GERMPLASM COLLECTION, CHARACTERIZATION, AND ENHANCEMENT OF EASTERN

DISSERTATION

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

By

Peter Jeffrey Zale, M.S.

Graduate Program in Horticulture and Crop Science

The Ohio State University

2014

Dissertation Committee:

Dr. Pablo Jourdan, Advisor

Dr. Mark Bennett

Dr. David Francis

Dr. John Freudenstein

Copyrighted by

Peter J. Zale

2014

ABSTRACT

Plants from the genus Phlox have become a staple of gardens worldwide since their

introduction to cultivation about 200 years ago, and are admired as versatile garden

and landscape , container plants, and cut flowers. Their long-lasting, intensive

floral displays in a variety of colors, forms, and seasons rank them among the most

widely recognized hardy perennials and annuals. The floral beauty has resulted in

extensive breeding and selection in three primary species: the annual Phlox

drummondii, and the hardy perennials P. paniculata, and P. subulata. Among these

three, hundreds of have been selected over the last 200 years, but the genus

includes other species that may have ornamental potential. Phlox L. ()

consists of ca. 65 species primarily endemic to ; 20-23 species occur in

the eastern U.S.A. and 40-45 in the west. The western taxa occur primarily in arid,

mountainous habitats whereas the eastern taxa occur in a different range of diverse

ecosystems and habitats. The eastern species are a polymorphic group arranged into

6 subsections within the genus; this group includes the three main cultivated species

and also up to 20 related species that exhibit ornamental characteristics, yet are rarely

cultivated.

Horticultural and botanical interest in Phlox has resulted in a large volume of literature pertaining to the , ecology, and cultivation, but there is

ii

recognition that the genetic diversity of the genus has barely been tapped for

ornamental use or as a source of desirable traits, such as disease resistance, and that

more widespread availability of diverse germplasm can contribute not only to new

cultivated forms but also to a greater understanding of species diversity and

interspecific relationships. Such interest has made Phlox a priority genus for

germplasm development and conservation at the Ornamental Germplasm Center

of The Ohio State University. The work presented here describes the development,

characterization and manipulation of such germplasm.

The development of a Phlox germplasm collection was initiated in 2010 with an effort to collect seeds and plants of the eastern species from natural populations throughout their distributional range. A total of 187 accessions were collected from wild populations of 22 eastern Phlox species during a series of expeditions to diverse ecosystems throughout the eastern United States. Another 166 accessions were of cultivated origin, acquired from nurseries and individuals; these represented a mix of selections from wild populations, selected seedlings, or hybrids and were used for comparison to wild-collected accessions. The 353 accessions represent the most comprehensive germplasm collection of Phlox to date that include taxa related to the three horticulturally important species. An additional 8 species of western USA

Phlox were also acquired as a source of comparative material. This germplasm was characterized in three primary ways.

The characterization of the germplasm collection focused on estimation of genome size using flow cytometry, coupled with validation of ploidy estimates by

iii chromosome counts. Genome size was surveyed in 287 accessions derived from 29 species (45 taxa); of these, 165 accessions were of wild origin and the rest were cultivated material. The majority of taxa were diploid, but genome size was variable.

Genome size consistent with tetraploid and hexaploid levels were found in species which had not been previously reported. Polyploidy in natural Phlox populations seems to occur primarily in populations at the distribution edges suggesting that it may be associated with genetic differentiation, diversification, and adaptation to marginal habitats. The 122 cultivars and hybrids analyzed for genome size were also primarily diploid, but anueploids were discovered in P. paniculata suggesting the use of polyploid breeding in this species. These data have important implications in the conservation, evolution, adaptation, and commercial breeding of Phlox.

Characterization of the germplasm collection also included a limited application of microsatellite (SSR) marker analysis to 8 populations, 5 diploid and 3 tetraploid, and 61 individuals from the P. pilosa complex. The data indicated that tested populations of P. pilosa had high genetic diversity and moderate population structuring, typical of outcrossing species with gametophytic self - incompatibility.

These preliminary data indicate that P. pilosa is a rich source of genetic diversity for

Phlox germplasm collections and enhancement.

The manipulation of germplasm involved a series of interspecific hybridization experiments; in two studies there was at least one recurring parent and the other study was performed in a partial diallel. A total of 157 unique hybrids were produced. Interspecific hybridization was most successful among closely related

iv

species, and the rate of success decreased with increasing phylogenetic distance;

however, even crosses between closely related species failed to produce seeds, and

indicate that selection of parental taxa is critical. Attempts at recreating previously described hybrid combinations were of limited success, but at least one was successfully resynthesized. Previous reported attempts at interploid crosses failed, but were successful and resulted in morphologically intermediate, aneuploid taxa with intermediate genome sizes to parental taxa.

In conclusion, the development of a comprehensive germplasm collection enabled a broad characterization and enhancement of eastern U.S.A Phlox. These initial studies have been necessary to provide baseline information for more focused studies pertaining to future Phlox characterization and hybridization studies.

v

Dedication

This work is dedicated to Simon P. Zale

vi

ACKNOWLEDGMENTS

The sum of my graduate experience has amounted to much more than the research and stories presented in the following pages, and was only possible though interactions with many advisors, colleagues, and friends that enriched the entire experience.

I extend thanks to my committee members Dr. Mark Bennett, Dr. David Francis, and Dr. John Freudenstein for faithfully attending committee meetings and providing feedback and ideas.

I thank my advisor, Pablo Jourdan for the privileges and autonomy I enjoyed during the research process and the extraordinary priviledge of pursuing my own interests within the context of the OPGC mandate. Many of the germplasm collection trips rank among the finest and most influential experiences of my life.

Gratitude is expressed to the staff of the OPGC. In Particular, Susan Stieve, Eric

Renze, Russell Eckley, and Steven Haba. Thanks for helping to create an environment conducive to my research. Two student workers, Elizabeth Reeder and Marcus Nichols, were intimately involved in Phlox research and magnitude of what was accomplished would not have been possible without them, and their devotion to thhe numerous facets of the project.

My deepest gratitude is expressed to my wife, Kate, for supporting my schooling for the last 7 years. I have been in graduate school longer than we have been married,

vii

and the completion of this document is also the delivery of a promise made long ago. I

thank my family for their unwavering support of my pursuit.

Many individuals helped me obtain germplasm samples, permits, and/or provided information, discussions, greenhouse space, and inspiration that were vital the synthesis of this project. Thanks are extended to my friend Dr. Daniel W.H. Robarts, who was a common companion in the field and lab, and linked with many of my best memories of graduate school. I thank Dr. Dan Struve, Charles and Martha Oliver, Dr. Warren

Stoutamire, Jason Parrish, Julian Campbell, Ron Determann, Matt Richards, Tom

Mitchell, Aaron Floden, Mike Graziano, Marianne Casey, Art Evans, Jim Rodgers, Tony

Avent, Carolyn Ferguson, Jim Ault, Shannon Fehlberg, Rick Gardner, Alan Blowers, Bill

Martin, Michael Jenkins, Ron Miller, Brandon Sinn, Barrie Francis, Rimmer de Vries,

Jim Vent, David Snodgrass.

Several institutions contributed to this work in numerous ways and I thank the following; The staff at the The Columbus Metroparks, Atlanta Botanical Garden, The

Cincinnati Zoo and Botanical Garden, The Greater Des Moines Botanical Center, Oak

Openings preserve, LadyBird Johnson Wildflower Center, Mount Cuba Center, The Ohio

State University Herbarium. Many others also helped in the completion of this work and

I thank you.

viii

VITA

June 1997 ...... Olmsted Falls High School

1997 - 2001...... B.S. Horticulture, The Ohio State University

2001 - 2007...... Nursery Production and Retail Manager. Marvin’s Organic Gardens, Lebanon, Ohio 2007 - 2009...... M.S., The Ohio State University

2009 -Present...... Ph.D Candidate, The Ohio State University

PUBLICATIONS

Zale, Peter J.,and P. Jourdan. 2012. Germplasm Development and Preliminary Interspecific Hybridization in Phlox. Acta Horticulturae 953:71-78

Zale, Peter J., Struve, D.K., Jourdan, P., and D.M. Francis. 2011. Rapid assessment of genetic variation for morphological traits in Sweetbay Magnolia using a container production system. J. Am. Soc. Hort. Sci. 136: 135-144

Zale, Peter J. 2012. Using X-ray technology to study secret lives of Clematis carrizoensis seeds. Yearbook of the British Clematis Society. 2012

Zale, Peter J., and P. Jourdan. 2012. Work with the threatened species Lilium iridollae at the Ornamental Plant Germplasm Center. The Quarterly Bulletin of the North American Lily Society.

Zale, Peter J. and P. Jourdan.”Educational Update - Phlox 101: Perspectives on a underutilized genus of native plants.” The Buckeye. Oct. 2011: 25-31. Print.

Zale, Peter J. 2009. On the Trail of Magnolia virginiana: A Trip to Western Tennessee and the Gulf Coast. J. Mag. Soc. Int. 44(85): 5-17

Zale, Peter J. Field Notes: Stachys monieri ‘Hummelo’, Hummelo Wood Betony. American Nurseryman, April 15, 2009 pg. 71 ix

Zale, Peter J. Field Notes: Cypripedium parviflorum var. pubescens, Large Yellow Lady’s Slipper. American Nurseryman, June 15, 2008 pg. 66

Zale, Peter J. Field Notes: Styrax americanus, American, Snowbell. American Nurseryman, October 1, 2007 pg. 74

FIELDS OF STUDY

Major Field: Horticulture and Crop Science

x

TABLE OF CONTENTS

Page

Abstract…………………………………………………………………………………….ii Dedication………………………………………………………………………………....vi Acknowledgements…………………………………………………………………...... vii Vita…………………………………………………………………………………...... ix List of figures…………………………………………………………………………….xvi List of tables ……………………………………………………...... xxvi

Chapters:

1. Development of a Comprehensive Phlox Germplasm Collection…………………..1

Introduction………………………………………………………………….. 1 Methodology……………………………………………………………….... 4 Locating Phlox populations…………………………………………………. 6 Phlox seed development, dispersal, and collection………………………….. 7 Collection and propagation of vegetative germplasm ………………………. 11 Seed increase of Phlox germplasm accessions………………………………. 12 Taxonomy and Phylogenetic relationships of eastern Phlox taxa………….... 13 Notes on Field Identification of Phlox………………………………………. 19 Summary of Collections……………………………………………………... 20 Summary of germplasm collections of Eastern Phlox taxa by Subsection, Species, and Subspecies……………………………………………………... 21 Subsection Cluteanae Wherry………………………………………………. 21 Phlox buckleyi Wherry - Greenbrier phlox, swordleaf phlox……………….. 22 Subsection Divaricatae (Wherry) Prather…………………………………… 24 Taxonomy and Phylogenetics of Subsection Divaricatae………………...... 24 Sims - hairy phlox, chalice phlox……………………………. 26 Phlox divaricata L. ssp. divaricata and Phlox divaricata ssp. laphamii (Wood) Wherry - wood phlox………………………………………………. 29 The Texas annual of subsection Divaricatae………………………. 33 Phlox drummondii Hooker - Drummond’s phlox………………………..…. 33 Phlox cuspidata Scheele - Navasota phlox……………………………….… 36 Phlox roemeriana Scheele - golden-eye phlox……………………………... 37 Phlox floridana Bentham - florida phlox……….………………………….. 38 Phlox pattersonii - Coahuila phlox………………………..…….………...... 40 The complex…………………………………………………... 41 xi

Phlox pilosa L. spp. pilosa - downy phlox………………………………..… 42 Phlox pilosa ssp. deamii (Deam) Levin - Deam’s downy phlox……………. 48 Phlox pilosa ssp. fulgida (Wherry)Wherry - Dakota downy phlox…………. 53 Phlox pilosa ssp. longipilosa (Waterfall) Locklear) - Kiowa downy phlox.... 54 Phlox pilosa ssp. ozarkana (Wherry) Wherry, Ozark downy phlox………… 56 Phlox pilosa ssp. sangamonensis Levin and D.M. Smith - Sangamon River downy phlox………………………………………………………………..... 58 Phlox pulcherrima (Lundell)Lundell (syn. Phlox pilosa ssp. pulcherrima) - big thicket phlox…………………………………………………………...... 60 Phlox villosissima Turner (Syn. P. pilosa ssp. latisepala and P. pilosa ssp. riparia) - Comanche phlox………………………………………………….. 62 Subsection Paniculatae Wherry……………………………………...... 65 Phlox amplifolia Britton - broadleaf phlox……….…………………………. 65 L. - summer or border phlox……………………………… 68 Subsection Phlox Ferguson………………………………………………….. 70 Taxonomy and phylogenetics of subsection Phlox………………………….. 70 Horticultural taxonomy of subsection Phlox………………………………... 74 L. ssp. carolina - carolina or thick-leaved phlox…………... 76 Phlox carolina ssp. angusta Wherry / narrow leaved variants of P. glaberrima ssp. glaberrima Wherry………………………………………… 79 Phlox carolina ssp. alta Wherry…………………………………………….. 82 Phlox carolina ssp. turritella Wherry……………………………………….. 83 L. ssp. glaberrima Wherry - Smooth Phlox……………... 83 Phlox glaberrima ssp. interior Wherry - Wabash smooth phlox……………. 85 Phlox glaberrima ssp. triflora Wherry (syn. P. triflora Michaux) - three flower smooth phlox………………………………………………………… 87 L. - meadow phlox…………………………………………. 90 Phlox ovata (L.) Locklear - mountain phlox………………………………… 93 Phlox pulchra Wherry - Alabama Phlox…………………………………….. 96 Subsection Stoloniferae……………………………………………………… 99 Phlox stolonifera Sims - Cherokee phlox…………………………………... 99 Subsection Subulatae………………………………………………………... 101 Phlox bifida Beck ssp. bifida - cleft, sand, or ten-point phlox………………. 102 Phlox bifida ssp. arkansana Marsh………………………………………….. 103 Phlox bifida ssp. stelleria (Gray) Wherry - Kentucky cleft phlox…………... 104 Phlox nivalis Loddiges ssp. nivalis Wherry - trailing phlox………………… 106 Phlox nivalis ssp. texensis Lundell - Texas trailing phlox…………………... 108 Phlox oklahomensis Wherry - Oklahoma Phlox…………………………….. 108 L. ssp subulata - creeping phlox - moss phlox, thrift………. 109 Phlox subulata L. ssp. brittonii (Small) Wherry - shalebarren creeping phlox………………………………………………………………………… 111 Phlox subulata ssp. setacea (L.) Locklear - Blue Ridge creeping phlox……. 112 Future Phlox Collections…………………………………………………….. 113 Characterization of Phlox Germplasm………………………………………. 115 Figures……………………………………………………………………….. 118

xii

Tables………………………………………………………………………... 154

2. Genome size and ploidy level in Phlox species and cultivars: Upright and long styled taxa from subsection Paniculatae and Phlox…………………………….. 164

Abstract……………………………………………………………………... 164 Introduction…………………………………………………………………. 165 Materials and Methods……………………………………………………… 171 Phlox collection……………………...……………………………………… 171 Tissue Sampling…………………………………………………………….. 172 Chromosome counts………………………………………………………… 172 Isolation of nuclei and genome size analysis……………………………….. 173 Statistical Analysis………………………………………………………….. 174 Results………………………………………………………………………. 174 Subsection Paniculatae……………………………………………………... 175 Subsection Phlox……………………………………………………………. 176 Discussion…………………………………………………………………... 178 Conclusion………………………………………………………………….. 187 Figures………………………………………………………………………. 188 Tables……………………………………………………………………….. 191

3. Genome size and ploidy level in Phlox species and cultivars: Mat forming taxa from eastern and western North America………………………………………. 198

Abstract……………………………………………………………………... 198 Introduction…………………………………………………………………. 199 Materials and methods……………………………………………………… 205 Phlox collection……………………………………………………………... 205 Tissue sampling, flow cytometry, chromosome counts, and statistical analysis……………………………………………………………………… 205 Results………………………………………………………………………. 205 Western U.S. mat forming Phlox…………………………………………... 206 Eastern U.S. mat forming Phlox…………………………………………… 208 Microsteris gracilis…………………………………………………………. 209 Cultivars and hybrids……………………………………………………….. 210 Discussion…………………………………………………………………... 211 Conclusion.…………………………………………………………………. 219 Figures………………………………………………………………………. 220 Tables……………………………………………………………………….. 221

4. Genome size, genetic diversity,and population structure in Phlox pilosa L. and related taxa in subsection Divaricatae…………………………………………... 226

Abstract…………………………………………………………………….. 226

xiii

Introduction………………………………………………………………… 227 Materials and methods……………………………………………………… 233 Phlox accessions……………………………………………………………. 233 Tissue sampling, flow cytometry analysis, and chromosome counts………. 234 DNA extraction……………………………………………………………... 234 Genotyping and amplififcation……………………………………………... 234 Genetic analysis…………………………………………………………….. 235 Results………………………………………………………………………. 238 Cytotypic variation and distribution in subsection Divaricatae……………. 238 Cytotypic variation among taxa in subsection Divaricatae………………... 239 Allelic variation at microsatellite loci………………………………………. 244 Genetic diversity……………………………………………………………. 244 Genetic structure……………………………………………………………. 245 Discussion…………………………………………………………………... 246 Figures………………………………………………………………………. 254 Tables……………………………………………………………………….. 258

5. Interspecific hybridization in Phlox for germplasm enhancement: Hybridization potential of Phlox paniculata L. and related species…………….. 268

Abstract……………………………………………………………………... 268 Introduction…………………………………………………………………. 269 Materials and methods……………………………………………………… 275 Phlox collection……………………………………………………………... 275 Controlled ……………………………………………………… 276 Data analysis………………………………………………………………... 278 Results………………………………………………………………………. 279 Parents for hybridization……………………………………………………. 279 Overview of hybridization………………………………………………….. 280 Hybridization with taxa in subsection Paniculatae………………………… 281 Hybridization between subsections Paniculatae and Phlox………………... 282 Hybridization between subsections Paniculatae and Divaricatae…………. 283 Discussion…………………………………………………………………... 284 Figures………………………………………………………………………. 292 Tables……………………………………………………………………….. 297

6. Interspecific hybridization in Phlox for germplasm enhancement: Hybridization potential of Phlox ‘Minnie Pearl’………………………………………………... 302

Abstract……………………………………………………………………... 302 Introduction…………………………………………………………………. 303 Materials and methods……………………………………………………… 307 Parent taxa and pollinations………………………………………………… 307 Growing conditions and germination……………………………………….. 309 Flow cytometric analysis…………………………………………………… 309

xiv

SRAP markers………………………………………………………………. 309 Data analysis………………………………………………………………... 310 Results ……………………………………………………………………… 310 Parents for hybridization……………………………………………………. 310 Overview of hybridization………………………………………………….. 311 Discussion…………………………………………………………………... 314 Figures………………………………………………………………………. 321 Tables……………………………………………………………………….. 322

7. Interspecific hybridization in Phlox for germplasm enhancement: Interploid crosses in Phlox subsection Divaricatae………………………………………… 326

Abstract……………………………………………………………………... 326 Introduction…………………………………………………………………. 327 Materials and methods……………………………………………………… 333 Phlox collection…………………………………………………………….. 333 Controlled pollinations……………………………………………………… 333 Flow cytometric analysis, chromosome counts, and SRAP analysis……….. 335 Pollen staining………………………………………………………………. 335 Results………………………………………………………………………. 336 Discussion…………………………………………………………………... 338 Figures………………………………………………………………………. 347 Tables……………………………………………………………………….. 352

Bibliography………………………………………………………………………… 355 .

xv

LIST OF FIGURES

Figure Page

1.1 Map of the natural range of eastern Phlox species and collection sites. Collection expeditions were focused in three sub-centers of Phlox species diversity. Sub-centers were defined based on the number of taxa reported from each state within the range of eastern species………………………..... 118

1.2 Physiogeographic provinces of the United States. Map obtained from http://apeoplesconstitution.wikispaces.com/Regions+Maps..... …………...... 119

1.3 The mature capsule of Phlox. This is the stage at which mature seed can be collected. a. The mature capsules of P. bifida ssp. bifida PZ10-149 on May 19, 2010 in Newton County, Indiana…………………………………...... 120

1.4 Differences in style length of Phlox taxa. a. The long style (10-25 mm) of P. paniculata is representative of subsections Paniculatae and Phlox. b. The long style (6-12 mm) of P. subulata is representative of long-styled in subsection Subulatae. c. The short style (1-4 mm) of P. villosissima is representative of subsection Divaricatae. The scale bar is 10 mm………….. 121

1.5 Flowers and habit of Phlox buckleyi in Greenbrier County, West Virginia. a. Detail of inflorescence and flowers of P. buckleyi. Note the characteristic narrow, lanceolate leaves of this taxon. b. Detail of the prolifertation of loosely rhizomatous sterile stems characteristic nof this species. …………... 122

1.6 Phlox amoena in the wild and in cultivation at the OPGC. a. A flower color variant of P. amoena with flowers that open white with lavender striae fused into a star that fades to pink. b. Two different forms of P. amoena at the OPGC showing variation in flower color. c. Phlox amoena in situ along a roadbank in Benton County, Tennessee. This species frequently inhabits disturbed areas……………………………………………………………….. 123

xvi

1.7 Flowering habit of Phlox divaricata ssp. divaricata in Hocking County, Ohio. a. Typical individual. b. A white-flowered individual. Such flower color variants of this taxon are easily found in large populations within the Great Plains regions of this taxon and P. divaricata ssp. laphamii……..…. 124

1.8 Flower detail and natural habitat of Phlox divaricata ssp. laphamii in Gadsden County, Florida, at the southern edge of this taxon’s range. a. The flowers of P. divaricata ssp. laphamii lack the notch in the in the tip of the corolla lobe that is present in P. divaricata ssp. divaricata. b. The steep, north facing bluff habitat of P. divaricata ssp. laphamii, where it grows with Adiantum capillus-veneris, Decumaria barbara, and Trillium lancifolium…………………………………………………………………… 125

1.9 Flowering characteristics of two subspecies of Phlox drummondii in situ. a. The natural habitat of P. drummondii in Caldwell County, Texas on May 28, 2010. b. A typical, flowering individual of P. drummondii ssp. drummondii PZ10-161 flowering on May 28, 2010 in Caldwell County, Texas. b. A typical, flowering individual of P. drummondii ssp. mccallisteri flowering along a roadside in Wilson County, Texas on May 29, 2010………………………………………………………………………….. 126

1.10 Phlox roemeriana PZ10-165 (TX-057) in situ, Comal County, Texas. a. The xeric, limestone habitat of P. roemeriana. b. A typical, flowering individual on May 29, 2010. c. Another flowering individual illustrating the range of the flower color variation…………………………………………... 127

1.11 Phlox floridana PZSH2011-010. a. The xeric, upland pine (Pinus spp.) habitat of P. floridana in Jackson County, Florida. The plants were most abundant on the steep slope of the road cut (not in flower) and growing with Lupinus perennis, April 4, 2011. b. Phlox floridana PZSH2011-010 flowering in cultivation at the OPGC. When grown under long days, the plants will flower throughout the year……………………………………….. 128

1.12 The ecotype of Phlox pilosa ssp. pilosa that occurs in the southeastern Great Lakes region and Ohio River valley of the Great Plains province. a. The limestone barren/cedar glade habitat of P. pilosa ssp. pilosa in Hardin County, Kentucky in the Interior Low Plateau province. b. A flowering individual of P. pilosa ssp. pilosa in Hardin County, Kentucky on April 24, 2012. c. The habitat of P. pilosa ssp. pilosa PZ12-060 is a tall grass prairie remnant in the Great Plains province in Lake County, Indiana. Plants grow in nutrient rich soils associated with this habitat. d. A flowering individual of P. pilosa ssp. pilosa………….……………………………………………. 129

xvii

1.13 Variation in Phlox pilosa ssp. pilosa PZSH2011-020 from Forest County, Mississippi. Phenotypic variation is greater in populations from the Coastal Plain and Interior Low Plateaus than in populations from the Great Plains. a. An exceptional color variant displaying white flowers with pink striae fused into a star. b. A richly colored individual that also displays c. Uniquely purple flushed foliage……………………………………………... 130

1.14 Phlox pilosa ssp. deamii PZ11-026 in situ, Christian County, Kentucky. a. The open, mesic forest habitat of P. pilosa ssp. deamii. b. Detail of inflorescence architecture showing the long, white, eglandular calyx pubescence of this taxon. c. A typical individual of P. pilosa ssp. deamii at anthesis on April 25, 2012……………………………………………...... 131

1.15 Phlox pilosa ssp. deamii PZ11-026 in cultivation at the OPGC. a. Detail of the dense, mounding habit of this taxon in cultivation. b. Close-up of flowers. c. Common garden study of P. pilosa complex accessions. The P. pilosa ssp. deamii are the plants in the forefront………………………...... 132

1.16 Phlox amoena x Phlox pilosa ssp. pilosa hybrid swarm (P. pilosa ssp. deamii) PZ12-054. a. The mowed, sloping roadside habitat. Note that P. amoena, P. pilosa ssp. pilosa, and hyrbids are growing side by side. b. Habit and flower color variation of hybrid individuals. The degree of variation suggested hybridization and varying degrees of introgression. c. A unique flower color variant found among hybrids. Flowers are salmon-pink……… 133

1.17 Phlox pilosa ssp. fulgida PZ12-092 in situ in Story County, Iowa. a. A completey white-flowered individual. b. A typical pink-flowered individual. c. A white-flowered individual with prominent purple striae. Photos and Germplasm collection (PZ12-092) courtesy of Jeffrey Carstens……………………………………………………………………… 134

1.18 Phlox pilosa ssp. longipilosa in situ and in cultivation at the OPGC. a. A close-up of the inflorescence showing the aglandular, comparatively long hairs that are longer than any other member of the P. pilosa complex. b. The natural habitat in the Quartz Mountains. c. Detail of the corolla, note the rounded, overlapping , and distinctive, vibrant coloration. d. The mounding habit of P. pilosa ssp. longipilosa, in early anthesis, at the OPGC………………………………………………………………………... 135

xviii

1.19 Phlox pilosa ssp. ozarkana in situ and in cultivation at the OPGC. a. The heavily forested roadside habitat of P. pilosa ssp. ozarkana in Iron County, Missouri. b. A typical, flowering individual of P. pilosa ssp. ozarkana in roadside habitat on April 27, 2012. c. Variation among individuals of P. pilosa ssp. ozarkana PZ10-227 collected in Johnson County, Arkansas, and cultivated at the OPGC. d. A selected white- flowering individual of P. pilosa ssp. ozarkana PZ10-227………………….. 136

1.20 Phlox pulcherrima PZSH2011-033 in situ and in cultivation. a. The mowed roadside habitat of P. pulcherrima in Shelby County, Texas. b. The flowers of P. pulcherrima. c. The habit of P. pulcherrima when grown in container. The plants are taller than typical Phlox pilosa ssp. pilosa, and differ in flower color and floral fragrance…………………………………………….. 137

1.21 Phlox villosissima in situ and in cultivation. a. The mowed, roadside habitat of P. villosissima PZSH2011-040 in Kerr County, Texas. b. A typical, flowering individual of P. villosissima PZSH2011-037 in Kerr County, Texas on April 13, 2011. c. Phlox villosissima PZSH2011-036 in cultivation at Natives of Texas Nursery in Kerr County, Texas…………… 138

1.22 Phlox amplifolia PZ11-050 in situ in Cocke County, Tennesssee. a. The steep, talus slope habitat along the French Broad River. b, A flowering individual on October 8, 2010. c. An infructescence with mature seed capsules and evidence of dehisced capsules (empty, star-shaped calyces)... 139

1.23 Phlox paniculata in situ. a. P. paniculata PZ10-209 from Scioto County, Ohio along a roadside bordering Rocky Fork Creek flowering on August 1, 2010. b. A typical flowering individual of P. paniculata PZ11-040 flowering along the Cheat River in Preston County, West Virginia on September 14, 2011. c. The roadside population of P. paniculata PZ11-043 in County, West Virginia. The population had more flower color variation than any other population discovered…………………………...... 140

1.24 Populations referred to as Phlox carolina ssp. alta in situ. a. The high elevation (ca. 1500 m) habitat of P. carolina ssp. alta PZ11-045 in Haywood County, North Carolina. b. An individual of P. carolina ssp. alta PZ11-045 flowering on September 16, 2011. This specimen does not accurately portray a typical flowering from this population, but demonstrates the potentially long flowering season of this taxon. Plants were seen flowering in this population in mid-July, 2011………………...... 141

xix

1.25 Phlox glaberrima ssp. interior PZ12-063 in situ. a. In Lewis County, Tennessee, P. glaberrima ssp. interior grows in a limestone seep habitat among various species of sedges (Carex spp.). b. An individual of P. glaberrima ssp. interior. Plants were photographed April 27, 2012. There were no flowering plants in this population on this date. c. The paniculate inflorescence of P. glaberrima ssp. interior PZ12-063 flowering in cultivation at the OPGC……………………………………………………. 142

1.26 Phlox glaberrima ssp. triflora PZ12-046 in situ in Lyon County, Kentucky where it grew in a seepage area in mixed hardwood forest………………... 143

1.27 Phlox maculata PZ10-208 in Adams County, Ohio. a. The roadside habitat of P. maculata. In this area, this taxon occurs in wet ditches and seepages in disturbed sites. b. Flowering individuals of P. maculata on September 1, 2010. Plants can begin to flower in June in this region...... 144

1.28 Phlox ovata in situ and in cultivation. a. The population of Phlox ovata PZ12-077 in Allegheny County, Virginia flowering on May 16, 2012. Note the white-flowered variant in the upper left corner of the photo growing among mostly typical pink-forms of the species. b. A close-up of the white flowered variant of P. ovata PZ12-077. This form has been propagated and given the name ‘White Mountainside’. c. Phlox ovata PZ11-014 flowering in cultivation at the OPGC. d. The foliage of P. ovata PZ12-076. This taxon has the largest leaves of all taxa in subsection Phlox……………. 145

1.29 Phlox pulchra in cultivation at the OPGC. This collection was made and distributed by Plant Delights Nursery, Raleigh, North Carolina…………...... 146

1.30 Phlox stolonifera in situ. a. A pink-flowering form of P. stolonifera PZ10- 095 from Hocking County, Ohio. b. A lavender-purple flowering form of P. stolonifera PZ10-125 growing along Panther Creek, in Habersham County, Georgia. c. The distinctive foliage and habit of P. stolonifera distinguishes it from all other species………………………...... 147

1.31 Phlox bifida ssp. bifida PZ10-149. a. The oak savannah habitat of P. bifida ssp. bifida in Newton County, Indiana. b. A typical flowering individual of P. bifida ssp. bifida May 19. 2010. c. Seedlings of P. bifida ssp. bifida PZ10-149 flowering at the OPGC on May 1, 2014. d. Detail of flower color variation among seedlings of P. bifida ssp. bifida PZ10-149………………... 148

xx

1.32 Phlox bifida ssp. stelleria in situ and in cultivation. a. The habitat of P. bifida ssp. stelleria PZ10-201 at the type location in Jessamine County, Kentucky. b. In this habitat, plants grow as chasmophytes in rock crevices of dolomite cliffs. c. The cedar glade habitat of P. bifida ssp. stelleria PZ10-018 in Wilson County, TN. d. A typical individual of P. bifida ssp. stelleria PZ10-018 flowering on April 7, 2010……………………………… 149

1.33 Phlox nivalis in situ and in cultivation. a. The habitat of P. nivalis PZSH2011- along a roadside in Bay County, Florida. Note the flower color variation among individuals. b. An exceptional color flower variant of P. nivalis. c. Phlox nivalis ‘Eco Flirty Eyes’, a selection made by Dr. Don Jacobs in Franklin County, Georgia. d. A plant received as P. nivalis ‘Camla’ that matches the description of Phlox xhenryae described by Wherry (1935) ………………………………………………………...... 150

1.34 Phlox subulata was frequently planted in cemeteries and around former homesteads. In these situations, it is likely that a variety of available cultivars and/or, local selections were planted. Cross- between different forms results in unique populations that can be variable in color and habit and have potential for selection and collection of unique germplasm. Such populations can become adventive and are responsible for the often exaggerated native range of this species…………………………... 151

1.35 Phlox subulata ssp. subulata PZ12-066 in situ and in cultivation at the OPGC. a. The serpentine barren habitat of P. subulata ssp. subulata PZ12- 066 in Chester County, Pennsylvania. b. A typical flowering individual of P. subulata ssp. subulata PZ12-066 flowering in situ on May 15, 2012. Most flowers had passed by this time, and many plants had mature seed. c. The shale cliff habitat of P. subulata ssp. subulata PZ12-091 on the Delaware/Franklin County, Ohio border. The plants colonize the xeric oak woods on the top of the cliff and cliff edge. d. A typical flowering individual P. subulata ssp. subulata PZ12-091 at the OPGC…...... 152

1.36 The habitat and flowering characteristics of Phlox subulata ssp. brittonii. a. The habitat of P. subulata ssp. brittonii PZ10-074 is often vertical, shale cliffs b. A flowering individual of P. subulata ssp. brittonii PZ10- 074 in Botetourt County, Virginia. The flower fade from pale lavender to lavender-white. c. Various accessions of P. subulata ssp. brittonii in cultivation at OPGC showing the lavender-white flower color of this taxon. The bright pink plant in the right is P. subulata ‘McDaniel’s Cushion, and on the left is P. bifida ssp. stellaria ‘Glade Blue’. d. Differences in flower shape among different individuals of P. subulata ssp. brittonii PZ12-074….. 153

xxi

2.1 Figure 2.1 a-l. Leaf, pollen, and style length variation of Phlox taxa. a. P. paniculata PZ10-209. Note the areolate leaf veins. b. P. carolina ssp. alta PZ10-045. Note obscure leaf veins. c. P. glaberrima ssp. triflora PZ12- diploid (n=7). d. P. glaberrima ssp. triflora PZ11-019 tetraploid (n=14). e. P. maculata PZ12-107. f-i. Differences in pollen color between subsection Paniculatae and Phlox. f. The yellow pollen of P. glaberrima ssp. triflora. g. The yellow pollen of P. maculata PZ12-107. h. The cream colored pollen of P. paniculata PZ10-209. i. The cream colored pollen of P. amplifolia PZ11-050. j-k. Differences in style length among subsections. j. Subsection Paniculatae and Phlox. k. Subsection Subulatae. l. Subsection Divaricatae. The scale bar is 1 cm...... 188

2.2 Concurrent comparison of four Phlox species from three sections and four different subsections of the genus; a. diploid P. subulata, b. diploid P. pilosa ssp. pilosa, c. diploid P. paniculata, and d. tetraploid P. buckleyi (from left to right). G1 peak placement and values support data from analyses with an internal standard…………………………………………… 189

2.3 Determination of DNA content and ploidy level of diploid (2n) and aneuploid (2n-2) cultivars of Phlox paniculata and P. amplifolia using flow cytometry with Pisum sativum ‘Ctirad’ as the internal standard. a. Histogram and meiotic metaphase chromosome counts for the diploid (n=7) P. paniculata ‘Delta Snow’ (14.90 pg), b. Histogram and mitotic metaphase chromosome counts (400X) for the diploid (2n=2x=14) P. amplifolia PZ11- 050 (15.21 pg), and c. Histogram and meiotic metaphase chromosome count of the aneuploid (n=12) P. paniculata ‘John Fanick’ (22.57 pg) are shown…………………………...... 190

3.1 a. Concurrent comparison of P. subulata ssp. subulata showing the serial increase in genome size with each increase in ploidy, A. diploid (2n=2x=14) PZ12-065 with a mean genome size of 7.78 pg; B. tetraploid (2n=4x=28) PZ12-066 with a mean genome size of 15.65, and; C. hexaploid (2n=6x=42) PZ12-067 with a mean genome size of 23.72 pg. b. Mitotic metaphase chromosome count of tetraploid P. subulata ssp. subulata PZ12-066 related to peak “B” in Figure 1a……………………………………………………... 220

4.1 Location and county level geographic distribution of accessions and cytotypes from the Phlox pilosa complex. The outline corresponds to the reported geographic distribution of the P. pilosa complex and all related taxa. All recognized subspecies and related species are included in the flow cytometry analysis (Table 4.1) …………………………………...... 254

xxii

4.2 Flow cytometry histograms and corresponding chromosome counts of diploid and tetraploid members of the Phlox pilosa complex. A. Flow cytometry histogram for diploid P. pilosa ssp. longipilosa PZSH2011-043. B. Meiotic metaphase chromosome count of P. pilosa ssp. longipilosa PZSH2011-043 with n=7 chromosomes. c. Flow cytometry histogram for tetraploid P. pilosa ssp. pilosa PZ12-124 collected in Polk County, NC. d. Mitotic metaphase chromosome counts of P. pilosa ssp. pilosa PZ12-124 with 2n=4x=28 chromosomes……………………………………………… 255

4.3 Population clustering analysis suggests significant structure among members of the Phlox pilosa complex. ΔK scores for each value of K genetic clusters following Evanno et al. (2005). The figure was generated using Structure Harvester (Earl et al., 2011) ……………………………..... 256

4.4 Barplots depicting STRUCTURE results of K=4 clusters of 61 samples from 8 populations of the Phlox pilosa complex. Populations are grouped by cluster, in order of highest mean Q score (cluster identity proportion). Ploidy differences are expressed on the top of the barplot...... 257

5.1 Histograms showing relative genome of diploid (n=7) parental taxa and diploid (n=7) F1 hybrids from 4 interspecific crosses; a. P. divaricata ssp. laphamii PZSH11-022 (11.03 pg), P. divaricata ssp. laphamii x P. paniculata F1 hybrid PZ11-199 (12.31 pg), and P. paniculata PZ12-106 (15.01 pg); b. P. ovata PZ10-167 (12.16 pg), P. amplifolia x P. ovata F1 hybrid PZ11-176 (14.24 pg), and P. amplifolia PZ11-050 (15.21 pg); c. P. amplifolia PZ11-050 (15.21 pg), P. amplifolia x P. paniculata ‘Delta Snow’ F1 hybrid PZ11-176 (15.34 pg), and P. paniculata ‘Delta Snow’ PZ10-027 (14.90 pg); d. P. paniculata PZ10-231 (14.14 pg), P. ‘Minnie Pearl’ x P. paniculata F1 hybrid PZ11-061 (14.62 pg), and P. ‘Minnie Pearl’ PZ11-012 (14.61 pg) ………………………………………………... 292

5.2 Comparison of parental taxa and their F1 hybrid. a. P. paniculata PZ10- 231(♀). b. P. paniculata x P. ‘Minnie Pearl’ F1 hybrid, and. c. P.’Minnie Pearl’ PZ11-012 (♂). Note the intermediacy of the leaf veins and petiole. The scale bar is 1 cm.……………………………………...……………….. 293

5.3 Comparison parental taxa and their F1 hybrid. a. P. amplifolia PZ11-050 (♀). b. P. amplifolia x P. ovata F1 hybrid, and. c. P. ovata PZ10-167 (♂). Note the intermediacy of the leaf veins and petiole. The scale bar is 1 cm…………………………………………………………………………... 294

xxiii

5.4 Comparison parental taxa and their F1 hybrid in the remake of P. xarendsii. a. P. divaricata ssp. laphamii PZSH2011-022 (♂). b. P. divaricata x P. paniculata F1 hybrid, and. c. P. paniculata PZ12-106 (♀). Note the intermediacy of the leaf veins and petiole. The scale bar is 1 cm... 295

5.5 Summary of the interspecific compatibilities between the species in subsection Paniculatae and subsections Phlox and Divaricatae. Areas blacked out indicate crosses that were not attempted. ✓ = successful combination; ✕= failed combination; PD = partial development of fruit….. 296

6.1 a. Flowers of P. carolina ssp. carolina PZ11-036. b. Flowers of P. ‘Minnie Pearl’ PZ11-012. c,d,e. Different flower phenotypes of F1 hybrids produced from the cross Phlox carolina ssp. carolina PZ11-036 x Phlox ‘Minnie Pearl’ PZ11-012………………...... ………...………...... 321

6.2 SRAP banding showing presence of shared alleles between F1 hybrid of P. paniculata PZ10-231 x P. ‘Minnie Pearl’ PZ11-012 and parental taxa. The primer pair me3-em3 was used for amplification of DNA fragments… 322

7.1 Flow cytometry histogram showing concurrent comparison of a. diploid (n=7) P. pilosa ssp. longipilosa (12.55 pg) b. alloaneuploid (n=13) P. floridana x P. pilosa ssp. longipilosa F1 hybrid (17.28 pg), and the c. tetraploid (n=14) P. floridana (21.85 pg) ………………………….……... 348

7.2 Histograms showing relative genome sizes of parental taxa and F1 hybrids from 4 interploid, interspecific crosses. a. Diploid (n=7) P. pilosa ssp. longipilosa PZSH11-043 (12.55 pg), alloaneuploid (n=13) P. floridana x P. pilosa ssp. longipilosa F1 hybrid PZ11-107 (17.28 pg), and tetraploid (n=14) P. floridana PZSH11-010 (21.85 pg). b. diploid (n=7) P. drummondii PZ10-161 (12.22 pg), alloaneuploid (n=13) P. drummondii x P. pulcherrima F1 hybrid PZ11-175 (17.78 pg), and tetraploid (n=14) P. pulcherrima PZSH11-034 (22.05 pg). c. diploid (n=7) P. divaricata PZSH11-022 (11.03 pg), alloaneuploid (n=13) P. divaricata x P. villosissima F1 hybrid PZ11-202 (17.79 pg), and tetraploid (n=14) P. villosissima PZSH11-042 (26.58 pg). d. diploid (n=7) P. xglutinosa PZ11-127 (10.90 pg), alloaneuploid (n=13) P. xglutinosa x P. villosissima F1 hybrid PZ11-212 (19.47 pg), and tetraploid (n=14) P. villosissima PZSH11-042 (26.58 pg). Error bars indicate the standard deviation of flow cytometry samples…………...... 349

7.3 Meiotic metaphase chromosomes of aneuploid F1 progeny generated from interploid interspecific crosses. a. P. floridana x P. pilosa ssp. longipilosa F1 hybrid with n=2x-1=13 chromosomes. b. P. drummondii x P. pulcherrima F1 hybrid with n=2x-2=12 chromosomes. The scale bar (lower portion of image) indicates 10 µm…………...……………………... 350

xxiv

7.4 SRAP banding showing presence of shared alleles between F1 hybrid of P. floridana x P. pilosa ssp. longipilosa PZ11-107 and parental taxa. The primer pair me3-em3 was used for amplification of DNA fragments…………………………………………………………………… 351

7.5 Electropherogram demonstrating SRAP fragment analysis and banding patterns of the F1 hybrid P. floridana x P. pilosa ssp. longipilosa PZ11- 107 and parental taxa on an ABI 3100 genetic analyzer. The primer pair me3-em3 was used for amplification of DNA fragments, and the green circles correspond to shared alleles between the hybrid and parental taxa (Figure 7.4) ………………………………………………………...... 352

xxv

LIST OF TABLES

Table Page

1.1 Taxonomic arrangement of eastern U.S.A. Phlox species showing the sections, subsections, species, and subspecies recognized……………….. 154

1.2 The number of eastern Phlox taxa listed by state in descending order….... 155

1.3 Relative flowering period and time to seed maturity of eastern Phlox taxa. This information derived from field observation of listed populations from 2010-2013……………………………………………...... 156

1.4 Taxonomic arrangement, and numbers of wild collected and cultivar germplasm accessions made of eastern U.S.A. Phlox species from 2010- 2013……………………………………………………………………...... 157

1.5 Botanical and common names of taxa in subsection Cluteanae and number of germplasm accessions obtained from natural plant populations and nursery sources during the development of a Phlox germplasm collection at the OPGC………………………………………………….… 158

1.6 Botanical and common names of taxa in subsection Divaricatae and number of germplasm accessions obtained from natural plant populations and nursery sources during the development of a Phlox germplasm collection at the OPGC……………………………………………….…… 159

1.7 Botanical and common names of taxa in subsection Paniculatae and number of germplasm accessions obtained from natural plant populations and nursery sources during the development of a Phlox germplasm collection at the OPGC………………………………………………….… 160

1.8 Botanical and common names of taxa in subsection Phlox and number of germplasm accessions obtained from natural plant populations and nursery sources during the development of a Phlox germplasm collection at the OPGC…………………………………………………………...... 161

xxvi

1.9 Botanical and common names of taxa in subsection Stoloniferae and number of germplasm accessions obtained from natural plant populations and nursery sources during the development of a Phlox germplasm collection at the OPGC…………………………………………………… 162

1.10 Botanical and common names of taxa in subsection Subulatae and number of germplasm accessions obtained from natural plant populations and nursery sources during the development of a Phlox germplasm collection at the OPGC………………………………………………………………. 163

2.1 Distinguishing vegetative and floral morphological characteristics and habitat preferences of taxa from Phlox subsections Paniculatae and Phlox as described by Wherry (1955) that were used in this study………………. 191

2.2 Summary of means and ranges of 2C Holoploid genome size (pg) of Phlox species grouped by section, subsection, and ploidy level…………... 192

2.3 Relative Holoploid (2C) genome sizes and ploidy levels, determined using flow cytometry and chromosome counts, for a diverse collection of cultivars and hybrids of Phlox……………………………………………... 193

2.4 Relative Holoploid (2C) genome sizes and ploidy levels, determined using flow cytometry and chromosome counts, for a diverse collection of Phlox germplasm accessions from natural plant populations……………...... 196

3.1 Summary of means and ranges of 2C, holoploid genome size (pg) of Phlox taxa grouped by section, subsection, and ploidy level……………… 221

3.2 Collection site, genome size, and ploidy level for a collection of 4 eastern and 7 western Phlox species from subsection Albomarginatae, Canascentes, Longifoliae, Stoloniferae, and Subulatae, and one species from the sister genus Microsteris………………………………………….. 222

3.3 Genome sizes, ploidy levels, and nursery sources of creeping Phlox cultivars and hybrids from subsections Stoloniferae and Subulatae………. 224

4.1 Taxonomic overview of perennial taxa in subsection Divaricatae within or related to the P. pilosa complex. The classification of Wherry (1955) is on the left, and on the right a modern classification system derived from several sources (Ferguson et al., 1999; Levin 1963; Levin, 1966; Locklear, 2011a; Prather, 1994) ……………………………………………………... 258

xxvii

4.2 Relative Holoploid (2C) genome sizes and ploidy levels of a diverse collection of taxa from Phlox subsection Divaricatae collected from natural plant populations (Figure 7.2) …………………………………….. 259

4.3 Relative Holoploid (2C) genome sizes and ploidy levels of cultivars selected from Phlox subsection Divaricatae……………………………..... 262

4.4 Phlox pilosa complex sampling localities, number of individuals per sampling site, and ploidy. These data are a subset of the taxa used in the genome size analysis………………………………………………………. 263

4.5 Characterization of the six SSR markers based on eight P. pilosa complex populations…………………………………………………………………. 264

4.6 Descriptive statistics and genetic diversity measures for eight populations of taxa in the P. pilosa complex subject to SSR analysis. Genetic diversity measures were calculated from a binary data set where microsatellite alleles were coded as present or absent…………………….. 265

4.7 Hierarchical analysis of molecular variation (AMOVA) among populations from the Phlox pilosa complex illustrating the proportion of variation attributable to differences among populations and within populations………………………………………………………………..... 266

4.8 Genetic differentiation and genetic diversity among eight populations of taxa from the P. pilosa complex…………………………………………… 267

5.1 Putative hybrids involving Phlox paniculata with the corresponding reported parentage, ploidy, and style length of the parents.……………….. 297

5.2 Phlox taxa used in the interspecific hybridization experiments organized by subsection, specific germplasm accessions, collection sites, relative genome sizes, ploidy levels, and style lengths.……………………….….... 298

5.3 Parental combinations, seeds obtained (No.), and germination (%) of successful and unsuccessful, reciprocal interspecific crosses in which either Phlox amplifolia or P. paniculata served as a parent……………….. 300

5.4 Mean plant height (cm), flowering stems per pot (No.), leaf length (cm), and leaf width (cm) for parental species and interspecific F1 hybrids of the cross Phlox divaricata ssp. laphamii PZSH2011-022 x P. paniculata PZ12-106…………………………………………………………………... 301

xxviii

6.1 Distinguishing morphological characteristics, flower color, and habitat preferences of Phlox subsections Paniculatae and Phlox taxa used in this study………………………………………………………………………. 323

6.2 Specific germplasm collections, internal accession numbers, collection sites, Relative genome sizes, ploidy levels, and style lengths of Phlox taxa used in interspecific hybridization experiments…………………………… 324

6.3 Parental combinations, seeds obtained (No.), and germination (%) of successful and unsuccessful, reciprocal interspecific crosses in which P. ‘Minnie Pearl’ serves as a parent…………………………………………... 325

6.4 Mean plants height (cm), flowering stems per pot (No.), leaf length (cm), and leaf width (cm) for parental species and interspecific F1 hybrids among Phlox species………….…………………………...... 326

7.1 Distinguishing morphological characteristics, habitat preferences, and geographic distribution of Phlox subsection Divaricatae taxa used in this study………………………………………………………………………... 353

7.2 Specific germplasm collections, internal accession numbers, collection sites, relative genome sizes, and ploidy levels of Phlox taxa used in interploid interspecific hybridization experiments………………………… 354

7.3 Parental combinations, cross ploidy, seeds obtained (No.), and germination (%) and internal accession numbers of successful and unsuccessful interploid interspecific crosses………………………………. 355

xxix

CHAPTER 1

DEVELOPMENT OF A COMPREHENSIVE PHLOX GERMPLASM COLLECTION

Introduction

Phlox is a diverse genus of U.S. native plants that are popular floriculture and

nursery crops. Cultivars and hybrids have been primarily derived from 3 species, P.

drummondii, P. paniculata, and P. subulata; however, up to 20 additional, related species

have been described, but many have not been introduced to horticulture and others are

only rarely utilized. These related species are a potentially important germplasm

resource of increased interest to the floriculture industry for the continued development

of Phlox as an ornamental crop. Availability of a comprehensive germplasm collection

focused on these related species could help define crop-breeding pools, result in selection

of novel cultivars, introduce new genetic variation into breeding programs, and be

important sources of disease and insect resistance (Harlan and de Wet, 1971; Tay, 2003).

Among U.S. native genera, focus on crop related germplasm has revolutionized breeding

and development of important genera such as Echinacea, Heuchera, Tiarella, and

Baptisia (Hawkins et al., 2013; Otteson, 2003; Schoellhorn and Richardson, 2004). This successful breeding has resulted in greater commercial interest in these genera and the formation of public and private germplasm collections. The development of germplasm collections has not only facilitated commercial development of these genera, but also

1 enhanced the study of ecological, phylogenetic, and medicinal attributes of these plants; thus, a well-documented germplasm collection can have far reaching effects (Tay, 2006;

Widrlechner and McKeown, 2002). The Convention on Biological Diversity (CBD) also highlights the need and opportunity for germplasm collections of U.S. native genera.

Whereas plant germplasm was once considered the property of all humankind regardless of political boundaries, ratification of the CBD has resulted in the view that plant germplasm is an individual country’s natural resource, and extensive documentation of germplasm collection and profit-sharing is now required for germplasm collections in such countries (Tay, 2005). This new regulatory environment has restricted or even halted germplasm acquisition from countries that are centers of origin for important ornamental crops, such as Brazil, Ecuador, Peru, and South Africa, where germplasm of traditionally important floriculture crops such as Petunia, Impatiens, and Begonia, is found (Heywood, 2003; Tay, 2005). These and other factors have resulted in renewed interest in floriculture and nursery crops that are native to the United States, where it is possible to collect and develop a comprehensive germplasm collection without the restrictions of the CBD. Among native US genera, Phlox is often a prominent component of different native ecosystems throughout the United States, and some species have been popular ornamental plants in temperate climates throughout the world. Although a relatively minor crop, the Census of Horticultural Specialties (USDA-NASS, 2010) reported that 2009 sales of Phlox totaled more than $16.2 million in the United States.

The ornamental use of Phlox and its extensive diversity led to its designation in 2006 as a priority genus for germplasm development and conservation at the Ornamental Plant

2

Germplasm Center (OPGC) of the United States Department of Agriculture

(USDA)/Agricultural Research Service (ARS).

The genus Phlox was named by Linnaeus in reference to the abundant, vibrantly

colored floral displays, that lead him to describe them in Latin as “floris flammeo igneoque colore” – flowers the color of a glowing flame (Locklear, 2011a). Phlox is a

North American genus of ca. 65 species with two primary centers of distribution; ca. 45

species occur in the west, and 20-23 species are native to the eastern U.S. (Figure 1.1,

Table 1.1). The eastern species P. drummondii, P. paniculata, and P. subulata are some

of the most widely recognized and cultivated flowering plants in temperate regions of the

world; western species are highly ornamental and of botanical interest, but less tractable

when cultivated in the eastern United States (Table 1.1) (Bendtsen, 2009; Levin, 1975;

Locklear; 2011a; Symons-Jeune, 1953; Wherry, 1935; Wherry, 1955). The first

commercial Phlox cultivar was released in 1824. Since then, intensive breeding and

selection of eastern taxa has resulted in the introduction of hundreds of cultivars,

primarily of P. paniculata, but the scope of breeding efforts have been relatively limited

(Symons-Jeune, 1953). Breeding programs dedicated to P. paniculata still exist in the

United States and Europe, and there is renewed interest in the use of wild germplasm to

meet diverse industry desires for new or improved uses of Phlox as bedding/garden

plants, potted herbaceous perennials, potted flowering plants, cut flowers, components of

native plant gardens, and in ecological restoration projects. Taxa such as P. amplifolia,

P. floridana, P. villosissima, and P. pilosa ssp. deamii, and others, are related to commercially important species, but only rarely, if ever, have been introduced into

3 cultivation, but have potential for breeding and selection (Deam, 1941; Locklear, 2011a;

Wherry, 1955). These taxa tend to be geographically remote endemics or relicts with restricted natural distributions in places of low populations density, have only rarely been collected by botanists, and may be rare in the wild (Wherry, 1955). Furthermore, nearly all species exhibit extraordinary phenotypic and genetic diversity among populations that has resulted in a confusing taxonomic history for some taxa. Bearing these factors in mind, a well-documented germplasm collection is warranted as it could be used to broaden the development of Phlox as commercial crop, to initiate interspecific hybridization and characterization experiments, contribute to phylogenetic and taxonomic resolution of the genus, and help conserve and disseminate native plant genetic resources.

Methodology

In 2010, the National Plant Germplasm System’s (NPGS) Genetic Resources

Information Network (GRIN) listed only three accessions of Phlox, and at that time the

OPGC began efforts to assemble a comprehensive germplasm collection both from natural plant populations of a variety of Phlox taxa (i.e., crop wild relatives), and from commercially available selections of known provenance as well as cultivars and hybrids.

Collection efforts of crop wild relatives was emphasized when the USDA/ARS sponsored a series of multi-state collection expeditions that were conducted from 2010-2013. Initial exploration efforts were focused on the 8 species of Phlox in Ohio, and then expanded to the 23 eastern species through a series 12 germplasm collection expeditions (Table 1.1,

Table 1.2).

4

Exploration of natural populations of Phlox in Ohio permitted observation and study of variation in phenology and seed development and maturation in 8 different taxa on a frequent basis (Table 1.2). In combination with study of specimens from herbaria throughout the eastern U.S., this critical information allowed me to assess and predict flowering and seed availability periods for related species in other parts of the country.

This information allowed me to organize germplasm collection expeditions performed during spring, summer, and autumn to permit observation of phenological variation, floral phenotypic diversity, and timing of seed maturation in taxa of interest. Specific locations of Phlox populations were also established from contact with regional field botanists, botanical gardens and arboreta, historical accounts of species from the literature, and exploratory fieldwork. Collection permits were obtained where necessary from a variety of federal and state governmental and private institutions. Permits are on file at the OPGC and with the USDA-ARS National Germplasm Resources office in

Beltsville, Maryland.

Interstate collection expeditions were based on similar information as above.

Species counts of Phlox taxa for each state indicated that there are three sub-centers of species diversity within the entire range of eastern species (Figure 1.1) (Locklear, 2011a;

Wherry, 1955). There is an eastern group centered on the central and southern Valley and Ridge province and Appalachian Highlands province; a central group centered on the

Interior Low Plateaus, southern Allegheny Highlands and Coastal Plain; and a western group consisting of taxa in the Interior Highlands, southern Great Plains Region, western

Coastal Plain, and Edward’s Plateau. (Figures 1.1, 1.2; Table 1.1). Each of these regions

5 contains at least one endemic species not found in the other region; for example, P. buckleyi is endemic to the central Valley and Ridge province of Virginia and West

Virginia, but is not found elsewhere (Wherry, 1930). A fourth group, not represented on the map, would be the eastern Coastal Plain of the Carolinas, Georgia, Northern Florida, and southern Virginia. Focused collection efforts within these regions during different seasons permitted collection of the maximum number of eastern species, and fulfilled collection goals related to specific taxa.

Within these regions, collection goals included attempts to obtain all 23 recognized eastern Phlox taxa; to focus on species that are not known to be in cultivation; the collection of unique floral variants for immediate use in breeding experiments; and to establish preliminary guidelines for timing of seed collections. Specific objectives for collection within each subsection and taxon are described below.

Locating Phlox populations

Most eastern Phlox species are members of successional ecological communities, and grow in habitats that experience regular disturbance (Amason, 2000; Locklear,

2011a; Wherry, 1955). Different eastern taxa respond to various types of disturbance based on geographical distribution and ecosystem preference; disturbance types include fire, mammalian tunneling, steep slopes with a surface layer of shifting soil or stone, flooding, and tree fall (Heikens, 2003; Hendrix and Kyhl, 2000; Locklear, 2011a;

Kritzman, 1974). Response to disturbance has resulted in many Phlox populations that are established in areas of anthropomorphic disturbance, and the most robust populations

6

were often encountered along infrequently mowed right-of-ways and similarly disturbed

habitats (P. Zale, personal observation, April 13, 2011). The infrequent mowing reduces

competition from sympatric species and often results in greater numbers of individual

Phlox plants. The altered environmental conditions, such as increased solar radiation,

can result in a greater abundance of flowering plants, a greater number of flowers per

plant, and potentially increased seed production (Forman and Alexander, 1998).

Phlox Seed Development, Dispersal, and Collection

Seed collection is the primary method of germplasm acquisition, and affords the

maximum capture of genetic diversity, and greatest potential for long-term storage and

maintenance of genetic diversity (Hay and Probert, 2013). However, seed development

and dispersal patterns in Phlox present challenges that are unique among other priority

genera at the OPGC. Phlox species are out-crossing, highly heterozygous taxa pollinated

by a variety of Lepidoptera (Clay and Levin, 1989; Grant and Grant, 1965). Most taxa

have a gynoecium (capsule) of three united carpels with one ovule each, therefore, a

single Phlox flower can produce up to three seeds per pollination event, although up to 5

ovules per carpel are reported for P. roemeriana (Wherry, 1955), and similar seed set per ovule was obtained from controlled pollinations of one accession of P. nivalis at the

OPGC. The number of seeds per fruit is not well characterized in most species, and may be more variable that originally indicated. Seed production under natural conditions is typically reduced, and natural populations of P. pilosa were shown to have an average 1.2 seeds per flower (Grant and Grant, 1965; Levin and Kerster, 1967). Seed set and

7

fecundity are linked to population size and the number and abundance of pollinators

available (Hendrix and Kyhl, 2000). The small number of seeds produced per Phlox

flower may reduce the ability to collect sufficiently large seed samples from wild Phlox

species, but other intrinsic factors may also influence seed production.

Explosive dehiscence, or ballistic seed projection, is a characteristic of all eastern

Phlox species (Levin and Kerster, 1967; Levin and Kerster, 1968). Observation of

natural populations of eastern taxa indicated that Phlox seeds dehisce soon after reaching

physiological maturity (Baskin and Baskin, 2005). This process results in a narrow

window for seed collection of several taxa and complicates the timing of germplasm

collection efforts (Stamp and Lucas, 1983; Wherry, 1955). Although data is limited, P.

drummondii was shown to have short-range ballistic seed dispersal, capable of projecting

seeds an average of 77.5 cm, but distances of 5 m were possible; similar seed projection

distances were reported for P. pilosa (Levin and Kerster, 1968). Species with short-

distance dispersal typically experience secondary dispersal (Stamp and Lucas, 1983).

Since Phlox seeds have been observed to float on water, it has been suggested that

rainwash may be responsible for long-distance seed dispersal, although this hypothesis has not been systematically tested (Stamp and Lucas, 1983). Variation in the distance of seed dispersal has not been tested for most species, and is likely to vary among taxa.

Since the timing of physiological maturity in Phlox seeds is closely linked to the timing of dispersal, collection of Phlox seeds must take place within a narrow window of time that can be difficult to estimate due to a variety of factors. In some cases, it may be necessary to bag developing infructescences to insure seed collection.

8

Differences in seed maturation can be attributed to differences in development

rates among species, phenology, and inflorescence type (Locklear, 2011a; Wherry, 1955).

Preliminary observations of seed development in natural Phlox populations have resulted in the formation of initial guidelines for germplasm collection (Table 1.3). Species in subsection Subulatae have a determinate, few flowered (3-12) cymose inflorescence with early-spring (late March-early May) phenology (Locklear, 2011a; Wherry, 1955). Seeds

of these species mature around 30 days after peak flowering and pollination (Table 1.3).

Observation indicated that seeds of taxa in subsection Subulatae might ripen and disperse within a period of a few days in small, isolated populations. Seed maturation of taxa in subsection Divaricatae is similar to those of subsection Subulatae although flowering and seed set may occur over a longer period; they have a determinate, corymbose

inflorescence with up to 100 flowers, but typically fewer (Wherry, 1955). Taxa in

subsection Divaricatae have an early to mid-spring (Late March-early June) flowering period, and seeds ripen 30-45 days after peak flowering (Table 1.3) (Levin, 1966).

Members of subsection Phlox can be divided into two groups based on inflorescence type and phenology. Phlox glaberrima ssp. triflora, P. ovata, and P. pulchra have determinate inflorescences with up to 50 flowers (typically 15-30) and mid-late spring

(May-June) phenology (Wherry, 1955). Seed maturation in these species occurs 40-60 days after flowering (Table 1.3). Although there appears to be a range of time for seed collection in this group, particular attention must be given to the peak flowering season of the population in question. Seeds ripen and disperse within a limited window of time, and field observation suggests that this window may be as small as one week.

9

The other taxa in subsection Phlox (P. carolina, P. glaberrima ssp. glaberrima

and ssp. interior, and P. maculata) as well as those in subsection Paniculatae differ from previously mentioned eastern taxa and have a determinate, compound-paniculate inflorescence capable of producing up to hundreds of flowers over a long period (Wherry,

1955). These species flower in mid-late summer (July-September), and in favorable climatic years, flowering will continue until onset of frost in autumn. Since these species flower over a long period of time, there can be considerable variation in the stage of seed maturity within a given infructescence. Field observations indicate that these species have seeds that ripen approximately 60 days after the beginning of anthesis (Table 1.3).

Generally plants that begin flowering in July will bear mature seeds in September, but the extended flowering period results in a broader window of time for seed collection, and mature seeds can be found well after the initial seeds mature and disperse (Table 1.3).

For example, seed collections of P. paniculata (PZ10-209, PZ10-231) from Erie and

Scioto counties in Ohio were made twice during the autumn of 2010, once in September and again from the same populations in November. This highlights the range of time from which seed can collected from these taxa, however, the duration of flowering may be variable dependent upon seasonal climatic patterns and flowering and seed set may be reduced in unfavorable years. Despite this, seeds of these species may are generally more readily collected than those with spring phenology and fewer flowers per infloresence.

Seed predation appears to be rare in natural Phlox populations, but in one population of P. ovata the developing capsules had been infected by an unidentified, seed

10

scavenging insect species that fed on developing seeds (P. Zale, personal observation, 18

June 2011). This pest noticeably reduced the overall seed production in this small, isolated population. This infestation was only seen one year, and seed production of the same population was not affected the succeeding year (P. Zale, personal observation, 11

July 2012). However, a florivorous beetle has been reported to affect seed set in populations of the federally endangered P. hirsuta, and it is possible that previously unknown Phlox seed predators may exist in other populations of other species (Ruane et al., 2013).

Field observations indicate that timing of seed dispersal appears to be closely linked with the coloration of the developing capsule. During the early stages of development, the capsule is green, but as seed maturity approaches, the capsule begins to change from green to straw color as it dries. The point when the capsule turns completely straw colored, seed is mature and can be harvested (Figure 1.3). The seeds are dark brown or charcoal colored, and ovoid in shape with a rough-textured testa.

Within a given accession of Phlox seeds, there will be a variety of seeds at different levels of maturity. Seeds that remain green in color when dried are not likely to be viable.

Collection and Propagation of Vegetative Germplasm

Phlox plants are easily propagated by asexual means (Bendtsen, 2009; Fuchs,

1994). Collection of vegetative germplasm material allows for propagation of selected

Phlox clones and the preservation of unique phenotypes discovered in the field. Division

11 of rhizomatous or stoloniferous taxa can take place at nearly anytime of year, but is best performed from March until October. Phlox species can be readily propagated by stem cuttings. The timing of cutting collection for efficient rooting is specific to the subsection of the taxon of interest. All Phlox species produce varying amounts of sterile stems after flowering that can be used as material for stem cutting propagation. Two to three node cuttings taken from sterile stems can be rooted in a variety of propagation media. Propagators at the OPGC have found particular success in rooting of Phlox cuttings using Oasis® root cubes made of expanded foam (E. Renze, personal communication, 10 June 2013). Although not necessary, use of rooting hormones may hasten the onset of callus formation and rooting of cuttings (Bendtsen, 2009; Fuchs,

1994). There is also evidence that P. paniculata can be propagated from root cuttings, although this method was not tested during the course of this study (Perry and Adam,

1994). This may also apply to the related P. amplifolia and some taxa in subsection

Phlox due to the similarities in habit. Generally, vegetative propagation of Phlox is a routine and easily accomplished task.

Seed Increase of Phlox Germplasm Accessions

It is not always possible to collect enough seeds from natural plant populations for immediate distribution from germplasm centers. Therefore, many accessions must undergo seed increase in order to generate the minimum seed quantity thresholds necessary to allow seed distribution (Widrlechner and McKeown, 2002). In this situation, pollinators, such as bumble bees, are introduced to accessions of interest that

12

have been caged to exclude contaminating pollinators. This results in genetically diverse

seed samples that are representative of the original wild collection. In 2013, the OPGC

began a project to produce significant quantities of seeds of germplasm aquisitions made

from 2010-2013, both to conserve the populations and to make seeds available to the

floriculture and research communities. At the OPGC, are used to enact

pollination of Coreopsis and Rudbeckia, but the pollination syndrome of these species

differs from Phlox (Widrlechner and McKeown, 2002). Phlox are pollinated only by lepidopterans; a protocol is being developed that involves rearing cabbage white butterflies (Pieris rapae L.) to serve as pollination vectors of cultivated Phlox accessions.

This project is still in the initial phases of development. Attempts at pollination using bumble bees to pollinate Phlox resulted in flowers being robbed of nectar and enactment of pollination without successful fertilization.

Taxonomy and Phylogenetic relationships of eastern Phlox taxa

Phlox is a taxonomically difficult genus. Wherry’s (1955) monograph formed the modern taxonomic basis of the genus Phlox; he noted that this classification was

“admittedly rather artificial and polyphyletic”, and was formulated using primarily morphological and geographic information based on field observation and herbarium specimens. Although morphology still forms the basis of Phlox taxonomy, additional fieldwork, discovery of new species, and molecular phylogenies have helped refine the classification of the genus. Despite this work, there are still some subsections in which a

13

well-resolved taxonomy and phylogeny does not exist (Ferguson et al., 1999; Ferguson and Janssen, 2002; Levin, 1966; Locklear, 2011a)

All Phlox species are herbaceous or suffrutescent perennials or annuals (Wherry,

1955). The majority of western species and 3 eastern species are suffrutescent, pulvinate,

caespitose, or mat-forming, evergreens that grow in xeric habitats (Locklear, 2011;

Wherry, 1955). Herbaceous species are generally rhizomatous and produce adventitious

roots that result in a fibrous root system. All plant parts may be glabrous, glabrescent,

pubescent, or hirsute. The pubescence can be glandular or eglandular, and hairs can vary

in length and number of cells. All species have oppositely arranged leaves initially, but

some taxa, such as the Texas annual species and P. pilosa ssp. longipilosa have

alternately arranged leaves on the upper portion of flowering stems (Locklear, 2011a;

Wherry, 1955). The leaves of eastern taxa are generally subulate, linear, lanceolate or

elliptical in shape, but leaf shape can be highly variable within and between taxa.

Individual plants are comprised of one to many stems that terminate in single to multi-

flowered, determinate inflorescences. Flowers are fragrant. The inflorescence is cymose;

cymes may be arranged into a corymb or panicle in some species, but in other species the

inflorescence is reduced to a single flower (Wherry, 1955). The corolla is salverform,

actinomorphic, and composed of 5 corolla lobes, a corolla tube, and calyx. The corolla is

showy and the color is generally “Phlox-purple”, but corolla color is highly variable and

can range from pink, lavender, purple, red, white, or rarely yellow (Wherry, 1955). The

base of each corolla lobe may be paler in color and bear one or two deeply hued striae.

This feature is variable and gives rise to the characteristic ‘eye’ or star-shaped pattern at

14

the center of the Phlox flower. Flowers are protandrous and the five are fused to

the inside of the corolla tube. Among species, the style can vary in length from 1-25 mm

and is of diagnostic value for distinguishing between ‘short-styled’ and ‘long-styled’

Phlox taxa (Figure 1.4).

Throughout this work, the various Phlox species used in all studies will be referred to as either long-styled or short styled (Figure 1.4). Style length is an important morphological character in the taxonomy of Phlox species that is used to delineate the genus at the sectional level. Short-styled species have a pistil that is 1-5 mm in length and has a tripartite with lobes equal to or longer than the length of the style

(Ferguson et al. 1999; Wherry, 1955). Additionally, the pistil of short-styled species is shorter than the length of and is completely contained within the calyx (Figure

1.4). In short-styled species, stamens are shorter than the corolla tube; therefore the anthers develop within the corolla tube, in close proximity to the stigma (Wherry, 1955).

Long-styled Phlox species have a style that is 8-25 mm in length, and the pistil is several times longer than the lobes of the tripartite stigma. In long-styled Phlox species, the pistil is longer than the sepals, and sometimes, exserted from the corolla tube.

Additionally, long-styled species also have long stamens that are generally equal to the length of the corolla tube, resulting in development of the anthers at the opening of the corolla tube, or exserted from the corolla tube (Wherry, 1955).

Style length can provide initial delineation of taxa, but calyx morphology is typically used to segregate species and intraspecific taxa (Wherry, 1955). The calyx consists of 5 sepals united for up to ¾ of their length and the tissue at their juncture is

15

membranous or scarious. This tissue can be flattened or pleated. The free portions of the

calyx are known as the calyx lobes and are variable in shape, length, and degree of

pubescence among different taxa (Wherry, 1955).

Based on morphological characteristics, the genus Phlox is divided into three

Sections: Annuae, Occidentales, and Phlox (Ferguson et al. 1999; Ferguson and Jansen

2002; Locklear, 2011a; Prather, 1994; Wherry, 1955). The three sections are composed of as many as 16 subsections that have been formulated on the basis of morphology and geographical distribution. The eastern species are placed into 6 subsections; Cluteanae,

Divaricatae, Paniculatae, Phlox, Stoloniferae, and Subulatae (Table 1.1).

Section Annuae is composed of ca. 4 subsections: Divaricatae, Nanae, Speciosae, and Tenuifoliae. The monophyletic subsection Drummondianae, composed only of annual species, is included in a broader concept of subsection Divaricatae (Ferguson et al., 1999; Prather, 1994; Wherry, 1955). Members of section Annuae exhibit similar morphology and all have styles shorter than the sepals, stamens shorter than the corolla tube, an herbaceous habit, and may be either perennials or annuals (Ferguson et al. 1999;

Werry, 1955). These taxa are native to eastern North America, but are found as far west as northeastern Mexico and Central Texas, along the edge of the great plains in the south and as far north as Canadian province of Quebec. These species grow in a variety of ecosystems, but are primarily found in xeric habitats with little competition from sympatric species.

Section Phlox is composed of ca. 6 subsections: Cluteaneae, Longifoliae,

Paniculatae, Phlox, Stoloniferae, and Subulatae. With the exception of subsection

16

Longifoliae, and one species each in subsections Cluteanae (P. cluteana) and Phlox (P.

idahonis) all species are native east of the front range of the Rocky Mountains (Locklear,

2011a; Wherry, 1955). These species all possess a style longer than the sepals (long-

styled), stamens equaling the length of the corolla tube, and an herbaceous habit

(Ferguson et al., 1999; Wherry, 1955). Molecular and karyotypic evidence suggests that

members of subsection Subulatae are more closely related to members of the primarily western section Occidentales (Ferguson et al. 1999; Ferguson and Jansen, 2002; Smith and Levin, 1967). Although no formal taxonomic change has been described, subsection

Subulatae taxa are treated here as members of section Occidentales.

Phlox section Occidentales is comprised of 6 subsections of western species:

Aculeatae, Albomarginatae, Caespitosae, Canascentes, Douglasianae, Multiflorae, and

Sibiricae (Wherry, 1955). These species all possess a short style and caespitose or pulvinate habit, and are restricted to western North America and Eastern Asia. A complete molecular phylogeny has yet to be developed for the majority of western species and the taxonomic status of some species remains in question. Most of these species occur in xeric alpine habitats may have highly specific climate requirements for successful cultivation, although there are exceptions. Many of these taxa have ornamental characteristics and are worthy of evaluation. Only a small subset of western taxa was

available for inclusion in this study.

Further taxonomic insight on the eastern U.S. Phlox was provided by two

molecular phylogenies (Ferguson et al., 1999; Ferguson and Janssen, 2002). Data from

ITS1 and ITS2 nuclear ribosomal regions (internal transcribed spacer) and cpDNA

17 restriction sites were analyzed, but failed to provide a well-resolved phylogeny for the genus Phlox (Ferguson et al. 1999; Ferguson and Janssen, 2002). Even with sampling of all taxa described by Wherry (1955), the various taxa in the P. pilosa complex and Phlox carolina-glaberrima complex formed a paraphyletic group and did not resolve evolutionary relationships of the genus at the level of species or subsection (species complexes are discussed below). The data indicated that hybridization could have resulted in this pattern, but it is likely that these marker systems did provide sufficient numbers of phylogenetically informative sites for delimitation of taxa in these groups

(Ferguson et al. 1999; Ferguson and Janssen, 2002). Despite this, the study did clarify relationships in some groups. Subsection Subulatae was monophyletic and shown to contain the species P. nivalis and P. oklahomensis. These species were previously placed in subsection Speciosae on the morphological basis of having short styles. Furthermore, subsection Subulatae was nested in the same clade as western pulvinate species from section Occidentales, and although sampling of these species was limited, this suggests that eastern mat-forming species may be more closely related to western caespitose species than to other eastern species as suggested by Wherry (1955). The Texas annual species formed a monophyletic group; P. roemeriana clustered with P. drummondii and

P. cuspidata, despite morphological differences that led Wherry (1955) to place them in separate subsections. The molecular phylogeny of Phlox would benefit from genus-wide sampling of more taxa at a finer geographic scale and a different, low copy number nuclear marker that has sufficient variation to distinguish between taxa. A study of genetic variation, genome size, and ploidy variation in the narrowly distributed P.

18

amabilis-P. woodhousei complex revealed a range of ploidy levels and complicated

genetic structure (Fehlberg and Ferguson, 2012). Additionally, it has recently been

demonstrated that ploidy is more widespread and variable than previously reported for

Phlox taxa (Fehlberg and Ferguson, 2012; Worcester et al., 2012). However, ploidy and

genome size variation have yet to be defined for most species, and it remains unknown

how ploidy variation has affected that evolutionary and phylogenetic relationships of

Phlox species.

The following description of Phlox taxa and collections is based primarily on the

eastern taxa described in “The Genus Phlox” (Wherry, 1955). This is the most recent

monograph of the genus, and still remains the primary taxonomic treatment of eastern

and western species. Since that time, phylogeneticists and field botanists have refined the

classification of some groups, and relevant changes are included to reflect the modern

classification of the genus. The analysis resulted in the recognition of 22 species and 28

subspecies (Table 1.1) that are included in this study. The following is a description of

these species, the germplasm collections made, and the range of phenotypic and

geographic variation within them.

Notes on field Identification of Phlox

Field identification of some Phlox taxa is considered difficult (Locklear, 2011a;

Wherry, 1933; Wherry, 1955). Even with precise knowledge of key morphological features and geographic distributions for each taxon, some populations may be morphologically intermediate, a fact that can obscure the boundaries of some subspecific

19 taxa (Wherry, 1945; Wherry, 1955). Attention to style length, features of the calyx, leaf, and inflorescence vesture are the primary distinguishing attributes, but for some subspecific taxa, careful attention to the size, shape, and distribution of the leaves, and length of nodes is necessary to identify accessions to the subspecies level. Knowledge of inflorescence architecture and the relative placement of the involucre, and size of involucral bracts may also be used to identify some accessions (Wherry, 1955). Even if all of these factors are considered, identification of subspecific taxa in the P. carolina-P. glaberrima complex may prove difficult (Locklear, 2011a; Wherry, 1945, Wherry, 1955).

Dichotomous keys for identification of taxa can be found in Wherry (1955) and Locklear

(2011a).

Summary of Collections

I was able to make collections from 187 natural plant populations. In addition,

166 commercial species selections and cultivars were obtained for comparative and characterization purposes. Phlox collections from natural plant populations represented the diversity in 19 of 22 eastern species; Phlox cuspidata, P. pulchra, and P. pattersonii were not observed in situ, but accessions of P. pattersonii and P. pulchra of known provenance were obtained from nursery sources (Table 1.4). Several subspecies of P. drummondii were also not observed or collected, but collections were made of other selected subspecies (Table 1.1). Taxa in subsection Divaricatae were the most frequently collected, with a total of 96 accessions, but only 18 cultivated selections could be found.

20

In contrast, accessions from subsection Paniculatae included 21 wild collected accessions, yet 103 cultivars were obtained from nursery sources for characterization.

Collection numbers were given to all Phlox germplasm accessions. Collections made during USDA sponsored collection expeditions were given unique identification numbers that pertain to that specific trip. Other accessions made during OPGC sponsored collection expeditions were assigned unique accession numbers; an example is PZ10-001.

Numbers describe the accession in the following way: PZ for Peter Zale, the primary collector of Phlox taxa; 10-13 denote the year of the collection, which was followed by a three-digit individual accession number identifying all individuals within a given accession. A decimal point and additional numbers were given to each clone within an accession. Once entered in the NPGS/GRIN database, each accession was given an

OPGC number.

Summary of germplasm collections of Eastern Phlox taxa by subsection, species, and subspecies

Subsection Cluteanae

Phlox subsection Cluteanae is composed of two species, the rare eastern, P. buckleyi, and the western endemic Phlox cluteana (Wherry, 1955). These taxa were grouped together on the basis of similarities in plant habit; both species are colony forming perennials bearing rosettes with persistent leaves developed on long rhizomes.

Both species occur within small geographic regions, and are among the Phlox species with the most restrictive range. Neither species is known to be in general cultivation, but

21

P. buckleyi has been collected and introduced on a limited basis (Locklear, 2011a).

During May of 2012, a germplasm collection expedition was conducted specifically to

locate and study populations of P. buckleyi in its natural habitat. This fieldwork indicated

that P. buckleyi occurred as small, isolated populations in Virginia, but one population in

Greenbrier County, West Virginia was continuous in several miles of roadside habitat.

Phlox buckleyi Wherry - Greenbrier phlox, swordleaf phlox

Phlox buckeyi is a tetraploid (n = 14), perennial species that is narrowly endemic

to a small area of the Valley and Ridge province in Virginia and West Virginia (Figure

1.5); this species was described after the original collection sat in a herbarium for nearly

100 years (Wherry, 1930; Wherry, 1955). This area is partially characterized by the

discontinuous presence of Devonian and Ordovician age shale formations in various

stages of disintegration; the island-like pattern of shale outcroppings gives rise to a xeric

shale barren ecosystem that supports several microendemic species, but P. buckleyi is found on or near Brallier shale (Wherry, 1953). Greenbrier phlox grows in oak forests on soils derived from shale in a more advanced stage of decomposition, and generally in absence of the shale barren endemic plant species (Wherry, 1955). Most populations occur in small, isolated patches with relatively few flowering plants (<200), although it can be abundant in some regions. Phlox buckleyi is considered a globally rare species

(G2/S2)(Virginia Dept. Conservation, 2010).

The distinctive characters of P. buckleyi are the long style, sharply acuminate, lanceolate leaves, glandular-pubescent inflorescence herbage, foliose rosettes, long

22

rhizomes, and limited geographic distribution (Wherry, 1930; Wherry, 1955). This

species grows to 35 cm tall and produces pink flowers from May-June.

Phlox buckleyi has not been widely tested in horticulture, but has been introduced

into the gardens of native rare plant enthusiasts (Locklear, 2010). Nursery sources may

list this species, but such plants are incorrectly identified (The Flower Farm, Wisconsin,

listed as P. buckleyi is actually P. glaberrima ssp. triflora). The adaptability of either

species to landscape or garden environments is unknown, but habitat preferences suggest

that it may require specialized conditions for successful cultivation. However, the

compact, mounding habit, late spring flowering season, and deep green glossy foliage

warrant further evaluation of this taxon. This species also deserves evaluation for

potential resistance to powdery mildew, as it has not been observed to infect plants in

natural populations. Phlox buckleyi has potential as a parent in interspecific

hybridization improvement of Phlox, but there is no information about the ability of P.

buckleyi to hybridize with other species.

In total, 7 collections (PZ12-068, PZ12-070, PZ12-079, PZ12-081, PZ12-084,

PZ12-087, and PZ12-088) were made of P. buckleyi from Allegheny, Craig, and Roanoke

counties in Virginia, and from Greenbrier County, West Virginia in May of 2012 (Table

1.5). In Virginia, populations of P. buckleyi were small (<200 flowering plants), isolated

occurrences; plant habit and flowering characteristics were relatively uniform within this

region. In Roanoke County, Virginia, a population of P. buckleyi was found that was sympatric with P. ovata. These species have similar phenology and timing of seed maturation, despite this, interspecific hybrids were not observed at this site. Phlox

23

buckleyi was most abundant in Greenbrier County, West Virginia, and in one area, a

large, continuous population was discovered that stretched several miles of roadside.

Plants from these collections have smaller leaves, more compact rosettes, and a shorter

height at anthesis. This latter population represents an ideal region for seed collection of

this taxon. The P. buckleyi collections at the OPGC has been difficult to maintain in cultivation and most accessions failed to persist when planted outdoors or even when kept in the greenhouse.

Subsection Divaricatae

Taxonomy and Phylogenetics of Subsection Divaricatae

Subsection Divaricatae is the most diverse and widely distributed among eastern

subsections of Phlox. As currently circumscribed, subsection Divaricatae is composed of

up to 10 species and 16 subspecies (Table 1.1). Taxa are noteworthy for their extensive

phenotypic diversity within and among populations (Levin, 1966; Wherry, 1955; Smith

and Levin, 1967). Such diversity has resulted in the description of several subspecies,

primarily in the “Phlox pilosa complex” and P. drummondii (Wherry, 1955). While the

intraspecific classification of P. drummondii has remained relatively stable, the P. pilosa

complex has experienced significant revision and expansion (Ferguson, 1998; Levin,

1966; Levin and Smith, 1965; Smith and Levin, 1967). The previously described

subspecific taxa, P. pilosa ssp. pulcherrima, has been elevated to P. pulcherrima, and P.

pilosa ssp. latisepala, and P. pilosa ssp. riparia, have been elevated to Phlox villosissima

(Ferguson et al., 1999; Turner, 1998; Wherry, 1955). Three additional subspecies from

24

the P. pilosa complex, P. pilosa ssp. deamii, P. pilosa ssp. longipilosa, and P. pilosa ssp. sangamonensis, have been described and broadly accepted since Wherry’s monograph

(Levin, 1965; Levin, 1966; Locklear, 2009). One species, P. pattersonii, was discovered in Northeastern Mexico and is included in subsection Divaricatae; this taxon extends the

known geographic range of the subsection (Prather, 1994).

Many taxa in this subsection are not in cultivation despite the fact that all are

considered to be showy native wildflowers. In cultivation, different geographic races of

taxa in the P. pilosa complex, such as P. pilosa ssp. longipilosa and P. villosissima, may

have potential as ornamental, disease resistant plants (Locklear, 2011a; Wherry, 1955).

This has been shown to be the case with the rare, and narrowly distributed P. pilosa ssp.

sangmonensis; since introduction it has become established in gardens and considered

one of the more persistent forms of P. pilosa in cultivation (Locklear, 2011a).

Furthermore, past studies have shown that there are few barriers to interspecific

hybridization in the P. pilosa complex, and use of selected genotypes to create

interspecific hybrids may result in broadly adapted, ornamental plants (Levin, 1966).

Several taxa in this subsection have highly restricted natural distributions, are of

conservation concern, and are considered imperiled with Natural Heritage rankings of

G2/S2 (Buthod and Skvaria, 2014). Since some taxa occur in geographically remote

regions of the U.S., they are not commonly collected, and relevant information pertaining

to initial characterization may be lacking (Levin, 1966; Locklear, 2011a). Taxa in the P.

pilosa complex tend to occur in ecosystems of exceptional plant species diversity, and

may be considered indicator species for pristine natural areas and ecosystems (Locklear,

25

2011a). Many species may be important nectar plants for a wide range of Lepidoptera

(Grant and Grant, 1965). Ex situ Conservation and multiplication of these species through responsible germplasm collection is needed for potential restoration efforts and further research projects.

A total of 114 collections were made of taxa in subsection Divaricatae (Table

1.4). Of these, 96 were collected from natural plant populations, and 20 were cultivated

selections (Table 1.6). The taxa in subsection Divaricatae were a priority on all

collection expeditions and the focus was on collection of members of the Phlox pilosa

complex. In particular, I targeted P. pilosa ssp. pilosa from the entire east-west range of

the species, as well as the narrow endemic and poorly understood P. pilosa ssp. deamii

and P. pilosa ssp. longipilosa, P. floridana P. pulcherrima, and P. villosissima. In addition to these taxa, at least one accession of all other taxa in subsection Divaricatae was obtained (Table 1.6).

Phlox amoena Sims - hairy phlox, chalice phlox

Phlox amoena is a diploid (n = 7), perennial species of the Interior Low Plateaus,

Valley and Ridge, and Coastal Plain provinces closely related to P. divaricata and P. pilosa (Figure 1.2, Table 1.1) (Ferguson et al., 1999; Levin, 1966). Wherry (1955) described two subspecies, P. amoena ssp. amoena and P. amoena ssp. lighthipei Wherry

but Ferguson (1998) treated both as P.amoena, citing P. amoena ssp. lighthipei as nothing more than a glabrous form of P. amoena. It typically grows in disturbed, xeric sites on bare soil of sloping ground where there is little competition from sympatric

26

species, accumulation of organic matter, or overhead shade. For this reason, this species

is a frequent invader of roadside right-of-ways and other areas subject to anthropogenic

disturbance (Figure 1.6). In such areas, P amoena can become quite abundant; it benefits

from the infrequent mowing that reduces competition from neighboring plants.

Phlox amoena is readily distinguished from other taxa in subsection Divaricatae.

The inflorescence is unique in that the involucre tightly surrounds the congested, densely

hirsute inflorescence (Locklear, 2011a; Wherry, 1955). In most individuals, there is a long scape separating the involucre and inflorescence architecture. The common name

“hairy phlox” is derived from the presence of abundant, multi-cellular white hairs on the stems, leaves, and inflorescence. Phlox amoena has a mounding habit, produces copious sterile stems, and grows to 30 cm in height. The flowers are typically Phlox-purple, but there are populations with bright pink flowers, and rare white-flowered forms (Figure

1.6). Compared with levels of phenotypic variation in the P. pilosa complex, this species

is relatively invariant. This species is not commonly cultivated, but appears in nursery

catalogs occasionally, and local forms may be cultivated in areas where it is native. It is

best suited to sterile, acid soils and limited competition from neighboring plants in the

garden. It may have potential as a rock garden plant, or for roadside right-of-way restoration. Initial evaluation suggests that individual plants may be short-lived, and that establishment in garden conditions may require regular seedling recruitment. There are no known cultivars (Bendtsen, 2009; Fuchs, 1994).

Twelve Phlox amoena accessions were made during the course of germplasm collection expeditions, and 3 accessions were obtained from nursery sources (Table 1.6).

27

Accessions from nurseries were of clonal material collected from natural plant

populations. Eleven collections were morphologically similar (PZ11-025, PZ11-062,

PZ11-068, PZ12-042, PZ12-053, PZ12-057, PZ12-094, PZ12-102, PZ12-132,

PZSH2011-001). Some collections (PZ10-111, PZ10-112, PZ10-113, and PZ10-122) perished before any evaluation or analyses could be performed. Two collections made during germplasm collection expeditions were distinct from others.

The accession PZ11-032 was the most phenotypically and ecologically unique collection of Phlox amoena. It was collected in McCreary County, Kentucky from river scour prairies (also known as Cumberlandian cobble bars) along the Big South Fork of the Cumberland River (Campbell, 2012). This unique, glabrous from of P. amoena occurs in the damp, peaty soils associated with this habitat. Typical individuals of P. amoena are comparatively few-stemmed, short-lived plants, but in comparison, PZ11-032 is distinct; this form produced large, domed hummocks to 1 m in diameter that were much larger than individual plants of P. amoena from any other observed population.

Individuals also appeared to be long-lived perennials that produced an abundance of evergreen, sterile stems. Flowers of this form are of heavier substance, and rich, bright pink color that is different from typical upland forms. Based on morphological attributes, this form is referable to the now unrecognized taxon Phlox amoena ssp. lighthipei, but this population does not occur within the range presented for that taxon (Wherry, 1955).

This taxon has been described as P. amoena “var. nova” in Atlas of Kentucky Flora, but further research is needed to determine the relationship of this taxon and typical P. amoena (Campbell, 2012).

28

An individual clone in the accession P. amoena PZ12-041 was the only flower

color variant found in all P. amoena populations that were observed. Flowers open

lavender-white with purple striae, fade to deep lavender (Figure 1.6). This was found in a

Logan County, Kentucky populations at the northern range limit of P. amoena. A single

individual was found at the edge of a Juniperus virginiana L. glade near the roadside.

Vegetative propagation of this clone proved unsuccessful, however this population of P.

amoena is extensive and further searching for additional plants might reveal more similar

variants.

Phlox divaricata L. ssp. divaricata and Phlox divaricata L. ssp. laphamii

(Wood)Wherry Phlox - timber phlox, wild sweet william

Phlox divaricata is a diploid (n=7), perennial species and, after P. pilosa, the most widely distributed of all Phlox species (Levin, 1967; Locklear, 2011a; Wherry, 1955).

Phlox divaricata is most abundant in the Great Plains and Interior Highlands regions, but can also be found in the Appalachian highlands, Coastal Plain, and Valley and Ridge provinces (Figure 1.2). Two subspecies are recognized by the presence of an apical notch in each corolla lobe (P. divaricata ssp. divaricata), or absence of an apical notch (P. divaricata ssp. laphamii) (Figures 1.7, 1.8). Phlox divaricata ssp. divaricata generally occurs in the eastern portion of the species range, and P. divaricata ssp. laphamii to the south and west (Wherry, 1955). However, extensive populations with both flower phenotypes can be found where the ranges of the subspecies overlap (Levin, 1967).

Phlox divaricata occurs in a variety of ecoystems, but is most abundant in hydric or

29

mesic soils in mesophytic hardwood forest and forested riparian zones. In many places,

P. divaricata forms large populations consisting of thousands individuals and can be a

dominant aspect of the spring flora (Levin, 1967; Wherry, 1955).

Phlox divaricata is distinguished from other taxa in subsection Divaricatae by the broadly elliptical, glabrous or pubescent leaves on the upper portion of flowering stems and second flush of sterile stems, the presence or absence of a notch in the distal edge of the corolla lobe, the early spring flowering phenology, and habitat preference. Phlox divaricata grows to 40 cm high, and has an erect, mounding habit. The lavender-purple flowers are produced in late March-early May; color is distinctly different from that of P. amoena and P. pilosa. Compared to the P. pilosa complex, P. divaricata is a relatively morphologically invariant species given its large geographical range, although there can be siginificant variation between individuals within populations, and there are unique regional phenotypes (Levin, 1967). White flowered variants of P. divaricata ssp. divaricata are particularly abundant within the central portion of the range, and can generally be found in any large population (Figure 1.7). When sympatric, P. divaricata can hybridize with P. pilosa, known as Phlox xglutinosa, and P. amoena, known as Phlox

xrugellii (Levin, 1966; Wherry, 1955).

Several cultivars of P. divaricata have been introduced into horticulture. Most

selections have been made from P. divaricata ssp. laphamii, because of the somewhat

larger, more rounded flowers. Observation of plants at the OPGC suggests that plants are highly susceptible to rodent damage. Individual cultivars may be short-lived under landscape conditions. Planting several cultivars together will encourage pollination (if

30

proper pollinators are present) and seed development that may result in seedling

recruitment and establishment of a persistent garden population.

A total of 22 wild P. divaricata populations were sampled, and 7 cultivars were

obtained from commercial sources (Table 1.6). Descriptions and names of cultivars can be found in several sources and are not repeated here (Bendtsen, 2009; Fuchs, 1994;

Locklear, 2011a; Oliver, 2011).

Collections from the Great Plains province (PZ10-143, PZ10-145, PZ11-003,

PZ11-004, PZ11-006, PZ11-008, PZ11-023, PZ11-074, and PZ12-142) in Ohio, Indiana,

Illinois, Wisconsin, and Minnesota consist of both subspecies, but aside from the presence or absence of the corolla lobe notch, these taxa are similar in plant habit, foliage, calyx features, and range of flower color. Large continuous populations can be found in some regions: In Ohio, this taxon has been reported from all 88 counties, and it is likely that much larger populations existed prior to colonization (Levin, 1967; Wherry,

1955). However, because of the density with which this species occurs in some regions, it is possible that novel phenotypic variants may exist in some populations. Collections of P. divaricata from the Valley and Ridge province and Allegheny plateaus (PZ11-073

Botetourt County, Virginia, PZ12-113 Union County, Tennessee, PZ12-133 Polk County,

Tennessee) were phenotypically similar to collections from the Great Plains province.

Populations of P. divaricata ssp. laphamii from the coastal plain are phenotypically distinct from populations to the north of that region, and exist as comparatively small, fragmented populations where suitable habitat occurs. Populations from Gadsden and Jackson Counties, Florida (PZSH2011-004, PZSH2011-007,

31

PZSH2011-008, PZSH2011-012) occurred at the southern edge of the natural distribution on dolomite outcroppings. These populations have a unique flower color with richly colored, lavender-purple flowers with deep, red-purple striae fused into a star–shaped pattern at the center (Figure 1.8). Two collections from southern Alabama, PZSH2011-

018 from Wilcox County, and PZ12-039 from Montgomery County, are similar to populations from Florida, but the flower color is lighter and ranged from light to dark lavender, only some individuals have the distinctive striae coloration and pattern, and the plants grow taller at anthesis. Similar to the above collections, PZSH2011-026 from

Wilkinson County, Mississippi had the most richly colored lavender-purple flowers of all

P. divaricata collections. Flower color and size in these populations is superior to all other accessions of P. divaricata made during germplasm collection expeditions.

Phlox divaricata ssp. laphamii PZSH2011-022 was collected from Jasper County,

Mississippi. This collection was noteworthy for the large, rounded flowers with larger than normal, overlapping corolla lobes and unique lavender-pink flower color. The accession PZ11-035 was also collected in Mississippi, and appears to be a similar phenotype to the collections from the Great Plains province.

Two accessions of P. divaricata ssp. laphamii were obtained from the Interior

Highands region (PZ10-224 from Reynolds County, Missouri, and PZ13-008 from

Newton County, Arkansas). The first collection perished before any analysis could take place, and the latter collection is under evaluation

32

The Texas Annual Phlox of Subsection Divaricatae

Phlox cuspidata, P. drummondii, and P. roemeriana form a monophyletic group that is essentially endemic to Texas, but outlying populations may exist in surrounding states or Mexico. There species are morphologically similar; they are all winter annuals that flower in March-May with variable flower color. Initially all taxa have opposite

vernation, which becomes alternate on the upper portions of flowering stems. Seven

accessions of P. drummondii were acquired from natural plant populations, and 2 seed

strains were acquired from seed producers (Table 1.6). One accession of P. roemeriana

was acquired from a natural population (Table 1.6). Phlox cuspidata germplasm was not

collected or analyzed during the course of this study. The latter two species are not

known to be in cultivation.

Phlox drummondii Hooker - Drummond’s phlox

Phlox drummondii is a diploid (n = 7), annual species nearly endemic to Texas

that occurs in a variety of habitat types and ecosystems, but is typically found on sandy

soils (Locklear, 2011a; Wherry, 1955). Although it occurs over limited geographical area

in comparison with the widepsread P. pilosa, this taxon exhibits tremendous phenotypic

variation within and among populations, particularly in regards to flower color, over its

relatively limited range (Figure 1.9). Phlox drummondii has been divided into as many as

3 species and 7 subspecies, but the most recent treatment of P. drummondii (including

Phlox glabriflora) by Locklear (2011a) where 6 subspecies are recognized (P.

drummondii ssp. drummondii, ssp. mccallisteri, ssp. glabriflora, ssp. johnstonii, ssp.

33

peregrina and ssp. tharpii) has been followed in this dissertation (Levin, 1977; Locklear,

2011a; Wherry, 1955). The distribution of the different subspecies has been shown to

relate to geography and soil type, but the variation is clinal, and morphologicaly

intermediate populations occur between populations referable to subspecies (Schwaigerle

and Levin, 1990). Phlox drummondii has been an important model species for plant

evolution, hybridization, and domestication studies, and is the most thoroughly

characterized Phlox species (Levin, 1976).

This taxon has an annual life cycle that distinguishes it from perennial taxa in

subsection Divariacatae. All taxa have leaves that are initially opposite, but become

alternate in the distal portions of the stem. The plants have a mounding habit, and flower

from late winter (February) until late spring (June) in favorable seasons with abundant

rainfall. Flower color ranges from red, pink, white, and purple (Figure 1.9); some

populations have relatively uniformly colored flowers, but some are extremely variable

and several color forms can be found (Wherry, 1955).

Phlox drummondii has been a popular garden plant since its discovery; the original introduction of these species into cultivation has resulted in the development of over 125 distinct breeding lines selected for differences in flower color and size, habit, and adaptation (Levin, 1976a; Levin, 1976b). This species has been widely used in highway right-of-way plantings in the Southeastern United States, and planted as an ornamental in other regions of the world.

A total of 7 collections of P. drummondii were made (Table 1.6). Four of these came from documented naturally occurring populations, one was collected from an

34

adventive population along a Texas roadside, and two were obtained as commercial seed

strains.

Two collections of P. drummondii ssp. drummondii, PZ10-160 (syn. TX-035) and

PZ10-161 (syn. TX-037), were made from Caldwell County cemeteries, in remnant post oak savannah habitat. These collections are noteworthy for their scarlet color with deeper red striae fused into a star-shaped pattern; this combination of floral features is indicative of this subspecies (Figure 1.9). Wherry (1955) described such populations as Phlox drummondii ssp. wilcoxiana. The bright red flower color found in this taxon is among the most intense and unique of all Phlox species.

One collection of P. drummondii ssp. peregrina was collected from adventive populations at an abandoned homestead along a Bastrop County, Texas roadside. Since

P. drummondii is frequently cultivated, Locklear (2011a) described P. drummondii ssp. peregrina to account for adventive or naturalized populations that result from cultivated plantings. Even within this small population, flower color was variable, with pink, white, and red flowered individuals present.

One collection of P. drummondii ssp. maccallisteri, PZ10-163 (syn. TX-48) was made from Wilson County, Texas. Flower color in this population was a comparatively uniform pink, with prominent striae and a white flush at the center of of each flower.

Habit and foliage characteristics are superficially similar to other collections.

The collection of P. drummondii PZ10-164 (syn. TX-56) was made from a roadside easement in Gonzales County, Texas. In this population most plants produced bright red flowers, but there were a considerable number of individuals with white and

35 pink flowers. The other intraspecific taxa of P. drummondii were not collected during the course of germplasm collection expeditions.

Phlox cuspidata Scheele - Navasota phlox

Phlox cuspidata is a diploid (n = 7), annual species that is endemic to Texas. This species reportedly grows in post oak (Quercus stellata Wangenh.) savannah on clay-rich soils. It is closely related to other Texas endemic annual species (Ferguson et al., 1999).

It is not known to be in cultivation, but may be involved in the production of interspecific hybrids with P. drummondii (Erbe, 1960).

There were no collections made of this taxon from either natural plant populations or nursery sources. However, because this species is known to hybridize in the wild with

P. drummondii, germplasm collection and conservation may provide an important genetic resource for development of hybrids; these populations have already been used to study the effects of hybridization on speciation, fecundity, and trait inheritance (Erbe, 1960;

Ruane, 2009). Populations of P. cuspidata should be targeted for future seed collection to conserve potentially novel variation that could be used for plant breeding and species characterization purposes. During a botanical expedition to Texas in May 2010, an attempt was made to locate a population of P. cuspidata, but the roadside habitat of the plants had been converted to turfgrass and the population was no longer extant.

36

Phlox roemeriana Scheele - golden-eye phlox

Phox roemeriana is a diploid (n=7) annual, species now placed in subsection

Divaricatae (Ferguson et al., 1999: Prather, 1994). Wherry (1955) placed this species in subsection Nanae on the basis of leaf and morphology, but more recent molecular studies indicate that it forms a monophyletic group with other annual species in subsection Divaricatae. This species is endemic to the Edward’s Plateau in central Texas

(Figure 1.2). In this region it grows in xeric habitat in limestone glades with sparse grasses near limestone outcroppings (Figure 1.10) where it is associated with cactus such as the horse crippler (Echinocactus texensis Hopffer), Texas live oak (Quercus fusiformis

Small), and small ball moss (Tillandsia recurvata (L.)L).

Phlox roemeriana can be distinguished from other annual Phlox species and taxa in subsection Divaricatae based on differences in leaf and sepal morphology, and by the distinctly tri-colored flowers that are gold at the center, fade to white, and have a pink edge (Wherry, 1955). This color combination does not appear to be found in any other eastern or western Phlox species. The overall habit is smaller than P. drummondii and the flowers are larger in size. Wherry (1955) reports that this species is difficult to maintain under landscape conditions, even within its native range. However, initial cultural evaluations at the OPGC indicate that this species is readily grown in containers.

It does not appear that this species has been used in interspecific hybridization experiments.

37

One seed collection of this species (TX-058) was made from a large population in

Comal County, Texas (Table 1.6). This large population appeared to be typical of the species, and was in full flower on May 29, 2010 (Figure 1.10).

Phlox floridana Bentham - Florida phlox

Phlox floridana is a tetraploid (n = 14), perennial species of restricted distribution in the Coastal Plain of Alabama, Georgia, and Florida (Wherry, 1955). It occurs in xeric sandhill communities within upland pine ecosystems (Figure 1.11) (Locklear, 2011a).

Frequent fires, every 1-3 years, are necessary to maintain these ecosystems and it is

possible that the persistence of P. floridana in these areas may be linked to the prevailing

fire management practices in a given region (Florida Natural Areas Inventory, 2010).

Wherry (1943b) described Phlox floridana ssp. bella as a dwarf form of P. floridana, but

this taxon was not mentioned in his later monograph or in a later treatment of Phlox

(Locklear, 2011a; Wherry, 1955). It is likely that this taxon was an environmentally

induced variant, or a local population of genetic dwarfs within the range of variation of

the species. Levin (1968) considered P. floridana to be an amphidiploid that originated via interspecific hybridization between P. carolina and P. pilosa, but this has not been verified.

The distinctive morphological features of P. floridana are the glabrous corolla tube, relatively short, glabrous, or glabrescent dark green leaves in which the longest leaves are found near the middle of the stem, and small involucral bracts (Locklear, 2011;

Wherry, 1955). This species is sympatric with P. pilosa, but can be distinguished by

38 morphological and phenological traits; in the same habitat, P. floridana flowers later than

P. pilosa and the species occupy different ecological niches (Wherry, 1943b; Wherry,

1955). Phlox floridana grows to 40 cm in height with an erect, mounding habit. The flowers are heavily textured, and occur from Mid-April through the end of May in the wild. The deep green foliage is glabrous, and the foliage on emerging shoots can be pigmented with copper-red hues.

Phlox floridana has only rarely been cultivated, but initial reports indicate that this species is more cold hardy than would be expected (Wherry, 1943b). Initial cultural experiments within the native range of the species suggested that the species is difficult to grow and lacking vigor (Booth, 2008). However the experiments were not replicated, a limited number of collections were evaluated, and the species was not exhaustively tested

(Booth, 2008). Seed germination has not been characterized for P. floridana, and it possible that seed germination may be correlated to the frequency of fires in its habitat.

One collection of P. floridana PZSH2011-010 was made from a mowed roadside right-of-way in Jackson County, Florida (Table 1.6). The increased abundance of this species in mowed area suggests that mowing may simulate the effect of fire by reducing competition from neighboring vegetation in a similar manner. In container cultivation, these plants have been vigorous and floriferous. Under long days, the plants will flower year-around in the greenhouse, and do not exhibit an obligate vernalization requirement to induce flowering (Figure 1.11). This accession has not been tested for cold hardiness at the OPGC, although P. floridana was reported to be hardy as far north as Philadelphia,

PA (Wherry, 1943b)

39

Phlox pattersonii Prather - Coahuila phlox

Phlox pattersonii is a diploid (n=7), perennial species from high elevations (1200-

2200 m) in the Sierra Madre Oriental in Coahuila and Nuevo Leon States in Northern

Mexico (Ferguson, 1998; Prather, 1994). This species occurs in mesic microclimates of an otherwise xeric region, and can be primarily found in oak (Quercus spp.) forest or open ridges at higher elevations. This is the most recently described Phlox species in subsection Divaricatae, and has only been collected a few times from a relatively limited geographic region, but the remote mountainous landscape in that region suggests it may be more widespread than currently known.

This species bears affinities to P. villosissima, P. pilosa ssp. longipilosa, and P. pilosa ssp. ozarkana, but differs in several features. Phlox pattersonii is an evergreen suffruticose, perennial with many nodes per stem, alternate upper leaves on flowering stems, and summer flowering period (Prather, 1994). This combination of morphological features is intermediate between the annual species formerly in subsection

Drummondianae, and perennial species in subsection Divaricatae. Discovery of this taxon resulted in the unification of these subsections (Prather, 1994). This taxon grows to

40 cm in height and flowers during the summer months. Cold hardiness and landscape adaptability are unknown, but the occurrence of the species at high elevations suggests that there may be cold hardiness in some populations.

The species is virtually unknown in cultivation, but one collection, made by the author of the species, is distributed by Arrowhead Alpines nursery in Fowlerville,

40

Michigan is available for garden use (Alan Prather, Personal Communication, 11 October

2012). One accession, PZ10-247, was obtained from Arrowhead alpines (Table 1.6). The morphological characteristics are typical of the species. However it should be noted that the flowers of this taxon are among the largest in subsection Divaricatae. This accession appears to consist of a single clone, as pollination between these plants has failed to produce seeds.

The Phlox pilosa complex

Collectively, the P. pilosa complex is the most geographically widespread and phenotypically variable group of Phlox taxa. It has also been one of the most taxonomically confused groups in the genus (Table 1.1). Historically, as many as 11 subspecies have been described, but in the current circumscription, 6 subspecies are recognized and three former subspecies are elevated to two species (P. pulcherrima, P. villosissima) (Ferguson, 1998; Ferguson et al., 1999; Locklear, 2011a; Wherry, 1955).

One taxon, Phlox pilosa ssp. detonsa was reduced to a synonym of P. pilosa ssp. pilosa;

Ferguson (1998) treated it as a glabrous form of P. pilosa ssp. pilosa.

Members of the Phlox pilosa complex were a priority for collection and germplasm development because of the extensive phenotypic diversity, potential for interspecific hybridization, and potential as research subjects for describing patterns of ecotypic differentiation in a widespread species. A total of 48 accessions from species within the P. pilosa complex were collected from natural plant populations and 6

41

cultivars were also obtained (Table 1.4, 1.6). There were more collections of this taxon than any other during the course of germplasm collection expeditions.

Phlox pilosa L. ssp. pilosa - downy phlox

Phlox pilosa ssp. pilosa is a diploid (n = 7) or tetraploid (n = 14) short to long- lived perennial that has the widest geographic distribution within this group (Smith and

Levin, 1967; Wherry, 1955; Worcester et al., 2012). Phlox pilosa is most abundant in the

Coastal Plain, Great Plains, Interior Highlands, and Interior Low Plateaus, but also occurs in Piedmont (Figure 1.2). This taxon is distinctly absent from the southern Appalachian

Highlands, and Valley and Ridge provinces (Figure 1.2). Phlox pilosa grows in a variety of mesic to xeric habitats and occurs in tall grass prairie, oak savannah, limestone barrens, cedar glades, and open woods (Figure 1.12) (Locklear, 2011a). Ecotypic differentiation and adaptation has resulted extensive phenotypic diversity in this species that may be worthy of botanic recognition, and there is vast potential for horticultural selection (Figures 1.11, 1.12) (Levin, 1966; Worcester et al., 2012).

Phlox pilosa ssp. pilosa is distinguished from other taxa in the complex by the presence of glandular pubescence on the leaves, stems and inflorescence, subulate sepal blades with a short awn tip (Wherry, 1955). However, it should be noted that these features can be variable, and not all of them may be present within individuals of a given population. Phlox pilosa grows from 10-40 cm in height, has linear to lanceolate leaves that can be glabrous, pubescent, or hirsute. Foliage is generally a shade of green, but some populations have purple-flushed foliage (Figure 1.12). Flowers are produced from

42 mid-spring to early summer, and color can be a range of pink, purple, and white hues

(Figure 1.12, 1.13).

Phlox pilosa is regarded as an unsatisfactory garden and landscape plant that has been described as short-lived with a poor ability to reseed and persist (Wherry, 1935a;

Wherry, 1955). However, many of the subspecies have rarely or never been introduced into cultivation, and ecotypic differentiation may influence adaptation to landscape conditions; evaluation of a wide variety of genotypes may result in selection of particularly adaptable clones or seed populations with potentially novel ornamental traits

(Locklear, 2011a; Wherry, 1935a). There is some evidence that P. pilosa ssp. ozarkana is a satisfactory garden plant that is long-lived and persistent in landscape conditions

(Oliver, 2011; Wherry, 1955). Based on habitat preference, P. pilosa ssp. deamii has been suggested for horticultural evaluation (Locklear, 2011a). Other subspecies, such as

P. pilosa ssp. fulgida have considerable ornamental merit, but were reportedly short-lived under garden conditions, but might have merit as components of native landscapes and ecological restoration projects within its native range (Locklear, 2011; Wherry, 1935;

Wherry, 1955). However, cultural information pertaining to P. pilosa is highly anecdotal in nature and has not been documented in a wide array of different climates and soil types. The limited Phlox evaluation trials have not tested a wide variety of P. pilosa subspecies, ecotypes, or cultivars (Hawke, 1999; Hawke, 2011).

Initial studies indicate extensive potential for interspecific hybridization of P. pilosa with other taxa in subsection Divaricatae. Biosystematic studies of Phlox subsection Divaricatae revealed the potential of this species to hybridize with other

43

members of the subsection, but the resultant F1 hybrids were not evaluated (Levin, 1966).

Since these initial studies, attempts to define the potential for interspecific hybridization

have not been reported. Interspecific and intersubspecific hybridization with other taxa

known or found to be adaptable to landscape conditions may result in hybrids that are

widely adaptable and persistent (Levin, 1966; Levin, 1975). One putative hybrid, Phlox

x ‘Chattahoochee’, is reportedly a hybrid of P. pilosa ssp. pilosa and P. divaricata, and

systematic plant evaluation trials have shown this taxon to be one of the top performers

among a variety of eastern Phlox taxa (Hawke, 1999; Locklear, 2011).

Phlox pilosa ssp. pilosa is listed as an endangered species in Maryland, New

Jersey, New York, and Pennsylvania (USDA Plants Database, 2014), giving it additional

interest as a conservation subject in those states. Throughout its range, P. pilosa is

associated with ecosystems of high biodiversity (Locklear, 2011; Wherry, 1955). The

widespread occurrence of P. pilosa in these areas renders it of interest to restoration

ecologists and native plant enthusiasts.

Given the extensive geographic range and phenotypic variation of this species, it

was a natural priority for collection; 35 collections of P. pilosa ssp. pilosa were made

throughout the distribution range (Table 1.6; Chapter 4). Of these, 30 were collected

from natural plant populations and 5 were obtained as cultivar selections that were

originally made from natural plants populations (Table 1.6). Collections were made from

a variety of different habitat types (Figure 1.1, 1.2)

Phlox pilosa ssp. pilosa is widespread and abundant in the Great Plains province

(Figure 1.2). Populations in this region are morphologically similar and appear to

44

comprise a broadly distributed metapopulation (Figure 1.12). Additionally, two

populations from the Interior Low Plateau province are included in this group. The

accessions PZ10-154, PZ10-192, PZ12- 040, PZ12-048, PZ12-061, PZ12-063 from Ohio,

Indiana, and Kentucky are included in this group. Within the Oak Openings of region of

Lucas County, Ohio, a population of this taxon at Lou Campbell Prairie in the Oak

Openings Preserve, collected initially as PZ10-192, was visited several times to study flower color, size, and pattern variations, phenology, and seed production. Plants from this population grow to ca. 40 cm tall, bear pink flowers with red striae from May until late June or early July, and have linear-lanceolate leaves, and lanceolate involucral bracts

(Figure 1.12). Initial plantings of PZ10-192 have been persistent in constructed landscapes and are worthy of further evaluation for adaptation to gardens and production systems.

Perhaps the most distinctive accession of the Great Plains form of P. pilosa ssp. pilosa was PZ12-013. This was collected from a small, isolated population that occurs in remnant prairie patches distributed along west-facing bluffs of the Big Darby Creek in

Franklin County, Ohio. Although this collection can be keyed to P. pilosa ssp. pilosa, it appears to be a locally adapted phenotype that differs from other P. pilosa ssp. pilosa in several regards; the foliage is glabrescent (compared to pilose/pubescent), there is a proliferation of sterile, persistent stems after flower, the habit is mounding compared to erect, resulting in a shorter overall plant height.

Populations of P. pilosa ssp. pilosa found in the Interior Low Plateau and Coastal

Plain from the Tallahassee Florida region westward to western Louisiana and Oklahoma

45 are more phenotypically variable than populations to the north. Population level differences in flower color, phenology, leaf color and shape, and habitat preference are more pronounced than in populations to the north. Several horticulturally distinct collections were made within this region.

One of the largest, most phenotypically diverse populations of P. pilosa ssp. pilosa was encountered in Forest County, Mississippi. This population was among the largest of all Phlox populations seen and consisted of thousands of individuals. The collection PZSH2011-36 is exceptional for two reasons: almost all individuals exhibited unique, purple flushed foliage, and the two of the most unique flower color variants of P. pilosa were discovered here. In one individual, the flowers were white with pink striae fused into a star, and other had dark-purple, richly colored flowers that differed from the typical bright pink (Figure 1.13). Additionally, plants from this population have deeply colored maroon purple foliage on the persistent sterile stems during the fall, winter and early spring months (Figure 1.13). The white flowered phenotype was not found in all subspecies of P. pilosa, and thus far has only been found in P. pilosa ssp. fulgida and P. pilosa ssp. ozarkana. However, these differ in the range of color saturation and color patterning.

The collections of Phlox pilosa ssp. pilosa (PZ12-049, PZ12-051, PZ12-052) from Trigg County, Kentucky, and Benton and Henry Counties in Tennessee in the vicinity of the Land Between the Lakes region appear to be a unique ecotype based on preliminary evaluations. In cultivation, plants from this region produce abundant pale pink flowers that fade to white, giving the impression of two different colored flowers on

46 the same individual; this trait was not seen in any other accessions of P. pilosa. Flowers are produced over an exceptionally long period from April until June in cultivation.

Plants are large, growing to ca. 60 cm in height and grow in mesic deciduous forest near streams. The phylogenetic distinctiveness of these populations requires further testing.

Tetraploid populations of P. pilosa ssp. pilosa (PZ11-069, PZ11-070) were discovered in Bossier Parish, Louisiana (Chapter 4). Tetraploid populations of P. pilosa were discovered further west of this region; the populations in Bossier Parrish represent an eastern range extension of these populations (Worcester et al., 2012). These plants are

‘cryptic’ polyploids; there are no obvious macromorphological differences between tetraploid and diploid populations, and the two ploidy levels do not occur in mixed populations (Worcester et al., 2012; Chapter 4). Three other accessions of P. pilosa ssp. pilosa obtained from nursery sources were tetraploid; one of these was collected by Tony

Avent of Plant Delights Nursery (Raleigh, North Carolina) in Polk County, North

Carolina from the disjunct populations at the eastern edge of the P. pilosa geographical distribution, another was collected at an unknown location in Alabama (PZ11-057), and one, ‘Eco Happy Traveler’, was selected by Don Jacobs, of the former Eco gardens in

Decatur, Georgia, but the origin of this clone is unknown, but presumed to have been selected from a wild population. The distinctiveness of tetraploid populations is discussed in Chapter 4.

Phlox pilosa PZ10-023 was collected from DeKalb County, Georgia, but perished before subsequent analysis could take place. Initial observation showed this plant to have copiously hirsute leaves and stems, with multi-cellular, eglandular hairs similar to P.

47

pilosa ssp. longipilosa. Flowers were pale pink, with a white eye and lacked striae. This accession was collected in the vicinity of Arabia Mountain, an area known for having a

unique flora, and this may represent a local ecotype; plants from this unique habitat

require further study.

Additional cultivars of P. pilosa were obtained for this study and include ‘Forest

Frost’, ‘Lavender Cloud’, and ‘Racy Pink’. All of these cultivars were collected from

natural plant populations. Morphological characters of cultivars are consistent with those

of P. pilosa ssp. pilosa.

Phlox pilosa ssp. deamii (Deam) Levin - Deam’s downy phlox

Phlox pilosa ssp. deamii is a diploid (n=7), perennial species from the Interior

Low Plateau (Levin, 1966; Smith and Levin, 1967). This taxon was described by Levin

(1966) based on material originally collected by Charles Deam in southern Indiana.

Wherry (1955) considered these populations as part of his Phlox pilosa ssp. pulcherrima,

but they occupy distinctly different geographic ranges and are readily separated by

morphological features. The morphology of this taxon is intermediate between P.

amoena and P. pilosa ssp. pilosa, and it is thought to be of hybrid origin (Levin and

Smith, 1966). It is narrowly endemic to the Tennessee River Drainage in Indiana,

Kentucky, and Tennessee. Locklear (2011a) reported the habitat as post oak (Quercus

stellata) flat woods, but field observation during germplasm collection expedition

indicated P. pilosa ssp. deamii does not occur in that habitat. This species occurs in

mesic forest on moderate slopes of small streams with a rich assemblage of spring

48 ephemerals and prairie forbs (Figure 1.14). Although this species is not formally listed by any state or federal agencies, P. pilosa ssp. deamii has been reported as a globally rare species, and is considered endangered in Indiana and Kentucky (Campbell, 2012; Indiana

Heritage Division, 2013). Numerous attempts were made to relocate populations during the course of OPGC collection expeditions; only two populations were discovered, and the third in Tennessee does not appear to be the same taxon. My field experience indicates that Phlox pilosa ssp. deamii should be considered critically endangered throughout its range.

Phlox pilosa ssp. deamii is distinguished from other taxa in subsection

Divaricatae by features of the calyx and inflorescence. The calyx is characteristically hirsute, and is similar to that of P. amoena in that in bears long, multi-cellular white hairs that are egandular, as opposed to the short, glandular hairs of P. pilosa. The sepal lobes of P. pilosa ssp. deamii are cuspidate and erect like P. amoena, whereas those of P. pilosa are subulate and somewhat reflexed. (Figure 1.14) (Levin, 1966; Levin and Smith

1966). The leaves of P. pilosa ssp. deamii are intermediate in shape and size, and intermediate between P. amoena and P. pilosa, but favor those of P. pilosa (Levin, 1966;

Levin and Smith, 1966).

Phlox pilosa ssp. deamii has not been introduced into cultivation and the horticultural value remains unknown and untested. Anecdotal evidence, and habitat preference suggested that it may be more adaptable to garden and landscape settings that other members of the P. pilosa complex (Locklear, 2011a). Initial observations indicate that this taxon is vigorous and hardy in outdoor conditions, and produces a large number

49

of flowering stems per plant, and have a shorter, mounded habit (to 30 cm tall) that

exhibits greater flower coverage than other members of the P. pilosa complex (Figure

1.14). Flower color is a uniform bright pink with red striae. Flowering occurs from late

April to Mid-June (Figure 1.14).

The collection P. pilosa ssp. deamii PZ11-026/PZ12-045 from Christian County,

Kentucky represents the typical form of this species (Table 1.6). These populations are

morphologically homogenous and occur in open woods and roadsides. In cultivation,

individual plants flower during April and May (Figure 1.15). Plants have a distinctive

mounding habit that is more compact than most subspecies of P. pilosa in cultivation.

Plants also tend to produce more sterile stems per plant as compared to other P. pilosa

accessions.

Phlox pilosa ssp. deamii PZ12- 054 is perhaps is the most interesting population

of this subspecies, and perhaps even the most interesting of all my Phlox collections; it

was found in the Western Highland Rim of Benton County, Tennessee. Unlike the

morphologically uniform populations of P. pilosa ssp. deamii in Indiana and Kentucky,

this morphologically diverse population appears to have arisen via hybridization and

introgression between P. amoena and P. pilosa ssp. pilosa (Figure 1.16). The plants of

“P. pilosa ssp. deamii” occur in the contact zone of discreet populations of P. amoena

and P. pilosa ssp. pilosa along a sloping roadside where both species are abundant.

Although these species occupy distinctly different habitats in this region, anthropogenic

disturbance from roadway and power line construction resulted in the clearing of a

marginal habitat, which both species could inhabit (Figure 1.16). In populations of P.

50

pilosa ssp. deamii to the north, typical P. amoena and P. pilosa are not present. This

raises a question about the relationship of this population to other populations of P. pilosa

ssp. deamii.

In his description of Phlox pilosa ssp. deamii, Levin (1966) designated a

specimen from Benton County, Tennessee as the paratype of Phlox pilosa ssp. deamii

based on a 1948 collection by A.J. Sharp (PH 00075931); this was also considered the

southernmost occurrence of the taxon. Prior to his description of P. pilosa ssp. deamii,

Levin described naturally occurring populations of P. amoena x P. pilosa ssp. pilosa further south of the range of P. pilosa ssp. deamii in Alabama and Tennessee in the

Tennessee River drainage (Levin and Smith, 1966). He mentions that hybrid populations occur along “sloping roadcuts” and that hybrids only form when individuals of P. amoena and P. pilosa ssp. pilosa come into contact and are separated by distances less than 25 m; the population in Benton County matches this decription. Although he states that individuals resembling P. pilosa ssp. deamii and P. amoena ssp. lighthipei occur in

the hybrid swarms, he did not include these populations when he described P. pilosa ssp.

deamii (Levin, 1966; Levin and Smith, 1966). The location reported on the Sharp

specimen is reported as “South of Mt. Moriah, Birdsong Landing”, but Wherry’s

comment “Birdsong Landing shown to the North of Mt. Moriah on the map” suggests

that the original location was incorrectly described on the specimen and that this

population was included in his description of P. pilosa ssp. deamii without visiting the

population.

51

The description suggests that Levin did not actually visit this population, and

based part of his description of P. pilosa ssp. deamii solely on herbarium specimens. My

personal observation of the herbarium specimen from Mt. Moriah revealed that only a

single stem was pressed, and while the specimen is attributable to P. pilosa ssp. deamii,

the single pressed stem does not convey the range of phenotypic variation in the

population. It is likely that the hybrid population I discovered in April 2012 on Birdsong

road to the southwest of Mt. Moriah, is the same population from which this herbarium

specimen was made; it is clearly distinct from P. pilosa ssp. deamii to the North. Since

hybrid populations of P. amoena x P. pilosa ssp. pilosa further south were not included in

the original description of P. pilosa ssp. deamii, it is apparent that Levin (1966b)

considered these populations to be different from P. pilosa ssp. deamii to the north, and

that the population from Mt. Moriah should not be included in the range of that taxon.

This potential reduction in the known range of P. pilosa ssp. deamii further demonstrates

the rarity of this taxon.

Morphological data suggest that P. amoena x P. pilosa hybrid swarms and P.

pilosa ssp. deamii populations to the north of Mt. Moriah are two distinct taxa. It is possible that hybrid swarms to the south are progenitor populations of northern populations, and that fluctuations in the pattern of the Tennessee River, differential adaptability of certain genotypes to soils, different climatic conditions further north may have resulted in a genetic bottleneck in which only certain genotypes could persist in this environment. The relationship of hybrid swarms and P. pilosa ssp. deamii need to be tested in a biogegraphical and phylogenetic context.

52

Other hybrids involving taxa in subsection Divaricatae have been formally

described: P. xglutinosa (P. divaricata x P. pilosa ssp. pilosa) and P. xrugelii (P. amoena x P. divaricata) suggests that populations of P. amoena x P. pilosa ssp. pilosa may warrant formal recognition as a distinct taxon to follow convention in describing naturally occurring Phlox hybrids. However, before such recognition, additional hybrid populations of P. amoena x P. pilosa ssp. pilosa reported from further south in the

Tennessee River drainage in Alabama and Tennessee, as described by Levin (1966b), need to be located and analyzed to determine the extant of similar hybrid populations.

Aside from taxonomic questions concerning this population, there is tremendous potential for selection of unique phenotypic variants for horticultural purposes.

Hybridization and introgression appear to have resulted in novel genetic variation that has resulted in unique phenotypes. So far one potential cultivar has been selected from this population for its unique peach-pink coloration producing an overall salmon like color

(Figure 1.16). The foliage is intermediate between P. amoena and P. pilosa and newly emerging stems are flushed with maroon pigment. The opportunity for further selection from the population is tremendous, and attempts should be made to collect seed from this population.

Phlox pilosa ssp. fulgida (Wherry) Wherry - Dakota downy phlox

Phlox pilosa ssp. fulgida is a diploid (n = 7), perennial species that occurs over a vast area at the edge of the western and northern distribution of the P. pilosa complex. It

53 is restricted to the Great Plains province, and is a major spring-flowering component of the tall grass prairie region (Hendrix and Kyhl, 2000).

The primary distinguishing feature of P. pilosa ssp. fulgida are egladnular, pilose hairs on the stem, leaves, inflorescence architecture, and calyx (Wherry, 1955). Plant habit and flower characteristics are similar to those of P. pilosa ssp. pilosa, but the range of flower color variation appears greater in P. pilosa ssp. fulgida. In one population, white and pink flowered forms with variably color striae were abundant (Figure 1.17).

This taxon is not widely cultivated and Wherry (1955) indicated that it was difficult to cultivate over a long period outside of its native range

Three collections of this taxon were made from natural populations by collaborating botanists (Table 1.6). Phlox pilosa ssp. fulgida PZ10-144 was collected in

Southeastern Minnesota, but the plants perished before further evaluation could be performed. Two individuals of P. pilosa ssp. fulgida PZ11-066 were collected by Daniel

Robarts in Crawford County, Wisconsin and represent the typical form of this species.

The flowers were pink in both collected individuals. Phlox pilosa ssp. fulgida PZ12-092 was collected by Jeffrey Carstens in Story County, Iowa from a population with variably colored flowers (Figure 1.17). Cultivated seedlings from this collection have the same range of flower colors as plants in the wild.

Phlox pilosa ssp. longipilosa (Waterfall) Locklear - Kiowa downy phlox

Phlox pilosa ssp. longipilosa is a diploid (n = 7), perennial taxon that is endemic to southwestern Oklahoma (Locklear, 2009; Locklear, 2011a). It occurs at the western

54

limit of distribution of the P. pilosa complex and in one the most extreme environments;

it grows in a notably xeric habitat on eroded granite slopes in remnants of the Wichita

mountains co-ocurring with Texas live oak (Quercus fusiformis), cacti (Echinocereus

reichenbachii ssp. baileyi), and xeric ferns (Cheilanthes lanosa) (Figure 1.18). Wherry

(1955) included these populations as part of Phlox pilosa ssp. riparia from the Edwards

Plateau of Texas. However P. pilosa ssp. riparia has been elevated to P. villosissima, and P. pilosa ssp. longipilosa is recognized as a morphologically and geographically distinct taxon (Locklear, 2009). This species is now very rare in the wild and is only present in the Quartz Mountain Nature Area (Bruce Hoagland, Pers. Comm., Jan. 2011).

It is a globally rare species G2/S2 (Locklear, 2009).

Phlox pilosa ssp. longipilosa is distinguished from other members of the P. pilosa complex in being copiously hirsute, and both the inflorescence and stem are covered in white, multicellular hairs that are longer than any other member of the P. pilosa complex

(Figure 1.18). This taxon also has a unique geographical distribution and is not sympatric with other phlox taxa. Morphological characteristics of P. pilosa ssp. longipilosa indicate adaptation to extreme drought: The densely pilose vesture, and presence of tap- root (as opposed to a fibrous root system). In addition to potential adaptability, the plants exhibit superior ornamental characteristics when compared to other accessions from the

P. pilosa complex. The comparatively large, round flowers exhibit overlapping corolla lobes and are a distinctively vibrant, bright pink color (Figure 1.18). The taprooted habit of this plant results in a mounded habit that is more uniform than other accessions of P. pilosa. This species also shows fewer tendencies to produce elongated rhizomes unlike

55 some accessions of P. pilosa ssp. pilosa. Overall garden adaptation and hardiness is still being evaluated, but plants have grown well in containers and raised beds at the OPGC and appear to be hardy to at least USDA zone 5. I have used P. pilosa ssp. longipilosa in preliminary interspecific hybridization experiments; it has potential for creating hybrids with larger, more vibrantly colored flowers due to an apparent intensifying factor, and potentially increased adaptation to landscape and garden settings (Chapter 7).

One collection of this taxon (PZSH2011-043) was made from the Quartz

Mountain Nature Park in Greer County, Oklahoma (Table 1.6). Because of the rarity of this species and potential conservation issues, special permission was obtained from the

Oklahoma heritage program state botanist Bruce Hoagland and the staff at the Quartz

Mountain Nature Center. The initial germplasm collection has been used to generate seed that display non-deep physiological dormancy, and requires 6-8 weeks of stratification for germination.

Phlox pilosa L. ssp. ozarkana (Wherry) Wherry - Ozark downy phlox

Phlox pilosa ssp. ozarkana is a diploid (n = 7), perennial species first described by Wherry (1955); it is a distinct taxon of the interior highlands of southern Missouri and northern Arkansas (Figure 1.2). Within the interior highlands, plants can be found in several habitat types, but is most frequent in dry, rocky woods dominated by Quercus and

Carya, and at the edges of Pinus echinata forests at higher elevations (Locklear, 2011a).

This taxon is abundant within its range and shows extensive phenotypic variability in flower color and plant habit (Figure 1.19).

56

Phlox pilosa ssp. ozarkana is distinguished from other taxa in the P. pilosa complex on the basis of large, cordate-ovate bracts of the involucre, and dense, glandular pubescence of the inflorescence, stems, and leaves. Geographical distribution in the

Interior Highlands is also distinct, but it is also reported to occur at the southern edge of the Valley and Ridge province in southern Tennessee, Northwestern Georgia, and

Northeastern Alabama, but this occurrence requires verification (Locklear, 2011;Wherry,

1955).

This taxon offers attributes not found in other members of the P. pilosa complex, and has potential for use in cultivation. A single introduction of P. pilosa ssp. ozarkana called ‘Ozark Rose’ is reportedly a longer-lived garden plant in comparison to other P. pilosa (Oliver, 2011); other observations also support its potential as a garden-worthy plant (Wherry, 1935a; Wherry, 1955). In cultivation at the OPGC, P. pilosa ssp. ozarkana accessions have a mounding habit and abundant flowering that lasts for several weeks. The foliage is larger, denser, and gray green in appearance due to dense pubescence. Flower color is also variable, from white with purple striae to pale pink and lavender (Figure 1.19).

Six collections were made in several locations throughout its range (Table 1.6).

The most noteworthy accession, PZ10-228, was made in Johnson County, Arkansas.

Plants from this population were exceptionally variable in flower color, ranging from pale pink, to white with purple striae (Figure 1.19). One particular clone from this collection is unique from others in this population and the nearby collection site of accession PZ10-227 in having white flowers with contrasting purple striae fused into a

57 star at the center of the corolla. Plants from these populations have smaller, deeper green leaves that can be more purple-flushed than plants from southeastern Missouri.

Two collections, PZ12-058 and PZ12-059, were made in southeastern Missouri

(Figure 1.18). Plants in these populations differ from those collected in Arkansas in having larger, light green, more pubescent leaves, and uniformly pale pink flowers, with a more compact mounding habit. These plants appear to be a slower-growing, perhaps longer-lived form of P. pilosa ssp. ozarkana, and although they lack some of the ornamental characteristics of Arkansas populations, they deserve further trial for general landscape and garden adaptation, and as parents in an interspecific breeding program.

Two additional collections made by an associate in north-central Arkansas are still under evaluation. PZ13-009, has particularly large, ovate leaves, and may represent introgression between P. pilosa ssp. ozarkana and P. divaricata ssp. laphamii.

Phlox pilosa L. ssp. sangamonensis Levin and D.M. Smith - Sangamon River downy phlox

Phlox pilosa ssp. sangamonensis is a diploid (n=7), perennial endemic that occurs within the Sangamon River drainage system in east-central Illinois and was originally described from only Champaign and Piatt counties (Levin and Smith, 1965). It was originally thought to be a hybrid involving P. divaricata ssp. laphamii, P. glaberrima ssp. interior, and P. pilosa ssp. pilosa, but this hypothesis was rejected based on morphological comparisons, and the taxon is now considered a locally divergent subspecies of P. pilosa (Levin and Smith, 1965). This subspecies is reported to grow in

58

prairie remnants, woodland edges, and floodplains associated with the Sangamon River

(Locklear, 2011a). Habitat destruction in combination with a narrow geographic range

have resulted in reduction of population numbers and sizes and it is now considered to be

an endangered species in Illinois (USDA Plants Database, 2014)

Phlox pilosa ssp. sangmonensis can be distinguished from other members of the

P. pilosa complex on the basis of glabrous to glabrescent leaves and unique geographical

distribution (Levin and Smith, 1965). Without knowledge of the collection site,

collections of P. pilosa ssp. sangamonensis could be mistaken for the now synonymized

taxon P. pilosa ssp. detonsa, the glabrous form of P. pilosa ssp. pilosa from the gulf coast

(Wherry, 1955).

This plant was introduced into horticulture by the former Seneca Hill Perennials

nursery of Oswego, New York; the accession was received from Dr. Wesley Whiteside of

Charleston, Illinois, a former botany professor at Eastern Illinois University, who

collected samples from the type location and established it in his private botanical garden

(Wesley Whiteside, Pers. Comm., 1 May 2011). Phlox pilosa ssp. sangmonensis has

persisted in his garden for over thirty years where it has outlasted other forms of P. pilosa

found in Illinois. In these conditions, P. pilosa ssp. sangamonensis has thrived due to

high seedling recruitment. The adaptability of this taxon in the garden is attributed to the

plant’s preference for growing in the richer, alluvial soils of the river, which are more

similar to the soils found in gardens (Wesley Whiteside, Pers. Comm., May 1, 2010)

One accession, PZ10-233, was obtained. Seedling material was donated to the

OPGC from the private botanical garden of Dr. Wesley Whiteside, the same source as the

59

original introduction to horticulture (Table 1.6). Given the endangered status of this

species, it is an ideal candidate for seed collection, so that seeds can be conserved long-

term and may serve as a backup for wild populations.

Phlox pulcherrima (Lundell) Lundell (syn. Phlox pilosa ssp. pulcherrima) - Big Thicket phlox

Phlox pulcherrima is a tetraploid (n = 14), perennial species found in the coastal plain of eastern Texas (Figure 1.2) (Ferguson et al., 1999). Wherry (1955) described this taxon as a subspecies of the P. pilosa complex from eastern Texas, southwestern

Arkansas, and western Louisiana with remote populations in Indiana, Kentucky, and

Tennessee. Populations from the Interior Low Plateau province are now referred to as P.

pilosa ssp. deamii; subsequent work has shown P. pulcherrima to be endemic to Texas,

and morphologically distinct (Ferguson, 1998; Ferguson et al., 1999; Locklear, 2011a). It

was therefore elevated to species status (Ferguson, 1998). This species inhabits xeric to

mesic, open oak (Quercus spp.) and pine (Pinus spp.) dominated forest, and forest edges,

usually on acidic, sandy soils (Figure 1.20). This taxon is rare or absent in cultivation.

Phlox pulcherrima can be distinguished from other members of the Phlox pilosa

complex by the glabrous calyces, inflorescence architecture, stems, and leaves (Locklear,

2011a). It also has broader sepal blades than P. pilosa ssp. pilosa and is superficially similar to P. pilosa ssp. ozarkana and P. villosissima in this character (Wherry, 1955).

This taxon is among the tallest members of subsection Divaricatae and individuals in some populations can reach 60 cm tall. The flowers are a pale lavender-pink to bright pink, and are produced in cultivation at the OPGC during April and May (Figure 1.20).

60

This species exhibits a strong fragrance reminiscent of hyacinth flowers (Hyacinthus spp.) that is distinct from all other forms of the P. pilosa complex except P. villosissima

(see below).

Only a single commercial cultivar P. pulcherrima is known in the trade, however

the species is not generally known in horticulture, but possesses several ornamental

attributes that make it worthy of introduction. It is a striking, large Phlox pilosa-like

plant with large, fragrant flowers. Flower shape is variable and color ranges from pale- lilac to pastel pink (Figure 1.20). Despite its southern origin, this species has proved hardy for two winters at the OPGC, and is worthy of further trialing in northern climates and as a parent in a breeding program to produce sterile hybrids with potentially longer flowering periods

Five collections of this species were made from populations in eastern Texas

(Table 1.6). Two initial collections (TX-12, TX-28) in Newton and Sabine Counties,

Texas perished before they could be evaluated. Subsequent collections were made from

Shelby and Jasper counties in Texas. Two collections from Shelby County (PZSH2011-

033, PZSH2011-034) were phenotypically similar. These plants reach up to 60 cm in

height, and have flowers that are star-shaped and pale lilac in color (Figure 1.20). The flowers of plants from these populations are notably fragrant, and are glabrous, but have fine, ciliate pubescence on the leaf margins.

One collection (PZSH2011-035) was made from Jasper County, Texas, and differed from the Shelby county collections. Plants are of shorter stature (to 30 cm tall), are completely glabrous, and have pastel pink flowers with large, rounded, overlapping

61

corolla lobes. Plants from this collection are not as fragrant as those from Shelby county,

and do not appear to be as persistent in cultivation as other collections.

One cultivar, ‘Slim Jim’, introduced by Plant Delights Nursery, Raleigh, NC

from samples collected in eastern Texas, appears to be this species, but the calyx

characters do not match the description of P. pulcherrima (Wherry, 1955). However, this

genotype is a tetraploid like P. pulcherrima (Chapter 4). Phlox pilosa ssp. pilosa is

known to be sympatric with P. pulcherrima in eastern Texas, and ‘Slim Jim’ may be an

intermediate form.

Phlox villosissima Turner (syn. P. aspera, P. pilosa ssp. latisepala, P. pilosa ssp. riparia) - Comanche phlox

Phlox villosissima is a tetraploid (n=14), perennial species centered on the

Edwards Plateau of south-central Texas, but is also reported to grow in Northeast Texas

in the vicinity of Dallas/Forth Worth (Figure 1.2) (Ferguson et al., 1999; Locklear,

2011a; Turner, 1998). This species grows in mesic to xeric, rocky habitats along rivers,

smaller tributaries, and rock outcroppings (Figure 1.21). Although recognized as two

subspecies by Wherry (1955), P. pilosa ssp. latisepala and P. pilosa ssp. riparia, these

taxa have a history of taxonomic confusion, and current thinking suggests they are

distinct enough to warrant a separate taxon, named P. villosissima (Locklear, 2011a;

Turner, 1998). Details of the taxonomic history can be found in Locklear (2011a) and

Wherry (1955). More recently, P. villosissima has been further split into two subspecific taxa, ssp. villosissima and ssp. riparia by Locklear (2009), however, without further sampling, we cannot differentiate the two taxa based on cultural experiments at the

62

OPGC, and choose to recognize germplasm collections as P. villosissima. Despite differences, the Comanche phlox is one of the most distinctive members of the P. pilosa complex from a botanical and horticultural standpoint (Figure 1.22).

Phlox villosissima can be distinguished from other of the P. pilosa complex by features of the calyx; this taxon has linear sepal blades, obscure costa, and a short awn tip

(Locklear, 2009; Wherry, 1955). Wherry (1955) separated P. pilosa ssp. latisepala and

P. pilosa ssp. latisepala using minor differences in leaf morphology, number of nodes per stem, and pubescence type, but these characters appear less distinctive when plants are grown in cultivation, and some of the differences seen on herbarium specimens and in wild populations may reflect environmental effects, and not heritable differences

(Locklear, 2009; Locklear, 2011a). This taxon can also be distinguished by having a unique geographic range and is not sympatric with other members of the P. pilosa complex. The horticultural distinctiveness of this further distinguishes it from other taxa.

This species is not known to be in cultivation; however, it may have been previously introduced by local nurseries on a regional scale under the name “Phlox pilosa”. This taxon has numerous ornamental characteristics to recommend it for further evaluation. In our evaluations, it is the shortest, most compact member of the P. pilosa complex; flowering plants range from 10-20 cm in height. The plants exhibit a dense, tidy, mounding-spreading habit, and abundant flowering, and thus far have not exhibited any tendency to spread by production of long rhizomes (Figure 1.22). The abundance and density of flower production is similar to that exhibited by P. subulata, rendering it among the best in the P. pilosa complex. The foliage is among the most disease resistant

63 of the P. pilosa complex and is a pleasing and unique blue-gray-green color. This species inhabits thin dry soils, commonly called “caliche” in central Texas and in limited trials appear to be one of the most drought tolerant and persistent of the P. pilosa complex.

Our initial trials indicate it may be one of the most garden-worthy taxa in subsection

Divaricatae.

A total of 5 collection of P. villosissima were made from the Edward’s Plateau portion of its distribution (Table 1.6). Two collections from Texas were similar to each other and typical of the species, but two collections were variable enough to warrant further attention from horticulturists. One collection was obtained from a nursery source from plants collected from a natural plant population in Kerr County, Texas.

The collection PZSH2011-040 from a large roadside population of P. villosissima was discovered growing on top of a limestone ridge near the border of Kerr and Gillespie counties in Texas (Figure 1.21). Most plants in this population displayed typical pink flowers, but several plants were of a deeper richer magenta pink that stood out from the rest of plants. Divisions from these plants were collected and have retained their distinct coloration in cultivation and are the richest most intensively colored plants of all of our collections of this species. This plant reaches only 15 cm in height and has a tidy, mounding spreading habit.

Phlox villosissima PZSH2011-042 was collected from a roadside in San Saba

County, Texas. These represent the northernmost collection of P. villosissima and are the shortest, most compact form of this species. The flowers are round with overlapping corolla lobes that are pale lavender-pink with a white eye; this color combination is

64 distinct from other collections. These plants were collected from a mowed roadside location where they grew with other showy natives species. Despite occasional mowing at the original collection site, the plants have retained their dwarf stature in cultivation. I consider these accessions some of the most promising germplasm I obtained, and several individuals have potential for use in cultivation.

Subsection Paniculatae Wherry

This subsection contains two well-defined species, P. amplifolia and P. paniculata (Wherry, 1933; Wherry, 1955). They differ from all others by having easily visible, impressed (areolate) leaf veins, the largest leaves of all Phlox species, and being completely deciduous (Wherry, 1933). Within this subsection, a total of 124 accessions were obtained (Table 1.4). Twenty-one were collected from natural plant populations, and the remaining 103 were cultivars. Cultivars are described and discussed in detail in other sources (Bendtsen, 2009; Fuchs, 1994).

Phlox amplifolia Britton - broadleaf phlox

Phlox amplifolia is a diploid (n = 7), perennial species that is patchily distributed in the Appalachian Highlands, and Interior Low Plateaus, and limitedly in the Interior

Highlands (Figure 1.2) (Wherry, 1933; Wherry, 1955). It occurs in mesic woodland habitats of upland deciduous forest and is frequently associated with limestone outcroppings and talus slopes (Figure 1.22) (Locklear, 2011a). The species is patchily distributed it its range, and preliminary observation indicates that population are

65

relatively small and fragmented. This species is listed as threatened in Indiana (USDA

Plants Database, 2014).

Phlox amplifolia can be distinguished from P. paniculata by the glandular

pubescence of the inflorescence, glabrous corolla tube, larger leaves, and fewer nodes per

flowering stem (Wherry, 1933; Wherry, 1955). Like P. paniculata, it is distinguished

from other phlox species by its cream-colored pollen. Phlox amplifolia also grows in

distinctly drier, steeply sloping habitats, whereas P. paniculata grows in mesic or hydric

alluvial soils on level ground. Phlox amplifolia and P. paniculata are the only Phlox

species that are truly deciduous; all other taxa are evergreen (Wherry, 1955). Phlox

amplifolia is a tall-growing species, reaching up to 150 cm in height. Flowering occurs during summer (June-October), although flowering can occur until the onset of frost in a given location. Flowers are generally pale pink, but are known to vary in depth of coloration in some populations (Figure 1.22). Seeds are produced over a long period from late summer until late mid-autumn (October) (Figure 1.22).

This species is unknown to cultivation in the United States. Deam (1940) reported that the taxon was easily cultivated, but these assertions have yet to result in widespread cultivation, perhaps owing to the rarity of this species in the wild.

Considering the popularity and large number of cultivars developed in the closely related

P. paniculata, it is surprising that this species has not been utilized more. Considering its

adaptation to drier habitats, P. amplifolia may serve as a source of disease resistance and

impart increased adaptability when crossed with P. paniculata. Given the range of

variation in size, habit, and flower color in P. paniculata, it is possible that such variation

66

could exist in P. amplifolia. This species warrants further testing to determine

widespread adaptability to constructed landscapes.

Five collections of this species were made from populations throughout its natiave

range (Table 1.7). Four cultivars were obtained from a commercial source in The

Netherlands, but they have yet to be confirmed as P. amplifolia. This species was the

target of some regional collection expeditions. The first collection (PZ11-050) was made in extreme eastern Cocke County, Tennessee on the bluffs of the French Broad River.

When observed in early October 2010, plants could be found in flower and with mature seed, suggesting that this species begins flowering during mid-summer (early July) and can continue, given favorable environmental conditions, until the onset of frost (Figure

1.2). The flowers of individuals in this population were a uniform pale-pink, and some plants reached 1.2 m in height. The collections PZ11-022 from Woodford County,

Kentucky, and PZ12-109 from Claiborne County, Tennessee include plants that are all morphologically similar.

PZ12-106 was collected in Campbell County, Tennessee on the Cumberland

Plateau. The characters of the inflorescence suggest this is P. amplifolia, but the foliage is intermediate between P. amplifolia and P. paniculata. A similar collection from

Tucker County, West Virginia, PZ11-010, appears to be morphogically intermediate as well. This collection flowers as early as the end of May in cultivation at the OPGC; it is the earliest flowering accession of any collected in subsection Paniculatae.

PZ10- 229 was collected from the western limit of the species in Barry County,

Missouri. This small population was growing in talus at the base of a steep slope in a

67 mesic, forested habitat. Unfortunately these plants perished before further evaluation could take place.

Phlox paniculata L. - summer phlox, border phlox

Phlox paniculata is a primarily diploid (n = 7), perennial species. However, there is evidence of triploid and aneuploid cultivars that suggest ploidy may be more variable than currently known (Chapter 2). This species is distributed in the Appalachian highlands, interior plains, and interior highlands, but is reportedly most abundant on the interior low plateau region of southern Indiana and adjacent glacial till plain (Figure 1.2)

(Wherry, 1933; Wherry 1955). This species occurs in mesic or hydric environments in alluvial soils, but also occurs in disturbed areas such as roadside easements mowed fields

(Figure 1.23) (Locklear, 2011a). Phlox paniculata is the most commercially successful

Phlox species. Hundreds of cultivars have been selected over the past 150 years, and there is continued interest in breeding new forms (Bendtsen, 2009; Fuchs, 1994)

Phlox paniculata can be distinguished from P. amplifolia on the basis of having a pubescent corolla tube, smaller, elliptical leaves, more nodes per flowering stem, and by habitat preference (Wherry, 1933; Wherry, 1955). This species can grow to 150 cm in height and produces flowers from mid-summer until fall in a wide variety of colors, but is typically Phlox purple (Wherry, 1955).

Sixteen accessions of this species were collected from natural plant populations

(Table 1.7). Most collections are phenotypically similar. Observation of this taxon confirmed the previous observations of Wherry (1955) that while phenotypic

68 differentiation occurs within the broad range of P. paniculata, there is nothing that approaches the flower color forms developed by horticulturists (Figure 1.23). Ohio populations are relatively uniform in flower color, generally being the “dingy magenta” as described by one author, or more appropriately “Phlox purple” (Wherry, 1933;

Wherry, 1955). More prominent differences can also be found in the shape and color of the striae, and the amount of color clearing near the base of the corolla lobes (Figure

1.23). Phlox paniculata populations studied across a north south transect of Ohio in

Adams, Delaware, Erie, Franklin, Highland, Hocking, Scioto, and Warren counties have revealed that populations are relatively uniform in color, and vary slightly in height and leaf shape. Observation of the extensive populations of P. paniculata at Clear Creek

Metro Park in Hocking County, Ohio resulted in the discovery of infrequently distributed white flowered clones of P. paniculata. These have also been reported by Wherry (1955) and are known to occur throughout the range of the species. Similar results were found in

P. paniculata populations in southwestern Pennsylvania, West Virginia, and Kentucky.

The most variable population of P. paniculata discovered was PZ11-043, in

Preston County, West Virginia (Figure 1.23). This small roadside population was located near an abandoned homestead and may be an adventive population that arose from a diverse cultivated planting. Flower color variation in this population was greater than any other population (Figure 1.23). Seedlings from this population have produced a similar range of variants.

Other accessions may have originated as collections from adventive populations.

The collections PZ10-109, PZ10-110 collected near Dahlonega, Georgia, by Jim Rodgers

69

of Nearly Native Nursery, Fayetteville, Georgia may fall into this category. These plants

are mildew resistant; PZ10-109 has pink flowers, and PZ10-110 has white flowers. A

similar population was found near a homestead in Avery, County, North Carolina (PZ11-

044).

Subsection Phlox Ferguson

Taxonomy and phylogenetics of subsection Phlox

This subsection is placed in section Phlox with other long-styled (>15 mm),

eastern taxa (Wherry, 1945). It contains 6 species and up to 9 subspecies that are

distributed primarily in the southeastern United States, except for P. idahonis, which is endemic to Idaho (Table 1.1) (Ferguson et al., 1999; Wherry, 1955). The subsection was originally described as subsection Ovatae by Wherry (1955), but later changed to Phlox

(Ferguson, 1998; Ferguson et al., 1999) because P. glaberrima is the type species for the

genus Phlox and the rules of International Code of Botanical Nomenclature (ICBN)

indicate that higher order classification must bear the name Phlox in reference to the original description (Ferguson et al., 1999). Wherry’s monograph still forms the modern concept of species classification in subsection Phlox. Following an “unparalleled” history of taxonomic confusion of all taxa in this group by several authors working from herbarium specimens, Wherry (1945, 1955) reformulated the taxonomy of this subsection based on Linneaus’ original descriptions of P. carolina, P. glaberrima, and P. ovata. He then modified the classification by describing intraspecific taxa that he felt were reflective of geographic patterns of variation he observed during field and herbarium

70

work (Wherry, 1955). Taxa in subsection Phlox are grouped based on common

morphological features: all have long styles (>15 mm), comparatively large leaves with

obscure leaf veins, calyces with sepal blades united for at least half their length, and are

generally taller, glabrous plants. Species are differentiated on the basis of additional

calyx features, leaf shape and size, number of nodes per flowering stem, phenology, and

habitat preference. All taxa are readily differentiated using these characters except for

members of the “Phlox carolina-glaberrima complex” (Wherry, 1955).

Members of the P. carolina-glaberrima complex form a morphologically

variable, widespread group of taxa that are among the most difficult phlox plants to

positively identify (Ferguson and Janssen, 2002; Wherry, 1932a; Wherry, 1932b;

Wherry, 1955). Phlox carolina and P. glaberrima are distinguished from each on the

basis of calyx features: P. glaberrima has sepal blades with prominent costa and flattened

membranes, while P. carolina has obscure costa and plicate, or folded, calyx membranes

(Wherry, 1955). Intraspecific taxa were distinguished using differences in leaf size,

number of nodes per flowering stem, distribution of different sized leaves on flowering

stems, and inflorescence architecture. Using calyx features to distinguish P. carolina and

P. glaberrima is difficult, and essentially unworkable. It is unclear at what stage in flower development Wherry was describing when he noted differences in calyx features; prior to anthesis, the differences are more distinct than at anthesis. At anthesis, the differences in calyx features are not as apparent, as costa can be present in both taxa and the calyx membrane can be plicate or folded in both taxa. Furthermore, the leaf size, leaf arrangement, and number of nodes per stem used by Wherry (1955) to delineate

71

intraspecific taxa do not exhibit broad geographic partitioning, and some populations

share suites of characters to define two or more intraspecific taxa (Locklear, 2011a;

Wherry, 1955). Furthermore, Phlox species can also exhibit phenotypic plasticity in natural populations and in cultivation that may influence the expression of characters used to segregate intraspecific taxa, and can further contribute to the difficulty in identifying them. Despite Wherry’s attempt to resolve this variable and confusing group,

a working taxonomy has yet to be developed, and a well-resolved phylogeny does not

exist. Field observation suggests that slight differences in climatic and edaphic

conditions may result in fine-scale ecotypic differentiation that promote the formation of

narrowly distributed ecotypes of more broadly distributed genetic races. It has also been

suggested that differentiation of taxa in the P. carolina-glaberrima complex may have

occurred in glacial refugia, and as the glaciers receded and taxa began to colonize new

areas, that contact between different genetic races resulted in hybridization that has

obscured species boundaries (Ferguson et al., 1999).

Two molecular phylogenies have been produced for eastern Phlox species. One

study used ITS (internal transcribed spacer) sequence data, and in a subsequent

congruency study, cpDNA restriction site data was added (Ferguson et al., 1999;

Ferguson and Janssen, 2002). Neither study successfully resolved the phylogeny of the

P. carolina-glaberrima complex, or subsection Phlox. In fact, members of subsection

Divaricatae and Phlox formed a paraphyletic group, and neither subsection was

adequately resolved. However, these studies were focused on resolution of subsection

Divaricatae and the P. pilosa complex, and sampling from subsection Phlox was

72 minimal. The taxonomy of the Phlox carolina-glaberrima group would benefit from fine-scale field sampling across the range of the species including all proposed subspecies. Phylogenetic analysis using fast-evolving, low-or-single copy nuclear- encoded genes, whole chloroplast genome sequencing, or next-generation sequencing

(NGS) may provide resolution of these species that may result in a well-resolved phylogeny and useful taxonomy (Small et al., 1998; Shaw et al., 2005; Zimmer and Wen,

2012).

Difficulty in identifying some accessions collected during the course of field work has resulted in my adaptation of a modified taxonomy of this group. Locklear (2011a) proposed a generalized classification of these taxa, but his classification may be over generalized, and does not accurately described the variation of intraspecific taxa. Within

P. carolina, I recognize P. carolina ssp. carolina and P. carolina ssp. alta. I have not made any collections that are referable to P. carolina ssp. turritella, so this taxon is only mentioned in reference (Wherry, 1955). Plants have been collected that are referable to

P. carolina ssp. angusta, however, the similarities between this and narrow-leaved forms of P. glaberrima ssp. glaberrima from the coastal plain render these taxa impossible to decipher from each other based on available botanical keys (see below). Based on my field work, and subsequent cultivation at the OPGC, Phlox glaberrima ssp. glaberrima,

P. glaberrima ssp. interior, and P. glaberrima ssp. triflora are recognized for the purposes of this study.

Given the extensive phenotypic and ecotypic diversification of populations at the local level, it is likely that this complex could be composed of several species. Accurate

73 resolution will require extensive and detailed sampling, as germplasm collection expeditions revealed that all taxa have sympatric natural distributions, and as many as 5 taxa can found in a single county in states such as Tennessee and North Carolina. The phenotypic diversity found within the subsection is also of great interest, and additional germplasm accessions could facilitate the discovery of novel populations and individuals that may be used in breeding programs. The need for a well-resolved taxonomy is underpinned by the rarity of some taxa and the need to establish appropriate conservation guidelines for taxa that may be exceptionally rare in their native habitat.

Horticultural Taxonomy of Subsection Phlox

Formulation of horticultural groups of taxa in subsection Phlox in cultivation can help distinguish accessions when botanical characteristics do not clarify the identification of a particular collection. These groupings are intended as a preliminary method for organizing these taxa based on the collections made and evaluated during the course of this study. Given the complex variation exhibited in the group, the classification will likely need to be revised as additional collections are made.

From a horticultural standpoint, species can be classified into three broad groups based on phenology, superficial morphological resemblance, and horticultural usage.

Horticultural Group 1 contains members of the P. carolina ssp. alta, P. carolina ssp. angusta, some accessions of P. carolina ssp. carolina, P. glaberrima ssp. glaberrima, P. glaberrima ssp. interior, and P. maculata; these taxa have upright stems that can reach from 45 to 100 cm in height, have stems with many nodes, a large, compound

74

inflorescence, and flowers produced over an extended period starting in mid-summer and

often lasting until fall. These are the largest plants of the three groups. The leaves may

be broad lanceolate-ovate or narrow linear. Sometimes upright types have a few winter

persistent, decumbent, sterile stems, but these can be absent and are, if present, generally

few in number. In nature, these species mostly occur in naturally hydric and mesic sites,

such as fens, wet meadows, seepages, river scour habitat, riparian areas, and canopy gaps

in deciduous forest. In cultivation, these species have been used in rain gardens, and in

others areas where they are not subject to seasonal drought. Based on tested accessions, these taxa are all diploid (n=7), and show variable susceptibility to powdery mildew,

however some accessions appear to exhibit exceptional resistance or tolerance to the

disease (Chapter 2).

The second Horticultural Group consists of taxa that can be identified as P.

carolina ssp. carolina and includes cultivars such as ‘Kim’ (PZ11-017), and ‘Minnie

Pearl’ (PZ11-010). The primary differentiating factors of this group are that plants are

shorter in stature and begin to flower earlier, but still have the potential to re-flower

throughout the growing season until growth ceases due to onset of cold weather. These

taxa can produce large numbers of sterile stems; and although all may be regarded as

evergreen, the display is of variable or uncertain ornamental interest. Some collections

exhibit resistance to powdery mildew, and the cultivar ‘Minnie Pearl’ is widely regarded

as mildew resistant. Opportunity exists to select equally superior clones from nearly all

taxa in this group. In the wild, these species occur in mesic to hydric sites, and are

75

adaptable to a wide variety of climatic conditions and soil types. All taxa in this group

are diploid (n=7) (Chapter 2).

The third group consists of the morphologically distinctive P. ovata, P. pulchra,

and the poorly understood P. glaberrima ssp. triflora (syn. P. triflora). These species

exhibit a distinct mounding habit, a discreet mid-to-late spring phenology (May-

June/July). These species produce copious evergreen, sterile stems after flowering, and

taxa such as P. ovata and P. glaberrima ssp. triflora exhibit noteworthy evergreen foliage during the winter months. The determinate inflorescence is generally fewer flowered than taxa in preceding groups, and seed ripens within a narrow window of time. These taxa grow in the driest habitats of any in subsection Phlox. Although further testing of P. ovata and P. pulchra is required, several accessions of P. glaberrima ssp. triflora have been persistent, adaptable plants in a variety of local landscape conditions, and are among the best phlox specimens for diverse landscape usages. This group also contains the only known tetraploids described from subsection Phlox; tetraploid accession of P. glaberrima

ssp. triflora and P. pulchra were discovered during the course of this study (Chapter 2).

Phlox carolina L. ssp. carolina - Carolina or thick-leaved phlox

Phlox carolina is a diploid (n = 7) species distributed in the southeastern United

States in the Appalachian Highlands, Valley and Ridge province, Coastal Plain, and

Interior Low Plateaus (Figure 1.2) (Locklear, 2011a; Wherry, 1945; Wherry, 1955). It

occurs in a variety of ecosystems and soil types and tends to grow in mesic or hydric

sites. This taxon is widespread and ranges from North Carolina to Texas, although the

76

exact distribution may be unknown due to the uncertain botanical history. The taxon

exhibits extensive phenotypic variation in flowering time, habit, and foliage length and

width and may be easily confused with P. glaberrima. Despite the difficulty in

distinguishing between P. carolina and P. glaberrima, all are highly ornamental plants,

and the inability to accurately identify accessions from this group should not inhibit

further collection and evaluation.

For this study, the concept of P. carolina proposed by Locklear (2011a) is used.

He defines these plants as having narrow linear leaves near the base of flowering stems

that develop into broadly linear, linear-ovate, or elliptical leaves towards the middle and

distal portions of the stem. Phlox glaberrima is recognized as having linear leaves that

may be broadly lanceolate or ovate-lanceolate on the upper portion of the stem. He

further uses internode length to distinguish between the nominate expression of P.

carolina and P. carolina ssp. angusta; the latter supposedly having longer nodes, but this

character is highly variable in both taxa, and appears to exhibit plasticity in cultivated

settings and different environmental conditions. The primary differentiating factor

between these taxa is leaf length and width.

Phlox carolina has been an important plant in the history of Phlox cultivation and

has been implicated as a parent in interspecific hybridization events with P. paniculata.

However, relatively few cultivars of P. carolina have been selected or remain in cultivation, but the number may be inaccurate due to the historical lack of taxonomic consistency in this group (Bendtsen, 2009; Fuchs, 1994). Taxa in this group have noteworthy ornamental characteristics that make them exemplary choices for a variety of

77 uses in constructed landscapes and restoration projects. Furthermore, the local adaptation of certain populations to different soil types suggests that selection based on habitat and soil preference may result in genotypes with increased adaptability to a variety of different constructed landscape conditions. Phlox carolina can reach up to 50 cm in height, but can be taller or shorter, have an erect habit, and glabrous to lustrous, comparatively broad upper leaves. Flowering may start as early as May, but can begin later, and plants exhibit the ability to re-flower in cultivation and in the wild. Flowers are generally pink, but white flowered mutants can occur (‘Minnie Pearl’, see below). The full range of flower color variation for this variable taxon is not known. The cultivar

‘Minnie Pearl’ has exceptional powdery mildew resistance, and suggests that other collections may exhibit similar resistance.

Two wild collections of this species were obtained for this study. Collaborators collected both accessions (Table 1.8). The collection PZ11-036, made by Julian

Campbell in Clay County, Mississippi, most closely matches the original description of

P. carolina. The lower leaves are linear, and gradually widen. The leaves subtending the involucre are cordate-ovate in shape, bearing an acuminate tip. Most members of the subsection have glabrous leaves, but this collection consisted of two individuals, one with typical glabrous leaves, and one with pubescent leaves. These plants have the potential to repeat flower if environmental conditions are optimal. The collection PZ10-114 was made on Currahee Mountain, by Jim Rodgers at Nearly Native Nursery, Fayetteville,

Georgia, but the plants perished before further assessment and identification could take place.

78

The cultivars ‘Kim’ (PZ11-017) and ‘Minnie Pearl’ (PZ11-010) can be assigned to P. carolina ssp. carolina based on leaf shape and width. Both of these cultivars originated as selected clones from natural populations. Phlox carolina ‘Minnie Pearl’ was collected from a roadside in Kemper County, Mississipi, and introduced by Tony

Avent of Plant Delights Nursery, Raleigh, North Carolina (Plant Delights Nursery website, 2014). This selection is widely cultivated for its long flowering season and resistance to powdery mildew. Phox carolina ‘Kim’ was selected by Alabama native plant enthusiast, Jan Midgely, from a Tuscaloosa County, Alabama population (Niche

Gardens website, 2014). This selection has pastel pink flowers, a compact habit, linear leaves, and the ability to flower throughout the summer.

Phlox carolina ssp. angusta Wherry / Narrow leaved variants of P. glaberrima ssp. glaberrima Wherry

This is a highly variable group of diploid (n = 7) taxa that occur in a wide range of ecosystems over a broad geographic area, but generally occur in hydric sites within the

Appalachian Plateaus, Coastal Plain, and Interior Low Plateaus (Figure 1.2) (Ferguson and Janssen, 2002; Wherry, 1945; Wherry, 1955). These taxa are perhaps the most confused of the P. carolina-glaberrima complex, and an objective method of distinguishing them has yet to be formulated on the basis of morphology. Collectively, they represent a group of narrow leaved taxa with leaves of varying length; Wherry

(1945, 1955) attempted to use the relative placement of different sized leaves along flowering stems to delineate variation in this group, but this treatment is unworkable.

79

This character appears to be highly plastic. Cultivated plants can exhibit many more

nodes per stem, and have leaf arrangements that do not correlate to published descriptions

of these taxa (Wherry, 1955). The difficulty in distinguishing these taxa is highlighted

by the differential usage of these names by herbaria and published sources to refer to the

same populations of a given collection. Further analysis using molecular markers is

needed to objectively differentiate the taxa within this group.

A total of 6 collections were keyed to this group (Table 1.8). This grouping could

change pending further taxonomic revision of the genus.

Despite the difficulty in distinguishing these, they represent a highly ornamental

group of taxa that merit further collection and evaluation. This taxon, under the name P.

carolina ssp. angusta ‘Gypsy Love’ was introduced into cultivation by North Creek

Nursery (Landenberg, PA). The accession PZ10-034 was obtained from Growild

Nursery in Fairview, Tennessee. They obtained plants from North Creek, and it appears that they are the cultivar ‘Gypsy Love’. These form large plants in cultivation to 0.90 m tall, have deep green, linear leaves, and a large compound inflorescence composed of numerous pink flowers.

Accession PZ11-067 was collected from Bibb County, Alabama near the Kathy

Stiles Freeman Nature preserve. This taxon was originally identified as P. carolina ssp.

angusta, and a commercially available cultivar called ‘Little Cahaba’ was collected from

a similar site and is sold under the name P. carolina ssp. angusta (Plant Delights Nursery,

2014). However, two separate references refer to similar plants from a similar site in

Bibb County as P. glaberrima ssp. glaberrima (Alabama Plant Atlas, 2014; Locklear,

80

2011a). Phlox PZ11-067 is one of the most distinctive accessions made from subsection

Phlox. Individuals have a mounding habit, bright green linear leaves, and a compound

inflorescence held on a long pedicel, well above the foliage. This accession was closely

associated with the Ketona dolomite found in Bibb County. Despite the southern origin,

this accession is winter hardy in Columbus, OH.

Accession PZ12-093 was collected from river scour habitat in Cumberland

County, Tennessee by Kentucky botanist Julian Campbell and identified by him. This

collection exhibits similar characteristics to the preceding, but was collected from a

unique and different habitat. It has a distinctly more upright habit, and the inflorescence

is held on a short scape. This accession is also described as P. glaberrima ssp.

glaberrima by Locklear, and further highlights the confusing nature of these collections

(2011).

Accession PZ12-047 was collected from seasonally wet, oak (Quercus spp.) flat

woods in Marshall County, KY. The associated flora suggested that soils in the area are

acidic. These plants were originally included in P. glaberrima ssp. interior, but are

morphologically more similar to the previous collections than to the latter. These plants

are still under evaluation. Further collection of similar taxa and evaluation with appropriate molecular markers and common garden studies are needed to assess the phylogenetic relationships of these taxa.

81

Phlox carolina ssp. alta Wherry

Phlox carolina ssp. alta is a diploid (n = 7), perennial species from the

southeastern United States that occurs in the Appalachian highlands region; it grows at

higher elevations than most eastern Phlox taxa (Figures 1.2, 1.24). The taxon was originally described as P. carolina ssp. altissima, and later renamed P. carolina ssp. alta

(Wherry, 1945; Wherry, 1955). While the taxon may not warrant taxonomic recognition, it is horticulturally distinct in being of greater stature than P. carolina ssp. carolina, having more nodes per stem, and a larger, compound inflorescence that is capable of producing flowers over a long season. Both taxa have the broadly elliptical, upper leaves characteristic of P. carolina (Locklear, 2011a). Wherry (1945) also noted that the

presence of dark purple color near the nodes of specimens belonging to this taxon, but

field observation indicates the degree of stem coloration is variable among taxa in

subsection Phlox and of little diagnostic value.

This taxon is not known to be in cultivation, but exhibits many ornamental

characteristics. Four collections were made; PZ11-045 and PZ11-072 were collected in

western North Carolina, and PZ11-048 and PZ11-049 were both collected from different

sites in Union County, Georgia (Table 1.8). All collections are phenotypically similar.

The plants are tall (up to 1 m), erect, bearing a large, conical inflorescence beginning in

July that remains in bloom over a long season. At one population (PZ11-045) flowering plants were still evident on September 16, 2011 (Figure 1.24). Flower color ranges from dull to bright pink, with individual striae, or striae fused into a star-shaped pattern.

82

Foliage color is a distinctive grey-green that is different from other collections in this subsection.

Phlox carolina ssp. turritella Wherry

This obscure taxon was described on the basis of the inflorescence shape

(Wherry, 1945; Wherry, 1955); it has been synonymized with P. carolina ssp. carolina by Locklear (2011a), although it was included in previous phylogenetic studies (Ferguson et al., 1999; Ferguson and Janssen, 2002: Locklear, 2011a). None of the Phlox germplasm collections can be keyed to this species. The status of this taxon needs to be reinvestigated.

Phlox glaberrima L. ssp. glaberrima Wherry - smooth phlox

Phlox glaberrima is a diploid (n = 7), perennial species and the type species for the genus Phlox (Wherry, 1955). Originally, Wherry (1932a) described it as occurring throughout the Coastal Plain, Interior Low Plateaus, and Piedmont, but it was later reported only from the Piedmont (Figure 1.2) (Wherry, 1932a; Wherry, 1955). Further confusion has ensued, as Locklear (2011a) combined P. glaberrima and P. glaberrima ssp. triflora, so the exact geographic range of this species remains unknown and unverified without further collection. The phylogenetic relationship with P. carolina needs to be clarified (Ferguson and Janssen, 2002).

Phlox glaberrima ssp. glaberrima is differentiated from P. carolina primarily on the basis of leaf and calyx shape (Locklear, 2011a; Wherry, 1955). Phlox glaberrima

83

ssp. glaberrima has leaves that are linear on the lower part of the stem that gradually

become broader, and lanceolate or ovate on the upper portion of a flowering stem.

Wherry (1955) also reports calyx differences that distinguish the two taxa, however this

characteristic is of little diagnostic value (described above).

One wild collected accession included here is Phlox PZ12-131 collected from

Polk County, Tennessee in the Valley and Ridge province (Table 1.8). This collection bears comparatively few nodes per stem, has ovate upper leaves, and compound inflorescence. It differs from P. carolina ssp. alta, which is sympatric in the region, in the number of nodes per flowering stem, but differs from other accessions of P. carolina as described above.

Two cultivars are included here: P. glaberrima ‘N3 Springfall’ (PZ10-103) and P. glaberrima ‘N3 Hvtke Lemke’ (PZ10-241). Both cultivars originated from collections

made by Jim Rodgers of Nearly Native Nursery, Fayetteville, Georgia and ascribed by

him to P. glaberrima. Phlox glaberrima ‘N3 Springfall’ was collected from a natural plat

population in Georgia and was selected for its long flowering period from May until frost. This selection produces many flowering stems throughout the season. The flowers are pink with red striae fused into a star. The lower foliage is linear, and upper foliage narrow lanceolate. Phlox glaberrima ‘N3 Hvtke Lemke’ was selected from a natural

population, is unique in having pure white flowers without striae. The foliage of this

plant is ovate to elliptical. Phenotypic variation seen in these 2 cultivar selections serves

to highlight the variation within this taxon and the need for further germplasm

collections.

84

Phlox glaberrima L. ssp. interior Wherry - Wabash smooth phlox

Phlox glaberrima ssp. interior is a morphologically distinct, diploid (n = 7)

perennial taxon indigenous to calcareous limestone seeps and hydric environments in tall

grass prairies of the Great Plains province and Interior Low Plateau province of the

central United States (Figure 1.2) (Locklear, 2011a; Wherry, 1955). The rich soils

associated with these habitats typically support an abundance of sympatric species

(Figure 1.25). This species is common throughout its range, but has not been widely

evaluated for horticultural purposes, although Wherry (1955) reported that is was a

relatively uniform taxon.

Phlox glaberrima ssp. interior can be distinguished from other subspecies in the

P. carolina-glaberrima complex on the basis of the narrow, linear leaves which remain equally sized and shaped on all portions of the stem, having more nodes per flowering stem than other members of this complex, by geographical distribution, and preference for calcareous soils (Locklear, 2011a; Wherry, 1955). Geographically, it is the northernmost, most interior taxon of the P. carolina-glaberrima complex, suggesting that they may be more cold hardy than populations and subspecies to the south. This species ranges from 0.5-1.0 m tall, with an erect, clumping habit, and flowers from June until frost in favorable conditions (Figure 1.25). Flower color is pink, and relatively invariant, without prominent striae, although Wherry (1955) indicates that rare, white-flowered individuals may occur in natural populations, and there can be variation in the presence and patterning of the striae (Figure 1.25). This species produces a large, compound

85

inflorescence and is capable of flowering over a long period with repeat flowering. It may

exhibit greater resistance to powdery mildew than P. paniculata, and suggests that further

collection and evaluation of different collections is warranted. Its large size may limit

use in smaller gardens, but the plant has potential for perennial borders, in rain gardens,

habitat restoration, and in breeding phlox for cut flower use. This taxon can be placed in

the previously described Horticultural Group 1.

Phlox glaberrima ssp. interior is common in the Midwestern U.S. and has been introduced into cultivation by at least one source (Prairie Moon Nursery, Winona,

Minnesota), but it is not common in horticulture (Table 1.8). Unlike other members of the P. carolina-glaberrima complex, wild collected and cultivated selections are morphologically uniform and preliminary assessment of collections at the OPGC

indicated that populations are relatively uniform. Sampling of additional populations of

this taxon from the central portion of the distribution in Illinois and Indiana, and from the

western edge of the species range in Missouri and Arkansas should be included among

future Phlox germplasm collection priortities.

Two collections were made from natural plant populations (Table 1.6). The

accession PZ12-055 was collected at the southeastern range limit of this species and was

found in a botanically diverse limestone seepage area in Lewis County, Tennessee. In

this habitat, plants occur as scattered, single stemmed individuals, but in cultivation form

dense, rhizomatous clumps that may be become quite large. The flowers are typical of

the species.

86

The second was PZ12-063 was collected from the eastern edge of the species range of in White County, Indiana. Habit and flower characteristics are similar to the previously mentioned collection, however the plants are still under evaluation.

Phlox glaberrima ssp. triflora Wherry (syn. P. triflora Michaux) – three flower smooth phlox

Phlox glaberrima ssp. triflora is a diploid (n = 7) or tetraploid (n = 14) species that occurs in the Allegheny Plateaus, Appalachian Highlands, and Interior Low Plateaus, and Valley and Ridge province (Figure 1.2). This taxon occurs in more xeric habitats than other member s of the P. carolina-glaberrima complex, and can be found in a variety of habitats ranging from limestone barrens, rocky woods, and forest gaps in mixed mesophytic hardwood forest (Locklear, 2011a; Wherry, 1955). This taxon was originally described by Michaux (1803) as P. triflora, then transferred to other taxa and renamed by subsequent authors. It was first recognized by Wherry (1945) as a subspecies of P. carolina, and then finally assigned by him as a subspecies of P. glaberrima (1955) on the basis of calyx morphology. More recently Locklear (2011a) lumped this species into synonymy with P. glaberrima ssp. glaberrima. This subspecies is considered distinct and treated here following Wherry (1955).

Observation of this taxon at field sites and in cultivation at OPGC from across an east-west transect of the range of its geographic distribution suggests that this is one of the most distinct and easily identified members of the P. carolina-glaberrima complex on the basis of plant habit, production of sterile stems, phenology, and ovate-oblong leaves.

87

This taxon appears to be more closely related to P. ovata and P. pulchra than to taxa in the P. carolina-glaberrima complex (Campbell, 2012; Wherry. 1955).

Phlox glaberrima ssp. triflora exhibits several ornamental characters and adaptability; at least four cultivars introduced as P. glaberrima are attributable to this taxon (see below). This species belongs to Horticulture Group 3; it has a mounding habit to 50 cm tall, a discreet May-June flowering period, production of numerous sterile stems that provide winter interest, and tends to occur in more xeric habitats than species in

Horticulture Groups 1 and 2.

Four accessions were collected from natural plant populations, and four cultivars were obtained (Table 1.8). All accessions are phenotypically similar; however, there are some differences among populations in plant height, phenology, and in leaf shape, size, and coloration. Such diversity increases the potential for selection. There were also ploidy differences between a population from the Valley and Ridge province (n = 14) and from the Allegheny Plateau and Interior Low Plateaus (n = 7) (Chapter 2). All are highly ornamental plants.

PZ10-193 was collected in Scioto County, Ohio at the northern edge of the species natural distribution (Cooperrider, 1986). Populations in the region are morphologically uniform. Accession PZ11-027 was collected in Bullitt County,

Kentucky from rocky areas in mesic hardwood forest. Numerous individuals of this accession have characteristically purple-flushed, newly emerging foliage that has ornamental potential. Only one other collection, the single individual of PZ12-135 from

Hawkins County, Tennessee shares this characteristic. Accession PZ12-046 was made in

88

Lyon County, Kentucky, where it was initially identified as P. glaberrima ssp. interior; this represents a taller form of P. glaberrima ssp. triflora with pale pink flowers (Figure

1.26). In the wild, these taxa have few nodes per flowering stem, but in cultivation there are many more nodes per flowering stem, but the plants from all collections maintain a comparatively short, mounding habit.

Four accessions were obtained from commercial sources (Table 1.6). Phlox glaberrima ssp. triflora PZ11-019 was collected in Bath County, Virginia by nurseryman

Don Hackenberry (Oliver, 2011). Two cultivars, ‘Anita Kistler’ and ‘Bill Baker’ appear to have been collected from the same population, and they may actually be the same plant. The popular cultivar ‘Morris Berd’ is morphologically similar to the previous accessions may have also originated from the same region. The aforementioned taxa are all tetraploid (n=14). One of the few variegated phloxes, P. glaberrima ssp. triflora

‘Triple Play’ PP 21,329 is a chimera selected for leaves bearing an irregular, creamy white margin, however is has shown a tendency to revert to non-variegated leaves.

Reversion can have normal green foliage, or produce at least one other variegate. The accession PZ10-233 was obtained from the Flower Factory, Wisconsin under the name P. buckleyi. The distinctive morphological characteristics indicate that it is P. glaberrima ssp. triflora. This form produces abundant, sterile, leafy shoots that provide winter interest.

89

Phlox maculata L. – meadow phlox

Phlox maculata is a diploid (n = 7), perennial species that occurs over a wide range; it has the northernmost distribution of taxa in subsection Phlox (Locklear, 2011a;

Wherry, 1932b; Wherry, 1955). The species is primarily found in the Great Plains province spanning from Southeastern Minnesota to western Vermont, and southward to

Northern Georgia in the Allegheny Plateaus, Interior Low Plateaus, and Valley and Ridge province (Figure 1.2). Wherry (1932b, 1955) recognized two subspecies, P. maculata ssp. maculata and P. maculata ssp. pyramidalis, which were distinguished on the basis of inflorescence shape, number of nodes per flowering stem, and geographical distribution, but variation in morphological differences between populations of both taxa is minimal and here they are considered synonymous. Phlox maculata is a variable species that occurs in a wide variety of habitats, which led Wherry (1955) to describe two intraspecific taxa. Phlox maculata ssp. maculata occurs primarily at the northern edge of the species range and generally inhabits hydric soils of fens, seepages, and sedge meadows. Phlox maculata ssp. pyramidalis grows in the southern part of the species range and tends to inhabit mesic or hydric sedge meadows, riparian zones, seepages, and forest gaps. All forms of P. maculata have an affinity for hydric soils (Figure 1.27).

Phlox maculata can be distinguished from other members of subsection Phlox by the typically red-maculate stems (although this character can sometimes be found in members of the P. carolina-glaberrima complexes), lanceolate to oblong-ovate upper leaves, and cylindric to pyramidal inflorescence (Locklear, 2011a; Wherry, 1955). Sepal and calyx characters do not provide an accurate method of identifying this species, and

90

are similar to members of the P. carolina-glaberrima complex. The leaves are typically

glabrous to lustrous, and rarely pubescent (Wherry, 1955). The species typically begins

to flower in June, and recurrent flowering can occur if environmental conditions are

suitable. Flowers are typically pink, but rare white-flowered individuals can occur, and at

least one cultivar, ‘Flower Power,’ has been selected for its white flowers (Figure 1.27).

Plants range in height from 35-125 cm, but this can be variable in the wild due to

phenotypic plasticity and environmental factors.

Phlox maculata is well known in cultivation and several cultivars have been selected. This taxon is also one of the parents of the Phlox Suffruticosa Group (syn. P.

xdecussata). The best-known cultivar is ‘Natascha’ which was discovered at The

Central Botanical Garden of the National Academy of Sciences Belarus, in Minsk by Luc

Klinkhamer, and named for the curator of collections (Luc Klinkhamer, Personal

Communication, 14 May 2014). This is one of the few Phlox cultivars that exhibited bicolored corolla lobes. Thorough reviews of P. maculata cultivars and hybrids can be found in Bendtsen (2009), Fuchs (1994), and Symons-Jeune (1953). It appears that cultivars have been selected from both of the former subspecies. Phlox maculata is placed in Horticulture Group 1 based on the tall, upright stems, recurrent flowering beginning in late spring, and preference for wet soils. Three commercial cultivars were obtained for subsequent comparative analysis: ‘Natascha’ (PZ10-035), ‘Flower Power’

(PZ10-235), and ‘Omega’ (PZ10-051).

Eight accessions were collected from natural populations, primarily within the southern part of the species range (Table 1.8). PZ10-198 was collected from a calcareous

91 fen at Cedar Bog in Champaign County, Ohio. Of all P. maculata accessions, this one most closely matches the description of P. maculata ssp. maculata; the plants bear a long, narrowly cylindric inflorescence, are relatively uniform, peoduce narrow linear to oblong leaves, and occur within the Great Plains province. In cultivation, it forms large clumps, as opposed to the single-stemmed flowering individuals found in the wild. Plants exhibit a tendency to produce long rhizomes and send up stems up to 1 m or more away from the original planting.

PZ10-208, PZ10-209, PZ13-004 were collected from Adams and Scioto Counties in Southern Ohio. These populations grow at the western edge of the Allegheny plateau, and are keyed to P. maculata ssp. pyramidalis. All accessions are similar and exhibit longer, broader leaves, a broader inflorescence with a pyramidal shape, and grow primarily along roadsides, in wet disturbed sites, or in forest gaps. One individual from the collection PZ10-208 exhibited somewhat scalloped leaves when grown in cultivation.

This phenomenon has not been observed in other accession; F1 hybrids using this specimen as a parent have inherited this leaf character.

PZ12-103 and PZ12-104 were collected in Campbell County, Tennessee from a unique habitat that differs from all other P. maculata habitat types observed. They were found in areas with wet, sandy soils and the associated flora suggests that the soil at these sites was acidic. These were tall plants to 125 cm tall, with maculate stems, and broad ovate-oblong leaves and coriaceous texture. The leaf shape and texture was more similar that of taxa in the P. carolina-glaberrima complex and the relationship of these accessions with those taxa needs to be assessed.

92

PZ12-105 was collected in Campbell County, Tennessee; it occurred in the

riparian zone of a small stream, in similar habitat to the accessions PZ10-208 and PZ10-

209 in southern Ohio (see above). The morphology of this collection is comparable to

Ohio collections, but different from the previously mentioned accessions from Campbell

County. PZ12-107 and PZ12-108 were collected from limestone seepages in Campbell

County, Tennessee. Despite the habitat differences compared to other collections in this

county, these collections are morphologically similar to collection from southern Ohio

and to PZ12-105.

Phlox ovata (L.) Locklear (Syn. P. latifolia) - mountain phlox

Phlox ovata is a diploid (n = 7), perennial species distributed primarily in the

central and southern Valley and Ridge province, but also recorded as occurring in the

Piedmont, with an outlying, disjunct occurrence in the Interior Plains, where it is found in

sand prairie and oak savannas in the Oak Openings region of northwest Ohio and

southeastern Michigan (Figure 1.2) (Cooperrider, 1986; Wherry, 1955). Phlox ovata is

associated with xeric environments were there is natural or anthropogenic disturbance in

the form of fire or infrequent mowing, and occurs in the driest habitat of any taxa in

subsection Phlox (Figure 1.28) (Locklear, 2011a).

Phlox ovata was until recently known as P. latifolia and several references still refer to it by that name (Locklear, 2011b). The species was originally named P. ovata L., but Reveal et al., (1982) found that the lectotype was actually based on a specimen of

Ruellia caroliniensis, and argued that nomenclatural preference be given to P. latifolia

93

(Michaux, 1803). Through detailed nomenclatural research, Locklear (2011b) ultimately

resolved the issue and concluded that P. ovata was a valid name. While the resurrection

of P. ovata has been generally accepted, the name change is being slowly integrated back

into popular use.

I place P. ovata in subsection Phlox Horticultural Group 3. It is readily distinguished from other members of subsection Phlox by the relatively large, oblong,

and long-petioled rosette leaves that are among the largest in the genus. Also, there are

comparatively few nodes per flowering stem, a shorter height at anthesis, mid to late

spring (May-June) phenology, and early maturation of seeds in mid-summer (July),

which ripen as species in Horticultural Group 1 are beginning to flower (Figure 1.28).

Morphologically, P. ovata is similar to P. glaberrima ssp. triflora and P. pulchra (which

used to be considered a southern form of P. ovata). All of these species grow in mesic or

xeric habitats that are drier than those of taller species of subsection Phlox, they have

prolific production of evergreen, sterile stems, a compact, mounding habit with relatively

few nodes per flowering stem at anthesis, mid to late spring (May-June) phenology, and

similar timing of seed maturation in mid summer (July) (Table 1.3). Despite similarities,

the distinctive long-petioled rosette leaves of P. ovata distinguish this species from all

others.

Phlox ovata has never been a common garden plant, but in my opinion exhibits

enormous ornamental potential. The long-petioled, glossy green basal leaves have led

one breeder to describe P. ovata as the “Hosta” of the Phlox species, and suggests that

this feature alone warrants further cultivation and evaluation (Figure 1.28) (Oliver, 2011).

94

The species also has a shorter height (to 45 cm), and can have a more compact mounding

habit than related species. The flowers are large, and variably shaped, but many forms

have overlapping corolla lobes that result in desirable, rounded flowers. Perhaps most

importantly, P. ovata appears not to be affected by powdery mildew (Erysiphe

cichoracearum) that can be problematic on related species. Initial propagation and

cultivation trials at the OPGC suggest that this species can be cultivated in a variety of

constructed landscape settings and that clonal selection and hybridization could result in

cultivars that are more readily adapted to landscape conditions.

Prior to this study, no interspecific hybrids have been reported involving P. ovata.

Karyotypic and molecular evidence suggests that this species might be more closely

related to P. stolonifera than to some members of subsection Phlox. Interspecific hybridization between P. ovata and other long-styled phlox species deserves experimentation (Ferguson et al., 1999; Smith and Levin, 1967). Among long-styled

Phlox taxa, this species exhibits tremendous potential as a parent in breeding programs and may be suitable for creating, disease-resistant, compact plants with tolerance to adverse environmental conditions and large, brightly colored flowers.

Phlox ovata was a germplasm collection priority; 11 collections of this species were made from throughout the species range in the Valley and Ridge province and Oak

Openings region of Ohio, and two accessions were obtained from nursery sources that grew plants from seed collected from natural plant populations within the geographical range of the species (Table 1.8). Phlox ovata is the only state-endangered Phlox in Ohio,

and special effort was made to obtain permits, study known populations, and collect

95 germplasm from these sites. Although other sites are recorded for this species, germplam collection efforts were made from primarily within the Oak Openings Preserve and

Meilke Road savannah in Lucas County, Ohio. Collections from these sites are uniform in habit and size with slight variations in flower color and flower shape (Figure 1.28).

Germplasm collection expeditions in the Valley and Ridge province of Virginia and West Virginia revealed this species to be particularly abundant in this region, and a large degree of variation was found. At one collection site in Allegheny County, Virginia

(PZ12-077), a population was discovered with novel, white-flowered variants described by Wherry (1955) as “rare”. These floral variants ranged from white with a pink eye, to pale pink, with some approaching salmon-pink in color (Figure 1.28). Special attention was paid to vegetative propagation of floral variants. Later, seed was collected from this population. In terms of floral distinctiveness, this was the most variable population of P. ovata discovered. Further study of this and proximal populations may reveal additional variants of this species.

Collections in the Valley and Ridge province of Tennessee resulted in an accession of P. ovata with smaller than average leaves; so far this feature has been retained in cultivation.

Phlox pulchra Wherry - Alabama phlox

Phlox pulchra was previously reported as a diploid, but the results of this research indicated that this species occurs in both diploid (n = 7) and tetraploid (n = 14) forms

(Smith and Levin, 1967; Chapter 2). Further sampling and study is needed to determine

96 the range of ploidy variation in this taxon. Phlox pulchra was formerly considered to represent the southernmost populations of P. ovata, but Wherry (1935b) separated it on the basis of morphological characters. This species is endemic to Alabama, and was originally reported from 5 counties. Most populations are associated with the southern edge of the Cumberland Plateau, but populations have also been reported from the Valley and Ridge province, and Ketona Dolomite Glades in Bibb County; it is a plant of open, mesophytic forest and occurs in open glade conditions with other endemic taxa in Bibb

County (Figure 1.2) (Allison, 2014). Although formal designation does not exist, this species has become rare since Wherry’s description and may be endangered. Further fieldwork is necessary to inventory all extant populations of this taxon.

Wherry (1935b) distinguished P. pulchra from P. ovata on the basis of morphological features: elongate sterile stems, more nodes per flowering stem, and short- petiolate, as opposed to long-petiolate, lower leaves. This taxon appears to form a natural group with other early flowering (May), short statured (>40 cm), members of subsection

Phlox, such as P. glaberrima ssp. triflora and P. ovata, that have a distinctly mounding habit at anthesis. These taxa grow in more xeric habitats than taller, later flowering taxa in the Phlox carolina-glaberrima complex, and from a horticultural standpoint, for a morphologically, and perhaps genetically, similar group (Ferguson et al., 1999; Smith and Levin, 1967). The potential for hybridization of this species with others remains untested.

Phlox pulchra has been introduced into cultivation, but is not a commonly seen in gardens. Despite this, it exhibits many ornamental characteristics that make it worthy of

97

expanded collection, selection, and breeding. Generally, the flowers are pastel pink,

exhibit overlapping corolla lobes, and are large in comparison with other members of

subsection Phlox (Figure 1.29). The foliage is similar to that of P. ovata, but is smaller

and blue-green in color with a short petiole. This species may represent a source of

powdery mildew (Erysiphe cichoracearum) resistance; it is not known to become infected with this pathogen, but there has not been and extensive trial and evaluation of this taxon. At least one cultivar attributed to this species, ‘Morris Berd’, has been rated as an excellent performer in one evaluation, however, the morphological characteristics of ‘Morris Berd’ indicate that it is a selection of P. glaberrima ssp. triflora (Hawke,

1999).

Phlox pulchra was the only species not studied during the course of botanical fieldwork carried out in this study. Four accessions were obtained from nursery sources

(Table 1.8). Phlox pulchra ‘Eco Pale Moon’ PZ10-102 was selected by Don Jacobs, Eco

Gardens, Decatur, Georgia. This selection was made and named in reference to the rounded flowers with overlapping corolla lobes and a pale, pastel-pink flower color that differs from the typical form. This selection has been difficult to establish at the OPGC and appears to be infected by virus. Phlox pulchra PZ10-242 was obtained from Tripple

Brook Farms in Southhampton, Massachusetts. These plants most closely matched the

botanical description of P. pulchra, but were unfortunately also lost. PZ11-021 was

collected by Tony Avent of Plant Delights Nursery, in Bibb County, Alabama (Figure

1.29). The leaves of this collection are narrowly ovate, that differ from the more ovate

leaves of PZ10-242, and have an affinity to P. glaberrima ssp. triflora. PZ11-058 was

98 obtained from Mulberry Woods Nursery in Garden City, Alabama. This selection appears to be intermediate in morphology between P. pulchra and P. glaberrima ssp. triflora. Given the range of morphological variation in cultivated selections of P. pulchra, the need to study and analyze the variation in natural populations becomes more apparent.

Subsection Stoloniferae

This subsection consists of two well-defined, geographically remote species with unique morphological attributes; P. adsurgens is native to the Pacific Northwest, and P. stolonifera is found in the eastern North American mountains. Both species have long styles (>6 mm), and a unique, stoloniferous growth habit (Table 1.1). A total of 14 collections were made from this subsection (Table 1.4). One accession of P. adsurgens, the cultivar ‘Wagon Wheel’, was obtained from a commercial source.

Phlox stolonifera Sims - Cherokee phlox

Phlox stolonifera is a diploid (n = 7), perennial species that is distributed in the

Allegheny Plateaus, Valley and Ridge province, and Appalachian Highlands; it is found in rich mesic hardwood forest where it grows with a diverse array of spring ephemerals.

Phylogenetic studies involving molecular and karyotypic data have suggested that P. stolonifera is the most primitive Phlox species and basal to the rest of the genus

(Ferguson and Jansen, 2002).

99

Phlox stolonifera can be differentiated from P. adsurgens on the basis of having sparsely pilose, obovate leaves with ciliate margins, and a natural distribution in the

Appalachian region of the eastern U.S. (Wherry, 1955). The leaves of P. adsurgens are glabrous and elliptical. The flowers and leaves of P. stolonifera are larger than those of

P. adsurgens. In cultivation, P. stolonifera bears a carpeting, wide-spreading growth habit that is different from all other species of Phlox (Figure 1.30). The leaves are evergreen and can vary in size, but always have a characteristic obovate shape.

Flowering occurs in April and May and the flowers can be lavender-purple, pink, or white. Striae are uncommon or absent from the flowers of this taxon (Figure 1.30)

(Wherry, 1955).

A total 13 accessions of P. stolonifera were obtained for evaluation (Table 1.9).

Seven collections were made from natural populations. Six cultivars were obtained from commercial sources, and included: ‘Bruce’s White’, ‘Eyes Have It’, ‘Fran’s Purple’,

Home Fires’, ‘Pink Ridge’, ‘Sherwood Purple’, and ‘Weesie Smith’. Descriptions of these cultivars can be found in Bendtsen (2009) and Fuchs (1994).

Accessions collected from natural populations were relatively invariant; cultivars available in the nursery trade appear to capture the breadth of phenotypic diversity seen in natural populations of this species. Populations from Ohio and Pennsylvania have large pink flowers lacking striae. These collections have the largest leaves of all accessions, reaching 3 cm in diameter. Southern populations from Tennessee and

Northern Georgia have lavender-purple flowers. PZ12-134 collected from Polk County,

Tennessee has exceptionally small leaves reaching 1.5 cm in diameter. Populations seen

100

flowering along Highway 441 in Great Smoky Mountains National Park in late April

2011 were also pale lavender, but exceptionally floriferous, however no accessions were made in this area due to collection restrictions. One accession from Georgia (PZ10-125) perished before further analysis could take place.

Subsection Subulatae

Phlox subsection Subulatae consists of 4 species and up to 7 subspecies (Table

1.1) (Locklear, 2011a). Wherry (1955) placed these taxa in section Phlox, but morphological comparison and molecular evidence have suggested that subsection

Subulatae may be more closely related to taxa in section Occidentales, which consist primarily of western Phlox species, than to other eastern taxa (Ferguson and Janssen,

2002; Smith and Levin, 1967). Species in section Occidentales can have long or short

styles (less than 6 mm or greater than 6 mm), generally have a pulvinate or caespitose,

suffruticose habit, and are associated with xeric rock outcroppings or alpine

environments. Taxa in subsection Subulatae occupy similar xeric habitats found

throughout the eastern United States, and are similar in morphology to western taxa. The

relationships of these taxa highlight the need for a more thorough and exhaustive genus-

wide molecular phylogeny.

A total of 22 accessions were collected from natural populations and 20 obtained

from commercial sources (Table 1.10). Members of subsection Subulatae are

collectively known as the “moss-phloxes” and are widely cultivated for their intense

floral spring displays. Most cultivars are listed as selections or hybrids of P. subulata,

101 but it is likely that several taxa have been involved in the breeding and selection of cultivars.

Phlox bifida Beck ssp. bifida Wherry - cleft phlox, sand phlox, ten-point phlox

Phlox bifida ssp. bifida is a diploid (n = 7), perennial species distributed primarily within the Great Plains province; it is a characteristic plant of xeric sand prairies and oak savannah ecosystems throughout the Midwestern U.S. (Figure 1.2, 1.31) (Locklear,

2011a; Wherry, 1955). This plant is listed as extirpated in Michigan (Michigan Flora,

2014).

Phlox bifida ssp. bifida is distinguished from other subspecies by having a style that is 6-12 mm in length, glandular pubescence, and relatively variable flower color that is typically pale-violet, but ranges from violet purple to pure white (Figure 1.31). This species is also segregated from other subspecies on the basis of geographical distribution and occurs to the north of other subspecies (Wherry, 1955). All subspecific taxa of P. bifida bear the characteristic notch in the corolla lobe, but this character is also shared with some forms of P. nivalis and P. subulata ssp. brittonii. In cultivation, P. bifida grows 15-20 cm in height, has a pulvinate habit, and flowers from April to early May

(Figure 1.31). This species prefers a well-drained site and performs particularly well on sandy soils, and in raised beds (Figure 1.31). Plants are short lived when planted in nutrient-rich soils with high levels of clay and silt. One cultivar, ‘Betty Blake’, was selected for its violet-purple flowers that are a richer color than the typical form. This cultivar was selected from a now extirpated Michigan population (Bendtsen, 2009).

102

Two accessions of this taxon were obtained; one was the cultivar ‘Betty Blake’ and the other was collected in Indiana (Table 1.10). The collection from Newton County,

Indiana (PZ10-149) was made in remnants of oak savannah habitat dominated by black oak (Quercus velutina) (Figure 1.31). Seeds were collected on May 19, approximately one month after flowering (Table 1.5). Most Phlox seed accessions tested in this study displayed non-deep physiological dormancy, but seeds of PZ10-149 only germinated at high percentages when treated with 500 ppm Giberellic Acid (GA3) (Baskin and Baskin,

2004). Seed germination of other taxa in subsection Subulatae have not required similar treatment and exhibit non-deep physiological dormancy comparable to other accessions.

Phlox bifida Beck ssp. arkansana Marsh – Arkansas cleft phlox

Phlox bifida ssp. arkansana is a perennial taxon endemic to 5 northwestern

Arkansas counties in the Interior Highlands province (Marsh, 1960). Populations occur within the Interior Highlands province and grow in xeric habitats (Marsh, 1960). This is the most recently described taxon of P. bifida, but has not been universally recognized as a distinct subspecies (Marsh, 1960; USDA Plants database, 2014). This uncommon taxon was originally described from 8 Arkansas counties, but has become rarer since the original description (Locklear, 2011a; Marsh, 1960). The taxon was not observed or collected during the course of this study, but I believe it has horticultural potential that warrants sampling. Given its rarity, germplasm collections should also be made in the interest of long-term conservation.

103

Phlox bifida ssp. arkansana is distinguished from other P. bifida taxa by having a short style that reaches 5 mm in length (Marsh, 1960). Flower color was reported to be a deeper, reddish-purple than other forms of P. bifida. The corolla lobes are narrower than other subspecies, giving rise to flowers that are star-shaped (Marsh, 1960). No germplasm collections of Phlox bifida ssp. arkansana were made during the course of this study.

Phlox bifida Beck ssp. stellaria A. Gray - Kentucky cleft phlox

Phlox bifida ssp. stellaria is a diploid (n = 7), perennial species that occurs at the southeastern edge of the natural distribution of P. bifida in the Interior Low Plateau province (Figure 1.2). Unlike P. bifida ssp. bifida, its habitat is limestone barrens, and rocky bluffs and cliffs along major rivers (Figure 1.32) (Campbell, 2012; Locklear,

2011a; Wherry, 1955). It grows in a xeric habitat, and occurs in rocky soils or as a chasmophyte (Figure 1.32). Field observation indicated that populations are small, highly fragmented, and patchily distributed in suitable habitat that occurs within large expanses of unsuitable habitat.

Phlox bifida ssp. stellaria is distinguished from other P. bifida taxa by the presence of eglandular pubescence, a style 6 -12 mm in length, and flowers that are typically pale-violet, and only rarely white (Figure 1.33). This subspecies is further differentiated from other taxa by its unique geographic range, and habitat preference.

Plants reach up to 12 cm in height at flowering and flower during April and May (Figure

1.32). The plants are evergreen.

104

From a horticultural standpoint, this species has a more compact, uniform habit

than P. bifida ssp. bifida in cultivation; P. bifida ssp. stelleria is a more desirable

ornamental plant. My field observations indicated that flower color of P. bifida ssp.

stellaria is less variable than other subspecies, and all collections of this subspecies have uniformly pale-violet flowers. This subspecies may also prove to be more persistent in garden and landscape conditions, but suffers from competition when crowded by neighboring plants.

Four accessions were obtained (Table 1.10). Two were collected in Kentucky, one in Tennessee, and one is a cultivar selected from a wild population in Tennessee.

PZ10-018 was made from Wilson County, Tennessee from cedar glade/limestone barren habitat in the Tennessee Central Basin. This species was irregularly distributed throughout large expanses of seemingly suitable habitat, and was only found in roadside right-of-ways, where road construction appears to have created an ecotone favorable to long-term persistence of this taxon. Flower color was slightly variable, but generally pale-violet (Figure 1.32). The only known cultivar of this species, ‘Glade Blue’ (PZ10-

026), was selected from the same region by Growild Nursery, Fairview, Tennessee.

PZ10-201 was collected from Jessamine County, Kentucky, from type locality for the species (Gray, 1870). At this location, plants occur on vertical dolomite cliffs and grow as chasmophytes in the cracks of these cliffs (Figure 1.32). Flower color of these plants is an invariable pale-violet, but the plants do vary in form, and some individuals have a tight, caespitose habit and evergreen leaves that are ornamental desirable features. A second collection made by Julian Campbell originated in Hart County, Kentucky, where

105 it was found on limestone cliffs above the Green River. This form is similar PZ10-201, but has spreading, caespitose habit, and more deeply hued, violet flowers. This selection is among the most superior collections of P. bifida ssp. stelleria and should be registered as a cultivar.

Phlox nivalis Loddiges ssp. nivalis Wherry - trailing phlox

Phlox nivalis is a diploid (n = 7) or tetraploid (n = 14) species of the Coastal Plain and occurs in xeric upland pine ecosystems in sandy soils, but also occurs on heavier, richer soils in the Piedmont (Figure 1.2) (Smith and Levin, 1967; Chapter 3). Wherry

(1955) divided P. nivalis into three subspecies on the basis of geographical distribution and morphology (Wherry, 1955). He recognized Phlox nivalis ssp. nivalis, P. nivalis ssp. hentzii, and P. nivalis ssp. texensis, but Ferguson (1998) included P. nivalis ssp. hentzii in

P. nivalis ssp. nivalis, citing the former as environmentally induced variants.

Furthermore, Wherry (1955) included this taxon, along with P. oklahomensis, in subsection Speciosae with primarily western taxa on the basis of having short style.

However, molecular phylogenetic studies have concluded that P. nivalis and taxa in subsection Subulatae for a monophyletic group, and that P. nivalis is member of subsection Subulatae (Ferguson, et al., 1999; Ferguson and Janssen, 2002). This taxon occurs at the southern edge of distribution of subsection Subulatae, and like other taxa in this subsection, occur as fragmented, local populations separated by expanses of unsuitable habitat (Wherry, 1955).

106

The short style (> 4 mm) of P. nivalis can be used to distinguish this taxon from

other members of subsection Subulatae, which have either longer or shorter styles. This

species also has a different natural distribution than other taxa, and there does not appear

to be any overlap between this and related taxa (Wherry, 1955). Phlox nivalis is an

evergreen suffruticose perennial with a mounding or mat forming habit. The plants can

grow to 20 cm in height at flowering but are typically less than 10 cm tall. Flower color

is variable, and can range from pink to white, and some individuals have unique flowers

with white centers and a pink edge (Figure 1.33). Striae may be present or absent.

A total of 8 accessions of P. nivalis were obtained; two of these were collected from natural plant populations, and the remaining accessions were obtained from commercial sources (Table 1.10). PZSH2011-013 was collected from Bay County,

Florida at the southern edge of the distribution of this species; it was found in a roadside easement. Plants were flowering on April 5, 2011, and flower color ranged from pink to white, with several intermediate colors (Figure 1.33). The flower color variation in this

population was extraordinary among populations of related taxa, however, since this was

the only population seen, I do not know if this is typical of the species. PZ11-051 was

obtained from the Mt. Cuba Center, Wilmington, Delaware. This accession was collected

from Durham County, North Carolina. The flower color among individuals from this

accession was a uniform bright pink. Individuals of two different ploidy levels were

found in this population (Chapter 3).

Dr. Don Jacobs of Eco Gardens, Decatur, Georgia selected two cultivars from a

population in Franklin County, Georgia: ‘Eco Brilliance’ and ‘Eco Flirty Eyes’. The

107

differences in these individuals suggest that they were selected from an exceptionally

variable population. One accession, obtained from Plants Delights Nursery as P. nivalis

‘Camla’ (PZ10-129) is misidentified and appears to be the plant described as Phlox xhenryae, a hybrid between P. bifida and P. nivalis (Wherry, 1935). The original

‘Camla’ is described as having large, rounded, salmon colored flowers with red striae, but the plant obtained has pale-lavender flowers with notched corolla lobes that result in a star-shaped flower (Figure 1.33) (Wherry, 1955). Another accession (PZ12-128) obtained from Plant Delights Nursery under the collection number “A2GA-005” was collected from a natural population in Georgia and bears bright pink flowers. Two additional selections were obtained from Nearly Native Nursery, Fayetteville, Georgia, but perished before further evaluation could take place.

Phlox nivalis ssp. texensis Lundell - Texas trailing phlox

This Texas endemic is one of the most rare Phlox taxa and is a federally endangered species (USDA Plants Database, 2014). It grows in open pine forest in sandy soils and is only known to occur in the Roy E. Larsen Sandyland Preserve in Hardin

County, Texas. This species was seen at this preserve during a 2010 germplasm collection expedition to Texas, but the endangered status prohibited collection.

Phlox oklahomensis Wherry - Oklahoma phlox

Phlox oklahomensis is a perennial species from Oklahoma and Kansas that grows in the Great Plains province and is associated with prairie habitat (Figure 1.2) (Springer

108 and Tyrl, 1989; Springer and Tyrl, 2003; Wherry, 1955). Populations of P. oklahomensis from northwestern Arkansas are now known as P. bifida ssp. arkansana based on comparison of morphological, habitat, and geographical differences (Marsh, 1960;

Wherry, 1955). This species was not collected and there as no reference to any cultivars of this species. It is not known to be in cultivation.

Phlox subulata L. ssp. subulata - creeping phlox, moss phlox, thrift

Phlox subulata ssp. subulata is a well-known perennial taxon; it is primarily diploid (n=7), but tetraploid (n=14) and hexaploid (n=21) populations have been found

(Chapter 3). It is distributed in primarily in the Allegheny Plateau province, but is also found in the Great Plains, Valley and Ridge, and Northeastern provinces (Figure 1.2)

(Wherry, 1929; Wherry, 1955). It is associated with a variety of soils types and habitats, but is most frequent on shale-derived soils where competition from sympatric plants is reduced or eliminated by period disturbance; it is frequently found on rock outcroppings derived from numerous geologic formations (Locklear, 2011a). Geographical variation in P. subulata has resulted in the description of three subspecies: P. subulata ssp. subulata, P. subulata ssp. brittonii, and P. subulata ssp. setacea (Locklear, 2011a;

Wherry, 1955). Phlox subulata ssp. setacea was named P. subulata ssp. australis by

Wherry (1955), but Locklear (2009) reinstated the name P. subulata ssp. setacea, and claimed that this name was used earlier to describe the same taxon. Phlox subulata ssp. subulata is a widely planted, popular ornamental plant in temperate regions of the world.

109

Phlox subulata ssp. subulata can be distinguished from other taxa in having eglandular pubescence of the inflorescence (Wherry, 1955). Other subspecies of P. subulata have glandular pubescence. This taxon is the most widespread of the group and occurs in the northern, eastern, and western portions of the species range.

Cultivars of P. subulata and its hybrids have been widely planted throughout the eastern United States; cultivated plantings are long-lived and may give rise to spontaneous, adventive populations that persist around abandoned homesteads and cemeteries (Figure 1.34). Plants from such populations are sometimes collected by botanists and catalogued in states well outside of the native range of P. subulata. The resultant maps greatly exaggerate the original range of the species as described by

Wherry (1929, 1955), and special attention must be paid to where the plants are collected in order to determine whether they represent a natural or adventive population. Such plantings offer potential as genetic resources and may consist of previously selected ornamental forms that were never designated as cultivars.

Four accessions were collected from natural populations (Figure 1.10). PZ12-

065, PZ12-066, and PZ12-067 were collected at the eastern edge of the geographical distribution of this subspecies in serpentine barrens, and were found to represent polyploid complex with diploid (PZ12-066, n=7), tetraploid (PZ12-065, n=14), and hexaploid (PZ12-067, n=21) plants (Figure 1.35). PZ12-065 was collected near Chester,

Pennsylvania and is diploid; to the south, in the Nottingham Barrens in Chester County,

Pennsylvania, a tetraploid population is found, and the plants around the Soldiers Delight barren in Maryland are hexaploid (Figure 1.35; Chapter 3).

110

The remaining collection (PZ12-091) was made on the border of Delaware and

Franklin counties in Ohio from shale cliff habitat bordering Hoover Reservoir (Figure

1.35). Morphological attributes are comparable to the previously described collections,

however this population is very close to the north-south divide of the two subspecies

(Figure 1.35). This taxon has richly colored, contrasting striae, and the shape and size of

the corolla lobes was more variable than P. subulata ssp. setacea PZ13-009 from

southern Ohio (see below).

Eleven cultivars referable to P. subulata were obtained for comparative analysis.

Some of these may represent interspecific hybrids with related species. Descriptions of

P. subulata cultivars can be found in previously published references (Bendtsen, 2009;

Fuchs, 1994).

Phlox subulata L. ssp. brittonii Small (Wherry) - shalebarren creeping phlox

Phlox subulata ssp. brittonii is a tetraploid (n=14), perennial species with a

restricted distribution in the Valley and Ridge province in West Virginia, Virginia,

Maryland and Pennsylvania where it is closely associated with Devonian age shale

deposits (Figure 1.2, 1.36) (Wherry, 1929; Wherry, 1955). This taxon is frequent within

its range, and is not sympatric with other subspecies of P. subulata.

Phlox subulata ssp. brittonii can be distinguished from P. subulata ssp. subulata on the basis of having glandular (versus eglandular) pubescence on the calyx and of inflorescence architecture. It is distinguished from P. subulata ssp. setacea by the longer corolla tube (averaging 10.5 mm in length), and white or pale lavender, as opposed to

111

pink, flowers (Figure 1.36) (Wherry, 1929; Wherry, 1955). It is a smallest taxon in the P.

subulata complex, with a pulvinate habit averaging 25-30 cm in diameter.

From a horticultural standpoint, P. subulata ssp. brittonii is the least showy

subspecies in regards to flower color (Figure 1.36). However, its dwarf habit is desirable

for rock gardens. Plants grow to 15 cm in height and flowering during April and early

May. Flower size and shape can be variable within and among populations (Figure 1.36).

Attempts should be made to find forms with differently, or more richly colored flowers.

Nine wild collections were made of this taxon in Allegheny and Botetourt

counties in Virginia, and Grant and Greenbrier counties in West Virginia; cultivars of this subspecies have not been introduced (Table 1.10). There is limited phenotypic variation among accessions in habit, flower color, and phenology, but some individuals of PZ11-

037 had pale pink, rather than pale lavender flowers. Several populations were seen on road cuts along the Maryland-Pennsylvania border that fit within the general concept of this taxon. Collections were not made from these populations.

Phlox subulata ssp. setacea (L.) Locklear - Blue Ridge creeping phlox

Phlox subulata ssp. setacea is a diploid (n=7), perennial species that occurs in the

Allegheny Plateaus, Valley and Ridge province, and Appalachian Highlands provinces where it grows on a variety of soil types, ecosystems, and elevations (Figure 1.2)

(Locklear, 2011a; Wherry, 1955).

This subspecies is distinguished from P. subulata ssp. subulata in having glandular (versus eglandular) pubescence of the calyx and inflorescence architecture.

112

Otherwise the differences between the two taxa are minimal. It is distinguished from P. subulata ssp. brittonii by a shorter (ca. 7.5 mm long versus 10.5 mm long) corolla tube

(Wherry, 1929; Wherry, 1955).

Two collections of this taxon were made (Table 1.10). One collection (PZ12-095) was made by Julian Campbell in Estill County, Kentucky where it was found in transitional areas between shale and limestone bedrock. Plant habit and flowering characteristics are similar to that of P. subulata ssp. subulata. An additional collection

(PZ13-009) was made in Vinton County, Ohio, where it was found growing on sandstone outcrops under sparse eastern white pine (Pinus strobus L.) forest. Plant habit and flowering characteristics are similar to those of the aforementioned collection and those of typical P. subulata ssp. subulata.

Future Phlox collections

Future Phlox germplasm collection expeditions should serve several purposes: continued acquisition of commercially important Phlox taxa, collection of rare or threatened taxa, and collection of taxa for which taxonomic confusion persists.

Observation of natural Phlox populations permitted preliminary assessment of a significant portion of phenotypic and genetic variation of eastern Phlox species. It has also allowed us to determine priorities for future collection efforts through identification of regions and populations of interest. Future collections should be prioritized based on the nature of the study being performed.

113

Several taxa were not observed or collected, and are of interest for seed conservation and hybridization studies. Phlox bifida ssp. arkansana (and adjacent populations of P. bifida in Missouri), P. cuspidata, P. floridana ssp. bella, P. oklahomensis, P. carolina ssp. turritella, and several subspecies of P. drummondii were not observed during the course of this study, and are not cultivated so germplasm of these taxa was not available for this study. Furture studies should focus on these species so that germplasm characterization initiated during the course of this study can be further advanced. Other taxa, such as P. pattersonii and P. pulchra, were obtained as cultivars or nursery selections of wild collections, but these species were not observed in the wild and they were not studied in their natural habitat.

Further collections of P. paniculata should be focused in two areas. The original introductions of P. paniculata were made from Virginia (Symons-Jeune, 1953). Many taxa were described to encapsulate the range of variation in the species, and it is possible that the initial horticultural selections for floral variants may have been developed from collections in this region. A second region should be the Cumberland Plateau of

Tennessee.

The extensive morphological variation and lack of taxonomic resolution in some species warrant further collection of certain taxa. This is particularly true of species in the P. carolina-glaberrima complex, where fine-scale geographic sampling and choice of appropriate molecular markers could help inform phylogenetic studies.

114

Characterization of Phlox Germplasm

Traditional characterization of germplasm involves the measurement and description of phenotypic traits from replicates of plants increased from seeds. Due to the difficult nature of Phlox seed collection, few previous NPGS collections prior to the onset of this study, and the difficulty in increasing phlox seeds in a timely manner under previously described protocols, initial characterization of Phlox germplasm focused on the crop descriptors than could be achieved with comparatively few plants than other similar studies within the USDA crop germplasm system. Driven by industry interest and limited prior information, initial characterization was focused on assessment of interspecific compatibility among a subset of collected taxa, estimation and analysis of genome size and ploidy variation for the entire collection, and use of two molecular marker systems to describe patterns of genetic variation for a subset of collections from natural plant populations.

Interspecific hybridization relationships remain only partially characterized or anecdotal in the genus Phlox, and there is a need for additional knowledge (Levin, 1966;

Locklear, 2011a; Symons-Jeune, 1953). Several potential prezygotic barriers to interspecific hybridization have been identified, among them are: protandry, differences in style length (long vs. short styled), and differences in phenology. Tentative post- zygotic barriers have also been identified and include: a strong self-incompatibility system, differences in style length, differences in genome size and ploidy, and endosperm imbalance (Levin, 1966; Levin and Smith, 1966; Wherry, 1935). Germplasm accessions were screened for unique phenotypes such as different flower colors, phenology, and

115 style length, among others. Depending upon the inheritance of such traits, it is possible that they could serve as markers of successful hybridization in F1 progeny. This information is critical for the delineation of Phlox crop gene pools that is needed to create novel and adaptable hybrids.

Genome size and ploidy are considered one of the most potent barriers to interspecific hybridization of ornamental plants (Eeckhaut et al., 2006). Flow cytometry analysis allows for rapid assessment of genome size in experiment taxa when compared with a standardized, internal reference genome. In combination with chromosome counts, ploidy can be calibrated to genome size, and used to infer ploidy among taxa where chromosome counts have not been made. Genome size and ploidy have only been partially characterized in eastern Phlox species. Initial karyotypic studies determined that most taxa were diploid, however recent flow cytometry studies have revealed polyploid complexes in several studies, even with sampling from a limited geographic region

(Flory, 1931, 1933; Fehlberg and Ferguson, 2012; Smith and Levin, 1967; Meyer, 1944;

Worcester et al., 2012).

Microsatellite (SSR) markers remain a powerful tool to describe patterns of genetic variation in species of interest to germplasm scientists. Tools such as this may be particularly useful in genera like Phlox, where some species boundaries are still not well defined and a poorly resolved taxonomy can affect the efficiency of germplasm collection and conservation. Preliminary studies have revealed cryptic ploidy and genetic variation that supported the delineation of two closely related and traditionally taxonomically disputed species from the western United States, but have not been applied to

116

taxonomically difficult eastern species complexes, such as the P. carolina-glaberrima

complex or P. pilosa complex. We selected a subset of taxa in subsection Divaricatae, with an emphasis on P. pilosa across an east west transect at the southern edge of the species range along the Gulf Coast of the southern United States. Such studies can help clarify patterns of genetic diversity in widespread taxa and provide baseline information for determining collection protocols and priorities.

117

Figure 1.1: Map of the natural range of eastern Phlox species and collection sites. Collection expeditions were focused in three sub- centers of Phlox species diversity. Sub- centers were defined based on the number of taxa reported from each state within the range of eastern species.

118

Figure 1.2: Physiogeographic provinces of the United States. Map obtained from http://apeoplesconstitution.wikispaces.com/Regions+Maps.

119

a.

Figure 1.3: The mature capsule of Phlox. This is the stage at which mature seed can be collected. a. The mature capsules of P. bifida ssp. bifida PZ10-149 on May 19, 2010 in Newton County, Indiana.

120

a. b. c.

Figure 1.4: Differences in style length of Phlox taxa. a. The long style (10-25 mm) of P. paniculata is representative of subsections Paniculatae and Phlox. b. The long style (6-12 mm) of P. subulata is representative of long-styled in subsection Subulatae. c. The short style (1-4 mm) of P. villosissima is representative of subsection Divaricatae. The scale bar is 10 mm.

121

a.

b.

Figure 1.5: Flowers and habit of Phlox buckleyi in Greenbrier County, West Virginia. a. Detail of inflorescence and flowers of P. buckleyi. Note the characteristic narrow, lanceolate leaves of this taxon. b. Detail of the prolifertation of loosely rhizomatous sterile stems characteristic nof this species.

122

a.

b. c.

Figure 1.6: Phlox amoena in the wild and in cultivation at the OPGC. a. A flower color variant of P. amoena with flowers that open white with lavender striae fused into a star that fades to pink. b. Two different forms of P. amoena at the OPGC showing variation in flower color. c. Phlox amoena in situ along a roadbank in Benton County, Tennessee. This species frequently inhabits disturbed areas.

123

a. b.

Figure 1.7: Flowering habit of Phlox divaricata ssp. divaricata in Hocking County, Ohio. a. Typical individual. b. A white-flowered individual. Such flower color variants of this taxon are easily found in large populations within the Great Plains regions of this taxon and P. divaricata ssp. laphamii.

124

a.

b.

Figure 1.8: Flower detail and natural habitat of Phlox divaricata ssp. laphamii in Gadsden County, Florida, at the southern edge of this taxon’s range. a. The flowers of P. divaricata ssp. laphamii lack the notch in the in the tip of the corolla lobe that is present in P. divaricata ssp. divaricata. b. The steep, north facing bluff habitat of P. divaricata ssp. laphamii, where it grows with Adiantum capillus-veneris, Decumaria barbara, and Trillium lancifolium.

125

a.

b. c.

Figure 1.9: Flowering characteristics of two subspecies of Phlox drummondii in situ. a. The natural habitat of P. drummondii in Caldwell County, Texas on May 28, 2010. b. A typical, flowering individual of P. drummondii ssp. drummondii PZ10-161 flowering on May 28, 2010 in Caldwell County, Texas. b. A typical, flowering individual of P. drummondii ssp. mccallisteri flowering along a roadside in Wilson County, Texas on May 29, 2010.

126

a.

b. c.

Figure 1.10: Phlox roemeriana PZ10-165 (TX-057) in situ, Comal County, Texas. a. The xeric, limestone habitat of P. roemeriana. b. A typical, flowering individual on May 29, 2010. c. Another flowering individual illustrating the range of the flower color variation.

127

a.

b.

Figure 1.11: Phlox floridana PZSH2011-010. a. The xeric, upland pine (Pinus spp.) habitat of P. floridana in Jackson County, Florida. The plants were most abundant on the steep slope of the road cut (not in flower) and growing with Lupinus perennis, April 4, 2011. b. Phlox floridana PZSH2011-010 flowering in cultivation at the OPGC. When grown under long days, the plants will flower throughout the year. 128

a. b.

d. c.

Figure 1.12: The ecotype of Phlox pilosa ssp. pilosa that occurs in the southeastern Great Lakes region and Ohio River valley of the Great Plains province. a. The limestone barren/cedar glade habitat of P. pilosa ssp. pilosa in Hardin County, Kentucky in the Interior Low Plateau province. b. A flowering individual of P. pilosa ssp. pilosa in Hardin County, Kentucky on April 24, 2012. c. The habitat of P. pilosa ssp. pilosa PZ12-060 is a tall grass prairie remnant in the Great Plains province in Lake County, Indiana. Plants grow in nutrient rich soils associated with this habitat. d. A flowering individual of P. pilosa ssp. pilosa.

129

a. b.

c.

Figure 1.13: Variation in Phlox pilosa ssp. pilosa PZSH2011-020 from Forest County, Mississippi. Phenotypic variation is greater in populations from the Coastal Plain and Interior Low Plateaus than in populations from the Great Plains. a. An exceptional color variant displaying white flowers with pink striae fused into a star. b. A richly colored individual that also displays c. Uniquely purple flushed foliage.

130

a.

b.

c.

Figure 1.14: Phlox pilosa ssp. deamii PZ11-026 in situ, Christian County, Kentucky. a. The open, mesic forest habitat of P. pilosa ssp. deamii. b. Detail of inflorescence architecture showing the long, white, eglandular calyx pubescence of this taxon. c. A typical individual of P. pilosa ssp. deamii at anthesis on April 25, 2012. 131

a.

c.

b.

Figure 1.15: Phlox pilosa ssp. deamii PZ11-026 in cultivation at the OPGC. a. Detail of the dense, mounding habit of this taxon in cultivation. b. Close-up of flowers. c. Common garden study of P. pilosa complex accessions. The P. pilosa ssp. deamii are the plants in the forefront.

132

a.

b. c.

Figure 1.16: Phlox amoena x Phlox pilosa ssp. pilosa hybrid swarm (P. pilosa ssp. deamii) PZ12-054. a. The mowed, sloping roadside habitat. Note that P. amoena, P. pilosa ssp. pilosa, and hyrbids are growing side by side. b. Habit and flower color variation of hybrid individuals. The degree of variation suggested hybridization and varying degrees of introgression. c. A unique flower color variant found among hybrids. Flowers are salmon-pink.

133

a.

b. c.

Figure 1.17: Phlox pilosa ssp. fulgida PZ12-092 in situ in Story County, Iowa. a. A completey white-flowered individual. b. A typical pink-flowered individual. c. A white- flowered individual with prominent purple striae. Photos and Germplasm collection (PZ12-092) courtesy of Jeffrey Carstens.

134

b. a.

c. d.

Figure 1.18: Phlox pilosa ssp. longipilosa in situ and in cultivation at the OPGC. a. A close-up of the inflorescence showing the aglandular, comparatively long hairs that are longer than any other member of the P. pilosa complex. b. The natural habitat in the Quartz Mountains. c. Detail of the corolla, note the rounded, overlapping petals, and distinctive, vibrant coloration. d. The mounding habit of P. pilosa ssp. longipilosa, in early anthesis, at the OPGC.

135

a.

b.

c. d.

Figure 1.19: Phlox pilosa ssp. ozarkana in situ and in cultivation at the OPGC. a. The heavily forested roadside habitat of P. pilosa ssp. ozarkana in Iron County, Missouri. b. A typical, flowering individual of P. pilosa ssp. ozarkana in roadside habitat on April 27, 2012. c. Variation among individuals of P. pilosa ssp. ozarkana PZ10-227 collected in Johnson County, Arkansas, and cultivated at the OPGC. d. A selected white-flowering individual of P. pilosa ssp. ozarkana PZ10-227. 136

a.

b. c.

Figure 1.20: Phlox pulcherrima PZSH2011-033 in situ and in cultivation. a. The mowed roadside habitat of P. pulcherrima in Shelby County, Texas. b. The flowers of P. pulcherrima. c. The habit of P. pulcherrima when grown in container. The plants are taller than typical Phlox pilosa ssp. pilosa, and differ in flower color and floral fragrance.

137

a. b.

c.

Figure 1.21: Phlox villosissima in situ and in cultivation. a. The mowed, roadside habitat of P. villosissima PZSH2011-040 in Kerr County, Texas. b. A typical, flowering individual of P. villosissima PZSH2011-037 in Kerr County, Texas on April 13, 2011. c. Phlox villosissima PZSH2011-036 in cultivation at Natives of Texas Nursery in Kerr County, Texas. 138

a.

b.

c.

Figure 1.22: Phlox amplifolia PZ11-050 in situ in Cocke County, Tennesssee. a. The steep, talus slope habitat along the French Broad River. b, A flowering individual on October 8, 2010. c. An infructescence with mature seed capsules and evidence of dehisced capsules (empty, star-shaped calyces).

139

b. a.

c.

Figure 1.23: Phlox paniculata in situ. a. P. paniculata PZ10-209 from Scioto County, Ohio along a roadside bordering Rocky Fork Creek flowering on August 1, 2010. b. A typical flowering individual of P. paniculata PZ11-040 flowering along the Cheat River in Preston County, West Virginia on September 14, 2011. c. The roadside population of P. paniculata PZ11-043 in County, West Virginia. The population had more flower color variation than any other population discovered.

140

a. b.

Figure 1.24: Populations referred to as Phlox carolina ssp. alta in situ. a. The high elevation (ca. 1500 m) habitat of P. carolina ssp. alta PZ11-045 in Haywood County, North Carolina. b. An individual of P. carolina ssp. alta PZ11-045 flowering on September 16, 2011. This specimen does not accurately portray a typical flowering from this population, but demonstrates the potentially long flowering season of this taxon. Plants were seen flowering in this population in mid-July, 2011.

141

a.

b. c.

Figure 1.25: Phlox glaberrima ssp. interior PZ12-063 in situ. a. In Lewis County, Tennessee, P. glaberrima ssp. interior grows in a limestone seep habitat among various species of sedges (Carex spp.). b. An individual of P. glaberrima ssp. interior. Plants were photographed April 27, 2012. There were no flowering plants in this population on this date. c. The paniculate inflorescence of P. glaberrima ssp. interior PZ12-063 flowerig in cultivation at the OPGC.

142

a.

Figure 1.26: Phlox glaberrima ssp. triflora PZ12-046 in situ in Lyon County, Kentucky where it grew in a seepages in mixed hardwood forest.

143

a. b.

Figure 1.27: Phlox maculata PZ10-208 in Adams County, Ohio. a. The roadside habitat of P. maculata. In this area, this taxon occurs in wet ditches and seepages in disturbed sites. b. Flowering individuals of P. maculata on September 1, 2010. Plants can begin to flower in June in this region.

144

b. a.

b. . d.

Figure 1.28: Phlox ovata in situ and in cultivation. a. The population of Phlox ovata PZ12-077 in Allegheny County, Virginia flowering on May 16, 2012. Note the white- flowered variant in the upper left corner of the photo growing among mostly typical pink- forms of the species. b. A close-up of the white flowered variant of P. ovata PZ12-077. This form has been propagated and given the cultivar name ‘White Mountainside’. c. Phlox ovata PZ11-014 flowering in cultivation at the OPGC. d. The foliage of P. ovata PZ12-076. This taxon has the largest leaves of all taxa in subsection Phlox.

145

Figure 1.29: Phlox pulchra in cultivation at the OPGC. This collection was made and distributed by Plant Delights Nursery, Raleigh, North Carolina.

146

a. b.

c.

Figure 1.30: Phlox stolonifera in situ. a. A pink-flowering form of P. stolonifera PZ10- 095 from Hocking County, Ohio. b. A lavender-purple flowering form of P. stolonifera PZ10-125 growing along Panther Creek, in Habersham County, Georgia. c. The distinctive foliage and habit of P. stolonifera distinguishes it from all other species.

147

b.

a.

d.

c.

Figure 1.31: Phlox bifida ssp. bifida PZ10-149. a. The oak savannah habitat of P. bifida ssp. bifida in Newton County, Indiana. b. A typical flowering individual of P. bifida ssp. bifida May 19. 2010. c. Seedlings of P. bifida ssp. bifida PZ10-149 flowering at the OPGC on May 1, 2014. d. Detail of flower color variation among seedlings of P. bifida ssp. bifida PZ10-149.

148

b.

a.

d.

c.

Figure 1.32: Phlox bifida ssp. stelleria in situ and in cultivation. a. The habitat of P. bifida ssp. stelleria PZ10-201 at the type location in Jessamine County, Kentucky. b. In this habitat, plants grow as chasmophytes in rock crevices of dolomite cliffs. c. The cedar glade habitat of P. bifida ssp. stelleria PZ10-018 in Wilson County, TN. d. A typical individual of P. bifida ssp. stelleria PZ10-018 flowering on April 7, 2010.

149

a. b.

d. c.

Figure 1.33: Phlox nivalis in situ and in cultivation. a. The habitat of P. nivalis PZSH2011- along a roadside in Bay County, Florida. Note the flower color variation among individuals. b. An exceptional color flower variant of P. nivalis. c. Phlox nivalis ‘Eco Flirty Eyes’, a selection made by Dr. Don Jacobs in Franklin County, Georgia. d. A plant received as P. nivalis ‘Camla’ that matches the description of Phlox xhenryae described by Wherry (1935).

150

a.

Figure 1.34: Phlox subulata was frequently planted in cemeteries and around former homesteads. In these situations, it is likely that a variety of available cultivars and/or, local selections were planted. Cross-pollination between different forms results in unique populations that can be variable in color and habit and have potential for selection and collection of unique germplasm. Such populations can become adventive and are responsible for the often exaggerated native range of this species.

151

a. b.

d.

c.

Figure 1.35: Phlox subulata ssp. subulata PZ12-066 in situ and in cultivation at the OPGC. a. The serpentine barren habitat of P. subulata ssp. subulata PZ12-066 in Chester County, Pennsylvania. b. A typical flowering individual of P. subulata ssp. subulata PZ12-066 flowering in situ on May 15, 2012. Most flowers had passed by this time, and many plants had mature seed. c. The shale cliff habitat of P. subulata ssp. subulata PZ12-091 on the Delaware/Franklin County, Ohio border. The plants colonize the xeric oak woods on the top of the cliff and cliff edge. d. A typical flowering individual P. subulata ssp. subulata PZ12-091 at the OPGC.

152

b. a.

d.

c.

Figure 1.36: The habitat and flowering characteristics of Phlox subulata ssp. brittonii. a. The habitat of P. subulata ssp. brittonii PZ10-074 is often vertical, shale cliffs, b. A flowering individual of P. subulata ssp. brittonii PZ10- 074 in Botetourt County, Virginia. The flower fade from pale lavender to lavender-white. c. Various accessions of P. subulata ssp. brittonii in cultivation at OPGC showing the lavender-white flower color of this taxon. The bright pink plant in the right is P. subulata ‘McDaniel’s Cushion, and on the left is P. bifida ssp. stellaria ‘Glade Blue’. d. Differences in flower shape among different individuals of P. subulata ssp. brittonii PZ12-074.

153

Eastern U.S.A. Phlox Taxaz,y Section Subsection Species Phlox amoena Sims Phlox divaricata L. ssp. divaricata Phlox divaricata L. ssp. laphamii (Wood) Wherry Phlox cuspidata Scheele Phlox drummondii Hooker ssp. drummondii Phlox drummondii var. peregrina Shinners Phlox drummondii ssp. glabriflora Brand, Pflanzer Phlox drummondii ssp. johnstonii (Wherry) Wherry) Phlox drummondii ssp. mccallisteri (Whitehouse) Wherry Phlox drummondii ssp. tharpii (Whitehouse) Wherry Annuae Divaricatae Phlox floridana Bentham ssp. floridana (Wherry) Prather Phlox floridana Bentham ssp. bella Wherry Phlox pattersonii Prather Phlox pilosa L. ssp. pilosa Phlox pilosa L. ssp. deamii Levin Phlox pilosa L. ssp. fulgida (Wherry) Wherry Phlox pilosa L. ssp. longipilosa (Waterfall) Locklear Phlox pilosa L. ssp. ozarkana (Wherry) Wherry Phlox pilosa ssp. sangamonensis Levin Phlox pulcherrima (Lundell) Lundell Phlox roemeriana Scheele Phlox villosissima Turner

Phlox bifida Beck ssp. bifida Phlox bifida Beck ssp. arkansana Marsh Phlox bifida Beck ssp. stellaria (Gray) Wherry Occidentales Subulatae Phlox nivalis Loddiges ssp. nivalis Wherry (Peter) Wherry Phlox subulata L. ssp. subulata Phlox subulata L. ssp. brittonii (Small) Wherry Phlox subulata L. ssp. setacea (L.) Locklear (syn. australis Wherry)

Phlox carolina L. ssp. carolina Phlox carolina L. ssp. alta Wherry Phlox carolina L.ssp. angusta Wherry Phlox Phox glaberrima L. ssp. glaberrima Ferguson Phlox glaberrima L. ssp. interior Wherry (syn. Ovatae Wherry) Phlox glaberrima L. ssp. triflora (Michaux) Wherry Phlox maculata L. Phlox ovata (L.) Locklear Phlox Phlox pulchra Wherry

Paniculatae Phlox amplifolia Britton Wherry Phlox paniculata L.

Stoloniferae Phlox stolonifera Sims Wherry Cluteanae Phlox buckleyi Wherry Wherry zIncludes all phlox taxa described that occur within the geographical limits described in Figure 1.1. yTaxa are derived mainly from Ferguson et al. (1999), Locklear (2011), Turner (1998), and Wherry (1955).

Table 1.1: Taxonomic arrangement of eastern U.S.A. Phlox species showing the sections, subsections, species, and subspecies recognized.

154

Table 1.2: The number of eastern Phlox taxa listed by state in descending order.

155

Table 1.3: Relative flowering period and time to seed maturity of eastern Phlox taxa. This information derived from field observation of listed populations from 2010-2013.

156

Table 1.4: Taxonomic arrangement, and numbers of wild collected and cultivar germplasm accessions made of eastern U.S.A. Phlox species from 2010-2013.

157

Table 1.5: Botanical and common names of taxa in subsection Cluteanae and number of germplasm accessions obtained from natural plant populations and nursery sources during the development of a Phlox germplasm collection at the OPGC.

158

Table 1.6: Botanical and common names of taxa in subsection Divaricatae and number of germplasm accessions obtained from natural plant populations and nursery sources during the development of a Phlox germplasm collection at the OPGC.

159

Table 1.7: Botanical and common names of taxa in subsection Paniculatae and number of germplasm accessions obtained from natural plant populations and nursery sources during the development of a Phlox germplasm collection at the OPGC.

160

Table 1.8: Botanical and common names of taxa in subsection Phlox and number of germplasm accessions obtained from natural plant populations and nursery sources during the development of a Phlox germplasm collection at the OPGC.

161

Table 1.9: Botanical and common names of taxa in subsection Stoloniferae and number of germplasm accessions obtained from natural plant populations and nursery sources during the development of a Phlox germplasm collection at the OPGC.

162

Table 1.10: Botanical and common names of taxa in subsection Subulatae and number of germplasm accessions obtained from natural plant populations and nursery sources during the development of a Phlox germplasm collection at the OPGC.

163

CHAPTER 2

GENOME SIZE AND PLOIDY LEVEL IN PHLOX SPECIES AND CULTIVARS: UPRIGHT AND LONG STYLED TAXA IN SUBSECTIONS PANICULATAE AND PHLOX

Abstract

Phlox paniculata L. is the most widely grown and intensively hybridized Phlox species, but little is known about variation in genome size and ploidy in this species and related taxa. The objective of this study was to perform an extensive survey of genome sizes and ploidy levels for a diverse collection of cultivars and wild collections Phlox paniculata and related taxa in subsections Paniculatae and Phlox. Holoploid (2C) genome sizes and ploidy levels were estimated by flow cytometry for 84 Phlox cultivars and a collection of 49 germplasm accessions obtained from natural plant populations.

Meiotic chromosome counts were made to calibrate ploidy with genome size for a subset of taxa. The majority of cultivars were diploid (n=7) and had mean genome sizes that did not vary between the two subsections (Paniculatae=14.76 pg, Phlox=14.84 pg), although intra-subsection variation was greater among subsection Phlox cultivars. Four cultivars of P. paniculata had a mean genome size of 22.16 pg and meiotic chromosome counts of n=2x-2=12 and n=2x-3=11, indicating aneuploidy, the first time such aneuploidy has been identified in P. paniculata; other ploidy levels were not observed among cultivars or wild collections. Five tetraploid (n=14) cultivars were found in subsection Phlox, all

164

selections of Phlox glaberrima ssp. triflora, with a mean genome size of 26.91 pg;

chromosome counts in one of these confirmed tetraploidy. Phlox (Suffruticosa Group)

‘Miss Lingard’ has an intermediate genome size of 22.04 pg, consistent with the triploid, hybrid origin of this taxon as previously reported. Mean 2C genome sizes among wild collected accessions were similar to values reported for cultivars (Paniculatae=14.59 pg,

Phlox=14.23 pg), but taxa in subsection Phlox exhibited greater intra-subsection variation. Genome size could not be used to distinguish among taxa in subsections

Paniculatae and Phlox. Two tetraploids were found among wild collected accessions of subsection Phlox; P. pulchra had a mean genome size of 27.16 pg and Phlox glaberrima

ssp. triflora fromm Bath Co., VA had a genome size of 26.65 pg. Genome size and

ploidy reports are the first for the majority of taxa; these results provide further insights

into the relative genome size and cytogenetics of phloxes and are intended to be of use to

plant breeders and systematists.

Introduction

The genus Phlox is an important ornamental group of herbaceous plants that has

been targeted for germplasm development, characterization, and enhancement by the

USDA National Plant Germplasm System. Phlox is a highly heterozygous,

phenotypically diverse genus in the Polemoniaceae with substantial inter-population

differentiation for numerous morphological characters and ploidy; it includes

approximately 65 species with centers of diversity in the eastern and western United

States (Wherry, 1955). The most important horticultural forms of Phlox are found in the

165

eastern species, which are taxonomically distributed among 3 sections (Annuae,

Occidentales, Phlox) and 6 subsections (Divaricatae, Subulatae, Phlox, Paniculatae,

Stoloniferae, Cluteanae) based on variations in the calyx, style length, and geographic

distribution (Ferguson et al. 1999; Wherry, 1955; Chapter 1). Subsections Paniculatae

and Phlox include species that have a style that exceeds 15 mm in length and is as long

as, or longer than, the corolla tube (Figure 1.1; Wherry, 1955). Style length differentiates

the taxa in these two subsections from all other Phlox taxa; collectively they are known

as “long-styled” phloxes. Style length has been implicated in restricting interspecific

hybridization among species of Phlox that differ for the trait, but this has not been

rigorously tested and documented (Levin, 1966; Locklear, 2011a; Symons-Jeune, 1953).

Among the phlox with long styles, P. amplifolia and P. paniculata comprise

subsection Paniculatae, and are distinguished from taxa in subsection Phlox by their

areolate leaf veins, and white or cream-colored pollen (Figure 1.1, Table 2.1; Wherry,

1933; 1955). Subsection Phlox contains as many as 6 species, although the taxonomy of this group has changed several times, and a well-resolved taxonomy does not yet exist for infraspecific variation prevalent in P. carolina and P. glaberrima which is frequently

referred as the P. carolina-P.glaberrima complex (Ferguson et al., 1999; Ferguson and

Jansen, 2002; Wherry, 1932a, 1932b, 1945, 1955). These species are phenotypically

similar to those in subsection Paniculatae, but are differentiated by having yellow pollen

and obscure leaf veins (Figure 1.1, Table 2.1; Wherry, 1955). To some degree, the taxa

in these subsections can be further distinguished from each other using morphology,

geographic distribution, and habitat preference (Table 2.1). The large phenotypic

166 variation within these two subsections has resulted in the selection of many cultivars and some putative interspecific hybrids, but the potential for further breeding and selection in this group has barely been tapped. There is a need to further examine the taxonomic relationships of these ornamental species and to assess the role, if any, that polyploidy had in the development of cultivars.

Differences in ploidy between species are a potent pre-zygotic barrier to hybridization, and are known to affect speciation and evolution by contributing to ecotypic isolation, novel gene expression, and divergence (Eeckhaut et al., 2006;

Hogenboom and Mather, 1975; Parris, et al., 2010; Ranney, 2006). Knowledge of ploidy is crucial for characterization of germplasm collections and for any attempt at interspecific hybridization. While the ploidy of many species has been documented, and the list of species analyzed keeps increasing, many others remain untested (Bennett and

Leicht, 2010). In some genera, the positive correlation generally established between genome size and ploidy has facilitated rapid and extensive study of ploidy variation, not only between species, but also between populations of a species (Greilhuber and Leitch,

2013). Flow cytometry is now well-established as a rapid, high-throughput method for estimating genome size in plants and such estimates can be used to calibrate ploidy, when chromosome counts for various taxa are limited or difficult to generate (Dolezel et al.,

1998; Dolezel, 2009; Parris, 2010). Analyses of genome size in populations of Dianthus broteri (Balao et al., 2009) Phlox pilosa (Worcester et al., 2012), Ranunculus parnassifolius (Cires et al., 2009) and others have demonstrated the existence of intraspecific cytotypic variation between populations that may have an adaptive value

167

(Greilhuber and Leitch, 2013; Smarda and Bures, 2010). Thus, it cannot be assumed that individuals in different populations of a taxon necessarily have the same genome size or ploidy level.

The taxonomy of subsection Paniculatae has remained stable since formal recognition, but phenotypic variation in P. paniculata has resulted in selection of over

800 cultivars (Bendtsen, 2009; Fuchs, 1994; Locklear, 2011a; Symons-Jeune, 1953;

Wherry, 1955). Far more cultivars have been selected from this species than any other

Phlox, and many of these are still grown, but some have been lost to cultivation or reintroduced under different names (Bendtsen, 2009; Hawke, 2011; Hawke, 2013;

Locklear, 2011a; Symons-Jeune, 1953; Wherry, 1955). Cultivars have been generally selected for variations in flower color, duration of flowering period, and foliar variegation; one cultivar, ‘David’, was the Perennial Plant Association Perennial of the year in 2002. This species continues to be the focus of breeding programs. The other species in the subection, P. amplifolia, has rarely been introduced into cultivation, but

European nurseries have described cultivars. Because of the morphological similarities and sympatric relationship of P. amplifolia and P. paniculata populations, the true identity of these cultivars has yet to be demonstrated.

In contrast to the taxonomic agreement on subsection Paniculatae the taxonomic standing of taxa within the morphologically similar P. carolina-P. glaberrima of subsection Phlox have been the subject of debate (Ferguson et al., 1999; Ferguson and

Jansen, 2002; Wherry, 1935; 1955). Wherry (1935; 1955) recognized two primary species, P. carolina and P. glaberrima and six intraspecific taxa that were described on

168

the basis of morphology, geographic distribution, and habitat preference. More recent

molecular analysis of the P. carolina-P.glaberrima complex and related eastern taxa

from other subsections using restriction site data from the internal transcribed spacer

region of ribosomal DNA (ITS) and cpDNA resulted in a paraphyletic grouping, and

included species from at least 3 subsections (Ferguson et al., 1999; Ferguson and Jansen,

2002). These studies underscore the concept of a complex where a single, widespread,

but polymorphic taxon is recognized (the “Phlox glaberrima complex” of Ferguson).

More recently, Locklear (2011a) recognized P. carolina and P. glaberrima as distinct, and described one subspecies for each species based on morphology. In our experience, the dichotomous keys based on morphological features do not consistently separate taxa at the intraspecific level, complicating germplasm collection development and characterization efforts. Thus, there is no resolved phylogeny for morphologically similar taxa in this group. Such confusion may adversely affect breeding efforts and highlights the need for a resolved molecular phylogeny.

A potentially confounding factor in the taxonomic relationship of these species is the recent evidence for the existence of ploidy variation in populations of other Phlox

including tetraploids (2n=4x=28) in P. pilosa and both tetraploids and hexaploids

(2n=6x=42) in P. amabilis and P. woodhousei (Fehlberg and Ferguson, 2012; Worcester

et al., 2012). Polyploid populations of P. pilosa and P. subulata were also discovered

during the course of this study (Chapters 3, 4). In all cases where polyploid Phlox have

been identified, the plants did not exhibit significantly altered morphology (e.g. the

‘gigas’ effect), and could thus be classified as ‘cryptic’ polyploids (Contreras et al., 2009;

169

Fehlberg and Ferguson, 2012; Worcester et al., 2012). Both the taxonomically uncertain

status of P.carolina and P.glaberrima, coupled with the possibility that cytotypic

variation is more common in Phlox than previously thought, suggest that an analysis of genome size and inferred ploidy measurements could further enhance understanding of the breeding potential in Phlox.

Polyploidy has played an important role in the development of many ornamental plant cultivars (Contreras et al., 2009; Parris et al., 2010; Ranney, 2006) but the extent to which it has been used in Phlox is uncertain. There are no published reports on genome sizes for any Phlox cultivars and comparatively few for any Phlox species although chromosome counts have been reported for the principal species (Flory, 1931; Flory,

1934; Meyer, 1944). A survey of the chromosome number of nearly 100 cultivars of P. paniculata using meristematic tissue isolated from branch tips determined that all had the diploid (2n=14) number, but that certain cultivars exhibited varying numbers of chromosomal fragments (Flory, 1931; Flory, 1934; Meyer, 1944). The number of chromosomal fragments varied between squash preparations and was not repeatable which makes the origin, confirmation, and significance of these fragments unclear. There are no published reports of polyploidy in P. paniculata, but only a small percentage of the introduced cultivars has been tested. Observations and comments by various growers and gardeners indicate that fruit and seed production in many of these cultivars is absent to very low, which could be a consequence of different factors, one of them, alterations in ploidy. Sterile flowers are desirable in some ornamental plants, as the reduction in seed set tends to increase the duration of flowering. Therefore, it is plausible that some

170

cultivars of P. paniculata may have altered ploidy. Determination of genome size and

ploidy level in a wide range of cultivars would be of benefit to P. paniculata breeders and

may have the potential to produce breeding lines with novel, useful traits (Hawke, 2011;

Locklear, 2011a).

The objectives of this study were to characterize the diverse germplasm from

subsections Paniculatae and Phlox by determining genome sizes and relationships to ploidy levels in order to 1) develop baseline information about genome sizes and ploidy for use by Phlox breeders; 2) Determine chromosome counts for genome sizes that could not be inferred from those previously reported in the literature.

Materials and Methods

Phlox collection

Living plants were obtained from nursery sources and from collection sites in the

east-central U.S. (Table 2.3, 2.4). Emphasis was placed on gathering a diverse collection

of P. paniculata cultivars readily available in the U.S. Many older cultivars cited in the

literature are either not available in commerce, or have been altogether lost to cultivation.

Cultivar names were verified using a variety of sources (RHS Plant Finder, 2013;

Wherry, 1955). The identity of wild collected material was confirmed using the keys of

Locklear (2011a) and Wherry (1955), information from the phylogenetic analyses of

Ferguson et al. (1999, 2002), and comparison with herbarium specimens from regional

herbaria. Herbarium vouchers are maintained at the Ornamental Plant Germplasm Center

(OPGC), Columbus, OH.

171

Tissue Sampling

All plants were mantained in 16 x 14.5 cm round, plastic containers at the OPGC

greenhouses. These served as tissue donors for flow cytometry and chromosome counts.

Plants were grown on greenhouse benches in a temperature regime of 22.2 ± 6 °C during

daytime and 18 ± 3 °C at night. Nutrients were supplied using slow release fertilizer containing micronutrients (Osmocote 15N-9P-12K; Scotts Co., Marysville, OH). Leaf

tissue for flow cytometry analysis was collected the day of the analysis and consisted of fully expanded, disease-free leaves from the upper one-half of the plants.

Cytological Analysis

Both meiotic (flower bud) and mitotic (root tip) tissues were used to determine chromosome numbers. Root tips were first treated with 0.008 µM 8-hydroxyquinoline

(Fisher Scientific, Suwannee, GA) for 2-4 h at room temperature to accumulate

condensed mitotic metaphase chromosomes, then rinsed with distilled water. Treated root

tips or freshly harvested immature flower buds were fixed in Carnoy’s solution (100%

EtOH: glacial acetic acid, 1:1, v/v) for 1-2 h at room temperature. They were then rinsed

in distilled water and either used immediately or stored in 95% EtOH (v/v) at 2°C.

Excised anthers were placed on a microscope slide; 2 drops of 1% acetocarmine solution

was added, the anthers were macerated with a dissecting needle and allowed to stain for 5

m. A cover slip was then added and slides were given gentle heat over a flame until the

cover slip became cloudy. After cooling, the slide was squashed with direct pressure for 172

at least 1 m before visualization. Before mounting, root tips were hydrolysed 1 M HCl

for 10-20 m depending on taxon; softened root tips were then stained in 1% acetocarmine

solution for a minimum of 1 h. Root tips were trimmed under a dissecting microscope in

2 drops of 1% acetocarmine on a slide and then squashed. A minimum of 10 cells

displaying metaphase chromosomes were used to confirm chromosome counts.

Isolation of Nuclei and Genome Size Analysis

Flow cytometry was used to estimate genome size from fresh tissue of Phlox

species (Arumuganathan and Earle, 1991; Gailbraith et al., 1983; Dolezel and Bartos,

2005). Samples were prepared using the two MgSO4 buffer solutions described by Van

den Engh et al. (1984) and Gailbraith et al. (1997). Propidium iodide (PI; Sigma-Aldrich,

St. Louis, MO) was used as the fluorochrome at a concentration of 5 mgml-1 for analysis

of all samples. Approximately 2 cm2 of each Phlox sample and 0.5 cm2 of Pisum sativum

‘Ctirad’ leaf tissue were placed in a plastic 9cm petri dish and 1.0 ml of cold buffer

added. The pea serves as an internal standard (Dolezel et al. 1998; Dolezel et al., 2007).

The leaf tissue was chopped with a single-edge, straight razor blade in cold buffer until all the tissue had been macerated. The nuclei suspensions were then filtered through a 70

µm nylon cell strainer (BD Falcon, Bedford, MA) and centrifuged at 15,000 rpm to remove debris. The supernatant was discarded and the remaining pellets were re-

suspended in 200 µl of MgSO4 buffer containing DNAase-free RNase A (Sigma-Aldrich,

St. Louis, MO) and PI (100 mgml-1). Samples incubated in a water bath for 15 m at 37

°C to digest RNA. Fluorescence was measured using a BD FACSCalibur Flow

173

Cytometer (Franklin Lakes, NJ). A sample analysis was considered complete once

10,000 events had been reached, and a minimum of 3,000 nuclei were analyzed. At least

3 samples were prepared for each accession. Data collected from each sample was analyzed with Cellquest™ Pro V (BD Biosciences, San Jose, CA). The following equation (Dolezel, 2009) was used to estimate genome size from the flow cytometry measurements: 2C DNA (pg) = {[Sample G1 peak mean x Standard 2C DNA content

(pg)]/Standard G1 peak mean}. A value of 9.09 pg (Dolezel et al., 1998) was used as the standard G1 peak mean (Pisum sativum ‘Ctirad’) 2C DNA content, where “2C” represents the holoploid, or complete, genome size (Greilhuber et al., 2007).

Statistical Analysis

Data for species were subjected to analysis of variance and means separation using the Waller Procedure (Proc GLM, SAS Verison 9.3; Cary, NC).

Results

The analysis of 369 samples representing 123 accessions of 15 Phlox taxa from two subsections (Phlox and Paniculatae; Locklear, 2011a; Wherry, 1955) revealed the presence of plants with DNA content or genome size equivalent to diploid, triploid and tetraploid levels (Table 2.2). The genome sizes did not vary significantly within each of the ploidy levels or between the subsections at each ploidy level except for the wild species in subsection Phlox where the variation between accessions greater than 10%

(24.6%). There was little variation between samples of a single accession, and none of

174

the samples had a coefficient of variation (CV) greater than 5% (Tables 2.3, 2.4). The

mean CV of all 369 samples was 1.65%. The flow cytometry histograms for every Phlox

sample consistently displayed two peaks, and normal readouts included a well-defined G1 peak and a reduced G2 peak for the Phlox, and also a prominent G1 peak of the internal

standard Pisum sativum ‘Ctriad’ (Figure 2.2). Four species with different genome sizes,

P. buckleyi, P. pilosa ssp. pilosa, P. paniculata, and P. subulata, were analyzed

concurrently to verify the position and placement of peaks in relation to each other

(Figure 2.3). Such concurrent analysis of samples in the absence of a reference genome has been shown to reduce variation that may be found in individual samples and reinforces calibration with the internal standard (Dolezel, 2009). The analysis confirmed the spacing of measured peaks with the Pisum standard and correlates with estimated genome sizes in Table 2.3 and Table 2.4.

Subsection Paniculatae

We measured genome size in 85 accessions of species and cultivars from Phlox

subsection Paniculatae; a subset of these was used for chromosome counts to confirm

ploidy levels (Tables 2.3, 2.4). Of the 72 cultivars of P. paniculata and P. xarendsii (P.

paniculata x P. divaricata), sixty eight had a mean genome size of 2C=14.76 pg with a

range of 14.22 pg to 15.57 pg, that varied by 9.50 %; these sizes correspond to a diploid

level (n=7). Chromosome counts from root tips or pollen mother cells confirmed the

diploid level (Figure 2.2). Interestingly, cultivars listed as P. xarendsii have genome size

essentially indistinguishable to that of P. paniculata even though the putative parental

species differ in genome size (Chapter 3). Four cultivars of P. paniculata (‘Blushing

175

Bride’, ‘Dick Weaver’, ‘John Fanick’, and ‘Robert Poore’) had a mean genome size of

2C=22.16 pg, with a range 21.74 pg to 22.57 pg and that varied by 3.8 %. These sizes

are intermediate to the diploid (14.76 pg) and calculated tetraploid (29.52 pg) levels

suggesting they are possible triploids. Analysis of chromosomes in pollen mother cells at

metaphase I showed that chromosome number of these taxa is variable, ranging from

n=2n-3=11 or 2n-2=12, indicating these cultivars are most likely aneuploids (Figure 2.2;

Table 2.3).

A collection of 13 accessions obtained from natural populations of P. amplifolia

(5 different sources) and P. paniculata (8 different sources) was also analyzed for genome size. Mean genome size of P. amplifolia (2C=14.81 pg) was equivalent to that of

P. paniculata (2C=14.50 pg ). Overall, the 13 accessions of these two species had a mean genome size of 14.59 pg with a range of 13.72 pg to 15.22 pg, and 10.80 % variation (Table 2.4). These wild-derived collections are consistent with diploid levels

(n=7). Both the mean and range of genome size found in these samples were equivalent to those of the diploid cultivars. Thus far, tetraploids have not been identified in wild- collected or cultivated accessions of taxa in subsection Paniculatae although artificial

tetraploids have been created (Matiska and Vejsaova, 2010).

Subsection Phlox

The number of cultivars and wild-collected accessions of taxa in subsection Phlox

(sensu Ferguson et al., 1999; syn. subsection Ovatae, sensu Wherry, 1955) is smaller than

that of subsection Paniculatae even though the number of known species in this

176 subsection is greater. Nevertheless, 54 accessions from 5 species were analyzed; 36 accessions were collected from natural plant populations, and 16 accessions were cultivated varieties available in commerce (Table 2.3, Table 2.4). The 5 species differed only slightly in mean genome size: P. carolina (2C=15.30 pg), P. glaberrima

(2C=14.31), P. maculata (2C=14.69), P. ovata (2C=12.44) and P. pulchra (2C=14.40, one accession only). These genome sizes are proportional to a diploid level. Combined, these wild-collected accessions had a mean genome size of 14.23 pg and a range of 12.16

– 16.13 pg. The 24.6% variation in genome size among the diploid taxa in this subsection is larger than that found in subsection Paniculatae. Among the 36 wild populations, we found 2 accessions that had genome sizes suggestive of polyploidy, most likely tetraploid, one in P. glaberrima ssp. triflora (2C=26.65 pg) and the other in P. pulchra

(2C = 27.16pg).

The 16 cultivated varieties examined included cultivars of P. carolina, P. glaberrima, P. maculata, and P. pulchra as well as 2 cultivars in the Suffruticosa Group, described as hybrids between P. carolina and P. maculata (Locklear, 2011a). The cultivars of P carolina, P. maculata, P. pulchra and one of the Suffruticosa Group

(‘Monica Linden Bell’) had genome size equivalent to the diploid level (2C=14.78pg; range: 13.96-15.66 pg; 10.8% variation). The 8 cultivars of P. glaberrima differed in genome size such that 3 of them had diploid-equivalent genome sizes (2C=14.93 pg) and the remaining 5 had tetraploid-equivalent sizes (2C=26.40 pg). All five polyploid cultivars could be keyed to P. glaberrima ssp. triflora. Chromosome counts confirmed that these accessions were tetraploid (n = 14). Phlox (Suffruticosa Group) ‘Miss

177

Lingard,’ had a mean genome size of 22.04 pg, intermediate between a diploid and

tetraploid level. Previous studies (Meyer, 1944) indicated that this taxon is triploid; the

intermediate genome size and chromosome count indicate that this cultivar is an aneuploid (n = 2n-2 = 12; Table 2.3).

Discussion

The goal of this study was to determine the genome size and ploidy of a closely related group of Phlox species and cultivars that are part of a developing germplasm collection at the Ornamental Plant Germplasm Center. This group of species is classified within subsections Phlox and Paniculatae, all eastern North American natives characterized broadly, among other traits, by having long-styles. The characterization of genome size within this long-styled group has been a prelude to interspecific hybridization studies that are part of a Phlox germplasm enhancement program (Chapters

5-7). The following conclusions may be drawn form this study; 1) Wild populations of the 7 species surveyed were primarily diploid, but 2 populations of tetraploids were identified, one in P. glaberrima ssp. triflora, the other in P. pulchra. 2) The 2C genome

size for the taxa in the two subsections is similar, in the range of 2C=13-15 pg, but

consistent differences among some of the species provide additional evidence for

distinctiveness of the taxa although the taxonomic uncertainty of the P.

carolina/glaberrima complex is borne out by the similar, yet variable size among the

accessions. 3) Cultivars of taxa in the subsections were diploid, tetraploid and aneuploid,

although the majority was diploid. More than 90% of the P. paniculata cultivars

178 examined were diploid, but four novel aneuploid cytotypes were identified. 4) Cultivars of P. glaberrima have apparently been selected at the tetraploid level. Although polyploidy appears to be rare in subsection Phlox, the discovery of polyploid cultivars suggests that increased ploidy may be associated with enhanced ornamental attributes, evidenced by the inadvertent selection of such lines, and may have greater, perhaps adaptive, importance among these taxa than previously described.

The Phlox species analyzed in this study represent the most diverse group within this genus whose genome sizes have been estimated to date. Our results show that the 1C genome size of Phlox, in the range of 4-13 pg (Table 2.3, 2.4; Chapter 3), is consistent with that of the vast majority of angiosperms whose average 1C value is 5.9 pg; although there is a nearly 2,400-fold range in genome size within the clade (Leitch and Leitch,

2013). The few reports that have examined genome size in other species of Phlox show very similar results: P. pilosa of subsection Divaricatae has a diploid genome size of 2C

= 9.2 – 13.3 pg (Worcester et al., 2012), and P. amabilis-P. woodhousei of subsection

Speciosae has a 2C= 8.36-9.01 pg (Fehlbergh and Ferguson, 2012). Within the

Polemoniaceae, the only other reports of genome size include squarrosa

(1C=1.32 pg) and Collomia grandiflora (1C=2.08) (Bennett and Leitch, 2012). Thus, this study adds significantly to our knowledge of genome size in Phlox, both for species forms and cultivars.

Mean genome sizes of the two species in subsection Paniculatae were indistinguishable (2C=14.50 pg for P. paniculata and 2C=14.81pg for P. amplifolia) whereas genome size variation in subsection Phlox was more pronounced. While the

179

average 2C values at the diploid level (2x=14) for P. carolina, P. glaberrima, P. maculata and P. pulchra were within the range of 14.3 – 14.7 pg, the variation within the

P. carolina-P.glaberrima complex, at 20.5%, represents considerable within-species

variation and may be related to morphological similarities and ecotypic divergence seen

in these and other taxa (Ceccarelli et al., 2011; Cires et a., 2009). Such variation

underscores the taxonomic uncertainty in this group and precludes use of the estimates of

genome size to distinguish between members of the complex. Comparable variation in

genome size was also found among taxa in the P. pilosa complex, but not among other taxa analyzed in this study (Table 2.2, 2.3; Chapter 3; Worcester et al., 2012). In contrast, a study of genome size in Penstemon showed that it was possible to use this

measure to distinguish among the many species in this large genus (Broderick et al.,

2011). Several factors appear to contribute to cytotype variation within a taxon, but the

differential proliferation of transposable elements, in particular LTR (long terminal

repeat)-retrotrasposons, can develop large copy numbers in non-protein coding regions of

plant genomes and represent a considerable portion of the large genomes found in some

plant species (Grover and Wendel, 2010). Breeding system, chromosome pairing, and

effective population size can exaggerate the effect of transposable elements within the

genome of a given species, particularly in out-crossing species, with variable effective

population sizes. An outcrossing species with variable effective population in

heterogeneous environmental conditions may experience differential selection pressure

that can result in fine-scale accumulation or deletion of transposable elements at different

rates (Grover and Wendel, 2010). Furthermore, evidence from Arabidopsis suggests that

180

genome size variation is generated by lineage specific differences in the molecular

mechanisms of DNA amplification and removal, creating variation in DNA content that

can serve as a template for genotypic and phenotypic selection in diverse environments;

changes in genome size may be the result of heritable “bursts” of DNA accumulations

and deletions that can take effect within a single generation (Bennetzen et al., 2005;

Ceccarelli et al., 2011). In other species, similar interpopulation differences in genome

size were correlated with phenotypic changes and adaptation to different environments.

This suggests that the extensive phenotypic and genetic diversity in the P. carolina-

glaberrima complex may have resulted from similar genetic and environmental

processes, and represents an important mode of evolution in this group. Phylogenetic

resolution of this group may require further analysis of the mechanisms that contribute to

genome size variation, and selection of appropriate markers to objectively compare

differences. The taxonomy used to describe the accessions in this study is purely provisional, and pending a well-resolved phylogeny of Phlox, should be interpreted as major phenotypic groups.

The analysis of genome size of P. paniculata cultivars revealed that 4 of 66 cultivars (6%) had genome sizes larger than diploid and the chromosome analysis indicated these were aneuploid. Such aneuploid cytotypes in P. paniculata have not been previously described, but evidence from related species and from hybridization experiments suggests they are derived from crosses between diploid and tetraploid cytotypes (Fehlberg and Ferguson, 2012; Flory, 1933; Chapter 5). A tetraploid cytotype of P. paniculata was not discovered among the wild collected accessions analysed in this

181 study, but population sampling was not exhaustive (Table 2.4). The recent discovery of polyploid populations of other eastern Phlox taxa in related subsections and that occur at the edges of their respective geographic distribution suggests that targeted sampling at the edges of the geographic range of P. paniculata may reveal polyploid populations

(Fehlberg and Ferguson, 2012; Worcester et al., 2012; Wright and Ferguson, 2013;

Chapter 3). The east-west geographic distribution of P. paniculata stretches from the eastern Pennsylvania/ New Jersey border to Northwestern Arkansas; this species can be abundant in certain localities and extensive phenotypic variation has been noted in wild populations (Locklear, 2011a; Symons-Jeune, 1953; Wherry, 1955). Wild sourced polyploid germplasm, should it be found, could be used to verify the products of F1 interploid crosses and have the potential to benefit Phlox breeding programs.

Introduction of wild collected germplasm can also result in genotypes not currently in cultivation and can be useful in systematic studies to discern the evolutionary history of the species.

The increase in genome size of aneuploid P. paniculata cultivars is due to the presence of additional chromosomes, not chromosome fragments. Mitotic and meiotic chromosome counts performed for over 140 cultivars and accessions of P. paniculata revealed that all tested taxa were diploids (Flory, 1931; Flory, 1934; Meyer, 1944; Smith and Levin; 1967). One study reported diploid chromosome counts for 126 taxa, but noted the addition of 1-13 chromosomal “fragments”. These fragments were found in shoot tip preparations, a less common tissue source for chromosome analysis, and it was admitted that the number of fragments per cell varied and was not repeatable between different

182 preparations (Meyer, 1944). We could locate only four of the cultivars (‘Leo P.

Schlageter’, ‘Jules Sandeau’, ‘New Bird’, and ‘Rijnstroom’) used in those previous studies for the current study. Of these, only ‘Leo P. Schlageter’ exhibited the chromosome fragments in that study (Meyer, 1944). The genome sizes of these cultivars fell within the range of diploid P. paniculata and meiotic chromosome count for ‘Leo P.

Schlageter’ confirmed a count of n=7; no chromosome fragments were observed. In a separate karyotyping study using root tip preparations, chromosome fragments were not observed among the three accessions of P. paniculata analyzed (Smith and Levin, 1967).

It appears that the original report of chromosome fragments may have been an artifact of the branch tip preparations and suggests chromosome spreads generated from anther or root tip squashes are preferable for determining chromosome counts of Phlox taxa.

An interesting observation about the aneuploid cultivars is that ‘Robert Poore’ and ‘John Fanick’ are consistently rated among the P. paniculata cultivars most resistant to powdery mildew (Hawke, 1999). Susceptibility to powdery mildew (Erysiphe cichoracearum) is the foremost factor limiting more widespread usage of P. paniculata in landscapes, but despite the popularity of this species, a definitive mode of resistance has yet to be identified (Hawke, 1999; Jorosz et al., 1982). It is noteworthy that alteration in ploidy may influence plant-pathogen interactions, thus indicating a potential avenue for exploring powdery mildew resistance. There are highly mildew resistant diploid cultivars such as ‘Shortwood’, ‘David’, and ‘Delta Snow’ (Hawke, 1999; 2011), so resistance in this species may have different pathways that potentially involve morphological adaptations, physiologic alterations, or specific R-gene action (Jorosz et al., 1982). While

183

no studies directly link polyploidy to disease resistance in plants, genetic variation

generated through interploid crosses may provide a strategy to develop mildew

resistance. Data from a variety of ornamental and agronomic crops indicate that

resistance to this obligate pathogen is controlled by one to several genes, and that

different genes may be responsible for resistance of different plant organs (Cohen et al.,

2003; Kessler et al., 2010). Further testing for modes of mildew resistance is needed for

P. paniculata, but the results presented here suggest that polyploidy breeding may offer

an option.

Imbalanced polyploids, such as triploids and pentaploids are rare in Phlox

(Fehlberg and Ferguson, 2012; Flory, 1933; Meyer, 1944). Only the hybrid Phlox

(Suffruticosa Group) ‘Miss Lingard’, reportedly the product of a cross between P.

carolina x P. maculata, has been shown to be triploid (Meyer, 1944) although the

parentage has not been confirmed (Locklear, 2011a). Our analysis of genome size for

this cultivar shows that the 2C=22.04 pg size is consistent with a triploid level,

supporting a hybrid origin from a cross of a tetraploid and diploid parent. While no

tetraploids were found among our survey of accessions from wild populations and

cultivars of plants keyed to P. carolina and P. maculata, the morphological similarity and

historical taxonomic confusion of taxa of the Phlox glaberrima complex suggests that

‘Miss Lingard’ may have resulted from a cross between a tetraploid P. glaberrima and diploid P. maculata. Tetraploid cytotypes were found among P. glaberrima cultivars tested in this study and are known to occur in natural populations in the southern

Appalachian Mountains; it is possible that early horticultural introductions originated in

184 this area (Wherry, 1955). The hybrid origin of Phlox ‘Miss Lingard’ is also supported by pollen sterility in the tested clones and failure to set seed in hybrid crosses (Wherry,

1935).

The cultivars listed as Phlox xarendsii had genome size that was the same as the rest of P. paniculata cultivars. This hybrid species is considered to have originated from a cross of P. paniculata with P. divaricata (Arends, 1912; Wherry, 1955). These species differ in genome size with P. divaricata having a 2C size of 9-10 pg (Chapter 3) which is approximately 30% lower than P. paniculata. Controlled hybridization between these two species (Chapter 5) shows the hybrids to have a genome intermediate between the parents and are pollen sterile. Thus, it appears that current cultivars labeled as P. Xarendsii may either be variants of P. paniculata or are backcross progeny to P. paniculata; this is further supported by the observation that viable seed was produced after open-pollination in such cultivars (Wherry, 1955). Despite this information, only a small subset of P. xarendsii cultivars was tested. Many more exist and it is possible that some may represent interspecific hybrids or P. paniculata selections derived from multiple origins.

There has been one previous report of polyploidy in taxa of subsection Phlox, a tetraploid individual of P. glaberrima collected from the southern Appalachians

(Ferguson, 1998). Our study has found 6 tetraploid accessions in the subsection: five are identified as P. glaberrima ssp. triflora, and one as P. pulchra. This is the first report of polyploidy among cultivars of subsection Phlox, and the first report of a tetraploid cytotype of P. pulchra (Table 2.3). The tetraploid P. pulchra has the largest genome size measured to date (2C= 27.16 pg) among a large survey of such genomes in wild and

185

cultivated Phlox (Chapter 3, 4). The P. glaberrima tetraploid cultivars ‘Anita Kistler’

and ‘Morris Berd’, are known to have been collected directly from natural populations

and the remaining tetraploid cultivars are not likely to be more than one generation

removed from wild collected progenitors (Bendtsen, 2009; Fuchs, 1994). All of these

selections were made without knowledge of ploidy, and suggests inadvertent selection for

polyploid plants due to novel phenotypic traits or increased adaptation and persistence in

a range of environmental conditions. Furthermore, all of the P. glaberrima cultivars can

be attributed to the morphologically distinct taxon P. glaberrima ssp. triflora (Table 2.1).

Together with P. ovata and P. pulchra, this taxon displays mid-spring (May-June) phenology, a proliferation of sterile, evergreen stems, and a preference for xeric habitats

(Table 2.1). The similarities of these species, and differences from others in subsection

Phlox, have been suggested by other authors; the presence of polyploidy among these, and not other taxa in subsection Phlox, suggests that ploidy may be an additional differentiating factor between these and other species (Michaux, 1803; Ferguson et al.,

1999; Smith and Levin, 1967). These discoveries highlight the importance of introducing wild collected Phlox germplasm to increase the genetic diversity of available ornamental crops and identify avenues for polyploid breeding of Phlox.

Identification of aneuploid P. paniculata cultivars indicates that ploidy manipulation has already occurred in Phlox, either deliberately or inadvertently. Induced tetraploids have been developed in P. drummondii (Raghuvanshi and Pathak, 1975;

Tiwari and Mishra, 2012; Vyas et al., 2007), P. paniculata (Matiska and Vejsadova,

2010) and in P. subulata (Zhang et al., 2008), but the extent to which these have

186 contributed to cultivars is unclear. There is room for further experimentation and manipulation of polyploidy in Phlox, perhaps to alter flowering characteristics (flower size, longevity, flowering period) and size/vigor. Sterile interspecific hybrids may also be restored to fertility by allopolyploidization (Leus et al., 2012).

Conclusion

This study provides new and pertinent information on genome sizes and ploidy levels for cultivars, species, and hybrids of long-styled phloxes. This data also indicates that flow cytometry can be used to confirm hybridity among phloxes with different genome sizes and ploidy. These results provide further insight into the cytogenetics, systematics, reproductive biology, and crossability of phloxes and contribute to the larger census of angiosperm genome sizes.

187

b d a e

c.

1

f. g. h. i. j. k. l.

Figure 2.1: a-l. Leaf, pollen, and style length variation of Phlox taxa. a. P. paniculata PZ10-209. Note the areolate leaf veins. b. P. carolina ssp. alta PZ10-045. Note obscure leaf veins. c. P. glaberrima ssp. triflora PZ10-193 diploid (n=7). d. P. glaberrima ssp. triflora PZ11-019 tetraploid (n=14). e. P. maculata PZ12-107. f-i. Differences in pollen color between subsection Paniculatae and Phlox. f. The yellow pollen of P. glaberrima ssp. triflora. g. The yellow pollen of P. maculata PZ12-107. h. The cream colored pollen of P. paniculata PZ10-209. i. The cream colored pollen of P. amplifolia PZ11- 050. j-k. Differences in style length among subsections. j. Subsection Paniculatae and Phlox. k. Subsection Subulatae. l. Subsection Divaricatae. The scale bar is 1 cm.

188

a. b. c. d.

Number of nuclei Number

Relative DNA fluorescence

Figure 2.2: Concurrent comparison of four Phlox species from three sections and four different subsections of the genus; a. diploid P. subulata, b. diploid P. pilosa ssp. pilosa, c. diploid P. paniculata, and d. tetraploid P. buckleyi (from left to right). G1 peak placement and values support data from analyses with an internal standard.

189

Figure 2.3: Determination of DNA content and ploidy level of diploid (2n) and aneuploid (2n-2) cultivars of Phlox paniculata and P. amplifolia using flow cytometry with Pisum sativum ‘Ctirad’ as the internal standard. a. Histogram and meiotic metaphase chromosome counts for the diploid (n=7) P. paniculata ‘Delta Snow’ (14.90 pg), b. Histogram and mitotic metaphase chromosome counts (400X) for the diploid (2n=2x=14) P. amplifolia PZ11-050 (15.21 pg), and c. Histogram and meiotic metaphase chromosome count of the aneuploid (n=12) P. paniculata ‘John Fanick’ (22.57 pg) are shown.

190

Table 2.1: Distinguishing vegetative and floral morphological characteristics and habitat preferences of taxa from Phlox subsections Paniculatae and Phlox as described by Wherry (1955) that were used in this study.

191

Table 2.2: Summary of means and ranges of 2C Holoploid genome size (pg) of Phlox species grouped by section, subsection, and ploidy level.

192

(Continued)

Table 2.3: Relative Holoploid (2C) genome sizes and ploidy levels, determined using flow cytometry and chromosome counts, for a diverse collection of cultivars and hybrids of Phlox.

193

Table 2.3: Continued

(Continued)

194

Table 2.3: Continued

195

(Continued)

Table 2.4: Relative Holoploid (2C) genome sizes and ploidy levels, determined using flow cytometry and chromosome counts, for a diverse collection of Phlox germplasm accessions from natural plant populations. 196

Table 2.4: Continued

197

CHAPTER 3

GENOME SIZE AND PLOIDY LEVELS IN PHLOX GERMPLASM: MAT FORMING TAXA FROM EASTERN AND WESTERN NORTH AMERICA

Abstract

Phlox subulata L., creeping phlox, is an eastern North American native species that is also a widely cultivated groundcover with prolific flower displays and a long history of breeding and selection. There is very limited information about the genome size and ploidy of P. subulata and related taxa from eastern and western North America.

This information could be used to inform conservation efforts, germplasm collection, and further breeding. The objective of this study was to analyze genome size and ploidy using flow cytometry in a diverse germplasm collection consisting of accessions from natural plant populations, cultivars, and hybrids obtained from nursery sources. Nuclear genome sizes (2C-values) and ploidy levels were determined for approximately 10% of the genus Phlox represented by 52 accessions of 11 morphologically similar Phlox species, cultivars, and hybrids from 3 sections (Annuae,Phlox, Occidentales), 5 subsections (Albomarginatae, Canascentes, Longifoliae, Speciosae, Stoloniferae,

Subulatae), and the closely related Microsteris gracilis using Pisum sativum ‘Ctirad’ as a reference genome. Three different mean genome sizes were found across three (2x=14,

4x=28, 6x=42) ploidy levels for all species, however, hexaploids were limited to one

198

population of P. subulata, which was part of a polyploid complex from the eastern

geographical limit of the species range. The mean genome size for diploid taxa in subsections Canascentes, Longifoliae, Speciosae, and Subulatae ranged from 8.05-9.30

pg whereas the genome of species in subsections Albomarginatae and Stoloniferae was

larger, 10.41-11.38 pg. Tetraploids were identified in two taxa from subsections

Canascentes and Subulatae; genome size was similar and ranged from 15.34-15.86 pg.

All 25 cultivar accessions were diploid, with mean genome size from 7.17 pg to 8.50 pg.

The interspecific hybrid Phlox xprocumbens has a mean genome size of 9.12 pg which is

intermediate between the parental taxa P. stolonifera x P. subulata. Genome size and

ploidy variation was greater among accessions from natural plant populations than

cultivated taxa. The genome sizes of taxa in subsection Subulatae are comparable to

those of western taxa with similar morphology, suggesting that these taxa are closely

related. This survey provides the first report of a hexaploid in P. subulata, and the first

for any eastern Phlox species.

Introduction

Phlox species, cultivars, and interspecific hybrids are widely grown as utilitarian,

flowering perennials that serve a variety of landscape functions (Bendtsen, 2009; Fuchs,

1994; Locklear, 2011a). The genus has important economic value; according to the

Census of Horticultural Specialties (USDA-NASS, 2010), sales of phlox plants in 2009

amounted to more than $16.2 million in the United States. Phlox has been recognized as a

genus with significant untapped horticultural potential and the National Plant Germplasm

199

System designated it a priority for germplasm development. Among the various

cultivated Phlox, the “creeping phloxes” are a morphologically distinct group of

evergreen subshrubs grown for their luxuriant, mat-forming growth habit, and “carpet of

color” flower displays. As many as 80 cultivars of “creeping phlox” have been

developed, and most are attributed to P. subulata; however, the breeding history of these

selections is largely anecdotal although it is apparent that the various species in this

horticultural category are related and likely compatible (Bendtsen, 2009; Foley, 1972;

Fuchs, 1994; Locklear, 2011a). Genome size and ploidy information about phlox

cultivars and related species are important characteristics of germplasm that could guide

plant breeders in appropriate manipulation and also provide insight about the history of

hybridization among various taxa.

Phlox (Polemoniaceae) is a phenotypically and ecologically diverse genus of ca.

65 species native to temperate regions of North America, with one outlying species

occurring over a vast area of Russia (Ferguson et al., 1999; Locklear, 2011a; Wherry,

1955). Within the United States there are two centers of diversity (Fig 1.1; Chapter 1);

the eastern region includes 20-23 species, and the rest in the western region (Ferguson, et

al., 1999; Wherry, 1955). Wherry, (1955) divided the genus into 3 sections and 16

subsections on the basis of morphology and geographic distribution, but admitted that his

classification was “rather artificial” and that certain taxa “bridged key gaps” between subsections. In Wherry’s classification, the eastern species are grouped into sections

Annuae and Phlox and further assigned to 7 subsections (Table 1.1; Chapter 1). Among these, subsection Subulatae contains 4 phenotypically distinctive species (P. bifida, P.

200

nivalis, P. oklahomensis, P. subulata) and up to 8 subspecies, all known commonly as the

creeping or moss phloxes. These taxa are distinguished from other eastern species in

being suffrutescent, caespitose or pulvinate, sclerophyllous subshrubs that are

superficially similar to many western taxa; all taxa exhibit a carpeting or mounding habit

often described as “mat forming” (Ferguson et al., 1990; Ferguson and Jansen, 2002;

Wherry, 1929; Wherry, 1955). Initially, Wherry (1929, 1955) split these taxa among

subsections Speciosae and Subulatae on the basis of differences in style length (Wherry,

1955). Phylogenetic analysis of ITS and cpDNA restriction sites revealed that all eastern

creeping phlox species formed a well-supported monophyletic group, regardless of style

length variation, and there was strong support for the placement of subsection Subulatae,

which were nested among a clade containing western taxa with similar morphology

(Ferguson and Jansen, 2002). Although the study supports the evolutionary relationship

of eastern and western mat forming taxa, few western taxa were used in the analysis and

the phylogenetic relationships of eastern and western Phlox species have yet to be thoroughly analyzed and resolved.

Polyploidy has been a potent mechanism in genetic diversification, reproductive

isolation and genetic divergence, implicated a well in adaptation to marginal ecosystems

and exploitation of new ecological niches (Levin, 1983; Soltis et al., 2009; Theodoridis et

al., 2013). Polyploidy is also an important factor in plant breeding because it can

influence reproductive compatibility, fertility, and expression of phenotypic traits

(Ranney, 2006). Historically, polyploidy was considered rare in the genus Phlox (Flory,

1933, 1934), but recent studies have indicated that it is more prevalent among Phlox taxa

201

than previously reported (Fehlberg and Ferguson, 2012; Flory, 1933, 1934; Meyer, 1944;

Worcester et al., 2012). Nevertheless, information about genome size and ploidy is not

available for most Phlox species, and there are no reports of genome size for Phlox cultivars (but see Chapter 2).

Within subsection Subulatae and related western taxa, few species have had detailed studies of chromosome counts and general ploidy levels (Flory, 1934; Meyer,

1944). Phylogenetic analyses of the genus Phlox based on molecular markers, led

Ferguson et al. (1999) to conclude that “More detailed examination of phylogenetic

relationships with relation to chromosome numbers is warranted…” particularly in

reference to subsection Subulatae and the potential phylogentic relationship of

morphologically similar western taxa. In more recent population studies of eastern and

western taxa, cytotype variation has become evident, with diploids, tetraploids, and

hexaploids identified even with limited sampling, and suggests that there may be similar

ploidy variation in the related subsection Subulatae (Fehlberg and Ferguson, 2012;

Worcester et al., 2012). Chromosome counts done in the 1930’s indicated that two

cultivars of P. subulata and two wild collections labeled as P. subulata brittonii were all

diploid (Flory, 1931; 1934). The most comprehensive karyological studies of P. subulata

to date showed that both diploid and tetraploid forms of the species exist and that

specimens labeled as P. nivalis x P. subulata hybrids were both diploid and tetraploid

(Meyer, 1944). However, there was no locality information for the collections and

cultivars were not identified, therefore the data could not be used to delineate ploidy

patterns of wild populations or cultivated taxa. A later study reported the chromosome

202 number for a collection from Coshocton County, Ohio as “14+28”, suggesting a mixed ploidy population, but the significance of this find was not discussed in detail or in relation to ploidy variation in other taxa (Smith and Levin, 1967). A few studies have indicated that ploidy variation among western species is greater than in eastern species, and there are reports of tetraploids and hexaploids (Eater, 1967; Flory, 1937, 1948;

Strakosh, 2004). Most notably, a study of genome size and ploidy variation in the P. amabilis-P. woodhousei complex revealed extensive variation in ploidy levels, ranging from diploid to hexaploid, with a serial increase in genome size (Fehlberg and Ferguson,

2012). It is evident that ploidy can be quite variable among western taxa and it is possible the same degree of variation might be found among taxa in subsection

Subulatae.

Taxa is subsection Subulatae occur in xeric habitats associated with rock outcroppings, barrens, and deep sand deposits, or short-grass prairies although some forms of P. nivalis can inhabit more mesic environments (Marsh, 1960; Springer and

Tyrl, 1989; Wherry, 1929, 1955). These habitats experience a variety of disturbances

(fire, shifting soil, floods, etc.) that eliminate competition from sympatric plant species, and results in increased solar radiation (Wherry, 1929). Preference for this type of habitat by Phlox taxa results in localized, fragmented populations that are surrounded by stretches of unsuitable habitat within comparatively large overall geographic ranges for most species (Fehlberg and Ferguson, 2012; Marsh, 1960; Wherry, 1929). Since Phlox are highly heterozygous, out-crossing species, the island-like distribution of populations within these ranges may limit gene flow and contribute to genetic diversification that may

203 impart adaptive advantage to species in marginal habitats or in regions of physiographic diversity.

Creeping habit is not restricted to taxa in subsection Subulatae. The two species that comprise subsection Stoloniferae, P. stolonifera, and P. adsurgens also exhibit a caespitose habit, but have very distinct leaf and flower characteristics that distinguish them from other creeping Phlox (Locklear, 2011a; Wherry, 1955). Whereas P. stolonifera is an evergreen species that inhabits mesic sites in climax deciduous forest within the

Allegheny Plateau and Appalachian Highland provinces, P. adsurgens is endemic to the

U.S. Pacific Northwest (Wherry, 1955). Molecular and karyotypic analyses indicated that the phylogenetic placement of P. stolonifera is basal to the rest of the genus; however, the taxonomic placement of this species remains unclear. Curiously, both taxa in subsection Stoloniferae have been successfully hybridized with taxa from subsection

Subulatae. Two hybrids have been reported: P. stolonifera x P. subulata, known as

Phlox xprocumbens, and P. adsurgens x P. nivalis, known as Phlox xoliveri (Locklear,

2011a; Wherry, 1935a). Flory (1934) studied meiosis of P. xprocumbens and discovered several irregularities that are consistent with plantas of interspecific hybrid origin.

The goals of this study were to determine 2C holoploid genome sizes and ploidy levels of caespitose taxa in Phlox subsections Subulatae and Stoloniferae, as well as some species from the western U.S. The taxa sampled included cultivars and interspecific hybrids from subsection Subulatae.

204

Materials and Methods

Phlox collection

Live plants were obtained from nursery sources and from collection sites in the

east-central United States visited during a series of germplasm collection expeditions

from 2010 to 2013. Taxa collected from natural plant populations were identified using

the keys of Locklear (2011a) and Wherry (1955). Cultivars and hybrids were obtained

from nursery sources. Over 80 cultivars of P. subulata and its hybrids have been

described, and an attempt was made to acquire a representative sample of readily

available selections from U.S. nurseries (Bendtsen, 2009; Fuchs, 1994; Wherry, 1935a).

Herbarium vouchers for the different taxa in this study are held at the Ornamental Plant

Germplasm Center (OPGC) in Columbus, OH.

Tissue Sampling, flow cytometry analysis, cytological analysis, statistical analysis

Tissue sampling, flow cytometry analysis, cytological analysis, and statistical

analysis are described in Chapter 2.

Results

The 12 species surveyed for genome size in this study included members from the

3 Phlox sections and from 6 subsections that included 28 taxa of wild origin as well as 25 cultivars. The group shared a common caespitose and/or creeping habit although they originated from diverse habitats and regions of North America. The genome size measurements revealed that the most frequent ploidy level was diploid (2n=2x);

205

polyploids were identified only in subsections Canescentes and Subulatae, and there were significant differences in genome sizes among the three different ploidy levels (2x-4x-6x) identified (Table 3.1). The 4 species that differed from diploid were P. opalensis and P. pungens from Section Occidentales, Subsection Canescentes, and P. subulata and P. nivalis from Section Phlox, Subsection Subulatae. In the diploid (2x) forms, the genome size ranged from 2C= 7.74 pg to 12.54 pg; the smallest size in this Phlox group, 7.89 pg, was 24% higher than from that of the outlying genus, Microsteris, 5.86 pg (Table 3.1).

The smallest genome at the 2n=2x level in this survey was found in P. subulata ssp. subulata whereas the largest was in P. adsurgens (12.54 pg). The high repeatability and consistency of the flow cytometric measurements (Figure 3.1) indicate that these differences reflect the distinctiveness of the corresponding taxa.

Western U.S. mat forming Phlox

Our sampling of Phlox from the western U.S. was constrained by limited availability and we could only obtain single samples for the 7 species tested; these included representatives of the three Sections of the genus and from subsections

Albomarginatae, Canescentes, Longifoliae and Speciosae (Table 3.2). Within Section

Occidentales, the only species from subsection Albomarginatae available for analysis was P. alyssifolia; its mean genome size of 2C=10.41 pg was the highest among western mat-forming taxa at the 2n=2x level. Four species from subsection Canascentes were analyzed. The widely distributed taxa P. austromontana and P. muscoides had a mean genome size of 2C=8.05 pg and were diploid (2n=2x=14). These species occur over

206 large geographic areas that include a wide range of elevations. In contrast, the narrowly endemic and more recently described species P. opalensis and P. pungens (Locklear,

2011a) had a mean genome size of 2C=15.87 pg and were tetraploid (2n=4x=28). These species have restricted geographic ranges and occur within narrow altitudinal ranges. No reports of chromosome counts exist for these species, so additional accessions from these and similarly narrow endemic species are needed to confirm the distinctiveness of these data.

The sole Western U.S. representative of section Annuae in this study was P. woodhousei from Subsection Speciosae. This taxon had a mean genome size of 2C=8.15 pg that is very similar to the mean value (2C=8.88 pg) reported for individuals from a population in Coconino County, Arizona in a study of genetic structure of the P. amabilis-P. woodhousei complex (Fehlberg and Ferguson, 2012).

A single sample of the Pacific Northwest-native P. adsurgens was included in this analysis. This species, together with P. stolonifera, are the only members of subsection

Stoloniferae. The cultivar P. adsurgens ‘Wagon Wheel’ was sampled and found to have a genome size of 2C=12.54 pg, the largest of the diploid western USA Phlox we surveyed.

Within the primarily eastern U.S. section Phlox, one western species, P. stansburyi of subsection Longifoliae, was sampled for analysis. The mean genome size of

9.30 pg was unique among other western taxa sampled in this study, and most similar to values reported for P. divaricata. This limited information highlights the need for genus- wide genome size sampling of Phlox.

207

Eastern U.S. mat forming Phlox

The eastern taxa in section Phlox represent the largest group of accessions in the germplasm sampled for this study. Different populations of the species were surveyed

(Table 3.2). All 5 accessions of P. stolonifera (subsection Stoloniferae) collected from wild populations were diploid (n=7) with a mean genome size of 2C=11.15 pg. Three cultivars of this species had a mean genome size of 2C=11.14 pg (Table 3.3). This is the first report of genome size of P. stolonifera and the consistency of the 2C value in both wild populations and cultivars suggests that the genome size of this species is relatively invariant.

Subsection Subulatae contains four species and up to 8 subspecies that include P. subulata, often called the moss-phlox, the most commonly grown horticultural form within the mat-forming taxa. Most accessions in this subsection were diploid (n=7) with a mean genome size of 2C=7.78 pg. The diploid genome in P. subulata ssp. subulata was remarkably consistent, with mean size among accessions that ranged narrowly from 7.74-

7.86 pg (Table 3.2). Three polyploid populations were found in this taxon. One accession of P. subulata ssp. subulata was tetraploid (2n=4x=28) with a genome size of

15.65 pg, and another was hexaploid (2n=6x=42) with a genome size of 23.72 pg (Figure

3.1). One population (PZ10-049) collected from a roadside in Scioto County, Ohio had both diploid and tetraploid cytotypes; such mixed ploidy populations were previously described by Smith and Levin (1967) for this species but the significance was not addressed. This population was found near an abandoned homestead and may be

208 adventive, perhaps indicating that ploidy may influence the ability of cultivated plantings to escape from constructed landscapes, and persist in marginal habitats with human- mediated disturbance. Potential differences in ploidy among adventive and wild populations suggest that germplasm collection efforts should seek to distinguish between these two types of populations. However, collection from adventive populations may yield novel variants not seen in sympatric wild taxa, and could be a useful source of germplasm for plant breeding efforts. The two accessions of P. subulata ssp. brittonii analysed were both found to be tetraploid with a mean genome size of 15.98 pg, a value similar to tetraploids of P. subulata ssp. subulata and to tetraploid western taxa such as P. opalensis and P. pungens (Table 3.2).

All accessions of P. bifida ssp. bifida and P. bifida ssp. stellaria were diploid with a mean genome size of 2C=8.59 pg. The genome of the taxa was relatively invariant, but sampling was limited within P. bifida and subspecies.

Both accessions of P. nivalis from wild populations were diploid with a mean genome size of 2C=8.98 pg. However, the P. nivalis from Durham Co. North Carolina was found to contain individuals with two genome sizes (Table 3.2). The larger mean genome size was 13.75 pg, and suggested that these taxa may be aneuploid or triploid, but assessment of chromosome counts is necessary to confirm the ploidy level.

Microsteris gracilis

This taxon was at one time included in the genus Phlox, but molecular analysis has shown that it is a distinct genus, despite morphological similarities to Phlox

209

(Ferguson et al., 1999; Ferguson and Jansen, 2002; Wherry, 1943a, 1955). The mean genome of M. gracilis was 5.86 pg, a value outside of the range of all reported Phlox

taxa. The presented data supports the differences between the two genera and reinforces

the currently recognized taxonomy.

Cultivars and hybrids

All tested cultivars of P. bifida, P. nivalis, P. subulata, P. stolonifera and Phlox

Scotia Alpines Group (syn. Phlox douglasii) had a mean genome size of 2C=8.52 pg

(Table 3.3) and were diploid (n=7; Locklear, 2011a). Among these taxa, genome size was

relatively invariant and ranged from 7.17-8.50 pg. Although cultivars are often attributed

to P. subulata, they contain several putative hybrid combinations involving P. subulata,

other members of subsection Subulatae, and western taxa, but primarily Phlox douglasii

(Bendtsen, 2009; Fuchs, 1994; Locklear, 2011a).

Phlox xprocumbens is reported to be the hybrid P. stolonifera x P. subulata

(Wherry, 1935a). One taxon, propagated as a non-variegated reversion of P.

xprocumbens ‘Variegata’ had a mean genome size of 9.12 pg. This value is intermediate

between the mean values of putative diploid parents P. subulata (7.78 pg) and diploid P.

stolonifera (11.15 pg). Genome size provides information that can be used to confirm successful hybridization because of the divergent sizes between parental genomes; this has been shown with other artificially created hybrids using parents with different genome sizes (Chapters 5, 7). Another cultivar, described as Phlox xoliveri ‘Sunrise’ is

210

reported to be a hybrid of P. nivalis x P. adsurgens; however, this cultivar could not be located for analysis.

One accession, obtained as P. nivalis ‘Camla’ appears to be the hybrid Phlox

xhenryae (P. bifida x P. nivalis). The original P. nivalis ‘Camla’ is described as having

large, rounded, salmon-colored flowers with un-notched, overlapping corolla lobes, but

the plant obtained under this name had lavender flowers with notched corolla lobes in the

fashion of P. bifida, and narrow corolla lobes that did not overlap, indicating a possible

relationship to P. bifida. The mean genome size of this accession was 7.82 pg, but does

not differentiate this taxon from other diploid hybrid accessions.

Discussion

The purpose of this study was to characterize accessions of a developing

germplasm collection of Phlox in order to provide information for breeders and

conservation workers. The goal of this study was to examine the relative genome size

and inferred ploidy levels of a diverse group of mat forming and creeping phlox taxa

collected from natural plant populations and obtained from nursery sources. The

following conclusions may be drawn: 1) Genome sizes proportional to a diploid level

were the more frequent among the taxa tested, but tetraploid species as well as tetraploid

and hexaploid populations of predominantly diploid taxa were identified. 2) The genome

size of P. subulata is more variable than previously reported and the species appears to

form a polyploid complex at the eastern edge of is range consisting of diploid, tetraploid,

and hexaploid cytotypes; this is the first report of a hexaploid cytotype of P. subulata and

211

the first report of a hexaploid eastern Phlox taxon. 3) All cultivated varieties, including

putative interspecific hybrids had genome sizes equivalent to the diploid level, and

indicated that previous breeding efforts have occurred among diploid taxa. The similar

genome sizes and the occurrence of insterspecific hybrids suggests there may be few

barriers to interspecific hybridization in this group.

Three accessions of P. subulata (PZ12-065, PZ12-066, PZ12-067) constitute a polyploid complex (2x-4x-6x) collected from serpentine barren ecosystems on the

Pennsylvania-Maryland (PA-MD) border (Lookingbill et al., 2007). In eastern North

America, this region of serpentine bedrock stretches from Alabama to Quebec, but 90% of these occur along the PA-MD border and have a fragmented distribution within the

eastern deciduous forest region resulting in an island-like distribution pattern (Brooks,

1987). Serpentine barrens have infertile soils with high concentrations of chromium,

cadmium, nickel, and cobalt, low concentrations of necessary plant nutrients such as

nitrogen and potassium, and a low calcium/magnesium ratio (Lookingbill et al., 2007).

In order to maintain the characteristic serpentine barren vegetation consisting of native

grassland with scattered pines (Pinus sp.) and oaks (Quercus sp.), sclerophyllous shrubs,

endemic herbaceous species, and rock outcroppings, frequent fire disturbance is also

necessary (Baskin and Baskin, 1988; Brooks, 1987; Lookingbill et al., 2007). Phlox

subulata is one of the characteristic species of serpentine barrens, but its presence in the

barrens represents a unique ecosystem association, and possibly marginal habitat for this

species at the eastern edge of its geographical distribution (Wherry, 1955). However,

these populations are in close proximity to the center of distribution of P. subulata just

212

north of this region on the Pennsylvania-New Jersey border (Wherry, 1929; Wherry,

1955). While there have been numerous studies of patterns of genetic variation and

modes of adaptation in sympatric serpentine endemics, there has been little attention paid

to the role of ploidy variation among species from serpentine habitats. According to

Raven (1964), a species growing at the geographical margins of its range may be growing

in edaphic conditions that are unusual for the species, and suggested a connection

between edaphic endemism and catastrophic (disruptive) selection. The distribution of P.

subulata in this region illustrates this hypothesis; serpentine barren populations occur just south of the center of distribution for P. subulata, and the relative isolation of these

habitats may have contributed to the extraordinary ploidy variation. The center of

distribution along the Pennsylvania-New Jersey border occurs in a region of

physiographic diversity where the Highlands, Valley and Ridge, Inner Coastal Plain, and

Piedmont provinces occur in close proximity. Within this region, P. subulata is reported

from all physiographic regions, and is found on rock outcroppings and bare soils derived

from sandstone and basalt formations, but are known to colonize rock outcrop habitats

derived from numerous types of bedrock throughout its range (Locklear, 2011a; Wherry,

1929, 1955). Despite this, the range of polyploidy discovered from serpentine barrens

was not seen in other P. subulata accessions from different regions of the species

distribution (Table 3.2). It has recently been shown that autopolyploid taxa can

experience genomic reorganizations that result in novel genetic variation that may permit

colonization of new habitats or ecological niches, and ultimately, divergence and

speciation (Weiss and Maluszynska, 2000; Weiss-Schneeweiss et al., 2007). If a similar

213 situation had occurred in P. subulata, it may account for successful colonization of serpentine barren habitats, and provide an effective means of preventing gene flow between populations with different ploidy levels that are adapted to different soil types.

Reproductive isolation of polyploid populations is supported by crossing data from interploid, interspecific hybridization experiments in which tetraploid x diploid crosses among taxa in subsection Divaricatae resulted in the production of sterile aneuploid F1 hybrids (Chapter 7). However, populations north of the serpentine barrens were not sampled and analysed in this study, and the range of ploidy variation among these populations is unknown, but data from other collections indicate that ploidy variation in the serpentine barrens in unique.

Both allopolyploidy and autopolyploidy have been proposed to explain the occurrence of polyploid Phlox, but neither has been definitively demonstrated (Fehlberg and Ferguson, 2012; Levin, 1966; Levin, 1968; Smith and Levin, 1967; Worcester et al.,

2012). Biosystematic studies, including early molecular phylogenies, concluded that patterns of morphological, genetic, and ploidy variation in Phlox were the result of hybridization and the formation of allpolyploids; however, Levin (1966) refers to tetraploid genotypes of P. pilosa ssp. pilosa as autotetraploids, but there is no further discussion of the origin of these taxa (Ferguson et al., 1999; Levin, 1966, 1968). Two recent studies, one concerning genome size and variation at microsatellite (SSR) loci in the P. amabilis-P. woodhousei complex, and the other, genome size and ploidy variation in P. pilosa ssp. pilosa, both indicate that autopolyploidy may be more widespread in the genus than previously demonstrated, but this hypothesis was not explicitly tested in either

214 study (Fehlberg and Ferguson, 2012; Ferguson and Jansen, 2002; Worcester et al., 2012).

While the genome size data for the polyploid complex of P. subulata from the serpentine barrens do not definitively demonstrate autopolyploidy, there is no evidence of allopolyploidy due to lack of sympatric Phlox species or putative aneuploidy populations or individuals. Furthermore, autopolyploidy would result in maintenance, or creation, of novel genetic diversity that may be of selective advantage on marginal sites. These populations provide an ideal template for assessing the origins of ploidy in Phlox using fluorescent in situ hybridization (FISH).

Several plant species have been shown to consist of mixed cytotype populations and exhibit variation in genome size and ploidy levels as a result of autopolyploidization

(Burton and Husband, 1999; Kolar et al., 2009; Sabara et al., 2013). Two Phlox accessions in this study were found to be mixed ploidy populations consisting of diploids and tetraploids (Table 3.2). One accession, PZ10-049, was collected from a small (>20 individuals) population in a mowed right of way along a state highway; the plants were located near a former homestead. Such populations may be indicative of adventive, or naturalized populations that became established from a former landscape planting. P. subulata has been widely and long-cultivated; in many regions it has escaped to become established in disturbed areas and adventive populations are common over a wide region of the U.S. (Wherry, 1929; 1955). However, these populations have not been studied in any capacity, but may provide interesting insights about the evolution and adaptation of

P. subulata. One accession of P. nivalis originally collected from a natural population in

Durham County, NC and maintained at Mt. Cuba Center in Hokessin, DE was found to

215 consist of diploid and tetraploid individuals, but information pertaining to the size of the progenitor population was not available. The absence of triploids suggests no interbreeding between cytotypes, although a larger sample of individuals from each population is needed to clarify this. However, a recent study of interploid interspecific hybridization among taxa in subsection Divaricatae demonstrated that sterile aneuploids develop from interploid crosses and suggests that ploidy differences are a major barrier to hybridization in Phlox (Chapter 7). Although further testing is required, the initial data further suggest that increased ploidy plays an adaptive role in Phlox, and supports the claim that ploidy differences are an important trait in the maintenance of genetic diversity, diversification, and speciation of caespitose Phlox species (Table 3.3). As the first report of mixed ploidy populations in any Phlox species, my data builds further the case for autopolyploidy in Phlox; allopolyploids would have morphological characteristics indicative of interspecific hybridization, and would likely occur in a contact zone of two species.

Mean genome sizes and distribution of polyploidy in subsection Subulatae are similar to values reported for taxa in section Occidentales analyzed in this and a related study (Fehlberg and Ferguson, 2012). Molecular evidence from ITS and chloroplast restriction sites indicated that subsection Subulatae was a monophyletic group that included P. nivalis, but that this clade is more closely related to morphologically similar pulvinate or caespitose, subulate-leaved, western North American species of section

Occidentales, than to other eastern species. According to the classification of Wherry

(1955), subsection Subulatae is placed in section Phlox because of their long styles and

216 geographic distribution. While some subspecies of P. bifida and all of P. subulata may have long styles in comparison to P. nivalis and subsection Divaricatae taxa, the styles average 7-12 mm in length; these values are more similar to those for western taxa in subsection Occidentales than they are to other long-styled eastern taxa. The style length of taxa in section Phlox is greater than 15 mm in length and provides a fundamental morphological distinction between subsection Subulatae and members of section Phlox.

Wherry opined in his monograph that his classification was “...rather artificial”: a statement that reflects the need for objective markers to accurately describe phylogenetic relationships within the genus Phlox. While previously reported data and data presented here do not provide definitive evidence for the phylogenetic relationships between subsection Subulatae and section Occidentales, it does point to the need to for a genus- wide phylogenetic study that includes all eastern and western species. Data from ITS and chloroplast restriction sites were focused on eastern species, and few western taxa were included, and did not adequately resolve evolutionary relationships of the genus. Further analysis of the ITS region with increased sampling is needed for greater resolution of the genus. In conclusion, data here support the recommendation of previous studies

(Ferguson and Jansen, 2002; Smith and Levin, 1967) that subsection Subulatae be included within section Occidentales.

Although limited sampling of Western mat forming species was possible, the information on genome sizes of this large and thus far little explored group provides a baseline for future studies. Of the four taxa analyzed from subsection Canascentes, two were diploid and two were tetraploid. The diploid taxa P. austromontana and P.

217

muscoides are geographically widespread species that occur in a wide altitudinal range on

a variety of geologic formations and habitat types, but the tetraploid species P. opalensis

and P. pungens occur within limited altitudinal ranges within narrow geographic ranges

(Locklear, 2011a). This fact suggests that, like the polyploidy complex of P. subulata,

higher ploidy may impart an important adaptation advantage over diploid progenitor

species.

This study is based on broad, but relatively limited sampling of natural

populations. Furthermore, two recently described, rare taxa, P. bifida ssp. arkansana and

P. oklahomensis, were not available for analysis. It does not include accessions from a

major part of the comparatively large range of P. bifida ssp. bifida and P. subulata ssp.

subulata; given the range of these species, polyploidy may be more widespread than

previously known (Marsh, 1960; Wherry, 1955; Springer and Tyrl, 1989). Further

collections are necessary to accurately describe relationships of taxa in subsection

Subulatae.

Given the range of genome size and ploidy variation in naturally occurring phlox,

the lack of genome size and ploidy variation in cultivated varieties is surprising and

indicates that the majority of breeding and selection efforts have occurred at the diploid

level, but further testing is warranted (Table 3.3). Although commonly attributed to

being cultivars of P. subulata, this species has a history of breeding and selection, and

many selections may be the result of directed or inadvertent hybridization. Previous

chromosome counts of putative P. nivalis x P. subulata hybrids were reported to be tetraploids, and were proposed to be allpolyploids (Flory, 1933; Flory, 1934). At least

218

one cultivar in this study, Phlox ‘Scarlet Flame’ is often labeled and sold as a cultivar of

P. subulata, but evidence from the literature and the intermediate style length imply that it is a hybrid of P. nivalis x P. subulata, but it is a diploid homoploid hybrid.

Conclusion

This study adds to the now abundant evidence that flow cytometric analysis using the flourochrome propidium iodidide is a rapid and effective method of estimating genome size among eastern and western creeping phlox taxa. Genome size variation in accessions collected from natural plant populations were the most variable of all eastern

Phlox species, and suggests that polyploidy may be more important for adaptation, diversification, and speciation than previously realized. This knowledge is also of use to plant breeders and indicated that all breeding and selection done among cultivars tested in this study have occurred at the diploid level, and that further advances in breeding and selection among creeping phloxes may be viable through ploidy manipulation.

219

PZ12 subulata ssp. count ofmetaphaseof tetraploid 23.72pg. P. subulata chromosome b.Mitotic genome size of 15.65, and; C. genome of 15.65,and; size agenome ofwith mean7.78pg; B. size ploidy in increase each with size genome in increase Figure 3.1:

subulata showing comparison subulata ssp. of P. subulata a.Concurrent the serial

-066 related to peak to -066 “B” related Figure 1 in Number of nuclei

a. a.

b. b. hexaploid (2n=6 hexaploid P. subulata A . RelativefluorescenceDNA tetraploid (2 ssp. ssp. 220 2 n =2 subulata B. x

x =42) PZ12-067 =28

a. n=4 A. PZ12 diploid (2n=2 diploid

x a mean mean a =28) PZ12-066with - C.

066

with a mean genome size size genome mean a with

x =14) PZ12-065=14)

Table 3.1: Summary of means and ranges of 2C, holoploid genome size (pg) of Phlox taxa grouped by section, subsection, and ploidy level.

221

(Continued)

Table 3.2: Collection site, genome size, and ploidy level for a collection of 4 eastern and 7 western Phlox species from subsection Albomarginatae, Canascentes, Longifoliae, Stoloniferae, and Subulatae, and one species from the sister genus Microsteris.

222

Table 3.2: Continued

Phlox subulata L. ssp. subulata Chester Co. PA PZ12-065 15.65±0.22 1.81 4x 28 Phlox subulata L. ssp. subulata Chester Co. PA PZ12-066 7.78±0.05 2.04 2x 14 Phlox subulata L. ssp. subulata Baltimore Co. MD PZ12-067 23.72±0.46 1.18 6x 42 Phlox subulata L. ssp. subulata Delaware Co. OH PZ12-091 7.78±0.05 2.34 2x 14 Phlox subulata L. ssp. subulata Scioto Co. OH PZ10-049 7.86, 16.58 1.70 2x , 4x 28 Phlox subulata L. ssp. subulata Vinton Co. OH PZ13-007 7.74±0.12 1.51 2x 14 Phlox subulata L. ssp. subulata Estill Co. KY; C PZ12-095 7.75±0.07 2.21 2x 14

Outgroup species for Phlox

Microsteris gracilis (Hook.) Greene E W6 27249 5.86±0.67 4.71 2x −−− zNote: Plants not collected by first author obtained from various sources: A, plant obtained from Alplains; B, plant obatined from Primrose Path Nursery; C, plant obtained from J. Campbell, Bluegrass Woodland Restoration Center; D, Plant obtained from Mt. Cuba Center, Hockessin, DE; E, plant obatined from Ornamental Plant Germplasm Center. yMean 2C DNA content (mean ± standard deviation) was determined with Propidium Iodide as the stain. xPloidy level and chromosome counts inferred from Flory (1931), Flory (1934), Meyer (1944), Smith & Levin (1967). wchromosome counts were made for taxa in which no genome size could be inferred.

223

(Continued)

Table 3.3: Genome sizes, ploidy levels, and nursery sources of creeping Phlox cultivars and hybrids from subsections Stoloniferae and Subulatae.

224

Table 3.3: Continued

225

CHAPTER 4

GENOME SIZE, GENETIC DIVERSITY, AND POPULATION STRUCTURE IN PHLOX PILOSA L.AND RELATED TAXA IN SUBSECTION DIVARICATAE

Abstract

Germplasm collection efforts are influenced by genetic factors that contribute to diversity and differentiation of plant populations, however, studies pertaining to these factors are lacking in many U.S. native species. Phlox pilosa is the most geographically widespread Phlox species and is of interest to restoration biologists and horticulturists, yet it remains poorly characterized on many levels. The objectives of this study were to assemble a diverse collection of Phlox germplasm from natural plant populations and representative commercial cultivars, use flow cytometry and chromosome counts to analyze genome size and ploidy, and to use six microsatellite (SSR) markers to estimate levels of genetic diversity and structure in eight populations of taxa from the P. pilosa complex. A total of 89 accessions were analyzed for genome size; among these 75 populations were diploid (n=7) with a range of genome sizes from 9.71 to 14.89 pg. The remaining 14 populations were tetraploid (n=14) and genome size ranged from 21.85 to

26.88 pg. Tetraploids and associated large genomes were only found among members of the P. pilosa complex and closely related taxa, and occurred only in the southwestern, southern, and southeastern range limits of P. pilosa. A subset of eight populations

226

collected from an east to west gradient at the southern and western range limits (five diploid and three tetraploid) from the P. pilosa complex was analyzed with microsatellite markers. A high level of genetic diversity and moderate population differentiation (mean

FST = 0.131) was found among populations. Cluster analysis resulted in four clusters (K

=4), in which all diploid populations formed a single cluster, and each tetraploid population was a distinct cluster. There was admixture between all diploid populations, and between diploid and tetraploid populations. Data suggests that patterns of genetic diversity may be more complicated than taxonomic circumscription implies, and that polyploidy is an important evolutionary mechanism in the diversification and differentiation of taxa in the P. pilosa complex.

Introduction

Germplasm collection efforts influence the production and use of U.S. native species for gardening, habitat restoration, plant breeding, and numerous other applications in a variety of crop, medicinal, ornamental species and their wild relatives

(Widrlechner and KcKeown, 2002). Development of germplasm collections requires knowledge of factors that influence diversification and local adaptation in a wide variety of species with differing life history characteristics to insure proper conservation of genetic resources (Widrlechner and KcKeown, 2002). Despite this, much of this information is only partially characterized, and relatively little is known about patterns of ploidy, genetic diversity, and genetic structure in a wide variety of species, and information is especially lacking in geographically widespread species (Robarts, 2013;

227

Soltis et al., 2006). Recognized patterns of regional diversification in geographically widespread taxa are used to make recommendations for the use of native plants in restoration efforts, and the use of proper regional ecotypes can provide ecological services and have potential for local adaptation that might not be found in closely related, but geographically disparate populations. Phlox pilosa is the most geographically widespread Phlox species, and is a complex of closely related taxa that exhibit extensive morphological variation. As many as 10 subspecies have been described, and the delimitation of taxa in this group has long been debated, and continues to be refined and expanded (Table 4.11) (Ferguson et al., 1999; Locklear, 2011a; Wherry; 1955). Several studies have described phenotypic and ecotypic segregation in Phlox, only recent studies have shown that ploidy is far more variable than previously known, but widespread patterns of genome size, ploidy variation, and genetic structure remain largely unknown

(Levin, 1975; Levin et al., 1979; Levy and Levin, 1974; Levy, 1983; Fehlberg and

Ferguson, 2012). Polyploidy is also a main factor in plant breeding because it can influence fertility, reproductive compatibility, and expression of phenotypic traits, and inform decisions on direction of plant breeding and germplam collection efforts. To begin to assess these questions, a diverse germplasm collection of taxa in the subsection

Divaricatae, with an emphasis on P. pilosa and relatives, was assembled from throughout the range of the species. Genome size and ploidy was analysed in all populations using

Propidium Iodide flow cytometry and chromosome counts, and a subset of 8 populations of P. pilosa were subject to microsatellite analysis to determine preliminary patterns of genetic diversity and structure.

228

Phlox (Polemoniaceae) is a genus of ca. 65 species with a center of origin and distribution in North America; one species is widespread in Siberia (Ferguson et al.,

1999; Locklear, 2011a; Wherry, 1955). The current taxonomic circumscription is based on morphological and geographic differences; and two major groups, eastern and western, are recognized. Wherry (1955) initially described two sections; Annuae and

Occidentale, and 19 subsections. Among these, the Eastern species are divided among 5 subsections: Cluteanae, Divaricatae, Paniculatae, Phlox (syn. Ovatae), Subulatae (Gray,

1870; Locklear, 2012; Prather, 1994; Wherry, 1955). These are distinguished primarily by differences in the shape and vesture of the calyx and style length (Ferguson et al.,

1999; Wherry, 1955). Subsection Divaricatae is differentiated from other eastern subsections by having a short style that is 1-3 mm in length with a tripartite stigma with lobes equal to or longer than the style (Wherry, 1955). Subsection Divaricatae sensu

Wherry contained 4 species and 13 subspecies (Table 4.1) (Wherry, 1955). However, taxonomic modifications have been made by a number recent authors, and as currently circumscribed subsection Divaricatae contains 10 species and as many as 15 subspecies, making it the most speciose subsection among eastern Phlox taxa (Table 4.1) (Erbe and

Turner, 1962; Ferguson et al., 1999; Levin, 1963; Levin, 1966; Locklear, 2011a;

Locklear, 2011b). The primary reason for the increase in number of species is due to the inclusion within subsection Divaricatae of the species formerly in subsection

Drummondianae. Wherry (1955) placed the morphologically similar short-styled, annual species into subsection Drummondianae, but this subsection was moved into subsection

Divaricatae upon the discovery of Phlox pattersonii, a species with morphology

229

intermediate to subsections Divaricatae and Drummondianae, which resulted in an

expanded view of subsection Divaricatae (Prather, 1994). This modern classification will

be referred to as sensu Zale. Further expansion of subsection Divaricatae has come

through the description of 3 new subspecies in the P. pilosa complex (P. pilosa ssp.

deamii, P. pilosa ssp. longipilosa, and P. pilosa ssp. sangamonensis) and the elevation of

P. pilosa ssp. pulcherrima to P. pulcherrima and the merging and elevation of P. pilosa ssp. latisepala and P. pilosa ssp. riparia into P. villosissima. Despite a large volume of work pertaining to numerous ecological, evolutionary, and horticultural aspects of P. drummondii, comparatively little similar information exists for the P. pilosa complex.

Phlox pilosa is the most geographically widespread Phlox species and occurs from the gulf coast of Florida from the Appalachicola River westward to the Edwards

Plateau in Texas, and northward along the western edge of the tall grass prairie regions eastward to the southern Great Lakes region. This species also occurs in the Coastal

Plain and northern Piedmont, but is curiously absent from the Appalachian Plateaus and

Highlands (Figure 4.1) (Ferguson et al., 1999; Wherry, 1955). As described by Wherry

(1955), Phlox pilosa consisted of 7 subspecies that were distinguished based on morphological features of the calyx and plant vesture, and were also roughly correlated to geographic distribution (Table 4.1) (Wherry, 1955). Despite this several new taxa have been described based on morphological and geographical differences, and some taxa that were previously subspecies are now considered species; even with these changes, the taxonomy of this group still remains unresolved, and as currently circumscribed, the P. pilosa complex is now comprised of 4 species and 5 subspecies (Table 4.1) (Ferguson et

230

al., 1999; Locklear, 2011a; Wherry, 1955). Phlox pilosa flowers from early April until

July throughout its range, and there is indication that it may be among the most important

nectar plants for several Lepidoptera. In Ohio, P. pilosa ssp. pilosa may serve as a nectar

plant for the state-endagered Karner blue butterfly (Melissa lycaeoides samuelis), and they are important in maintaining populations of the endangered Phlox moth ( indiana), which co-occur with Karner blue butterflies in parts of their range. (Grant and

Grant, 1965; Wisconsin Deprtment of Natural Rersources, 2014; Wiggam and Ferguson,

2005). Such ecological services only serve to bolster the importance of these plants in restoration efforts.

Two molecular phylogenetic studies have attempted to resolve the taxonomy and evolutionary relationships of the P. pilosa complex using variation in the internal transcribed spacer (ITS) and 18 chloroplast restriction sites (Ferguson et al., 1999;

Ferguson and Jansen, 2002). The initial ITS study concluded that all annual species endemic to Texas were a monophyletic group and included the previously separated P. roemeriana. Discovery of the morphologically intermediate P. pattersonii resulted in the merging of subsections Divaricatae and Drummondianae (Prather, 1994; Wherry, 1955).

However, both studies failed to resolve relationships among the P. pilosa complex, which formed a polyphyletic group with members of subsection Phlox, and primarily taxa in the

P. carolina-P. glaberrima complex. In an ITS and cpDNA congruency study, the most parsimonious trees exhibited similar branching patterns and resulted in the same conclusions presented in the ITS study, and failed to further resolve the phylogenetic realtionships of the P. pilosa complex. Characteristics of the phylogenetic trees, such as

231 short branch lengths, suggest that the markers are not variable enough to be used for taxon delimitation and that other molecular markers need to be evaluated. Short branch lengths and paraphyletic groupings could also be explained by a history of hybridization

(Ferguson et al., 1999; Ferguson and Jansen, 2002). While hybridization has been implicated as a major driver of evolution in several biosystematic studies of Phlox, the role of hybridization and evolution has yet to be fully determined in the diversification within the P. pilosa complex.

Patterns of ploidy variation among subsection Divaricatae have only been partially characterized, but it has already been shown that there is dramatic ploidy variation within a small portion of the range of P. pilosa, indicating that genome size and ploidy were associated with divergence at the population level (Worcester et al., 2012).

In an initial study, extensive genome size and ploidy variation was found among populations of P. pilosa ssp. pilosa at the southwestern edge of the species range in

Arkansas, Kansas, Oklahoma, and Texas (Worcester et al., 2012). However, further study of populations from the P. pilosa complex has not ensued; available data highlights the need for extensive geome size and ploidy analysis of this widespread species complex. In a study of the P. amabilis-P. woodhousei complex, genome analysis and ploidy resulted in the discovery of diploid, tetraploid, and hexaploid cytotypes. In this study, genome size alone could not be used to distinguish between the species, but microsatellite analysis of 18 populations supported the taxonomic split between P. amabilis and P. woodhousei, and suggests that the combination of genome size, ploidy,

232

and microsatellite variation can be used to detect patterns of fine-scale variation in Phlox species.

The objectives of this study were to; 1) Develop a diverse collection of Phlox taxa from subsection Divaricatae from natural plant populations that represented all members of the P. pilosa complex and related taxa. 2) To estimate the nuclear genome size and ploidy for these collections. 3) To use microsatellite markers to estimate genetic diversity and structure in a subset of phenotypically diverse populations of P. pilosa from the

United States Gulf coast, Interior Highlands of Arkansas, and Edward’s Plateau of Texas.

Materials and Methods

Phlox accessions

Germplasm accessions used for this study were collected from 76 natural Phlox populations between 2010-2013 (Figure 4.1; Table 4.2). Some accessions were obtained from nursery sources when the wild origin of the taxon was verified (Table 4.2).

Identifications of all taxa were determined using the keys of Locklear (2011a) and

Wherry (1955). Herbarium vouchers have been deposited at the Ornamental Plant

Germplasm Center (OPGC), in Columbus, OH. In addition to the accessions of wild origin, thirteen cultivars were obtained from a variety of commercial sources (Table 4.3).

All 89 accessions were examined for genome size by flow cytometry (Tables 4.2, 4.3), and a subset of 8 populations of P. pilosa from the southern U.S. was used in a genetic diversity analysis using microsatellites (Table 4.4).

233

Tissue sampling, flow cytometry analysis, and chromosome counts

Sample preparation for flow cytometric analysis was described in Chapter 2.

Selected samples with genome sizes corresponding to diploid and tetraploid ploidy levels were examined for chromosome counts following procedures also detatiled in Chapter 2.

DNA Extraction

Genomic DNA was isolated from 30-50 mg of silica-dried leaf tissue using a

modified CTAB mini-prep protocol (Doyle and Doyle, 1987; Robarts, 2013). DNA

products were visualized in a 1% agarose gel, 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).

Amplification and genotyping

Eleven simple sequence repeat primer pairs developed for Phlox pilosa ssp. pilosa

(Fehlberg et al., 2008) were used for amplification. Two additional primers were

obtained directly from Fehlberg (Shannon Fehlberg, Pers. Comm. 20 January 2011).

These were screened for variability in a subset of Phlox subsection Divaricatae samples,

and six were selected due to the diversity and repeatability of alleles produced. The

primers used were Phl33, Phl68, Phl84, Phl13, Phl137, and M4M5 (Table 4.5). The

234 template for PCR amplification consisted of raw DNA extract diluted 1:20 in TAE buffer, yielding ~20-50 ng genomic DNA. PCR amplifications were performed in a reaction volume of 10µL, comprised of 0.05µl (0.25U) of ExTaq (Clontech Laboratories, Inc.,

Mountain View, CA), 1µl of 10X ExTaq PCR Buffer, 0.8µl of ExTaq dNTP Mix (2µM),

0.5 µl of fluorescence-labeled forward strand primer (0.5µM), and 0.5L of reverse strand primer (0.5µM) in 5.1ul UV-irradiated HPLC pure water (Fisher Scientific, Pittsburgh,

PA). The amplification was carried out in an Eppendorf Mastercycler® ep gradient S thermal cycler (Eppendorf North America, New York, NY) following optimized PCR protocols previously described (Fehlberg et al., 2008). The presence of PCR product was confirmed by running 2µL 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 prepared by mixing 1 µL of PCR product with

9.5 µL of HiDi Formamide (Life Technologies, Grand Island, New York) and 0.5 µL

GeneScan ROX500 size standard (Life Technologies, Grand Island, New York).

Resulting electropherograms were scored with GeneMapper analysis software (version

3.7, Applied Biosystems), and manually reviewed and adjusted.

Genetic analysis

Flow cytometry and microsatellite data confirmed that tested populations were both diploid and tetraploid (Figures 4.1, 4.2; Tables 4.2, 4.3). In polyploids, it is not possible to determine the number of copies of each allele because of the unknown origin

235

and dosage levels of alleles in heterozygotes. Therefore, it is not always possible to

determine the genotype of polyploid taxa. The methods of Fehlberg and Ferguson

(2012), Robarts (2013), and Sampson and Byrne (2012), which provide strategies for

dealing with mixed ploidy and polyploid data, were followed. All alleles at a locus were

combined to form multilocus allele phenotypes for each individual. Allele copy numbers

were converted to presence-absence data and were treated as dominant markers in which

the presence-absence of bands was used to calculate statistics to assess intraspecific

differentiation and diversity. Therefore, data cannot be treated as genotypes, and are

referred to as ‘allelic phenotypes’ (Robarts, 2013). To describe diversity, the allelic

phenotype scoring methods of Sampson and Byrne (2012) were followed: allelic

diversity was measured as the total number of alleles seen over all loci (A), the number of

different alleles seen in a population, averaged over loci (A’), as the proportion of

heterozygous individuals averaged over loci (Ho), and as the number of alleles seen in an

individual per locus, averaged over loci (H’). Additional measures of diversity included the number of alleles per individual, averaged over loci (X), and the number of single- locus allele phenotypes observed, averaged over loci (N’p).

Genetic structure was analyzed in three ways. Analyses of molecular variance

(AMOVA) was performed in GenAlEx, yielding estimates of PhiST: a statistic analogous

to FST (Peakall and Smouse, 2012; Wright 1969, 1978), dividing the variance into the among-population and within-population components. Additionally, the software

SPAGeDi 1.3 (Hardy, O. J. and X. Vekemans, 2002) was used to generate population FIS,

236

FIT, and FST values to further determine the degree of population structure. SPAGeDi was

also used to generate Nei’s diversity (h) measures.

STRUCTURE 2.2.3 software was used to estimate the assignment of individuals

to a specific genetic cluster (K). Individuals from related origins can be detected, as the

probability of ancestral identity will be divided between specified (K) genetic groups.

Individuals are assigned to clusters and the proportion of an individual’s genome (Q) that

originated from each cluster is determined. The STRUCTURE program was run with no

prior knowledge, under the admixture ancestry model, with ploidy set to 4 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 8 with three replicate runs for each value.

Probability of K values was calculated using the method of Evanno et al. (2005) in the program STRUCTURE Harvester (Earl and von Holdt, 2011). Review of analyzed K values indicated the highest likelihood for values under 4, at which time STRUCTURE analysis was rerun with a burn-in period of 105 with 5x105 iterations, testing K values 1

to 4, with 10 replicate runs for each value. When the optimal cluster value was

determined, STRUCTURE was run a final time with 105 and 5x105 iterations, 10

repetitions with 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. This program calculates a similarity coefficient, H’, a statistic assessing

the congruence of replicate runs, and also calculates a mean cluster assignment

237

probability per individual. The average clustering probabilities were then visualized

using the program DISTRUCT 1.1 (Rosenberg, 2004).

Results

Cytotypic variation and geographic distribution in subsection Divaricatae

Measurements of nuclear genome size by flow cytometry were made for all

currently recognized species in subsection Divaricatae, except P. cuspidata (Ferguson et al., 1999; Locklear, 2011a; Wherry, 1955). The mean genome size was calculated from two to 24 samples per population. The most extensive sampling was made in the P. pilosa complex (34 accessions) while P. divaricata taxa consisted of 19 accessions and P. amoena of 9 accessions; the remaining species all included four or fewer accessions

(Table 4.2). Measurements confirmed that all accessions were either diploid (n=7) or tetraploid (n=14). Typical flow cytometry histograms and chromosome preparations obtained in this survey are shown on Figure 4.2. Within population variation in genome size and ploidy was not detected for any populations (Table 4.2). Genome size of diploid taxa ranged from 9.23 pg to 14.87 pg. The genome size of tetraploid taxa was approximately twice that of diploid taxa and ranged from 21.77 pg to 26.88 pg.

Tetraploids were only identified in P. pilosa ssp. pilosa, P. floridana, P. pulcherrima, and

P. villosissima. The coefficient of variation for samples ranged from 0.95% to 4.66 %.

Chromosome counts were used to confirm ploidy levels of several samples (Table 4.2).

The geographic distribution of cytotypes for the different taxa is shown in Figure 4.1; the

238

four tetraploid taxa were distributed along the southern, southwestern, and southeastern

edges of the distribution of the P. pilosa complex.

Cytotypic variation among taxa in subsection Divaricatae

Nine populations of P. amoena were collected from throughout the northern and

central range of the species (Wherry, 1955). All accessions were diploid (n=7) with a

mean genome size of 12.01 pg. Most populations showed little variation and possessed

the heavily pubescent leaves, stems, and pilose calyces typical of P. amoena (Wherry,

1955). Populations with glabrous herbage from the southeastern portion of the species distribution, named P. amoena ssp. lighthipei by Wherry, were not sampled, and the distinctiveness of this taxon is not currently recognized (Ferguson et al., 1999; Locklear,

2011a). However, accession PZ11-032 displayed the glabrous leaves, stems, and calyces of Wherry’s P. amoena ssp. lighthipei, but was collected from a unique river scour habitat in Kentucky that differs from the upland, xeric habitats this species normally inhabits (Wherry, 1995). Individuals from this collection displayed phenotypic characteristics often associated with polyploidy, such as thickened leaves and flower petals, but were determined to be diploid with a mean genome size of 11.58 pg (Table

4.2). These populations may represent a new taxon but the flow cytometry data does not

provide indication of distinctiveness; further analysis of these taxa using appropriate

molecular markers is warranted.

Phlox divaricata is the second most widely distributed species in subsection

Divaricatae; two subspecies have been described, P. divaricata ssp. divaricata and P.

239

divaricata ssp. laphamii. They are distinguished by the presence (ssp. divaricata) or absence (ssp. laphamii) of a cleft in the apex of the corolla lobes (Wherry, 1955).

Populations can be found where both phenotypes occur and intergrade (Table 4.2). The germplasm for this species consisted of 7 accessions of ssp. divaricata, 11 accessions of ssp. laphamii and one accession that appeared to be a mixture of the two. All accessions of both subspecies were diploid (n=7) with a mean genome size of 10.22 pg (range: 9.18-

11.61 pg). Three cultivars of P. divaricata were also diploid (n=7) and the genome sizes

were within the range of natural populations, and varied from 9.66 pg to 9.89 pg (Table

4.3).

The monophyletic annual Phlox species from Texas (P. drummondii,

P.roemeriana), formerly of subsection Drummondianae (Ferguson et al., 1999), had genome sizes ranging from 12.21 pg to 13.04 pg, and were diploid (n=7). Three

accessions representing two subspecies of P. drummondii had a mean genome size of

12.12 pg (Table 4.2). At least four more subspecies of P. drummondii are recognized

(Locklear, 2011a; Wherry, 1955) but were not available for testing. Phlox roemeriana is

endemic to Texas and easily separated from P. drummondii and P. cuspidata on the basis

of morphology and geographic distribution; it is the only known phlox with tri-colored

flowers. The mean genome size of 13.04 pg was slighty larger than, but similar to P.

drummondii, supporting the relationship of these taxa as indicated by molecular data

(Ferguson et al., 1999). The closely related, self-pollinating P. cuspidata also was not

included in the flow cytometry analysis. Five cultivars described as P. drummondii

(although some may be interspecific hybrids involving P. drummondii and P. cuspidata)

240 were analyzed. All accessions were diploid (n=7) with genome size that ranged from

12.24 to 12.91 pg (Table 4.3); these sizes are within the range of values reported for other annual taxa of wild origin.

Phlox pattersonii is the most recently described member of subsection

Divaricatae, and may be one with the most southern distribution of all Phlox species

(Prather, 1994). This accession was obtained from the original introduction of this taxon to cultivation, made by the author of the species (Prather, 1994). The mean genome size is 14.87 pg and represents the largest of all diploid (n=7) taxa in subsection Divaricatae measured in our study (Table 4.2).

P. pilosa ssp. pilosa was the only taxon in this survey that consisted of both diploid and tetraploid populations; these results corroborate previous reports of such polyploid populations in this species (Smith and Levin, 1967; Worcester et al., 2012).

Most populations of P. pilosa ssp. pilosa were diploid (n=7) with a mean genome size of

11.98 pg. However, genome sizes ranged from 11.05 pg to 13.02 pg, the widest range among all eastern Divaricatae species at the diploid level (Table 4.2). Three cultivars of

P. pilosa were also diploid and genome size ranged from 11.16 pg to 12.55 pg (Table

4.3). Of the four tetraploid (n=14) accessions of P. pilosa ssp. pilosa identified, two were discovered in Bossier Parish, Louisiana, and represent an extension eastward of tetraploids reported at the nearby western edge of the species distribution (Worcester et al., 2012). A collection of P. pilosa ssp. pilosa from Polk County, North Carolina, occurring at the eastern edge of the species range was also determined to be tetraploid and suggests that similar patterns of ploidy as seen in populations at the western edge of

241

the species range may also exist at the eastern extreme of the range (Figure 4.1). Among

the 5 cultivars of P. pilosa examined, two were tetraploid and had genome sizes 22.14-

24.47 pg (Table 4.3).

All currently recognized subspecies of P. pilosa had genome sizes ranging from

10.52 to 13.12 pg and were diploid (Table 4.2). Three populations of the narrowly distributed endemic P. pilosa ssp. deamii were analyzed (Table 4.2). Two accessions,

PZ12-045 and PZ12-090, were morphologically uniform and closely matched the description of this taxon and occurred in the absence of the putative parental taxa P. amoena and P. pilosa ssp. pilosa (Levin, 1966). The populations had mean genomes sizes of 11.26 pg and 11.96 pg. However, the collection PZ12-024, referred to as P. pilosa ssp. deamii by Levin (1966), was discovered to be a hybrid swarm that occurred in the presence of P.amoena and P. pilosa ssp. pilosa. This population was clearly of hybrid origin, and distinctly different from populations to the north. Twenty-four individuals, representing various degrees of hybrids and intergrades were diploid with a mean genome size of 12.04 pg.

Two populations of P. pilosa ssp. fulgida were analyzed and both accessions were diploid with a mean genome size of 12.50 pg (Table 4.2). One accession of P. pilosa ssp. longipilosa had a mean genome size of 12.76 pg, equivalent to the diploid level.

Phlox pilosa ssp. ozarkana had the smallest genome size of all members of the P. pilosa complex with a mean genome size of 10.73 pg. One accession of P. pilosa ssp. sangamonensis was found to have a mean genome size of 11.03 pg. Genome size could not be used to distinguish any of the diploid P. pilosa subspecies from each other. The

242

range of genome sizes suggested that genome size variation may be an important factor in

the ecotypic diversification of the P. pilosa complex.

Two former subspecies of the P. pilosa complex, P. pulcherrima and P.

villosissima had genome sizes ranging from 21.77 pg to 26.88 pg and were tetraploid

(n=14). The mean genome sizes of three accessions of P. pulcherrima were in the

smaller range of the tetraploids (21.77-23.66 pg). These values are within the range of

the genome size for P. floridana and for tetraploid accessions of P. pilosa ssp. pilosa, and

suggest that these polyploids may share a common origin (Table 4.2). Phlox villosissima,

known formerly as P. pilosa ssp. latisepala and P. pilosa ssp. riparia, had the largest

genome size of all taxa in subsection Divaricatae at 26.88 pg. Genome size was uniform

for the 4 accessions examined (26.58-26.88 pg; Table 4.2). This larger genome size is

distinctive and is not shared with other members of the P. pilosa complex, and is among

the largest of all genome sizes found in a large survey of eastern Phlox taxa; similar

genome sizes were only found among tetraploids taxa in the Phlox carolina/glaberrima

complex and P. pulchra (Chapters 2, 3).

One accession of P. floridana, collected in Jackson County, Florida had a mean

genome size of 21.85 pg. This genome size supports previous reports that P. floridana is

tetraploid (n=14); this was the smallest genome size discovered for all tetraploids in

subsection Divaricatae (Table 4.2).

A sister subsection to Divaricatae in the section Annuae is subsection Nanae

which contains mainly western U.S. species (Chapter 1). Only one of these western

members was tested; Phlox nana had a mean genome size of 10.21 pg, consistent with

243 the diploid level (Table 4.2). This plant has a superficial morphological resemblance to taxa in the P. pilosa complex, and has a genome size that falls within the range of diploid taxa in subsection Divaricatae.

Allelic variation at microsatellite loci of P. pilosa

The availability of a variety of taxa within the P. pilosa complex made it possible to initiate a preliminary evaluation of genetic diversity in the group; a subset of 8 populations of 3 taxa within the complex was chosen for analysis (Table 4.4). This subset included the subspecies pilosa and ozarkana as well as the now species P. villosissima, formerly subspecies latisepala and riparia. A total of 154 alleles were scored across all 6 loci, and size of the alleles ranged from 121 bp in Phl33 to 554 bp in

Phl113. There were 1-4 bands per locus (Table 4.5). There was variation in the number of alleles per locus, which ranged from a low of 3 for M4M5 to a high of 39 in Phl33 and

Phl84.

Genetic Diversity

The analyzed populations displayed patterns and levels of genetic diversity comparable to that found in other Phlox taxa using allozyme and microsatellite markers

(Fehlberg and Ferguson, 2012; Levin, 1977). The results are summarized in Table 4.6. A total of 297 alleles were found among all populations. Within populations, the number of alleles (A) varied from 26 in MS to 48 in AL, with an average of 37.13 per population.

The number of alleles in a population, averaged over loci (A’) ranged from 3.25 in

244

population PZSH2011-020 to a high of 5.75 in PZSH2011-017, and the mean was 4.46.

The number of alleles per individual, averaged over loci (X) ranged from 7.56 in LA1 to

11.86 in TX, and the mean was 9.16. The number of alleles per locus, averaged over

individuals (H’) ranged from 1.26 LA1 to a high of 1.98 in TX, with a mean of 1.53. The

proportion of heterozygotes, averaged over loci (Ho) ranged from 0.43 in P. pilosa ssp.

ozarkana AR2 to 0.68 in the tetraploid P. pilosa ssp. pilosa LA3, with a mean of 0.54

(Table 4.6). Lastly, the number of different single-locus allelic phenotypes observed, averaged over loci ranged from 3.67 in diploid P. pilosa ssp. pilosa MS to 5.83 in P. pilosa ssp. ozarkana AR1, with a mean of 4.61 (Table 4.6).

Genetic Structure

Results from AMOVA indicated that most of the observed variation (79%) was due to differences among individuals; 21% of the variation was due to differences among

populations, indicating moderate population structure (Table 4.7). These values are

indicative of outcrossing species exhibiting a high degree of within population variation,

and are comparable to other Phlox species (Fehlberg and Ferguson, 2012; Levin, 1977).

Among the 8 populations, the mean ΦPT, a statistic analogous to FST, was 0.207 and was

significant (P<0.05). These values indicate moderate, but significant population differentiation. This conclusion is supported by Population FST, which had similar values and ranged from 0.062 to 0.201, with a mean FST of 0.131 (Table 4.8). The mean

inbreeding coefficient of the individual to the total population, FIT, was 0.212 and

inbreeding coefficient of the individual to the population, FIS was 0.243, suggesting some

245

inbreeding within populations, and supporting the differentiation of populations (Table

4.8). Levels of population differentiation were similar to those reported for P.

drummondii, but lower than those for the P. amabilis-P. woodhousei complex (Fehlberg

and Ferguson, 2012; Levin, 1977).

Bayesian clustering analysis by STRUCTURE indicated the most probable

clustering arrangement was K=4 (Figure 4.4), with significant admixture between populations within the same cluster, and between clusters 1 and 4. Support for K=4 was

evaluated using the methods of Evanno et al. (2005) with the software STRUCTURE

Harvester (Figure 4.3). Further analysis of 10 repeated runs of the K=4 assignments in

CLUMPP also supported the K=4 pattern (Figure 4.4). The H’ value of 0.957 indicated strong congruency between the 10 independent STRUCTURE runs. The four clusters were roughly associated with geographic distribution and ploidy level (Figure 4.4). The mean population proportion assignment over all individuals was Q=0.887, and between sample locations ranged from Q=0.785 for group one to Q=0.975 for the cluster containing tetraploid individuals from the accession PZ11-069. All diploid populations were assigned to the same cluster, with admixture from the tetraploid TX population.

The two tetraploid populations LA2 and LA3 were collected from close proximity in

Bossier Parish, yet each formed a separate cluster.

Discussion

The primary goal of this study was to assess patterns of genome size and ploidy variation among 76 naturally occurring Phlox populations, and 13 Phlox cultivars, within

246

taxa in subsection Divaricatae. A secondary goal was to define preliminary patterns of

genetic diversity and genetic structure in the large and diverse P. pilosa complex by

examining a subset of eight populations using 6 microsatellite (SSR) markers. The following conclusions can be drawn from this study: 1) Genome size in subsection

Divaricatae was variable for all tested taxa, but the highest degree of variation, and only occurrence of tetraploidy, was within the P. pilosa complex. 2) Tetraploid populations

tended to occur at the edges of the distribution range of the widely distributed species

supporting the conclusion that ploidy may be associated with dispersal and persistence in

marginal habitats, as has been shown for other plant species. 3) Cultivar development in

P. drummondii and P. divaricata appears to involve only diploid forms whereas ploidy

manipulation (or at least selection for higher ploidy) has occurred in some cultivars of P.

pilosa. 4) Microsatellite analysis revealed that each population within the P. pilosa

complex exhibit high genetic diversity, with moderate differentiation between

populations and additional population structure due to ploidy differences. 5) Structure

analysis indicated that each tested polyploid populations was assigned to a distinct

cluster, suggesting that these populations may harbor unique genetic variation that can

influence germplasm collection ans conservation.

Interpopulational genome size and ploidy variation exhibited by taxa in

subsection Divaricatae may be associated with an adaptive environmental response, and

such variation is a widespread phenomenon in other angiosperms, yet its significance

remains unknown (Greilhuber, 1998; Greilhuber, 2005; Greilhuber and Leitch, 2013).

The variation in genome size among diploid members of the P. pilosa complex (24.7%)

247

was similar to that for tetraploid taxa (23.4 %), but tetraploids appeared to be distributed

geographically in a non-random manner (Figure 4.1; Table 4.2). They were found

primarily along the southern, southeastern and southwestern margins of the distribution of the P. pilosa complex and exhibit some of the largest reported genome sizes in the genus Phlox; diploid populations tended to be centrally distributed within the range of the

P. pilosa complex (Figure 4.1). This likely non-random distribution suggests that genome

size and ploidy may be associated with adaptation to marginal habitats (Parisod et al.,

2010); however, it is not likely that ploidy changes are directly involved in the adaptive

process (Greilhuber and Leitch, 2013). For example, secondary metabolites such as

flavonoids showed both additive profiles as well as novel patterns as a consequence of

polyploidization in Phlox (Levin, 1968). Bayesian cluster analysis of SSR data indicated

that tetraploid populations of the P. pilosa complex exhibited unique genetic structuring

within and between ploidy levels, indicating that polyploidy plays a role in the

diversification of Phlox, regardless of the mode of polyploidization (Figure 4.4). In other

angiosperm species, induction of allopolyploids and autopolyploids was associated with

genomic reorganizations, and generation of novel genetic variation that altered genetic

and epigenetic effects resulting in physiological or phenotypic changes that enhance the

adaptive process and allow colonization of marginal habitats or exploitation of new

ecological niches (Levin, 2002; Weiss-Scheeweiss et al., 2007). These results suggest

that polyploidy is associated with the production of novel genetic variation, and leads to

differentiation and diversification (Soltis et al., 2009).

248

However, the origin of polyploidy in the P. pilosa complex is still unknown, and

SSR data cannot be used to determine whether or not polyploids have formed via allopolyploidy as a result of hybridization or are autopolyploids that arose by chromosome doubling (Levin, 1966; Levin, 1968; Worcester et al., 2012). Previous studies suggested interspecific hybridization was the major driver of polyploidization in the P. pilosa complex, but recent evidence suggests that autopolyploidy, or a combination of allo and autopolyploidy may be found in the P. pilosa complex (Ferguson et al., 1999;

Worcester et al., 2012). Triploid populations of P. pilosa have not been discovered, even when tetraploid and diploid populations have been found in close proximity, and the results of artificial interploid hybridization efforts indicate that triploid progeny result from interploid crosses and all crosses between diploid taxa resulted in homoploid hybrids (Worcester et al., 2012). Together, this information provides strong indication for autopolyploidy, but further confirmation is necessary. Despite the preliminary nature of this data, the evidence suggests that polyploid populations are genetically unique, and require further analysis to determine their role in the evolution of the P. pilosa complex.

Even with limited sampling, a high degree of genetic diversity and moderate genetic structure was discovered among populations of the P. pilosa complex from the southern United States (Table 4.7, Table 4.8). Members of the P. pilosa complex and subsection Divaricatae are outcrossing species and have a gametophytic, self- incompatibility system, and entomophilous pollination via lepidopterans (Grant and

Grant, 1965; Levin 1973; Levin, 1993). Species exhibiting similar outcrossing mating systems typically show high levels of genetic diversity, with moderate genetic structuring

249 among populations and regions as most of the diversity is held within populations

(Hamrick and Godt, 1996; Hereford, 2010). Previous studies of allozyme and microsatellite diversity in Phlox species support the findings here. Allozyme variation in

P. drummondii revealed a high degree of genetic diversity with little population structure: heterozygosity ranged from 0.338 to 0.625 and FST ranged from 0.087 to 0.228 indicating that genetic diversity was held within populations and population structuring was weak

(Levin, 1975). FST values generated from my study are comparable to those from the P. drummondii study, and structure analysis indicated that similar geographic partitioning of populations might be at play (Tables 4.7, Table 4.8). Furthermore, patterns of genetic diversity were not correlated with morphological subspecies delimitation in P. drummondii, and were more closely associated with particular geographic formations and edaphic differences (Levin, 1975). In a related study, microsatellite analysis of the narrowly geographically distributed Phlox amabilis-woodhousei complex had a mean He of 0.744, and Fst values that ranged from 0.086-0.247, indicating moderate, but significant population differentiation among diploids (Fehlberg and Ferguson, 2012).

There was also a significant Fst value (0.305) supporting the separation of P. amabilis and

P. woodhousei as independent species (Fehlberg and Ferguson, 2012). Previous authors had considered all populations to represent a widespread, geographically variable P. amabilis, but microsatellite analysis was robust enough to discern between the two taxa.

Patterns of genetic diversity and population structure in the P. pilosa complex exhibit similar patterns and values to previous studies and support the data presented here. The data generated in this study suggests that microsatellite markers can be used to discern

250

patterns of genetic variation within narrowly distributed endemics in the P. pilosa

complex and can contribute to an increased understanding of the evolutionary history of

this species.

Members of the P. pilosa complex are relatives of the important floriculture crop

P. drummondii (Schoellhorn and Richardson, 2004). Phlox pilosa and P. drummondii

hybridize under artificial conditions, and hybrids may have divergent phenotypes that are

superior to either P. drummondii or P. pilosa (Levin, 1966; Zale, unpublished data). The high degree of genetic variation and genetic structure within the P. pilosa complex has important implications for Phlox breeders wishing to improve P. drummondii via interspecific hybridization. Although there is extraordinary phenotypic variation within

P. drummondii that has given rise to more than 125 phenotypically distinct seeds strains, all known subspecies and related taxa have an annual life cycle (Levin, 1966; Levin,

1976a, 1976b). From a production standpoint, most P. drummondii are grown from seed that requires laborious hand pollination and seed collection, and resultant plants are not well adapted to modern production systems (Schoellhorn and Richardson, 2004). . The development of interspecific hybrids between P. drummondii and members of the P. pilosa complex could result in plants with the floral characteristics of P. drummondii with the perennial life cycle of P. pilosa that would allow for the development of individuals that could be maintained in a vegetative state and asexually propagated. The moderate population differentiation between populations suggest that using individuals from different populations could yield varied results, but results of the AMOVA indicate that

251

selection of individuals with characteristics of interest may be more important than using

plants from a particular population.

The results of this study may have several implications for further germplasm

collection and germplasm regeneration and conservation (Czarnecki II et al., 2008). The

availability of high levels of genetic diversity in natural populations of P. pilosa provides

a rich source of genetic variation and indicates that seed collection should be the primary

method of germplasm acquisition to establish genetically diverse collections for regeneration. However, the moderate population differentiation and clustering results indicate that there are important differences between populations, and that collection for restoration efforts should comprise material collected from the most geographically proximate population. Further delineation of patterns of genetic diversity and structure are needed to form guidelines for conservation and habitat restoration efforts involving P. pilosa.

Finally, the results and conclusions presented here should be considered preliminary because we were able to sample only a limited number of natural populations, especially in the microsatellite analysis. Confirmation of these findings will require sampling of a larger number of populations including all subspecies and species within and related to the P. pilosa complex, and a greater number of individuals (n≈25)

per population (Fehlberg and Ferguson, 2012). Because of the large natural range of the

P.pilosa complex, identification and testing of low copy number, fast evolving

chloroplast and nuclear genes may provide species level and subspecific delineation of

taxa. In particular, the chloroplast gene ycf1, has been shown to resolve phylogenetic

252 relationships in the closely related genus Polemonium. Additionally, the use of tools such as fluorescent in situ hybridizaion (FISH) could be employed to discern the origin of polyploidy in Phlox.

253

Figure 4.1: Location and county level geographic distribution of accessions and cytotypes from the Phlox pilosa complex. The outline corresponds to the reported geographic distribution of the P. pilosa complex and all related taxa. All recognized subspec ies and related species are included in the flow cytometry analysis (Table 4. 1).

254

b. a.

c. d.

Figure 4.2: Flow cytometry histograms and corresponding chromosome counts of diploid and tetraploid members of the Phlox pilosa complex. A. Flow cytometry histogram for diploid P. pilosa ssp. longipilosa PZSH2011-043. B. Meiotic metaphase chromosome count of P. pilosa ssp. longipilosa PZSH2011-043 with n=7 chromosomes. c. Flow cytometry histogram for tetraploid P. pilosa ssp. pilosa PZ12-124 collected in Polk County, NC. d. Mitotic metaphase chromosome counts of P. pilosa ssp. pilosa PZ12-124 with 2n=4x=28 chromosomes.

255

Figure 4.3: Population clustering analysis suggests significant structure among members of the Phlox pilosa complex. ΔK scores for each value of K genetic clusters following Evanno et al. (2005). The figure was generated using Structure Harvester (Earl et al., 2011).

256

Figure 4.4: Barplots depicting STRUCTURE results of K=4 clusters of 61 samples from 8 populations of the Phlox pilosa complex. Populations are grouped by cluster, in order of highest mean Q score (cluster identity proportion). Ploidy differences are expressed on the top of the barplot.

257

Subsection Divaricatae sensu Wherryz Subsection Divaricatae (Wherry) Pratherz

Taxon Taxon Common Name

Phlox amoena Sims ssp. amoena Wherry Phlox amoena Sims Chalice Phlox Phlox amoena Sims ssp. lighthipei (Small) Wherry

Phlox divaricata L. ssp. divaricata Phlox divaricata L. ssp. divaricata Wood Phlox Phlox divaricata L. ssp. laphami i (Wood) Wherry Phlox divaricata L. ssp. laphami i (Wood) Wherry

Phlox floridana Bentham ssp. floridana Phlox floridana Bentham Florida phlox Phlox floridana Bentham ssp. bella Wherryx

Phlox pattersonii Prather Coahuila phlox

Phlox pilosa L. ssp. pilosa Phlox pilosa L. ssp. pilosa Downy phlox Phlox pilosa L. ssp. detonsa (Gray) Wherryw Phlox pilosa L. ssp. deamii Levin Deam's downy phlox Phlox pilosa L. ssp. fulgida (Wherry) Wherry Phlox pilosa L. ssp. fulgida (Wherry) Wherry Dakota downy phlox Phlox pilosa L. ssp. ozarkana (Wherry) Wherry Phlox pilosa L. ssp. longipilosa (Waterfall) Locklear Kiowa downy phlox Phlox pilosa L. ssp. ozarkana (Wherry) Wherry Ozark downy phlox Phlox pilosa L. ssp. sangamonensis Levin Sangamon downy phlox

Phlox pilosa L. ssp .pulcherrima (Wherry) Wherry Phlox pulcherrima (Lundell) Lundell Big Thicket phlox

Phlox pilosa L. ssp. latisepala (Wherry) Wherry Phlox villosissima Turner Comanche phlox Phlox pilosa L. ssp. riparia (Wherry) Wherry

zThis classification is derived from the classification of Wherry, 1955. ySubsection Divaricatae was expanded to include subsection Drummondianae (Prather, 1994). This change was followed here but annual taxa were not included in this table. xWherry ()described P. floridana ssp. bella as a dwarf form of P. floridana , but did not include it in his monograph (1955). It is assume that this is an enviornmental variant not worthy of further recognition. w Now recognized as a glabrous form of P. pilosa ssp. pilosa (Ferguson et al., 1999).

Table 4.1: Taxonomic overview of perennial taxa in subsection Divaricatae within or related to the P. pilosa complex. The classification of Wherry (1955) is on the left, and on the right a modern classification system derived from several sources (Ferguson et al., 1999; Levin 1963; Levin, 1966; Locklear, 2011a; Prather, 1994).

258

z y x Chromosome Taxon Collection Site Accession (No.) 2C ± SD (pg) CV (%) Ploidy w No. (n )

Section Annuae

Subsection Divaricatae

Phlox amoena Sims Benton Co. TN PZ12-053 11.24±0.18 1.61 2x 7 McCreary Co. KY PZ11-032 11.58±0.33 4.19 2x w 7 Logan Co. KY PZ10-200 12.03±0.08 4.56 2x 7 Cumberland Co. TN PZ12-094 12.11±0.17 1.36 2x 7 Polk Co. TN PZ12-133 12.13±0.25 2.43 2x 7 Campbell Co. TN PZ12-102 12.14±0.23 2.18 2x 7 McCreary Co. KY PZ11-062 12.22±0.08 4.66 2x 7 Bibb Co. AL PZ11-068 12.26±0.11 3.15 2x 7 Tallapoosa Co. AL PZSH2011-001 12.42±0.11 2.1 2x 7

Phlox divaricata L. ssp. divaricata Hocking Co. OH PZ10-145 9.18±0.22 2.26 2x 7 Woodford Co. KY PZ11-074 9.23±0.10 2.98 2x 7 Clay Co. MS PZSH2011-035 9.43±0.27 2.58 2x 7 Polk Co. TN PZ12-133 9.75±0.11 1.65 2x 7 Union Co. TN PZ12-113 9.76±0.02 1.38 2x 7 Hardin Co. KY PZ12- 10.08±0.26 2.96 2x 7 McCreary Co. KY PZ11-003 10.51±0.01 2.19 2x 7

Phlox divaricata divaricata /laphamii mixed population Clark Co. IL PZ11-005 9.86±0.14 2.29 2x 7

Phlox divaricata L. ssp. laphami i (Wood) Wherry Wilkinson Co. MS PZSH2011-026 9.78±0.13 1.73 2x 7 Etowah Co. AL PZ13-002 9.93±0.09 1.96 2x 7 Gadsden Co. FL PZSH2011-008 10.02±0.48 2.15 2x 7 Gadsden Co. FL PZSH2011-004 10.37±0.13 2.03 2x 7 Jackson Co. FL PZSH2011-012 10.41±0.23 2.22 2x 7 Wilcox Co. AL PZSH2011-018 10.49±0.09 2.38 2x 7 Gadsden Co. FL PZSH2011-007 10.8±10.49 1.99 2x w 7 Walker Co. GA PZ13-001 10.81±0.06 2.03 2x 7 Montgomery Co. AL PZ11-039 10.98±0.07 1.77 2x 7 Newton Co. AR PZ13-008 11.24±0.28 2.55 2x 7 Garland Co. AR PZ13-010 11.61±0.05 2.65 2x 7

Phlox drummondi i ssp. mccallisteri Wilson Co. TX PZ10-163/TX-048 12.21±0.08 2x 7 Phlox drummondi i ssp. drummondii Caldwell Co. TX PZ10-160/TX-035 12.22±0.10 2.16 2x 7 Phlox drummondi i ssp. drummondii Caldwell Co. TX PZ10-161/TX-037 12.27±0.04 2.36 2x 7

Phlox floridana Bentham Jackson Co. FL PZSH2011-10 21.85±0.30 3.55 4x 14

(Continued)

Table 4.2: Relative Holoploid (2C) genome sizes and ploidy levels of a diverse collection of taxa from Phlox subsection Divaricatae collected from natural plant populations (Figure 7.2).

259

Table 4.2: Continued

Bustamante Canyon, Phlox pattersonii Prather Nuevo Leon, Mexico; PZ10-247x 14.87±0.09 2.26 2x 7 Ax

Phlox pilosa L. ssp. deamii Levin Spencer Co. IN PZ12-090 11.26±0.10 2.29 2x 7 Christian Co. KY PZ12-045 11.96±0.44 3.33 2x 7

Phlox pilosa L. ssp. deamii Levin Hybrid Swarm Benton Co. TN PZ12-054 12.04±0.50 2.08 2x 7

Phlox pilosa L. ssp. fulgida (Wherry) Wherry Crawford Co. WI PZ11-066 12.34±0.38 2.78 2x 7 Phlox pilosa L. ssp. fulgida (Wherry) Wherry Story Co. IA PZ12-092 12.76 2.18 2x 7

Phlox pilosa L. ssp. longipilosa (Waterfall) Locklear Greer Co. OK PZSH2011-043 12.55±0.23 4.53 2x w 7

Phlox pilosa L. ssp. ozarkana (Wherry) Wherry Iron Co. MO PZ12-059 10.52±0.005 3.49 2x 7 Phlox pilosa L. ssp. ozarkana (Wherry) Wherry Butler Co. MO PZ12-058 10.70±0.02 2.32 2x 7 Phlox pilosa L. ssp. ozarkana (Wherry) Wherry Johnson Co. AR PZ10-227 10.78±0.33 3.52 2x 7 Phlox pilosa L. ssp. ozarkana (Wherry) Wherry Newton Co. AR PZ13-009 10.79±0.17 3.12 2x 7 Phlox pilosa L. ssp. ozarkana (Wherry) Wherry Johnson Co. AR PZ10-228 10.92±0.34 2.58 2x w 7

Phlox pilosa L. ssp. pilosa Phelps Co. MO PZ11-071 11.05±0.03 4.56 2x 7 Phlox pilosa L. ssp. pilosa Rapides Parish LA PZSH2011-29 11.11±0.14 2.28 2x 7 Phlox pilosa L. ssp. pilosa Benton Co. TN PZ12-052 11.29±0.09 1.93 2x 7 Phlox pilosa L. ssp. pilosa Vernon Parish LA PZSH2011-31 11.34±0.32 2.74 2x 7 Phlox pilosa L. ssp. pilosa Scioto Co. OH PZSH2011-025 11.78±0.17 1.94 2x 7 Phlox pilosa L. ssp. pilosa Franklin Co. OH PZ11-013 11.82±0.67 1.91 2x w 7 Phlox pilosa L. ssp. pilosa Pulaski Co. IN PZ12-061 11.88±0.07 2.84 2x 7 Phlox pilosa L. ssp. pilosa Chickasaw Co. MS PZ11-034 11.88±0.19 2.16 2x 7 Phlox pilosa L. ssp. pilosa Hardin Co. KY PZ12-040 11.90±0.21 1.52 2x 7 Phlox pilosa L. ssp. pilosa Marshall Co. KY PZ12-048 12.01±0.13 1.77 2x 7 Phlox pilosa L. ssp. pilosa Scott Co. MS PZSH2011-21 12.17±0.10 4.05 2x 7 Phlox pilosa L. ssp. pilosa Wasington Parish LA PZSH2011-023 12.18±0.35 2.23 2x 7 Phlox pilosa L. ssp. pilosa Pearl River Co. MS PZSH2011-024 12.22±0.27 2.21 2x 7 Phlox pilosa L. ssp. pilosa Lucas Co. OH PZ10-192 12.34±0.17 3.90 2x 7 Phlox pilosa L. ssp. pilosa Scott Co. MS PZSH2011-020 12.40±0.40 2.74 2x 7 Phlox pilosa L. ssp. pilosa Crenshaw Co. AL PZSH2011-14 12.49±0.10 2.17 2x 7 Phlox pilosa L. ssp. pilosa Lowndes Co. AL PZSH2011-017 12.66±0.11 1.57 2x 7 Phlox pilosa L. ssp. pilosa Gadsden Co. FL PZSH2011-6 13.12±0.02 2.91 2x 7

Phlox pilosa L. ssp. pilosa Polk Co., NC; B PZ12- 22.71±20.13 1.75 4x w 14 Phlox pilosa L. ssp. pilosa Somewhere in AL; C PZ11-059 21.89±0.21 1.47 4x w 14

Phlox pilosa L. ssp. pilosa Bossier Parish LA PZ11-069 23.06±0.52 2.28 4x w 14 Phlox pilosa L. ssp. pilosa Bossier Parish LA PZ11-070 22.36±0.44 2.49 4x 14

Phlox pilosa ssp. sangamonensis Levin Champaign Co. IL; D PZ10-203 11.33±0.14 3.04 2x 7

(Continued)

260

Table 4.2: Continued

Phlox pulcherrima (Lundell) Lundell Shelby Co. TX PZSH2011-033 21.77±0.41 4.63 4x 14 Phlox pulcherrima (Lundell) Lundell Shelby Co. TX PZSH2011-034 22.05±0.07 2.46 4x 14 Phlox pulcherrima (Lundell) Lundell Jasper Co. TX PZSH2011-035 23.66±0.67 4.06 4x 14

Phlox villosissima San Saba Co. TX PZSH2011-42 26.58±1.56 4.55 4x w 14 Phlox villosissima Bandera Co. TX PZSH2011-39 26.79±0.12 0.95 4x 14 Phlox villosissima Kerr Co. TX PZSH2011-40 26.85±0.69 3.52 4x 14 Phlox villosissima Kerr Co. TX PZSH2011-36 26.88±0.75 1.42 4x 14

Phlox roemeriana Scheele Ferg mono Comal Co. TX PZ10-165/TX-057 13.04±0.06 2.11 2x 7

Subsection Nanae

Phlox nana Nuttall E PZ12-131 10.22±0.01 1.25 2x w 7 zNote: Collection site: A, plant obatined from Arrowheadalpines.com; B, plant donated by Plant Deights Nusery; C, plant obtained from MulberryWoodsN plants donated by Wesley Whiteside, Charleston, IL ; E, plants donated by Denver Botanic Garden. yMean 2C DNA content (mean ± standard deviation). xChromosome counts inferred from Flory, 1931; Flory, 1934; Flory, 1937; Meyer, 1944; Smith & Levin, 1967; Worcester et al., 2013. windicates that ploidy was determined with chromsome counts.

261

Taxon Cultivar Accession No. 2C ± SD (pg) CV (%) Ploidy Level no. Sourcez

Phlox drummondii Intensia® Blueberry −−− 12.54±0.08 1.49 2x 14 A Intensia® Orchid Blast −−− 12.91±0.05 1.40 2x 14 A Intensia® Cabernet −−− 12.76±0.07 1.49 2x 14 A Palona Deep Lavender with Eye PI601603 12.24±0.05 2.20 2x 14 B mixed colors seed strain PZ10-202 12.38±0.14 2.27 2x 14 C

Phlox divaricata 'Charleston Pink' PZ10-234 9.66±0.18 1.40 2x 14 D 'Manita' PZ10-036 9.83±0.10 1.57 2x 14 E 'Lemon Slice' PZ10-131 9.89±0.09 1.89 2x 14 F

Phlox pilosa 'Eco Happy Traveler' PZ10-241 22.14±0.45 1.38 4x 28 G 'Forest Frost' PZ10-022 12.55±0.08 2.99 2x 14 E 'Lavender Cloud' PZ10-025 11.42±0.26 2.10 2x 14 E 'Racy Pink' PZ10-024 11.16±0.17 2.74 2x 14 E 'Slim Jim' PZ10-133 24.47±0.31 1.63 4x 28 F zNote: Plant Source: A, plant obtained from Proven Winners; B, plant obtained fromOrnamental Plant Germplasm Center; C, plant obtaine Wildseed Farms; D, plant obatined from Seneca Hill Perennials; E, plant obtained from Growild Nursery; F, plant obtained from Plant De Nursery; G, plant obtained from Nearly Native Nursery.

Table 4.3: Relative Holoploid (2C) genome sizes and ploidy levels of cultivars selected from Phlox subsection Divaricatae.

262

Taxon Accession (No.) Population ID County/Parish State West North Ploidy

P. pilosa ssp. pilosa PZSH2011-017 AL Lowndes AL 86°71'76.7" 31°97'24.2" 2x P. pilosa ssp. pilosa PZSH2011-020 MS Forest MS 89°47'60.0" 32°33'86.8" 2x P. pilosa ssp. pilosa PZSH2011-029/030 LA1 Rapides LA 92°56'13.7" 31°07'78.6" 2x P. pilosa ssp. pilosa PZ11-069 LA2 Bossier LA 93°30'26.9" 32°51'53.9" 4x P. pilosa ssp. pilosa PZ11-070 LA3 Bossier LA 93°37'12.0" 35°53'09.7" 4x

P. pilosa ssp. ozarkana PZ10-227 AR1 Johnson AR 93°29'68.9" 35°65'12.9" 2x P. pilosa ssp. ozarkana PZ10-228 AR2 Johnson AR 93°31'43.1" 35°57'67.2" 2x

P. villosissima PZSH2011-040 TX Kerr TX 98°07'19.6" 30°11'69.4" 4x

Table 4.4: Phlox pilosa complex sampling localities, number of individuals per sampling site, and ploidy. These data are a subset of the taxa used in the genome size analysis.

263

z Locus Primer Sequence (5' →3') Alleles (No.)Ranges of alleles (bp

Phl33 F: GTCCCCCAGCTGATACTGG 39 121-183 R: CTTGCTTTGGCTTCTCCAAC

Phl68 F: ACGCAACAACCAAACTCCA 29 242-302 R: GATGCAGCCACACGAGTTTA

Phl84 F: TCAGACTAGGGGAGGCAGT 39 170-226 R: TTCTTTACCGCTGGCTGAAT

Phl113 F: TGTCCACATGGGCTTGACT 35 462-554 R: ACGTACACGCCCAACTAAGG

Phl137 F: TCGGGCACCAGATTTTATTC 9 202-228 R: TTCGACCCCCAGATAGTCAG

M4-M5 F: GTTCGCCGGAGATAGTTAC 3 158-164 R: GGTAAACCCACGGGAGAACT

zAll markers beginning with Phl are afrom Fehlberg & Ferguson (2008), M4-M obtained from S. Fehlberg (unpublished primer sequence, pers. comm. 2010).

Table 4.5: Characterization of the six SSR markers based on eight Phlox pilosa complex populations.

264

Population Population ID N micro AA ' X H o H ' N 'p

P. pilosa ssp. pilosa PZSH2011-017 AL 8 46 5.75 8.13 0.50 1.35 5.50 PZSH2011-020 MS 5 26 3.25 8.6 0.50 1.43 3.67 PZSH2011-029 LA1 7 37 4.63 7.56 0.44 1.26 4.50 PZ11-069 LA2 9 32 4.00 10.44 0.68 1.74 3.83 PZ11-070 LA3 9 30 3.75 10.00 0.65 1.67 4.17

P. pilosa ssp. ozarkana PZ10-227 AR1 8 45 5.63 8.36 0.50 1.40 5.83 PZ10-228 AR2 8 37 4.63 8.29 0.43 1.38 4.67

P. villosissima PZSH2011-040 TX 7 44 5.5 11.86 0.64 1.98 4.67

Total Mean/range 61 297 4.64 9.16 0.54 1.53 4.61

N micro = sample size per population. A = Number of alleles from all loci. A ' = Number of alleles observed in the population, averaged over loci. X = Number of alleles per individual, averaged over loci.

H o = Proportion of heterozygotes, averaged over loci. H ' = Number of alleles per locus in an individual, avergaed over loci. N 'p = Number of different single locus allelic phenotypes observed, averaged over loci.

Table 4.6: Descriptive statistics and genetic diversity measures for eight populations of taxa in the Phlox pilosa complex subject to SSR analysis. Genetic diversity measures were calculated from a binary data set where microsatellite alleles were coded as present or absent.

265

Variance Percentage of Source df Sum of squares component variation

Among Populations 7 21.957 0.274 21% Within Populations 53 55.748 1.052 79% Total 60 77.705 1.326 100%

Table 4.7: Hierarchical analysis of molecular variation (AMOVA) among populations from the Phlox pilosa complex illustrating the proportion of variation attributable to differences among populations and within populations.

266

Loci F IS F I T F ST h

Phl33 0.246 0.196 0.062 0.969 Phl68 0.395 0.308 0.126 0.974 Phl84 0.283 0.168 0.138 0.969 Phl113 0.502 0.421 0.139 0.971 Phl137 0.326 0.156 0.201 0.835 M4-M5 -0.291 -0.471 0.123 0.211

Mean 0.243 0.212 0.131 0.821

F IS = Inbreeding coefficient at the population level.

F IT = Inbreeding coefficient at the total sample level.

F ST = Proportion of differentiation among populations. h = Nei's gene diversity.

Table 4.8: Genetic differentiation and genetic diversity among eight populations of taxa from the P. pilosa complex.

267

CHAPTER 5

INTERSPECIFIC HYBRIDIZATION IN PHLOX FOR GERMPLASM ENHANCEMENT: HYBRIDIZATION POTENTIAL OF PHLOX PANICULATA AND RELATED SPECIES

Abstract

Phlox paniculata, the garden phlox, is probably the most widely cultivated

perennial Phlox species in the world. Interspecific hybrids involving this species have

been variously reported but conclusive evidence for such hybrids is lacking. The extent

to which interspecific hybridization can be used as a tool for improvement of garden

phlox is unclear. The objective of this study was to determine the potential for

interspecific hybridization between P. paniculata (Section Phlox, subsection

Paniculatae) and other Phlox species, primarily those in subsections Paniculatae, Phlox

and Divaricatae. Parental material included both cultivars and plants of wild origin.

Crosses of P. paniculata with a species within subsection Paniculatate were successful whereas those with species in subsections Phlox and Divaricatae were much more limited. Only one of 5 species in subsection Phlox and one of 4 species in subsection

Divaricatae yielded viable hybrids. Interspecific hybrids were possible between P. paniculata and P. amplifolia, P. carolina-P.glaberrima complex and P. divaricata.

Except for the reciprocal crosses possible with P. amplifolia, crosses with the other species displayed unilateral incongruity, reduced seed set and low germination. The 268

interspecific progeny that developed normally showed intermediate morphology and had

a genome size and ploidy consistent with that of a hybrid origin. The hybrids were

mostly sterile. P. amplifolia (subsection Paniculatae) was successfully crossed with P.

ovata (subsection Phlox) but P. paniculata was not. The cross P. divaricata x P. paniculata, reported as the source of the commercially popular hybrid P. xarendsii, was

successful but at a very low frequency, with only one of six cross combination resulting

in viable seeds. The hybrids had a mean genome size of 12.31 pg, intermediate to the

genome size of parental taxa, but different from the size of cultivars of P. xarendsii,

which are similar to that of P paniculata. Interspecific hybridization as a tool for P.

paniculata breeding appears to be limited although the few successful combinations

provide some opportunity for manipulation of desirable traits.

Introduction

Interspecific hybridization arguably plays an important a role in speciation

(Morgan et al. 2010; Rieseberg, 2009; Soltis et al., 2009) but is also considered the

primary method for generating genetic variation and novelty in ornamental crops

(Eeckhaut et al., 2006; Kato and Mii, 2012; Van Tuyl and de Jeu, 2005). Numerous pre-

and post-zygotic barriers exist that prevent the development of hybrids in several genera

of ornamental plants, but these barriers remain poorly defined, anecdotal in nature, or

unknown for several genera, including Phlox (Levin, 1966; Lierval, 1866; Locklear,

2011a; Symons-Jeune, 1953; Wherry, 1935a). The genus Phlox has played an important role in our understanding of plant speciation and evolution, including the possible role of

269 interspecific hybridization as a driver for speciation (Levin, 1966) but the extent to which such hybridization potential has been applied in a horticultural context appears limited.

Part of the limitation may be due to various factors such as a lack of appropriate germplasm resources, inherent incompatibilities between the horticultural forms of Phlox, but also due to the fact that most breeding activity in ornamental plants occurs in private companies where information about such hybridization is not published. This lack of information hinders advancement in more broad-based breeding efforts for the genus.

In Phlox, numerous prezygotic barriers to interspecific hybridization such as style length, ploidy, and phylogenetic distance have been suggested, but few of these have been systematically tested (Eeckhaut et al., 2006; Levin, 1966; Wherry, 1955). Presence of post zygotic barriers in Phlox is suggested by the partial development of fruit and seed after interspecific crosses, perhaps resulting from endosperm imbalance (Levin, 1966).

There are various reports or claims for the existence of interspecific hybrids in cultivated

Phlox, but evidence that supports such claims is scant (Flory, 1933; Flory, 1934;

Locklear, 2011a; Pridham, 1934; Symons-Jeune, 1953). The extent to which interspecific hybridization can be used to enhance ornamental attributes in this genus thus appears underexplored, but given the previous history of hybridization (Levin, 1966;

Pridham, 1934; Symons-Jeune, 1953) and an increased interest in developing novel combinations of traits in Phlox hybrids for use by the floriculture and nursery industry, a more comprehensive assessment of barriers to hybridization among species of Phlox is warranted.

270

Although a relatively minor crop, the Census of Horticultural Specialties (USDA-

NASS, 2010) reported that 2009 sales of cultivated phlox totaled more than $16.2 million

in the United States. Most of this value is attributed to P. paniculata, and is supported by

the presence of several commercial breeding and development programs dedicated to that

species in the United States and Europe. Other commercially important species include

P. subulata and P. drummondii.

Phlox L. (Polemoniaceae) is a North American genus of ca. 65 species of

ornamental, herbaceous plants known for their long-lasting floral displays in a variety of colors (Bendtsen, 2009; Wherry, 1955). Of the 65 species, 23 are native to the eastern

United States; they are organized into 3 sections (Annuae, Occidentales, Phlox) and 6 subsections that are divided primarily on the basis of style length. ‘Long-styled’ Phlox species have a style greater than 15 mm and belong to subsections Phlox and

Paniculatae; style length does not exceed 12 mm among taxa in the remaining subsections. Phlox paniculata and P. amplifolia comprise subsection Paniculatae; over

800 cultivars of P. paniculata have been described, but P. amplifolia is rarely cultivated.

Both species can be distinguished from taxa in subsection Phlox by their areolate leaf veins and white or cream colored pollen (Wherry, 1933, 1955; Chapter 2). Phlox paniculata has many nodes per flowering stem, eglandular pubescence, a pubescent corolla tube, and often cream-colored, or white pollen. It is a species of hydric environments and occurs in alluvial soils in riparian zones and is often closely associated with rivers and streams. Phlox amplifolia is morphologically similar to P. paniculata, but differs in having fewer nodes per stem, a glandular pubescent inflorescence, and a

271

glabrous corolla tube; it also occurs in more xeric habitats in open forests, frequently on

steep slopes (Deam, 1940; Wherry, 1955).

Subsection Phlox is comprised of 6 species that are morphologically similar to the

two species in subsection Paniculatae. Phlox carolina and P. glaberrima are widely distributed, phenotypically variable species that have been divided into up to 8 subspecies, but because of uncertain taxonomy, are also referred to as the P. carolina-P.

glaberrima complex (Wherry, 1955). The remaining taxa, P. idahonis, P. maculata, P.

ovata, and P. pulchra, are readily differentiated from each other by unique morphological

attributes (Chapter 1). The most commonly cultivated taxa in subsection Phlox are

members of the P. carolina-P. glaberrima complex and P. maculata (Bendtsen, 2009;

Fuchs, 1994; Locklear, 2011a). Taxa from subsections Paniculatae and Phlox are known

collectively as the ‘garden phloxes.’ Despite a history of reported interspecific

hybridization, fundamental questions remain about the potential for hybridization among

these species (Lierval, 1866; Locklear, 2011a; Pridham, 1934; Symons-Jeune, 1953).

Phlox paniculata has a long and complicated history of interspecific hybridization

with epicenters of activity on two continents and numerous countries (Bendtsen, 2009;

Fuchs, 1994; Lierval, 1866; Locklear, 2011a; Pridham, 1934). Various hybrids are

described in the literature, but three main groups share P. paniculata as a parent (Table

5.1). Circumscription of hybrid taxa has been plagued by taxonomic confusion in part

because phenotypic variation in geographically disparate populations of P. paniculata has

resulted in description of these populations as separate species (Pridham, 1934; Symons-

Jeune; 1953). Interspecific hybrids involving P. paniculata and species from subsections

272

Divaricatae and Phlox have been described but these have not been systematically

verified (Arends, 1912; Flory, 1931). For example, in the original description of Phlox

xarendsii (P. divaricata x P. paniculata), no crossing data was reported and it is not clear if the hybrid developed as chance seedlings from open pollinated plants, or by a deliberate effort to create hybrids using controlled pollinations (Arends, 1912). In another study, cultivars of P. xarendsii did not exhibit the meiotic irregularities seen in

the confirmed interspecific hybrid Phlox xprocumbens (P. subulata x P. stolonifera),

suggesting that the cultivars may not be interspecific hybrids (Lehmann, 1828; Flory,

1931). In addition, pollen of P. xprocumbens was generally sterile, whereas pollen of P.

xarendsii had viability comparable to that of P. paniculata, again questioning the hybrid

origin of putative P. xarendsii cultivars. Both P. xarendsii and P. xleopoldiana, if indeed

hybrids, would be the product of crosses between short-styled taxa from subsection

Divaricatae and long-styled Paniculatae (Table 5.1). One study of interspecific

hybridization involving long-styled P. paniculata and short-styled P. drummondii

reported that crosses failed when P. paniculata was used as the female parent and that

only 3 empty capsules were formed when P. drummondii served at the female parent;

however, there is at least one report of successful hybridization between these species

(Flory, 1931; Kelly, 1915; Wherry, 1955). Other reported P. paniculata hybrids such as

Phlox Suffruticosa Group (syn. P. xdecussata), have been described as products of

crosses with taxa in subsection Phlox, but again, these hybrids have not been systematically verified (Locklear, 2011a; Pridham, 1934; Symons-Jeune, 1953). Thus, the true potential for interspecific hybridization with P. paniculata remains unclear.

273

In a broader context, the use of related wild species in interspecific hybridization

experiments with a crop plant helps define the gene pools for genetic improvement of

important ornamental crop species such as P. paniculata (Harlan and De Wet, 1971).

Phlox amplifolia is rare in the wild and not widely cultivated, although there are some commercial cultivars attributed to this species (Bendtsen, 2009; Fuchs, 1994; Deam,

1940; Wherry, 1955). The widely grown P. paniculata ‘David’ is sometimes reported as

P. amplifolia instead, but examination of morphological characteristics indicate it is a P. paniculata selection. ‘David’ is grown especially for its resistance to powdery mildew

(Erysiphe cichoracearum). Limited field observations I have made of P. amplifolia in the wild indicated no presence of powdery mildew, so this species could be a source of resistance, if such resistance can indeed be shown under controlled conditions. There are no previous reports of successful interspecific hybridization between P. amplifolia and P. paniculata and related taxa.

Similarly, experimental hybridization between taxa in subsection Paniculatae and taxa in subsection Phlox has not been sufficiently tested and reported. The species in subsection Phlox, such as the P.carolina-P.glaberrima complex and P. ovata have particularly desirable habit and foliage attributes often described by growers as consistently free of powdery mildew; this trait may be of great value in P. paniculata if

interspecific trait transfer is possible and if it is heritable (Bir, 1999; 2003).

Among various factors that may influence the success of interspecific

hybridization is the genotype of the parents (Kato and Mii, 2012). It is common for

breeders of ornamental plants to use existing cultivars of different species in interspecific

274

hybridization attempts; such crosses are based on the narrow genetic base of the cultivars

and failures may not be due to species-specific incompatibilities, but rather due to unique

genetic differences of the chosen cultivars. The use of genetically diverse, wild collected

germplasm provides an opportunity to more efficiently test compatibility between species

that have not previously been evaluated. It also provides an opportunity to re-examine

previously reported hybrid combinations (Locklear, 2011a; Symons-Jeune, 1953)

The objectives of this study were: 1) To test the potential for hybridization between P. paniculata and a diverse selection of related Phlox germplasm accessions from subsections Paniculatae, Phlox and Divaricatae in an attempt to define P. paniculata gene pools and also recreate previously published hybrids. 2) To determine the rate of seed set and germination in successful crosses. 3) To verify formation of F1

hybrids using morphology and genome size as estimated by flow cytometry.

Materials and Methods

Phlox collection

Living plants were obtained from nursery sources and from collection sites in the

east-central United States that were visited during a series of botanical expeditions

carried out from 2010-2013 (Table 5.2). Emphasis was placed on obtaining cultivars that

were known to have originated as clones or seedlings selected directly from wild

populations. The identity of wild collected material was confirmed using the key of

Wherry (1955) and comparison with herbarium specimens at the Museum of Biological

275

Diversity of The Ohio State University. Herbarium vouchers are currently maintained at the Ornamental Plant Germplasm Center (OPGC) in Columbus, OH.

Controlled Pollinations

Plants to be used in hybridizations were grown in a glass greenhouse following standard greenhouse procedures at the OPGC in Columbus, OH; all growth compartments were screened to minimize the possibility of contamination from pollinating insects. Phlox taxa used in this study are protandrous, out-crossing species with gametophytic self-incompatibility (SI); so individual flowers can receive and donate pollen with limited risk of self-pollination (Levin, 1975; Ruane and Donohue, 2007;

Chapters 6,7). These taxa are described as “short-styled” and “long-styled” (Table 5.2;

Figure 2.1 in Chapter 2). “Short style” flowers have styles no more than 4 mm long, a tripartite stigma with lobes nearly equal to the length of the pistil, and the entirety of the pistil surrounded by the calyx lobes. In contrast, “long-style” flowers have a style up to

25mm long that is exerted from the calyx, and sometimes the corolla (Ferguson et al.,

1999; Wherry, 1955).

Pollinations were carried out on flowers with shedding pollen and with open tripartite stigma. The stigma of long-styled species is easily accessed and readily pollinated after removal of the corolla. To access the stigma of short-styled species, the corolla must be carefully separated from the corona and the calyx lobes removed with scissors. The calyx of several taxa in subsection Divaricatae are covered in glandular pubescence that becomes viscid when handled and can be easily damaged. Early

276

experimentation indicated that either removal or excessive handling of the calyx resulted in abortion of the flower and failure of the cross.

Fifty pollinations (individual flowers) were made per cross. Dehiscing anthers

were brought in contact with the stigma and a record made of each cross. Pollinations

were performed over a 1-2 week period. Successful crosses were defined as the

formation of at least 5 viable seeds per cross. Nearly all Phlox species can develop a

maximum of three seeds per ovary. Mature fruit have ballistic seed dispersal so

developing infructescences were bagged with nylon cloth (pantyhose) to aid seed

collection (Wherry, 1955). Seeds of species in subsections Paniculatae and Phlox mature

approximately 60-90 days after pollination. Seeds were harvested when they had been

released and were visible inside the nylon bag that surrounded the fruits.

Reciprocal crosses where one parent was either P. amplifolia or P. paniculata were performed during 2011 and 2012. The accessions, P. paniculata PZ10-109 with

white pink flowers and PZ10-231 with pastel pink flowers were used as recurring

parents; each of these selections were crossed with all of the accessions listed in Table

5.2, except P. amplifolia. Due to a limited number of both P. amplifolia plants and some accessions of P. paniculata, they were used in a partial diallel mating system when possible.

Though not carefully studied, the majority of eastern Phlox taxa appear to exhibit

non-deep physiological dormancy that is overcome by stratification (Baskin and Baskin,

2004; Ridout and Tripepi, 2009). Seeds were stored at 4C and 25% relative humidity

until sown in germination boxes (Tri-State Plastic, Henderson, KY) on blue blotter paper

277

(Anchor Paper Co., St. Paul, MN) and stratified at 7.2°C for 60-90 days. The boxes were monitored weekly and water was added as needed to make sure the paper was moist at all times. Following stratification, seeds were transferred to a growth chamber at 22°C in the dark. Seeds were considered to have germinated when a 2 mm radicle was evident.

Germinated seeds were transplanted to individual plastic containers (11 cm diameter round by 9.5 cm deep, Dillen Products, Middlefield, OH) in Metro mix 360

(Scotts-Miracle Gro Co., Marysville, OH) and grown in the greenhouse under natural light conditions at temperatures of 22.2 °C ± 6 °C 0800 to 1800 HR and 18.3 °C ± 3 °C from 1800 to 0800 HR.

Seedlings were given accession numbers as follows; PZ refers to the author, 11 refers to the year the pollinations were made, and the number following refers to a specific cross combination. Progeny from a cross were also given individual accession numbers to distinguish them from sibs in a given hybridization.

Data Analysis

Analysis of variance and means separation tests were performed in SAS (Proc

GLM, SAS Version 9.3; Cary, NC). Means separation were determined by Fisher’s least significant difference at P = 0.05.

278

Results

Parents for Hybridization

Hybridizations were attempted between P. paniculata and 10 species, one from subsection Paniculatae, five from subsection Phlox, and four from subsection

Divaricatae; in all, 38 different accessions were used in hybridization experiments (table

5.2). Taxa in subsections Paniculatae and Phlox are characterized by long styles whereas those of subsection Divaricatae have short styles. To confirm the ploidy of all taxa, genome size was measured with flow cytometry (Chapters 2-4). The majority of

accessions were diploid; the genome size for these ranged from 2C=9.18 pg of DNA to

16.13 pg. Nine accessions of P. paniculata were used; three were cultivars, and six were either directly collected from wild populations or were selections of wild origin made by

nurserymen. The accessions of P. paniculata have unique phenotypes specifically

chosen to provide possible morphological markers that could be used to interpret the

outcome of hybridization. For example, the cultivar ‘Delta Snow’ was chosen because it

has atypical white flowers with a pink eye. Similarly, two individuals of P. paniculata

accession PZ10-231 were selected because of atypical pure white and pastel pink flowers;

PZ10-109 was also chosen because of its pure white flowers (Table 5.2). Except for

Phlox ‘Minnie Pearl’ that also has white flowers, all other taxa had variously pink or purple-pigmented flowers that could be used as possible markers in the resulting progeny.

Two cultivars of P. paniculata, ‘John Fanick’ and ‘Robert Poore,’ are aneuploids with mean genome sizes intermediate to diploid and tetraploid accessions and are pollen sterile

(Chapter 2); these were included in some hybridization attempts to determine if they had

279

sufficient female fertility for seed production. From subsection Phlox, 13 accessions

representing 8 taxa were used in crosses; 11 accessions are diploid, but P. glaberrima

ssp. triflora ‘Bill Baker’ and ‘Anita Kistler’ are tetraploid with a mean genome size of

2C= 26.70 pg. From subsection Divaricatae, 14 accessions representing 4 species were

also used; 13 accessions were diploid and had the smallest genome sizes of all taxa in the

hybridizations, 2C= 9.18-12.66 pg. The tetraploid P. floridana had a mean genome size

of 2C=21.85 pg, the largest among subsection Divaricatae taxa used in this study.

Overview of Hybridization

Approximately 4000 pollinations were attempted among the various accessions

used in these experiments. There were far more unsuccessful than successful crosses.

Crosses between P. paniculata and 10 other species resulted in viable hybrids only with 3

species: P. amplifolia, P. carolina-P.glaberrima complex, and P. divaricata (Table 5.3).

The P. carolina-P.glaberrima complex included two taxa, accession PZ11-036 keyed to

P. carolina, and the cultivar Phlox ‘Minnie Pearl.’ Morphological evidence indicates that this cultivar has affinity to P. carolina but resolution requires a more thorough phylogenetic analysis. F1 individuals with P. paniculata as a parent were produced from

6 interspecific cross combinations, three with P. amplifolia, two with P

carolina/glaberrima and one with P. divaricata (Table 5.3). In addition, the other

species in subsection Paniculatae, P. amplifolia, could also be hybridized with P.

carolina and P. ovata (Table 5.3). The majority of hybrid combinations between

subsection Paniculatae and subsections Divaricatae and Phlox were unsuccessful.

280

Hybridizaton within subsection Paniculatae

When two accessions of P. paniculata were crossed, seed production and germination was, as expected, relatively high. This cross served as a general reference for comparison with the interspecific crosses. The 50 pollinations in each direction yielded 138 and 112 seeds, a pollination ‘efficiency’ 2.8 and 2.2 seeds per pollination.

This rate of seed set is comparable to rates presented from other species (Levin and

Kerster, 1975). The germination of these seeds averaged only about 56%, but there is very little information about overall germination rates in seed of P. paniculata, so this level may not be atypical for the species. Preliminary studies of P. paniculata seed at the

OPGC show that seed germination rates are generally lower than those of other Phlox species, perhaps due to seed quality parameters that are yet to be identified. Not surprisingly, crosses involving the aneuploid P. paniculata cultivars ‘John Fanick’ and

‘Robert Poore’ with the fertile cultivar ‘Delta Snow’ failed to produce any seed. Seed production has not been obtained in these aneuploidy cultivars even after repeated pollinations (data not shown).

The two species within subsection Paniculatae could be readily hybridized.

Crosses between P. amplifolia and P. paniculata were successful in both directions; each species could serve as a pollen and egg donor although there were differences in the efficiency of seed set (Table 5.3). In 2 of the 3 combinations of P. paniculata and P. amplifolia, higher seed set was obtained with the former was the female parent. In all these crosses, the relative pollination efficiency was less than 1 seed per pollination with

281

a range of 0.12 to 0.74 seeds per pollination; this represents a 70% reduction in

pollination efficiency in the interspecific cross. Morphological features of F1 progeny of

P. amplifolia x P. paniculata were intermediate between parental taxa. Hybrids displayed a glandular-pubescent inflorescence in combination with a pubescent corolla tube, combining features of both taxa (Wherry, 1955). The mean genome size of these hybrids

(2C= 15.09-15.34 pg) was similar to that of both parents that do not differ significantly in genome size (Figure 5.1). However, additional morphological and horticultural traits have yet to be measured for these hybrids. Initial efforts to generate an F2 generation

using sib-crosses resulted in very low seedset (2%), suggesting that F1 interspecific

progeny have reduced fertility or may be pollen sterile.

Hybridization between subsections Paniculatae and Phlox

Crosses of diploid P. paniculata with diploid members of subsection Phlox were

rarely successful (Table 5.3). All species in these subsections share the characteristic of

long styles. Successful cross combinations displayed unilateral incongruity, highly

reduced seed set, and poor germination (Table 5.3). Two accessions of P. paniculata

could be crossed with the diploids Phlox ‘Minnie Pearl’ and a wild accession of P.

carolina PZ11-036 from Mississippi, but few seeds were obtained (Table 5.3). When P.

paniculata was the female in a cross with Phlox ‘Minnie Pearl’, it was possible to obtain

approximately 1 seed per pollination; however, only 2 of the 51 seeds germinated even

though all seeds appeared normal. The F1 individuals obtained displayed morphology intermediate between the parental taxa, such as the lustrous leaves of Phlox ‘Minnie

282

Pearl’ and the areolate leaf veins of P. paniculata (Figure 5.2). Although both parents had white flowers, the flower color of the hybrids was variable and ranged from bicolor pink and white to ‘phlox’ purple. I was not able to obtain any seed in crosses between P. paniculata and P. maculata or P. ovata. The cross of diploid P. paniculata with diploid or tetraploid P. glaberrima also failed. This was the only interploid cross attempted between long-styled taxa, and it is not known whether the crosses failed due to differences in ploidy or to genetic incompatibility. Tetraploid P. paniculata has not been described, and was therefore not available to test intraploid crosses among P. paniculata.

P. amplifolia was successfully hybridized with P. ovata, although seed set was reduced (0.26 seeds per pollination) and only 2 of the 13 seeds germinated (Table 5.3).

Of the two resulting seedlings, one lacked vigor and died, and the other grew vigorously and reached flowering size. Morphological characteristics of this plant are intermediate to parental taxa and possess the petiolate leaves of P. ovata, but the pubescence of P. amplifolia (Figure 5.3). The mean genome size was 14.24 pg, a value that is intermediate between the sizes of parental taxa (Figure 5.1). A cross between P. carolina ssp. carolina

‘Kim’ and P. amplifolia resulted in partially developed seeds (Table 5.3).

Hybridization between subsections Paniculatae and Divaricatae

Twelve attempts were made to recreate P. xarendsii (P. paniculata x P. divaricata; Arends, 1912; Symons-Jeune, 1953) using different parental combinations, but only one combination was successful (Table 5.3). The cross P. divaricata ssp. laphamii PZSH2011-022 x P. paniculata PZ12-106 displayed unilateral incongruity, reduced seed set, and low germination. The short-styled P. divaricata served as the

283

successful female parent, and when P. paniculata served as the female, the crosses failed

and did not even set partially developed seed. All of the 6 F1 progeny obtained were

intermediate in height, number of flower stems per container, leaf length, and leaf width

(Figure 5.4, Table 5.4). The mean genome size of the F1 individuals was 12.31 pg, a value intermediate in size to those of the parental taxa. This genome size is different from those I found in the cultivars of P. xarendsii ‘Hesperis’ and ‘Ping Pong’ (Bendtsen, 2009;

Symons-Jeune, 1953) which were 14.42 pg and 14.31 pg, respectively (Chapter 2). The fertility and crossability of these hybrids has not yet been tested.

Discussion

The goal of this study was to determine the potential for interspecific hybridization among P. amplifolia, P. paniculata, and a diverse germplasm collection of eastern Phlox taxa from subsection Divaricatae and Phlox. The following conclusions may be drawn: 1) P. amplifolia and P. paniculata readily hybridize with each other, but hybridization of these two species with taxa in subsection Phlox is less successful and crosses exhibit unilateral incongruity, reduced seed set and low germination rates. 2) F1

hybrids developed from the few successful crosses involving taxa from subsection

Paniculatae and subsection Phlox were intermediate between parental taxa in gross

morphology. 3) Success in recreating P. xarendsii was limited to one pair of parental

genotypes, which displayed unilateral incongruity and reduced seed set. 4) The

intermediate DNA content of F1 hybrids between parents with divergent genome sizes serves as confirmation of successful hybridization.

284

Successful crosses between P. amplifolia and P. paniculata indicate that there are

few pre-zygotic barriers to interspecific hybridization between these species, but

selection of compatible genotypes appears critical (Table 5.3). The close taxonomic relationship between these species is supported by the relative compatibility between them, but the low rates of seed set and germination of the hybridization products also support their distinctiveness (Levin and Kerster, 1967). In addition, there appear to be post-zygotic barriers that limit the potential for creating advanced generation hybrids

(Eeckhaut et al., 2006; Morgan et al., 2011). Preliminary analysis of crossability between hybrid progeny indicates that fertility may be reduced and that creation of an F2

generation may be difficult. F1 individuals need to be screened for differences in pollen

fertility and unreduced gametes in order to continue the breeding line. However, further

characterization of F1 hybrids through parental backcrosses, three-way crosses, and

experimental interspecific crosses has yet to be done. Continued testing of a variety of

parental combinations between these two species might yield F1 hybrids with differential fertility that could be used to create advanced generation breeding lines. Based on preliminary evidence it is possible that F1 hybrids of P. amplifolia x P. paniculata might

exhibit increased environmental and biotic tolerances that warrant selection of unique

phenotypes from the F1 generation (Deam, 1940; Wherry, 1933). The majority of P. paniculata cultivars are vegetatively propagated. Should evaluation for powdery mildew tolerance in F1 hybrids identify individuals with increased resistance, these could be

vegetatively propagated and commercialized, or used as parents in advanced generation

breeding programs. Vegetative propagation of selected F1 individuals could negate the

285

need to create advanced generation hybrids when fertility is reduced or selected

individuals are sterile.

The rate of seed set for P. amplifolia x P. paniculata crosses was comparable to values reported for intraspecific crosses; this data supports the close evolutionary relationship of these taxa (Table 5.3) (Levin and Kerster, 1967). However, the rate of seed set was not equal for both taxa in a given cross, and P. paniculata was more successful as a female parent. This suggests that genotypic effects may influence the success and direction of success in a given cross. Although germination of hybrid seed was similar to that of the intraspecific cross, the fertility of F1 individuals was variable,

and progeny may require additional screening for fertility.

Successful crosses between taxa in subsections Paniculatae and Phlox were

limited and did not follow a discernable, predictable pattern suggesting the presence of a

variety of pre and post zygotic barriers that influence interspecific hybridization

(Eeckhaut et al., 2006; Morgan et al., 2010). Phlox species have gametophytic self- incompatibility that prevents inbreeding within species, but this system may also impact interspecific hybridization. Molecular phylogenetic analysis of the ITS (internal transcribed spacer) region and chloroplast restrictions sites revealed that taxa in subsection Divaricatae, Paniculatae, and Phlox formed a paraphyletic group that did not resolve relationships at the subsection level (Ferguson et al., 1999; Ferguson and Jansen,

2002). Such results suggests that species are of recent divergence or the product of hybridization, and that taxa may share S-locus haplotypes and alleles that prevent successful interspecific hybridization between some parental combinations (Edh, et al.,

286

2009). However, allelic variation at the same S-locus may not affect some parental

combinations that result in successful interspecific hybridization. Further

characterization of S-loci and corresponding allelic variation in a range of species would

be of benefit to plant breeders and could be used to predict the compatibility of certain

parental combinations. Failed cross combinations may also be impacted by genetic

incongruity (Hogenboom and Mather, 1975; Van Tuyl and De Jeu, 2005). In this case,

pollen tube growth may be arrested due to lack of genetic information necessary for

successful fertilization. Both hypotheses are supported by the complete failure of many

crosses; the lack of partially developed seed in the majority of failed crosses indicates

that pre-zygotic barriers are largely responsible for cross incompatibility.

The hybridization data presented here helps to clarify crossing relationships reported in the literature (Lierval, 1866; Locklear, 2011a; Symons-Jeune, 1953; Wherry,

1955). The hybrid P. xdecussata has been used to describe a variety of hybrids typically involving P. paniculata and subsection Phlox taxa, however, Locklear (2011a), included

P.xdecussata under the name Phlox Suffruticosa Group, and described them as hybrids between P. carolina-glaberrima and P. maculata. The resuts of my experiments suggest that plants under this name might be a series of different hybrids involving P. paniculata, and other members of subsection Phlox (Table 5.1). The data presented here likely underestimate the potential for successful interspecific hybridization among subsections

Paniculatae and Phlox. It may be possible that many of the crosses (reported and unreported) that failed to produce seed in this study would be successful with different genotypes or in different environmental conditions. Karyotype analysis revealed a high

287

degree of chromosomal variation among taxa in subsection Phlox (Smith and Levin,

1967). Chromosome number and morphology are known to affect interspecific

hybridization, and may influence the degree of compatibility in interspecific hybrids

(Smith and Levin, 1967). Unfortunately, these data do not indicate any systematic

method to predict interspecific hybridization potential within tested taxa. In order create

interspecific hybrids breeders will have to explore possible hybrid combinations through

trial and error.

Crossing and genome size data suggest that not all cultivars P. xarendsii may

have arisen from hybridization between P. divaricata and P. paniculata. The

intermediate genome size of P. xarendsii created from the cross P. divaricata ssp.

laphamii x P. paniculata PZ12-106 differed from the values reported for the cultivars of

this species by approximately 2 picograms of DNA (Figure 5.1). A genome size

intermediate to that of parents has been used as evidence of hybridization between other

Phlox taxa, and various other plant species (Morgan et al., 2011; Ozminkowski, Jr. and

Jourdan, 1994: Parris et al., 2010; Chapter 7). The genome size of cultivars listed under

P. xarendsii fall within the range of values reported for P. paniculata suggesting that the cultivars are not, in fact hybrids. This data supports the findings of Flory (1933) in which

P. xarendsii did not exhibit the meiotic irregularities reported for the Phlox interspecific hybrid P. xprocumbens. Furthermore, morphological examination of P. xarendsii plants shows that the two cultivars in this study have areolate leaf veins, cream colored pollen, a pubescent corolla tube, and a long (>18 mm) style; whereas all of these characters were intermediate to parental taxa in artificially created F1 hybrids (Figure 5.4, Table 5.4). All

288 of these characters suggest that tested P. xarendsii are selections of compact forms of P. paniculata, rather than interspecific hybrids (Wherry, 1935a). It is possible that the original plants for which this hybrid was named were chance seedlings that resulted from open pollination of P. paniculata or P. divaricata growing in close proximity to each other and related Phlox taxa but these taxa differ widely in phenology, and the flowering periods do not typically overlap. The early spring flowering P. divaricata usually finishes flowering before the mid-summer to autumn flowering P. paniculata begins

(Locklear, 2011a; Wherry, 1955). However, the duration of flowering may be influenced by climatic differences. Furthermore, these species are sympatric throughout there respective ranges and occur in similar ecosystems, where they are often growing side by side in the same habitat. At least two naturally occurring Phlox hybrids, P. xglutinosa (P. divaricata x P. pilosa) and P. xrugelii (P. divaricata x P. amoena) have been described among sympatric species with similar ecological tolerances, but no such hybrids have ever been found among P. divaricata and P. paniculata. Despite all of this, Wherry

(1955) reported that P. xarendsii was seed sterile and supported the hybrid origin of this taxon, but did not report the cultivar(s) upon which those observations were made. Since only two cultivars were tested here, analysis of more P. xarendsii is needed to clarify the origin of this taxon.

One of the most interesting results of the crossability study was the production of a successful hybrid between P. amplifolia and P. ovata (Figure 5.3, Table 5.3).

Karyotypic analysis of P. amplifolia revealed that this species had the most symmetrical chromosomes of all tested taxa, and differed from P. paniculata only in one chromosome

289

(Smith and Levin, 1967). This chromosome symmetry was reported as a primitive state in the genus Phlox, but this data is not supported by the molecular phylogenies, and it is not known how this could affect hybridization (Ferguson et al., 1999; Ferguson and

Jansen, 2002; Smith and Levin, 1967). The karyotype of P. ovata differed from other members of subsection Phlox, and was more similar to P. stolonifera than other taxa in the subsection; it is possible that these karotypic differences influced the outcome of the cross. These data also suggested that taxonomic descriptions of species based on morphology do not reflect evolutionary relationships, and that molecular phylogenies also have not accurately resolved these relationships (Ferguson et al., 1999; Ferguson and

Jansen, 2002; Smith and Levin, 1967; Wherry, 1955). Hence, information derived from taxonomic treatments and molecular phylogenies thus far may not help predict cross compatibility among eastern Phlox taxa. It is possible that the differences in these karyotypes influenced interspecific hybridization. From a plant development standpoint, the F1 hybrid of P. amplifolia x P. ovata is s desirable cross combination because it

combines the spreading-mounding habit of P. ovata with the larger inflorescence and

extended flowering time of P. amplifolia.

Availability of genome size information about the parents derived from flow

cytometric analysis makes it possible to confirm hybrids of crosses between taxa with

different genome sizes (Figure 5.1). Flow cytometry provides a rapid and efficient

method for screening putative hybrids, but does not immediately confirm hybridization

among species with similar genome sizes. However, it does indicate that crosses between

diploid taxa result in homoploid hybrids rather than allopolyploids. Several naturally

290 occurring allopolyploid Phlox hybrids been suggested, but data presented here do not support this claim (Levin, 1966; Levin, 1968).

In conclusion, crossability among species from subsection Paniculatae was high, but intersubsectional crosses of these taxa and members of subsections Divaricatae and

Phlox was low. However, there is potential for creating a diverse array of interspecific

Phlox hybrids, but the degree of success may depend on the genotypes of the parents used; this highlights the importance of using crop wild relatives in experimental hybridization.

291

18 16

16 14

14 12 12 10 10

8 8 6 6

4 4

2 2

0 0 DNA content (pg) a. P. amplifolia ♀ F1 P. paniculata ♂ b. P. ‘Minnie Pearl’ ♂ F1 P. paniculata ♀ ‘Delta Snow’

16 18

16 14

14 12

12 10

10 8

8 6 6 4 4 DNA content (pg) 2 2

0 0 d. P. divaricata ♀ F1 P. paniculata ♂ c. P. ovata ♂ F1 P. amplifolia ♀ ssp. laphamii

Figure 5.1: Histograms showing relative genome of diploid (n=7) parental taxa and diploid (n=7) F1 hybrids from 4 interspecific crosses; a. P. divaricata ssp. laphamii PZSH11-022 (11.03 pg), P. divaricata ssp. laphamii x P. paniculata F1 hybrid PZ11-199 (12.31 pg), and P. paniculata PZ12-106 (15.01 pg); b. P. ovata PZ10-167 (12.16 pg), P. amplifolia x P. ovata F1 hybrid PZ11-176 (14.24 pg), and P. amplifolia PZ11-050 (15.21 pg); c. P. amplifolia PZ11-050 (15.21 pg), P. amplifolia x P. paniculata ‘Delta Snow’ F1 hybrid PZ11-176 (15.34 pg), and P. paniculata ‘Delta Snow’ PZ10-027 (14.90 pg); d. P. paniculata PZ10-231 (14.14 pg), P. ‘Minnie Pearl’ x P. paniculata F1 hybrid PZ11-061 (14.62 pg), and P. ‘Minnie Pearl’ PZ11-012 (14.61 pg).

292

a. b. c.

Figure 5.2: Comparison of parental taxa and their F1 hybrid. a. P. paniculata PZ10- 231(♀). b. P. paniculata x P. ‘Minnie Pearl’ F1 hybrid, and. c. P.’Minnie Pearl’ PZ11- 012 (♂). Note the intermediacy of the leaf veins and petiole. The scale bar is 1 cm.

293

a. b. c.

Figure 5.3: Comparison parental taxa and their F1 hybrid. a. P. amplifolia PZ11-050 (♀). b. P. amplifolia x P. ovata F1 hybrid, and. c. P. ovata PZ10-167 (♂). Note the intermediacy of the leaf veins and petiole. The scale bar is 1 cm.

294

a. b. c. d.

Figure 5.4: Comparison parental taxa and their F1 hybrid in the remake of P. xarendsii. a. P. divaricata ssp. laphamii PZSH2011-022 (♂). b,c. P. divaricata x P. paniculata F1 hybrid, and. d. P. paniculata PZ12-106 (♀). Note the intermediacy of the leaf veins and petiole. The scale bar is 1 cm.

295

Figure 5.5. Summary of the interspecific compatibilities between the species in subsection Paniculatae and subsections Phlox and Divaricatae. Areas blacked out indicate crosses that were not attempted. ✓ = successful combination; ✕= failed combination; PD = partial development of fruit.

296

Parental Style Putative Hybrid Taxa Reported Parentage Reported ploidy Lengthz

Phlox paniculata x Phlox maculata or Phlox Phlox Suffruticosa Group (syn. P. xdecussata ) carolina 2x = 14 L x L

Phlox xarendsii ( Arends)Wherryy Phlox divaricata x Phlox paniculata 2x = 14 L x S

Phlox xleopoldiana (syn. Phlox 'Depressa', Phlox 'Criterion')w Phlox drummondi x Phlox paniculata 2x = 14 S x L zParental style length: L indicates a style longer than 12 mm, an S indicates a paret species with a style shorter than 6 mm. yArends (1912); Wherry (1955). xLierval (1866); Symons-Jeune (1953). wKelly (1915); Wherry (1955).

Table 5.1: Putative hybrids involving Phlox paniculata with the corresponding reported parentage, ploidy, and style length of the parents.

297

Subsectionz Taxon Accession no. Collection Sourcey Genome sizex CV (%) Ploidy Style lengthw

Divaricatae Phlox divaricata L. ssp. divaricata PZ10-145 Hocking Co. OH 9.18±0.22 2.26 2x S Phlox divaricata L. ssp. laphamii (Wood) Wherry PZSH2011-004 Gadsden Co. FL 10.37±0.13 2.03 2x S

P. divaricata ssp. laphamii v PZSH2011-022 Jasper Co. MS 11.03±0.03 1.53 2x S P. divaricata ssp. laphamii PZSH2011-026 Wilkinson Co. MS 9.78±0.13 1.73 2x S P. drummondii L. ssp. drummondii PZ10-161 Caldwell Co. TX 12.27±0.04 2.36 2x S P. drummondii L. ssp. mccallisteri PZ10-163 Wilson Co. TX 12.21±0.08 2.11 2x S P. drummondii PZ10-202 A 12.38±0.14 2.27 2x S P. drummondii Palona Deep Violet with Eye PI601603 B 12.24±0.05 2.20 2x S P. floridana Bentham PZSH2011-010 Jackson Co. FL 21.85±0.30 3.55 4x S P. pattersonii Prather PZ10-247 Nuevo Leon, Mexico 14.87±0.09 2.26 2x S

P. pilosa L. ssp. pilosa PZ10-192 Lucas Co. OH 12.34±0.17 3.90 2x S

P. pilosa L. ssp. pilosa PZSH2011-017 Lowndes Co. AL 12.66±0.11 1.57 2x S P. pilosa 'Racy Pink' PZ10-024 C 11.42±0.26 2.10 2x S

P. pilosa 'Lavender Cloud' PZ10-025 C 11.16±0.17 2.74 2x S

Phlox P. carolina L. ssp. carolina PZ11-036 Clay Co. MS 15.25±0.39 4.31 2x L P. carolina L. ssp. carolina 'Kim' PZ11-017 D 14.38±0.012 2.00 2x L P. carolina L. ssp. angusta Wherry PZ10-034 C 15.54±0.06 1.66 2x L

P. glaberrima sp. glaberrima PZSH2011-012 Gadsden Co. FL 16.13±0.27 1.79 2x L

P. glaberrima L. ssp. interior Wherry PZ10-248 E 13.97±0.43 2.38 2x L P. glaberrima ssp. triflora 'Anita Kistler' PZ10-175 Bath Co. VA; C 26.70±0.30 0.70 4x L P. glaberrima ssp. triflora 'Bill Baker' PZ11-030 D 26.37±0.03 0.79 4x L P. maculata PZ10-198 Champaign Co. OH 14.41±0.08 1.64 2x L P. maculata PZ10-208 Adams Co. OH 14.47±0.17 1.93 2x L P. maculata 'Natascha' PZ10-035 C 14.77 1.74 2x L P. maculata 'Omega' PZ10-051 F −−− −−− 2x L

P. 'Minnie Pearl' PZ11-012 Kemper Co. MS; A 14.61±0.14 3.08 4x L P. ovata (L.) Locklear PZ10-167 Lucas Co. OH 12.16±0.22 3.51 2x L

(Continued)

Table 5.2: Phlox taxa used in the interspecific hybridization experiments organized by subsection, specific germplasm accessions, collection sites, relative genome sizes, ploidy levels, and style lengths.

298

Table 5.2: Continued

Paniculatae P. amplifolia Britton PZ11-050 Cocke Co. TN 15.21±0.24 3.72 2x L

P. paniculata L. PZ10-109 G 14.74±0.29 1.62 2x L P. paniculata PZ10-110 G 14.77±0.41 1.81 2x L P. paniculata pastel pink flower PZ10-231 Erie Co. OH 14.61±0.17 2.65 2x L P. paniculata white flower PZ10-231 Erie Co. OH 14.61±0.17 2.65 2x L P. paniculata PZ11-043 Preston Co. WV 14.53±0.05 1.19 2x L P. paniculata PZ12-106 Campbell Co. TN 15.01±0.35 1.61 2x L P. paniculata 'Delta Snow' PZ11-027 C 14.90±0.30 1.64 2x L P. paniculata 'John Fanick' PZ12-012 H 22.57±0.13 1.48 4x -n L

P. paniculata 'Robert Poore' PZ10-028 C 21.74±0.13 1.38 4x -n L zTaxonomic classifications are based on Locklear (2011), Wherry (1955) and Ferguson et al. (1999). yNote: Collection source: A, Seeds obtained from Wildseed Farms, Fredericsburg, TX; B, Seeds obtained from Ornamental Plant Germplasm Center, Columbus, OH; C plant obtained from Growild Nursery, Fairview, TN; D, Plant obtained from Primrose Path Nursery, Scottdale, PA; E, Plant obtained Prairie Moon Nursery, Winona, MN; F, Plant obtained from Arrowhead Alpines Nursery, Fowlerville, MI; G, Plant obtained from Nearly Native Nursery Fayetteville, GA; H, Plant obtained from Plant Delights Nursery, Raleigh, NC. xGenome size is the 2C holoploid genome size ± the standard deviation. wNote: "L" denotes a style longer than 15 mm, and "S" denotes a style less then 15 mm in length. vTaxa highlighted in bold were successfully hybridized with either P. amplifolia or P. paniculata (Table 5.3).

299

Parental Combinationz

Seeds obtained Seeds germinated Subsection Female Parent (♀) Male Parent (♂) Germination (%)x Accession No. (No.)y (No.)

P. paniculata PZ10-110 P. paniculata PZ10-109 138 83 60.5 PZ11-096 P. paniculata PZ10-109 P. paniculata PZ10-110 112 56 50.5 PZ11-097

Paniculatae P. amplifolia PZ11-050 P. paniculata PZ11-043 6 −−− −−− PZ11-179 P. paniculata PZ11-043 P. amplifolia PZ11-050 31 27 87.1 PZ11-181

P. amplifolia PZ11-050 P. paniculata 'Delta Snow' 10-027 35 24 68.6 PZ11-180 P. paniculata 'Delta Snow' 10-027 P. amplifolia PZ11-050 37 7 18.9 PZ11-182

P. amplifolia PZ11-050 P. paniculata PZ10-109 8 3 37.5 PZ11-183 P. paniculata PZ10-109 P. amplifolia PZ11-050 29 19 65.5 PZ11-185

Phlox P. paniculata PZ10-231 P. carolina ssp. carolina PZ11-036 35 2 6.0 PZ11-091 P. carolina ssp. carolina PZ11-036 P. paniculata PZ10-231 0 −−− −−− −−−

P. paniculata PZ10-231 P. 'Minnie Pearl' 51 6 8.5 PZ11-060 P . 'Minnie Pearl' P. paniculata PZ10-231 0 −−− −−− −−−

P. amplifolia PZ11-050 P. carolina 'Kim' PZ11-017 −−− −−− −−− −−− P. carolina 'Kim' PZ11-017 P. amplifolia PZ11-050 3/PD −−− −−− PZ11-167

P. amplifolia PZ11-050 P. ovata PZ10-167 13 2 13.1 PZ11-176 P. ovata PZ10-167 P. amplifolia PZ11-050 0 −−− −−− −−−

Divaricatae P. divaricata ssp. laphamii PZSH2011-022 P. paniculata PZ12-106 11 −−− 55.5 PZ11-199 P. paniculata PZ12-106 P. divaricata ssp. laphamii PZSH2011-022 0 −−− −−− −−− z50 pollination per cross were performed over a period of 3-7 days, not all of any given cross was performed on a single day. ySeed was collected by bagging the inflorescence 3-4 weeks after completion of crosses and allowing the seeds to fall when mature. xGermination (%) calculated as the number of seeds that germinated and produced a seedling by the total number of seeds.

Table 5.3: Parental combinations, seeds obtained (No.), and germination (%) of successful and unsuccessful, reciprocal interspecific crosses in which either Phlox amplifolia or P. paniculata served as a parent.

300

Flowering stems Taxon Height (cm) Style length (mm) Leaf length (cm) Leaf width (cm) Pollen color per container (No.)

Phlox divaricata ssp.laphamii PZSH2011-022 x Phlox paniculata PZ12-106

P. divaricata PZSH2011-022 44.0cz 6.60a 3.3c 6.58b 2.50b Yellow P. paniculata PZ12-106 87.6a 2.0b 21.1a 10.54a 3.80a Cream

F1 67.0b 4.40a 12.7b 7.38b 2.66b Cream/Yellow zNumbers of followed by the same letter are not siginificantly different as determined using analysis of variance with means separation by Fisher's least significant difference at P = 0.05.

Table 5.4: Mean plant height (cm), flowering stems per pot (No.), leaf length (cm), and leaf width (cm) for parental species and interspecific F1 hybrids of the cross Phlox divaricata ssp. laphamii PZSH2011-022 x P. paniculata PZ12-106.

301

CHAPTER 6

INTERSPECIFIC HYBRIDIZATION IN PHLOX FOR GERMPLASM ENHANCEMENT: HYBRIDIZATION POTENTIAL OF PHLOX ‘MINNIE PEARL’

Abstract

There is increased interest in developing Phlox hybrids to meet the needs of the floriculture industry; despite a long history of hybridization, most information pertaining to Phlox hybridization is anecdotal or has never been verified. Phlox ‘Minnie Pearl’ is a

spring and summer flowering herbaceous perennial selected for its vigorous growth,

compact habit, recurrent white flowers, resistance to powdery mildew. These unique

characteristics make it a desirable parent in an interspecific breeding program, but

fundamental information regarding the identity, hybridization potential, and inheritance

are lacking. The objective of this study was to determine the interspecific crossability of

‘Minnie Pearl’ with related Phlox species, and to confirm hybridity using morphology

and SRAP (Sequence Related Polymorphism) markers, and determine inheritance of crop

quality traits in the F1 families. Among parents selected for hybridization, most were diploid (n=7) and genome size ranged from 9.78 – 16.13 pg, however the mean diploid

genome size of subsections Paniculatae (14.40 pg) and Phlox (14.89 pg) were similar,

but differed from subsection Divaricatae (11.51 pg). Two tetraploid taxa were also used.

The highest rates of successful interspecific hybridization were between ‘Minnie Pearl’

302

and related species from subsections Paniculatae and Phlox. Crosses between ‘Minnie

Pearl’ and taxa from subsection Cluteanae and Divaricatae were unsuccessful. All

successful crosses were between taxa with long styles, similar genome sizes, and the

same ploidy level, and displayed unilateral incongruity with reduced seed set compared

to intraspecific crosses. Seeds of successful crosses all exhibited varying degrees of

germination, but all successfully germinated seeds resulted in vigorous F1 progeny. F1

progeny were morphologically intermediate to parental taxa for height, number of

flowering stems per container, and leaf morphology, but 2 hybrid populations had mean

height measurements that were greater than either parent. All F1 populations derived from crosses between ‘Minnie Pearl’ and pink flowered taxa had pink displayed complex flower color inheritance. This study indicates that there is potential for using ‘Minnie

Pearl’ to create interspecific hybrids, but the choice of parents is critical for success.

Introduction

Phlox ‘Minnie Pearl’ was introduced by Plant Delights Nursery (Raleigh, NC) in

2003. It was selected from a natural roadside population in Kemper County, Mississippi

(Plant Delights Nursery, 2014). Since that time it has rapidly become known as one of the most dependable Phlox for general garden and landscape use in a variety of climates.

It is a unique phenotype with recurrent, white flowers, a compact, rhizomatous habit, and lustrous, disease-free foliage. These attributes make ‘Minnie Pearl’ a potentially useful parent in the creation of Phlox hybrids, but there are several issues that may impede use in breeding programs. The specific identity of this plant has been debated; it has been

303

reported as a hybrid between P. glaberrima, and P. maculata, however it also bears

similarities to P. carolina and P. paniculata. Furthermore, the putative hybrid origin suggests that it may be sterile or have reduced fertility. In absence of precise knowledge of phylogenetic relationships of ‘Minnie Pearl’, experimental hybridization can be used to define species relationships and reveal barriers to hybridization.

Phlox is a genus of ca. 65 species distributed throughout North America with two centers of species diversity in the eastern and western United States. There are 22 eastern species classified into 2 sections and 6 subsections primarily on the basis of style length and geographic distribution (Ferguson, et al., 1999; Wherry, 1955). Of these, three subsections in section Phlox, subsections Paniculatae, Phlox (Ovatae sensu Wherry), and

Stoloniferae are referred to as long-styled phloxes, and have a style greater than 15 mm in length. Taxa in three other eastern subsections have styles less than 12 mm in length, and usually much shorter (1 mm). The putative parental taxa of ‘Minnie Pearl’, P. glaberrima and P. maculata, are members of the taxonomically difficult subsection

Phlox. Members of this subsection are distinguished from other subsections by having obscure leaf veins, a glabrous inflorescence and calyx, and distinct calyx morphology

(Table 6.1) (Wherry, 1955). However, extensive ecotypic and phenotypic diversification in this group has resulted in the description of numerous subspecies, particularly within the geographically widespread taxa P. carolina and P. glaberrima (Wherry, 1932a;

Wherry, 1932b; Wherry, 1955). Variation in these taxa has caused taxonomic confusion,

the classification has changed significantly in the last 160 years, and phylogenetic studies

have not resolved relationships in this group (Bentham; 1845; Ferguson et al., 1999;

304

Ferguson and Jansen, 2002; Gray, 1870; Wherry, 1955). Currently, this group is known

as the P. carolina-P. glaberrima complex (Ferguson and Jansen, 2002;Wherry, 1955).

Subsection Paniculatae is most closely related to subsection Phlox, but differs in having

conspicuous areolate leaf veins, nearly white pollen color, inflorescence and calyx vesture, and calyx morphology (Table 6.1). Both subsection Paniculatae and Phlox are similar in having comparatively upright or erect plant habits, large leaves, mid-spring to early autumn flowering, and many nodes per stem; these differ substantially from the carpeting growth and early spring flowering members of subsection Stoloniferae.

Morphological comparisons suggest that ‘Minnie pearl’ may be a member of the P. carolina-P. glaberrima complex, and that it may most readily hybridize with other taxa in that subsection, and other long-styled taxa.

The difficulty of creating interspecific hybrids increases along with the phylogenetic distance between the parents and chromosome number (Eeckhaut et al.,

2006). In the absence of a well-resolved phylogeny, experimental interspecific hybridization among morphologically similar taxa from a diverse germplasm collection can help characterize gene pools for breeding efforts (Harlan and de Wet, 1971).

Comparatively few cultivars have been selected among taxa in subsection Phlox and these have not been extensively characterized and are of uncertain origin. Crop wild relatives can be novel sources of genetic variation and provide necessary germplasm to accurately define gene pools when there is inadequate representation among commercial cultivars (Tay, 2003). Although there have been extensive breeding efforts among long- styled phloxes, reports of successful hybridization are anecdotal in nature, and were

305

based on a complicated, outdated taxonomy that makes modern interpretation difficult

(Locklear, 2011a; Pridham, 1934; Symons-Jeune, 1953). Despite this, at least one group

of hybrids, the Phlox Suffruticosa Group (syn. P. xdecussata), are reportedly hybrids derived from taxa in the P. carolina-P. glaberrima complex and P. maculata, but the history of such hybrids remains unclear and unverified (Locklear, 2011a; Symons-Jeune,

1953).

Differences in genome size and ploidy can be major barriers to interspecific hybridization, however, relatively few Phlox taxa have been characterized for genome size and ploidy variation; this is especially true for those in subsection Phlox (Flory,

1933; Flory, 1934; Smith and Levin, 1967). Previous studies of chromosome numbers and karyology of subsection Phlox taxa revealed that all tested accessions were diploid

(n=7), but that Phlox Suffruticosa Group ‘Miss Lingard’, a supposed hybrid of P. carolina and P. maculata, was a sterile triploid that may have originated from an interploid cross (2n=21) (Flory, 1933; Flory, 1934). Tetraploids have been found among one taxon (P. glaberrima ssp. triflora) in the P. carolina-P. glaberrima complex, however the majority of tested taxa were diploid (Chapter 2).

Recent flow cytometric analysis of P.pilosa ssp. pilosa and the P. amabilis-P. woodhousei complex revealed that ploidy and genome size were more variable than previously thought and that polyploid populations can exist in close proximity (Fehlberg and Ferguson, 2011; Worcester et al., 2012). Previous investigations of genome size of

Phlox cultivars resulted in the discovery of aneuploid and polyploid taxa in species previously known only as diploids (Chapters 2, 3, 4). Phlox ‘Minnie Pearl was reported

306

as a diploid with a mean genome size of 14.61 pg (Table 6.2). Genome size was similar

to species in subsection Paniculatae, and to members of the P. carolina-P. glaberrima

complex and P. maculata (Table 6.2). The data, combined with morphological and

phylogenetic data, suggests that these species represent a logical starting place for initial

interspecific hybridization experiments.

SRAP (Sequence Related Polymorphism) molecular markers are convenient for

screening genetic diversity in plant species for which there is no prior molecular

information. These markers have been applied to a variety of crop and ornamental

species. The application of this marker system in Phlox has not been previously

demonstrated, but application of this technology may help characterize Phlox germplasm

collections, and has the potential to discriminate between parents and hybrid taxa.

To evaluate the hybridization potential and begin to define gene pools of ‘Minnie

Pearl,’ using crop wild relatives to provide preliminary information to plant breeders, our

objectives were to: 1) test the interspecific compatibility of ‘Minnie Pearl’ with a select

number of Phlox species from subsections Divaricatae, Paniculatae, and Phlox; 2)

determine rates of seed set and germination among successful crosses, and; 3) use

morphological markers, crop quality traits, and SRAP to confirm hybridity in F1 hybrids.

Materials and Methods

Parental taxa and controlled pollinations

Twelve Phlox species were obtained from nursery sources and collected from natural plant populations. Plants of ‘Minnie Pearl’ and P. maculata ‘Flower Power’ were

307

obtained from Nursery sources. Compared to other Phlox species, relatively few cultivar

selections have been made from species in subsection Phlox (Bendtsen, 2009; Fuchs,

1994). In order to obtain germplasm for experimentation, a series of botanical collection

trips to numerous areas of the United States resulted in collection of Phlox germplasm

over a 3-yr period. Since morphological comparisons suggest ‘Minnie Pearl’ is in the P. carolina-P.glaberrima complex, emphasis was placed on collecting members of that group and related taxa from subsections Divaricatae, Paniuclatae, and Phlox (Tables 6.1,

6.2) (Wherry, 1955). Plant identities were confirmed using the dichotomous keys provided by Wherry (1955), with modifications from various sources (Ferguson, 1999;

Prather, 1994; Locklear, 2011a). Herbarium specimens were made for each collection.

Crosses were performed during 2011 and 2012 in a reciprocal crossing scheme in

which ‘Minnie Pearl’ was the recurrent parent; with a minimum of 50 pollinations

(individual flowers) per cross. Nearly all Phlox species can develop a maximum of three

seeds per flower. Crosses were performed over a 1-2 week period. Successful

pollination was defined as the formation of at least 5 viable seeds per cross. All Phlox

species have ballistic seed dispersal and developing infructescences were bagged with

nylon pantyhose to aid seed collection (Wherry, 1955). Seeds of subsection Paniculatae

and Phlox taxa used in this study mature 60-90 days after pollination (Chapter 1). Seeds were harvested from parent plants when they had dehisced and could be seen inside the nylon bag that surrounded the infructescence.

308

Growing conditions and seed germination

Growing conditions, methods used for controlled pollinations, and seed

germination are described in Chapter 5.

Flow Cytometric Analysis

The details of flow cytometry analysis are discussed in Chapter 2.

SRAP markers

Genomic DNA was extracted from fresh young leaves using a modified CTAB

mini-prep (Robarts, 2013). The SRAP primer pair em3-me3 was selected from a series of 6 forward and 7 reverse primers because of the presence of scorable banding patterns.

DNA amplification was carried out in 10 µL reaction volume using an ExTaq (Clontech

Laboratories, Inc., Mountain View, CA) containing: 1 µL 10X PCR buffer, 1 µL genomic

DNA, 0.8 µL dNTP mixture, 0.25 µL of each primer, 0.05 µL Taq DNA polymerase, and

6.65 µL PCR pure water. The amplification was carried out using a Eppendorf

Mastercycler ep gradient S thermal cycler (Eppendorf North America, New York, NY) using the following thermal cycling profile used for all; 32 cycles of 1 min at 94°C, 1 min at 47°C, 1 min at 72°C, and ending with 5 min at 72°C (Budak et al., 2004). The presence of PCR amplicons was confirmed by running 2 µL of reaction product on a

1.0% agarose gel stained with ethidium bromide and quantified by UV spectrophotometry. PCR products were analyzed on an ABI 3100 genetic analyzer (Life

Technologies, Grand Island, New York).

309

Data Analysis

Analysis of variance, means separation and chi-square tests were performed in

SAS (Proc GLM, SAS Version 9.3; Cary, NC). Means separation were determined by

Fisher’s least significant difference at P = 0.05.

Results

Parents for hybridization

Twenty-four different taxa were used in hybridization experiments (Table 6.2).

Parental taxa were chosen that exhibit a broad diversity of phenotypes that could serve as markers of successful hybridization in F1 progeny. The majority of taxa used in this

study were diploid (n=7) and overall genome size ranged 9.78 – 16.13 pg (Table 6.2).

The mean genome size of diploid taxa in subsection Divaricatae was 11.51 pg, but was

14.89 pg for subsection Paniculatae, and subsection Phlox was 14.40 pg. Nine accessions representing 6 taxa from subsection Phlox were used in interspecific

hybridization with ‘Minnie Pearl’; all were diploid. Eight accessions representing 2 taxa from subsection Paniculatae were used; six of these were diploid and 2 were anueploids

(Chapter 2). The similarity in mean genome sizes between the long-styled subsections

Paniculatae and Phlox supports the close phylogenetic relationship of these taxa that was

formed using morphology (Smith and Levin, 1967; Wherry, 1955). The only member of

subsection Cluteanae, Phlox buckleyi, was tetraploid (n=14) with a mean genome size of

23.04 pg; this long-styled taxon is not closely related to other eastern taxa (Table 6.2).

Five accessions of 5 taxa from subsection Divaricatae were used; all these taxa are short-

310

styled, and 4 of them were diploid. The mean genome size of the tetraploid (n=14) P.

pulcherrima was similar, 22.05 pg.

Overview of Hybridization

A total of 2300 pollinations were made among 24 accessions comprising 12

species from 4 different subsections of eastern Phlox taxa (Table 6.2). Phlox ‘Minnie

Pearl’ was most readily hybridized with members of subsection Phlox; however not all cross combinations were successful, and all successful crosses displayed unilateral incongruity (Table 6.3). While this data helps confirm the taxonomic placement of

‘Minnie Pearl’ in subsection Phlox, it also suggests that crossing relationships are complicated. Overall seed set was lower in comparison to one intraspecific cross between individuals of P. carolina ssp. carolina PZ11-036; but the reciprocal cross resulted in seed set similar to ‘Minnie Pearl’ hybrids (Table 6.3). Despite low seed set among these crosses, they exhibited the highest germination percentage of all successful crosses (Table 6.3). The best germination occurred in crosses of ‘Minnie Pearl’ and two accessions of P. carolina PZ11-036 and P. carolina ‘Kim’. Other crosses within subsection Phlox yielded similar success rates, but decreased germination (Table 6.3).

Crosses of ‘Minnie Pearl’ with P. ovata failed to set seed.

Intersubsectional crosses were rarely successful (Table 6.3). The greatest seed set

of all crosses resulted when ‘Minnie Pearl’ was used as the male parent and crossed with

two accessions of P. paniculata. These crosses also displayed unilateral incongruity.

However, seed germination of these crosses was low, and seeds of one cross completely

311

failed to germinate (Table 6.3). This suggests that seeds may have been partially

developed due to endosperm deficiency or other factors. F1 progeny were intermediate in

morphology between the parents and displayed the areolate leaf veins of P. paniculata

and lustrous leaf vesture of ‘Minnie Pearl’ (Table 6.2). All crosses between ‘Minnie

Pearl’ and diploid taxa in subsections Cluteanae and Divaricatae failed to set seeds, but

two crosses resulted in partial development.

Crosses between ‘Minnie Pearl’ and P. buckleyi and P. pulcherrima produced small numbers of partially developed seeds. Partially developed seeds were obtained from reciprocal crosses with the long-styled P. buckleyi, but were only obtained when

‘Minnie Pearl’ was the female parent in a cross with the short-styled P. pulcherrima.

There were significant differences among families of F1 progeny of successful crosses between ‘Minnie Pearl and other subsection Phlox taxa. The average height for

F1 progeny of ‘Minnie Pearl’ x P. carolina ‘Kim’ and P. carolina ssp. alta PZ11-045 x

‘Minnie Pearl’ was greater than either parent, and both families had an increased number

of flowering stems, although the latter were not significantly different from parental taxa

(Table 6.4). Most were intermediate between parental taxon for plant height, number of

flowering stems per container, and leaf dimensions (Table 6.4).

Both white and pink flowered individuals were found in F1 families of all crosses

between the white flowered ‘Minnie Pearl’ and subsection Phlox accessions. Segregation

between white and pink flower color in F1 hybrids of ‘Minnie Pearl’ and subsection

Phlox taxa was complex. The cross P. maculata ‘Flower Power’ x ‘Minnie Pearl’, an

interspecific cross between two white flowered plants, resulted in 7 pink-flowered

312 progeny and 1 white-flowered plant. Although there were not enough progeny to adequately determine the inheritance ratio, the initial results suggest that 2 loci may be responsible for controlling white flower color; if one locus were responsible then flower color segregation ratios would have been different (Tobutt, 1993). Crosses between the white-flowered ‘Minnie Pearl’ and pink-flowered taxa from subsection Phlox, exhibited a general 1:1 ratio of pink to white flower color inheritance (Figure 6.1). Although small

F1 families were generated, this mode of inheritance was consistent among crosses of

‘Minnie Pearl’ x pink flowered subsection Phlox taxa. Preliminary results indicate that the white flower color of ‘Minnie Pearl’ can be used a marker of successful hybridization in certain cross combinations. However, inheritance ratios need to be verified with larger numbers of F1 progeny, F2 generations, and parental backcrosses to confirm initial results.

F1 progeny of the cross P. paniculata x ‘Minnie Pearl’ displayed intermediate morphology suggestive of successful hybridization, and was confirmed with SRAP markers (Figure 6.2). This initial screening using the primer pair me3-em3 showed the presence of shared bands among P. paniculata PZ10-231, P. longipilosa, and their F1 hybrid. SRAP banding patterns have been used to confirm hybridity in several horticultural genera: including Brassica, Coffea, and Lactuca (Li and Quiros, 2001; Liu et al., 2012; Mishra et al., 2011). Electropherograms confirmed the presence of shared alleles; peak height of shared alleles was similar among parents and hybrid. Although intermediate morphology can be used to readily identify Phlox hybrids, these results suggest that the SRAP markers could be used to confirm hybridity between closely

313

related taxa or cultivars and may be of value for characterizing genetic diversity in a

Phlox germplasm collection (Budak et al., 2004).

Discussion

The goal of this study was to determine the potential for the development of

interspecific hybrids for crop improvement using ‘Minnie Pearl’ as a recurrent parent in

experimental crosses with taxa in subsections Cluteanae, Divaricatae, Paniculatae, and

Phlox obtained from nursery sources and natural plant populations. The following

conclusions may be drawn from study: 1) Phlox ‘Minnie Pearl’ hybridizes most readily with taxa in subsections Paniculatae and Phlox, however, seed germination is reduced in intersubsectional crosses. 2) All successful crosses displayed unilateral incongruity. 3) F1

progeny were intermediate between parental taxa in gross morphology. 4) SRAP markers

can be adapted to Phlox taxa. These results suggest that a variety of factors determine the success of a given cross, and that incompatibility and incongruity contribute to several pre and post zygotic phases of hybrid development (Eeckhaut et al., 2006; Hogenboom and Mather, 1975).

Self-incompatibility (SI) systems meant to prevent inbreeding may also serve as a prezygotic barrier to interspecifc hybridization. Phlox species are highly heterozygous, obligate out-crossing species that have well-developed gametophytic, self-incompatibility

(SI) system (Levin, 1975, 1993). The SI system is known to affect interspecific

hybridization through inhibition of pollen tube germination when the same SI locus and

S-gene allele is found in the pistil and pollen of parental taxa. In one natural population

314

of P. drummondii, the SI system was regulated by a single gene at the S-locus, but as

many as 30 different S-gene alleles were found, and suggests that extensive genotypic

diversification at the S-locus is found in P. drummondii, and other Phlox species (Levin,

1993). A study of variation in S-locus haplotypes in Brassica and Arabidopsis indicated

a single evolutionary origin of SI in both genera, and suggests that similar cases may

exist in other genera (Edh, et al., 2009). While similar studies have not been attempted in

Phlox, phylogenetic studies of subsection Phlox suggested that most taxa are of recent

evolutionary divergence or have experienced recent natural hybridization events

(Ferguson et al., 1999; Ferguson and Jansen, 2002). This may have resulted in taxa that

share S-locus haplotypes and S-locus alleles that prohibit successful interspecific

hybridization, but divergent populations may have sufficient allelic variation to cause

inactivation the SI system. Several different genotypes from subsection Phlox and P.

paniculata were crossed with ‘Minnie Pearl’ with divergent outcomes in regard to seed set and germination, and all displayed unilateral incongruity. Therefore, the SI system

may serve as a pre-zygotic barrier to interspecific and that shared allelic variation at S-

gene locus among specific Phlox clones may prevent complete pollen tube growth.

Prevention of hybridization by the SI system is further supported by the fact that majority

of unsuccessful crosses completely failed to develop seed, and only two crosses resulted

in partially developed seed, suggesting a pre-zygotic barrier.

Difference in style length among species has been considered to be a pre-zygotic

barrier to interspecific hybridization in several ornamental plant genera, and has been

reported as a barrier to interspecific hybridization in Phlox (Eeckhaut et al., 2006; Levin,

315

1966). Incongruity between taxa with different style lengths may result in breakdown of

the genetic mechanisms required to complete successful fertilization. However,

numerous Phlox hybrids involving species with different style lengths have been

reported; P. xarendsii (P. divaricata x P. paniculata), P. xleopoldana, P. nivalis x P.

subulata, and P. nivalis x P. bifida ssp. bifida (Bendtsen, 2009; Locklear, 2011a;

Symons-Jeune, 1953; Zale, unpublished data). Genotypic effects, specific to an

individual or clone, may be critical to the success of a given cross between taxa with

different style lengths (Chapter 5). To support this claim, the cross ‘Minnie Pearl’ x P.

pulcherrima produced partially developed seeds, despite having different length styles

and ploidy levels, but all crosses involving diploid, short-styled species and ‘Minnie

Pearl’ were unsuccessful (Table 3). The use of cut-styles has been shown to be a mechanism for overcoming differences in style length in a variety of genera, and may have application in Phlox (Eeckhaut et al., 2006). However, the comparatively thin style

of all Phlox taxa may pose a problem to this method as excessive handling may cause

irreparable damage. In this case, it may be appropriate to experiment with removing the

entire style and applying pollen to the cut surface of the ovary (Eeckhaut et al., 2006).

Further testing of these pre-zygotic barriers through use of pollen/pistil staining using

aniline blue will elucidate whether incompatibility or incongruity affects the success of

interspecific Phlox crosses (Van Tuyl and Lim., 2003). This data also points to the

importance of developing a diverse Phlox germplasm collection and the testing of

different clones in successfully developing interspecific Phlox hybrids.

316

The hybridization data indicates there may also be post-zygotic barriers to hybridization. All successful crosses displayed reduced seed set and capacity to germinate compared to reports of Phlox species (Madeiras et al., 2007). Partial development of seeds in two crosses suggests that seeds fail to mature properly due to endosperm deficiency or failure of the endosperm to support hybrid embryo development

(Morgan et al., 2010). In such instances, embryo rescue may be needed to obtain seedlings of given cross combinations.

Successful cross combinations resulted in the formation of homoploid, diploid hybrids from successful interspecific crosses indicated that crossing over and chromosome recombination is occurring in F1 progeny and may lead to increased

heterozygosity and associated changes in gene expression, altered regulatory pathways,

and rapid and epigenetic changes could also contribute to novel phenotypes, hybrid vigor,

increased environmental adaptation, and disease resistance (Ranney, 2006). The initial

results suggest that interspecific hybridization has potential for creating genetically

diverse, potentially fertile progeny that are worthy of horticultural evaluation in the

development of advanced generation hybrids and as vegetatively propagated individuals

that display unique phenotypes or exceptional environmental adaptation.

Mean height measurements of 2 F1 hybrid sib families exceeded that of either

parent and suggests that some individuals may be exhibiting heterosis as a result of

interspecific hybridization (Table 6.4). However, some individuals were shorter than

either parent and suggest that transgressive segregation may also affect phenotypes of F1

individuals. Although number of flowering stems was intermediate between parental taxa

317

in all sib families, it should be noted that all F1 sib families have the potential to produce flowers throughout the growing season, and there may be variation among individuals in the number of flowering stems produced in a given length of time. Flowering potential of

F1 individuals may also exhibit heterosis (Morgan et al., 2010).

The inheritance of white flower color in ‘Minnie Pearl’ hybrids is complex

(Figure 6.1). These flower color inheritance patterns are in contrast to homozygous inbred lines of P. drummondii, where crosses between plants with carmine colored flowers and white flowers resulted in F1 progeny with uniformly pink flowers, however,

the F2 generation was highly variable and contained a wide-range of flower color

phenotypes (Kelly, 1915). Further crosses between F1 progeny and F1 parental backcrosses are necessary to further elucidate the inheritance of white flower color in

‘Minnie Pearl’ hybrids.

Successful crosses between ‘Minnie Pearl’ and P. paniculata resulted in F1

progeny that were morphologically intermediate between parental taxa and were pollen

sterile (Table 6.4). The areolate leaf veins of P. paniculata were inherited in all F1

progeny, and since this character is unique to subsection Paniculatae, it is a useful

character rapidly identifying putative successful hybrids (Table 6.1). Sterility of F1

hybrids limits their use as parents in a breeding program, but might be possible if individuals produce unreduced gametes. Restoration of fertility may also be possible through artificial conversion of ploidy. Given the susceptibility of P. paniculata to

powdery mildew (Erysiphe cichoracearum) and the noted resistance of ‘Minnie Pearl’, sterile F1 individuals may exhibit the disease resistance; this needs to be systematically

318 verified. Since most Phlox are propagated by asexual means, sterile hybrids lacking the disease susceptibility of P. paniculata could serve as improved cultivars selected directly from the hybrid population.

The results of this study help define crop gene pools for the genus Phlox (Harlan and De Wet, 1971). While no systematic evidence of interspecific hybridization within and between subsection Phlox taxa and other members of the genus exists, anecdotal evidence has provided some indication of initial gene pools (Locklear, 2011a; Symons-

Jeune, 1953). Initial studies indicate that the diploid taxa of subsection Phlox, P. carolina, P. glaberrima, and P. maculata, are in the primary gene pool of ‘Minnie Pearl’, but further testing with other taxa in subsection Phlox is needed. Phlox paniculata would be placed in the secondary gene pool; although these crosses resulted in healthy, vigorous

F1 hybrids, seed germination was poor, and the resultant progeny were sterile. Phlox buckleyi and P. pulcherrima and the remainder of tested taxa would be placed in the secondary gene pool; however, successful hybridization might require embryo rescue techniques (Harlan and De Wet, 1971). Further testing with wider diversity of all tested taxa has the potential to alter the presented results.

In conclusion, this study provides preliminary information for creating interspecific hybrids using Phlox ‘Minnie Pearl’. Although all successful hybrid combinations were unilaterally successful, progeny were vigorous and provided initial information the degree of success expected from interspecific cross combinations, and the inheritance of crop quality traits.

319

b. a.

c . d. e.

Figure 6.1: a. Flowers of P. carolina ssp. carolina PZ11-036. b. Flowers of P. ‘Minnie Pearl’ PZ11-012. c,d,e. Different flower phenotypes of F1 hybrids produced from the cross Phlox carolina ssp. carolina PZ11-036 x Phlox ‘Minnie Pearl’ PZ11-012.

320

’

Minnie Pearl Minnie

‘ 1 F P. paniculata P. B. Gene ladder Gene A. C.

Figure 6.2: SRAP banding showing presence of shared alleles between F1 hybrid of P. paniculata PZ10-231 x P. ‘Minnie Pearl’ PZ11-012 and parental taxa. The primer pair me3-em3 was used for amplification of DNA fragments.

321

Subsection Phlox taxa

P. 'Minnie P. carolina ssp. P. carolina ssp. P. carolina ssp. P.glaberrima ssp. P. maculata P. P. glaberrima ssp. P. maculata Characterz Pearl' PZ11- carolina PZ11- carolina 'Kim' alta glaberrima 'Flower Power' paniculata interior PZ10-249 PZ10-208 019 036 PZ11-036 PZ11-045 PZSH2011-004 PZ10-235 PZ10-249

Stem Height 40-50 cm 60-80 cm 30-45 cm 40-100 cm 50 - 75 cm 60 -120 cm 60-90 cm 75-100 cm 75-200 cm

Ovate- Oblong- elliptic- Leaves lanceolate Ovate-lanceolate Ovate-lanceolate Ovate-elliptical Lanceolate linear to lanceolate linear Oblong-linear ovate

Leaf Vesture Lustrous Pubescent Glabrous Glabrous Glabrous Glabrous Glabrous Glabrous glabrescent

Leaf Veins Obscure Obscure Obscure Obscure Obscure Obscure Obscure Obscure Areolate

Inflorescence Vesture Glabrous Glabrous Glabrous Glabrous Glabrous Glabrous Glabrous Glabrous Pubescent

Calyx Vesture Glabrous Glabrous Glabrous Glabrous Glabrous Glabrous Glabrous Glabrous Pubescent

Flower color White Pink with red striaePink with red striaePink with red striaePink with red striae Pink with red striae Pink white with pink striae Phlox purple

Pollen Color Yellow Yellow Yellow Yellow Yellow Yellow Yellow Yellow Cream -white

sterile stems present Many Few Many Few Few No No No No

Habitat −−− Mesic/Hydric −−− Mesic Hydric Hydric Hydric −−− Hydric

zMorphological characters derived from the keys of Wherry (1935, 1955) and from field observation of naturally occurring Phlox populations and common garden observation at the Ornamental Plant Germplasm Center in Columbus, OH.

Table 6.1: Distinguishing morphological characteristics, flower color, and habitat preferences of Phlox subsections Paniculatae and Phlox taxa used in this study.

322

Style Subsection Taxon Internal accession no Collection Sitey 2C ± SD (pg) CV (%) Ploidy Lengthx

Phlox P. 'Minnie Pearl' PZ11-012 A 14.61±0.14 3.08 2x L

P. carolina L. ssp. carolina PZ11-036 Clay Co. MS 15.25±0.39 4.31 2x L P. carolina ssp. carolina 'Kim' PZ11-017 Tuscaloosa Co. AL 14.38±0.012 2 2x L

P. carolina L. ssp. alta Wherry PZ11-045 Haywood Co. NC 15.67±0.17 1.97 2x L P. glaberrima L. ssp. glaberrima PZSH2011-005 Gadsden Co. FL 16.13±0.27 1.79 2x L

P. glaberrima L. ssp. interior Wherry PZ10-249 B 13.97±0.43 2.38 2x L

P. maculata L. PZ10-208 Adams Co. OH 14.47±0.17 1.93 2x L

P. maculata 'Flower Power' PZ10-235 C 14.85±0.09 1.42 2x L

P. ovata (L.) Locklear PZ12-077 Allegheny Co. VA 12.55±0.09 1.73 2x L

P. ovata PZ10-167 Lucas Co. OH 12.16±0.22 3.51 2x L

Paniculata P . aff. amplifolia Britton PZ11-010 Tucker Co. WV 14.78±0.76 1.94 2x L

P. amplifolia PZ12-106 Campbell Co. TN 15.01±0.13 1.61 2x L

P. amplifolia PZ11-050 Cocke Co. TN 15.21±0.24 3.72 2x L

P. paniculata L. PZ10-231 Erie Co. OH 14.61±0.17 2.65 2x L P. paniculata PZ10-209 Scioto Co. OH 14.97±0.14 1.17 2x L P. paniculata PZ10-109 D 14.74±0.29 1.62 2x L P. paniculata 'Delta Snow' PZ10-027 E 14.90±0.30 1.64 2x L P. paniculata 'Robert Poore' PZ10-028 E 21.74±0.13 1.38 3x L

Cluteanae P. buckleyi Wherry PZ12-088 Greenbrier Co. WV 23.04±0.07 1.06 4x L

Divaricata P. divaricata L. ssp. laphamii (Wood)W PZSH2011-022 Jasper Co. MS 11.03±0.03 1.53 2x S

P. divaricata ssp. laphamii PZSH2011-026 Wilkinson Co. MS 9.78±0.13 1.73 2x S

P. drummondii ssp. drummondii PZ10-161 Greer Co. OK 12.55±0.23 4.53 2x S

P. pilosa L. ssp. pilosa PZSH2011-016 Lowndes Co. AL 12.66±0.11 1.57 2x S

P. pulcherrima (Lundell) Lundell PZSH2011-034 Shelby Co. TX 22.05±0.07 2.46 4x S zTaxonomic classifications are based on Wherry (1955) and Ferguson et al. (1999). yNote: Collection site: A, Plants obtained from Plant Delights Nursery, Raleigh, NC; B, Plant obtained from Prairie Moon Nursery, Witoka, MN obtained from Forestfarm, Williams, OR; D, Plant obtained fromNearly Native Nursery, Fayetteville, GA; E, Plant obtained from Growild Nurs Fairview, TN. xNote: "L" denotes a style longer than 15 mm, and "S" denotes a style less then 15 mm in length.

Table 6.2: Specific germplasm collections, internal accession numbers, collection sites, Relative genome sizes, ploidy levels, and style lengths of Phlox taxa used in interspecific hybridization experiments.

323

Parental Combinationz

Seeds Seeds Germination Subsection Female Parent (♀) Male Parent (♂) obtained germinated Accession no. (%)x (No.)y (No.)

Phlox P. carolina ssp. carolina PZ11-036.1 P. carolina ssp. carolina PZ11-036.2 128 99 77.3 PZ11-090 P. carolina ssp. carolina PZ11-036.2 P. carolina ssp. carolina PZ11-036.1 22 14 63.6 PZ11-092

P. carolina ssp. carolina PZ11-036 P. 'Minnie Pearl' PZ11-012 21 16 76.2 PZ11-131 P. 'Minnie Pearl' PZ11-012 P. carolina ssp. carolina PZ11-036 0 −−− −−− −−−

P. carolina ssp. carolina 'Kim' PZ11-0P. 'Minnie Pearl' PZ11-012 0 −−− −−− −−− P. 'Minnie Pearl' PZ11-012 P. carolina ssp. carolina 'Kim' PZ11-017 17 16 94.1 PZ11-154

P. carolina ssp. alta PZ11-045 P. 'Minnie Pearl' PZ11-012 16 11 68.8 11-128 P. 'Minnie Pearl' PZ11-012 P. carolina ssp. alta PZ11-045 0 −−− −−− −−−

P. 'Minnie Pearl' PZ11-012 P. glaberrima ssp. interior PZ10-249 0 −−− −−− −−− P. glaberrima ssp. interior PZ10-249 P. 'Minnie Pearl' PZ11-012 15 10 66.6 11-123

P. maculata 'Flower Power' PZ10-235 P. 'Minnie Pearl' PZ11-012 12 8 66.6 11-124 P. 'Minnie Pearl' PZ11-012 P. maculata 'Flower Power' PZ10-235 0 −−− −−− −−−

P. maculata PZ10-208 P. 'Minnie Pearl' PZ11-012 18 6 33.3 PZ11-142 P. 'Minnie Pearl' PZ11-012 P. maculata PZ10-208 14 −−− 0 PZ11-155

Paniculatae P. paniculata PZ10-209 P. 'Minnie Pearl' PZ11-012 5 2 40.0 11-081 P. 'Minnie Pearl' PZ11-012 P. paniculata PZ10-209 0 −−− −−− −−−

P. paniculata PZ10-231 P. 'Minnie Pearl' PZ11-012 51 8 15.7 11-060 P. 'Minnie Pearl' PZ11-012 P. paniculata PZ10-231 0 −−− −−− −−−

P. 'Minnie Pearl' 11-012 P. paniculata 'Delta Snow' PZ10-027 0 −−− −−− −−− P. paniculata 'Delta Snow' PZ10-027 P. 'Minnie Pearl' 11-012 27 −−− 0 −−−

Cluteanae P. 'Minnie Pearl' PZ11-012 P. buckleyi PZ12-088 8/PD −−− −−− 11-222 P. buckleyi PZ12-088 P. 'Minnie Pearl' PZ11-012 7/PD −−− −−− 11-223

Divaricatae P. 'Minnie Pearl' PZ11-012 P. pulcherrima PZSH2011-034 5/PD −−− 0 −−− P. pulcherrima PZSH2011-034 P. 'Minnie Pearl' PZ11-012 0 −−− −−− −−−

z50 pollination per cross were performed over a period of 3-7 days, not all of any given cross was performed on a single day. ySeed was collected by bagging the inflorescence 7-10 weeks after completion of crosses and allowing the seeds to fall when mature. xGermination (%) calculated as the number of seeds that germinated and produced a seedling by the total number of seeds.

Table 6.3: Parental combinations, seeds obtained (No.), and germination (%) of successful and unsuccessful, reciprocal interspecific crosses in which P. ‘Minnie Pearl’ serves as a parent.

324

flowering stems Genotype Height (cm) Leaf length (cm)Leaf width (cm) per pot (No.)

Phlox 'Minnie Pearl' x Phlox carolina 'Kim'

P. carolina 'Minnie Pearl' 45.7az 8.50b 7.00a 1.97a P. carolina 'Kim' 33.5b 12.83a 6.03b 1.22b

F1 49.0a 13.75a 6.71ab 1.63a

P. maculata 'Flower Power' x Phlox 'Minnie Pearl'

P. carolina 'Minnie Pearl' 45.7c 8.50a 7.00a 1.97a P. maculata 'Flower Power' 73.2b 4.60b 6.32a 1.46b

F1 45.67c 7.43a 4.26b 1.57b

Phlox glaberrima ssp. interior PZ10-249 x Phlox 'Minnie Pearl'

P. carolina 'Minnie Pearl' 45.7c 8.50a 7.00a 1.97a P. glaberrima ssp. interior PZ1 84.8a 5.0b 6.08ab 1.42b

F1 64.2b 7.40ab 5.66b 1.42b

Phlox carolina ssp. alta PZ11-045 x Phlox 'Minnie Pearl'

P. carolina 'Minnie Pearl' 45.7ab 8.50a 7.00a 1.97b P.carolina ssp. alta PZ11-045 43.6b 6.00a 5.46b 2.42a

F1 50.7a 9.00a 4.483c 1.63b

Phlox carolina ssp. carolina PZ11-036 x Phlox 'Minnie Pearl'

P. carolina 'Minnie Pearl' 45.7b 8.50a 7.00a 1.97a P.carolina ssp. carolina PZ11-0 59.2a 7.00a 6.32b 2.20a

F1 48.4b 7.8a 6.68ab 1.96a

zNumbers of followed by the same letter are not siginificantly different as determined using analysis of variance with means separation by Fisher's least significant difference at P = 0.05.

Table 6.4: Mean plants height (cm), flowering stems per pot (No.), leaf length (cm), and leaf width (cm) for parental species and interspecific F1 hybrids among Phlox species.

325

CHAPTER 7

INTERSPECIFIC HYBRIDIZATION IN PHLOX FOR GERMPLASM ENHANCEMENT: INTERPLOID CROSSES IN SUBSECTION DIVARICATAE

Abstract

Development of new breeding lines is vital to the development of novel hybrids of ornamentals. Despite a history of hybridization, there is modern interest in Phlox interspecific hybridization, but development is hindered by a lack of information about crossing relationships in the genus. Ploidy differences have previously been considered a barrier to interspecific hybridization in the genus Phlox L, but this was determined from few attempts at interploid crosses that used a limited subset of polyploid taxa. The objective of this study was to determine chromosome counts and DNA content (genome size) of selected taxa, to evaluate crossability among species of different ploidy in Phlox subsection Divaricatae, and to confirm hybridity using chromosome counts, flow cytometry, and Sequence Related Amplified Polymorphism (SRAP) markers with the goal of facilitating future breeding endeavors and development of new hybrids. Phlox floridana, P. pulcherrima , and P. villosissima were tetraploid (2n=4x=28) and 2C genome size ranged from 21.85–26.58 pg. Phlox divaricata, P. drummondii, and P. pilosa were diploid (2n=2x=14) and genome sizes ranged from 9.78–12.55 pg. Interploid crosses were successful when tetraploids served as the female parent in crosses with diploid, perennial species, but could only serve as the male parent in crosses with the

326

diploid, annual P. drummondii. Interploid crosses resulted in the formation of

morphologically intermediate, aneuploid F1 progeny with n=2x-1=13, n=2x-2=12 and

n=2x-3=11 chromosomes, and genome sizes intermediate to parental taxa that ranged

from 17.28 – 17.78 pg. SRAP markers also confirmed hybridity of interploid crosses.

Aneuploid hybrids were sterile and exhibited low to no pollen staining; sib and parental

backcrosses of F1 hybrid individuals were unsuccessful. This study demonstrates

successful interploid crosses among Phlox taxa, but indicates that further development of hybrids may only be possible through ploidy conversion or via discovery of unreduced gametes among F1 progeny.

Introduction

Interspecific hybridization is considered the primary method for development of

genetic diversity in ornamental crops (Eeckhaut et al., 2006; Van Tuyl and De Jeu, 2005).

Allopolyploid, aneuploid, and triploid hybrids can differ from diploid and tetraploid

progenitor species in various ways, and offer numerous advantages to plant breeders. As

a result of the increase in the number of chromosomes and DNA content, and

recombination of parental genomes, allopolyploid hybrids of different origins may have

altered physiological features such as increased and novel enzyme and secondary

metabolite production, altered water relations and photosynthesis, flower initiation,

duration of flowering, and alteration of gross morphological characters (Eeckhaut et al.,

2006; Zhou et al., 2008). The genus Phlox represents a highly heterozygous,

phenotypically diverse range of taxa; introgression of desirable traits from multiple

327

species could result in new breeding lines that improve upon commercially available

selections.

Phlox contains ca. 65 species of annuals and perennials grown for their long-

lasting, terminal inflorescences of showy flowers born in prodigious displays and a range

of colors from April until October (Wherry, 1955). Superficially, these species comprise

two groups: a western group consisting primarily of pulvinate or caespitose,

suffrutescent species adapted to specialized climatic conditions in mountain ecosystems,

and an eastern group consisting of a diverse array of primarily herbaceous species

adapted to a variety of ecosystems and habitat types. The eastern species form the basis

of horticultural selections and continue be a source of new horticultural introductions

(Wherry, 1935a). The ca. 22 eastern species are distinguished on the basis of

morphology and geographical distribution and divided into 6 taxonomic subsections:

Cluteanae, Divaricatae, Paniculatae, Phlox, Stoloniferae, and Subulatae (Locklear,

2011a; Prather, 1994; Wherry, 1955). Phlox subsection Divaricatae consists of as many as 10 species and 11 subspecific taxa; extensive population level phenotypic variation is prevalent in the subsection, and identification of taxa requires precise knowledge of distinguishing morphological characteristics (Table 7.1) (Ferguson et al, 1999; Smith and

Levin, 1965; Levin, 1966; Prather, 1994; Wherry, 1955). Aside from morphological characters of phylogenetic value, there can be extensive variation in flower color, plant habit, and foliage characteristics that are of interest to horticulturists.

Most species in subsection Divaricatae are diploid, but tetraploid taxa have been described and used to delineate some species (Flory, 1933; Flory, 1934; Levin, 1964;

328

Smith and Levin, 1967; Worcester et al., 2012). Recent cytological evidence

demonstrated that geographically proximal populations of P. pilosa ssp. pilosa from the

western edge of the species range are comprised of both diploid and tetraploid individuals

(Worcester et al., 2012). Populations consisted only of individuals of a single ploidy

level, but genome size was variable and no macromorphological differences were noted

between populations with different ploidy levels. (Levin, 1968; Worcester et al., 2012;

Smith and Levin, 1967). While triploid populations have not been found, sampling in this study was not exhaustive, and the range in genome sizes suggest that triploid individuals may exist as a result of gene flow between populations.

The only previous studies of hybridization in Phlox subsection Divaricatae concluded that differences in ploidy were an absolute barrier to interspecific hybridization (Levin, 1966). This is contradicted by earlier evidence in which chromosome counts indicated that Phlox Suffruticosa Group ‘Miss Lingard’ was triploid, indicating a hybrid origin for this cultivar (Flory, 1933; Flory 1934; Locklear, 2011a).

‘Miss Lingard’ is also reportedly sterile, and cannot set seeds, which further supports an origin via interploid hybridization (Wherry, 1935a; Wherry, 1955). In a corollary study of genome size variation among P. paniculata cultivars, four cultivars with intermediate genome sizes approximating triploids were found that suggest past interploid hybridization, although the plants do not show morphological irregularities (Chapter 2).

There is also evidence that pentaploid populations of P. amabilis arose via interploid hybridization between geographically proximate tetraploid and hexaploid populations

(Fehlberg and Ferguson, 2012). While pollen viability of this population was not

329 reported in the study, natural interploid hybridization appears possible between geographically proximate natural populations that leads to development and establishment of aneuploid hybrid populations (Fehlberg and Ferguson, 2012). Thus, not only is interploid hybridization possible in Phlox, but it may represent an important evolutionary pathway for speciation in natural populations. Experimental, artificial interploid hybridization studies in Phlox may yield novel information about reproductive pathways and result in F1 generations for further testing.

Current knowledge of interspecific hybridization in subsection Divaricatae was derived from a few studies (Levin, 1966; Levin, 1968). Despite the exhaustive nature of this research, the biosystematic and floristic nature of this work leave fundamental questions unanswered particularly in relation to more applied interspecific hybridization efforts. Crossing data is presented for only two polyploid taxa, P. floridana and P. pilosa ssp. pilosa, despite the occurrence of other tetraploid members of the subsection.

Among intra-subsectional, interploid crosses, data is presented for only one genotype each of P. amoena, P. divaricata, and P. pilosa ssp. pilosa despite the fact that numerous genotypes of each of these species were used in crosses at the diploid level (Levin, 1964;

Levin, 1966). Additionally, ploidy was not described for the majority of parental taxa, possibly confounding some of the reported effects of hybridization, since polyploidy has been found to be more widespread among natural Phlox populations than previously thought. Perhaps the most important reason to reassess the potential for interpecific interploid hybridization among Phlox, was that the only criterion for successful hybridization was formation of seed, and the author cryptically stated that “…of the

330

hundreds of hybrid seeds obtained…only a small percentage germinated and a smaller

portion reached the flowering stage” (Levin, 1966). Therefore, living plants were never

obtained for the majority of crosses, and definitive confirmation of successful

hybridization remains in doubt for some cross combinations (Levin, 1966). Further

systematic testing of barriers to interspecific hybridization among Phlox taxa could yield

additional information about the potential for creating interploid hybrids.

Subsection Divaricatae is noted among other Phlox subsections for the high

degree of phenotypic and genetic diversity, yet few cultivar selections have been made

and some species have never been successfully introduced into cultivation (Bendtsen,

2009; Hawke, 1999; Locklear, 2011a; Wherry, 1932). These species inhabit a diverse

range of ecosystems and geographical regions. Numerous taxa occur in markedly xeric

habitats, in addition to being desirable ornamental plants; those from marginal habitats

and the perimeter of the geographic distribution of the subsection are generally tetraploid

(n=14) (Wherry, 1955). While some taxa in this group have been described as poor

performers in garden and landscape settings, at least one putative hybrid taxon, Phlox

‘Chattahoochee’, considered a hybrid between P. divaricata and P. pilosa, has persisted in horticulture since its introduction in the 1930’s and was one of the most highly rated taxa in at least one Phlox evaluation trials (Bendtsen, 2009; Hawke, 2011). Phlox

‘Chattahoochee’ and the parental taxa are diploids. Phlox divaricata is a species of mesic or hydric habitats on alluvial soils, while P. pilosa ssp. pilosa grows in xeric, disturbed habitats on thin soils (Wherry, 1955). The success of Phlox ‘Chattahoochee’ suggests that crossing species with different ecological tolerances can result in hybrids that are

331 more adaptable to landscape and garden settings (Wherry, 1935a). Further development of persistent, garden worthy Phlox hybrids from species in subsection Divaricatae may best be accomplished by interspecific hybridization between selected species and/or cultivars, rather than selection of cultivars from species. This approach to ornamental crop improvement has been successful with a variety of genera, including Echinacea,

Heuchera, and Coreopsis (Otteson, 2003; Schoellhorn and Richardson, 2004).

Interploid, interspecific hybridization has been used to develop novel cultivars and breeding lines in a variety of ornamental plant genera, but has yet to be broadly tested in

Phlox, although there is potential for the development of such breeding lines (Cisneros and Tel-zur, 2012; Palmer et al., 2008; Zhou et al., 2008).

The objective of this research was to; 1) collect and introduce a diverse range of diploid and tetraploid Phlox germplasm from a natural plant populations 2) determine the potential for interploid interspecific crossability between diploid and tetraploid members subsection Divaricatae 3) to use flow cytometry and cytology to determine holoploid genome sizes and ploidy level of selected Phlox species used in novel experimental crossability experiments and their putative F1 hybrids and to (4) determine fertility of putative Phlox F1 hybrids through use of sib, parental, and experimental, interspecific 3- way crosses, and pollen staining.

332

Materials and Methods

Phlox collection

To obtain Phlox species not available from nursery sources and to insure proper

identity of taxa, we conducted a series of botanical explorations to numerous areas of the

United States to collect germplasm from natural plant populations. Collection

expeditions were made over a 3-yr period from 2010-2013. Plant identities were confirmed using the dichotomous keys provided by Wherry (1955), with modifications from various sources (Ferguson et al., 1999; Prather, 1994; Locklear, 2011a). Herbarium specimens were made for each collection and are currently housed at the Ornamental

Plant Germplasm Center (OPGC) in Columbus, OH.

Controlled pollinations

Phlox taxa used in this study are protandrous, out-crossing species with gametophytic self-incompatibility; so individual flowers can receive and donate pollen without risk of self-pollination (Levin, 1975; Ruane and Donohue, 2007). Flowers in which pollen was shedding and the tripartite stigma was open were chosen for the pollination. All taxa used in this study are “short-styled” and have a style no more than

4mm long, a tripartite stigma with lobes nearly equal to the length of the pistil, and the entirety of the pistil surrounded by the calyx lobes (in contrast, “long-styled” Phlox have a style up to 25mm long that is exerted from the calyx, and were not used in this study)

(Ferguson et al. 1999; Wherry, 1955). To access the stigma, the corolla was carefully removed from the calyx and the sepal lobes trimmed with scissors. The calyx of several

333 taxa in subsection Divarcatae is covered in glandular pubescence that becomes viscid when handled and can be easily damaged. Early experimentation indicated that removal or excessive handling of the entire calyx resulted in abortion of the flower and failure of the cross.

Crosses were performed during 2011 and 2012 in a partial diallel with a minimum of 50 pollinations (individual flowers) per cross. Nearly all Phlox species can develop a maximum of three seeds per flower. Crosses were performed over a 3-6 week period in the greenhouses at the Ornamental Plant Germplasm Center, in Columbus, OH.

Greenhouse vents were screened to eliminate potential contamination from insect pollinators. Successful pollination was defined as the formation of at least 5 viable seeds per cross. All Phlox species have ballistic seed dispersal and developing infructescences were bagged with nylon pantyhose to aid in seed collection (Wherry, 1955). Seeds of species in subsection Divaricatae typically mature 3-5 weeks after pollination; they were harvested when they had dehisced and could be seen inside the nylon bag that surrounded the infructescence.

Interspecific interploid hybridization was considered successful if viable seeds were germinated and resulted in living plants. In general, seeds of taxa in subsection

Divaricatae exhibit non-deep morphological dormancy (Baskin and Baskin, 2004).

Seeds were sown in germination boxes (Tri-State Plastic, Henderson, KY) on blotter paper (Anchor Paper Co., St. Paul, MN) and placed in a walk-in cooler at 7°C for a 60-90 day cold stratification period. Enough water was applied to make sure the paper was

334 moist at all times. After this period, seeds were placed in growth chamber at 22°C in the dark. Seeds were considered to have germinated when a 2 mm radicle was evident.

Germinated seeds were transplanted to individual plastic containers (11 cm diameter round by 9.5 cm deep, Dillen Products, Middlefield, OH) in Metro mix 360

(Scotts-Miracle Gro Co., Marysville, OH) and grown on a greenhouse bench under natural light conditions at temperatures of 22.2 °C ± 6 °C 0800 to 1800 HR and 18.3 °C ±

3 °C from 1800 to 0800 HR.

Seedlings were given unique accession numbers (Table 2). The numbers were assigned as follows; PZ refers to the first author, 11 refers to the year the pollinations were made, and the number following refers to a specific cross combination. Progeny within a given cross combination were also given individual accession numbers to distinguish them from sibs in a given cross.

Flow cytometric analysis, chromosome counts, and SRAP analysis

Methodology for flow cytometric and cytological analysis were described in

Chapter 2. SRAP (sequence related amplified polymorphism) analysis is described in

Chapter 6.

Pollen Staining

Pollen was collected from F1 hybrid progeny and stained with 1% acetocarmine, heated over an open flame until the cover slip turned cloudy, squashed, and visualized

335 under 400x magnification. Replications of 50 pollen grains were made on at least 3 different slides on at least 3 different days.

Results

Parents used for hybridization were individuals selected from germplasm collection expeditions introduced from natural plant populations. Although cultivars of

P. divaricata and P. pilosa were available from nursery sources, novel variants collected during botanical field trips were used instead and allowed us to generate completely novel hybrids separate from other Phlox breeding efforts. The tetraploid species P. floridana, P. pulcherrima, and P. villosissima were not available from nurseries and had to be collected from natural plant populations (Chapter 4). Morphology, ploidy, genome size and geographic distribution were used to correctly identify each collection (Table 2)

Phlox divaricata, P. drummondii, P. pilosa ssp. longipilosa, and P.pilosa ssp. pilosa, were confirmed as diploids (2n=2x=14) with genome size ranging from 9.78-

12.58 pg. Phlox floridana, P. pulcherrima, and P. villosissima were tetraploid

(2n=4x=28) with a genome size that ranged from 21.85-26.85 pg (Table 7.2). Genome size data and chromosome counts are similar to those presented in previous studies, and support the use of Propidium Iodide as the fluorochrome in flow cytometry of Phlox taxa

(Flory, 1931; Flory 1934; Smith and Levin, 1967; Worcester et al., 2012). No hexaploids have been found in subsection Divaricatae, but they exist in other subsections of the genus (Flory, 1934; Chapter 4)

336

Interspecific interploid Phlox parental combinations display unilateral incongruity; they were only successful in one direction (Table 7.3). The tetraploid individual served as the female in three of four successful interploid cross combinations.

For the fourth cross, the diploid annual species P. drummondii served as the female

parent in a cross with tetraploid P. pulcherrima. One cross resulted in partially developed

seeds, and seed failed to develop for several crosses on either parent. Regardless of

parental combination, all F1 progeny had a genome size intermediate to both parents and

ranged from 17.28-17.58 pg (Figure 7.1, 7.2). Meiotic metaphase chromosome counts

confirmed that F1 progeny were alloaneuploids with varying chromosome counts between

n=2x-1=13, n=2x-2=12, and n=2x-2=12 (Figure 3). The intermediate genome size of F1

hybrids confirmed successful interploid interspecific hybridization and highlights the

utility of flow cytometric analysis to rapidly screen hybrid populations derived from

parental taxa with disparate genome sizes (Chapter 5).

The hybridity of plants derived from the cross P. floridana x P. longipilosa was

first observed based on intermediate phenotype and then confirmed with genome size and

SRAP markers (Figure 7.4, 7.5). This initial screening using the primer pair me3-em3

showed the presence of shared bands among P. floridana, P. longipilosa, and their F1

hybrid. SRAP banding patterns have been used to confirm hybridity in several

horticultural genera: Brassica, Coffea, and Lactuca (Li and Quiros, 2001; Liu et al., 2012;

Mishra et al., 2011). Electropherograms confirmed the presence of shared alleles; peak

height of shared alleles was similar among parents and hybrid. Although intermediate

morphology can be used to readily identify Phlox hybrids, these results suggest that the

337

SRAP markers could be used to confirm hybridity between closely related taxa or

cultivars and may be of value for characterizing genetic diversity in a Phlox germplasm

collection (Hudak et al., 2004).

The progeny of the crosses P. floridana x P. longipilosa, and P. drummondii x P.

pulcherrima exhibited low to no pollen staining (data not presented). F1 individuals used

in sib, and parental backcrosses were unsuccessful and further confirmed the sterility of

alloaneuploid F1 hybrids.

Discussion

The goal of this study was to determine the potential for the development of

interploid, interspecific Phlox hybrids for crop improvement among taxa in subsection

Divaricatae using germplasm collected from natural plant population as parental taxa.

The following conclusions may be drawn from this study: 1) Differences in ploidy are

not a barrier to interspecific hybridization among Phlox species in subsection

Divaricatae, and F1 progeny from interploid crosses resulted in the formation of sterile aneuploid or triploid individuals. 2) All interploid crosses displayed unilateral incongruity; F1 progeny were intermediate between parental taxa in genome size,

chromosome number, and gross morphology; these results suggest that ploidy alone does

not determine the success of a given cross, and that incompatibility and incongruity

contribute to several pre and post zygotic phases of hybrid development (Eeckhaut et al.,

2006; Hogenboom and Mather, 1975).

338

Phlox species are highly heterozygous, obligate out-crossing species that have

well-developed gametophytic, self-incompatibility (SI) system (Levin, 1975; Levin,

1993). The SI system is known to affect interspecific hybridization through inhibition of pollen tube germination when the same SI locus and S-gene allele is found in the pistil

and pollen of parental taxa. In one natural population of P. drummondii, the SI system

was regulated by a single gene at the S-locus, but as many as 30 different S-gene alleles were found, indicating that extensive genotypic diversification at the S-locus occurs in P. drummondii, and other Phlox species (Levin, 1993). A study of variation in S-locus haplotypes in Brassica and Arabidopsis indicated a single evolutionary origin of SI in both genera, and suggests that similar cases may exist in other genera (Edh et al., 2009).

While similar studies do not exist in Phlox, phylogenetic studies of Phlox taxa in subsection Divaricatae suggest that some taxa may be of recent divergence or have experienced natural hybridization events; these taxa may share S-locus haplotypes and S- locus alleles that may prevent successful interspecific hybridization, or have sufficient allelic variation that does not inhibit pollen tube germination, but restricts pollen tube growth. To support this hypothesis, certain individuals of taxa involved in successful interploid crosses failed in different cross combinations when used in the same or similar parental combinations when another clone of the other parent was used; this suggests that shared allelic variation at S-gene locus among specific Phlox clones may prevent pollen tube germination or growth, and represent a barrier to interspecifc hybridization regardless of differences in ploidy. Prevention of hybridization by the SI system is further supported by the fact that most unsuccessful crosses completely failed to develop

339

even partially developed seed, suggesting a pre-zygotic barrier. Careful assessment of pollen/pistil interaction will elucidate whether incompatibility or incongruity affects the success of interploid Phlox crosses (Van Tuyl and Lim, 2003). This data also points to the importance of developing a diverse Phlox germplasm collection and the testing of different clones in successfully developing interploid interspecific Phlox hybrids.

Incongruity occurs in interspecific crosses as a result of a lack of genetic information in one partner to complete pre and post pollination processes in the other

(Eeckhaut et al., 2006; Hogenboom and Mather, 1975; Van Tuyl, 2005). Pre and post fertilization success is regulated by a series of complicated developmental steps during which the pollen tube interacts with the pistil, the micropylar end of the ovule, and the

egg. In order for these processes to continue unimpeded, genetic recognition of both

parents is required for successful fertilization. The cross of P. pulcherrima x P.

xglutinosa was only successful with the tetraploid as the parent, but the reciprocal cross result in partially developed seed, suggesting that incongruity in seed development may

constitute a barrier to hybridization. Interpretation of successful and unsuccessful

parental cross combinations using the endosperm balance number theory (EBN) may

explain the unilateral incongruity of interploid interspecific crosses (Carputo et al., 1999;

Johnston et al., 1980). Regardless of ploidy, the EBN theory proposes that the most

important component in the success of an interploid cross is maintenance of the 2:1,

maternal: paternal, endosperm: embryo contribution. Deviations from the 2:1 ratio result

in the irregular development of endosperm and abnormal or arrested development of the

embryo in interploid crosses involving a wide array of crop and wild species (Carputo et

340

al., 1999; Johnston et al., 1980). Additional post-zygotic effects, such as cytoplasmic

inheritance, expression of deleterious genes in the gametophytes and gametes, and

differential expression of alleles in the developing seeds have been described (Dilkes and

Comai, 2004). These pre and post-zygotic barriers provide a potential framework for understanding the mechanisms influencing the success and failure of given crosses.

Despite the biological factors that contribute to these potential limitations, interploid crosses, when successful, may significantly contribute to the development of adaptable and ornamental Phlox hybrids.

The formation of alloaneupolyploids from interploid interspecific crosses indicated that crossing over and partial chromosome recombination is probably occurring in F1 progeny and may lead to increased heterozygosity and associated changes in gene

expression, altered regulatory pathways, and epigenetic changes that could also

contribute to novel phenotypes, hybrid vigor, and increased environmental adaptation

(Ranney, 2006; Soltis et al., 2006). At least two previously unidentified aneuploid P.

paniculata cultivars (‘Robert Poore’, John Fanick’) were consistently rated among the

top-performing plants in evaluation trials in geographically disparate regions of the

United States, and suggests that partial polyploids may have an adaptive advantage over

diploid cultivars (Hawke, 1999). Early botanists considered P. floridana and P. villosissima as subspecies of P. pilosa that represented divergent ecotypes adapted to climatic and ecological extremes at the southern and western edge of the range of P. pilosa (Locklear, 2011a; Wherry, 1955). Biosystematic studies suggest that polyploidization in these species arose through interspecific hybridization of P. pilosa

341 with sympatric Phlox taxa, but more recent molecular evidence suggests that autopolyploidization may also have contributed to the evolution and diversification of

Phlox species (Levin, 1966; Levin, 1968; Fehlberg and Ferguson, 2012; Worcester et al.,

2012). In either scenario, it is likely that altered or increased ploidy levels contributed to increased adaptation to marginal ecosystems and suggested that artificial interploid interspecific hybridization could result in highly adaptable hybrids for differen hardiness zones and geographical regions.

Susceptibility to disease and rodent predation are limiting factors in the widespread cultivation of Phlox (Hawke, 2011). Interspecific hybrids may offer increased adaptability to garden and landscape environments by production of novel secondary metabolites that contribute to increased adaptability (Zhou et al., 2002).

Recombination of parental genomes may result in production of parental and hybrid enzymes and secondary metabolites that may have a role in the increased adaptation of hybrid taxa to a wider diversity of climates and soil types than parental taxa (Levin, 1968;

Ranney, 2006). Previous studies have shown that chromatographic patterns of phenolic compounds of P. floridana, P. pulcherrima, and P. villosissima displayed novel flavonoid variation indicative of allopolyploiziation and contained a greater diversity of novel flavonoid compounds in comparison with diploid Phlox taxa (Levin, 1968; Levy and

Levin, 1974). Flavonoid compounds are involved in numerous plant defense systems and hybridization among polyploidy taxa has potential in producing broadly adapted, pest resistant phloxes (Levin, 1976; Wink, 1988).

342

Another post fertilization barrier is the sterility of the interploid hybrids.

Although such sterility can prevent the further development of breeding lines, it can also

have numerous benefits in ornamental plants. Sterile hybrids may exhibit hybrid vigor

that is manifest as increased plant growth, flower production, and longevity of flowering that is superior to either parent or related species. Additionally, progeny of interploid crosses may be more adaptable to vegetative propagation of production systems (Hawke,

1999). There is additional pressure on the floriculture and nursery industries for breeding sterile plants to reduce the potential for invasiveness (Ranney, 2006). There is evidence for invasive potential in P. drummondii; naturalized populations have been reported in

India and outside of the species natural geographic distribution in the United States.

Vegetatively propagated, sterile Phlox cultivars could be used in place of fertile seed strains in regions where invasive potential could be problematic.

Interploid interspecific hybridization can be used to introgress desirable traits into

popular floriculture species where there is potential for improvement of existing seed

grown F1 hybrids. Phlox drummondii is a popular bedding plant because of the unique flower color variants available from large number of seed strains and a few vegetatively propagated series of cultivars (Intensia ®, Astoria ®) (Levin, 1978; Schoellhorn, 2005).

Production of F1 seeds is labor intensive, seed grown plants enter a rosette phase that increases the number of days to flowering, and plants are not tolerant of heat and humidity (Schoellhorn, 2005). Although vegetatively propagated cultivars are an

improvement over F1 seed strains, interspecific hybridization could result in the

production of plants that combine the range and depth of coloration in P. drummondii

343

with improved foliage and habit characteristics of related taxa that are hardy, adaptable

perennials in a wide range of climatic zones. Previous reports of hybrids between P.

drummondii and other diploid members of subsection Divaricatae (P. divaricata, P. pilosa) are likely to exhibit the annual life-cycle of P. drummondii, but variation for this

trait is likely dependent upon parental combination (Charles Oliver, Personal

communication 10 June 2012; Levin, 1966). Preliminary evaluation of interploid hybrid

progeny of the P. drummondii x P. pulcherrima suggests that some of individuals from this breeding line exhibit a perennial life cycle, continuous flowering when grown under long days, and adaptation to vegetative propagation. Despite these benefits of hybridization, the flower color of F1 individuals is intermediate between the bright red

flowers of the female P. drummondii and pale pink P. pulcherrima; the flower color is

inferior to the P. drummondii and suggests that using F1 progeny in further breeding

efforts is necessary to acquire hybrids with the desired suite of morphological characters.

Ploidy conversion of aneuploid Phlox hybrids may result in production of fertile

polyploids that can be used in the development of advanced generation hybrids (Eeckhaut

et al., 2006; Ranney, 2006). Colchitetraploids of P. drummondii and P. subulata were

generated from treatment of seedlings of the former and in vitro stem cuttings of the

latter, and suggested that other Phlox taxa can be successfully converted using colchicine

or related spindle formation inhibitors (Meyer, 1944; Zhang et al., 2008). This technique

has also been useful for restoration of fertility in a variety of other ornamental crops

(Contreras et al., 2009; Olsen et al., 2006). However, the nature of chromosome pairing and recombination in Phlox alloaneuploids is not known, further application of polyploid

344

breeding using this methodology should proceed after GISH analysis is used to reveal

chromosomal inheritance and recombination in alloaneuploid hybrids (Dar et al., 2013).

Colchiploids may have “fixed heterozygosity” that limits downstream breeding efforts

due to lack of genomic recombination in advanced generation hybrids (Barba-Gonzalez

et al., 2008).

Unreduced gametes have been shown to occur in most angiosperms, and there is

evidence of increased formation of unreduced gametes in highly sterile triploid or

aneuploid hybrids (Harlan and De Wet, 1975; Levin, 2012). Alloaneuploid Phlox

hybrids are highly sterile, and analysis of pollen genome size may help detect the

presence of diploid pollen and the stage or stages at which it develops. Protocols exist

for harvesting pollen, creating nuclear suspensions, and utilizing flow cytometry to detect

2n pollen (Kron and Husband, 2012; Pan et al., 2004). The stage or stages of meiosis in

which unreduced gametes are produced, influences the breeding behavior of the parents and influences the genetic constitution of advanced generation hybrids. The key advantage of using unreduced gametes over colchiploids is that genomic recombination and introgression occur and can create novel genetic variation as compared to induced polyploids (Barba-Gonzalez et al., 2008).

The results of this study help define crop gene pools for the genus Phlox (Harlan and De Wet, 1971). Several studies indicate that crossing barriers among diploid taxa in subsection Divaricatae are weak and readily breached and essentially all diploid members of the subsection could be considered in the primary gene pool of broadly distributed species such as P. divaricata and P. pilosa. (Levin, 1966; Levin, 1993;

345

Chapter 4). In relation to diploid taxa, tetraploid taxa would be placed in secondary gene pool because only some parental combinations are successful and the reported hybrids are sterile (Harlan and De Wet, 1971). However, all alloaneuploid plants are vigorous plants, and do not appear to have the weak growth or lethal alleles that might be encountered in alloaneuploid hybrids of other crop species. Further analysis of hybridization between subsection Divaricatae taxa and taxa from other subsections will help refine the overall crop gene pools for the genus Phlox.

In conclusion, it appears that interploid interspecific breeding has potential for creating genetically diverse, sterile alloaneuploids that are worthy of horticultural evaluation and probably further development of advanced generations. Although unilateral incongruity may influence the success of a given cross, the formation of alloaneuploids suggests that parental genomes experience some degree of recombination and that desirable traits from both parents may be introgressed to create novel genetic diversity. Given the range of genetic phenotypic variation within many Phlox species, careful selection of parental taxa from natural Phlox populations could be used to create myriad interploid interspecific breeding lines.

346

a. b. c.

Number of nuclei of Number

Relative DNA fluorescence

Figure 7.1: Flow cytometry histogram showing concurrent comparison of a. diploid (n=7) P. pilosa ssp. longipilosa (12.55 pg) b. alloaneuploid (n=13) P. floridana x P. pilosa ssp. longipilosa F1 hybrid (17.28 pg), and the c. tetraploid (n=14) P. floridana (21.85 pg).

347

25 25

20 20

15 15

10 10

5 5 DNA content (pg) content DNA 0 0 a. P. longipilosa F1 P. floridana b. P. drummondii F1 P. pulcherrima

♂ ♀ ♀ ♂

30 30

25 25

20 20

15 15

10 10

5 5 DNA content (pg) content DNA

0 0 d. P. xglutinosa F1 P. villosissima c. P. divaricata F1 P. villosissima ♂ ♀ ♂ ♀

Figure 7.2: Histograms showing relative genome sizes of parental taxa and F1 hybrids from 4 interploid, interspecific crosses. a. Diploid (n=7) P. pilosa ssp. longipilosa PZSH11-043 (12.55 pg), alloaneuploid (n=13) P. floridana x P. pilosa ssp. longipilosa F1 hybrid PZ11-107 (17.28 pg), and tetraploid (n=14) P. floridana PZSH11-010 (21.85 pg). b. Diploid (n=7) P. drummondii PZ10-161 (12.22 pg), alloaneuploid (n=13) P. drummondii x P. pulcherrima F1 hybrid PZ11-175 (17.78 pg), and tetraploid (n=14) P. pulcherrima PZSH11-034 (22.05 pg). c. diploid (n=7) P. divaricata PZSH11-022 (11.03 pg), alloaneuploid (n=13) P. divaricata x P. villosissima F1 hybrid PZ11-202 (17.79 pg), and tetraploid (n=14) P. villosissima PZSH11-042 (26.58 pg). d. diploid (n=7) P. xglutinosa PZ11-127 (10.90 pg), alloaneuploid (n=13) P. xglutinosa x P. villosissima F1 hybrid PZ11-212 (19.47 pg), and tetraploid (n=14) P. villosissima PZSH11-042 (26.58 pg). Error bars indicate the standard deviation of flow cytometry samples.

348

a. b.

Figure 7.3: Meiotic metaphase chromosomes of aneuploid F 1 progeny generated from inte rploid interspecific crosses. a. P. floridana x P. pilosa ssp. longipilosa F 1 hybrid with n=2 x-1=13 chromosomes. b. P. drummondii x P. pulcherrima F1 hybrid with n=2 x-2=12 chromosomes. The scale bar (lower portion of image) indicates 10 µm.

349

longipilosa

ssp.

pilosa pilosa 1 F P. P. floridana floridana P. B. A. C. Ladder Gene

Figure 7.4: SRAP banding showing presence of shared alleles between F1 hybrid of P. floridana x P. pilosa ssp. longipilosa PZ11-107 and parental taxa. The primer pair me3- em3 was used for amplification of DNA fragments.

350

Figure 7.5: Electropherogram demonstrating SRAP fragment analysis and banding patterns of the F 1 hybrid P. floridana x P. pilosa ssp. longipilosa PZ11 -107 and parental taxa on an ABI 3100 gene tic analyzer. The primer pair me3 -em3 was used for amplification of DNA fragments, and the green circles correspond to shared alleles between the hybrid and parental taxa (Figure 7. 4).

351

Subsection Phlox taxa Phlox divaricata Phlox Phlox pilosa ssp. Phlox pilosa ssp. Phlox Characterz Phlox floridana Phlox villosissima ssp. laphamii drummondii pilosa longipilosa pulcherrima

Life History Perennial Annual Perennial Perennial Perennial Perennial Perennial Stem Height 15-30 cm 15-30 cm 15-30 cm 20-40 cm 20-40 cm 40-60 cm 40-60 cm linear to linear to linear to Leaves elliptic Oblanceolate Linear linear to lanceolate lanceolate lanceolate lanceolate

Glabrous, Glabrescent to Glabrescent to Glabrous to Glabrescent to Leaf Vesture Glabrescent Glabrous Glabrescent, Glabrous Hirsute glabrescent pilose Pilose Leaf Margin Glabrescent Glabresecent Glabrous Pilose densely Pilose Ciliate Pilose Inflorescence Glandular Glandular Glandular Glabrous to pubescent- Pubescent Densely Pilose Vesture pubecent pubescent pubescent glabresecent glandular pubescent Corolla Color Lavender-purple Red Pink Pink-white Pink Pink Pink Corolla tube Glandular Glabrous Glabrous Pilose Pilose Pilose to glabrous Pubescent vesture pubescent Habitat Mesic/Hydric Xeric Xeric Mesic/Xeric Xeric Mesic/Xeric Xeric Geographic Southwestern Edwards Plateau, Eastern U.S. Texas Gulf Coast Eastern U.S. Eastern Texas distribution Oklahoma Central Texas zMorphological characters derived from the keys of Wherry (1935, 1955) and from filed and common garden observation at the Ornamental Plant Germplasm Center

Table 7.1: Distinguishing morphological characteristics, habitat preferences, and geographic distribution of Phlox subsection Divaricatae taxa used in this study.

352

Genome size Taxon Accession (no.) Collection Site Ploidyy (pg)z

Phlox divaricata L. ssp. laphamii (Wood) Wherry PZSH2011-018 Wilcox Co. AL 10.49±0.09 2x

P. divaricata ssp. laphamii PZSH11-022 Jasper Co. MS 11.03±0.03 2x

P. pilosa L. ssp. longipilosa (Waterfall) Locklear PZSH11-043 Greer Co. OK 12.55±0.23 2x

P. pilosa L. ssp. pilosa PZSH11-020 Scott Co. MS 12.40±0.40 2x P.drummondii Hooker ssp. drummondii Wherry PZ10-161 Caldwell Co. TX 12.22±0.10 2x

P. xglutinosa (Phlox divaricata ssp. laphamii PZSH2011-022 PZ11-127 Artificial Hybrid 10.90±0.09 2x x Phlox pilosa ssp. pilosa PZSH2011-020)

P. floridana Bentham PZSH11-010 Jackson Co. FL 21.85±0.30 4x

P. pulcherrima (Lundell) Lundell PZSH11-034 Shelby Co. TX 22.05±0.07 4x P. villosissima (A. Gray) Small PZSH11-042 San Saba Co. TX 26.58±1.56 4x zGenome size is the holoploid (2C) value in picograms and the standard deviation (Chapter 4). yPloidy of these taxa discussed in Chapter 4.

Table 7.2: Specific germplasm collections, internal accession numbers, collection sites, relative genome sizes, and ploidy levels of Phlox taxa used in interploid interspecific hybridization experiments.

353

Parental Combinationz

Seeds obtained Seeds germinated Female Parent (♀) Male Parent (♂) Cross Ploidy Germination (%)x Accession No. (No.)y (No.)

P . pulcherrima PZSH2011-034 P. villosissima PZSH11-042 4 x 4 23 11 47.8 PZ11-211 P. villosissima PZSH2011-042 P . pulcherrima PZSH2011-034 4 x 4 0 −−− −−− −−−

P. divaricata ssp. laphamii PZSH2011-032 P. villosissima PZSH11-042 2 x 4 0 −−− −−− −−− P. villosissima PZSH11-022 P. divaricata ssp. laphamii PZSH2011-022 4 x 2 28 5 17.8 11-210

P. divaricata ssp. laphamii PZSH11-032 P. villosissima PZSH11-042 2 x 4 0 −−− −−− −−− P. villosissima PZSH2011-042 P. divaricata ssp. laphamii PZSH11-032 4 x 2 42 26 61.9 11-202

P. drummondii ssp. drummondii PZ10-161 P . pulcherrima PZSH2011-034 2 x 4 46 46 100 11-175 P . pulcherrima PZSH2011-034 P. drummondii ssp. drummondii PZ10-161 4 x 2 0 −−− −−− −−−

P. floridana PZSH2011-010 P. pilosa ssp. longipilosa PZSH2011-043 4 x 2 30 29 96.7 11-107 P. pilosa ssp. longipilosa PZSH2011-043 P. floridana PZSH2011-010 2 x 4 0 −−− −−− −−−

P. xglutinosa PZ11-127 P . pulcherrima PZSH2011-034 2 x 4 2 (PDw) −−− −−− −−− P . pulcherrima PZSH2011-034 P. xglutinosa PZ11-127 4 x 2 5 3 60 PZ11-215

P. xglutinosa PZ11-127 P. villosissima PZSH2011-042 2 x 4 0 −−− −−− −−− P. villosissima PZSH2011-042 P. xglutinosa PZ11-127 4 x 2 12 4 33.3 PZ11-212 z50 pollination per cross were performed over a period of 3-7 days, not all of any given cross was performed on a single day ySeed was collected by bagging the inflorescence 3-4 weeks after completion of crosses and allowing the seeds to fall when mature xGermination (%) calculated as the number of seeds that germinated and produced a seedling by the total number of seeds wPD= formation of partially developed seeds. Pollination was successful but seed development arrested before maturity.

Table 7.3: Parental combinations, cross ploidy, seeds obtained (No.), and germination (%) and internal accession numbers of successful and unsuccessful interploid interspecific crosses.

354

BIBLIOGRAPHY

Alabama Plant Atlas Editorial Committee. 2014. Alabama Plant Atlas. [S.M. Landry and K.N. Campbell (original application development), Florida Center for Community Design and Research. University of South Florida]. University of West Alabama, Livingston, Alabama.

Allison, J. 2010. A botanical lost world in Bibb county Alabama. Web. 12 Jan. 2010.

Amason, C. 2000. Some notes on Phlox pilosa. Louisiana Native Plant Society, Spring issue. Pg. 7.

Arends, J. 1912. Proc. Royal Hort. Soc. Fig. 114. 38: 151.

Arumuganathan, K., and E.D. Earle. 1991. Estimation of nuclear DNA content of plants by flow cytometry. Plant Mol. Biol. Rep. 9: 229–241.

Balao, F., Casimiro-Soriguer, R., Talavera, M., Herrera, J., and S. Talavera. 2009. Distribution and diversity of cytotypes in Dianthus broteri as evidenced by genome size variations. Ann. Bot. 104: 965–973.

Barba-Gonzalez, R., Lim, K.B., Zhou, S., Ramanna, M.S., and J.M. van Tuyl. 2008. Interspecific hybridization in lily: The use of 2n gametes in interspecific lily hybrids. Floriculture, ornamental and plant biotechnology Volume 5. Global Science Books, London, UK.

Baskin, J.M. and C.C. Baskin. 1988. Endemism in rock outcrop plant communities of unglaciated eastern United States: an evaluation of the roles of the edaphic, genetic, and light factors. J. Biogeogr. 15: 829-840.

Baskin, J.M. and C.M. Baskin. 2004. A classification system for seed dormancy. Seed Sci. Res. 14: 1-16.

Baskin, C.C. and J.M. Baskin. 2005. Seed dormancy in wild flowers, p. 163–185. In: McDonald, M.B. and F.Y. Kwong (eds.). Flower seeds: Biology and technology. CABI Publishing, Cambridge, MA.

Bendtsen, B.H. 2009. Phloxe fur den garten. 256 pages. Forlaget Geranium Verlag, Denmark. 355

Bennett, M.D. and Leitch, I.J. 2010. Plant DNA C-values Database (Release 5.0, Dec. 2010).

Bennett, M.D. and Leitch I.J. 2012. Plant DNA C-values Database (Release 6.0, Dec. 2012).

Bennetzen, M.D. and I.J. Leitch. 2005. Plant genome size research: A field in focus. Ann. Bot. 95: 1-6.

Bentham, G. 1845. Polemoniaceae. p. 302-322. In A.P. de Candolle (ed.), Prodomus systematis naturalis regni vegetabilis. Paris, France.

Bir, Richard. 1999. Phlox without fail. Organic Gardening 46: 52-55.

Bir, Richard E. 2003. Phlox Get Humidity Test. Am. Gardener: 82: 19.

Booth, J. 2008. Phlox floridana research grant 2007 final report. Florida Wildflower Foundation. 10 January 2011.http://flawildflowers.org/resources/pdfs/REPORTS- 07/R-004-07%20Phlox%20Floridana%20Research.pdf.

Broderick, S.R., Stevens, M.R., Geary, B., Love, S.L., Jellen, E.N., Dockter, R.B., Daley, S.L., Lindgren, D.T. 2011. A survey of Pensemon’s genome size. Genome 54: 160-173.

Brooks, R. R. 1987. Serpentine and its Vegetation: a Multi-disciplinary Approach. Dioscorides Press, Portland, OR.

Budak, H., Shearman, R.C., Parmaksiz, I., Gaussoin, R.E., Riordan, T.P., and I. Dweikat. 2004. Molecular characterization of buffalograss germplasm using sequence related amplified polymorphism markers. Theor. Appl. Genet. 108: 328-334.

Burton, T. L., and B. C. Husband. 1999. Population cytotype structure in the polyploid Galax urceolata (Diapensiaceae). Heredity 82: 381–390.

Buthod, A.K. and J. J. Skvarla. 2014. Pollen morphology of the Oklahoma endemic plants Leavenworthia aurea (Brassicaceae/Cruciferae) and Phlox pilosa ssp. longipilosa (Polemoniaceae), with special reference to their natural history. Rhodora. In-Press.

Campbell, J.J.N. 2012. The atlas of the flora of Kentucky. 485 pp. Lexington, KY. http://www.bluegrasswoodland.com/uploads/Index-Alphabetic.pdf.

Carputo, D., Monti, L., Werner, J.E., and L. Frusciante. 1999. Uses and usefulness of endosperm balance number. Theor. Appl. Genet. 98: 478-484.

356

Ceccarelli, M., Sarri, V., Caceres, M.E., and P.G. Cionini. 2011. Intraspecific genotypic diversity in plants. Genome 54:701-709.

Cires, E, Cuesta, C., Peredo, E.L., Revilla, M.A., and J.A.F. Prieto. 2009. Genome size and morphological differentiation within Ranunculus parnassifolius group (Ranunculaceae) from calcareous screes in the northwest of Spain. Plant Syst. Evol. 281: 193-208.

Cisneros, A., and N. Tel-Zur 2012. Evaluation of Interspecific-Interploid Hybrids (F1) and Back Crosses (BC1). In: A. Swan (Ed.) Hylocereus Species (Cactaceae), Meiosis - Molecular Mechanisms and Cytogenetic Diversity. ISBN: 978-953-51- 0118-5, InTech, DOI: 10.5772/32435. http://www.intechopen.com/books/meiosis- molecular-mechanisms-and-cytogenetic-diversity/evaluation-of-interspecific- interploid-hybrids-f1-and-back-crosses-bc1-in-hylocereus-species-cactace

Clay, K., and D. A. Levin. 1989. Quantitative variation in Phlox: Comparison of selfing and outcrossing species. Am. J. Bot. 76: 577-588.

Cohen, R., Hanan, A., and H.S. Paris. 2003. Single-gene resistance to powdery mildew in zucchini squash (Cucurbita pepo). Euphytica 130: 433-441.

Contreras, R.N., Ruter, J.M., and W.W. Hanna. 2009. An oryzalin-induced autoallooctoploid of Hibiscus acetosella ‘Panama Red’. J. Amer. Soc. Hort. Sci. 134: 553-559.

Cooperrider, T.S. 1986. The genus Phlox (Polemoniaceae) in Ohio. Castanea 51: 145- 148.

Czarnecki II, D.M., Rao, M.N., Norcini, J.G., Gmitter Jr., F.G., and Z. Deng. 2008. Genetic diversity and differentiation among natural, production, and introduced populations of the narrowly endemic species Coreopsis leavneworthii (Asteraceae). J. Am. Soc. Hort. Sci. 133: 233-241.

Dar, T.H., Raina, S.N., and S. Goel. 2013. Molecular analysis of genomic changes in synthetic autotetraploid Phlox drummondii Hook. Biol. J. Linnean Soc. 110: 591- 605.

Deam, C.C. 1940. The flora of Indiana. Indianapolis, Indiana. Department of Conservation, Division of Forestry.

Dilkes, B.P., and L. Comai. 2004. A differential dosage hypothesis for parental effects in seed development. Plant Cell 16: 3174-3180.

357

Dolezel, J., Greilhuber, J., Lucretti, S., Meister, A., Lysak, M.A., Nardi, L., and Obermayer, R. 1998. Plant genome size estima- tion by flow cytometry: inter- laboratory comparison. Ann. Bot. 82(Suppl. A): 17–26.

Dolezel, J. and J. Bartos. 2005. Plant DNA flow cytometry and estimation in nuclear genome size. Ann. Bot. 95: 99–110.

Dolezel, J., J. Greilhuber, and J. Suda. 2007. Flow cytometry with plant cells: analysis of genes, chromosomes, and genomes. Wiley-VCH, Weinheim, Germany.

Dolezel, J. 2009. Determination of nuclear genome size. 10 Sept. 2011. http://olomouc.ueb.cas.cz/book/determination- nuclear-genome-size.

Doyle, J.J. and J.L. Doyle. 1987. A rapid DNA isolation procedure from small quantities of fresh leaf tissues. Phytochemistry Bull. 19: 11-15.

Earl, D.A. and B.M. von Holdt. 2012. STRUCTURE HARVESTER: A website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation Genet. Res. 4: 359-361.

Eater, J.W. 1967. A systematic study of subsection Nanae of the genus Phlox. M.A. Thesis, University of California, Santa Barbara.

Edh, K., Widen, B., and A. Ceplitis. 2009. The evolution and diversififcation of s-locus haplotypes in the Brassicaceae family. Genet. 181: 977-984.

Eeckhaut, T., Van Laere, K., De Riek, J., and J. van Huylenbroek. 2006. Overcoming barriers in floriculture, ornamental and plant biotechnology Volume 1 Global science books. Cambridge University Press, Cambridge, UK.

Erbe, L.W. 1960. A biosystematics study of the Phlox cuspidata-Phlox drummondii complex. Ph.D Thesis, University of Texas at Austin.

Erbe, L.W. and B.L. Turner. 1962. A biosystematics study of the Phlox cuspidata-Phlox drummondii complex. Am. Midl. Nat. 67: 257-281.

Evanno, G., Regnaut, S., and J. Goudet. 2005. Detecting the number of clusters of individuals using the STRUCTURE software: A simulation study. Mol. Ecol. 14: 2611-2620.

Fehlberg, S.J., Ford, C.A., Ungerer, M.C., and C.J. Ferguson. 2008. Development, characterization and transferability of microsatellite markers for the plant genus Phlox (Polemoniaceae). Mol. Ecol. Res. 8: 116-118.

358

Fehlberg, S.J. Personnal Communication. 15 January 2011.

Fehlberg, S.D. and C.J. Ferguson. 2012. Intraspecific cytotypic variation and complex genetic structure in the Phlox amablis-P. woodhousei (Polemoniaceae) complex. Am. J. Bot. 99: 865-874.

Ferguson, C.F. 1998. Molecular systematics of the eastern Phlox (Polemoniaceae). PhD Thesis, University of Texas at Austin.

Ferguson, C. J., F. Krämer and R. K. Jansen. 1999. Relationships of eastern North American Phlox (Polemoniaceae) based on ITS sequence data. Systematic Bot. 24: 616-631.

Ferguson, C.J. and R.K. Jansen. 2002. A chloroplast DNA phylogeny of eastern Phlox (Polemoniaceae): Implications of congruence and incongruence with the ITS phylogeny. Am. J. Bot. 89: 1324-1335.

Florida Natural Areas Inventory (FNAI). 2010. Guide to the natural communities of Florida: 2010 edition. Florida Natural Areas Inventory, Tallahassee, FL.

Flory, Jr., W.S. 1931. Chromosome Numbers in Phlox. Am. Nat. 65: 473-476.

Flory, Jr., W.S. 1934. A cytological study on the genus Phlox. Cytologia 6:1–18.

Flory, Jr., W.S. 1937. Chromosome numbers in the Polemoniaceae. Cytologia Fujii Jubilaei 171–180.

Flory, W.S., JR. 1948. The chromosomes of a tetraploid Phlox from the Chisos Mountains. Proc. West Virginia Acad. Sci. 26:85.

Foley, D.J. 1972. Ground covers for easier gardening. USA, Dover Publishing Inc. New York, NY.

Forman, R.T.T., and L.E. Alexander. 1998. Roads and their major ecological effects. Ann Rev. Ecol. Syst. 29: 207-231.

Fuchs, H. 1994. Phlox stauden und polsterphloxe. Eugen Ulmer GmbH and Co. Stuttgart, Germany.

Galbraith, D.W., Harkins, K.R., Maddox, J.M., Ayres, N.M., Sharma, D.P., and Firoozabady, E. 1983. Rapid flow cytometric analysis of the cell cycle in intact plant tissues. Science 220: 1049–1051.

Galbraith, D.W., G.M. Lambert, J. Macas, and J. Dolezel. 1997. Analysis of nuclear DNA and ploidy in higher plants. Curr. Protoc. Cytom. 7.6.1–7.6.22. 359

Grant, V. and K.A. Grant. 1965. Flower pollination in the Phlox family. Columbia University Press, New York and London.

Gray, A. 1870. Revision of the North American Polemoniaceae. Proc. Am. Acad. Arts Sci. 8: 247-282.

Greilhuber, J. 1998. Intraspecific variation in genome size: A critical reassessment. Ann. Bot. 82(Suppl. A): 27–35 doi:10.1006/anbo.1998.0725.

Greilhuber, J. 2005. Intraspecific variation in genome size in angiosperms: identifying its existence. Ann. Bot. 95: 91–98.

Greilhuber, J., Temsch, E.M., and J.C.M. Loureiro. 2007. Nuclear DNA content measurements. In: (eds.) J. Dolezel, J. Greilhuber, J. Suda Flow cytometry with plant cells: Analysis of genes, chromosomes and genome. Wiley-VCH Verlag GmbH and Co. KGaA, Weinheim, Germany pp. 67–101.

Greilhuber, J. Leitch, I.J. 2013. Genome size and the phenotype. In; (ed.) I.J. Leitch Plant Genome Diversity Volume 2 (ed.) pgs. 320-344. Springer-Verlag, Wien.

Grover, C.E. and J. F. Wendel. 2010. Recent insights into the mechanisms of genome size change in plants. J. Bot. 2010: 1-8.

Hadley, E.B. and D.A. Levin. 1969. Physiological Evidence of Hybridization and Reticulate Evolution in Phlox maculata. Am. J. Bot. 56: 561-570.

Hamrick, J.L. and M.J.W. Godt. 1996. Effects of life history traits on genetic diversity in plant species. Philosophical Transactions of the Royal Soc. London Biol. Sci. 351: 1291–1298.

Harlan, J.R., and J.M.J de Wet. 1971. Toward a rational classification of cultivated plants. Taxon 20-509-517.

Hardy OJ, and X. Vekemans. 2002. SPAGeDi: a versatile computer program to analyse spatial genetic structure at the individual or population levels. Mol. Ecol. Notes 2: 618-620.

Hawke, R.G. 1999. Plant evaluation notes: An evaluation report of selected Phlox species and hybrids. Chicago Hort. Soc. 13: 1-4.

Hawke, R.G. 2011. A comparative study of Phlox paniculata cultivars. Plant evaluation Notes. Issue 35 Chicago Botanic Garden pgs. 1-10.

360

Hawke, R.G. 2013. Fabulous Phlox. Greenhouse Management October 29 2013. 2 pgs.

Hawkins, S.M., Ruter, J.M., and C. Robacker. 2013. Seed Set and Germination for Interspecific and Intergeneric Hybrids in Two Genera of Fabaceae. Proc. IPPS Southern Region North America.

Hay, F.R. and R.J. Probert. 2013. Advances in seed conservation of wild species: A review of recent research. Conservation Physiol. Rev. 1: 1-11.

Heikens, A.L. 2003. Conservation Assessment for Broad-leaved Phlox (Phlox amplifolia Britt). USDA Forest Service Bulletin, Eastern Region. Hoosier National Forest 9 pp.

Hendrix, S.D., and J.F. Kyhl. 2000. Population size and reproduction in Phlox pilosa Conservation Biol. 14: 304-313.

Hereford, J. 2010. Does selfing or outcrossing promote local adaptation? Am. J. Bot. 2: 298-302.

Heywood, V. 2003. Conservation and sustainable use of wild species as sources of new ornamentals. In: (ed.) E. Düzyaman andY.Tüzel. Sustainable use of plant biodiversity. Acta Hort. 598, ISHS 2003.

Hoagland, B. Personnal Communication. 15 February 2011.

Hogenboom, N.G., and K. Mather. 1975. Incompatibility and incongruity; Two different mechanisms for the non-functioning of intimate partner relationships (and comments). Proc. R. Soc. Lond. 188: 361-375.

Hufford, K.M. and S.J. Mazer. 2003. Plant Ecotypes: genetic differentiation in the age of ecological restoration. Trends Ecol. Evol. 18: 147-155.

Indiana Division of Nature Preserves (IDNR). 2012. Annual report. Indianapolis, Indiana.

Iverson, R.R. and T.C. Weiler. 1994. Strategies to force flowering of six herbaceous garden perennials. HortTechnology 4: 61-65.

Jakobsson, M. and N.A. Rosenberg. 2007. CLUMPP: A cluster matching and permutation program from dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23: 1801-1806.

361

Johnston, S.A., den Nijs, T.P.M., Peloquin, S.J., and R.E. Hannon, Jr. 1980. The significance to genic balance to endosperm development in interspecific crosses. Theor. Appl. Genet. 57: 5-9.

Jorosz, A.M., Sheets, M., and M. Levy. 1982. Cuticle thickness in Phlox and resistance to powdery mildew: An unreliable line of defense. Am. J. Bot. 69: 824-828.

Kato, J. and M. Mii. 2012. Production of interspecific hybrids in plants. Methods Mol. Biol. 877: 233-245.

Kelly, J.P. 1915. The cultivated varieties of Phlox drummondii. J. N.Y. Bot. Gard. 16: 179-191.

Kelly, J.R. 1944. Chromosome studies of Phlox. Genet. 29: 199-216.

Kessler, S.A., Shimosato-Asano, H., Keinath, N.F., Wuest, S.E., Ingram, G., Panstruga, R. and U. Grossniklaus. 2010. Conserved molecular components for pollen tube reception and fungal infection. Science 330: 968-971.

Klinkhamer, L. Personnal Communication. 14 May 2014.

Kolar, F., M. Štech, P. TravvnÍicek, J. rauchova, T. Urfus, P. VÍt, M. Kubesova, and J. Suda. 2009. Towards resolving the Knautia arvensis agg. (Dipsacaeae) puzzle: Primary and secondary contact zones and ploidy segregation at landscape and microgeographic scales. Ann. Bot. 103: 963–974.

Kritzman, E. B. 1974. Ecological relationships of Peromyscus maniculatus and Perognathus parvus in eastern . J. Mammalogy. 55: 172-188.

Kron, P., and B.C. Husband. 2012. Using flow cytometry to estmate pollen DNA content: Improved methodology and applications. Annals Bot. 110: 1067-1078.

Lehmann. 1828. Phlox procumbens. Sem. Bot. Hamburg. 17.

Leitch, I.J. Greilhuber, J., Dolezel, J., and J Wendel (eds.) 2013. Plan genome diversity Vol.2: Physical behavior and evolution of plant genomes. Springer-Verlaug, Wien.

Leitch I.J., and A.R. Leitch. 2013. Genome size diversity and evolution in land plants. In: (eds.) I.J. Leitch, J. Greilhuber, J. Dolezel, J. F. Wendel. Plant genome diversity,vol 2. Physical structure, behaviour and evolution of plant genomes. Vienna: Springer-Verlag, 307–322.

362

Leus, L., Eeckhaut, T., Dhooghe, E., Van Labeke, M.C., Van Laere, K. and Van Huylenbroeck, J. 2012. Polyploidy breeding in vitro: Experiences with ornamentals. Acta Hort. (ISHS) 961:235238 http://www.actahort.org/books/961/961_29.htm.

Levin, D.A. 1963. Natural hybridization between Phlox maculata and Phlox glaberrima and its evolutionary significance. Am. J. Bot. 50: 724-729.

Levin, D.A. 1964. Variation and evolution in Phlox subsection Divaricatae. Ph.D. dissertation, University of Illinois, Urbana.

Levin, D.A. 1966. The Phlox pilosa complex: crossing and chromosome relationships. Britt 18: 142-162.

Levin, D.A. 1967. Variation in Phlox divaricata. Evolution. 21: 92-108.

Levin, D.A. 1968. The genome constitutions of eastern North American Phlox amphiploids. Evolution 22: 612–632.

Levin, D.A. 1973. Polymorphism for interspecific cross-compatibility in Phlox. Proc. Nat. Acad. Sci. 70: 1149-1150.

Levin, D.A. 1975. Interspecific Hybridization, Heterozygosity and Gene Exchange in Phlox. Evolution 29: 37-51.

Levin, D. A. 1976a. The consequences of long-term artificial selection, inbreeding, and isolation in Phlox. I. The evolution of cross‐incompatibility. Evolution 30: 335- 344.

Levin, D. A. 1976b. The consequences of long‐term artificial selection, inbreeding, and isolation in Phlox. II. The organization of allozymic variation. Evolution 30: 463- 472.

Levin, D.A. 1977. The organization of genetic variability in Phlox drummondii. Evolution 31: 477-479.

Levin, D.A. 1978. Genetic variation in annual Phlox: self-compatible vs. self- incompatible species. Evolution 32-245-263.

Levin, D.A. 1983. Polyloidy and novelty in flowering plants. Am. Nat. 122: 1-25.

Levin, D.A. 1993. S-gene polymorphism in Phlox drummondii. Heredity 7: 193-198.

363

Levin, D. A. 2002. The role of chromosomal change in plant evolution. Oxford University Press, New York.

Levin, D. A. and H. W. Kerster. 1967. Natural selection for reproductive isolation in Phlox. Evol. 21: 679‐687.

Levin, D. A. and H. W. Kerster. 1968. Local gene dispersal in Phlox. Evol. 22: 130‐139.

Levin, D. A. and H. W. Kerster. 1975. The effects of gene dispersal on the statics and dynamics of gene substitutions in plants. Heredity 35: 317-336.

Levin, D.A and D.M. Smith. 1965. An enigmatic Phlox from Illinois. Britt. 17: 254-266

Levin, D. A. and D. M. Smith. 1966. Hybridization and evolution in the Phlox pilosa complex. Amer. Natur. 100:289‐302.

Levin, D.A., Torres, A.M., and M. Levy. 1979. Alcohol dehydrogenase in diploid and autotetraploid Phlox. Biochemical Genetics 17: 35-42.

Levy, M. 1983. Flavone variation and subspecific variation in Phlox pilosa (Polemoniaceae). Systematic Bot. 8: 118-126. Levy, M. and D.A. Levin. 1974. Novel flavonoids and reticulate evolution in the Phlox pilosa-P. drummondii complex. Am. J. Bot. 65: 156-167. Li, G., and C.F. Quiros. 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. Theor. Appl. Genet. 103: 455-461. Lierval, M. 1866. Culture Pratique des Phlox, par Lierval. E. Dounnaud, Paris, France.

Liu, L., Liu, Z., Chen, H., Zhou, L., Liu, Y., and L. Luo, 2012. SRAP markers and morphological traits that could be used in test of distinctiveness, uniformity, and stability (DUS) of lettuce (Lactuca sativa) varieties. J. Ag. Sci. 4: 227-236.

Locklear, J. H. 2009. Nomenclatural innovations in Phlox (Polemoniaceae), with updated circumscription of P. caespitosa, P. douglasii, P. missoulensis, and P. richardsonii. J. Bot. Res. Inst. Texas 3: 645-658.

Locklear, J.H. 2011a. Phlox: A natural history and gardener’s guide. Timber Press, Portland. 284 pp.

364

Locklear, J.H. 2011b. Phlox ovata L. (Polemoniaceae): Clarification of the nomenclature of the Allegheny phlox. Castanea. 76: 116-117.

Lookingbill, T.R., Engelhardt, K.A.M., Florkowski, L.N., Churchill, J.B., and L.J. Ashley. 2007. Evaluation of the Nottingham Park serpentine barrens. University of Maryland Center for Environmental Science 54pp.

Madeiras, A.M., Boyle, D.H., and W.R. Autio. 2007. Germination of Phlox pilosa L. seeds is improved by gibberellic acid and light, but not stratification, light, or surface disinfestation. Hortscience 42: 1263-1267.

Marsh, D.L. 1960. Relationship of Phlox oklahomensis to the Phlox bifida complex: Including a new subspecies of Phlox bifida. Trans. Kans. Acad. Sci. 63: 12-19.

Matiska, P., and H. Vejsadova. 2010. Polyploidy induction in Phlox paniculata under in vitro conditions. Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis 58: 101-106.

Meyer, J.R. 1944. Chromosome studies of Phlox. Genetics 29: 199-216.

Michaux, A. 1803. Flora Boreali-Americana, Sistens Characteres Plantarum. 2 vols. Paris and Strasbourg.

Michigan Flora. 2014. University of Michigan Herbarium. Accessed February 2014. http://michiganflora.net/species.aspx?id=2253

Minnesota Dept. Natural Resources. 2014. Accessed February 2014. http://www.dnr.state.mn.us/rsg/profile.html?action=elementDetail&selectedElem ent=IILEYMP130.

Mishra, K. M., Suresh, N., Baht, A.M., Suryaprakash, N., Kumar, S.S., and A.K. Jayarama. 2011. Genetic molecular analysis of Coffea arabica (Rubiaceae) hybrids using SRAP markers. Rev. Biol. Trop. 59: 607-616.

Morgan, E. R., Timmerman-Vaughan, G. M., Conner, A. J., Griffin, W. B. and R. Pickering. 2011. Plant Interspecific Hybridization: Outcomes and Issues at the Intersection of Species, in Plant Breeding Reviews, Volume 34 (ed.) J. Janick, John Wiley and Sons, Inc., Hoboken, NJ, USA.

Oliver, C. 2011. Primrose Path Nursery, Phlox: An exploration. Web. Accessed 2011- 2014. http://www.theprimrosepath.com/phlox/introduction.htm.

Oliver, C. 2012. Personnal communication. 10 June 2012.

365

Olsen, Richard T., T.G. Ranney, and Z. Viloria. 2006. Reproductive behavior of induced allotetraploid ×Chitalpa and in vitro embryo culture of polyploidy progeny. J. Amer. Soc. Hort. Sci. 131:716-724.

Otteson, C. 2003. Heuchera explosion. Am. Gardener Pgs. 42-47.

Ozminkowski, Jr., R.H. and P. Jourdan. 1994. Comparing the resynthesis of Brassica napus L. by interspecific somatic and sexual hybridization. I producing and identifying hybrids. J. Amer. Soc. Hort. Sci. 119: 808-815.

Palmer, I.E., Ranney, T.G., Lynch, N.P., and R.E., Bir. 2008. Crossability, cytogenetics, and reproductive pathways in Rudbeckia subgenus Rudbeckia. Hortscience 44: 44-48.

Pan, G., Zhou, Y, Fowke, L.C., and H. Wang. 2004. An efficient method for flow cytometric analysis of pollen and detection of 2n nuclei in Brassica pollen. Plant Cell Rep. 23: 196-202.

Parisod, C., Holderegger, R., and C. Brochmann. 2010. Evolutionary consequences of autopolyploidy. New Phytologist 186: 1-17

Parris, J.K., Ranney, T.G., Knap, H.T., and W.V. Baird. 2010. Ploidy levels, genomes sizes, and base pair composition in Magnolia. J. Am. Soc. Hort. Sci. 135: 533- 547.

Peakall, R. and Smouse P.E. 2012. GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research-an update. Bioinformatics 28: 2537- 2539.

Perry, L.P., and S.A. Adam Jr. 1994. ‘David’ Phlox. Hortsci. 29: 713.

Plant Delights Nursery. 2014. Nursery Sales Catalog. Web. Accessed 2010-2014.

Plant DNA C-values database release 6.0 December 2012. http://data.kew.org/cvalues/.

Prather, A.L. 1994. A new species of Phlox (Polemoniaceae) from Mexico with and expanded circumscription of subsection Divaricatae. Pl. Syst. Evol. 192: 61-66.

Prather, A. Personal communication, 11 October 2012.

Pridham, A.M.S. 1934. History, culture, and varieties of summer-flowering phloxes. Bulletin No. 588 Cornell University. Agriculture Experiment Station.

366

Raghuvanshi, S.S., Pathak, C.S. 1975. Polyploid Breeding and Possibility of Raising Double Varieties in Phlox drummondii, Hook Cytologia 40: 355-363.

Ranney, T.G. 2006. Polyploidy: From evolution to new plant development. Proc. Intl. Plant Prop. Soc. 56: 137–142.

Raven, P.R. 1964. Catastrophic selection and edaphic endemism. Evo. 18: 336-338.

Renze, E. 2013. Personnal communication. 10 June 2013.

Reveal, J.L., C.R. Broome, M.L. Brown, and G.F. Frick. 1982. Comments on the typification of two Linnaean species of Phlox (Polemoniaceae). Taxon 31: 733- 736.

Ridout, M.E., and R.R. Tripepi. 2009. Improving seed germination of native perennial Phlox longifolia. Native Plants Journal 10: 80-88.

Royal Horticulture Society (RHS). 2014. Plant Locator. Accessed 2014. https://www.rhs.org.uk/plants/search-form

Rieseberg, L. H., and Willis, J. H. 2007. Plant Speciation. Science 317: 910–914.

Robarts, D.W.H. 2013. The phytogeography of Viola pedata L. (Violaceae). Ph.D dissertation. The Ohio State University.

Rosenberg, N.A. 2004. Distruct: a program for the graphical display of population structure. Mol. Ecol. Notes 4: 137-138.

Ruane, L.G. and K. Donohue. 2007. Pollen Competition and Environmental Effects on Hybridization Dynamics Between Phlox drummondii and Phlox cuspidata. Evol. Ecol. 2007: 229–241.

Ruane, L.G. 2009. Mating system and hybridization between self-compatible Phlox cuspidata and self-incompatible Phlox drummondii. Evol. Ecol. 23:791-805.

Ruane, L.G., Rotzin, A.T., and P.H. Congleton. 2013. Floral display size, conspecific density, and florivory affect fruit set in natural populations of Phlox hirsuta, an endangered species. Ann. Bot. 113: 887-893.

Sabara, H.A., Kron, P., and B.C. Husband. 2013. Cytotype coexistence leads to triploid hybrid production in a diploid-tetraploid contact zone of Chamerion angustifolium (Onagraceae). Am J. Bot. 100: 962-970.

367

Sampson, J. F. and M. Byrne. 2012. Genetic diversity and multiple origins of polyploid Atriplex nummularia Lindl. (Chenopodiaceae). Biol. J. Linnean Soc. 105: 218– 230.

Schoellhorn, R. and Richardson, A.A. 2005. Warm Climate Production Guidelines for Echinacea. University of Florida IFAS Extension Bulletin.

Schoellhorn, R. 2005. A state of Phlox. Greenhouse Product news January 2005. http://www.gpnmag.com/sites/default/files/p%2018vegetative%20matters.pdf.

Schwaigerle, K. E., and D. A. Levin. 1990. Environmental effects on growth and reproduction in Phlox drummondii. J. Ecol. 78: 15‐26.

Shaw, J., Lickey, E.B., Schilling, E.E., and R.L. Small. 2005. Comparison of whole chloroplast genome sequencing to choose noncoding regions for phylogenetic studies in angiosperms: The tortoise and the hare III. Am. J. Bot. 97: 274-288.

Small, R.L., Ryburn, J.A., Cronn, R.C., Seelanan, T. and J.F. Wendel. 1998. The tortoise and the hare: Choosing between noncoding plastome and nuclear Adh sequences for phylogeny reconstruction in a recently diverged plant group. Am. J. Bot. 85: 1301-1315.

Smarda, P., and P. Bures. 2010. Understanding intraspecific variation in genome size in plants. Preslia 82: 41-61.

Smith, D.M. and D.A. Levin. 1967. Karyotypes of Eastern North American Phlox. Am. J. Bot. 54: 324-334.

Soltis, D. E. Morris, A. B., Lachlan, J. S., Manos, P. S., and P.S. Soltis. 2006 Comparative phylogeography of unglaciated eastern North America. Mol. Ecol. 15: 4261–4293.

Soltis, D.E., Albert, A.A., Leebens-Mack, J., Bell, C.D., Paterson, A.H., Zheng, C., Sankoff, D., DePamphilis, C.W., Wall, P.K., and P.S. Soltis. 2009. Polyploidy and angiosperm diversification. Amer. J. Bot. 96: 336-348.

Strakosh, S.C. 2004. Systematic studies in Phlox (Polemoniaceae) with a focus on P. dolichantha, P. superba, P. stansburyi, and P. grayi. M.S. thesis, Kansas State University, Manhattan.

Springer, T.L., and R.J. Tyrl. 1989. Distribution, habitat, and reproductive biology of Phlox oklahomensis Wherry (Polemoniaceae). Proc. Okla. Acad. Sci. 69: 15-21.

368

Springer, T.L., and R.J. Tyrl. 2003. Status of Phlox oklahomensis (Polemoniaceae) in North-western Oklahoma and adjacent Kansas: Assessment 20 years later. Proc. Okla. Acad. Sci. 83: 89-92.

Stamp, N.E. and J.R. Lucas. 1983. Ecological correlates of explosive seed dispersal. Oecologia 59; 272-278.

Symons-Jeune, B.H.B. 1953. Phlox: A Flower Monograph. D. Van Nostrand, Inc.

Tay, D. 2003. Herbaceous ornamental plant genebank: Its roles in the floriculture industry. Acta Hort. 624: 29-36.

Tay, D. 2005. Ornamental plant genetic resources conservation and utilization in CBD era Acta Hort. 683: 233-242.

Tay, D. 2006. Herbaceous ornamental plant germplam conservation and use; Theoretical and practical use (pgs.113-176). In: N.O. Anderson (ed.) Flower breeding and genetics: Issues, challenges and opportunities for the 21st century. Springer Verlag, Netherlands.

Taylor, E., Cartwright, R., Robbins, J., Klingiman, G., and J. Lindstrom. 2002. Evaluation of Phlox paniculata L. cultivars for susceptibility to powdery mildew. Southern Nurserymans Assoc. research conf. proceedings vol. 47.

Theodoridis, S., Randin, C., Broenninmann, M., Patsiou, T., Conti, E. 2013. Divergent and narrower climatic niches characterize polyploid species of European primroses in Primula sect. Aleuritia. J. Biogeogr. 40: 1278-1289.

Tiwari, A.K., Mishra S.K. 2012. Effect of colchicine on mitotic polyploidization and morphological characteristics of Phlox drummondii. African Journal of Biotechnology 11: 9336-9342.

Tobutt., K.R. 1993. Inheritance of white flower color and congested growth in certain Buddleia progenies. Euphytica 67: 231-235.

Turner, B.L. 1998. Atlas of the Texas species of Phlox (Polemoniaceae). Phytologia 85: 309-326.

United States Department of Agriculture Plants. 2014. Web. http://plants.usda.gov/core/profile?symbol=PHPI.

USDA-NASS. 2010. Statistics of fruits, tree nuts, and horticultural specialties. 2010 Agricultural statistics annual. http://www.nass.usda.gov/Publications/Ag_Statistics/2010/. 369

Van den Engh, G., Trask, B., Cram, S., and K. Bartholdi. 1984. Preparation of chromosome suspensions for flow cytometry. Cytometry 5: 108-117.

Van Tuyl, J.M. and K.B. Lim. 2003. Interspecific hybridization and polyploidisation as tools in ornamental plant breeding. Acta Hort. 612: 13-22.

Van Tuyl, J.M., and M.J. De Jeu. 2005. Methods for overcoming interspecific crossing barriers. (eds.) V. K. Sawhney and K. R. Shivanna. In: Pollen Biotechnology for Crop Production and Improvement. Cambridge University Press, Cambridge, UK.

Vyas, P., Bisht, M.S., Miyazawa, S., Yano, S., Noguchi, K., Terashima, and S. Funayama-Noguchi. 2007. Effects of polyploidy on photosynthetic properties and anatomy in leaves of Phlox drummondii. Functional Plant Biol. 34: 673-682.

Virginia Department of Conservation and Recreation. Summer 2010. E-newsletter 7 pp. http://www.dcr.virginia.gov/natural_heritage/documents/enewssum10.pdf.

Weiss, H., and Malyuszynska, J. 2000. Chromosomal rearrangement in autotetraploid plants of Arabidopsis thaliana. Hereditas 133: 255-261.

Weiss-Schneeweiss, H., Schneeweiss, G. M., Stuessy, T. F., Mabuchi, T., Park, J.M., Jang, C.G. and B.Y. Sun. 2007. Chromosomal stasis in diploids contrasts with genome restructuring in auto- and allopolyploid taxa of Hepatica (Ranunculaceae). New Phytologist, 174: 669–682.

Wherry, E.T. 1929. The eastern subulate-leaved phloxes. Bartonia 11: 5-35.

Wherry, E.T. 1930. A long lost Phlox. J. Wash. Acad. Sci. 20: 25-28.

Wherry, E.T. 1932a. The eastern long-styled phloxes, part 1. Bartonia 13: 18-37.

Wherry, E.T. 1932b. The eastern long-styled phloxes, part 2. Bartonia 14: 14-26.

Wherry, E.T. 1933. The eastern veiny-leaved phloxes. Bartonia. 15:14-26.

Wherry, E.T. 1935a. Our native phloxes and their horticultural derivatives. Nat. Hort. Mag. 14: 209-231.

Wherry, E.T. 1935b. A new variety of Phlox ovata from the Alabama mountains. Bartonia 16:37-39.

Wherry, E.T. 1943a. Microsteris, Phlox, and an intermediate. Brittonia 5: 60-63.

370

Wherry, E.T. 1943b. Variation in Phlox floridana. Bartonia 22: 1-2.

Wherry, E.T. 1945. The Phlox carolina complex. Bartonia 23: 1-9

Wherry, E.T. 1953. Shale-barren plants on other geological formations. Castanea 18: 64- 65.

Wherry, E.T. 1955. The Genus Phlox. Morris Aboretum Monographs Philadelphia, PA.

Whiteside, Wesley. Personal Communication. 1 May 2011.

Widrlechner, M.P., and K.A. McKeown. 2002. Assembling and characterizing a comprehensive Echinacea germplasm collection. p. 506–508. In: J. Janick and A. Whipkey (eds.), Trends in new crops and new uses. ASHS Press, Alexandria, VA.

Wiggam, S., and C.J. Ferguson. 2005. Pollinator importance and temporal variation in a population of Phlox divaricata L. (Polemoniaceae). Am. Mid. Nat. 154: 42-54.

Wilkins, H. and Anderson, N.O. 2006. Creation of new floral products; Annualization of perennials – Horticultural and commercial significance. (ed.) N.O. Anderson. In: Flower breeding and genetics: Issues, challenges and opportunities for the 21st century pgs.49-64 Springer, Netherlands.

Wink, M. 1988. Plant breeding: Importance of plant secondary metabolites for protection against pathogens and herbivores. Theor. Appl. Gent. 75: 225-233.

Wisconsin Dept. Natural Resources. 2014 http://dnr.wi.gov/topic/EndangeredResources/Animals.asp?mode=detail&SpecCo de=IILEYMP130&Type=Description.

Worcester, L., Mayfield, M.H., and C.J. Ferguson. 2012. Cytotypic variation in Phlox pilosa ssp. pilosa (Polemoniaceae) at the western edge of its range in the central United States. J. Bot. Res. Inst. Texas. 6: 443-451.

Wright, B., and C.J. Ferguson. 2014. Polyploidy in Phlox nana Nutt. (Polemoniaceae): documentation of cytotypic variation, and relationship to geography and taxonomic recognition. Botany 2013. http://2013.botanyconference.org/engine/search/index.php?func=detail&aid=433>

Wright, S. 1969. Evolution and the Genetics of Populations: The Theory of Gene Frequencies. The University of Chicago Press, Chicago, Illinois.

Wright, S. 1977. Evolution and the Genetics of Populations: Experimental results and evolutionary deducations. The University of Chicago Press, Chicago, Illinois.

371

Zale, P.J. Personnal observation. 13 April 2011. 18 June 2012. 11 July 2012.

Zhang, Z., Dai, H., Xiao, M., and X. Liu. 2008. In vitro induction of tetraploids in Phlox subulata L. Euphytica 159: 59-65.

Zhou, S. Barba-Gonzalez, R., Lim, K.B., Ramanna, M.S., and J.M. Van Tuyl. Floriculture, Ornamental, and Plant Biotechnology. Global Science Books, United Kingdom Chapter 15 volume 4: 152-156.

Zimmer, E.A., and J. Wen. 2012. Using nuclear gene data for plant phylogenetics: Progress and prospects. Mol. Phyl. Evol. 65: 774-785

372