EMPHASIZES FRAGARIA VESCA AS A BRIDGE SPEClES

Transcription

1 A STREAMLINED SYNTHETlC OCTOPLOID SYSTEM THAT EMPHASIZES FRAGARIA VESCA AS A BRIDGE SPEClES A Thesis Presented to The Faculty of Graduate Studies of The University of Guelph by ROBERT HAROLD BORS In partial fuifilment of requirements for the degree of Doctor of Philosophy Robert Bors O June, 2000

2 l*m National Library of Canada Acquisitions and Bibliographie Services Bibliothèque nationale du Canada Acquisitions et services bibliographiques 395 Wellington Street 395, nie Wellington OttawaON KIAON4 Ottawa ON K1 A ON4 Canada Canada The author has granted a nonexclusive licence allowing the National Library of Canada to reproduce, loan, distribute or sel1 copies of this thesis in microform, paper or electronic formats. The author retains ownership of the copyright in this thesis. Neither the thesis nor substantial extracts fiom it may be printed or otherwise reproduced without the author's permission. L'auteur a accordé une licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, distribuer ou vendre des copies de cette thèse sous la forme de microfiche/film, de reproduction sur papier ou sur format électronique. L'auteur conserve la propriété du droit d'auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.

3 ABSTRACT A STREAMLINEO SYNTHETIC OCTOPLOID SYSTEM T HAT EMPHASIZES FRAGAR~A VESCA AS A BRIDGE SPECIES Robert Harold Bors Advisor: 3. A. Sullivan University of Guelph The synthetic octoploid (SO) system is a method of combining diploid, tetraploid and hexaploid Fragaha species into hybrid octoploids for introgression into octoploid strawberry cultivars of F.x ananassa. I n th is stud y, (d iplo id x diploid) and (diploid x hexaploid) crosses were used to create tetraploids followed by (tetraploid x tetraploid) crosses to create SOS from a wide germplasm base. Techniques refined for the SO system included Row cytometry protocols for identification of ploidy levels, in vitro germination and in vitro delivery of colchicine using the dropper method. Of the 36 species combinations investigated, 20 resulted in viable F, plants, 13 bloomed and produced seeds, and seven did not bloom. When nine species were intercrossed at the diploid level, the best crossability occurred with F.pentaphylla and F.vesca; each fomed fertile Ç, hybrids with five other species. No sig nificant differences in interspecific crossability were found among

4 F-vesca subspecies bractacea, americana, vesca and var. sempeflorens when pollinated with Enilgerensis, F.nubico/a, Epentaphylla, and Eviridis. These hybrids were treated with colchicine to create tetraploids. Tetraploids were also created from Fmoschata (6x) hybridized with F. viridis or F mbicola. The crossing strategy at the tetraploid level ernphasized F. vesca as a bridge species for F.nilgerrensis. Enubicola, F.pentaphy/la, and F. viridis. A collection of 1 O4 tetraploids inct uding (F. vesca x diploid) hybrids, F. vexa (4x), F. onentalis, (F.nubicola x Fmoschata), (Fmoschata x F. vindis), [F.orientalis x, (Fmoschata x diploid species)] and (F.viridis x F.nubicola) were intercrossed and doubled with colchicine to create SOS. When the (F. vesca x diploid species) colchi-tetraploids were intercrossed, only 37% fruit set with an average of three seedslflower. When these colchiploids were pollinated with tetraploids derived from F.orientaîis or (Emoschata x diploid species), 52% fruit set and 10 seeds/flower were obtained. A total of 192 SOS were created from 25 combinations of species. Of the 98 Sots planted in 1997, 28 produced fruit in The seven species successfully incorporated into SOS were F.vesca, F.nilgerrensis, F.nubicola, F.pentaphylla, F-viridis, F. orientalis, and F.moschata. The results suggests Fragana species are closely related and have great potential for breeding at the diploid and tetraploid levels and for introgression into 8x material.

5 ACKNOWLEDGEMENTS I wish to express my sincere appreciation to my wife, Loretta, and my children, Kirk, Aurora, and Nicholas who made many sacrifices to allow me to continue rny studies. Their support and enthusiasm was invaluable. I gratefully acknowledge my advisor, Dr. J. Alan Sullivan for his guidance, support. and encouragement over the entire period of my Ph.D. program. I am tremendously indebted to committee member Dr. K. Kasha whose Cytogenetics course formed the theoretical backbone on which this thesis was inspired. I thank my advisory cornmittee members Drs. D. Wolyn and T. Michaels for their valuable suggestions and encouragement. I acknowledge Sue Couling and Ron Dutton for assisting the care and culture of field and greenhouse plants. The technical expertise of Jan Brazelot, Dr. Peter Pauls. and Shawn Rogers is appreciated in utilizing flow cytometry in this thesis.

6 ACKNOWLEDGMENTS TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES TABLE OF CONTENTS i ii vi x LIST OF APPENDICES... xii CHAPTER 1 : LITERATURE REVIEW... 1 INTRODUCTION... 1 FRAGARIA SPECIES... 1 INTERSPECIFIC CROSSABILITY IN FRAGARIA... 7 Crossability arnong diploid species lncompatibility mechanisms in diploid Fragah Crossability arnong polyploids Outcrossing mechanisms in polyploids Crossability among dipioids and polyploids Surnmary of interspecific crossability GENOMlC INVESTIGATIONS Karyotyping Investigation of chloroplast DNA Investigations of nuclear DNA Isozyme studies Chromosome pairing in diploid hybrids Chromosome pairing in tetraploid hybrids Chromosome pairing in octoploid cultivars Genomic designations proposed for Fragaria species Genomic investigations and introgression strategy Pivotal Genome Theory ADVANTAGES OF USlNG FRAGARIA VESCA AS A BRIDGE SPEClES Introduction F.vesca is widely adapted and rnay be a progenitor species F. vesca is self-compatible Some forms of F. vesca are adapted to cultivation Fragana vesca var. sempeflorens has useful traits that facilitate breeding... 42

7 Fruitquality Lack of inbreeding depression F. vesca is amenable to chromosome doubling F. vexa cytoplasrn F-vesca as a bridge species INTROGRESSION OF FRAGARIA SPECIES USlNG SYNTHETIC OCTOPLOID SYSTEMS Crossing strategies to create SOS lmproving SO production lmproved strategy for SO production 57 CHAPTER 2: GENERAL MATERIALS AND METHODS Introduction ln vitro germination Colchicine treatment for chromosome doubling Acclimatization of seedlings Handling of parent plants for crossing Floral induction Pollination Organization of crosses Seed collection and storage Data collection and analysis Visual screening of plants following colchicine treatment Flow cytometry protocol Flow cytometry standards and the estimation of ploidy level CHAPTER 3: FRAGARIA SPECIES DIFFER IN DNA CONTENT 78 ABSTRACT INTRODUCTlON MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSION CHAPTER 4: HYBRlDlZATlON OF FRAGARIA VESCA SUBSPECIES AS FEMALE PARENTS WlTH F.NlLGERRENSIS, F*NUBICOLA. F.PENTAPHYLLA AND F*VIRIDIS rn m a 91 ABSTRACT... INTRODUCTION... Advantages of using Evesca as a bridge species

8 Advantages of using F.vesca as a female parent Evidence that F. vesca subspecies have diverged Goals of this study MATERIAL AND METHODS Experimental design and data analysis RESULTS Fruit set (%) and seeds/fiower Seeds with embryos and embryo germination Differences in intraspecific crosses Differences among kvesca accessions DISCUSSION 'Il7 CONCLUSION... CHAPTER 5: INTERSPECIFIC CROSSABILITY OF NlNE DlPLOlD FRAGANA SPECIES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION SUMMARY CHAPTER 6: INTERSPEClFlC HYBRlDlZATlON OF Fragaria moschata (6x) WlTH F.nubicola (2x) AND F.viridis (2x) 149 ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION CHAPTER7 :THE USE OF F.vesca, TETRAPLOID AND HEXAPLOID FRAGAHA SPECIES TO INCORPORATE DlPLOlD FRAGARIA SPECIES INTO SYNTHETIC OCTOPLOIDS ABSTRACT INTRODUCTION 171 Previous crossing strategies to create SOS Tetraploid breeding in Fragaria

9 Emp hasis on F. vesca. F. orientalis and F.moschata in crossing strategy Objectives and hypothesis MATERIALS AND METHODS Germplasm Crosses Techniques Data analysis RESULTS Differences among fernale parents Differences among male parents Interaction among males and fernales Influence of number of F.vesca genomes Lineage and fertility of SOS DISCUSSION SUMMARY... CHAPTER 8 :GENEML DISCUSSION AND SUMMARY REFERENCES APPENDICES

10 List of Tables 1.1 lnterspecific crossability of Fragana species with self-compatible diploid species used as males lnterspecific crossability of Fragaria species with self-incompatible diploid species used as males Interspecific crossability of Fragaria species with tetraploid and hexaploid species used as males Interspecific crossability of Fragana species with octoploid species used as males S. Gametophytic system of unilateral interspecific incompatibility of Fragaria species, based on Evans and Jones, Combinations of Fragaria species that have resulted in synthetic octoploids with different fertility levels Results of studies on hybridization of Fragaria tetraploids Results of attempts to create synthetic octoploids (SOS) from 2x. 4x and 6x Fragana species Nuclei isolation solutions , Nuclei isolation procedure Settings for the Coulter Epics Elite ESP flow cytometer when running Fragana nuclei with trout red blood cells (TRBC) as an internai standard Calculations used to determine chromosome numbers for Fragana hybrids from flow cytometry channel numbers Gemplasm used for fiow cytometry testing of Fragana species Variation in DNA content of various Fragaria species. accessions, and blood cellstandards Results of previous research using Fragaha vexa as a female in crosses with F. vindis, F. nubicola, and F. nilgerensis

11 4.2. Fragana vesca accessions used as fernale parents in crosses with F. nilgenensis, F. nubicola, F.pentaphylla, F. vindis and F. vesca Fragana accessions used as male parents in crosses with F.vesca Fruit set for crosses among Fragana vesca subspecies when pollinated with Ç. nilgerrensis (1 accession), F. nubicola (2), F. pentaphylla (1 ). F. viridis (2) and F-vesca (1) Seedslflower for crosses among Fragaia vesca subspecies when pollinated with F. nilgerrensis (1 accession), Enubicola (2), Epentaphylla (1 ), F. vindis (2) and F. vesca (1 ) Seeds with embryos (%) and embryo germination (Oh) for interspecific crosses among Fragana vesca accessions as fernales and six diploid accessions as males Num ber of reports for interspecific hybridization of diploid Fragana species, listed by crossability categories Diploid Fragada germplasm used in interspecific crosses Crossability categories for interspecific crosses among several diploid Fragana species Results of in vitro germination of seed from interspecific crosses among diploid Fragada species Results of interspecific crosses among several diploid Fragana species, showing differences among SCZ x SC, SC x SI, and SI x SI combinations ofspecies Outcomes of interspecific crosses among Fragaria moschata and diploid species Accessions of Fragana moschata (6x), F.nubicola (2x) and F-vindis (2x)used in crossing program for tetraploid production Fruit set, seedsmower, embryos. healthy plants, and one year old hybrids that bloomed, for crosses among accessions of Fragana moschata and F.nubicola or F. viridis Published reports of lnterspecific crosses of Fragana species and hybrids at the tetraploid level

12 7.2. Studies that created Synthetic Octoploids (SOS) from crosses among tetraploid... Fraganà species and hybrids Fragaria accessions that were founders of tetraploid germplasm for this experiment Tetraploid Fragarja germplasm derived from diploid species used for synthetic octoploid production Tetraploid Fragana germplasm derived from hexaploid, tetraploid and diploid species without the use of colchicine Effect of number of Fragana vesca genomes and female parentage on fruit set, seeds/pollination, germination, and synthetic octoploids (SOS) for crosses among tetraploid Fragaria hybrids Effect of nurnber of Fragana vesca genomes and male parentage on fruit set, seeds/pollination. germination, and synthetic octoploids (SOS) for crosses among tetraploid Fragana hybrids Fruit set, seedsl Rower, germination and synthetic octoploids (SOS) resulting from interspecific crosses among tetraploid Fragaria hybrids with different numbers of F.vesca genomes Details of the 14 most productive crosses among tetraploid Fragaria which resulted in 128 synthetic octoploids (SOS) Combinations of synthetic and natural Fragana tetraploids used to produce synthetic octoploids (SOS) A2.1. ANOVA for fruit set (%) of interspecific crosses with Evesca subspecies as female parents and six accession of other diploid species A2.2. ANOVA for seedslflower of interspecific crosses with Evesca subspecies as female parents and six accession of other diploid species A2.3. ANOVA for Seeds with embryos (%) of interspecific crosses with Evesca subspecies as female parents and six accession of other diploid species A2.4. ANOVA for embryo germination (%) of interspecific crosses with F.vesca subspecies as female parents and six accession of other diploid species

13 A2.5. ANOVA for fruit set (%) of intraspecific crosses with Evesca subspecies as female parents and the control pollinizer w ,426. ANOVA for seedsmower of intraspecific crosses with F.vesca subspecies as female parents and the control pollinizer vw A3.l. ANOVA for fruit set (%) of intraspecific crosses comparing (F.moschata x F.nubicola), (F-moschata x F. viridis), (F.nubico/a x F.moschata) and (F. viridis xf.moschata) A3.2. ANOVA for seeds per pollinated flower of intraspecific crosses comparing (F. moschata x F.nubicola), (F.moschata x F. viridis). (F. nubicola x Fmoschata) and (F. vindis x F.moschata) A3.3. ANOVA for seeds containing embryos (%) of intraspecific crosses comparing (F. moschata x F. nubicola), (F. moschata x F. viridis), (F. nubicola x Fmoschata) and (F. vindis x F.moschata) A3.4. ANOVA for healthy plants per pollination of intraspecific crosses comparing (F-moschata x F.nubicola), (F.moschata x F. vindis), (F. nubicola x F.moschata) and (F. vindis x Fmoschata) A3.5. ANOVA for hybrids that bloomed the following year resulting from intraspecific crosses comparing (F-moschata x F.nubicola), (Emoschata x F. vindis), (F. nubicola x F. moschata) and (F. vindis x F. moschata) A3.6. Fruit set, seeds, embryos, healthy plants, and one year old hybrids that bloomed, for crosses among accessions of Fragaha moschata and F.nubicola or F. viridis A4. Fruit set, seedsl flower, germination and synthetic octoploids (SOS) resulting from interspecific crosses among tetra ploid Fragana accessions A5.1. ANOVA for fruit set (%) of tetraploid crosses for data in Chapter A5.2. ANOVA for seeds per pollinated flower of tetraploid crosses for data in Chapter A5.3. ANOVA for germination (Oh) of tetraploid crosses for data in Chapter A5.4. ANOVA for synthetic octoploid production (%) of tetraploid crosses for data inchapter

14 List of Figures Leaf morphology of diploid and polyploid Fragaria species Mature fruit of several Fragaria species Geographic distribution of diploid Fragana species. based on Evans (1964). Reed (1966). Staudt (1989). and Hummer (1996) Geographic distribution of polyploid Fragaria species. based on Evans (1 964). Reed (1966). Staudt (1989). and Hummer (1996) Crossability of diploid Fragaria species Interspecific crossability of polyploid Fragana Unbalanced ploidy strategies for creating synthetic octoploids Balanced ploidy strategy for introgression of lower ploidy species into octoploid cultivars A simplified Balanced Ploidy Strategy for synthetic octoploid production In vitro treatment of Fragana vesca seedlings with colchicine solution Pollen storage method based on procedures of MacFarlane-Smith and Jones (1989) Diploid (left) and tetraploid (right) forms of the "Alexandria" variety of F. vesca ssp. vexa var. sempeflorens at identical ages Leaf morphology of interspecific hybrids of Fragada species Example of fiow cytometer output for a 4x-8x rnixiploid Fragaria hybrid with an interna1 standard Example of flow cytometer output for a hexaploid Fragaria plant Example of a secondary peak in flow cytometer output Regression of channel numbers from secondary peaks ont0 channel numbers of primary peaks for Fragana nuclei... 84

15 X Distribution of Fragaria vesca subspecies (Evans, 1964; Reed, 1966; Staudt, 1989) Hybrid vigour for vegetative growth of F. vexa x F.daltoniana F, hybrids Diploid hybrids shown one year after planting Variation in fruit size and yield of interspecific diploid hybrids Geneflow potential among diploid Fragaria species, showing results of previous investigators combined with this study Combinations of diploid Fragana species that have not resulted in viable progeny, but might be successful if more accessions were investigated Strategies used by Sebastiampillia and Jones (1 976) and Evans (1977) to create synthetic octoploids with the hexaploid Fragana moschata The three Fragada species used in this study: F.moschata, F.virids and F.nubicola Crossing strategy: a simplified balanced ploidy strategy for synthetic octoploid production Examples of tetraploid hybrids and fruit set resulting from interspecific hybridization Crossing plans using Fragaria species that were successful in creating synthetic octoploids in this study Synthetic octoploid fruit produced under open pollination conditions.. 206

16 .. y LIST OF APPENDICES 1. Abb reviations and add resses of gerrnplasm sources ANOVAs for Chapter ANOVAs and data for Chapter Data from tetraploid crosses in Chapter ANOVAs for Chapter

17 I CHAPTER 1 LITERATURE REVIEW INTRODUCTlON A need to expand the narrow germplasm base of octoploid strawberry cultivars (Sjulin and Dale, 1987; Daubney, 1990; Galletta and Maas, 1990; Hancock et al., 1990, 1996; Shaw, 1991; Sullivan ; Dale et al., 1993; Harrison et al., 1993; McNichol and Graham, 1992; Sangiacorno and Sullivan, 1994; Davidson, 1995) has prompted this investigation into the genetics and breeding of Fragana species. Previous reviews of strawberry breeding (Darrow, 1 966; Scott and Lawrence, 1975; Hancock et al., ) have indicated limited investigation has occurred on interspecific hybridization in Fragana. Besides providing more details regarding d iploid, tetraploid and hexaploid Fragana species, t his review reflects a desire by the author to develap a greater understanding of relationships within the genus. The areas of evolution and interspecific crossability are interconnected in this review. A greater understanding of evolution in Fmgana could help in development of an introgression strategy for bringing wild Fragana gemplasm into strawberry cultivars. Conversely, studies involving interspecific hybridization may help define evolution of the species. FRAGARIA SPECIES The commercial strawberry Fragana x ananassa Duch. was the result of hybridization of two wild octoploid species, F. chiloensis L. and F. virginiana Mill.,

18 2 during the mid 1700's (Darrow,I 966). While technically not a wild species, F. x ananassa is included in this review because of its economic importance and because it has been crossed to more species than either of its progenitor parent species. Fragaria species have a basic chromosome number of x = 7 and diploids are Zn = 14. Of the known Fragana species, there are 11 diploid, three tetraploid, one hexaploid and four octoploid species. Figures 1.1 and 1.2 show morphological differences present in leaves and fruit of the species. Botanical descriptions of these species have been published by several researchers (Reed, 1966; Staudt.1989; Hummer, 1996). It has been suggested that the diploid F. yezoensis, should be classified as a variety of F. nipponica based on chromosome karyotyping (Iwatsubo and Naruhashi, 1989), but a recent review (Hancock et al., 1996) lists it as a distinct species. Figures 1.3 and 1.4 shows distribution and ploidy levels Fragana. There are probably many Fragana species to be discovered in Asia (Jones, 1976). Numerous intemediate forrns of Fragana exist in herbariums, originally collected in eastern Asia, that cannot be classified botanically because they lack distinct differences (Staudt, 1989). A botanical expedition by the University of Tokyo reported Fragana species to be well differentiated in the eastern Hirnalayas (Hara, 1966) thus providing a western border to the location of undifferentiated Fragana. Eastern China appears to be the centre of diversity for Fragana as more species were located there than other areas.

19 Figure 1.l. Leaf morphology of diploid and polyploid Fragaria species.

20 Figure 1.2. Mature fruit of several Fragana species.

21 5 Fig Geographic distribution of diploid Fragada species, based on Evans (1 964). Reed (1 966), Staudt (1 989). and Hummer (1 996).

22 Fig.1.4. Geographic distribution of polyploid Fragana species, based on Evans (1 964), Reed (1 966), Staudt (1 989), and Hurnrner (1 996).

23 7 While disease resistance, pest resistance, cold hardiness, desirable flaveurs, specialized ecological adaptations and other valuable traits can be found in wild Fragana (Darrow, 1949,1966; Scott, 1951 ; Reed. 1960; Jones, l966,1976; Evans, 1974; Bauer, 1976; Chabot and Chabot, 1977; Kantor, 1984; Galletta and Maas, 1990; Trajkovski, ; Sullivan, 1991 ; Ahmadi and Bringhurst, 1992; Sangiacomo and Sullivan, 1994), this literature review will not focus on these characteristics in detail. This thesis focuses on introgression and evolution of Fragana. Previous attempts to introgress 2x, 4x and 6x species into cultivated octoploids (Darrow, 1966; Evans, 1977,1982~; Kantor 1984) had very low success rates and present an opportunity for improvement of introgression methodology. INTERSPECIFIC CROSSABILITY IN FRAGARIA lnterspecific hybridization studies in Fragana have been summarized in Tables Fragana gracilis Losinsk., F.mandschu#ca Staudt, Epentaphylla Losinsk. and Eiturpensis Staudt have not been included in these tables since no data could be found on their interspecific crossability. Of the 255 possible interspecific crosses between Fragaria species 54% have not been reported, 17% have been reported by only one researcher, and 29% have been reported by two or more researchers. Researchers have had variable results with interspecific crosses. For example, crosses between F. vesca L. x F.niigerrensis Schltdl. ex J.Gay have resulted in no seed set, no seed germination, death after germination, sterile plants, or viable plants (Table 1.1).

24 Table 1.1 lnterspecific crossability of Fragaria species with self-compatible diploid species used as males. Female Species used as males in crosses F. daltoniana Eiinurnae F. nilgerensis F. vesca Crossability categories' Ploidy OCMYABCDEF ABCDEF ABCDEF ABCDEF Fdaltoniana 2x SC F. iinumae 2x SC F. nilgeuensis 2x SC 1 F. vesca 2x SC F. nipponica 2x SI F. nubicola 2x SI 1 F. viridis 2x SI 1 F. orientalis 4x Tri 1 F. moupinensis 4x Di 1 Emoschata 6x Tri 1 F. chiloensis 8x Tri 1 F. virginiana 8x Tri 1 F. x ananassa 8x H+f 1 II ': Crossability categories: ': OCM outcrossing rnechanism: A = no set SC = self compatible B = seed, no germination SI = Self incompatible C = death after germination D = sterile plants Di = dioecious Tri = trioecious (males, fernales, or hemaphphrodites) E = plants produced, fertility not reported H+F = hermaphrodites and fernales F = fertile plants ': Numben indicate number of reports by researchers. note: Tables are based on: Yarnell, 1931 a; Mandelsdorf & East, 1 927; Federova, 1946; Williams, 1959; Staudt, 1959; Evans, 1964,1974,1977; Fadeeva, 1966; Evans and Jones, 1967; Senanake & Bringhurst, 1967; Staudt, 1967b. 1989; Burnagina & Mesaros, 1969; Sebastiampillia and Jones,1976; Kantor, 1984; Trajkovski, 1993; Noguchi et al., 1997.

25 9 In al1 these studies researchers were working with a limited number of accessions which may be a limiting factor to their success. A well-established tenet in plant breeding was that interspecific crosses often require the use of many accessions to obtain interspecific hybrids (Burbank, 1921 ; Stoskopf et al., 1993). "When initiating a program of using wild species in plant breeding, the first step is to collect as many accessions as possible of related species... "(Hermsen,1992). Variation between accessions for interspecific hybridization has been mentioned for Fragana species (Mangelsdorf and East. 1927; Williams, 1959; Evans. 1964). An example of the value of having many accessions was provided by Williams (1959) who reported that of the many accessions of F.vesca used. only one F.vesca accession from eastern Europe could successfully hybridize and form viable offspring with F. nilgenensis. Clearly, this hybridization was successful only because several accessions from a variety of locations were used. In addition to increasing the probability of success for interspecific hybridization, many accessions may also be important for creation of fertile hybrids. Sterile progeny has been reported 34 times and fertile progeny 32 times (column's D & F in Tables ) in interspecific Fragana hybrids. Contrad ictory reports exist for fertility of ten combinations of species. Genetic variation between accessions seems a plausible explanation for difiering results reported by many researchers. The possibility exists that environmental conditions rnay also influence interspecific crossability. Octoploid hermaphrodites could Vary in sex expression becauçe of environmental factors (Ahmadi and Bringhurst, 1991). If a plant appears to be hermaphrodite but is either

26 Table 1.2 lnterspecific crossability of Fragaria species with self-incompatible diploid species used as males. Species used as males in crosses F. nipponica F. nubicola F. vin'dis Crossa bility categories' Fernale Ploidy OCMy ABCDEF ABCQEF ABCDE F F.daltoniana 2x SC 1 1 F. iin urnae 2x SC F.nilgetrensis 2x SC F. vesca 2x SC F.nipponica 2x SI 7 1 F. nubicola 2x SI F. vitidis 2x SI F. orientalis 4x Tri F.moupinensis 4x Di 1 1 F-moschata 6x Tri F.chiloensis 8x Tri Evirginiana 8x Tri F. x ananassa 8x H+F ': Crossability categories: A = no set B = seed, no germination C = death after germination D = sterile plants E = plants produced, fertility not reported F = fertile plants OCM outcrossing mechanism: SC = self compatible SI = Self incompatible Di = dioecious Tri = trioecious H+F = hermaphrodites and fernales ': Numben indicate number of reports by researchers.

27 Table 1.3 lnterspecific crossability of Fragaria species with tetraploid and hexaploid species used as males. Species used as males in crosses F. corymbosa F. moupinensis F. orientalis F. moschata (44 (4~) (4~) (sx) Crossability categoriesz Fernale Ploidy OCMY ABCDEF ABCDEF ABCDEF A BCDE F Edaltoniana 2x SC F. iinumae 2x SC F.nilgenensis 2x SC F. vesca 2x SC F.nipponica 2x SI 1 F. nubicola 2x SI F. viridis 2x SI F. otien talis 4x Tri F.rnoupinensis 4x Di F.moschata 6x Tri F.chiloensis 8x Tri F-virginiana 8x Tri F.x ananassa 8x H+F ': Crossability categories: ': OCM outcrossing mechanism: A = no set B = seed, no germination SC = self compatible SI = Self incompatible C = death after germination Di = dioecious D = sterile plants Tri = trioecious E = plants produced, fertility not reported H+F = hermaphrodites and females F = fertile plants ': Numben indicate nurnber of reports by researchers.

28 Table 1.4 Interspecific crossability of Fragaria species with octoploid species used as males. Species used as males in crosses F. chiloensis F. virginiana F.x ananassa Crossability categoriesz Female PloidyOCMYAEjCDEF ABCDEF ABCDE F F. daltoniana 2x SC 1 F.iinumae 2x SC F. niigerrensis 2x SC F. vesca 2x SC F. nipponica F. n ubicola F. vin'dis F. orientalis F.moupinensis F. moschata F. chiloensis F. virginiana 2x SI 2x SI x SI x Tri x Di x Tri x Tri M M 8x Tri M M F. x ananassa 8x H+Ç M M ': Crossability categories: A = no set Y: OCM outcrossing mechanism: SC = self compatible 8 = seed, no germination SI = Self incompatible C = death after germination D = sterile plants E = plants produced, fertility not reported F = fertile plants Di = dioecious Tri = trioecious H+F = hermaphrodites and females ': Numbers indicate number of reports by researchers. M = many reports, crossability is well established in literature

29 13 male or female sterile, this would affect the success of interspecific hybridization. Since Fragana species grow at high elevations in Asia (Reed, 1966; Staudt, 1989), conditions may not be optimal for complete fertility at other locations with different environmental conditions. Temperature is "unquestionably the most critical factor affecting hybridization" (Layne, 1983), for example 60% seed set occurred for interspecific hybridization crops of peach and apricot with optimum temperature. As these two crops are also in the Rosaceous family, such information may be relevant to Fragana species. Factors such as humidity, light, wind, soi1 fertility, diseases, topography, and plant vigour can also influence the result of hybridization (Layne, 1983). If hybridization is diffwlt or impossible for certain species, mimicking the natural ecosystem of the species rnay be beneficial. Besides using many accessions and optimizing environmental conditions, several techniques have been used to en hance crossa bility. Emasculation 1-3 days before bud break prevented self-fertilization (Scott and Lawrence, 1975). Pollinating on the first and fourth day after emasculation was used to increase seed set in interspecific crosses of Fragana (Evans. 1964). Proper hand ling of pollen was also important to the success of interspecific hybridization (Galletta, 1983; Layne, 1983). Pollen viability can be maintained for 4 to 7 years if fresh anthers are placed in cotton stoppered vials in air tight jars containing desiccant and placed in the dark at 3-4'~ (MacFarlane-Smith et al ). Additional detail on pollen management such as checking for pollen viability was provided in Galletta's review (1983). Low seed nurnbers and poor germination rates have been common for interspecific crosses for Fragana. When 1256 interspecific pollinations were done

30 13 involving 2x, 4x and 6x Fragaia species, averages of 1.2 seeds per pollination and a 43% germination rate were achieved (Evans, 1964). When Fragaria species and hybrids were intercrossed in 15 different combinations, averages of 21 seeds per fruit and a 27% germination rate were achieved (Fadeeva, 1 966). As Fadeeva does not report the unsuccessful crosses, the actual yield of seedlpollination would be expected to be much lower. With low seed numbers likely from interspecific crosses, maximizing germination rates is important. Of the 16 germination methods tested on F. x ananassa seeds, the best rate (97%) occurred with in vitro germination of cut achenes on MS medium with no hormones (Miller et al., 1992). Only 27% of whole seeds on soilless mix germinated. The most stringent definition of interspecific crossability requires that two or more species create fertile offspring. Several rnethods such as use of mutagenic compounds (Kantor. l984), chromosome doubling (Scott, 1951; Evans, 1964, 1967,1974,1977,1982~; Fadeeva, 1966; Staudt, 1967; Sebastiampillia and Jones, 1976; Bauer, 1976, 1993; Trajkovski, 'l990, 1993), and unreduced gametes (Darrow, 1966; Bringhurst and Gill, IWO; Bauer, 1976; Trajkovsk,; l990, 1993; Bauer, 1993) have been used by researchers to overcome sterility in progeny. These techniques will be discussed in more detail in a later section. Crossability between di ploid species Less than one third of the possible crossing combinations between diploid species have been attempted and reported (Fig. 1.5). Of the seven diploids involved in interspecific crossa bility studies, only F. dalioniana J.Gay has not

31 15 successfully produced progeny with other species (Evans, 1964). Fragaria iinumae Makino, F. nilgenensis, F.nipponica Makino, F. nubicola Hook., F. vesca, and F. viridis Duchesne have been successfully hybridized to at least one other diploid species, and several resulting hybrids were fertile (Fadeeva, 1966; Evans and Jones, 1967; Sebastiarnpillia and Jones, 1976; Davis et al., 1996) showing potential for gene transfer at the diploid level with these species. Diploid crossability of F. nilgerrensis, F. nubicola, F. vesca, and F. virids has been studied by several researchers (Tables 1.1 and 1.4). Fragaria nilgenensis was the most dificult to hybridize, confiming cataplastic DNA analysis (Harrison et al., 1993, 1997) suggesting this diploid was distantly related to other Fragana species. The other three widely used diploids, F. nubicola, F. vesca, and F. viridis, have been intercrossed to produce fertile hybrids. This suggests that F. nubicola, F. vesca, and F. vindis are more closely related to each other than to F.nilgerensis. Differences exist in crossability of these three species to F. daltoniana, F. nipponica and F. nilgenensis (Fig. 1.5). It is difficult to make conclusions regarding crossability of F. daltoniana, F.iinumae and F. nipponica since there has been little research to document their crossability (Tables 1.1 and 1.2). While several researchers have focused on the introgression of diploid Fragaria species into cultivars (Scott, 1951 ; Fadeeva, 1966; Bringhurst and Gill, 1970; Sebastiampillia and Jones, 1976; Bauer, 1976, 1993; Evans, 1977; Trajkovski, 1990,1993) it is surprising that prebreeding at the diploid level between species has not been mentioned, since fertile diploid hybrids are possible. Studies involving breeding of F, interspecific hybrids have had the goal of understanding

32 Figure 1.5. Crossability of diploid Fragaria species. This shows the best results obtained by researchers as outlined in Tables 1.1 and 1.2. Arrows originate at males and teminate at females. Lines with two arrow heads indicate similar results for reciprocal crosses. fertile progeny created m F. penta p h yi/a 1 no viable progeny Ul--IIIHIYll----- C progeny created -,---,,-,- O self-compati ble self-incompatible unknown compatibility

33 17 genetic relationships but not improvement of diploid Fragaria (Evans and Jones, 1967; Davis et al., 1996). Diploid level prebreeding would have advantages over breeding at higher ploidy levels for combining desirable and, breaking undesirable, genetic linkages. Diploids have fewer alleles than polyploids, producing simpler segregation ratios and making the selection process easier, especially for recessive traits such as dayneutrality and fruit firmness. Incornpaübility mechanisms in diploid Fragaria Similar to other fruit crops in the Rosaceae (Lapins, l983), F.mandschunca, F-nipponka, F.nubicola, and F. vindis possess gametophytic self-incompatibility (SI) alleles that prevented selfing but allow crossing to genotypes with different SI alleles (Evans, 1964; Evans and Jones, 1967; Staudt, 1967a, 1989). Fragaria daltoniana. F.iinumae, Enilgenensis, and F. vesca are self-compatible (Evans, 1964; Staudt, 1989) while the compatibility statuses of Egracilis and F.pentaphy/la were not reported. Incompatibility mechanisms in Fragaria were due to inhibition of pollen tube growth in the styles, are garnetophytic in nature (Evans and Jones, l967), and corresponded with Lewis and Crowe's (1958) descriptions of unilateral selfincompatibility. These incompatibility relationships are summarized in Table 1.5. Variation in style length was not a factor effecting success of interspecific hybridization of Fragaria species (Evans, 1964). Self-incompatibility mechanisms have been shown to break down on occasion when self-incompatible F. vindis occasionally sets seeds with selfcompatible pollen from F. vexa (Williams, 1959; Fadeeva. 1966). It is questionable

34 Table 1 5. Gametophytic system of unilateral interspecific incornpatibility of Fragaia species, based on Evans and Jones lncompatibility relationships: Fernale Male Outcome cornpatibility compatibility type t~ Pe normal pollen tube growth nonal pollen tube growth pollen tube growth arrested pollen tube growth depends on proper combination of SI alleles of gametes Examples of SI x SI combinations: Fernale Male Outcome SI alleles Y SI alleles SI,S2 SI,S2 pollen tube growth arrested SI,S2 SI $3 pollen tubes with SI alleles have arrested growth those with S3 alleles grow norrnally SI,S2 S3,S4 normal pollen tube growth ': SC = self-compatible species, SI = self-incompatible species Y: S1, S2, S3, S4 = different self-incompatibility alleles

35 19 if these were true hybrids since F. vesca and F.vindis have similar leaf morphology (Staudt 1989) and the studies did not mention how they determined the resulting progeny were hybrids. Mutations are a plausible explanation for seeds from SI x SC crosses and for selfing of self-incompatible species (Lapins, 1983). Seif-incompatible alleles have neither been specifically identified nor categorized for Fragana species. It is not known if different species contain the same SI alleles or loci, which could have had an impact on success of interspecific crosses. There is no indication of how many SI alleles may be operating within Fragada species. Therefore the use of many accessions is recomrnended for crosses between self-incompatible species. As a self-compatible species, F. vexa has advantages over SI diploids, such as F.nubicola and F.viridis, for use in interspecific crosses. Since compatibility barriers do not operate when SC species are used as fernales, Evesca could be especially useful when lirnited accessions of SI species are available. Backcrossing SC x SI hybrids to SC species could eliminate SI alleles and make it easy to create isogenic lines and aid the isolation and transference of specific genes. As Evesca has the best record of SC species for interspecific hybriditation (Tables ), Evesca has clear potential for use as a bridge species. Crossability between polyploids Polyploid species are more likely to produce interspecific progeny that are fertile when the parents are at the same ploidy level (Fig. 1.6). All tetraploid Fragaria species have been hybridized and fertile progenies have been produced (Staudt,

36 20 Fig lnterspecific crossability of polyploid Fragana. Hybrids resulting from crosses of species with the same ploidy level are generally inter-fertile. Hybridization between species with different ploidy levels has been less successful than species with the same ploidy levels. Arrows originate at males and terminate at females. Lines with two arrow heads indicate similar results for reciprocal crosses. Based on the best results of crosses listed in Tables 1.3 and 1.4. Not al1 crosses have been attempted as in the case of F.iturpensis. 1 pmgeny created -----O--. * no viable progeny -) fertile progeny created

37 ). Crosses between three of the four octoploid species, F. virginiana, F.chiloensis, and F. x ananassa, also result in fertile progeny (Evans, 1964; Darrow, 1966; Staudt, 1989; Hancock et al., 1996) but no reports were found regarding crossability of F. iturpensis (8x). Crosses between polyploid species were more likely to produce viable seed when the higher ploidy species were used as the female parent (Evans,1974). In such crosses, the resulting embryo : endosperm chromosome ratio in progeny is closer to a normal ratio, and thus is more likely to survive than if the lower ploidy parent was used as the female. Unreduced gametes rnay have better survival potential than reduced gametes, particularly if a normal 2:3 embryo : endosperm ratio occurs (Evans, 1974). In such crosses, progeny frequently have higher ploidy levels than expected. Crosses resulting in odd numbers of genomes are usually sterile (Darlington, 1973; Evans, 1974; Sanford, 1983) although sterility of F. x ananassa (8x) X F. moschata (6x) hybrids was overcorne by applying the chernical mutagens H-nitrosoethylurea and 1,4-bis-diazoacetlybutane to seedlings (Kantor, 1984). It was suggeçted that the mutagens altered genes involved in meiosis and that odd numbers of chromosomes were not a factor causing sterility in Fragana hybrids. The ability of polyploid Fragaria to hybridize and produce fertile offspring suggested that polyploids were closely related (Federova, 1946; Fadeeva, 1966; Bringhurst and Senanayake, 1966; Bringhurst 1 WO), through common genomes, which is surprising because of their morphological differences (Fig. 1.1 and 1.2).

38 Outcrossing mechanisms in polyploids 22 In the first half of the 19th century, al1 F. x ananassa (8x) cultivars were dioecious, requiring growers to grow staminate plants to pollinate pistillate varieties. Pistillate varieties were considered superior in yields to hermaphrodites by early breeders (Roe, 1881). With the discovery and incorporation of productive hemap hrodites into breeding programs, pistillate varieties were rapid ly replaced during the late 1800's (Roe, 1881) and virtually eliminated in modern times (Darrow, 1966). Cross-pollination was not required for fruit set of cultivars, but was important for breeding programs to obtain hybrid vigour since octoploid cultivars are subject to inbreeding depression (Spangelo et al., 1971). Successive generations of inbred lines have lower vigour and fertility than parental lines (Scott and Lawrence, 1975). There were three major alleles that control sex expression: female (F), male (M) and hermaphrodite (H), which express dominance in the order of F > M > H in F. onentalis Losinsk. (4x), F. moschata (6x), F. chiioensis (8X) and F.virginiana (8x) (Staudt, 1968). Fragana orientalis and F. moschata were found to have tetrasomic and hexasomic inheritance ratios, respectively, for sex determination. Sex ratios in octoploids were not as clear. Staudt (1968) theorized that genes controlling female and male expression were similar in tetraploids and hexaploids but suggested octoploid species may have different sex-detenining genes. Further evidence to support Staudt's theory that sex inheritance differed in octoploids was provided when octoploid species were crossed to the diploid F. vesca and diploid Potentilla glandulosa (Ahmed and Bringhurst, 1991). Sex was controlled by octoploid inheritance in the order of F > H > M. The three sex genes were allelic in

39 2 3 octoploids but environmental conditions could influence hermaphroditic expression toward predominantly male or female expression (Ahmed and Bringhurst, 1991). Environmental influences were discussed as a plausible explanation for unclear sex ratios in octoploids. It has been difficult to differentiate visually between males and hermaphrodites as some males have pseudo-pistils. A method to differentiate between simiiar phenotypes was to treat emasculated flowers with phytohormones, such as NAA, whereby only female flowers set fruit (Staudt, 1968). Dioecious forms promoted outcrossing in polyploid Fragaria species (Staudt, 1989). It may be that separate sexes evolved in Fragana polyploids as an alternative rnechanism to self-incornpatibiiity to maintain heterozygosity. Many diploid Fragaria species were self-incompatible and were likely to have been involved in the evolution of polyploids (Staudt, 1989). However, if polyploid species have self-incornpatibility alleles, they are probably inactive. Self-incompatibility mechanisms broke down in synthetic tetraploids derived from self-incompatible diploid species (Evans and Jones, 1967) and hermaphrodite forms of polyploid species were self-compatible (Darrow, 1966). Of the diploid species, only Evesca subspecies bracteata had female flowers (Ahmedi and Bringhurst, 1991). As F. vesca ssp. bracteata was located in Western North Arnerica, it is unlikely to be the source of genes determining sex expression in Asian or European polyploids. Perhaps, this F.vesca subspecies and polyploid species share a common dioecious diploid ancestor or dioecy evolved separately in polyploids and diploids.

40 24 Crossability between diploids and polyploids Crosses between diploid and polyploid species often gave endosperm that failed to develop properly (Allard, 1960). It has been suggested that the irnbalances of chromosomes (Stephens, 1942) or gene dosages (Allard, 1960) were responsible for difficulties in obtaining hybrids behrveen diploids and polyploids. In most plant species, the genetic material of the endosperm was two parts fernale parent and one part male parent (North, 1979), thus, the chromosome ratio between endosperm and embryo was 3:2 for crosses between parents with the same number of chromosomes. This ratio becomes altered when crosses occur between plants of different ploidy levels. For example, diploid x tetraploid and tetraploid x diploid crosses have endosperm / embryo chromosome ratios of 4:3 and 5:3 respectively. It was suggested that endospem:embryo chromosome ratios were responsible for the failure of many diploid x polyploid crosses in Fragaria, and recommended that polyploid parents be used as females when crossed to diploids (Evans, 1974). Nevertheless, diploids species have often been hybridized as females to polyploid Fragana (Tables 1.1 to 1.4). A notable exception to this recommendation was the tetraploid F. moupinensis that çuccessfully crossed as a male to two diploid species. It is surprising that F. moupinensis and F.nilgenensis did not hybridize. These two species have similar morphology which has led to the theory that F. moupinensis may be a polyploid form of F. nilgerensis (Staudt, 1989). Limited numbers of accessions may have been responsible for the failure of these two species to hybridize or F.nilgerrensis and F. moupinensis are not related. As F.nilgemnsis is SC, it is doubfful that self-incompatibility alleles were involved in the

41 failure to hybridize these species. Crossability data supported the theory that F.vesca was a progenitor to polyploid Fragaria (Federova, 1946; Bringhurst and Gill, 1970; Bringhurst, 1990) but many other diploid species have been untested (Tables ). Fragaria vesca (Zr) has often been hybridized to octoploid species by breeders (Darrow, 1966; Bauer, 1976; 1993; Trajkovski. 1990, 1993, Ahmadi and Bringhurst, l992), and hybrids between F.vesca and F-chiloensis (8x) have beei: iound in the wild (Bringhurst and Khan, 1963; Bringhurst and Senanayake, 1966). Fragana vesca and F. vindis have been the only non-octoploid species to have produced progeny when used as females in crosses with F. x ananassa. Thus, these two species may have potential for contributing cytoplasmic DNA to 8x cultivars. Some F. vesca x F. x ananassa hybrids have been partially fertile while only sterile hybrids have been reported with F. viridis x F. x ananassa. Crossability studies indicate that F. vesca may have the greatest potential of the diploid species to act as a bridge species to octoploid cultivars. Although not mentioned in the literature, it is likely that recessive detrimental or lethal alleles present in polyploids may be a factor that lirnits the success of creating polyploid x diploid hybrids. As polyploid Fragana species were susceptible to inbreeding depression (Spangelo et al., 1971; Scott and Lawrence, 1975) polyploid x diploid crosses may have reduced vigour since heterozygosity would be decreased by the reduction of the number of genomes. The mechanism of self-incompatibility promotes heterozygosity, which may obscure the presence of detrimental or lethal alleles. Conversely, self-compatible

42 26 species, such as F. vesca, tend to be more homozygous and less likely to harbour lethal alleles. Thus, crossing polyploid species with self-compatible species may have a greater chance of success than crossing polyploid species with selfincompatible species. Summary of interspecific crossability Based on current Fragaria literature, F. vesca may have the greatest potential of the diploid species for acting as a bridge species to cultivars. Fragana vesca alone has successfully crossed as a male and female with F. x ananassa producing semi-fertile progeny and may offer the possibility to transfer unique cytoplasm to cultivars. As F.vesca is self-compatible, it can be used as a female in crosses with self-incompatible species without the interference of self-incornpatibility mechanisms. As previous research has not used an extensive collection of Fragana species and most had only one or two accessions for each species studied, crossing studies have been unable to elucidate the evolution of Fragaria. Since several Fragana at both the diploid and polyploid levels can be hybridized to produce progeny, it would seem that many species in Fragan'a are closely related. This could facilitate the introgression of the wild species of Fragafia into cultivars. GENOMIC INVESTIGATIONS Genomic investigations in Fragana may offer insight into evolutionary relationships and aid in developing introgression strategies. Studies involving

43 27 chromosomes, cytoplasrnic DNA, nuclear DNA, and isozymes have been loosely defined as 'genornic investigation' and included in this section, as these studies are useful in discerning relationships between various Fragaria species. The genus of Fragana is divided into 2x, 4x, 6x, and 8x species. The diploids species are likely progenitors of higher ploidy species and may have contributed different genomes, complicating investigation of evolutionary relationships. Much of the genomic research in Fragana has focused on discovering progenitors of the commercial strawberry, F.x ananassa. As evolution is an ongoing process, the discovery of progenitor or intermediate species may not be possible if they no longer exist (Bringhurst and Gill, 1970). The search for a species similar or less divergent to the progenitor is perhaps a more realistic goal. Although genomic investigations of Fragaria do not offer conclusive evidence for identification of a progenitor species, the information is beneficial for developing an introgression strategy. Karyotyping Fragana chromosomes were very srnall, ranging in size from 0.9 to 1.7 microns (Yarnell, 1929) making differences between chromosomes difficult to discem with light microscopy. Chromosome banding techniques would be difficult to do with such small chromosomes, and has not been reported for Fragana. Karyotypes were developed for some diploid species. Fragana vesca, F. nipponica, F. nubicola and F. daltoniana were very similar in karyotype morphology (Iwatsubo and Naruhashi, 1989,1991). Fragana iinumae differed from other diploids because it had more chromosomes with median centromeres and a large satellite on

44 28 chromosome seven. F. yezoensis and F. nipponica chromosomes were so similar it was recommended that F. yezoensis be reclassified as a variety of F. nipponica (Iwatsubo and Naruhashi, 1989). Investigation of chloroplast DNA Investigation of chloroplast DNA (cpdna) has limited use for determining evolutionary relationships since cpdna is inherited rnaternally. Chloroplast DNA may be highly conserved in Fragaia but some differences existed. The cpdna of nine Fragana species had "strikingly little cpdna variability" (Harrison et al., 1993, 1997). In these studies it was determined that F. iinumae, F. nilgenensis, F. nubicola, F. vesca, and F. vidis were not materna1 ancestors of octoploid species. Fragana virginiana and F. chiloensis had no detectable differences in cpdna and s hared four unique restriction fragment length pol ymorp hism (RFLP) bands not present in other Fragada species. Some species, which were widely separated geographically, were found to have similar cpdna. Fragaria nubicola (from India), and F. vesca (from North American) and had no detectable differences in cpdna. These two species had more (RFLP) bands in common with polyploid species than other diploid species. Fragana nubicola diverged from F. viridis (from Europe) by one RFLP band, F. vesca (from Europe) by eight bands, F. nilgerrensis by 11 bands and F. iinumae by 21 bands. These results compliment crossing studies (Tables ), morphology (Fig 1-1 and 1.2; Reed, 1966; Staudt, 1989) and karyotyping (Iwatsubo and Naruhashi 1989, 1991) which suggest F.nilgerrensis and F.iinumae are divergent from other Fragana species.

45 2 9 Evesca accessions displayed a wider range of diversity in cytoplasrn than other Fragaria species (Harrison et al. 1993, 1997). The octoploid species, F. X ananassa, F. chiloensis and F. virginiana d iffered f rom Ame rican F. vesca s pecies by only four mutations, and eight mutations differentiated Arnerican and European F. vesca. This could support a theory of F.vesca as a progenitor species. Little is known, however, about diversity of cytoplasm within other diploid species in the study since they were each represented by only one or two accessions. Unfortunately, F.daltoniana, F.gracilis, F. mandschuica, Fmpponica, F.pentaphylla, were not included in the studies by Harrison et al. (1993, 1997). Perhaps one of these species has a cytoplasm similar to that of octoploid species. Investigations of nuclear DNA Most investigations of nuclear DNA in Fragaria have been concerned with fingerprinting octoploid cultivars, estimating relatedness of cultivars, or using marker assisted selection (Davis, 1996a). Only the study of nuclear DNA by Davis et a1.(1 996b) had implications for Fragaria evolution. When progenies of the cross (F. vexa x F. vindis) x Evesca were investigated, only four linkage groups were found instead of the expected seven (Davis et al., 1996b). It seems likely that a selfincompatibility allele from Eviridis may have been responsible for the 16 markers in Mage group IV that had no recombination since the (F. vesca x F. vindis) hybrid was used as the female in crosses (see Table 1.5). The female gametes containing self-incompatibility alleies would have been selected against since these gametes would have prevented growth of SC pollen tubes from F.vesca. Despite

46 30 difficulties with linkage group IV, evidence provided indicated chromosomal translocations have occurred in the divergence of F. vesca and F. vindis. lsozyme studies Glucose phosphate isomerase -2 (GPI-2) isozymes were studied in many accessions of F.chiloensis, F. virginiana, F.x ananassa and in a few accessions of F. vesca, F. vindis, F. nubicoia, F.nilgemnsis, F.iinumae and F. moschata (Arulseker and Bringhunt, 1983) but not al1 Fragana species were investigated. Each genorne canied genes that encoded for synthesis of GPI-2 isozymes. Thus, polyploid accessions could have more isozyme bands than diploid accessions. Since GPI-2 isozymes were expressed as dimers, three isozymes bands were possible in diploid accessions. Fifteen bands were isolated which proved useful for fingerprinting and inheritance studies. The diversity of GPI-2 isozymes in octoploids showed broader biochemical diversity that may be related to adaptations evolved by octoploids (Arulseker and Bringhurst, 1983) but implications concerning Fragada evolution were not fully discussed. If hybrid dimer bands are excluded, F. chiloensis had the greatest diversity followed by F. vesca, F. virginiana, F. viridis, Emoschata and F.nilgenensis with 6, 4, 3, 2, 2, and 1 GPI-2 bands, respectively. Octoploid accessions had GPI-2 bands corresponding to bands found in several diploid species, the exception being band #25 that was unique to some Echiloensis accessions. Thus, it is possible that F.chiloensis was derived in part from another species not present in the study. Fragana daltoniana (2x) would be an excellent candidate for further investigation of band #25 since leaf morphology suggests it

47 3 l may have been a progenitor to F.chiloensis (Fig. 1.1, Staudt, 1989). Since Fragana viridis had three GPI-2 bands (one being a dimer hybrid) that were unique, it seems unlikely it was involved in evolution of octoploid species. Conversely, F.vesca, Ç.iinumae, F.nubicola, and Fmoschata had GPI-2 bands that corresponded to bands in octoptoid Fragana, so these four species may be considered as possible progenitors of octoploids. Fragana iinumae and F. nubicola had a GPI-2 bands that corresponded to F. vesca ssp. califomica and F. vexa ssp. vexa respectively, which suggests a degree of divergence in some F.vesca subspecies. It has been suggested that F. x ananassa was highly diploidized. based on isozymes studies (Arulsekar et al., Arulseker and Bringhurst, 1983). The monohybrid and dihybrid segregation of GPI isozymes indicated that chromosomes carrying GPI genes are diploidized, but such results do not show that al1 chromosomes were diploidized. Conclusive proof of diploidization would require genetic markers for each chromosome that make up Fragana genomes. Genetic control was suggested as the cause of diploidization in Fragana octoploids (Byrne and Jalenkov, 1976; Arulsekar et al ; Bringhurst, 1990). An alternative explanation is octoploids are allopolyploids derived from several diploid species with homologous pairing occurring preferentially to homeologous pairing. Glucose phosphate isomerase-one (GPI-1) bands were also studied by Anilseker and Bringhurst (1983). Only two different GPI-1 bands were found in Fragana species; F.nilgemnsis had one unique GPI-1 band while al1 other species had the other GPI-1 band. This finding showed further evidence of the divergence of F.nilgerrensis from other species in its genus.

48 32 Chromosome pairing in diploid hybrids The ability of homeologous chromosomes to pair during meiosis is, with few exceptions, a prerequisite for fertility of interspecific hybrids. While interspecific hybridization between di ploid species may be difficult, chromosome pairing in meiosis was normal for interspecific hybrids of Fragada diploids (Staudt, 1959; Jones, 1976). Pairing of homeologous chromosomes occured in hybrids derived from crossing remotely related Evesca and Enilgenensis (Jones, 1976). Jones suggests that little differentiation has occurred in the homology of chromosome sets in Fragana, thus preventing any simple analysis of evolutionary relations. Chromosome pairing in tetraploid hybrids Fragaria orientalis (4x) formed rnultivalents in diakinesis for 50% of cells studied, but by anaphase 1 bivalent formation with late separation of one bivalent was noted for 15% of the cells (Staudt. 1952). Some chromosomes tend to associate as quadrivalents more ofien than others in autotetraploids derived from four diploid Fragarfa species (Bhanthumnavin, 1965) indicating that diploidization was probably not under genetic control in these species. Evans and Jones (1967) observed 9 to 14% meiotic quadrivalent chromosome associations for three 4x (F. vesca x F. virids) hybrids and 26 to 37% quadrivalents for two 4x (F. nilgerensis x F. vindis) hybrids. In these putative allotetraploids there was tetrasomic inheritance of self-incompatibility alleles. Quadrivalent formation in F. nilgerrensisx F. vinds hybrids was especially surprising since the parental species were thought to be distantly related. These two species

49 33 differed in geographical location (Reed, 1966), morphological characteristics (Staudt, 1989). chloroplast DNA mutations (Harrison et al., 1993, 1997) and isozyme banding patterns (Arulsekar and Bringhurst, 1983). In another study (Sebastiampillia and Jones, 1977) an average of 10.4 bivalents in diakinesis and 11.5 bivalents in metaphase I were found in eight tetraploid forms of F-vesca, Enipponica and Eiinumae. Thus, it was shown that diploidization may be only partial in autotetraploids. Perhaps small chromosome size (Darlington and Ammal, 1932; White, 1961) or low chiasmatic frequency (Upcott. 1938) found in Fragaria were responsible for the majority of these chromosomes pairing as diploids. Chromosome pairing in octoploid cultivars Examination of nine octoploid cultivars found on average between four and seven chromosomes involved in multivalent formation in diakinesis but the remainder of the cell cycle appeared normal (Mok and Evans, 1971). It was concluded that tetrasomic inheritance could occur in octoploid cultivars. Other studies detenined chromosomes in cultivars paired as bivalents (Byrne and Jelenkovic, 1976; Ibrahim et al., 1981). It was suggested that genetic control causes preferential pairing within genomes but if homologues were not present, pairing could occur with homoeologous chromosomes (Byrne and Jelenkovic, 1976). As pointed out by Galletta and Maas (1990), whatever the cause of diploidization, 'selection by breeders for high fertility types probably autornatically results in a correlated selection for regular bivalent pairing and disjunction'. If

50 34 pairing in octoploid species is indeed under genetic control, then introgression of other species should be easy. Genornic designations proposed for Çragaria species The following genomic relationships have been proposed, based on homology of pairing in interspecific crosses. with each genome represented by a capital letter: F.nipponica = BB, F. vesca = CC. Eorientaiis = AAAA, Fmoschata = AAAABB, and F. x ananassa = AAAABBCC (Federova, 1946). The designation for F. nipponica was not based on pairing studies, but on its ability to hybridize with F. moschata to produce fertile tetraploids (Federova, 1946). The Federova study had the disadvantage of using hybrid species F. x ananassa (8x) rather than its progenitor species F. chiloensis (8x) and F. virginiana (8x). lt was suggested that F. chiloensis (8x) and F. virginiana (8x) could be classified as AAAABBCC since Fragana octoploids easily intercross and have fertile progeny (Fadeeva, 1966). Inferring that F. virginiana and modern F.x ananassa have similar genomic constitution may be reasonable since F. x ananassa is more closely related to F. virginiana than F. chiloensis. Fragana virginiana was extensively used in early breeding programs of F. x ananassa cultivars and F.x ananassa x F.chiloensis hybrids display heterosis while F.x ananassa x F.virginiana hybrids do not (Darrow, 1966). However, it may be erroneous to assume F.chiloensis had the same genomic constitution as other octoploid species. Recent reviews (Scott and Lawrence, 1975; Hancock et. al, 1990,1996) accepted that F. chiloensis, F. virginiana, and F. xananassa = AAA'A'BBBB, F.

51 35 vesca = AA, and F. vindis = AA designations suggested by Senanayake and Bringhurst (1967) but do not discuss Federova (1946) or Fadeeva (1966). Bringhurst (1990) modified his earlier hypothesis based on evidence that octoploid Fragada species are hig hly diploidized (Byrne and Jalen kovic, 1976; Arulseker and Bringhurst, 1981, 1983) and concluded that octoploid Fragaria genomes should be designated as AAA'A'BBB'B'. These designations were based an limited interspecific crosses between diploid and tetraploid forms of F. vexa with both F. chiloensis and F. virginiana as well as F. chiloensis x Potentilla glandulosa and F. chiloensis x F. viridis. While Senanaya ke and Bring hurst (1967) discuss Federova's (1946) genomic designations (erroneously altering Federova's AAAABBCC to become AABBBBCC) at length to present evidence to re-designate CC to A'A', they do not discuss the bulk of Federova's work which designated possible origins for each of the genomes in F. x ananassa. Their suggestion that the A' genorne originated from an extinct species and B genomes originated from an undiscovered species is not well substantiated since only two diploids, F.vesca and F-vindis were included in their study. Research on karyotyping and chromosome pairing of diploids seerned at variance with the above octoploid genomic designations. Thus far, only F.inumae genomes were noticeably divergent from other fragada genomes, and it has not been established or negated that F.iinumae genomes were present in octoploids. Genomic investigations and introgression strategy Fragana chromosomes exhibited high hornology with a strong tendency

52 3 6 towards diploidization at al1 ploidy levels. The diploidization tendency could be due to small chromosome size, low chiasma frequency, or genetic control. Morphology and isozyme studies indicate octoploid species rnay share some common as well as unique diploid progenitors. This possibility will be discussed in more detail in the pivotai genome section. Hornoeologous pairing and chromosome substitutions could have occurred at the diploid, tetraploid or octoploid levels during the evolution of F.chiioensis and Evirginiana so there is no guarantee that genomes in these species can be identified with diploid progenitors. If F.chiloensis and F.virginiana descended in part from different species, the hybrid Ex ananassa could have recombined genomes derived from two or more diploid species. While the similarity of Fragaria genomes may obscure evolutionary relationships, octoploid Fragaria species may have the capacity to integrate genomes from other species in the genus. Thus, there may be great potential for broadening the germplasm base of F. x ananassa by integrating genomes from species that were not involved in the evolution of octoploid species No clear strategy for integration of 2x, 4x. and 6x species into octoploids is indicated by previous genomic studies in Fragaria. It seems likely a number of integration strategies may be possible (Jones. 1966). While chromosome pairing is not likely to be difficult, factors such as crossability and genetic differences could play important rotes in developing an introgression strategy. While F.vesca has been designated as an A genome and is considered by many to be a progenitor to the octoploid species, it seems rnost diploid species in Fragana could also be

53 37 designated as A genomes. The choice to investigate Evesca in this thesis is made with the knowledge that F. vesca genomes are probably present in octoploid species but that genomes from other diploids may also be present. While genomic differences between species can cause serious barriers for introgression strategies, this is not the case in Fragaria. Pivotal Genome Theory The Pivotal Genome Theory proposed that allopolyploid species may result from crosses between polyploid species that shared at least one genome in common (i.e. pivota1 genomes), yet differed by at least one genome (Le. differential genomes) (Zohary and Feldman, 1961). In polyploid interspecific hybrids, pivota1 genomes helped maintain fertility so that recombination could occur between differential genomes. Cytological, morphological, and taxonomic evidence from studies of Aegilops and Tdticurn and interspecific hybrids provided the basis for the theory (Zohary, 1965). Although several studies suggested that evolution of new species in nature through the mechanism of some pivotal-differential genomes was probably rare (Zhang and Dvorak, 1992; Zhang et al., 1992; Talbert et al., 1993; Dubcovsky and Dvorak, 1995), the Pivota1 Genome Theory has been quite useful for breeden involved with interspecific hybridization of polyploid grain crops. This theory has shaped introgression strategies and allowed chromosomes from distantly related species to be recombined and introgressed (Gustafson and Dera 1989; Armstrong et al., 1992; Gustafson and Sears, 1993; Lelley et. al, 1995). In wheat breeding the use of the Pivotal Genome Theory led to many new varieties with

54 38 increased yield, quality and adaptability because related çpecies could be used as sources of new genes (Feldrnan and Sears, 1981). The Pivotal Genorne theory could be used to enhance the introgression of chromosomes from other Fragada species into F.x ananassa cultivars. If genomes similar to those in F. x ananassa, were used in interspecific hybridization perhaps crossability and fertility would be enhanced during the introgression process. The most likely diploid species to have been involved in the creation of both F.virginiana and Ç.chiloensis is Evesca (Bringhurst and Senanayake, 1966; Bringhurst, 1990). Fragana vesca is the only diploid species that shared the same geographic range as the octoploid species (Evans, 1964; Reed, 1966; Staudt, 1989). It also hybridized with F. virginiana and F.chiloensis and often resulted in fertile progeny when ploidy levels of hybrids were doubled (Yarnell, 1931a; Bringhurst and Khan, 1963; Evans, 1964; Senanayake and Bringhurst, 1967; Bauer, l976,l993; Ahmadi and Bringhurst, 1992). In addition to F.vesca, two other species have been shown to have common genornes with the octoploid species. Chromosome pairing studies suggested Eorientalis (4x) and Emoschata (6x) share genomes with F.x ananassa (Federova, 1946). These latter two species rnay also contain F. vesca genomes (Federova, 1946; Staudt, 1959; Darrow, 1966). If F.vesca is present in 4x, 6x and 8x species, it may have been a pivotal genome in the evolution of polyploid Fragana s pecies.

55 3 9 ADVANTAGES OF USlNG Fragaria vesca AS A BRIDGE SPECIES Introduction An important element to any interspecific introgression program is the proper knowledge and use of germplasrn. Of al1 the species in the genus Fragana, F. vesca may have the greatest potential for use as a bridge to introgress lower ploidy species into octoploid cultivars. In a broader sense, a bridge species is defined as one which allows the combining of genetic material from species which cannot directly hybridize. In regards to the SO system a bridge species could one that crosses with several other species and could allow the introgression of dificult to hybridize species into synthetic octoploids. This section will review the unique attributes of F.vesca and its potential as a bridge species for both geneflow and the SO system. F.vesca is widely adapted and rnay be a progenitor species Fragaria vesca is widely distributed and partially overlaps geog rap hically with al1 of the other Fragaria species (Figs. 1.3 and 1.4) which has led several researchers to conclude that it rnay be the oldest Fragaba species and progenitor of the genus (Duchesne in Darrow, 1966; Staudt, 1953; Nurnberg-Kruger;l958; Longley in Darrow 1966). Morphological traits have been used to suggest that Evesca is closely related to, or a progenitor of other species (Federova, 1946; Darrow, 1966; Reed, 1966; Staudt, 1959, 1989, Ahmed and Bringhurst, 1991). Since hybrids can be made between Evesca and polyploid species, several researchers have suggested F. vesca is at least a progenitor to polyploid species

56 4 O (Darrow, 1949, 1966; Federova, 1946; Ellis, 1958; Staudt, 1959, 1984; Bringhunt and Senanayake, 1966; Bringhurst and Gill, IWO; Bringhurst, 1990; Ahmed and Bringhunt, 1991). If F.vesca is closely related to other Fragaria species, it may have the greatest potential of al1 diploid species for use as a bridge species. A factor contributing to F. vesca's widespread distribution was its ability to photosynthesize at low lig ht levels, which helped it to adapt to forest ecosysterns (Chabot and Chabot, 1977). It may be better adapted to habitats in cooler climates since it is found primarily in northern areas, and at higher altitudes in equatorial locations (Staudt, 1989). Adaptability of F. vesca was a major reason for attempting to introgress it into cultivars (Darrow, 1966). Decaploids (called F. x vescana) derived from crosses between F. vesca var. sempeflorens and F. x ananassa were noted to possess broader climatic adaptation than F. x ananassa cultivars (Bauer, 1976). These hybrids showed increased vegetative growth during long daylengths compared to pure F. x ananassa cultivars. These hybrids conti nued vegetative growth later in the fall and had flower initiation at 5-10 C in the field rather than entering dormancy at the same time as F. x ananassa cultivars (Bauer, 1976). This attribute could be especially valuable at northern latitudes where low temperatures limit the duration of flower bud initiation and reduce yield potential (Chercuitte et al., 1992). Delayed dormancy was attributed to the day neutral gene of F. vesca var. sempedlorens (Bauer, 1976) although wild types of F. vesca without the day neutral gene were not tested. As the day neutral gene of F. vesca was recessive (Darrow, 1966; Ahmadi and Bringhurst, 1990) it seems unlikely that this gene was responsible for delayed dormancy. Fragaria vesca's ability to carry out

57 4 1 photosynthesis under low light levels (Chabot and Chabot, 1977) rnay have contributed to increased flower bud initiation during the shorter day lengths and low light levels in autumn. As F. vesca seemed well adapted to cooler climates, a variety of other genetic factors could be involved in F. x vescana's improved climatic adaptation. F. vesca is self-compatible Evans and Jones (1967) found diploid Fragada species had a gametophytic system of unilateral interspecific incompatibility as described by Lewis and Crowe (1958). Whiie F. vesca was rarely successful when pollinating diploid species (Table 1.1 and 1.2), it can be used as a female in crosses with self-incompatible species to circurnvent self-incornpatibility mechanisms. Only four of the1 0 diploid Fragana species were self-corn patible: F. vesca, F. daltoniana, F.iinumae and F. nilgemnsis. Of these, little data was available on crossing F. daltoniana and F.iinumae with other species and F.nilgenensis has rarely crossed with other species or has produced stunted interspecific hybrids (Reed, 1966). In contrast, F. vesca has been hybridized with several Fragana species (Mangelsdorf and East, 1927; Yarnell, 1931 ; Scott, 1951; Bringhurst and Khan, 1963; Evans, 1964; Jones, 1966; Fadeeva, 1966; Evans, 1974; Bauer, 1976; Sebastiampillia and Jones, 1976; Hancock et al., l99o,'i996; Ahmadi and Bringhurst, 1992; Bauer, 1993; Trajkovski, 1993) and the resulting hybrids are often viable (Tables ). Thus, the self-compatibility trait of F.vesca could be of assistance in diploid hybridization.

58 4 2 Some forms of F. vesca are adapted to cultivation Fragaria vesca, also known as the Alpine Strawberry, has been cultivated for at least 500 years in Europe (Darrow, 1966; Wilhelm and Sagen,1974). More than three hundred years ago the everbearing non-runnering form, F. vesca var. sempemorens, was discovered in the Alps which formed the basis for modern F. vesca cultivars. (Darrow, 1966). As alpine strawberries were propagated primarily from seed and need to be replanted every few years, it is possible Evesca cultivars have had approximately 100 generations of active andlor passive selection for adaptation to cultivation. Fragaria vesca var. sempedloiens has useful traits that facilitate breeding Fragaria vesca var. sempedorens is both day neutral (continuously blooming) and non-runnering. No other diploid species in Fragada was known to carry non-runnering or day neutral genes, although a dominant day-neutral gene existed in some octoploid cultivars (Bringhurst and Voth, 1984). These traits were recessive, and governed by independent genes (Darrow, 1966; Ahmadi et al., 1990) making them idealiy suited for use as genetic markers to verify hybridization has taken place. Fragaia vesca, F. vindis, F.nipponica, and F.nubicola have similar morphology (Reed, 1966; Staudt, 1989; Hancock et al., 1990; Figure 1.l) and hybrids between these species could be difficult to identify visually. The day neutral trait of F. vesca var. sempeflorens could be especially useful logistically when working with other species. As species derived from various latitudes and altitudes are very likely to bloorn at different times, everblooming

59 4 3 F.vesca cultivars could be continuously available for hybridization. Fruit quality The aroma and flavour of Evesca is considered superior to other Fragana species and several breeding programs have attempted to introgress these two attributes into F. x ananassa cultivars (Darrow, 1949, 1966; Jones, 1966; Ahmadi and Bringhurst, 1992; Trajkovski, 1993). The pleasing aroma from F. vesca was expressed in decaploids derived from F. vesca and F. x ananassa (Darrow, 1949; Bauer, 1993). Some characteristics of F. vesca fruits were undesirable. Small, very soft fruit with large intercellular air spaces (Darrow, 1966) couid make hawesting both delicate and labour intensive. Lack of inbreeding depression Unlike octoploid cultivars (Morrow and Darrow, 1952; Spangelo et al., 1971 ; Niemirowicz-Szczytt, 1989), diploid F. vesca cultivars did not have in breeding depression and were true from seed (Darrow, 1966). It is, therefore, unlikely that undetected deleterious recessive traits would be introduced from F. vesca gennplasm. Even at the autotetraploid level, F. vesca was fertile (Yarnell, 1931 b; Scott, 1951 ; Evans, 1964; Bumagina, 1974; Bauer, 1976; Trajkovski, 1993; Sebastiampillia and Jones, 1976,1977; Silva and Jones, 1996). These factors may make F. vesca more fit at higher ploidy levels than other Fragaria diploids, which rnay be important in introgression strategies.

60 4 4 Fm vesca is amenable to chromosome doubling Fragaria vesca was not only more responsive to colchicine's chromosome doubling action but it was less sensitive to colchicine toxicity than other species. Chromosome doubling occurred more readily for F. vesca than for F.nubicola, Eiinumae, Enilgerrensis, F. vindis, F. moschata, or F. x ananassa (Fadeeva, 1966; Sebastiampillia and Jones. 1976) and had a higher survival rate after colchicine treatment (Sebastiampillia and Jones, 1976). The superiority of Evesca seemed to be due to bath genetic and cytoplasrnic factors (Fadeeva. 1966). This peculiarity may be due in part to phenolics or biochemical adaptations of different species. Porebski et al. (1997) found it necessary to modify CTAB DNA extraction protocols for Fragana species because of high polysaccharide, tannins and polyphenol components. It seems likely that concentrations of these compounds differ among Fragana species and could interfere with the alkaloid activity of colchicine or perhaps be connected with increasing the toxic effects of colchicine. Whatever the mechanism, Evesca's cytoplasrn and genomes may transmit ease of chromosome doubling to hybrid progeny. Thus, the use of F.vesca as a female parent would be particularly beneficial for introgression strategies that require chromosome doubling. F. vesca cytoplasm Cytoplasmic differences have been important in the success of interspecific hybridization (North, 1979; Stoskopf et al., 1993). Too few synthetic octoploids have been created to judge how important cytoplasmic differences are for synthetic octoploid production, but fertile synthetic octoploids were created that had F. vesca

61 4 5 cytoplasm (Ellis, in Darrow, 1966; Sebastiarnpillia and Jones, 1976; Evans, 1982b). Synthetic octoploids have also been created with F. moupinensis and F. moschata cytoplasm (Sebastiampillia and Jones, 1976; Evans, 1977). Although rnost of Fadeeva's (1966) interspecific crosses were unbalanced between species and hybrids of differing ploidy level, good seed set was achieved when F. vesca was the female parent. Investigations of RFLPs from chloroplast genomes revealed cytoplasmic variation was greater among three subspecies of F. vexa than among six subspecies of F.virginiana and F.chiloensis (Harrison et al., 1993, 1 997). While it is not known if different cytoplasm would be of benefit to cultivars, the use of F. vesca in introgression strategies could offer researchers an opportunity to explore this possibility. F.vesca as a bridge species As will be seen in the section that follows, Evesca has been used successfully as a bridge species in the synthetic octoploid system. However, such researc h did not emp hasize F. vesca, so the potential of F. vesca as a bridge species in the synthetic octoploid system has not been fully investigated.

62 4 6 INTROGRESSION OF FRAGARIA SPECIES USiNG SYNTHETIC OCTOPLOID SYSTEMS Only a few accessions of F. vesca, F. vindis, F-nubicola, Fmoupinensis, and Fmoschata have been introgressed into octoploid cultivars (Darrow, 1 966; Evans, 1976; Kantor, 1984; Ahmadi and Bringhurst, 1992; Bauer, 1993; Sangiacomo and Sullivan, 1994). There is no current evidence to suggest that the nine other non- octoploid species have been introgressed into 8x cultivars. While fertile heptaploids and decaploids resulted from crosses between F. x ananassa and lower ploidy species, these groups were reproductively isolated from octoploid varieties because of differences in ploidy levels (Kantor, 1984; Bauer, 1976, 1993; Bringhurst and Voth, 1984; Ahmadi and Bringhurst, 1992; Noguchi et al., 1997). It was proposed that lower ploidy species (2, 4, and 6x) be intercrossed and have ploidy levels doubled with colchicine to create synthetic octoploids (SOS) (Evans, 1974,1977). This strategy would overcome ploidy level barriers and facilitate introg ression of lower ploidy species into octoploid cultivars. While creation of synthetic octoploids may be diffkult, once created they would not require specialized cytogenetic techniques, such as doubling or counting chromosomes, ta be used in strawberry breeding programs (Evans 1977,1982~). Despite potential benefits, SO production had been abandoned for more than twenty years. The SOS created by Sebastiampillia and Jones (1 976) were the rnost recent attempts in this area. Evans had several papers in the 1980's on the synthetic octoploid system but he last created SOS in 1972 (Evans, 1982~). Guelph SOI, and Guelph S02 have been the only SOS released for use by breeders

63 4 7 (Evans, 1982bc), and two breeding programs, one at Guelph and one at Sweden, have reported using them (Sangiacomo, 1989; Sangiacomo and Sullivan, 1994; Trajkovski; ). lmproving upon the SO system requires elucidation of reasons for difficulties in obtaining SOS. A critical examination of SO literature must consider the level of technology in past investigations and new techniques that have become available in the last twenty years. Crossing strategies to create SOS Several researchers have attempted to create SOS by crossing lower ploidy species directly ta octoploid cultivars (Federova. 1946; Ellis, 1958; Bhanthumnavin, 1965; Jones, 1966; Ellis, in Darrow 1966; Kantor, 1984). Such strategies often relied on colchicine or unreduced gametes and required several generations of crosses between different ploidy levels to restore hybrids to the octoploid level (Fig. 1.5). This type of strategy is designated 'Unbalanced Ploidy Strategy' in further discussion. An alternative strategy will be designated the 'Balanced Ploidy Strategy' in further discussion as almost al1 crosses leading to SO creation were between plants of the same ploidy. The Balanced Ploidy Strategy was adopted by Evans (1977, l982a, l982b) and Sebastiampillia and Jones (1976) (Fig.l.6). They intercrossed lower ploidy species and applied colchicine to increase chromosome numbers and created SOS, which could then be crossed into octoploid cultivars.

64 4 8 Fig Unbalanced ploidy strategies for creating synthetic octoploids. Octoploid cultivars were used in initial crosses, followed by crosses with other species to restore ploidy level to 8x. The diagrams are based on Ellis (1 958), Bhanthumnavin (1 965), Darrow (1 966), and Kantor (1 984). unreduced ametes mutagenic chemicals unreduced 5x gametes - 8x 16x colchicine - a colchicine 2x / colchicine colchicine 8x 8x le)

65 4 9 Figure 1.6. Balanced ploidy strategy for introgression of lower ploidy species into octoploid cultivars. Lower ploidy species are used in initial crosses, and the resulting synthetic octoploid is crossed to cultivars as the last step. Except for crosses using hexaploid species, al1 crosses are between plants having the same ploidy level. Diagrams are based on Sebastiampillia and Jones (1 976) and Evans (1976). -- 4x 8x colchicine 4~ / colchicine 4x / colchicine - 8x colchicine *'> 4x 8x colchicine 52% unreduced - 2x 4x - 8x colchicine colchicine 2x =chicine 4x - 8x colchicine 2~ 4~ / colchicine colchicine

66 50 Germination rates ranged from O to 4.4% for unbalanced ploidy crosses of ten researchers (Evans, 1974). When Evans made diploid x hexaploid and diploid x octoploid crosses, set was 2 seeds / pollination, germination was 25%, but seedlings that resulted did not appear to be hybrids. It was suggested chromosome imbalances between the zygote and endosperm occur in unbalanced ploidy crosses resulted in low seed set and poor germination. Using the parent with the higher ploidy level as the female in crosses was recommended (Evans. 1974). Reciprocal crosses were not done in the Evan's study (1974), but diploid and tetraploid foms of (F.vesca x F.vindis) were crossed as females to F.chiloensis, F.virginiana, and F. x ananassa. Only the combination (F. vesca x F. vindis) 4x X (Enubicola x F.moschata) 8x produced hybrids while similar crosses using diploid (Evesca x F.viridis) as females were not successful. The tetraploids were crossed to 14 octoploid accessions while diploids were crossed to only seven octoploid accessions, so the variation between these two ploidy levels rnight be due to too few accessions in the diploid segment of the study. Despite difficulties, Unbalanced Ploidy Strategies have resulted in a limited number of fertile synthetic octoploids (Table 1.6). Using a unique approach, Kantor (1 984) used mutagenic chemicals (H- nitrosoethylurea and l,4-bis-diazoacetlybutane) to overcome sterility of 7x hybrids produced by crossing F-moschata (6x) to F. x ananassa (8x). He was attempting to transfer desirable flavour from F.moschata to F. x ananassa., but this flavour disappeared during backcrossing so Kantor abandoned introgression in favour of developing 7x iines that were F, hybrids of F. x ananassa X F. moschata.

67 Table Combinations of Fragaria species that have resulted in synthetic octoploids with different fertility levels. Sterile: F. moupinensis x F. moupinensis (Evans, 1964, 1 977) F. vesca x F. moschata (Ellis, in Darrow, 1966) (F. nilgenensis x F. virids) 4x x F. moupinensis (Evans, 1 977) Ferülity unknown: F. moschata X F. viridis (Bhanthumnavin, 1965) F. vesca X (F. vexa 4x x F. x ananassa) 6x (Ellis, in Darrow 1966) F. moupinensis x F. corymbosa (Sebastiampillia and Jones, 1976) (F. rnoschata X F. vindis) X F. corymbosa (Sebastiampillia and Jones, 1976) F. vesca 4x X (Ç. moschata X F. vindis) (Sebastiampillia and Jones, 1976) Partial Fertility : (F. x ananassa X F. moschata) 7x, selfed and backcrossed to Ex ananassa (Federova, 1946; Kantor, 1984) F. moschata X F. nubicola (Bhanthumnavin, 1965; Evans, 1977) (F. x ananassa x Ç. vesca)lux X F. moschata (Nicholson, in Jones, 1966) ( F. vesca 4x x F. x ananassa)6x X(4x F. vesca X 76x F. x ananassa) 70x (Ellis, in Darrow, 1966) F. vesca 8x (Dermen, in Darrow,1966) (F. vesca x F. vindis)4x x (F. moupinensis) (Evans, 1977)

68 52 Balanced Ploidy Strategy usually has tetraploid crossing as the final step prior to chromosome doubling to create octoploids. Tables 1.7 and 1.8 show results of balanced ploidy crosses among tetraploid species and hybrids. Note that germination rates for balanced ploidy crosses are much higher in these tables than the O % germination range reported for unbalanced ploidy crosses (Evans, 1964). All three tetraploid species and seven 'synthetic' tetraploids have been used in crosses at the tetraploid level. Of these, F.vesca (4x), (F.vesca x F. vindis) 4x and (F.nilgenensis x F. vindis) 4x are the only synthetic tetraploids derived solely from diploid species. Clearly, further research is warranted regarding use of tetraploids derived from diploid species. Strategies to create tetraploids from diploid species include: creation of auto- and allotetraploids, and crossing diploids to F.moschata (6x). Many. but not all, combinations of tetraploid species and hybrids have been hybridized to one another. Besides yielding higher numbers of seeds, the Balanced Ploidy Strategy required only 1-2 generations to achieve an octoploid, while two to four generations are required for an Unbalanced Ploidy Strategy. Using a Balanced Ploidy Strategy, Evans (1982~) successfully introgressed three SOS into octoploid cultivars; no reports were published on the eight SOS created by Sebastiampillia and Jones (1976). Only Evans (1977, 1982a, l98zb, 1982~) has given a full account of creation, fertility and introgression of synthetic octoploids. When the two synthetic octoploids, Guelph SOI and Guelph S02, were backcroçsed into F.x ananassa it was demonstrated that important SO traits could be introgressed into cultivars and that fertility was rapidly restored in two or three crosses to cultivars (Sangiacomo and Sullivan, 1994). Thus, it has been

69 53 Table 1.7. Results of studies on hybridization of Fragaia tetraploids. These studies did not atternpt to create synthetic octoploids, but are included since tetraploid hybridization is part of Balanced Ploidy Strategy for synthetic octoploid production. Evans,l964: Fernales' F-moupinensis (1) * F.moupinensis (7) poilin- combin- seeds % Males ations ations Y Ipo11. gem. F.nilgemnsis x F. viridis (3) 91 F. vexa x F. viridis (2) da F.nilgemnsis x F. vitidis (4) F.nilgemnsis x F. viridis (3) F.nilgemnsis x F. vindis (3) F.orientalis (7) F-orientalis (1) F-orientalis (7) F. vesca x F. vindis (5) F. vesca x F. vindis (5) F. vesca x F. vindis (5) F.moupinensis (7) F. orientalis (7) F.vesca x F. viridis (4) F.moupinensis (1) F.nilgenensis x F. vindis (1) F. vexa x F. vindis (3) F.moupinensis (1) F.nilgenensis x F. viridis (3) F. orientalis (1) da nia nia II Females Males seedibeny % germination F. vesca (4x1 F. viridis x F. moschata F. vesca (4x) F.moschata x F. viridis Females Males Progeny F. orientalis F- corymbosa 765 F. vesca (4x1 F. onentalis 67 F. vesca (4x) F. vesca (4x) X F. orientalis 44 F. vesca (4x1 X F.orientalis F. orientalis 499 2: Except for F.moschata (6x1 x F-nubicola (24 from Evans 19ï7, al1 female and male parents in this table are tetraploids Y: combinations are the number of crosses with dierent accessions as parents. X: number in ( ) indicates number of accessions available

70 54 Table 1.8. Results of attempts to create synthetic octoploids (SOS) from 2x, 4x and 6x Fragana species. The resulting progeny from these crosses were treated with colchicine to create synthetic octoploids which could eventually be hybridized with 8x cultivars of F. x ananassa. Except for the Emoschata x F.nubicola crosses of Evans (1977) the final crosses made were between tetraploids. Females Males seedlings survivors SOS Fertile reference after SOS colchicine F.moupinensis (4x) F. corymbosa (4x) Sebastiampillia and F. onentalis (4x) [F. moschata (6x) x F. nubicola (2x114x O Jones, 1976 IF. vesca (2x) x F. viridis (2x)] 4x [F.moschata (6x) x F.nubicola (2x114x O IF. moschata (ex) x F. viridis (2x)] 4x F. orientalis (4x1 8 4 O [F.moscheta (6x) x F. vindis (2x)]4x F. corymbosa (4x) F. vesce (4x) [Emoschata (6x) x F. viridis (2x11 4x F,moschata (6x) F. nubicola (2x) Evans, 1977 [F.moschata (6x) x F.nubicola (2x11 4x F.orienta/is (4x) O O [F. moschata (6x) x F.nubicola (2x11 4x [F. vesca (2x) x F. vindis (2x11 4x O O F.moupinensis (4x) [F.nilgerrensis(2x) x F. viridis (2x)] 4x O O F. moupinensis (4x) F.moupinensis (4x) O If.ni/gemnss (2x1 x F. viridis (2x11 4x F.moupinensis (4x) O /F. vesca (2%) x F. vindis (2x11 4x F. moupinensis (4x) [F. vesca (2x) x F. vindis (2x)]4x [F. vesca (2x) x F. viridis (Zx)] 4x O O

71 5 5 demonstrated that the synthetic octoploid system can be used to introgress traits from lower ploidy species into octoploid cultivars. The group lead by J.K.Jones at the University of Reading (Ellis, Bhanthumnavin, Evans. and Sebastiampillia) probably used the same group of accessions for their research. Thus, synthetic octoploids produced by this group may have been derived from a very narrow germplasm base. Evans' (1 964) thesis lists the available accessions as follows: F. vesca (1 3), F.nubicola (2), F. vindis (1 ), F.nilgerrensis (1 ), F. daitoniana (1 ), F. orientalis (1 ), F.moupinensis (2) and F.moschata (4). This group originally used Unbalanced Ploidy Strategy (Fig. 1.5) but later abandoned it in favour of Balanced Ploidy Strategy (Fig. 1.6). While Balanced Ploidy Strategy rnay have been a more successful approach, research did not continue beyond a few attempts. Only a few of the possible combinations of species have been investigated for creation of SOS. With ten diploids, three tetraploids and one hexaploid species, theoretically 12,769 combinations of species could result in SOS within two generations. A simplification of these possibilities, including creation of tetrahaploids is presented in Figure 1.7. This figure is similar to the crossing strategy for this thesis except that tetrahaploids derived from octoploid species (tetra-haploids) will not be investigated. lmproving SO production While sorne researchers developed techniques that potentially could improve SO production, there has not been an effort to bring these techniques together for SO production. For instance, Sebastiarnpillia and Jones (1 W6), developed a

72 56 Fig A simplified Balanced Ploidy Strategy for synthetic octoploid production. This diagram simplifies crossing strategies depicted in Fig 1.6 but is broader in scope. Breeding at diploid and tetraploid levels as well as creation of tetrahaploids is incorporated into this strategy. Diagrams show ploidy levels of species and hybrids involved in crosses. Parents are on left and progeny on right. Tetraploid creation Diploid breeding chromosome doubling Tetraploid breed ing Octoploid creation 4x 8x -- chromosome doubling anther culture androgenesis chr. elimination

73 57 superior colchicine application technique that was not used by Evans (1977, 1982~). Test crosses to discover combinations of accessions with high seed set was likely used by Evans (1 977) who produced 6564 hybrid tetraploid seed by using the successful combinations of parents identified in an earlier study (Evans, 1964). A wider germplasm base increased the likelihood of successful interspecific hybridization (Burbank,1921 ; Galletta and Maas, 1990; Stoskopf et al., 1993) and is an obvious way to improve SO production. In addition to test crossing, pollination of the same flowers on different days has increased seed set (Evans, 1964). lmproved germination methods (Jonkers, 1958; Scott and Draper, 1967; Wilson et al., 1973; Miller et ai., 1992) were also not employed in SO research. If al1 these techniques were incorporated into one rnethodology, it seems likely that the efficiency of the SO system would be greatly improved. lmproved strategy for SO production Previous research showed that a Balanced Ploidy Strategy was more likely to be successful in the creation of synthetic octoploids than an Unbalanced Ploidy Strategy. Such a strategy requires fewer generations and gives higher seed production than an Unbalanced Ploidy Strategy but also requires the sometimes difficult step of chromosome doubling. Previous investigations of SO production have been hampered by lirnited availability of gennplasrn, particularly diploid species. As more species become available through increasing holdings of genebanks, intercrossing of al1 available accessions becomes less feasible. There has not been an attempt to focus on specific species to improve SO

74 58 production. As mentioned earlier, Fragana vesca not only possesses traits that would be beneficial to introgress into cultivars, but has several unique attributes that could improve introgression strategies. Its ability to hybridize with octoploid cultivars has been shown but it has not been used as a bridge to aid introgression of other species into cultivars. It is hypothesized that a SO strategy focusing on use of Evesca as a bridge species would be an improvement over earlier attempts to create SOS.

75 59 Chapter 2 General materials and methods Introduction This chapter details the general materials and methods used throughout the thesis. Most of these techniques were adapted for this study and represent technology not used in earlier attempts to create synthetic octoploids (SOS) in Fragaria. Synthetic octoploid production has been abandoned for more than 20 years, largely because the system was inefficient. Difficulties previously encountered such as poor fruit set, low germination rates, chromosome doubling and small chromosomes that were difficult to count, have been addressed by the methodology presented in this chapter. In this study, several modern techniques have been applied simultaneously to improve the SO system. ln vitro germination The in vitro germination technique O f Miller et al. (1992) was modified for use in this research. This technique was originally developed for use with F.x ananassa seed, but was adapted for seeds resulting frorn interspecific hybridizations. Murshige and Skoog medium (Sigma-Aldrich Canada Ltd., Oakville, ON.) was prepared with no hormones as descrîbed by Miller et al. (1992) but Gellan gurn agar (Schweizerhall, South Plainfield, N.J.) at 2.5gi1 was substituted for agar. Acid washed, activated charcoal (Gibco Laboratories 1 Life Technologies Inc., Grand Island, NY) was added at 3.0 g/l to darken the medium. In a previous study, shoots

76 60 of seedlings germinated on clear medium often grew upside down or laterally. Shoots in darkened media had a higher frequency of upright growth, which is an important characteristic for colchicine treatment. Medium (7.5mls) was placed in 30ml glass French squares, autoclaved (mode! 3021, AMSCO, Erie, Pa) at 121OC for 20 minutes and aliowed to solidify in a slanted position on racks. For surface sterilization, achenes were wrapped in cheese cloth held by a paper clip, dipped in a 0.5% (vlv) sodium hypochlorite (Lilo products, Hamilton, ON) solution containing tween 20 (2~111) and placed on an orbital shaker (Labline Instruments, Melrose Park, III.) for mins. Agitation continued through three transfers into vials of sterile deionized water to remove the sodium hypochlorite solution. Achenes were placed on moistened sterilized filter paper to prevent desiccation, and the distal third of each achene was removed with a scalpel and discarded. Cutting occurred under a dissecting microscope (AusJENA, Gei-many) in a larninar flow hood (Canadian Cabinets Co. Ltd., Ottawa, ON). Cut achenes were placed in vials on the surface of the medium described above. These vials were placed on racks and in a tissue culture growth room maintained at 20 C, 30cm below 40 watt cool white flourescent lights (Sylvania, Mississauga, ON) at 70 pmol m-2 - s*~. Seeds usually germinated and had fully expanded cotyledons within 10 days of being cut and placed on the media. Colchicine treatrnent for chromosome doubling Colchicine (SigrnaAldrich Canada Ltd., Oakville, ON.) was dissolved in 95%

77 61 ethanol and diluted with distilledideionized water for final concentration of 50 gm colchicine / litre. This solution was applied to seedlings days after placement of cut achenes on media, which coincided with full expansion of cotyledons and before expansion of the first true leaf. Colchicine solution was applied as drops (10ullseedling) with a micropipettor to the meristematic region between cotyledons of in vitro seedlings (Fig. 2.1). The sterile environment was not maintained for colchicine treatment since seedlings were to be removed from tissue culture within 20 hours. Vials were sealed to rnaintain humidity, and returned to the growth room. Vials were filled with deionized water to provide one quick rinse after hours of colchicine treatment. Acclimatization of seedlings ARer in vitro germination or colchicine treatrnent, seedlings were transplanted to plug trays (72 plugs / tray, Kord Products Ltd., Blarnalea, ON) for acclimatization. The plug trays contained soilless mix (Promix BX, Premier Horticulture Inc., Red Hill, Pa.). Plug trays were placed for 3-4 weeks in a growth chamber (Model EY8M, Controlled Environrnents Ltd., Winnipeg, Manitoba) with continuous light provided by 40 watt cool white flourescent lights (Sylvania, Mississauga, ON) at 15-17C before being moved to the greenhouse. Plants were bottorn watered with deionized water. When plants grew larger, they were transplanted into larger pots containing Promix BX. Seedlings were raised in the greenhouse until planted into the field. Greenhouse conditions consisted of : 18/14 C daylnight temperature; 16 hour daylength provided by 400W high pressure sodium supplemental lighting (PL

78 62 Figure 2.1. In vitro treatment of Fragaria vexa seedlings with colchicine solution. Note the droplets of colchicine solution placed on cotyledons.

79 63 Lighting, Grimsby, ON); and 60% relative humidity during growth stage. Plantings into the field occurred in either early June or late August of 1995, 1996, and Plants were field grown at Cambridge Research Station, Cambridge, ON in Fox sandy loam soil. Handling of parent plants for crossing In early November, dormant plants were dug from the field and stored bare rooted in a freezer at -1 C. From late December to early March, plants were potted with a soilless mix (Promix BX, Premier Horticulture Inc., Red Hill, Pa.) in either 3 or 6 litre pots (Plant Products, Brampton, ON) and grown in the greenhouse at Guelph, ON. Plants were watered as required with deionized water and fertilized with an all-purpose fertilizer solution [20N - 8P - 20K (Plant Products, Brampton, ON ) used in the Guelph greenhouse at a concentration of 267mgllitre solution that had been brought to a ph of 6.0 with phosphoric acid]. Floral induction F.nubicola, F.pentaphy//a, F. daltoniana, F.iinurnae, F. orientalis, and some hybrids derived from these species did not bloom under the greenhouse conditions described above. They required additional short day and temperature treatments (four weeks, 10 hr photoperiod, 18C day, 15C night) (Jonkers, 1958) in a growth chamber (Model EY8M, Controlled Environments Ltd., Winnipeg, Manitoba) to induce bloom. The flower induction treatment was applied in late winter or early spring to greenhouse plants that showed no signs of blooming.

80 64 Pollination Emasculation was carried out according to Scott and Lawrence (1975). Each moming, anthers and petals were removed with sharp tweezers when buds were in the "popcorn" stage (i.e., buds were in the early stages of opening with white petals emerging). Anthers were placed on petri dishes in the dark at C and allowed to dry and dehisce overnight. Pollen was stored according to the protocol shown in Fig Microcentrifuge tubes (1.%?.O mls.) containing pollen were placed in a Magenta box (#GA7-3. Magenta Corp., Chicago, III.) half filled with selfindicating desiccant (# & Drierite, Fisher Scientific. Nepean, ON). Magenta boxes were stored in styrofoam boxes and placed in a refrigerator at 3 to 4 C. Often, dried anthers were crus hed to release pollen within microcentrifuge tubes using the handle end of paint brushes with an action similar to a rnortar and pestle. Flowers were emasculated and pollinated the same day. Camel hair brushes were used to spread pollen ont0 stigmas of emasculated fiowers. For diploid crosses, only one primary bud was pollinated per truss. As tetraploid flowen were larger than diploid fiowers, primary, secondary, and occasionally tertiary fiowers were also used, if receptacles were greater than 3 mm in diameter after emasculation. Small buds often showed signs of wilting on sunny days and were not used. High humidity (80-90%) in the greenhouse. and removal of fully opened flowers reduced chances of selfing. Flowers were not bagged after emasculation.

81 65 Fig Pollen storage method based on procedures of MacFarlane Smith and Jones (1989). Pollen was protected from light by placing the storage container in a box kept in a refrigerator at 3C. Air tight Container - u-- - '-C % - --' Desiccant Microcentrifuge Hole in top Non-absorbant cotton Anthers and Pollen

82 66 Organization of crosses Crosses were done over a period of time as different genotypes flowered. Within each group of species or hybrids, pollen from various genotypes was randomized each day and used in crosses. An attempt was made to pollinate each female with equal number of the different species used as pollinizers when possible. Parents with common ancestry were not crossed together, thereby fostering greater heterozygosity in offspring. Choice of parents was made to maximize the number of male/female combinations and, thus, sample a greater amount of diversity. Pollen from genotypes that had never been used was given prîority over previously used genotypes. Seed collection and storage Ripe berries were either crushed on petri dishes or placed intact into airtight jars containing desiccant. Crushed berries were air-dried at room temperature. Once dry, achenes were rubbed off. counted and stored until use. The conventional method of extracting seed in a blender was not used to avoid damage to achenes which might allow penetration of bleach into embryos during the in vitro sterilization step. Only large plump seed were collected and stored. Intermediate and smallsized seeds contained no embryos in as determined in previous investigations. Seeds were stored at room temperature or in a refrigerator at 3 to 5 C. Data collection and analysis Variables studied included % fruit set, seedslflower, % embryos, and %

83 67 embryo germination. Percent set referred to the number of flowers that produced at least one seed divided by the number of flowers pollinated. Large. plump seeds were counted for each flower to provide seed numbers. Percent ernbryos was the percentage of seeds that contained fully developed embryos with a healthy appearance. These data were obtained when achenes were cut in half for in vitro germination. Only seeds containing white coloured embryos and cotyledons were considered healthy. As semi-filled seed with white interiors were capable of germination, they were classified as developed ernbryos. Previous results had shown that achenes with yellowish or light brown interiors were not viable and were categorized as dead. Emb~yo germination was the ratio of healthy seedlings 1 embryos. Although almost al1 healthy embryos grew when placed on MS medium, only those that developed into normal green seedlings were counted. Abnormal seedlings that were albinos, lacked hypocotyls or roots, or failed to develop true leaves were occasionally observed but were not counted as geninated. Data were analyzed with the GLM program of PC SAS (Sas system for Windows V Cary. N.C.) and means separation was accomplished using the LS means function. Nomality of data was checked with SAS and arc-sine transformations were done on those data which did not appear to have normal distribution. Visual screening of plants following colchicine treatment Visual screening was used to identify hybrid progeny which showed morphological evidence of chromosome doubling at four to six months after colchicine treatment. Hybrids that had slower rates of development (see Fig. 2.3)

84 68 and larger, thicker, darker leaves with more pronounced serrations (Fig. 2.4) compared to their parents and siblings were set aside for flow cytometry testing as described in the next section. Flow cytometry protocol Flow cytometry has been used to estimate ranges of 8x to 16x for F x ananassa accessions derived from protoplasts (Nyman and Wallin, 1992), but a flow cytometry protocol has not been developed for Fragaha with ploidy levels lower than 8x. Nyrnan and Wallin's (1992) flow cytometry protocol required creation of strawberry protoplasts as a first step. This step would have been quite difficult to replicate on a variety of Fragana species and would have increased the time required to obtain ploidy level estimates. A protocol being used at the University of Guelph for analysis of Lycopersicon esculentum (tomato) was adapted for use with Fragana species. The nuclei isolation solutions (Table 2.1) and nuclei isolation protocol (Table 2.2) were derived from Arumuganthan and Earle (1991) and Bino et al. (1992) as modified by Shawn Rogers (unpublished). Sample size, chopping time, centriiugation tirne. propidium iodide concentration, and standards were refined for use with Fragaria species in this study and have been included in Table 2.2. The Coulter Epics Elite ESP Row cytometer, manufactured in 1994, was used to estimate ploidy levels. Flow cytometer settings are provided in Table 2.3. The flow cytometer was set to read a maximum of 5000 cells for each sample. When intemal standards were used 10,000 cells were counted. Often, distinct peaks were

85 69 Figure 2.3. Diploid (left) and tetraploid (right) forms of the "Alexandria" variety of F. vesca ssp. vesca var. sempeflorens at identical ages. The diploid has ripe fruit (indicated with arrows) and greater number of leaves, indicative of a faster growth rate. The tetraploid plant has just begun to bloom.

86 70 Figure 2.4. Leaf morphology of interspecific hybrids of Fragaria species. Diploids, tetraploids and octoploids are located on the lower, middle and upper rows, respectively. Leaf serrations along leaf edges and leaf sire increase at higher ploidy levels.

87 Table 2.1. Nuclei isolation solutions. These solutions are used during chopping, staining, and running of samples. Solution A: 15 mm HEPES (H-3375 Sigma, Oakville, ON) 1 mm disodiurn EDTA (BP Fisher Scientific, Fairlawn, N. J.) 80 mm KCI (3040 J.T.Baker, Phillipsburg, N.J.) 20 mm NaCl (S-9625 Sigma) 300 mm Sucrose (Redpath, Toronto, ON.) 0.2% Triton X-1 O0 (T-9284 Sigma) 0.5% Spermine tetrahydrochloride (S-2876 Sigma) 1.O% PVP (MW=40,000) (PVP-40 Sigma) ph adjusted to 7.5 store in aliquots at -20C add 0.1% 2-mercapto-ethanol (R-4875 Sigma) prior to use Solution B: 2.5 ul RNAaseI ml of Solution A (10mglml stock RNAase) (R-4875 Sigma) 10 ul Propidium lodide Solution per sample (P-3566 Molecular Probes, Eugene, N.J.) Note: solution is 1.0 rnglrnl stock)

88 Table 2.2. Nuclei isolation procedure. 1. Leaf tissue (4 cm2 per plant) was harvested, placed in a in petri dish and put on ice. Fully expanded young leaves were preferred but old leaves were occasionally used. 2. During maceration, petri dishes containing the leaf sample, centrifuge tubes and filters were on top of ice. Solution A (1.O ml) was added to leaf samples that were then chopped with a razor blade until leaf tissue attained a liquid consistancy and the suspension started momentarily sticking to the razor blade foming strings. After 'string' fornation was noted, chopping continued for about 15 seconds. Often, chopping lasted minutes per sample. but this varied according to speed of chopping and species. Blades were used on a maximum of 5 samples. 3. Homogenate was filtered through 100 Fm and then a 30 pm nitride mesh into a microcentrifuge tube. 4. Optional step: Interna1 standard: Diploid trout red blood cell solution (Ten cytornetry control, #IOOï, Biosure, Grass Valley, California) was diluted 30 MI per Iml of solution B and filtered through a 30 pm nitride mesh. External standard: Desiccated chicken red blood cells (CRBC) (#R0504, Sigma, Oakville, Ontario) were suspended in distilled HZ0 at 25mglml. This suspension could be stored for a few months if refrigerated. CRBC suspension was mixed at a rate of 20 pl 1500 pl solution B. 5. Samples (from step 3) were vortexed for 5 seconds and then centrifuged for four to five seconds (time from start, not from peak speed). Supernatant was discarded. 6. Solution B (200 pl) was added to each sample. 7. Optional step: 20 pl of interna1 standard solution (step 4) was to each sample. 8. For extemal standards, 20 pl of intemal standard stock solution (step 4) were added to 200 pl of solution B. 9. Samples and exterml standards were vortexed and incubated at 37' C in darkness. After 15 minutes, they were removed from the incubator and placed in the darkness at room temperature. Sarnples were used within three houn and vortexed immediately before nins through the Row cytometer.

89 Table 2.3. Settings for the Coulter Epics Elite ESP fiow cytometer when running Fragana nuclei with trout red blood cells (TRBC) as an interna1 standard. For the PMT3 sensor a red FI filter was used. The 488 nm argon laser was used as the light source. PROTOCOL: STRMJBERRY NUCLeI srnmom FS PHT1 PlIl2 PHT3 PHT4 PM6 AüXl AUX2 Uoltm f e 409 1Q.h t.e a f.0 POIin e L.0 1. O p--p oœly LOU LOU LOU LOU LOU LOU LOU LOU Flwrercrnci Conpmnmat ion mtmd ~ m p Si-1 out (Y> Y - XX PARAPI A PHT~ LOG y X ~ PMI ~ mf3 l PHT~ PP~TE QUXL ISUXZ PC~RA~? I P~TS LOO Pm C3WOi.c OFF PM fr Sourc FS PEAK PPnr 0.0 9s 0 Clock S. 0 nt4 O O Un UIdth 25.0 PHTE B Un Omlru 60.0 AUX U T Oimc 09 R W a.0 O. B tr Omin 8 Sq Olhy 20 ui Lao Lack OFF Scmpm Trmcm CnbiOibl OFF Pluidici (psi) tlist811an~our f.iph Shmath arnctwidtk LOU 11.a t2. OB 0i.c 9rt ~ x t 6.0- Timiûamm La.@ us FLR(CT1OHS Ckting: ON nutonmlyria: OFF Eurnt Colar: OFF S toptimm t OPf ortsmttingr: OFF Expart: NoSaum Lfmhadi: NoSauo Cytokttlngm: Of P

90 74 observed before maximum cell count was achieved and a run could be terminated early. Examples of flow cytometry output c m be seen in Figures 2.5 and 2.6. Flow cytometry standards and the estimation of ploidy level For most of this thesis the flow cytometer protocol was used with external standards consisting of known 2x, 4x, 6x. and 8x Fragaria species that were run at the beginning and end of each session. Later, interna1 and external standards were adopted which simplified estimations of ploidy levels. Calculations used to estirnate ploidy levels of hybrids are presented in Table 2.4. These calculations were used in conjunction with standard channel numbers of Fragana species (Chapter 3) to estirnate ploidy levels.

91 Figure 2.5. Example of flow cytometer output for a 4x-8x mixiploid Fragana hybrid with an internal standard. Peak 'A' is the trout red blood cell internal standard peak. Peaks 'B' and 'C' correspond to tetraploid and octoploid peaks of the Fragana hybrid, respectively. Data below 50 on the X axis were not recorded because background noise increases geometrically as the x axis approached zero. Numerical output is presented below the graph. The mean value under the data column 'x channel' was used for estimation of ploidy levels. SXNQLE PARRMETER STATISTICS.O*.roPmmk+* O...t..o...X Channml ID Pmt Amr Position Hmiaht Hœrn 90 FulLCü HmlfCU Min Mmx CI % a TI C

92 Figure 2.6. Example of flow cytometer output for a hexaploid Fragaha plant. Peak 'A' is the trout red blood cell infernal standard peak. Peak '8' is the hexaploid peak from the plant sample. Data below 50 on the X axis were not recorded because background noise increases geornetrically as the x axis approached zero. ~umerical output iç presented below the graph.... Pmak... SINQLE PARA?iET R STRTISTICS-... X Chamml Pcnt Wrr Padtion Hright Mian 50 FullCU HalfCV Min Max CI X f s 1.18 lai 246 R Lou Chrnnri = 742, Hiph Chmnnrl = 871 B B Low Channal = 161, High Chrnnmi = 205 8

93 Table 2.4. Calculations used to determine chromosome numbers for Fragana hybrids from Row cytometry channel numbers. samples with multiple external standards XE Sarnple's channel number Standards Correction C Standards' channel numbers from table 2.5 D'A contents x = sample pg DNA sarnples with intemal standards Sarnple's channel number Standard's DNA content Correction = çample ~CJ DNA Standard's channel number from table 2.5 factor expected DNA level Add together the DNAlgenome values from Table 2.5 according to species in lineage of hybrid king tested. Sam~le DNA chromosome number estimate ExpeCted Sample X chromosome = chromosome Expected number number Z The correction factor for 2,4.6, and 8 x samples is , 1.045, and 1.060, respectively.

94 78 Chapter 3 Fragaria species differ in DNA content ABSTRACT Flow cytometry was used to detenine DNA content of diploid ( F. daltoniana, F.gracilis, F. Nnumae, F. nilgenensis, F. nipponica, Enubicola, F.pentaphylla, F. vindis, F. vesca), tetraploid (F. onentalis), hexaploid (F.moschata) and octoploid (F. chiloensis, F. virginiana) Fragana species to provide a baseline for determining ploidy level of interspecific hybrids and colchiploids. Fragana species ranged frorn 0.62 pg (octoploid species and F.nubicola) to 0.82 pg DNN genome (F. daltoniana). This variability among species indicates that ploidy levels of interspecific tetraploids and octoploids need to be based on a cornparison for DNA levels with their parental species rather than natural tetraploids or octoploids. Cornparison of 242 samples having primary (GA period cells) and secondary (G2 period plus mitotic cells) peaks indicated that a doubling in ploidy results in flow cytometry readings being approximately 3% lower than expected. This discrepancy should be taken into account when estimating ploidy levels of colchiploids. The flow cytometry protocol developed was successfully used to differentiate ploidy levels of interspecific hybrids and colchiploids.

95 INTRODUCTION Understanding DNA content of Fragaria species is important for the development of synthetic polyploids. Differences in DNA content among species would be important to consider in estimating ploidy levels of hybrids and may also provide insight into evolution of Fragana species. Although flow cytometry has been used to estimate ploidy of 8x and 16x Fragana x ananassa protoplasts (Nyman and Wallin, 1992) this technique has not been used to rneasure nuclear DNA content of other Fragana species. Chromosome counts using root tip squashes have been used by many researchers for ploidy verification in Fragah but the number of cells counted was not included in their materials and methods (Mangelsdorf and East, 1927; Scott, 1951; Niemirowicz-Szczytt, 1987; Sayeg h and Hennerty, 1989; Rose et al., 1993). Other researchers have published chromosome counts based on 16 to 26 cells (Evans and Jones, 1967) or 83 to 253 cells per accession (Mok and Evans, 1971). A flow cytometer can estimate DNA levels of thousands of cells in a few minutes and is an efficient way to estimate ploidy levels; it can also be used to estimate genome size. Quantification of Fragaria genome size with flow cytornetry has not been reported. Species of Malus (Dickson et al., 1992). Honleum (Kankanpaa et al., 1996) and Eucalyptus (Grattapaglia and Bradshaw, 1994) have been shown to differ in their DNA contentlgenome. The objective of this study was to establish DNA levels of several Fragana species and determine if flow cytometry output was linear as ploidy level increases,

96 80 (Le. if the channel readings of a tetraploid nuclei are twice the channel readings of a diploid nuclei). These data would be needed to estimate ploidy levels of interspecific hybrids. MATERIALS AND METHODS Plant material was grown in the greenhouse at the University of Guelph according to procedures described in Chapter 2. Table 3.1 lists the accessions used in the experirnent. Isolation of nuclei and measurement of DNA content were done with the procedure and flow cytometry settings described in Chapter 2. The linearity of DNA content was tested using channel readings of primary and secondary peaks of 242 samples of Fragana nuclei. These samples were obtained from various hybrids and species with different ploidy levels produced at the University of Guelph. In fiow cytometry output, primary peaks are larger than secondary peaks and represent nuclei from cells in a resting stage. Secondary peaks occur for nuclei which have replicated their DNA but have not yet divided (Dickson et al., 1992). Cells which have only partially replicated their DNA would have fiow cytometry readings between the values of primary and secondary peaks. An example of a secondary peak can be found in Figure 3.1. To calculate linearity of flow cytometry readings, the mean channel number of the pnmary peak was plotted as the independent variable and secondary peaks were plotted as the

97 Table 3.1. Gennplasrn used for fiow cytometry testing of Fragana species. Species Guelph #Identifier used by source SourceZ F. daltoniana J. Gay F.gracilis Losinsk. F.iinumae Makino F. nilgenensis Schltd. ex J. Gay da12 gra 1 gra2 ii3 ni dal-2 Pl CFRA CFRAI OO8.OOO CFRA RBG Edin burg h NCGR NCGR NCGR NCGR ni12 Mt. Omei, frch059 Forest Farms ni14 CFRA NCGR ni15 CFRA NCGR ni16 CFRA NCGR F. nipponica Makino F.nubicola Hook. n ip4 nubl nub2 CFRA PI PI NCGR NCGR NCGR F.pentaphy/la Losinsk. penl pen RBG Edinburgh RBG Edinburgh F. viridis Duchesne virl vir2 PI PI NCGR NCGR vir RBG Edinburgh F. vesca L. vcvl 1 vcv10 Ruegen Alexandria Richters Thompson and Morgan vw12 PI NCGR vw39 CFW 988 NCGR vw53 Newcastle, Ontario collected by author Eorientalis Losinsk. F.moschata Duchesne ori? mcvl mcv x 4940 PI PI Gunter Staudt NCGR NCGR rnwl PI NCGR F. chiloensis L., BCy sm366 Fra 366 Ag Canada, Harrow sm777 Fra 777 Ag Canada, Harrow F. chiloensis, SAX sm883 Fra 883 Ag Canada. Harrow cl 077 CFRA 1077 NCGR cl O89 CFRA 1089 NCGR F. virginiana Mill. cl 110 CFRA 1110 NCGR vg BRB Blind River Beach, Ontariocollected by author vg4 Spanish, Ontario collected by author v ~ 7 Manitoulin Island. Ontario collected by author 2: details regarding sources of gennplasm can be found in appendix 1. Y: F.chiloensis from British Columbia, Canada X: F.chiloensis from South Arnenca

98 Fig Example of a secondary peak in fiow cytometer output. The peak designated 'A' is the TRBC interna1 standard peak. 'B' is the primary peak while 'C' is the secondary peak.

99 83 dependant variable. Regression analysis was done with Microsoft ExceIl97 SR-1. To determine if Fragana species Vary in amount of DNA, nuclei from the accessions listed in Table 3.1 were isolated and analyzed by flow cytometry. Samples of the accessions were run as a group on three separate occasions and treated as three replications. A minimum of 5 x 103 nuclei were analyzed per sample. Calculations converting channel numbers into pgdna were based on chicken and trout red blood cells as external standards as described in Chapter 2. Also, DNA estimates were adjusted to account for the distortion in flow cytometry output detemined by comparing primary and secondary peaks. The GLM prograrn in SAS (release 6.12, SAS lnstitute Inc., Cary, N.C., USA) was used for data analysis with means separation done with LS means at a 95% confidence level. RESULTS AND DISCUSSION Although fiow cytometry channel numbers were significantly linear when comparing prirnary to secondary peaks (Figure 3.2), the value derived for the slope was less than expected. Since the y-intercept value of is negligible when channel numbers are in the range of , it would be reasonable to expect secondary peaks to have twice the channel numbers of primary peaks, or a dope of Channel numbers of secondary peaks would be expected to be twice amount of primary peaks, however, the x coefficient is with a standard error of Thus, the channel nurnbers of secondary peaks were 3.0% % less than expected. This distortion could be particularly important when calculating expected channel numbers for polyploid hybrids. The standard error of the y

100 84 Figure 3.2. Regression of channel numbers from secondary peaks ont0 channel numbers of primary peaks for Fragada nuclei. Channel readings of 242 samples were used in the calculation of this regression. Channel numbers of primary peaks

101 85 estimate was A possible explanation for this phenomenon lies in the dynamics of a sphere, and light passing through, or reflecting upon a suspension within it. Two spheres will reflect more light than a larger sphere that has the same volume and concentration of a suspension. In a similar manner, tetraploid nuclei rnay contain the same amount of DNA as two diploid nuclei but its surface area would be less, and thus would be less efficient at reflecting or absorbing laser beams from the flow cytometer. Nuclei moving througn a suspension in a Row cytometer are probably not exactly spherical, nor is it likely tetraploid nuclei have exactly twice the volume of two diploid nuctei, so it would be impractical to derive a mathematical model describing this phenomenon until more data become available. To adjust for the 3% distortion, channel numbers of 4x, 6x, and 8x species were multiplied by 1.03, 1.O45 and 1.O6 when calculating pgdnngenome for Table 3.2. More study is required to determine if this 3% distortion is a universal phenomenon with flow cytornetry studies or if this is specifically applicable to Fragana nuclei or the flow cytometer used in this study. it might be suggested that the multiple peaks of preparations of chicken red blood cells do not show this distortion, but it should be kept in mind that such multiple peaks are created by the dumping of individual cells. A clump of unifon individual cells is not the same phenornena as smaller and larger sized nuclei with variable DNA content. While a 3% distortion in channel numbers when ploidy levels are increased may not be important for many flow cytometry applications, it could be important in Fragana research. In particular, an attempt to create synthetic octoploids from diploid species would result in ploidy bels being doubled twice, which could result in a 6%

102 Table 3.2. Variation in DNA content of various Fragaria species, accessions and blood cell standards. DNA content (pg DNA) / genome Accession Species means Ploidy S pecies Ploidy Accession means SEZ mean F. daltoniana 2x da abdy 0.70 F. gracilis 2x gra b,d-f,g-i q ra Eiinumae 2x ii cd,f-l F. nilgerrensis 2x ni c-f ni Fnpponica 2x nip a-f F.nubicola 2x nubl i-l nub F. pen tap h ylla 2x penl b-h F. vesca 2~ vcv f-k vcvl O vw vw vw F. vindis 2x virl d,g-l vir F. orientalis 4x oril b-e,gi 0.74 F-moschata 6x mcvl e,g-j 0.69 mwl F.chiloensis 8x cl i-l 9.62 sm F. virginiana 8x ' i-l vg vg BRB mean Blood cell standards Chicken red blood celts, reconstituted 0.58 Chicken red blood cells, Biosure 2.50 Trout red blood cells, Biosure 6.13 ': SE based on three samples letters are LS means groupings at 0.05 level ': DNA values derived by the author *: DNA values frcm Vinogradov (1998)

103 87 distortion. Flow cytometry results indicate some Fragaria accessions (ie. sri777 and da12) differ by as much as 42% in amount of nuclear DNAIgenome (Table 3.2). As DNA content varies in Fragaria species, calculations of expected DNA Row cytometry readings of Fragana hybrids levels should be based on DNA levels of species in the lineage of hybrids. For example. a synthetic polyploid derived from the cross F-moschata x F.vidis should not be compared to the 8x species Evirginiana to determine if it is an octoploid, rather the DNA levels of Emoschata and Evindis should be added to derive an expected value for an 8x hybrid. Octoploid Fragana and the diploid Fmbicola have the lowest DNA/genome of Fragaria species. Fragana octoploids may have lost DNA in their evolution. Perhaps at the octoploid level. eight copies of some alleles are disadvantageous and a genotype with deletions of certain alleles rnay have a survival advantage. It has been proposed that smaller sized chromosomes rnay have evolutionary advantages for polyploids when it was observed with microscopy that in Dianthus, Chrysanthemum, Narcissus, and Tulipa genera polyploid species have smaller chromosomes than diploid species (Darlington, 1973). Conversely, perhaps there is an advantage for lower ploidy species having larger genomes. Fragana daltoniana is native to elevation of 5000 meter above sea level in the Himalayan mountains (Staudt, 1989) which may be the harshest environment where Fragana species exist. The high DNA content of this species may be indicative of mutations acquired that have allowed it to adapt to such an environment. Perhaps the inability of many researchers to produce tetrahaploids from octoploid cultivars (Hughes and

104 88 Janick, 1974; Predieri et al., l989b; Sayegh and Hennerty, 1989, 1993; Jelenkovic et al., 1983; Quarta et al., 1991 ; Rose et al., 1993; Svensson and Johansson, 1994) could be due to having insufficient doses of some alleles when ploidy levels are reduced. Among the diploid species in the study there is considerable overlap of LS means groupings (Table 3.2). Since six of the nine diploid species in the study were represented by Mo or fewer accessions, conclusions that can be drawn regarding the evolution of DNA content within diploids is speculative. Within the two species having five accessions, F.nilgerrensis, and F. vesca had ranges of 0.12 and 0.05 pg DNAIgenome, respectively. Presumably, such variation would be synonymous with genetic variation, if indeed the extra genetic material encodes active genes. The 21 diploid and 9 octoploid accessions have ranges of 0.24 and 0.13 pgdnaigenorne, respectively. This fact parallels studies of chloroplast DNA (Harrison et al., 1993, 1997) which showed more variability among diploid species than among octoploid species. The mean DNA content of F.onentalis (4x) and Fmoschata (6x) is closer to that of diploid species than octoploid species. Perhaps at 4x and 6x levels there is less selection pressure for reduced genome site. Fragana rnoschata is alleged to be a polyploid hybrid between F.vesca and F.viridis (Staudt, 1959; Ahmed and Bringhurst, 1991). The data of this experiment supports this theory since the DNA content of F.moschafa is in the range of what could occur if diploid accessions were corn bined. DNA content can Vary among samples of the same accession. While the

105 89 accessions of ni15, nip4 and cl 110 had rather large SE values of 0.11, 0.18, and 0.14, respectively, other accessions had SE values in the range of 0.02 to 0.06 with a mean SE of.05. On average, the mean SE was 7% of the mean for pg DNAlgenorne which gives a reasonable estimate of the precision of the fiow cytometry technique. Put in terms of the 56 chromosomes found in octoploid Fragana species, the SE would account for approximately +/- four chromosomes. This degree of precision is adequate for identihing differences among ploidy levels such as 4x, 6x and lx, but would not be precise enough to identify additions or deletions of a few chromosomes. If greater precision is required additional flow cytometry runs could be performed, but even one sample run provides information on 1000s of cells compared to 10s of cell through root tip squashes. For the purposes of synthetic octoploid identification, the flow cytometry technique outlined offers an efficient way to screen many accessions for ploidy levels. As Fragana chromosomes are extremely small, from 0.9 to 1.7 microns (Yarnell, 1929; lwatsubo and Naruhashi, 1989, 1991 ), the flow cytornetry method would be much easier to use than root tip squashes. CONCLUSION Fragana species Vary in nuclear DNA per genome. Ploidy levels of intenpecific hybrids should be calculated using DNA levels of individual species in their lineage. Also, the flow cytorneter underestimated DNA content by 3% when an increase in ploidy occurred, and such distortion should be accounted for when calculating DNA content of polyploid Fragana. It is not known if this underestimation

106 90 is universal among al1 types of cells and al[ flow cytometers, but a comparison of primary and secondary peaks may be useful in other situations. The octoploid species have a lower DNA contentlgenome than most other Fragana species, which suggests DNA has been lost in their evolution. Diploid species are more variable in their DNA content/ genome than octoploid species but more accessions are needed to study differences among diploids. The above factors of flow cytometry linearity and differences of DNAfgenome in Fragaria species can be used to interpret flow cytornetry results and identify ploidy levels of interspecific hybrids and colchiploids.

107 91 Chapter 4 Hybridization of Fragaria vesca subspecies as female parents w it h FMgerrensis, F.nubicola, Epentaphylla and F. viridis ABSTRACT The potential of using F.vesca as a bridge species to F.nilgenensis, F.nubicola, F.pentaphylla, and F. vindis was investigated using a wide gerrnplasm base of 40 F.vesca accessions. This study was successful in producing many hybrids between F.vesca and other diploid species. Of the species used as pollenizers, F.nubicola, F.pentaphylla, and F. viridis accessions were more successful, averaging 85% to 100% fruit set and 12 to 25 seedslflower. It was most dificult to obtain hybrids with Fragada nilgerrensis, which had only 38% fruit set and 1.2 seedslflower. Differences among pollenizers were minimal when hybrid seeds were germinated in vitro. For different species combinations, 75 to 99% of seeds had embryos and 77% to 85% of these embryos geninated. The lack of significant differences in crossability variables among the four types of F. vesca studied (F. vesca subspecies americana. bracteata, vexa and vesca var. sempemorens) demonstrated that no F.vesca subspecies was superior for interspecific hybridization with any of the other diploid species used in this study. As European and North American F.vesca subspecies are not sufficiently divergent to differ in interspecific hybridization, it seems most likely that F-vesca is a young species rather than an old progenitor species. The high crossability of Evesca indicates that strong potential exists for geneflow between F.vesca and

108 other diploid species. INTRODUCTION The synthetic octoploid (SO) system was developed to incorporate germplasm of 2x, 4x and 6x Fragaria into commercial 8x cultivars of F. x ananassa. This bypasses ploidy level differences and eases introgression of 2x. 4x and 6x species (Evans, 1977). The first step of a synthetic octoploid system is the intercrossing of diploid species to create diploid hybrids (Fig. 1.7). The use of a bridge (or pivotal genome) to facilitate interspecific crosses has been useful in introgressing diploid species into polyploid cereals (Feldman and Sears, 1981; Gustafson and Dera, 1989; Armstrong et al., 1992; Gustafson and Sears, 1993; Lelley et. al, 1995) but no suggestions for bridge species have been mentioned in previous investigations of SO systems with Fragana (Sebastiampillia and Jones, 1976; Evans, 1977). Fragana vesca is a logical choice to use as a bridge species for other diploid Fragada in the diploid step of the SO system because it has been successfully hybridized with several Fragana species including octoploid cultivars (Yarnell, 1931 a; Mandelsdorf and East, 1927; Federova, 1946; Williams, 1959; Staudt, 1959, 1967b, 4989; Evans, 1964, 1974, 1977; Fadeeva, 1966; Evans and Jones, 1967; Senanayake & Bringhunt, 1967; Sebastiampillia & Jones, 1976; Ahrnadi and Bringhurst, 1992; Trajkovski, 1993). Fragah vesca may be progenitor of the genus Fragana (Duchesne in Darrow, 1966; Staudt, 1953; Numberg-Kruger, 1958; Longley in Darrow, 1966) and rnay, therefore, be closely related to other diploid species. This close relationship rnay be exploited to facilitate

109 93 interspecific crosses. The use of F. vesca subspecies in interspecific crosses in this study may also aid in understanding divergence and evolution in Fragaria. Since F. vesca is distributed across five continents (see Figure 1.3), crossability differences among F.vesca subspecies could also be exploited to optimize an SO system. Advantages of using F.vesca as a bridge species The primary importance of using F.vesca as a bridge species is its known ability to hybridize with other species. Table 4.1 provided details on crosses beheen F. vexa as a female parent with three of the pollenizer species also used in this study. No data were available regarding the cross F. vesca x F. pentaphylla. The crossability of /=.vexa with other species can be seen in Table and 1.4. in Chapter 1. In many species, using several accessions increased the probabiiity of success that interspecific hybridization would occur (Burbank, 1921 ; Galletta and Maas, 1990; Stoskopf et al., 1993). However, very few accessions of self- incompatible Fragaria species were available for this experiment. Of al1 diploid species, the most accessions were available for F. vesca (ie. approximately 120). In corn panson, when this study began, only 20 accessions of diploid species (other than F.vesca) were listed at various genebanks and most of these either were unavailable, incorrectly labeled or would not germinate. Fragaria vesca's adaptation to low light levels (Chabot and Chabot,1977) and cool temperatures (Bauer, 1976) may be especially useful in northem areas

110 Table 4.1. Results of previous research using Fragaria vexa as a female in crosses with F. vindis, F. nubicola, and F. nilgenensis. All subspecies rnentioned refer to subspecies of F. vesca Species combinations F. vesca x F. viridis F. vesca x F. nubicola F. vesca x F. nilgenensis 16 plants partially fertile ssp. vesca: 13 pollinations, 6 seed, 4 germinated, al1 dwarfs Reference Yarnell, 1931 a ssp. bracteata: 90% germinated, 19 plants dwarf, al1 died F. vesca from Hawaii: 50 seeds, 18 plants, vigourous full seed set, plants full seed set,plants full seed set, viable hybrids Williams, produced produced only with eastern European 1959 ssp. vesca. Most F. vesca accessions were unsuccessful ssp. vesca: 17 ssp. vesca: 21 ssp. vesca: 18 pollinations, Evans, 1964 pollinations, 90 pollinations, 201 seed, O seed seed, 62 germinated 152 germinated accession #1: 25 seedlberry, 383 seed, 231 germinated accession #2: 7 seed/berry, 337 seed, 200 germinated Only stunted hybrids Fadeeva, obtained 1966 viable seed viable seed produced viable seed produced Evans and produced Jones, 1967

111 95 and could enhance late season growth (Bauer, 1976), fall Rower bud initiation, and yield potential if these adaptations were successfully transferred into cultivars. Its worldwide distribution shows broad adaptation to diverse conditions (Darrow, 1966; Fig. 4.1). Fragaria vesca also has a very desirable aroma that could be used to improve fruit quality of cultivars (Darrow, 1949, 1966; Jones, 1966; Ahmadi and Bringhurst, 1992; Trajkovski, 1993). Advantages of using F.vesca as a female parent Fragaria vesca is more successfully used as a female parent in crosses with other diploid species. When F.vesca is used as a male in crosses with self- incompatible species, such crosses often produce no seed (Evans and Jones, 1967). However, when the self-compatible F. vesca is used as a female in crosses with self-incompatible diploid species (ie. F. vindis, F. nubicola and F. penfaphylia) (Evans, 1964; Staudt, 1989), incompatibility mechanisms are bypassed (Lewis and Crowe, 1958; Evans and Jones, 1967). Cornpared to other Fragaria species, F. vesca has more variable cytoplasm (Harrison et al., 1993; 1997) which could be important to the success of interspecific hybridization (Stoskopf et al., 1993). This variability can be exploited if F.vesca is the female parent. Also, F. vesca cytoplasm and gemmes are more amenable than other Fragafia species to chromosome doubling by colchicine (Fadeeva, 1966; Sebastiampillia and Jones, 1976). Thus, F. vesca could be of special benefît to the Synthetic Octoploid system which requires two stages of chromosome doubling. Accessions of Evesca ssp. vesca var.

112 96 Figure 4.1. Distribution of Fragana vesca subspecies (Evans, 1964; Reed. 1966; Staudt, 1 989). ssp. vexa var. ssp.

113 97 sempeflorens, when used as a female in crosses, have recessive everblooming and non-runnering traits (Ahmadi et al., 1990) that can be used as genetic markers to facilitate sorting of interspecific hybrids from selfed progeny. This everbearing trait also simplifies coordinating the timing of pollination with other species and could be used to breed multiple generations per year provided selfing or backcrossing is used to restore the homozygous condition to the recessive alleles that govern the everblooming trait. Evidence that F.vesca subspecies have diverged If F. vesca is the progenitor species to other diploids (Duchesne in Darrow, 1966; Staudt, 1953; Nurnberg-Kruger; 1958; Longley in Darrow, 1966), then F. vesca subspecies may differ in relatedness and hence crossability with other diploid species. With a range of distribution over five continents (see Fig 4.1). F. vesca subspecies could have become geographically isolated and diverged sufficiently to differ in interspecific crossability. Although it has been suggested that F.vesca is not highly variable genetically (Reed, 1966), there have been several studies that have highlighted morphological differences of plants, fruit, leaves and flowers of F. vesca subspecies (Williams, 1959; Staudt, 1989; Ahmadi et al., 1990; Harrison et al., 1993, 1997). Fragana vexa subspecies were also found to differ in chloroplast DNA RFLPs (Harrison et al., 1997), and the number of genes goveming photosensitivity (Ahmadi and Bringhurst, 1990).

114 98 Goals of this study In this study it is hypothesized that F. vesca subspecies have diverged sufficiently to cause differences in interspecific crossability. In this experiment, divergence among F. vexa su bspecies amencana, bracteata, vesca and ssp. vesca var. sempeflorens is examined. The latter group, var. sempeflorens, is the cultivated form of F.vesca, which rnay have diverged through artificial selection. The objective of this study is to detenine which of these four Evesca types is best suited for interspecific hybridization with the six accessions of other diploid species. This information could be important for developing a crossing strategy for early stages of the SO system and may provide insight into the evolution of Fragana species. MATERIAL AND METHODS General material and methods including induction of flowering, pollination methods, seed storage and germination are presented in Chapter 2. Additional details for this experiment are covered in the remainder of this section. A total of 40 accessions of Evesca (Table 4.2) were hybridized with seven pollenizers (Table 4.3). Crosses were with Evesca as a female pollinated with either F.nilgenensis, Enubicola, F-pentaphylla, F. vindis, or F. vesca. Fragana vesca subspecies amencana, bracteata and vesca were represented by 12, 10 and 8 accessions, respectively. To maximize time available to facilitate crossing, one accession of each of these three subspecies was placed in the greenhouse in early

115 Table 4.2. Fragada vesca accessions used as female parents in crosses with F. nilgen-ensis, F. nubicola, F. pentaphylla, F. viridis and F. vesca. Accessions are grouped by subspecies. Page 1 of 2. Guelph id Identifier used by source SourceL F. vesca ssp. vesca var. sempemorens (dayneutral cultivars) vcvo2 PI , Alpine NCGR vcvo4 PI , Yellow Wonder NCGR vcv06 CFRA , Golden Alpine NCGR vcv07 Pineapple Crush Park Seed vcvl O Alexand ria Thompson & Morgan vcvl 1 Ruegen Richters vcvl3 Mignonette Park Seed vcvl4 Golden Sprite Ornamental Edibles vcv20 Alpine Strawberry White Flower Fams vcv22 FRA 2193 IPK Genbank F. vesca spp. vesca (wild type) vw 07 Pl NCGR vw 09 Pl NCGR vwl O Pi NCGR vwl 1 PI NCGR vwl2 PI NCGR vwl3 PI NCGR vwl4 PI NCGR wu16 PI NCGR vw22 FRA IPK Genbank vw38 PI NCGR F. vesca spp. bracteata (wim type) vw 02 PI NCGR vw 06 PI NCGR vw23 PI NCGR vw24 PI NCGR vw25 PI NCGR vw26 PI NCGR vw57 PI 148 C.C.G. wu58 PI 418 C.C.G. 2: full narnes and addresses for germplasm sources can be found in Appendix 1.

116 Table 4.2. continued. Guelph id Identifier used bv source Source F. vesca ssp. americana (wild type) vw17 Pl NCGR vw29 PI NCGR vw30 PI NCGR vw32 PI NCGR vw33 Grand River, Cambridge, ON Spencer Murch vw34 Cambridge, ON collected by author vw35 Pl NCGR vw36 Pi NCGR vw47 PI NCGR vw48 PI NCGR vw49 Srnithfield, ON collected by author vw53 Newcastle, ON collected by author

117 Table 4.3. Fragaria accessions used as male parents in crosses with Evesca. Accessions are grouped by species. Guelph id Identifier used by source SourceL F. nilgerrensis niil CFRA NCGR nubl nub2 penl F.nubicola Pi PI F. pen tap h ylla PI NCGR NCGR RBG Edinburgh NCGR NCGR F. vesca (control) vw69 Fra IPK Genbank Z: full names and addresses for gemplasm sources can be found in Appendix 1.

118 102 January and another in early February. Three plants each of 10 cultivars of F-vesca ssp. vesca var. sempeflorens were also included. Since sempeflorens accessions bloom continuously, they were brought out of cold storage only in January. The recessive everbearing and non-runnering traits of sempeflorens cultivars were also useful in this study because they provided easily identifiable morphological markers to identify true hybrids from selfs. Accessions used as males were removed from cold storage in early Decernber and placed in the greenhouse. Initially, accessions of F. nubicoia did not bloom in the greenhouse although they received the same growing conditions as other plants in the study. However, they were induced to bloom after being subjected to a shorter day length and cooler temperatures in a growth chamber (see Chapter 2). Experimental design and data analysis The female parents were grouped into four F. vesca su bspecies: amencana, bracteata, vesca, or vexa var. sempeflorens. There were six male parents for interspecific crosses: nubl, nub2, penl, virl, vie, and Nil1 and one F.vesca accession used as a pollen source for control crosses. Ten F.vesca var. sempeflorens cultivars were assigned to each of the three blocks but accessions of the three wild subspecies (ie amerkana, bracteata, vesca) were randomly distnbuted among the blocks. Control crosses with F.vesca pollen were done to ver@ that emasculation

119 1 O3 and pollen storage techniques were correct and for comparison with interspecific crosses. VW69 was used for control crosses because it was native to Asia (unlike other F.vesca accessions in the study) and produced abundant pollen. Data from test crosses were discussed in results but were not included in the data set of inters pecific crosses. Pollinations for interspecific crosses were done to maximize the number of parental combinations. Pollination was randomized each day according to available pollen but attempts were made to equalize the various combinations of species. Details of pollination methods are in Chapter 2. Fruit set (%), seeds/flower, seeds with ernbryos (%) and germination (%) were measured. Where the same combination of parents occurred in the same block, data were pooled for that combination before statistical analysis. Data were analyzed with the GLM program of PC SAS (Sas system for Windows V.6.12, Cary, N.C.). Mean separation was accomplished using the LS means function of PC SAS. Analysis of fruit set and seedslflower was based on each combination of parents crossed, while seeds with embryos (%) and germination (%) was based on a subset of those cornbinations that received in vitro germination. RESULTS Flowering time influenced the number of potential combinations of parents. There were 191 different cornbinations of parents achieved by 360 interspecific pollinations. The males, nill, nubl, nub2, penl, virl and vir2 were used as pollenizers in 37, 38, 28, 27, 35 and 36 hybrid combinations, respectively. Pen1

120 1 O4 produced trusses with only one to three flowers and was used in fewer combinations. Virl and vir2 were the first pollenizer species to bloom and were involved in more crosses. While nill bloomed about three weeks after virl and vir2, its total blooming period was about two weeks longer than that of any of the other pollenizers. It was noted that ssp. bracteata plants had fewer fiowers per plant than other F. vesca subspecies in the study so low nurnbers of crosses occurred with this group. Fruit set (Oh) and seedslflower Pollenizer accessions were a significant source of variation for fruit set and seedslflower but Evesca subspecies and the interaction sources of variation were not significant (see ANOVAs in Appendix 2). The F. nilgerrensis accession, nill, had the lowest flower set, 38%, while cther pollenizers averaged between 76% and 100% (Table 4.4). Crosses having the nill accession as a pollenizer also had the lowest seedslfiower with an average of 1.2. Pollenizers nub2, penl virl and vir produced similar seedslfiower averaging from 11.6 to 16.9 while nub 1 was significantly higher with 24.7 seedslflower (Table. 4.5). The two North American subspecies, amencana and bracteata were similar with 69 and 65% fruit set, respectively. However, these averages were not statistically different from the European su bspecies, vesca (wild type) and vesca var. semperflorens, which had 85 and 83% fruit set, respectively. Means of seedsmower for Evesca subspecies varied by less than two seeds, having a range

121 Table 4.4. Fruit set for crosses among Fragaria vesca subspecies when pollinated with F. nilgenensis (1 accession), F.nubicola (2). F.pentaphylla (1) and Eviridis (2) and F.vesca (1). lnterspecific crosses were analyzed separately from the control crosses and means separation was done with LS means at p = F. vesca ssp. Species used as male parents (accessions) Control potlenizer & F. nilgenensis F. nubicola F. ~ 8nta~h~//a F.. - viridis Mean F. vesca accessions inii ) (nubi) (nub2) (~enl) (virl) (vir2) (Mg) used as fernales N Fruit N Fruit N Fruit N Fruit N Fruit N Fruit N Fruit N Fruit set (74) set {%) set (9%) set (Yo) set ( Oh) set ("/O) set ( Oh) set (%) ssp. americana VW17 2 O 1 IO0 1 O O0 VW30 2 O O O VW33 1 O 1 1 O O0 VW O VW35 3 O O 1 O O0 VW36 3 O O O0 w VW48 1 O 1 IO IO O0 VW49 3 O O0 VW53 4 O Mean 26 20ns 8 1OOns s 7 loons 31 70ns s 96 69ns 17 1 OOns ssp. vesca VW O0

122 Table 4.4. Continued. F. vesca ssp. Species used as male parents (accessions) Control ~ollenizer & F. nilgenensis F. nubicola F.pentaph ylla F. vjridjs Mean F. vesca -. - accessions (ni11 ) (nubl) (nu b2) (pen 1 (virl) (vi r2) (M9) used as females N Fruit N Fruit N Fruit N Fruit N Fruit N Fruit N Fruit N Fruit set (%) set (%) set (%) set [%) set (%) set (%) set (%) set (%) ssp. vesca var. sempeflorens VCV IO vcv04 VCVO6 vcv07 VCVI 0 VCVI 1 VCVI 3 vcv14 VCV20 ssp. bracteata VW02 1 O O 1 IO O Overall OOab 40 76bc 33 93abc 75 88abc 68 9labc

123 Table 4.5. Seeds/flower for crosses among Fragaria vesca su bspecies w hen pollinated with F, nilgenensis (1 accession), F. nubicola (2), F.pentaphyl/a (1) and F. vjridis (2) and Evesca (1). lnterspecifc crosses were analyzed separately from the control crosses and means separation was done with LS means at p = F. vesca ssp. Species used as male parents (accessions) Control pollenizer F. nilgerrensis F. nubicola F. pentaphylla F. viridis Mean F. vesca accessions (ni11 ) (nubl) (nub2) (~en1) (virl) (vi R) (vw69) used as fernales N Seeds N Seedsl N Seedsl N Seedsl N Seedsl N Seedsl N Seedsf N Seedsl / flower flower flower flower flower flower flower flower ssp, americana WVI 7 2 O O VW30 2 O O O VW33 1 O O VW O WU VW VW VW O. O VW VW Mean ns ns ns ns ns ~s ns ns ssp. vesca VW O VW VWIO O VWI O VW O W W VW VW VW Mean ns ns ns ns ns ns ns ns

124 99 0 =? a &O0 CVCV - CC) Y- Nrr N a e i P N N

125 1 O9 of 13.2 to 14.7 seedsmower (Table 4.4). Although fruit set for control crosses was 9% higher than the overall mean for interspecific crosses, four of the six pollenizers had between 1 and 13% hig her fruit set than controls. Control crosses averaged twice as many seeds/flower than interspecific crosses and al1 interspecific pollenizers had lower numbers for this variable. Seeds with embryos and embryo germination InsufFicient numbers of hybrid seeds were produced with Evesca ssp. bracteata fernales to allow meaningful ANOVA analysis of differences among F.vesca subspecies for the proportion of seeds with embryos and percent embryo germination. Therefore, only the effect of pollenizer accessions was incorporated into ANOVAs of these two variables (Appendix 2). Pollenizer accessions were a significant source of variation for percent of seeds with embryos but not for germination (Table 4.6, ANOVAs in Appendix 2). For percent of seeds with ernbryos, ni11 had the lowest numbers but this was not significantly different from vir2. Other pollenizers were at least 13% higher then nill or vir2 for percent of seeds with embryos (Table 4.6). Percent embryo germination was similar for al1 pollenizen, with only a 12% difference between the lowest (vir2) and highest (virl) pollenizer. Differences in intraspecific crosses Percent flower set did not differ significantly among F. vesca subspecies

126 Table 4.6. Seeds with ernbryos (%) and ernbryo genination (%) for interspecific crosses arnong Fragaria vesca accessions as females and six di ploid accessions as males. Pollinizer Accession Number of Total no. of Seeds In vitro em bryo species female seeds with embryos germination accessions germinated (%) (%) in vitro F. nilgerrensis ni b" 85ns F. nubicola nubl a 81 ns nub a Wns F. pentaphylla penl a 81 ns F. vindis virl a 89ns vi R ab 77ns Means separation with LS means at p = y Sums. Means Y 1868 Y 90 ' 82 '

127 111 when crossed as fernales to vw69 (control). F.vesca ssp. amencana had the highest values, which averaged 43% higher than F. vesca ssp. bracteata (Table 4.4). Fragana vesca subspecies d id not differ significantly for seedslflower with averages within a 27 seed range and the lowest and highest values occurring for F. vesca ssp. bracteata and F. vesca ssp. vesca var. sempefiorens, respectively. Surprisingly, some accessions of F. vesca ( ~02, vw24, vw25 and vw38) did not set fruit when pollinated with vw69. However, each of these four accessions did produce seed in interspecific crosses. Vw26 was the only F.vesca accession that produced seeds with the control pollenizer but was not successful when crossed with other species (Table 4.5). However, it was involved in only five interspecific crosses. Differences among Evesca accessions Althoug h variability existed within F. vesca subspecies, this study was not designed to allow for statistical analysis of crossability of individual F.vesca clones. Further pollinations would be required to insure replication for each F.vesca clone x pollenizer combination. Superior combinations of F.vesca accessions and pollenizen for interspecific hybridization can be found in Tables 4.4 and 4.5. Such data could be useful for future investigation to produce interspecific hybrids. For example, superior accessions for hybridizing with nill were vcv07, vw 23 and vw47 (Table 4.4). Some accessions of F-vesca had superior crossability with most of the pollenizers in this study. Of the 15 Evesca accessions pollinated with al1 six

128 -Il2 interspecific pollenizers, vw32, vw07, vcv10, vcvl3 and vw09 were very successful ranging between 15 and 22 seeds/flower. Other notable accessions were vw53, w49, vw14, and vw25, although these were not crossed with al1 pollenizers. The superior genotypes mentioned in this paragraph are fairly evenly distributed amongst the F.vesca groups in this study. Each group had three superior genotypes except for F. vesca ssp. vesca var. sempeflorens which had only two. DISCUSSION This study used 39 accessions of F.vesca in interspecific hybridizations, a number fa greater than those used in other reported studies, and it is likely that this study provides a better representation of the diversity within F.vesca compared to those earlier studies. Despite isolation in widely separated areas of North America and Europe as well as the isolation that occurred when Evesca varsempeflorens was domesticated, F. vesca subspecies did not differ sig nificantly in interspecific crossability in this study. although considerable variation was present within subspecies. The interactions among F. vesca subspecies and pollenizers were not statistically different for any of the variables investigated. The hypothesis that F-vesca subspecies differ in interspecific crossability (Williams, 1959; Fadeeva, 1966) was not supported by the results of this study. Since a few superior genotypes were found within each of the F-vesca subspecies. it is quite possible that smaller collections of F.vesca subspecies in previous studies might be skewed resulting in some su bspecies being found su perior. For example, observations that

129 113 indicated cultivars of F.vesca var. sempemorens were superior over wild types for interspecific crossability was based on only two accessions of F.vesca var. sempeflorens and an undisclosed number of wild accessions of F.vesca (Fadeeva. 1966). The ability of F. vesca to hybridize with four other diploid species validates the hypothesis that F.vesca could be very useful as a bridge species for the first steps of the synthetic octoploid system. Fragana breeders wishing to use F. vesca as a bridge species could use any subspecies of Evesca with a high likelihood of success. Since most F. vesca accessions hybridized with most pollenizers. with the exception of F.niigenensis, it may be advantageous to focus on F. vesca accessions with superior horticultural qualities when introgressing other species. Qualities such as larger fruit size or disease resistance may make an accession desirable to use, even if only a few seeds were produced in interspecific hybridization. Fadeeva (1966) observed a correlation between fruit size of F.vesca lines and fruit size of intenpecific hybrids. king a locally adapted subspecies would also be a recommended strategy. The fiexibility to choose among many genotypes of F. vesca would be especially useful if prebreeding were to occur at the diploid level as part of the synthetic octoploid system. The data from this experiment can neither support nor refute the theory (Duchesne in Darrow, 1966; Staudt, 1953; Numberg-Kruger, 1958; Longley in Darrow, 1966) that F. vesca is the progenitor species of the genus Fragana. Trying to elucidate evolution in Fragana is very problematic since many species in the genus are closely related (Jones, 1966). The lack of significant differences among

130 Il4 F.vesca subspecies suggests F.vesca is a younger species rather than an ancestral species undergoing differentiation into other species. However, F.vesca successfully hybridized with four other diploid species in this study, which supports the progenitor theory. Until more is known about the crossability among diploid species, determining Evesca's role in the evolution of its genus is dificult. It is not known if F.vesca is the most easily intercrossed of the diploid species. Even if F.vesca is not the progenitor of diploid species, it could be an ancestor of some species. The ability of F.vesca to hybridize with other diploid species in this study tends to support the theory (Darrow, ; Federova. 1946; Ellis, 1958; Staudt, 1959, 1984; Bringhurst and Senanayake, 1967; Bringhurst and Gill, 1970; Bringhurst, 1990; Ahmadi and Bringhurst. 1991) that Evesca could have been involved in the formation of polyploid species. A theory that could explain the similarities among F.vesca subspecies is that Evesca's adaptation to low light levels, which enables it to survive in forest ecosystems (Chabot and Chabot, 1977), is responsible for its wide distribution (Reed, 1966; Staudt, 1989), despite being a young species. As F.vesca is widespread, particularly throughout the northern continents (Reed, 1966; Staudt, 1989), it could have rnigrated during the Miocene ( million years ago) with numerous forest species across intercontinental land bridges connecting Asia, North America and Europe (Stebbins, 1950; Hara, 1956, 1972; Constance, 1972; Li, 1972; Wood, 1972; Lowery and Jones, 1984). This would explain the morphological similarities between F.vesca ssp. vesca of Europe and Evesca ssp. arnericana of eastern North America noted by Fletcher (1917) and Staudt (1989).

131 115 It would also explain why North American and European subspecies of F. vesca did not differ in interspecific crossability in this study. Fragana species that were not adapted to northern latitudes would not have been able to migrate across the northern land bridges. Thus. the presence of F. vexa in North America may be due to its specialized adaptations to forest ecosystems rather than the species being ancient. This study shows that F. pentaphylla. F.nubicola. and F. vidis could be closely related since rnost accessions formed interspecific hybrids with F-vesca. However, intercrossing of F. pentaphylla, F.nubicola, and F. vindis would provide more information regarding the relatedness of these three species. Morphology (Reed, 1966; Staudt. 1989). and chloroplast DNA analysis (Harrison et al., 1993, 1996) also suggest F. vesca, F. nubicola and F. vindis are closely related. Fragana vexa x F. viridis and F. vexa x F. nubicola crosses prod uced similar results to those of other researchers (Table 4.1 ) but F vesca x F.pentaphylla crosses have not been previously reported. It was surprising that F. pentaphylla had similar crossing results to that of F. vindis and F. nubicola as this species has unique leaf (Fig 1.l) and fruit (Fig 1.2) morphology suggesting at least some divergence. The results obtained in this study concur with that other research (Williams, 1959; Evans, 1964), that Evesca x F. nilgemnsis hybrids are difficult to obtain. An explanation for these findings is that F. nilgerrensis is divergent from other diploid species (Harrison et al., 1993,1997). Since F. vesca accessions produced few seeds when pollinated with F.nilgemnsis, averaging only 1.2 seedslflower, it may

132 Il6 be of benefit to use the more successful F.vesca accessions in Table 4.5 to cross with other F.nilgerrensis accessions. Earlier studies identified only three F.vesca accessions with potential to produce healthy hybrids with F.nilgenensis (Yarnell, 1931 a; Williams, 1959; Evans, 1964) but in this study, 19 accessions were found that produced hybrid seeds and three superior accessions that produced greater than five seeds per flower. While stunting (Yarnell, 1931 a; Fadeeva, 1966) and lethal seedlings (Williams, 1959) have been observed in F. vexa x F. nilgenensis hybrids, the lack of these phenornena in this study indicates variability in the species of F.nilgenensis and perhaps that nill is superior for interçpecific hybridization with F.vesca. While it has been suggested that F.nilgerrensis cannot be introgressed into the genepool of cultivated strawberries (Hancock et al., 1 W6), the current study shows encouraging progress toward this goal. A few other differences existed among this study and earlier studies. While fruit set was not reported in previous studies, seedstflower and germination rates were reported. The seedslflower produced in this study were higher than the average of other researchers (Table 4.1). Averages for F. vesca x F. vindis, F. vesca x F.nubicola, and F-vesca x F-nilgerrensis were 2, 7, and 1 seedsmower higher in this study, respectively. The results of Evesca x F.niigenensis crosses were particularly noteworthy in this study since very few successful crosses between these two species had been reported previously. If percent seeds with embryos is rnultiplied with percent embryo germination (Table 4.6), the resulting numbers can be compared to germination rates in Table 4.1. For the above three combinations, this study had germination rates 13, 5, and 0% higher, respectively, than earlier

133 117 studies. While the use of in vitro germination did not dramatically increase germination rates compared to traditional rnethods, it provided additional information regarding the germination of interspecific hybrids. Sig nificant differences among the various combinations of species (Table 4.6) were due to ernbryo abortion during seed formation (ie. seeds with embryos) and not germination rates of fully formed embryos (ie. embryo germination). CONCLUSION This study was successful in producing many hybrids between Evesca and other d iploid species, including the difficult to obtain F. vesca x F. nilgerensis hybrids. The lack of significant differences in crossability variables arnong the four types of F. vesca demonstrated that no F.vesca subspecies was superior for interspecific hybridization with any of the pollenizers of this study. As European and North American Evesca subspecies are not divergent enough to differ in interspecific hybridization, it seems more likely that F.vesca is a young species rather than a progenitor species. An explanation for the global distribution of F.vesca is that specialized adaptations have allowed F.vesca to spread in forest ecosysterns. This theory is counter to previous theories of F. vesca evolution (Duchesne in Darrow, 1966; Staudt. 1953; Nurnberg-Kruger, 1958; Longley in Darrow, 1966). Using a wide gemplasrn base of F.vesca was significant for the success of this investigation corn pared to previous research. Crossability of F. vesca indicates that strong potential exists for geneflow between F.vesca and other diploid species.

134 118 Although the number of accessions of other diploid Fragana species was limited, Evesca hybridized with the four other diploid species used in this study, indicating good potential as a bridge species. Limitations of gerrnplasm make it impossible to conclude that F.vesca is the best bridge species available, but certainly F. vexa has potential value in initial steps of creating of synthetic octoploids.

135 CHAPTER 5 Interspecific crossability of nine diploid Fragaria species ABSTRACT Accessions of the diploid species F. daltoniana, F.iinumae, F.nilgerrensis, F. vesca, F. gracilis, F. nipponica, F. nubicola, F.pentaphylla and F. viridis were crossed in 36 interspecific cornbinations. Of the 20 combinations of species that resulted in viable F, plants, 13 bloomed and produced seeds and seven did not bloom. Seven combinations of species failed to produce seeds and two additional combinations produced seeds that would not geminate. With the exception of F.iinumae, al1 species in the study produced fertile F, hybrids with at least one other species. This demonstrated the significant potential for geneflow at the diploid level. Fragana pentaphylla was the most successful species for interspecific hybridization followed by F. vesca. Fragana pentaphylla formed hybrids with seven species in the study, and five of the seven hybrid types were fertile. Fragada vesca foned fertile hybrids with five other species, but fomed non-viable hybrids with three species. The superiority of Epentaphylla and F.vesca for interspecific hybridization indicates they have the greatest potential as bridge species for diploid Fragaia. Hybrids were readily obtained when intercrossing self-incompatible species fie. F.gracilis, F.nipponica, Fmbicola, Fpentaphylla amd F. vindis), but were vety difficult to obtain when intercrossing self-compatible species fie. Edaltoniana, Ç.iinumae, F. nilgerrensis, F. vesca). In contrast to previous research which showed

136 120 F. vesca was the most likely progenitor of the genus Fragaria, this study indicates F. pentaphyiia rnay be the progenitor. It is also likely that self-incompatible Fragana species are more closely related and ancestral to self-compatible species. lntroductlon Previous research on interspecific hybridization with Fragatia diploids has been based on a limited number of accessions and only 28 of the approximately 90 possible combinations of diploid species have been reported (Table 5.1). For example, no reports have been made regarding interspecific crossability of F.pentaphylia, Ç.gracilis, or F.manchunka while interspecific hybridization involving F.iinumae (Sebastiampillia and Jones, 1976), Enipponica (Williams, 1959) and Edaltoniana (Evans, 1974) has received only limited attention. In contrast, the diploid interspecific crossability of F.nilgerrensis, F.nubicola, F. vesca, and F. vitidis has been investigated by many researchers in several reports (Table 5.1 ). An early investigation (Chapter 4) focussed on an introgression strategy to broaden the germplasm base of the cultivated strawberry that used Evesca as a bridge species with F.niigenensis, F.nubicola, F.pentaphylla and F.vindis. That study cast some doubt on a theory of Fragana evolution (Duchesne in Darrow, 1966; Staudt, 1953; Nurnberg-Kruger, 1958; Longley in Darrow, 1966) that suggested F.vesca was the progenitor species of the genus. Following that study. Evesca was crossed with F.iinumae and Egracilis but the hybrid seedlings failed to develop beyond the cotyledon stage despite the use of 15 Evesca accessions and in vitro (Miller et al., 1992) germination. Since F. vesca may not be sufficient as

137 Table 5.1. Nurnber of reports for interspecific hybridization of diploid Fragana species, listed by crossability categories. Based on Yarnell, 4931 a; Mandelsdorf and East, 1927; Federova, 1946; Williams, 1959; Staudt, 1959, 1967, 1989; Evans, 1964,1974,1977; Fadeeva, 1966; Evans and Jones, 1967; Bumagina and Mesaros, 1969; Sebastiampillia and Jones, 1976; and Trajkovski Self-compatible males Self-incompatible males F. daltoniana F. iinumae F. nilgerrensis F. vesca F. nipponica F. nubicola F. viriciis Fernales ~ornpatibility~~'~~~~~ ABCDEF ABCDEF ABCDE F ABCDEF ABCDEF ABCDE F F. daltoniana SC F. nilgerrensis SC F. vesca SC F. nipponica SI F. nubicola SI F. viridis SI =:SC = self compatible SI = Self incompatible Crossability categories: A = no set 6 = seed, no germination C = death after germination D = sterile plants E = plants produced, fertility not reported F = fertile plants

138 122 a bridge for al1 Fragana diploids, this study was initiated to investigate other diploid species in Fragana that may have potential as a bridge species. Although investigation continued using colchi-tetraploid F. vesca x diploid species hybrids as parents in tetraploid crosses (Chapter 7), this study was undertaken to determine which species combinations may be most useful for breeding at the diploid level and to gain an insight into the evolution of Fragana species. The synthetic octoploid system originally proposed by Evans (1977) required crossing amongst diploid species as the first step. Identification of bridge species would greatly facilitate hybridization and broaden the range of germplasm advanced to the octoploid level. This study examines the crossability of nine diploid Fragana species including multiple accessions of some species. Thus, this study includes more species and a broader genetic base than previous studies. it was hypothesised that if another species was similar to the progenitor species, it would cross more readily to other species than F.vesca. Such a species could also have potential as a genetic bridge ai the diploid level. MATERIALS AND METHODS Details regarding handling of plants for crosses, pollination, in vitro germination, and growing methods are provided in Chapter 2. Over a period of three years, Fragana vesca, F. viridis, F.nubicola, F. nipponica, F. nilgerrensis, and F. penfaphylla were crossed in the greenhouse and field in most pair-wise combinations. Fragaria dalloniana, Egracilis and F.iinumae produced only a few flowers each and were included in only a partial set of crosses. Many species used

139 --- 1 L 3 in this study were represented by only a few accessions (Table 5.2). Fragaria vesca was represented by the greatest number of accessions. Forty F.vesca accessions were crossed to F. viridis, Fmbicola, Epentaphylla and F.ni/gerrensis, in a related study (chapter 4) but other species were crossed to at least 10 F. vesca accessions. Du ring bloom, species were intercrossed at random using fres h pollen but some pollen was stored (MacFarlane Smith et al., 1989) for use with species not in bloom. Self-compatible species were not used to pollinate self-incompatible species. As the compatibility status of F.pentaphylia and F.gracilis was unknown, these species were allowed to self pollinate under greenhouse conditions. Crossing strategy was infuenced by time of bloorn and availability of stored pollen. Not all species survived in the field each year, thus limiting field crosses. Preference was given to combinations of species that had not been hybridized or had been unsuccessful in previous studies. Most seeds were germinated with the in vitm rnethod (Miller et al., 1992) while some crosses with good seed set were germinated in the greenhouse. Progenies were compared to both parents to confirm that the hybridization was successful. Hybrid progenies were planted at Cam bridge Research Station, Ontario where female fertility was judged based on fruit and seed set under open-pollinated conditions. The results of interspecific crosses were categorised into six groups: A = no seeds from cross, B = seeds produced but no germination, C = seedlings died after germination, D = F, plants bioomed but had no seeds, E = F, plants have not bloomed, F = F, plants bloomed and produced seeds. Group E included those plants that have not bloomed or did not survive field conditions. As many of the

140 124 Table 5.2. Diploid Fragaria germplasm used in interspecific crosses. Species Guelph # Identifier used by Sourcez source F. daltoniana da RBG Edinburgh NCGR gra2 CFRA NCGR F. iinumae F. nilge rensis ii3 CFRAI NCGR ni1 1 CFRA II NCGR ni12 Mt. Omei, frch059 Forest Farm nip4 CFRA NCGR nubl PI NCGR nub2 PI NCGR F. pentaphylla penl RBG Edinburgh pen RBG Edinburgh virl PI NCGR vir2 PI NCGR vir3 Oberau 2 IPK Genebank vir RBG Edinburgh F. vesca 40 see Table 4.2 accessions Cha~ter 4 ' : See Appendix 1 for full names and addresses of germplasm sources

141 125 species used in this study have not bloomed consistently at this latitude, further observation may be required to judge fertility of hybrids which have not yet bloomed (ie. Group E). Each species combination was assigned to the group that showed the greatest potential for geneflow displayed by its hybrid progeny. For example, if some crosses between two species produced no seeds but other crosses yielded fertile F, progeny, that combination of species was categorized as 'FI = F, plants bloomed and produced seeds. Also, general observations were made concerning general health and vigour of F, hybrid progeny grown in the field. Some species combinations were judged only by the qualitative index mentioned above. Data were recorded for number of seedslflower, % seeds with embryos and % embryo germination for species combinations treated with the in vitro method of germination. Data were analyzed with the GLM program of PC SAS (SAS systern for Windows V.6.12, Cary, N.C.). Mean separation was accomplished using the LS rneans function of PC SAS. RESULTS The crossability categories of 36 combinations of species are presented in Table 5.3. F, plants were produced from 20 combinations of species, of which 13 bloorned and produced seeds (Group F) and seven did not bloom (Group E). The other 16 combinations of species failed to result in viable progeny (Groups A. B, or C). No combinations were categorized Group D (i.e., blooming but producing no seed). Wlh the exception of F.iinumae, al1 species in the study produced fertile F,

142 126 Table 5.3. Crossability categories for interspecific crosses among several diploid Fragada species. Self-incompatible (SI) females were not poliinated with selfcompatible (SC) males as this type of cross is unlikely to produce seeds. Females Compat- SC species as males i bility F. dalloniana F. iinumae F. n~lqenensis F. vesca F. daitoniana SC E A F. iinumae SC A F.nilgenensis SC A~ A F. daltoniana SC F. iinumae SC F.nilgemnsis SC SI species as males F.qraciiIs F.nipponIca F.nubicoia F.~eniaphylla Evindis F. vesca SC C C F F F F. gracilis SI E F F. nipponica S 1 C B F. nubicola SI A C E F F. pentaphyila SI F F F F F. viridis S 1 E E F F Crossability categories A = no seeds from cross B = seeds produced but no germination C = seedlings died after germination O = F, bloorned but had no seeds under open-pollinated conditions E = F, plants have not bloomed F = F, bloomed and produced seeds under open-pollinated conditions

143 127 hybrids with at least one other species. Fragaria pentaphylla, and F-vexa formed fertile F, (Group F) hybrids with five out of seven and five out of eight species, respectively. However, Fragaria pentaphylla was more successful than F. vesca for interspecific hybridization if other groups of F, hybrids were considered. Two additional species, F. daltoniana and F. nilgerrensis, hybridized with F. pentaphylla forming viable piants which did not bloom (Group E). When F. vesca was pollinated with F.iinumae, F.nipponica, orf.gracilis, it resulted in unviable seedlings (Group C). When Evesca was used as a male parent with the other three self-incompatible species no seed set occurred. Unlike F.pentaphy//a, F.vesca was not a parent in any Group E hybrids. Fragaria viridis was almost as successful as F.vesca for producing interspecific hybrids. It produced Group F hybrids with four other species and Group E hybrids with one other species. The next best results were obtained with F. nubicola, F.gracilis, F.nipponica and F. daltoniana which had Group F hybrids with three, two, one and one species. respectively. Besides being the most successful at interspecific hybridization, Epentaphylla, F. vesca, F. vindis and Enubicola, when intercrossed, resulted in fertile progeny. Other species included in this study were only successful in hybridizing with two or fewer species. Detailed information on those interspecific crosses subjected to in vitro germination are shown in Table 5.4. Of the three variables studied (Le. seeds~ower, % seeds with embryos, % embryos developing into plants) seedsmower was most variable, having a standard error approximately the same magnitude as the mean. SeedsMower ranged from 3 to 96 for the various species cornbinations, most of which had greater than 50% of seeds with embryos.

144

145 129 Notable exceptions were crosses involving Edaltoniana as a female and the combination F-nubicola x F.nipponica, which had less than 23% seeds with embryos. In 89% of the species combinations that produced seeds, hybrid plants were obtained through in vitro germination. Hybrid seedlings failed to develop from two crosses: F.nilgenensis x F. vindis and F. vesca x Enipponica. It was noted that vegetative growth of most interspecific hybrids was very vigourous. Some of the crosses involving F.nubicola and F.daltoniana, the smallest and least vigourous species in this study, were particularly vigourous. When these two species were crossed with other species, hybrid progeny had larger leaves and more vigourous plants than the rnost vigourous parent (Fig. 5.1). In the field, F, hybrids were usually vigourous and healthy (Fig. 5.2). The exceptions were hybrids of F. daltoniana x F. vindis, F. nilgerrensis x F-nubicola and F-nubicola x Enipponica, which were stunted and only lived for several months before dying. most hybrids typically varied within the range shown in Figure 5.3. Fruit set of Notable exceptions were hybrids resulting from the cross F.vesca x F.daltoniana, which produced only one to five seeds on a small percentage of flowers. It is too early to categorize non-blooming hybrids as sterile since most have had only one year's growth in the field. Fragada iinumae, Fdaltoniana, Egracilis, and F.nubicola have not bloomed consistently at Cambridge Research station, so hybrids with these species in their lineage may not be well adapted for Rower initiation at this location. Fraganapentaphylla and F. gracilis accessions did not set fruit under isolated greenhouse conditions and were deemed self-incompatible. The few flowers of F.iinumae allowed to self-pollinate did not set seed. Pollen from F.iinumae (ii3) was

146 Figure 5.1. Hybrid vigour for vegetative growth of Evesca x F.daltoniana F, hybrids. The hybrid is depicted with selfed seedlings of its parents. All seedlings are three months ald. Note: F-vesca and F.daltoniana are self-compatible and not subject to inbreeding depression, so these seedlings are indicative of vigour present in the parents.

147 Figure 5.2. Diploid hybrids shown one year after planting. Hybrids resulting from these combinations of species have not been previously reported. A = F.vesca x F.pentaphyiia, B = F.pentaphyiia x F.nipponica, C = F.vesca x F.daltoniana and D = F.daltoniana x F.pentaphylla.

148 Figure 5.2 continued. E = Enubicola x Epentaphylla, F = F.pentaphylla x Enubicola, G = F.pentaphyila x F.viridis, and H = F.vjridis x F.pentaphyl/a.

149 Figure 5.3. Variation in fruit size and yield of interspecific diploid hybrids. The photo shows variability among lines arising from hybridization of different accessions of Fragaria pentaphylla (P 1 & P2) and F. viridis (V2 & V5).

150 134 viable because seed set occurred when ii3 pollen was used to pollinate Evesca accessions. Self-cornpatibility status of al! other species was consistent with Table 5.4. Reciprocal combinations differed for seed set arnong self-compatible (SC) species but were relatively less important among self-incompatible (SI) species (Table 5.3). Crosses among F.vesca and other SC species produced seeds and plants when F.vesca was used as a female parent, but produced no seeds when used as a male parent. Also, crosses between Fdaltoniana and F.nilgerrensis produced plants when Fdaltoniana was the female but did not produce seeds when F. nilgenensis was the female. Reciprocal combinations were identical for three combinations of SI species (F. vindis and Epentaphylla, F. vindis and Fmbicola, and F.nubicola and Enipponica) while other corn binations of SI species (ie. F. viridis x F.graciik and Enubicola x F.pentaphylla) were classified into group E with reciprocal combinations classified into group F. One combination involving SI species, F. vindis x Enipponica produced health y plants but reciprocal crosses produced seeds that did not geminate. Table 5.5 combines data from reciprocal crosses showing those combinations of species where geneflow has potential. SI species have the greatest potential for gene flow as 60% of the combinations resulted in fertile FI hybrids. Although fertile FI hybrids were produced by 50% of SC combinations, many different F.vesca accessions were used as parents to produce a few seeds of F. vesca x F. daltoniana and F. vesca x F-nilgerrensis. As noted earlier, one of these combinations, Evesca x F.daltoniana, was observed to have the lowest fertility of

151 Table 5.5. Results of interspecific crosses among several diploid Fragaria species, showing differences among SCZ x SC, SC x SI, and SI x SI combinations of species. Reciprocal combinations from Table 5.3 are combined into one value for each pair of species. SC x SC F. daltoniana F. iinumae F. nilgen-ensis F. vexa F. gracilis F. nipponica F. nubicola F. pentaphylla F. iinumae F. nilgen-ensis EY F. vesca F C F SC x SI SI x SI F. gracilis C F. nipponica C C E F. nubicola C A C F A C F. penlaphylla E E F F F F F. viridis C A B F F E F F = Self-compatible species = Self-incompatible species no seeds from cross seeds produced but no germination seedlings died after germination F, bloomed but had no seeds under open-pollinated conditions F, plants have not bloomed F, bloomed and produced seeds under open-pollinated conditions

152 13 6 any hybrid in Group F. Although seed set occurred for rnost combinations among SC and SI species, 43% of the combinations produced seedlings that died after germination (Group C). SC x SI hybrids which developed into plants had either Evesca or Epentaphylla as one of their parents. Combinations between SC and SI species produced fertile hybrids only 21 % of the time, and al1 fertile combinations had F. vesca as a female parent. Hybrids of F. daltoniana x Epentaphylla and F.ni/gerrensis x Epentaphylla were produced but have not bloomed, but six other combinations of SC and SI species produced unhealthy seedlings that died. Fragaria vesca was the cornmon denorninator for producing fertile hybrids between SI and SC species, but F.vesca was not able to provide a bridge to F.iinumae, F.gracilis, or F. nipponica. However, F. pentaphylla can produce fertile hybrids with al1 other SI species (which includes F.graci/is and F.nipponica) as well as the SC species F.vesca. Thus. potential exists for geneflow at the diploid level among eight diploid species, if Evesca and F.pentaphy/la are used as genetic bridges. In this study, no single species was identified with the potential to provide a bridge to F.iinumae, but this study was limited by having only one accession of Einumae that rarely produced flowers. Fragaria pentaphylla was superior to F. vesca for interspecific hybridization since it formed viable hybrids with eight other species in the study, compared to five with F. vesca. This is particularly surprising since far greater numbers of Evesca accessions were available for crossing and only two accessions of Epentaphylla were available.

153 DISCUSSION With the discovery that octoploid strawberry cultivars have a very narrow germplasm base (Sjulin and Dale, 1987), research in low ploidy species may be particuiarly beneficial for broadening the germplasm base in Fragaria. Researchers have suggested that tremendous advances could be made if strawberries could be bred at the tetraploid level instead of the octoploid level (Barrientos and Bringhurst, 1973; Niemirowicz-Szczytt, 1987; Owen and Miller, 1996). The arguments advanced for breeding with tetraploids instead of octoploids included simpler inheritance patterns, ease of selecting recessive traits, and the ability to create homozygous lines w hen chromosome dou bling occurs. Breeding and selecting of diploids would be even simpler than tetraploids or octoploids. Such breeding would require use of the SO systern (Evans, 1977) to facilitate geneflow between low and high ploidy Jevels. This study has a bearing on investigations to improve the synthetic octoploid (SO) system (Evans, 1977) by using F-vesca as a bridge to diploid species. Fragana vesca showed potential to act as a bridge byforming fertile hybrids with five diploid species but formed only lethal hybrids with three other species (Table 5.5). In addition, Fragana pentaphylla had greater potential as a bridge species than Evesca. It also formed fertile hybrids with five species, but in addition, viable hybrids were formed with two other species. It is interesting to note that viable progeny resulted from SC x SI crosses only when F.vesca or F.pentaphylla was one of the parents. The abilrty to serve as a genetic bridge among groups with different compatibility mechanisrns is indicative

154 138 of potential to assist breeding at the diploid level. Both F. vesca and F.pentaphylla are required to transfer genes among eight diploid species since only the former species formed fertile hybrids with Edaltoniana and F-nilgenensis and only the latter species formed fertile hybrids with F.graciiis and F.nipponica. The identification of 11 combinations of species (13 if reciprocal combinations are counted) that can produce fertile offspring, could be used for prebreeding of wiid species prior to use in the SO system. Pre-breeding could concentrate desirable traits and greatly increase the value of wild germplasm prior to introgression into commercial cultivars and is a strategy employed with polyploid species such as wheat (Gill et al ), soybeans (Lewers et al., 1998) and potatoes (Jacobsen and Jansky, 1989). Since chromosome doubling could restore fertility of interspecific hybrids (Eigsti and Dustin, 1955), fertility of F, hybrids at the diploid level is not an absolute requirement for the SO system. Thus, al1 20 combinations of species listed in Table 5.3 that produced viable plants could potentially be used for SO production. If al1 tetraploid hybrids were fertile, several diploid species could be used as genetic bridges. For example, F.pentaphyiia could substitute for F. vesca as a bridge to F. daltoniana or F. nilgerrensis. Similarly, F. vindis could replace F.pentaphylla as a bridge to F. nipponica. This strategy would significantly increase genetic diversrty in germplasm derived from the SO system. Previous studies (Table 5.1) attempted 15 combinations (28 counting reciprocals) of diploid species but only eight combinations (14 with reciprocals) produced viable plants. Of these, only three combinations of species (five with

155 135 reciprocals) produced fertile interspecific hybrids. This study investigated 28 combinations of species (38 with reciprocals) and found approximately three tirnes as many combinations that could produce viable plants and more than three times as many fertile combinations. This success was primarily due to having a greater number of species and accessions. Fragana pentaphylla had the highest level of crossability in this study but had not been previously investigated. Crossability of F.graci/îs had also not been studied previously. Using a large number of Evesca accessions to obtain a few rare seeds and then using an in vitro germination technique (Miller et. al., 1992) was important to obtain interspecific hybrids from SC x SC crosses. As many of the species in this (Table 5.2) and previous studies were represented by only a few accessions it seems likely that additional successful combinations will be found in the future, if more accessions become available. Certainly, differences among accessions may be partially responsible for variation of results obtained by earlier researchers (Table 5.2). In this study, accession differences may have contributed to the large standard errors for traits such as seeds I flower (Table 5.4). Of the 34 combinations of species in this study, 19 combinations had not been previously reported by other researchers. The results of previous studies (Tables 5.1) and this study (Table 5.3) are cornbined to show successful combinations in Figure 5.5, and those combinations which have not produced viable progeny are presented in Figure 5.6. The combined data indicates more possibilities for geneflow at the tetraploid level if hybrids become fertile when chromosome numbers are doubled.

156 Figure 5.5. Geneflow potential among diploid Fragana species, showing results of previous investigators combined with this study. Note that only F.inumae has not produced fertile progeny with other species. AH other species have produced fertile progeny with F.vesca andfor F.pentaphy//a. The interspecific progenies that have not been shown to be fertile may yet have genefiow potential if chromosome numbers are doubled, or after a longer period of observation. Based on Tables 5.1 and I.-*. Legend O self-compatible self-incompatible fertile progeny progeny created, fertility not yet demonstrated arrows go from males to females

157 Figure 5.6. Combinations of diploid Fragana species that have not resulted in viable progeny, but might be successful if more accessions were investigated. Combinations that have not been investigated or have been unsuccessful are depicted. Since self-compatible species cannot usually pollinate self-incompatible species, this combination of parents is not recommended. Based on Tables 5.1 and 5.3. F. viridis a HI---* cross Legend O self-compatible self-incompatible no viable progeny not attempted barrows go from males to females

158 142 Two combinations of species from previous research are worth mentioning. Plants had been created with combinations of F.nilgerrensis x Eiinumae (Sebastiampillia and Jones, 1976) and F-nilgerrensis x F. viridis (Evans, 1967, 1977), thus, indicating a possible bridge to F.inumae and an additional bridge to F.niigerrensis. In neither of these combinations was fertility of F, hybrids reported at the diploid level, but the latter combination was fertile at the tetraploid level. Figure 5.6 shows additional combinations of species that should be further investigated if more accessions become available, since increased numbers of accessions improve the likelihood of creating interspecific hybrids (Burbank, 1921 ; Hermsen, 1992). While nine diploid species were investigated in this study, more combinations of species could be crossed if F.manchunca (Staudt, 1989) or newly discovered species are made available. Another factor in the success of this study was that one type of cross was deliberately avoided, and is not recommended in Figure 5.6. Self-compatible (SC) species were not used to pollinate self-incompatible (SI) species. This type of cross is extremely unlikely to be successful (Evans and Jones, 1967). However, in rare instances where SI x SC crosses have resulted in progeny (Yarnell, 1931a; Williams, 1959; Fadeeva, 1966), mutations involving self-incompatibility alleles are a plausible explanation for success (Lapins, 1983). Since SC x SI crosses can be quite successful (Evans and Jones, 1967; Lapins, 1983), this type of cross was perfonned in this study while SI x SC crosses were avoided. Differences among SC x SC, SI X SI and SC X SI hybrids (Table 5.5, Figures 5.5, 5.6) may have implications for evolutionary theory of Fragaria. These data

159 143 suggest SI species are easier to intercross than SC species and that SI species are closely related while SC species are distantly related. Of the 15 SI x SI combinations attempted, 15 have resulted in hybrids and seven were fertile to date. Of the nine SC x SC combinations attempted, only four have resulted in progeny and only two of these four combinations have been fertile. It has been speculated that self-incompatibility arose from self-compatibility in Fragana species (Darlington, 1973) and that the widespread self-compatible F.vesca is a progenitor to other diploid species (Staudt, 1953; Nurnberg-Kruger, 1958; Darrow, 1966). However, these two theories are contrary to classic studies (Lewis and Crowe, 1958) that proposed SI mechanisrns orig inated early in the establishment of angiospems to further promote cross-fertilization and that most SC species have SI ancestors. Lewis and Crowe (1958) speculated that two types of SC species existed in plants. one ancient, that could pollinate styles of SI species and one more recent, derived from SI species, which would not pollinate SI species (Lewis and Crowe, 1958). Self-compatible species F.nilgerrensis and F. vesca would not pollinate selfincompatible species but reciprocal crosses were successful (Evans and Jones.1967). Although no assertions were made regarding evolution in Fragaria, the research by Evans and Jones (1967) supports the theory that F.nilgemnsis and F. vexa were likely derived from SI species. Geographical isolation not only leads to formation of new species but may cause SC to arise from SI species in order to continue seed production (Stebbins, 1950; Lewis and Crowe, 1958; Darlington, 1973). The four self-compatible species occupy unique ecological niches that may have lead to isolation and speciation:

160 i44 F.vesca is adapted to low light levels of forest ecosysterns (Chabot and Chabot, 1977; Angevine, 1983); Enilgerrensis has the rnost southern range of diploid Fragaria (Evans, 1964); Edaltoniana is found at the highest elevations (ie. Himalayas) for diploid Fragaria, (Reed, 1966; Staudt, 1989) and Eiinumae is found in the isolated mountain regions of northern Japan (Reed, 1966; Staudt, 1989). As these self-compatible species, F vesca, F. nilgenensis, F. iinumae and Fdaltoniana, occupy different ecological niches. have divergent morphology (Fig. 1.1, Reed, 1966; Staudt. 1989) and are difficult to intercross, they probably do not originate from a common self-compatible ancestor. It is more likely that mutations arose independently in each of these species during isolation which broke down SI mechanisrns, thereby becoming SC, in a manner similar to that discussed for other species by Lewis and Crowe (1958). It is noted that F.pentaphylla does not have the wide geographical distribution of Evesca. It may be that the progenitor species was restricted in its distribution and that new species were formed through adaptation to new environrnents. It was suggested that F.vesca rnight not be the progenitor species of the Fragada genus and that intercrossing of diploid species would be needed to provide additional insight into the identification of the progenitor species (Chapter 4). It was hypothesized that the progenitor species would cross more readily to other species, The results of this study indicate that the SI species, F.pentaphylla, rnay be more similar to the progenitor species of the Fragana genus than F.vesca. It is conceivable that F.vesca could have evolved from Epentaphylla or that bath species are intermediate steps in the evolution of other diploid species. Since fertile

161 14 5 interspecific hybrids can occur among many diploid species, an alternative theory is that F.penfaphylla and F.vesca are descendants of two or more species, which has resulted in improved crossability. Although F. pentaphylla may be superior to other diploids for interspecific hybridization, the SO system could have enhanced flexibility by using two species as a bridge to other species. For example, if F.pentaphyila is narrowly adapted. it may be preferential to use F-vexa as a bridge species whenever possible. Alternatively, attributes may be discovered in Epentaphylla (eg. pest resistance, cold hardiness, etc) that could make it preferable to use as a bridge species. In light of the earlier discussion of SC species evolving from SI species, it is interesting that the SC F.vesca crossed readily with other species. This suggests F.vesca has diverged more recently from SI species than other SC species. More similarities (ie. leaf morphology, flavours, chloroplast DNA) exist among F. vesca and the SI species than with other self-compatible species (Fig. 1.1, Evans, 1964; Darrow, 1966; Reed. 1966; Staudt. 1989; Harrison et al ). Fragaria vesca also produces more seeds when pollinated with SI than with SC species (Chapter 4; Table 5.4). While previous crossing studies revealed limited information on fertility of interspecific hybrids (Table 5.1), this study revealed that many combinations of hybrids were, in fact. fertile (Table 5.3). This supports other studies involving karyotyping (Iwatsubo and Naruhashi, 1989, 1991). chromosome pairing (Staudt, 1959; Jones,I 976) and DNA analysis (Harrison et al., 1993, 1997) that suggested Fragaria species are closely related. Except for F.iinumae which has one

162 146 noticeably different chromosome (Iwatsubo and Naruhashi, 1989), and has not produced progeny when crossed with other species, it seems likely that similartypes of genornes exist among the other eight species in the study. It has been suggested that F-vesca and F.vindis have A genomes while octoploid Çragaria has AAA'A'BBBB or AAA'A'BBB'B' genomes (Senanayake and Bringhurst, 1967; Bringhurst, 1990). The assignment of genomic designations in these two studies was mainly based on the ability to hybridize octoploids with F-vesca and Evindis. The data of this study suggests that diploid Fragana are very closely related and, therefore, most would possess A genomes. Without a confirmed B genome species at the diploid level, perhaps a genomic designation of AAA'A'A"A"A"'Anl would be more appropriate to explain genomic makeup of octoploid Fragana. Thus, it may be less important that cornmon, or pivotal, genomes be used in polyploid crosses for SO production. There may already be enough genomic homology among species to make this precaution unnecessary. It may be more important to insure that diverse genotypes are used in crossing so that multivalent pairing does not occur, particularly when chromosome doubling is used. Almost 50% of the potential combinations of species (Figure 5.6) have not resulted in viable offspring. About 20% of the total combinations have not been investigated while the remaining 30% did not result in progeny, emphasizing that crossability differences do exist among Fragada species. Certain genetic combinations may be lethai. This would seem to be the case in those combinations where seeds gerrninated but did not develop, or had embryos that did not geminate. However, only a few accessions for most species have been available

163 147 in this and previous studies. It is possible that additional successful combinations will be found if a wide gerrnplasm base (ie. more accessions) of these rare species is used. SUMMARY While the germplasm base of octoploid Fragaria cultivars is narrow, there is great potential for broadening the germplasm base at the diploid level. This study provides additional evidence suggesting diploid Fragaria species are closely related, and is the first study to show that geneflow could occur among most diploid Fragaria species. While only five combinations of species were previously shown to have fertile F, hybrids, this study has expanded that number to 14 combinations. Among the nine diploid species studied, F. vesca and F.pentaphyl/a are the most promising species to use as bridges, at the diploid level. With these two species geneflow could theoretically occur among al1 the diploid species in this study, except for F.iinumae. The successful combinations identified in this study could serve as an outline for prebreeding and improving diploid germplasm prior to use in the synthetic octoploid system. As this study had only one or two accessions for many species, it is quite possible that more successful combinations of species will be found as more accessions become available. The similarities among diploid species could become a factor in creating synthetic octoploids from diploid species. It may be not be important to design crossing strategies with common, or pivotal, genomes since the diploid species may already be closely related. This study also has implications for the evolutionary theory of Fragaria

164 i48 species. While Evesca produces interspecific progeny more readily than other SC species, the results of this study cast doubt on the assertion that it is the progenitor species of the genus. Instead, it is more likely that the progenitor species was similar to F.pentaphyila and that SC species evolved from SI species in Fragana.

165 149 Chapter 6 lnterspecific hybridization of Fragaria moschata (6x) with Enubicola (2x) and Eviridis (2x) ABSTRACT The crosses Fragana rnoschata x F.nubicola, Emoschata x F.vindis, and their reciprocal combinations were done to create tetraploids for eventual introgression into octoploid cultivars of Ex ananassa via the synthetic octoploid system. The combination F.nubicola x F.moschata averaged 3.3 healthy plantslpollination, which was 2.3 times more than the next best combination, Fmoschata x F.vindis. Fragaria viridis x Fmoschata crosses averaged only 0.1 heaithy plantslpollination and no plants were obtained from F.moschata x Enubicola hybridizations. The success of the Fmbicola x Fmoschata crosses in this study indicate that embryo 1 endospen balance ratios were not as important a factor in producing hybrids with F.moschata x diploids species, as previously suggested. Although 90% fruit set and 19 seeds/pollination occurred for the experiment, only 15% of the seeds contained healthy, white embryos and only 38% of these developed into healthy seedlings. The remaining 62% of embryos lacked pigment, had short hypocotyls or did not continue to develop after slight expansion out of the seed coat. Only 39% of the crosses resulted in viable progenies but al1 six diploid and six of the seven hexaploid accessions were successfully incorporated into 89 hybrids. After one year, 27% of these hybrids bloomed and 71% of the blooming

166 hybrids were tetraploids. WTRODUCTION This experiment contributes to the Synthetic Octoploid (SO) system (Evans, 1977) by focusing on creating tetraploids from the hexaploid Çragaria moschata for eventual introgression of this species into octoploid cultivars of F.x ananassa. Since F. rnoschata may be a naturally occurring hybrid of F.vesca and either F.viridis or Enubicola (Federova, 1946; Staudt, 1959; Darrow, 1966), the creation of hybrids in this experiment is useful for later research focusing on the use of Evesca to create synthetic octoploids (Chapter 7). Of the 15 SOS created by Sebastiampillia and Jones (1976) and Evans (1 977), nine had F. moschata in their lineage. The two strategies for incorporating hexaploid F. moschata into SOS are depicted in Figure 6.1. Only a very limited amount of Fmoschata germplasm was introgressed into cultivars. A fertile SO released by Evans was created by crossing F.moschata (6x) x F.nubico/a (2x) and applying colchicine to seeds to create an octoploid hybrid (Evans 1982a, 1982~). Guelph SOI was semi-fertile and was used in breeding programs in Guelph (Sangiacomo and Sullivan, 1994) and Sweden (Trajkovski, 1990, 1993). Kantor (1984) introgressed F.moschata into F. x ananassa by crossing Emoschata directly to cultivars and using rnutagens to overcome sterility in the 7x hybrids. Kantor abandoned this line of research because restoration to the 8x level required several

167 Fig Strategies used by Sebastiampillia and Jones (1 976) and Evans (1 977) to create synthetic octoploids with the hexaploid Fragaria moschata. Numbers represent ploidy levels of species and hybrids. Parents are on the left and progenies are on the right. Strategy A involved applying colchicine directly to hybrid seedlings. Strategies B,C and D used F.moschata x diploids hybrids in crosses with other tetraploids. 6x 'Ti->*' 8x colchicine / colchicine 2x 4x / colchicine colchicine

168 152 generations of backcrossing, after which the backcrossed progeny no longer possessed the desirable traits from F. moschata. Several desirable traits have been identified in F. moschata that have potential for introgression into cultivars. These include resistance to leaf diseases (Evans, 1982~; Kantor, 1984; Trajkovski, 1993; Maas et al., 1995), resistance to spider mites (Gimenez-Ferrer, 1993), cold hardiness (Reed, 1 966; Kantor, 1 984), large plants more vigorous than octoploid species (Reed, l966), unique, hig hly aromatic Ravour (Reed, 1 966; Kantor, 1 984), upright flower stalks that hold fruit above the canopy (Evans, 1982; Kantor, 1983) and unifom ripening (Reed. 1966). For SO production, the hexaploid Fmoschata can be crossed to a diploid species to create tetraploid hybrids. Table 6.1 summarizes investigations involving crosses among F.moschata and diploid species. Crosses between Fdalfoniana, F.nilgerensis, F.nipponica, or F. vesca, and accessions of F.moschata were unsuccessful but fertile hybrids were produced when F.moschata was crossed with Eviridis (2x) or Enobicola (2x) (Evans, 1964, 1977; Bhanthurnnavin, 1965; Fadeeva, 1966; Bumagima and Mesaros. 1969). Thus, this study focused on using F. vindis and Fmbicola. Althoug h F.nubicoia x F.moschata crosses have not been successful (Table 6.1), this combination of species was included in this experirnent, since the reciprocal combination was successful. This experiment was done with increased numben of accessions (compared to earlier studies) to improve the likelihood of creating

169 Table 6.1. Outcornes of interspecific crosses between Fragana moschata and diploid species. Species combination Reference F. moschata x F. vesca F. vesca x F. moschata Mangelsdorf and East, 1927 Federova, Evans, 1964, 1974 Fadeeva, 1966 Bumagima and Mesaros, 1969 no set no seedlings good set. 600 seed, 4 germinated, al1 hybrids were weak and died within 2 years 5430 seed, 15 plants, 1 1 materna1 diploids, 2 tetraploid hybrids, 2 pentaploid hybrids 15 pollinations, 1 seed, O 6 pollinations, O seed gerrninated no germination or lethal seediings sterile plants and seedlings that died shortly after germination 0.04% set, 2 seedlings Ç. moschata x F. viridis F. viridis x F. moschata Evans, 19 pollinations, 18 seed, 3 7 pollinations, 17 seed, 2 germinated but 1964,1974 germinated died soon after Fadeeva, seedlpol., 38 seed, seedlpol., 64 seed, 4 germinated, germinated, hybrids were hybrids were fertile fertile Bhanthumnavin, 1965 obtained hybrid plants Bumagima and Mesaros, Oh set, S. 1-3% gem. 5 hybrids F. moschata x F. nubicola F. nubicola x F. moschata Evans, 1964, 28 pollinations, 85 seed, 26 3 pollinations,l4 seed, 1974 germinated O germination Bhanthumnavin, 1965 Evans, 1977 obtained hybrid plants obtained fertile 4x and 8x hybrids

170 Table 6.1. continued. Species combination Reference F. moschata x F. nipponica F. nipponica x F. moschata Fadeeva, 1966 Menfield, 1933 sterile plants fertile hybrids 273 seed, 1 sterile plant F. moschata x F. nilgerrensis F. nilgerrensis x F. moschata Evans, pollinations, no set 6 pollinations, no set F. rnoschata x F. daltoniana F. daltoniana x F. moschata Evans, pollinations, no set 2 pollinations, no set

171 15 5 interspecific hybrids (Burbank, 1921 ; Hermsen, 1992). The objective of this study was to examine crossability of a larger number of accessions and to determine optimum combinations of species to create tetraploids with Fmoschata parentage using the germplasm available. It was hypothesized that either Fmbicola or F. vindis would be superior for crossing with F.moschata. MATERIALS AND METHODS The three species used in this study are depicted in Fig and sources of germplasm are listed in Table 6.2. The Fmoschata accessions originate from diverse regions, in France, Italy, Germany and Russia. The F.nubicola accessions are from Pakistan, while three of the four F.viridis accessions are from Germany. The fourth F.vjridis accession (vir-5) was obtained from a botanical garden but its geographical origin is unknown. Plants were field grown at Cambridge Research Station, Cambridge, Ontario, dug when dormant in late fall, and forced in the University of Guelph's greenhouse during winter. Fragarfa nubicola accessions required short day treatment in a growth chamber (i.e. four weeks at 10 hr. day length, C daylnight temperature) to encourage bloom. A total of 127 pollinations were made including 47 different combinations of parents. An effort was made to have similar nurnbers of crosses for each species combination. For many combinations of parents, one or two pollinations were made. Additional crosses were made for those accessions producing abundant fiowers. As Fragana moschata is triecious with plants being either male, female, or (rarely)

172 156 Figure 6.2. The three Fragaria species used in this study: Emoschata, F. vindis and F. nubicola.

173 Table 6.2. Accessions of Fragana moschata (6x), F.nubico/a (2x) and F.vindis (2x)used in crossing program for tetraploid production Species Guelph # Identifier used by SourceY Se? accession source t~ PeS Fmoschata mcvl "Capron" PI "Profumata di Tortona" PI NCGR NCGR NCGR NCGR h f f f+m clone clone clone family FRA 9193 IPK f clone PI NCGR m clone mw5- abcdefik #2 "Cottan IPK f+m family F. nubicola nu b 1 PI NCGR h clone Pl NCGR h clone F. viridis virl PI NCGR h clone Pl NCGR h clone #2 "Oberau 2" IPK h clone RBGS h clone ': mcv = Fmoschata cultivars nub = Fmbicola mw = F.moschata wild vir = F. viridis Y: See Appendix A for full names and addresses of germplasm sources. X: f = female m = male h= hermaphrodite W: Letten listed after numbers are individuals of a family. For example, four members are in the mw-2 family.

174 158 hermaphrodite (Staudt, 1989). reciprocal crosses cannot be done with most F. moschata accessions. The only hermaphrodite F-moschaia in the study. mcvl, was included in additional crosses since it was perceived to be valuable germplasm. As F.nubicola required forcing to bloom, it flowered later than many F.moschata accessions and consequently participated in only 10 crosses as a male with Emoschata. Since Fmoschaia pollen was stored, Enubicola could be pollinated with earlier blooming F-moschata accessions. Reciprocal crosses were not possible with Fmoschata accessions except for the hermaphrodite clone mcvl. Fragana nubicola and F.vindis accessions are hermaphroditic and were used as both male and female parents. Seeds were genninated in vitro according to Miller et al. (1992) with slight modification. Rather than cutting cotyledons in half. only 30% was removed. Observations were made during dissection of achenes and only white healthy embryos were placed on media for germination. Small white embryos that did not completely fil1 were included in the healthy category since they were found to be capable of geninating in vitro. Seeds were discarded if embryos were brown, offwhite or yellowish. Previous research had shown that these three types of embryos would not germinate. Further details regarding pollination and in vitro germination of seedlings have been presented in Chapter 2. Five dependant variables were studied: percent fruit set, seedslpollination, embryoslpollination, heathy plantslpollination and one year old hybrids that bloomed/pollination. Descriptions of the first four variables are contained in Chapter

175 Ploidy levels of hybrids that bloorned after one year were estimated with fiow cytometry (see Chapter 3 for details). Means for each combination of parents were used in statistical analysis of differences among species. Data were analyzed with the GLM program of PC SAS (SAS System for Windows V.6.12, Cary, N.C.). Mean separation was accomplished using the LS means function at the 5% level of significance. Statistical analysis was used to determine differences among species combinations and to determine if a species was more successful as a male or female parent for interspecific crosses. RESULTS Means for species combinations are presented in Table 6.3 whiie data for individual crosses and ANOVAs are presented in Appendix 3. Percent fruit set averaged 90% and seed production averaged 19 seedslpollination for the experiment. From a total of 2032 seeds that were dissected for in vitro germination, 296 healthy, white embryos were obtained. Of the remaining seeds, 1493 and 243 contained brown or discoloured embryos, respectively, and were considered nonviable. Alrnost al1 healthy embryos showed signs of growth when placed in vitro but only 38% developed into healthy seedlings. The remaining 62% of in vitro embryos had lethal deformities such as lack of pigment (albinos), lack of hypocotyls, and arrested development after slight expansion out of the seed coat. In total, 89 healthy hybrids were produced, 24 of these bloomed when approximately one year old, and 17 of the blooming hybrids were tetraploids. The plants that were not

176 160 tetraploids consisted of a pentaploid, a 4x16~ mixaploid, two hexaploids from 2x x 6x crosses and a diploid also from a 2x x 6x cross. Plant morphology of al1 hybrids, except a diploid resulting from nub2 x mwzd, was indistinguishable from F.moschata, regardless of ploidy level. This exception had leaves similar in width and pubescence to F.nubico/a and was intermediate in length and shape when corn pared to Fmbicola and Fmoschata. Means of % fruit set. seedslpollination, embryoslpollination and healthy plants/pollination, and one year old hybrids that bloomedlpollination were calculated for four combinations of species: Emoschafa x F.nubicola, F-nubicola x Fmoschata, Fmoschata x F. virids and F. vindis x Emoschata (see Table 6.3). Only the means of embryosfpollination and healthy plantslpollination were significantly different among the four species combinations used in crosses. The combination Fmbicola x Fmoschata produced significantly larger numbers of healthy embryos and healthy plants than other species combinations. This combination produced almost three times as many healthy embryos and 2.3 times more healthy plants than the next best combination of Fmoschata x F.vindis. Although fruit set and seeds were produced, no healthy plants were obtained from the combination of F-moschata x F.nubicola. Although F. vindis males produced 1.3 more healthy plantsfpollination than F. vindis females in crosses with F. moschata, this difference was not significant. While 90% of various combinations of parents used in Fmbicola x F.moschata crosses resulted in healthy plants, only 25% and 16% of combinations of parents used in F-moschata x F.vindis and F-vindis x F.moschata crosses. respectively, produced any healthy plants.

177

178 102 There were only a few combinations of accessions that were superior in creating hybrids. It is interesting to note the best three combinations of accessions for production of healthy plants were crosses arnong nub2 and three accessions in the mw5 family (ie. mw5d, mw5f and mw5i). These three combinations averaged seven to 10 healthy plants per pollination. The crosses mcvl x vir2, mcv2 x virl, mw2c x vir and nubl x mw2d were the second most successful, producing five to six healthy plants per pollination. All other combinations of accessions produced on average one or less progeny per pollination. Although vir5 x mw5f and mw5k x nubl produced the most seeds, (ie. 65 and 62 seeds per pollination, respectively), neither combination produced healthy plants. Although only 18 of 46 combinations of accessions were successful in creating viable progenies, al1 of the diploid species and most of the Fmoschata accessions were successfully incorporated into hybrids. The Emoschata accessions that were not incorporated into hybrids included mw3b and some members of the mw2 and mw5 families. Most of these unsuccessful accessions were only used in one or two crosses. DISCUSSION The methodology and hybrids resulting from this study greatly expands the genepool of Fmoschata that could be used for introgression into strawberry cultivars. Although, several authors have created tetraploids with F.moschata, none of the hybrids mentioned in Table 6.1 were listed in genebanks contacted (see Appendix l), and the tetraploids that Evans had created, no longer existed at the

179 163 University of Guelph. Only two F-moschaia accessions have been introgressed into cultivars (Evans, 1982c; Kantor, 1984). The 89 hybrids in this study were derived from five clones and two open-pollinated families of F.moschata (Table 6.1) Compared to previous crossing (Table 6.1), this study used a broader collection of Emoschata and created more interspecific hybrids. While Evans (1964) obtained 31 hybrids between F.moschatû and diploid species, al1 other reports either did not show how many hybrids were created or have reported zero to seven hybrids. Much of the success of this study may be due to adopting a crossing strategy of maximizing the number of combinations of parents. Having a wide genetic base is especially important for interspecific crosses (Herrnsen, 1992). To increase success of wide crosses, using many parental genotypes for a few interspecific hybrids is superior to using a few parents for many crosses (Burbank, 1921; Galletta and Maas, 1990). An in vitro technique for improved germination (Miller et al., 1992) was important for obtaining hybrids. Previous research involving Emoschata hybrids probably germinated whole seed on soil-less mix. When used on F. x ananassa seeds, the whole seed method had 27% germination but the in vitro method had 97% germination (Miller et al. 1992). The use of the in vitro germination (Miller et a1.j 992) technique in this study provided information not reported by previous researchers. This stud y s howed that embryo abortion is a primary cause of low germination rates with seed produced by crossing F.moschafa to either F.nubicola or F-vindis. The prevalence of normal sized seeds that contained brown shriveled embryos, indicates embryos aborted at

180 104 a stage sometime after the outer seedcoat had reached maximum size. Discoloured embryos which did not develop were likely to have aborted in the later stages of seed development. Disruption of normal development was obsewed beyond the seed development stage. During in vitro germination a number of aberrations were observed including albinism, lack of hypocotyls and arrested development that could have prevented germination with traditional methods. It would appear there is no specific developmental stage in which Emoschata x diploid hybrids are likely to abort. Rather, a number of developmental events may be affected. As fruit set averaged 90% for the experirnent, there apparently is little or no influence of selfincompatibility rnechanisms. Ideas regarding interspecific crossability often change as more accessions become available or as new techniques such as embryo rescue are used (Hermsen, 1992). In this study, a diploid x hexaploid combination, Enubicola x Fmoschata, resulted in the highest numbers of embryos and healthy plants. Evans (1974) suggested embryolendosperm balance was responsible for difficulties in obtaining diploid x hexaploid hybrids and recommended that the hexaploid species be used as a female in crosses with diploid species. Although, Evans (1974) based his recommendation on several combinations of species, including octoploids and tetraploid hybrids, he had used only one combination of parents for each of Enubicola x Fmoschata and F.viridis x Emoschata crosses. While Evans was not successful in obtaining diploid x Fmoschata hybrids, other researchers were successful with F.vjrids x F.moschata crosses (Fadeeva,l966; Bumagima and Mesaros, 1969). Researchers crossing parents with different ploidy levels often use

181 165 the higher ploidy parent as the female parent (Stoskopf et al., 1993), but a notable exception is in Helianthus species where accessions of the lower ploidy species can be the most effective female (Georg ieva-todorova, 1984). Other factors could be more important than ernbryo/endosperrn ratios when Emoschata is crossed with F. vindis or F.nubicola. An explanation for embryo abortion and malformed hybrid seedlings in this study may be chromosome instability or chromosome elimination. Chromosome elimination likely occurred with the diploid hybrid that had some characteristics of F.moschata (Table 6.4). A combination of chromosome elimination and spontaneous chromosome doubling may have occurred with the 4x16~ mixaploid and the two hexaploids derived from diploid x hexaploid crosses. The pentaploid hybrid could have been derived from an unreduced gamete of the diploid parent that fused with a 3x gamete from the hexaploid parent similar to hexaploids produced when F. vesca was hybridized with Fmoschata in an earlier study (Federova, 1934). Fragana moschata rnay also harbour lethal or sub-lethal alleles. The occurrence of lethal and semi-lethal genes at high frequency in obligate outcrossen is well known to geneticists (Stebbins,l950). As F.moschata is polyploid and an obligate outcrosser (except for the rare hermaphrodite forms), recessive lethal alleles could easily accumulate in natural populations of this species. Song et al. (1995) showed that rapid genome change occurs in derived polyploids of Brassica and suggested that polyploid species can generate extensive genetic diversity in a short period. If deletions or mutations have occurred in F.moschata, a reduction in ploidy level from 6x to 4x would reduce the buffering capacity offered by the

182 106 additional two sets of chromosome. Differences in cytoplasm or interactions between cytoplasm and nuclear DNA could be involved in obtaining hybrids (Stoskopf et al., 1993) between F. moschata and diploid species. Unlike previous attempts (Table 6. l), the F. nubicola accessions in this study were particularly suited as female parents when crossed to Fmoschata. Conversely, no hybrids were obtained from Fmoschata x F.nubicola crosses, althoug h other researchers were successful with that combination (Evans, 1964, 1974; Bhanthumnavin, 1965). Both this and Evans' (1964,1974) studies used only two accessions of F.nubicola, so it is not surprising that results would Vary among these studies. The 1.4 plantslpollination for Fmoschata x F. vifidis crosses in this study were much higher than the 0.16 and 0.05 plantslpollination obtained by Evans (1964) and Fadeeva (1966). The reciprocal combination, F.vitidk x F+moschata, had somewhat similar results to previous research with 0.1 plantslpollination in this study compared to earlier studies averaging O (Evans, 1964), 0.36 (Fadeeva, 1966) and 0.07 (Bumagima and Mesaros, 1969) plantslpollination. The putative diploid ancestors, F.nubicola, F. vindis or F. vesca (Staudt, 1959, 1984, 1989), may not be fully compatible with F. rnoschata if the latter has undergone genomic changes. Recombination among different genomes may have occurred in Emoschata since pairing readily occurç in diploid interspecific hybrids (Jones, 1976) and the three putative ancestors are similar enough to create fertile hybrids (Chapter 5). Fragaria moschata accessions may be highly heterozygous so that interspecific crossability varies greatly among accessions.

183 167 If it is accepted that Fmoschata has a genomic constitution of AAAABB (Federova, 1946; Fadeeva, 1966), it seems likely that the B genome in Fmoschata is similar to F.nubicoia or F-viridis genomes, since this study and others (Evans, 1974, 1977; Fadeeva, 1966) obtained fertile tetraploid hybrids when Emoschata was crossed with F-vindis or F-nubicola. These resulting tetraploid hybrids would be fertile because they have two corn plete sets of different genomes (je. AABB). Hybrids with F.moshata in their lineage may be reflecting the low germination inherent in Fmoshata. The accessions of mw2, mw3 and mw5 were originally obtained from seed that had in vitro germination rates of 16%, 1.3% and 8%, respectively (un pu blis hed data by author). Mean germination rate of these accessions was 6.0% while the germination rate of hybrids in this study was 5.8%. The germination rates of the original F.moschata parents was much lower than the 21 other accessions of various Fragaria species from the same genebanks, which ranged from 12 to 96% with an average of 58% germination (unpublished data by author). It is unlikely that detrimental alleles were introduced into hybrids from either F.nubicola or F.vindis. Malformations in seedlings did not occur when F.nubicola and F.viridis were crossed to diploids in other experiments (Chapters 4 and 5). Although this study used more combinations of accessions than previous studies, far too few accessions were used to make generalizations regarding interspecific relationships. Of the accessions used in this study, the most productive crosses were male F.moschata plants crossed ont0 F.nubicola and female F.moschata plants pollinated with F.vindis. The data of this study and earliet

184 10s research combined indicates F.moschata can be crossed both as a male and as a female with Enubicola and F. viridis. CONCLUSION This study was successful in obtaining more Fmoschata x diploid hybrids with a larger gemplasm base than previously reported. In particular, this study is the first to show that viable hybrids of F.nubicola x F.moschata can be created. While low germination of F, seeds was noted in previous research, it was not known that aborted embryos and malformations in seedling progeny were major factors limiting Fmoschata x diploid hybrid production. The success of the Enubicola x F.moschata crosses in this study showed that embryo I endospem balance ratios were not as important a factor in F.moschata x diploids crosses as previousiy suggested by Evans (1974) and that there does not seem to be any advantage in using Fmoschata as a female parent when crossing to Ç.nubicola. Chromosome elimination, presence of lethal or sub-lethal alleles in F.moschata, and a reduction in buffering capacity associated with a decrease in ploidy from 6x to 4x may be important factors in disrupting growth of hybrid embryos and germinating seedlings. A total of 18 of the 46 combinations of parents were identified as capable of creating hybrid progeny. While many combinations of accessions did not create viable offsprhg, most available germplasm was successfully incorporated into at least one tetraploid progeny. Thus, this study demonstrated the importance of attempting many different crosses and using a wide gemplasm base during intenpecific hybridization. The diversity of hybrids created in this study could be

185 169 especially beneficial for research into the final stages of the SO system (ie. intercrossing of tetraploids with colchicine treatment of progeny) since earlier attempts of intercrossing Fragada tetraploids had a narrow genetic base (Chapter 1). In particular, the tetraploids created in this study are allopolyploids and, thus, differ from autoploids and naturally occurring tetraploids, which will be used for tetraploid breeding in the next chapter.

186 170 Chapter 7 The use of Evesca, tetraploid and hexaploid Fragaria species to incorporate diploid Fragaria species into synthetic octoploids ABSTRACT The synthetic octoploid (SO) system is a method of combining diploid, tetraploid and hexaploid Fragaria species into hybrid octoploids. The SOS produced could then be introgressed into 8x cultivars of the commercial strawberv, F.x ananassa. This study used F-vesca, Eonentalis and Emoschata as genetic bridges, incorporated the pivotai genome theory in its crossing strategy, used a wide gerrnplasrn base, maximized heterozygosity in the crossing design, and used advanced techniques for germination and colchicine delivery to improve the SO system. In this study 192 SOS were created from 25 combinations of species, with 11 fertile combinations. Of the 98 SO's planted in 1997,28 produced fruit in The number of SOS created in this study represents a significant improvement over previous attempts to create SOS. This strategy used colchi-tetrapioids derived from F. vesca crossed with F.nilgerrensis, F.nubicola, Epentaphylla. and F. vindis for crosses at the tetraploid Ievel. These tetraploids produced, on average, only 37% fruit set and three seedsmower when intercrossed. When the colchiploids were pollinated with tetraploids derived from F.orientaiis or F.moschata, 52% fruit set and 10 seedslflower were obtained. Tetraploids containing two genomes of F.vesca were not as fertile as tetraploids without any Evesca genomes, but F.vesca was useful

187 171 in introgressing four diploid species into SOS. The fact that so many combinations of species were capable of producing progeny at the tetraploid level suggests that the species involved in this study are closely related and that there may be great potential for cornbining many species and breeding at the tetraploid level in Fragana. INTRODUCTION The commercial strawberry, Fragada x ananassa, is an octoploid species. Within the genus of Fragana there are ten diploid, three tetraploid and one hexaploid species known, which are generally not used by breeders because crosses among species with different ploidy levels result in poor seed set, poor germination andfor sterile progeny (Evans, 1976; Kantor, 1984). By crossing lower ploidy species and applying colchicine, synthetic octoploids (SOS) can be created which can be introgressed into cultivars (Evans, 1977,1982a, 1982b, 1982~). There are 12,769 combinations of Fragana species which could be investigated to create SOS in two generations. Previously, SOS were created from available accessions, without attention to species combinations that might have a high rate of success. Rather than attempt al1 possible combinations of species, this study focused on the use of F. vesca, F-moschata, and F.onentalis in SO production as bridges to incorporate some of the d iploid Fragana species (F. nilgerrensis, Enubicola, F.pentaphyila, and F. vindis) into synthetic octoploids.

188 172 Previous crossing strategies to create SOS One strategy to create SOS was to cross 2x. 4x. or 6x species directly to octoploid cultivars (Figure 1.5). This strategy relied on unreduced garnetes, colchicine, or required several generations of crosses among different ploidy levels to restore hybrids to the octoploid level. These unbalanced ploidy crosses rarely produced viable seed (Evans, 1974) or resulted in progeny with odd numbers of chromosomes which had dysfunctional chromosome pairing in meiosis and sterility (Darrow, 1966; Kantar, 1984). The Synthetic Octoploid (SO) method involves creating tetraploids from 2x, 4x, and 6x species and treating tetraploid hybrids with colchicine to double chromosome numbers (Evans, 1976; Figure 1.6). Advantages to this approach include: parents have the same ploidy level, embryolendosperm ratios are the same as intraspecific hybrids, rare unreduced gametes are not required, hybrids have even numbers of chromosomes, and only one or two generations of crosses are required. Tetraploid breeding in Fragaria Previous studies investigated tetraploid breeding (Table 7.1 ) while others attempted to create SOS from tetraploid crosses (Table 7.2). In this latter group, only three SOS were introgressed into octoploid cultivars (Evans, 1982~) and two Sots were released for use by breeders (Evans, 1982a.b). There is no further report on the eight SOS created by Sebastiampillia and Jones (1976). The University of Reading germplasm used by Sebastiampillia and Jones

189

190

191 Table 7.2. Studies that created Synthetic Octoploids (SOS) from crosses among tetraploid Fragana species and hybrids. Note that Evans (1977) reported on fertility of SOS but Sebastiampillia and Jones (1976) did not. Female Male # of # of sun/ivors # of SOS # of fertile seedlings after colchicine SOS Reference F. moupinensis F. corymbosa 6 F. orientalis F. moschata x F. nubicola 16 Sebastiarnpillia and Jones, 1976 F. (vesca x viridis) F. moschata x F. nubicola 16 F. (moschata x viridis) F. orientalis 8 F, (moschata x viridis) F. corymbosa 16 F. vesca (4x) F. moschata x F. viridis 4 F. moschata F. nubicola F. (moschata x nubicola) F. orientalis Evans, O O F, (moschata x nubicola) F. vesca x F. viridis O O F. moupinensis F. nilgerrensis x F. vindis O O Fmoupinensis F. moupinensis O F. (nilgerrensis x viridis) F. moupinensis O F, (vesca x viridis) F. moupinensis F. (vesca x viridis) F. vesca x F. viridis O O

192 176 (1976) and Evans (1964, 1977) was derived from the Max Planck Institute in Gerrnany (Evans, l964), where Staudt (1959) researched. Thus, it would appear that previous studies involving tetraploid crosses are in fact based on closely related if not identical germplasm. Evans (1964) lists 13 accessions of Evesca, four of Fmoschata, two of F.moupinensis, two of Enubicola, and one each of F. vindis, F.niigenensis and F-orientalis as being the founders of his hybrids. The hybrid tetraploid material created by previous research was not available for use in this study. It most likely no longer exists, or at least was not listed at genebanks contacted in the U.S., Canada, U.K., or Russia. Thus, it was necessary to create tetraploids for this study (Chapters 2, 4 and 6). Emphasis on F. vesca, F. orientalis and Emoschata in crossing strategy The relationship among crossing strategies of Chapters 4, 5, 6 and this experiment is shown in Figure 7.1. Those three chapters involved diploid breeding and tetraploid creation while this study focused on tetraploid breeding and octoploid creation. Figure 7.1. also encompasses the strategies presented by Evans (1977 ; Figure 1.6) ). Although this and earlier studies are similar in the way ploidy levels are combined, this study is unique in emphasizing F.vesca, Eorientalis and F.moschafa in crossing strategy. Fragana vesca possesses many worthwhile attributes (Chapter 1) that rnay be useful to improve SO production, particularly if F.vesca could act as a bridge species to facilitate gene fiow. As a diploid, tetraploid or interspecific tetraploid hybrid Fragaria. vesca has hybridized with many species including octoploid

193 177 Figure 7.1. Crossing strategy: a simplified balanced ploidy strategy for Synthetic Octoploid production. Diagrams show ploidy levels of species and hybrids involved in crosses. Parents are on left and progeny on rig ht. Diploid breeding Tetraploid creation 2x 4x chromosome dou bling Tetraploid breeding Octoploid creation chromosome doubling

194 178 cultivars (Yarnell, 1931 a; Federova, 1946; Bringhurst and Khan, 1963; Evans, 1964; Darrow. 1966; Senanayake and Bringhurst, 1967; Bauer, 1976,1993; Ahmadi and Bringhurst, 1992). Fragada vesca was more responsive than other Fragana species to chromosome doubling with colchicine (Sebastiarnpillia and Jones, 1976) and both nuclear and cytoplasmic inheritance were involved (Fadeeva, 1966). The use of F.vesca as a bridge species in SO production might be enhanced if it were used as a pivotal genome in crosses at the tetraploid level [ie. (F.vesca x diploid) (4x) x (F-vesca x diploid) (4x)I. A pivotal genome is one shared by polyploid species that enhances interspecific crossability in the presence of uncommon genomes, (Zohary and Feldman, 1961). While the pivota1 genome theory has been beneficial for introgression of wild species in grain crops (Feldman and Sears. 1981; Gustafson and Dera, 1989; Gustafson and Sears, 1993) it has not been exploited for fruit crops. The (F. vesca x diploid) tetraploids used in this experiment were colchiploids. Colchiploids of fruit species often have reduced fertility and may require crossing with d issim ilar genotypes to maximize heterozygosity and restore fertility (Sanford, 1983). Thus, it may not be feasible to use colchiploids as both the female and male parents in a cross since fertility problems would be compounded. Instead, it may be advantageous to cross colchi-tetraploids with (hexaploid x diploid) hybrids or naturally occurring tetraploid species so that at least one of the parents will be highly heterozygous and fertile. For this reason, tetraploids derived from Emoschata x diploids crosses (Chapter 6) and the naturally occurring tetraploid, F.onenta/is, were included to cross with the (Evesca x diploid) colchi-tetraploids.

195 Objectives and hypothesis The choice to emphasize Evesca, F.orientalis, and F-moschata in this study limits the possible combinations of species that could be investigated, but further explores Evesca's potential as a bridge species and increases the potential for production of fertile progeny. The objective of this study is to determine the potential of using Evesca in the SO system in combination with F-orientalis (4x) or tetraploid hybrids of Fmoschata x diploid species. In this study it was hypothesized that colchiploid hybrids derived from F. vesca x diploid species crosses could be used as bridges for diploids species in the SO system. It was also hypothesized that crossing these colchiploids to F-onentalis or Emoschata x diploid (4x) accessions will be more successful than intercrossing colchiploid accessions. MATERIALS AND METHODS Gerrnplasm The 46 accessions listed in Table 7.3 were used to create 104 tetraploids for this experiment (Tables 7.4,7.5). Five types of tetraploids were used in this study: 2x x 2x hybrids (Chapters 4 and 5) treated with colchicine, 6x x 2x hybrids (Chapter 6),4x x 4x hybrids, colchiploid Evesca (4x). and F-orientalis (4x). The ploidy level of each parent used in the study was verified by fiow cytometry (Chapter 3).

196 Table 7.3. Fragaria accessions that were founders of tetraploid gennplasm for this experiment. This table can be used to trace the lineage of tetraploids in Tables 7.4 and 7.5. Note that many accessions of Fragana vesca were used but other species had limited representation. Continued on next page. Species Ploidy Guelph Source identifier Source identifier Fmoschata 6x mcvl Capron,PI NCGR mcv2 Profumata di NCGR Tortona,P mw2c,dy PI NCGR mw4 PI NCGR mw5e,f Cotta #2 IPK F.nilgerensis 2x ni1 1 CFRA NCGR ni12 Mt. Omei Forest Farm F. nubicola 2x nubl PI NCGR nub2 PI NCGR F. orientalis 4x oril,2,13, 4947 x 4940 Dr. Gunter Staudt 20,a,b,k F. pentaphylla 2x pen RBG Edinburgh pen RBG Edinburgh F. vesca 2x vcv2 vcv6 vcv7 vcvl O vcvl 1 vcv13 vcv15 vcv20 Alpine, PI Golden Alpine, CFRA Pineapple crush Alexand ria Ruegen Mignonette Crimson Sprite A02 alpine Strawberry Baron Solemacher FRA 2193 Berry Sweet Reins des Vallees Pl Pi PI PI PI PI PI PI NCGR NCGR wvl7 Pl NCGR Park Seed Thompson & Morgan R ickte rs Park Seed Ornamental edibles White Flower Farm U. of N.H. IPK Henry F. Michel1 Co. SAPM NCGR NCGR NCGR NCGR NCGR NCGR NCGR NCGR

197 Table 7.3. Continued. Species Ploidy Gueip h Source identifier Source identifier F. vesca 2x vw19 East Malling F. vesca U. of Guelph vw22 FRA IPK vw25 PI NCGR vw34 Cambridge, ON collected by author vw35 PI NCGR vw39 CFRA 988 NCGR vw47 PI NCGR vw50 Rockwood, ON collected by author vw53 Newcastle, ON collected by author vw6 1 Niagara Falls, ON collected by author F. viridis 2x virl PI NCGR vir2 PI NCGR vir RBGS ': Full names and addresses of germplasm resources are listed in Appendix 1. Letters following numbers indicate members of a family. For example, mw2c and mw2d are derived from the sarne seed lot.

198 Table 7.4. Tetraploid Fragana germplasm derived from diploid species used for synthetic octoploid production. All tetraploids in this table are colchiploids. Accessions are grouped by species combination. Tetraploid accessions and their parents are listed with Guelph identifiers. Additional information on parents of these tetraploids can be found in Table 7.3. Tetraploid Female parent Male parent accession species accession species accession tl5-2 F. vesca vcvl5 F. vesca vcvl vcv7 x vw2 vcv7 x vw2 7 F. vesca vcv2 F. nubicola nub2 139 vcv6 nubl 138 vcv7 nubl 155 vcvl O nub2 11 vcvl 1 nub2 184 vcv22 nubl 9 vcv23 nub2 12 vw7 nub2 149 vw7 nub2 170 vw9 nub2 120 vwl3 nubl 133 vwl3 nub2 1 O0 vw? 6 nubl 59 vw34 nub2 135 vw39 nubl 108 vw47 nub2 157 vw47 nub2 168 vw47 nub2 175 vw47 nub2 177 vw47 nub2 137 vw53 nubl

199 Table 7.4. Continued. Tetraploid Female parent Male parent accession species accession species accession 411 F. vesca vcv7 x vw2 F.pentaphyl/a pen2 vcvl O vcv20 x vw3 vcv22 x vw34 vcv23 vw22 vw25 wu25 vw25 vw35 vw39 vw47 vw50 pen 1 pen2 penl penl penl pen l pen 1 penl penl pen2 pen 1 pen Ml penl 421 F.vesca vcv20 x vw3 F. vindis vir5 vir5 vir2 vir2 virl virl virl virl 15 vw33 vir 45 F. vesca vcvl3 F. nilgerrensis niil 427 vcv20 x vwl7 n il2 2 vcv23 ni wu15 niil 4 vw47 niil 46 vw47 nill 130 vw47 n il F. viridis virl F-nubicola nub virl nubl

200 Table 7.5. Tetraploid Fragana germplasm derived from hexaploid, tetraploid and diploid species without the use of colchicine. Accessions are grouped by species combination with ploidy levels in brackets. Except where noted, tetraploid accessions and their parents are listed with Guelph identifies. Additional information on parents is in Table 7.3. Tetraploid Female parent Male parent accession s pecies accession species accession 190 F.nubicola(2x) nubl Fmoschata(6x) mcvl 192 nub? mcvl 110 nubl mw2dz 131 nubl mw2d 169 nubl mw2d 182 nubl mw2d 186 nubl mw4 17 nub2 mw2d 47 1 F.moschata(6x) mw5e F.nubicola(2x) nub2 n4 m b F. pentaphylla (2x) pen2 F. rnoschata (6%) bu l k F. rnoschata F.moschata (6x) mcvl mcvl mcvl mcvl mcvl mcv2 mcv2 mcv2 mcv2 mcv2 mw2c mw2c F. vindis (2x) vir2 v1r2 v1r2 v1r2 v1r2 virl vir? vi rl virl v1r2 v1r2 v1r2 77 mw2c vir2 38 F. vnidis (2x) vir 1 Emoschata (6x) mw2d virl vit2 vir5

201 Table 7.5. Continued. Tetraploid Female parent Male parent accession s pecies accession species accession 'oril,2,13,20, F. orientalis (4xj 4947 F.onentalis (4x) 4940 a,u 877 F. orientalis (4x) ori b F. vindis (2x) x virl x mw4 F-moschata (6x F.orientaiis(4x) orib F-moschata mcvl x vir2 (ex) x F. vindis Px) Z* Letters following numbers indicate members of a family. For example, mw2c and mw2d are full sibs. All F.onentalis accessions are full sibs. The parents of these accession (ie and 4940) are accessions numbers of Dr. Gunter Staudt. He provided the author with seeds resulting from 4947 x 4940 crosses.

202 186 Crosses A total of 464 different combinations of parents were crossed. Additional crosses were done for some combinations of parents with 647 pollinations in total. Crosses provided a variable num ber of F. vexa genomes in the resulting hybrids. Tetraploid parents had either zero, MO, or four Evesca genomes and will be referred as O-ves, 2-ves or 4-ves, respectively. The O-ves and 4-ves groups were mostly pollinated with O-ves, although a few were pollinated with 2-ves. The 2-ves plants were pollinated with both O-ves and 2-ves pollen. No groups were pollinated with 4-ves. The 2-ves plants were hermaphrodites and could be used as either males or females in crosses. However, the O-ves group was monecious (except for 110, a hermaphrodite) so individual accessions could be crossed in only one direction. Pollination was randomized each day (see Chapter 2) according to available supply of pollen with the following qualifications: half or full sibs were not intercrossed and new combinations of parents were emphasized to increase the number of crosses and diversity of potential hybrids. An attempt was made to pollinate each female parent with equal numbers from the various combinations of species, but this was not always possible due to different flowering times, different numbers of accessions and limited numbers of Rowers per accessions. Techniques Hybridizations were done from February to April in 7996 and Plants were brought into the greenhouse in two groups, three weeks apart for both 1996

203 187 and General procedures for handling of plants, emasculation, pollination, raising seedlings, colchicine application, and verification of ploidy level are described in Chapter 2. Pollen was collected during the emasculation procedure, categorized according to species combination, dried and stored. Most flowers were pollinated twice in this study with the second pollination occurring 24 hours after the first pollination. A total of 8403 seeds were harvested of which 5846 were germinated in vitro (Miller et al., 1992; see Chapter 2) and the remainder were saved for future experiments. The resulting hybrid seedlings were treated with a 5% colchicine treatment for 16 to18 hours as outlined in Chapter 2. Hybrids with morphological attributes of higher ploidy level than their parents (ie. thicker leaves and stems, wider leaf angles, larger plant size, slower rate of development) were verified with flow cytometry (Chapter 3). Hybrids identified as octoploids were planted (Cam bridge Research Station, Cambridge, Ontario) in close proximity to other octoploids (ie. F x ananassa, Evirginiana, and F.chiloensis) and allowed to set fruit under open pollinated conditions. SOS which set fruit in 1998 were noted for future study. The other SO's produced by this study were not planted in the field until 1998, so fertility data for these hybrids were not available. Data analysis When more than one pollination occurred for the same combination of parents, results were averaged prior to statistical analysis. Fruit set (%), seedsflower, % germination and % SOS (ie. SOs/seedlings treated with colchicine

204 188 x 100) were measured. Means for each combination of parents were used in statistical analysis of the various species combinations. Parents with reciprocal lineage (ie F. vindis x Fmoschata and Fmoschata x F. vindis) were grouped together for data analysis. The influence of female and male parents and their interaction were analyzed with the GLM program of PC SAS (SAS System for Windows V.6.12, Cary, N.C.). Mean separation was accomplished using the LS means function at the 5% level of significance. Fertility of hybrids was not measured due to lack of time in the field although SOS with fruit set were noted. RESULTS Ali hybrids described, for exarnple F.vesca x F.nilgenensis, are tetraploids unless othenivise stated. A total of 192 Sois were created in this study, and F. vesca, F. vindis, F. nubicola, F. pentaphylla, F. nilgerrensis, F. orientalis, and F. moschata were successfully incorporated into 101, 71, 105, 12, 39, 95, and 139 synthetic octoploids, respectively. Of the crosses attempted, 51 O/O produced fruit with an average of 24 seeds per fruit (ie. 12 seeds per fiower pollinated). Examples of tetraploid parents with fruit set are shown in Figure 7.2. Of the seeds that received in vitro treatment, 65% germinated, and 4.1 % of geminated seedlings treated with colchicine were verified as synthetic octoploids. Female and male parents were important sources of variation for % fruit set and seedslflower (see Appendix 5). Male parents were a significant source of variation for germination, but female parents were not significant (P = ). Female and male parents were not

205 Figure 7.2. Examples of tetraploid hybrids and fruit set resulting from interspecific hybridization. Tetraploid plants were vigourous and readily produced fruit for many combinations of parents. From lefi to right: T70 (Fmoschata x Evindis), Tl70 (F. vesca x F. nubicola), and T4 (F. vesca x F. nilgerrensis).

206 190 significantly different in SO production and there was no interaction among females and males for any of the characters studied. The data for individual crosses are presented in Appendix 4 while ANOVAs are presented in Appendix 5. Differences among female parents The O-ves and 4-ves females had 20% more fruit set than 2-ves females (Table 7.6). The only naturally occurring tetraploid in the study, F. orientalis, had the highest fruit set at 86%. The 2-ves type derived from self-compatible (SC) species (ie. F.vesca x Ç.ni/gerrensis) had 70% fruit set which was within 2% of F.vesca x F. vesca, and F.moschata x F.vMis female parents. Among the O-ves females, F.moschata x F.nubicola had the lowest fruit set (41%) which was very similar to those 2-ves types derived from F-vesca and self-incompatible (SI) species. These latter mentioned 2-ves types were F. vesca x F. vindis, F. vesca x F.pentaphylla, and F.vesca x Fmbicola with 37%. 43%, and 41% fruit set, respectively. For seedsmower O-ves and Cves produced 3.5 times more seedslflower than 2-ves. The hig hest seedslflower occurred for F. onentalis, F moschata x F. vindis and F.vesca x F-vesca with 38, 30 and 24 seedslflower, respectively. Again, Emoschata x F.nubicola females had the lowest numbers (six seedslflower) of the O-ves types, which was similar to 2-ves fernales. Means were closely grouped among 2-ves females ranging from F-vesca x F.pen@hy/ia with four seedsmower to F. vesca x F.nilgerrensis with 10 seedslflower. Although in vitm germination of embryos for O-ves females ranged from 82

207 5 3 O E cn m Q) 2 $ 'C $9 ; u C.C 2.'F + c s 8 c c O u. 8-c cal c n 2 & g9 Q, c ca C E E r z ou u g: E L g.r Q) ' cn:

208 192 to 88% this was not statistically different from Zves (range 53 to 65%) or 4-ves (53%) fernales. There were no significant differences among fernales for % SOS. Fragana vexa x F-pentaphylla, F.vesca x Enilgerrensis and F.orienta/is had the hig hest proportions with 6.2, 5.5, and 5.4% SOS produced, respectively. 6 Differences among male parents Overall, O-ves males were superior to 2-ves males in every character measured (Table 7.7). The O-ves males had 23% more fruit set, five times more seedslflower, 17% higher genination and almost four times more % SOS than 2- ves males. Among O-ves types, Fmoschata x F.nubico/a, F.moschata x F. vindis, F. pentaphylia x F.moschata, and Eorientalis were the best males if al1 four characters are considered. These aforementioned O-ves types had 53% to 73% fruit set surpassed only by Evesca x F.vindis males with 82% fruit set. These four species combinations also had the highest means for seeddfiower and were among the best for genination. For percent SOS, Emoschata x Enubicola and Eonentalis x (F-maschata x F. vindis) were exceptional with 11.7% and 8.8%, respectively, followed by F. vesca x Fmbicola, Epentaphylla x F.moschata, F.orienta/is, and Emoschata x F. vindis with 3.7%, 3.1 %, 2.3% and 2.0%, respectively. Other species combinations used as males had O to 1.1 % SOS. Despite the fecundity of the O-ves males already mentioned, some O-ves males were also the poorest parents used. The O-ves males derived from two diploid species, F.vindis x F.nubico/a and the male derived from three species, F. onentalis x (F. moschata x F. vindis)), had the

209 Table 7.7. Effect of number of Fragana vexa genomes and male parentage on % fruit set, seeds/flower, % germination, and % synthetic octoploids (SOS) for crosses arnong tetraploid Fragana hybrids Evesca genomes Male lineage NZ Fruit set Seeds/flower N Germination N SOS (%) (%) ('w 2 F. vesca x F. viridis aby 3 b-f a-c ns F. vesca x F. nubicola bd-f 6 b-f b-e IO 3.7ns F. vexa x F. pentaphylla 56 Il e-f 0.5 d-f a-c 9 1.1ns F. vesca x F. nilgerrensts bd-g 4 bd-f 7 14 c-e 3 O ns F. viridis x F. nubicola 14 7e-g 0.2 bd-f 1 33 a-e 1 O ns F. moschata x F. viridis ' ab 17 a-d a-c ns F. nubicola x F. moschata " bc 18 a-c a-c ns Fpntaphylla x F. moschata a-ce 14 a-e 3 58 a-d ns F. orientalis x (F. moschata x F. viridis) d-g 3 d-f 4 50 a-e ns F. orientalis ab 19 a-c a-c s mean N is the number of unique combinations of parents that were crossed, N was the same for % fruit set and seedslflower but differed for other variables. Letters following means denote Ismeans groupings at 5%. lncludes reciprocal combination

210 194 lowest values for each character when compared to other O-ves. In most cases, there were no significant differences within the group of 2-ves males. Each combination of species tended to have the highest value for at least one of the characters measured. Fragana vesca x F.viridis had 44% more fruit set. F. vesca x F.nubicola had 50% more seeds/flower, F. vesca x F.pentaphylla had 39% higher germination and F-vesca x Enubicola had Z.60h more SOS than the next best 2-ves males. Fragaria vesca x F.ni/gerensis had the poorest overall record with only 14% germination and no SO production. Interaction among males and females In this study, 52 of the 66 (79%) combinations of species were succesçful in producing viable tetraploid progeny, and 25 of these combinations resulted in SOS (Table 7.8). The interaction among males and fernales was not significant (Appendix 5) for any of the characters in the study. Thus, the greatest values for each trait were obtained when the best females were pollinated with the best males. For example, F.onentalis accessions were among the best of the female parents (Table 7.6) while F.moschata x F.viridis and Enubicoia x Fmoschata accessions were good male parents (Table 7.7) When Ç.orientalis was hybridized with these two groups of males, % fruit set, seedsmower, % germination, and % SOS were much higher than the average for the experiment (See Table 7.8).

211 Table 7.8. Fruit set, seedsl flower, germination and synthetic octoploids (SOS) resulting from interspecific crosses arnong tetraploid Fragana hybrids with different numbers of F.vesca genornes. Y Female parent Male parent Seeds No, of Species No. of Species No. of NZ Fruit set Seeds receiving N Germ. N SOS (%) F. vesca F. vesca flowers (Oh) lflower invitro (Oh) genomes genornes pollina ted treatment 4 F. vesca 2 F. vesca x F. viridis O F. vesca x F. nubicola F. vesca x F. pentaphylla 6 4 O 0.0 O O nla nla F. vesca x F. nilgerrensis 2 2 O O. 0 O O nla nla Total N & rneans IO F. vesca O f. moschata x F. viridis Y F.nubic01a x F.moschata Y F.pentaphylla x Fmoschata F. orientalis x (F. rnoschata x F. viridis) F. orientalis Total N & means : N = number of combinations of accessions used in crosses. N is the same for fruit set and seedslflower but differs for other variables. lncludes reciprocal combination

212 Table 7.8. Continued. 196 Female parent Male parent Seeds No. of Species No. of Species No. of N Fruitset Seeds receiving N Gerrn. N SOS (%) F. vesca F. vesca flowers (%) / flower in vitro (%),cienornes qenomes poliinated treatment 2 F. vesca x F. viridis 2 F. vesce x F. viridis F. vesca x F. nubicola 13 IO F. vesca x F. pentaphylla F. vesca x F. nubicola F. vesca x F. viridis F. vesca x F. nubicola F. vesca x F. pentaphy//a O F. vesca x F. nilgerrensis F, vesca x F. pentaphyila F. vesca x F. viridis F. vesca x F. nubicola F. vesca x F. pentaphylla F. vesca x F. nilgerrensis n/a F. vesca x F. nilgenensis F. vesca x F. virjdis II F. vexa x F. nubicola F. vesca x F. pentaphylla F. vesca x F. nilqerrensis total N & means

213 Table 7.8. Continued Female parent Male parent Seeds No. of Species No. of Species No. of N Fruitset Seeds receiving N Germ. N SOS (%) F. vesca F. vesca flowers (%) l flower in vitro ("10 genornes qenomes pollinated treatment 2 F. vesca x F. viridis O F. viridis x F. nubicola Y O nla nla F. moschata x F. viridis F. nubicola x F. moschata Y F. pentaphylla x F.moschata F. orientalis x nla (F. moschata x F. viridis) F. orientalis F. vesca x F. nubicola F. viridis x F. nubicola 2 1 O 0.0 O n/a nla F. moschata x F. viridis Y F. nubicola x F. moschata y F.pentaphylla x Emoschata O nla nla F, orientalis x 6 6 O 0.0 O nia nla (F. moschata x F. viridis) F. orientalis F. vesca x F. pentaphylla F. viridis x F. nubicola Y F. moschata x F. v~ridis y F. nubicola x F. moschata F-pentaphylla x F. moschata O nla nla F. orientalis x 2 2 O 0.0 O nla nla (F. moschata x F. viridis) F. vesca x F. nilgerrensis F. orientalis F. viridis x F. nubicola 4 3 O O. O O nla nia F. moschata x F. viridis Y Y F.nubicola x F. moschata F. orientalis x nia (F. moschata x F. viridis) F. orientalis total N & means

214 J?$2 C C C 322 CCC

215 Table 7.8. Continued Fernale parent Male parent Seeds No. of Species No. of Species No. of N Fruitset Seeds receiving N Germ. N SOS (Oh) F. vesca F. vesca flowers (%) 1 fiower in vitro (NI genomes clenornes potlinated treatment O F. moschata x F. viridis O F. moschata x F. viridis Y F.nubicola x F. moschata Y Epentaphylla x F. moschata F. orientalis x (F. moschata x F. viridis) F. orientalis nla F. nubicola x F. moschata F. viridis x F. nubicola F. rnoschata x F. viridis Y F. nubicola x F. moschata Y F.pentaphylla x F. moschata F. orientalis x (F. moschata x F. viridis) F. orientalis O nla O nla O nla nla n/a nla 7.3 F. orientalis F. viridis x F, nubicola nla F. moschaia x F. viridis Y 4.5 F.nubicola x F.moschata Y total N & means summary for experiment

216 200 Influence of number of Evesca genomes The 4-ves x O-ves crosses resulted in the highest fruit set (92%) and seedsiflower (32). When 4-ves females were pollinated with 2-ves males, fruit set was 20% lower than with O-ves males and only two seedsiflower were produced. A similar decrease was noted with 2-ves females. When pollinated with O-ves, fruit set and seedsmower were 52% and 10 respectively, but when pollinated with 2-ves the result was 37% and 3, respectively. O-ves x O-ves had similar numbers of seeds/flower (27) to 4-ves x O-ves crosses, although fruit set was only 69%. The O- ves x 2-ves and 2-ves x O-ves crosses produced similar results with 44% fruit set and 10 seeds per flower. Too few seeds were obtained from 4-ves x 2-ves and O-ves x 2-ves crosses to make germination numbers rneaningful for these two cornbinations. The hybrid combinations with a total of four F.vesca genomes (ie. 4-ves x O-ves or 2-ves x 2- ves) had 53% germination. Crosses involving zero Evesca genomes had a much higher rate of germination (87%) while crosses involving two F. vesca genomes (ie. 2-ves x O-ves) were intermediate with 65% germination. The highest and lowest % SOS were produced with Cves x Zves and O-ves x 2-ves crosses, 10% and 0% SOS, respectively. The data of these two groups were based on in vitro treatment of only 40 and 27 seeds, respectively. The next hig hest SO production occurred with 2-ves x O-ves with 6.8% followed by O-ves x O- ves with 4.1 % which were based on a much larger arnount of seeds, 1733 and 2442 respectively. The 4-ves x O-ves crosses and 2-ves x 2-ves crosses were similar with 1.3% and 0.8% SOS, respectively.

217 201 Within many combinations of species there were differences in seedsfiower for certain combinations of accessions. For individual crosses, only 215 of the 464 combinations of accessions (ie. 46%) resulted in seed production (Appendix 4). An extreme exam p le of varia bility for d ifferen t accessions was (F. vexa x F. nilgerrensis) x (F.nubico/a x Emoschata) crosses, which had a range of O to 131 seedsmower for different combinations of accessions (Appendix 4). Most of the 464 combinations of accessions were based on only one pollinated flower. However, number of flowers did not appear to have a major confounding effect. When 139 accession combinations involving two or more pollinated flowers (Appendix 4) were examined, adding more flowers did not affect fruit set in 107 (ie. 77%) crosses. Lineage and fertility of SOS The most productive combinations of accessions for SO production are shown in Table 7.9. These 14 crosses represent only 3% of the combinations of accessions crossed, yet they resulted in 128 of the 198 (ie. 67%) SOS produced. The most common species in the lineage of superior fernales were, F.vesca, F.orientalis, F.vindis, and F.nubico/a, which were present in 72, 39, 31 and 24 SOS, respectively. In the male lineage of the superior combinations, F. moschata was the most common, being present in 105 SOS. Fragaria nubicola was present in male lineage of 91 SOS in the combination F.nubicola x Emoschata. None of the 2-ves male parents were present in this group of productive crosses. Twenty-five combinations of species are represented in the 192 synthetic

218 Table 7.9. Details of the 14 most productive crosses among tetraploid Fragaria which resulted in 128 synthetic octoploids (SOS). Crosses are grouped by the nurnber of F.vesca genomes and species combinations of the parents. Each cross is shown with numbers of flowers pollinated (N), seeds/flower, percent germination in vitro, percent SOS and total SOS produced. Female Parent Male Parent N Seeds Germ.' SOS Total F. vexa Species corn bination Acc. Evesca Species com bination Acc. /flower (%) (%) SOS genomes Id? genomes Id F. vesca x F. vesca 4181 O F. nubicola x F. rnoschata F. vesca x F. viridis 106 O F. nubicola x F. moschata F. pentaphylla x F. moschata n4mb F. vesca x F. nubicola 177 F. nubicola x F. moschata F. orientalis ori F. vesca x F. ntlgerrensis F. nubicola x F. moschata i F. orientalis ori O F. moschata x F. viridis 70 O F. rnoschata x F, viridis F. nubicola x F. moschata F. orien talis oril 3 F. viridis x F. moschata oril 3 F. nubicola x F. moschata ori ori Acc. Id. = Accession identification, see Tables 7.4 and 7.5 for more details on individual accessions. Y Gerrn. = Germination

219 203 octoploids produced in this study (Table 7.10) with eight different combinations of ploidy levels (Figure 7.3). The resulting genepool of SOS created includes al1 accessions of F.moschata, F. nilgerrensis, F. nubicola, F. orientalis, F.pentaphylla, and F-vin'dis and IO of the F.vesca accessions listed in Table 7.3. The successfully incorporated F. vesca accessions encompassed three subspecies, ssp. vesca var. semperflorens (vcv 20,22.23,24), ssp. bracieata (vw 3), and ssp. amencana (vw17,34,47,50,61),and an accession from Ecuador (~15). Each of the various types of tetraploids, except F. viridis x F.nubicola, were incorporated into SO's. Only four SO's were created with 2-ves pollen while 1 90 SO 's were created with O-ves pollen. The female lineage of 106 SOS contained F.vesca in the tetraploid form or in combination with F.nilgerrensis, F.nubicola, Epentaphylla, or F. viiîdis. Of the 97 SOS planted in 1997, 28 fruited in Fruits from two haivests of open pollinated SOS are shown in Figure 7.4. All species in the study except Epentaphylla were incorporated into SOS capable of setting fruit. There are however 11 SO's with F.peniaphylla lineage which were not planted until Fertility data is considered preliminary since most of the SOS were not planted until 1998, and fruit production is an indication of female fertility and not male fertiliity. Full study of the fertility of these SOS was not undertaken as part of this study due to tirne constraints. DISCUSSION A need to expand the narrow germplasm base of octoploid strawberry

220 Table Combinations of synthetic and natural Fragaia tetraploids used to produce synthetic octoploids (SOS) Female parent Male parent No. SOS No. with No. of No. of of fruit in F. vesca Species combination F. vesca Species combination SOS 1998 ~enomes qenomes 4 F.vesca(4x) 2 (F. vesca x F.nubicola) Total 1 O 4 F. vesca (4x) O (F. nubicola x F. moschata) Total (F. vesca x F. nilgerrensis) 2 (F. vesca x F.nubicola) 1 (F. vesca x F. nilgenensis) (F. vesca x F. vinais) 1 1 (F. vesca x F. nubicola) (F. vesca x F.nubico/a) 1 Total (F. vesca x F. nilgenensis) (F. vesca x F.nilgenensis) (F. vesca x F. ndgenensis) (F. vesca x F.nilgemnsis) (F. vesca x F.nubicola) (F. vesca x F. nubicola) (F. vesca x Fmbicola) (F. vesca x F.pentaphyila) (F. vesca x F-pentaphylla) (F. vesca x F. vindis) (F. vesca x F. vindis) O (F.moschata x F. vindis) 4 (F. viridis x F. moschata) 1 (F. nubicola x F. moschata) 13 2 F. onenfalis 19 (F. nubicola x Fmoschata) 18 (F. vindis x Emoschata) 2 2 F. orientalis S (F. nubicola x F. rnoschata) 3 F. orientalis 1 (F.nubicola x Fmoschata) 1 O 1 (F.pentaphylla x F. moschata) 4 Total 84 5 O (F. moschata x F. viridis) O (F. moschata x F. vinüis) (Fmoscha ta x F. viridis) (F. moschata x F. vin'ds) (F. moschata x F. viridis) (Enubicola x F.moschata) F. orientalis F. orientalis F- orientalis (F. moschata x F. viridis) (F. viridis x F. moschata) (Fmbicola x Fmoschata) F. orientalis F. orien talis x (F. moschata x F. viridis) F. orientalis (F. moschata x F. viridis) (F. viridis x F. moschata) (F. nubicola x F. moschata) Total Overall Total

221 Figure 7.3. Crossing plans using Fragana species that were successful in creating synthetic octoploids in this study. Ploidy levels of parents and colchicine application for doubling of chromosomes are shown. Numbers In parenthesis indicate synthetic octoploids produced. zx 4x / colchicine colchicine 2x 4x Bx (29) colchicine (3) 25.- (3) 2x colchicine - 4x colchicine 2x 4x / colchicine \ I colchicine

222

223 207 cultivars (Sjulin and Dale, 1987; Daubney, 1990; Galletta and Maas, 1990; Hancock et al , 1996; Shaw, 1991 ; Sullivan, 1991 ; Dale et al., 1993; Harrison et al., 1993; McNichol and Graham. 1992; Sangiacorno and Sullivan, 1994; Davidson, 1995) prompted this investigation into the creation of SOS. Previous reviews of strawberry breeding (Darrow, 1966; Scott and Lawrence, 1975; Hancock et al., 1990, 1996) have indicated limited investigation has occurred on interspecific hybridization in Fragaria. The 192 SOS produced in this study represent a much larger number of interspecific hybrids than al1 previous attempts combined (Table 7.1). While this and Evans' (1977) studies produced similar nurnbers of tetraploid seeds, 5846 vs respectively. this study produced 25 times more synthetic octoploids from a wide genetic base. After five years of observations only three male fertile SOS were identified by Evans (1982a,b,c) but after only one season in the field 28 SOS were identified as female fertile in this study. Furthermore, it is likely that additional SOS will be identified as fertile when the second planting of SOS has a full season in the field and when male fertility is investigated. Previously, SOS have been created from just seven combinations of species, with only two combinations producing fertile plants (Table 7.2) but this study created SOS from 25 combinations of species, with 11 combinations producing plants identified, to date, as fertile. As in Chapters 4, 5 and 6, this study used a very wide germplasm base. Thus, the many SOS created in this study have the potential to greatly expand the genetic base of Fragana available to breeders. This study demonstrated that F. vesca could be used to introgress the diploid species F. nilgerrensis, F. nubicola, F-pentaphylla, and F. vindis into synthetic

224 208 octoploids. Although F. vexa has been used to create synthetic octoploids (Sebastiampillia and Jones, 1976; Evans, 1977) and has much potential as a bridge species (Chapters 1 and 4) previous research had not used the concept of a bridge species or F. vexa as that bridge species. This study investigated alternative ways for combining Evesca genomes when creating SOS. When considering the results, it is important to focus upon crosses having 2-ves parents as these have excellent potential to transfer genes from other diploid species. The optimum combinations involving 2-ves parents was 2-ves x the best O-ves males (either F.odentalis or Emoschata x diploids), which gave the highest values for % fruit set, seedslflower, % gemination and % SO (Table 7.8). The O-ves x 2-ves crosses may also have potential as they averaged the same number of seeddflower but not enough seeds were germinated to detenine germination rate and SO production. There was a dosage effect whereby increasing the number of genomes derived from F.vesca and diploid species decreased fruit set and seedslflower. For example, O-ves x O-ves, O-ves x 2-ves, 2-ves x O-ves, 2-ves x 2-ves, and Cves x 2- ves had 69%. 52%, 44%, 37%, and 20% fruit set and 27, 10,10,3,2, seedsflower, respectively. This effect could be due to decreased heterozygosity (ie. hybrid vigour) in the gametes from 2-ves and 4-ves plants. The 4-ves x 2-ves combination would have had less heterozygosity than other combinations and it had lower numbers than other combinations. It had been suggested that pivota1 genome crosses (ie. 2-ves x 2-ves) would be more productive than other crosses (Chapter 1, Zohary and Feldman, 1961; Zohary, l965), but this does not seem to be universal for the Fragaria tetraploids used in this study. Fragana genomes may be so similar (Jones,

225 ; lwatsubo and Naruhashi, 1989, 1991) that a strategy using pivotal genomes might not be beneficial for tetraploid breeding. In Fragaia, the exact genomic constitution of polyploid species is unknown. As polyploid species were derived from diploid species, some tetraploid parents may have shared common or pivota1 genomes. Only 2 of the 31 combinations of tetraploids reported by previous researchers (listed in Tables 7.1 and 7.2) had common genomes [ie. (F.vesca x Eviridis) x (F.nilge~ensis x F.vindis) and the reciprocal] which had similar seed production(2.9 and 1.8 seeds/flower) to the 2-ves x 2-ves crosses in this study. Also, since the 2-ves hybrids were colchiploids, there may be a need to maximize heterozygosity for interspecific crosses (Sanford, 1983). In Aegilops and Tnticum, where the species have distinctly different genomes, it is very important for interspecific crosses to have parents that share at least one common genome (Feldman and Sears, 1981). While it is likely that similar genomes are important for producing interspecific hybrids in Fragaria, it may be that there are not many different genornes in the genus. The natural tetraploid, Fragana onentalis, and (6x x 2x) hybrids (F. moschata x diploids) were very important in this study. One of these was present in 98% of the SOS produced (Figure 7.3). In previous studies (Tables 7.1 and 7.2), F. orientalis and (F. moschata x diploids) have been successful in tetraploid crosses. However, this study was the first to incorporate F. orientalis and F.pentaphylla x Fmoschata into SOS (Table 7.10). These types of O-ves parents may have been superior to 2-ves parents since the latter were colchiploids. Colchiploids are highly homozygous (EingstJ959) and are known to be subject to fertility problems especially in fruit

226 21 O species (Sanford, 1983). As both F.orientalis and Fmoschata are dioecious and obligate outcrossers (Staudt, 1989) they would have been more heterozygous than 2-ves parents, and would have produced genetically variable gametes. The garnetes from 2-ves and Cves colchiploids would have had much less genetic variability, and variability is very important for interspecific hybridization (Burbank, 1929; Galetta and Maas,1990). Fragana orientalis and Fmoschata rnay be most genetically fit for high ploidy levels. It was noted that 6x and 8x Fragana tend to have less DNNgenorne than 2x Fragana (Chapter 3). Perhaps in polyploids, multiple copies of some genes are detrimental, and have been selected against during evolution. Thus, it may beneficial to hybridize colchi-tetraploids derived from diploid species with tetraploids derived from higher ploidy species to mask deleterious genes and maximize heterozygosity. Fragana moschata (6x) was identified as a bridge for three diploid species (ie. Fmbicola, F.pentaphylla, and F.vin'dis) to be introgressed into SOS. However, recombination among the three Emoschata genomes and the one diploid genome in the tetraploid gametes likely provided only a few chromosomes from the diploid species into SOS. The strategy of using F.vesca to cross with other diploids has the advantage of delivering an intact genome of the other diploid into an SO. An alternative use of F.moschata would be to treat F 1 (F.moschata x diploid) hybrids (4x) with colchicine (Evans, 1977) to produce an octoploid and bypass tetraploid crossing. This method may be very difficult because of the low success rate of crossing Emoschata with diploids compounded with low chromosome doubling rates. However, the strategy couid be improved by simply using the best combinations of

227 21 1 accessions (Chapter 6) in the crossing program. At the tetraploid level, most combinations of species can intercross, particularly if severai combinations of accessions are tested. Only 14 combinations of species did not produce seeds in this study but nine of these involved only one combination of accessions for parents. The value of having many accessions has been a recurring theme at other ploidy levels (Chapters 4, 5, and 6). In previous studies, 88% of 25 combinations of tetraploid hybrids resulted in progeny (Tables 7.1 and 7.2). This study was similar to previous studies when 79% of 66 combinations of species produced seeds and 90% of the combinations germinated in vitro resulted in viable offspring (Table 7.8). This was much higher than crosses at the diploid level in which 55% of 36 combinations of species resulted in progeny (Chapter 5). With so many combinations of tetraploids possible, there is excellent potential to breed at the tetraploid level prior to making synthetic octoploids. Although, the two accessions of Eodentalis x (Fmoschata x F. vindis) averaged only three seedsflower (Table 73, they demonstrated that tetraploids derived from three species could be at least partially fertile. The successful intercrossing of the tetraploids verifies that self-incompatibility mechanisms begin to break down at the tetraploid level in Fragaria because of tetrasomic segregation of self-incompatibility alleles (Evans and Jones,1967). A partial breakdown could explain why the colchiploid fernales derived from self-compatible species (ie. F. vesca x F. vexa and F. vesca x F.niigerrensis) had the highest fruit set and seedslflower (Table 7.6). These two types would have presented no incompatibility barriers while the other 2-ves types would have reduced crossability. Conversely, self-incompatibility alleles could have been responsible for

228 212 reduced crossability when the F.vesca x F.ni/genensis hybrids were used as males (Table 7.7). The current study compliments cytogenetic evidence (Staudt, 1959; Jones, 1976; lwatsubo and Naruhashi, 1989, 1991) and DNA analysis (Harrison et a1.j 993, 1997) that suggests Fragaria species are closely related. Since many combinations of species were used to create fertile SOS, it is diffîcult to suggest which species might have been involved in the evolution of natural occurring octoploid species. The results of this study tend to expand rather than limit the number of possibilities for progenitors of the polyploid species. The two polyploid species used in this study, F.orienta/is and (Fmoschata x diploid) accessions sften had similar results (Table7.8, Appendix 4), which supports the theory that Emoschata and Eonentalis may have been derived from the same diploid species (Federova, 1946; Staudt, 1959, 1989). The F.vesca x F. vindis and F-vesca x F-nubicola tetraploids bore a strong resemblance to F.moschata while the F.vindis x F.nubicola hybrids most closely resem bled F. orientalis. As the geograp hical location of F-moschata overlaps with Evesca and F.viridis but not with F.nubicola (Figures 1.3 and 1.4) it is more likely that F.vesca and F.vinüis fomed F.moschata. Although the main habitat of F.odentalis lies to the east of F.vindis and Enubicola, they cannot be ruled out as possible progenitors of F.odentalis since the exact locations of Fragana species in Asia requires further research (Staudt, 1989). Difficuities encountered in creating tetraploid hybrids from diploid or hexaploid species may not be present when these hybrids are used as parents. For example, Evesca (2x) x F.nilgemnsis (2x) and F-moschata (6x) x diploid hybnds were difficult

229 213 to obtain because of poor seed production of the former (Chapter 4) and poor germination of the latter (Chapter 6). At the diploid level, only 1.2 seeds/flower were obtained for F.vesca x Enilgerensis (2x). while F.vesca pollinated with either F.nubicola, Epentaphylla, or F. viridis averaged at least 10 times more seeds/flower (Table 4.4, Chapter 4). At the tetraploid level, Evesca x F.nilgenensis (4x) females had the highest numbers of 2-ves females with 10 seedsmower (Tables 7.6). This, and the occurrence of fertile F. vexa x Enilgenensis diploids (Chapter 5) indicates F. vesca and F.nilgerrensis are genetically closer than previous research suggested. When Emoschata x diploid (4x) hybrids were intercrossed germination rate was 79% (Table 7.8). This is considerably higher than the 6% germination from seeds obtained by intercrossing F-moschata (6x) and diploid species (Chapter 6). It was suggested that Fmoschata could harbour lethal alleles that surfaced when crossed to diploids (Chapter 6). If so, such alleles may not be present in viable F.moschata x diploid hybrids (ie. parents used in this study) which could explain the higher germination rates and the lack of albinos or stunted seedlings at the tetraploid level. Endosperm balance may also be an important factor (Evans,? 977) in the improved germination of tetraploid crosses compared to hexaploid x diploid crosses. Data analysis indicated that males and females were important sources of variation but that the interaction among males and females was not significant (Appendix 5). This would be similar to stating that general combining ability is much more important than specific cornbining ability for interspecific hybridization of Fragaria tetraploids. It wouid, therefore, be more important to identify good male and females parents for crossing rather than attempt to intercross many inferior

230 214 tetraploids in the hopes of finding a good combination. As F. orientalis performed well as a male and female, it might be valuable for use in test crosses to identify other superior tetraploids. While this study shows that some combinations of species are more successful as tetraploid parents, it may also be true that certain accessions are superior parents. Among the 14 superior crosses identified in this study (Table 7.9), 50% of the parents appear more than once. Tetraploid '1 10' was particularly noticeable as it appeared in the lineage of three of the best crosses. Further investigation revealed that nine of the 14 best crosses resulted in 95 SOS with the Emoschata accession 'mw2dp in their lineage via the tetraploids 18, 110, 169, and 182 (Tables 7.8 and 7.4). The 2-ves accessions 106, 151, and 171 appeared twice in Table 7.8, but each of these originated from a different source of F. vesca. With these superior crosses identified, it may be possible to do similar crosses to introduce new gerrnplasm into the SO collection. For example, mw2d could be crossed with a new F.nubicola accession and the resulting tetraploid could be used to pollinate ori20. Again. including a large variety of accessions in this type of research is very important to its success or failure (Burbank, 1921, Galletta, 1990). Cornparison of % fruit set, seedslflower, % germination and Oh SOS data indicated that chromosome doubling was the most limiting step for synthetic octoploid production in this study. The in vitro colchicine method of this study created 4.1% SOS per treated seedling, which was higher than the 0.1 % SOS obtained with the immersion method (Evans, 1977) but less than the 7% SOS obtained by the dropper method (Sebastiampillia and Jones,l976). However, al1 but two of the most productive crosses (Table 7.9) had higher chromosome doubling rates than the

231 215 Sebastiampillia and Jones study (1976). The evidence suggesting F. vesca was more amenable to chromosome doubling than other Fragana species (Fadeeva, 1966; Sebastiampillia and Jones, 1976) was one of the considerations for emphasizing F.vesca in this SO system. As males and fernales did not significantly influence percent SOS (Appendix 5) it would seern that F.vesca genomes had little bearing on chromosome doubling at the tetraploid level. Too few O-ves x 2-ves seedlings were produced to make a meaningful cornparison with 2-ves x O-ves seedlings or to draw conclusions regarding Evesca cytoplasm. It may be that other factors such as heterozygosity were more important factors influencing the survival of colchicine treated plants. The success of colchicine treatment that could be improved by treating several runners from mature plants so chromosome doubling could be repeated many times for the same genotype. SUMMARY The 192 SOS created in this study represent a tremendous improvement over previous techniques and considerable potential to broaden the gemplasm base of the cultivated strawberry. While tetraploids containing two genomes of F. vesca did not produce as many hybrids as those tetraploids without any F-vesca genomes, this does not diminish the usefulness of using F.vesca to introgress other diploid species into SOS. This study showed that tetraploid hybrids resulting from crosses among Evesca and either F.nilgerrensis, F.nubicola, F.pentaphy/la, or F. vindis are fertile, and that such hybrids could be used to create SOS through additional crosses with other tetraploids. Rather than intercrossing 2-ves tetraploids, a better strategy would

232 216 be to cross Z-ves tetraploids with tetraploids derived from Fmoschata x diploids of F.orientalis. Heterozygosity, presence of lethal alleles, and genetic fitness of hybrids derived from polyploid species are important factors involved in crossability of tetraploid hybrids. The fact that so many combinations of species were capable of producing progeny at the tetraploid level suggests that the species involved in this study are closely related with similar genomes and that there may be great potential for cornbining many species and breeding at the tetraploid level in Fragada. While many combinations of accessions can intercross, it is important to use a wide gemplasm base to obtain accessions with the highest potential success rates. Screening tetraploids by crossing to highly fertile accessions is recommended to identify those accessions that are superior male or fernale parents.

233 217 Chapter 8 General discussion & summary The need for a wider genetic base for breeding of octoploid cultivars of strawberries (Sjulin and Dale, 1987) prompted this investigation of introgression of wild Fragada species using the synthetic octoploid system. By hybridizing 2x, 4x and 6x species and using colchicine to create synthetic octoploids, interspecific barriers due to differences of ploidy level were overcome and introgression into octoploid cultivars was facilitated (Evans, 1976). There have only been a few reports regarding the creation of synthetic octoploids (Evans, 1976; Sebastiampillia and Jones, 1976; Jones 1977) and only two fertile SOS were released (Evans, 1982ab). In previous reports, crossing strategies have been outlined to combine species of various ploidy levels to produce octoploids (Darrow, 1966: Evans. 1976) but actual investigation of a planned crossing strategy had not been done. With 10 diploid, three tetraploid and one hexaploid Fragana species identified (Staudt, 1989; Hummer, 1996), there are over combinations of species that could be investigated for creation of SOS. Rather than attempt ail possible combinations of species, this investigation focused on a strategy of using F.vesca as a bridge or pivotal genome to bring other diploid species to the tetraploid level. This was followed by intercrossing of tetraploids including Eonentalis (4x) and F.moschata (6x) x diploid species. Thus, diploid, tetraploid and hexaploid species were incorporated into the crossing strategy used in this study. The crossing strategy also increased heterozygosity by including as many different com binations of

234 218 parents as possible and by avoiding crosses between full or half sibs. This aspect of the crossing strategy was especially recommended for interspecific crosses where it is important to have as wide a germplasrn base as possible (Burbank ; Galletta, 1990). The decision to ernphasize F.vesca as a bridge to diploid species in this crossing strategy was based on several factors. Fragaria vexa was alleged to be the progenitor species of the genus of Fragana (Staudt, 1953; Nurnberg-Kruger; Darrow, 1966) or at least the progenitor of polyploid species (Darrow, 1966; Ellis, 1958; Staudt, 1984; Bringhurst and Senanayake, 1967; Bringhurst, 1990) and has been hybridized with many other species (Tables 1.1, 1.2, 1.3, 1.4). Also F.vesca is self-compatible (Staudt, 1989), has a desirable fiavour and aroma (Darrow, 1966), is adapted over a wide geographical area (Reed, 1966; Staudt, 1989) and many accessions were available from geneban ks. In addition to crossing strategies. this study adapted advanced techniques and obtained germplasm which had not been available during earlier research into SO production (Evans, 1976; Sebastiampillia and Jones, 1976). These techniques included flow cytometry for ploidy verification, in vitro germination. and in vitm colchicine application. Genebanks that were not in existence in the 1970's were used in this study. providing higher number of accessions than in previous studies. The results of interspecific crosses between diploid species in this study cast doubt on F.vesca as the progenitor of its genus. Fragana vesca subspecies from Europe and North America did not significantiy differ in crossability with

235 219 F.nilgerrensis, F. nubicola, F.pentaphyl/a or F. viridis ind icating that F. vesca may be younger than previously supposed. Fragana vesca and F-pentaphylla were found to be highly crossable and each able to fonn fertile hybrids with five other diploid species, but F. pentaphylla had a better crossing record than Evesca. Fragana pentaphylla formed non-blooming hybrids with two other species (F.daltoniana and F. nilgemensis) while F. vesca formed let ha1 hybrids with three other species (F.iinumae, Egracilis, and /=.nipponka). The superior crossability of F.pentaphy/la rnay indicate it is the progenitor of the genus of Fragana. As self-incompatible species were more readily intercrossed than self-compatible species and are located in the centre of Fragaria's Asian centre of diversity, it was theorized that self-incompatible species are ancestral to self-compatible species in Fragana. Selfcompatible species have often arisen frorn self-incompatible species when isolation restricted opportunities for cross pollination (Stebbins, 1950; Lewis and Crowe, 1958). Thus, F.pentaphyl/ats self-incompatibility, crossability with other species, and central location in Fragana's centre of diversity in central Asia, further support the theory that it is the progenitor of the Fragana genus, and not the self-compatible F. vesca. Previously, only five combinations of diploid species were reported as fertile but this study increased the number of known fertile combinations to 13. All diploid species, except Eiinumae, formed fertile hybrids with at least one other diploid species indicating favourable potential of interspecific breeding at the diploid level. Hybridization between diploid species could be further improved if more accessions

236 220 of some species were available. The results of diploid crosses indicated that F.vesca is valuable for the SO system, particularly as a bridge for self-compatible species F-nilgenensis and Fdaltoniana. The availability of many accessions of Evesca was an important factor in obtaining hybrids with these two species. Fragaria vesca probably is not the progenitor of the genus Fragana, but it could be a progenitor of polyploid Fragada. With its widespread range (Staudt, 1989) there would have been opportunity for Evesca to hybridize with other species and form polyploid species. In this study it was noted that F.vesca x F-viridis and Evesca x F.nubicola tetraploids resemble Fmoschata (6x). The successful integration of F.vesca into many SOS in this study suggests that F.vesca has potential to form polyploids in nature, To incorporate F.moschata (6x) into the SO system it was hybridized with the diploid species F.nubicola and F.vindis to create tetraploids. Fruit set (90%) and seeds/pollination (19) were quite good for these combinations, but embryo abortions and malfoned seedlings reduced the outcome to only 1.1 plants/pollination. The best combinations for viable plants were found to be Fmbicola x Emoschata followed by Emoschata x F-vindis. This was contrary to earlier investigations that did not obtain any F.nubicola x F.moschata hybrids (Evans, 1964, 1974). These results contradict the idea that the higher ploidy species should be used as the female parent (Evans, 1974). As most Emoschata accessions are dioecious (Staudt, 1989), they cannot be used as both male and females in crosses so it was

237 221 impossible to do reciprocal crosses. It was suggested that the early findings on crossability may have been biased by use of limited germplasm. It was also noted that having many accessions was important for incorporating the F.moschata germplasm into tetraploid hybrids. A new theory was proposed that Emoschata harbours lethal alleles that become exposed when crossing to lower ploidy species. The resulting tetraploids created from crosses of Emoschata with diploids may be more genetically fit, which explains why embryo abortion and malformed seedlings were rare when F.moschata x diploid hybrids (4x) were intercrossed with other tetraploids. At the tetraploid level, 52 out of 66 hybrid combinations of species resulted in viable progeny. A total of 192 SOS were created from 25 different combinations of species and at least 28 of these SOS were fertile. With so many combinations resulting in progeny, there appears to be great potential for breeding at the tetraploid level when using the crossing strategy of this study. Each of the four species which were hybridized with F. vesca at the diploid level were fertile at the tetraploid level and successfully incorporated into SOS. However, F. vesca x diploid colchi-tetraploids have reduced fertility and best results were obtained when these hybrids were crossed with highly fertile F. orientalis (4x) or Fmoschata x diploid species (4x). The results of this study show that certain combinations of species and accessions are good as either males or females but that the interaction between males and female was not significant. Thus, it would be more important to identify tetraploid accessions with good general crossability to serve as either

238 222 males or females rather than attempt to identify specific combinations of parents that are successful. A recommended strategy for future use of this system is to identify the best parents by test crossing with highly fertile F.odentalis andlor F.moschata x diploid accessions. The success of tetraploid crosses in this study makes it difftcult to speculate which species rnay have been involved in the creation of naturally occurring polyploids. This study broadens rather than narrows the number of species that could have been involved in creation of the polyploid species. Certainly, F. vesca could have been involved in evolution of octoploid species and evidently many diploid species have potential to make octoploids when hybridized with F.vesca. The success of F. moschata x diploid tetraploids in crosses indicates that F.moschata could be used as a genetic bridge to incorporate F-nubicola, F.pentaphylla or F.vindis into the SO system. However, the Fmoschata x diploid (Le. 6x x 2x) strategy may not be as desirable as using F.vesca as a genetic bridge to diploids as recombination between the three F. moschata and one diploid genornes in the tetraploid hybrid rnay allow only incomplete genomes of the diploid species to be transmitted into SOS, while the latter would likely transmit intact genomes of the component diploids. The crossing strategy, wide germplasm base, and techniques developed each contributed to the overall success of this study. These results indicate great potential for breeding at the diploid and tetraploid levels and success of the synthetic octoploid system for widening the genetic base of octoploid cultivars.

239 REFERENCES Ahmadi, H., and R.S. Bringhurst Modes of inheritance of photoperiodisrn in Fragana. J. Amer. Soc. Hort. Sci. 1 l5(l): Ahmadi, H., and R.S. Bringhurst Genetics of sex expression in Fragan'a species. Amer. J. of Bot. 78(4): Ahmadi, H., and R.S. Bringhurst Breeding strawberries at the decaploid level. J. Amer. Soc. Hort. Sci. 117(5): Allard, R.W Principles of plant breeding. John Wiley, New York and London. Angevine, M.W Variations in the demography of natural populations of the wild strawberries Fragana vesca and Evirginiana. J. of Ecology 71 : Armstrong, K.C., S. J. Molnar, and G. Fedak Differential transmission of homoeologous chromosomes in heterozygous amphiploids: rrna loci of Thinopyrum distichum in an amphidiploid with Tnticum durum. Genorne 35: Arulsekar, S., R.S. Bringhurst and V. Voth lnheritance of PGI and LAP isozymes in octoploid cultivated strawberries. J. Amer. Soc. Hort. Sci. 1 O6: Arulsekar, S., and R.S. Bringhurst Strawberries, p In: S.D. Tanksley and T.J.Orton (eds.). lsozymes in plant genetics and breeding part B. Elsevier Science Publishers, New York. Arumuganathan, K., and E.D. Earle Estimation of nuclear DNA content of plants by flow cytometry. Plant Molecular Biology Reporter 9(3): Bhanthumnavin, K A cytological study of polyploidy in the genus Fragana. Ph. D. Thesis, University of Reading, England. Barrientos, F., and R.S. Bringhurst A haploid of an octoploid strawberry cultivar. HortScience 8 (1):44. Bauer, R New patterns of heterosis in Fragana hybrids, p In: A. Janossy and F.G.H. Lupton (eds.). Heterosis in plant breeding, proceedings of the 7th congress of Eucarpia. Elsevier Scientific Publishing Company, New York. Bauer, A Progress in breeding decaploid Fragaria x vescana. Acta Hort. 348:60-64.

240 Bino. R.J., J.N. De Vries, H.L. Kraak, and J.G. Van Pijlen Flow cytometric determination of nuclear replication stages in tornato seeds during priming and germination. Annals of Bot. 69: Blankeslee, A.F Redoublement du nombre de chromosomes chez les plants par traitment chimique. Comptes Rendus Acad. Sci. Paris 205: Bringhurst, R.S., and D.A. Khan Natural pentaploid Fragaria chiloensis - F.vesca hybrids in coastal california and their significance in polyploid Fragana evolution. Amer. J.of Bot. 50: Bringhurst. R.S., and Y.D.A. Senanayake The evolutionary significance of natural Fragana chiioensis x F. vesca hybrids resulting from unreduced gametes. Amer. J. of Bot. 53: Bringhurst, R.S., and Gill. T Origin of Fragana polyploids. II. Unreduced and double reduced gametes. Amer. J. of Bot. 57(8): Bring hurst, R.S., and V. Vot h Breeding octoploid strawberries. Iowa State J. of Research 58(4): Bringhurst, R.S Cytogenetics and evolution in Fragana. HortScience 25(8): Bumagima. SI.. and E. Mesaros (Evaluation of the results of hybridization of the strawberry species Fragaria vesca L., F. vindis Duch. and F. rnoschata Duch. by the set of hybrid seed and their viability). Genetika, Moskva 5 (5): Bumagina. S.I Meiosis in interspecific hybrids of strawberry in relation to problems of genome analysis. Dep , lopp (Russian). In Plant Breeding Abstracts 1977,44:398. Burbank, L How plants are trained to work for men. P.F. Collier & Son Company, New York. Byrne, D., and G. Jelenkovic Cytological diploidization in the cultivated octoploid strawberry Fragana xananassa. Can. J. Genet. Cytol. 18: Chabot, B.F., and J.F. Chabot Effects of light and temperature on leaf anatomy and photosynthesis in Fragana vesca. Oecologia 26: Chercuitte, L., J.A. Sullivan. Y.D. Desjardins, and R. Bedard Yield potential and vegetative growth of summer-planted strawberry. J. Amer. Soc. Hort. Sci. 116 (6):

241 Constance, L Patterns in the distribution of Japanese-American Umbelliferae, p In: G. Graham (ed.). Floristics and Paleofloristics of Asia and Eastern North America. Elsevier Pub. Co., New York. Dale, A., H.A. Daubeny, M. Luffrnan. and J.A. Sullivan Development of Fragana germplasrn in Canada. Acta Hort. 348: Darlington, C.D., and E.K. Janaki Ammal The origin and behaviour of chiasrnata. Bot. Gaz. 93: Darlington. C.D Chromosome botany and the origins of cultivated plants. George Allen and Unwin Ltd.. London. Darrow, G.M Polyploidy in fruit development. Proc. Amer. Soc. for Hort. Sci Darrow, G.M The Strawberry. Holt, Rinehart, and Winston, New York. Daubeny, H.A Strawberry breeding in Canada. HortScience 25(8): Davidson, CG Canadian wild plant germplasm of economic significance. Can. J. of Plant Sci. 75(1): Davis, T.M. 1996a. The use of molecular markers to accelerate cultivar development in Fragana, p In:M.P.Pritts, C.K.Chandler, and T.E.Crocker (eds.). Proceedings of the fourth North American strawberry conference. University of Florida. Davis, T.M., H. Yu, and K.M. Haymes. 1 W6b. Genetic linkage group structure in the diploid strawberry: evidence for chromosome differentiation between Fragana vexa and F. viridis, p In: M.P.Pritts, CKChandler, and T.E.Crocker (eds.). Proceedings of the fourth North American strawberry conference. University of Florida. Dickson, E.E., K. Arurnuganathan, S. Kresovich, and J.J. Doyle Nuclear DNA content variation within the Rosaceae. Amer. J. of Bot. 79(9):lO8I-l086. Dubcovsky, J., and J. Dvorak Genome identifications of the Tnticum crassum complex (Poaceae) with the restriction patterns of repeated nucleotide sequences. Amer. J. of Bot. 82(l):l3l-l4O. Eigsti, O.J., and P. Dustin Colchicine in agriculture, medicine, biology and chemistry. Iowa State College Press.

242 Ellis, J.R Cytogenetic studies in genera Fragaria and Potentilla. PhD Thesis. University College, London. Evans, W.D Cytogenetic investigations in the genus Fragana. PhD thesis. University of Reading, England. Evans, W.D., and J.K. Jones Incompatibility in Fragana. Can. J. Genet. Cytol Evans, W.D Evidence of a crossability banier in diploid x hexaploid and diploid x octoploid crosses in the genus Fragaria. Euphytica Evans, W.D The use of synthetic octoploids in strawberry breeding. Euphytica 26: Evans, W.D. l982a. Guelph SOI synthetic octoploid breeding clone. HortScience l7(5):833. Evans, W.D. l982b. Guelph S02 synthetic octoploid breeding clone. HortScience 17(5):834. Evans, W.D. 1982~. The production of multispecific octoploids from Fragana species and the cultivated strawberry. Euphytica 31 :go Fadeeva, T.S Problems of comparative plant genetics: Communication 1. Principles of genome analysis (with reference to the genus Fragana) Genetika 216-q6. Federova, N.J Crossability and phylogenetic relations in the main European species of Fragana. Compt. Rend. (Doklady) Acad. Sci. U.S.S. R , Feldman, M Further evidence of natural hybridization between tetraploid species of Aegilops section Pleionathera. Evolution, 1 9: Feldman, M., and ER. Sears The wild gene resources of wheat. Scientific Amer. 244(l): Fletcher, S The strawberry in North America; history, origin, botany and breeding. Macmillan Company, New York. Galletta, G.J Pollen and seed management, p In: J.N. Moore and J. Janick (eds.). Methods in fruit breeding. Purdue University Press, Indiana. Galletta, G.J., and J.L. Maas Strawberry genetics. HortScience 25(8):87l- 879.

243 Georgieva-Todorova, J lnterspecific hybridization in the genus Helianthus L. Zeitschrift fur Pflanzenzuchtung 93: Gill, K.S., H.S. Dhaliwal, and H. Singh Cataloguing and pre-breeding of wheat genetic resources. Biotechnology Centre, Punjab Agricultural University, Ludhiana, India. Gimenez-Ferrer, R.M., Scheerens, J.C., and W.A. Erb In vitro screening of 76 strawberry cultivars for two spotted spider mite resistance. HortScience 28 (8) : Grattapaglia, D., and H.D. Bradshaw Nuclear DNA content of cornrnercially important Eucalyptus species and hybrids. Can. J. for Res. 24: Gustafson, J.P., and E.R. Sears An effective wheat gene manupulation system: problems and uses. Plant breeding reviews 11 : Gustafson, J.P., and A.R. Dera Alien gene manipulation and expression in wheat. Genome 31 :l Hancock, J.F., J.L. Maas, C.H. Shanks, P.L. Breen, and J.J. Luby Strawberries (Fragada spp.), p In: J.N. Moore and J.R. Ballington (eds.). Genetic resources in temperate fruit and nut crops. Int. Soc. Hort. Sci. Wageningen, Netherlands. Hancock, J.F., D.H. Scott, and F.J. Lawrence Strawberries, In J. Janick, and J.N. Moore (eds.) Fruit Breeding, vol 2. Wiley, New York. Hara, H Contributions to the study of variations in the Japanese plants closely related to those of Europe and North America. I 1. J. Fac. Sci. Univ. Tokyo, Sect. 3, Bot. 6: Hara. H The Flora of Eastern Himalaya. The University of Tokyo Press. Japan. Hara, H Corresponding taxa in North America, Japan and the Himalayas, p In: D.H. Valentine (ed.). Taxonomy, phytogeograp hy and evolution. Academic Press, New York. Harrison, R.E., J.J. Luby, and GR. Fumier Molecular investigation of the chloroplast genome in Fragada spp. Acta Hort. 348: Harrison, R.E., J.J. Luby, and G.R. Fumier Chloroplast DNA restriction fragment variation among strawberry (Fragana spp.) taxa. J. Amer. Soc. Hort. Sci. 122(1):63-68.

244 Hermsen, J.G.Th lntroductory considerations on distant hybridization, p ln: G. Kalloo, and J.B. Chowdhury (eds.). Distant hybridization of crop plants. Springer-Verlag. New York. Hughes, HG., and J.Janick Production of tetrahaploids in the cultivated strawberry. HortScience 9(5) Hummer, K What's new in strawberry genetic resources: Raw materials for a better berry, p In: M.P. Pritts, C.K. Chandler, and T.E. Crocker (eds.). Proceedings of the fourth North American strawberry conference. University of Florida. Ibrahim, A.M.F., K. Sadanaga, and E.L. Denisen Chromosomal behavior in octoploid strawberry progenies and their parental clones during meiosis. J. Am. Soc. Hort. Sci.106 (4): Iwatsubo, Y., and N. Naruhashi Karyotypes of three species of Fragana (Rosaceae). Ctyologia 54: Iwatsubo, Y., and N. Naruhashi Karyotypes of Fragana nubicola and F. daitoniana (Rosaceae). Cytologia 56: Jacobsen. T.L., and S.H. Jansky Effects of pre-breeding wild species on tuberization of Solanum tuberosum haploid-wild species hybrids.66 (1 2: Jelenkovic, G., M.L. Wilson, and P.J. Harding An evaluation of intergeneric hybridization of Fragana spp. x PotenMa spp. as a means of haploid production. Euphytica 33: Jones, J.K Evolution and breeding potential in strawberries. Sci. Hort. 18: Jones, J.K Strawberry, Fragana x ananassa (Rosaceae), p In N.W. Simmonds (ed.). Evolution of crop plants. Longman Group Ltd., London; UK Jonkers, H Accelerated flowering of strawberry seedlings. Euphytica 7:M- 46. Kankanpaa, J., L. Mannonen, and A.H. Schulman The genome size of Hordeum species show considerable variation. Genome 39: Kantor, T.S Results of breeding and genetic work aimed at creating economicaliy valuable cultivars from incongruent crossing of Fragana x ananassa Duch. X Fragaia moschata Duch. Genetika 19(l2):

245 Kornas, J Corresponding taxa and their ecological background in the forests of temperate Eurasia and North Arnerica, p In: D.H.Valentine (ed.). Taxonomy, phytogeography and evolution. Academic Press, London and New York. Lapins, K.O Mutation breeding, p In: J.N. Moore, and J. Janick (eds.). Methods in fruit breeding. Purdue University Press, Indiana. Layne, R. E.C Hybridization. In Methods in fruit breeding, p In: J. N. Moore and J. Janick (eds.). Methods in fruit breeding. Purdue University Press, Indiana. Lelley. T., E. Kazman, K.M. Devos, and M.D. Gale Use of RFLPs to determine the chromosome composition of tetraploid triticale (AIB)(AIB)RR. Genome 38: Lewers,K.S., S.K. St-Martin, 8.R Hedges, and R.G. Palmer Testcross evaluation of soybean germplasm. Crop. Sci. 38 (5): Lewis, D.. and L.K. Crowe Unilateral interspecific incompatibility in fiowering plants. Heredity 12: Li, H.L Eastern Asia-eastern North American species-pairs in wide ranging genera, p In: G. Graham (ed.). Floristics and Paleofloristics of Asia and Eastern North America. Elsevier Pub. Co., New York. Lilienfeld, F.A Cytological and genetical studies in Fragana. 1. A fertile tetraploid hybrid between F. nipponica (n=7) and F. elaitor (n=21). Japan. J. Bot Lowery, P.P., and AG. Jones Systematics of Osmohiza raf. (Apiaceae: Apioideae). Annals of the Missouri Botanical Gardens 71 : Maas, J.1, M.R. Pooler and G. J. Galletta Bacterial angular leafspot disease strawberry: present status and prospects for control. Advances in Strawberry Res. 14: MacFarlane Smith, W.H., J.K. Jones, and A.R. Sebastiampillai Pollen storage of Fragana and Poten filla. Euphytica 41 :6569. Mangelsdorf, A.J., and E.M. East Studies on the genetics of Fragana. Genetics 12: