ii

Dedication ...... v

Acknowledgments...... vi

Vita ...... vii

Publications ...... vii

Fields of Study ...... viii

Table of Contents ...... ix

List of Tables ...... xii

List of Figures ...... xiv

Chapter 1. Morphological variation in a collection of pansies (Viola sect. Melanium) detected by digital image analysis ...... 1

Abstract ...... 1

Introduction ...... 3

Methods ...... 10

Results ...... 15

Discussion ...... 17 ix

Tables and Figures ...... 30

Chapter 2. Molecular variation in a collection of pansies (Viola sect. Melanium) characterized by sequence-related amplified polymorphism (SRAP) markers ...... 36

Abstract ...... 36

Introduction ...... 38

Methods ...... 44

Results ...... 50

Discussion ...... 56

Tables and Figures ...... 67

Chapter 3. Phylogeography of Viola pedata L. (Violaceae) ...... 86

Abstract ...... 86

Introduction ...... 88

Methods ...... 95

Results ...... 102

Discussion ...... 106

Chapter 4. Sequence-related amplified polymorphism (SRAP) markers: A potential resource for studies in plant molecular biology ...... 143

Abstract ...... 143

Introduction ...... 144

x

Methods ...... 148

Results and Discussion ...... 149

Population level (intraspecific) systematics ...... 150

Hybridization ...... 154

Higher-order (interspecific) systematics ...... 158

Biogeography...... 162

Conservation genetics ...... 165

Ecology ...... 168

References ...... 172

Appendix A. Accession list and description of Melanium violets ...... 235

Appendix B. Morphological characters and descriptions for Melanium violets ...... 240

Appendix C. Tomato Analyzer calibration information ...... 243

Appendix D. One-way ANOVA tables comparing molecular measures between clusters of Viola pedata ...... 249

Appendix E. Reviewed SRAP publications ...... 258

xi

List of Tables

Table 2.1. Sequence-related amplified polymorphism primers screened to characterize

Melanium violets ...... 67

Table 2.2. Descriptive results of fragments generated by SRAP primers and calculated error rates ...... 68

Table 2.3. Student’s t-test comparing SRAP fragment number between classes of

Melanium violets ...... 71

Table 2.4. AMOVA results for STRUCTURE clusters of Melanium violets ...... 72

Table 3.1. Locality information of Viola pedata collections ...... 129

Table 3.2. Descriptive statistics and missing data of microsatellite markers in Viola pedata ...... 131

Table 3.3. Mean differentiation, diversity, and heterozygosity variables of Viola pedata populations ...... 132

Table 3.4. Significant results from Pearson's correlation analysis of geographic and molecular variables in Viola pedata ...... 134

Table 3.5. ANOVA results of cluster-level molecular variables in Viola pedata ...... 141

Table 3.6. AMOVA results by locus in microsatellite analysis of Viola pedata ...... 142

Table A.1. Accession list and description of Melanium violets ...... 235

xii

Table B.1. Characters and descriptions for morphological analysis of Melanium violets

...... 240

Table C.1. Descriptions of 72 Munsell color chips and their interpretation by colorimeter and flat-bed scanner...... 243

Table D.1. ANOVA tables of molecular measures for clusters of Viola pedata ...... 249

Table E.1. Descriptive information for reviewed SRAP publications ...... 258

xiii

List of Figures

Figure 1.1 Anatomy of Melanium violet flowers...... 30

Figure 1.2. Examples of morphological analysis in Tomato Analyzer software ...... 31

Figure 1.3. Principal coordinates scatterplot of QT Viola data...... 32

Figure 1.4. Cluster analysis dendrograms of QT Melanium violet data ...... 33

Figure 1.5. Principal components analysis scatterplot of 30 morphometric points ...... 35

Figure 2.1. ROX799 size standard for scoring fragments in capillary electrophoresis .... 69

Figure 2.2. Example test gel of SRAP amplicons in Melanium violets ...... 70

Figure 2.3. Principal coordinates analysis of SRAP fragments in Melanium violets with superimposed STRUCTURE results...... 74

Figure 2.4. Cluster analysis of SRAP results in Melanium violets with superimposed

STRUCTURE results ...... 76

Figure 2.5. STRUCTURE Harvester results for SRAP fragments in Melanium violets .. 80

Figure 2.6. STRUCTURE barplots of K=6 and K=8 in Melanium violets ...... 81

Figure 2.7. Mantel test of molecular vs. morphometric similarity matrices in Melanium violets ...... 82

Figure 2.8. Principal coordinates analysis scatterplot of combined morphological and

SRAP data in Melanium violets ...... 83

Figure 2.9. Cluster analyses of combined data in Melanium violets ...... 84

xiv

Figure 3.1. County-level distribution of Viola pedata ...... 125

Figure 3.2. Examples of seasonal variation in Viola pedata leaf morphology ...... 126

Figure 3.3. Flower variants of Viola pedata ...... 127

Figure 3.4. Distribution map of Viola pedata varieties ...... 128

Figure 3.5. Example electropherogram of microsatellite data for Viola pedata suggesting higher-order ploidy ...... 130

Figure 3.6. Mantel test scatterplots of relationships between pair-wise geographic distance and fixation indices in Viola pedata accessions ...... 135

Figure 3.7. Consensus neighbor-joining tree of 42 Viola pedata populations...... 136

Figure 3.8. Principal coordinates analysis scatterplots of Viola pedata accessions labeled by STRUCTURE cluster...... 137

Figure 3.9. STRUCTURE Harvester results indicating likely K values for 43 Viola pedata populations ...... 138

Figure 3.10. STRUCTURE barplots of 43 Viola pedata populations arranged by

STRUCTURE cluster...... 139

Figure 3.11. Geographic representation of STRUCTURE (K=4) clustering of Viola pedata populations ...... 140

Figure 4.1. Subset (n=188) of publications employing SRAP markers by year of publication...... 171

Figure C.1. Regression of colorimeter and flat-bed scanner values of L* for 72 Munsell color chips...... 246

xv

Figure C.2. Regression of colorimeter and flat-bed scanner values of a* for 72 Munsell color chips...... 247

Figure C.3. Regression of colorimeter and flat-bed scanner values of b* for 72 Munsell color chips...... 248

xvi

Chapter 1. Morphological variation in a collection of pansies (Viola sect. Melanium) detected by digital image analysis

Abstract

The domestication of plants has brought about distinct changes in plant morphology, enhancing characters such as shape, size, and color, relative to their wild progenitors. These morphological changes are well known in agronomic crops, but are often more apparent in ornamental taxa. Violets (Viola spp., Violaceae) have been cultivated since ancient times for their beauty, and today advanced hybrids from the Old

World “pansy group” (section Melanium Ging.) maintain international horticultural ubiquity. Hundreds of polyploid, F1 hybrids currently exist in the marketplace, and have been among the highest-value annual bedding plants for decades. In order to enhance breeding efforts and help characterize large germplasm collections of these hybrid violets, efficient methods of morphological characterization are critical.

Digital images were collected of flowers from a large collection of contemporary hybrid (V. × wittrockiana Gams., VW; V. cornuta L., VC) and species type Melanium violets (n=127) using a flat-bed scanner. Twenty flowers were collected from each of five replicate plants, and characterized by 78 traits associated with shape, size, and color through novel use of Tomato Analyzer (TA) software. Characters were scored as both continuous (QT) and categorical (QL), and the two approaches were compared through examination of variation captured in ordination (principal coordinates) analysis, and the 1 degree of bootstrap support and cophenetic correlation coefficients (CPCC) in clustering analyses (UPGMA, Ward). Traditional morphometric points (MP; n=30) were also implemented to characterize flower shape, assessed via principal components analysis.

Principal coordinates analysis of QT data explained 51.86% of the variation in the first three axes, while QL captured 33.2%. Clustering via Ward’s method gave a higher

CPCC for the QT dendrogram (0.9264) than the QL (0.6379), but UPGMA analysis of the two yielded comparable CPCC scores (QT, 0.7306; QL, 0.7621). As analysis of QL data explained significantly less variation than the QT across all analyses, it was not pursued further. The low support scores for the QL dataset may have been due to creation of biologically invalid category bins. A primary trend for clustering by blotch presence was seen in all QT analyses. Secondary clustering of hybrids was apparent by color and then flower shape. Species-types clustered apart from hybrids, regardless of color, due to their distinctive flower shapes. Cluster analyses tended to group accessions by horticultural class (VW vs. VC), but this may have been due to the bias in VW cultivars to be blotched, and for VC to be unblotched. Principal components analysis of

MP indicated a single, undivided cluster, with the roundest-flowered (widest-petaled) cultivars on one end of the spectrum, and narrow-petaled species at the other. Higher variation contributions by MP 10-22 suggested the shape of the lateral and inferior petals were the most important in differentiating accessions. No patterns of clustering by hybrid origin were detected, as had been previously reported. Overall, this study described the diversity of morphology in pansy group hybrids, and their divergence from wild types.

Further, it expanded the applicability of TA to morphometric analysis of floral structures.

2

Introduction

The domestication of plants has brought about distinct changes in plant morphology, expanding the degree of variation in characters such as shape, size, and color, relative to their wild progenitors (Perry & McIntosh, 1991; Ferriol, Pico, & Nuez,

2004a; Doebley, 2004; Li, Zhou, & Sang, 2006; Cuevas et al, 2007; for reviews see

Gepts, 2011; Miller & Gross, 2011). Morphological diversification in the plant structures is well-known to be one of major elements in the ‘domestication syndrome’ of crop plants (Hawkes, 1983; Gepts, 2004; Purugganan & Fuller, 2009; Zohary et al., 2012), and is equally apparent in ornamental taxa (Schmid, 1992; Van Tuyl, 1996; Horn, 2004; Petit

& Peat, 2004; Anderson, 2006). Selection of ornamentals mirrored that of crop plants, but weight of yield was given to their desirable, aesthetic traits. Maintenance and accumulation of valuable characters were primarily accomplished by cultivation of seed- derived lines (Paxton, 1849). Unfortunately, little verified information has been recorded of historical, horticultural endeavors to cultivate new traits in ornamental plants, providing little background for contemporary and future generations of flower breeders and geneticists (Anderson, 2006).

Gregor Mendel (1866) was the first to apply scientific methods to examine the inheritance of flower color. Empirical studies of deliberate hybridization of divergent morphotypes became of interest for both ornamental and systematic purposes in subsequent years (East, 1916; Kristofferson, 1923; Clausen, 1926). Although many genera from the Middle East and Asia such as Paeonia (Harding, 1917; Xu, 2005), Rosa

3

(Goody, 1993), Tulipa (Segal, 1993), orchids (Hew, 2001), and Narcissus (Wylie, 1952;

Nunez et al., 2003) have been cultivated for millennia, the directed breeding of numerous contemporary, herbaceous ornamental plants in commerce began in western Europe, during the Victorian era of the nineteenth century.

Violets (Viola spp., Violaceae) have also been cultivated since ancient times for their beauty and utility (Coombs, 1981). Today, taxa from the section Melanium Ging., known as the “pansy group”, may be the best known of the genus due to their broad, Old-

World distribution and modern horticultural ubiquity. The section Melanium is primarily a Eurasian group, with a few representatives in northern Africa and one disjunct species in North America (Clausen et al., 1964). It is comprised of approximately 80-100 species

(Erben, 1985; Ballard, 1996), the majority of which are herbaceous, with distinctive floral morphological characteristics (Clausen, 1927; Erben, 1985). Like all Viola, the pansies’ zygomorphic flowers produce five petals: two symmetrical pairs and one inferior petal

(Figure 1.1). Specific to the Melanium group are frontally flattened flowers, with lateral petals upturned (whereas these petals are downturned in most other sections), and a single, inferior petal that is enlarged, probably acting as landing space for insect pollinators (Beattie, 1971). Flowers also display an expanded yellow throat, typically demarcated by contrasting, dark venation serving as a nectary guide. Most taxa in this group also produce large, leaf-like stipules, and demonstrate the absence of the cleistogamous syndrome typical of the genus (West, 1930; Meyers & Lord, 1983;

Herrera, 1993; Culley, 2000, 2002; Winn & Moriuchi, 2009).

4

The earliest monographic works addressing the Melanium violets were provided by Wittrock (1895, 1897), followed by others focusing on V. tricolor L. and its allies

(Clausen, 1921; Schindler, 1923). Taxonomic identification of species within this section has been disputed (Gay, 1832; Clausen, 1921, 1931), potentially due to the inherent variability and adaptability of many of the more broadly distributed taxa (Clausen, 1921) as well as the general ease of interspecific hybridization, leading to localized, fertile hybrid complexes (Wittrock, 1895; Kristofferson, 1923; Clausen, 1924, 1926; Fothergill

1938, 1939, 1941; Pettet, 1964). Genetic introgression has been furthered by directed hybridization efforts of horticultural enthusiasts, and the cultivation of interspecific hybrid varieties.

Ornamental interest in breeding pansies began in earnest in the early nineteenth century, as cultivation of wild species in English estate gardens became en vogue (Quest-

Ritson, 2003). In the gardens of two great estates west of London, wild forms of Viola, primarily the locally available V. tricolor and V. lutea Huds., were collected for bedding plants around 1813 (Cook, 1903). Spontaneous seedlings arose in gardens where these wild-type plants were grown together, which demonstrated a diversity of flower color, pattern, and size of flowers, previously unseen. Discovery of these novel recombinants led to selection and subsequent hand pollinations, attempting to stabilize and further enhance these characters. One of these two cultivated lineages was developed by the gardener William Thompson, who has been credited for the greatest improvements from the wild-type collections and was known as “the Father of the Heart’s-ease”

(Cuthbertson, 1910). His efforts to create unique and large-flowered pansies were well

5 documented in horticultural literature (e.g., Sinclair & Freeman, 1835), but the importance of his contributions to the modern garden flower were captured in his frequently republished letter to the Flower Gardeners Library and Floricultural Cabinet, which was published in 1841 (as referenced in Cook, 1903; Crane, 1908; Cuthbertson,

1910; Genders, 1958). In this letter, he recounted the early years of his hybridization efforts, describing the form of the flowers to be “little more symmetrical than a child’s windmill” and “as lengthy as a horse’s head”, indicating the relative narrowness of wild- type conspecifics to the wide-petaled forms he was selecting from his hybridization efforts. Thompson also made note of the unexpected appearance of what was later called the “face” or “blotch,” which became a defining character of modern pansy-type Viola hybrids. This characteristic can be described as a darkened area below the (yellow) throat, or “honey spot” (Clausen, 1926), arising from the lateral expansion of the deeply- colored nectary guides to form an area of continuous color within the lighter-colored area beyond the throat (see Figure 1.1):

“Still up to this period, a dark eye, which is now considered one of the chief

requisites in a first-rate flower, had never been seen. Indeed such a feature had

never entered my imagination, nor can I take any merit to myself for originating

this peculiar property, for it was entirely the offspring of chance. In looking one

morning over a collection of heaths, which had been some time neglected, I was

struck, to use a vulgar expression, all of a heap, by seeing what appeared to me

a miniature cat's face steadfastly gazing at me.”

6

Thompson and contemporary horticulturalist James Lee were responsible for development some of the earliest and most reputed hybrid Viola cultivated varieties

(cultivars) through directed breeding for wider petals and the bold, contrasting blotch.

Visitors to the gardens of these Viola breeders in the 1830’s gave reports of numerous species being grown, including V. tricolor, V. lutea, V. grandiflora L. (V. lutea var. grandiflora), V. altaica Ker Gawler, V. amoena Leconte (Sinclair & Freeman, 1835).

Species-specific characters of other Melanium taxa were also detected in garden hybrids, including long spurs and mat-forming, perennial habit. It was presumed these traits were introduced through addition of V. cornuta L., V. calcarata L. (Wittrock, 1896), or V. gracilis Sm. (Genders, 1958) to the breeding stocks, though it was generally thought that the primary lineage could be attributed to V. tricolor and V. lutea (Darwin, 1868;

Wittrock, 1896; Crane, 1908). This supposition of a “two-parent” lineage was supported by their floral color and patterns, as these taxa demonstrated significantly more diversity in terms of coloration, which ranged from near white and bright yellows to rich, dark purples and combinations thereof, in contrast to flowers of other implicated progenitors that were often monochromatic. Of these two species, Viola tricolor has a wider range, endemic to Britain and north-central Europe into Russia and Siberia (Clausen, 1921) as compared to V. lutea which is restricted to the British Isles and north-central Europe.

With the ready availability of germplasm, the enthusiasm for pansy hybridization spread through Europe and soon expanded into a global phenomenon. This horticultural fervor

7 brought new appreciation to the genus, and initiated the first systematic attempts to describe and identify discrete taxonomic groups of species in sect. Melanium.

Plant breeders have cultivated and exploited many Viola species in their hybridizing programs to produce varieties demonstrating exaggerated traits of parental lineages. Interspecific hybrids have the potential to capture hybrid vigor as well as combine traits that do not occur within a single species (Volker & Orme, 1988). In many cases, developing novel traits has been achieved through thorough exploration of recombination and polyploidization. Polyploidization of somatic tissues typically leads to augmented characters, including larger, thicker flowers and leaves (Briggs & Knowles,

1967), and allowing for significantly more phenotypic variation via increased allelic recombination. Though the rise of the most naturally occurring polyploid hybrids can be traced to interspecific hybridization (Soltis et al. 1993, 2009), including the Melanium violets (Kristofferson, 1923; Clausen, 1931; Fothergill, 1938), many ornamental plants in cultivation today are derived from induced polyploidization events (e.g., Iris,

Hemerocallis, Hosta, Chrysanthemum, Lilium, Narcissus).

Modern commercial Melanium hybrids display tremendous variety of flower colors, patterns, and sizes as a result of exploitation of the breadth of underlying genetic diversity. Two primary cultivated lineages exist: the large-flowered tetraploid V. × wittrockiana (VW) cultivars, known as the “garden pansy”, widely grown as a cool season bedding plant, and the diploid V. cornuta (VC) hybrids, or “violas”, which have smaller flowers but are more floriferous and generally more adaptable and environmentally-tolerant. Hundreds of cultivated varieties exist on the market today

8

(Peter Stefany, personal communication, March 2, 2008) produced by numerous domestic and international companies, sold as F1 progeny derived from highly inbred lines. Advanced hybrids are continually developed, and carry great horticultural and monetary value, representing over 7.5% of annual bedding plants sold in the United

States (including vegetables) (Jerardo, 2007). These floriferous garden plants have created more than $100 million in annual wholesale value for more than a decade (United

States Department of Agriculture, National Agriculture Statistics Service website, accessed October 28, 2011), with even higher sales in Europe (Peter Stefany, personal communication, March 2, 2008). The United States horticultural market has proven pansies have been one of the top-selling annual bedding plants since the middle of the twentieth century (Genders, 1958).

Though cultivated Melanium hybrids have existed for nearly two hundred years, there have not been any recent reviews directed at quantitative floral variation attained via selective breeding in Viola. Investigations of diversity within ornamental crops have important consequences for genetic conservation and breeding (Tay, 2005, 2006), and provide opportunities to identify how wild ancestral populations have contributed to germplasm collections (Gepts, 1993). Further, detailed comparison of valuable, floral characteristics in cultivated and wild populations may yield important information on historical breeding efforts and traditional consumer preferences. The diverse pool of

Viola hybrids is worthy of examination due to its economic importance alone, but paired with the number of similar cultivars and industry interest for within-variety uniformity of modern F1 products, these hybrids are ideal subjects for inter- and intra-varietal

9 morphometric analysis. That said, to describe the discrete levels of quantitative variation that exist in modern cultivars, advanced tools for objective measurement would be necessary.

In this study, the morphologic variation in a large germplasm collection of hybrid

Viola cultivars from the section Melanium were examined in order to evaluate variation within modern commercial cultivars. To further review the range of characters in this sample and gauge the efforts of ornamental plant breeders hybridizing efforts in the

Melanium violets, wild-type accessions of proposed progenitor taxa were included in this analysis. Morphological traits were measured from digital images. Two regimes of morphological characters scoring (binary and continuous) were employed to evaluate their relative significance, descriptive capability, and effectiveness through different phonetic analyses.

Methods Examined accessions

Seed accessions of 127 Viola hybrids and species types were acquired from commercial breeding programs and horticultural sources (Appendix A). This collection was considered to be a representative sample of commercial cultivars on the market at the time of the study, including both industry-classified hybrid groups: V. × wittrockiana

(VW, “pansies,” presumed tetraploid) and V. cornuta (VC, “violas,” presumed diploid).

Accessions of multiple species mentioned in the literature as contributing to early hybridization efforts such as V. tricolor, V. altaica, V. corsica, and others (all presumed diploid), were also included. As per the request of private, donating entities, the origins 10 of proprietary cultivars were not identified by name or origin, but by color, pattern, and trade group (VW or VC).

Plant growth

Accessions were sown on ddH2O-saturated, 4x4 inch Steel Blue Germination

Blotter (Anchor Paper Company, St. Paul, MN) in an inverted 4.3125x4.3125x1.325 inch clear plastic case (model 156-C, Pioneer Packaging, Dixon, KY), and germinated in a

19C Percival Scientific growth chamber (model GR-36; Percival Scientific, Perry,

Iowa). After ~75% of seedlings had grown to the cotyledon stage, they were removed by hand and planted into 4x4x4” square pots in medium comprised (by volume) of 75%

SureClone 360 (Michigan Grower Products Inc., Galesburg, MI), 10% native soil, and

10% pearlite. Plants were arranged in a complete randomized design and grown in a

Conviron E15 growth chamber (Controlled Environments Limited, Winnipeg, Manitoba) under conditions of 71º F (21.7º C) day, and 65º F (18.3º C) night with 17 h days maintained by a 1:1 ratio of fluorescent and incandescent lights (350 mol/m2/sec; adapted from Bailey & Garner, 1995; Adams, Pearson, & Hadley, 1997). Lighting was checked with a digital light meter weekly to confirm intensity. Plants were watered every four days with 100 ppm 20-10-20 (N-P-K) liquid fertilizer, and supplementally watered with local, tap water as needed. Under these conditions, most plants flowered in 10-12 weeks.

Imaging and calibration

Five flowers from each of four plants were examined from each accession. The first three to five blooms on each plant were not selected for analysis, but were observed

11 from anthesis to senescence in order to appropriately gauge collection of similar, mature flowers, in terms of size, color, and shape, as these characteristics can be variable over a plant’s flowering period (Peter Stefany, personal communication, March 2, 2008).

Digital images of Viola flowers were generated by pressing the blooms to a flatbed scanner (Microtek ScanMaker 5900, Microtek, Inc.) held in place by a black polycarbonate plate (approx. 6” x 6” x 0.25”) and scanned at 300 dpi. The five petals

(two inferior, two lateral, and one inferior) were then separated by hand and similarly scanned, though only one petal of each type was scored. Deformation of flowers and petals was negligible due to their two-dimensional nature. Flowers and petals of some accessions were not accurately or reliably detected by analysis software (those with an area of less than 0.5 inches2 and those with comparable darkness in color to the polycarbonate plate background), and were scanned at progressively higher resolution

(600 dpi then 800 dpi) as needed for detection and measurement.

Images of flowers and petals were initially scored for 100 quantitative traits involving shape, size, and color (Appendix B) with Tomato Analyzer ver. 2.2.0 software

(TA; Brewer et al., 2006) and its Color Test function (TACT; Darrigues et al., 2008).

Within each image, flowers were rotated and boundary corrections (described in Brewer et al., 2006) were made such that the perimeter intersection of the two superior petals and the mid-vein of the inferior petal corresponded to the 12 o'clock (0/360) and 6 o'clock positions (180). The distal and proximal end points were set at these landmarks, respectively, to orient morphometric points (Figure 1.2-A). Thirty of these points were equidistantly set by TA, following the perimeter of each flower/petal, starting at the 12

12 o'clock position and circumscribing objects in a counterclockwise direction. In order to further characterize accessions that demonstrated the conspicuous floral color pattern known as a blotch, the “pericarp boundary” function was used. Images of whole flowers and petals were manually adjusted to circumscribe this feature, allowing for subsequent analysis in TA (Figure 1.2-B).

For varieties that demonstrated a blotch, the boundary was adjusted to exclude this feature when measuring ground color. To compensate for the inherent biases in color capture of flat-bed scanners (factors including make, model, bulb age, duration of scanner use, etc.), calibration of the scanner was accomplished by capturing 72 color chips representing the color diversity of colors seen in Viola accessions from the Munsell Book of Color (Neighboring Hues Edition, 1965; Appendix C) via scanner and Minolta CR300 colorimeter (Minolta, Ramsey, NJ). Microsoft Excel (Microsoft Corp., Redmond, WA) was used to plot the regression of TACT and colorimeter values in order to correct for discrepancies in the TACT dialog box (see Appendix C for details), as described in the

TACT manual. To determine the temporal color-scoring precision of the scanner as the bulb was used, one red color chip (Chip ID #13; Munsell 5R 4/14) was scored every 30 minutes during use, and also at the beginning and end of each session.

Data transformation

To avoid biases by including correlated traits, Pearson’s r was calculated to examine the linear dependence between the 100 traits. Traits scoring higher than r=0.9 in any pair-wise comparisons were excluded from the analysis, which yielded 78 characters for further examination (3 qualitative, 75 quantitative - see Appendix B for details). In

13 order to examine the contributions of these characters to differentiation between accessions, two transformation regimes were employed. Under the first regime, quantitative (QT) character means were standardized by a log10 transformation. In instances where these means were negative values (i.e. color traits, L*, a*, b*), the character means were numerically scaled up so the lowest accession mean was equal to one, thus allowing for log10 transformation. Under the second regime, raw, quantitative means were reviewed by trait, and divided into subgroups, and scored as present or absent (1, 0) within those categories creating a binary, qualitative (QL) data matrix.

Subdivisions were determined by examining means in ascending order, and calculating the step-wise differences between values. The largest, single difference between two accessions was determined, and presumed to be a “biological division”, and at every pair- wise difference of 90% or more of this difference, a subcategory boundary was created.

Number of subcategories per trait ranged from 3 to 10 per character.

Data Analysis

The two datasets were subjected to multiple phenetic analyses. Principal coordinates analysis (PCoA) was used to as it does not require the a priori classification of specimens, and can incorporate qualitative data, which is inapplicable in principal components analysis (PCA; Gower, 1966). Data were also subjected to cluster analysis via unweighted pair group method with the arithmetic mean (UPGMA; Sokal &

Michener, 1958) and Ward’s minimum variance method (Ward, 1963) in PAST software

(Hammer, Harper, & Ryan, 2001). Gower and Jaccard similarity coefficients were used for mixed and binary datasets, respectively, and a Euclidean distance approach was

14 employed under Ward’s method. Each cluster analysis was resampled with1,000 bootstrap replicates to estimate support for topologies.

In order to assess the relationships between the original pair-wise distance between accessions (“true distances”) and the predictions suggested by cluster analyses’ dendrograms, the cophenetic correlation coefficient (CPCC; Sokal & Rohlf, 1962) was also calculated. The CPCC has been shown to be an appropriate measure to assess the validity of clustering assumptions made by different algorithms. The CPCC has been specifically used to determine the most meaningful clustering assessment when comparing UPGMA and Ward’s methods, depending on the relative value difference between the two approaches (Odong et al., 2011). This measure was compared over cluster analyses to determine which dataset (QT or QL) was to be pursued for further examination, though some have found the CPCC to have limited value in assessing patterns suggested by a tree (Holgersson, 1978).

Morphometric (X, Y) points were averaged over plants within accessions and subjected to PCA under UPGMA constraints.

Results

The PCoA imposing Gower’s similarity measure (Figure 1.3) indicated 31.7% of the variation was expressed by the first axis, and 51.86% expressed by the first three axes. The first and third axes separated accessions primarily by horticultural class

(assumedly correlated with size) and blotch presence, while color dominated the pattern in the second axis. UPGMA (Figure 1.4-A) and Ward (Figure 1.4-B) clustering analyses

15 of the QT data demonstrated a strong signal for distinction between cultivars based on the presence of the blotch, and to lesser degrees grouping by color and shape. Across analyses, some colors, including white with and without blotch, yellow with blotch, and orange, formed more cohesive clusters, whereas the purple, blue, and lavender flowered accessions were generally intermixed. Due to their size and shape, species-type accessions tended to group together as well. Measures of shape including ratios of flower and petal perimeter : area (P:A) and height : width (H:W, “shape index 1”) played significant roles in clustering, though generally secondary to color. Comparison of

CPCC between the UPGMA and Ward analyses indicated that the latter produced a tree more representative of the data (higher CPCC scores), though patterns of clustering results were comparable between the two.

Analyses of the QL data yielded similar patterns of relationships between accessions to those suggested by analysis of the transformed, QT values, clustering accessions based on blotch, size / shape, and color. The axes that represented the associations with these traits mirrored that of the QT analysis. PCoA of QL data yielded a similar distribution of accessions in the first three axes when represented as a scatter plot, which was apparent by greater spacing of points, and also it’s decreased capture of the total variation in the first three axes (33.2%). Similarly, the UPGMA and Ward dendrograms indicated a greater level of dissimilarity manifested as a greater distance between accessions relative to the QT analysis. Further, these ordinations demonstrated lower CPCC scores than their respective QT analyses as well (UPGMA=0.7621,

Ward=0.6379).

16

Principal components analysis of 30 morphometric points from 127 accessions of

Melanium violets suggested the first axis explained 93.6% of the variation and 98.1% by the first three axes (Figure 1.5). No single points yielded high levels of variation, but contributions were higher for points between 10 and 22, which suggested the shape of the lateral and inferior petals produced the highest levels of variation. Analysis of the data collected from the control, red color chip (#13 - see Appendix C) scored for scanner consistency, indicated no significant difference (no between measure difference greater than 4%) of the scored color traits (L*, a*, b*, hue, or chroma), suggesting the scanner utilized for this study did not vary significantly over the duration of digital image capture

(data not shown).

Discussion Digital morphometric analysis

Historically, morphometric analyses have involved time-consuming, manual measurements and subjective scoring of traits that restrict or bias the degree of heterogeneity within characters being examined. Digital imaging technology and computer processing power have increased dramatically over the last few decades. These advancements have allowed researchers to develop software to overcome individual subjectivity and bias in evaluation of discrete, quantitative traits at high speed, which has been critical for continued progress in detection and selection of valuable characters

(Soille, 2003).

The use of flat-bed scanners have been shown to be an inexpensive, efficient, and standardizable approach for capturing digital images of plant structures (Kleeberger & 17

Moser, 2002; Metzing, 2004). Many studies have demonstrated that analysis of many two-dimensional structures including leaves (Olmstead, Lang, & Grove, 2001; Anderson,

Reng, & Kirk, 2005; Wijekoon et al., 2008), seeds (Sako et al., 2001; Marcos Filho et al.,

2010), roots (Pan & Bolton, 1991; Himmelbauer & Kastanek, 2004), and flowers

(Yoshioka et al., 2006; Wang & Bao, 2007) can be done using this imaging technology.

Many specialized applications have been specifically designed to quantify desired aspects of objects captured in these digital fields, but many are limited in the characters they can easily score.

One of the most widely used and freely available software packages for digital image analysis is ImageJ, developed by the National Institutes of Health (Rasband,

1997). This application has been applied for digital image analysis in many aspects of biological research (for review, see Schneider et al., 2012). While ImageJ is a versatile program that allows the user to make minor adjustments to examined images (i.e. rotate, transform color to grayscale) and develop macros for the specific needs of individual projects, these require additional user input. Though ImageJ is not designed specifically for morphological analysis, it is capable of calculating measures of length, width, area, etc., with limited user demand.

Many programs have been developed for shape analysis of organismal structures.

LeafAnalyser, developed by Weight et al. (2008), scores traditional morphometric points

(evenly spaced points, set on object perimeters) in detected shapes (up to 256 evenly spaced points) and allows for processing of numerous files under one set of predetermined parameters. Programs such as MORPHOJ (Klingenberg, 2011) that

18 calculate geometric morphometrics (relative position of biologically significant landmarks) are widely used in evolutionary and developmental biology (Parsons et al.,

2011; Brusatte et al., 2012; Lewton, 2012), including plant sciences (Viscosi & Cardini,

2011; Silva et al., 2012; Viscosi et al., 2012) to compare taxa. Some pipelines have been created specifically to examine flower development via geometric morphometrics (AAM

Toolbox for MATLAB, Langlade et al., 2005). Alternative methods of shape analysis of objects in digital images have been addressed by programs such as SHAPE (Iwata &

Ukai, 2002), which calculates elliptic Fourier descriptors (EFDs; Kuhl & Giardina,

1982).

Other publically available software packages have since been developed specifically for the plant sciences, for morphometric analysis of hypocotyl emergence and expansion (HYPOTrace, Wang et al., 2009), root structure (EZ-Rhizo, Armengaud et al.,

2009), leaf (LAMINA, Bylesjo et al., 2008), and fruit shape and size (Tomato Analyzer,

Brewer at al., 2006). Though all include convenient graphical user interfaces (GUIs), many have specific, limited applications or require supplementary input for image scaling or manual image modification in order to capture desired data. These adjustments may become very laborious for projects examining large volumes of images, and dramatically increasing the time expended per morphological data point generated.

Tomato Analyzer (Brewer et al., 2006; Rodriguez et al., 2011) is a free-to- download program, developed to score morphological characters from scanner-generated images of tomato fruit. This program calculates a suite of measures from digital objects typically generated by other applications, such as perimeter, area, height and width.

19

Beyond these standard measures, TA creates indices of objects’ correspondence to generic shapes including squares, circles, or ellipses, measures differences in shape of user-designated proximal and distal ends, and positions up to 200 traditional morphometric points per object. Unlike many other programs, TA requires virtually no scaling data, beyond setting image dpi for determination of size and shape characters.

TA can also be easily calibrated for color analysis to recognize and correct for scanner interpretation of color (Darrigues et al., 2008) in order to output RGB and CIELab color profiles (Commission Internationale de l’Eclairage; CIE, 1978).

This program also demonstrated its great utility in high throughput capacity with its “batch mode” where over 100 images can be processed at one time. As the current study analyzed over 4,000 images and scored nearly15,000 objects, this option allowed for great efficiency in data collection, and output as data formatted as comma-delimited,

Excel-ready files. The ease of TA use is evident in its broad application in plant morphology, including fruits (Depypere et al., 2007; Gonzalo et al., 2009; Mazhar et al.,

2011), seeds (Marcos Filho et al., 2010), and flowers (present study). Our novel application of Tomato Analyzer software to describe floral traits underscores the broad applicability of this free, publicly-available, easy-to-use software in analysis of plant morphology. TA software could become a fantastic resource for plant breeders, to quantify visually-unscorable gains in development of elite hybrid lines. Further, germplasm banks of ornamental and agronomic crops could similarly apply this program for low-cost, rapid data collection, to create detailed, reference profiles of maintained accessions.

20

Though the most recently released version TA (v. 3.0, Rodriguez et al., 2011) was available, it was not used for this study. This updated version could produce scores of up to 200 morphometric points per object (previous version generated only 30), but consistently crashed the program when used in conjunction with “batch mode” function for image analysis. Further, the “pericarp boundary” function to score blotch traits similarly crashed the program.

Character calling (QL vs. QT)

Though the relative results of both character scoring methods were very similar, the artificial superimposition of discrete states in this large germplasm collection appears to be a less meaningful approach. In PCoA analysis, the QT and QL methods yielded comparable scatterplots in terms of relative point position, though the absolute distance between points in QL analysis were almost twice as distant as the QT points. Likewise, the fraction of the variation captured in the first three axes, which was almost twice as high in the QT method (52%) as in the QL approach (33%). The low levels of captured variation may be a function of artificial character states superimposed onto the data.

Increased levels of complexity and divergence were suggested through UPGMA analysis via QL data, while it maintained a comparable, moderate, level of matrix correspondence to the QT approach (QL=0.7621, QT=0.7306). Significant divergence was seen via

Ward’s method between the two datasets (QT=0.9264, QL 0.6379). Through examination of these two clustering techniques in synthetic and real data, Ward’s method has been better at recovering the original subgroups in the data in situations where

UPGMA analysis yields a low CPCC (Odong et al., 2011). Together, the results of these

21 phenetic analyses suggest that characters scored in the present study were better scored as continuous, and that accessions should not fall into apparent classes. Therefore, we contend that the QT results are more biologically significant, and further discussion of morphological differentiation in hybrid Viola will pertain to these results only.

QT analysis

Comparison of clustering and ordination analysis suggested highly similar grouping patterns. In all phenetic analyses, two strongly supported, primary groups of accessions were created based on presence of blotches, with secondary trends within these partitions of clustering by color and shape/size. In the cluster analyses, there was little bootstrap support relating clusters of accessions with different-colored flowers.

Within the blotched clade, Ward’s approach grouped accessions more on shape and size relative to color, which further led to fewer, larger clusters of VC accessions (Figure 1.4-

A). In the UPGMA analysis, grouping by color was more apparent, with substructure by size and shape within like-colored blocks (Figure 1.4-B). In the non-blotched clade, species types were distinctly separated, especially following Ward’s method, which gave great weight to the shape, including more narrow petals and distinctly smaller lateral petals than hybrid accessions.

There was higher bootstrap support and greater consistency for clustering of white, yellow, and orange flowered accessions across analyses, while purple and blue flowers tended to intermix. This is most likely due to a number of factors, including a greater diversity of size and shape in the purple group, and also more uniform flower color in the white- and yellow-based cultivars. Further, there was a higher prevalence of

22 bicolorous flowers (superior petals different color than lateral and inferior) in the purple group. Alternatively, the patterns of stronger clustering within like-colored violet accessions may be due simply to limited sampling of their diversity, or possibly lack of introduction of diverse cultivars in all color classes.

There were some outliers in these more strongly clustered groups, in which general patterns of blotch and color were displaced by stronger distinction of characters related to relative size and shape. Some accessions with blotches appeared in the non- blotched group, and vice versa. Of the white, blotched accessions, one cultivar

(accession 124) did not cluster strongly with the others in its class. This was probably due to its having the narrowest petals of the blotched-white VW cultivars (low P:A ratio) and also due to its larger blotch size relative to petal size. Further, this large blotch also appeared on all three petal types, whereas the majority of blotched accessions only demonstrated blotch markings on two. Within the orange class, all accessions were VC types, but one (accession 306) that was distinctly smaller flowered, though had relatively wider petals than the other orange cultivars. There was support for separation of one purple, blotched accession (252) in the UPGMA analysis, though in review of the raw data no single trait obviously separated it from any others.

Many of these trends may arise from inherent biases in the collection and characters examined. For example, the trend for separating cultivated varieties by trade classification (VW vs. VC) may have been influenced by many factors: the higher number of pansies (n=70) versus violas (n=40), or the higher prevalence of blotches in the pansy (63.3%) than viola (29.2%) accessions. For example, seven out of eight

23 accessions with white, blotched flowers were VW cultivars. While small, unblotched pansies and large, blotched violas exist in the horticultural realm, they were not well represented in the collection examined here. Further, the majority of blotched pansy cultivars examined had petals often wider than long, which is not a trait that can be generalized across all pansy cultivars. The dominant pattern of clustering by blotch may also be due to the effect of blotch traits in the analysis. Of the 78 characters scored, 17 referred to a blotch-related character (color, size, etc.). Though none of these were strongly correlated, they provided some bias in clustering due to absences of scores for all non-blotched individuals. Principal components analysis of QT data indicated that blotch characters were the primary source of variation in the first component, though they contributed little to the second or third (analysis not shown).

It is important to evaluate the relevance of analyses attempting to differentiate between cultivars falling into the horticultural groups VW and VC. These two categories do have historical, horticultural significance, but in terms of hybridization they are not distinct and are sexually compatible, even with wild types (Kroon, 1972). It is common for breeders of these commercial products to interbreed the two categories for advancement and enhancement of desired characters across ploidy levels (Peter Stefany, personal communication, March 2, 2008), a practice that may perpetuate the continuous nature of phenotypic (and presumedly genotypic) characters in these cultivated hybrids.

Previous morphological studies in cultivated Viola hybrids have found results similar to those of the present study in terms of clustering by horticultural classifier (VW vs. VC) and presence of blotch. Analysis of 32 morphological characters in 18 inbred

24 pansy lines by Wang and Bao (2005) suggested clustering by both blotch and trade group

(ploidy) in UPGMA analysis. Trends of like-colored flowers to group together were also evident in some color classes, including accessions with and without blotches, though floral characters represented a small percentage relative to vegetative traits examined

(~28%). Disproportionate sampling by horticultural class was apparent in this study, as there was a significantly greater representation of VW than VC types (8:1), as well as blotched to unblotched cultivars (13:5). Further, neither of the VC accessions examined had blotches, which would further segregate the two horticultural groups based on the results in the current study.

The ability to determine geographic origin of cultivated Viola by their morphology has also been suggested by previous investigations. Du et al. (2011) scored

13 morphological characters in 28 violet accessions, acquired from five geographically distinct regions, and reported significant correlation between UPGMA clustering patterns of accessions and origin. Also, the lowest levels of genetic distance were detected between the clusters of Chinese cultivars, relative to those cultivars originating from the

Netherlands. Authors suggested this was due to a higher degree of diversity from a variety bred in Europe, where pansy hybrids were originally developed. While it was suggested distance measures corresponded to region-specific diversity, examination of the UPGMA dendrogram proposed some clustering by color as well, involving red, orange, and yellow flowered cultivars, though no apparent grouping by blotch presence.

Another study found similar, origin-based clustering patterns (Du, Liu, Zhu, & Zhang,

2011), but again the sampling strategy in this investigation may have affected the results.

25

Only five of the 28 accessions were not Chinese (3 Netherlands, 2 USA), and within these western varieties, color may have biased clustering of those from these marginally represented groups as both of the accessions from the USA were orange, and those from the Netherlands were yellow and/or purple. Many companies breeding Viola develop

“series” of both VW and VC cultivars, in which varieties’ morphological characters are very uniform except for flower color, and have been developed to fit specific niches based on flower size or season planted. As these are cultivated, hybrid taxa, not wild, naturally distributed species, the relationship of morphology to region of origin may be associated with horticultural selection for adaptability to regional, environmental conditions or local consumer preferences.

Morphometric analysis

The morphometric analysis of flower shape suggested differentiation between horticultural groups. Although accessions do not appear to fall into distinct clusters, PCA placed the species-types at one end of the scatterplot, and the VW cultivars at the other.

The distribution of accessions along the primary axis was related to multiple points associated with the shape of the lower three petals. The trend for higher levels of variation to be associated with the shape of the lateral and inferior petals reflects the efforts of plant breeders to enhance flower size and area. It has been well documented that since the earliest years of breeding Melanium violets, hybridizers have worked to select forms with wider petals (see introduction). Placement of accessions also suggested a pattern of increasing ratios of individual petal H:W (“fruit shape index 1”), P:A ratio,

26 and concurrently decreasing lateral : inferior petal area ratio, as well as circular and ellipsoid shape scores. The VW (polyploid) accessions exemplified this pattern in flower traits as they have much larger petals, relative to VC (diploid) accessions. On average,

VW petal width was distinctly larger than length (23.5% across petals), whereas VC cultivars averaged equal width and length, and species accessions were 20% longer than wide. The pattern of petal size ratios suggests that the horticultural class, a proxy for ploidy, of a cultivar can be estimated by these relative measures. This pattern may hold true for all cultivars, and may simply be a function of the recent market interest. Further, given the observed diversity and apparent plasticity in the Melanium group, future changes in consumer interest could easily redirect the efforts of plant breeders, and thus the distribution of relative size characters within this hybrid pool.

Characterization of flower shape in pansies and violas via traditional morphometrics, as in the analysis presented here, or with geometric (landmark) morphometric analysis, has not been described in previous studies. Analysis of relative petal size and shape in hybrid violets has been carried out via elliptic Fourier descriptors

(Yoshioka et al., 2006). To describe specific combining ability of 25 inbred lines of pansies and F1 progeny, digital images were generated by flat-bed scanners, and utilized

SHAPE software (Iwata & Ukai, 2002) to determine associations of relative size and petal shape via elliptic Fourier descriptors. Here, analysis by principal components suggested a wide range of petal shapes, and ANOVA indicated the significant differences in variation associated with different lineages of petal types.

27

The current study provides the most detailed digital analysis of pansy floral morphology to date. It incorporates the greatest number of varieties studied, as well as number of traits examined. The ability of TA to discern between fine levels of color and shape variation in the continuum of Viola flower morphology supports previous findings that the quantitative analysis of digital images are of great value for the morphological assessment of cultivated plants (Yoshioka et al., 2004a). Image analysis with TA could have been easily applied to previous plant science studies using flatbed scanners directed at germplasm collection characterization (Bacchetta et al., 2010), inheritance of flower shape (Yoshioka et al., 2006) disease resistance (Falloon, 1995; Olmstead et al., 2001), insect herbivory (Kerguelen & Hoddle, 1999; O’neal, Landis, & Issacs, 2002), and cultivar identification (Yoshioka et al., 2004b; Venora et al., 2007).

The novel use of digital image analysis with Tomato Analyzer demonstrated the ability to describe morphological diversity present in the flowers of modern, Viola cultivars and a selection of wild-type taxa. As it is likely that advanced hybrids examined in this study were based on the genetic contributions of only two, or potentially three, species (Wittrock, 1896), our findings of diversity in color, shape, and size underscore the extreme plasticity and variation that exists in the Melanium violets. Further, it appeared that hybrid-derived diversity was magnified when paired with polyploidy, as seen by divergence between horticultural classes (VW vs. VC), highlighting the efforts of plant breeders and the general effects inherent in the domestication and directed breeding processes. Future examinations in the cultivated pansy group should be conducted with awareness of the potential divide between horticultural classes, and cognizance of the

28 potential effects of related sampling biases when interpreting their results. Also, examination of other measures of shape, such as EFDs or geometric morphometric analyses, may prove to be of value for large collections.

29

Tables and Figures

Figure 1.1 Anatomy of Melanium violet flowers

Flowers of two Melanium violets: a species-type accession, V. tricolor (accession 240), at left, and a cultivated hybrid, V. × wittrockiana (accession 64), at right. Representative scale bars are placed to the lower left of their associated flower. Descriptive accession information can be found in Appendix A.

30

A)

B)

Figure 1.2. Examples of morphological analysis in Tomato Analyzer software Examples of flowers from a cultivated hybrid V. × wittrockiana (accession 62) seen in the split-window of Tomato Analyzer software depicting (A) flower orientation and placement of counter-clockwise counted morphometric points (n=30) including the 12 o’clock (0º/360 º) and 6 o’clock (180 º) positions (teal points), and (B) the “pericarp boundary” function (yellow line), adjusted to circumscribe the blotch (for additional details, see Brewer et al., 2006, and Gonzalo et al., 2009). 31

32

Figure 1.3. Principal coordinates scatterplot of QT Viola data

Principal coordinates analysis scatterplot of first three coordinates (PCo) depicting the relationships between quantitative (QT; continuous) scoring of morphological characters from flowers of 127 Viola accessions. Individual accessions are represented by colored points which represent their horticultural class, V. × wittrockiana, (VW-red), V. cornuta (VC-green), and species-type (S-yellow). A filled square indicates the presence of blotch in that accession. Circles around groups of squares indicate examples of accessions with similar color, and are individually labeled. Arrows indicate gradients in morphological traits, as labeled. Descriptive accession information can be found in Appendix A. 32

A)

Figure 1.4. Cluster analysis dendrograms of QT Melanium violet data Dendrograms depicting the relationships between quantitative (QT; continuous) scoring of morphological characters from flowers of 127 Viola accessions under (A) UPGMA (Gower) and (B) Ward constraints. Individual accessions are represented by colored points, indicating their horticultural class: V. × wittrockiana, (VW-red), V. cornuta (VC- green), and species-type (yellow). Bootstrap values are based on 1,000 resamplings of the dataset. Symbols above branches indicate bootstrap support: asterisks represent >75% support, and ampersands represent >90% support. Brackets indicate groups with similar traits, as labeled. Descriptive accession information can be found in Appendix A. continued 33

Figure 1.4 continued

B)

34

3

5

Figure 1.5. Principal components analysis scatterplot of 30 morphometric points

Principal components analysis scatterplot of the first three components depicting the relationships between 30 morphometric points (mean values) from flowers of 127 Viola accessions. Individual accessions are represented by colored points, indicating their horticultural class: V. × wittrockiana, (VW-red), V. cornuta (VC-green), and species-type (yellow). A filled triangle indicates the presence of blotch in that accession. Descriptive accession information can be found in Appendix A.

35

Chapter 2. Molecular variation in a collection of pansies (Viola sect. Melanium) characterized by sequence-related amplified polymorphism (SRAP) markers

Abstract

Domestication of ornamental crops has been marked by increases in morphological diversity and reductions in genetic diversity. Market demands and breeding trends have transitioned away from seed-strains and landraces to F1 breeding techniques, limiting incorporation of genetically diverse, wild-type material in many ornamental taxa. These practices have led to increased likelihood of genetic homogenization in many valuable ornamental crops, including the hybrids from the

“pansy group” (sect. Melanium) of Viola. To assess the molecular variation in the modern hybrid pool, a germplasm collection of 124 Viola accessions including putative, progenitor taxa, and the two horticultural classes (V. × wittrockiana, VW; V. cornuta,

VC) were characterized with sequence-related amplified polymorphism (SRAP) markers.

Eight SRAP primer pairs were selected, which amplified 470 distinct markers. Accession profiles were compared using ordination (PCoA), clustering methods (Ward, UPGMA), and Bayesian (STRUCTURE) analysis. AMOVA was used to determine the variation between and within determined STRUCTURE clusters. A Mantel test was performed to compare the molecular similarity matrix to the corresponding morphometric matrix of quantitative data generated in Chapter 1.

36

Phenetic analysis of fragment patterns did not distinguish between horticultural classes. Species types generated significantly more amplicons over all loci than cultivated accessions of the same ploidy (VC). There was grouping by blotch presence and petal color within cultivated types, especially accessions with white, orange, white and yellow with blotch flowers. Relationships between these blotch/color themed subgroups were unclear due to low bootstrap support and conflicting topologies between analyses. STRUCTURE analysis indicated substructure of both K=6 and K=8 groups, producing large clusters dominated by the presence or absence of blotch and, secondarily, color, while creating small clusters of species-types including separate groups of V. lutea and V. tricolor. AMOVA indicated that 90% of the variation was within clusters, and

10% among clusters for the K=6 projection, while for K=8, 93% of the variation was within clusters and 7% among. One locus (Me1-Em1) detected >39% variation between clusters in both scenarios, distinctly more than other loci. A Mantel test indicated significant correlation between the two similarity matrices (p<0.01), and phenetic analyses of a combined dataset (n=117) maintained patterns of grouping by color and blotch, though topology of cluster analyses had distinctly lower levels of bootstrap support.

These results suggest that for advanced, hybrid violas, morphology alone (e.g., color, blotch) may allow for distinction between genetic lineages, which may be of value for maintenance of germplasm collections involving cultivated varieties. The association of molecular data to morphological markers and the high level of inter-cluster variation detected at the Me1-Em1 locus strongly suggested the potential for development of

37 character-related, SRAP-derived markers, valuable for future horticultural improvement.

Further, the distinctive molecular phenotypes of species-type accessions indicate that unutilized genetic resources are available to plant breeders via wide outcrosses.

Introduction

Many cultivated ornamental crops have undergone changes in genetic diversity due to domestication (Powell, 1997), and are often derived from limited collections of wild taxa (Levin, 1977; Tomkins et al., 2001). Horticulturalists have exploited these limited genetic resources and developed hybrids that demonstrate enhanced morphologic variation with exaggerated, diverse, and previously-undescribed traits (Darwin, 1868;

Chittenden, 1928; Petit & Peat, 2004; Chao et al., 2005). In many cases, these pursuits to develop novelty have been attained through directed, interspecific hybridization (Janick,

1998; Tay, 2006). Yet over time, genetic diversity has been reduced due to use of a limited pool of elite germplasm and repeated selection for desirable, ornamental traits in plant breeding programs (Ogisu, 1999; Linde, Yan, & Debner, 2007). Many ornamental taxa demonstrate this genetic homogenization process, such as Impatiens wallerana,

Petunia × hybrida, Begonia × tuberhybrida, (Tay, 2006) and Viola × wittrockiana (Peter

Stefany, personal communication, March 2, 2008), where complex interspecific hybrids have been exploited for desirable horticultural traits instead of maintenance for genetic diversity.

Traditional selection has long been the standard approach for developing new varieties of ornamental plants. Historically, individual breeders have been independent

38 farmers, gardeners, and enthusiasts that have been responsible for the development of new horticultural varieties, for both ornamental and agronomic crops. While responsible for the releases of countless cultivars, in many instances their work was seldom recorded and/or distributed for future development. Because ornamental plants represent a small portion of the agricultural market relative to crops for food or fodder, there has been limited effort to collect and conserve their historical or genetic resources (Heywood,

2003a, 2003b). As a result, there is a dearth of peer-reviewed literature involving the origins and genetic history for many contemporary ornamental crops. Though recent efforts have been made to publish the methods of ornamental crop development

(MacDonald & Kwong, 2005; Anderson, 2006), and conserve germplasm (Tay, 2003), the proprietary efforts of the private sector have maintained a shroud over the lineage of many commercial cultivated varieties.

The globalization of the agricultural industry has played a significant role in the depletion of genetic diversity in numerous ornamental plants. The demand of consumers and competitive nature of the marketplace apply pressure to continually develop new cultivars, creating a marketplace where new herbaceous varieties maintain a very short lifetime before the next generation is offered. Competing companies release similar cultivars, often derived from related parental lines, leading to expectations of appearance and performance. For seed-generated ornamentals, the rise of F1 hybrids worldwide has replaced the propagation of many traditional, seed-strain, heirloom cultivars and lesser- grown species. This may lead to their disappearance from cultivation, and further decrease overall genetic diversity within associated groups. Additionally, the race for

39 innovation and improvement in ornamental plant introductions, paired with ex situ micropropagation techniques, has led to the introduction of many herbaceous ornamentals with significant flaws in terms of disease resistance and environmental adaptability (Bennett, 2006; Frett, 2009). While development of some ornamental crops incorporate wide backcrosses to species to capture valuable traits associated with environmental and pathogen tolerance (Seiler, 1992; Van Tuyl & Van Holsteijn, 1994;

Onozaki et al., 1996), this a secondary thought for many breeders who are narrowly focused on developing novelty in traits such as flower color or foliar pattern. As a result of a narrowing genetic base, the potential for development of horticultural innovation and improvement is reduced, while the potential for widespread decline is simultaneously increased (Raver, 2013).

Efforts to describe the discrete variation between advanced hybrid ornamental plants have been attempted in many taxa, due to the substantial numbers of induced hybrids that have been developed for the horticultural trade (Petit & Peat, 2004; Zilis,

2009). For example, within the Orchidaceae, there are over 20,000 species and 100,000 registered hybrid cultivars (American Orchid Society website, accessed 12 September

2013), and in Hemerocallis around 20 species exist (Stout, 1934) and over 75,000 named cultivars (American Hemerocallis Society website, accessed 23 September 2013). Many of these texts have been guides for horticultural audiences, with cultivars described in morphological terms. Molecular analyses of large collections of ornamental plants have been carried out in some taxa (Hu et al., 2005; Bruna et al., 2006; Palumbo et al., 2007;

Zhao et al., 2007; Han et al., 2008; Ma, Olsen, & Pooler, 2009), but due to many factors

40 including time, cost, facility or interest, many of the most valuable ornamental crops have not been extensively studied.

In recent decades, molecular tools have made it possible to make systematic assessments of diversity in germplasm collections, for breeding or conservation purposes

(Bretting & Widrlechner, 1995; de Vicente et al., 2006; Barcaccia, 2009). Such analysis of large ornamental plant collections provides the opportunity to examine inter- and intra- specific introgression, diversity, and population structure (Tay, 2005). Simultaneously, investigations into the genetic underpinnings of ornamental plants can provide valuable feedback for use in both commercial (Speijer et al., 1997; Hill et al., 2004) and academic arenas (Jan et al., 1999; Pankhurst et al., 2001). The value of these genetic analyses lies in their ability to provide important information on how to efficiently manage and maintain germsplasm collections (Brown & Briggs, 1991; Bataillon et al., 1996; Gilbert et al., 1999; Palumbo et al., 2007) and develop markers for valuable, heritable traits within breeding programs (Ballard et al., 1995; Batlle & Alston, 1996; Bi et al., 1999;

Von Malek et al., 2000).

Dominant molecular markers, such as random amplified polymorphic DNA

(RAPD; Williams et al., 1990), inter-simple sequence repeat (ISSR; Zietkiewicz et al.,

1994), and amplified fragment length polymorphism (AFLP; Vos et al., 1995), have been important for assessing diversity and identifying the lineages for many ornamental crops

(Debener & Mattiesch, 1999; Lesur et al., 2000; Kochieva et al., 2004; Hou et al., 2006).

This is due to the ability of non-specific primers to amplify multiple variable markers without the need for a priori primer sequence information, thus elucidating previously

41 ambiguous relationships between many cultivated selections and hybrids (Friesen et al.,

1997; Jiajue, 1998; Besnard et al., 2002). Rose (Rosa, Zhang et al., 2000), peony

(Paeonia, Han et al., 2008), daylily (Hemerocallis, Tomkins et al., 2001), geranium

(Pelargonium, Palumbo et al., 2007), chrysanthemum (Dendranthema grandiflora, Scott et al., 1996), impatiens (Impatiens hawkeri, Carr et al., 2003), and pansy (Viola, Wang,

2005) are examples of economically valuable ornamental plants evaluated via dominant markers.

Hybrid violets, Viola × wittrockiana Gams. (VW) and V. cornuta L. (VC) derivatives, known as garden pansies and violas, respectively, are together some of the most popular herbaceous bedding plants in the world market (see Chapter 1 for historical details). These floriferous, interspecific hybrids from the section Melanium have created over $100 million in annual wholesale value for more than a decade (United States

Department of Agriculture, National Agriculture Statistics Service website, accessed

October 28, 2011), and represent over 7% of annual bedding plants sold in the US, including vegetables (Jerardo, 2007). Even higher estimates of sales have been reported for Europe (Peter Stefany, personal communication, March 2, 2008). The consumer market in the United States has shown that pansies have been one of the top-selling annual bedding plants since the middle of the twentieth century (Genders, 1958). Many private companies have worked to improve the garden value of violets (including flower size, color clarity/diversity, and plant adaptability), releasing hundreds of F1 varieties over the last half century (Peter Stefany, personal communication, March 2, 2008).

42

These Melanium (“pansy group”) violet hybrids have ambiguous hybrid origin, and are presumedly derived from many European species including V. tricolor, V. cornuta, and V. altaica (see Chapter 1 for details), but limited molecular characterization of the diverse, cultivated types or parental taxa has been published to date. Yockteng et al. (2003) assessed ITS sequence divergence and (RFLP) markers in Melanium taxa, but the variation detected was not sufficient to substantively distinguish their relative phylogenetic positions. More recently, the relationships between cultivated, commercial

Melanium hybrids (including both violas and pansies) have been assessed using dominant markers including RAPD (Wang, 2005; Wang & Bao, 2005; 2007) and SRAP (Wang et al., 2012). These investigations examined relatively small numbers of cultivars (17-43 accessions), and similarly encountered low post-hoc support for their findings.

Contrasting conclusions have been drawn as to which factors play the most important roles in grouping reviewed accessions. Examination of larger and more diverse collections of the two cultivated types and their ancestral taxa using highly variable markers may enhance the understanding of diversity and relatedness between and within these classes. Furthermore, marker information could enhance efficiency in managing large germplasm collections, and the discovery of character-associated markers could better direct commercial breeding endeavors.

In this study, molecular variation in a large germplasm collection of Viola from the section Melanium (n=124), including both cultivated hybrids (VW and VC), and presumed ancestral wild type taxa (S), was characterized using hierarchical and Bayesian approaches. These results were compared to those attained through morphological

43 analysis in Chapter 1, and were combined to explore the relationships between morphologic and molecular characters. The present study also utilized capillary electrophoresis to score fluorescently-labeled SRAP fragments (a practice almost rarely used with this marker) and the results of this approach relative to previous SRAP marker analysis were addressed.

Methods Accessions and plant growth

Seed accessions of 124 Viola hybrids and species types were acquired from commercial breeding programs and horticultural sources (Appendix A). These represent the subset of the accessions examined in Chapter 1 that yielded amplifiable DNA. These also include some accessions that did not produce flowers of quality or number for morphological analysis. As requested by donating companies, the origins of these commercial cultivars are not identified by name or parent company. Accessions were grown as previously described (see Chapter 1: Methods).

Amplification and genotyping

Eight forward (Me1-8) and eleven reverse (Em1-11) SRAP primer pairs (Table

2.1; see Budak et al. 2004a for details) were screened with all possible primer combinations in a randomly-selected subset (n=31) of Viola accessions to ascertain variation and amplification. This subset of 31 samples was independently run in duplicate to assess band pattern reproducibility of the selected markers. Of these, eight primer combinations were selected based on repeatability and degree of variation (Table

44

2.2), and were similarly applied to all 124 accessions following this primer pair selection and protocol optimization.

The SRAP marker system consists of primers targeting open reading frames

(ORFs) via PCR (Li & Quiros, 2001). Based on their high relative G-C content, forward primers preferentially target regions of exon (G-C rich), whereas reverse primers are designed to target introns and promoters (G-C poor). The template for PCR amplification consisted of raw DNA extract diluted 1:9 in TAE buffer, yielding ~20-50 ng genomic

DNA. PCR amplifications were performed in a reaction volume of 10 µL, comprised of

0.05 µl (0.25 U) of ExTaq (Takara), 1 µl of 10X ExTaq PCR Buffer, 0.8 µl of 2.5 µM

ExTaq dNTP Mix (2 µM), 0.25 µl of FAM-labeled forward strand primer (0.5 µM), and

0.25 L of reverse strand primer (0.5 µM) in 7.15 ul UV-irradiated, HPLC pure water

(Fisher Scientific, Pittsburgh, PA).

Though multiple PCR protocols exist for SRAP marker systems (for review see

Aneja et al., 2012), the primary, two-phase method of Li and Quiros (2001) yielded the best products in Viola. Samples were subjected to the following amplification program in an Eppendorf Mastercycler Pro S thermocycler (Eppendorf North America, New York,

NY): 5 min of denaturing at 94º C, five cycles of three steps: 1 min of denaturing at 94º

C, 1 min of annealing at 37º C and 1 min of elongation at 72º C. This was followed by

35 cycles with an annealing temperature increased to 53º C, with a final elongation step of 5 min at 72º C. The presence of PCR product was confirmed by running 4 L of reaction product on a 1% agarose gel and imaged as above.

45

Primer combinations produced fragments varying in size from less than 50 bp to over 3000 bp. Of the eight primer pairs utilized, seven produced bands primarily below

1000 bp, which were separated via capillary electrophoresis (CE) on an ABI Prism 3100

Genetic Analyzer (Life Technologies, Grand Island, New York) with a custom program utilizing ABI POP-6 polymer (Life Technologies, Grand Island, New York). Samples were made by mixing 0.5-1 L of PCR product with 9.5 L of HiDi Formamide (Life

Technologies, Grand Island, New York) and 0.2 L of a lab-made ROX799 size standard adapted from DeWoody et al. (2004) (Figure 1.2). Samples were pipetted into 96-well plates and denatured for five minutes at 95º C, then immediately placed on ice for five minutes, and stored in the dark at 4º F until loaded into the genetic analyzer. Resulting

SRAP fragments were scored from 100-800 bp with GeneScan analysis software (version

3.7, Applied Biosystems) as present (1) or absent (0). Fragments scored as present had electropherogram peaks with minimum amplitudes of 200 relative fluorescence units

(rfu). Following this screening and optimization, all 124 accessions were run as described above.

Samples that were unscorable due to low peak quality of some PCR products were subjected to a 20% polyethylene glycol (PEG) precipitation procedure to eliminate signal interference by small fragments generated by primer dimerization. This was accomplished by adding equal volume PEG solution (20 g PEG 8000, 5 ml 5M NaCl, with HPLC pure, UV-irradiated water to a final volume of 100 ml) to PCR reactions and storing samples on ice for one hour, then centrifuging at 15,000 rpm for 15 min, removing the supernatant and resuspending the pellets in 100% EtOH. Samples were

46 subjected to two additional washes, but using 80% then 75% EtOH, with only seven minute spins at 15,000 rpm. The resulting pellets were dried in an unheated vacuum centrifuge for 15 minutes then resuspended in UV-irradiated HPLC pure water at 75% of the initial PCR sample volume and subjected to CE as described above.

One forward primer (Me8) consistently yielded fragments in the 1000-3000 bp range when paired with most of the reverse primers tested, making analysis via CE impossible due to fragment size scoring capabilities of CE equipment. The most consistent and variable pair was selected for analysis (Me8-Em8) and amplified in all accessions. Four microliters of PCR product were run on 2% agarose gels with two reference samples of 1 kb Plus DNA ladder (Life Technologies, Grand Island, NY) at

40V for approximately 2.5 h, followed by staining and imaging as above (Figure 2.2).

Kodak 1-D software (Eastman Kodak, Rochester, NY) was used to score the resultant fragments, scoring bands from 1000-3000 bp.

Descriptive statistics were calculated, including the rate of polymorphism per locus, the number of fragments per locus, the number of fragments per individual per locus, the range of fragment sizes per locus. The degree of repeatability of the 31 screening samples was calculated in two ways: the proportion of similarity between the two runs of each of the tested (31 samples per run) per primer, and between the total number of fragments over all individuals (31 samples, 2 replicates, 8 primer pairs each with unique amplicon yields).

47

Molecular analysis

Fragments resulting from SRAP analysis were subjected to principal coordinates

(PCoA) and UPGMA analyses in PAST software (Hammer, Harper, & Ryan, 2001).

Dice’s Similarity Coefficient (Dice, 1945) was selected for these analyses as it includes only character matches between samples, and is identical to the Nei and Li distance coefficient (1979), typically used for genetic similarity. Ward’s (1963) minimum variance method was also utilized to determine clustering, using pair-wise

Euclidean distance. This second clustering approach was employed due to its distinct algorithm, relative to UPGMA, and reported concordance with results of Bayesian methods (Odong, 2011). Both clustering analyses were carried out with1,000 bootstrap replicates, and cophenetic correlation coefficient (CPCC; Sokal & Rohlf, 1962) scores were calculated to assess the relationship between the pair-wise differences and resultant clustering projections. Student’s t-test (two-tailed) was used to determine if there was a difference in number of bands between (presumed) ploidy classes of accessions (VW and

VC+S), as increases in SRAP fragment number with ploidy has been documented (Budak et al., 2005; Gulsen et al., 2009). Cluster analysis of a concatenated molecular (SRAP) and morphological dataset (see Chapter 1 for details) was also done using Gower’s similarity coefficient, which has been shown appropriate for mixed qualitative- quantitative data (Dunn & Everitt, 1982), and resampled as described above.

The Bayesian model-based clustering program STRUCTURE v. 2.3.4 (Pritchard,

Stephens, & Donnelly, 2000) was used to estimate the most probable number of clusters

(K), given the SRAP data. The proportion of an individual’s genome (Q) that originated

48 from each cluster was recorded. The STRUCTURE program was run with no prior knowledge, under the admixture ancestry model, with ploidy set to diploid (2). Though it has been documented that VW and VC cultivars in the trade have different ploidy levels

(Kroon, 1972; Peter Stefany, personal communication, March 2, 2008), without specific quantification it was unclear how best to classify cultivars into one group or another.

Also, hybridization between the two groups has been done regularly for advancement of specific horticultural goals (Peter Stefany, personal communication, March 2, 2008), and breeding between groups of different ploidy results in swarms of fertile intermediates

(Dalbato, Kobza, & Karlsson, 2013). As cytological examination was beyond the purposes of the current study, a conservative approach was applied and all accessions were presumed diploid. In all STRUCTURE analyses, the number of populations was set to one because the two horticultural classes are freely interbred and could not be easily grouped by morphological characters. As a preliminary screen to estimate K, the Markov chain Monte Carlo (MCMC) parameters were set to a burn-in period of 105 followed by

105 iterations, testing K values 1 to 35 with three replicate runs for each value.

Probability of K values was calculated using the method of Evanno, Regnaut, and Goudet

(2005) in the web-based program STRUCTURE Harvester (Earl & von Holdt, 2011).

Superficial review of probabilities indicated the highest probability of fewer than 10 clusters. STRUCTURE analysis was then rerun with a burn-in period of 106 with 5x106 iterations, testing K values 1 to 10 with three replicate runs for each value. When the optimal cluster value was determined, STRUCTURE was run a final time with 106 and

5x106 iterations, 10 repetitions with that value. Similarity in STRUCTURE runs was

49 calculated by the method described by Jakobsson and Rosenberg (2007) via their computer program CLUMPP 1.1.2, using the fullsearch option. This method calculates a similarity coefficient, H, providing for the assessment of the similarity of individual runs, and also calculates a mean probability per individual of the input matrices. This average clustering of accessions determined was then visualized using the program DISTRUCT

1.1 (Rosenberg, 2004).

Analyses of molecular variance (AMOVA) were executed in GenAlEx 6.4

(Peakall & Smouse, 2006), to assess the partitioning of variation, variation per locus, and estimate PhiST: a value analogous to Wright’s (1969) fixation index FST . Samples were analyzed with the STRUCTURE-generated K value used as a proxy for populations in order to assess the contributions of clusters to overall variation.

To assess the relative similarity of the molecular information generated by SRAP, and the morphometric data examined in Chapter 1, a Mantel test was used to compare the correlation of the matrices generated by Dice similarity (SRAP markers) to the Gower similarity estimates (transformed, quantitative (QT) data - see Chapter 1 for details). If these matrices demonstrated significant correlation, they would be combined to create a supermatrix, and be reviewed by ordination and cluster analysis as described above:

PCoA (Gower), UPGMA (Gower) and Ward’s (Euclidean) approaches.

Results SRAP diversity

Analysis of 124 Viola accessions with eight SRAP primer pairs yielded a total of

470 amplicons (Table 2.2) ranging from 101bp to 813bp for all primer combinations 50 except Me-8-Em8, for which fragments 1117bp to 2937bp were scored. Of these 470 fragments, 432 (92%) were polymorphic. The number of fragments produced by any one primer combination ranged from 50 (for combinations Me1-Em1) to 72 (Me3-Em5) with an average of 58.75, while levels of polymorphism ranged from 83.3% (Me5-Em2) to

98.4% (Me8-Em8) with an average of 91.9%. Individual accessions produced from 171 to 216 total fragments, with an average of 205.79, with as few as 11 fragments per primer pair (Em1-Me1) per individual, to as many as 41 (Me1-Em2). The number of polymorphic fragments varied per locus from 46 (Me1-Em1 and Me1-Em5) to 65 (Me3-

Em5) with an average of 54.125. Reproducibility rates were very high in the 31 random individuals tested for each of the eight primer pairs. On average, 91% of the 31 individual accession replicates yielded identical fragment patterns, and in examination over total fragments within individuals (14,570), more than 99% identity was recorded.

Analysis of groups

Examination of the number of fragments between the two groups of taxa, separated by presumed ploidy from commercial trade classifications VW, believed to be tetraploid, and VC, diploid, detected no difference in number of fragments (t(122)=

0.85023255, two-tailed, p>0.19; Table 2.3). Division of the VC group into groups of

“species types” and “cultivated hybrids” found that the species-type accessions had significantly more fragments per individual than cultivated varieties (t(28)= -3.31, two- tailed, p<0.001). Analysis by locus did not indicate any significant difference between groups (results not shown).

51

Principal coordinates analysis (Figure 2.3) depicted less variation than in either morphological approach reviewed in Chapter 1, with 19.9% in the first three axes. The primary axis generally separated accessions by the presence of blotches, and loosely by color. The second coordinate has an association with color traits and described 4.8% of the variation. Similar to the morphometric analyses (see Chapter 1 for details), wild-type accessions migrated to the periphery of the scatter, where like species were associated.

The third axis (~4.6% variation) did not appear to group accessions based on dominance of any individual morphological character.

Analysis of pair-wise similarity measures (Dice) varied widely, from 0.183 between V. tricolor (acn. 239) and V. dubyana (acn. 227) to 0.741 between two V. lutea accessions (acn. 909, 910). The mean maximum pairwise similarity between accessions was 0.626, while the average minimum similarity was 0.294, with a grand average of

0.448 over all comparisons. UPGMA analysis (Figure 2.4-A, B) clustered accessions by color, with some more tightly grouped (white, white with blotch, orange) than others

(blue, purple, yellow, yellow with blotch). There was also bootstrap support (>50%) for separation of some lavender and orange cultivars. Clustering via Ward’s method (Figure

2.4-C, D) was very similar to UPGMA analyses, as accessions with blotches were segregated with very strong support, with like-colored flowers clustered together. Within these two primary groups, there is very little support for topology beyond one or two steps from branch tips. Species-specific clades were again apparent, with V. tricolor and

V. lutea being located the most distantly from other blotchless accessions (VW vs. VC).

Cultivars of similar color and blotch status were often divided by their trade

52 classifications. All wild-type accessions clustered together by species, but with little support of how they were related to each other or cultivars beyond being distantly related to all hybrid cultivars. A low correlation between the data and the resulting tree was found under the WARD regime (CPCC=0.5601) compared to UPGMA (CPCC=0.7018).

Examination of probable clusters within the tested accessions by STRUCTURE software indicated that the optimal number groups was K=6, but secondarily supported

K=8 (Figure 2.5). Clusters demonstrated significant admixture, as virtually all accessions have contributions from each of the other groups in both the K=6 (Figure 2.6-A) and K=8

(Figure 2.6--B) scenarios. Within both schemes, all clusters contained individuals with identity over 79%, but also contained individuals with as low identity as 27%, with an average individual cluster identity of 59%. The optimal placement of many accessions was not clear, as many clusters contained individuals with nearly equal Q scores for multiple clusters (e.g., acn. 270: 0.4633 for cluster 3 and 0.4601 for cluster 5; acn. 302:

0.2504 for cluster 2, 0.2809 for cluster 4, and 0.2689 for cluster 5), suggesting a complex, shared lineage. Further, accessions without strong cluster membership typically differed morphologically from those with highly probably placement in the same cluster. This was apparent when Bayesian diagnoses were superimposed onto the clustering and ordination analyses, as the STRUCTURE groupings were divided into strongly overlapping groups in both analyses. Previous comparison of clustering approaches, reviewing UPGMA, Ward’s method, and STRUCTURE analysis, found that while

UPGMA topologies generally had much higher CPCC but occasionally unbalanced

53 clusters, Ward results agreed much more strongly with STRUCTURE results (Odong,

2011).

Considerable admixture within clusters was also demonstrated by AMOVA results. In the K=6 analysis, 90% of the variation was within clusters, and the remaining

10% was attributed to variation between clusters (PhiPT=0.100, p<0.01; Table 2.4).

Analysis by primer combination indicated a range of values of PhiPT, from 0.022 (Me5-

Em2) to 0.408 (Me1-Em1). The K=8 results indicated 93% of the variation was within clusters, and the remaining 7% was among clusters (PhiPT =0.068, p<0.01). Individual primer combinations (see Table 2.4 for details), values of PhiPT ranged from 0.020 (Me5-

Em2) to 0.386 (Me1-Em1). Only one primer combination (Me1-Em1) indicated more than 15% of the variation between STRUCTURE defined clusters in both K scenarios, suggesting that the Me1-Em1 locus produced descriptive markers in identifying variation between clusters. The grand average of between-cluster variation over both clustering projections was only 8.5%, indicating that clusters are highly heterogeneous and interrelated. The divergent genotypes possessed by wild-type taxa (i.e. V. tricolor, V. lutea) were confined to separate, small clusters, distant from clusters of cultivated hybrids.

Combined analysis

Of the 124 accessions characterized by SRAP markers, seven individuals were not scored in the morphological analysis (described in Chapter 1). These accessions were removed from this combined analysis due to their incomplete profiles. Results of the

Mantel test, indicated significant correlation between the two datasets (r=0.21, p<0.01;

54

Figure 2.7), but with a low explanatory relationship. Examination of 117 Viola accessions via PCoA utilizing Gower’s similarity coefficient found little variation relative to the independent morphometric analyses, with 10.7% of the variation in the first axis, and 21.5% in the first three axes (Figure 2.8). The first three axes all indicated some separation of accessions by class and blotch. Clustering by color was also maintained, demonstrated again by accessions with flowers of orange, white, and white and yellow with blotch, though groupings of other colors were small and intermixed.

Clustering by UPGMA (Figure 2.9-A) did not divide accessions by the presence of the blotch as seen in other clustering analyses, but did generally group accessions by flower color. For the majority of internal branches, there was very little to no support, but within clusters fewer than three steps from the branch tips there was generally good

(>50%) support. There was very strong support that V. altaica is distinct from all other clusters, and also for a secondary separating of white-flowered cultivars with blotches.

This analysis provided moderate conservation of the distance matrix (CPCC=0.6976).

The tree generated by the Ward paradigm (Figure 2.9-B) indicated there were two well- supported clusters, divided by blotch presence, with a similar pattern of little internal branch support. Clustering by color was less clear under this regime. Viola altaica was not segregated in this dendrogram, but clustered with V. corsica in the blotchless clade.

A low CPCC value was calculated for this tree (0.5066), which suggested divergence between the morphologic and genetic datasets.

55

Discussion Comparison of SRAP results to previous studies

Advancements in the development of dominant markers in recent years have provided researchers with new molecular tools (for reviews see Agarwal, Shrivastava, &

Padh, 2008; Jonah et al., 2011), including sequence-related amplified polymorphism

(SRAP; Li & Quiros, 2001) markers. This molecular marker has proven to be extremely valuable when assessing diverse germplasm collections (Ferriol, Pico, & Nuez, 2003;

Ferriol, Pico, Córdova, & Nuez, 2004b; Tam et al., 2005; Liu et al., 2006; Wang, Yang,

& Shen, 2011) and well as within highly derived hybrid lines (Liu et al., 2007; Yan et al.,

2009). Due to their low cost, repeatability, and ease of use, they have been extensively used to explore molecular variation in crops that have little available genetic data (for review, see Chapter 4). The SRAP markers are similar to AFLPs in that they were designed to be highly customizable with numerous ambiguous primers, though they have a distinct advantage of requiring fewer, more simplified steps to complete.

While SRAP markers have been compared to AFLP in terms of their utility (Li &

Quiros, 2001), some assert that SRAP markers are less useful and provide lower levels of polymorphism (Nybom, 2003). This suggestion could be premature, due to the limited literature base available in the first few years following SRAP development. A recent literature search (see Chapter 4) yielded over 300 studies published since 2004 in plant breeding and germplasm characterization, presenting SRAP primers yielding levels of polymorphism anywhere from 4% to 99%, using from five to 1634 primer pairs. Budak et al. (2004) compared four marker systems in buffalograss (Buchloe dactyloides) and found SRAP had more power revealing genetic diversity than SSR, ISSR, and RAPD. 56

Relative to AFLP, results of SRAP are mixed in terms of differentiating fragments produced, though in comparative studies, SRAP markers have provided a higher rate of polymorphism in many cases (Ferriol, Pico, & Nuez, 2003, 2004a; Li et al., 2007; Lin et al., 2010; Youssef et al., 2011).

As SRAP markers target coding regions, many studies have found them to be useful when paired with morphological analyses to address broader questions of phenetic clustering, involving measures of shape (Zang et al., 2010; Meng et al., 2012; Zhang et al., 2012), size (Ferriol et al., 2004a; Nagl et al., 2007), color (Gao et al., 2007; Liu et al.,

2007; Rahman, McVetty, & Li, 2007; Han et al., 2008; Chen et al., 2010; Wang et al.,

2012; Han et al., 2012), or geographic origin (Budak et al., 2005; Wang et al., 2009;

Wang et al., 2012). Color and size were primary characters that clustered a group of 43

Viola cultivars subjected to SRAP analysis (Wang et al., 2012). In the present study, morphological and molecular data present similar results in terms of clustering by blotch presence, color, and horticultural class (ploidy).

Molecular characterization of 124 accessions via SRAP markers demonstrated a significant amount of diversity relative to previous studies employing the same marker system. A review of SRAP literature to date is presented in Chapter 4, where the descriptive statistics of SRAP amplicons from 188 studies were reviewed. Here, an average of 17.7 SRAP fragments were scored per locus, with a range of 1.1 (Chen et al.,

2007) to 74 per locus (Alghamdi et al., 2012), as well as an average polymorphism rate of

69%, ranging from of 4% (Chen at al., 2010) to 99.7% (Jia et al., 2011). Scoring an average of 58.75 fragments per primer combination in the present study was more than

57 twice that seen in previous analysis of V. × wittrockiana using a similar suite of SRAP primers, which only discovered 23.81 fragments per locus (Wang et al., 2012). Similar levels of polymorphism were detected in these two studies, with mean polymorphic rate of 92% here (range of 83-98% per primer), relative to previously reported average of

93% (range of 79-100% per primer). One primer pair used in the current study (Me8-

Em8) was scored via digital analysis of samples in agarose gels and still yielded more than twice as many bands as previous work with the same primers (61 vs. 28), though it produced a comparable level of polymorphism (98.36% vs.100%).

Many factors may have contributed to the variability in the results of molecular analyses, including different PCR reagents, cycling protocols, visualization systems, and scoring techniques (including error rate calculation). Scoring the majority of PCR products in the present study was done via CE, in conjunction with band scoring software set to specific parameters. Most other studies found in the current literature have used traditional polyacrylamide gel analysis (PAGE), generally partnered with computer- based, digital image analysis software for band visualization, while some have visually scored agarose gels. Previous characterization of pansies with SRAP markers scored acrylamide gels by eye (Wang et al., 2012), potentially adding some degree of subjectivity to band scoring. In the present study, the CE technique resulted in high repeatability and robust amplicon detection, which may have enabled detection of fragments not apparent in manual scoring.

In some cases, the number of SRAP fragments has been able to distinguish between ploidy levels in related taxa. Examination of taxa within the Poaceae spanning

58 different ploidy levels has shown that the number of fragments produced increases with ploidy (Budak et al., 2005; Gulsen et al., 2009), which may have been a consequence, in part, of their potential for co-dominant nature. In the present study, comparison of Viola accessions by assumed ploidy did not demonstrate significant variation, though there were more bands scored per accession on average in VW than VC types (183.02 vs.

164.1, respectively). In partitioning the VC (diploid) group into cultivated versus wild- type taxa, a significant difference between the numbers of SRAP fragments produced by the two groups was detected (157.64 vs. 172.55, respectively). This high average of fragments in unimproved species was due in part to the greater number generated by V. lutea accessions, which averaged 186 scored bands per individual. This may be due to the higher number of chromosomes maintained by V. lutea, relative to other taxa examined

(Clausen, 1927). The V. lutea accessions in this study were metallophytes endemic to

Poland, which have been shown to have distinct cytology (Nauenburg, 1986; Hildebrandt et al., 2006) and may not be representative of British forms used in early pansy breeding.

Patterns of Viola clustering

Examination of the cluster and ordination analyses suggested clustering by color, blotch presence, and to a lesser degree, class (VW, VC, S). This trend was seen in all analyses, though it is unclear which of the morphological factors played a more significant role due to the high proportion of VW cultivars with blotches, and the low occurrence of this trait in the VC group (see Chapter 1 for discussion of potential sampling biases). Still, the clustering of accessions with similar floral morphologies, regardless of horticultural class suggests that they may share a common lineage of

59 blotched ancestors. All patterns in cluster analyses are speculative at best, based on the lack of resampling support (Figure 2.4). Low branch support of other analyses of cultivated hybrid Viola incorporating dominant markers has been reported (Wang & Bao,

2005; Wang, 2012), as have other interspecific-hybrid ornamentals characterized by

SRAP markers (Han et al., 2008).

After nearly two centuries of selective breeding exploring the plasticity of pansies, primarily investigating floral traits including color, pattern and size, a spectrum of floral color morphs exists. Comparison of all SRAP analyses in this study demonstrated general similarities in clustering, through grouping of accessions by color and blotch presence. Separation between blotched and non-blotched cultivars was a dominant, well supported feature, as was grouping of color patterns to varying degrees.

Distinction between other Viola species has also indicated genetic separation by flower color (Mereda et al., 2008).

The development of clear flower color is considered paramount by professional flower breeders, and is generally a function of combining inbred lines developed to carry these and linkages to other desirable traits. In modern Viola breeding programs, crosses between lines of dissimilar color, form, and habit are infrequently attempted, in order to preserve advances of previously captured traits for future generations. This practice appears evident in the clustering of accessions with white or yellow, blotched flowers, and white, unblotched flowers. Also, breeders freely incorporate germplasm from competing companies into their lines (Peter Stefany, personal communication, March 2,

2008), which would be expected to lead to increases both morphological and genetic

60 homogenization over time. Though results indicate clustering relationships based on color classes, it is not likely that the 470 scored SRAP amplicons are associated with genes related to flower color. The inheritance of color in different morphotypes of the presumed diploid V. tricolor has been shown to be affected by more than four genes

(Clausen, 1926), but more simple in other violets (Mereda et al., 2008, 2011).

Accessions with flowers of certain colors and patterns distinctly clustered in all

SRAP analyses. White cultivars with blotches were predominantly large-flowered VW accessions and were highly similar morphologically, representing a line that has been continuously developed since the late-19th century (Cuthbertson, 1910). Two accessions identified as white with blotch did not cluster with these, and varied in terms of assumed ploidy (278) and morphology (124). Accession 278 was a VC type that clustered with blue blotched accessions in UMPGA (Figure 2.4-A/B), and with yellow blotched cultivars following Ward’s method and the K=6 Bayesian clustering scenario (Figure 2.4-

C/D). In the K=8 analysis it was placed in the same group as the other blotched white cultivars (Figure 2.6-B). Accession 124 was morphologically distinctive within the white with blotch cohort, having a blotch on all 5 petals instead of the typical three. This accession also demonstrated a distinctive trailing growth habit relative to the mounding habit of the other white with blotch accessions. These traits were notably distinct, and suggested a divergent lineage. A line of pure white-flowered violets, without ray-like nectary guides, was a horticultural group also developed in the late 19th century (Cook,

1903). Of the seven rayless, white-flowered cultivars examined in this study, accessions

041, 042, and 043 were highly related, representing two inbred lines and their F1 product.

61

This subgroup of whites had high bootstrap support in all cluster analyses, providing validation of the ability of the SRAP technique to detect relationships between closely related individuals while distinguishing between them.

Similar to long-established lineages, accessions with novel colors, such as orange, also clustered together. Orange is a relatively new color class to the commercial hybrid pool, having existed only a few decades, and is not found in wild Melanium taxa. It has been suggested that the rise of this color probably stemmed from a single, diploid breeding program, and therefore has a restricted genetic base (Peter Stefany, personal communication, March 2, 2008). RAPD analysis of pansies by Wang et al. (2005) showed very strong support for relatedness of orange-flowered accessions. Orange- flowered cultivars in the present study were closely allied over all analyses, except for one accession (148) which deviated from the primary orange cluster in the UPGMA analysis (Figure 2.4-A/B). This may be due to its divergent origin, as it was a product of a European-based company whereas all others originated from companies in the United

States. Bayesian analysis similarly placed orange accession 148, as well as 78, apart from the other three orange-flowered accessions in both K scenarios.

Other colors appeared to be more loosely grouped, including purples, blues, and lavenders, which were intermixed in both cluster analyses. The presence of the blotch was a stronger segregating factor than these colors under Ward’s method (than UPGMA.

Accessions in this range were the majority in this analysis, and were a heterogeneous mix of ploidy, shape, and size. Accessions with bicolorous flowers, individuals sporting

62 superior petals of a different color than the lateral and inferior, loosely clustered in both methods, but this may have been biased by the fact VC cultivars dominated this group.

Previous analyses of genetic diversity in pansies have come to varying conclusions about the basis for clustering and relatedness of the two horticultural groups of cultivated Viola. Analysis by Wang et al. (2012) suggested that SRAP markers provided strong evidence for separation by trade classification. This was evident in

UPGMA analysis, as six of eight VC cultivars clustered, but no degree of bootstrap support was presented to bolster this finding. Similarly, Wang and Bao (2005) found segregation of presumed VC cultivars though cluster analysis of RAPD markers, but VC accessions were likewise a small minority of cultivars examined. Although the results of cluster analysis in the present study did not find a consistent molecular distinction between horticultural groups, the fragment number by trade classification interaction seen in the present data, suggests the potential for divergence between VW and VC cultivars.

When examining relationships between VW and VC cultivars and wild-type taxa, it is important to recognize that the Melanium group suffers virtually no hybridization boundaries by species or ploidy. Kroon (1972) found that hybridization between wild- type V. tricolor (2×=23) and cultivated VW cultivars (4×=52) was easily done, and downstream generations proved fertile though were primarily cytological intermediates and aneuploids. Similar results were also seen in crossing V. arvensis with commercial, polyploid hybrids (Novotna, 1978; 1981). Further obscuring cytological boundaries between horticultural groups, it has been shown that naturally-occurring hybrid swarms of interspecific pansy hybrids have demonstrated higher chromosome counts than either

63 parent (Becker, 1910; Clausen, 1924; Fothergill, 1938). Diploid and tetraploid lines of elite hybrids are often are intercrossed (Peter Stefany, personal communication, March 2,

2008), as isolation and linkage of novel traits is generally a faster process with diploids than tetraploids (for reviews see Nei, 1963; Khush, 1996). Thus, many advances in morphology of VW cultivars may arise from breeders’ efforts in developing VC lines.

The relationship between geographic origin of cultivars and SRAP markers has been mixed, depending on the crop and its history. Analysis of RAPD markers in Viola cultivars (Wang & Bao, 2005) has shown support for origin-specific markers in UPGMA analysis, but not under maximum parsimony constraints. In morphological and molecular analysis of pea cultivars (Pisum sativum L.), Esposito et al. (2007) determined there was no relationship between the origin of the accessions and the molecular clusters.

Such results may reflect the exchange of germplasm among breeding programs in different countries. Conversely, work by Ferriol et al. (2004b) saw clustering of cultivated varieties of Cucurbita moschata L. by country of origin. In this instance, many of these accessions were considered landraces, with few varieties receiving additional germplasm inputs from secondary domestication events or other means. The trend of

SRAP markers suggesting separation by geographic origin has been seen in many taxa, though many of these studies have focused on long lived organisms with long generation time (Han et al., 2008) or on regionally adapted landraces (Zaefizadeh & Rauf Goliev,

2009; Alghamdi et al., 2012; Zhang et al., 2012). Understanding the history of cultivation and hybridization, directed and natural, is critical for correctly interpreting genetic relatedness and origins.

64

Combined analysis

Previous studies in the pansies have incorporated both morphological and molecular analyses, though on smaller scales in terms of both approaches (Wang & Bao,

2005, 2007; Du et al., 2011). Typically, these data are reviewed independently, with assumptions that the two are not strongly correlated and their concatenation would lead to poorly supported associations (“noisy data”). Systematic analyses in other organisms have suggested that combination of these two types of data can be advantageous for describing the true relationships between individuals in complex systems (Huelsenbeck et al., 1996; Gatesey et al., 2000; Puorto et al., 2001; Giribet & Wheeler, 2002). In the present study, the morphological arrangement based on blotch presence, corolla shape, and petal color appeared to be mirrored in molecular clustering patterns. In analysis of the combined dataset, the presumption of congruency was evident in the segregation based on these traits. While there was distinctly less bootstrap support (>50%) for any of the branches under either UPGMA or Ward algorithms relative to either individual analysis, as well as low CPCC scores (suggesting disagreement in phylogenies between the two approaches), the maintenance of clustering by class (cultivated versus species types), blotch, and in some cases, color, suggest linkages between the morphology and

SRAP markers. Grouping of cultivars of some colors were distinct in both cluster analyses, including white with blotch and orange. While no specific markers were found to be uniquely associated with specific traits in cultivated types, unique bands or banding patterns were amplified in all species types at some loci. In conjunction with their morphological distinction from the cultivated types, it is not surprising that species

65 accessions clustered by taxon, specifically V. tricolor and V. lutea. The UPGMA approach distinguished V. altaica as distantly related to all other accessions, suggesting its limited contribution to hybrid violas, as has been previously suggested (Wittrock,

1896). Similarly, PCoA separated accessions by blotch presence and also by hybrid or wild-type origins, but yielded a small proportion of the variation within the first three axes suggesting a complex pattern of interrelation.

The finding that molecular information corresponds to morphometric characters measures suggests the potential for development of markers for floral traits, useful in plant breeding. Trait-associated SRAP markers could be used to develop sequenced- characterized amplified region (SCAR) or target-region amplified polymorphisms

(TRAP; Hu & Vick, 2003) markers for early screening for desirable traits in Viola breeding programs. The correlation of SRAP markers to easily-observable, morphological traits could also be employed for germplasm collection maintenance, allowing ornamental germplasm curators to maintain high levels of the extant genetic diversity in commercial hybrids by selecting on the basis of morphology alone. Still, due to the high levels of admixture and low support for cluster analyses, these relationships likely involve a very complex association between numerous traits, and examination via other molecular measures may increase resolution of the relationships between these plants.

66

Tables and Figures

Table 2.1. Sequence-related amplified polymorphism primers screened to characterize Melanium violets

Forward primers Reverse primers

Me1 TGA GTC CAA ACC GGA TA Em1 GAC TGC GTA CGA ATT AAT

Me2 TGA GTC CAA ACC GGA GC Em2 GAC TGC GTA CGA ATT TGC

Me3 TGA GTC CAA ACC GGA AT Em3 GAC TGC GTA CGA ATT GAC

Me4 TGA GTC CAA ACC GGA CC Em4 GAC TGC GTA CGA ATT TGA

Me5 TGA GTC CAA ACC GGA AG Em5 GAC TGC GTA CGA ATT AAC

Me6 TGA GTC CAA ACC GGA CA Em6 GAC TGC GTA CGA ATT GCA

Me7 TGA GTC CAA ACC GGA CG Em7 GAC TGC GTA CGA ATT CAA

Me8 TGA GTC CAA ACC GGA CT Em8 GAC TGC GTA CGA ATT CAC

Me9 TGA GTC CAA ACC GGA GG

Me10 TGA GTC CAA ACC GGA AA

Me11 TGA GTC CAA ACC GGA AC

The forward and reverse sequence-related amplified polymorphism (SRAP) primers tested for variation in a subset of 31 Melanium violets. Bold, red text indicates the primers selected as the most variable and repeatable primers used to characterize 124 Melanium violets (specific primer pairings given in Table 2.2)

67

Table 2.2. Descriptive results of fragments generated by SRAP primers and calculated error rates

Fragments / Proportion mean frag. / range frag. range frag. repeatability by repeatability by frag.: (total Locus locus polymorphic individual (low) (high) rxn (% of 31) frag. errors / total frag.) 1: Em1-Me1 50 0.92 19.07 14 29 0.94 0.9936 2: Em1-Me5 51 0.9216 20.45 11 30 0.94 0.9924 3: Em2-Me1 65 0.9385 25.81 17 38 0.87 0.9921 4: Em2-Me5 60 0.8333 27.65 19 38 0.9 0.9930 5: Em5-Me1 52 0.8846 25.38 19 36 0.9 0.9926 6: Em5-Me3 72 0.9028 22.21 15 31 0.9 0.9884 7: Em7-Me7 59 0.9661 19.15 14 30 0.87 0.9902 8: Em8-Me8 61 0.9836 21.54 16 33 0.97 0.9968 TOTALS 470 0.92 171.27 129 198 0.91 0.99

68

68

Figure 2.1. ROX799 size standard for scoring fragments in capillary electrophoresis

A test gel demonstrating the 17 individual markers of the ROX799 size standard used as reference for capillary electrophoresis, labeled with their size (bp) in relationship to a 1kb plus size standard (Invitrogen). Adapted from DeWoody et al. (2004), this size standard utilized a well-known viral cloning plasmid (pUC19) to develop primers for amplification of specifically sized fragments. Markers were individually amplified and combined to create a size standard for capillary electrophoresis in the present study.

69

Figure 2.2. Example test gel of SRAP amplicons

This 1% agarose test gel was run with SRAP fragments generated by primer combination Me8-Em8 from 16 accessions of Melanium violets with 1kb+ ladder as reference (Invitrogen). This primer combination consistently produced fragments in the 1000- 3000bp range, above the size capacity of available capillary electrophoresis equipment. Samples were instead run in 2% agarose gels at 40V for ~2.5h, and scored in Kodak 1-D software (Eastman Kodak, Rochester, NY).

70

Table 2.3. Student’s t-test comparing SRAP fragment number in classes of Melanium violets

Group n Mean frag. SD t df p

VW 63 183.02 26.5695 0.85 122 0.198

VC + species 61 164.1 24.8637

VC – cultivated hybrids 39 157.64 17.8716 -3.31 28 0.001

VC – species types 22 172.55 29.3592

TOTAL VW + VC 124 171.27 25.7156

Results of Student’s t-tests examining the difference between the number of SRAP fragments (frag.) amplified from classes (VW, VC) of Melanium violets. No difference was found between VW or VC classes, but within the VC group, species generated significantly more fragments over all loci and cultivated types. Bold, red text indicates a significant t score and its associated p value.

71

Table 2.4. AMOVA results for STRUCTURE clusters of Melanium violets A)

Locus Source df SS MS Est.Var % Stat Value p<

locus 1 Among Pops 5 4.083 0.817 0.039 41%

Within Pops 118 6.756 0.057 0.057 59% PhiPT 0.408 0.010

locus 2 Among Pops 5 2.417 0.483 0.010 5%

Within Pops 118 20.833 0.177 0.177 95% PhiPT 0.055 0.030

locus 3 Among Pops 5 3.570 0.714 0.020 15%

Within Pops 118 13.204 0.112 0.112 85% PhiPT 0.152 0.010

locus 4 Among Pops 5 1.235 0.247 0.003 2%

Within Pops 118 17.499 0.148 0.148 98% PhiPT 0.022 0.170

locus 5 Among Pops 5 2.262 0.452 0.012 11%

Within Pops 118 11.673 0.099 0.099 89% PhiPT 0.106 0.010

locus 6 Among Pops 5 1.483 0.297 0.006 4%

Within Pops 118 14.606 0.124 0.124 96% PhiPT 0.044 0.050

locus 7 Among Pops 5 3.974 0.795 0.023 17%

Within Pops 118 13.469 0.114 0.114 83% PhiPT 0.165 0.010

locus 8 Among Pops 5 3.113 0.623 0.014 7%

Within Pops 118 22.435 0.190 0.190 93% PhiPT 0.070 0.030

TOTAL Among Pops 5 22.137 4.427 0.113 10%

Within Pops 118 120.476 1.021 1.021 90% PhiPT 0.100 0.010

Analysis of molecular variance results for 470 SRAP markers generated by 8 loci in 124 accessions of pansy group Viola. STRUCTURE results with A) K=6 and B) K=8 groups were used as proxies for populations. Significant contributions of among population variation were detected at all loci except locus 6 (Em5-Me3) in the K=8 scenario, and locus 4 (Em2-Me5) in both clustering arrangements. The component of variance attributed to among population variation was significant measured over all loci in both clustering arrangements. For information on primers used at each locus, see Tables 2.1 and 2.2.

continued

72

Table 2.4 continued

B)

Locus Source df SS MS Est.Var % Stat Value p<

locus 1 Among Pops 7 4.090 0.584 0.037 39%

Within Pops 116 6.749 0.058 0.058 61% PhiPT 0.386 0.010

locus 2 Among Pops 7 3.703 0.529 0.011 6%

Within Pops 116 19.547 0.169 0.169 94% PhiPT 0.063 0.020

locus 3 Among Pops 7 4.803 0.686 0.018 15%

Within Pops 116 11.971 0.103 0.103 85% PhiPT 0.151 0.010

locus 4 Among Pops 7 1.690 0.241 0.003 2%

Within Pops 116 17.044 0.147 0.147 98% PhiPT 0.020 0.110

locus 5 Among Pops 7 1.786 0.255 0.005 4%

Within Pops 116 12.149 0.105 0.105 96% PhiPT 0.043 0.010

locus 6 Among Pops 7 1.622 0.232 0.003 3%

Within Pops 116 14.467 0.125 0.125 97% PhiPT 0.026 0.060

locus 7 Among Pops 7 2.819 0.403 0.009 6%

Within Pops 116 14.624 0.126 0.126 94% PhiPT 0.064 0.010

locus 8 Among Pops 7 3.276 0.468 0.009 4%

Within Pops 116 22.273 0.192 0.192 96% PhiPT 0.043 0.030

TOTAL Among Pops 7 23.790 3.399 0.074 7%

Within Pops 116 118.823 1.024 1.024 93% PhiPT 0.068 0.010

73

A)

74

Figure 2.3. Principal coordinates analysis of SRAP fragments in Melanium violets with superimposed STRUCTURE results Scatterplots demonstrating the first three axes of principal coordinates analysis from 470 SRAP fragments generated by eight primer pairs in 124 accessions of pansy-group Viola. Accessions’ horticultural class is indicated by shape: VW (squares), VC (circles), or species type (triangles). Filled shapes indicate the presence of blotch, and open shapes indicate absence of blotch. Circles around groups of squares indicate examples of accessions with similar color, and are individually labeled. Arrows indicate gradients in morphological traits, as labeled. The A) K=6 and B) K=8 STRUCTURE predictions are also superimposed and indicated by color, corresponding to clusters in the barplots in Figure 2.6. K=6 Color Key: cluster 1, red; cluster 2, yellow; cluster 3, green; cluster 4, blue; cluster 5, purple; cluster 6, black (in STRUCTRE barplot, white). continued 74

Figure 2.3 continued

B)

75

K=8 Color Key: cluster 1, red; cluster 2, yellow ; cluster 3, blue; cluster 4, orange; cluster 5, green; cluster 6, purple; cluster 7, grey; cluster 8, black (in STRUCTRE barplot, white).

75

A)

Figure 2.4. Cluster analysis of SRAP results in Melanium violets with K=6 and K=8 Cluster analysis dendrograms based on markers generated by eight SRAP primer pairs in 124 pansy group Violas following UPGMA approach employing Dice’s similarity coefficient with overlaid STRUCTURE scenarios of A) K=6, B) K=8, and Ward’s minimum variance method with STRUCTURE scenarios of C) K=6, and D) K=8. Bootstrap values are based on 1,000 resamplings of the dataset. Symbols above branches indicate support: asterisks represent >75% support, and ampersands represent >90% support. Clusters of morphologically similar accessions are indicated with brackets at right. Descriptive accession information can be found in Appendix A. K=6 Color Key: cluster 1, red; cluster 2, yellow; cluster 3, green; cluster 4, blue; cluster 5, purple; cluster 6, gray (in STRUCTRE barplot, white). continued 76

Figure 2.4 continued

B)

K=8 Color Key: cluster 1, red; cluster 2, yellow; cluster 3, blue; cluster 4, light green (STRUCTURE, orange); cluster 5, dark green; cluster 6, purple; cluster 7, black (STRUCTRE barplot, white); cluster 8, gray. continued 77

Figure 2.4 continued

C)

continued 78

Figure 2.4 continued

D)

79

A)

B)

Figure 2.5. STRUCTURE Harvester results for SRAP fragments in Melanium violets Clustering analyses suggest significant population structure among pansy group Viola accessions. Graphs show A) the likelihood scores for each value of K genetic clusters from STRUCTURE (Pritchard et al., 2000) and B) ΔK scores for each value of K genetic clusters following Evanno et al. (2005). Figures were generated using STRUCTURE Harvester (Earl & van Holdt, 2011).

80

A)

B)

Figure 2.6. STRUCTURE barplots of K=6 and K=8 in Melanium violets

STRUCTURE results indicated A) K=6 and B) K=8 were likely clustering arrangements. Clusters within barplots are labeled to indicate general morphological trends of accessions. Significant admixture was apparent in all clusters of hybrid types, but to lesser degrees in clusters dominated by species.

81

0.400

y = 0.1813x + 0.0681 R² = 0.044 0.300

0.200

0.100 Gower similarity Gowersimilarity (morphological)

0.000 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 0.900 Dice similarity (SRAP)

Figure 2.7. Mantel test of molecular vs. morphometric similarity matrices in Melanium violets

Mantel test indicated a significant correlation (r=0.21, p<0.01) between similarity matrices generated from 78 morphological traits and 470 SRAP markers in 117 Melanium violets.

82

83

Figure 2.8. Principal coordinates analysis scatterplot of combined morphological and SRAP data in Melanium violets Scatterplots demonstrating the first three axes of Principal components analysis from combined molecular (470 SRAP fragments) and morphometric (78 traits, see Chapter 1, Appendix B) characters, scored in 117 accessions of pansy group Viola. Accessions’ horticultural class is indicated by shape and color: VW (red squares), VC (green circles), or species type (yellow triangles). Filled shapes indicate the presence of blotch, and open shapes indicate absence of blotch. Circles around groups of squares indicate examples of accessions with similar color, and are individually labeled.

83

A)

Figure 2.9. Cluster analyses of combined data in Melanium violets

Cluster analysis of combined molecular (470 SRAP fragments) and morphometric (78 traits, see Chapter 1, Appendix B) data, scored in 117 accessions of Melanium violets under A) UPGMA (Dice) and B) Ward constraints, each carried out with 1,000 bootstrap replicates. Colored text indicates if an accession belongs to the V. × wittrockiana (VW, red), V. cornuta (VC, green), or species (S, yellow) class. Clusters of morphologically similar accessions are indicated with brackets at right. Descriptive accession information can be found in Appendix A. continued 84

Figure 2.9 continued

B)

85

Chapter 3. Phylogeography of Viola pedata L. (Violaceae)

Abstract

Phylogeography has become a broad area of study in the last few decades, primarily attributed to detection and analysis of discrete, molecular and genetic variation.

Such studies shed light on the effects geographic elements and climatological events have had on historical dispersal and modern distributions of biota. North American’s Pacific

Northwest and central Europe have been extensively studied, but eastern North America has received less attention, especially involving studies of herbaceous plants. Viola pedata L. is a perennial forb widely distributed in eastern North America, often in prairies and oak-pine savannahs, including areas associated with refugia during the last glacial maximum. In order to assess the diversity and phylogeographic relationships in this species, 458 individuals were sampled from 42 populations and one cultivar of V. pedata and subjected to characterization at eight microsatellite loci. Measures of allelic differentiation (X, NpL), diversity (A, A’), and heterozygosity (Ho, H’, H”), were calculated. To further examine population-level differentiation, neighbor-joining analysis and principal coordinates analysis were carried out. Calculation of FST and

STRUCTURE analysis were performed to assess the level of differentiation and clustering within sampled populations. Population cluster information was incorporated into a hierarchical AMOVA analysis, which estimated PhiST and assessed variation 86 between and among clusters (“regions”) as well as populations. Mantel tests were carried out to determine the relationship between FST and PhiST to measures of geographic distance, and correlation analyses were carried out to determine the relationships between molecular measures (X, A, H, Np, etc.) and geographic variables. One-way ANOVAs were employed to determine significance of variation between all of these measures of differentiation, diversity, and heterozygosity between STRUCTURE-suggested clusters.

PCR amplification generated 97 distinct alleles and up to eight per individual at some loci, which confirmed V. pedata is a high-order polyploid. Neighbor-joining and principal coordinates analyses indicated populations could be separated into two primary groups (eastern and western). STRUCTURE analysis suggested four groups, with clusters in the Driftless Area (DA), northeast (NE), southeast (SE), and interior lowlands

/ Ozark Plateau (IO). AMOVA indicated 88% of the variation detected was among populations, 8% within populations, and 4% among regions. Values of FST and PhiST were lower than expected for outcrossing, insect pollinated plants, but may have been affected by the presence of some fixed alleles at all loci. Analysis of variance indicated the DA cluster had the highest levels of diversity and heterozygosity. This cluster also generated two of the three cluster-level, private alleles detected in this study. Mantel tests indicated correlations between genetic and molecular distance, and also correlations between longitude and latitude with some molecular variables, due primarily to the divergence of DA populations. These results agree with molecular findings for other taxa, that the unglaciated DA was a refuge for many organisms during previous glacial maxima, and that the Appalachian Mountains play a significant role in dividing lineages

87 in their post-glacial dispersal to northern areas. Future investigations of V. pedata phylogeography should consider greater sampling in southern lowlands for refugial signals, and exploration of dominant markers such as AFLP or SRAP, to avoid concerns of fixed alleles and increased polymorphic amplicons.

Introduction

Phylogeography - the relationship between molecular phylogeny and geographic distribution (Avise et al., 1987) - has become an expansive and diverse field in the last few decades. The advances in molecular techniques have allowed for the discovery of trends in historical movement of numerous organisms, and the ability to estimate their past distributions. The summation of population-level genetic diversity findings has led to recognition of many well-supported dispersal routes following major, prehistoric climatic events in North America (Soltis et al., 1997; Carstens et al., 2005; Soltis et al., 2006), and have simultaneously suggested that some contemporary taxa with similar distributions may have experienced these historical processes differently, depending on their biological and ecological needs (Clark, 1998; Clark et al., 1998; Arbogast & Kegany,

2001; Martínez‐Meyer & Peterson, 2006). In order to understand the complex nature of previous organismal distribution and dispersal, it is important to investigate inter- population variation in a diverse collection of organisms over a wide and variable geographic range.

Genetic variation and structure between and within populations are functions of evolutionary forces including natural selection, genetic drift, and gene flow. In many

88 cases, these processes have been mediated by colonization and migration patterns

(Wright, 1951; Barrett et al., 1990) while directed by geo-historical events, including the movements of glacial ice (Hewitt, 2004). The climatic patterns leading to the multiple glacial cycles of the Quaternary Period caused shifts in habitat availability for terrestrial organisms. Subsequent migration was widespread, and in some cases, extinction of local populations or species was the end result (Jackson & Overpeck, 2000).

The bio- and phylo-geographic histories of contemporary taxa on all major land masses have been pursued in order to identify the effects of the last glacial maximum

(LGM) on terrestrial biota. In many regions, molecular data have led to fairly generalized trends of Pleistocene refugia and subsequent reclamation of suitable habitat. The examination of the numerous phylogenetic studies of taxa in Europe has revealed similar biogeographic patterns in plants and animals (Avise, 1998; Taberlet, 1998; Hewitt, 2000;

Petit et al., 2003; Tribsch & Schonswetter, 2003; Schmitt, 2007), owing in part to the effects of the east-west nature of major mountain ranges and north-south-oriented peninsular formations. Similarly, the phylogeography of the Pacific Northwest region of

North America has also been heavily studied, and patterns of genetic diversity in many taxa have been associated with the LGM (Soltis et al., 1997; Brunsfeld et al., 2001;

Jacobs et al., 2004; Carstens et al., 2005; Shafer et al., 2010).

The effects of historical glacial cycles on genetic diversity of broadly distributed taxa in Eastern North America have received significantly less review relative to other these other geographic regions (Soltis et al., 2006). In contrast to Europe and the Pacific

Northwest, this region has significantly greater land area, and is also more geologically

89 and ecologically complex. Molecular studies have shown that many plants that now occupy previously glaciated regions of North America have migrated north from southern refugia following glacial retreat (Parker et al., 1997; Dorken & Barrett, 2004; Griffin &

Barrett, 2004). While multiple corridors of colonization have been identified for many animals (Soltis et al., 2006), trends maintained by plants have not been so defined, due in part to the limited literature base (Schaal et al., 1998).

Although it has been shown that patterns of genetic diversity in seed plants have varied dramatically based on their life history characteristics, including growth form, mating system, and seed dispersal mechanism (Hamrick & Godt, 1996; Nybom &

Bartish, 2000; Nybom, 2004), the bulk of phylogenetic studies describing eastern North

American plants have focused on long-lived woody species (Soltis et al., 2006; Jaramillo-

Correa et al., 2009). This focus has been due to numerous factors, including species’ longevity, year-round identification and sampling potential, and the availability of historical data via fossil records such as pollen from sediment cores. Of these, few utilize species or sampling regimes that are very widespread both longitudinally and latitudinally in eastern North America (McLachlan et al., 2005). Though general trends exist, the research to date supports numerous hypotheses about prior distributions of plants and their dispersal since the LGM, suggesting reclamation routes and timeframes were complex, diverse, and directed by many factors including life history and ecological requirements. This underscores the need for broader examination of plants with diverse biological characters and needs, and for greater expansion into forbs.

90

The few studies that have focused on herbaceous plants as subjects of research investigating glacial refugia and recolonization events in the North American southeast have supported previous findings of phylogeographic discontinuity (Joly & Bruneau,

2004; Gonzales et al., 2008). Cryptic northern refugia have also received limited research attention (Griffin & Barrett, 2004; Li et al., 2013), though are supported by analysis of variation in many arboreal taxa (Parker et al., 1997; MacLachlan et al., 2005;

Saeki et al., 2011). Further analysis of the regional genetic variation of widely adapted, herbaceous taxa could significantly enhance the understanding of the continent’s biological and ecological history, as well as provide information valuable for maintenance of local genotypes.

In North America, the genus Viola includes about 70 endemic species (United

States Department of Agriculture, Germplasm Research Information Network website, accessed 23 March, 2012), which are both morphologically and biogeographically diverse. Distinctive vegetative characters can be used to describe representatives of this genus, including acaulescent habit (producing prostrate, fleshy rhizomes), and cordate leaves, though there are some examples of species that have upright stems and dissected leaves. Throughout the genus, five-petaled, zygomorphic flowers are typically white or purple, though yellow-flowered taxa are not uncommon. Seasonally dependent, mixed mating systems also dominate reproductive habit in this group. Many of the stemless taxa of eastern North America have very broad and diverse distributions, including V. blanda Willd., V. cucullata Aiton, V. palmata L., and V. sororia Willd.

91

Viola pedata L. is one of the most distinctive of North American Viola. Widely dispersed in eastern North America, it is most common in short grass prairies, oak-pine savannahs, and open, dry, oak-pine forests from New Hampshire west to Minnesota, south to Texas and east to Georgia (Figure 3.1). This species typically grows in well- drained, acidic, nutrient-depleted soils (Molano-Floroes, 1999), where disturbance, such as fire, maintains an open or discontinuous canopy. This violet has been documented as having great abundance in a number of phylogenetic regions of interest, including the DA of southwest Wisconsin and bordering states, the Ozark Plateau of northern Arkansas and south-central Missouri, and the south and central Appalachian Mountains (Klaber, 1976;

Steven Carroll, personal communication, April 13, 2010). It is also common in many

New England states (Haines, 2011), and some small, isolated populations north of the

Great Lakes (Kavanagh et al., 1989; Hutchison & Kavanagh, 1994). Although broadly distributed, this violet is typically locally abundant and demonstrates a strong discontinuity in its range, as it is virtually absent in the Ohio and Mississippi River

Valleys. It is locally abumdant over much of its range due to its ability to readily colonize poor or disturbed soils in roadside margins where grasses and woody species are maintained (Burns & Cusick, 1984).

Viola pedata also carries a number of morphological and cytological characters rare to the genus. Its acaulescent, erect caudexes produce white, fibrous roots. Typically overwintering as a condensed rosette, the first flush of foliage produces palmate laminae with three to five divided lobes, coarsely ciliate to fimbriate-margined, bourne on short

(2-4cm) petioles, which are held below flowers (Figure 3.2). In the warmer months of

92 summer following flowering, new flushes of foliage have notably larger laminae (2-3x), and have more distinctly oblanceolate to narrowly obovate lobes (Figure 3.2). These leaves are also presented on longer petioles (~6-12cm). Flowering occurs in early spring, with (nearly) self-incompatible flowers presented above the foliage (Becker & Ewart,

1990). Single, pedunculate inflorescences emerge from the caudex, presenting large

(~3.5cm) corollas, commonly shades of lavender to purple, leading to the common name,

“mountain pansy.” The distinctly unbearded petals are variably broad and reflexed, and exserted, bright orange anthers make these reproductive structures unique in eastern Viola species.

Two co-occurring color morphs exist, but have notably distinct patterns of distribution. Viola pedata var. lineariloba is the most common form, occurring over the entirety of the species range. This variety displays concolorous petals, ranging from purple to light lavender (Figure 3.3-A). Viola pedata var. pedata is specific to the southwest and central part of the species’ range (though rare reports of outliers do exist), where it occurs with var. lineariloba, and is most abundant in the Ozark Plateau and east toward the Cumberland Plateau (personal observation) (Figure 3.4). Variety pedata has strongly bicolorous flowers, sporting dark purple superior petals and light lavender lateral and inferior petals (Figure 3.3-B). It is not uncommon for individual clones to sport incomplete, dark coloration over the superior petals (Figure 3.3-C), or have irregular patterns of dark coloration on lateral and inferior petals (Figure 3.3-D, E). Albino forms,

V. pedata var. alba (Figure 3.3-F), are rare and seemingly spontaneous in occurrence, though some populations they have been reported as to occurring in a relatively high

93 proportion (Steven Carroll, personal communication, April 13, 2010; William

Tannenberger, personal communication, May 3, 2010). Other stemless taxa with similar color morphs exist in North America, including V. beckwithii Torr. & A. Gray, V. hallii

A. Gray, and V. trinervata (Howell) Howell ex A. Gray, but these species maintain strictly bicolorous flowers and occur in very limited distributions west of the Rocky

Mountains. Unlike the majority of North American representatives in the genus, V. pedata does not demonstrate a mixed-mating system, producing only chasmogamous flowers (Becker & Eward, 1990). Cytologically, V. pedata has been shown to be a polyploid (2n=54), though there is not agreement in the literature on how best to describe the level of polyploidy (Canne, 1987; Marcussen et al., 2012).

Few North American violets have been the subject of molecular, systematic, or ecological studies (Culley, 2001; Culley & Wolfe, 2001), although some rare endemic taxa have been the focus of sequence-based, systematic studies (Ballard & Sytsma, 2000;

Havran et al., 2009; Marcussen et al., 2011). Other PCR-based marker systems have been optimized to identify variation within the North America Viola, but these have not been applied to taxa over broad geographic areas (ISSR, Culley & Wolfe, 2001; Culley,

Sbita, & Wick, 2007; SSR, Culley et al., 2005). The availability of targeted, molecular tools, widespread distribution, and distinctive life histories suggest that examination of many North American violets may provide valuable information on the effects of the

LGM on herbaceous plants endemic to eastern North America.

Because V. pedata is present over such a broad geographic area, including multiple proposed glacial refugia, this species is an ideal model for describing post-

94 glacial phylogenetic patterns of eastern North America. The simple-sequence repeat

(microsatellite) markers developed by Culley (2005) were employed to evaluate populations of V. pedata over its range. The objectives of this study were to determine if these markers could provide support for previous findings of polypoidy, describe patterns of genetic diversity in V. pedata, or detect regional variation and correspondence to previously-reported glacial refugia and genetic discontinuities, with an overall goal of enhancing the understanding of phylogeographic tends in eastern Northern American.

Methods Study Sites and Sampling

Leaf tissue was collected from flowering plants in 42 naturally-occurring populations of Viola pedata over the course of three field seasons in 2009-2011 (Table

3.1). Tissue from two clones of a cultivated variety of V. pedata was included as well, yielding a total of 458 individuals. Newly expanded leaves of 7-19 flowering individuals

(average10.65 individuals/population, range of 7-19) were collected from each population and dried in paper coin envelopes with silica gel (Chase & Hills, 1991). The species variety (pedata, lineariloba, alba) of each sampled individual was recorded.

Voucher specimens are deposited at the Ohio State University Herbarium.

DNA Extraction

Genomic DNA was isolated from 30-50 mg of silica-dried leaf tissues in a randomized design (as to avoid batch effects) using a modified mini-prep protocol derived from the CTAB extraction method of Doyle and Doyle (1987; sensu Wolfe,

2005). As Viola tissues typically contain high levels of polysaccarides and some 95 phenolics (Bubenchikov & Goncharov, 2005), extracted Viola DNA is often viscous, making it challenging to pipette and difficult to amplify by PCR. Modifications to the

CTAB method made by Porebski, Bailey, and Baum (1997) to reduce the presence of secondary compounds were utilized: an addition of 3 µl RNAse A added to each sample before the 65º C incubation phase; a second chloroform / isoamyl alcohol wash; and the addition of 200 µl of 5 M NaCl with the first cold precipitation in 100% isopropanol. To ensure quality of resultant DNA, products were visualized in a 1% agarose gel (made with 1X Tris Acetate EDTA buffer, TAE, pH 8.0), stained in 0.1% ethidium bromide, and imaged under UV irradiation. Quantity and quality of whole DNA was assessed via comparison to reference bands of a 1kb plus marker (Life Technologies, Grand Island,

NY). Previous work in Viola has demonstrated 30-50 mg of dry leaf material yields from

75 to 300 ng/μg whole DNA (unpublished data; Harvey Ballard, personal communication, August 15, 2009).

Amplification and Genotyping

Twelve simple sequence repeat primer pairs developed for the V. pubescens genome (Culley, 2005) were screened for variability in a subset of 31 V. pedata samples, and of these, eight were selected due to the diversity and repeatability of alleles produced.

The primers used targeted the Vpub7, Vpub9, Vpub11, Vpub16, Vpub21, Vpub57,

Vpub60, and Vpub69 loci. The template for PCR amplification consisted of raw DNA extract diluted 1:19 in TAE buffer, yielding ~20-50 ng genomic DNA. The PCR amplifications were performed in a reaction volume of 10 µL, comprised of 0.05 µl (0.25

U) of ExTaq (Clontech Laboratories, Inc., Mountain View, CA), 1 µl of 10X ExTaq PCR

96

Buffer, 0.8 µl of ExTaq dNTP Mix (2 µM), 0.5 µl of fluorescence-labeled forward strand primer (0.5 µM), and 0.5 L of reverse strand primer (0.5 µM) in 5.1 µl UV-irradiated

HPLC pure water (Fisher Scientific, Pittsburgh, PA). The amplification was carried out in a randomized block design using an Eppendorf Mastercycler ep gradient S thermocycler

(Eppendorf North America, New York, NY) following optimized PCR protocols previously described (Culley, 2005). The presence of PCR product was confirmed by running 2L of reaction product on a 1% agarose gel.

Genotyping was done via capillary electrophoresis on an ABI Prism 3100 Genetic

Analyzer (Life Technologies, Grand Island, New York) using POP-6 polymer and an in- house developed program. Samples were arranged in a randomized block design and prepared by mixing 1 L of PCR product with 9.5 L of HiDi Formamide (Life

Technologies, Grand Island, New York) and 0.5 L GeneScan ROX500 size standard

(Life Technologies, Grand Island, New York). Samples were denatured for five minutes at 95º C, then immediately placed on ice for 5 minutes, and stored in the dark at 4º F until loaded into the genetic analyzer. Resulting electropherograms were scored with

GeneScan analysis software (version 3.7, Applied Biosystems), and manually reviewed and adjusted in the absence of sample origin information. Samples that were unscorable due to high rates of dimerization in some PCR products were subjected to a 20% polyethylene glycol (PEG) precipitation procedure to eliminate signal interference by small fragments from primer dimerization and nonspecific binding. This was accomplished by adding equal volume 20% PEG solution (292.5 g/L NaCl 200 g/L PEG

8000, ddH20 to volume) to PCR reactions and storing samples on ice for 1 h, then

97 centrifuging samples at 15,000 rpm for 15 minutes, followed by removal of the supernatant from the pelletized sample in resuspension in 100% EtOH. This spinning/washing process was repeated in two additional cycles, but with resuspensions in 80% and 75% EtOH, and 7 min centrifuging cycles at 15,000 rpm. Following alcohol washes, the pellet was resuspended in UV-irradiated, HPLC-pure water at 75% of the initial PCR sample volume.

Genetic Analysis

Data confirmed the polyploid status of V. pedata, as multiple loci produced up to eight allelic fragments from single accessions. Further, this finding indicates a higher order ploidy than the tetraploid condition (2n=4x=54) previously reported (Canne, 1987), and supports the decaploid claim of Marcussen et al. (2012). Thus, it was not possible to determine the number of copies of each allele because of the unknown origin and dosage levels of alleles in heterozygotes. Further, we could not estimate traditional heterozygosity levels or Hardy-Weinberg equilibrium. The data were therefore treated as dominant markers, and the presence/absence of bands was used to calculate statistics to assess intraspecific differentiation and diversity. To describe diversity, the allelic phenotype scoring methods of Sampson and Bryne (2012) were followed: diversity was measured as the total number of alleles seen over all loci (A), as the number of different alleles seen in a population, averaged over loci (A'), as the proportion of heterozygous loci within individuals, averaged over loci (Ho), and as the number of alleles seen in an individual per population, averaged over loci (H′). Further, the proportion of individuals heterozygous at all loci (H'') was calculated. The number of different allelic phenotypes

98 at each locus (L) in a population was counted (NpL) and an average was calculated across loci (Np). The number of different allele phenotypes was also calculated by locus for the total sample. Also scored were the number of alleles present in an individual (X) and the adjusted number of alleles (X’) which took into consideration the incomplete data

(completely missing loci) of some individuals. This was calculated by determining the proportion of alleles present of all alleles at scored loci (<8 loci), divided by the total number of alleles over all possible loci (8 loci, 97 total alleles). These molecular measures (X, A, H, etc.) were also calculated for each of the two varieties sampled

(lineariloba and pedata), which were treated as populations for secondary analysis.

Estimation of the genetic similarity and distance between “partial heterozygotes”

(sensu Bruvo et al., 2004) in polyploid organisms has proven problematic. Following

Kosman and Leonard (2005) and Sampson and Bryne (2012), we agree banding patterns in polyploids from co-dominant or dominant marker systems are best represented as molecular phenotypes. Nei’s (1972) genetic distance was calculated using the program

AFLP-SURV v. 1.0 (Vekemans et al., 2002), and a majority rule consensus neighbor- joining (NJ) tree was constructed using Nei’s distance by the NEIGHBOR program in

PHYLIP 3.69 (Felsenstein, 1989) from 1,000 bootstrap replicates generated using the

CONSENSE program. The final tree was drawn with UNROOTED, software developed at PBIL (Pôle Bioinformatique Lyonnais, http://pbil.univ-lyon1.fr/). To investigate the potential for differentiation between populations due to isolation by distance, Mantel tests were utilized to calculate correlations between pairwise "genetic" measures and geographical distance measures between populations in GenAlEx v. 6.5 (Peakall &

99

Smouse, 2006, 2012). Further, correlation analyses were performed between all population diversity measures (X, A, H, etc.) and geographic variables (longitude, latitude, elevation, STRUCTURE cluster – described below). To visualize the relationships of individuals, principal coordinates analysis (PCoA) of binary allelic data using Dice’s genetic distance was calculated by PAST software (Hammer, Harper, &

Ryan, 2001).

The R package Polysat (Clark & Jasieniuk, 2011) is one of the few programs to date that will appropriately analyze ambiguous polyploid data, and was used to calculate pairwise FST values. Polysat was also employed to generate a data file formatted to run in the Bayesian model-based clustering program STRUCTURE v. 2.3.4 (Pritchard,

Stephens, & Donnelly, 2000) following the approach of Falush et al. (2007) for ambiguous genotypes from dominant marker data. As the input format of STRUCTURE requires that every locus for each individual has as many values as the estimated ploidy, which is based on the highest number of alleles at any locus, individuals’ genotypes had one arbitrary allele duplicated at each locus, as necessary, up to the ploidy of the sample

(up to x=10, in this study). Individuals with a lower estimated ploidy than the level of user directed ploidy (< x=10, due to multiple copies of the same allele) received a missing data symbol (-9) to fill in the extra values.

STRUCTURE analysis estimates the probability of an individual's assignment to a specific genetic cluster. Individuals from related origins can be detected, as the probability of ancestral identity will be divided between specified K groups. Individuals are assigned to clusters and the proportion of an individual’s genome (Q) that originated

100 from each cluster is determined. The STRUCTURE program was run with no prior knowledge, under the admixture ancestry model, with ploidy set to 10 and putative population information given for each individual. As a preliminary screen to estimate K, the Markov chain Monte Carlo (MCMC) parameters were set to a burn-in period of 104 with 105 iterations, testing K values 1 to 44 with three replicate runs for each value.

Probability of K values was calculated using the method of Evanno et al. (2005) in the web-based program STRUCTURE Harvester (Earl & von Holdt, 2011). Review of analyzed K values indicated the highest likelihood for values under 10, at which time

STRUCTURE analysis was rerun with a longer burn-in period of 105 followed by and increased number of iterations (5x105), testing K values 1 to 10 with 10 replicate runs for each value. When the optimal cluster value was determined, STRUCTURE was rerun with 106 and 5x106 iterations, 10 repetitions at that value. Similarity in STRUCTURE runs was calculated by the method described by Jakobsson and Rosenberg (2007) via their computer program CLUMPP 1.1.2, using the fullsearch option. This program calculates a similarity coefficient, H, a statistic assessing the congruence of individual, replicate runs, and also calculates a mean cluster assignment probability per individual derived from all of the input matrices. The average clustering probabilities were then visualized using the program DISTRUCT 1.1 (Rosenberg, 2004). The STRUCTURE calculated membership assignments of each of the populations was similarly analyzed in

CLUMPP 1.1.2 and visualized via DISTRUCT. Further analysis of determined clustering patterns included one-way analysis of variance (ANOVA) for all diversity

101 measures (X, A, H, etc.) over the K clusters and also recognized varieties (lineariloba and pedata) using PAST software.

Analyses of molecular variance (AMOVA) were executed in GenAlEx, yielding estimates of PhiST: a statistic analogous to FST (Wright 1969, 1978), dividing the variance into the among-population (AP) and within-population (WP) components. GenAlEx also provided hierarchical partitioning of variance though second variable to group populations, known as ‘regions’, allowing analysis to attribute variation to an among- region (AR) source. Here, the clustering scenario suggested by STRUCTURE was incorporated, and the AMOVA was run with K clusters (used as a proxy for regions) as a source of variation, as well as populations. Due to the high number of alleles and ambiguous number of total alleles per individual, the data were considered dominant markers and analyzed as haploid. As a means to visualize the relationships of grouped populations, the STRUCTURE-determined cluster identity information was superimposed onto the previously calculated PCoA of individual samples. A second

AMOVA separating individuals by taxonomic variety into two metapopulations (ignoring sampling locality), was also carried out without STRUCTURE clustering information.

Results

The microsatellite primers developed by Culley (2005) amplified well in V. pedata. Genetic phenotypes consisted of one to eight bands per locus, within the range expected for a decaploid genome (Table 3.2). The maximum number of alleles amplified at each locus per individual ranged from five at locus Vpub69 to eight in Vpub21 and

102

Vpub57 (Figure 3.5). In total, 97 different alleles were amplified from eight microsatellite loci, with a range of five (Vpub69) to twenty (Vpub21) per locus. Of these alleles, nine were invariant over the samples examined and considered “fixed” or null alleles. Each individual examined demonstrated a unique molecular phenotype, when examined across all loci, with the exception of the two accessions of the cultivar ‘Eco

Artist’s Palette’, which were identical. The amount of missing data varied by primer, with a high of 42 individual accessions (9.17%) unscored for Vpub7 and a low of five

(1.09%) missing for Vpub69. Over the eight independent loci examined in each of the

458 accessions (3,664 total loci), 161 loci (4.39%) did not yield scorable data (for missing data information, see Table 3.2).

Measures of genetic diversity between and within populations indicated dramatic levels of variation between populations (Table 3.3). The total number of alleles per population (A) varied from 33 in MA-1 to 60 in WI-6, as well as the lowest (4.125) and highest (7.5) number of alleles per locus (A’), respectively. The number of alleles per individual per locus (H’) ranged from 2.34 (MA-1) to 4.25 (EcoArt), while the proportion of heterozygous loci per individual (Ho) ranged from 0.625 (VA-1) to 0.9125 (WI-3).

Only a small fraction of populations (3/43) maintained individuals that were heterozygous at all loci, ranging from 10% (MO-1) to 44% (WI-4). The average number of different alleles per individual per locus also varied by population, with the lowest value demonstrated by population VA-1 (3.75), and the highest by KY-1 (8.75). The number of unique alleles per locus per population is described in Table 3.3. No single population demonstrated any private alleles. There were also significant correlations of

103 latitude with X’, and longitude with A, A’, and Ho (Table 3.4). Mantel test results indicated that there were significant correlations between FST (r=0.503, p<0.01) and PhiST

(r=0.16, p<0.01) with geographic distance (Figure 3.6), but with a low explanatory relationships.

The NJ consensus tree based on a matrix of Nei’s (1972) distance between populations suggested divergence of populations based on an east-west divide (Figure

3.7). Populations from the Appalachians and southeast were grouped, while those from the Midwest were clustered together. There was no strong (>50%) support for any of the proposed branching patterns. Likewise, PCoA based on Dice similarity index did not clearly separate populations. When analyzed as individuals, only 7.47% of the variation was captured by the first axis, and 16% in the first three (Figure 3.8).

Bayesian clustering analysis by STRUCTURE indicated the most probable clustering arrangement was at K=4 (Figure 3.9), with significant admixture between individuals within populations, and between populations within the same cluster (Figure

3.10). There was strong support for K=4, as calculated in the method of Evanno et al.

(2005). Review of eight repeated runs of this K-value assignments via CLUMPP further supported the K=4 pattern, as the H’ value was high (0.8513) indicating strong congruence between independent runs. The four clusters were roughly associated with the northern Appalachian Mountains (NE), southern Appalachian Mountains (SE), the

Driftless Area (DA), and interior lowlands and Ozark Plateau (IO) (Figure 3.11), and agreed with results of neighbor-joining analysis. The mean population proportion assignment over all individuals was Q = 0.76, and within clusters ranged from Q = 0.694

104

(IO) to Q = 0.826 (DA). A second, less likely proposition of K=8 also received some support via Evanno’s method. CLUMPP analysis of repeated runs at this value indicated there was significant variation between runs (H=0.5182), suggesting this clustering pattern was much less likely than the K=4 arrangement. While this higher K appointment of populations maintained similar patterns to the K=4 scenario, many populations with higher rates of missing data were separated to form new, smaller clusters (data not shown).

Variation in diversity measures between STRUCTURE-defined regions was also detected by ANOVA (Table 3.5; see Appendix D for full analysis details). One-way

ANOVA indicated that the DA region had the greatest variability, in terms of total number of alleles overall and at specific loci, and also the highest levels of diversity and heterozygosity across the different measures. No significant differentiation was detected between the taxonomic varieties.

AMOVA results totaled over all loci indicated 88% of the variation rested within populations, 8% among populations, and 4% among regions (Table 3.6), and a significant

PhiPT score of 0.184 (p<0.010). Results of variation for individual loci suggested markers differed in the amount of variation they explained. Variation described by individual primers ranged from 83-99% within populations, 1-16% between populations, and 0-10% between regions, while PhiPT ranged from 0.003 to 0.175. These PhiPT values may underestimate the level of differentiation in the data due to calculation of negative values of PhiPT between populations. The analysis software employed for this calculation

(GenAlEx) converted to zero by default, which lowered average pairwise distances.

105

Analysis based on taxonomic variety indicated there was no significant difference between them at any individual locus or over all loci.

Discussion Variation within and between populations

The primary objective of this study was to describe the genetic structure of Viola pedata, a widespread perennial forb. Overall genetic diversity and heterozygosity in the species was high, though the majority of this variation (88%) was detected within populations. Findings of among-population diversity relative to within-population diversity were lower than estimates detected by other dominant markers, but low values were generally expected for outcrossing perennial taxa relative to those with other reproductive or persistence schemes (Hamrick & Godt, 1996; Nybom, 2004).

Outcrossing between populations has been shown to increase levels of genetic variation within a population, even if the gene flow is occasional or infrequent (Wright,

1951; Hamrick et al., 1990). Overall, populations of outcrossing species, such as Viola pedata (Becker & Ewart, 1990), may be genetically similar to one another since they share a high number of alleles (Loveless & Hamrick, 1984; Hamrick et al., 1990;

Hamrick & Godt, 1996). It has been demonstrated that gene flow plays a very important role in conserving genetic diversity in outbreeding species (Ellstrasnd, 1992; Hamrick,

1987; Govindaraju, 1988; Schoen et al., 1991; Hamrick et al., 1992; Hamrick & Godt,

1997). Such species have been shown to be more vulnerable to the loss of genetic variation through habitat fragmentation than self-compatible species (Honnay &

Jaquemyn, 2007). The present finding of low inter-population and high intra-population 106 variation indicates limited or recent division or fragmentation between populations or regions. Comparison of diversity measures suggests that recent limitations to gene flow, including the significant degradation of V. pedata habitat (tall grass prairies, pine-oak savannahs, and other fire-mediated ecosystems) since the westward colonization of

European settlers in North America (Samson & Knopf, 1994) have had little effect on genetic transmission between populations. On a geological time scale, the Quaternary glacial advances may have constricted or condensed the habitat suitable for V. pedata, facilitating gene flow for persistent conspecifics through increased proximity of surviving populations.

The low level of among population variation was enhanced by the presence of some minimally variable, “fixed” (null) alleles. Examination of allelic variability showed that nine alleles (9.5%) were fixed (>97% presence across all individuals), with at least one invariable allele occurring at each locus, and 17 alleles (17.4%) appeared in >75% of individuals. The presence of invariable markers generated by this primer set has been previously documented within the genus Viola, where it was shown that octo- and higher- ploid taxa had at least two fixed, heterozygous alleles (Hepenstrick, 2009). Further, no private alleles were discovered for any population, and only two cluster-specific alleles were found in across all loci (via Vpub69).

The level of allelic fixation may reflect a number of historical hybridization events. It has been determined that V. pedata is an intersectional polyploid, the function of multiple inter-specific hybridization events occurring ~1 million YBP (Marcussen et al., 2012). Through limited hybridization events, it may be that few alleles were

107 contributed to genomic recombination, possibly resulting from the cleistogamous nature of proposed ancestral taxa. Low fertility of hybrid offspring may have played a role in limiting inter-specific genetic contributions of unique alleles, which has been seen in other plant examples (Stebbins, 1958; Rieseberg et al., 1995; Rieseberg, 1997), especially in the case of polyploidization (Soltis & Soltis, 2009), though it does vary by species and environmental constraint (Arnold & Hodges, 1995). Alternatively, the fixation of certain alleles may have been generated by historical bottlenecks within V. pedata, or contributing taxa, following glacial or oceanic advances mediated by ancient climatic events.

Suggestion of ploidy via Vpub markers

Previous research examining V. pedata concluded that this species was a polyploid organism (Clausen, 1964; Canne, 1987) based on chromosome number relative to taxa presumed related (2n=4x=54). The findings in the present study (mean of 3.07 alleles per individual, per SSR locus) confirm the polyploid status of this species. While the Vpub69 locus produced an average of 1.16 alleles per individual, other loci (Vpub16,

Vpub60, Vpub1, and Vpub4) produced up to four alleles per individual, and some (Vpub7,

Vpub9, Vpub11, and Vpub57) even generated more than four in numerous individuals across populations (Figure 3.5). These findings agree that V. pedata is a higher-order polyploid (>4x). Marcussen et al. (2012) investigated the origins of presumed high polyploid Viola (including V. pedata) as part of a larger study, to determine species networks and elucidate the origins of the duplicated genomes in these taxa.

Characterization of their ancestral genomes with both the low-copy nuclear gene glucose-

108

6-phosphate isomerase and the chloroplast trnL-F region led to the conclusion that all high-polyploids were closely related, although many had been previously classified as originating from distinct sections of the genus.

Evidence exists suggesting that the use of the Vpub loci may not be appropriate to determine ploidy. Hepenstrick (2009) applied the SSR primers developed by Culley

(2005) to investigate the origins of allopolyploidization events in European Viola through examination of 38 species and seven interspecific hybrids. In optimizing reactions to target these SSR in a diverse, intersectional group of polyploid Viola, fragment patterns were generated that differed at low and high annealing temperatures with some primer pairs. Due to their locus-specific length differences, these fragments were presumed to be artifacts and were not scored. For example at Vpub9, an additional fragment was detected, 15bp shorter than the presumed target alleles, but only amplified at annealing temperatures both lower and higher than reported as optimal (Culley, 2005). The use of

Vpub9 primers in V. pedata in the present study generated up to six fragments per sample at the recommended annealing temperature, all of which were scored. Also, two loci examined in the current study (Vpub7 and Vpub69) were both dismissed from

Hepenstrick’s final analysis due to ambiguous artifact peaks in some species.

Hepenstrick also reported no more than four alleles at any locus in any of the polyploid taxa investigated, though many were of octoploid or higher ploidy. Further, the present analysis of Vpub9 amplicons led to the scoring of 12 different alleles in V. pedata

(though not more than six in any individual), whereas Hepenstrick reported no more than six alleles per taxon (V. riviniana). This underscores that application of SSR primers

109 within related taxa often generate different numbers and patterns of amplicons, characterized by species-specific alleles and allele size ranges (Peakall et al., 1998;

Yamamoto et al., 2001; Clauss et al., 2002; Zhao et al., 2008; Fehlberg & Fergeson,

2012), though this is not the case for all groups (Barbara et al., 2007). Still, the inclusion of artifacts, leading to scoring a higher number of alleles than previously reported should not be cause for concern considering the randomization of extraction, amplification, and scoring (including manual review). Under these protocols we expect any possible artifacts to be effectively neutral.

Patterns of diversity in V. pedata, and correspondence to previously-reported glacial refugia and genetic discontinuities

Patterns of genetic diversity and differentiation of V. pedata populations and regions were consistent with previous findings of genetic discontinuities, post-glacial expansion and proposed glacial refugia in other taxa (Soltis et al., 2006). The most prevalent pattern detected across analyses was the genetic distinction of eastern populations of the Appalachian Mountain continuum (NE and SE) and western/central populations (IO and DA). Results also suggested divergence of northern and southern populations along the Appalachian Mountain corridor, and the distinctiveness of populations from the DA relative to other interior populations. These regional groupings of populations were not apparent via cluster analysis without incorporation of Bayesian- suggested arrangements. While the results of AMOVA indicated variation between these four STRUCTURE-determined regions significantly contributed to their differentiation, the amount of inter-regional variation also varied by locus. The pattern of four clusters

110 was also supported by AMOVA results, indicating significant differences between diversity measures primarily associated with cluster and latitude, though a significant correlation was detected between these two factors.

Driftless Area populations

Diversity measures demonstrate genetic variation of V. pedata populations between some regions. Allelic patterns indicated that individuals from the DA region had the greatest variability in terms of number of alleles, measured over all loci, as well as the highest levels of heterozygosity across the different diversity measures. Also, this area had one of the two regionally-specific alleles detected in this study. Bayesian analysis supported the finding that this group was the most distinctive, having the highest average

Q value of any of the four regions projected. The clustering and segregation of the DA populations is in agreement with previous findings suggesting the upland plateau in this area served as a Pleistocene glacial refugium.

Many phylogeographic studies of North American flora have suggested a picture of post-glacial recolonization more complex than south-north migrations (Abbott et al.,

2000; Griffin & Barrett, 2004; McLachlan et al., 2005; Gonzales et al., 2008), with a growing body of evidence supporting the existence of a number of putative refugia across the continent, with some very near the glacial front (Kennedy & Walker, 2007; Pielou,

2008; Jaramillo-Correa et al., 2009). Geological evidence indicates that much of the upland area commonly called Driftless Area (also known as the Paleozoic Plateau) remained ice-free during the last glacial despite the southern expansion of the Laurentide

Ice Sheet to lower latitudes (Flint, 1957; Dyke & Prest, 1987; Eyles & Westgate, 1987).

111

It has been well established that many relict and rare plant (Hansen, 1939;

Hartley, 1966; Read, 1976; Cole et al., 1979; Glenn-Lewin, et al., 1984; Schwartz, 1985;

Pleasants & Wendel, 1989) and animal (Mossman & Hine, 1985; Maxwell & Young,

1998; Ross, 1999; Hyde & Staffen, 2012) species occur in this region of southwest

Wisconsin and adjacent Minnesota, Iowa, and Illinois. Numerous phylogeographic studies employing cytoplasmic (Breen et al., 2012; Li et al., 2013) and nuclear (Chung et al., 2004; Li et al., 2013) markers have found evidence that supports the hypothesis that this region within the upper Midwest served as glacial refugium for a diverse collection of taxa during the LGM. For example, Driftless Area populations of Monotropa hypopitys exhibited particularly high levels of genetic diversity and several private chloroplast haplotypes (Beatty & Provan, 2011). Similarly, populations of herbaceous

Smilax spp. from southwest Wisconsin uplands maintained all of the haplotypes detected in this study of representatives in eastern North America, one of which was unique (Li et al., 2013).

While supported by molecular evidence, the suggestion of a DA refugium is surprising, as this upland region is hypothesized to have experienced harsh, sub-arctic conditions and consisted of tundra or taiga-like vegetation during the LGM. Fossil pollen from the northern extent of the unglaciated DA suggests that the dominant plants were non-arboreal in the late Pleistocene (~12,000 YBP, Fries, 1962), and other records indicate prairie–deciduous forest persisted throughout the Holocene (~10,000 YBP,

Davis, 1977), and that American beech (Fagus grandifolia) colonized this region more than 2,000 years before latitudinally similar populations have been identified in eastern

112

North America (Webb, 1987). Phylogeographic examination of plant taxa, tolerant of the cold and dry conditions (Tremblay & Schoen, 1999; Eidesen et al., 2007; Ehrich et al.,

2008), that would have existed in this refugium and glacial boundaries may be one of the best approaches to gain insights into the extent of ice sheets at the LGM, as they are more likely to have maintained populations in northern refugia where conditions would exclude many temperate species. As V. pedata is today a widespread forb, capable of persisting through drought and extreme cold, as well as diverse edaphic conditions

(personal observation), its patterns of distribution and diversity described in the present study are valuable in supporting the DA refugial hypothesis.

It is also possible, however, that the violets from the DA region may have been recolonized, at least in part, from a southern refugium in the Ozark Plateau. Multiple studies have suggested this upland region was a stronghold for many taxa during the

Pleistocene, including species of fish (Strange & Burr, 1997), amphibian (Austin,

Lougheed & Boag, 2004), reptile (Walker & Avise, 1998; Howes et al., 2006), and woody plants (Johnson, 1988; Shaw & Small, 2005; Dane, 2009). Examination of the patterns of V. pedata population clustering between independent runs by STRUCTURE in the K=8 scenario (data not shown) suggested collections from the Ozark region (AR*,

MO*) formed a distinct cluster, while others grouped these populations with either populations within the DA region or southern lowland populations (MS).

113

Appalachian corridor populations

Principal coordinates, Bayesian, and NJ analysis suggest distinct separation of the

Appalachian SE and NE populations from the western IO and DA groups. These results suggest the Appalachian Mountains acted as a barrier between these groups, as all sampling of populations east of this mountain range (SE and NE groups) while sampling immediately to the west included populations of the IO group in the adjacent

Appalachian foothills (OH-1) and uplands of the Cumberland Plateau (KY-1). It is possible that the present clusters of V. pedata east (SE and NE) and west (IO) of the

Appalachian Mountains arose from two distinct, post-glacial migration routes from what may have been single, or multiple, southern refugia. As discussed in Soltis et al. (2006), the “Appalachian Mountain discontinuity” appears to have segregated the lineages of numerous organisms as they recolonized northward following glacial contraction.

Several plant species exhibit similar division in extant genetic patterns, including Atlantic white cedar (Chamaecyparis thyoides; Mylecraine et al., 2004), tulip tree (Liriodendron tulipifera; Parks et al., 1994; Sewell et al., 1996), American sweetgum (Liquidambar styraciflua; Morris et al., 2008), Virginia pine (Pinus virginiana, Parker et al., 1997), and

Smilax spp. (Li et al., 2012).

Among the Appalachian clusters of V. pedata, the data support the potential for several genetic patterns. According to the STRUCTURE analysis, the NE populations share the most genetic similarity (based on Q partition) with the SE group, while having a small fraction of total alleles detected (0.612) compared to all other clusters. Populations of the SE region demonstrated a significantly higher proportion of scored alleles (0.865),

114 and had higher scores for all diversity measures relative to NE populations. The trend of decreased genetic diversity in more recently founded (presumed northern) populations of

V. pedata would be expected in this Appalachian continuum, following both the long- distance and stepping-stone models of migration (Hewitt, 2000) from southern refugia.

This directional trend is in congruence with the patterns of diversity found in shorter- range expansion of organisms endemic to the southeast region of North America (for review, see Soltis et al., 2006) and is supported by phylogeographic analysis on other continents, postulating post-glacial recolonization patterns (Lewis & Crawford, 1995;

Petit et al., 1997; Taberlet et al., 1998; Hewitt, 2000; Petit et al., 2001).

An alternative hypothesis is that the NE populations reflect a northern glacial refugium, situated near the ice sheet edge. Examination of the mean values of diversity in the NE group showed comparable or slightly lower levels of almost all measures of allelic variability and heterozygosity to those of the SE group. Another distinctive character of the NE group is the presence of one regionally-unique marker at locus

Vpub69. Other phylogenetic evidence (high levels of genetic diversity, unique genotypes or haplotypes) suggests more northerly refugia in eastern North America, just south or east of the limits of the Laurentide Ice Sheet (Walter & Epperson, 2001; Jaramillo-Correa et al., 2004; Godbout et al., 2005; Beatty & Provan, 2010; de Lafontaine et al., 2010;

Godbout, Beaulieu, & Bousquet, 2010) and also areas along the Atlantic Coastal Plain

(Wall et al., 2010). Such suggestions may be bolstered by the proposed increases in available (refugial) land area along the Atlantic continental shelf following the significant recession of sea levels during the LGM of up to 100 meters below current levels (Donn et

115 al., 1962; Schlee, 2000). Further, indication that V. pedata persisted in the Driftless Area of the upper Midwest during the LGM supports the capacity of this species to exist in the harsh conditions at other points near the previous glacial boundary. The Q score (cluster identity) for the NE group (0.813) was comparable to that of the DA region, which had the highest value (0.826), suggesting differentiation. The populations of the SE and IO clusters had lower levels of identity, indicating higher levels of ad-mixture. Based on the phylogeographic literature of plants to date, it may be more likely that this north-south division within the Appalachian corridor is simply an indication of a loss in allelic diversity from long-distance migration from southern populations.

Interior lowland and Ozark Plateau populations

The IO cluster extended from gulf coastal plain into the Great Lakes region, pairing populations from east and west of the Mississippi River. This grouping was counterintuitive, based on multiple factors. The first being the strongly-evidenced effects this major river system has on a phylogenetic scale, and the second follows the regional morphological patterns of recognized botanical varieties. The contrast of these trends underscores the complexity of migration and gene flow in these grouped populations.

It has been suggested that the major river systems of eastern North America, including the Mississippi (Gornitz, 2009), and to a lesser extent the Ohio River drainage

(White et al., 2005), have caused genetic discontinuities between populations of many organisms (Soltis et al., 2006). As global temperatures began to warm following the

LGM, glacial melt was released at fluctuating rates from the receding Laurentide Ice

Sheet and Lake Agassiz over thousands of years. Uneven discharge patterns may have

116 caused periodic, catastrophic outbursts downstream, drastically scoring downstream valleys (Teller, 1995; Teller, Leverington, & Mann, 2002) and submersing lowland areas.

Such tumultuous releases of prodigious volumes of water and glacial drift could easily thwart recolonization efforts of species – plant or animal. Even if such violent outpourings had not occurred regularly or at all, the Mississippi is, and has historically been, one of the widest rivers in North America. It measures over a mile wide at its widest point. In recent floods, it has demonstrated the capacity to cover surrounding lowlands for over sixty miles (Daniel, 1996) - a force which could displace or extirpate proximal migrant populations. Some phylogeographic data strongly agree with proposals of such forceful, periodic events following the last glacial maximum (Near, Page, &

Mayden, 2001; Brant & Orti, 2003). Distinct genetic divergence was detected between loblolly pine (Pinus taeda) east and west of the Mississippi River (Al-Rabab’ah &

Williams, 2002). In the current study, though, Bayesian analysis of contemporary populations of V. pedata implied that the Mississippi and Ohio River Valleys may have played some role in defining boundaries of eastern and western populations, but have not caused historical limitation of regional gene flow. Genetic patterns in this area may reflect more ancient migrations or stochastic dispersal via other means.

The distinct paucity of V. pedata populations in the Ohio and southern Mississippi

River valleys may be due to the lack of suitable habitat available. While generally quite adaptable to soil type, V. pedata does not frequently occur in rich, silty soils, such as those deposited in lowland floodplains, nor does it persist in low-lying environments with heavily saturated soils. Though some of the populations of V. pedata in the current study

117 existed in fine, clay-like soils, these were typically steeply graded sites which allowed for excellent drainage. Also, heavy use of lowland, flood plains for agricultural and commercial purposes due to even grades and nutrient-rich alluvial deposits (Sparks,

1995; Craft & Cassey, 2000) may have extirpated local populations.

In populations east of the Mississippi, a decrease in diversity along a south-north gradient was not apparent, as would be generally expected following northward migration from southern refugia (Hewitt, 2000). Such a post-glacial expansion has been postulated in many organisms, and supported by molecular analysis (Soltis et al., 2006), but the reverse was true in the current study, as levels of allelic diversity and heterozygosity were notably lower in the MS population than almost all the MI populations. This may suggest refugia closer to the ice sheet than the present, southern distribution of this species suggests, or, potentially, gene flow from proximal, genetically rich, DA populations. The limited sampling in the southern portion of V. pedata range in the current study may have limited the ability of analyses to detect potential southern refugia and true post-glacial migration patterns.

Evidence for V. pedata habitat during the LGM based on paleontological records of associated species is ambiguous. Difficulty in determining abundance via association may be confounded due to the nature of prehistoric communities. Ancient floras surely exhibited species composition unlike current groups (e.g., having mixed compositions of species from divergent ecological communities or biomes), and such environments are well documented in the Southeast (Jackson & Williams, 2004). While it has been suggested that many contemporary, ecologically-associated pines may have maintained a

118 more southerly distribution during the LGM (Parker et al., 1997; Schmidtling et al., 2000;

Schmidtling, 2003), co-occurring oaks had broad range during the LGM and have shown to have highest levels of genetic diversity in northern populations, including unique genotypes (Magni et al., 2005). As many herbaceous taxa have not left significant markers in the fossil record, and few taxa associated with V. pedata been examined from a phylogeographic perspective, it is difficult to determine with which species this violet has been historically associated. Other, non-molecular approaches, including ecological niche modeling, may be of value in determining the historical distributions of V. pedata.

But due to its ability to persist in diverse environmental conditions, this may prove difficult as well.

Interestingly, both the STRUCTURE and NJ analyses did not place the MS population with the SE group, which did have a geographically proximal population, AL-

2. Some previous analyses of eastern North American taxa have found genetic discontinuities created by major river systems (for review, see Soltis et al., 2006). Along with the Apalachicola and Mississippi Rivers acting as barriers creating distinctive east- west genotypes, the Alabama River system (Mobile Basin) involving the Coosa, Cahaba, and Tombigbee tributaries has been suggested to be the reason for population division in many taxa including birds (Gill et al., 1993; 1999; McKay, 2009), fish (Bermingham &

Avise, 1986) and reptiles (Lawson, 1987; Jackson & Austin, 2010), though many of these divisions are estimated to predate the Pleistocene (~1 MYBP). Trillium cuneatum populations in the southeast demonstrated a similar pattern to that of the current study, demonstrating an east-west separation of haplotypes along the Tombigbee River

119

(Gonzales et al., 2008; Wallace & Doffitt, 2013). Other major river discontinuities may be important in determining the patterns of genetic and geographic distribution of V. pedata was mediated by bio- and geo-paleontological events, and could be aided by further sampling populations in the southeast and gulf states.

Of the four Bayesian-defined clusters, the IO group maintains the highest number of total number of alleles present (91.6%). Examination of populations herein indicate higher levels of allelic diversity (A, A’) and heterozygosity (Ho, H’, H”) in Ozark Plateau

(AR / MO) populations than populations east of the Mississippi. This suggests that populations of V. pedata may have found refuge here during previous glacial cycles.

As the southern highlands of Missouri and northern Arkansas were not glaciated in the Quaternary Period (Federal Writers' Project, 1941), the Ozark Plateau has acted as a refugium for many species of fish (Strange & Burr, 1997), amphibians (Austin,

Lougheed, & Boag, 2004), reptiles (Walker & Avise, 1998; Howes, et al., 2006), and woody plants (Shaw & Small, 2005). Some species that sought shelter here during the

LGM have since dispersed, including many species of darters and minnows (Berendzen,

Dugan, & Gamble, 2010), which show post-glacial advancement into the upper Midwest.

Antithetically, Castanea pumila var. ozarkensis, the Ozark chinkapin, is a woody species endemic to this upland region that exemplifies restricted dispersal and local adaptation as it displays distinctive foliage from conspecifics from east of the Mississippi (Johnson,

1988; Dane, 2009), purportedly due to long-term isolation. This regional isolation of

Arkansas and Missouri may explain the locally common V. pedata var. pedata, the bicolored morphotype of the species, which is most common east of these highlands to

120 the Cumberland Plateau, but is rare to the north and south. It co-occurs here with the solid-colored form, var. lineariloba, though intermediate forms have not been described, presenting the hypothesis that these two varieties may be separated by a simple (single gene) mutation.

The present analysis did not find observed flower morphology to be significantly correlated with any measure of population- or cluster-level variation. Viola pedata var. pedata, the bicolorous form of V. pedata, has a distinctive range, most abundant and seemingly centered in the Ozark Plateau highlands (OP) and extending east toward the

Cumberland Plateau. In the OP region, the bicolorous form appeared in every population observed in the present study, in most cases in equal number to the concolorous, var. lineariloba. This finding is consistent with previous studies of the region (Carroll &

Goldman, 1994; Morehouse & Carroll, 2001). While the bicolor form has been reported from other localities including the DA region of Wisconsin (Ted Cochrane, personal communication, May 15, 2010) the pine barrens of New Jersey (Renee Brecht, Associate

Director of Citizens United to Protect the Maurice River & Its Tributaries, personal communication, April 5, 2011) and western Michigan (Ballard, personal communication,

November 13, 2013), it is exceeding rare. Previous analysis of the European species V. suavis M. Bieb. was able to separate conspecifics based on flower color (purple / white), but differentiation was attributed to long-term isolation and survival in distinct glacial refugia (Mereda et al., 2008). The occurrence of these white morphotypes from distinct lineages in two different locations suggested related genetic origins.

121

Although local abundance or co-occurrence of the sampled varieties was not significantly associated with any specific marker of measure in this study, previous research suggests flower color influences pollinator preference in V. pedata. In a study conducted within the OP region (Missouri) investigating pollinator visitation, results indicated that although pollinators spent more time, on average, visiting bicolouous flowers than concolorous types, this higher rate of visitation was correlated to their relative abundance (Morehouse & Carroll, 2001). Though this finding would suggest bicolored forms had the potential for higher fruit set, it has been shown that there is a trend for higher seed production with increasing proportions of concolorous conspecifics within a population (Carroll & Goldman, 1994).

Some discrepancies were apparent between the Bayesian and NJ clustering patterns. Within the Appalachian corridor, two populations from the northern extent of the SE cluster (VA-1 and VA-2), were assigned to the IO cluster by STRUCTURE. As these populations were on the eastern side of the Appalachian Mountains at this point, it would not be expected that they were derived from the interior lowlands or Ozark

Plateau. It is of note though, that the VA-2 population had significant proportion of association with the SE region (Q=0.282) and that NJ analysis clustered all Virginia populations together. It is possible that this finding is a function of missing data from informative loci. These two populations were both in the top four in this study for missing data, in terms of markers unscored (of 97 markers, VA-2, 12.04%; VA-1,

9.04%). In more westerly populations, one IO outlier (AR-2) was associated by

STRUCTURE with the SE cluster. This cannot be explained by missing data, as this

122 population was missing an average of less than 1.2% of the total alleles scored per individual. Examination of the Bayesian output indicated a mean probability of 0.49 (Q) for AR-2 to belong to the IO group, while the cluster average was 0.708. The AR-2 population demonstrated high levels of ad-mixture, sharing high proportions of identity with both the SE (0.222) and DA (0.277) clusters. This may suggest a level of gene flow between regions, or may also be an artifact of limited sampling from southern populations.

Interspecific hybridization may also play a role in the differences seen between and within regions. Because of the intersectional hybrid origin of V. pedata, and overlapping or similar distributions of species of common ancestry and cytology (e.g. V. sagittata), it is possible that interspecific hybridization introduced unique alleles to local populations of V. pedata. It is possible that such hybridization events led to the addition of rare alleles to populations examined in this study (within DA and NE clusters) or populations discordant with intuitive, regional clustering (VA-1, VA-2, AR-2). While V. pedata has not been reported as being able to hybridize with other species, it does share a common ancestor with many contemporary taxa, as evidenced by congruity in glucose-6- phosphate isomerase homologs (Marcussen et al., 2012). Further, Russell (1965) suggested that V. sororia (which also maintains 2n=54) has the capacity to hybridize with any acaulescent blue violet. Neither the discovery of interspecific hybrids, nor successful hand pollinations, has been reported to verify the ability of V. pedata to produce fertile, hybrid progeny, though preliminary results in our lab indicate crossings of V. pedata and

V. sororia can produce viable seed.

123

Sampling limitations, future studies

The populations of V. pedata sampled for this study were not evenly distributed.

Biases in sampling, including higher-density sampling in the upper Midwest and the

Appalachians may have overestimated their diversity, relative to interior and gulf lowlands. While current distribution and genetic measures of V. pedata suggest otherwise, it is possible that the post- glacial expansion from refugia was more complex than suggested here, and that some contemporary populations of this species thus show reductions in genetic diversity relative to now-extinct refugial populations. Even if all present-day populations are derived migrants, both long-distance and stepping-stone models of migration predict a loss of genetic diversity in the most recently founded (i.e., northern) populations (Hewitt, 2000). Increased sampling of populations from the gulf lowlands and immediately east/west of the Appalachian Mountains may help clarify the patterns seen in the present study. Further, examination of cytoplasmic sequence (i.e. chloroplast haplotypes) could be of use in determining levels of admixture, and routes of dispersal and migration.

124

Tables and Figures

Figure 3.1. County-level distribution of Viola pedata

County-level information of Viola pedata, adapted from historical collection information by the United Stated Department of Agriculture’s Germplasm Research Information Network (USDA-GRIN) and Biota of North America Program (BONAP) databases, as well as personal observation.

125

Figure 3.2. Examples of seasonal variation in Viola pedata leaf morphology

Early and late-season leaf samples of Viola pedata (images by Dan Tenaglia, www.missouriplants.com, reproduced with permission).

126

Figure 3.3. Flower variants of Viola pedata

Variation of flower coloration in Viola pedata: (A) Viola pedata var. lineariloba, (B) Viola pedata var. pedata – typical form, (C) var. pedata – incomplete superior coloration, (D and E) var. pedata – lateral petal coloration, and (F) var. alba (images A, D, F by Dan Tenaglia, www.missouriplants.com, reproduced with permission).

127

Figure 3.4. Distribution map of Viola pedata var. pedata and var. lineariloba

Estimated distribution of concolorous Viola pedata var. lineariloba (blue) and bicolorous Viola pedata var. pedata (red) based on personal observation.

128

Table 3.1. Locality information of Viola pedata collections

Population ID Individuals analyzed Abundance Locality Longitude and Latitude Elevation (ft.) CT-5 10 200-300 CT, Windham Co. 41' 40"N 072' 7"W 153 CT-4 11 50-100 CT, Tolland Co. 41' 45"N 072' 11"W 255 NH-1 9 100-200 NH, Hillsborough Co. 42' 46"N 071' 22"W 423 CT-8 12 300-400 CT, Windham Co. 41' 41"N 071' 55"W 196 NJ-1 11 100-200 NJ, Burlington Co. 39' 52"N 074' 33"W 126 CT-69 11 25-50 CT, Hartford Co. 41' 45"N 072' 51"W 229 CT-2 12 50-100 CT, Tolland Co. 41' 56"N 072' 21"W 737 MA-2 9 75-100 MA, Barnstable Co. 41' 38"N 070' 34"W 92 NJ-2 12 300-400 NJ, Cumberland Co. 39' 20"N 075' 9"W 47 RI-1 12 100-200 RI, Kent Co. 41' 41"N 071' 45"W 443 MA-1 10 300-400 MA, Hampden Co. 42' 9"N 072' 43"W 236 GA-1 9 300-400 GA, Rabun Co. 34' 58"N 083' 11"W 2306 AL-2 9 25-50 AL, Bibb Co. 33' 5"N 087' 2"W 570 GA-2 13 1000-1500 GA, Habersham Co. 34' 43"N 083' 24"W 1427 Eco-Art 2 - cultivated variety - - NC-2 9 200-300 NC, Buncombe Co. 35' 33"N 082' 30"W 3235 NC-3 11 200-300 NC, Jackson Co. 35' 7"N 082' 58"W 3186 NC-1 11 400-500 NC, Ashe Co. 36' 19"N 081' 22"W 2799 VA-3 8 100-200 VA, Bedford Co. 37' 32"N 079' 31"W 3871 AR-1 9 600-700 AR, Pope Co. 35' 58"N 092' 11"W 892 VA-4 12 300-400 VA, Botetourt Co. 37' 24"N 079' 49"W 1932 nMI-2 16 2000+ MI, Crawford Co. 44' 39"N 084' 38"W 1183 VA-1 7 100-200 VA, Rockbridge Co. 37' 44"N 079' 19"W 1585 IN-1 12 300-400 IN, Jasper Co. 41' 11"N 086' 57"W 666 sMI-2 10 300-400 MI, Muskegon Co. 43' 13"N 086' 7"W 660 nMI-3 14 750-1000 MI, Alcona Co. 44' 29"N 083' 41"W 843 nMI-1 12 100-200 MI, Wexford Co. 44' 13"N 085' 48"W 922 sMI-3 12 1500-2000 MI, Allegan Co. 42' 36"N 086' 1"W 653 OH-1 10 100-200 OH, Scioto Co. 38' 42"N 081' 15"W 1056 MO-3 8 100-200 MO, Pulaski Co. 37' 47"N 092' 13"W 1082 VA-2 10 800-1000 VA, Amherst Co. 37' 39"N 079' 20"W 1467 sMI-1 13 500-750 MI, Washtenaw Co. 42' 25"N 083' 59"W 900 MS-1 9 50-100 MS, Lauderdale Co. 32' 21"N 088' 40"W 410 KY-1 19 1500-2000 KY, McCreary Co. 36' 45"N 084' 29"W 1270 MO-1 10 200-300 MO, Oregon Co. 36' 39"N 091' 13"W 505 AR-2 10 400-500 AR, Johnson Co. 35' 44"N 093' 25"W 1972 WI-5 13 50000+ WI, Sauk Co. 43' 12"N 090' 3"W 771 WI-4 9 500-600 WI, Richland Co. 43' 13"N 090' 17"W 705 WI-3 10 300-400 WI, Trempealeau Co. 44' 37"N 091' 15"W 1106 WI-6 10 300-400 WI, Dane Co. 43' 8"N 089' 46"W 889 WI-1 10 1500-2000 WI, Crawford Co. 43' 13"N 090' 52"W 755 MO-2 12 1000-1500 MO, Adair Co. 40' 6"N 092' 37"W 942 WI-2 10 100-200 MN, Houston Co. 43' 46"N 091' 26"W 1430

129

Figure 3.5. Example electropherogram of microsatellite data for Viola pedata suggesting higher-order ploidy

Electropherogram depicted in GeneScan v. 3.7 software, indicating 8 alleles (green peaks labeled with size, in base pairs) scored for a single Viola pedata sample (acn. 11.642, population CT-69) at the Vpub57 locus. The number of allelic peaks shown here suggests a high-order polyploid status of this species.

130

Table 3.2. Descriptive statistics and missing data of microsatellite markers in Viola pedata

Size sm Size lg Range individuals % total missing of total Locus (bp) (bp) (bp) alleles/locus alleles/individual missing data individuals

Vpub11 132 169 37 16 1-8 26 5.68%

Vpub16 191 206 15 7 1-4 34 7.42%

Vpub21 115 172 57 20 1-8 7 1.53%

Vpub57 90 148 58 16 1-8 6 1.31%

Vpub60 183 215 32 10 1-5 35 7.64%

Vpub7 193 219 26 11 1-6 42 9.17%

Vpub9 216 243 27 12 1-6 6 1.31%

Vpub69 140 156 16 5 1-3 5 1.09%

TOTALS 157.5 191 33.5 97 1-6 161 4.39%

131

Table 3.3. Mean differentiation, diversity, and heterozygosity variables of Viola pedata populations

POP ID X X' A A' Ho H' H'' Np1 Np2 Np3 Np4 Np5 Np6 Np7 Np8 Np CT-5 22.45 25.876 46 5.75 0.73 3.08 0 5 5 11 11 4 2 9 1 6 CT-4 25.5 23.608 45 5.63 0.65 2.81 0 6 4 10 7 1 3 8 1 5 NH-1 26 26.323 42 5.25 0.75 3.25 0 8 4 9 5 4 1 4 1 4.5 CT-8 24.6 25.254 47 5.88 0.67 2.77 0 6 5 11 9 6 3 9 2 6.38 NJ-1 22.15 25.251 43 5.38 0.68 3.00 0 5 5 9 8 5 2 9 1 5.5 CT-69 25.64 26.575 44 5.5 0.69 3.20 0 4 5 11 10 6 2 9 1 6 CT-2 24.42 26.227 46 5.75 0.74 3.19 0 5 5 12 10 7 3 8 2 6.5 MA-2 22.78 23.934 40 5 0.69 2.85 0 3 3 7 7 5 2 6 2 4.38 NJ-2 24 25.291 42 5.25 0.70 3.05 0 2 3 10 6 7 4 9 1 5.25 RI-1 22.92 22.917 47 5.88 0.69 2.86 0 8 4 12 9 5 4 11 2 6.88 MA-1 18.7 20.061 33 4.13 0.66 2.34 0 4 3 5 9 3 2 7 1 4.25 GA-1 28.44 29.333 52 6.5 0.82 3.56 0 6 5 9 9 2 4 7 1 5.38 AL-2 25.44 28.169 45 5.63 0.71 3.18 0 4 6 9 4 6 5 2 1 4.63 GA-2 31.15 31.416 56 7 0.84 3.89 0 8 9 12 10 6 3 7 1 7 EcoArt 34 34 34 4.25 0.75 4.25 0 1 1 1 1 1 1 1 1 1 NC-2 27.22 27.623 47 5.88 0.79 3.40 0 5 6 8 6 8 3 5 1 5.25 NC-3 28.55 28.545 54 6.75 0.76 3.57 0 7 7 10 8 7 2 8 1 6.25 NC-1 26.18 27.982 51 6.38 0.70 3.27 0 9 6 10 10 7 3 7 1 6.63 VA-3 23.63 24.707 48 6 0.72 2.95 0 5 4 8 5 4 4 7 1 4.75 AR-1 26 27.342 48 6 0.85 3.25 0 5 4 7 5 8 5 7 3 5.5 VA-4 23.67 25.241 53 6.63 0.71 2.96 0 8 5 6 11 10 7 8 1 7 nMI-2 21.14 23.279 42 5.25 0.70 2.72 0 8 5 8 9 4 6 5 1 5.75 VA-1 21.75 22.034 43 5.38 0.63 2.64 0 5 2 6 4 4 3 5 1 3.75 IN-1 21.5 22.694 39 4.88 0.68 2.69 0 5 2 5 6 5 5 4 1 4.13 sMI-2 22.9 23.929 40 5 0.74 2.86 0 6 4 7 6 4 4 4 1 4.5 nMI-3 20.86 21.99 42 5.25 0.88 2.61 0 5 4 6 9 6 7 6 1 5.5 nMI-1 23 21.691 43 5.38 0.65 2.88 0 10 5 7 6 6 6 7 1 6 sMI-3 22.17 22.167 44 5.5 0.71 2.77 0 9 7 8 9 9 7 6 1 7 OH-1 23.6 24.671 42 5.25 0.80 2.95 0 7 2 8 4 9 5 5 1 5.13 MO-3 24.63 25.553 43 5.38 0.83 3.08 0.13 5 2 7 6 6 5 5 2 4.75

Key: X, the average number of alleles per individual per population, X‘ the adjusted number of alleles per individual per population, A, mean number of different alleles per population; A’, mean number of different alleles per population per locus; Ho, mean proportion of heterozygous loci per individual per population; H', mean number of alleles per individual per locus per individual per population; H”, mean proportion of individuals heterozygous at all loci per population; NpL, mean number of different allelic phenotypes per locus (L) per individual per population; Np, mean number of different allelic phenotypes averaged over all loci.

continued 132

Table 3. continued

POP ID X X' A A' Ho H' H'' Np1 Np2 Np3 Np4 Np5 Np6 Np7 Np8 Np VA-2 20.1 22.616 44 5.5 0.65 2.51 0 7 2 7 8 6 3 6 1 5 sMI-1 21.62 23.136 41 5.13 0.71 2.70 0 7 5 7 10 6 4 5 1 5.63 MS-1 23.11 24.52 37 4.63 0.74 2.89 0 7 4 9 2 3 6 3 1 4.38 KY-1 21 22.28 51 6.38 0.68 2.63 0 12 7 11 11 15 7 6 1 8.75 MO-1 25.9 27.983 55 6.88 0.76 3.24 0.1 7 4 8 6 5 7 9 2 6 AR-2 24.4 24.628 48 6 0.83 3.05 0.3 7 3 8 4 7 6 7 3 5.63 WI-5 25.69 26.692 54 6.75 0.84 3.21 0.38 7 5 10 11 7 7 6 2 6.88 WI-4 27.78 27.111 49 6.13 0.85 3.47 0.44 6 4 6 8 6 7 6 3 5.75 WI-3 26.2 26.393 47 5.88 0.91 3.28 0.3 7 7 7 9 7 5 5 3 6.25 WI-6 29.2 29.232 60 7.5 0.88 3.65 0.3 7 4 9 8 7 6 7 3 6.38 WI-1 25.8 26.282 57 7.13 0.80 3.23 0.2 7 6 9 19 8 5 8 3 8.13 MO-2 23.08 23.208 51 6.38 0.85 2.89 0.33 8 4 5 8 10 8 10 3 7 WI-2 27.2 27.571 55 6.88 0.81 3.40 0.3 7 5 9 7 9 6 8 3 6.75

133

Table 3.4. Significant results from Pearson's correlation analysis of geographic and molecular variables in Viola pedata

Variable X' A A' Ho

Latitude r -0.38 -0.152 -0.152 0.037

t -2.601 -0.974 -0.974 0.231

Longitude r 0.209 0.394 0.394 0.627

t 1.354 2.711 2.711 5.086

Elevation r 0.279 0.423 0.423 0.249

t 1.836 2.95 2.95 1.623

STRUCTURE cluster r 0 0.34 0.34 0.525

t -0.002 2.288 2.288 3.902

Values in red, bold text indicates significant t scores (t(41) > 2.021, p < 0.05) for two- tailed Pearson’s correlation analysis.

134

A)

0.10

0.08

0.06

0.04 FST FST values 0.02 y = 2E-05x + 0.0264 r² = 0.2531 0.00 0 500 1000 1500 2000 2500 Geographic Distance

B)

1.0 y = 6E-05x + 0.1119 r² = 0.0257

0.8

0.6

0.4 PhiPT PhiPT values 0.2

0.0 0 500 1000 1500 2000 2500 Geographic Distance

Figure 3.6. Mantel test scatterplots of relationships between pair-wise geographic distance and fixation indices in Viola pedata accessions

Mantel test scatter plots demonstrating significant associations of geographic distance between 458 samples of Viola pedata from 42 naturally-occurring populations with (A) FST (r=0.503, p<0.01) and (B) PhiPT (r=0.160, p<0.01). Accessions of ‘Eco Artist’s Palette’ are not included, and the origins of the cultivar are unpublished.

135

Figure 3.7. Consensus neighbor-joining tree of 42 Viola pedata populations

Majority-rule consensus neighbor-joining tree of Nei’s (1972) genetic distance between 42 naturally-occurring populations of Viola pedata, based on 1,000 bootstrap replicates. Bootstrap support > 50% is shown on branches as number of bootstraps/1000. The red, hashed line suggests a divide between eastern from western populations, corresponding to eastern NE / SE and western IO / DA clusters, determined by STRUCTURE. See Table 3.1 for detailed population location information.

136

137

Figure 3.8. Principal coordinates analysis scatterplots of Viola pedata accessions labeled by STRUCTURE cluster

The first three coordinates of principal coordinates analysis describing microsatellite data from eight loci in 458 accessions of Viola pedata collected from 42 wild populations and including one cultivated variety. Dots represent individual accessions, and colors represent the four STRUCTURE suggested clusters of populations: NE (green), SE (red), IO (yellow), WI (blue). The black, hashed line suggests a break between the eastern and western populations.

137

A)

B)

Figure 3.9. STRUCTURE Harvester results indicating likely K values for 43 Viola pedata populations

Population clustering analyses suggest significant population structure among Viola pedata populations: A) Likelihood scores for each value of K genetic clusters from STRUCTURE (Pritchard et al., 2000), B) ΔK scores for each value of K genetic clusters following Evanno et al. (2005). Figures were generated using STRUCTURE Harvester (Earl et al., 2011) and are based on a burn-in of 105 followed by 5x105 iterations, testing K values 1 to 10, with 10 replicate runs for each value. 138

139

Figure 3.10. STRUCTURE barplots of 43 Viola pedata populations arranged by STRUCTURE cluster

Barplots depicting STRUCTURE results of (K=4) clusters of 458 samples from 42 populations of Viola pedata and one cultivated variety. Populations are grouped by cluster (NE, northeast; SE, southeast; IO, interior lowlands and Ozark Plateau; DA, Driftless Area), in order (left to right) of highest mean Q score (cluster identity proportion) within respective clusters. 139

Figure 3.11. Geographic representation of STRUCTURE (K=4) clustering of Viola pedata populations

Populations of Viola pedata sampled in the present study are represented by colored circles, indicating the four STRUCTURE clusters: DA (blue), IO, (yellow), SE (red), and NE (green). Estimation of species’ historical distribution is depicted by gray shading and the maximum extent of the last glacial maximum is depicted by light blue shading.

140

Table 3.5. ANOVA results of cluster-level molecular variables in Viola pedata cluster X X' A A' Ho H' H'' Np1 Np2 Np3 Np4 Np5 Np6 Np7 Np8 Np NE 23.56 24.67 43.18 5.40 0.70 2.94 0.00 5.09 4.18 9.73 8.27 4.82 2.55 8.09 1.36 5.51 SE 22.51 23.54 48.80 6.10 0.76 3.43 0.00 5.80 5.30 8.00 6.90 5.90 3.70 5.90 1.20 5.34 IO 27.43 28.44 43.64 5.46 0.72 2.80 0.03 7.29 4.00 7.50 6.71 6.36 5.43 5.57 1.21 5.51 DA 26.42 26.64 52.00 6.50 0.85 3.27 0.30 6.75 4.63 7.75 9.50 7.50 6.13 6.88 2.75 6.48

Results of one-way ANOVAs comparing mean molecular measures between STRUCTURE determined regions (K=4) of Viola pedata. Cluster names are abbreviated: northern Appalachian Mountains (NE), southern Appalachian Mountains (SE), the Driftless Area (DA), and interior lowlands and Ozark Plateau (IO). Values in red, bold font indicate significance at the p < 0.05 level. Detailed ANOVA results and mean- separation tests (Tukey’s HSD) are given in Appendix D.

141

Table 3.6. AMOVA results by locus in microsatellite analysis of Viola pedata

Treatment WP AP AR WP: PhiPT p AP: PhiPR p AR: PhiRT p

locus 1 % var. 99 1 0 0.003 0.411 0.01 0.275 -0.008 0.898

df 415 39 3

MS 0.056 0.062 0.014 locus 2 % var. 84 10 6 0.164 0.01 0.107 0.01 0.064 0.01 df 415 39 3 MS 0.301 0.682 3.306 locus 3 % var. 98 2 0 0.012 0.258 0.022 0.15 -0.011 0.987

df 415 39 3

MS 0.06 0.074 0.002

locus 4 % var. 83 11 6 0.175 0.01 0.12 0.01 0.062 0.01 df 415 39 3 MS 0.137 0.337 1.522 locus 5 % var. 84 15 1 0.16 0.01 0.15 0.01 0.012 0.08

df 415 39 3

MS 0.167 0.478 0.766

locus 6 % var. 99 1 0 0.015 0.227 0.013 0.267 0.002 0.293 df 415 39 3 MS 0.242 0.275 0.342 locus 7 % var. 89 11 0 0.096 0.01 0.107 0.01 -0.012 0.995

df 415 39 3

MS 0.093 0.21 0.073

locus 8 % var. 84 6 10 0.155 0.01 0.065 0.01 0.096 0.01 df 415 39 3 MS 0.114 0.197 1.666 TOTAL (all % var. 88 8 4 0.118 0.01 0.086 0.01 0.035 0.01 loci) df 415 39 3

MS 1.169 2.311 7.692

Results of hierarchical AMOVA, attributing variation to within population, among population, and among region (STRUCTURE clusters) sources. Values in red, bold font indicate significant values of Phi at the p < 0.05 level. Key: AR = Estimate (Est.) of variation (Var.) Among Regions; AP = Est. Var. Among Populations; WP = Est. Var. Within Populations.

142

Chapter 4. Sequence-related amplified polymorphism (SRAP) markers: A potential resource for studies in plant molecular biology

Abstract In the last few decades, many investigations in the fields of plant biology have employed selectively-neutral, multilocus, dominant markers such as ISSR, RAPD, and

AFLP to address hypotheses at lower taxonomic levels. More recently, sequence-related amplified polymorphism (SRAP) markers have been developed amplifying coding regions of DNA with primers targeting open reading frames. These markers have proven to be robust and highly variable, on par with AFLP, and are attained through a process significantly less technically-demanding. Since their inception, SRAP markers have been used primarily for agronomic and horticultural purposes, developing quantitative trait loci in populations of advanced hybrids and assessing genetic diversity of large germplasm collections. Here, we suggest that SRAP markers should be more broadly appreciated, and used in research addressing hypotheses in plant systematics, biogeography, conservation, ecology, and beyond. We provide an overview of the SRAP literature to date, review descriptive statistics of SRAP markers in a subset of 188 publications, and present relevant case studies to demonstrate the applicability of SRAP markers to the diverse field of plant biology. Results of these selected works indicate that

SRAP markers have the potential to complement and enhance the current suite of molecular tools in a diversity of fields, by providing an easy-to-use, highly variable

143 marker with inherent biological significance. In conjunction with next-generation sequencing techniques, genetic evaluation via SRAP may play a major role in future studies of plant differentiation.

Introduction

Ecological and systematic studies often depend on the use of molecular tools to address questions regarding genetic relatedness among individuals, population structure, and phylogenetic relationships, as well as mapping loci and tracking patterns of quantitative traits. Molecular systematic studies of plants have frequently depended on chloroplast DNA (cpDNA) sequences as a primary source of data for investigations in delimitation of high-order taxonomic groups (for reviews, see Olmstead & Palmer, 1994;

Petit et al., 2005). Some universal primers for noncoding regions have been developed

(i.e., trnL-trnF and trnK/matK) and widely applied in such studies. More recently, primers for previously unexploited, fast-evolving cpDNA regions (for reviews see Small et al., 1998; Shaw et al., 2005; Shaw et al., 2007), and nuclear loci have provided increased resolution for taxonomic and systematic investigations. PCR-based dominant marker systems, such as random amplified polymorphic DNA (RAPD; Williams et al.,

1990), inter-simple sequence repeat (ISSR; Zietkiewicz et al., 1994), and amplified fragment length polymorphism (AFLP; Vos et al., 1995), have shown systematic utility in discerning between higher-order relationships as well (Brauner et al., 1992; Federici et al., 1998; Nair et al., 1999; Smissen et al., 2003; Pang et al., 2007) though are typically used for investigating more shallow taxonomic levels of variation. Since the 1990’s,

144 these multi-locus marker systems have been used to estimate genetic variation in plants because they produce numerous amplicons and do not require a priori sequence information for molecular characterization (Fang et al., 1997; Wolfe, Xiang, & Kephart,

1998a; Aggarwal et al., 1999; Culley & Wolfe, 2001; Koopman et al., 2008).

Despite the potential for these dominant markers to produce useful data, some researchers have been hesitant to utilize them in many investigations due to technical limitations. In the case of RAPDs, many studies have shown inconsistencies in data replication (Jones et al., 1997; Perez, 1998). Though generally thought to produce more repeatable results, ISSRs have been shown to be less productive in terms of polymorphisms detected for some primer combinations (Meksem & Kahl, 2005), while the AFLP technique involves numerous, time-demanding steps (Vos et al., 1995).

Further drawbacks of these methods include the ambiguity of fragment homology, or whether banding patterns represent hetero- or homozygous loci. The lack of identified homology between bands of identical length generated by the same primer has also been reason for concern (Rieseberg, 1996; Vekemans et al., 2002; Simmons et al., 2007;

Caballero & Quesada, 2010), but there is evidence that fragments of the same calculated size originate from the same locus and are homologous, at least at the intraspecific level

(Koopman 2005; Gort et al., 2006; Ipek et al., 2006). Therefore, the complications of homoplasy should not be a significant factor in dominant molecular analyses, if restricted to one species, or a group of taxa that are closely related. Although use of these molecular techniques to resolve phylogenetic relationships at higher taxonomic levels may be considered controversial, corrections for the presumably high proportion of

145 homoplastic fragments have been developed to reduce the rate of fragment incursion

(Caballero et al., 2008; Gort et al., 2009; Caballero et al., 2010; Paris et al., 2010).

Though no single technique is universally applicable (for reviews see Maheswaran, 2004;

Arif et al., 2010), and all molecular marker approaches have inherent strengths and weaknesses, the selection of a particular method ultimately depends on the research question at hand and the degree of resolution needed.

A more recently developed, dominant marker technique, sequence-related amplified polymorphism (SRAP; Li & Quiros, 2001) PCR, is simple, inexpensive, and effective for producing genome-wide fragments with high reproducibility and versatility.

These markers were originally developed for gene tagging in Brassica oleracea to specifically target coding regions of the genome with ambiguous primers developed to target GC-rich exons (forward primers) and AT-rich promoters, introns, and spacers

(reverse primers). The primers are 17 or 18 nucleotides long and consist of core sequences, which are 13 to 14 bases long, with the first 10 or 11 bases starting at the

5′ end, are “filler” sequences, maintaining no specific constitution. These are followed by the sequence CCGG (forward) or AATT (reverse). This core is followed by three selective nucleotides (random) at the 3′ end (see Li & Quiros, 2001, or Budak et al.,

2004a, for sequence details). PCR amplification techniques (specifically termperature) have been modified by many to suit specific research needs (for review, see Aneja et al.,

2012), but the original protocol was described as a two-phase process: 1 min cycles including a denaturation (94º C), five cycles at a low annealing temperature to generate

146 template fragments (~35º C) followed by an elongation step (72º C). This template then amplified over 35 similar cycles, employing a higher annealing temperature (~50º C).

Like other dominant markers, SRAPs have demonstrated the ability to elucidate genetic variation at a variety of taxonomic levels (Uzun et al., 2009), but are often used for fine-scale analyses of populations of inter- and intra-specific hybrids (Hale et al.,

2007; Liu et al., 2008). Analysis of SRAP data has frequently been employed for the construction of linkage maps (LM; Lin et al., 2003; Yeboah et al., 2007; Lin et al., 2009;

Levi et al., 2011; Gao et al., 2012) and identification of quantitative trait loci (QTL; Chen et al., 2007; Yuan et al., 2008; Zhang et al., 2009). Consequently, this system has been valuable for the improvement of agronomic crops (Zhang et al., 2005; Brown et al., 2007;

Zhao et al., 2010; Wright & Kelly, 2011). Many comparative studies have found SRAP markers provide comparable levels of variation to AFLP markers, but with significantly less technical effort and cost for similar levels of band-pattern variability and reproducibility (Li & Quiros, 2001; Levi & Thomas, 2007; Liu et al., 2007; Wang et al.,

2007; Lou et al., 2010). These factors highlight the value of SRAP markers for examination of previously unexplored, non-model systems, and also for experimental ventures in developing countries. Limitations of SRAP marker application have not yet been described, as these markers are relatively new and their use is still in its early stages.

Innovations in molecular marker systems are employed by many branches of the plant sciences to elucidate diversity, including both applied and academic pursuits. The distinction between approaches often does not lie in the systems examined, but in the nature of hypotheses addressed on the processes of evolution - natural versus directed.

147

Over the past decade, application of SRAP markers has gained momentum, especially in the applied plant sciences (for reviews, see Aneja et al., 2012; Wang et al., 2012). Based on the rapidly growing body of literature, we propose that the SRAP marker system could, and should, be applied to the fields of systematics, ecology, and evolutionary biology. Here, we review the potential for SRAP markers in these areas of research.

Methods

In a review of the current literature, using all available databases, we found more than 350 peer-reviewed references to SRAP application in the biological sciences. These were predominantly associated with plant breeding, and demonstrated very little exploration of utility beyond this field. As many of these articles were directed at developing LM and QTL for agronomic crops, they presented little information on general statistics of amplified SRAP products and measures of diversity within accessions. Examination of these works revealed that only a subset (n=188; Appendix E) provided descriptive data of SRAP markers of value to the broader biology community.

We recorded information on taxon, number of individuals, number of primer pairs, number of fragments generated per primer pair, number of polymorphic fragments generated per primer pair, and polymorphic fragment rate. Not all data necessary for calculation of these statistics were presented in all references. In some instances, the number of fragments generated was not explicitly described as polymorphic or otherwise.

In these cases, we report this value as the total number of bands, as to not overestimate

148 the true rate of polymorphism, though this practice could underestimate the total number of bands generated per primer.

Results and Discussion

Since the introduction of SRAP markers in 2001, the application of SRAP markers has increased dramatically, especially in the last few years (Figure 4.1). Review of the taxa evaluated and hypotheses addressed during this recent spike in SRAP marker employment indicates a rise in awareness of the utility of these markers and a broadening of diversity in organisms examined. There was considerable variation in values scored for all manuscripts examined (Appendix E). The mean number of polymorphisms was

12.08 per primer combination of the 174 studies reporting this statistic, and of these, 76

(43.7%) reported more than 10 polymorphisms per primer. The rate of polymorphism averaged 69.69% across studies, but of the 132 citations explicitly reporting polymorphic rates, 32 (24.2%) indicated scored bands were over 90% polymorphic, and 64 (48.5%) studies noted over 80%.

To further assess and demonstrate the general applicability and value of SRAP, we have highlighted case studies that exemplify the potential of these molecular markers in specific research areas of plant biology: population-level (intraspecific) systematics, hybridization, higher-order (interspecific or higher) systematics, biogeography, conservation, and general ecology.

149

Population level (intraspecific) systematics

Through examination of models estimating genetic differentiation, studies of variation between populations may shed light on evolutionary processes affecting the relationships of analyzed groups. Due to their high yield of polymorphism examined via distance-based or probabilistic approaches, multi-locus, dominant markers have been invaluable for describing relatedness within and among populations (Nybom, 2004).

The majority of SRAP studies have focused on intraspecific relations of highly- related, cultivated taxa. This has been due to the nature of dominant markers, but also to the primary users of SRAP markers, plant breeders, working to enhance the productivity and efficiency of selection through review of variation within populations of hybrid progeny (Li et al., 2007; Levi et al., 2011; Guo et al., 2012; Li et al., 2012) and germplasm collections (Ferriol et al., 2003, 2004a, 2004b; Chen et al., 2010; Amar,

2012). In contrast with phenotypic data, molecular variation has generally decreased with cultivation (for reviews, see Duvick, 1984; Clegg et al., 1984; Sonnante et al., 1994;

Lashermes et al., 1996; Sun et al., 2001; Whitt et al., 2002; Yang et al., 2006), as many modern agricultural crops are derived from limited selections of wild progenitors (Ferriol et al., 2004a; Wills & Burke, 2006), though this trend has been reversing in recent years with application of molecular tools (Reif et al., 2005). The majority of SRAP investigations have focused on agronomic applications, and demonstrated the capacity to uncover discrete genetic diversity within elite hybrid lineages. Such genetic distinction between different cultivated lineages within taxa can be considered analogous to projects interested in describing intra-specific or population-level variation found in wild- 150 collected accessions. SRAP markers are often used in conjunction with other markers, including RAPD (Zeng et al., 2008; Xue et al., 2010; Comlekcioglu et al., 2010), ISSR

(Yeboah et al., 2008; Wu et al., 2010; Liu et al., 2011; Zhang et al., 2012), SSR (Wang et al., 2010; Isk et al., 2011; Guo et al., 2012), or AFLP (Ferriol et al., 2004b; Zhao et al.,

2007; Devran et al., 2011; Youssef et al., 2011), often proving their utility by elucidating greater levels of variation within groups of highly related individuals.

Case studies

Euodia rutaecarpa (Juss.) Benth. (Rutaceae) is a woody, Asian shrub, the fruit of which has been used in traditional Chinese medicine. Recognized varieties of this species have been cultivated widely, and distinction between regional populations has been noted (Liang, 2011). Subspecific classifications have been based on macro-level morphology (Miki et al., 1999), though distinguishing traits demonstrate significant variation due to human selection and hybridization, seasonal, and environmental effects.

This regional differentiation has been confirmed on the molecular level with the internally transcribed spacer (ITS; Huang et al., 2008; Liu et al., 2009). Though cluster analysis could not effectively distinguish between all taxonomic varieties, Euodia rutaecarpa (Juss.) Benth var. rutaecarpa appeared distinct from varieties officinalis

(Dode) Huang and bodinieri (Dode) Huang. To further examine the potential for geographic and varietal genetic variation, Wei et al. (2011) employed AFLP and SRAP to examine the relationship between cultivated accessions of E. rutaecarpa var. rutaecarpa and E. rutaecarpa var. officinalis, and to compare the utility of the two marker systems.

SRAP primer combinations produced fewer scorable fragments than AFLP (188 vs. 353),

151 but yielded a higher rate of polymorphism (77.1% vs. 64.6%) and more variety-specific bands. Calculations of Nei’s genetic diversity and Shannon’s Information Index demonstrated that SRAP revealed higher levels of variation within and between these taxa than AFLP. Examination of genetic distance over populations (accessions grouped by province) suggested a higher level of divergence in populations from SRAP banding patterns than AFLP in both var. rutaecarpa (0.1002 vs. 0.0498) and var. officinalis

(0.2086 vs. 0.1231). Similarly, unweighted pair-group method (UPGMA) analyses of both individual accessions and groups clustered by provincial origin yielded nearly identical topologies for both markers, but the SRAP-based tree had slightly higher correspondence to the distance measure (rSRAP=0.98, rAFLP=0.94). While both markers successfully distinguished between the varieties of E. rutaecarpa examined, the authors proposed SRAP was the preferred method, due to higher genetic distance and polymorphism.

More widely grown crops have been examined by SRAP analysis as well, including the cultivated grape, Vitis vinifera L. (Vitaceae). Variation in accessions of this species, interspecific hybrids, and other horticulturally important grape taxa was illuminated by SRAP markers by Guo et al. (2012). Three primary groups were revealed by SRAP analysis, based on taxonomic status (V. vinifera vs. related taxa) and usage

(table grape vs. wine grape within V. vinifera). Of the two V. vinifera subgroups, one included all the table grape varieties, separating two established European cultivars and their known hybrid offspring from clusters of ancient Chinese varieties. The second V. vinifera assemblage demonstrated substructure as well, in the form of two primary sub-

152 clusters: Euro-America varieties in one, and a second that congregated all of the wine grape accessions. The third distinct group included primarily wild taxa, as well varieties used as rootstocks for grafting. This clade also demonstrated structure, with two subgroups segregated by geographic origins - species of Euro-American descent in one, and wild taxa from China in the second. SRAP phylogenies demonstrated agreement with previous molecular data describing geographic origin, horticultural use, and hybrid lineage among recently developed Vitus varieties and related species (Arroyo-Garcia et al., 2002; Aradhya et al., 2003).

Representatives from the genus Curcurbita, along with Zea, were the earliest domesticated crops in the New World, ca. 9000 years before present (YBP) (Smith, 1997;

Piperno et al., 2009), and is represented by numerous, diverse landraces (Lira-Saade,

1995). Examination of the economically valuable and morphologically diverse C. maxima Duchesne species complexes was carried out by Ferriol et al. (2004a).

Germplasm from 50 C. maxima landraces was assessed using SRAP and AFLP markers.

Results yielded similar rates of polymorphism, though AFLP yielded a greater number of polymorphic bands. Correlation between the AFLP and SRAP clustering patterns was very low (r=0.31), potentially due to the differences in nature of the markers’ genetic targets. Principal coordinates analyses (PCoA) of these markers clustered accessions based on different characters. The first two axes in SRAP scatterplots clustered some taxa by morphology, including flesh color, and shape, to a lesser degree (i.e. “heart shaped” varieties), whereas AFLP markers clustered accessions by region of origin.

Some groups of morphologically similar accessions, including “turban” and “banana”

153 types, clustered strongly together in both molecular approaches. Further, both markers indicated greater genetic distance in the nine accessions from the Americas, than the 39 from Europe, suggesting a fraction of the C. maxima diversity has been exported and exploited by Old World agriculture.

Hybridization

Hybridization is thought to be one of the driving factors behind speciation

(Mallet, 2007), and most significant in the genesis of novel taxa when combined with polyploidization (for reviews, see Soltis & Soltis, 1989; Soltis et al., 2003; Rieseberg &

Willis, 2007; Wissemann, 2007; Kim et al., 2008; Soltis & Soltis, 2009). Allopolyploid progeny of hybridization events typically exhibit intermediate characters, but in some instances demonstrate either increases or decreases in fitness relative to parental taxa, which may be stabilized or enhanced in subsequent generations (for review, see

Rieseberg, 1995). Homoploid hybridization is also an important mechanism for introducing genetic diversity into populations via introgression, and has been evident through examination of dominant markers (Arnold et al., 1991; Ungerer et al., 1998;

Wolfe, Xiang, & Kephart, 1998a; Wolfe, Xiang, & Kephart, 1998b; O’hanlon et al.,

1999). Many different hypotheses of hybridization have been tested using SRAP markers at different taxonomic levels, though most have addressed variation at the population level in intra-specific hybrids (Hale et al., 2006; Hale et al., 2007; Liu et al., 2007; Xuan et al., 2008). Some have relied on SRAP markers to depict inter-specific hybridization as well (Ren et al., 2010; Yu et al., 2009), and even inter-generic hybrids (Sheng et al.,

154

2002), but the majority apply these markers to closely related taxa, where they are most unbiased.

Case studies

Selections of Paoenia L. (Paeoniaceae) have been cultivated in Eurasia for nearly three millennia, primarily for medicinal and ornamental purposes (Halda & Waddick,

2004). Thousands of hybrid cultivars exist today, including a significant proportion of interspecific and intersectional hybrids. The taxonomy of the genus is still in contention

(Halda & Waddick, 2004; Hong, 2010). To examine the relationships between cultivated, intersectional Itoh hybrids and their parental lines from sections Paeonia

(herbaceous) and Moutan (woody), Hao et al. (2008) assessed diversity of a collection of

29 cultivars with SRAP. Although overall banding patterns showed high levels of polymorphism (~ 95%), some primers demonstrated robust amplification in sect. Paeonia while yielding no amplicons in Moutan. The UPGMA tree (CPCC =0.86) revealed two primary clusters separating the two taxonomic sections, and allied the examined hybrids with (maternal) sect. Paeonia. Sub-clusters within the Moutan and Paeonia groups were generally ambiguous, as the origins of the examined cultivars were not known, but there were some loose trends of grouping by flower color. One sub-cluster of the Moutan taxa was distinct and had high levels of genetic similarity – a group of hybrids arising from the morphologically distinctive P. rockii. Similar to UPGMA analysis, PCoA aligned the

Itoh hybrids closer to section Paeonia than Moutan. The placement of the intersectional hybrids with the Paeonia group may be due to the ploidy level of parental taxa.

According to Saunder and Stebbins (1938), generation of triploid hybrids derived from

155 diploid (Moutan) and tetraploid (Paeonia) species was more fruitful, even across sections, than between homoploids of the same section. Results of SRAP similarity add weight to this finding by demonstrating strong genetic affinity of primary hybrids to the maternal genotypes. A greater maternal contribution has been demonstrated in many hybrids derived from parental types of mixed ploidy (Lashermes et al., 2000; Haw, 2001;

Horita et al., 2003; Lim et al., 2003; Singh, 2006).

Coffea arabica L. (Rubiaceae) is another agriculturally valuable polyploid, grown today under varying regimes on all continents with frost-free agricultural zones. The identification of varieties via morphological and genetic traits is extremely difficult, due to the limited variability of the narrow genetic base of genotypes grown. Coffea arabica is self-fertile, but produces significantly greater yields of larger seeds with outcrossing

(Philpott et al., 2006), demonstrating the importance of genetic diversity in Coffea spp. seed set in both natural and cultivated environments. Decreases in seed production have been noted due to loss of pollinators (Klein et al., 2003) and the shallow genetic base of widely-grown, cultivated varieties (Anthony et al., 2002), underscoring the need for methods to distinguish between plants of common lineage for both ecological and agricultural investigations. In order to enhance productivity of cultivated C. arabica,

Mishra et al. (2011) employed SRAP markers to identify cultivated genotypes and their progeny in commercial hybrids. Selecting four cultivated varieties, researchers produced

60 inter-varietal hybrid seedlings from six unique crosses, including two reciprocal crosses. Of the SRAP markers documented, 90 were shared exclusively with maternal parents, and 48 with the paternal parents. Examination of banding patterns of reciprocal

156 crosses suggested parental contributions varied depending on the direction of the hybrid cross. This result may suggest that SRAP primers amplify some non-nuclear material, or alternatively that there may have been differences in levels of heterozygosity between parental genotypes. Also, a small proportion of the markers scored were hybrid specific, suggesting recombination events.

Breeding approaches of annual, agronomic crops are distinct from the methods used for long-lived, perennial species such as Paeonia and Coffea, which are slow to mature and yield harvestable products. Commercial production of seed shifted away from open-pollinated methods to favor development of multiple, distinct inbred lines to interbreed for enhancement of heterosis and uniformity (Tay, 2002). To assess the effectiveness of this hybridization strategy in production of an F1 cabbage hybrid

(Brassica oleracea var. capitata ‘Zaoxia 16’), and the purity of resultant hybrid seed, Liu et al. (2007) examined 210 F1 seedlings, as well as parental genotypes, with SRAP,

RAPD, ISSR, and SSR markers. Results indicated that all tested markers were able to assess the purity of the study population at similar rates (94.3% to 94.8%). Analysis of

SRAP markers indicated that these were the most robust relative to other markers, detecting more maternal- and paternal-specific markers per primer. The different molecular assays each flagged either 11 or 12 false hybrids, whereas field observation of morphological characters only exposed seven of these. Nine of the 12 hybrids exhibited nearly identical banding patterns to the female parent, suggesting that they were probably derived from self-pollination.

157

Higher-order (interspecific) systematics

Higher-order systematic investigations, examining variation at or above the species level, have typically been done through comparison of cytoplasmic DNA sequences or highly conserved regions of the nuclear genome, but dominant markers have demonstrated value in estimating relative distance between taxa. Variation between species has been assessed by the use of dominant markers, including RAPD (Graham et al., 1995; Rodriguez et al., 1999; Sun et al., 2001), ISSR (Wolfe et al., 1998a; Wolfe &

Randle, 2001; Kumar et al., 2008), RFLP (Federici et al., 1998; Lakshmi et al., 2000), and AFLP (Loh et al., 2000; Sasa et al., 2009). SRAP markers have similarly demonstrated value in detecting interspecific variation in many cases (Fan et al., 2010;

Wang et al., 2010; Tabebi et al., 2012), though, like other dominant markers, they are primarily used to address hypotheses involving lower-order systematics.

Case studies

Taxonomic and phylogenetic relationships within Citrus L. (Rutaceae) are unclear at best, due primarily to the promiscuity of related species and genera, centuries of cultivation, and wide geographic dispersion (Nicolosi et al., 2000). It has been proposed that there are only three “true” species of Citrus, and that cultivated varieties are a function of hybridization between closely related genera (i.e. Citrus with Poncirus or

Fortunella) or spontaneous somatic mutations (Barrett & Rhodes, 1976). Uzun et al.

(2009) investigated the systematics of Citrus using SRAP markers, utilizing a diverse group of 83 taxa dominated by cultivated species and interspecific hybrids, but representing all tribes and subtribes within the Rutaceae. Banding patterns demonstrated

158 a very high recovery of rare fragments - over 63% were found in 5% or fewer of the taxa studied. Cluster analysis via UPGMA resulted in a strongly supported dendrogram

(CPCC=0.93) that maintained the arrangement of the currently accepted subtribes

(Barrett and Rhodes 1976), though relation of the “primitive citrus” and “near citrus” groups was less distinct relative to the “true citrus” clade. Intergeneric hybrids were genetically intermediate, but tended to cluster with the maternal parent. The genus Citrus was separated into two primary groups, one of which was subdivided into two distinct clusters, congruent with previous molecular work (Federici et al., 1998; Liang et al.,

2007) in supporting proposals of few ancestral species (“three species” concept; Barrett

& Rhodes, 1976). Accessions from the genus Fortunella clustered within the Citrus group, a pattern without consensus in current literature, though it has been suggested that this genus is the closest relative to Citrus (Nicolosi et al., 2000; Barkley et al., 2006).

Following the discovery of SRAP utility in Citrus, SRAP markers were further exploited to address systematic hypotheses within Citrus. Investigations of distinction between groups as defined by the “three species” concept were conceived, involving the citron (Citrus medica L.; Uzun et al., 2011a), mandarin (Citrus reticulata Blanco; Uzun et al., 2011b), and pummelo (Citrus maxima (Burm.) Merr.; Uzun et al., 2011c). To examine the relative value of markers within this group, Amar et al. (2011) compared

SRAP, simple sequence repeats (SSR; microsatellites), and cleaved amplified polymorphic sequences from single nucleotide polymorphic (CAPS-SNP) markers in a

Citrus germplasm collection. Of the markers, the highest number of total markers and polymorphic markers per primer combination were produced by SRAP, followed by

159

SSR, and CAP-SNP markers. SRAP primers yielded more than twice the number of polymorphic bands per primer relative to other markers, and also demonstrated the highest average polymorphism information content (PIC) of the molecular descriptors employed.

The genus Hedychium L. (Zingiberaceae), known as the ginger lily group, is closely allied to true gingers (genus Zingiber), and is similarly used in regional, traditional medicine, perfume fragrance, but is also cultivated for large, boldly-colored flowers. Some of the ornamental taxa within Hedychium have become naturalized in regions outside their native ranges, including Australia (Csurhes & Hannan-Jones, 2008) and Hawaii (Funk, 2005), where they hybridize with endemic ginger lilies and/or displace native flora and are considered aggressive weeds. Though over 100 species have been historically identified in this genus, the definitive number of taxa is still disputed (Wood,

2000). Gao et al. (2008) addressed this issue of species delimitation in Hedychium with one of the first molecular phylogenetic analyses, comparing 22 accessions with SRAP markers. Genetic similarities between individuals ranged 0.08 to 0.97, though most were near 0.5. The similarity coefficient of H. flavescens and H. chrysoleucum was 0.97, supporting previous propositions of lumping of both taxa (among others) into the H. coronarium complex (Baker, 1892), though this evaluation has been debated by many

(Turrill, 1914), it is not currently accepted (Burtt & Smith, 1972).

The genus Panicum L. is a cosmopolitan group within the Poaceae, comprised of more than 450 species (Warmke, 1951; Webster, 1988). Though many of the taxa within this large group have not been subjected to comparative molecular analysis, switchgrass

160

(Panicum variegatum L.) has been the focus of many groups due to its adaptability and promise for large-scale biomass production (McLaughlin et al., 1999; McLaughlin &

Kszos, 2005; Bouton, 2007). To examine genetic diversity in cultivated switchgrass accessions and related taxa, Huang et al. (2011) applied SRAP and EST-SSR markers to a collection of 91 accessions of Panicum, representing 22 distinct species and 34 samples of P. variegatum. SRAP primers produced more polymorphic bands per primer combination (18 vs. 11.5), but markers demonstrated similar rates of polymorphism

(~95%). Comparison of genetic distance calculated for the two marker systems suggested they were highly similar (Mantel r = 0.874, p < 0.01). Concordant with this result, AMOVA results attributed congruent levels of the variation (70.02% vs. 73.35%) among species with SRAP and EST-SSR markers. Differentiation between markers was seen in their effectiveness to discover polymorphism, as SRAP markers had a marker index (sensu Powell & Morgante, 1996) almost twice as high as EST-SSR. Phylogenies indicated by UPGMA analysis of a combined dataset agreed with recognized systematic arrangement (Aliscioni et al., 2003), as well as geographic origin of accessions. While both SRAP and EST-SSR distinguished Panicum species and related taxa with similar discriminating power, SRAP bands were presumed more efficient than EST-SSR for examining large Panicum collections, due to the higher rate of polymorphic band discovery. Although SRAP markers demonstrated an advantage in terms of polymorphism, it is important to note that the EST-SSR markers were developed specifically for P. variegatum (Kantety et al., 2002; Tobias et al., 2005), and interspecific

161 variation of SSR flanking regions may have decreased primer affinity at some loci in related taxa.

Biogeography

The distribution of plants examined through molecular analysis of genetic diversity has drastically enhanced our understanding of historical climatic, geographic, and ecological events. Analysis of population-level genetic variation has been examined with dominant markers within many contexts, including the effects of glacial refugia (for reviews, see Soltis & Soltis, 1998; Schonswetter et al., 2005; Soltis & Soltis, 2006), systematics (Perrie et al., 2003; Smissen et al., 2003; McKinnon et al. 2008), exotic plant invasion (Meekins et al., 2001; Riaz et al., 2008), etc. SRAP markers have been similarly employed to describe variation over broadly distributed populations, though examples are limited. These investigations have typically been undertaken to detect novel variation for breeding purposes (Feng et al., 2009), or in other cases, conservation of genetic resources in taxa threatened by human disturbance or overharvesting (Qian et al., 2009, Zhao et al., 2012). Still, these studies demonstrate the applicability of SRAP markers to unveil molecular variation correlated to a geographic scale.

Case studies

Some of the earliest applications of SRAP markers beyond the Brassicaceae (Li

& Quiros, 2001), for which it was developed, were in buffalograss (Buchloe dactyloides

L.). The relative value of SRAP to other commonly employed molecular markers

(ISSR, RAPD, and SSR) in this species was examined by Budak et al. (2004a),

162 comparing the ability of these approaches to detect genetic patterns in 15 accessions of buffalograss. SRAP markers produced a significantly greater number of variable fragments per primer pair (>+1.25 polymorphisms per primer pair) and rate of polymorphism (>13%) relative to other markers tested. The average genetic similarity was also highest as detected by SRAP markers, which was strongly correlated to RAPD results (r=0.73). UPGMA dendrograms of SRAP-determined relatedness also had much higher bootstrap support relative to other markers. The authors concluded that SRAP markers “demonstrated better approximations to true variation within and among buffalograss cultivars”, and further pursued examination of this North American endemic herb with a larger germplasm collection of 53 accessions (Budak et al., 2004b).

Samples in this study represented a combination of geographically diverse genotypes and cytotypes (2x, 4x, 5x, 6x), selected from both wild and hybrid lineages. Principal components and cluster analysis of SRAP banding patterns created groupings that generally agreed with the ploidy level, more so than geographic association. Due to the general trend of a south-north gradient of increasing ploidy level, and the sampling strategy employed, the relationship of genetic diversity to geographic region was unclear, though there was some support for both correlations. Other examinations of

SRAP diversity in taxa with multiple ploidy levels have also found associations with fragment diversity and ploidy (Gulsen et al., 2009).

Celosia argentea L. is a cosmopolitan herb used in traditional Chinese medicine, while closely allied C. cristata L. has been developed as an ornamental flowering plant.

Due to a long history of both natural and artificial selection, and broad natural

163 distribution, notable differentiation has occurred in Celosia populations on a geographic scale. Feng et al. (2009) examined 330 plants from 14 natural and two cultivated C. argentea populations as well as six cultivated varieties of C. cristata with SRAP markers. Between populations, diversity indices were high (Nei=0.375;

Shannon=0.348) relative to other outcrossing annual species (Nybom & Bartish, 2000).

Cluster analysis divided accessions into two primary groups by species. Within the C. argentea group, populations collected from geographically proximal regions clustered together, suggesting a relationship between genetic variation and geographic distribution. Additionally, the genetic distances of wild populations of C. argentea are greater relative to cultivated accessions, suggesting human selection and agricultural systems have selected only a fraction of the diversity occurring naturally in C. argentea.

Herbaceous taxa have dominated applications of published SRAP analyses, with only a small fraction investigating woody plant species. Of these, most examine insect- pollinated fruiting taxa, and few investigate wind-pollinated species. In a study of Pinus koraiensis Siebold and Zucc., Feng et al. (2009) used SRAP analysis to examine genetic diversity and gene flow in undisturbed Chinese populations of this conifer. Employing

SRAP markers, seedlings derived from 24 provinces were characterized. Genetic diversity indices were low (Nei=0.1274, Shannon=0.1937), though predictable based on pollination system (Hamrick & Godt, 1996). Reciprocally, inter-population similarity indices were very high, ranging from 0.9590 to 0.9903, and AMOVA indicated that genetic variation primarily originated from intra-provenance differentiation (92.35%).

These results, paired with high calculated levels of gene flow (Nm=2.9), are congruent

164 with other studies utilizing dominant markers in detecting extremely low variation and diversity between geographically distant populations of P. koraiensis (Potenko &

Velikov, 1998; Kim et al., 2005; Feng et al., 2006). Other long-lived temperate conifers reproducing via airborne pollen have demonstrated similar levels of genetic identity

(Latta & Mitton, 1997; Echt et al., 1998; Yeaman & Jarvis, 2006; Robledo-Arnuncio,

2011; for review see Hamrick et al., 1992).

Conservation genetics

One critical application of molecular markers is to assess the genetic variation in the context of conservation (Karp, 1997). In order to develop an effective strategy to protect, promote, and maintain genetic diversity, it must first be quantified. For instances in which only limited protective measures are possible, it is essential to have this information in order to best direct conservation efforts (Petit et al., 1998). These genetic data provide relevant information for identifying units of conservation and illuminate the genetic processes that take place in the populations such as patterns of genetic flux, bottlenecks, and genetic drift.

In most cases, anthropogenic impacts have led to the decline of taxa, whether through overharvesting, habitat destruction, or more recently, changes in climate. Long- lived, slow-reproducing species are exceptionally vulnerable to such disturbances, including many plants in the Orchidaceae. Within this family, taxa have been pillaged from native habitats for ornamental, agronomic, and medicinal purposes, with significantly detriment to their abundance in the wild. Thanks to recent, commercial-

165 scale development of tissue culture propagation of these species, effective ex situ methods of production may decrease wild collection. Still, the detection of remaining genetic diversity is paramount to maintain these taxa over time, though habitat loss via human encroachment is always a lingering concern for endemic populations.

Case studies

Dendrobium loddigesii Rolfe is a Chinese epiphytic orchid that has been overharvested for medicinal uses (Hu, 1970), and is today very rare in the wild. Though endemic populations of many Dendrobium have been critically dwindling for decades

(Siu, 2001; Chen et al., 2009), the vast majority of studies involving these taxa, including

D. loddigesii, are related to commercially valuable, chemical isolates (Li et al., 1991;

Zhang et al., 2003; Wenyun, 2006; Yin, 2007) and production (Zhang et al., 2001; Liang et al., 2010). To assess the genetic diversity in extant populations of D. loddigesii, Cai et al. (2011) collected tissues from 92 individuals across seven populations in southern

China. Although the levels of SRAP polymorphism varied significantly between populations (27.71% to 68.83%), there was no correlation between molecular variation and geographic location. Analysis of molecular variance did suggest that there was genetic variation among populations (GST = 0.304), which is higher than the average outcrossing monocot (Hamrick & Godt, 1996) and member of the Orchidaceae (Forrest et al., 2004), but such diversity indices can vary widely between species.

A closely allied species, Dendrobium officinale Kimura et Migo, is a similarly imperiled Chinese endemic, suffering from overharvesting for anthropocentric purposes.

Ding et al. (2008) collected 84 individuals from nine selected populations in southeast

166

China for SRAP analysis. Analysis of molecular variance also indicated that there was significantly higher variation within than between populations (72.95% vs. 27.05%), which is comparable to that reported by Ding et al. (2009) for ISSR (78.84%) and RAPD

(78.88%) markers. The amount of variation (GST) among D. officinale populations was comparable to D. loddigesii (0.3484 vs. 0.304), which is not surprising as the taxa are closely related and maintain sympatric distribution and pollination strategies (Rong,

2008). Analysis of Nei’s genetic distance strongly supported two primary clusters, but

Mantel test revealed that no significant correlation was detected between genetic and geographic distance. The strategies of pollination and seed dispersal of Dendrobium, and many orchids, often lead to this pattern of high levels of variation within and low levels between populations. Wind-mediated dispersal of propagules promotes outbreeding and gene flow between distant populations (Ridley, 1905) as seen in many species of

Dendrobium (Arditti &Ghani, 2000), and such traits are associated with high levels of within population genetic diversity relative to differentiation between populations

(Hamrick & Godt, 1996).

Lycium ruthenicum Murr. is an endangered species within this Solanaceae, threatened due to habitat depletion and overharvesting for use in traditional medicine

(Chen et al., 2008). Liu et al. (2012) examined variation in 174 individuals collected from 14 wild populations in northwestern China using SRAP markers and five environmental variables. Results of AMOVA analysis and Nei’s gene diversity indicated that genetic variation was primarily found within populations (PhiST=0.155; GST=0.216).

Correlations between genetic measures and ecogeographic variables showed positive

167 relationships between altitude and annual number of sunshine hours with genetic diversity, while no relationship between and latitude or longitude. Diversity measures for

L. ruthenicum were considerably higher when compared to the average of plants within the Solanaceae (Hamrick & Godt, 1996) and long-lived perennial woody plants (Hamrick et al., 1992). These high measures of population-level diversity were attributed to long lifespan and wind-mediated pollen dispersal.

Ecology

Numerous ecological applications for the SRAP marker system have been explored, though have not been pursued beyond the interests of crop development. For example, SRAP markers have been devoted to the identification of markers for pathogen resistance in agronomic crops (Liu et al., 2008; Li et al., 2010; Saha et al., 2010a, 2010b;

Zhao et al., 2010; Devran et al., 2011), as well as variation in those pathogens (Baysal et al., 2009; Pasquali et al., 2010; Devran et al., 2011; Devran & Baysal, 2012; Irzykowska et al., 2012). Such studies suggest the value of SRAP applications in broader, ecological contexts, including description of the evolutionary interactions between microbes and their hosts. Similarly, employment of SRAP markers to distinguish between flower color

(Fu et al., 2008; Han et al., 2008; Guo et al., 2011; Wang et al., 2012; Yang et al., 2012) or fruit morphology (Yuan et al., 2008; Chen et al., 2010; Han et al., 2012) could be used in conjunction with data of preference or relative behavior of pollinators or seed- dispersing organisms. Some genetic information regarding onset or duration of flowering

(Zhang et al., 2011) could be used in studies of pollinator-interaction, but further in

168 analyses examining phenology, and potentially the selective effects of global climate change. Associations of SRAP markers and environmental adaptations have also been discovered, pairing specific banding patterns with increased cold tolerance (Castonguay et al., 2010) and other ecological factors including seasonal temperature variation, annual rainfall variability, and mean annual humidity (Dong et al., 2010). Any and all of the underlying functional genes uncovered by SRAP analysis that control variation in morphology, phenology, or adaptability could also be applied to investigations of ecological niche modeling and further, ecological speciation.

Conclusion

In the last few decades, dominant marker systems have become the standard tool for investigations of genetic variation in plant biology (reviews in Westman et al., 1997;

Glaubitz & Moran, 2001; Nybom, 2001, 2004; Agarwal et al., 2008). The presumption that these markers are selectively neutral due to their amplification from random primers has been essential for their use in systematic investigations. As dominant marker analyses have been described as a comparison of “molecular phenotypes” (sensu

Zhivotovsky, 1999), we propose that use of SRAP markers should be viewed as an analogy of morphological analysis, which have been widely used in species delimitation and assessment of variation. As SRAP markers target coding regions of the genome (Li

& Quiros, 2001), they have potential beyond more commonly applied multi-locus markers, possessing the capacity to elucidate markers with inherent biological significance. Hypotheses addressing ecological processes including niche adaptation and

169 interspecific interaction could be addressed via SRAP, as well as investigations of taxonomic affinities and conservation genetics. Further, this technique has been shown to robustly distinguish between taxa as well as, or better than, previously developed dominant marker techniques. Paired with the large-scale capability of next-gen techniques, these easy-to-use, highly-variable markers may prove to be invaluable for discovery of polymorphism for development of novel variable loci (SNP, SSR, QTL, etc.) for investigations of plant biology.

170

Tables and Figures

40

30

20 Number Number referencesof 10

0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Year

Figure 4.1. Subset (n=188) of publications employing SRAP markers by year of publication

171

References

Abbott, R. J., Smith, L. C., Milne, R. I., Crawford, R. M., Wolff, K., & Balfour, J. (2000). Molecular analysis of plant migration and refugia in the Arctic. Science, 289(5483), 1343-1346.

Abedian, M., Talebi, M., Golmohammdi, H. R., & Tabatabaei, B. E. (2012). Genetic diversity and population structure of mahaleb cherry (Prunus mahaleb L.) and sweet cherry (Prunus avium L.) using SRAP markers. Biochemical Systematics and Ecology, 40, 112-117.

Adams, S. R., Pearson, S., & Hadley, P. (1997). The effects of temperature, photoperiod and light integral on the time to flowering of pansy cv. Universal Violet (Viola × wittrockiana Gams.). Annals of Botany, 80(1), 107-112.

Agarwal, M., Shrivastava, N., & Padh, H. (2008). Advances in molecular marker techniques and their applications in plant sciences. Plant Cell Reports, 27(4), 617- 631.

Aggarwal, R. K., Brar, D. S., Nandi, S., Huang, N., & Khush, G. S. (1999). Phylogenetic relationships among Oryza species revealed by AFLP markers. Theoretical and Applied Genetics, 98, 1320-1328.

Ahmad, R., Potter, D., & Southwick, S. M. (2004). Genotyping of peach and nectarine cultivars with SSR and SRAP molecular markers. Journal of the American Society for Horticultural Science, 129(2), 204-210.

Ai, P. F., Lu, L. P., & Song, J. J. (2012). Cryopreservation of in vitro-grown shoot-tips of Rabdosia rubescens by encapsulation-dehydration and evaluation of their genetic stability. Plant Cell, Tissue and Organ Culture, 108(3), 381-387.

Ai, P. F., Zhen, Z. J., & Jin, Z. Z. (2011). Genetic diversity and relationships within sweet kernel apricot and related Armeniaca species based on sequence-related amplified polymorphism markers. Biochemical Systematics and Ecology, 39(4), 694-699.

Al-Rabab’ah, M. A.., & Williams, C. G. (2002). Population dynamics of Pinus taeda L. based on nuclear microsatellites. Forest Ecology and Management, 163, 263-271.

Alasaad, S., Li, Q. Y., Lin, R. Q., Martin-Atance, P., Granados, J. E., Diez-Banos, P., Perez, J.M., & Zhu, X. Q. (2008). Genetic variability among Fasciola hepatica samples from different host species and geographical localities in Spain revealed by the novel SRAP marker. Parasitology Research, 103(1), 181-186.

172

Alghamdi , S. S., Al-Faifi, S. A., Migdadi, H. M., Khan, M. A., EL-Harty, E. H., & Ammar, M. H. (2012). Molecular diversity assessment using sequence related amplified polymorphism (SRAP) markers in Vicia faba L. International Journal of Molecular Sciences, 13, 16457-16471.

Aliscioni, S. S., Giussani, L. M., Zuloaga, F. O., & Kellogg, E. A. (2003). A molecular phylogeny of Panicum (Poaceae: Paniceae): tests of monophyly and phylogenetic placement within the Panicoideae. American Journal of Botany, 90(5), 796–821.

Alwala, S., Kimbeng, C. A., Veremis, J. C., & Gravois, K. A. (2008). Linkage mapping and genome analysis in a Saccharum interspecific cross using AFLP, SRAP and TRAP markers. Euphytica, 164(1), 37-51.

Amar, M. H. (2012). Comparative analysis of SSR and SRAP sequence divergence in Citrus germplasm. Biotechnology, 11, 20-28.

Amar, M. H., Biswas, M. K., Zhang, Z., & Guo, W. (2011). Exploitation of SSR, SRAP and CAPS-SNP markers for genetic diversity of Citrus germplasm collection. Scientia Horticulturae, 128, 220–227.

Aneja, B., Yadav, N. R., Chawla, V., & Yadav, R. C. (2012). Sequence-related amplified polymorphism (SRAP) molecular marker system and its applications in crop improvement. Molecular Breeding, 30(4), 1635-1648.

Anderson, N. O. (Ed.). (2006). Flower Breeding and Genetics: Issues, Challenges and Opportunities for the 21st Century. Dordrecht, Netherlands: Springer.

Anthony, F., Combes, M.-C., Astorga, C., Bertrand, B., Graziosi, G., & Lashermes, P. (2002). The origin of cultivated Coffea arabica L. varieties revealed by AFLP and SSR markers. Theoretical and Applied Genetics, 104(5), 894–900.

Aradhya, M. K., Dangl, G. S., Prins, B. H., Boursiquot, J. M., Walker, M. A., Meredith, C. P., & Simon, C. J. (2003). Genetic structure and differentiation in cultivated grape, Vitis vinifera L. Genetical Research, 81(3), 179-192.

Arbogast, B. S., & Kenagy, G. J. (2001). Comparative phylogeography as an integrative approach to historical biogeography. Journal of Biogeography, 28(7), 819-825.

Arditti, J., & Ghani, A. K. A. (2000). Tansley Review No. 110. Numerical and physical properties of orchid seeds and their biological implications. New Phytologist, 367- 421.

Arif, I. A., Bakir, M. A., Khan, H. A., Al Farhan, A. H., Al Homaidan, A. A., Bahkali, A. H., Al Sadoon, M., & Shobrak, M. (2010). A brief review of molecular techniques 173

to assess plant diversity. International Journal of Molecular Science, 11, 2079- 2096.

Ariss, J. J., & Vandemark, G. J. (2007). Assessment of genetic diversity among nondormant and semidormant alfalfa populations using sequence-related amplified polymorphisms. Crop Science, 47(6), 2274-2284.

Armengaud, P., Zambaux, K., Hills, A., Sulpice, R., Pattison, R. J., Blatt, M. R., & Amtmann, A. (2009). EZ-Rhizo: Integrated software for the fast and accurate measurement of root system architecture. The Plant Journal, 57(5), 945-956.

Arnold, M. L., Buckner, C. M., & Robinson, J. J. (1991). Pollen-mediated introgression and hybrid speciation in Louisiana irises. Proceedings of the National Academy of Sciences, 88(4), 1398-1402.

Arnold, M. L., & Hodges, S. A. (1995). Are natural hybrids fit or unfit relative to their parents?. Trends in Ecology & Evolution, 10(2), 67-71.

Arroyo-Garcia, R., Lefort, F., Andrés, M. T. D., Ibanez, J., Borrego, J., Jouve, N., Vabello, F., & Martínez-Zapater, J. M. (2002). Chloroplast microsatellite polymorphisms in Vitis species. Genome, 45(6), 1142-1149.

Austin, J. D., Lougheed, S. C., & Boag. P. T. (2004). Discordant temporal and geographic patterns in maternal lineages of eastern North American frogs, Rana catesbeiana (Ranidae) and Pseudacris crucifer (Hylidae). Molecular Phylogenetics and Evolution, 32, 799-816.

Avise, J. C. (1998). Pleistocene phylogeographic effects on avian populations and the speciation process. Proceedings of the Royal Society of London. Series B: Biological Sciences, 265(1395), 457-463.

Avise, J. C., Arnold, J., Ball, R. M., Bermingham, E., Lamb, T., Neigel, J. E., Reeb, C. A., & Saunders, N. C. (1987). Intraspecific phylogeography: The mitochondrial DNA bridge between population genetics and systematics. Annual Review of Ecology and Systematics, 18, 489–522.

Bacchetta, G., Grillo, O., Lovicu, G., Orrù, M., Piazza, G., Ravalli, C., & Venora, G. (2010, August). Pips image analysis to support cultivar identification of Vitis vinifera L. In Abstract Book of CIGR Workshop on Image Analysis in Agriculture (pp. 26-28). Budapest: CIGR.

Bailey, D. A., Baker, J. R., & Jones, R. K. (1996). A penny for your pansees. North Carolina Flower Growers' Bulletin: Official Publication of the NC Commercial Flower Growers' Association, 44(3), 2-14. 174

Baker, J. G. (1892). Scitamineae. In Hooker, J. D., The Flora of British India (Vol. VI., pp. 198-264). London: Reeve.

Ballard Jr., H. E. (1996). Phylogenetic relationships and infrageneric groups in Viola (Violaceae) based on morphology, chromosome numbers, natural hybridization and internal transcribed spacer (ITS) sequences. (PhD thesis). ProQuest Dissertations and Theses.

Ballard, H. E., & Sytsma, K. J. (2000). Evolution and biogeography of the woody Hawaiian violets (Viola, Violaceae): Arctic origins, herbaceous ancestry and bird dispersal. Evolution, 54(5), 1521-1532.

Ballard, R., Rajapakse, S., Abbott, A., & Byrne, D. H. (1995). DNA markers in rose and their use for cultivar identification and genome mapping. Acta Horticulturae, 424, 265-268.

Barcaccia, G. (2009). Molecular markers for characterizing and conserving crop plant Germplasm. In S. M. Jain & D. S. Brar (Eds.), Molecular Techniques in Crop Improvement (pp. 231-254). Dordrecht, Netherlands: Springer.

Barbara, T., Palma-Silva, C., Paggi, G. M., Bered, F., Fay, M. F., & Lexer, C. (2007). Cross-species transfer of nuclear microsatellite markers: Potential and limitations. Molecular Ecology, 16(18), 3759-3767.

Barkley, N. A., Roose, M. L., Krueger, R. R., & Federici, C. (2006). Assessing genetic diversity and population structure in a Citrus germplasm collection utilizing simple sequence repeat markers (SSRs). Theoretical and Applied Genetics, 112, 1519-1531.

Barrett, H., & Rhodes, A. (1976). A numerical taxonomic study of affinity relationships in cultivated Citrus and its close relatives. Systematic Botany, 1, 105-136.

Barrett, S. C., & Husband, B. C. (1990). The genetics of plant migration and colonization. In A. H. D. Brown, M. T. Clegg, A. L. Kahler., & B. S. Weir (Eds.), Plant Population Genetics, Breeding, and Genetic Resources (pp. 254-277). Sunderland, MA: Sinauer Associates. Inc.

Bataillon, T. M., David, J. L., & Schoen, D. J. (1996). Neutral genetic markers and conservation genetics: simulated germplasm collections. Genetics, 144(1), 409- 417.

Batlle, I., & Alston, F. H. (1996). Genes determining leucine aminopeptidase and mildew resistance from the ornamental apple,‘White Angel’. Theoretical and Applied Genetics, 93(1-2), 179-182. 175

Baysal, O., Siragusa, M., Ikten, H., Polat, I., Gümrukcu, E., Yigit, F., Carimi, F., & Teixeira da Silva, J. A. (2009). Fusarium oxysporum f. sp. lycopersici races and their genetic discrimination by molecular markers in West Mediterranean region of Turkey. Physiological and Molecular Plant Pathology, 74(1), 68-75.

Beattie, A. J. (1971). Pollination mechanisms in Viola. New Phytologist, 70(2), 343-360.

Beatty, G. E., & Provan, J. (2011). Phylogeographic analysis of North American populations of the parasitic herbaceous plant Monotropa hypopitys L. reveals a complex history of range expansion from multiple late glacial refugia. Journal of Biogeography, 38(8), 1585-1599.

Beatty, G. E., & Provan, J. (2010). Refugial persistence and postglacial recolonization of North America by the cold‐tolerant herbaceous plant Orthilia secunda. Molecular Ecology, 19(22), 5009-5021.

Becker, W. (1910). Violae Europeae, Dresden, Germany: Verlag von C. Heinrich.

Becker, W.A. & Ewart, L. C. (1990). Pollination, seed set and pollen tube growth investigations in Viola pedata L. Acta Horticulturae, 272, 33-36.

Bennett, P. J. (2006). 2005 Herbaceous Ornamental Field Trial Results. In J. A. Chatfield, E. A. Draper, D. E. Dyke, J. P. Bennett, & J. F. Boggs (Eds.) Ornamental Plants; Annual Reports and Research Reviews 2005. Retrieved from http://ohioline.osu.edu/sc197/pdf/sc197.pdf

Berendzen, P.B., J. F. Dugan, & T. Gamble. 2010. Post-glacial expansion into the Paleozoic Plateau: Evidence of an Ozarkian refugium for the Ozark minnow Notropis nubilus (Teleostei: Cypriniformes). Journal of Fish Biology, 77, 1114- 1136.

Bermingham E., & Avise J. C. (1986). Molecular zoogeography of freshwater fishes in the southeastern United States. Genetics, 113, 939–965.

Besnard, G., Tagmount, A., Baradat, P., Vigouroux, A., & Berville, A. (2002). Molecular approach of genetic affinities between wild and ornamental Platanus. Euphytica, 126(3), 401-412.

Bi, Y. M., Cammue, B. P. A., Goodwin, P. H., KrishnaRaj, S., & Saxena, P. K. (1999). Resistance to Botrytis cinerea in scented geranium transformed with a gene encoding the antimicrobial protein Ace-AMP1. Plant Cell Reports, 18(10), 835- 840.

176

Bouton, J. H. (2007). Molecular breeding of switchgrass for use as a biofuel crop. Current Opinion in Genetics & Development, 17(6), 553-558.

Brant, S. V., & Ortí, G. (2003). Phylogeography of the Northern short-tailed shrew, Blarina brevicauda (Insectivora: Soricidae): past fragmentation and postglacial recolonization. Molecular Ecology, 12, 1435-49.

Brauner, S., Crawford, D., & Stuessy, T. (1992). Ribosomal DNA and RAPD Variation in the Rare Plant Family Lactoridaceae. American Journal of Botany, 79(12), 1436-1439.

Breen, A. L., Murray, D. F., & Olson, M. S. (2012). Genetic consequences of glacial survival: the late Quaternary history of balsam poplar (Populus balsamifera L.) in North America. Journal of Biogeography, 39(5), 918-928.

Bretting, P. K., & Widrlechner, M. P. (1995). Genetic markers and horticultural germplasm management. HortScience, 30(7), 1349-1356.

Brewer, M. T., Lang, L., Fujimura, K., Dujmovic, N., Gray, S., & van der Knaap, E. (2006). Development of a controlled vocabulary and software application to analyze fruit shape variation in tomato and other plant species. Plant Physiology, 141, 15-25.

Briggs, F. N., & Knowles, P. F. (1967). Introduction to Plant Breeding. New York, NY: Reinhold Publishing Corporation.

Brown, A. F., Jeffery, E. H., & Juvik, J. A. (2007). A polymerase chain reaction-based linkage map of boccoli and identification of quantitative trait loci associated with harvest date and head weight. Journal of the American Society of Horticultural Science, 132(4), 507–513.

Brown, A. H. D., & Briggs, J. D. (1991). Sampling strategies for genetic variation in ex situ collections of endangered plant species. Genetics and Conservation of Rare Plants, 99, 119.

Bruna, S., Giovannini, A., De Benedetti, L., Principato, M. C., & Ruffoni, B. (2006, February). Molecular analysis of Salvia Spp. through RAPD markers. Acta Horticulturae, 723, 157-160.

Brunsfeld, S. J., Sullivan, J., Soltis, D. E., & Soltis, P. S. (2001). Comparative phylogeography of northwestern North America: A synthesis. Special Publication-British Ecological Society, 14, 319-340.

177

Brusatte, S. L., Sakamoto, M., Montanari, S., & Harcourt Smith, W. E. H. (2012). The evolution of cranial form and function in theropod dinosaurs: Insights from geometric morphometrics. Journal of Evolutionary Biology, 25(2), 365-377.

Bruvo, R., Michiels, N. K., D’Souza, T. G., & Schulenburg, H. (2004). A simple method for the calculation of microsatellite genotype distances irrespective of ploidy level. Molecular Ecology, 13, 2101–2106.

Bubenchikov, R. A., & Goncharov, N. F. (2005). HPLC analysis of phenolic compounds in field violet. Pharmaceutical Chemistry Journal, 39(3), 143-144.

Budak, H., Shearman, R. C., Parmaksiz, I., & Dweikat, I. (2004a). Comparative analysis of seeded and vegetative biotype buffalograsses based on phylogenetic relationship using ISSRs, SSRs, RAPDs, and SRAPs. Theoretical and Applied Genetics, 109(2), 280-288.

Budak, H., Shearman, R. C., Gulsen, O., & Dweikat, I. (2005). Understanding ploidy complex and geographic origin of the Buchloe dactyloides genome using cytoplasmic and nuclear marker systems. Theoretical and Applied Genetics, 111(8), 1545-1552.

Budak, H., Shearman, R. C., Parmaksiz, I., Gaussoin, R. E., Riordan, T. P., & Dweikat, I. (2004b). Molecular characterization of buffalograss germplasm using sequence- related amplified polymorphism markers. Theoretical and Applied Genetics, 108(2), 328-334.

Burns, J. F., & Cusick, A. W. (1984). Viola pedata L. Bird-foot Violet. Columbus, OH. Ohio Department of Natural Resources Division of Natural Areas and Preserves.

Burtt, B. L., & Smith, R. M. (1972). Tentative keys to the subfamilies, tribes and genera of Zingiberaceae. Notes Royal Botanical Garden Edinburgh, 31, 171-176.

Bylesjo, M., Segura, V., Soolanayakanahally, R. Y., Rae, A. M., Trygg, J., Gustafsson, P., Jansson, S., & Street, N. R. (2008). LAMINA: A tool for rapid quantification of leaf size and shape parameters. BMC Plant Biology, 8(1), 82.

Caballero, A., Quesada, H., & Rolán-Alvarez, E. (2008). Impact of amplified fragment length polymorphism size homoplasy on the estimation of population genetic diversity and the detection of selective loci. Genetics, 179(1), 539-554.

Caballero, A., & Quesada, H. (2010). Homoplasy and distribution of AFLP fragments: an analysis in silico of the genome of different species. Molecular Biology and Evolution, 27(5), 1139-1151.

178

Cai, X., Feng, Z., Zhang, X., Xu, W., Hou, B., & Ding, X. (2011). Genetic diversity and population structure of an endangered orchid (Dendrobium loddigesii Rolfe) from China revealed by SRAP markers. Scientia Horticulturae, 129(4), 877-881.

Canne, J. M. (1987). Determinations of chromosome numbers in Viola (Violaceae). Canadian Journal of Botany, 65(4), 653-655.

Cao, K., Wang, L., Zhu, G., Fang, W., Chen, C., & Zhao, P. (2011). Construction of a linkage map and identification of resistance gene analog markers for root-knot nematodes in wild peach, Prunus kansuensis. Journal of the American Society for Horticultural Science, 136(3), 190-197.

Carr, J., Xu, M., Dudley, J. W., & Korban, S. S. (2003). AFLP analysis of genetic variability in New Guinea impatiens. Theoretical and Applied Genetics, 106(8), 1509-1516.

Carroll, S. B. & Goldman, P. C. (1994, October 15-16). Analysis of flower color in Viola pedata (birdfoot violet). Paper presented at the North American Conference on Savannas and Barrens, Illinois State University. Normal, Illinois. doi: http://www.epa.gov/ecopage/upland/oak/oak94/Proceedings/Carroll.html

Carstens, B. C., Brunsfeld, S. J., Demboski, J. R., Good, J. M., & Sullivan, J. (2005). Investigating the evolutionary history of the Pacific Northwest mesic forest ecosystem: hypothesis testing within a comparative phylogenetic framework. Evolution, 59(8), 1639-1652.

Castonguay, Y., Cloutier, J., Bertrand, A., Michaud, R., & Laberge, S. (2010). SRAP polymorphisms associated with superior freezing tolerance in alfalfa (Medicago sativa spp. sativa). Theoretical and Applied Genetics, 120(8), 1611–1619.

Chao, C. C. T., Devanand, P. S., & Chen, J. (2005). AFLP analysis of genetic relationships among Calathea species and cultivars. Plant Science, 168(6), 1459- 1469.

Chaotian, X., Changsheng, C., Yan, X., & Dehua, J. (2010). Construction of a genetic linkage map for Porphyra haitanensis (Bangiales, Rhodophyta) based on sequence-related amplified polymorphism and simple sequence repeat markers. Journal of Phycology, 46(4), 780-787.

Chase, M. W., & Hills, H. H. (1991) Silica gel: An ideal material for field preservation of leaf samples for DNA studies. Taxon, 40, 215-220.

179

Chaudhary, A. A., Dhar, U., Ahmad, A., Bhatt, I. D., Jugran, A., & Kaur, G. (2012). Analysis of genetic diversity in Hedychium spicatum Buch. Ham. ex. D. Don in West Himalaya, India. Journal of Medicinal Plants Research, 6(23), 3984-3991.

Chen, D., Peng, R., Li, L., Sun, N., Zhong, G., & Cai, Y. (2009). Study on genetic diversity of Codonopsis tangshen by SRAP and ISSR markers. China Journal of Chinese Materia Medica, 34(3), 255-259.

Chen, G., Geng, J., Rahman, M., Liu, X., Tu, J., Fu, T.,Li, G., McVetty, P. B., & Tahir, M. (2010). Identification of QTL for oil content, seed yield, and flowering time in oilseed rape (Brassica napus). Euphytica, 175(2), 161-174.

Chen, H. K., Pu, L. K., Cao, J. M., & Ren, X. (2008). Current research state and exploitation of Lycium ruthenicum Murr. Heilongjiang Agricultural Sciences, 5, 155-157.

Chen, M. M., Feng, F., Sui, X., Li, M. H., Zhao, D., & Han, S. (2010). Construction of a framework map for Pinus koraiensis Sieb. et Zucc. using SRAP, SSR and ISSR markers. Trees, 24(4), 685-693.

Chen, M. X., Wei, C. L., Qi, J. M., Chen, X. B., Su, J. G., Li, A. Q., Tao, A. F., & Wu, W. R. (2011). Genetic linkage map construction for kenaf using SRAP, ISSR and RAPD markers. Plant Breeding, 130(6), 679-687.

Chen, S., Meng, H., Chenga, Z., Li, Y., Shaanxi, Y., & Chai, D. (2010). Identification of two SRAP markers linked to green skin color in cucumber. Acta Horticulturae, 859, 351-356.

Chen, W., Zhang, Y., Liu, X., Chen, B., Tu, J., & Tingdong, F. (2007). Detection of QTL for six yield-related traits in oilseed rape (Brassica napus) using DH and immortalized F(2) populations. Theoretical and Applied Genetics, 115(6), 849- 858.

Chen, X. Q., Luo, Y. B., & Wood, J. J. (2009). Dendrobium. In: Wu, Z.Y., Raven, P.H., Hong, D.Y. (Eds.), Flora of China, vol. 25 (Orchidaceae). Bejing: Science Press, and St. Louis, MO: Missouri Botanical Garden Press.

Chen, X., Xia, Z., Fu, Y., Lu, C., & Wang, W. (2010). Constructing a genetic linkage map using an F1 population of non-inbred parents in cassava (Manihot esculenta Crantz). Plant Molecular Biology Reporter, 28(4), 676-683.

Chen, Y., Li, G., & Wang, X. L. (2010). Genetic diversity of a germplasm collection of Cucumis melo L. using SRAP markers. Hereditas (Beijing), 32(7), 744-751.

180

Chi, L., Teng Z., Li, X., Sui, S. C.,, & Li, M. (2011). SRAP analysis of 23 species of wild lily genetic relationships. Journal of Agricultural Biotechnology, 19(4), 677-684.

Chittenden, R. J. (1928). Notes on species crosses in Primula, Godetia, Nemophila, and Phacelia. Journal of Genetics, 19(3), 285-314.

Chung, M., Gelembiuk, G., & Givnish, T. J. (2004). Population genetics and phylogeography of endangered Oxytropis campestris var. chartacea and relatives: arctic‐alpine disjuncts in eastern North America. Molecular Ecology, 13(12), 3657-3673.

Clark, J. S. (1998). Why trees migrate so fast: confronting theory with dispersal biology and the paleorecord. The American Naturalist, 152(2), 204-224.

Clark, J. S., Fastie, C., Hurtt, G., Jackson, S. T., Johnson, C., King, G. A., Lewis, M., Lynch, J., Pacala, S., Pretentice, C., Schupp, E. W., Webb, T., & Wyckoff, P. (1998). Reid's paradox of rapid plant migration. BioScience, 48(1), 13-24.

Clausen, J. (1921). Studies on the collective species Viola tricolor 1; Preliminary notes. Botanisk Tidsskrift, 37, 205-221.

Clausen, J. (1924). Increase of chromosome numbers in Viola experimentally induced by crossing. Hereditas, 5(1), 29-32.

Clausen, J. (1926). Genetical and cytological investigations on Viola tricolor L. and V. arvensis Murr. Hereditas, 8(1-2), 1-156.

Clausen, J. (1927). Chromosome number and the relationship of species in the genus Viola. Annals of Botany, 41(4), 677-714.

Clausen, J. (1931). Cyto-genetic and taxonomic investigations on Melanium violets. Hereditas, 15(3), 219-308.

Clausen, J. (1964). Cytotaxonomy and distributional ecology of western North American violets. Madrofno, 17, 173-204.

Clauss, M. J., Cobban, H., & T. Mitchell-Olds, T. (2002). Cross-species microsatellite markers for elucidating population genetic structure in Arabidopsis and Arabis (Brassicaeae) Molecular Ecology, 11, 591-601.

Clegg, M. T., Brown, A. H. D., & Whitfeld, P. R. (1984). SHORT PAPER: Chloroplast DNA diversity in wild and cultivated barley: Implications for genetic conservation. Genetical Research, 43, 339-343.

181

Cole, M., Bowers, F. D., & Freckmann, S. (1979). Bryophytes of the Kickapoo River Valley, southwestern Wisconsin. The Bryologist, 82(2), 273-276.

Comlekcioglu, N., Simsek, O., Boncuk, M., & Aka-Kacar, Y. (2010). Genetic characterization of heat tolerant tomato (Solanum lycopersicon) genotypes by SRAP and RAPD markers. Genetics and Molecular Research, 9, 2263-2274.

Commission Internationale de l’Eclairage (CIE) (1978). Recommendations on uniform color spaces: Color-difference equations, psychometric color terms. CIE Publication 15 Supplement 2. Commission Internationale de le E´ clairage, Paris, France.

Cook, E. T. (1903). Sweet Violets and Pansies: And Violets from Mountain and Plain. London: C. Scribner's Sons.

Coombs, R. E. (1981). Violets: the history and cultivation of scented violets. London: Croom Helm.

Craft, C. B., & Casey, W. P. (2000). Sediment and nutrient accumulation in floodplain and depressional freshwater wetlands of Georgia, USA. Wetlands, 20(2), 323-332.

Crane, H. H. (1908). The Book of the Pansy, Viola, & Violet. London: J. Lane Company.

Cravero, V., Martin, E., & Cointry, E. (2007). Genetic diversity in Cynara cardunculus determined by sequence-related amplified polymorphism markers. Journal of the American Society for Horticultural Science, 132(2), 208-212.

Csurhes, S. & Hannan-Jones, M. (2008). Pest plant risk assessment: Kahili ginger Hedychium gardnerianum, White ginger Hedychium coronarium, Yellow ginger Hedychium flavescens. The State of Queensland, Department of Primary Industries and Fisheries. 20pp.

Cuevas, H. E., Song, H., Staub, J. E., & Simon, P. W. (2011). Inheritance of beta- carotene-associated flesh color in cucumber (Cucumis sativus L.) fruit. Euphytica, 171(3), 301-311.

Cui J. J., Song, J., Bao, L. M., Fang, L., & Hu, K. (2012). Genetic diversity analysis of germplasm resources of towel gourd based on SRAP markers. Jounral of Plant Genetic Resources, 13(6). 1061-1066.

Culley, T. M., & Wolfe, A. D. (2001). Population genetic structure of the cleistogamous plant species Viola pubescens Aiton (Violaceae), as indicated by allozyme and ISSR molecular markers. Heredity, 86(5), 545-556.

182

Culley, T. M. (2005). Characterization of newly developed microsatellite loci in the stemmed yellow violet, Viola pubescens (Violaceae). Molecular Ecology Notes, 5(4), 882-884.

Culley, T. M., Sbita, S. J., & Wick, A. (2007). Population genetic effects of urban habitat fragmentation in the perennial herb Viola pubescens (Violaceae) using ISSR markers. Annals of Botany, 100(1), 91-100.

Culley, T. M. (2000). Inbreeding depression and floral type fitness differences in Viola canadensis (Violaceae), a species with chasmogamous and cleistogamous flowers. Canadian Journal of Botany, 78(11), 1420-1429.

Culley, T. M. (2002). Reproductive biology and delayed selfing in Viola pubescens (Violaceae), an understory herb with chasmogamous and cleistogamous flowers. International Journal of Plant Sciences, 163(1), 113-122.

Cuthbertson, W. (1910). Pansies, Violas & Violets. London: TC & EC Jack.

Dalbato, A. L., Kobza, F., & Karlsson, L. M. (2013). Effect of polyploidy and pollination methods on capsule and seed set of pansies (Viola × wittrockiana Gams). Horticultural Science, 40(1), 22-30.

Dane, F. (2009). Comparative phylogeography of Castanea species. Acta Horticulturae 844, 211-222.

Daniel, P. (1996). Deep'n As It Come: The 1927 Mississippi River Flood. New York, N.Y.: Oxford University Press.

Darrigues, A., Hall, J., van der Knaap, E., & Francis, D. M. (2008). Tomato Analyzer- Color Test: A new tool for efficient digital phenotyping. Journal of the American Society of Horticultural. Science, 133, 579-586.

Darwin, C. (1868). The Variation of Animals and Plants Under Domestication. London: John Murray. de Lafontaine, G., Turgeon, J., & Payette, S. (2010). Phylogeography of white spruce (Picea glauca) in eastern North America reveals contrasting ecological trajectories. Journal of Biogeography, 37(4), 741-751. de Vicente, M. C., Guzmán, F. A., Engels, J., & Rao, V. R. (2006). 12. ‘Genetic characterization and its use in decision-making for the conservation of crop germplasm’. In Ruane, J. and A. Sonnino (Eds.), The Role of Biotechnology in Exploring and Protecting Agricultural Genetic Resources (pp. 129-138). Rome, Italy: FAO. 183

Debener, T. (2002). Molecular markers as a tool for analyses of genetic relatedness and selection in ornamentals. Acta Horticuturae, 522, 121-127.

Debener, T., & Mattiesch, L. (1999). Construction of a genetic linkage map for roses using RAPD and AFLP markers. Theoretical and Applied Genetics, 99(5), 891- 899.

Depypere, L., Chaerle, P., Vander Mijnsbrugge, K., & Goetghebeur, P. (2007). Stony endocarp dimension and shape variation in Prunus section Prunus. Annals of Botany, 100(7), 1585-1597.

Devran, Z., & Baysal, O. (2012). Genetic characterization of Meloidogyne incognita isolates from Turkey using sequence-related amplified polymorphism (SRAP). Biologia, 67(3), 535-539.

Devran, Z., Firat, A. F., Tor, M., Mutlu, N., & Elekcioglu, I. H. (2011). AFLP and SRAP markers linked to the mj gene for root-knot nematode resistance in cucumber. Scientia Agricola, 68(1), 115-119.

DeWoody, A. J., Schupp, J., Kenefic L., Busch, J., Murfitt, L, & Keim, P. L. (2004). Universal method for producing ROX-labeled size standards suitable for automated genotyping. BioTechniques, 37, 348-350.

Dice, L. (1945). Measures of the amount of ecologic association between species. Ecology, 26(3), 297-302.

Ding, G., Zhang, D., Ding, X., Zhou, Q., Zhang, W., & Li, X. (2008). Genetic variation and conservation of the endangered Chinese endemic herb Dendrobium officinale based on SRAP analysis. Plant Systematics and Evolution, 276(3-4), 149-156.

Ding, G., Li, X., Ding, X., & Qian, L. (2009). Genetic diversity across natural populations of Dendrobium officinale, the endangered medicinal herb endemic to China, revealed by ISSR and RAPD markers. Russian Journal of Genetics, 45(3), 327-334.

Ding, W., Cao, Z., & Cao, L. (2010). Molecular analysis of grass carp (Ctenopharyngodon idella) by SRAP and SCAR molecular markers. Aquaculture International 18(4), 575-587.

Dinler, G., & Budak, H. (2008). Analysis of expressed sequence tags (ESTs) from Agrostis species obtained using sequence related amplified polymorphism. Biochemical Genetics, 46(9-10), 663-676.

184

Doebley, J. (2004). The genetics of maize evolution. Annual Reviews of Genetics, 38, 37- 59.

Duvick, D. N. (1984). Genetic diversity in major farm crops on the farm and in reserve. Economic Botany, 38(2), 161-178.

Dong, A., Wang, T., Xu, J., & Zhao, L. (2012). Genetic relationship of germplasm resources of Salvia splendens using SRAP marker. Journal of Beijing Forestry University, 34(5), 134-138.

Dong, P., Wei, Y. M., Chen, G. Y., Li, W., Wang, J. R., Nevo, E., & Zheng, Y. L. (2010). Sequence-related amplified polymorphism (SRAP) of wild emmer wheat (Triticum dicoccoides) in Israel and its ecological association. Biochemical Systematics and Ecology, 38(1), 1-11.

Dong, P., Wei, Y. M., Chen, G. Y., Li, W., Wang, J. R., Nevo, E., & Zheng, Y. L. (2010). Sequence-related amplified polymorphism (SRAP) of wild emmer wheat (Triticum dicoccoides) in Israel and its ecological association. Biochemical Systematics and Ecology, 38(1), 1-11.

Donn, W. L., Farrand, W. R., & Ewing, M. (1962). Pleistocene ice volumes and sea-level lowering. The Journal of Geology, 70(2), 206-214.

Dorken, M. E., & Barrett, S. C. H. (2004). Chloroplast haplotype variation among monoecious and dioecious populations of Sagittaria latifolia (Alismataceae) in eastern North America. Molecular Ecology, 13(9), 2699-2707.

Dou, M., Qiu, W., Wu, Q., & Sun, W. (2010). Genetic diversity analysis on pineapple by SRAP markers. Journal of Fruit Science, 27(6), 930-937.

Doyle, J. J. & Doyle, J. L. (1987). A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin, 19, 11-15.

Du, P., Cui, B. K., & Dai, Y. C. (2011). Genetic diversity of wild Auricularia polytricha in Yunnan Province of South-western China revealed by sequence-related amplified polymorphism (SRAP) analysis. Journal of Medicinal Plant Research, 5, 1374-1381.

Du, X. H., Liu, H. C., Zhu, F. F., & Zhang, F. J. (2011). Clustering of pansy accessions based on principal component analysis. Northern Horticulture, 5, 180-184.

Dunn, G. & Everitt, B.S. (1982). An Introduction to Mathematical Taxonomy. Cambridge University Press.

185

Dyke, A. S., & Prest, V. K. (1987). Late Wisconsinan and Holocene history of the Laurentide Ice Sheet. Géographie Physique et Quaternaire, 41(2), 237-263.

Earl, D. A., & von Holdt, B. M. (2012). STRUCTURE HARVESTER: A website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation Genetics Resources, 4(2), 359-361.

East, E. M. (1916). Studies on size inheritance in Nicotiana. Genetics, 1(2), 164-176.

Echt, C., DeVerno, L., Anzidei, M., & Vendramin, G. (1998). Chloroplast microsatellites reveal population genetic diversity in red pine, Pinus resinosa Ait. Molecular Ecology, 7(3), 307-316.

Ehrich D., Alsos I. G., & Brochmann, C. (2008). Where did the northern peatland species survive the dry glacials: cloudberry (Rubus chamaemorus) as an example. Journal of Biogeography, 35, 801-814.

Eidesen, P. B., Carlsen, T., Molau, U., & Brochmann, C. (2007). Repeatedly out of Beringia: Cassiope tetragona embrace the Arctic. Journal of Biogeography, 34, 1559-1574.

Ellstrand, N. C. (1992). Gene flow by pollen: implications for plant conservation genetics. Oikos, 77-86.

Erben, M. (1985). Cytotaxonomische untersuchungen an sudosteruropaischen Viola- Arten der section Melanium. In D. Podlech (Ed.), Mitteilungen der Botanischen Staatssammlung Munchen (pp. 339-740). Munich.

Esposito, M. A., Martin, E. A., Cravero, V. P., & Cointry, E. (2007). Characterization of pea accessions by SRAP's markers. Scientia Horticulturae, 113(4), 329-335.

Evanno, G., Regnaut, S., & Goudet, J. (2005). Detecting the number of clusters of individuals using the software STRUCTURE: A simulation study. Molecular Ecology, 14(8), 2611-2620.

Eyles, N., & Westgate, J. A. (1987). Restricted regional extent of the Laurentide Ice Sheet in the Great Lakes basins during early Wisconsin glaciation. Geology, 15(6), 537-540.

Falloon, R. E., Viljanenes-Rollinson, S. L. H., Coles, G. D., & Poff, J. D. (1995). Disease severity keys for powdery and downy mildews of pea, and powdery scab of potato. New Zealand Journal of Crop and Horticultural Science, 23(1), 31-37.

186

Falush, D., Stephens., M., & Pritchard, J. K. (2007). Inference of population structure using multilocus genotype data: Dominant markers and null alleles. Molecular Ecology Notes, 7(4), 574–578.

Fan, H. H., Li, T. C., Qiu, J., Lin, Y., & Cai, Y. P. (2010). Comparison of SRAP and RAPD markers for genetic analysis of plants in Dendrobium sw. Chinese Traditional and Herbal Drugs, 41(4), 627-632.

Fang, D. Q., Roose, M. L., Krueger, R. R., & Federici, C. T. (1997). Fingerprinting trifoliate orange germ plasm accessions with isozymes, RFLPs, and inter-simple sequence repeat markers. Theoretical and Applied Genetics, 95(1-2), 211-219.

Fang, H. L., Guo, Q. S., Shen, H. J., & Li, Y. C. (2012). Genetic diversity evaluation of Chrysanthemum indicum L. by medicinal compounds and molecular biology tools. Biochemical Systematics and Ecology, 41, 26-34.

Federal Writers' Project. (1941). Arkansas: A Guide to the State. New York, NY. Hastings House Publishers.

Federici, C. T., Fang, D. Q., Scora, R. W., & Roose, M. L. (1998). Phylogenetic relationships within the genus Citrus (Rutaceae) and related genera as revealed by RFLP and RAPD analysis. Theoretical and Applied Genetics, 96(6-7), 812-822.

Felsenstein, 1. (1989). PHYLIP - phylogeny inference package (version 3.2). Cladistics, 5, 164-166.

Feng, F. J., Han, S. J., & Wang, H. M. (2006). Genetic diversity and genetic differentiation of natural Pinus koraiensis population. Journal of Forestry Research, 17(1), 21-24.

Feng, F., Chen, M., Zhang, D., Sui, X., & Han, S. (2009). Application of SRAP in the genetic diversity of Pinus koraiensis of different provenances. African Journal of Biotechnology, 8(6). 1000-1008.

Feng, N., Xue, Q., Guo, Q., Zhao, R., & Guo, M. (2009). Genetic diversity and population structure of Celosia argentea and related species revealed by SRAP. Biochemical Genetics, 47(7-8), 521-532.

Ferriol, M., Pico, B., & Nuez, F. (2003). Genetic diversity of a germplasm collection of Cucurbita pepo using SRAP and AFLP markers. Theoretical and Applied Genetics, 107(2), 271–282.

187

Ferriol, M., Pico, B., & Nuez, F. (2004a). Morphological and molecular diversity of a collection of Cucurbita maxima landraces. Journal of the American Society for Horticultural Science, 129(1), 60-69.

Ferriol, M., Pico, B., Cordova, P., & Nuez, F. (2004b). Molecular diversity of a germplasm collection of squash (Cucurbita moschata) determined by SRAP and AFLP markers. Crop Science, 44(2), 653-664.

Forrest, A., Hollingsworth, M., Hollingsworth, P., Sydes, C., & Bateman, R. (2004). Population genetic structure in European populations of Spirathes romanzoffiana set in the context of other genetic studies on orchids. Heredity, 92, 218-227.

Fries, M. (1962). Pollen profiles of late Pleistocene and recent sediments from Weber Lake, northeastern Minnesota. Ecology, 43(2), 295-308.

Fothergill, P. G. (1938). Studies in Viola, I: The cytology of a naturally-occurring population of hybrids between Viola tricolor L. and Viola lutea Huds. Genetica, 20(1), 159-186.

Fothergill, P. G. (1939). Studies in Viola, II: Some irregularities in natural Viola hybrids. Genetica, 21(3), 153-176.

Fothergill, P. G. (1941). Studies in Viola. III. An Account of the inheritance of certain characters in the progeny of plants of a wild population of Viola hybrids. New Phytologist, 40(2), 139-151.

Fothergill, P. G. (1944). Studies in Viola: IV. The somatic cytology and taxonomy of our British species of the genus Viola. New Phytologist, 43(1), 23-35.

Frett, J. (2009). Mt . Cuba Center Research Report; Coneflowers for the Mid-Atlantic Region. Greenville,DE: Mt. Cuba Center, Inc. 12pp. doi: http://www.mtcubacenter.org/images/PDFs-and- SWFs/Mt_Cuba_Report_Coneflowers_for_Mid-Atlantic.pdf

Friesen, N., Fritsch, R., & Bachmann, K. (1997). Hybrid origin of some ornamentals of Allium subgenus Melanocrommyum verified with GISH and RAPD. Theoretical and Applied Genetics, 95(8), 1229-1238.

Fu, F. Y., Liu, L. Z., Chai, Y. R., Chen, L., Yang, T., Jin, M. Y., Zhang, Z. S., & Li, J. N. (2007). Localization of QTLs for seed color using recombinant inbred lines of Brassica napus in different environments. Genome, 50(9), 840-854.

Fu, L. Z., Zhang, H. Y., Wu, X. Q., Li, H. B., Wei, H. L., Wu, Q. Q., & Wang, L. A. (2010). Evaluation of genetic diversity in Lentinula edodes strains using RAPD, 188

ISSR and SRAP markers. World Journal of Microbiology and Biotechnology, 26(4), 709-716.

Fu, W., Ma, X., Tang, Q., & Mo, C. (2012). Karyotype analysis and genetic variation of a mutant in Siraitia grosvenorii. Molecular Biology Reports, 39(2), 1247-1252.

Fu, X., Ning, G., Gao, L., & Bao, M. (2008). Genetic diversity of Dianthus accessions as assessed using two molecular marker systems (SRAPs and ISSRs) and morphological traits. Scientia Horticulturae, 117(3), 263-270.

Funk, J. L. (2005). Hedychium gardnerianum invasion into Hawaiian montane rainforest: interactions among litter quality, decomposition rate, and soil nitrogen availability. Biogeochemistry, 76(3), 441-451.

Gao, L., Liu, N., Huang, B., & Hua, X. (2008). Phylogenetic analysis and genetic mapping of Chinese Hedychium using SRAP markers. Scientia Horticulturae, 117(4), 369–377.

Gatesy, J., & Arctander, P. (2000). Hidden morphological support for the phylogenetic placement of Pseudoryx nghetinhensis with bovine bovids: A combined analysis of gross anatomical evidence and DNA sequences from five genes. Systematic Biology, 49(3), 515-538.

Gay, M. (1832). Observations on Viola lutea, and V. grandiflora. Annales Des Sciences Naturelles 26, 159-160.

Genders, R. (1958). Pansies, Violas, and Violets. London: John Gifford.

Geng, J. F., Zhu, C. S., Zhang, X. W., Cheng, Y., Zhang, Y. M., & Hou, X. L. (2007). A genetic linkage map of nonheading Chinese cabbage. Journal of the American Society for Horticultural Science, 132(6), 816-823.

Gepts, P. (2004). Crop domestication as a long-term selection experiment. Plant Breeding Reviews, 24(2), 1-44.

Gilbert, J. E., Lewis, R. V., Wilkinson, M. J., & Caligari, P. D. S. (1999). Developing an appropriate strategy to assess genetic variability in plant germplasm collections. Theoretical and Applied Genetics, 98(6-7), 1125-1131.

Gill, F. B., & Slikas, B. (1999). Speciation in North American chickadees: II. Geography of mtDNA haplotypes in Poecile carolinensis. The Auk, 116(1), 274-277.

Gill, F. B., Mostrom, A. M., Mack, A. L. (1993). Speciation in North American chickadees. I. Patterns of mtDNA genetic divergence. Evolution, 47, 195–212. 189

Giribet, G., & Wheeler, W. (2002). On bivalve phylogeny: a high‐level analysis of the Bivalvia (Mollusca) based on combined morphology and DNA sequence data. Invertebrate Biology, 121(4), 271-324.

Glaubitz, J. C., & Moran, G. F. (2000). Genetic tools: The use of biochemical and molecular markers. In A. Young,, D. Boshier, & T. Boyle (Eds.), Forest Conservation Genetics: Principles and Practice (pp. 39-59). Collingwood, Australia: CSIRO Publishing.

Glenn-Lewin, D. C., Laushman, R. H., & Whitson, P. D. (1984). The vegetation of the Paleozoic Plateau, northeastern Iowa. Proceedings of the Iowa Academy of Scrience, 91(1), 22-27.

Godbout, J., Beaulieu, J., & Bousquet, J. (2010). Phylogeographic structure of jack pine (Pinus banksiana; Pinaceae) supports the existence of a coastal glacial refugium in northeastern North America. American Journal of Botany, 97(11), 1903-1912.

Godbout, J., Jaramillo-Correa, J.P., Beaulieu, J. & Bosquet, J. (2005) A mitochondrial DNA minisatellite reveals the postglacial history of jack pine (Pinus banksiana), a broad-range North American conifer. Molecular Ecology, 14, 3497–3512.

Gonzalo, M. J., Brewer, M. T., Anderson, C., Sullivan, D., Gray, S., & van der Knaap, E. (2009). Tomato fruit shape analysis using morphometric and morphology attributes implemented in Tomato Analyzer software program. Journal of the American Society of Horticultural Science, 134, 77-87.

Gonzales, E., Hamrick, J. L., & Chang, S. M. (2008). Identification of glacial refugia in south-eastern North America by phylogeographical analyses of a forest understory plant, Trillium cuneatum. Journal of Biogeography, 35(5), 844-852.

Goody, J. (1993). The Culture of Flowers. Chicago, IL: Cambridge University Press.

Gornitz, V. (Ed.). (2009). Encyclopedia of Paleoclimatology and Ancient Environments. Dordrecht, Netherlands: Springer.

Gort, G., Koopman, W.J., & Stein, A. (2006). Fragment length distributions and collision probabilities for AFLP markers. Biometrics, 62(4), 1107-1115.

Gort, G., van Hintum, T., & van Eeuwijk, F. (2009). Homoplasy corrected estimation of genetic similarity from AFLP bands, and the effect of the number of bands on the precision of estimation. Theoretical and Applied Genetics, 119(3), 397-416.

190

Govindaraju, D. R. (1988). A note on the relationship between outcrossing rate and gene flow in plants. Heredity, 61, 401-404.

Gower, J.C. (1966). Some distance properties of latent root and vector methods used in multivariate analysis. Biometrika, 53, 325-338.

Graham, J., & McNicol, R. J. (1995). An examination of the ability of RAPD markers to determine the relationships within and between Rubus species. Theoretical and Applied Genetics, 90(7-8), 1128-1132.

Griffin, S. R., & Barrett, S. C. (2004). Post-glacial history of Trillium grandiflorum (Melanthiaceae) in eastern North America: Inferences from phylogeography. American Journal of Botany, 91(3), 465-473.

Gulsen, O., Karagul, S., & Abak, K. (2007). Diversity and relationships among Turkish okra germplasm by SRAP and phenotypic marker polymorphism. Biologia, 62(1), 41-45.

Gulsen, O., Sever-Mutlu, S., Mutlu, N., Tuna, M., Karaguzel, O., Shearman, R. C., Riordan, T. P., & Heng-Moss, T. M. (2009). Polyploidy creates higher diversity among Cynodon accessions as assessed by molecular markers. Theoretical and Applied Genetics, 118(7), 1309-1319.

Gulsen, O., Uzun, A., Canan, I., Seday, U., & Canihos, E. (2010). A new citrus linkage map based on SRAP, SSR, ISSR, POGP, RGA and RAPD markers. Euphytica, 173(2), 265-277.

Guo, D. L., & Luo, Z. R. (2006). Genetic relationships of some PCNA persimmons (Diospyros kaki Thunb.) from China and Japan revealed by SRAP analysis. Genetic Resources and Crop Evolution, 53(8), 1597-1603.

Guo, D., Zhang, J., Liu, C., Zhang, G., Li, M., & Zhang, Q. (2012). Genetic variability and relationships between and within grape cultivated varieties and wild species based on SRAP markers. Tree Genetics & Genomes, 8, 789–800.

Guo, L., Liu, X., Liu, X., Yang, Z., Kong, D., He, Y., & Feng, Z. (2012). Construction of genetic map in barley using sequence-related amplified polymorphism markers, a new molecular marker technique. African Journal of Biotechnology, 11(74), 13858-13862.

Guo, X., Wang, X., Yu, X., Sun, X., & Zhou, T. (2011). Molecular identification and genetic variation of 13 herbaceous peony cultivars using SRAP markers. Acta Horticulturae, 918, 333-340.

191

Haines, A. (2011). Flora Novae Angliae. New Haven, CT: Yale University Press.

Halda, J. J., & Waddick, J. W. (2004). The Genus Paeonia. Portland, OR: Timber Press.

Hale, A. L., Farnham, M. W., & Menz, M. A. (2006). Use of PCR-based markers for differentiating elite broccoli inbreds. Journal of the American Society for Horticultural Science, 131, 418-423.

Hale, A. L., Farnham, M. W., Nzaramba, M. N., & Kimbeng, C. A. (2007). Heterosis for horticultural traits in broccoli. Theoretical and Applied Genetics, 115(3), 351-360.

Hammer, O., Harper, D. A. T., & Ryan, P. D. (2001). PAST: paleontological statistics software package for education and data analysis. Palaeontologia Electronica, 4(1), 1-9.

Hamrick, J. L. (1987). Gene flow and distribution of genetic variation in plant populations. In K. M. Urbanska (Ed.) Differentiation Patterns in Higher Plants (pp. 53-67). New York, NY: Academic Press.

Hamrick, J. L., Godt, M. J. W., Brown, A. H. D., Clegg, M. T., Kahler, A. L., & Weir, B. S. (1990). Allozyme diversity in plant species. In Plant Population Genetics, Breeding, and Genetic Resources (pp. 43-63). Sunderland, MA: Sinauer Associates, Inc.

Hamrick, J. L., Godt, M. J. W., & Sherman-Broyles, S. L. (1992). Factors influencing levels of genetic diversity in woody plant species. In W. T. Adams (Ed.), Population Genetics of Forest Trees: International Symposium on POPULATION Genetics of Forest Trees (pp. 95-124). Dordrecht, Netherlands: Springer.

Hamrick, J. L., & Godt, M. J. W. (1996). Effects of life history traits on genetic diversity in plant species. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences, 351(1345), 1291-1298.

Han, J., Liu, G., Chang, R., & Zhang, X. (2012). The sequence-related amplified polymorphism (SRAP) markers linked to the color around the stone (Cs) locus of peach fruit. African Journal of Biotechnology, 11(42), 9911-9914.

Han, X. Y., Wang, L. S., Shu, Q. Y., Liu, Z. A., Xu, S. X., & Tetsumura, T. (2008). Molecular characterization of tree peony germplasm using sequence-related amplified polymorphism markers. Biochemical Genetics, 46, 162–179.

Han, X., Wang, L., Liu, Z. A., Jan, D. R., & Shu, Q. (2008). Characterization of sequence-related amplified polymorphism markers analysis of tree peony bud sports. Scientia Horticulturae, 115(3), 261-267. 192

Hansen, H. P. (1939). Postglacial vegetation of the Driftless Area of Wisconsin. American Midland Naturalist, 21(3), 752-762.

Hao, Q., Liu, Z. A., Shu, Q. Y., Zhang, R. E., De Rick, J., & Wang, L. S. (2008). Studies on Paeonia cultivars and hybrids identification based on SRAP analysis. Hereditas, 145(1), 38-47.

Harding, A. H. (1917). The Book of the Peony. Philadelphia, PA: JB Lippincott Company.

Hartley, T. G. (1966). The Flora of the Driftless Area (Vol. 21, No. 1). University of Iowa Press.

Hastorf, C. A. (2009). Rio Balsas most likely region for maize domestication. Proceedings of the National Academy of Sciences, 106(13), 4957-4958.

Haw, S. G. (2001). Tree peonies: a review of their history and taxonomy. New Plantsman, 8(3), 156-171.

Hawkes, J. G. (1983). The Diversity of Crop Plants. Cambridge, MA: Harvard University Press.

Havran, J. C., Sytsma, K. J., & Ballard, H. E. (2009). Evolutionary relationships, interisland biogeography, and molecular evolution in the Hawaiian violets (Viola: Violaceae). American Journal of Botany, 96(11), 2087-2099.

He, D. H., Lin, Z. X., Zhang, X. L., Nie, Y. C., Guo, X. P., Zhang, Y. X., & Li, W. (2007). QTL mapping for economic traits based on a dense genetic map of cotton with PCR-based markers using the interspecific cross of Gossypium hirsutum × Gossypium barbadense. Euphytica, 153(1-2), 181-197.

He, F. F., Yang, Z. P., Zhang, Z. S., Wang, G. X., & Wang, J. C. (2007). Genetic diversity analysis of potato germplasm by SRAP markers. Journal of Agricultural Biotechnology, 15(6), 1001-1005.

He, Y. H., Ning, G. G., Sun, Y. L., Qi, Y. C., & Bao, M. Z. (2009). Identification of a SCAR marker linked to a recessive male sterile gene (Tems) and its application in breeding of marigold (Tagetes erecta). Plant Breeding, 128(1), 92-96.

Hepenstrick, D. (2009). Tracing allopolyploid evolution with microsatellites in Viola (Master’s thesis). Zürich: Ecological Genetics and Evolution, WSL Swiss Federal Research Institute, Biology.

193

Herrera, C. M. (1993). Selection on complexity of corolla outline in a hawkmoth- pollinated violet. Evolutionary Trends in Plants, 7, 9-13.

Hew, C. S. (2001). Ancient Chinese orchid cultivation: A fresh look at an age-old practice. Scientia Horticulturae, 87(1), 1-10.

Hewitt, G. (2000). The genetic legacy of the Quaternary ice ages. Nature, 405(6789), 907-913.

Heywood, V. H. (2003a). Conservation and sustainable use of wild species as sources of new ornamentals. Acta Horticulturae 598, 43-53.

Heywood, V. H. (2003b). Conservation strategies, plant breeding, wild species and land races. In K. Ammann, Y. Jacot, and R. Braun (Eds.), Methods for Risk Assessment of Transgenic Plants IV. Biodiversity and Biotechnology (pp. 143– 159). Switzerland: Birkhäuser Verlag Basel.

Hildebrandt, U., Hoef-Emden, K., Backhausen, S., Bothe, H., Bozek, M., Siuta, A., & Kuta, E. (2006). The rare, endemic zinc violets of Central Europe originate from Viola lutea Huds. Plant Systematics and Evolution, 257(3-4), 205-222.

Hill, C. B., Li, Y., & Hartman, G. L. (2004). Resistance to the soybean aphid in soybean germplasm. Crop Science, 44(1), 98-106.

Himmelbauer, M. L., Loiskandl, W., & Kastanek, F. (2004). Estimating length, average diameter and surface area of roots using two different image analyses systems. Plant and Soil, 260(1-2), 111-120.

Holgersson, M. (1978). The limited value of cophenetic correlation as a clustering criterion. Pattern Recognition, 10(4), 287-295.

Hong, D. (2010). Peonies of the World: Taxonomy and Phytogeography. Royal Botanic Gardens.

Horita, M., Morohashi, H., & Komai, F. (2003). Production of fertile somatic hybrid plants between Oriental hybrid lily and Lilium formolongi. Planta, 217, 597–601.

Horn, W. (2004). The patterns of evolution and ornamental plant breeding. Acta Horticulturae, 651, 19-31.

Hou, X. G., Yin, W. L., Li, J. J., & Wang, H. F. (2006). AFLP analysis of genetic diversity of 30 tree peony (Paeonia suffruticosa Andr.) cultivars. Scientia Agricultura Sinica, 39(8), 1709-1715.

194

Howes, B. J., Lindsay, B. & Lougheed, S. C. (2006). Range-wide phylogeography of a temperate lizard, the five-lined skink (Eumeces fasciatus). Molecular Phylogenetics and Evolution, 40,183-194.

Hu, D., Zhang, Z., Zhang, D., Zhang, Q., & Li, J. (2005). Genetic relationship of ornamental peach determined using AFLP markers. HortScience, 40(6), 1782- 1786.

Hu, J., & Vick, B. A. (2003). Target Region Amplification Polymorphism: A novel marker technique for plant genotyping. Plant Molecular Biology Reporter, 21, 289-294.

Hu, Q., Andersen, S,. Dixelius, C., & Hansen, L. (2002). Production of fertile intergeneric somatic hybrids between Brassica napus and Sinapis arvensis for the enrichment of the rapeseed gene pool. Plant Cell Reports, 21(2), 147-152.

Hu, S. Y. (1970). Dendrobium in Chinese medicine. Economic Botany, 24(2), 165-174.

Hu, Z., Lv, Y., Wang, C., & Xiang, S. (2012). Analysis on ‘Whole Red’ of patterns Oujiang colour carp (Cyprinus carpio var. color) by SRAP and SCAR markers. Aquaculture Research 43, 588–594.

Huang, D., Li, S. X., Cai, G. X., Yue, C.H., Wei, L.J., & Zhang, P. (2008). Molecular authentication and quality control using a high performance liquid chromatography technique of fructus Evodiae. Biological & Pharmaceutical Bulletin, 31, 312-315.

Huang, C. Q., Zhang, Y. F., Liu, G. D., Bai, C. J., & Wang, W. Q. (2012). Genetic diversity of Cynodon radiatus assessed by sequence-related amplified polymorphism markers. Biochemical Systematics and Ecology, 40, 56-61.

Huang, L. K., Bughrara, S. S., Zhang, X. Q., Bales-Arcelo, C. J., & Bin, X. (2011). Genetic diversity of switchgrass and its relative species in Panicum genus using molecular markers. Biochemical Systematics and Ecology, 39(4), 685-693.

Huelsenbeck, J. P., Bull, J. J., & Cunningham, C. W. (1996). Combining data in phylogenetic analysis. Trends in Ecology & Evolution, 11(4), 152-158.

Hutchison, L. J., & Kavanagh, K. (1994). The bird's foot violet (Viola pedata L.) in Canada: population biology and ecology of a threatened species. Michigan Botanist, 33(1), 3-16.

Hyde, T. & Staffen, A. (2012). Rapid Ecological Assessment for Driftless Area Study Streams: A Rapid Ecological Assessment Focusing on Rare Plants, Selected Rare 195

Animals, and High-quality Natural Communities. Wisconsin Natural Heritage Inventory, PUB-ER-836, 193pp.

Ikten, H., Mutlu, N., Gulsen, O., Kocatas, H., & Aksoy, U. (2010). Elucidating genetic relationships, diversity and population structure among the Turkish female figs. Genetica, 138(2), 169-177.

Ipek, M., Ipek, A., & Simon, P. W. (2006). Sequence homology of polymorphic AFLP markers in garlic (Allium sativum L.). Genome, 49(10), 1246-55.

Irzykowska, L., Weber, Z., & Bocianowski, J. (2012). Comparison of Claviceps purpurea populations originated from experimental plots or fields of rye. Central European Journal of Biology, 7(5), 839-849.

Isk., N., Doganlar, S., & Frary, A. (2011). Genetic diversity of Turkish olive varieties assessed by simple sequence repeat and sequence-related amplified polymorphism markers. Crop Science, 51(4), 1646-1654.

Iwata, H., & Ukai, Y. (2002). SHAPE: A computer program package for quantitative evaluation of biological shapes based on elliptic Fourier descriptors. Journal of Heredity, 93(5), 384-385.

Jackson, N. D., & Austin, C. C. (2010). The combined effects of rivers and refugia generate extreme cryptic fragmentation within the common ground skink (Scincella lateralis). Evolution, 64(2), 409-428.

Jackson, S. T., & Overpeck, J. T. (2000). Responses of plant populations and communities to environmental changes of the late Quaternary. Paleobiology, 26(sp4), 194-220.

Jackson, S. T. & Williams, J. W. (2004). Modern analogs in Quaternary paleoecology: here today, gone yesterday, gone tomorrow?. Annual Review of Earth and Planetary Sciences 32, 495–537.

Jacobs, D. K., Haney, T. A., & Louie, K. D. (2004). Genes, diversity, and geologic process on the Pacific coast. Annual Review of Earth and Planetary Sciences, 32, 601-652.

Jakobsson, M., & Rosenberg, N. A. (2007). CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics, 23, 1801-1806.

Jan, C. H., Byrne, D. H., Manhart, J., & Wilson, H. (1999). Rose germplasm analysis with RAPD markers. HortScience, 34(2), 341-345. 196

Janick, J. (1998). Hybrids in horticultural crops. In K. R. Larnkey and J. E. Staub (Eds.), Concepts and Breeding of Heterosis in Crop Plants (pp. 45-56). Madison, WI: Crop Science Society of America.

Jaramillo-Correa, J. P., Beaulieu, J., & Bousquet, J. (2004). Variation in mitochondrial DNA reveals multiple distant glacial refugia in black spruce (Picea mariana), a transcontinental North American conifer. Molecular Ecology, 13(9), 2735-2747.

Jaramillo-Correa, J. P., Beaulieu, J., Khasa, D. P., & Bousquet, J. (2009). Inferring the past from the present phylogeographic structure of North American forest trees: seeing the forest for the genes. Canadian Journal of Forest Research, 39(2), 286- 307.

Jerardo, A. (2007). Floriculture and Nursery Crops Yearbook. Economic Research Service, United States Department of Agriculture, Washington, DC, USA.

Jian, L.., & Zhu, L. (2010). Analysis on Cymbidium kanran with SRAP markers. Scientia Agricultura Sinica, 43, 3184-3190.

Jiajue, L. (1998). Studies on the origin of Chinese Mudan (Tree Peony). Journal of Beijing Forestry University, 20(2), 22-26.

Johnson, G. P. (1988). Revision of Castanea sect. Balanocastanon (Fagaceae). Journal of the Arnold Arboretum, 69, 25-49.

Jonah, P. M., Bello, L. L., Lucky, O., Midau, A., Moruppa, S. M. (2011). Review : The importance of molecular markers in plant breeding programmes. Global Journal of Science Frontier Research, 11(5), 1-9.

Joly, S., & Bruneau, A. (2004). Evolution of triploidy in Apios Americana (Leguminosae) revealed by genealogical analysis of the histone H3-D gene. Evolution, 58, 284- 295.

Jones, C., Edwards, K., Castaglione, S., Winfield, M., Sala, F., van deWiel, C., Bredemeijer, G., Vosman, B., Matthes, M., Daly, A., Brettschneider, R., Bettini, P., Buiatti, M., Maestri, E., Malcevschi, A., Marmiroli, N., Aert, R., Volckaert, G., Rueda, J., Linacero, R., Vazquez, A., & Karp, A. (1997). Reproducibility testing of RAPD, AFLP and SSR markers in plants by a network of European laboratories. Molecular Breeding, 3, 381-390.

Kantety, R. V., Rota, M. L., Matthews, D. E., & Sorrells, M. E. (2002). Data mining for simple sequence repeats in expressed sequence tags from barley, maize, rice, sorghum and wheat. Plant Molecular Biology, 48(5-6), 501-510. 197

Karp, A. (1997). Molecular Tools in Plant Genetic Resources Conservation: A Guide to the Technologies (No. 2). Rome, Italy: International Plant Genetic Resources Institute.

Kavanagh, K. C., Hutchison, L., & Varga, S. (1989). Status report on bird's-foot violet Viola pedata L. in Canada. Prepared for the Committee on the Status of Endangered Willdlife in Canada (COSEWIC).

Kennedy, A. H., & Walker, G. L. (2007). The population genetic structure of the showy lady's-slipper orchid (Cypripedium reginae Walter) in its glaciated and unglaciated ranges. Castanea, 72(4), 248-261.

Kerguelen, V., & Hoddle, M. S. (1999). Measuring mite feeding damage on avocado leaves with automated image analysis software. The Florida Entomologist, 82(1), 119-122.

Kim, S. T., Sultan, S. E., & Donoghue, M. J. (2008). Allopolyploid speciation in Persicaria (Polygonaceae): Insights from a low-copy nuclear region. Proceedings of the National Academy of Sciences, 105(34), 12370-12375.

Kim, Z. S., Hwang, J. W., Lee, S. W., Yang, C., & Gorovoy, P. G. (2005). Genetic variation of Korean Pine (Pinus koraiensis Sieb. et Zucc.) at allozyme and RAPD markers in Korea, China and Russia. Silvae Genetica, 54(4-5), 235-245.

Khush, G. S. (1996). Haploids in plant breeding. In J. S. Mohan, S. K. Sopory, and R. E. Veilleux (Eds.), In Vitro Haploid Production in Higher Plants: Volume 1: Fundamental Aspects and Methods. Dordrecht, Netherlands: Springer.

Klaber, D. (1976). Violets of the United States. New York, NY: AS Barnes.

Kleeberger, K. S., & Moser, B. C. (2002). Flatbed scanners: An alternative tool for gathering horticultural data. HortTechnology, 12(3), 444-446.

Klein, A-M., Steffan-Dewenter, I., & Tscharntke, T. (2003). Fruit set of highland coffee increases with the diversity of pollinating bees. Proceedings of the Royal Society of London. Series B: Biological Sciences, 270(1518), 955-961.

Klingenberg, C. P. (2011). MorphoJ: An integrated software package for geometric morphometrics. Molecular Ecology Resources, 11(2), 353-357.

Kochieva, E. Z., Ryzhova, N. N., Molkanova, O. I., Kudryavtsev, A. M., Upelniek, V. P., & Okuneva, I. B. (2004). The genus Syringa: Molecular markers of species and cultivars. Russian Journal of Genetics, 40(1), 30-32. 198

Koopman, W. J. (2005). Phylogenetic signal in AFLP data sets. Systematic Biology, 54(2), 197-217.

Koopman, W. J., Wissemann, V., De Cock, K., Van Huylenbroeck, J., De Riek,J., Sabatino, G. J., Visser, D., Vosman, B., Ritz, C. M., Maes, B., Werlemark, G., Nybom, H., Debener, T., Linde, M., Marinus J. M., & Smulders M. J. (2008). AFLP markers as a tool to reconstruct complex relationships: A case study in Rosa (Rosaceae). American Journal of Botany, 95(3), 353–366.

Kosman, E., & Leonard, K. J. (2005). Similarity coefficients for molecular markers in studies of genetic relationships between individuals for haploid, diploid, and polyploid species. Molecular Ecology, 14(2), 415-424.

Kristofferson, K. B. (1923). Crossings in Melanium violets. Hereditas, 4(3), 251-289.

Kroon, G. H. (1972). Reduction of ploidy level of tetraploid large-flowered garden pansies (Viola × wittrockiana Gams.) to diploid level after crossing with diploid V. tricolor L. Euphytica, 21(2), 165-170.

Kuhl, F. P., & Giardina, C. R. (1982). Elliptic Fourier features of a closed contour. Computer Graphics and Image Processing, 18(3), 236-258.

Kumar, S. A., Sudisha, J., & Sreenath, H. L. (2008). Genetic relation of Coffea and Indian Psilanthus species as revealed through RAPD and ISSR markers. International Journal of Integrative Biology, 3(3), 150.

Lakshmi, M., Senthilkumar, P., Parani, M., Jithesh, M. N., & Parida, A. K. (2000). PCR- RFLP analysis of chloroplast gene regions in Cajanus (Leguminosae) and allied genera. Euphytica, 116(3), 243-250.

Langlade, N. B., Feng, X., Dransfield, T., Copsey, L., Hanna, A. I., Thebaud, C., Bangham, A., Hudson, A., & Coen, E. (2005). Evolution through genetically controlled allometry space. Proceedings of the National Academy of Sciences of the United States of America, 102(29), 10221-10226.

Lashermes, P., Andrzejewski, S., Bertrand, B., Combes, M.-C., Dussert, S., Graziosi, G., Trouslot, P., & Anthony, F. (2000). Molecular analysis of introgressive breeding in coffee (Coffea arabica L.). Theoretical and Applied Genetics, 100(1), 139-146.

Lashermes, P., Trouslot, P., Anthony, F., Combes, M. C., & Charrier, A. (1996). Genetic diversity for RAPD markers between cultivated and wild accessions of Coffea arabica. Euphytica, 87(1), 59-64.

199

Latta, R., & Mitton, J. (1997). A comparison of population differentiation across four classes of gene marker in limber pine (Finus flexilis James). Genetics, 146, 1153- 1163.

Lawson, R. (1987). Molecular studies of thamnophiine snakes. I. The phylogeny of the genus Nerodia. Journal of Herpetology, 21, 140–157.

Lesur C., Boury S., Weising K., Kahl G., Wolff, K., Becher S. A. & Peltier D. (2000). Comparison of seven molecular techniques for Pelargonium cultivar identification. Acta Horticulture, 508, 297-301.

Levi, A., Harris, K. R., Wechter, W. P., Kousik, C. S., & Thies, J. A. (2010). DNA markers and pollen morphology reveal that Praecitrullus fistulosus is more closely related to Benincasa hispida than to Citrullus spp. Genetic Resources and Crop Evolution, 57(8), 1191-1205.

Levi, A., & Thomas, C. E. (2007). DNA markers from different linkage regions of watermelon genome useful in differentiating among closely related watermelon genotypes. HortScience, 42(2), 210-214.

Levi, A., Thomas, C. E., Trebitsh, T., Salman, A., King, J., Karalius, J., Newman, M., Reddy, O. U., Xu, Y., & Zhang, X. (2006). An extended linkage map for watermelon based on SRAP, AFLP, SSR, ISSR, and RAPD markers. Journal of the American Society for Horticultural Science, 131(3), 393-402.

Levi, A., Wechter, P., Massey, L., Carter, L., & Hopkins, D. (2011). An extended genetic linkage map for watermelon based on a testcross and a BC2F2 population. American Journal of Plant Sciences, 2(2), 93-110.

Levin, D. A. (1976). Consequences of long-term artificial selection, inbreeding and isolation in Phlox. II. The organization of allozymic variability. Evolution, 30(3), 463-472.

Lewis, P. O., & Crawford, D. J. (1995). Pleistocene refugium endemics exhibit greater allozymic diversity than widespread congeners in the genus Polygonella (Polygonaceae). American Journal of Botany, 82(2), 141-149.

Lewton, K. L. (2012). Evolvability of the primate pelvic girdle. Evolutionary Biology, 39(1), 126-139.

Li, C., Zhou, A., & Sang, T. (2006). Rice domestication by reducing shattering. Science, 311(5769), 1936-1939.

200

Li, G., & Quiros, C. F. (2001). Sequence-related amplified polymorphism (SRAP), a new marker system based on a simple PCR reaction: its application to mapping and gene tagging in Brassica. Theoretical and Applied Genetics, 103(2-3), 455-461.

Li, H., Yin, Y., Zhang, C., Zhang, M., & Li, J. (2007). Evaluation of application of SRAP on analysis of genetic diversity in cultivars of Allium fistulosum L. Acta Horticulturae Sinica, 34(4), 929-934.

Li, H., Chen, H., Zhuang, T., & Chen, J. (2010). Analysis of genetic variation in eggplant and related Solanum species using sequence-related amplified polymorphism markers. Scientia Horticulturae, 125(1), 19-24.

Li, H., Ruan, C., Teixeira da Silva, J., & Liu, B. (2010). Associations of SRAP markers with dried-shrink disease resistance in a germplasm collection of sea buckthorn (Hippophae L.). Genome, 53(6), 447-57.

Li, L., Peng, J., & Bai, R. (2009). Analysis on the genetic relationships of Chinese Ziziphus with SRAP markers. Scientia Agricultura Sinica, 42(5), 1713-1719.

Li, M. F., Hirata, Y., Xu, G. J., Niwa, M., & Wu, H. M. (1991). Studies on the chemical constituents of Dendrobium loddigesii Rolfe. Acta Pharmaceutica Sinica, 26(4), 307.

Li, P., Li, M., Shi, Y., Zhao, Y., Wan, Y., Fu, C., & Cameron, K. M. (2013). Phylogeography of North American herbaceous Smilax (Smilacaceae): Combined AFLP and cpDNA data support a northern refugium in the Driftless Area. American Journal of Botany, 100(4), 801-814.

Li, Q., Dong, S., Zhang, W., Lin, R., Wang, C., Qian, D., Lun, Z., Song, H., & Zhu, X. (2009). Sequence-related amplified polymorphism, an effective molecular approach for studying genetic variation in Fasciola spp. of human and animal health significance. Electrophoresis, 30, 403-409.

Li, Q., Liu, X., Su, J., & Tao, J. (2010). Genetic diversity of Acer L. germplasm revealed by sequence-related amplified polymorphism (SRAP) markers. Jiangsu Journal of Agricultural Sciences, 26(5), 1032-1036.

Li, S., Jia, J., Wei, X., Zhang, X., Li, L., Chen, H., Fan, Y., Sun, H., Zhao, X., Lei, T., Xu, Y., Jiang, F., Wang, H., & Li, L. (2007). An intervarietal genetic map and QTL analysis for yield traits in wheat. Molecular Breeding, 20(2), 167-178.

Li, W., Zhang, J., Mou, J., Geng, J., McVetty. P., Hu, S., & Genyi Li, G. (2011). Integration of Solexa sequences on an ultradense genetic map in Brassica rapa L. BMC Genomics, 12, 249. 201

Li, Y., Shen, J., Wang, T., Fu, T., & Ma, C. (2007). Construction of a linkage map using SRAP, SSR and AFLP markers in Brassica napus L. Scientia Agricultura Sinica, 40(6), 1118-1126.

Li, Y., Fan, X., Shi, T., Zhang, Q., & Zhang, Z. (2009). SRAP marker reveals genetic diversity in tartary buckwheat in China. Frontiers of Agriculture in China, 3(4), 383-387.

Li, Y., Liu, Z., Wang, Y., Yang, N., Xin, X., Yang, S., & Feng, H. (2012). Identification of quantitative trait loci for yellow inner leaves in Chinese cabbage (Brassica rapa L. ssp. pekinensis) based on SSR and SRAP markers. Scientia Horticulturae, 133, 10-17.

Li, Z., ZhongHua, T., XianYuan, L., ShunZhao, S., & MingYang, L. (2011). Phylogenetic relationship analysis of 23 wild species of Lilium by SRAP markers. Journal of Agricultural Biotechnology, 19(4), 677-684.

Liang, C. (2011). Genetic Diversity Analysis and Suitability Evaluation and Regionalization of Euodia rutaecarpa (Masters thesis). Hunan University of Traditional Chinese Medicine: College of Pharmacy, Changsha, Hunan, China.

Liang, G., Xiong, G., Guo, Q., He, Q. & Li, X. (2007). AFLP analysis and the taxonomy of Citrus. Acta Horticulturae, 760, 137-142.

Liang, J., Teng, J. B., & Lao, W. R. (2010). Determination of polysaccharides in Dendrobium loddigesii from different cultivation and growth time. West China Journal of Pharmaceutical Sciences, 1, 27.

Liao, L., Guo, Q. S., Wang, Z. Y., Liu, L., & Zhu, Z. B. (2012). Genetic diversity analysis of Prunella vulgaris in China using ISSR and SRAP markers. Biochemical Systematics and Ecology, 45, 209-217.

Lim, K.-B., Ramanna, M., Jacobsen, E., & van Tuyl, J. (2003). Evaluation of BC2 progenies derived from 3x-2x and 3x-4x crosses of Lilium hybrids: A GISH analysis. Theoretical and Applied Genetics, 106(3), 568-574.

Lin, X., Lou, Y., Zhang, Y., Yuan, X., He, J., & Fang, W. (2011). Identification of genetic diversity among cultivars of Phyllostachys violascens using ISSR, SRAP and AFLP markers. The Botanical Review, 77(3), 223-232.

Lin, Z., Zhang, X., & Nie, Y. (2004). Evaluation of application of a new molecular marker SRAP on analysis of F2 segregation population and genetic diversity in cotton. Acta Genetica Sinica, 31(6), 622-626. 202

Lin, Z., Zhang, X., Nie, Y., He, D., & Wu, M. (2003). Construction of a genetic linkage map for cotton based on SRAP. Chinese Science Bulletin, 48(19), 2064-2068.

Lin, Z., Zhang, Y., Zhang, X., & Guo, X. (2009). A high-density integrative linkage map for Gossypium hirsutum. Euphytica, 166(1), 35-45.

Linde, M., Yan, Z., & Debner, T. (2007). Ornamentals. In C. Kole. (Ed.), Technical Crops (Genome Mapping and Molecular Breeding in Plants). Dordrecht, Netherlands: Springer.

Lira-Saade, R. (1995). Estudios Taxonómicos y Ecogeográficos de las Cucurbitaceae Latinoamericanas de Importancia Económica. Systematic and Ecogeographic Studies on Crop Genepools. No. 9. International Plant Genetic Resources Institute, Rome, Italy.

Liu, H. J., Xu, X. H., & Jiang, T. B. (2011). Genetic diversity analysis of wild Auricularia auricula strains in Daxing’anling area using SRAP Markers. Scientia Agricultura Sinica, 44(13), 2641-2649.

Liu, L., Zhu, X., Gong, Y., Song, X., Wang, T., Zhao, L., & Wang, L. (2007). Genetic giversity analysis of radish germplasm with RAPD, AFLP and SRAP markers. Acta Horticulturae, 760, 125-130.

Liu, L., Liu, G., Gong, Y., Dai, W., Wang, Y., Yu, F., & Ren, Y. (2007) Evaluation of genetic purity of F1 hybrid seeds in cabbage with RAPD, ISSR, SRAP, and SSR markers. Hortscience, 42(3), 724–727.

Liu, L. J., Peng, D. X., & Wang, B. (2008). Genetic relation analysis on ramie [Boehmeria nivea (L.) Gaud.] inbred lines by SRAP markers. Agricultural Sciences in China, 7(8), 944-949.

Liu, L. W., Zhao, L. P., Gong, Y. Q., Wang, M. X., Chen, L. M., Yang, J. L., Wang, Y., Yu, F. M., & Wang, L. Z. (2008). DNA fingerprinting and genetic diversity analysis of late-bolting radish cultivars with RAPD, ISSR and SRAP markers. Scientia Horticulturae, 116(3), 240-247.

Liu, L., Ma, X., Wei, J., Qin, J., & Mo, C. (2010). The first genetic linkage map of Luohanguo (Siraitia grosvenorii) based on ISSR and SRAP markers. Genome, 54(1), 19-25.

Liu, L., Liu, Z., Chen, H., Zhou, L., Liu, Y., & Luo, L. (2011). SRAP markers and morphological traits could be used in test of distinctiveness, uniformity, and

203

stability (DUS) of lettuce (Lactuca sativa) varieties. Journal of Agricultural Science, 4(3), 227-236.

Liu Y., Yan H., Yang W., Li X., Li Y., Meng Q., Zhang L., Liu D., & Zhang T. (2008). Molecular diversity of 23 wheat leaf rust resistance near-isogenic lines determined by sequence-related amplified polymorphism. Scientia Agricultura Sinica, 41(5), 1333-1340.

Liu, Z., Liu, Y., Liu, S., Ding, X., & Yang, Y. (2009). Analysis of the sequence of ITS1- 5.8 S-ITS2 Regions of the three species of fructus Evodiae in Guizhou Province of China and identification of main ingredients of their medicinal chemistry. Journal of Computer Science & Systems Biology, 2, 200-207.

Liu, Z., Shu, Q., Wang, L., Yu, M., Hu, Y., Zhang, H., Tao, Y., & Shao, Y. (2012). Genetic diversity of the endangered and medically important Lycium ruthenicum Murr. revealed by sequence-related amplified polymorphism (SRAP) markers. Biochemical Systematics and Ecology, 45, 86-97.

Loh, J. P., Kiew, R., Hay, A., Kee, A., Gan, L. H., & Gan, Y. (2000). Intergeneric and interspecific relationships in Araceae tribe Caladieae and development of molecular markers using amplified fragment length polymorphism (AFLP). Annals of Botany, 85, 371-378.

Lou, Y., Lin, X., He Q., Guo, X., Huang, L., & Fang, W. (2010). Analysis on genetic relationship of Puji-bamboo species by AFLP and SRAP. Molecular Plant Breeding, 8(1), 83-88.

Loveless, M. D., & Hamrick, J. L. (1984). Ecological determinants of genetic structure in plant populations. Annual review of Ecology and Systematics, 15, 65-95.

Lu, F., Yu, N., Zhao, X., Zhang, J., Xiao, W., Li, W., Liu, R., Zhu, J., & Peng, C. (2011). Genetic diversity analysis of Phalaenopsis 'Frigdaas Oxford' using SRAP markers with reference to those genes responsible for variations in the pigmentation of petals and sepals. Journal of Horticultural Science and Biotechnology, 86(5), 486-492.

Lu, J., Zhao, H., Suo, N., Wang, S., Shen, B., Wang, H., & Liu, J. (2012). Genetic linkage maps of Dendrobium moniliforme and D. officinale based on EST-SSR, SRAP, ISSR and RAPD markers. Scientia Horticulturae, 137, 1-10.

Ma, D., Yang, G., Mu, L., & Song, Y. (2010). Application of SRAP in the genetic diversity of Tricholoma matsutake in northeastern China. African Journal of Biotechnology, 9(38), 6244-6250.

204

Ma, H., Olsen, R., Pooler, M., & Kramer, M. (2009). Evaluation of flowering cherry species, hybrids, and cultivars using simple sequence repeat markers. Journal of the American Society for Horticultural Science, 134(4), 435-444.

Ma, X., Wang, G., & Han, L. (2012). Genetic diversity in lilyturf germplasm detected by SRAP markers. Pratacultural Science, 29(11), 1686-1691.

Magni, C. R., Ducousso, A., Caron, H., Petit, R. J., & Kremer, A. (2005). Chloroplast DNA variation of Quercus rubra L. in North America and comparison with other Fagaceae. Molecular Ecology, 14(2), 513-524.

Maheswaran, M. (2004). Molecular markers: History, features and applications. Advanced Biotech, 17-34.

Mallet, J. (2007). Hybrid speciation. Nature, 446(7133), 279-283.

Marcos Filho, J., Gomes Junior, F. G., Bennett, M. A., Wells, A. A., & Stieve, S. (2010). Using Tomato Analyzer software to determine embryo size in x-rayed seeds. Revista Brasileira de Sementes, 32(2), 146-153.

Marcussen, T., Blaxland, K., Windham, M. D., Haskins, K. E., & Armstrong, F. (2011). Establishing the phylogenetic origin, history, and age of the narrow endemic Viola guadalupensis (Violaceae). American Journal of Botany, 98(12), 1978- 1988.

Marcussen, T., Jakobsen, K., Danihelka, J., Ballard, H., Blaxland, K., Brysting, A., & Oxelman, B. (2012). Inferring species networks from gene trees in high-polyploid North American and Hawaiian violets (Viola, Violaceae). Systematic Biology, 61(1), 107–126.

Margesin, R., Neuner, G., & Storey, K, B. (2007). Cold-loving microbes, plants, and animals-fundamental and applied aspects. Naturwissenschaften 94, 77–99.

Martínez-Meyer, E., & Peterson, A. T. (2006). Conservatism of ecological niche characteristics in North American plant species over the Pleistocene-to-recent transition. Journal of Biogeography, 33(10), 1779-1789.

Maxwell, J. A., & Young, D. K. (1998). A significant eastern range extension for Eleodes tricostatus (Say)(Coleoptera: Tenebrionidae). The Coleopterists' Bulletin, 52(1), 90-92.

Mayers, A. M., & Lord, E. M. (1983). Comparative flower development in the cleistogamous species Viola odorata. I. A growth rate study. American Journal of Botany, 70(10), 1548-1555. 205

Mazhara, M., Joycea, D., Hofmanb, P., Collinsa, R., Brereton, I., Cowin, G., & Guptaa, M. (2011). Bruising in the Queensland scupply chains of Hass avocado fruit. Proceedings VII World Avocado Congress 2011 (Actas VII Congreso Mundial del Aguacate 2011). Cairns, Australia.

McKay, B. D. (2009). Evolutionary history suggests rapid differentiation in the yellow- throated warbler Dendroica dominica. Journal of Avian Biology, 40(2), 181-190.

McKinnon, G. E., Vaillancourt, R. E., Steane, D. A., & Potts, B. M. (2008). An AFLP marker approach to lower-level systematics in Eucalyptus (Myrtaceae). American Journal of Botany, 95(3), 368-380.

McLachlan, J. S., Clark, J. S., & Manos, P. S. (2005). Indicators of tree migration capacity under rapid climate change. Ecology, 86(8), 2088-2098.

McLaughlin, S., Bouton, J., Bransby, D., Conger, B., Ocumpaugh, W., Parrish, D., Taliaferro, D., Vogel, K., & Wullschleger, S. (1999). Developing switchgrass as a bioenergy crop. In J. Janick (Ed.), Perspectives On New Crops and New Uses (pp. 282-299). Alexandria, VA: ASHS Press.

McLaughlin, S., & Kszos, L. (2005). Development of switchgrass (Panicum virgatum) as a bioenergy feedstock in the United States. Biomass and Bioenergy, 28(6), 515– 535.

Meekins, J. F., Ballard Jr, H. E., & McCarthy, B. C. (2001). Genetic variation and molecular biogeography of a North American invasive plant species (Alliaria petiolata, Brassicaceae). International Journal of Plant Sciences, 162(1), 161- 169.

Meksem, K., & Kahl, G. (Eds.). (2006). The Handbook of Plant Genome Mapping. Ames, Iowa: Wiley-Blackwell.

Mendel, G. (1866). Versuche über Plflanzenhybriden. Verhandlungen des Naturforschenden Vereines in Brünn, Bd. IV für das Jahr 1865, Abhandlungen, 3- 47.

Meng, H., Chen, S., Cheng, Z., Chai, D., & Li, Y. (2012). SRAP markers for fruit shape in cucumber. Pakistan Journal of Botany, 44(4), 1381-1384.

Mereda, P., Hodalova, I., Martonfi, P., Kucera, J., & Lihova, J. (2008). Intraspecific variation in Viola suavis in Europe: Parallel evolution of white-flowered morphotypes. Annals of Botany, 102(3), 443-462.

206

Mereda, P., Hodalova, I., Kucera, J., Zozomova-Lihova, J., Letz, D. R., & Slovak, M. (2011). Genetic and morphological variation in Viola suavis s.l. (Violaceae) in the western Balkan Peninsula: Two endemic subspecies revealed. Systematics and Biodiversity, 9(3), 211-231.

Merotto Jr, A., Jasieniuk, M., & Fischer, A. J. (2009). Estimating the outcrossing rate of Cyperus difformis using resistance to ALS‐inhibiting herbicides and molecular markers. Weed Research, 49(1), 29-36.

Metzing, D. (2004). A quick and easy imaging method for documenting flower characters. Haseltonia, 10, 30-37.

Miki E., Meng, Y. Q., Terabayashi, S., Higuchi, M., Sasaki, H., Okada, M., Shu, G. M., & Huang, H. (1999). Morphological identification of fruits of three original plants of "Wu Zhu Yu". Natural Medicines, 53(6), 275-282.

Milla-Lewis, S. R., Kimball, J. A., Zuleta, M. C., Harris-Shultz, K. R., Schwartz, B. M., & Hanna, W. W. (2012). Use of sequence-related amplified polymorphism (SRAP) markers for comparing levels of genetic diversity in centipedegrass (Eremochloa ophiuroides (Munro) Hack.) germplasm. Genetic Resources and Crop Evolution, 59(7), 1517-1526.

Miller, A. J. & Gross, B. L. (2011). From forest to field: Perennial fruit crop domestication. American Journal of Botany, 98(9), 1389-1414.

Mishra, M., Suresh, N., Bhat, A., Suryaprakash, N., Kumar, S., Kumar, & Jayarama, A. (2011). Genetic molecular analysis of Coffea arabica (Rubiaceae) hybrids using SRAP markers. Revista de Biología Tropical, 59(2), 607-617.

Mokhtari, N., Rahimmalek, M., Talebi, M., & Khorrami, M. (2012). Assessment of genetic diversity among and within Carthamus species using sequence-related amplified polymorphism (SRAP) markers. Plant Systematics and Evolution, 299(7), 1285-1294.

Molano-Flores, B. (1999). Biological Assessment; Bird's-foot Violet, Viola pedata. Champaign, IL. Illinois Natural History Survey.

Morehouse, S. & Carroll, S. B. (2001, August 12-16). Effect on pollinator behavior of a flower-color polymorphism in Viola pedata. Poster presented at Botany 2001 on Ecological Section, BSA, Albuquerque, NM. doi: http://2001.botanyconference.org/section3/abstracts/31.shtml

207

Mossman, M. J., & Hine, R. L. (1985). Wisconsin's Frog and Toad Survey, 1984(Vol. 16). Bureau of Endangered Resources, Wisconsin Department of Natural Resources. 14pp.

Munsell Color Co. (1969). Munsell Book of Color, Neighboring Hues Edition, Mattefinish Collection. Baltimore, MD: Munsell Color Co.

Mutlu, N., Boyaci, F. H., Gocmen, M., & Abak, K. (2008). Development of SRAP, SRAP-RGA, RAPD and SCAR markers linked with a Fusarium wilt resistance gene in eggplant. Theoretical and Applied Genetics, 117(8), 1303-1312.

Nagl , N., Weiland, J., & Lewellen, R. (2007). Detection of DNA polymorphism in sugar beet bulks by SRAP and RAPD markers. Journal of Biotechnology, 131S, S32- S35.

Nair, N. V., Nair, S., Sreenivasan, T. V., & Mohan, M. (1999). Analysis of genetic diversity and phylogeny in Saccharum and related genera using RAPD markers. Genetic Resources and Crop Evolution, 46(1), 73-79.

Nauenburg, J. D. (1986). Untersuchungen zur Variabilitat, Okologie und Systematik der Viola tricolor -Gruppe in Mitteleuropa. (PhD thesis). Gottingen, Germany.

Near, T. J., Page, L. M., & Mayden, R. L. (2001). Intraspecific phylogeography of Percina evides (Percidae: Etheostomatinae): an additional test of the Central Highlands pre-Pleistocene vicariance hypothesis. Molecular Ecology, 10, 2235- 2240.

Nei, M. (1963). The efficiency of haploid methods of plant breeding. Heredity, 18, 95- 100.

Nei, M. (1972). Genetic distance between populations. American Naturalist, 106(949), 283-292.

Nei, M., & Li, W. H. (1979). Mathematical model for studying genetic variation in terms of restriction endonucleases. Proceedings of the National Academy of Sciences, 76(10), 5269-5273.

Nicolosi, E., Deng, Z., Gentile, A., La Malfa, S., Continella, G., & Tribulato, E. (2000). Citrus phylogeny and genetic origin of important species as investigated by molecular markers. Theoretical and Applied Genetics, 100, 1155–1166.

Novotna, I. (1978). Analysis of flower size in self-pollinated and open-pollinated individuals which differ as a result of mixed pollination of garden pansy (Viola ×

208

wittrockiana) and wild species V. arvensis or tricolor. Sbornik UVTIZ, Zahradnictvi, 5/6(3/4), 283-290.

Novotna, I. (1981). Interspecific crossing of garden pansy (Viola wittroc kiana) with the wild species arvensis, in relation to flower colour and size. Sbornik UVTIZ, Zahradnictvi 8(2), 131-156.

Nybom, H. (2001). DNA markers for different aspects of plant breeding research and its applications, Acta Hortorticulturae, 560, 63–67.

Nybom, H. (2003) Navigating the ’Smorgasbord’ of DNA marker methods for improvement of horticultural crops. Acta Horticulturae, 625, 15-28.

Nybom, H. (2004). Comparison of different nuclear DNA markers for estimating intraspecific genetic diversity in plants. Molecular Ecology, 13(5), 1143–1155.

Nybom, H., & Bartish, I. (2000). Effects of life history traits and sampling strategies on genetic diversity estimates obtained with RAPD markers in plants. Perspectives in Plant Ecology, Evolution and Systematics, 3(2), 93–114.

O’hanlon, P. C., Peakall, R., & Briese, D. T. (1999). Amplified fragment length polymorphism (AFLP) reveals introgression in weedy Onopordum thistles: hybridization and invasion. Molecular Ecology, 8(8), 1239-1246.

O'neal, M. E., Landis, D. A., & Isaacs, R. (2002). An inexpensive, accurate method for measuring leaf area and defoliation through digital image analysis. Journal of Economic Entomology, 95(6), 1190-1194.

Odong, T. L., van Heerwaarden, J., Jansen, J., van Hintum, T. J., & van Eeuwijk, F. A. (2011). Determination of genetic structure of germplasm collections: Are traditional hierarchical clustering methods appropriate for molecular marker data?. Theoretical and Applied Genetics, 123(2), 195-205.

Ogisu, M. (1999). Japanese Hepaticas. Curtis's Botanical Magazine, 16(2), 136-138.

Okazaki, K., Sakamoto, K., Kikuchi, R., Saito, A., Togashi, E., Kuginuki, Y., Matsumoto, S., & Hirai, M. (2007). Mapping and characterization of FLC homologs and QTL analysis of flowering time in Brassica oleracea. Theoretical and Applied Genetics, 114(4), 595-608.

Olmstead, J. W., Lang, G. A., & Grove, G. G. (2001). Assessment of severity of powdery mildew infection of sweet cherry leaves by digital image analysis. HortScience, 36(1), 107-111.

209

Olmstead, R. G., & Palmer, J. D. (1994). Chloroplast DNA systematics: A review of methods and data analysis. American Journal of Botany, 81(9), 1205-1224.

Onozaki, T., Ikeda, H., Yamaguchi, T., & Himeno, M. (1996). Introduction of bacterial wilt (Pseudomonas caryophylli) resistance in Dianthus wild species to carnation. Acta Horticulturae, 454, 127-132.

Ortega, O. R., Duran, E., Arbizu, C., Ortega, R., Roca, W., Potter, D., & Quiros, C. F. (2007). Pattern of genetic diversity of cultivated and non-cultivated mashua, Tropaeolum tuberosum, in the Cusco region of Peru. Genetic Resources and Crop Evolution, 54(4), 807-821.

Palumbo, R., Hong, W. F., Wang, G. L., Hu, J., Craig, R., Locke, J., Krause, C., & Tay, D. (2007). Target region amplification polymorphism (TRAP) as a tool for detecting genetic variation in the genus Pelargonium. HortScience, 42(5), 1118- 1123.

Pan, W. L., & Bolton, R. P. (1991). Root quantification by edge discrimination using a desktop scanner. Agronomy Journal, 83(6), 1047-1052.

Pang, X., Hu, C., & Deng, X. (2007). Phylogenetic relationships within Citrus and its related genera as inferred from AFLP markers. Genetic Resources and Crop Evolution, 54(2), 429-436.

Paris, M., Bonnes, B., Ficetola, G. F., Poncet, B. N., & Després, L. (2010). Amplified fragment length homoplasy: In silico analysis for model and non-model species. BMC Genomics, 11(1), 287-299.

Pasquali, M., Komjati, H., Lee, D., & Bayles, R. (2010). SRAP Technique efficiently generates polymorphisms in Puccinia striiformis isolates. Journal of Phytopathology, 158(10), 708–711.

Parker, K., Hamrick, J., Parker, A., & Stacy, E. (1997). Allozyme diversity in Pinus virginiana (Pinaceae): Intraspecific and interspecific comparisons. American Journal of Botany, 84(10), 1372-1372.

Parsons, T. E., Schmidt, E. J., Boughner, J. C., Jamniczky, H. A., Marcucio, R. S., & Hallgrímsson, B. (2011). Epigenetic integration of the developing brain and face. Developmental Dynamics, 240(10), 2233-2244.

Paxton, J. (Ed.). (1849). Paxton's Magazine of Botany, and Register of Flowering Plants (Vol. 16). London: Orr and Smith.

210

Peakall, R., Gilmore, S., Keys, W., Morgante, M., & Rafalski, A. (1998). Cross-species amplification of soybean (Glycine max) simple sequence repeats (SSRs) within the genus and other legume genera: implications for the transferability of SSRs in plants. Molecular Biology and Evolution, 15(10), 1275-1287.

Peakall, R., & Smouse, P. E. (2006) GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes. 6, 288-295.

Peakall, R., & Smouse, P. E. (2012) GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research – an update. Bioinformatics, 28, 2537- 2539.

Perez, T., Albornoz, J., Domínguez, A. (1998). An evaluation of RAPD fragment reproducibility and nature. Molecular Ecology, 7(10), 1347-1357.

Perrie, L. R., Brownsey, P. J., Lockhart, P. J., Brown, E. A., & Large, M. F. (2003). Biogeography of temperate Australasian Polystichum ferns as inferred from chloroplast sequence and AFLP. Journal of Biogeography, 30(11), 1729-1736.

Perry, M. C., & McIntosh, M. S. (1991). Geographical patterns of variation in the USDA soybean germplasm collection: I. Morphological traits. Crop Science, 31(5), 1350-1355.

Petit, R. J., Duminil, J., Fineschi, S., Hampe, A., Salvini, D., & Vendramin, G. G. (2005). Invited review: Comparative organization of chloroplast, mitochondrial and nuclear diversity in plant populations. Molecular Ecology, 14(3), 689-701.

Petit, R. J., El Mousadik, A., & Pons, O. (1998). Identifying populations for conservation on the basis of genetic markers. Conservation Biology, 12(4), 844-855.

Petit, R. J., Aguinagalde, I., de Beaulieu, J. L., Bittkau, C., Brewer, S., Cheddadi, R., Ennos, R., Fineschi, S., Grivet, D., Lascoux, M., Mohanty, A., Müller-Starck, G., Demesure-Musch, B., Palmé, A., Martín, J. P., Rendell, S., & Vendramin, G. G. (2003). Glacial refugia: Hotspots but not melting pots of genetic diversity. Science, 300(5625), 1563-1565.

Petit, R. J., Bialozyt, R., Brewer, S., Cheddadi, R., & Comps, B. (2001). From spatial patterns of genetic diversity to postglacial migration processes in forest trees. Special Publication-British Ecological Society, 14, 295-318.

Petit, R. J., Pineau, E., Demesure, B., Bacilieri, R., Ducousso, A., & Kremer, A. (1997). Chloroplast DNA footprints of postglacial recolonization by oaks. Proceedings of the National Academy of Sciences, 94(18), 9996-10001.

211

Petit, T. L., & Peat, J. P. (2008). The New Encyclopedia of Daylilies: More Than 1700 Outstanding Selections. Portland, OR: Timber Press.

Pettet, A. (1964). Studies on British pansies the staus of some intermediates between Viola tricolor L. and V. arvensis Murr. Watsonia, 6(1), 51-69.

Philpott, S., Uno, S., & Maldonado, J. (2006). The importance of ants and high-shade management to coffee pollination and fruit weight in Chiapas, Mexico. Biodiversity and Conservation, 15(1), 487-501.

Pielou, E. C. (2008). After The Ice Age: The Return of Life to Glaciated North America. Chicago, IL: University of Chicago Press.

Piperno, D. R., Ranere, A. J., Holst, I., Iriarte, J., & Dickau, R. (2009). Starch grain and phytolith evidence for early ninth millennium B.P. maize from the Central Balsas River Valley, Mexico. Proceedings of the National Academy of Science, 106(13), 4957-4958.

Pleasants, J. M., & Wendel, J. F. (1989). Genetic diversity in a clonal narrow endemic, Erythronium propullans, and in its widespread progenitor, Erythronium albidum. American Journal of Botany, 76(8), 1136-1151.

Polat, I., Kacar, Y. A., Yesiloglu, T., Uzun, A., Tuzcu, O., Gulsen, O., Incesu, M., Kafa, G., Turgutoglu, E., & Anil, S. (2011). Molecular characterization of sour orange (Citrus aurantium) accessions and their relatives using SSR and SRAP markers. Genetics and Molecular Research, 11(3), 3267-3276.

Potenko, V. V., & Velikov, A. V. (1998). Genetic diversity and differentiation of natural populations of Pinus koraiensis (Sieb. et Zucc.) in Russia. Silvae Genetica, 47(4), 202-207.

Powell, W. (2006). Germplasm Resources and Conservation' In D. A. Sleper and J. M. Poehlman (Eds.), Breeding Field Crops. Fifth edition. Ames, Iowa: Wiley- Blackwell.

Powell, W., & Morgante, M. (1996). The comparison of RFLP, RAPD, AFLP and SSR (microsatellite) markers for germplasm analysis. Molecular Breeding, 2(3), 225– 238.

Pritchard, J. K., Stephens, M., & Donnelly, P. (2000). Inference of population structure using multilocus genotype data. Genetics, 155(2), 945–959.

Puorto, G., da Graca Salomao, M., Theakston, R. D. G., Thorpe, R. S., Warrell, D. A., & Wuster, W. (2001). Combining mitochondrial DNA sequences and morphological 212

data to infer species boundaries: Phylogeography of lanceheaded pitvipers in the Brazilian Atlantic forest, and the status of Bothrops pradoi (Squamata: Serpentes: Viperidae). Journal of Evolutionary Biology, 14(4), 527-538.

Purugganan, M. D., & Fuller, D. Q. (2009). The nature of selection during plant domestication. Nature, 457(7231), 843-848.

Qian, D., Huang, L., Cui, G., & Chen, M. (2009). Study on genetic relationship of Astragalus membranaceus var mongholicus in different producing area using SRAP. China Journal of Chinese Materia Medica, 34(4), 382-385.

Qiao, L. X., Liu, H. Y., Guo, B. T., Weng, M. L., Dai, J. X., Duan, D. L., & Wang, B. (2007). Molecular identification of 16 Porphyra lines using sequence-related amplified polymorphism markers. Aquatic Botany, 87(3), 203-208.

Qiao, Y. C., Lin, S. Q., & Liu, C. M. (2010, May). SRAP markers in genus Eriobotrya. In III International Symposium on Loquat 887 (pp. 39-42).

Quest-Ritson, C. (2003). The English Garden: A Social History. Boston, MA: David R. Godine.

Rahman, M., McVetty, P., & Li, G. (2007). Development of SRAP, SNP and multiplexed SCAR molecular markers for the major seed coat color gene in Brassica rapa L. Theoretical and Applied Genetics, 115, 1101–1107.

Rasband, W. S. ImageJ, United States National Institutes of Health, Bethesda, Maryland, USA, http://imagej.nih.gov/ij/, 1997-2012.

Raver, A. (2013, April 13). A Pox on Your Flowers. The New York Times. Retrieved from http://www.nytimes.com/2013/04/18/garden/a-mold-devastates- impatiens.html?pagewanted=all

Read, R. H. (1976). Endangered and Threatened Vascular Plants in Wisconsin. Wisconsin Dept. of Natural Resources Technical Bulluletin No. 92, 58 pp.

Reif, J. C., Zhang, P., Dreisigacker, S., Warburton, M. L., van Ginkel, M., Hoisington, D., Bohn, M, & Melchinger, A. E. (2005). Wheat genetic diversity trends during domestication and breeding. Theoretical and Applied Genetics, 110(5), 859-864.

Ren, N., Liu, J., Yang, D., Chen, J., Luan, M., & Hong, J. (2012). Sequence-related amplified polymorphism (SRAP) marker as a new method for identification of endophytic fungi from Taxus. World Journal of Microbiology and Biotechnology, 28(1), 215-221. 213

Ren, X., Huang, J., Liao, B., Zhang, X., & Jiang, H. (2010). Genomic affinities of Arachis genus and interspecific hybrids were revealed by SRAP markers. Genetic Resources and Crop Evolution, 57(6), 903-913.

Riaz, A., Liow, P., Spencer, D., & Jasieniuk, M. (2008). Molecular evidence for a single genetic clone of invasive Arundo donax in the United States. Aquatic Botany, 88, 113-120.

Ridley, H. N. (1905). On the dispersal of seeds by wind. Annals of Botany, 3, 351-364.

Rieseberg, L. H. (1995). The role of hybridization in evolution: Old wine in new skins. American Journal of Botany, 82(7), 944-953.

Rieseberg, L. H. (1996). Homology among RAPD fragments in interspecific comparisons. Molecular Ecology, 5(1), 99-105.

Rieseberg, L. H., & John H. Willis, J. H. (2007). Plant speciation. Science 317, 910-914.

Rieseberg, L. H. (1997). Hybrid origins of plant species. Annual review of Ecology and Systematics, 28, 359-389.

Rieseberg, L. H., Linder, C. R., & Seiler, G. J. (1995). Chromosomal and genic barriers to introgression in Helianthus. Genetics, 141(3), 1163.

Rivera Nunez, D., De Castro, C. O., R os Ruiz, S., & Alcaraz Ariza, F. (2003). The origin of cultivation and wild ancestors of daffodils (Narcissus subgenus Ajax) (Amaryllidaceae) from an analysis of early illustrations. Scientia Horticulturae, 98(4), 307-330.

Robledo-Arnuncio, J. J. (2011). Wind pollination over mesoscale distances: an investigation with Scots pine. New Phytologist, 190(1), 222-233.

Rodriguez, G., Francis, D., van der Knaap, E., Strecker, J., Njanji, I., Thomas, J., & Jack, A. (2011). New features and many improvements to analyze morphology and color of digitalized plant organs are available in Tomato Analyzer 3.0. MAICS (pp. 160-163).

Rodriguez, J. M., Berke, T., Engle, L., & Nienhuis, J. (1999). Variation among and within Capsicum species revealed by RAPD markers. Theoretical and Applied Genetics, 99(1-2), 147-156.

214

Rong, H. (2008) Study on Orchids Diversity in Northern Hainan Island and Pollination Biology of Dendrobium loddigesii (Masters thesis). Guangxi Normal University, Guilin, Guangxi, China.

Rosenberg, N. A. (2004). Distruct: a program for the graphical display of population structure. Molecular Ecology Notes, 4, 137-138.

Ross, T. K. (1999). Phylogeography and conservation genetics of the Iowa Pleistocene snail. Molecular Ecology, 8(9), 1363-1373.

Ruiz, J. J., Garcia-Martinez, S., Pico, B., Gao, M., & Quiros, C. F. (2005). Genetic variability and relationship of closely related Spanish traditional cultivars of tomato as detected by SRAP and SSR markers. Journal of the American Society for Horticultural Science, 130(1), 88-94.

Russell, N. H. (1965). Violets (Viola) of central and eastern United States: An introductory survey. Sida, 2, 1-113.

Saeki, I., Dick, C. W., Barnes, B. V., & Murakami, N. (2011). Comparative phylogeography of red maple (Acer rubrum L.) and silver maple (Acer saccharinum L.): impacts of habitat specialization, hybridization and glacial history. Journal of Biogeography, 38(5), 992-1005.

Saha, G. C., Sarker, A., Chen, W., Vandemark, G. J., & Muehlbauer, F. J. (2010a). Identification of markers associated with genes for rust resistance in Lens culinaris Medik. Euphytica, 175(2), 261-265.

Saha, G. C., Sarker, A., Chen, W., Vandemark, G. J., & Muehlbauer, F. J. (2010b). Inheritance and linkage map positions of genes conferring resistance to Stemphylium blight in lentil. Crop Science, 50(5), 1831-1839.

Sako, Y., Regnier, E. E., Daoust, T., Fujimura, K., Kent Harrison, S., & McDonald, M. B. (2001). Computer image analysis and classification of giant ragweed seeds. Weed Science, 49(6), 738-745.

Sampson, J. F., & Byrne, M. (2012). Genetic diversity and multiple origins of polyploid Atriplex nummularia Lindl.(Chenopodiaceae). Biological Journal of the Linnean Society, 105(1), 218-230.

Sasa, O., Vladislava, G., Miroslav, Z., Branislav, K., Andrej, P., & Galic, Z. (2009). Evaluation of interspecific DNA variability in poplars using AFLP and SSR markers. African Journal of Biotechnology, 8(20), 5241-5247.

215

Saunders, A. P., & Stebbins Jr, G. L. (1938). Cytogenetic studies in Paeonia I. The compatibility of the species and the appearance of the hybrids. Genetics, 23(1), 65-82.

Schaal, B. A., Hayworth, D. A., Olsen, K. M., Rauscher, J. T., & Smith, W. A. (1998). Phylogeographic studies in plants: Problems and prospects. Molecular Ecology, 7(4), 465-474.

Schindler, T. E. (1923). A monograph and studies on Viola tricolor and closely related species (Unpublished doctoral dissertation). Cornell University, Ithica, NY.

Schlee, J. S. (2000). Our changing continent. U.S. Department of the Interior, U.S. Geological Survey. SuDoc I 19.2:C 76/991. doi: http://pubs.usgs.gov/gip/continents

Schmid, W. G. (1992). The Genus Hosta. Portland, OR: Timber Press.

Schmidtling, R. C. (2003). The southern pines during the Pleistocene. Acta Horticulturae, 615, 203-209.

Schmidtling, R. C., Hipkins, V., & Carroll, E. (2000). Pleistocene refugia for longleaf and loblolly pines. In A. K. Mitchell, P. Puttonen, M. Stoehr, & B. J. Hawkins, Frontiers of Forest Biology: Proceedings of the 1998 Joint Meeting of the North American Forest Biology Workshop and the Western Forest Genetics Association (pp. 349-354). New York, NY: Haworth Press.

Schmitt, T. (2007). Molecular biogeography of Europe: Pleistocene cycles and postglacial trends. Frontiers in Zoology, 4(11), 1-13.

Schneider, C. A., Rasband, W. S., & Eliceiri, K. W. (2012). NIH Image to ImageJ: 25 years of image analysis. Nature Methods, 9(7), 671-675.

Schonswetter, P., StehlikK, I., Holderegger, R., & Tribsch, A. (2005). Molecular evidence for glacial refugia of mountain plants in the European Alps. Molecular Ecology, 14(11), 3547–3555.

Schwartz, O. A. (1985). Lack of protein polymorphism in the endemic relict Chrysosplenium iowense (Saxifragaceae). Canadian Journal of Botany, 63(11), 2031-2034.

Scott, M. C., Caetano-Anollés, G., & Trigiano, R. N. (1996). DNA amplification fingerprinting identifies closely related Chrysanthemum cultivars. Journal of the American Society for Horticultural Science, 121(6), 1043-1048.

216

Segal, S. (1993). Tulips portrayed. The tulip trade in Holland in the 17th century. In: M. Roding and H. Theunissen (Eds.), The Tulip: A Symbol of Two Nations (pp. 9-24). Utrecht: M. Th, Houtsma Stichting; Istanbul: Turkish–Netherlands Association.

Seiler, G. J. (1992). Utilization of wild sunflower species for the improvement of cultivated sunflower. Field Crops Research, 30(3), 195-230.

Shafer, A., Cullingham, C. I., Cote, S. D., & Coltman, D. W. (2010). Of glaciers and refugia: A decade of study sheds new light on the phylogeography of northwestern North America. Molecular Ecology, 19(21), 4589-4621.

Shao, Y., Xu, L., & Chen, F. (2011). Genetic diversity analysis of Monascus strains using SRAP and ISSR markers. Mycoscience, 52(4), 224-233.

Shaw, J., Lickey, E. B., Beck, J. T., Farmer, S. B., Liu, W., Miller, J., Siripun, K. C., Winder, C. T., Schilling, E. E & Small, R. L. (2005). The tortoise and the hare II: Relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. American Journal of Botany, 92(1), 142-166.

Shaw, J., & Small, R. L. (2005). Chloroplast DNA phylogeny and phylogeography of the North American plums (Prunus subgenus Prunus section Prunocerasus, Rosaceae). American Journal of Botany, 92(12), 2011-2030.

Shaw, J., Lickey, E. B., Schilling, E. E., & Small, R. L. (2007). Comparison of whole chloroplast genome sequences to choose noncoding regions for phylogenetic studies in angiosperms: The tortoise and the hare III. American Journal of Botany, 94(3), 275-288.

Sheng, X., Liu, F., Zhu, Y., Zhao, H., Zhang, L., & Chen, B. (2002). Production of fertile intergeneric somatic hybrids between Brassica napus and Sinapis arvensis for the enrichment of the rapeseed gene pool. Plant Cell Reports, 21(2), 147-152.

Simmons, M. P., Zhang, L. B., Webb, C. T., & Muller, K. (2007). A penalty of using anonymous dominant markers (AFLPs, ISSRs, and RAPDs) for phylogenetic inference. Molecular Phylogenetics and Evolution, 42(2), 528-542.

Silva, M. F. S., De Andrade, I. M., & Mayo, S. J. (2012). Geometric morphometrics of leaf blade shape in Montrichardia linifera (Araceae) populations from the Rio Parnaíba Delta, north-east Brazil. Botanical Journal of the Linnean Society, 170(4), 554-572.

Sinclair, J., & Freeman, J. (1835). A History and Description of the Different Varieties of the Pansey, or Heartsease, Now in Cultivation in the British Gardens: Illustrated with Twenty-four Coloured Figures, of the Choicest Sorts. London: E. Wilson. 217

Singh, R. J. (2006). Genetic Resources, Chromosome Engineering, and Crop Improvement: Vegetable Crops (Vol. 3). CRC Press.

Siu, G. (2001). Crisis and opportunity for Dendrobium species. Living Forests,3, 3-7.

Small, R. L., Ryburn, J. A., Cronn, R. C., Seelanan, T., & Wendel, J. F. (1998). The tortoise and the hare: choosing between noncoding plastome and nuclear Adh sequences for phylogeny reconstruction in a recently diverged plant group. American Journal of Botany, 85(9), 1301-1315.

Smissen, R., Breitwieser, I., Ward, J., McLenachan, P., & Lockhart, P. (2003). Use of ISSR profiles and ITS-sequences to study the biogeography of alpine cushion plants in the genus Raoulia (Asteraceae). Plant Systematics and Evolution, 239(1- 2), 79-94.

Smith, B. D. (1997). The initial domestication of Cucurbita pepo in the Americas 10,000 years ago. Science, 276(5314), 932-934.

Soleimani, M. H., Talebi, M., & Tabatabaei, B. E. (2012). Use of SRAP markers to assess genetic diversity and population structure of wild, cultivated, and ornamental pomegranates (Punica granatum L.) in different regions of Iran. Plant Systematics and Evolution, 298(6), 1141-1149.

Soille, P. (2003). Morphological Image Analysis: Principles and Applications. New York, NY: Springer-Verlag.

Sokal, R. R., & Michener, C. D. (1958). A statistical method for evaluating systematic relationships. The University of Kansas Scientific Bulletin, 38(2), 1409-1438.

Sokal, R. R., & Rohlf, F. J. (1962). The comparison of dendrograms by objective methods. Taxon, 11, 33–40.

Soltis, D. E., & Soltis, P. S. (1989). Allopolyploid speciation in Tragopogon: Insights from chloroplast DNA. American Journal of Botany, 1119-1124.

Soltis, D. E., Soltis, P. S., & Rieseberg, L. H. (1993). Molecular data and the dynamic nature of polyploidy. Critical Reviews in Plant Sciences, 12(3), 243-273.

Soltis, D. E., Gitzendanner, M. A., Strenge, D. D., & Soltis, P. S. (1997). Chloroplast DNA intraspecific phylogeography of plants from the Pacific Northwest of North America. Plant Systematics and Evolution, 206(1-4), 353-373.

218

Soltis, D. E., Soltis, P. S., & Tate, J. A. (2003). Advances in the study of polyploidy since plant speciation. New Phytologist, 161, 173–191.

Soltis, D. E., Morris, A. B., McLachlan, J. S., Manos, P. S., & Soltis, P. S. (2006). Comparative phylogeography of unglaciated eastern North America. Molecular Ecology, 15(14), 4261-4293.

Soltis, P. S., & Soltis, D. E. (2009). The role of hybridization in plant speciation. Annual Review of Plant Biology, 60, 561-588.

Song, H., Mo, X., Zhao, G., Li, J., Zou, F., Liu, W., Wu, X., Lin, R., Weng, Y., & Zhu, X. (2011). Electrophoretic detection of genetic variability among Schistosoma japonicum isolates by sequence-related amplified polymorphism. Electrophoresis, 32, 1364-1370.

Song, Z., Li, X., Wang, H., & Wang, J. (2010). Genetic diversity and population structure of Salvia miltiorrhiza Bge in China revealed by ISSR and SRAP. Genetica, 138(2), 241-249.

Sonnante, G., Stockton, T., Nodari, R. O., Velásquez, V. B., & Gepts, P. (1994). Evolution of genetic diversity during the domestication of common-bean (Phaseolus vulgaris L.). Theoretical and Applied Genetics, 89(5), 629-635.

Sparks, R. E. (1995). Need for ecosystem management of large rivers and their floodplains. BioScience, 45(3), 168-182.

Speijer, P. R., De Waele, D., & Fogain, R. (1997). Screening of Musa germplasm for resistance and tolerance to nematodes (Vol. 1). Heverlee, Belgium: IITA.

Stebbins, G. L. (1958). The inviability, weakness, and sterility of interspecific hybrids. Advances in Genetics, 9, 147-215.

Stout, A. B. (1934). Daylilies: The Wild Species and Garden Clones, Both Old and New, of the Genus Hemerocallis. New York, NY: Macmillan Company.

Strange, R. M., & Burr, B. M. (1997). Intraspecific phylogeography of North American highland fishes: A test of the Pleistocene vicariance hypothesis. Evolution, 51, 885-897.

Suman, A., Kimbeng, C. A., Edme, S. J., & Veremis, J. (2008). Sequence-related amplified polymorphism (SRAP) markers for assessing genetic relationships and diversity in sugar cane germplasm collections. Plant Genetic Resources: Characterization and Utilization, 6(3), 222-231.

219

Sun, C. Q., Wang, X. K., Li, Z. C., Yoshimura, A., & Iwata, N. (2001). Comparison of the genetic diversity of common wild rice (Oryza rufipogon Griff.) and cultivated rice (O. sativa L.) using RFLP markers. Theoretical and Applied Genetics, 102(1), 157-162.

Sun, J., Zhang, X., Zhang, Y., Zhuo, C., & Bo, H. (2009). Analysis on genetic diversity and genetic basis of main sesame cultivars released in China. Scientia Agricultura Sinica, 42(10), 3421-3431.

Sun, M., & Wong, K. C. (2001). Genetic structure of three orchid species with contrasting breeding systems using RAPD and allozyme markers. American Journal of Botany 88(12), 2180-2188.

Sun, S. J., Gao, W., Lin, S. Q., Zhu, J., Xie, B. G., & Lin, Z. B. (2006). Analysis of genetic diversity in Ganoderma population with a novel molecular marker SRAP. Applied microbiology and Biotechnology, 72(3), 537-543.

Sun, Z., Wang, Z., Tu, J., Zhang, J., Yu, F., McVetty, P. B. E., & Li, G. (2007). An ultradense genetic recombination map for Brassica napus, consisting of 13551 SRAP markers. Theoretical and Applied Genetics, 114, 1305–1317.

Taberlet, P., Fumagalli, L., Wust-Saucy, A. G., & Cosson, J. F. (1998). Comparative phylogeography and postglacial colonization routes in Europe. Molecular Ecology, 7(4), 453-464.

Talebi, M. B., Mokhtari, N., Rahimmalek, M., & Sahhafi, S. (2012). Molecular characterization of Carthamus tinctorius and C. oxyacanthus germplasm using sequence related amplified polymorphism (SRAP) markers. Plant Omics, 5, 136- 142.

Talebi, M., Hajiahmadi, Z., & Rahimmalek, M. (2011). Genetic diversity and population structure of four Iranian alfalfa populations revealed by sequence-related amplified polymorphism (SRAP) markers. Journal of Crop Science and Biotechnology, 14(3), 173-178.

Talebi, M., Kazemi, M., & Sayed-Tabatabaei, B. E. (2012). Molecular diversity and phylogenetic relationships of Pistacia vera, Pistacia atlantica subsp. mutica and Pistacia khinjuk using SRAP markers. Biochemical Systematics and Ecology, 44, 179-185.

Tam, S. M., Mhiri, C., Vogelaar, A., Kerkveld, M., Pearce, S. R., & Grandbastien, M. A. (2005). Comparative analyses of genetic diversities within tomato and pepper collections detected by retrotransposon-based SSAP, AFLP and SSR. Theoretical and Applied Genetics, 110(5), 819-831. 220

Tan, Z., Li, Y., Hu, Q., Mei, D., Xu, Y., & Li, Y. (2009). Genetic diversity of parental lines of rapeseed hybrids based on SSR and SRAP markers. Journal of Agricultural Biotechnology, 17(5), 882-890.

Tang, L., Xiao, Y., Li, L., Guo, Q., & Bian, Y. (2010). Analysis of genetic diversity among Chinese Auricularia auricula cultivars using combined ISSR and SRAP markers. Current Microbiology, 61(2), 132-140.

Tang, Q., Zeng, X., Liao, M. A., Pan, G., Zha, X., Gong, J., & Ciren, Z. (2012). SRAP Analysis of genetic diversity of Paeonia ludlowii in Tibet. Scientia Silvae Sinicae, 48(1), 70-76.

Tay, D. (2002). Vegetable Hybrid Seed Production. Proceedings International Seed Seminar: Trade Production and Technology, pp. 128-139. doi: http://www.seedconsortium.org/PUC/pdf%20files/23- Vegetable%20Hybrid%20Seed%20Production.pdf

Tay, D. (2003). The herbaceous ornamental plant genebank: its role in the floriculture industry. Acta Horticulturae, 624, 29-36.

Tay, D. (2005). Conserving herbaceous ornamental plant germplasm. In M. B. MacDonald, & F. Y. Kwong (Eds.), Flower Seeds; Biology and Technology (pp. 333-356). Cambridge, MA: CABI.

Tay, D. (2006). Herbaceous ornamental plant germplasm conservation and use. In Anderson, N. (Ed.) Flower Breeding and Genetics (pp. 113-175). Dordrecht, Netherlands: Springer.

Teller, J. T. (1995). History and drainage of large ice-dammed lakes along the Laurentide Ice Sheet. Quaternary International, 28, 83-92.

Teller, J. T., Leverington, D. W., & Mann, J. D. (2002). Freshwater outbursts to the oceans from glacial Lake Agassiz and their role in climate change during the last deglaciation. Quaternary Science Reviews, 21, 879-887.

Tobias, C. M., Twigg, P., Hayden, D. M., Vogel, K. P., Mitchell, R. M., Lazo, G. R., Chow., E. K., & Sarath, G. (2005). Analysis of expressed sequence tags and the identification of associated short tandem repeats in switchgrass. Theoretical and Applied Genetics, 111(5), 956-964.

Tomkins, J. P., Wood, T. C., Barnes, L. S., Westman, A., & Wing, R. A. (2001). Evaluation of genetic variation in the daylily (Hemerocallis spp.) using AFLP markers. Theoretical and Applied Genetics, 102(4), 489-496. 221

Tremblay, N. O., & Schoen, D. J. (1999). Molecular phylogeography of Dryas integrifolia: glacial refugia and postglacial recolonization. Molecular Ecology, 8, 1187-1198.

Tribsch, A., & Schonswetter, P. (2003). Patterns of endemism and comparative phylogeography confirm palaeoenvironmental evidence for Pleistocene refugia in the Eastern Alps. Taxon, 53(3), 477-497.

Ungerer, M. C., Baird, S. J., Pan, J., & Rieseberg, L. H. (1998). Rapid hybrid speciation in wild sunflowers. Proceedings of the National Academy of Sciences, 95(20), 11757-11762.

Uzun, A., Yesiloglu, T., Polat, I., Aka-Kacar, Y., Gulsen, O., Yildirim, B., Onder Tuzcu, O., Tepe, S., Canan, I., & Anil, S. (2011a). Evaluation of genetic diversity in lemons and some of their relatives based on SRAP and SSR markers. Plant Molecular Biology Reporter, 29(3), 693-701.

Uzun, A., Yesiloglu, T., Aka Kacar, Y., Tuzcu, O., Gulsen, O., Seday, U., & Canan, I. (2011b). Genetic diversity and relationships of clementine mandarin accessions as assessed using SRAP markers. Acta Horticulturae, 892, 65-72.

Uzun, A., Yesiloglu, T., Aka Kacar, Y., Tuzcu, O., Gulsen, O., Kafa, G., Yildirim, B., & Anil, S. (2011c). Genetic diversity of grapefruit and pummelo accessions based on SRAP markers. Acta Horticulturae, 892, 49-56.

Uzun, A., Yesiloglu, T., Aka-Kacar, Y., Tuzcu, O., & Gulsen, O. (2009). Genetic diversity and relationships within Citrus and related genera based on sequence related amplified polymorphism markers (SRAPs). Scientia Horticulturae, 121, 306–312.

Van Tuyl, J. M. (1996). Interspecific hybridization of flower bulbs: A review. Acta Horticulturae, 430, 465-476.

Van Tuyl, J. M., & Van Holsteijn, H. C. (1996). Lily breeding research in the Netherlands. Acta Horticulturae, 414, 35-46.

Vandemark, G. J., Ariss, J. J., Bauchan, G. A., Larsen, R. C., & Hughes, T. J. (2006). Estimating genetic relationships among historical sources of alfalfa germplasm and selected cultivars with sequence related amplified polymorphisms. Euphytica, 152(1), 9-16.

Vekemans, X. (2002). AFLP-surv version 1.0. Distributed by the author. Laboratoire de Génétique et Ecologie Végétale, Université Libre de Bruxelles, Belgium. 222

Vekemans, X., Beauwens, T., Lemaire, M., & Roldán-Ruiz, I. (2002). Data from amplified fragment length polymorphism (AFLP) markers show indication of size homoplasy and of a relationship between degree of homoplasy and fragment size. Molecular Ecology, 11(1), 139-151.

Venora, G., Grillo, O., Shahin, M. A., & Symons, S. J. (2007). Identification of Sicilian landraces and Canadian cultivars of lentil using an image analysis system. Food Research International, 40(1), 161-166.

Viscosi, V., & Cardini, A. (2011). Leaf morphology, taxonomy and geometric morphometrics: A simplified protocol for beginners. PLoS One, 6(10), e25630.

Viscosi, V., Antonecchia, G., Lepais, O., Fortini, P., Gerber, S., & Loy, A. (2012). Leaf shape and size differentiation in white oaks: Assessment of allometric relationships among three sympatric species and their hybrids. International Journal of Plant Sciences, 173(8), 875-884.

Volker, P. W. & Orme, R. K. (1988). Provenance trials of Eucalyptus globulus and related species in Tasmania. Australian Forestry, 51(4), 257-265.

Von Malek, B., Weber, W. E., & Debener, T. (2000). Identification of molecular markers linked to Rdr1, a gene conferring resistance to blackspot in roses. Theoretical and Applied Genetics, 101(5-6), 977-983.

Vos, P., Hogers, R., Bleeker, M., Reijans, M., Lee, T., Hornes, M., & Zabeau, M. (1995). AFLP: A new technique for DNA fingerprinting. Nucleic Acids Research, 23(21), 4407-4414.

Turrill, W. B. (1914). Hedychium coronarium and allied species. Bulletin of Miscellaneous Information (Royal Gardens, Kew), 10, 368-372.

Walker, D., & Avise, J. C. (1998). Principles of phylogeography as illustrated by freshwater and terrestrial turtles in the southeastern United States. Annual Review of Ecology and Systematics, 29, 23-58.

Wall, W. A., Douglas, N. A., Xiang, Q. Y., Hoffmann, W. A., Wentworth, T. R. & Hohmann, M. G. (2010). Evidence for range stasis during the latter Pleistocene for the Atlantic coastal plain endemic genus Pyxidanthera Michaux. Molecular Ecology, 19, 4302-4314.

Wallace, L., & Doffitt, C. (2013). Genetic structure of the mesic forest-adapted herbs Trillium cuneatum and Trillium stamineum (Melanthiaceae) in the south-central United States. Castanea, 78(2), 154-162. 223

Walter, R., & Epperson, B. K. (2001). Geographic pattern of genetic variation in Pinus resinosa: Area of greatest diversity is not the origin of postglacial populations. Molecular Ecology, 10(1), 103-111.

Wang, B., Guo, W., Zhu, X., Wu, Y., Huang, N., & Zhang, T. (2006). QTL mapping of fiber quality in an elite hybrid derived-RIL population of upland cotton. Euphytica, 152, 367-378.

Wang, C. (2012). Review on application of SRAP marker in plant trait marker and linkage map construction. Journal of Henan Agricultural Sciences, 41(9), 1-5, 21.

Wang, D., Wu, J., Tan, W., Pu, Z., & Yan, W. (2007). Genetic relationship of high starch sweet potato germplasm of Sichuan by SRAP. Southwest China Journal of Agricultural Sciences, 20(3), 506-509.

Wang, G. X., Tang, Y., Yan, H., Sheng, X. G., Hao, W. W., Zhang, L., Lu, K., & Liu, F. (2011). Production and characterization of interspecific somatic hybrids between Brassica oleracea var. botrytis and B. nigra and their progenies for the selection of advanced pre-breeding materials. Plant Cell Reports, 30(10), 1811-1821.

Wang, J. (2005). Pansy crossbreeding, RAPD marker-assisted breeding and tissue culture studies. (PhD thesis). Retrieved from http://www.dissertationtopic.net/

Wang, J., & Bao, M. Z. (2005). Characterization of genetic relationships in pansy (Viola × wittrockiana) inbred lines using morphological traits and RAPD markers. Journal of Horticultural Science & Biotechnology, 80(5), 537-542.

Wang, J., & Bao, M. Z. (2007). Inheritance of several quantitative traits in pansy (Viola × wittrockiana). Acta Horticulturae Sinica 34(2), 449-454.

Wang, J. S., Yao, J. C., Liu, L., Wang, Y. J., & Li, W. (2008). Construction of a molecular genetic map for melon (Cucumis melo L.) based on SRAP. Frontiers of Agriculture in China, 2(4), 451-455.

Wang, K., Sha, Y., Gong, Y., Xu, L., Liu, L., Zhu, X., & Huang, D. (2012). ‘Karyotype analysis and molecular markers identification of radish (Raphanus sativus L.). Acta Horticulturae, 929, 215-222.

Wang, L., Uilecan, I. V., Assadi, A. H., Kozmik, C. A., & Spalding, E. P. (2009). HYPOTrace: Image analysis software for measuring hypocotyl growth and shape demonstrated on Arabidopsis seedlings undergoing photomorphogenesis. Plant Physiology, 149(4), 1632-1637.

224

Wang, M. M., Chen, H. B., Wang, J. H., Song, Z. Q., & Li, S. B. (2010). Genetic relationship of Chaenomeles cultivars revealed by SRAP analysis. Scientia Agricultura Sinica, 43(3), 542-551.

Wang, S., Yang, W., & Shen, H. (2011). Genetic diversity in Apium graveolens and related species revealed by SRAP and SSR markers. Scientia Horticulturae, 129(1), 1-8.

Wang, T., Xu, J., Zhang, X. X., & Zhao, L. J. (2012). Genetic relationship of 43 cultivars of Viola tricolor and Viola cornuta using SRAP Marker. Scientia Agricultura Sinica 45(3), 496-502.

Wang, W., Zhang, F., Yu, X., Shen, X., Ge, Y., & Zhang, Z. (2012). Genetic analysis and associated SRAP markers for horticultural traits of Aechmea bromeliads. Scientia Horticulturae, 147, 29-33.

Wang, Y., Li, X., Xu, J., Xu, Y., & Liu, L. (2011). Cultivar identification and genetic diversity analysis of culiflower with molecular markers. Acta Horticulturae, 918, 315-322.

Wang, Y., Sun, X., Tan, B., Zhang, B., Xu, L. A., Huang, M., & Wang, M. (2010). A genetic linkage map of Populus adenopoda Maxim.× P. alba L. hybrid based on SSR and SRAP markers. Euphytica, 173(2), 193-205.

Wang, Z., Yuan, X., Zheng, Y., & Liu, J. (2009). Molecular identification and genetic analysis for 24 turf-type Cynodon cultivars by sequence-related amplified polymorphism markers. Scientia Horticulturae, 122(3), 461-467.

Wang, Z., Wang, J. E., Wang, X. M., Gao, H. W., Dzyubenko, N. I., & Chapurin, V. F. (2012). Assessment of genetic diversity in Galega officinalis L. using ISSR and SRAP markers. Genetic Resources and Crop Evolution, 59(5), 865-873.

Wangsomnuk, P. P., Khampa, S., Jogloy, S., Srivong, T., Patanothai, A., & Fu, Y. B. (2011). Assessing genetic structure and relatedness of Jerusalem artichoke (Helianthus tuberosus L.) germplasm with RAPD, ISSR and SRAP Markers. American Journal of Plant Sciences, 2(6), 753-764.

Ward, J. (1963). Hierarchical groupings to optimize an objective function. Journal of the American Statisticians Association, 58, 236-244.

Warmke, H. E. (1951). Cytotaxonomic investigations of some varieties of Panicum maximum and P. purpurascens in Puerto Rico. Agronomy Journal, 43(1), 143- 149.

225

Webb, S. L. (1987). Beech range extension and vegetation history: Pollen stratigraphy of two Wisconsin lakes. Ecology, 68(6), 1993-2005.

Webster, R. D. (1988). Genera of the North American Paniceae (Poaceae: Panicoideae). Systematic Botany, 13(4), 576-609.

Wei, B., Cao, L., Li, S., Huang, D., Cai, G. & Lu, X. (2011). Population structure of Euodia rutaecarpa in China revealed by amplified fragment length polymorphism (AFLP) and sequence-related amplified polymorphism (SRAP). Journal of Medicinal Plants Research, 5(30), 6628-6635.

Wei, J. Y., Liu, D. B., Wei, S. X., Xie, Z. S., Xu, G. Y., & Chen, Y. Y. (2009, September). Analysis of genetic diversity in banana cultivars (Musa spp.) using sequence-related amplified polymorphism markers. Acta Horticulturae, 897, 263- 265.

Weight, C., Parnham, D., & Waites, R. (2008). TECHNICAL ADVANCE: LeafAnalyser: A computational method for rapid and large-scale analyses of leaf shape variation. The Plant Journal, 53(3), 578-586.

Wen, Y. C., Wang, H. Z., Shen, J. X., Liu, G. H., & Zhang, S. F. (2006). Analysis of genetic diversity and genetic basis of Chinese rapeseed cultivars (Brassica napus L.) by sequence-related amplified polymorphism markers. Scientia Agricultura Sinica, 39(2), 246-256.

Wenyun, L. (2006). Compared analysis of total alkaloids and polysaccharides in tubing shoot and wild plant of Dendrobium loddigesii. Journal of Anhui Agricultural Sciences, 34(8), 1606.

West, G. (1930). Cleistogamy in Viola riviniana with especial reference to its cytological aspects. Annals of Botany, 1, 87-109.

Westman, A. L., Kresovich, S., Callow, J. A., Ford-Lloyd, B. V., & Newbury, H. J. (1997). Use of molecular marker techniques for description of plant genetic variation. In J. A. Callow, B. V. Ford-Lloyd, & H. J. Newbury, Biotechnology and Plant Genetic Resources: Conservation and Use (pp. 9-48). Cambridge, MA: CABI.

White, D., Johnston, K., & Miller, M. (2005). Ohio River Basin. In: A. C. Benke and C. E. Cushing (Eds.), Rivers of North America. (pp. 375-412). London: Elsevier Academic Press.

226

Whitt, S. R., Wilson, L. M., Tenaillon, M. I., Gaut, B. S., & Buckler, E. S. (2002). Genetic diversity and selection in the maize starch pathway. Proceedings of the National Academy of Sciences, 99(20), 12959-12962.

Wijekoon, C. P., Goodwin, P. H., & Hsiang, T. (2008). Quantifying fungal infection of plant leaves by digital image analysis using Scion Image software. Journal of Microbiological Methods, 74(2), 94-101.

Williams, J., Kubelik, A., Livak, K., Rafalski, J., & Tingey, S. (1990). DNA polymorphisms amplified by arbitrary primers are useful as genetic markers. Nucleic Acids Research, 18, 6531–6535.

Willis, D., & Burke, J. (2006). Chloroplast DNA variation confirms a single origin of domesticated sunflower (Helianthus annuus L.). Journal of Heredity, 97(4), 403- 408.

Winn, A. A., & Moriuchi, K. S. (2009). The maintenance of mixed mating by cleistogamy in the perennial violet Viola septemloba (Violaceae). American Journal of Botany, 96(11), 2074-2079.

Wissemann, V. (2007) Plant evolution by means of hybridization. Systematics and Biodiversity, 5(3), 243-253.

Wittrock, V. B. (1895). Viola-Studier II: Morphological, biological and systematic studies over Viola tricolor (L.) and its related species. Acta Horti Bergiani, 2(7), 3-78. (in Swedish)

Wittrock, V. B. (1897). Viola-Studier I: A contribution to the history of the pansies having special reference to their origin. Acta Horti Bergiani, 2(1), 1-122. (in Swedish)

Wittrock, V. B. (1896). A contribution to the history of pansies. The Gardeners’ Chronicle; A Weekly Illustrated Journal of Horticulture and Allied Subjects. 492(19), 684, 494(19), 726-728, 495(19), 754-755.

Wolfe, A. D., Xiang, Q. Y., & Kephart, S. R. (1998a). Diploid hybrid speciation in Penstemon (Scrophulariaceae). Proceedings of the National Academy of Sciences, 95(9), 5112-5115.

Wolfe, A. D., Xiang, Q. Y., & Kephart, S. R. (1998b). Assessing hybridization in natural populations of Penstemon (Scrophulariaceae) using hypervariable inter simple sequence repeat markers. Molecular Ecology, 7, 1107-1125. 227

Wolfe, A. D., & Randle, C. P. (2001). Relationships within and among species of the holoparasitic genus Hyobanche (Orobanchaceae) inferred from ISSR banding patterns and nucleotide sequences. Systematic Botany, 26(1), 120-130.

Wolfe, A. D. (2005). ISSR techniques for evolutionary biology. Methods in Enzymology, 395, 134-144.

Wood, T. H., Whitten, W. M., & Williams, N. H. (2000). Phylogeny of Hedychium and related genera (Zingiberaceae) based on ITS sequence data. Edinburgh Journal of Botany, 57(2), 261-270.

Wright, E. M. & Kelly, J. D. (2011). Mapping QTL for seed yield and canning quality following processing of black bean (Phaseolus vulgaris L.). Euphytica, 179, 471– 484.

Wright, S. (1951). The genetical structure of populations. Annals of Eugenics, 15, 323- 354.

Wright, S. (1969). Evolution and the Genetics of Populations: Vol. 2. The Theory of Gene Frequencies. Chicago, IL: University of Chicago Press.

Wright, S. (1978). Evolution and the Genetics of Populations: Vol. 4. Variability Within and Among Natural Populations. Chicago, IL: University of Chicago Press.

Wu, X. M., Chen, B. Y., Lu, G., Wang, H. Z., Xu, K., Guizhan, G., & Song, Y. (2009). Genetic diversity in oil and vegetable mustard (Brassica juncea) landraces revealed by SRAP markers. Genetic Resources and Crop Evolution, 56(7), 1011- 1022.

Wu, Y. G., Guo, Q. S., He, J. C., Lin, Y. F., Luo, L. J., & Liu, G. D. (2010). Genetic diversity analysis among and within populations of Pogostemon cablin from China with ISSR and SRAP markers. Biochemical Systematics and Ecology, 38(1), 63-72.

Wylie, A. P. (1952). The history of the garden Narcissi. Heredity, 6, 137–156.

Xie, W. G., Zhang, B. Y., Zhang, X. Q., Liu, W., Chen, Y. X., & Zhou, H. (2010). Identification and genetic analysis of orchardgrass (Dactylis glomerata L.) hybrids by SRAP molecular markers. Scientia Agricultura Sinica, 3(16), 3288- 3295.

228

Xie, W., Zhang, X., Cai, H., Huang, L., Peng, Y., & Ma, X. (2010). Genetic maps of SSR and SRAP markers in diploid orchardgrass (Dactylis glomerata L.) using the pseudo-testcross strategy. Genome, 54(3), 212-221.

Xie, X., Zhou, F., Zhang, X., & Zhang, J. (2009). Genetic variability and relationship between MT-1 elephant grass and closely related cultivars assessed by SRAP markers. Journal of Genetics, 88(3), 281-290.

Xu, M. (2005). Yangzhou Chinese herbaceous peony (Paeonia lactiflora). Landscape Plant, 4, 52-54.

Xu X., Yu, M., Luo, X., Jiang, N., Zeng J., Wei, W., & Zheng L. (2008). Molecular diversity of Auricularia polytricha revealed by AFLP and SRAP. Southwest China Journal of Agricultural Sciences, 21 (1), 121-124.

Xu, Z. D., Zhao, L. Y., Zhang, L., & Yang, Z. Y. (2011). Analysis of genetic diversity and construction of fingerprint of Rosa rugosa by SRAP. Scientia Agricultura Sinica, 44(8), 1662-1669.

Xuan, C., Guo H. Xue, D., & Liu J. (2008). Identification of Zoysia hybrids by SRAP Analysis. Molecular Plant Breeding, 6, 1233-1238.

Xue, D., Feng, S., Zhao, H., Jiang, H., Shen, B., Shi, N., Lu, J., Liu, J., & Wang, H. (2010). The linkage maps of Dendrobium species based on RAPD and SRAP markers. Journal of Genetics and Genomics, 37(3), 197-204.

Yamamoto, S., Kikuchi, T., & Handa, T. (2012, September). Genetic Diversity of Lilium auratum var. platyphyllum. Act Horticulturae, 953, 97-102.

Yan, X., Li, J., Fu, F., Jin, M., Chen, L., & Liu, L. (2009). Co-location of seed oil content, seed hull content and seed coat color QTL in three different environments in Brassica napus L. Euphytica, 170, 355–364.

Yang, S., Pang, W., Ash, G., Harper, J., Carling, J., Wenzl, P., Huttner, E., Zong, X., Kilian, A. (2006). Low level of genetic diversity in cultivated pigeonpea compared to its wild relatives is revealed by diversity arrays technology. Theoretical and Applied Genetics, 113(4), 585-595.

Yang, X., Yu, Y. J., Zhang, F. L., Zou, Z. R., Zhao, X. Y., Zhang, D. S., & Xu, J. B. (2007). Linkage map construction and quantitative trait loci analysis for bolting based on a double haploid population of Brassica rapa. Journal of Integrative Plant Biology, 49(5), 664-671.

229

Yang, M., Han, Y., Xu, L., Zhao, J., & Liu, Y. (2012). Comparative analysis of genetic diversity of lotus (Nelumbo) using SSR and SRAP markers. Scientia Horticulturae, 142, 185–195.

Yeaman, S., & Jarvis, A. (2006). Regional heterogeneity and gene flow maintain variance in a quantitative trait within populations of lodgepole pine. Proceedings of the Royal Society of London. Series B: Biological Sciences, 273(1594), 1587–1593.

Yeboah, M. A., Chen, X., Feng, C., Liang G., & Gu, M. (2007). A genetic linkage map of cucumber (Cucumis sativus L.) combining SRAP and ISSR markers. African Journal of Biotechnology, 6(24), 2784-2791.

Yeboah, M. A., Xuehao, C., Feng, C. R., Alfandi, M., Liang, G., & Gu, M. (2008). Mapping quantitative trait loci for waterlogging tolerance in cucumber using SRAP and ISSR markers. Biotechnology, 7(2), 157-167.

Yamamoto, T., Kimura, T., Sawamura, Y., Kotobuki, K., Ban, Y., Hayashi, T., & Matsuta, N. (2001). SSRs isolated from apple can identify polymorphism and genetic diversity in pear. Theoretical and Applied Genetics, 102(6-7), 865-870.

Yildiz, M., Ekbic, E., Keles, D., Sensoy, S., & Abak, K. (2011). Use of ISSR, SRAP, and RAPD markers to assess genetic diversity in Turkish melons. Scientia Horticulturae, 130(1), 349-353.

Yin, B., Wang, W., Bao, Y. (2007) Comparative study on tissue structures and content of polysaccharides in stem of Dendrobium loddigesii. Journal of Beijing University of Traditional Chinese Medicine, 30, 196-199.

Yockteng, R., Ballard, H. J., Mansion, G., Dajoz, I., & Nadot, S. (2003). Relationships among pansies (Viola section Melanium) investigated using ITS and ISSR markers. Plant Systematics and Evolution, 241(3-4), 153-170.

Yoshioka, Y., Iwata, H., Ohsawa, R. Y. O., & Ninomiya, S. (2004a). Analysis of petal shape variation of Primula sieboldii by elliptic Fourier descriptors and principal component analysis. Annals of Botany, 94(5), 657-664.

Yoshioka, Y., Iwata, H., Ohsawa, R. Y. O., & Ninomiya, S. (2004b). Quantitative evaluation of flower colour pattern by image analysis and principal component analysis of Primula sieboldii E. Morren. Euphytica, 139, 179-186.

Yoshioka, Y., Iwata, H., Hase, N., Matsuura, S., Ohsawa, R., & Ninomiya, S. (2006). Genetic combining ability of petal shape in garden pansy (Viola × wittrockiana Gams.) based on image analysis. Euphytica, 151(3), 311-319.

230

You, Y., & Hongping, D. (2012). Analysis of genetic diversity of the rare and endangered species Cibotium barometz by SRAP. Acta Botanica Boreali-Occidentalia Sinica, 32(4), 688-692.

Youssef, M., James, A. C., Rivera-Madrid, R., Ortiz, R., & Escobedo-GraciaMedrano, R. M. (2011). Musa genetic diversity revealed by SRAP and AFLP. Molecular Biotechnology, 47(3), 189-199.

Yu, C., Yin, Y., & Xu, J. (2009). Identification of Taxodium hybrids by SRAP analysis. Scientia Silvae Sinicae, 45(2), 142-146.

Yu, M., Ma, B., Luo, X., Zheng, L., Xu, X., & Yang, Z. (2008). Molecular diversity of Auricularia polytricha revealed by inter-simple sequence repeat and sequence- related amplified polymorphism markers. Current Microbiology, 56(3), 240-245.

Yu, S., Chai, M., & Jiang, L. (2005). Identification of main varieties of cherry tomato using molecular markers. Acta Agriculturae Boreali-Sinica, 20(5), 42-46.

Yuan, X. Z., Li, J. S., Pan, G., Wang, S., Jiang, X. H., Li, S. L., Deng, H. L., He, M. X., Si, L., Lai , A. Z., Wu, L. H., Zhu, & Cai, R. (2008). Genetic linkage map construction and location of QTLs for fruit-related traits in cucumber. Plant Breeding, 127(1), 180-188.

Zaefizadeh, M., & Goliev, R. (2009). Diversity and relationships among durum wheat landraces (subconvars) by SRAP and phenotypic marker polymorphism. Research Journal of Biological Sciences 4(8), 960-966.

Zeng, B., Zhang, X., & Lan, Y. (2008). Assessment of genetic diversity among Dactylis glomerata L. as determined by RAPD and SRAP. Acta Horticulturae, 783, 291- 302.

Zhang, C., Chen, X., Zhang, Y., & Yuan, Z. (2009). A method for constructing core collection of Malus sieversii using molecular markers. Scientia Agricultura Sinica, 42(2), 597-604.

Zhang, D., Meynet, J., Jacob, Y., Gudin, S., & Crespel, L. (2000). Identification of apomictic plants in Rosa hybrida L. by AFLPs. Acta Horticulturae, 547, 51-55.

Zhang, F., Chen, S., Chen, F., Fang, W., Chen, Y., & Li, F. (2011). SRAP-based mapping and QTL detection for inflorescence-related traits in chrysanthemum (Dendranthema morifolium). Molecular Breeding, 27(1), 11-23.

231

Zhang, F., Chen, S., Chen, F., Fang, W., Deng, Y., Chang, Q., & Liu, P. (2011). Genetic analysis and associated SRAP markers for flowering traits of chrysanthemum (Chrysanthemum morifolium). Euphytica, 177(1), 15-24.

Zhang, F., Ge, Y., Wang, W., Shen, X., & Yu, X. (2012). Assessing genetic divergence in interspecific hybrids of Aechmea gomosepala and A. recurvata var. recurvata using inflorescence characteristics and sequence-related amplified polymorphism markers. Genetics and Molecular Research, 11(4), 4169-4178.

Zhang, F., Ge, Y., Wang, W., Shen, X., Liu, X., Liu, J., Tian, D., & Yu, X. (2012). Genetic diversity and population structure of cultivated bromeliad accessions assessed by SRAP markers. Scientia Horticulturae, 141, 1-6.

Zhang, G. N., Bi, Z. M., Wang, Z. T., Xu, L. S., & Xu, G. J. (2003). Advances in studies on chemical constitutents from plants of Dendrobium. Chinese Traditional and Herbal Drugs, 34, S5–S8.

Zhao, K., Zhou, M., Chen, L., Zhang, D., & Robert, G. W. (2007). Genetic diversity and discrimination of Chimonanthus praecox (L.) link germplasm using ISSR and RAPD markers. Hortscience, 42(5), 1144–1148.

Zhang, M., Shang, X., Li, Y., Guo, Q., Chen, M., & Tan, Q. (2009). Evaluation of Grifola frondosa germplasm using ITS-RFLP and SRAP markers. Acta Edulis Fungi, 16(3), 5-10.

Zhang, Q. S., Xu, B. L., Yuan, Q. Q., Dong, H. X., Cheng, X. H., & Lin, D. L. (2012). Analysis of genetic diversity among Chinese Pleurotus citrinopileatus Singer cultivars using two molecular marker systems (ISSRs and SRAPs) and morphological traits. World Journal of Microbiology and Biotechnology, 28(5), 2237-2248.

Zhang, S., Fu, T., Li, Y., Zhu, J., Wang, J., & Ma, C. (2006). Polymorphism analysis of SRAP (sequence-related amplified polymorphism) markers in rapeseed [Brassica napus]. Acta Agriculturae Boreali-Sinica, 21(1), 54.

Zhang, T., Liu, S., & Dong, Y. (2011). Analysis of genetic basis of Camellia oleifera from Hubei province. Journal of Henan Agricultural Sciences, 40(11), 53-56.

Zhang, W., He, H., Guan, Y., Du, H., Yuan, L., Li, Z., Yao, D., Pan, J., & Cai, R. (2010). Identification and mapping of molecular markers linked to the tuberculate fruit gene in the cucumber (Cucumis sativus L.). Theoretical and Applied Genetics, 120(3), 645-54.

232

Zhang, Y., Fu, J. A., Li, Q. S., & Ye, D. J. (2001). Preliminary study on the tissue culture of Dendrobium nobile Lindl. Lishizhen Medicine and Materia Medica Research, 2.

Zhang, Y., Zhang, X., Wei, H., Wang, L., & Zhuo C. (2012). Analysis of genetic diversity among indigenous landraces from sesame (Sesamum indicum L.) core collection in China as revealed by SRAP and SSR markers. Genes and Genomics, 32(3), 207-215.

Zhang, Z. S., Xiao, Y. H., Luo, M., Li, X. B., Luo, X. Y., Hou, L., Li, D. & Pei, Y. (2005). Construction of a genetic linkage map and QTL analysis of fiber-related traits in upland cotton (Gossypium hirsutum L.). Euphytica, 144(1-2), 91-99.

Zhang, Z., Hu, M., Zhang, J., Liu, D., Zheng, J., Zhang, K., Wang, W., & Wan, Q. (2009). Construction of a comprehensive PCR-based marker linkage map and QTL mapping for fiber quality traits in upland cotton (Gossypium hirsutum L.). Molecular Breeding, 24(1), 49-61.

Zhao, F., Cai, Z., Hu, T., Yao, H., Wang, L., Dong, N., Wang, B., Ru, Z., & Zhai, W. (2010). Genetic analysis and molecular mapping of a novel genetic resistance to rice stripe virus. Plant Molecular Biology Reporter, 28, 512-518.

Zhao, L. P., Liu, L. W., Gong, Y. Q., Wang, M. X., Yu, F. M., & Wang, L. Z. (2007). Cultivar fingerprinting in radish (Raphanus sativus) with SRAP and AFLP marker. Bulletin of Botanical Research, 27(6), 687-693.

Zhao, W., Fang, R., Pan, Y., Yang, Y., Chung, J., Chung, I., & Park, Y. (2009). Analysis of genetic relationships of mulberry (Morus L.) germplasm using sequence- related amplified polymorphism (SRAP) markers. African Journal of Biotechnology, 8(11), 2604-2610.

Zhao, L. P., Liu, Z., Chen, L., Yao, M. Z., & Wang, X. C. (2008). Generation and characterization of 24 novel EST derived microsatellites from tea plant (Camellia sinensis) and cross-species amplification in its closely related species and varieties. Conservation Genetics, 9(5), 1327-1331.

Zhivotovsky, L. A. (1999). Estimating population structure in diploids with multilocus dominant DNA markers. Molecular Ecology, 8(6), 907-913.

Zhou, J. Fu., J., Wu, Z., Huang, S., Zhang, Y., Wang, Y., Hu, Y., Hu, G., & Liu. C. (2011). Genetic diversity in litchi and longan germplasm as determined by SRAP markers. Acta Horticulturae, 918, 799-806.

233

Zhou, Y. J., Wang, X. G., & Zhang, X. Q. (2011). Development and application of a SRAP marker for the identification of sex in Buchloe dactyloides. Euphytica, 181(2), 261-266.

Zhou, Y., Gu, F., Zhou, C., Yao, H., Duan, H., Wang, F., Liu, Y., Xing, Y., & Chu, S. (2010). Genetic diversity of Rehmannia glutinosa cultivars based on sequence- related amplified polymorphism markers. Scientia Horticulturae, 125(4), 789- 794.

Zietkiewicz, E., Rafalski, A., & Labuda, D. (1994). Genome fingerprinting by simple sequence repeat (SSR)-anchored polymerase chain reaction amplification. Genomics, 20, 176-183.

Zilis, M. R. (2009). The Hostapedia. Rochelle, IL: Q&Z Inc.

Zohary, D., Hopf, M., & Weiss, E. (2012). Domestication of Plants in the Old World: The Origin and Spread of Domesticated Plants in Southwest Asia, Europe, and the Mediterranean Basin. Oxford University Press.

Zou, M., Xia, Z., Lu, C., Wang, H., Ji, J., & Wang, W. (2012). Genetic diversity and differentiation of Aquilaria sinensis (Lour.) Gilg revealed by ISSR and SRAP markers. Crop Science, 52(5), 2304-2313.

234

Appendix A. Accession list and description of Melanium violets

Table A.1. Accession list and description of Melanium violets

Accession ID ID # Color Origin Type Ch. 1 Ch. 2 (SRAP) Ch. 2 (combined) p001Y 001 yellow USA VW x x x p002W 002 white USA VW x x x p003Y-b 003 yellow with blotch USA VW x x x p004W-b 004 white with blotch USA VW x x x p005Bl-b 005 blue with blotch USA VW x x x p006Y-b 006 yellow with blotch USA VW x x x p007W-b 007 white with blotch USA VW x x x p008Bl-b 008 blue with blotch USA VW x x x p009P-b 009 purple with blotch USA VW x x x p010Rs-b 010 rose with blotch USA VW x x x p011Y-b 011 yellow with blotch USA VW x x x p012Y-b 012 yellow with blotch USA VW x x x

Viola accessions examined in Chapter 1 and Chapter 2 of this dissertation. Accession types include Viola × wittrockiana hybrids (VW), Viola cornuta hybrids (VC) and species types (S). The first character in accession IDs (p,v) indicate presumed ploidy: VW ("pansy") types (p:tetraploid, 4x) and VC ("viola") types (v:diploid, 2x). Numerical values indicate the number given to each accession as they we attained. Following numerican terms are alphabetical descriptors of blotch presernce (-b) preceeded by causal descriptions of flower color: Y, yellow; W, white; P, purple; Rs, rose; Bl, blue; Blk, black; Lav, lavender; Brnz; bronze; Mar, marina (light blue); Br, brown; O, orange; Pk, pink. For bicolorous accessions (superior petals differ from later and inferior petals), both colors are presented in the accession ID as "Color 1.Color 2". The "Origin" column indicates the location of parent company. Columns also indicate which accessions were sampled for morphologic (Ch.1), molecular (Ch. 2), or combined (Ch. 2) analyses.

continued 235

Table A.1 continued

Accession ID ID # Color Origin Type Ch. 1 Ch. 2 (SRAP) Ch. 2 (combined)

p013W 013 white USA VW x x x

p014W-b 014 white with blotch USA VW x x x

p015Bl-b 015 blue with blotch USA VW x x x

p016Y-b 016 yellow with blotch USA VW x x x

v017P 017 Viola tricolor - purple USA S x x x wild- v018P 018 Viola tricolor - purple collected S x x x wild- v019Y 019 Viola tricolor - yellow collected S x x x Viola tricolor - lavender wild- v020Lav.W 020 and white collected S x x x Viola tricolor - yellow wild- v021Y.W 021 and white collected S x x

v039Y 039 Viola tricolor - yellow USA S x x x

p040Y-b 040 yellow with blotch Germany VW x x x

p041W 041 white, F1 USA VW x x x p042W 042 white, maternal parent USA VW x x x

p043W 043 white, paternal parent USA VW x x x wild- v044Bl 044 Viola corsica - blue collected S x x x wild- v045Y 045 Viola stanjowii - yellow collected S x x x

p046Y-b 046 yellow with blotch Germany VW x x x

p047P 047 purple USA VW x x x

p048Blk 048 black UK VW x x x

p049Bl 049 blue USA VW x x x

p050P-b 050 purple with blotch UK VW x x x

p051Y-b 051 yellow with blotch UK VW x x x

p052W.Rs 052 white and rose UK VW x x x yellow and bronze with p053Y.Brnz-b 053 blotch UK VW x x x yellow and rose with p055Y.Rs-b 055 blotch Canada VW x x x purple and white with p057P.W-b 057 blotch UK VW x x x

p060Br.Y-b 060 yellow with blotch Japan VW x x x

p061W-b 061 white with blotch Japan VW x x x

p062Y-b 062 yellow with blotch Japan VW x x x

p063W-b 063 white with blotch Japan VW x x x

continued

236

Table A.1 continued

Accession ID ID # Color Origin Type Ch. 1 Ch. 2 (SRAP) Ch. 2 (combined)

p064Mar-b 064 marina with blotch Japan VW x x x yellow and rose with p065Y.mix-b 065 blotch Japan VW x x x

p066Bl 066 blue Japan VW x x x bronze and yellow with p067Br.Y-b 067 blotch Japan VW x x x

p068B.P 068 blue and purple Japan VW x x x

p069P-b 069 purple with blotch Japan VW x x x

p070Bl 070 blue Japan VW x x x

p071Mar-b 071 marina with blotch Japan VW x x x white and rose with p072W.Pk-b 072 blotch Japan VW x x x

p073Rs-b 073 rose with blotch Japan VW x x x

p074O 074 orange Japan VW x x x

p078O 078 orange USA VW x x x

p097Y-b 097 yellow with blotch Japan VW x

p099Rs-b 099 rose with blotch Japan VW x x

p101Bl 101 blue Japan VW x x x

p102W 102 white UK VW x x x

p105Rd-b 105 rose with blotch UK VW x x x

p109Y-b 109 yellow with blotch UK VW x x x

p112P-b 112 blue with blotch UK VW x

p114W-b 114 white with blotch USA VW x x x cream-yellow with p115Y.mix-b 115 blotch USA VW x x x

v116Lav 116 purple USA VC x

p124W-b 124 white with blotch USA VW x x x

v125P.Y 125 purple and yellow USA VC x x x

p132Blk 132 black Poland VW x x

p135P.W 135 purple and white Poland VW x

p136Bl 136 blue Poland VW x wild- v140P.Y 140 purple and yellow collected VC x x x

p147P 147 purple Poland VW x x x

continued

237

Table A.1 continued

Accession ID ID # Color Origin Type Ch. 1 Ch. 2 (SRAP) Ch. 2 (combined)

p148O 148 orange Poland VW x x x purple and yellow with p149P.Y-b 149 blotch Poland VW x x x Viola altaica - blue; wild- v220Bl-1 220 sample 1 collected S x x x Viola altaica - blue; wild- v220Bl-2 220 sample 2 collected S x wild- v224Bl 224 Viola corsica - blue collected S x x x wild- v227Bl 227 Viola dubyana - blue collected S x x x Viola tricolor - purple wild- v239P.Rd 239 and red collected S x x x wild- v240P 240 Viola tricolor - purple collected S x x x wild- v241P 241 Viola tricolor - purple collected S x x x wild- v242P 242 Viola tricolor - purple collected S x x x

v250Lav 250 lavender UK VC x x x

v251Y-b 251 yellow with blotch UK VC x x x

v252P-b 252 purple with blotch UK VC x x x lavender and yellow v254Lav.Y-b 254 with blotch UK VC x x x

v256Lav 256 lavender UK VC x x x

p257Y-b 257 yellow with blotch USA VW x

p258P-b 258 blue with blotch USA VW x

v259Bl 259 blue USA VC x x x

v260Y-b 260 yellow with blotch UK VC x x x

v261Lav.W 261 lavender and white UK VC x x x

v262Lav.Y 262 lavender and yellow UK VC x x x

v263Lav.Y 263 lavender and yellow UK VC x x x

p264P.O 264 purple and orange UK VW x x x

v268R.Y 268 red and yellow Japan VC x x x

v269B.Y 269 brown and yellow Japan VC x x x

v270P.Y 270 purple and yellow Japan VC x x x

v271Y 271 yellow Japan VC x x x purple and yellow with p274P.Y-b 274 blotch USA VW x x x

v277Bl-b 277 blue with blotch USA VC x x x

v278W-b 278 white with blotch USA VC x x x

continued

238

Table A.1 continued

Accession ID ID # Color Origin Type Ch. 1 Ch. 2 (SRAP) Ch. 2 (combined)

v279Y-b 279 yellow with blotch USA VC x x x

v280Bl 280 blue USA VC x x x

v281W 281 white USA VC x x x

v282O 282 orange USA VC x x x

v283Rd 283 red USA VC x x x

v284P.Y 284 purple and yellow USA VC x x x bronze and yellow with v285Brnz.Y-b 285 blotch USA VC x x x bronze and yellow with v286Brnz.Y-b 286 blotch USA VC x x x

v287Mar-b 287 marina with blotch USA VC x x x

v288W.P 288 white and purple USA VC x x x

v289Lav 289 lavender Japan VC x x x

p290BlY 290 blue and yellow USA VW x x x

v291B.P 291 purple and blue USA VC x x x

p292Lav 292 lavender USA VW x x x

v293Lav 293 lavender Netherlands VC x x x

v294Bl 294 blue Netherlands VC x x x

p297Bl 297 blue USA VW x x x

v300P 300 purple USA VC x x x

v301P.W 301 purple and white USA VC x x x

v302P.Y 302 purple and yellow USA VC x x x

v303P 303 purple USA VC x x x

v304Y-b 304 yellow with blotch USA VC x x x

v305Y 305 yellow USA VC x x x

v306O 306 orange USA VC x x x

v307P.w 307 purple and white USA VC x x x

v308P-b 308 purple with blotch USA VC x x x

v325Lav 325 lavender USA VC x

v888O.Br-b 888 light brown with blotch Netherlands VC x Poland:wild- v909Y 909 Viola lutea - yellow collected S x Poland:wild- v910Y 910 Viola lutea - yellow collected S x Poland:wild- v911Y 911 Viola lutea - yellow collected S x Poland:wild- v912Y 912 Viola lutea - yellow collected S x

239

Appendix B. Morphological characters and descriptions for Melanium violets

Table B.1. Characters and descriptions for morphological analysis of Melanium violets

Trait Description

wiskers presence (1) or absence (0) of nectary guides

ruffled presence (1) or absence (0) of ruffled petal edges presence (1) or absence (0) of strongly demarkated (lighter) throat, relative to base expanded throat ("watermark") petal color

S/whole area ratio of superior petal area to whole flower area

L/Whole area ratio of lateral petal area to whole flower area

I/whole area ratio of inferior petal area to whole flower area

S/L area ratio superior petal area / lateral petal area

S/I area ratio superior petal area / inferior petal area

L/I area ratio lateral petal area / inferior petal area

I-Perimeter perimeter of inferior petal (cm)

I-Area area of inferior petal (cm2)

I-P/A ratio perimeter / area ratio inferior petal

I-sup. self L* L* of inferior petal ground color

I-sup. Self a* a* of inferior petal ground color

I-sup. self b* b* of inferior petal ground color

I-sup. self hue hue of inferior petal ground color

I-sup. self chroma chroma of inferior petal ground color

Description of qualitative (n=3) and quantitative (n=75) traits scored in morphologic study of 127 accessions of Viola sect. Melanium. Symbols following morphological trait names (e.g., “$”) suggest reviewing Brewer et al. (2006) for details.

continued 240

Table B.1 continued

Trait Description

I-Blotch area blotch area inferior petal

I-Blotch area ratio blotch area of inferior petal / area of inferior petal

I-Blotch L*/self L* blotch L* of inferior petal / L* of inferior petal ground color

I-Fruit Shape Index 1 (Hmax/Wmax) ratio maximum height of inferior petal / maximum width of inferior petal

I-Fruit Shape Index 2 (H50%/W50%) ratio of 50% inferior petal height/ 50% inferior petal width

I-Ellipsoid$ fitting precision R2 of inferior petal to of ellipse

I-Circular$ fitting precision R2 of inferior petal to circle

I-Rectangular$ fitting precision R2 of inferior petal to of rectangle

L-P/A ratio perimeter / area ratio lateral petal

L-sup. self L* L* of lateral petal ground color

L-sup. self a* a* of lateral petal ground color

L-sup. self b* b* of lateral petal ground color

L-sup. self hue hue of lateral petal ground color

L-sup. self chroma chroma of lateral petal ground color

L-Blotch area blotch area lateral petal

L-Blotch area ratio blotch area of lateral petal / area of lateral petal

L-Blotch L*/self L* blotch L* of lateral petal / L* of lateral petal ground color

L-Fruit Shape Index 1 (Hmax/Wmax) ratio maximum height of lateral petal / maximum width of lateral petal

L-Fruit Shape Index 2 (H50%/W50%) ratio of 50% lateral petal height/ 50% lateral petal width

L-Ellipsoid$ fitting precision R2 of lateral petal to of ellipse

L-Circular$ fitting precision R2 of lateral petal to circle

L-Rectangular$ fitting precision R2 of lateral petal to of rectangle

S-P/A ratio perimeter / area ratio superior petal

S-sup. self L* L* of superior petal ground color

S-sup. self a* a* of superior petal ground color

S-sup. self b* b* of superior petal ground color

S-sup. self hue hue of superior petal ground color

S-sup. self chroma chroma of superior petal ground color

S-Blotch area blotch area superior petal

S-Blotch area ratio blotch area of superior petal / area of superior petal

S-Blotch L/self L blotch L* of superior petal / L* of superior petal ground color

S-Fruit Shape Index 1 (Hmax/Wmax) ratio maximum height of superior petal / maximum width of superior petal

continued 241

Table B.1 continued

Trait Description

S-Fruit Shape Index 2 (H50%/W50%) ratio of 50% superior petal height/ 50% superior petal width

S-Ellipsoid$ fitting precision R2 of superior petal to of ellipse

S-Circular$ fitting precision R2 of superior petal to circle

S-Rectangular$ fitting precision R2 of superior petal to of rectangle

Wh-Perimeter perimeter of inferior petal (cm)

Wh-Area area of inferior petal (cm2)

Wh-P/A ratio perimeter / area ratio inferior petal

Wh-Width Mid-height width of whole flower at 50% height

Wh- Maximum Width maximum width of whole flower

Wh-Height Mid-width height of whole flower at 50% width

Wh-Maximum Height maximum height of whole flower

Wh-sup. self L* L* of whole flower ground color

Wh-sup. self a* a* of whole flower ground color

Wh-sup. self b* b* of whole flower ground color

Wh-sup. self hue hue of whole flower ground color

Wh-sup. self chroma chroma of whole flower ground color

Wh-Blotch area blotch area whole flower

Wh-Blotch area ratio blotch area of whole flower / area of whole flower

Wh-Blotch L* blotch L* of whole flower blotch

Wh-Blotch a* blotch a* of whole flower blotch

Wh-Blotch b* blotch b* of whole flower blotch

Wh-Blotch hue hue of whole flower blotch

Wh-Blotch chroma chroma of whole flower blotch

Wh-Blotch L/self L L* of whole flower blotch / L* of whole flower ground color ratio maximum height of whole flower petal / maximum width of whole flower Wh-Fruit Shape Index 1 (Hmax/Wmax) petal

Wh-Fruit Shape Index 2 (H50%/W50%) ratio of 50% whole flower petal height/ 50% whole flower petal width

Wh-Ellipsoid fitting precision R2 of whole flower petal to of ellipse

Wh-Circular fitting precision R2 of whole flower petal to circle

Wh-Rectangular fitting precision R2 of whole flower petal to of rectangle

242

Appendix C. Tomato Analyzer calibration information

Table C.1. Descriptions of 72 Munsell color chips and their interpretation by colorimeter and flat-bed scanner.

Chip ID Plate Chip L*c L*s a*c a*s b*c b*s 1 2.5R 9.5/0 94.67 94.2967 -0.34 -3.93963 3.31 4.72462 2 2.5R 9/0 89.59 89.6424 0.03 -3.2686 1.5 3.11904 3 2.5R 8/0 80.16 78.8954 -0.22 -2.91826 1.31 2.14013 4 2.5R 7/0 70.08 66.2394 -0.34 -2.33157 0.92 1.38623 5 2.5R 6/0 60.22 54.4172 -0.68 -1.72916 0.73 0.799851 6 2.5R 5/0 49.92 43.8241 -0.13 -0.892455 0.6 0.352585 7 2.5R 4/0 39.93 33.9243 0.46 -0.21432 0.19 -0.118569 8 2.5R 3/0 30.52 25.1025 -0.13 0.46043 0.78 0.282396 9 2.5R 2/0 15.63 14.0217 -0.31 1.86975 0.78 0.481 10 2.5R 1/0 15.68 13.3989 -0.36 1.92031 2.42 0.827399 11 2.5R 6/10 61.83 58.6 37.55 39.8197 16.82 20.0545 12 2.5R 3/6 32.22 28.2378 28.02 22.8872 11.33 10.1003 13 5R 4/14 43.97 39.9118 52.99 43.2789 30.02 26.9094 14 5R 3/2 30.68 25.1457 8.38 6.9136 6.66 3.90441 15 7.5R 8/2 80.45 78.9511 5.6 3.83046 4.84 4.76037 16 10R 9/2 89.28 90.2228 5.85 4.40652 7.4 8.02956 17 10R 5/10 53.58 46.3719 33.5 34.7559 42.49 31.7067 18 2.5YR 6/14 63.15 58.1387 36.74 39.8743 62 49.731 19 2.5YR 4/6 44.76 38.1169 21.75 19.5188 30.12 21.2426 20 5YR 4/8 46.17 37.0106 15.97 15.4775 31.3 19.7518 21 5YR 7/2 70.91 67.3966 5.61 4.54501 12.12 12.374 22 7.5YR 8/4 80.51 77.8432 11.07 13.2396 34.51 31.4231 23 7.5YR 5/8 59.33 51.6343 13.05 11.1755 50.97 37.4049 24 10YR 8/8 80.7 77.2981 9.73 11.5733 51.63 43.8258

Subscript “c” indicates measurements made by colorimeter, and subscript “s” indicates measurement made by flat-bed scanner.

continued

243

Table C.1 continued

Chip ID Plate Chip L*c L*s a*c a*s b*c b*s 25 10YR 4/2 41.02 33.9447 2.54 2.6988 13.29 6.16259 26 2.5Y 9/6 90.15 92.3334 -1.55 -8.94913 43.16 43.8606 27 2.5Y 9/4 90.93 91.713 -0.78 -2.80766 30.08 30.2914 28 2.5Y 8/12 82.41 81.33 1.42 0.131207 83.66 63.7602 29 2.5Y 4/4 46.46 40.1884 2.2 -0.704639 31.82 20.6443 30 5Y 9/8 89.36 90.5022 -6.41 -9.31688 58.38 52.8448 31 5Y 5/2 50.42 42.9048 -1.64 -1.86152 13.77 9.11928 32 7.5Y 9/10 73.64 69.2249 -6.2 -6.99461 28.68 24.0114 33 7.5Y 7/4 90.3 91.5269 -9.05 -11.9792 32.37 33.0802 34 10Y 9/4 50.57 43.6899 -5.82 -4.83386 14.14 11.162 35 2.5GY 5/2 84 81.7991 -22.28 -22.2351 65.7 52.894 36 2.5GY 8/10 60.95 57.7809 -28.6 -34.8198 53.68 40.0833 37 5GY 6/8 46.87 41.1977 -22.5 -18.9433 28.89 20.0072 38 7.5GY 4/6 66.34 61.5601 -43.47 -34.1242 38.3 29.2938 39 10GY 6/10 35.83 30.0568 -17.91 -13.3061 13.58 10.6673 40 10GY 3/4 78.93 77.7612 -12.56 -14.1389 8.01 8.51701 41 2.5G 8/2 88.8 90.647 -7.47 -10.252 4.07 6.77988 42 5G 9/2 50.88 45.5817 -27.92 -15.8717 -2.74 -2.52822 43 10G 5/6 69.4 67.4595 -16.85 -16.5085 -4.62 -1.4508 44 5BG 7/4 33.31 28.9087 -11.86 -1.04269 -17.72 -14.9605 45 10BG 3/6 89.41 89.4192 -4.33 -7.25115 -2.91 -0.846679 46 5B 9/2 32.94 28.5658 -5.09 1.96138 -25.38 -19.2856 47 5B 3/6 60.01 56.6172 -6.6 0.47735 -33.05 -29.4928 48 10B 6/8 35.23 30.333 -2.36 -0.476151 -8.91 -7.81106 49 10B 3/2 60.01 56.5767 2.9 3.42517 -22.37 -19.5937 50 5PB 6/6 68.69 66.8824 9.97 8.3766 -28.16 -25.16 51 7.5PB 7/8 30.6 27.967 28.2 20.5329 -36.45 -26.8615 52 10PB 3/10 25.44 22.3194 9.75 4.77287 -13.97 -4.83512 53 10PB 2/4 20.55 19.3668 9.7 4.82902 -8.1 -6.47854 54 2.5P 2/2 59.63 68.5882 19.23 13.453 -22 -13.6619 55 2.5P 6/8 42.29 39.2869 39.67 31.7699 -27.8 -19.9893 56 5P 4/12 71.03 68.5617 9.3 8.17306 -9.14 -7.5596 57 5P 7/4 82.72 82.5165 11.8 11.1641 -4.13 -2.62976 58 10P 8/4 80.26 80.366 12.18 11.3886 -6.03 -4.3522 59 7.5P 4/10 39.08 36.5642 36.65 25.6581 -21.28 -15.7098 60 7.5P 2/4 23.77 21.9063 16.66 10.7015 -8.59 -6.50666 61 10P 8/1 79.49 76.3623 2.86 -0.392367 0.21 1.00999 62 10P 2/4 22.19 22.4512 23.27 14.2849 -8.39 -5.34775

continued

244

Table C.1 continued

Chip ID Plate Chip L*c L*s a*c a*s b*c b*s 63 10P 3/10 30.33 38.6415 37.95 28.0905 -18.19 -10.0141 64 10P 6/2 60.27 66.0314 6.93 3.92285 -2.63 -1.95181 65 2.5RP 4/8 40.95 37.1506 35.17 27.5816 -8.68 -3.84249 66 2.5RP 2/2 22.12 18.814 11.09 7.9782 -2.12 -0.539731 67 5RP 3/2 31.47 26.7036 10.8 8.60966 -0.1 1.21417 68 5RP 6/10 60.68 58.0927 43 42.8455 -0.91 6.25413 69 7.5RP 8/2 79.81 79.008 5.24 3.05209 1.74 2.58766 70 7.5RP 4/8 43.82 39.5495 34.9 31.0204 3.45 5.66134 71 10RP 3/4 32.71 26.3908 16.69 14.6151 3.95 4.09663 72 10RP 7/8 70.59 67.2152 33.02 34.956 9.63 13.2791

245

L* Graph

100

90 80 y = 1.0498x - 5.6059 R² = 0.9827 70 60

50 bed scanner bed values

- 40

30 Flat 20 10 0 0 20 40 60 80 100 Colorimeter values

Figure C.1. Regression of colorimeter and flat-bed scanner values of L* for 72 Munsell color chips.

246

a* Graph

60 y = 0.8747x - 0.467 R² = 0.9487 50 40 30 20 10 0

bed scanner bed values -60 -40 -20 0 20 40 60 - -10

Flat -20 -30 -40 -50 Colorimeter values

Figure C.2. Regression of colorimeter and flat-bed scanner values of a* for 72 Munsell color chips.

247

b* Graph 80 y = 0.7918x + 1.022

R² = 0.978 60

40

20

0 bed scanner bed values - -60 -40 -20 0 20 40 60 80 100

Flat -20

-40 Colorimeter values

Figure C.3. Regression of colorimeter and flat-bed scanner values of b* for 72 Munsell color chips.

248

Appendix D. One-way ANOVA tables comparing molecular measures between clusters of Viola pedata

Table D.1. ANOVA tables of molecular measures for clusters of Viola pedata

X

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 180.528 3.000 60.176 11.830 0.000 Within groups: 198.395 39.000 5.087

Total: 378.923 42.000 omega^2: 0.430

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster1 cluster2 cluster3 cluster4 cluster1 0.003 0.728 0.035 cluster2 5.422 0.000 0.752 cluster3 1.468 6.890 0.002 cluster4 4.012 1.410 5.480

One-way analysis of variance tables with Tukey’s HSD test scores for (mean) molecular measures of STRUCTURE determined clusters (K=4; cluster1, NE; cluster2, SE; cluster3, IO; cluster4, DA). Bold, red text indicates significant p (ANOVA) or Q (Tukey) values for the traits and specific pair-wise comparisons.

continued

249

Table D.1 continued

X'

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 160.292 3.000 53.431 12.650 0.000 Within groups: 164.695 39.000 4.223

Total: 324.987 42.000

omega^2: 0.448

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster1 cluster2 cluster3 cluster4 cluster1 0.001 0.619 0.156

cluster2 5.799 0.000 0.224

cluster3 1.723 7.523 0.009

cluster4 3.039 2.760 4.763

A

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 528.201 3.000 176.067 6.664 0.001 Within groups: 1030.450 39.000 26.422

Total: 1558.650 42.000

omega^2: 0.283

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.997 0.078 0.002

cluster_2 0.288 0.120 0.004

cluster_3 3.512 3.224 0.498

cluster_4 5.513 5.225 2.001

continued

250

Table D.1 continued

A'

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 8.253 3.000 2.751 6.664 0.001 Within groups: 16.101 39.000 0.413

Total: 24.354 42.000

omega^2: 0.283

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.997 0.078 0.002

cluster_2 0.288 0.120 0.004

cluster_3 3.512 3.224 0.498

cluster_4 5.513 5.225 2.001

Ho

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 0.117 3.000 0.039 13.120 0.000 Within groups: 0.116 39.000 0.003

Total: 0.234 42.000

omega^2: 0.458

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.594 0.030 0.000

cluster_2 1.781 0.372 0.000

cluster_3 4.093 2.312 0.009

cluster_4 8.876 7.095 4.782

continued 251

Table D.1 continued

H'

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 2.846 3.000 0.949 12.080 0.000 Within groups: 3.062 39.000 0.079

Total: 5.908 42.000

omega^2: 0.436

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.622 0.002 0.050

cluster_2 1.717 0.000 0.002

cluster_3 5.534 7.252 0.612

cluster_4 3.794 5.511 1.740

H"

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 0.543 3.000 0.181 44.580 0.000 Within groups: 0.158 39.000 0.004

Total: 0.701 42.000

omega^2: 0.753

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.741 1.000 0.000

cluster_2 1.437 0.742 0.000

cluster_3 0.000 1.436 0.000

cluster_4 15.050 13.610 15.050

continued

252

Table D.1 continued

Np1

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 33.785 3.000 11.262 3.118 0.037 Within groups: 140.866 39.000 3.612

Total: 174.651 42.000

omega^2: 0.129

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.057 0.831 0.212

cluster_2 3.711 0.300 0.918

cluster_3 1.199 2.512 0.670

cluster_4 2.805 0.906 1.606

Np2

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 11.086 3.000 3.695 1.418 0.252 Within groups: 101.611 39.000 2.605

Total: 112.698 42.000

omega^2: 0.028

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.994 0.405 0.924

cluster_2 0.362 0.275 0.815

cluster_3 2.226 2.588 0.778

cluster_4 0.882 1.244 1.344

continued

253

Table D.1 continued

Np3

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 34.493 3.000 11.498 2.560 0.069 Within groups: 175.182 39.000 4.492

Total: 209.674 42.000

omega^2: 0.098

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.096 0.265 0.165

cluster_2 3.377 0.950 0.993

cluster_3 2.619 0.758 0.993

cluster_4 2.998 0.379 0.379

Np4

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 49.503 3.000 16.501 1.904 0.145 Within groups: 337.939 39.000 8.665

Total: 387.442 42.000

omega^2: 0.059

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.629 0.716 0.780

cluster_2 1.701 0.999 0.156

cluster_3 1.499 0.203 0.203

cluster_4 1.340 3.041 2.838

continued

254

Table D.1 continued

Np5

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 35.040 3.000 11.680 1.916 0.143 Within groups: 237.751 39.000 6.096

Total: 272.791 42.000

omega^2: 0.060

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.497 0.753 0.081

cluster_2 2.003 0.975 0.720

cluster_3 1.408 0.595 0.464

cluster_4 3.490 1.487 2.082

Np6

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 81.334 3.000 27.111 14.860 0.000 Within groups: 71.131 39.000 1.824

Total: 152.465 42.000

omega^2: 0.492

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.000 0.228 0.000

cluster_2 6.860 0.029 0.648

cluster_3 2.747 4.113 0.001

cluster_4 8.518 1.657 5.770

continued

255

Table D.1 continued

Np7

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 44.585 3.000 14.862 4.137 0.012 Within groups: 140.113 39.000 3.593

Total: 184.698 42.000

omega^2: 0.180

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.022 0.057 0.472

cluster_2 4.272 0.979 0.411

cluster_3 3.715 0.557 0.650

cluster_4 2.061 2.210 1.653

Np8

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 14.695 3.000 4.898 15.920 0.000 Within groups: 12.003 39.000 0.308

Total: 26.698 42.000

omega^2: 0.510

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 0.928 0.908 0.000

cluster_2 0.865 1.000 0.000

cluster_3 0.948 0.083 0.000

cluster_4 8.031 8.896 8.979

continued

256

Table D.1 continued

Np

One-way ANOVA

Mean Sum of sqrs df square F p(same) Between groups: 7.036 3.000 2.345 1.433 0.248 Within groups: 63.825 39.000 1.637

Total: 70.861 42.000

omega^2: 0.029

Tukey's pairwise comparisons: Q below diagonal, p(same) above diagonal

cluster 1 cluster_2 cluster_3 cluster_4 cluster 1 1.000 0.990 0.323

cluster_2 0.006 0.990 0.321

cluster_3 0.437 0.431 0.192

cluster_4 2.444 2.450 2.881

257

Appendix E. Reviewed SRAP publications

Table E.1. Descriptive information for reviewed SRAP publications

primer fragments polymorphisms % Taxon Common name Markers individuals pairs /primer pair / primer pair polymorphic Analysis methods Reference Abelmoschus esculentus okra SRAP 23 39 2.5 1.3 50.5 UPGMA Gulsen et al. (2007) Acer spp. maple SRAP 31 11 16.9 16.6 98.4 UPGMA Li et al. (2010) Aechmea hybrids bromilliad SRAP 42 16 - 9.7 - QTL mapping Wang et al. (2012) Aechmea spp. bromilliad SRAP 16 16 16.6 10.3 61.9 UPGMA, PCoA Zhang et al. (2012) Dinler & Budak Agrostis spp. bentgrass SRAP, EST 1000 56 1.4 - - Min evolution (2008) Allium fistulosum scallion, green onion SRAP 20 161 - 2.1 - UPGMA Li et al. (2007) Ananas comosus pineapple SRAP 61 35 16.2 11.0 67.7 UPGMA Dou et al. (2010)

258 UPGMA, Apium graceolens celery SRAP, SSR 60 40 22.2 20.9 94.1 STRUCTURE Wang et al. (2011)

Aquilaria sinensis Incense tree SRAP, ISSR 112 20 16.7 12.9 77.5 UPGMA Zou et al. (2012) Arachis and insp. hybrids peanut SRAP 48 60 10.5 5.9 55.9 UPGMA, PCA Ren et al. (2010) SRAP, TE-based Arundo donax giant cane markers 185 10 18.5 - - similarity indicies Riaz et al (2008) Auricularia aricula jelly ear fungus SRAP 34 11 14.0 13.5 96.1 UPGMA Tang et al. (2010) Auricularia auricula jelly ear fungus SRAP 24 9 10.0 8.7 86.7 UPGMA Liu et al. (2011) Auricularia polytricha cloud ear fungus SRAP 20 10 42.6 42.5 99.8 UPGMA, PCoA Du et al. (2011) Auricularia polytricha cloud ear fungus SRAP, AFLP 19 14 - 32.8 - parsimony Xu et al. (2008)

Subset (n=188) of total SRAP papers discovered in a literature search (n>350) that presented descriptive statistics of SRAP products. These references were used to calculate descriptive statistics reported in Chapter 4.

continued 258

Table E.1 continued

primer fragments polymorphisms % Taxon Common name Markers individuals pairs /primer pair / primer pair polymorphic Analysis methods Reference Auricularia polytricha cloud ear fungus SRAP, ISSR 19 14 32.8 31.4 95.9 parsimony Yu et al. (2008) Boehmeria nivea Ramie SRAP 35 33 10.1 8.6 85.5 UPGMA Liu et al. (2008) cauliflower x mustard SRAP, RFLP, Brassica hybrids hybrid CAPS 8 20 26.4 8.4 31.8 UPGMA Wang et al. (2011) Brassica juncea leaf mustard SRAP 95 8 40.8 20.1 49.4 UPGMA, PCoA Wu et al. (2009) Brassica napus oilseed rape SRAP 258 170 1.1 - - QTL mapping Chen et al. (2007) Brassica napus oilseed rape SRAP 58 1634 8.3 - - LM Sun et al. (2007) Brassica napus oilseed rape SRAP 130 25 20.4 4.9 24.2 UPGMA Wen et al. (2006) Brassica napus oilseed rape SRAP 4 20 21.0 5.8 27.7 - Zhang et al. (2006) SRAP, AFLP, Brassica napus oilseed rape SSR 184 61 2.5 - 91.3 LM Li et al. (2007) SRAP, RAPD, Brassica napus oilseed rape SSR 90 260 - 1.6 - QTL mapping Fu et al. (2007) Brassica napus oilseed rape SRAP, SSR 150 104 - 9.0 - LM Chen et al. (2010)

259 Brassica nupus oilseed rape SRAP, SSRq 51 25 - 7.9 - UPGMA Tan et al. (2009) broccoli, cauliflower,

Brassica oleracea kale SRAP, AFLP 28 16 8.1 - - LM Li & Quiros (2001) SRAP, AFLP, Brassica oleracea broccoli SSR 9 23 - - - cluster analysis Hale et al (2007) SRAP, AFLP, SSR, RFLP, Brassica oleracea broccoli CAPS 140 27 - 2.1 - LM, QTL Okazaki et al. (2007) Brassica oleracea broccoli SRAP, SSR 24 30 2.2 1.7 76.1 LM, QTL Brown et al. (2007) Brassica oleracea var. SRAP, RAPD, botrytis cauliflower ISSR 32 7 16.0 12.6 78.6 UPGMA Wang et al. (2011) SRAP, AFLP, RAPD, SSR, Brassica rapa Chinese cabbage ISSR, Isozyme - 33 - 4.4 - LM Yang et al. (2007) Brassica rapa L. ssp. SRAP, RAPD, chinensis bak choi SSR, ISSR 112 90 - 1.7 - LM Geng et al. (2007)

continued

259

Table E.1 continued

primer fragments polymorphisms % Taxon Common name Markers individuals pairs /primer pair / primer pair polymorphic Analysis methods Reference Brassica rapa L. ssp. Pekinensis Chinese cabbage SRAP, SSR 142 247 30.0 22.7 75.7 LM, QTL Li et al. (2012) Bromeliaceae hybrids bromilliad SRAP 39 11 42.6 39.6 93.0 UPGMA, PCoA Zhang et al. (2012) Buchloe dactyloides buffalograss SRAP 53 34 7.1 7.1 99.2 UPGMA, PCA Budak et al. (2004) Buchloe dactyloides buffalograss SRAP 46 207 13.0 0.0 0.0 QTL mapping Zhou et al. (2011) SRAP, RAPD, cluster analysis, Buchloe dactyloides buffalograss SSR 20 14 2.6 - - AMOVA Budak et al. (2005) SRAP, RAPD, UPGMA, Buchloe dactyloides buffalograss SSRr, ISSR 15 30 8.8 8.3 94.7 AMOVA Budak et al. (2004) Camellia oleifera tea oil SRAP, ISSR 60 11 5.8 5.5 9.1 cluster analysis Zhang et al. (2011) UPGMA, AMOVA, PCoA, Carthamus spp. distaff thistles SRAP 47 12 24.4 18.7 76.5 Shannon Talebi et al. (2012) Celosia populations UPGMA, cockscomb SRAP 22 10 50.7 27.4 54.0 Shannon, Nei Feng et al. (2009)

260 Chaenomeles spp. and hybrids flowering quince SRAP 32 22 - 6.9 - UPGMA Wang et al. (2010)

Chrysanthemum indicum chrysanthemum SRAP 360 24 9.9 8.5 86.1 UPGMA, GST Fang et al. (2012) Chrysanthemum morifolium Hardy garden mum SRAP 144 261 3.4 2.3 68.2 QTL mapping Zhang et al. (2011) Chrysanthemum morifolium Hardy garden mum SRAP 12 62 - 5.6 - QTL mapping Zhang et al. (2011) Cibotium barometz wolly fern SRAP 79 10 10.7 9.1 85.0 UPGMA You et al. (2012) SRAP, AFLP, RAPD, SSR, Citrullus lanatus watermelon ISSR 100 69 - 3.2 - UPGMA Levi et al. (2006) Citrullus spp. and SRAP, TRAP, hybrids watermelon HFO-TAG - 73 - 2.7 - LM Levi et al. (2011) Citrus aurantium bitter orange SRAP, SSR 51 21 9.1 8.0 87.0 UPGMA Polat et al. (2012) SRAP, RAPD, SSR, ISSR, Citrus hybrids citrus POGP, RGA 168 134 2.9 - - LM Gulsen et al. (2010)

continued

260

Table E.1 continued

primer fragments polymorphisms % Taxon Common name Markers individuals pairs /primer pair / primer pair polymorphic Analysis methods Reference SRAP, SSR - Citrus spp. citrus flaking regions 24 10 2.4 - - NJ, MP Amar (2012) Citrus spp. citrus SRAP, SSR, SNP 24 33 21.3 19.9 93.2 UPGMA Amar et al. (2011) Citrus spp. (clementine group) clementine SRAP 42 21 6.0 0.2 4.0 UPGMA Uzun et al. (2011b) Citrus spp. (grapefruit group) graprfruit / pomello SRAP 27 21 6.9 4.3 62.5 UPGMA Uzan et al. (2011c) Citrus spp. (lemon group) lemon SRAP, SSR 56 21 8.7 6.7 77.0 UPGMA Uzun et al. (2011a) cluster analysis, Codonopsis tangshen bellflower SRAP, ISSR 18 29 11.3 9.2 80.9 Shannon's Chen et al. (2009) Coffea arabica coffee SRAP 64 31 23.0 - - similarity indicies Mishra et al. (2011) Ctenopharyngodon UPGMA, idella grass carp SRAP 120 38 23.8 21.2 89.1 Shannon's Ding et al. (2010) Cucumis melo muskmellon SRAP 61 16 28.3 16.6 58.6 LM Chen et al. (2010) Cucumis melo muskmellon SRAP 116 29 7.6 6.4 85.4 LM Wang et al. (2008)

261 SRAP, RAPD, UPGMA, Cucumis melo muskmellon ISSR 82 8 3.4 2.3 66.7 Shannon, Nei Yildiz et al. (2011)

Cucumis sativus cucumber SRAP 125 5 - - 2.7 QTL mapping Chen et al. (2010) Cucumis sativus cucumber SRAP 130 34 5.4 2.6 48.4 QTL mapping Meng et al. (2012) Cucumis sativus cucumber SRAP, AFLP 20 4 27.3 24.0 88.1 LM, QTL Devran et al. (2011) Cucumis sativus cucumber SRAP, ISSR 112 26 - 2.3 - QTL mapping Yeboah et al. (2008) UPGMA, PCoA, Curcurbita maxima large winter squash SRAP, AFLP 120 10 8.8 5.0 56.8 Nei Ferriol et al. (2004) UPGMA, PCoA, Curcurbita moschata small winter squash SRAP, AFLP 47 11 13.5 8.9 66.2 Nei Ferriol et al. (2004) UPGMA, PCoA, Curcurbita pepo field pumpkin SRAP 69 11 8.0 5.8 72.7 Nei Ferriol et al. (2003) cold growing Cymbidium kanran cymbidium SRAP 51 11 53.3 45.8 86.0 UPGMA Jian and Zhu (2010) Cynara cardunculus artichoke thistle SRAP 26 8 - 34.4 - Ward's method Cravero et al. (2008) Cynodon hybrids Bermudagrass SRAP 24 30 9.1 8.3 90.9 UPGMA Wang et al. (2009)

continued 261

Table E.1 continued

primer fragments polymorphisms % Taxon Common name Markers individuals pairs /primer pair / primer pair polymorphic Analysis methods Reference Cynodon radiatus Bermudagrass SRAP 33 15 25.5 24.7 97.1 UPGMA, PCoA Huang et al. (2012) SRAP, RAPD, UPGMA, Cynodon spp. Bermudagrass ISSR, POGP 182 34 - 5.4 - AMOVA Gulsen et al. (2009) Cyperus difformis variable flatsedge SRAP 85 12 6.1 - - Outcrossing rates Merotto et al. (2009) Cyprinus carpio var. color Oujiang color carp SRAP, SCAR 50 12 8.7 - - QTL mapping Hu et al. (2012) Dactylis glomerata orchard grass SRAP 142 16 9.4 4.4 47.0 UPGMA Xie et al. (2010) Dactylis glomerata orchard grass SRAP, RAPD 45 21 22.9 19.3 84.4 UPGMA, Nei Zeng et al. (2008) Dactylis glomerata orchard grass SRAP, SSR 113 36 3.0 LM Xie et al. (2010) Dendrobium loddigesii Loddiges' dendrobium SRAP 92 17 13.6 11.0 81.0 UPGMA Cai et al. (2011) SCAR marker Dendrobium officinal white dendrobium SRAP 84 4 27.3 24.0 88.1 development Ding et al. (2008) SRAP, EST-SSR, Dendrobium spp. bamboo orchid ISSR, RAPD 92 28 - 5.1 - LM Lu et al. (2012) Dendrobium spp. bamboo orchid SRAP, RAPD 9 40 49.4 44.6 90.2 UPGMA Fan et al. (2010) 262 bamboo orchid, Dendrobium spp. Singapore orchid SRAP, RAPD 90 10 9.8 6.5 66.3 LM Xue et al. (2010) Dianthus spp. and cultivars pinks SRAP, ISSR 24 11 15.0 14.4 95.8 UPGMA Fu et al. (2008) Diospyros kaki and related taxa Japanese Persimmon SRAP 27 20 6.8 5.5 80.9 UPGMA, PCoA Guo and Luo (2006) endophytic fungi from Taxus - SRAP 20 24 24.3 18.6 76.4 UPGMA Ren et al. (2012) UPGMA, Milla-Lewis et al. Eremochloa ophiuroides centipedegrass SRAP 49 11 62.5 25.4 40.6 AMOVA, PCoA (2012) SRAP, AFLP, Eriobotrya loquat SSR, ISSR 46 - - 67.8 LM Qiao et al. (2011) UPGMA, NJ, Euodia rutaecarpa Evodia SRAP, AFLP 33 10 18.8 14.5 77.1 Shannon, Nei Wei et al. (2011) Fagopyrum tartaricum buckwheat SRAP 10 26 11.0 9.0 82.5 UPGMA Li et al. (2009) Fasciola hepatica common liver fluke SRAP 34 10 - 5.1 - UPGMA Alasaad et al. (2008)

continued

262

Table E.1 continued

primer fragments polymorphisms % Taxon Common name Markers individuals pairs /primer pair / primer pair polymorphic Analysis methods Reference Fasciola spp. liver fluke SRAP 120 5 15.0 11.8 78.7 UPGMA Li et al. (2009) NJ, AMOVA, SRAP, RAPD, PCA Ficus carica common fig SSR 96 87 - 1.5 - STRUCTURE, Fst Ikten et al. (2010) Panama disease Fusarium oxysporum fungus SRAP, ISSR 15 4 4.5 4.5 100.0 UPGMA Baysal et al. (2009) Galega officinalis goat's rue, French lila SRAP 35 7 12.6 8.4 67.0 UPGMA Wang et al. (2012) Ganoderma spp. Lacquered Bracket SRAP 31 6 - 14.2 - cluster analysis Sun et al. (2006) SRAP, SSR, EST-SSR, Gossypinum hirsutum cotton RAPD, RGAP 139 4096 1.2 0.1 6.9 LM Lin et al. (2009) Gossypinum hybrids cotton SRAP 13 26 - 6.6 - LM Lin et al. (2004) SRAP, RAPD, Gossypinum hybrids cotton SSR, REMAP 207 121 - 3.6 - QTL mapping He et al. (2007) Grifola frondosa hen of the wood, SRAP, ITS- maitake RFLP 25 47 - 2.9 - UPGMA Zhang et al. (2009)

263 Hedychinum spp. ginger lily SRAP 22 132 - 3.1 - LM Gao et al. (2008) SRAP, RAPD, Wangsomnuk et al.

Helianthus tuberosus Jerusalem artichoke ISSR 47 9 - 21.6 - NJ, STRUCTURE (2011) SRAP, RAPD, Hibiscus cannabinus kenaf ISSR 180 78 - 1.3 - LM Chen et al. (2011) Hippophae spp. sea buckthorn SRAP 77 17 16.9 11.2 66.6 UPGMA Li et al. (2010) Hordeum vulgare barley SRAP, SSR 152 21 - 3.7 - LM Guo et al. (2012) Lactuca sativa lettuce SRAP 50 18 30.5 25.8 84.5 UPGMA Liu et al. (2011) SRAP, RAPD, Lens culinaris lentil SSR 206 60 4.5 1.8 40.0 LM Saha et al. (2010)b SRAP, RAPD, SSR, RFLP, Lens culinaris lentil PSMPs 196 60 4.5 0.7 14.8 LM Saha et al. (2010)a SRAP, RAPD, Lentinula edodes Shiitake ISSR 23 23 6.3 3.5 56.3 UPGMA Fu et al. (2010)

continued

263

Table E.1 continued

primer fragments polymorphisms % Taxon Common name Markers individuals pairs /primer pair / primer pair polymorphic Analysis methods Reference Lilium auratum var. cluster analysis, Yamamoto et al. platyphyllum Golden-rayed Lily SRAP 65 5 - 7.0 - Nei (2012) Lilium spp. lily SRAP 23 19 12.4 12.3 98.7 UPGMA Chi et al. (2007) Lilium spp. lily SRAP 23 19 12.4 12.3 98.7 UPGMA Li et al. (2011) Litchi chinensis and lychee and false Dimocarpus confinis longan SRAP 32 9 37.7 37.6 99.7 UPGMA Zhou et al. (2011) Luffa spp. luffa SRAP 64 60 23.9 21.3 89.3 UPGMA Cui et al. (2012) UPGMA, AMOVA, Shannon, Nei, Lycium ruthenicum buckthorn SRAP 174 31 15.1 12.8 85.0 GST Liu et al. (2012) Malus sieversii wild apple SRAP, SSR 109 10 20.9 - - UPGMA Zhang et al. (2009) SRAP, AFLP, Manihot esculenta cassava SSR, EST-SSR 210 15 - 3.7 - LM Chen et al. (2010) Medicago sativa Ariss & Vandemark alfalfa SRAP 38 7 26.9 26.0 96.8 UPGMA (2010) 264 UPGMA, AMOVA, PCA,

Medicago sativa alfalfa SRAP 48 14 13.8 6.8 49.2 Fst Talebi et al. (2011) Medicago sativa alfalfa SRAP 3 14 17.8 16.1 90.4 UPGMA Vandemark (2006) Meloidogyne incognita Root Knot Nematode SRAP 20 20 8.8 1.0 11.4 UPGMA Devran et al. (2012) Monascus spp. yeast SRAP, ISSR 2011 8 22.9 21.6 94.5 UPGMA Shao et al. (2011) Morus spp. mulberry SRAP 23 12 6.9 4.9 71.1 UPGMA, Nei Zhao el al. (2009) Musa hybrids banana SRAP 29 25 - 13.0 - UPGMA Wei et al. (2011) Musa spp. banana SRAP, AFLP 40 10 40.3 35.3 87.6 UPGMA, PCoA Youssef et al. (2011) UPGMA, Nelumbo spp. lotus SRAP, SSR 43 16 11.4 7.4 65.0 STURCTURE Yang et al. (2012) Olea europaea oliva SRAP, SSR 66 13 14.2 7.9 55.7 UPGMA Isk et al. (2011) Ophiopogon spp. mondo grass SRAP 48 11 56.5 1.0 100.0 cluster analysis Ma et al. (2010) Paeonia hybrids Peony SRAP 29 24 8.2 7.8 94.9 UPGMA, PCA Hao et al. (2008)

continued

264

Table E.1 continued

primer fragments polymorphisms % Taxon Common name Markers individuals pairs /primer pair / primer pair polymorphic Analysis methods Reference common garden Paeonia lactiflora peony SRAP 13 8 10.1 8.5 84.0 UPGMA Guo et al. (2011) UPGMA, AMOVA, Paeonia ludlowii Ludlow's tree peony SRAP 79 17 23.3 21.0 90.2 Shannon, GST Tang et al. (2012) Paeonia suffruticosa tree peony SRAP 66 23 12.9 11.4 88.5 NJ Han et al. (2008) Paeonia suffruticosa tree peony SRAP 6 25 6.0 4.8 80.5 UPGMA Han et al. (2008) UPGMA, Panicum spp. switchgrass SRAP 91 28 18.6 18.0 96.7 AMOVA Huang et al. (2011) Pennisetum purpureum elephant grass SRAP 60 62 22.5 8.9 39.5 UPGMA, PCoA Xie et al. (2009) Phalaenopsis ‘Frigdaas Oxford’ moth orchid SRAP 159 14 - 5.7 80.8 PCoA Lu et al. (2011) SRAP, AFLP, Phyllostachys violascens bamboo ISSR 17 15 14.8 10.1 68.5 UPGMA, PCA Lin et al. (2011) Pinus koraiensis Korean pine SRAP 94 70 - 2.5 - QTL mapping Chen et al. (2010) Pinus koraiensis UPGMA, 265 AMOVA, Korean pine SRAP 480 9 27.7 15.9 57.4 Shannon, Nei, Nm Feng et al. (2009)

NJ, AMOVA, PCoA, Fst, Pistacia spp. pistacio SRAP 36 11 18.4 15.3 83.2 Shannon Talebi et al (2012) Euclidean, Ward, Pisum sativum garden pea SRAP 40 15 61.3 10.8 17.6 PCA Espsito et al. (2007) SRAP, ISSR, UPGMA, Pleurotus cultivars oyster mushroom morphometrics 20 8 8.6 8.1 94.2 Shannon, Nei Zhang et al. (2012) UPGMA, Pogostemon cablin patchouli SRAP, ISSR 192 18 16.3 13.4 82.0 Shannon, Nei Wu et al. (2010) Populus hybrids poplar SRAP, SSR 189 40 - 3.3 - LM Wang et al. (2010) Porphyra red algae SRAP 16 14 38.1 37.3 97.9 UPGMA Qiao et al. (2007) Porphyra haitanensis red seaweed SRAP, SSR 157 15 35.7 8.2 23.0 LM Chaotian et al. (2010) Praecitrullus fistulosus SRAP, SSR, and related taxa round melon EST-SSR, 16 18 - 34.9 - UPGMA Levi et al. (2010)

continued

265

Table E.1 continued

primer fragments polymorphisms % Taxon Common name Markers individuals pairs /primer pair / primer pair polymorphic Analysis methods Reference UPGMA, PCoA, Prunella vulgaris heal-all SRAP, ISSR 26 17 9.0 7.9 87.6 GST Liao et al. (2012) Prunus and Armeniaca spp. cherry and apricot SRAP 28 15 19.1 16.8 88.1 UPGMA Ai et al. (2011) UPGMA, Prunus spp. cherry SRAP 53 13 8.8 4.2 48.2 AMOVA Abedian et al. (2012) Prunus spp. wild peach and allies SRAP 190 53 - 12.0 - LM Cao et al. (2011) Prunus spp. peach / nectarine SRAP, SSR 38 10 - 4.9 - UPGMA Amhad et al. (2004) Puccinia striiformis stripe rust SRAP 16 9 29.4 13.1 44.5 NJ Pasquali et al. (2010) NJ, AMOVA, Soleimani et al Punica granatum pomegranate SRAP 63 13 19.2 10.2 53.2 Shannon, Nm (2012) Rabdosia rubescens blushred rabdosia SRAP 16 72 9.5 0.0 0.1 - Ai et al. (2012) Raphanus sativus radish SRAP, AFLP 7 11 14.1 7.6 54.2 UPGMA, Nei Zhao el al. (2007) SRAP, AFLP, Raphanus sativus radish RAPD 17 13 18.7 11.5 61.7 UPGMA Liu et al. (2007) SRAP, RAPD, Raphanus sativus radish ISSR 35 17 13.7 11.7 85.4 UPGMA Liu et al. (2008) Rehmannia glutinosa Chinese Foxglove SRAP 23 13 26.0 23.5 90.5 UPGMA, Shannon Zhou et al. (2010)

266 Rosa rugosa Japanese rose SRAP 50 15 17.6 15.1 86.0 Shannon, Nei Xu el at. (2011) Rutaceae subfamily

Aurantioideae citrus SRAP 86 21 - 17.9 - UPGMA Uzun et al. (2009) Saccharum spp. sugarcane SRAP 30 31 44.0 36.6 83.2 UPGMA Suman et al. (2008) SRAP, AFLP, Saccharum spp. sugarcane TRAP 100 32 - 5.0 - QTL mapping Alwala et al. (2007) UPGMA, Shannon, Nei, Salvia miltiorrhiza Chinese sage SRAP, SSR 48 6 20.3 18.3 90.2 GST, Nm Song et al. (2010) NJ, Shannon's Salvia splendens scarlet sage cultivars SRAP 24 24 - 12.8 - GST Dong et al. (2012) UPGMA< Schistosoma japonicum Blood fluke SRAP 21 5 16.6 12.2 73.5 sequence analysis Song et al. (2011) Sesamum indicum sesame SRAP 67 21 26.7 12.6 47.2 UPGMA Sun et al. (2009) Siraitia grosvenorii Monk's fruit SRAP 3 189 24.2 3.1 12.6 Sequence analysis Fu et al. (2012)

continued

266

Table E.1 continued

primer fragments polymorphisms % Taxon Common name Markers individuals pairs /primer pair / primer pair polymorphic Analysis methods Reference Siraitia grosvenorii Monk's fruit SRAP, ISSR 150 74 3.0 - - LM Liu et al. (2011) Comlekcioglu et al. Solanum lycopersicum tomato SRAP, RAPD 15 11 6.0 2.5 42.4 UPGMA (2010) Solanum lycopersicum and related taxa tomato SRAP, SSR 21 13 13.7 7.5 55.1 UPGMA, PCoA Ruiz et al. (2005) Solanum lycopersicum var. SRAP, RAPD, cerasiforme cherry tomato RGA 10 1 - 5.0 - UPGMA Yu et al. (2005) Solanum melongena eggplant SRAP 56 55 20.6 11.5 56.0 UPGMA, PCoA Li et al. (2010) SRAP, SRAP- SCAR marker RGA, RGA, development, QTL Solanum melongena eggplant RAPD, SCAR 720 208 2.9 - - mapping Mutlu et al (2008) Solanum tuberosum potato SRAP 44 20 6.1 5.2 85.2 cluster analysis He et al. (2007) SCAR marker Tagetes erecta marigold SRAP, ISSR 167 170 4.5 0.2 3.4 development He et al. (2009) Tricholoma matsutake pine mushroom SRAP 129 12 12.8 9.8 76.6 UPGMA Ma et al. (2012) 267 SRAP, EST-SSR, Triticum aestivum common wheat ISSR, TRAP 1064 19 - 1.4 - LM, QTL Li et al. (2007)

Triticum aestivum 'Thatcher' common wheat SRAP 23 41 26.5 13.1 49.4 UPGMA Liu et al. (2008) Triticum dicoccoides wild emmer SRAP 120 30 14.6 8.1 55.7 UPGMA Dong et al. (2010) Zaefizadeh et al. Triticum durum Durum wheat SRAP 40 12 5.4 3.3 60.0 UPGMA, Nei (2009) UPGMA, Tropaeolum tuberosum mashua (nasturtium) SRAP 153 8 - 15.0 - AMOVA, PCA Ortega et al. (2007) Alghamdi et al. Vicia faba fava bean SRAP 58 14 74.0 74.0 100.0 UPGMA (2012) Viola xwittrockiana garden pansy SRAP 43 21 25.5 23.8 93.5 UPGMA Wang et al. (2012) Vitis vinifera and related taxa grape SRAP 76 19 12.0 9.5 78.9 UPGMA, PCoA Guo et al. (2012) Ziziphus spp. jujube SRAP 26 19 30.5 30.0 98.3 UPGMA Li et al. (2009)

267