Departament de Biologia Facultat de Ciències

Hybridization patterns in Balearic endemic plants assessed by molecular and morphological markers

— Ph. D. Thesis —

Miquel Àngel Conesa Muñoz

Supervisors:

Dr. Maurici Mus Amézquita (Universitat de les Illes Balears)

Dr. Josep Antoni Rosselló Picornell (Universitat de València)

May 2010

Palma de Mallorca

El doctor Maurici Mus Amézquita, professor titular de la Universitat de les Illes

Balears, i el doctor Josep Antoni Rosselló Picornell, professor titular de la Universitat de

València,

CERTIFIQUEN:

Que D. Miquel Àngel Conesa Muñoz ha realitzat, baix la seva direcció en el

Laboratori de Botànica de la Universitat de les Illes Balears i en el Departament de

Botànica del Jardí Botànic de la Universitat de València, el treball per optar al grau de

Doctor en Biologia de les Plantes en Condicions Mediterrànies, amb el títol:

“HYBRIDIZATION PATTERNS IN BALEARIC ENDEMIC PLANTS ASSESSED BY

MOLECULAR AND MORPHOLOGICAL MARKERS”

Considerant finalitzada la present memòria, autoritzem la seva presentació amb la finalitat de ser jutjada pel tribunal corresponent.

I per tal que així consti, signem el present certificat a Palma de Mallorca, a 27 de maig de 2010.

Dr. Maurici Mus Dr. Josep A. Rosselló

1

2

A la meva família, als meus pares.

3

4 Agraïments - Acknowledgements

En la vida tot arriba. A moments semblava que no seria així, però aquesta tesi també s’ha acabat. Per arribar avui a escriure aquestes línies, moltes persones han patit amb mi, per mi, o m’han aportat el seu coneixement i part del seu temps. Així doncs, merescut és que els recordi aquí. Segurament deixaré algú, que recordaré quan ja sigui massa tard per incloure’l. Demano disculpes. De totes formes, els que m’heu donat una ma ho sabeu. I jo també.

Aquesta tesi s’ha dut a terme gràcies a la beca FPI (BES-2002-0541) associada al projecte REN2001-3506-C02-02, i gràcies al projecte CGL2007-60550/BOS.

Primer de tot, aquesta tesi ha estat possible gràcies a les ensenyances i la paciència dels meus dos directors, Maurici Mus i Josep Antoni Rosselló. Amb ells vaig aprendre a estimar les plantes, especialment les baleàriques. Vaig aprendre que el seu estudi es basa en la interpretació de diferents tècniques, com més millor; des de la morfologia clàssica a la molècula. Els tres hem après els uns dels altres, hem lluitat i hem patit. Però també queden les hores i hores de camp (de pujar i de suar!), de laboratori, de despatx... i per suposat les cerveses! Espero que això, més que un final, sigui un inici, prístin i productiu. Gràcies per tot, sincerament. En els darrers anys, la meva família de “Biologia de les Plantes en Condicions Mediterrànies” ha estat la que m’ha dut “en bolandas” fins al dia d’avui. Sense ells possiblement hauria quedat a mig camí. Als companys del “zulo”: Jeroni i Pep (gràcies als dos per les empentes fins al final), Xavi, Miquel, Jaume (gràcies als tres pel bon criteri i per solucionar-me tots els problemes que han sorgit), Joan, Alex, Pepe (gràcies pels ànims a diari). Com no, al “gran jefe” Hipólito, a qui, certament, se li ha de posar un confessionari (gracias por todo!). També als companys que “se lo ven venir”: Igor, Alícia, Magdalena, Enrico, Sebastià, Cristina, Ocho,... Gràcies a tots pels ànims, per les esperonades i les crítiques, i, sobretot, pel recolzament des del primer dia. Això és l’esperit de grup! En els primers anys al laboratori de botànica de la UIB, he d’agrair les hores plegats, les passejades per la muntanya i els bons moments de feina a: Xavi, Eva, Cristina, Lluís, Joan i Biel. Gràcies pels bons moments viscuts, pels viatges, pels illots,... i per tot el que he après de vosaltres.

5 En els següents anys, en el Jardí Botànic de València: Arantxa (gracias por absolutamente todo, todísimo, todo), Marcela, Mercedes i Raúl (gracias por la ayuda en todo momento y por sufrir conmigo), Dani, Pepe, Cristina, Ester, Javi, Sara, Elena, Pata, David,... (gràcies pels bons moments en el Jardí... i fora!). A Sari, tant a València com a Suècia, pel temps invertit, la cordura, els ànims, i per fer les coses senzilles. Gràcies. En el camp, per localitzar les plantes, he d’agrair tant la informació com la companyia a molta gent. A Mallorca: Guillem Alomar, Vicent Fortesa, Biel Bibiloni, Joan Rita, Llorenç Sáez, Iván Ramos,.... Tots m’han donat indicacions o han vingut a suar amb mi. A Menorca: Pere Fraga i Xavi Cardona. Sense ells encara cercaria. A Pitiüses: Nèstor Torres. Sense ell no hi ha manera. També al personal del Parc de Cala d’Hort (des de 2003... Carles, Miquel, Miguel,...). He d’agrair l’aportació de material i informació també a Josep Lluís Gradaille i Magdalena Vicens, del Jardí Botànic de Sóller, i a Xuso, del Jardí Botànic de València. Gràcies per les facilitats! En la recta final de la tesi: Maria Mayol i Miquel Riba (gràcies pel temps invertit, la sensatesa i el bon enteniment), Guillem Mateu i Guillem Pons (ja sabeu perquè!). Gràcies a tots pels ànims tan sincers. Also, in the final stretch of this thesis, I am indebted to the doctors Miquel Ribas- Carbó, Guillem Mateu-Vicens and Maxim Kapralov for their friendly comments and discussion, and for the accurate English improvement of part of the document. Gràcies, спасибо.

Per totes els moments que no he pogut estar amb ells, per haver estat partícips de tot el que ha suposat per mi aquesta tesi, pel recolzament incondicional, i pels moments en que m’han permès oblidar-me de tot això, he donar els més sincers agraïments als meus companys de sempre. Ja sabeu qui sou!

Finalment, a la meva família: Xavier, Toni, Cati, Jordi,... que tant m’heu recolzat, i especialment als meus pares, Maria del Carme i Salvador, que ho han donat tot per mi, que han patit molt més que jo per arribar fins avui, i a qui dedico especialment aquesta tesi. No pel que és, sinó pel que suposa. Gràcies per ser com sou i per tot el que m’heu ensenyat i ajudat. Gràcies!

Miquel Àngel Conesa Muñoz Palma de Mallorca, 25 de maig de 2010.

6 Hybridization Patterns in Balearic Endemic Plants Assessed by Molecular and Morphological Markers PhD. Thesis

CONTENTS OF THE THESIS

page

INTRODUCTION AND OBJECTIVES 13

Chapter 1.General introduction 13

1.1. Natural hybridization in plants: definitions and concepts 15

1.1.1. Hybrids and hybridization 15 1.1.2. Natural hybridization 16 1.1.3. Introgression and hybridization of the habitat 18 1.1.4. Barriers to natural hybridization and introgression 19 1.1.4.1. Cycle of hybridization and introgression 20 1.1.4.2. Barriers in the cycle of natural hybridization and introgression 22

1.1.5. Hybrid zones 24

1.1.5.1. Models of hybrid zones 25

1.2. Natural hybrids and plant evolution 29

1.2.1. Opposite points of view of the same natural process 29 1.2.1.1. Current hypotheses 30 1.2.2. Positive and negative effects of natural hybridization in plant evolution 31 1.2.3. Polyploid and homoploid hybrid speciation 34 1.2.4. Abundance and distribution of natural hybrids 36 1.2.5. Intermediacy: other options to hybridization 38

1.2.5.1. Primary intergradation of populations 38

1.2.5.2. Incomplete lineage sorting 39

1.2.5.3. Rapid divergence and diversification in different habitats 40

1.2.5.4. Random lineage sorting effects 40 1.2.5.5. Biased concerted evolution 41 1.2.5.6. Parallel genotypic adaptation 42 1.3. Hybridization and conservation 43 1.4. References 47

Chapter 2. Objectives of the thesis 57

2.1. Aims of the thesis 59

2.2. Selected case studies 61

2.2.1. Suspected cases of homoploid hybrids in the Balearic Islands 61 2.2.2. Selection of the case studies 63 2.3. References 64

7 Hybridization Patterns in Balearic Endemic Plants Assessed by Molecular and Morphological Markers PhD. Thesis

page

CASE STUDY I: Hybridization involving the endemic Viola jaubertiana 67 Chapter 3. The hybrid origin of the endemic Viola x balearica Rosselló, Mayol & Mus 67 3.1. Introduction 69 3.1.1. The genus Viola 69 3.1.2. Hybridization in the genus Viola 70 3.1.3. Hybrid sterility and its maintenance in nature 71

3.1.4. The Viola species in the Balearic Islands 72

3.2. The hybrid and its putative parental species 76

3.2.1. The Gorg Blau locality 77

3.3. Objectives 80 3.4. Materials and methods 82 3.4.1. Plant sampling 82 3.4.2. DNA extraction 82 3.4.3. Nuclear ribosomal ITS sequences 82

3.4.4. Phylogenetic analysis 83

3.4.5. TrnT-trnLchloroplast DNA sequences 84

3.4.6. Sterility evaluation 85

3.5. Results 87

3.5.1. ITS and cpDNA species-specific markers in Balearic Viola 87 3.5.2. Nuclear and cpDNA genotyping of Viola accessions at the Gorg Blau site 87 3.5.3. The phylogenetic relationships of V. jaubertiana 90 3.5.4. Sterility evaluation of V. x balearica 91 3.6. Discussion 93 3.6.1. Hybridization between V. jaubertiana and V. alba subsp. dehnhardtii 93

3.6.2. The phylogenetic relationships of V. jaubertiana 97

3.7. References 99

CASE STUDY II: Hybridization involving the endemic Lotus fulgurans 105 Chapter 4. Hybridization processes in the Balearic Islands between the endemic Lotus 105 fulgurans (Porta) D.D. Sokoloff and the Mediterranean distributed Lotus dorycnium L. 4.1. Introduction 107 4.1.1. The genus Lotus sect. Dorycnium 107

4.1.2. The species of Lotus sect. Dorycnium in the Balearic Islands 109

4.2. The hybrid and its putative parental species 111

4.2.1. Lotus x minoricensis (Fabaceae), a new hybrid from the Balearic Islands 111

4.2.2. The putative parental species 113 4.2.3. The Na Macaret locality 114

8 Hybridization Patterns in Balearic Endemic Plants Assessed by Molecular and Morphological Markers PhD. Thesis

page

4.3. Objectives 116

4.4. Materials and methods 118 4.4.1. Plant sampling 118 4.4.2. Morphometric analysis 118 4.4.3. Hybrid index 120 4.4.4. DNA extraction 121 4.4.5. Nuclear ribosomal ITS sequences 122

4.4.6. Restriction analysis 123

4.4.7. Chloroplast DNA sequences 123

4.5. Results 125

4.5.1. Morphometric analysis 125 4.5.2. Molecular markers 130 4.5.2.1. ITS variation and genotyping 130 4.5.2.2. CpDNA polymorphisms 133 4.6. Discussion 136 4.6.1. Overall evidence suggest interspecific gene flow between L. fulgurans 136

and L. dorycnium

4.6.2. CpDNA variation better explains a related origin of both species rather 137

than hybridization

4.6.3. Hybridization and the conservation of the endangered L. fulgurans 139 4.7. References 142

CASE STUDY III: Hybridization involving the endemic Helichrysum crassifolium 147 Introduction to the case study 149 The genus Helichrysum and the Mediterranean sect. Stoechadina 150 References 153

Chapter 5. Leaf shape variation and taxonomic boundaries involving the endemic 155 Helichrysum crassifolium (L.) D. Don 5.1. Introduction 157 5.2. Objectives 159 5.3. Materials and methods 161 5.3.1. Plant sampling 161 5.3.2. Environmental parameters from the sampled localities 164 5.3.3. Geometric morphometric data acquisition 165 5.3.4. Geometric morphometric analysis 166 5.3.5. Correlation between RWs data and abiotic parameters 167 5.3.6. Other statistical analyses 168

9 Hybridization Patterns in Balearic Endemic Plants Assessed by Molecular and Morphological Markers PhD. Thesis

page

5.4. Results 169

5.4.1. Leaf dimensions 169

5.4.2. Geometric morphometric analysis 170

5.4.2.1. Allometry in geometric morphometric data 170

5.4.2.2. Relative warp analysis 171

5.4.3. Shape differences among species, populations and individuals 173

5.4.4. Shape differences unrelated to species 174

5.4.5. Correlation between leaf shape and environmental variables 176

5.5. Discussion 177

5.5.1. Helichrysum crassifolium and H. pendulum show distinct patterns of leaf 177

shape variation

5.5.2. Leaf intermediacy is not the rule in H. crassifolium and H. pendulum 178

populations

5.5.3. The nature of intraspecific leaf shape variation in H. crassifolium and H. 178

pendulum

5.6. References 182

Appendix 5.1 187

Appendix 5.2 188

Appendix 5.3 189

Chapter 6. Widespread hybridization in the Balearic Islands involving the endemic 191

Helichrysum crassifolium (L.) D. Don

6.1. Introduction 193

6.2. Objectives 194

6.3. Materials and methods 195

6.3.1. Plant material 195

6.3.2. DNA extraction 197

6.3.3. Nuclear ribosomal ETS sequences 197

6.3.4. Phylogenetic analyses 198

6.4. Results 200

6.4.1. ETS ribotypes 200

6.4.2. ETS phylogenetic analyses 206

6.5. Discussion 208

6.5.1. ETS sequence variation in Helichrysum sect. Stoechadina 209

6.5.2. Testing for divergence vs. hybridization in the Balearic Helichrysum 210

6.5.3. Widespread natural hybridization as a cause of the observed variation in 215

the ETS sequences of the Balearic Helichrysum

6.6. References 218

10 Hybridization Patterns in Balearic Endemic Plants Assessed by Molecular and Morphological Markers PhD. Thesis

page GENERAL DISCUSSION AND CONCLUSIONS 223 Chapter 7. General discussion 223 7.1. Outcomes of the thesis 225 7.1.1. The putative origin of the morphologically intermediate plants between 225 Balearic endemics and more widespread species 7.1.2. The importance of hybridization processes in the evolution of Balearic 225 endemic taxa 7.1.3. The symmetry of the hybridization processes 228 7.1.4. The hybridization processes as a threat in the conservation of the 229 endemic flora of the Balearic Islands 7.2. References 231

Chapter 8. Conclusions of the thesis 233 8.1. Conclusions 235

11 Hybridization Patterns in Balearic Endemic Plants Assessed by Molecular and Morphological Markers PhD. Thesis

12 Chapter 1. General Introduction

INTRODUCTION AND OBJECTIVES

Chapter 1 General introduction

page

1.1. Natural hybridization in plants: definitions and concepts 15

1.1.1. Hybrids and hybridization 15

1.1.2. Natural hybridization 16

1.1.3. Introgression and hybridization of the habitat 18

1.1.4. Barriers to natural hybridization and introgression 19

1.1.4.1. Cycle of hybridization and introgression 20

1.1.4.2. Barriers in the cycle of natural hybridization and introgression 22

1.1.5. Hybrid zones 24

1.1.5.1. Models of hybrid zones 25

1.2. Natural hybrids and plant evolution 29

1.2.1. Opposite points of view of the same natural process 29

1.2.1.1. Current hypotheses 30

1.2.2. Positive and negative effects of natural hybridization in plant evolution 31

1.2.3. Polyploid and homoploid hybrid speciation 34

1.2.4. Abundance and distribution of natural hybrids 36

1.2.5. Intermediacy: other options to hybridization 38

1.2.5.1. Primary intergradation of populations 38

1.2.5.2. Incomplete lineage sorting 39

1.2.5.3. Rapid divergence and diversification in different habitats 40

1.2.5.4. Random lineage sorting effects 40

1.2.5.5. Biased concerted evolution 41

1.2.5.6. Parallel genotipic adaptation 42

1.3. Hybridization and conservation 43

1.4. References 47

13 Chapter 1. General Introduction

14 Chapter 1. General Introduction Definitions and concepts

1.1. NATURAL HYBRIDIZATION IN PLANTS: DEFINITIONS AND CONCEPTS

1.1.1. Hybrids and hybridization

In ordinary language, hybridization means the cross between two different things giving rise to an intermediate one, the hybrid. Nevertheless, the first accession of most dictionaries states a biological-related definition of hybrid and hybridization. The hybrid etymology comes from the Latin hybrida, hibrida or ibrida, related to the Greek hubris, an insult or outrage, with special reference to lust, hence, an outrage on nature (Warren, 1884;

Britannica, 2007). Romans understood under hybrida, strictly speaking, the progeny of a wild boar and a sow (Warren, 1884), however probably in interpreting such offspring as a result of an outrageous "interracial miscegenation" (OED Online, 2007).

The term hybrid entered into popular use in English in the 19th century, though examples of its use have been found from the early 17th century -1601- (Harper, 2001;

Merriam-Webster Online, 2007; OED Online, 2007). The modern meaning of the term is normally “an offspring of two animals or plants of different races, breeds, varieties, species, or genera” (Merriam-Webster Online, 2007). Therefore, it has a wider application than the terms “mongrel” or “crossbreed”, which usually refer to animals or plants resulting from a cross between two races, breeds, strains, or varieties of the same species (Britannica,

2007). The first common use of the word was breeders-related, as clearly shown in the

Oxford English Dictionary (Simpson & Weiner, 1989), in the “of animals” definition of hybrid, in the quotations of Holland, 1601; Cockeram, 1623; Webster, 1828; Darwin, 1859, or

Huxley, 1862; and in the “of plants” definition, in the quotations of J. Lee, 1788; Webster,

1828; J. Baxter, 1846, or Darwin, 1867 (see Simpson & Weiner, 1989, for these authors’ references).

15 Chapter 1. General Introduction Definitions and concepts

However, besides of its “outrage to nature” etymology and its breeders-related use, the importance of hybridization in the origin and evolution of species in the wild was noticed, and it was heavily related to the biological species concept. In his revision of the importance of the process in the New Zealand flora, Cockayne (1923: 108) stated that “Darwin in The

Origin of Species makes a distinction between ‘mongrels’ (crosses between varieties) and

‘hybrids’ (crosses between species), but he consider that, except in the matter of fertility

(mongrels being usually fertile), there was ‘the closest general resemblance’ [...] between them. Regarding this matter of fertility Darwin had animals chiefly in mind and was comparing the domestic with the feral. But, in the case of plants, the matter is different for, as every gardener knows, there are hundreds of self-fertile hybrids, and some breed true”.

1.1.2. Natural hybridization

Afterwards, these terms –i.e., hybrid and hybridization- were adopted commonly by evolutionary biologists to indicate the spontaneous crosses between individuals from different species in nature (reviewed in Harrison, 1993). Indeed, other species concepts but the biological species one (i.e., evolutionary, phylogenetic –diagnosability and apomorphic-, and taxonomic; reviewed in Soltis & Soltis, 2009) also had to fit the definition and thus, the definition of hybridization has been polished up constantly for different authors throughout the 20th century. Consequently, in contemporary days it can be found a different definition – even with slight differences- for each field of investigation or even author. Nevertheless, the interest in this thesis is in a current evolutionary biology definition of hybridization. Thus, from all definitions found in literature, the following, supported by many of the most important authors in this field, was the one that better fitted the present framework:

“…natural hybridization involves successful matings in nature between individuals from two

16 Chapter 1. General Introduction Definitions and concepts populations, or groups of populations, which are distinguishable on the basis of one or more heritable characters”. (Arnold, 1997; similar to Woodruff, 1973, and Harrison, 1990).

This definition has important words that serve to correctly frame the concept of natural hybridization and its possible evolutionary consequences. In order of appearance,

“successful matings” implies that at least some remain of fertility must exist in the first hybrid generations to grant the continuation of the hybrid (e.g., Grant, 1963). The condition “in nature” rejects all man-mediated crosses (such as breeders’ crosses). Note that this does not exclude man-activity derived hybrids, such as natural hybrids allegedly produced due to an alteration of a habitat (Anderson, 1949), since the arising of such hybrids depends on natural processes. Then, the use of “populations” avoids the indication of any taxonomic rank, emphasizing that the discussion of the species concept must be out of the definition.

This was yet explicit in Woodruff’s quotations: “…it has become increasingly apparent that the nature of the ineraction –involving intergradation and hybridization- between taxa may vary geographically […]. This suggests that our terminology should not depend on the taxonomic status of populations involved or on the probable outcome of their interactions”

(Woodruff, 1973: 213).

In this direction, some authors (e.g., Miller, 1949; Woodruff, 1973; Harrison, 1986;

Arnold, 1997; Mullen et al., 2008) used the terms “divergent populations / individuals” instead of taxonomic ranks when talking about the two (or more) units involved in a hybridization event. Oppositely, and depending on the used framework, sometimes authors use a more restrictive definition, stating “different species” rather than divergernt populations (e.g., Rieseberg, 1997).

Finally, two more concepts must be denoted in the above definition; those of

“distinguishable” and “heritable”, reflecting that at least a differential trait between hybridizing units must exist, and that it must be genetically controlled and by-generation

17 Chapter 1. General Introduction Definitions and concepts transferred, discarding non-inherited phenotypical traits due strictly to the environment. Both concepts were yet included in the definitions of Woodruff (1973) and Harrison (1990).

1.1.3. Introgression and hybridization of the habitat

Once the natural hybrid has been formed, its ongoing in the wild goes through its ability to get successfully stablished in an habitat, either in one of its parental species or in a new one; and its ability to reproduce, either with other hybrids or with any of the parental species. Therefore, Anderson and colleagues described some important concepts that feeded the theory, those of “introgression” and “hybridization of the habitat”: “…if hybrids are produced, they tend to cross back to the more abundant species. The progeny of these secondary hybrids are likewise crossed back again, and so on. […] while such is not the only effect of hybridization between species, it is certainly one of the commonest. We have therefore given it a distinctive name, introgressive hybridization. In discussing the effects of introgressive hybridization, we shall speak of the hybridization of one species into another rather than hybridization with another. […]. After a few back-crosses most of the individuals cannot be distinguished by morphological means from the pure species” (Anderson &

Hubritch, 1938: 396).

“It has been very generally recognized that if hybrids are to survive we must have intermediate habitats for them. It has not been emphasized, however, that if anything beyond the first hybrid generation is to pull through, we must have habitats that are not only intermediate but which present recombinations of the contrasting differences of the original habitats. [...] Only by a hybridization of the habitat can the hybrid recombinations be preserved in nature. […]. It is concluded that hybrid swarms can survive only in ‘hybridized habitats’” (Anderson, 1948: 5-7).

18 Chapter 1. General Introduction Definitions and concepts

1.1.4. Barriers to natural hybridization and introgression

It holds true that the highest chance for effective matings leading to fully viable and fertile offspring does exist among congeners (i.e., belonging to the same taxa or population). Therefore, the occurrence of interspecific hybrids –or hybrids between infraspecific taxa of the same species, and even different populations– must encounter limitations to the full or complete compatibility: the so called barriers to natural hybridization.

Moreover, once the fertile hybrid offspring occurs there is a great chance for introgression into each parental species (e.g., Anderson, 1949; Heiser, 1973; Arnold et al., 1991; Rhymer

& Simberloff, 1996; Burke et al., 1998; Ellstrand et al., 1999; Rieseberg et al., 1999; Grant et al., 2004; Cruzan, 2005; Whitney et al., 2006), nevertheless limited for the so called barriers to introgression.

Strictly speaking, the arising of a natural hybrid must include the formation of a viable seed of hybrid origin in nature. Therefore, two main stages in the cycle of formation of a hybrid plant can be separated (Figure 1.1), corresponding to before and after in the hybrid arising. Different barriers to natural hybridization and introgression (i.e., barriers against the natural hybrid formation and to its long term descendance success; Table 1.1) act in one or both stages. In fact, if interspecific hybrids are not found in a natural habitat it does not mean that such crosses do not occur. It only means that there is at least an unbridgeable barrier in one of both stages, and only if it is located in the first stage the hybrid is really not produced. Consequently, if the barrier is in the second stage, the viable hybrid seed/plant is produced but it is not able to grow, to compete or to reproduce in that habitat. Therefore, barriers affecting the first stage would be the ones responsible of the rarity in the hybrid formation, and the barriers in the second stage would be the ones broken by these positively selected hybrid genotypes existing in nature.

19 Chapter 1. General Introduction Definitions and concepts

1.1.4.1. Cycle of hybridization and introgression

An example of what would be a cycle of natural hybrid formation and success in the plant evolution process is represented in the Figure 1.1. The scenario must be that of two parental taxa (i.e., divergent populations) with some degree of sympatry, thus able to cross

(but also through long distance dispersions). After such mating, a hybrid zygote can be

formed and rising a viable seed able to germinate: the interspecific hybrid (F1). This would be the end of the first stage of the cycle, since the viable hybrid does exist; and it would start the second stage of the cycle, that of the hybrid success. If that hybrid is fertile and can

cross in turn with another hybrid plant, either F1 or another hybrid generation (Fn), the

hybridizing cycle could “roll” indefinitely. In any of these generations (F1-Fn), a new lineage could arise, that is, hybrid speciation.

On the other hand, in this second stage of the cycle (i.e., hybrid success) the fertile

hybrid (F1-Fn) could cross in turn with either of its parental taxa, thus starting the cycle of introgression. Like in the hybridization cycle, an introgressed zygote and then a viable

introgressant individual could arise (B1). When a fertile introgressant is created, the common situation would be that of further crosses with the same parental taxon, generation

by generation. Although, the possibility of cross with introgressants (Bn) rather than its parental taxon, or the introgressant speciation –i.e., new lineage arising– could also be possible. In addition, the repeated cross of the introgressed individuals with the same parental taxon should produce parental-like introgressants (e.g., Anderson & Hubritch,

1938; Arnold, 1997) resembling morphologically, ecologically and genetically (for many molecular markers) an individual of the original parental taxon. However, some barriers can block any of the steps occurring in these cycles (Figure 1.1).

A comprehensive summary of the barriers that can be found in the differents steps of the above stated cycle is shown in Table 1.1.

20 Chapter 1. General Introduction Definitions and concepts

Figure 1.1. Barriers to hybridization and introgression that can occur in the cycle of arising and success of natural hybrids. Barriers to hybridization in the first stage (in blue) occur before the formation of the hybrid (viable seed). Barriers to hybridization in the second stage occur from the viable hybrid formation to its cross.

The hybridization cycle (in red) can continue indefinitely, or lead to hybrid speciation in any generation (Fn). Also, the hybrid can cross with either parental individuals (B1), starting the cycle of introgression (in green). After many cycles (Bn), the fertile introgressant can be almost undistinguishable from the parental taxon to which backcrossed. Indeed, in any of the backcross generations (Bn), introgressant speciation is possible. See the text for more information.

21 Chapter 1. General Introduction Definitions and concepts

1.1.4.2. Barriers in the cycle of natural hybridization and introgression

When different authors summarize or compile barriers to hybridization, two common classifications can be found, differing on the location of the inflection point to separate the barriers. In the first case, the inflection point has been established in the mating success

(i.e., pollen arrival to the other species stigma in plants) and thus are the “premating and postmating barriers” (e.g., Rieseberg, 1997; Soltis & Soltis, 2009). In the second case, the inflection point is the formation of the zygote, and thus are the “prezygotic and postzygotic barriers” (e.g., Levin et al., 1996; Lowry et al., 2008). Nevertheless, a comprehensive classification of the barriers to natural hybridization and introgression from the viewpoint of the true formation of a natural hybrid can also be established (Table 1.1), mainly based in two premises. First, the inflection point to the hybrid formation is the arising of a hybrid viable seed (i.e., intrinsically able to germinate and develop a plant). Second, all the barriers to natural hybridization found to be after the arising of the hybrid plant would prevent both hybrid speciation and introgression. Therefore, a summary of most of the barriers to hybridization and introgression are shown in the Table 1.1. Those barriers are stated in a comprehensive sequence order related to the cycle shown above (Figure 1.1).

The importance of pre-hybridization barriers is not just that prevents the first hybrid

occurrence (F1) in the natural hybridization process, but it is suggested this to be the most difficult stage to achieve, due to the –relative- rarity of the formation of the first hybrid generation (e.g., Arnold, 1993, Carney et al., 1996; Emms et al., 1996; Cruzan & Arnold,

1999; Burke & Arnold, 2001). The recognition of the mechanisms that limit succesful interspecific crosses was already stated by Darwin (1859; cited in Arnold, 1997). On the other hand, the occurrence of these barriers against the arising of the hybrid favoured the hypotheses of natural hybrids as evolutionary “dead ends” (Mayr, 1942).

22 Chapter 1. General Introduction Definitions and concepts

Table 1.1. Barriers to natural hybridization and introgression. Major classifications of these barriers are shown in the left columns. (a) Separation based on pre- and post-hybridization barriers. (b) Classification proposed by Arnold (1997). (c) Classical divisions of the barriers in pre- and post-mating barriers and (d) pre- and post- zygotic barriers.

(a) (b) (c) (d) Barriers to natural hybridization and introgression Asexual reproduction (clonal, apomixis, inbreeding) Geographic Pollen delivery / pollination mechanisms (“pollinator behavior”) Ethological isolation Mechanical isolation Pre-mating Pre-mating Pollination syndrome Flower phenologies

Pollen stigma interactions Competitive interactions between interspecific pollen: the mentor effect Pollen germination Pollen tube penetrates stigma Pollen-style interactions

Pre-zygotic Pollen tube development Pollen tube growth through style Post-pollination Pollen tube penetration of the ovule micropyle Pollen tube discharges masculine gametes in embryo sac Pre-hybridization barriers Fertilization

Chromosome structural differences (low viability / fertility) Incongruity (hetero-incompatibility): too divergent gametes Self-incompatibility: male gamete too similar to the maternal genotype Gene recombinations (disruption of coadapted genomes) Zygote formation

Embryo abortion Low viability of the embryo Post-mating Endogenous selection Seed intrinsic capability to germinate properly

Seed extrinsic possibilities to germinate properly Finding of the appropriate habitat (dispersion/soil preferences) Growth (low capacity, environmental limitations)

1 Survivorship of the genotypes resulting from environment-dependent to F

Post-fertilization selection (herbivorie, parasitism, diseases) Fertility (production of viable gametes)

Post-zygotic Occurrence of interfertile individuals (able to allow successful crosses) − Barriers pre-hybridization-like −

− Barriers to F1 formation-like − Exogenous selection Fromation of novel evolutionary lineages Post-hybridization barriers , and later 1 Reticulation (cross among taxa derived from a common hybridization , B 2 process) to F Speciation without possibility of cross among related species

23 Chapter 1. General Introduction Definitions and concepts

As emphasized above, if a hybrid individual occurs in natural conditions, it must be assumed that all barriers to natural hybridization related to hybrid arising have been overcome. Therefore, after hybrid arising, two main aspects should be addressed: hybrid viability and fertility, and barriers to introgression and speciation. In the first case, if that hybrid is not able to break more barriers after its arising, it becomes a plant “mule”, since it can not succeed in the evolutionary process. Nevertheless, the hybrid is actually formed.

Although, compared to parental species, a hybrid can also be selected against.

Furthermore, in the case of the barriers to introgression and speciation, barriers avoiding introgression would favour hybrid isolation from its parental taxa and thus, barriers to speciation would be overcome. In addition, mainly the same barriers that have been

overcome in the F1 hybrid formation will be the ones to break again in the formation of the

first introgressant generation (B1) and if so, it will start the introgression cycle.

Summarizing, barriers affecting the first stage would be the ones responsible of that hybrid formation rarity, and barriers in the second stage would be the ones broken by these

positively selected F1 genotypes that give rise to subsequent generations: the hybrid success in nature.

1.1.5. Hybrid zones

The hybrid zone is known to be the physical region in which natural hybrid swarms

(or simply hybridization events) do occur (e.g., Arnold, 1997). Normally, it is a zone of contact among different habitats (i.e., different environmental conditions), a disturbed zone

(altered by man or by natural happenings), or just a zone with an environmental gradient, suitable for the occurrence of two (or more) taxa (i.e., divergent populations) allowing its hybridization and the success of at least part of its offspring (e.g., Anderson, 1948; and following references). Woodruff (1973) reviewed that “hybrid belt”, “tension zone” and

24 Chapter 1. General Introduction Definitions and concepts

“zones of intergradation, recombination and contact” also have been used instead of “hybrid zone”, suggesting the latter as the most useful of the above, and defining it as “an area in which hybrids occur” (Woodruff, 1973). Moreover, he stated that Short (1969) used “hybrid zone” restricted to zones where only hybrids occur –i.e., allopatric parents-, and “zone of overlap and hybridization” to zones where both parental types and hybrids occur –i.e., sympatric parents-.

1.1.5.1. Models of hybrid zones

When attempting to summarize hybrid zones, authors invoke different viewpoints of the relations among parental taxa and hybrids, such as geographical distribution (e.g.,

Woodruff, 1973); reproductive and ploidy level (e.g., Grant, 1953); or fitness (e.g., Arnold,

1997). Woodruff (1973) reviewed the hybrid zones studied by many authors, and summarized them in a scheme of the types of “hybrid zones”, shown in the Table 1.2. He stated that his suggested scheme for hybrid zones differs from the one proposed by Short

(1969) in that “it is independent of taxonomic considerations” (Woodruff, 1973) and thus, as for the definition of hybridization, this is the most suitable in the present framework.

With respect to the classification of Grant (1953), based on the reproductive and ploidy level, it really represents a classification of “hybrid complexes” arranged in a decreasing taxonomic complexity order, rather than a truly hybrid zone classification.

However, he stated that the homogamic compex (i.e., hybrids are sexual and diploids, with normal chromosomal cycle at meiosis) is the single “open system of evolution”, since it has much less restriction of gene recombination in the hybrid derivates, and has no restrictions on the free recombination of genes in the hybrid derivates. Yet, this situation is the most easily found in hybrid swarms.

25 Chapter 1. General Introduction Definitions and concepts

Table 1.2. Woodruff (1973) summarized scheme for hybrid zone types, independent of taxonomic considerations.

Hybrid zone Definition Intergradation or allopatric hybridization (Mayr, 1942) Hybrids constitute an annectant population between ranges of parental taxa Parapatric hybridization (Smith, 1955, 1965) “Situations wherein ranges are in contact and genic interchange is geographically possible even without sympatry”. No area populated exclusively by hybrids, but adjacent to contact; being a type of sympatry rather than allopatry since the two forms are in –even slight- contact” (Key, 1968) Sympatric hybridization Peripheral sympatric (Short, 1969) Hybrids occur at range periphery Widespread sympatric (Woodruff, 1973) Parents and hybrids together all along the overlap area Localized sympatric (Woodruff, 1973) Parents and hybrids together in specific zones in the overlap region, which are commonly disturbed areas Man-mediated hybridization (Woodruff, 1973) Naturally allopatric taxa mixed due to man-mediated taxa movement

More modern classifications of hybrid zones are based in the fitness relations among parental taxa and hybrids (e.g., Arnold, 1997). From this viewpoint, several theoretical frameworks can be found in hybrid-related literature to explain the suitability of the natural hybrid occurrence or the appearance of a hybrid zone. Examples of such frameworks are those of Darwin (1859); Dobzhansky (1937, 1940) Huxley, (1942); Mayr (1942); Wilson

(1965); Remington (1958); Endler (1973, 1977); Moore (1977); Barton (1979); Howard

(1982, 1986), and Harrison (1986), (reviewed in Arnold, 1997; see those references therein). One of the most important differences among these theories is the relative importance given to selection and/or dispersal in explaining the evolution of the hybrid zone.

In other words, those differ in the importance given to the main groups of barriers (e.g.,

Table 1.1) to explain the relations within the hybrid swarm.

26 Chapter 1. General Introduction Definitions and concepts

All these theories use to be standarized in three historical models for hybrid zones, those of the Bounded Hybrid Superiority model (Moore, 1977), Mosaic model (Howard,

1982, 1986; Harrison, 1986), and Tension Zone model (Huxley, 1942; Bigelow, 1965; Key,

1968; Barton, 1979). Afterwards, Arnold (1997) designed the Evolutionary Novelty model, chiefly to create a more suitable model than the three above to explain many results and conclusions achieved by very different authors that do not feed into the historical frameworks (Table 1.3). The Bounded Hybrid Superiority and the Mosaic models correspond to the “Environmental Gradients” (Endler, 1977), thus environment dependent models in contrast with the Tension Zone model. Even though, the fitness of hybrids related to its parental species differentiates the two former models, therefore becoming all three models differentiated from the join of two dissenting assumptions: environment-dependence and hybrid fitness related to parents.

Other models have also been suggested to explain specific situations rather than to

feed-in most of hybrid zone cases. One of these examples is the F1-dominated hybrid zone

model (Milne et al., 2003). This model was designed to explain situations where F1 genotype overcomes any other genotype in a specific habitat. Therefore, this model would be closely related to the Bounded Hybrid Superiority model of Moore (1977).

27 Chapter 1. General Introduction Definitions and concepts

Table 1.3. Summary of the hybrid zone models reviewed in Arnold (1997). Exogenous selection: environment- dependent or selection-gradents determine hybrid existence. Endogenous selection: environment-independent mechanisms determine hybrid existence. More fitness: hybrids more fit than parents at least in some habitats. Less fitness: hybrids less fit than parents in any habitat. Parental habitat: hybrids may be more fit than one of the parental forms in the parental habitat. Ecotones: hybrids normally restricted to ecotones or disturbed areas. No ecotones: not assumed that hybrids are normally restricted to ecotonal areas.

Hybrid zone models Assumptions Bounded Hybrid Superiority More fitness Ecotones (Moore, 1977) Exogenous selection Mosaic (Howard, 1982, 1986; Harrison, 1986) Less fitness Ecotones? Tension Zone (Dynamic Equilibrium) Endogenous selection (Huxley, 1942; Bigelow, 1965; Key, 1968; Barton, 1979) Evolutionary Novelty Exogenous and More fitness No ecotones (Arnold, 1997). endogenous selection Parental habitat

28 Chapter 1. General Introduction Natural hybrids and plant evolution

1.2. NATURAL HYBRIDS AND PLANT EVOLUTION

1.2.1. Opposite points of view of the same natural process

In natural hybridization-related literature it is patent the occurrence of primarily two opposite points of view concerning on the importance that natural hybridization has in the evolutionary theory. When early evolutionary biologists adopted the term “hybridization” it was due to the importance that some of them gave to that process in nature. That is, when authors like Linné (1760), Kölreuter and Gartner (cited in Roberts, 1929), Herbert (1847),

Naudin (1863) or Mendel (1866), first used this concept it was to indicate that natural hybridization could explain some of the observations in plant –and animal- species, as well as that this could be the origin of new species. In times of firm believes of God creation,

Linné identification of the so-called “mule species” was disappointed and thus, the author ommited it –and the whole process of hybridization- in his Species Plantarum (1753).

Indeed, later in time it favoured the appearance of some hypotheses pointing out the importance of natural hybridization in the evolution of the species and the widespread occurrence of this process, taking these first observations to the extreme: “Natural hybridization is the single most important factor in producing the necessary genetic variation for evolution (Lotsy, 1916)”, (cited in Arnold, 1994: 141).

During the second half of the 20th Century, two of the most active authors developing studies to understand the importance of hybridization in plant evolution were Edgar

Anderson and George Ledyard Stebbins Jr. As stated above, they described many theoretical concepts of this theory, and hypothesized that hybridization could be responsible of evolutive bursts along plant history (Anderson & Stebbins, 1954). Other authors endorsing their theories were Lotsy (1916, 1931), Cockayne (1923), Lowe (1936), Grant

(e.g., 1993; Grant & Grant, 1992, Grant et al., 2005), or Lewontin & Birch (1966).

29 Chapter 1. General Introduction Natural hybrids and plant evolution

On the other side, the opposite point of view in this debate was championed in the

20th century by more skeptical authors like Warren H. Wagner Jr. and Ernst W. Mayr.

Following Darwin’s ideas (1859), chiefly based on a zoological point of view, they supported the belief that natural hybridization processes have no great evolutionary importance, mainly due to the low or null fertility and/or viability of the first hybrid generations, especially in the parental species habitats (e.g., Wagner, 1992). They stated that parental species could be recognized as such –at least morphologically and with some molecular techniques

(e.g., Mayr, 1992)- after the hybridization process, thus not being affected, and maintaining the species integrity: “In the rare cases that two well differentiated species happen to be interfertile enough to produce fertile progeny, their hybrids will usually have to fit into some hybrid niche. Such fertile hybrids will therefore tend to be transient, disappearing once the differentiated community returns and the parental species re-occupy their normal habitats.

[...]. A kind of evolutionary noise is produced [...] (Wagner, 1970: 149-150). Other authors endorsing this more skeptical theory were also Dobzhansky (e.g., 1937), and Coyne & Orr

(2004).

1.2.1.1. Current hypotheses

Afterwards, many important scientific achievements and new hypotheses to understand the importance of natural hybridization in the evolution of plants came mainly from the work of Michael L. Arnold and Loren H. Rieseberg. In the last 30 years, their respective work with Iris and Helianthus hybrids lead to detailed knowledge of some examples of hybrid zones, introgression and hybrid speciation, besides of being the basis for many modern genetic models. Nevertheless, current studies also find weaknesses in the zoological point of view of a low frequency of fit hybrids leading to speciation (e.g., Dowling

& DeMarais, 1993; Mable, 2004; Morjan & Rieseberg, 2004). These renewed views

30 Chapter 1. General Introduction Natural hybrids and plant evolution especially reinforced the early views of Anderson & Stebbins: “Results from more recent analyses of purported cases of hybridization suggest that an intermediate stance must be taken -between Lotsy (1916) and Wagner (1970)- with regard to the evolutionary importance of natural hybridization. […] Natural hybridization is apparently not ‘evolutionary noise’ (Wagner 1970) that gets in the way of a clear definition of species diversification.

Rather, natural hybridization and its various outcomes appear to make a creative contribution to adaptation and speciation” (Arnold, 1994: 141, 146).

“In summary, hybrids typically are not morphologically intermediate, but rather are a mosaic of parental, intermediate, and novel characters. Likewise, hybrids may be less, more, or equally fit relative to the parents. Character coherence appears to be the exception rather than the rule. Finally, hybrids between closely related lineages are unlikely to cause major disruptions in the topology of phylogenetic trees” (Rieseberg, 1995: 947).

Indeed, their studies showed that even extremely low fertility or viability of early-

generation hybrids and introgressants (e.g., F1, F2, B1,...) does not necessarily prevent extensive gene flow and the establishment of new evolutionary lineages (Arnold et al.,

1999). On the other hand, their studies and other important contemporary works demonstrated that natural hybridization could be either a creative force, or a threat for some species (Ellstrand & Elam, 1993; Levin et al., 1996; Arnold et al., 1999).

1.2.2. Positive and negative effects of natural hybridization in plant evolution

Many case studies suggest a creative role for natural hybridization processes, including the origin of new ecotypes or species, the increasing genetic diversity within species, the origin and transfer of genetic adaptations, the reinforcement or breakdown of reproductive barriers, and the promotion of dispersion and colonization (Rieseberg, 1997;

Hardig et al., 2000). Thus, as stated above, hybridization is a relatively common feature of

31 Chapter 1. General Introduction Natural hybrids and plant evolution plant vascular species and it is an evolutionary key force shaping speciation through processes of homoploid hybrid speciation, introgression, and allopolyploidy (e.g., Anderson

& Stebbins, 1954; Grant, 1981; Abbott, 1992; Rieseberg, 1995; Arnold, 1997; Chapman &

Burke, 2007; Doyle et al., 2008; Soltis & Soltis, 2009). Rapid karyotypic and genomic changes (including elimination and doubling of specific DNA sequences, chromosome rearrangements, or loss of entire chromosomes), as well as epigenetic changes (such as differential DNA methylation, gene-dosage compensation, gene silencing) can occur in the hybrids (and the allopolyploid derivatives) after relatively very few generations (Ungerer et al., 1998; Nasrallah et al., 2000; Lai et al., 2006). These changes demonstrate that extensive genetic diversity can be generated in hybrid derivatives in a short period of time after their formation, contributing to the success and diversification of many plant lineages

(Arnold, 2006).

In some cases, hybridization leads to hybrid isolation from the parental taxa, favouring the arising of new lineages. However, in other cases hybridization leads to infertile/inviable offspring which, if frequent, can limit parental taxa reproduction because a part of its descendance will not succeed. Nevertheless, a more common result of hybridization processes is introgression (Rieseberg, 1997). Introgressive events involving narrow distributed species or isolated populations can dilute its genetic particularities or local adaptations through specific genetic interactions (Antonovics, 1976; Simberloff, 1988).

Yet, asymmetrical introgression through unidirectional backcrossing of partially fertile hybrids to the more abundant parent can lead to the loss of parental nuclear genotypes and can ultimately lead to local extinction of the rare species where the direction of gene flow is into the rarer species (Burgess et al., 2005). Thus, in contrast with its creative role in plant evolution, hybridization may contribute to the demise of rare species through demographic swamping and genetic assimilation by widespread congeners (Levin et al., 1996; Wolf et al.,

32 Chapter 1. General Introduction Natural hybrids and plant evolution

2001; Buerkle et al., 2003). Anthropogenetic disturbance promotes hybridization by fostering joint colonization by normally allopatric species, creating opportunities for interspecific gene flow and suitable habitats for the survival of hybrids (Arnold, 1997). The presence of hybrids could cause new competition with parental species and could be a threat to either of them by crowding them out (e.g., Milne et al., 2003). Therefore, hybridization is one of the biotic interactions promoting extinction that increases threats to rare species whose ecological barriers are being disrupted by human activities (Rhymer &

Simberloff, 1996; Seehausen et al., 2008). This can be particularly relevant in islands that are typically rich in endemic species showing ecological diversification (Brochmann, 1984;

Levin, 2000).

Nonetheless, hybridization is not an imposition of specific gene combinations but an increase of the genotypes available in a habitat to be naturally selected and thus, an increase of the genotypes available for that population adaptation to its changing habitat.

However, natural selection processes tend to eliminate the less competitive genotypes in each habitat, thus reducing the ones able to reproduce. In this respect, some experiments with Helianthus (Rieseberg et al., 1996a) suggested that selection would favor specific genomic compositions and epistatic interactions, which strongly suggests that it must be considered selection rather than chance, at least in some cases of hybrid species formation.

Thus, genomic structure and composition of hybrid species is essentially fixed within a few generations after the initial hybridization event and remain relatively static thereafter

(Rieseberg, 1997).

In contrast with the above stressed demise effect that natural hybridization can produce, there is also an increasing number of studies reporting the ancient or recent origin of plant species through hybridization (e.g., Arnold et al., 1990; Rieseberg, 1991; James &

Abbott, 2005; Paun et al., 2006; Devey et al., 2008; Peterson et al., 2009), including cases

33 Chapter 1. General Introduction Natural hybrids and plant evolution of homoploid hybrid speciation, in spite of its rarity (Chapman & Burke, 2007; Soltis & Soltis,

2009).

1.2.3. Polyploid and homoploid hybrid speciation

Polyploidy is an increase in the number of chromosomes, generally consisting in genome doubling. Two main types of polyploids are usually recognized in literature: auto- and allopolyploids, depending on the origin of such extra-cromosomes (e.g.,Grant, 1981).

Thus, autoployploids use to be formed after crosses involving unreduced gametes into the same taxon; whereas allopolyploids would arise after a cross between different taxa.

However, some authors consider hybridization as a cross between genetically different individuals instead of different taxa and thus, hybridization would also be a source of autopolyploids (e.g., Arnold, 1997; Soltis & Soltis, 1999). In any case, polyploid hybridization implies an increase in chromosome number of the offspring related to the parental plants. In contrast, homoploid hybridization produces hybrids with the same ploidy level as parents.

With regards to their frequency in nature, polyploid hybridization is a relatively more common process than homoploid hybridization (Soltis et al., 2004), since it is much easier to achieve (e.g., Grant, 1981; Rieseberg & Willis, 2007). Therefore, it is expected that most cases of speciation through hybridization would arise from polyploidy events. Polyploid hybrid speciation is favoured basically by two aspects: polyploid hybrids are usually completely fertile, and genome doubling is generally a barrier against back-cross with parental taxa (e.g., Rieseberg & Willis, 2007; Soltis & Soltis, 2009). Moreover, some lines of evidence suggest that most angiosperm lineages could have a more or less recent origin through polyploid speciation events (Soltis et al., 2008).

34 Chapter 1. General Introduction Natural hybrids and plant evolution

Oppositely, homoploid hybrid speciation is much less frequent than polyploid hybrid speciation because homoploid hybrids use to have reduced fertility in early generations, and have a high chance for introgression into parental taxa (e.g., Stebbins, 1950; Grant, 1981;

Soltis & Soltis, 2009). In fact, there are only about 20 well reported cases of homoploid hybrid speciation (Gross & Rieseberg, 2005). However, recent molecular techinques promise the discovery of many more examples (Hegarty & Hiscock, 2005). Usually, those species can be morphologically more difficult to discriminate from parental taxa than polyploid hybrid species (Rieseberg & Willis, 2007). Although, contrasting with polyploid species, the chance for homoploid hybrid speciation seems to decrease as phylogenetic distance between parental species increases (Buggs et al., 2008). Rapid karyotypic evolution, ecological divergence and spacial isolation, are factors favouring homoploid hybrid isolation from their parental taxa (Rieseberg & Willis, 2007). Therefore, colonizing a new habitat would favour both ecological and spatial isolation from the parental species, avoiding competence and introgression. However, in many cases such hybrid species occupy a habitat close to that of at least one of the parental species, commonly in intermediate habitats (ecotones), but also extreme habitats (Rieseberg & Willis, 2007).

Hence, in contrast to auto- or allopolyploid speciation, the challenge in homoploid hybrids to speciate is to scape the homogenizing effects of gene flow from parental taxa

(Buerkle et al., 2000). Current studies suggest that ecological selection is a major factor promoting homoploid hybrid speciation (Gross & Rieseberg, 2005), increasing in consonancy with the strength of ecological selection and novel habitats availability (Buerkle et al., 2000, 2003; Coyne & Orr, 2004). Moreover, parallel genotypic adaptation has also been demonstrated to be feasible, giving rise to the same adaptations (i.e., taxa) in independent homoploid hybrid-derivated populations especially by ecological selection

(Gross & Rieseberg, 2005; Wood et al., 2005).

35 Chapter 1. General Introduction Natural hybrids and plant evolution

1.2.4. Abundance and distribution of natural hybrids

Reports of the existence of natural hybrids come mainly from observation of morphological intermediacy, even in fossil records (Arnold, 1997). Also, karyological approaches and artificial production of such crosses have been used to verify some hybrid status. Yet, current molecular techniques allow the detection of hybrids and introgressants even if they do not show morphological intermediacy, also allowing to discern about the possible hybrid origin of the species (e.g., Hegarty & Hiscock, 2005). There are many studies reporting phylogenetic incongruences in some clades that could be explained by hybridization (Avise, 2000; Linder & Rieseberg, 2004). On the other hand, one must expect that many interspecific crosses are probably occurring, though not detected due to their low or null importance in plant evolution (Arnold, 1997).

Frequency of hybridization is expected to be defined by two main aspects: the territory and the plant group studied. With regards to the territory, higher habitat heterogeneity is expected to increase the chance for interspecific crosses and further hybrid establishment, since the ease in finding related species in sympatry and the existence of many zones of habitat intermediacy suitable for hybrid success (e.g., Anderson, 1948).

Also, human activity could increase such frequency either by disturbing natural habitats or by favouring contact among allopatric species through species introductions (e.g., Excoffier et al., 2009, and references therein).

On the other hand, some families or genera are expected to show higher interspecific cross frequencies than others merely due to having a higher number of species. Also, particular diversification histories could have favoured the arising of mechanisms to avoid interspecific crosses in specific groups rather than others, thus having less chance for hybridization. However, a higher frequency of studies based on a specific

36 Chapter 1. General Introduction Natural hybrids and plant evolution group could also biass the report of hybrids in such group, merely due to scientists interest

(e.g., Arnold, 1997).

There are some attempts to summarize or quantify all known plant hybrids, even for a specific region, such the list of Cockayne for the New Zeland flora (Cockayne, 1923); and the reviews of Ellstrand et al. (1996) and Mallet (2005). From Ellstrand et al. (1996) review, which includes the floras of British Isles, Scandinavia, USA Great Plains, USA

Intermountain West, and Hawaii Islands, the authors state an average of 11% for hybrid taxa. This implies that, considering 250,000 described plant species, an approximated number of 27,500 hybrid species would exist around the world (Ellstrand et al., 1996). This percentage ranges from the 22% of the British flora to 5.8% of the intermountains flora, being the former probably more intensely explored, and thus more complete than the latter

(Ellstrand et al., 1996). It must be noted that the above 11% average of hybrids among some major floras given by Ellstrand and colleagues (Ellstrand et al., 1996) is referred to relatively recent-origin hybrids, that is, hybrid species yet recognizable as such rather than species with a hybrid origin. In this case, some evidences indicate that many of the angiosperms could have a hybrid origin, since estimations suggest that approximately 70% of all flowering plants owe their existence to past natural hybridization between different species or genera (Grant, 1981; Whitham et al., 1991; Soltis & Soltis, 2009).

In parallel, some studies attempted to recognize families or genera with higher frequency than others, and even biological forms or pollination syndroms. Thus, higher frequency of hybrids in dicots vs. monocots, in animal-pollinated vs. wind-pollinated, and in perennial vs. annual plants have been suggested (e.g., Stace, 1975; Arnold, 1997). Thus, among flowering plant families, only an average of 16-34% of plant families and 6-16% of genera have one or more reported hybrid (Ellstrand et al., 1996). Therefore, hybridization appears to be concentrated in a small fraction of genera. In this respect, 5-21% of genera

37 Chapter 1. General Introduction Natural hybrids and plant evolution with hybrids accounted for more than half of a flora’s reported hybrids (Ellstrand et al.,

1996).

Hence, the frequency of natural hybridization can be subtantially underestimated due to the lack of studies in several regions or in segeral plant groups, to the discordance in the proportion of reported hybrids among different floras, to the ability in recognizing hybrids as such, or to the evolutionary impact of hybrids in the evolution of each plant group (e.g.,

Mayr, 1992; Rieseberg, 1997; Mallet, 2005).

1.2.5. Intermediacy: other options to hybridization

Intermediate morphology does not necessarily support the hypothesis of hybridity

(Rieseberg & Ellstrand, 1993). Molecular markers provide more reliable evidence for hybrid parentage than morphology alone (Rieseberg & Ellstrand, 1993; Ellstrand & Schierenbeck,

2000), allowing the demonstration that either genetic and morphological intermediacy can arise from several other evolutive processes but hybridization. Therefore, in the study of hybridization, one must be aware of other possible explanations to the observed variability.

Phylogenetic methods proved to be resolutive enough in such cases as they determined the relationship between these populations. Thus, the closer the relation between taxa or populations involved, the greater the chance for primary intergradation rather than hybridization (Rieseberg & Ellstrand, 1993).

1.2.5.1. Primary intergradation of populations

Genetic and morphological intermediacy can arise from crosses among partially diverged populations, resulting in recombination of such primarily diverged traits. This is referres in literature as primary intergradation, while hybridization and introgression are referred to as secondary intergradation (e.g., Rieseberg & Ellstrand, 1993). Distinction

38 Chapter 1. General Introduction Natural hybrids and plant evolution between cases of primary and secondary intergradation can be hard to achieve. Indeed, population intergradation is often observed in cases of divergence after hybridization events. Nevertheless, note that, depending on the definition of hybridization considered, the difference between primary and secondary intergadation can be confused.

As an example od primary intergadation, Jordan et al. (1993) found high and continuous variation ranges in capsule traits in the Eucalyptus globulus complex, concluding that high levels of intergradation among populations was leading to such scenario. Indeed, putative hybridization and speciation events were also hypothesized. Recently, Brown et al.,

(2009) showed that in the wide group Calystegia there was a high degree of geographical intergradation and morphological variation in Britain. Yet, hybridization was also occurring and leading to speciation.

1.2.5.2. Incomplete lineage sorting

In another scale, low divergence among populations, and even species, can arise from an incomplete lineage sorting situation. Poorly diverged species (both genetically and morphologically) are also expected to show intermediate and/or non-fixed traits. This can be especially important in genera where species are long-lived and show large population sizes, like Pinus. Willyard et al., (2009) demonstrated the existence of that situation in the ponderosa pines complex. Plastic, homoplasious, very variable morphology, and high inter- and intrapopulation genetic variability is occurring, blurring species discrimination and depicting a hybridization-like scenario. Moreover, long-term retention of ancestral polymorphisms has also been detected in that Pinus complex, probably enhanced by occasional interspecific hybridization allowing allele sharing among species (Willyard et al.,

2009). Thus, in this example reticulate evolution was also detected, entangling the

39 Chapter 1. General Introduction Natural hybrids and plant evolution relationship among species and preventing from finding a definitive resolution among the species’ complex.

1.2.5.3. Rapid divergence and diversification in different habitats

Senecio leucanthemifolius var. casablancae shows a conspicuous leaf morphology variation along its distribution in the western Moroccan coast, especially in the most isolated populations, where in turn inhabits the close relative S. glaucus subsp. coronopifolius.

Hybridization events involving both taxa in their zone of contact seemed a plausible explanation. However, Coleman & Abbot (2003) demonstrated that recent divergence rather than hybridization was responsible for the observed variation. The lack of gene flow among populations, and local adaptations favoured by the existence of different habitats, was promoting the arising of local geographic races. Nevertheless, ancient hybridization processes and incomplete lineage sorting were also inferred from the observed frequences in different cpDNA haplotypes.

1.2.5.4. Random lineage sorting effects

Several more processes are expected to produce inaccuracy in phylogenetic trees.

Either genetic and morphological intermediacy in two taxa or populations can be the result of (random) lineage sorting effects. Thus, trait variation observed into several populations could be actually intrinsic shared variation rather than variation due to primary or secondary intergradation. In this scenario, both divergent taxa or populations would retain a degree of variation already found in their common ancestor, without any degree of gene exchange.

Some examples can be found in literature. For example, in the European Quercus petrea and Q. robur, extensive hybridization has been considered to occur. However, Muir &

Schlötterer (2005) showed that gene flow should be much smaller than anticipated, allowing

40 Chapter 1. General Introduction Natural hybrids and plant evolution the maintenance of separated species observed in most loci. They invoked genetic drift operating against the homogenizing effects of gene flow to explain such differences instead of hybridization. However, their results did not detect hybridization, rather they demostrated shared ancestral variation (i.e., lineage sorting effects) in such species. Indeed, they detected a clear phylogeographic pattern of cpDNA haplotypes, also favouring the low effect of gene flow in this case.

Also, in the Mediterranean Senecio sect. Senecio, Comes & Abbott (2001) found evidences for the existence of ancestral cpDNA polymorphisms maintained for up to one million years. There was a lack of phylogeographic structure in the studied haplotypes, which confounded phylogenetic trees. Thus, hybridization and introgression could be invoked. However, their results also showed that reticulation and incomplete lineage sorting were feasible to occur in some degree into this species complex.

1.2.5.5. Biased concerted evolution

Several molecular markers can show unexpected patterns across populations due to factors othrer than hybridization alone. In the nrDNA markers, like the internal transcribed spacer (ITS) and the external transcribed spacer (ETS), concerted evolution phenomena are expected to produce sequence homogenization in relatively few generations (e.g.,

Fuertes-Aguilar et al., 1999; Álvarez & Wendel, 2003). This is feasible either in intragenomic polymorphisms due to hybridization, or due to random lineage sorting effects. On the contrary, the lack of concerted evolution would lead to the unexpected maintenance of such polymorphisms after hybrid speciation events. Also, ITS and ETS sequences, though occurring in the same tandem ribosomal unit repeat, can show very different patterns of concerted evolution (Okuyama et al., 2005).

41 Chapter 1. General Introduction Natural hybrids and plant evolution

In Armeria, Nieto-Feliner et al. (2001) noted that the similarity in ITS sequences detected among populations responding to their geographical origin was better explained by biased concerted evolution rather than hybridization.

1.2.5.6. Parallel genotypic adaptation

Convergence, and especially parallel genotypic adaptation (i.e., homoplasy), could also lead to two related taxa or populations showing a similar degree of variation (Wood et al., 2005). In this case (i.e., parallelism), shared variation would be the result of a common genetic and morphological response to common habitat features (through natural selection), rather than primary or secondary intergradation. Although, in very selective habitats, genotypic and phenotypic similarities could be displayed in species not only without hybridization processes involved, but also showing no common ancestry (e.g., succulent spiny stems in the Euphorbiaceae and Cactaceae; Wood et al., 2005).

In the Hawaiian violets, Havran et al. (2009) demostrated that leaf morphology responded to parallel evolution through habitat selection. Thus, molecular markers showed that analogous taxa reached the same ecological adaptations in different islands, resulting in dry and wet-adapted paraphiletic phenotypes. Yet, complex dispersal events and suspected hybridization in specific taxa were also bluring clear phylogeographical patterns.

42 Chapter 1. General Introduction Hybridization and conservation

1.3. HYBRIDIZATION AND CONSERVATION

Conservation of endemic taxa, due to their fragility and to their intrinsic narrow distribution range, is a major goal for the conservation of the World’s biodiversity. Thus, the assessment of the causes of species decline and extinction is crucial. Nevertheless, factors causing demise of populations can be gradual and thus not obvious on a short time scale.

An evaluation of the likelihood of plant species extinction usually involves the estimation of several demographic (number and size of populations, rates of demographic fluctuation) and ecological parameters (conservation of pollinators and plant-animal mutualisms) that could be involved in threats concerning rare plants. Moreover, interspecific gene flow can have severe genetic consequences on small populations of hybridizing taxa.

As seen above, natural hybridization is recognized as a key force in plant (and animal) evolution (e.g., Arnold, 1994; Rieseberg, 1995; Chapman & Burke, 2007; Soltis &

Soltis, 2009). However, it can be either a source of variation and novelty, contributing to the success and diversification of many plants lineages (Arnold, 2006); or otherwise can lead to introgressive events diluting novelties and adaptations gathered through divergence and isolation (Rieseberg, 1997). Therefore, in contrast with its creative role in plant evolution, hybridization may contribute to the demise of rare species through demographic swamping and genetic assimilation by widespread congeners (Levin et al., 1996; Wolf et al., 2001;

Buerkle et al., 2003).

Asymmetrical introgression through unidirectional backcrossing of partially fertile hybrids to the more abundant parent can lead to the loss of parental nuclear genotypes and can ultimately lead to local extinction of the rare species where the direction of gene flow is into the rarer species (Milne et al., 2003; Burgess et al., 2005). This can be particularly

43 Chapter 1. General Introduction Hybridization and conservation relevant on islands that are typically rich in endemic species that show ecological diversification (Brochmann, 1984; Levin, 2000).

The Mediterranean basin is recognized as one of the hot spots of speciation, and islands constitute major centers of endemicity in the Mediterranean basin (Médail & Quezel,

1999). Also, the existence of strong geographic barriers to dispersal heavily shape species distribution and divergence. Due to their high rates of habitat variation and ecogeographic heterogeneity, islands served as a refuge for several taxa during the glaciations in the

Quaternary (Hewitt, 2004; Thompson, 2005). Indeed, new lineages arose in this period through local adaptations and isolation (Hampe & Petit, 2005).

On the other hand, this habitat richness, added to anthropogenic disturbance, could ease the establishment of introduced species suitable to hybridize with local endemics.

Therefore, endemic species could be threatened by an introduced congeneric species in sympatry. This can be especially important if the latter has a population numerically larger and/or is reproductively more effective then the endemic species (Levin et al., 1996;

Ellstrand et al., 1999). Moreover, some studies also favour hybridization processes as promoters of invasiveness in plants (e.g., Ellstrand, 2003).

There are many examples in literature demonstrating the effects of hybridization of local taxa with more widespread congeners. The endangered North American Morus rubra was threatened by the higher fitness of the introduced M. alba and their hybrids. In fact, M. rubra proved to be less fit in their habitat than both M. alba and their hybrids (Burgess &

Husband, 2006). In the Alps, the native Taraxacum ceratophorum was shown to be threatened by genetic swamping by the invasive T. officinale (Brock, 2004). Butcher et al.

(2005) showed that Eucalyptus benthamii, endemic from an eastern Australia region, was threatened by introgression with other congeners. Indeed, this was facilitated by competition with introduced species and habitat disturbance (promoting higher fitness for the non-

44 Chapter 1. General Introduction Hybridization and conservation natives). This was a very similar case to the Tasmanian endemic Eucalyptus cordata

(McKinnon et al., 2004). Yet, in this latter case the gene flow was from the rare form into the more widespread species. Fossati et al. (2003) also demonstrated that the introgression of cultivated Populus clones into a natural population of P. nigra in northern Italy was not threatening the local population.

In other instances, early generation hybrids can even overcome both parental species in the intermediate habitat. In Turkey, Milne et al. (2003) described a hybrid zone

dominated by the F1 plants of Rhododendron. Those plants were more competitive than

further generation hybrids (i.e., F2,...), which prevented introgression. In this case, hybridization would not be a threat for the parental species in pristine habitats. However, anthropogenic disturbance was increasing the potential habitat for such hybrids and thus, promoting the demise of the parental species.

There are also several examples of island endemic plants threatened by hybridization with a congeneric species, many times involving habitat alteration. Rieseberg et al. (1989) demonstrated that Cercocarpus traskiae, endemic from Santa Catalina Island, was at the edge of extinction due to the assimilation by a the widespread C. betuloides var. blancheae. The latter species introduction was favoured by the introduction of large herbivores in the island. Morus boninensis, endemic to the Bonin Islands (Japan), was found to hybridize with the widespread congener M. acidosa (Tani et al., 2003). Both hybrid and widespread plants were threatening the native species survivirship. This is of particular importance in oceanic islands, such Hawaii. There, one of the genus showing higher radiation and also high rates of introduction of continental congeners is Rubus. For example, the endemic Rubus hawaiensis and the naturalized R. rosifolius were found to hybridize in on the island of Maui. This resulted in a conspicuous threat for the long term viability of endemic species (Randell et al., 2004).

45 Chapter 1. General Introduction Hybridization and conservation

Natural hybridization is an spontaneous event largely occurred in plant evolution, allowing faster rates of genotype recombinations and the arising of novel adaptations (e.g.,

Soltis & Soltis, 2009). Oppositely, it can also dilute the novel adaptations and lineages arose through isolation and divergence (e.g., Buerkle et al., 2003), allowing natural selection to test for the most fit genotypes in each habitat. Nevertheless, man activities are a major factor in habitat disturbance, promoting both habitat change and species translocation. Anthropogenic factors worldwide promote either rapid changes in natural habitats favouring new habitat arising, and the occurrence of normally allopatric species through the disruption of ecological and geographical barriers (Rhymer & Simberloff, 1996;

Seehausen et al., 2008). This creates many opportunities for interspecific gene flow, and the arising of suitable habitats for the survival of hybrids (Arnold, 1997). Therefore, the alteration of the pristine and unique habitats where endemic species use to evolve would heavily promote the demise of such endemisms by more widespread species, either through competition, demographic swamping, or genetic assimilation by their widespread congeners (Levin et al., 1996; Wolf et al., 2001; Buerkle et al., 2003).

46 Chapter 1. General Introduction References

1.4. REFERENCES

Abbott RJ. 1992. Plant invasions, hybridization and the evolution of new plant taxa. Trends in Ecology and

Evolution, 7: 401–405.

Álvarez I & Wendel JF. 2003. Ribosomal ITS sequences and plant phylogenetic inference. Molecular

Phylogenetics and Evolution, 29: 417–434.

Anderson E. 1948. Hybridization of the habitat. Evolution, 2: 1-9.

Anderson E. 1949. Introgressive hybridization. John Wiley and Sons: New York.

Anderson T & Hubritch L. 1938. Hybridization in Tradescantia. III. The evidence for introgressive hybridization.

American Journal of Botany, 25: 396-402.

Anderson T & Stebbins GL, Jr. 1954. Hybridization as an evolutionary stimulus. Evolution, 8: 378-488.

Antonovics J. 1976. The nature of limits to natural selection. Annals of the Missouri Botanical Garden, 63: 224-

247.

Arnold ML. 1993. Rarity of hybrid formation and introgression in Louisiana irises. Plant Genetics Newstletter, 9:

14-17.

Arnold ML. 1994. Natural hybridization and Louisiana irises. Bioscience, 44: 141-147.

Arnold ML. 1997. Natural Hybridization and Evolution. Oxford University Press: New York.

Arnold ML, Buckner CM & Robinson JJ. 1991. Pollen mediated introgression and hybrid speciation in Louisiana

irises. Proceedings of the National Academy of Sciences of the USA, 88: 1398-1402.

Arnold ML. 2006. Evolution through genetic exchange. Oxford University Press: New York.

Arnold ML, Hamrick JL & Bennett BD. 1990. Allozyme variation in Louisiana irises: a test for introgression and

hybrid speciation. Heredity, 65: 297-306.

Arnold ML, Bulger MR, Burke JM, Hempel AL & Williams JH. 1999. Natural hybridization: How low can you go

and still be important? Ecology, 80: 371-381.

Avise JC. 2000. Phylogeography: the history and formation of species. Harvard University Press: Cambridge.

Barton NH. 1979. The dynamics of hybrid zones. Heredity, 43: 341-359.

Bigelow RS. 1965. Hybrid zones and reproductive isolation. Evolution, 19: 449-458.

Britannica Online. 2007. Encyclopedia Britannica. Encyclopedia Britannica Inc.: Chicago. April 2007.

47 Chapter 1. General Introduction References

Brochmann C. 1984. Hybridization and distribution of Argyranthemum coronopifolium (Asteraceae-Anthemidae)

in the Canary Islands. Nordic Journal of Botany, 4: 729-736.

Brock MT. 2004. The potential for genetic assimilation of a native dandelion species, Taraxacum

ceratomorphum (Asteraceae), by the exotic congener T. officinale. American Journal of Botany, 91:

656-663.

Brown JM, Brummitt RK, Spencer M & Carine MA. 2009. Disentangling the bindweeds: hybridization and

taxonomic diversity in British Calystegia (Convolvulaceae). Botanical Journal of Botany, 160: 388-401.

Butcher PA, Skinner AK & Gardiner CA. 2005. Increased inbreeding and inter-species gene flow in remnant

populations of the rare Eucalyptus benthamii. Conservation Genetics, 6: 213-226.

Buerkle CA, Wolf DE, Rieseberg LH. 2003. The origin and extinction of species through hybridization. In:

Brigham CA, Schwartz MW (eds). Population viability in plants. Springer Verlag: Berlin.

Buggs RJA, Soltis PS, Mavrodiev EV, Symonds VV, Soltis DE. 2008. Does phylogenetic distance between

parental genomes govern the success of polyploids? Castanea, 73: 74–93

Burgess KS, Morgan M, DeVerno L & Husband BC. 2005. Asymmetrical introgression between two Morus

species (M. alba, M. rubra) that differ in abundance. Molecular Ecology, 14: 3471–3483.

Burgess KS & Husband BC. 2006. Habitat differentiation and the ecological costs of hybridization: the effects of

introduced mulberry (Morus alba) on native congener (M. rubra). Journal of Ecology, 94: 1061-1069.

Burke JM & Arnold ML. 2001. Genetics and the fitness of hybrids. Annual Review of Genetics, 35: 31-52.

Burke JM, Carney SE & Arnold ML. 1998. Hybrid fitness in the Louisiana irises: Analysis of parental and F1

performance. Evolution, 52: 37-43.

Burke JM, Bulger MR, Wesselingh RA & Arnold ML. 2000. Frequency and spatial patterning of clonal

reproduction in Louisiana iris hybrid populations. Evolution, 54: 137–144.

Carney SE, Hodges SA & Arnold ML. 1996. Effects of pollen-tube growth and ovule position on hybridization in

the Louisiana irises. Evolution, 50: 1871-1878.

Chapman MA & Burke JM. 2007. Genetic divergence and hybrid speciation. Evolution, 61: 1773-1780.

Cockayne L. 1923. Hybridism in the New Zeland Flora. New Phytologist, 22: 105-127.

Coleman M & Abbott RJ. 2003. Possible causes of morphological variation in an endemic Moroccan groundsel

(Senecio leucanthemifolius var. casablancae): evidence from chloroplast DNA and random amplified

polymorphic DNA. Molecular Ecology, 12: 423-434.

48 Chapter 1. General Introduction References

Comes HP & Abbott RJ. 2001. Molecular phylogeography, reticulation, and lineage sorting in Mediterranean

Senecio sect. Senecio (Asteraceae). Evolution, 55: 943–62

Coyne JA & Orr HA. 2004. Speciation. Sinauer & Associates: Sunderland.

Cruzan MB. 2005. Patterns of introgression across an expanding hybrid zone: analysing historical patterns of

gene flow using nonequilibrium approaches. New Phytologist, 167: 267-278.

Cruzan MB & Arnold ML. 1999. Consequences of cytonuclear epistasis and assortative mating for the genetic

structure of hybrid populations. Heredity, 82: 36-45.

Darwin C. 1859. On the origin of species by means of natural selection, or the preservation of favoured races in

the struggle for life. John Murray: London.

Devey DS, Bateman RM, Fay MF & Hawkins JA. 2008. Friends or relatives? Phylogenetics and species

deminitation in the controversial European orchid genus Ophrys. Annals of Botany ,101: 385-402.

Dobzhansky T. 1937. Genetics and the origin of species. Columbia University Press: New York.

Dowling TE & DeMarais BD. 1993. Evolutionary significance of introgressive hybridization in cyprinid fishes.

Nature, 362: 444–446.

Doyle JJ, Flagel LE, Paterson AH, Rapp RA, Soltis DE, Soltis PS & Wendel JF. 2008. Evolutionary genetics of

genome merger and doubling in plants. Annual Review of Genetics, 42: 443-461.

Ellstrand NC & Elam DR. 1993. Population genetic consequences of small population size: implications for plant

conservation. Annual Review in Ecology and Systematics, 24, 217–242.

Ellstrand NC, Prentice HC, Hancock JF. 1999. Gene flow and introgression from domesticated plants into their

wild relatives. Annual Review in Ecology and Systematics, 30: 539-563.

Ellstrand NC, Whitkus R & Riesberg LH. 1996. Distribution of spontaneous plant hybrids. Proceedings of the

National Academy of Sciences of the USA, 93: 5090-5093.

Ellstrand NC & Schierenbeck KA. 2000. Hybridization as a stimulus for the evolution of invasiveness in plants.

Proceedings of the National Academy of Sciences of the USA, 97: 7043–7050.

Ellstrand NC. 2003. Dangeous Liaisons? When cultivated plants mate with their wild relatives. John Hopkins

University Press: Baltimore.

Emms SK, Hodges SA & Arnold ML. 1996. Pollen-tube competition, siring success, and consistent asymmetric

hybridization in Louisiana irises. Evolution, 50: 2201-2206.

Endler JA. 1977. Geographic variation, speciation, and clines. Princeton University Press: Princeton.

49 Chapter 1. General Introduction References

Excoffier L, Foll M & Petit RJ. 2009. Genetic consequences of range expansions. Annual Review in Ecology,

Evolution and Systematics, 40: 481-501.

Fuertes-Aguilar J, Rosselló JA & Nieto-Feliner G. 1999. Nuclear ribosomal DNA (nrDNA) concerted evolution in

natural and artificial hybrids of Armeria (Plumbaginaceae). Molecular Ecology, 8: 1341-1346.

Grant PR, Grant BR, Markert JA, Keller LF & Petren K. 2004. Convergent evolution of Darwin's finches caused

by introgressive hybridization and selection. Evolution, 58: 1588-1599.

Grant V. 1953. The role of hybridization in the evolution of the leafty-stemmed Gilias. Evolution, 7: 51-64.

Grant V. 1963. The Origin of Adaptations. Columbia University Press: New York.

Grant V. 1981. Plant speciation. Columbia University Press: New York.

Grant PR. 1993. Hybridization of Darwin’s finches on Isla Daphne Major, Galapagos. Philosophycal

Transactions of the Royal Society of London, Series B, Biology, 340: 127-139.

Grant PR & Grant BR. 1992. Hybridization of bird species. Science, 256: 193-197.

Grant PR, Grant BR & Petren K. 2005. Hybridization in the recent past. The American Naturalist 166: 56-67.

Gross BL & Rieseberg LH. 2005. The ecological genetics of homoploid hybrid speciation. Journal of Heredity,

96: 241-252.

Hampe A & Petit RJ. 2005. Conserving biodiversity under climate change: the rear edge matters. Ecology

Letters, 8: 461-467.

Hardig TM, Brunsfeld SJ, Fritz RS, Morgan M & Orians CM. 2000. Morphological and molecular evidence for

hybridization and introgression in a willow (Salix) hybrid zone. Molecular Ecology, 9: 9-24.

Hegarty MJ & Hiscock SJ. 2005. Hybrid speciation in plants: new insights from molecular studies. New

Phytologist, 165: 411–423.

Harper D. 2001. Online etymology dictionary. < http://www.etymonline.com/>

Harrison RG. 1986. Pattern and process in a narrow hybrid zone. Heredity, 56: 337-349.

Harrison RG. 1990. Hybrid zones: windows on evolutionary process. Oxford Surveys in Evolutionary Biology 7,

69-128.

Harrison RG. 1993. Hybrids and hybrid zones: historical perspective. In: Harrison RG. (ed). Hybrid zones and

the evolutionary process. Oxford University Press: Oxford.

Havran JC, Sytsma KJ & Ballard HEJr. 2009. Evolutionary relationships, interisland biogeography and molecular

evolution in the Hawaiian violets (Viola: Violaceae). American Journal of Botany, 96: 2087-2099.

Heiser CB, Jr. 1973. Introgression re-examined. Botanical Review, 39: 347-366.

50 Chapter 1. General Introduction References

Herbert W. 1847. On hybridization amongst vegetables. Journal of the Horticultural Society of London, 2: 1-107.

Hewitt GM. 2004. Genetic consequences of climatic oscillations in the Quaternary. Phylosophical Transactions

of the Royal Society of London, series B Biological Sciences, 359: 183-195.

Howard DJ. 1982. Speciation and coexistence in a group of closely related ground crickets. PhD dissertation:

Yale University.

Howard DJ. 1986. A zone of overlap and hybridization between two ground cricket species. Evolution, 40: 34-43.

Huxley J. 1942. Evolution: the modern synthesis. Harper and Brothers: New York.

James JK & Abbott RJ. 2005. Recent, allopatric, homoploid hybrid speciation: the origin of Senecio squalidus

(Asteraceae) in the British Isles from a hybrid zone on Mount Etna, Sicily. Evolution, 59: 2533–2547.

Jordan GJ, Potts BM, Kirkpatrick JB & Gardiner C. 1993 . Variation in the Eucalyptus globulus complex revisited.

Australian Journal of Botany, 41: 763–785.

Key KHL. 1968. The concept of stasipatric speciation. Systematic Zoology, 17: 14-22.

Lai Z, Gross BL, Zou Y, Andrews J & Rieseberg LH. 2006. Microarray analysis reveals differential gene

expression in hybrid sunflower species. Molecular Ecology, 15: 1213–1227.

Levin DA, Francisco-Ortega J & Jansen RK. 1996. Hybridization and the extinction of rare plant species.

Conservation Biology, 10: 10-16.

Levin DA. 2000. The origin, expansion, and demise of plant species. Oxford University Press: Oxford.

Lewontin RC & Birch LC 1966. Hybridization as a source of variation for adaptation to new environments.

Evolution, 20: 315–336.

Linder CR & Rieseberg LH. 2004. Reconstructing patterns of reticulate evolution in plants. American Journal of

Botany, 91: 1700-1708.

Linné C. 1753. Species plantarum. Printed for the Ray Society: London.

Linné C. 1760. Disquisitio de sexu plantarum, ab Academia Imperiali Scientiarium Petropolitana praemio ornata.

Amoenitales Academicae, 10: 100-131.

Lotsy JP. 1916. Evolution by means of hybridization. Nijhoff M: The Hague.

Lotsy JP, 1931. On the species of the taxonomist in its relation to evolution. Genetica, 13: 1-16.

Lowry DB, Modliszewski JL, Wright KM, Wu CA & Willis JH. 2008 The strength and genetic basis of reproductive

isolating barriers in flowering plants. Philosophical Transactions of the Royal Society of London Series

B, Biological Sciences, 363: 3009–3022.

Lowe PR. 1936. The finches of the Galápagos in relation to Darwin’s conception of species. Ibis, 6: 310–321.

51 Chapter 1. General Introduction References

Mable BK. 2004. ‘Why polyploidy is rarer in animals than in plants’: myths and mechanisms. Biological Journal

of the Linnean Society, 82: 453–466.

Mallet J. 2005. Hybridization as an invasion of the genome. Trends in Ecology & Evolution, 20: 229-237.

Mayr E. 1942. Systematics and the origin of species. Columbia University Press: New York.

Mayr E. 1992. A local flora and the biological species concept. American Journal of Botany, 79: 222-238.

McKinnon GE, Vaillancourt RE, Steane DA & Potts BM. 2004. The rare silver gum, Eucalyptus cordata, is

leaving its race in the organellar gene pool of Eucalyptus globosus. Molecular Ecology, 13: 3751-3762.

Médail, F. & Quézel, P. 1999. Biodiversity hotspots in the Mediterranean Basin: setting global conservation

priorities. Conservation Biology, 13: 1510–1513.

Mendel G. 1866. Experiments on plant hybrids. English translation. In: Stern C, Sherwood ER (eds). The Origin

of Genetics: A Mendel Source Book. Freeman WH: San Francisco.

Merriam-Webster Online. 2007. Merriam-Webster’s Collegiate Dictionary. April 2007.

webster.com/>

Milne RI, Terzioglu S & Abbott RJ. 2003. A hybrid zone dominated by fertile F1s: maintenance of species

barriers in Rhododendron. Molecular Ecology, 12: 2719-2729.

Miller AH. 1949. Some concepts of hybridization and intergradation in wild populations of birds. Auk, 66: 388-

342.

Moore WS. 1977. An evaluation of narrow hybrid zones in vertebrates. Quarterly Review of Biology, 52: 263-

277.

Morjan CL & Rieseberg LH. 2004. How species evolve collectively: implications of gene flow and selection for

the spread of advantageous alleles. Molecular Ecology, 13: 1341–1356.

Muir G & Schlötterer C. 2005. Evidence for shared ancestral polymorphism rather than recurrent gene flow at

microsatellite loci differentiating two hybridizing oaks (Quercus spp.). Molecular Ecology, 14: 549–561.

Mullen SP, Dopman EB & Harrison RG. 2008. Hybrid zone origins, species boundaries, and the evolution of

wing-pattern diversity in polytypic species complex of North American asmiral butterflies (Nymphalidae:

Limenitis). Evolution, 62: 1400-1417.

Nasrallah ME, Yogeeswaran K, Snyder S & Nasrallah JB. 2000. Arabidopsis species hybrids in the study of

species differences and evolution of amphiploidy in plants. Plant Physiology, 124: 1605–1614.

Naudin C. 1863. Nouvelles recherches sur l'hybridité dans les végétaux. Annales des Sciences Naturelles,

Botanique et Biologie Vegetale, 19: 180-203.

52 Chapter 1. General Introduction References

Nieto-Feliner G, Fuertes-Aguilar J, Rosselló JA. 2001. Can extensive reticulation and concerted evolution result

in a cladistically structured molecular data set? Cladistics, 17: 301–312.

OED Online. 2007. The Oxford English Dictionary. Oxford University Press. April 2007.

Okuyama Y, Fujii N, Wakabayashi M, Kawakita A, Ito M, Watanabe M, Murakami N & Kato M .2005. Nonuniform

concerted evolution and chloroplast capture: heterogeneity of observed introgression patterns in three

molecular data partition phylogenies of Asian Mitella (Saxifragaceae). Molecular Biology and Evolution,

22: 285–296.

Paun O, Greilhuber J, Temsch EM & Hörandl E. 2006. Patterns, sources and ecological implications of clonal

diversity in apomictic Ranunculus carpaticola. Ranunculus auricomus complex (Ranunculaceae).

Molecular Ecology, 15: 897–910.

Peterson A, Harpke D, Peruzzi L, Levichev IG, Tison JM & Peterson J. 2009. Hybridization drives speciation in

Gagea (Liliaceae). Plant Systematics and Evolution, 278: 133-148.

Randell RA, Howarth DG & Morden CW. 2004. Genetic analysis of the natural hybrids between endemic and

allien Rubus (Rosaceae) species in Hawai’i. Conservation Genetics, 5: 217-230.

Rhymer JM & Simberloff D. 1996. Extinction by hybridization and introgression. Annual Review in Ecology and

Systematics, 27: 83-109.

Rieseberg LH. 1995. The role of hybridization in evolution: old wine in new skins. American Journal of Botany,

82, 994-953.

Rieseberg LH. 1997. Hybrid origins of plant species. Annual Review in Ecology and Systematics, 28. 359-389.

Rieseberg LH, Kim MJ & Seiler GJ. 1999. Introgression between the cultivated sunflower and a sympatric wild

relative, Helianthus petiolaris (Asteraceae). International Journal of Plant Sciences, 160: 102-108.

Rieseberg LH, Sinervo B, Linder CR, Ungerer MC & Arias DM. 1996. Role of gene interactions in hybrid

speciation: evidence from ancient and experimental hybrids. Science, New Series, 272: 741-745.

Rieseberg LH. 1991. Homoploid reticulate evolution in Helianthus (Asteraceae): evidence from ribosomal genes.

American Journal of Botany, 78: 1218-1237.

Rieseberg LH & Willis JH. 2007. Plant speciation. Science, 317: 910–14

Rieseberg LH & Ellstrand NC. 1993. What can molecular and morphological markers tell us about hybridization?

Critical Reviews in Plant Sciences, 12: 213-241.

Roberts HF. 1929. Plant Hybridization before Mendel. Princeton Univeristy Press: Princeton.

53 Chapter 1. General Introduction References

Seehausen O, Takimoto G, Roy D & Jokela J. 2008. Speciation reversal and biodiversity dynamics with

hybridization in changing environments. Molecular Ecology, 17: 30-44.

Short LL. 1969. Taxonomic aspects of avian hybridization. Auk, 86: 84-105.

Simberloff D. 1988. The contribution of population and community biology to conservation science. Annual

Review in Ecology and Systematics, 19: 473-551.

Simpson J & Weiner E (eds). 1989. Oxford English Dictionary. Second Edition. Clarendon Press.

Smith HM. 1955. The perspective of species. Turtox News, 33: 74-77.

Smith HM. 1965. More evolutionary terms. Systematic Zoology, 14: 57-58.

Soltis PS & Soltis DE. 2009. The role of hybridization in plant speciation. Annual Review of Plant Biology, 60:

561-588.

Soltis DE & Soltis PS. 1999. Polyploidy: Recurrent formation and genome evolution. Trends in Ecology and

Evolution 14: 348–352.

Soltis DE, Soltis PS, Pires JC, Kovarik A, Tate JA & Mavrodiev E. 2004. Recent and recurrent polyploidy in

Tragopogon (Asteraceae): cytogenetic, genomic and genetic comparisons. Biological Journal of the

Linnean Society, 82: 485–501.

Soltis DE, Mavrodiev EV, Doyle JJ, Rauscher J & Soltis PS. 2008. ITS and ETS sequence data and phylogeny

reconstruction in allopolyploids and hybrids. Systematic Botany, 33: 7–20.

Stace CA 1975. Hybridization and the flora of the British Isles. Academic Press: London.

Stebbins GL. 1950. Variation and evolution in plants. Columbia University Press: New York.

Tanin N, Kawahara T, Yoshimaru H & Hoshu Y. 2003. Development of SCAR markers distinguishing pure

seedlings of the endangered species Morus boninensis from M. boninensis x M. acidosa hybrids for the

conservation in Bonin (Ogasawara) Islands. Conservation Genetics, 4: 605-612.

Thompson JD. 2005. Plant evolution in the Mediterranean. Oxford University Press: New York.

Ungerer MC, Baird SJE, Pan J & Rieseberg LH. 1998 Rapid hybrid speciation in wild sunflowers. Proceedings of

the National Academy of Sciences of the USA, 9: 11757–11762.

Wagner WH. 1970. Biosystematics and evolutionary noise. Taxon 19: 146-151.

Wagner FS. 1992. Cytological problems in Lycopodium sens. lat. Annals of the Missouri Botanical Garden, 79:

718–729.

Warren M. 1884. On the etymology of hybrid (Lat. Hybrida). The American Journal of Philology, 5: 501-502.

Whitham TG, Morrow PA & Potts BM. 1991. Conservation of hybrid plants. Science, 254: 779-780.

54 Chapter 1. General Introduction References

Whitney KD, Randell RA &, Rieseberg LH. 2006. Adaptive introgression of herbivore resistance traits in the

weedy sunflower Helianthus annuus. American Naturalist, 167: 794-807.

Willyard A, Cronn R & Liston A. 2009. Reticulate evolution and incomplete lineage sorting among ponderosa

pines. Molecular Phylogenetics and Evolution, 52: 498-511.

Wolf DE, Takebayashi N & Rieseberg LH. 2001. Predicting the risk of extinction through hybridization.

Conservation Biology, 15: 1039–1053

Wood TE, Burke JM & Rieseberg LH. 2005. Parallel genotypic evolution: when evolution repeats itself.

Genetica, 123: 157–170.

Woodruff DS. 1973. Natural hybridization and hybrid zones. Systematic Zoology, 22: 213-218.

55 Chapter 1. General Introduction References

56 Chapter 2. Objectives of the thesis

Chapter 2 Objectives of the thesis

page 2.1. Aims of the thesis 59 2.2. Selected case studies 61 2.2.1. Suspected cases of homoploid hybrids in the Balearic Islands 61 2.2.2. Selection of the case studies 63 2.3. References 64

57 Chapter 2. Objectives of the thesis

58 Chapter 2. Objectives of the thesis Aims of the thesis

2.1. AIMS OF THE THESIS

This thesis is a first approach to the study of the importance of natural hybridization in the evolution of the endemic taxa of the Balearic Islands, focused on homoploid hybridization assessed from the molecular and morphological points of view.

Several situations make this approach suitable for study: (i) The importance of natural homoploid hybridization in worldwide plant evolution, either as a creative or a species demise factor (e.g., Soltis & Soltis, 2009). (ii) The importance of the Balearic

Islands as a part of the Mediterranean speciation hot spot, with about 7% endemic species

(Alomar et al., 1997) and thus, the high chance for the existence of hybrids and also for hybrid speciation. (iii) The existence of described homoploid hybrids involving endemic species from the Balearic Islands, and the suggestion about the existence of others, from the observation of character intermediacy (e.g., Chodat, 1924; Cardona et al., 1983;

Galbany-Casals et al., 2006). (iv) The frequent difficulty to demonstrate, and even to detect, the existence of homoploid hybrids from morphological approaches alone (e.g., Rieseberg

& Willis, 2007). (v) The existence of other factors promoting character intermediacy but hybridization (e.g., incomplete lineage sorting and parallel adaptation Gross & Rieseberg,

2005; Wood et al., 2005). (vi) The proved utility of molecular markers in the detection of hybridization processes, especially when enhanced with morphology (e.g.,Rieseberg &

Ellstrand, 1993).

In the last years, there have been an increasing number of genetic studies dealing with the Balearic endemic flora, especially from karyological (e.g., Castro & Rosselló, 2007;

Rosselló & Castro, 2008), and phylogeographical approaches (e.g, Cosín; Rosselló et al.,

2006; Rosselló et al., 2007; Molins et al., 2009; Molins, 2010). These studies detected

59 Chapter 2. Objectives of the thesis Aims of the thesis prominent cytotype variation within some genera (e.g., Limonium), but most Balearic endemics showed ploidy levels equal to their more widespread relatives. Thus, if endemic species of hybrid origin are expected to exist in the endemic Balearic flora, such results would favour a homoploid rather than polyploid origin in most genera.

On the other hand, genetic variability in narrowly distributed taxa of the Balearic endemic flora resulted much higher than expected (Molins et al., 2009; Molins, 2010). This would favour either hypotheses pointing to radiation and population divergence, or hypotheses pointing to the merge of those taxa with their widespread relatives to obtain such genotypic richness. Indeed, the large range of phenotypes frequently shown by some

Balearic endemics, with respect to their closest relatives, could have arisen either from natural selection favouring divergent individuals in extreme habitats, or from hybrid phenotypes showing enough fitness to outcompete their relatives in any habitat.

However, there is a lack of molecular studies aiming to detect or to test for the putative hybrid origin of intermediate or divergent plants or populations, or even of endemic species of the Balearic Islands. Therefore, this tesis is a first approach to assess the levels of interspecific gene flow in the endemic flora of the Balearic Islands. Specifically, the objectives of this thesis are to assess:

1) The putative hybrid origin of the morphologically intermediate plants between Balearic

endemic and more widespread species.

2) The importance of the hybridization processes in the evolution of the Balearic endemic

taxa, either as a factor promoting divergence and speciation, or as a factor swamping

possible divergent genotypes.

3) The symmetry of the hybridization processes.

4) The hybridization processes as a threat in the conservation of the endemic flora of the

Balearic Islands.

60 Chapter 2. Objectives of the thesis Selected case studies

2.2. SELECTED CASE STUDIES

2.2.1. Suspected cases of homoploid hybrids in the Balearic Islands

There are six hypothesized cases of endemic taxa putatively involved in homoploid hybridization events in the Balearic Islands : Viola jaubertiana Marès & Vigineix, Lotus fulgurans (Porta) D.D.Sokoloff, Rhamnus ludovici-salvatoris Chodat, Teucrium cossonii D.

Wood subsp. punicum Mayol, Mus, Rosselló & N. Torres, Helichrysum crassifolium (L.) D.

Don and Arenaria bolosii (Cañig.) L. Sáez & Rosselló.

In the first case, the hybrid was described by Chodat (1924), currently known as

Viola x balearica. However, only morphological and karyological approaches have been made to confirm the hybrid origin of such deviant plants (Chodat, 1924 ; Schmidt, 1961).

Viola x balearica is known to occurr in a single and single location in the Tramuntana range of Majorca (Gorg Blau), and it is considered to be highly sterile. Moreover, the existence of deviant plants in this specific locality date back about 130 years.

In the case of Lotus fulgurans, intermediate plants were reported from two localities in Minorca by Cardona et al. (1983). However, neither hybrid description nor studies aiming to discern about the origin of such intermediate plants were done before this thesis work.

Lotus fulgurans deviant plants seem to occur only in the zone of contact between the typical coastal habitat of the endemic species and the coastal garrigue inhabited by Lotus dorycnium L., considered the putative parental taxa of the hybrids.

Rhamnus x bermejoi P. Fraga and Rosselló is the name given to the recently described hybrid between the endemic R. ludovici-salvatoris and the widespread R. alaternus L., and it is only known to occur in a single locality of Minorca (Fraga & Rosselló,

2008). In fact, early descriptions of such hybrid as R. jacobi-salvatori O. Bolòs & Vigo from

Majorca (Bolòs & Vigo, 1974), corresponded to R. alaternus (Rosselló & Mus, 1988).

61 Chapter 2. Objectives of the thesis Selected case studies

Teucrium cossonii subsp. punicum is an Ibizan endemic taxa related to the Majorcan endemic T. cossonii subsp. cossonii. It shows a conspicuously long inflorescence related to the Majorcan plants, though many variable plants have been observed in northern Ibiza.

Since the close proximity of the related T. capitatum L. plants (from the T. polium L. complex) grows in the vicinities, it has been suggested that such variability could be due to hybridization processes with the latter taxon (Mus, 1992). However, the Majorcan subsp. cossonii plants also shows somehow such variability in several locations. Indeed, there is a lack of knowledge about the possible taxa of the T. polium complex inhabiting in the

Balearic Islands, thus deserving of a deep revision of the T. polium-T. cossonii complex.

Helichrysum crassifolium is the single Helichrysum taxa endemic from the Balearic

Islands. It is rupicolous and grows in calcareous crevices all over Majorca and in a single location in Minorca, showing a conspicuous leaf shape variation along its distribution range.

It has been suggested that this variation is due to hybridization events with the also rupicolous H. pendulum (C. Presl.) C. Presl. (wrongly known as H. rupestre (Rafin.) DC.), and such hybrids have been reported (but not formally described) from both islands (Cosín et al., 2002 ; Galbany-Casals et al., 2006).

Finally, Arenaria bolosii is a narrow endemic species known from a single location of the Tramuntana range (Massanella). It is a very related taxon to A. grandiflora L. subsp. glabrescens (Willk.) López & Feliner, also endemic from Majorca. The latter taxon inhabits either in rupicolous habitat in calcareous cliffs, or protected into spinous vegetation in the basis of such cliffs. Oppositely, A. bolosii occurs in a rocky area with low slope separated several hundred meters from A. grandiflora subsp. glabrescens plants in Massanella.

Although, plants showing intermediate morphology have been detected between both areas, resembling putative hybrid plants (Mus, 1992).

62 Chapter 2. Objectives of the thesis Selected case studies

Therefore, in order to test for the magnitude that hybridization processes could have on the evolution of the endemic Balearic flora, three of the above cases were selected to be studied in this thesis.

2.2.2. Selection of case studies

First, cases with a very narrow location of putative hybrid plants are those involving

V. jaubertiana, A. bolosii, and R. ludovici-salvatoris. However, the latter was described in the final stage of this thesis and thus, it could not be included. In the case of A. bolosii, preliminary molecular data (Rosselló et al., unpublished) denoted the lack of genetic markers to discriminate both putative parental taxa. Indeed, the characteristics of this case study deserve of very different molecular techniques others that those used in this thesis.

Therefore, this case was rejected. The opposite situation with regards to preliminary molecular data was obtained for V. x balearica and thus, this case study was selected as an example of a very narrowly located putative hybrid, from which high degree of sterility have been reported (Chodat, 1924 ; Schmidt, 1961).

Second, the case involving L. fulgurans showed suspicions of being a typical hybridization process occurring exclusively in a zone of contact between two different habitats (ecotone), as described in literature (e.g., Anderson, 1948 ; Woodruff, 1973).

Therefore, this was also selected.

Lastly, both cases involving Teucrium cossonii and Helichrysum crassifolium show morphologically deviant plants in a wide portion of its distribution range. With regards to the former, the lack of a deep revision of the genus in the Balearic Islands put in concern the selection of putative parental taxa involved in the suspected hybridization process.

Therefore, previous work on this direction should be performed. Since that, the third case study selected was that involving the endemic H. crassifolium.

63 Chapter 2. Objectives of the thesis References

2.3. REFERENCES

Alomar, G., Mus, M. & Rosselló, J.A. 1997. Flora endèmica de les Balears. Palma de Mallorca: Consell Insular

de Mallorca.

Anderson E. 1948) Hybridization of the habitat. Evolution 2:1-9.

Bolòs O de, Vigo J. 1974. Notes sobre taxonomia i nomenclatura de plantes, I. Butlletí de la Institució Catalana

d’Història Natural 38: 61-89.

Cardona MA, Llorens L, Sierra E. 1983. Étude biosystématique de Dorycnium pentaphyllum Scop. subsp.

fulgurans (Porta) comb. nova, endémique des Baléares orientales. Collectanea Botanica (Barcelona)

14: 133-150.

Castro M & Rosselló JA. 2007. Kariology of Limonium (Plumbaginaceae) species from the Balearic Islands and

the western Iberian Peninsula. Botanical Journal of the Linnean Society, 155: 257-272.

Chodat L. 1924. Contributions à la Géo-Botanique de Majorque. Thèse. Genève.

Cosín, R., M.À Conesa, M. Mus & J.A. Rosselló. 2002. Hibridación reticular en especies insulares de

Helichrysum Mill. (Compositae) detectada mediante ISSR y secuencias de ADN cloroplástico. 1er

Congreso de Biología de la Conservación de Plantas, Jardín Botánico de la Universidad de Valencia,

2-5 octubre 2002. Valencia. Spain. pàg. 93.

Fraga P & Rosselló JA. 2008. Rhamnus x bermejoi, a new hybrid between R. alaternus and R. ludovici-

salvatoris. Flora Montiberica, 40: 47-49.

Galbany-Casals, M., Sáez, L. & Benedí, C. 2006. A taxonomic revision of Helichrysum Mill. sect. Stoechadina

(DC.) Gren. & Godr. (Asteraceae, Gnaphalieae). Can. J. Bot. 84:1203-1232.

Gross BL & Rieseberg LH. 2005. The ecological genetics of homoploid hybrid speciation. Journal of Heredity,

96: 241-252.

Molins A. 2010. Estudio filogeográfico de especies vegetales del Mediterráneo Occidental. Tesis doctoral.

Departament de Biologia. UIB: Palma de Mallorca.

Molins A, Mayol M & Rosselló JA. 2009. Phylogeographical structure in the coastal species Senecio rodriguezii

(Asteraceae), a narrowly distributed endemic Mediterranean plant. Journal of Biogeography, 36: 1372-

1383.

64 Chapter 2. Objectives of the thesis References

Mus M. 1992. Estudio de la diversificación en la flora endémica de Baleares. Aspectos taxonómicos y

evolutivos. Tesi doctoral inèdita. Universitat de les Illes Balears, Palma de Mallorca.

Rieseberg LH & Ellstrand NC. 1993. What can molecular and morphological markers tell us about hybridization?

Critical Reviews in Plant Sciences, 12: 213-241.

Rieseberg LH,Willis JH. 2007. Plant speciation. Science 317:910–14.

Rosselló JA & Mus M. 1988. Estudio morfológico, anatómico y cromatográfico de Rhamnus x jacobi-salvadorii

O.Bolòs & J.Vigo, supuesto híbrido de Rhamnus alaternus por R. ludovici-salvatoris. Candollea 43:

199-207.

Rosselló JA & Castro M. 2008. Karyological evolution of the angiosperm endemic flora of the BAlearic

Islands. Taxon, 25: 259-273.

Rosselló JA, Cosín R, Boscaiu M, Vicente O, Martínez I & Soriano P. 2006. Intragenomic diversity and

phylogenetic systematics of wild rosmaries (Rosmarinus officinalis L. sl.l, Lamiaceae) assessed by

nuclear ribosomal DNA sequences (ITS). Plant Systematics and Evolution, 262: 1-12.

Rosselló JA, Lázaro A, Cosín R & Molins A. 2007. A phylogeographic split in Buxus balearica (Buxaceae) as

evidenced by ribosomal markers: when ITS paralogues are welcome. Journal of Molecular Evolution,

64: 143: 157.

Schmidt A. 1961. Zytotaxonomische Untersuchungen an europäischen Viola-Arten der Sektion Nomimium.

Oesterreichische Botanische Zeitschrift. Gemeinnütziges Organ für Botanik... 108: 20-88.

Soltis PS & Soltis DE. 2009. The role of hybridization in plant speciation. Annual Review of Plant Biology, 60:

561-588.

Wood TE, Burke JM & Rieseberg LH. 2005. Parallel genotypic evolution: when evolution repeats itself.

Genetica, 123: 157–170.

Woodruff DS. 1973) Natural hybridization and hybrid zones. Systematic Zoology 22, 213-218

65

.

66 Chapter 3. The hybrid origin of the endemic Viola x balearica

CASE STUDY I Hybridization involving the endemic Viola jaubertiana

Chapter 3 The hybrid origin of the endemic Viola x balearica Rosselló, Mayol & Mus

page 3.1. Introduction 69 3.1.1. The genus Viola 69 3.1.2. Hybridization in the genus Viola 70 3.1.3. Hybrid sterility and the maintenance in nature 71 3.1.4. The Viola species in the Balearic Islands 72 3.2. The hybrid and its putative parental species 76 3.2.1. The Gorg Blau locality 77 3.3. Objectives 80 3.4. Materials and methods 82 3.4.1. Plant sampling 82 3.4.2. DNA extraction 82 3.4.3. Nuclear ribosomal ITS sequences 82 3.4.4. Phylogenetic analysis 83 3.4.5. TrnT-TrnL chloroplast DNA sequences 84 3.4.6. Sterility evaluation 85 3.5. Results 87 3.5.1. ITS and cpDNA species-specific markers in Balearic Viola 87 3.5.2. Nuclear and cpDNA genotyping of Viola accessions at the Gorg Blau site 87 3.5.3. The phylogenetic relationships of V. jaubertiana 90 3.5.4. Sterility evaluation of V. x balearica 91 3.6. Discussion 93 3.6.1. Hybridization between V. jaubertiana and V. alba subsp. dehnhardtii 93 3.6.2. The phylogenetic relationships of V. jaubertiana 97 3.7. References 99

67 Chapter 3. The hybrid origin of the endemic Viola x balearica

THIS CHAPTER HAS BEEN PARTIALLY PUBLISHED AS AN ORIGINAL PAPER:

Conesa MÀ, Mus M & Rosselló JA. 2008. Hybridization between insular endemic and widespread

species of Viola in non-disturbed environments assessed by nuclear ribosomal and

cpDNA sequences. Plant Systematics and Evolution 273: 169-177.

68 Chapter 3. The hybrid origin of the endemic Viola x balearica Hybrid and parental species

3.1. INTRODUCTION

In this first case study new data regarding to the origin of the endemic Viola x balearica Rosselló, Mayol & Mus are presented. This taxon was described as an hybrid between the Balearic endemic V. jaubertiana Marès & Vigineix, and the Mediterranean distributed Viola alba subsp. dehnhardtii (Ten.) W. Becker (Chodat, 1924).

3.1.1. The genus Viola

The genus Viola L. comprises among 400 (Cronquist, 1981), and 600 species

(Clausen, 1964), mainly distributed in temperate regions of the North Hemisphere and tropical mountains. This genus has been usually divided into sections and subsections based on morphological traits, on the chorology of the different groups, and later on molecular systematics (e.g., Becker, 1925; Gershoy, 1934; Clausen, 1964; Ballard, 1996;

Marcussen & Borgen, 2000). Consequently, there are different proposals for the infrageneric taxonomy of the genus Viola. This thesis follows the criteria of Valentine et al.

(1968), Muñoz Garmendia et al. (1993), Marcussen & Borgen (2000), and Malécot et al.

(2007). Thus, all species studied in this thesis belong to sect. Viola L. subsect. Viola.

The type species of the genus is V. odorata L. (Haesler, 1982), previously considered as a diploid with 2n=20, which is the chromosome number of most species of subsect. Viola. However, V. ambigua Waldst. & Kit., and V. suavis M.Bieb. show 2n=40 and were considered tetraploids (Clausen, 1929; Schmidt, 1961a; Muñoz-Garmendia et al.

1993; Okamoto et al. 1993; Mered’a et al., 2006). Currently, it is widely accepted that the genus Viola is a palaeotetraploid complex with an ancient hybrid origin by several allopolyploid episodes (Marcussen & Nordal, 1998; Nordal & Jonsell, 1998; Marcussen &

69 Chapter 3. The hybrid origin of the endemic Viola x balearica Hybrid and parental species

Borgen, 2000). Thus, 2n=20 species would be tetraploids and 2n=40 octoploids

(Marcussen & Borgen, 2000), being x=6 the basic chromosome number of the genus

(Clausen, 1929; Miyaji, 1929; Valentine, 1962). This interpretation is in agreement with phylogenetic reconstructions based on ITS sequences(Ballard et al., 1999) which also suggest this number to be the ancient basic number of the whole Violaceae family, but stating the number of x=4 for some genera. Yet, other options can be found in literature for the genus Viola, such x=5 (Marcussen & Borgen, 2000) and, at least for sect. Melanium

DC. ex. Ging., x=5 or x=7 (Yockteng et al., 2003). Thus, originally diploid species are not found into sect. Viola. However, those (2n=10) exist into the close sect. Melanium (Schmidt,

1961a; Ballard et al., 1999).

3.1.2. Hybridization in the genus Viola

Over 60 hybrids have been described in Europe within the sect. Viola, half of which have been originated exclusively among taxa from subsect. Viola (Valentine et al., 1968).

As stressed by this latter author for Europe, N of Africa and W of Asia, no hybrids among taxa belonging to different sections of the genus have been reported in the literature.

However, hybrids exist among subsections of the same section. In the Iberian Peninsula and the Balearic Islands, at least 30 hybrids have been cited within the genus Viola, from which only one belongs to sect. Melanium DC. ex. Ging., and the remaining 29 to sect.

Viola (Muñoz Garmendia et al., 1993; Erben, 1996). From those, the species of the subsect.

Viola are parental species of 11 hybrids, five more are hybrids between a taxa of that subsection and another from subsect. Rostratae Kupffer, and the 13 remaining belong to the latter subsection (Muñoz Garmendia et al., 1993).

Many authors agree with the importance of hybridization and introgression in the evolution of the group. Although, hybridization in the genus Viola seems to be feasible

70 Chapter 3. The hybrid origin of the endemic Viola x balearica Hybrid and parental species when two or more taxa occur in the same region, also favoured by the absence of prezigotic barriers and anthropic activities allowing contact among normally allopatric species

(Marcussen & Borgen, 2000). In consequence, successive processes of isolation, speciation, migration, hybridization, and stabilization of the new hybrids along the

Pleistocene could have been of major importance in the evolutionary history of the subsect.

Viola (see references in Marcussen & Borgen, 2000). In contrast, reported cases of hybrid swarms are scarce, only involving V. hirta L., which hybridizes with V. alba Bess., V. collina

Bess., or V. odorata L. (Schöfer, 1954; Schmidt, 1961b).

3.1.3. Hybrid sterility and its maintenance in nature

Reported hybrids among species of subsect. Viola and subsect. Rostratae are always fully sterile (Marcussen & Borgen, 2000, and references therein). Oppositely,

hybrids between subsect. Viola species are characterized by the first hybrid generation (F1) showing a high descent in fertility related to their parental taxa. However, these hybrids frequently present more than 5% of pollen fertility (not considered totally sterile), and it has been reported the increase in the fertility of such hybrids generation by generation

(Marcussen & Borgen, 2000). This would allow hybrids to take part in the evolution of the genus (e.g., Arnold et al., 1999) through the creation of new lineages (by isolation), through introgression into any or both the parental species, or even through crosses with any other species of the subsection.

Cleistogamy is a characteristic trait in several sections of the genus, including sect.

Viola (e.g. Valentine et al., 1968; Nadot et al., 2000; Oakley et al., 2007). In his study of the species of the subsect. Rostratae (V. riviniana Reicheb. and V. reichenbachiana Jordan),

Beattie (1969) stated that most of the seeds produced came from cleistogamous flowers, suggesting that the lack of chasmogamous seeds was due to a lack in pollinators. Indeed,

71 Chapter 3. The hybrid origin of the endemic Viola x balearica Hybrid and parental species

Beattie (1971) explained the existence of such complex chasmogamous flowers, producing only a small proportion of seeds compared to the cleistogamous flowers, arguing that in years of high pollinator frequency such proportions could be inverted. He also emphasized the evolutive importance of these two mechanisms to produce seeds: one giving rise to less diverse seeds mainly from autopollination, and the other producing recombinant seeds through sexual reproduction. In fact, the existence of cleistogamous flowers would favour the stabilization and maintenance of low fertility hybrid lineages (Moore, 1959; Valentine,

1962; Gil-ad, 1997; cf. Marcussen & Borgen, 2000), favouring the reproductive isolation of the hybrid from the parental taxa. In turn, chasmogamous flowers are needed to allow the interspecific cross. Thus, a low frequency in chasmogamous seeds would mean a low possibility for hybrid arising, maybe explaining the above stated low frequency of hybrid swarms in the genus Viola.

3.1.4. The Viola species in the Balearic Islands

There are many taxa of the genus Viola cited and described in the Balearic Islands

(see below). However some of them are nomenclatural combinations, taxonomic misinterpretations, or not wrong names. With the aim of avoiding ambiguities later on in this thesis with respect to the species involved in the putative hybridization, a brief revision is performed. Following Muñoz Garmendia et al. (1993), and Rosselló & Sáez (2001), in the

Balearic Islands there are only six wild taxa correctly identified, summarized in the Table

3.1. Indeed, following the infrageneric classification criteria stressed above, all these species belong to sect. Viola subsect. Viola but V. arborescens L.

72 Chapter 3. The hybrid origin of the endemic Viola x balearica Hybrid and parental species

Table 3.1. Taxa of the genus Viola existing in the wild in the Balearic Islands. Sections are denoted for all species, following Muñoz Garmendia et al. (1993). Islands in the Balearic archipelago: MA, Majorca; CA, Cabrera; ME, Minorca; EI, Ibiza; FO, Formentera. An asterisk denotes endemicity from the Balearic Islands. 1 Some V. odorata plants from Minorca have been described as the endemic species V. stolonifera J.J.Rodr. Both species are cited in Minorca (Fraga et al., 2004).

Section Taxon MA CA ME EI FO Viola alba subsp. dehnhardtii (Ten.) W. Becker ✔ ✔ Viola jaubertiana Marès & Vigineix ✔* Viola L. Viola odorata L. ✔1 ✔ Viola x balearica Rosselló, Mayol & Mus ✔* Xylinosium W. Becker Viola arborescens L. ✔ ✔ ✔ ✔

Viola arborescens belongs to sect. Xylinosium W. Becker, a very different section from sect. Viola. This species was discarded in the present study according to the available literature stressing the lack of hybrids among the different sections of the genus Viola (e.g.,

Valentine et al., 1968; Ballard, 1996) and the lack of evidences of being involved in hybridization processes in the Balearic Islands.

With respect to V. odorata L., some authors consider the Minorcan plants as V. stolonifera J.J.Rodr., an endemic species from Minorca (e.g., Rodríguez, 1878; Barceló,

1879-1881; Burnat & Barbey, 1880; Porta, 1882; Chodat, 1924; Bonafè, 1976; Rosselló &

Sáez, 2001; Fraga et al., 2004). However, other authors include the latter into the former

(e.g., Muñoz Garmendia et al., 1993; Valentine et al., 1968). Indeed, many citations of V. odorata in Majorca, and many related taxa described, are no further considered (Muñoz

Garmendia et al., 1993; Rosselló & Sáez, 2001). In spite of the uncertainty on the real taxonomic status of the two species (suggesting a deeper revision of the complex), there are evidences neither for the occurrence of wild V. odorata - V. stolonifera in Majorca, nor plants with similar characteristics. In addition, in a wide sampling of the species of subsect.

Viola throughout Majorca carried out parallel to this study (Conesa et al., unpublished), no evidences for the existence of V. odorata or V. stolonifera in Majorca have been found. 73 Chapter 3. The hybrid origin of the endemic Viola x balearica Hybrid and parental species

Likewise, many dubious taxa related to V. alba Bess. have been cited from the

Balearic Islands, most of them not considered (e.g., Muñoz Garmendia et al., 1993;

Rosselló & Sáez, 2001). Indeed, following Marcussen & Borgen (2000), in the Western

Mediterranean it only occurs V. alba subsp. dehnhardtii (Ten.) W. Becker, however stressing these authors its parapatry with other subspecies. Although, the subsp. scotophylla (Jord.) Nyman, cited by Bolòs et al. (1990) from Majorca, is considered a synonym of the subsp. alba and thus, not occurring in the western Mediterranean

(Marcussen, 2003).

Another species cited in the Balearic Islands is Viola suavis M. Bieb. The first indication is that of Bolòs & Vigo (1974), describing the subsp. barceloi, after combined as subsp. sepincola (Jord.) Nyman var. barceloi (Bolòs & Vigo, 1990). Orell & Romo (1991) indicated it in Majorca (Gorg Blau), and it was also cited from Ibiza (Kühbier, 1978; Rivas-

Martínez et al., 1992). However, all reports of V. suavis from Majorca must be attributed to

V. alba subsp. dehnhardtii (Rosselló & Sáez, 2001) and those from Ibiza to V. odorata

(Muñoz Garmendia et al., 1993; N. Torres, unpublished data). Yet, the presence of this species in the Balearic Islands is not considered by many authors (e.g., Muñoz Garmendia et al., 1993; Marcussen & Borgen, 2000; Rosselló & Sáez, 2001). Although, one must be aware that all chromosome counts from Gorg Blau always showed 2n=20 (Schmidt, 1961b;

Guinochet & Lefranc, 1972; Castro & Rosselló, 2006); while V. suavis is 2n=40 (Miyaji,

1929; Clausen, 1929; Schmidt, 1961b, 1964; Muñoz Garmendia et al., 1993; Okamoto et al., 1993; Marcussen & Nordal, 1998).

Three more taxa cited in the Balearic Islands, probably also corresponding to V. alba subsp. dehnhardtii, are V. segobricensis Pau (e.g., Bonafè, 1979), V. ambigua Barceló

(1879) nom. illeg., and V. hirta (Marès & Vigineix, 1880). The former is considered a synonym of V. suavis (Muñoz Garmendia et al., 1993). Barceló (1879) described V.

74 Chapter 3. The hybrid origin of the endemic Viola x balearica Hybrid and parental species ambigua from Majorca, and it was renamed by Nyman (1889) as V. barceloi. In turn, this would illegitimate the description of Chodat (1924) of the hybrid of Gorg Blau (Rosselló,

Mayol & Mus, 1991). The citation of Barceló (1879) could be attributed to V. alba subsp. dehnhardtii due to the extent of the distribution stated in the description by this author.

However, the indication of its use in gardens (Barceló, 1879) could favour its correspondence to V. odorata. In the case of V. hirta, as well as V. ambigua, no further reports are known from the Balearic Islands (e.g., Valentine et al., 1968; Orell & Romo,

1991; Muñoz Garmendia et al., 1993; Marcussen, 2003). This would reinforce the correspondence of such citations to V. alba subsp. dehnhardtii.

With respect to the endemic taxa from Majorca, all consulted authors emphasize the clear differences among V. jaubertiana Marès & Vigin. and any other taxa of the subsection.

Indeed, many also consider it an isolated and relict taxon into the subsect. Viola (Tchourina,

1909; Chodat, 1924; Schmidt, 1961b, Marcussen & Borgen, 2000). Noticing its similarities with the species, Knoche (1922) combined V. jaubertiana as a subspecies of V. odorata

(i.e., corresponding to V. alba subsp. dehnhardtii in Majorca; Muñoz Garmendia et al.,

1993; Rosselló & Sáez, 2001).

75 Chapter 3. The hybrid origin of the endemic Viola x balearica Hybrid and parental species

3.2. THE HYBRID AND ITS PUTATIVE PARENTAL SPECIES

Viola x balearica, being this the accepted name (Mus et al., 2000), was described by

Chodat (1924) as V. barceloi (nom. illeg.) from the Gorg Blau locality, in Majorca. It was considered an hybrid between the western Mediterranean V. alba subsp. dehnhardtii and the Majorcan endemic V. jaubertiana and thus, it is only expected to occur in Majorca.

Viola jaubertiana is restricted to small, fragmented, and scattered populations of northern mountains, where the more distant populations are located only 24 km apart. As stated above, Viola jaubertiana is a separate and well-defined species that can be clearly distinguished from other European violets on morphological grounds. Further, isozyme analyses revealed that V. jaubertiana is also quite distinct from other species of Viola subsect. Viola (Marcussen & Borgen, 2000). The ecological requirements of V. jaubertiana are unique among European violets since it is the only species that is strictly rupicolous and lives in inaccessible rocky places and calcareous overhangs. Indeed, it is entirely glabrous and morphologically very uniform. However, several authors have reported that individuals collected at the type locality (Gorg Blau; Rosselló & Sáez, 2001) were also hairy (Burnat &

Barbey, 1882; Chodat, 1924; Schmidt, 1961b, Mus et al., 2000). It has been suggested that these rupicolous, pubescent plants resembling V. jaubertiana are hybrids with V. alba Bess. subsp. dehnhardtii, and so described by Chodat (1924).

The latter is a Mediterranean woodland species growing in the slopes of evergreen- oak forests, in shade environments, and is the only hairy violet growing in the area (Chodat,

1924; Schmidt, 1961b; Mus et al., 2000; see above). A summary of the morphological traits distinguishing both putative parental species and V. x balearica is shown in Table 3.2.

76 Chapter 3. The hybrid origin of the endemic Viola x balearica Hybrid and parental species

Some works dealing with this suspected hybrid from Gorg Blau involve caryological and fertility studies favouring the hybridization hypothesis, and arguing for the putative parental species (Chodat, 1924; Schmidt, 1961b). In addition, Schmidt (1961b) identified at

Gorg Blau a glabrous hybrid between V. jaubertiana and V. alba subsp. dehnhardtii from meiotic irregularities observed. Schmidt’s results, added to the degree of sterility observed by Chodat (1924), are indirect evidences for the hybrid origin of V. x balearica. Moreover, the high sterility of early generation hybrids among taxa of subsect. Viola (e.g., Marcussen

& Borgen, 2000) reinforces the hybrid origin hypothesis.

On the other hand, when describing V. jaubertiana, Marès & Vigineix (1880) reported that chasmogamous flowers were sterile. If these observations are correct (this is probably a misinterpretation of the fact that chasmogamous flowers frequently do not get pollinated and fail to set fruit), they compromise the hybrid status of the deviant pubescent plants, since there is no chance for pollen flow between cleistogamous and chasmogamous flowers.

3.2.1. The Gorg Blau locality

The Gorg Blau locality is an interesting place for the study of the violets. Some taxa of the genus have been described from that locality, including several more hybrids but V. x balearica (Table 3.3). Also, the endemic V. jaubertiana was described using specimens from this site (Marès & Vigineix, 1880; together with specimens from Cova de sa Botella,

Pollença). In addition, this is the only locality in the Balearic Islands where violet hybrids have been reported. It also represents a location where the rupicolous habitat of V. jaubertiana contacts widely with the oak forest habitat of V. alba subsp. dehnhardtii, situation that would favour the hybridization process (e.g., Schmidt, 1961b; Dizerbo, 1967,

1968; Valentine, 1975; Marcussen & Borgen, 2000).

77 Chapter 3. The hybrid origin of the endemic Viola x balearica Hybrid and parental species

Only four taxa of the genus Viola were detected at the Gorg Blau site in this study:

V. arborescens, V. alba subsp. dehnhardtii, V. jaubertiana, and those deviant plants putatively corresponding to V. x balearica. However, all the contradictory data found in literature from the Gorg Blau locality deserves of a more detailed study of the plants from this locality to ascertain the origin of the deviant plants reported.

Table 3.2. Diagnostic morphological features between V. jaubertiana, V. alba subsp. dehnhardtii (Balearic accessions) and the putative hybrid (V. x balearica). Extracted from Chodat (1924) and Muñoz-Garmendia, 1993.

Character V. alba subsp. dehnhardtii V. x balearica V. jaubertiana Habitat non-rupicolous rupicolous rupicolous Leaf Shape lower half narrow intermediate lower half wide Margin sinuouse, forward orientation not-sinuouse, not-sinuouse, curled teeth but not-curled teeth intermediate teeth Basis narrowly cordate or nearly intermediate (8mm) widely cordate (15mm) closed sinus Consistence non coriaceous non coriaceous more or less coriaceous Lamina dull, dark green dull, pale green glossy, dark green Indumentum hairy hairy glabrous Stipules Shape narrow, linear-lanceolate, lanceolate, ciliate margin triangular ciliate margin Fimbriae not-glandular glandular glandular Sepals Apex obtuse obtuse acute Indumentum hairy hairy glabrous Stamens smaller than V.j and V.b., with different from V.a. different from V.a. apendix nectariferous slim and capitated Carpel indumentum (capsule) Hairy Hairy Glabrous

78 Chapter 3. The hybrid origin of the endemic Viola x balearica Hybrid and parental species

Table 3.3. Taxa described from the Gorg Blau locality. Following Rosselló & Sáez (2001), only boldface ones are currently accepted. The synonymy is shown in the right column.

Described taxon Synonymy V. alba nothosubsp. bonaefidei Orell & Romo (1991) V. alba subsp. dehnhardtii (=Viola alba subsp. dehnhardtii x V. suavis subsp. barceloi, nom. illeg.) V. alba nothosubsp. neusii Orell & Romo (1991) V. x balearica (=Viola alba subsp. dehnhardtii x V. jaubertiana) V. x balearica Rosselló, Mayol & Mus (1991) - (=Viola alba subsp. dehnhardtii x V. jaubertiana) V. barceloi L. Chodat (1924), nom. illeg. V. x balearica - Homotipic synonym (=Viola alba subsp. dehnhardtii x V. jaubertiana) V. jaubertiana Marès & Vigin.(1880) - V. suavis nothosubsp. cardonae Orell & Romo (1991) V. jaubertiana (=Viola suavis subsp. barceloi, nom. illeg. x V. jaubertiana)

Three different Viola x balearica plants from the Gorg Blau locality, used in this study (October 2006).

79 Chapter 3. The hybrid origin of the endemic Viola x balearica Objectives

3.3. OBJECTIVES

Circumstantial evidence supports the hybrid origin of the hairy plants in the Gorg

Blau locality, like their low pollen and ovule fertility (Chodat, 1924), their irregular meiotic behaviour (Schmidt, 1961b), and the somewhat mixed morphological features of V. jaubertiana and V. alba subsp. dehnhardtii they present (Table 3.2). However, observations of the sterility of chasmogamous flowers in V. jaubertiana (Marès & Vigineix, 1880), the possible existence of hybrid variability or other hybrid combinations (Schmidt, 1961b), the high number of taxa cited in Gorg Blau (see above), and the description of other hybrid lineages at the same locality (Table 3.3), deserve of further analyses to confirm the origin of such deviant plants.

Other alternatives to the hybridization scenario are likely and should be considered.

First, pubescent specimens related to V. jaubertiana could merely represent infraspecific polymorphisms of labile characters (hairs) that are also present in other species of subsection Viola. Second, it could be argued that V. jaubertiana and V. alba are sister species, and that these individuals showing intermediate morphology are the result of incomplete primary divergence. Third, a cryptic, previously undetected, relict species could be present at the Gorg Blau site. Low pollen and ovule fertility could be explained by the effects of genetic drift and inbreeding depression in a low effective population size. In the short term, inbreeding can reduce fitness and may potentially lower population viability

(Ellstrand & Elam 1993).

Hybridization is not a rare event among violet species. However, most of the reported hybrids occur on ecotones, disturbed and artificial habitats. Neuffer et al. (1999) reported the presence of introgressive hybrids between V. riviniana and V. reichenbachiana

80 Chapter 3. The hybrid origin of the endemic Viola x balearica Objectives in non human-disturbed habitats (pine forests) but suffering from air pollution. Interestingly, there was a correlation of violet hybrid populations and the impact of pollution on habitat conditions (Neuffer et al., 1999). In contrast with these observations, putative hybrids of V. jaubertiana and V. alba subsp. dehnhardtii occur in the singular rupicolous and pristine habitats of V. jaubertiana.

Given these uncertainties it seemed appropriate to examine the status of the pubescent violets resembling the narrow endemic V. jaubertiana from the perspective of molecular systematics. Molecular markers have been successfully used to address taxonomic and evolutionary questions at lower taxonomic ranks among a wide array of organisms (Avise, 1994).

In this work, it is presented an analysis of the internal transcribed spacers region

(ITS) from the nuclear 45S ribosomal cistron, and a non-coding chloroplast region (trnT- trnL) from Balearic Viola accessions. The aims of the study were (i) to evaluate whether the morphological polymorphisms detected in V. jaubertiana are due to hybridization between this species and V. alba subsp. dehnhardtii, and (ii) to explore the phylogenetic relationships of the endemic V. jaubertiana using ITS sequences.

81 Chapter 3. The hybrid origin of the endemic Viola x balearica Materials and methods

3.4. MATERIALS AND METHODS

3.4.1. Plant sampling

Flowering specimens of Viola jaubertiana (n=15), V. alba subsp. dehnhardtii (n=16), and putative hybrids (n=4) were sampled at Gorg Blau site. A single accession of the putative hybrid collected at Gorg Blau was also available from the living collections of the

Botanical Garden of Sóller (Majorca). Reference specimens of V. jaubertiana and V. alba subsp. dehnhardtii were sampled from distant sites from Gorg Blau, at populations where no other species of subsection Viola do occur. A wild collection of V. odorata from Minorca

(considered V. stolonifera by some authors) was used for comparison purposes. For each individual, healthy young leaves were placed on polyethylene bags filled with silica gel.

Voucher herbarium specimens are kept at the herbarium of the University of the Balearic islands (UIB).

3.4.2. DNA extraction

Total DNA of all specimens was extracted from dried leaves using the CTAB protocol (Doyle & Doyle, 1987), scaled down to perform the process in 1.5 ml microfuge tubes.

3.4.3. Nuclear ribosomal ITS sequences

ITS1, 5.8S and ITS2 sequences were generated for the reference specimens of V. jaubertiana, V. alba subsp. dehnhardtii, V. odorata, and putative hybrids. The whole ITS region was amplified using the primer pair ITS-1/ITS-4 (White et al., 1990; Figure 3.1). PCR reactions were performed in a 20-μl solution, containing 75 mM Tris-HCl (pH 9.0), 5 mM

KCl, 20 mM (NH4)2SO4, 0.0001% BSA, 2 mM MgCl2, 200 μM of each dATP, dCTP, dGTP

82 Chapter 3. The hybrid origin of the endemic Viola x balearica Materials and methods and dTTP, 0.15 μM of each primer, 5% DMSO, approximately 50-100 ng of genomic DNA and 0.3 units of Netzime® DNA polymerase from Thermus thermophilus. Amplification comprised an initial denaturalization step of 3 min at 94ºC, and 40 cycles of 30 s at 94ºC, 30 s at 57ºC and 1 min at 72ºC, and a final extension step of 5 min at 72ºC, in a Px2 thermal cycler (Thermo Electron Corporation, Milford, MA, USA). The PCR products were separated on 1.0% agarose gels and purified using the High Pure PCR Product Purification Kit (Roche

Diagnostics). For sequencing, purified PCR products were reacted with BigDye Terminator

Cycle Sequencing Ready Reaction (Perkin-Elmer, Applied Biosystems) using both amplification primers. The boundaries of the ITS sequences and ribosomal coding regions were determined by comparison to published sequences of Viola species (Ballard et al.,

1999).

Figure 3.1. Schematic presentation of the nuclear 18S-26S ribosomal DNA repeat of the nucleolar organizer region (NOR) in plants. It is not scaled. Genes (blue boxes) are separated by the internal transcribed spacers (ITS, green boxes). These units, tandemly repeated, are separated by the intergenic spacer (IGS). The location of the primers ITS-1 and ITS-4 (White et al., 1990) is shown (in red). Primers ITS-1 and ITS-4 were used to amplify the ITS1+5.8S+ITS2 region with about 700 bp.

3.4.4. Phylogenetic analysis

Phylogenetic reconstructions were performed using ITS sequences from Balearic accessions of V. alba subsp. dehnhardtii, V. jaubertiana, V. odorata, sequences of Viola subsect. Viola obtained from several studies (Ballard et al., 1999; Yockteng et al., 2003;

Yoo et al., 2005; Malécot et al., 2007), and sequences available from GenBank (Table 3.4).

83 Chapter 3. The hybrid origin of the endemic Viola x balearica Materials and methods

V. scandens (sect. Leptidium) and V. hondoensis (sect. Viola, from Korea) were used as outgroups based on a previous ITS phylogeny (Ballard et al., 1999). Phylogenetic reconstructions were obtained by using a matrix of ITS sequences aligned with Clustal X

1.81 (Thompson et al., 1997), after manual adjustment of sequences, and implementing

PAUP*4.0b10 (Swofford, 2002). Phylogenetic analyses using heuristic searches were conducted using Fitch parsimony with equal weighting of all characters, polymorphic states were considered as uncertainties, and indels were treated as missing data. Support for monophyletic groups was assessed by fast bootstrapping (1000 resamplings) using the heuristic search strategy. In addition, phylogenetic reconstructions were also performed by distance-based methods. The Kimura two-parameter distance model (Kimura, 1980) and the Neighbor-Joining method (Saitou & Nei, 1987) were used to construct a bootstrapped tree (1000 resamplings).

3.4.5. TrnT-trnL chloroplast DNA sequences

The trnT-trnL chloroplast region of V. jaubertiana, V. alba subsp. dehnhardtii, and V. odorata was amplified using the primers a and b given in Taberlet et al. (1991). PCR reactions were performed as above, except that DMSO was not used, and the PCR program was set at an initial denaturalization step of 3 min at 94ºC, and 40 cycles of 50 s at

94ºC, 50 s at 53ºC and 1 min 30 s at 72ºC, with a final extension step of 3 min at 72ºC in a

Primus 25 thermal cycler (MGW AG Biotech, Ebersberg, Germany). Purified PCR products were sequenced as above.

84 Chapter 3. The hybrid origin of the endemic Viola x balearica Materials and methods

Figure 3.2. Schematic presentation of the chloroplast DNA portion H from Shaw et al. (2005). Genes are represented in black boxes, with the name below. Primers designed to amplify intergenic regions are denoted with small black arrows with the name above. In red, the trnT-trnL region amplified with the primers a and b (Taberlet et al., (1991). In this study, a ca. 400 bp fragment was obtained.

3.4.6. Sterility evaluation

Observation of ovule and pollen fertility was carried out according to Martin (1959).

The cleistogamous flowers were fixed in ethanol:acetic acid (3:1) and stored at 4ºC. Prior observation, the material was washed in distilled water and softened in 1M KOH solution at

60ºC for 15 min.

Pistils were stained within 0.1% solution of water-soluble aniline blue dye dissolved

in 0.1 N K3PO4 over night. Analyses of preparations were made in a dark room. Pistils were observed using an epifluorescence Olympus microscope with ultraviolet light filter (356 nm) equipped with an Olympus Camedia C-2000-Z digital camera. The callose fluoresces bright yellow-green and contrasts strongly with the bluish fluorescence of the pistilar tissue.

For pollen studies, anthers were stained in a drop of 2% of haematoxilin and the slides were analyzed using the phase contrast in the same microscopy above mentioned.

85 Chapter 3. The hybrid origin of the endemic Viola x balearica Materials and methods

Table 3.4. Species, provenances, and codes of the GenBank accessions used in the phylogenetic analysis of Viola subsect. Viola. V. hondoensis (sect. Viola subsect. Viola) and V. scandens (sect. Leptidium Ging.) have been used as outgroups.

Taxon Origin Code GenBank accession no. Reference

V. alba Bess. subsp. alba Cyprus CYP1 DQ358847, DQ358824 Malécot et al., 2007

Cyprus CYP2 DQ358848, DQ358825 Malécot et al., 2007

Greece GRE1 DQ358846, DQ358823 Malécot et al., 2007

Greece GRE2 DQ358850, DQ358827 Malécot et al., 2007

Not reported UNK DQ358849, DQ358826 Malécot et al., 2007 V. alba subsp. cretica (Boiss.& Heldren) Greece - DQ358851, DQ358828 Malécot et al., 2007 Marcussen V. alba Bess. subsp. dehnhardtii (Ten.) W. France FRA DQ358853, DQ358830 Malécot et al., 2007

Becker Greece GRE DQ358852, DQ358829 Malécot et al., 2007

Majorca MAJ EU430657 This study

Spain SPA1 DQ358855, DQ358832 Malécot et al., 2007

Spain SPA2 DQ358854, DQ358831 Malécot et al., 2007

V. hirta L. Not reported - DQ358856, DQ358833 Malécot et al., 2007

V. hondoensis W.Becker & H. Boissieu Korea KOR1 AY928296 Yoo et al., 2005

Korea KOR2 AY928272 Yoo et al., 2005 V. jaubertiana Marès & Vigin. Majorca - EU430659 This study V. odorata L. Germany GER AF097250, AF097296 Ballard et al., 1999 Minorca MIN EU430658 This study Not reported UNK AY148238, AY148258 Yockteng et al. 2003

USA USA AF097249, AF097295 Ballard et al., 1999

V. scandens Willd. Costa Rica - AF097221, AF097267 Ballard et al., 1999

V. sintenisii W.Becker Not reported - DQ358859, DQ358836 Malécot et al., 2007

V. suavis M. Bieb Not reported UNK1 DQ358861, DQ358838 Malécot et al., 2007

Not reported UNK2 DQ358860, DQ358837 Malécot et al., 2007

86 Chapter 3. The hybrid origin of the endemic Viola x balearica Results

3.5. RESULTS

3.5.1. ITS and cpDNA species-specific markers in Balearic Viola

Nuclear ribosomal ITS sequences were highly similar in the three Balearic species belonging to subsect. Viola. ITS sequences showed the same length (612 bp) and no indels were hypothesized when they were aligned. Viola alba subsp. dehnhardtii and V. jaubertiana differed by three nucleotides in the reference accessions analyzed (Table 3.5).

Viola alba subsp. dehnhardtii and V. odorata could be distinguished by only two site mutations, whereas three mutations discriminated V. jaubertiana and V. odorata. Thus, species-specific markers were identified for the three species of Viola. The T at the site 449 characterized V. alba subsp. dehnhardtii, the T and G at sites 168 and 183 were diagnostic for V. jaubertiana, and the G at site 103 identified V. odorata. Sequences of the trnT-trnL cpDNA spacer of V. alba subsp. dehnhardtii, V. jaubertiana and V. odorata differed (Table

3.5) by a single site mutation, by the variable length of a microsatellite (poly-A) motif, and by a 5 bp indel (GAAAA). Lack of the GAAAA motif and the 11 bp length of the microsatellite characterized V. alba subsp. dehnhardtii. V. jaubertiana showed a species-specific point mutation and a 10 bp length of the poly-A microsatellite, whereas the 12 bp length of the microsatellite was the single molecular marker singularizing V. odorata.

3.5.2. Nuclear and cpDNA genotyping of Viola accessions at the Gorg Blau site

ITS and trnT-trnL sequences of all V. alba subsp. dehnhardtii (n=16) and V. jaubertiana accessions (n=15) were identical to those retrieved from the Balearic reference individuals of both species. The ribosomal sequences of the five putative hybrids showed additive nucleotides at the three sites polymorphic for the pair V. alba subsp. dehnhardtii-V. jaubertiana as estimated from direct sequencing (Table 3.5). This is interpreted as the result of amplification of copies inherited from both progenitors. No evidence for concerted

87 Chapter 3. The hybrid origin of the endemic Viola x balearica Results evolution was found at any of the three polymorphic sites, although putative hybrids differed by the equimolecular amounts of progenitor repeats present. CpDNA sequences of all putative hybrids were uniform, and identical to those present in V. jaubertiana accessions.

No nuclear and chloroplast species-specific markers of V. odorata, or other unrelated Viola, was detected in any V. alba subsp. dehnhardtii, V. jaubertiana, and the putative hybrid accessions at Gorg Blau. This agrees with morphological observations indicating that only two species of subsection Viola were present at the Gorg Blau site, and involved in the hybridization scenario.

Table 3.5. Nuclear and chloroplast species-specific sites in taxa of Viola subsect. Viola from the Balearic Islands, and in putative hybrids between V. alba subsp. dehnhardtii and V. jaubertiana. Sequence ambiguities follow IUPAC codes: R denotes A or G, Y denotes C or T, W denotes A or T. The superindexes indicate the base with strongest signal in the electrophoretograms.

Population ITS trnT-trnL Taxon (GenBank accession no.) 103 168 183 449 117 151 376 V. alba subsp. dehnhardtii Gorg Blau (n=16) A C A T T A11 - EU430656, EU434926 Puig Major (n=1) A C A T T A11 - EU430657, EU434925 V. jaubertiana Gorg Blau (n=15) A T G A A A10 GAAAA EU430660, EU434923 Mortitx

(n=1) A T G A A A10 GAAAA EU430659, EU434922 V. odorata Algendar

(n=1) G C A A T A12 GAAAA EU430658, EU434924 Putative hybrids Gorg Blau H1 A Y RA WT A A10 GAAAA (n=5) EU430661, EU434927 Gorg Blau H2 A Y R WA A A10 GAAAA EU430662, EU434928 Gorg Blau H3 A YT RG WA A A10 GAAAA EU430663, EU434929 Gorg Blau H4 A YC R W A A10 GAAAA EU430664, EU434930 Botanical Garden of Sóller A YC RA WA A A10 GAAAA EU430665, EU434931

88 Chapter 3. The hybrid origin of the endemic Viola x balearica Results

Figure 3.3. Electrophoretograms showing homologous positions (arrows) in the ITS region for V. alba subsp. dehnhardtii (left), V. x balearica (center) and V. jaubertiana (right). A) Position 168 (C+T=Y). B) Position 183 (A+G=R). C) Position 449 (T+A=W). See Table 3.5 for further information.

Figure 3.4. Electrophoretograms showing (arrows) homologous positions in the trnT- trnL region for V. alba subsp. dehnhardtii (left), V. x balearica (center) and V. jaubertiana (right). A) Position 117. B) Position 151 (poli-A). C) Position 376 (GAAAA InDel). See Table 3.5 for further information.

89 Chapter 3. The hybrid origin of the endemic Viola x balearica Results

3.5.3. The phylogenetic relationships of V. jaubertiana

Phylogenetic analyses of Viola subsection Viola using parsimony and distance methods were highly congruent and only conflicted in the resolution of few terminal clades.

Parsimony analysis yielded 1620 trees of 172 steps (CI = 0.89, RI = 0.86, and RC = 0.77, including uninformative characters) and the 50% majority-rule consensus tree is shown in

Figure 3.5. Viola hondoensis (from Korea) appear as paraphyletic. One accession is sister to all other species (except V. scandens) and the other sample is part of the large polytomy including the remaining violet ingroup. Further resolution was not recovered and a basal polytomy including three highly supported clades was shown in all analyses (Figure 3.5). A first clade (BS=95%) included V. jaubertiana, V. odorata and all Western Mediterranean accessions of V. alba subsp. dehnhardtii samples. However, the sister relationships of these three taxa were not resolved using distance-based phylogenetic reconstructions. The monophyly of V. alba subsp. dehnhardtii and V. odorata was only moderately supported

(63% and 62% BS, respectively, in each case). A second clade (BS=92%) comprising members of the V. alba complex (V. alba subsp. alba, V. alba subsp. cretica) plus a Eastern

Mediterranean accession of V. alba subsp. dehnhardtii, and V. sintenisii V. alba subsp. cretica and V. sintenisii appeared as sister taxa, but this relationship was weakly supported

(BS= 55%), and was not supported in the NJ tree. The third clade included monophyletic V. suavis (BS=99%) samples.

90 Chapter 3. The hybrid origin of the endemic Viola x balearica Results

Figure 3.5. 50% majority-rule consensus tree of 1,620 most parsimonious trees of 172 steps from the analysis of members of Viola subsection Viola (CI = 0.89, RI = 0.86, RC = 0.77). Viola scandens served as the outgroup based on previous analyses (Ballard et al., 1999). Combinable components appearing over 50% in the consensus tree are shown above branches. Significant bootstrapping values obtained in parsimony and distance (Kimura two- parameter and NJ) reconstructions are shown below branches.

3.5.4. Sterility evaluation of V. x balearica

The difficulty in obtaining flowers from the four wild known individuals of V. x balearica prevented to evaluate fertility in all leaf-sampled individuals. However, cleistogamous flowers were obtained from a pot-planted specimen obtained from stolons of the specimen from the Botanical Garden of Sóller. Since most of explored material showed similar results, no quantitative measures were taken. Qualitative results suggested that ovules were aborted (Figure 3.6A-C). Also, most of pollen was malformed (Fig 3.6E), but rare ones were non-aborted (Fig 3.6D).

91 Chapter 3. The hybrid origin of the endemic Viola x balearica Results

Figure 3.6. Evaluation of ovule and pollen fertility in cleistogamous flowers of V. x balearica. A-C: Optical epifluorescence microscope images with aniline blue stain. Callose fluoresces green. A) Ovary of the cleistogamous flower. B) Ovules. C) Ovule detail. D,E: Phase-contrast microscope images, with haematoxilin stain. D) One of the scarce non-aborted pollen grains detected. E) Pollen grains more commonly detected, showing malformation.

92 Chapter 3. The hybrid origin of the endemic Viola x balearica Discussion

3.6. DISCUSSION

3.6.1. Hybridization between V. jaubertiana and V. alba subsp. dehnhardtii.

Field prospection in the Gorg Blau locality resulted in no suspicions of the presence of any other known taxa of the sect. Viola but those suggested as putative parental taxa: V. alba subsp. dehnhardtii, and the endemic V. jaubertiana. Moreover, chromosome counts done in that locality for V. jaubertiana (Schmidt, 1961b; Castro & Rosselló, 2006), V. alba subsp. dehnhardtii (Schmidt, 1961b), and putative hybrids (Schmidt, 1961b; Castro &

Rosselló, unpublished data) report no other number but 2n=20. This stresses no discordances in chromosome number in the putative parental species and the deviant plants.

The evaluation of the nuclear and chloroplast DNA markers shown by the rupicolous, hairy individuals resembling V. jaubertiana refute the hypotheses that they are

(i) morphological variants of the latter, (ii) the result of incomplete primary divergence between V. jaubertiana and V. alba subsp. dehnhardtii, (iii) another undetected violet species from subsection Viola. On the contrary, available evidence strongly suggest that these individuals are hybrids generated by the cross between the Balearic endemic V. jaubertiana and the Mediterranean V. alba subsp. dehnhardtii.

Violets usually have weak barriers to interspecific gene flow. Accordingly, hybridization has been frequently reported among species (e.g. Valentine, 1949, 1958,

1975; Schmidt, 1961b; Krahulcova et al., 1996; Neuffer et al., 1999), and it has been credited with an important role in hybridization and introgression in past and present evolution within subsection Viola (Marcussen & Borgen, 2000). In fact, the presence of many duplications in gene coding isozymes in members of subsection Viola led to the

93 Chapter 3. The hybrid origin of the endemic Viola x balearica Discussion suggestion that this subsection, and the entire section Viola, is paleotetraploid and of allopolyploid origin (Marcussen & Borgen, 2000).

In Viola, the identification of hybrids has been mainly based on the evaluation of morphological, cytogenetic, and embryological features (e.g. Schmidt, 1961b; Erben, 1996;

Siuta et al., 2005). In sharp contrast, the use of molecular markers to identify reticulation in violets has been to date scarce. Only a few studies using isozyme (Marcussen & Borgen,

2000; Marcussen et al., 2001) and RAPD (Oh et al., 1998; Neuffer et al., 1999) markers have documented (i) primary- and later-generation wild hybrids between homoploid species in Viola (V. alba, V. chaerophylloides, V. collina, V. hirta, V. mandshurica, V. odorata, V. patrini, and V. variegata), and (ii) recurrent hybridization and backcrossing between diploid and tetraploid species (V. reichenbachiana and V. riviniana).

The ribosomal region, composed by the 18S, 5.8S, and 25S genes, two internal spacers (ITS1 and ITS2) and the intergenic spacer (IGS), is formed by hundreds to thousands of tandem copies that are located in one or, more usually, several ribosomal loci.

Due to functional constraints the ribosomal multigene family usually evolves in concert

(Arnheim, 1983) and all copies within and among ribosomal loci are expected to be homogenized throughout genomic mechanisms of turnover like gene conversion and unequal crossing over (Dover, 1994). These mechanisms of concerted evolution tend to homogenize sequences in genomic nrDNA arrays in a process that operates within whole reproductive groups. However, this process is not always operating or found completed.

When different ITS repeats belonging to homeologous loci are merged within a single genome via hybridization (including allopolyploidy) or introgression, the speed and direction of homogenization cannot be predicted and is not consistent across different descendant

lineages (Alvarez & Wendel, 2003). F1 hybrids typically show ITS additivity, but biased fast

homogenization towards a parental ribosomal sequence is already detectable in artificial F2

94 Chapter 3. The hybrid origin of the endemic Viola x balearica Discussion hybrids in Armeria (Fuertes Aguilar et al., 1999). Thus, the positive identification of parental lineages in hybrid derivatives depends on whether the different ITS repeats are maintained without recombination or homogenization (e.g. Soltis and Soltis, 1991; Ritland et al., 1993).

When the different repeats undergo some degree of homogenization giving rise to chimeric sequences (e.g. Buckler et al., 1997; Nieto Feliner et al., 2004), or one of the parental repeats becomes dominant within the new genome (Wendel et al., 1995), the identification of plant hybrids and their progenitors using these ribosomal markers remains elusive.

The presence of two divergent ITS repeats in the hairy specimens of V. jaubertiana, showing the additive pattern of V. alba subsp. dehnhardtii and V. jaubertiana, strongly supports its hybrid origin from these progenitors, as previously hypothesized on morphological and cytogenetic evidence (Chodat, 1924; Schmidt, 1961b). The cpDNA data, of maternal transmission in Viola (Harris & Ingram, 1991), but this inheritance has been only checked in a single species, V. tricolor. If this maternal inheritance is proven to be the rule in the genus, then cpDNA data suggest that the gene flow between V. alba subsp. dehnhardtii and V. jaubertiana is unidirectional and identifies the endemic V. jaubertiana as the unique ovule donor. The additivity of the ITS sequences, together with the pollen and

ovule sterility, suggests that the sampled individuals are primary F1 hybrids, whereas no trace of introgressive hybridization or hybrid zone has been evidenced by the nuclear and plastidial markers used. Low hybrid fertility is apparently the most important constraint in allowing introgression with V. alba subsp. dehnhardtii and V. jaubertiana. This contrasts greatly with other studies reporting introgressive hybrids of V. riviniana and V. reichenbachiana, two species differing in ploidy level but sharing one genome (Neuffer et al., 1999).

Hybridization processes at Gorg Blau affecting V. alba subsp. dehnhardtii and V. jaubertiana are certainly not of recent origin. In fact, herbarium specimens identified

95 Chapter 3. The hybrid origin of the endemic Viola x balearica Discussion morphologically as hybrids of the two species were collected at this site as early as the

XIXth century, dating back about 130 years (Gorg Blau, 4 June 1881, Barbey, G). Collecting dates from herbarium sheets indicate that, since then, the hybrid was at least present in

1904 (Mus et al., 2000), 1924, and 1947 (unpublished data). These dates do not necessarily indicate simply an occasional presence of hybrids as individuals have been observed yearly since 1991 at Gorg Blau (Mayol, Mus, and Rosselló, unpublished data).

What should be stressed is that all reports, field observations, and labels of herbarium specimens, agree that the hybrid plants are always rupicolous, and grow near V. jaubertiana individuals. Both V. alba subsp. dehnhardtii and V. jaubertiana show secondary dispersal via myrmecochory (Beattie & Lyons, 1975), because the seeds have an elaiosome that attracts ants that collect the seeds and may disperse them a few meters away from the source plant. This strongly suggests that successful gene flow between both species is probably unidirectional, and that all hybrid individuals have originated from V. jaubertiana mother plants, in accordance with the results presented here for extant hybrid individuals. However, the occurrence of hybridization in the opposite direction cannot be completely discarded. In this case, the hybrid could be a weak competitor and therefore not persistent in the habitat of V. alba subsp. dehnhardtii, with more dense vegetation.

Moreover, Schmidt (1961b) reported a non-hairy hybrid from Gorg Blau based on meiotic irregularities, stressing this to be the reciprocal situation to the hybrid reported by L. Chodat

(1924). However, the author did not report the habitat occupied for that individual. The biological reasons behind this apparent unidirectional gene flow are by no means evident. In fact, asymmetric hybridization has been postulated for the pair of diploids V. hirta and V. collina and it has been assumed that V. collina alleles are more easily introgressed into V. hirta than the reverse (Marcussen et al., 2001). However, introgression takes place in both

96 Chapter 3. The hybrid origin of the endemic Viola x balearica Discussion parental populations (diploid V. reichenbachiana and tetraploid V. riviniana) of hybrid swarms in Central Germany (Neuffer et al., 1999).

Members of subsect. Viola live in habitats that are often disturbed, and at least some

European species seem to be favoured by direct or collateral human activity (Marcussen &

Nordal, 1998). Our study has shown that in violets, hybridization is not restricted to either ecotones, disturbed and artificial habitats, as usually reported. In fact, hybrids between V. alba subsp. dehnhardtii and V. jaubertiana are present in pristine habitats and are restricted to one of the parental environments (calcareous cliffs).

3.6.2. The phylogenetic relationships of V. jaubertiana

The Balearic endemic violet shows no strong morphological affinity with any species of subsection Viola (Marcussen & Borgen, 2000; Mus & Rosselló, unpublished data).

Several authors have postulated, on morphological (Schmidt, 1961b) and isozyme grounds

(Marcussen & Borgen, 2000), that V. jaubertiana represents a relict and isolated lineage among members of subsection Viola. Interestingly, the phylogenetic analyses of ITS sequences from available members of this subsection refute this view. The Balearic endemic never stands in a basal position in the phylogenetic analyses. Rather, it forms a well supported (terminal) clade with the sympatric V. odorata and V. alba subsp. dehnhardtii. However, further resolution between these three taxa is not possible due to the high sequence similarity shared, and thus the identification of the sister taxa of the Balearic taxa remains elusive. The choice of other molecular markers showing faster mutation rates are needed to solve this issue. Indeed, V. odorta from Minorca (considered as the endemic

V. stolonifera by some authors; e.g., Fraga et al., 2004) resulted in the same clade as other

European and American V. odorata accessions, thus stressing its close affinity.

97 Chapter 3. The hybrid origin of the endemic Viola x balearica Discussion

Evolutionary changes leading to contrasting morphological divergences could obscure sister relationships if not properly analyzed, as demonstrated in some species showing a high level of morphological change (e.g. Sytsma & Smith, 1992). The endemic flora of the Balearic archipelago is believed to contain a large number of old lineages

(palaeoendemic species, Contandriopoulos & Cardona, 1984). However, phylogenetic analysis of some genera including several seminal species has revealed that their stenochory is a consequence of a recent origin and no hypotheses relating them to relict lineages could be supported (e.g. Hypericum balearicum L., Crokett et al., 2004;

Helichrysum crassifolium (L.) G. Don, Galbany-Casals et al., 2004; Lysimachia minoricensis

J.J. Rodr., Rosselló, unpublished data). Available data suggest that the same also holds true for V. jaubertiana.

This species is restricted to chasmophyte habitats, including step-crevices, vertical cliffs, overhangs and step-slopes. These environments are usually characterized by singular plant communities of very limited distribution (e.g. Simmons et al., 1998). These escarpment biotopes are exceptional areas of species richness, hotspots of endemism and refuge centers for relict taxa. As a rule, exclusive chasmophytes tend to be suffruticose chamaephytes, apparently adapted to insect pollination, with wind being the prevalent method of dissemination of seeds and fruits (Davis, 1951). However, this biological syndrome does not apply to V. jaubertiana. It is likely that the speciation events leading to this endemic lineage were linked to concomitant habitat shifts, from a woodland habitat typical of the related V. alba subsp. dehnhardtii and V. odorata, to cliff environments. Thus, species boundaries could be later reinforced by the acquisition of sterility barriers restricting, but not annulling, interspecific gene flow among these close relatives.

98 Chapter 3. The hybrid origin of the endemic Viola x balearica References

3.7. REFERENCES

Álvarez I & Wendel JF. 2003. Ribosomal ITS sequences and plant phylogenetic inference. Molecular

Phylogenetics and Evolution, 29: 417–434.

Arnheim N. 1983. Concerted evolution of multigene families. In: Nei M, Koehn R (eds). Evolution of Genes and

Proteins. Sinauer: Sunderland.

Arnold ML, Bulger MR, Burke JM, Hempel AL & Williams JH. 1999. Natural hybridization: How low can you go and

still be important? Ecology, 80: 371-381.

Avise JC. 1994. Molecular markers, natural history and evolutiuon. Chapman & Hall: New York.

Ballard HE Jr, Sytsma KJ & Kowal RR. 1999. Shrinking the violets: phylogenetic relationships of infrageneric

groups in Viola (Violaceae) based on Internal Transcribed Spacer DNA sequences. Systematic Botany 23:

439-458.

Ballard HE. 1996. Phylogenetic relationships and infrageneric groups in Viola (Violaceae) based on morphology,

chromosome numbers, natural hybridization and ITS sequences. PhD. Thesis, University of Wisconsin:

Madison.

Barceló F. 1879-1881. Flora de las Islas Baleares. Editorial Moll: Palma de Mallorca.

Beattie AJ & Lyons N. 1975. Seed dispersal in Viola (Violaceae): adaptations and strategies. American Journal of

Botany, 62: 714-722.

Beattie AJ. 1969. Pollination ecology of Viola: I. Contents of stigmatic cavities. Watsonia, 7: ?

Beattie AJ. 1971. Pollination mechanisms in Viola. New Phytologist 70: 343-360.

Becker W. 1925. Viola L. In: Engler A (ed). Die natürlichen Pflanzenfamilien, vol. 21, 2nd ed.: 363-376. Wilhelm

Engelmann: Leipzig.

Bolòs O de & Vigo J. 1974. Notes sobre taxonomia i nomenclatura de plantes, I. Butlletí de la Institució Catalana

d’Història Natural, 38: 61-89.

Bolòs O de, Vigo J. 1990. Flora dels Països Catalans. Volum II (Crucíferes – Amarantàcies). Ed. Barcino:

Barcelona.

Bonafè F. 1979. Flora de Mallorca, vol. 3. Editorial Moll: Palma de Mallorca.

Buckler ESI, Ippolito A & Holtsford TP. 1997. The evolution of ribosomal DNA: divergent paralogous and

phylogenetic implications. Genetics, 145: 821–832.

99 Chapter 3. The hybrid origin of the endemic Viola x balearica References

Burnat E & Barbey W. 1882. Notes sur un voyage botanique dans les îles Baléares et dans la province de Valence

(Espagne). Mai–Juin, 1881. Genève & Bâle: H. George.

Castro M & Rosselló JA. 2006. New chromosome numbers for plant taxa endemic to the Balearic Islands. Folia

Geobotanica, 41: 433-451.

Chodat L. 1924. Contributions à la Géo-Botanique de Majorque. Thèse: Genève.

Clausen J. 1929. Chromosome number and the relationship of some North American species of Viola. Annals of

Botany, 43: 741-764.

Clausen J. 1964. Cytotaxonomy and distributional ecology of western North American violets. Madroño, 17: 173-

204.

Contandriopoulos J & Cardona MA. 1984. Caractère original de la flore endémique des Baléares. Botanica

Helvetica, 94: 101–131.

Crockett SL, Douglas AW, Scheffler BE & Khan IA. 2004. Genetic profiling of Hypericum L. (St. John's Wort)

species by nuclear ribosomal ITS sequence analysis. Planta Medica, 70: 929-935.

Cronquist AJ. 1981. An Integrated System of Classification of Flowering Plants. Columbia University Press: New

York.

Davis PH. 1951. Cliff vegetation in the Eastern Mediteranean. Journal of Ecology, 39: 63-93.

Dizerbo AH. 1967. Observations sur Viola odorata L., Viola hirta L. et lur hybride Viola permixta Jord. dans le

Massif Armoricain I-II. Bulletin de la Société Scientifique de Bretagne, 42: 113-124 & 215-223.

Dizerbo AH. 1968. Observations sur Viola odorata L., Viola hirta L. et lur hybride Viola permixta Jord. dans le

Massif Armoricain III-IV. Bulletin de la Société Scientifique de Bretagne, 43: 71-79 & 225-236.

Dover G. 1994. Concerted evolution, molecular drive and natural selection. Current Biology, 4: 1165–1166.

Doyle J & Doyle JJ. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical

Bulletin, 19: 11–15.

Ellstrand NC & Elam DR. 1993. Population genetic consequences of small population size: implications for plant

conservation. Annual Review in Ecology and Systematics, 24: 217–242.

Erben M. 1996. The significance of hybridization on the forming species in the genus Viola. Bocconea, 5: 113-118.

Fraga i Arguimbau P, Mascaró Sintes C, Carreras Martí D, Garcia Febrero O, Pallicer Allès X, Pons Gomila M,

Seoane Barber M & Truyol Olives M. 2004. Catàleg de la flora vascular de Menorca. Col·lecció recerca 9.

Institut Menorquí d’Estudis: Menorca.

100 Chapter 3. The hybrid origin of the endemic Viola x balearica References

Fuertes Aguilar J, Rosselló JA & Nieto Feliner G. 1999. Nuclear ribosomal DNA (nrDNA) concerted evolution in

natural and artificial hybrids of Armeria (Plumbaginaceae). Molecular Ecology, 8: 1341–1346.

Galbany-Casals M, Garcia-Jacas N, Susana A, Sáez L & Benedí C. 2004. Phylogenetic relationships in the

Mediterranean Helichrysum (Asteraceae, Gnaphalieae) based on nuclear rDNA ITS sequence data.

Australian Systematic Botany ,17: 241-253.

Gershoy A. 1934. Studies in North American violets. III. Chromosome numbers and species characters. Vermont

Agricultural Experiment Station Bulletin, 367: 1-91.

Gil-ad NL. 1997. Systematics of Viola subsection Boreali-Americanae. Boissiera, 53: 325-338.

Guinochet M & Lefranc M. 1972. In: Löve A, ed. IOPB Chromosome number reports. XXXVII. Taxon 21: 497.

Haesler I. 1982. Lectotypisierung der Arten Viola hirta L. und Viola odorata L. Mitt.Bot. München 18: 289-296.

Harris SA & Ingram R. 1991. Chloroplast DNA and biosystematics: the effects of intraspecific diversity and plastid

transmission. Taxon 40: 393-412.

Kimura M. 1980. A simple method for estimating evolutionary rate of base substitutions through comparative

studies of nucleotide sequences. Journal of Molecular Evolution, 16: 111-120.

Knoche H. 1922. Flora Balearica. Etude phytogéographique sur les îles Baléares, 2: Montpellier.

Krahulcova A, Krahulec F & Kirschner J. 1996. Introgressive hybridization between a native and and introduced

species: Viola lutea subsp. sudetica versus V. tricolor. Folia Geobotanica, 31: 219-244.

Kühbier H. 1978. Beiträge zur Flora der Pityusen. Veröffentlichungen aus dem Überseemuseum Bremen, Museum

für Natur-, Völker- und Handelskunde, Reihe A. Naturwissenschaften, 5: 1-23.

Malécot V, Marcussen T, Munzinger J, Yockteng R & Henry M. 2007. On the origin of the sweet-smelling Parma

violet cultivars (Violaceae): wide intraspecific hybridization, sterility, and sexual reproduction. American

Journal of Botany, 94: 29-41.

Marcussen T, Borgen L & Nordal I. 2001. Viola hirta and its relatives in Norway. Nordic Journal of Botany, 21: 5-17.

Marcussen T & Borgen L. 2000. Allozymic variation and relationships within subsection Viola (Violaceae). Plant

Systematics and Evolution, 223: 29–57.

Marcussen T & Nordal I. 1998. Viola suavis, a new species in the nordic flora, with analyses of the relation to other

species in the subsection Viola (Violaceae). Nordic Journal of Botany, 18: 221-237.

Marcussen T. 2003. Evolution, phylogeography, and taxonomy within the Viola alba complex (Violaceae). Plant

Systematics and Evolution, 237: 51-74.

101 Chapter 3. The hybrid origin of the endemic Viola x balearica References

Marès P & Vigineix G. 1880. Catalogue raisonné des plantes vasculaires des Iles Baléares. Edit. G. Masson:

Paris.

Martin FW. 1959. Staining and observing pollen tubes in the style by means of fluorescence. Stain Technology, 34,

125–128.

Mereďa, P Jr, Hodálová I, Mártonfi P & Kolarčik V. 2006. [Reports (17–22)]. In: Mráz P (ed). Chromosome number

and DNA ploidy level reports from Central Europe - 2. Biologia (Bratislava), 61: 116–117.

Miyaji Y. 1929. Studien über die Zahlen-verhältnisse der Chromosomen bei der Gattung Viola. Cytologia, 1: 28-58.

Moore DM. 1959. Population studies on Viola lactea Sm. and its wild hybrids. Evolution, 13: 318-332.

Muñoz Garmendia F, Montserrat P, Laínz M & Aldasoro JJ. 1993. Viola L. In: Castroviejo S, Aedo C, Cirujano S,

Laínz M, Montserrat P, Morales R, Muñoz Garmendia F, Navarro C, Paiva J, Soriano C (eds). Flora iberica

3: 276-317. Real Jardín Botánico, CSIC: Madrid.

Mus M, Sáez L & Rosselló JA. 2000. Viola barceloi Nyman, V. barceloi L. Chodat and V. suavis var. barceloi O.

Bolos & Vigo: too many eponyms or too many taxa? Taxon, 49: 785-788.

Nadot S, Ballard HE Jr, Creach JB & Dajoz I. 2000. The evolution of pollen heteromorphism in Viola: a

phylogenetic approach. Plant Systematics and Evolution, 223: 155.

Neuffer B, Auge H, Mesch H, Amarell U & Brandl R. 1999. Spread of violets in polluted pine forests: morphological

and molecular evidence for the ecological importance of interspecific hybridization. Molecular Ecology, 8:

365-377.

Nieto Feliner G, Gutiérrez Larena B & Fuertes Aguilar J. 2004. Fine scale geographical structure, intra-individual

polymorphism and recombination in nuclear ribosomal internal transcribed spacers in Armeria

(Plumbaginaceae). Annals of Botany, 93: 189–200.

Nordal I & Jonsell B. 1998. A phylogeographic analysis of Viola rupestris: Three postglacial immigration toutes into

the nordic area? Botanical Journal of the Linnean Society, 128: 105-122.

Nyman CF. 1889. Conspectus florae europaeae. Suppl. 2: 45. Orebro, Sueciae.

Oakley CG, Moriuchi KS & Winn AA. 2007. The maintenance of outcrossing in predominantly selfing species: ideas

and evidence from cleistogamous species. Annual Review in Ecology, Evolution and Systematics, 38: 437-457.

Oh Ko MK, Lee CH. 1998. Identification of the series-specific random amplified polymorphic DNA markers of Viola

species. Plant Breeding, 117: 295-296.

Okamoto M, Okada H & Ueda K. 1993. Morphology and chromosome number of Viola pilosa, and its systematic

position. Taxon, 42: 781-787.

102 Chapter 3. The hybrid origin of the endemic Viola x balearica References

Orell J & Romo AM. 1991. The genus Viola in the Balearic Islands. Butlletí de la Institució Catalana d’Història

Natural, 59: 81-88.

Porta P. 1887. Stirpium in insulis Balearium anno 1885 collectarum enumeratio. Nuovo Giornale Botanico Italiano

(vol. XIX), 4.

Ritland C.E., R.K. Ritland & N.A. Straus, 1993. Variation in the ribosomal internal transcribed spacers (ITS1 and

ITS2) among eight taxa of the Mimulus guttatus species complex. Molecular Biology and Evolution, 10:

1273–1288.

Rivas Martínez S, Costa M, Soriano P, Pérez R, Llorens L & Rosselló JA. 1992. Datos sobre el paisaje vegetal de

Mallorca e Ibiza (Islas Baleares, España). Itinera Geobotanica, 6: 5-89.

Rodríguez JJ. 1878 Additions à la flore de Minorque. Bulletin de la Société Botanique de France, 25: 238.

Rosselló JA & Sáez L. 2001. Index Balearicum: an annotated check-list of the vascular plants described from the

Balearic Islands. Collectanea Botanica (Barcelona), 25: 1-192.

Saitou N & Nei M. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees.

Molecular Biology and Evolution 4: 406-425.

Schmidt A. 1961a. Zytotaxonomische Untersuchungen an Viola-Arten der Sektion Melanium. Berichte der

Bayerischen Botanischen Gesellschaft, 34: 93–95.

Schmidt A. 1961b. Zytotaxonomische Untersuchungen an europäischen Viola-Arten der Sektion Nomimium.

Oesterreichische Botanische Zeitschrift. Gemeinnütziges Organ für Botanik... ,108: 20-88.

Schöfer G. 1954. Untersuchungen über die Polymorphie einheimischer Veilchen. Planta, 43: 537-565.

Shaw J, Lickey EB, Beck JT, Farmer SB, Liu W, Miller J, Siripun KC, Winder CT, Schilling EE & Small RL. 2005.

The tortoise and the hare II: relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic

analyses. American Journal of Botany, 92: 142-166.

Simmons RE, Griffin M, Griffin RE, Marais E & Kolberg H. 1998. Endemism in Namibia: Patterns, processes and

predictions. Biodiversity and Conservation, 7: 513-530.

Siuta A, Bozek M, Jedrzejczyk M, Rostanski A & Kuta E. 2005. Is the blue zinc violet (Viola guestphalica Nauenb.)

a taxon of hybrid origin? Evidence from embryology. Acta Biologica Cracoviensia, Series Botanica, 47:

237–245.

Soltis PS & Soltis DE. 1991. Multiple origins of the allotetraploid Tragopogon mirus (Compositae): rDNA evidence.

Systematic Botany, 16: 407–413.

103 Chapter 3. The hybrid origin of the endemic Viola x balearica References

Swofford DL. 2002. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods), 4.0 edn, Sinauer:

Sunderland.

Sytsma KJ & Smith JF. 1992. Molecular systematics of Onagraceae: examples from Clarkia and Fuchsia. In: Soltis

PS, Soltis DE, Doyle JJ, eds. Molecular Systematics of Plants. Chapman & Hall: New York & London.

Taberlet P, Gielly L, Patou G & Bouvet J. 1991. Universal primers for amplification of three non-coding regions of

chloroplast DNA. Plant Molecular Biology, 17: 1105–1109.

Tchourina O. 1909. Note sur le Viola jaubertiana Marès [sic]. Bulletin de la Société Botanique de Genève (serie 2),

1: 204-207.

Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F & Higgins DG. 1997. The ClustalX windows interface:

flexible strategies for multiple sequencealignment aided by quality analysis tools. Nucleic Acids Research,

25: 4876-4882.

Valentine DH, Merxmüller H & Schmidt A. 1968. Viola. In: Tutin TG, Heywood VH, Burges NA, Valentine DH,

Walters SM, Webb DA (eds). Flora Europaea, vol. 2.: Cambridge.

Valentine DH. 1949. The experimental taxonomy of two species of Viola. New Phytologist ,49: 194-212.

Valentine DH. 1958. Cytotaxonomy of the rostrate violets. Proceedings of the Linnean Society of London, 169:

132-134.

Valentine DH. 1962. Variation and evolution in the genus Viola. Preslia, 34: 190-206.

Valentine DH. 1975. Viola L. In: Stace CA (ed). Hybridization and the Flora of the British Isles. Academic Press:

London.

Wendel JF, Schnabel A & Seelanan T. 1995. Bidirectional interlocus concerted evolution following allopolyploid

speciation in cotton (Gossypium). Proceedings of the National Academy of Sciences, 92: 280–284.

White TJ, Bruns T, Lee S & Taylor J. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for

phylogenetics. In: Innis M, Gelfand D, Sninsky J, White T (eds). PCR Protocols: A guide to methods and

applications. Academic Press: San Diego.

Yockteng R, Ballard HE Jr, Mansion G, Dajoz I & Nadot S. 2003. Relationships among panises (Viola section

Melanium) investigated using ITS and ISSR markers. Plant Systematics and Evolution, 241: 153-170.

Yoo K-O, Jang S-K, Lee W-T. 2005. Phylogeny of Korean Viola based on ITS sequences. Sigmul Bunryu Hag-

hoeji, 35: 7-23.

104 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium

CASE STUDY II Hybridization involving the endemic Lotus fulgurans

Chapter 4 Hybridization processes in the Balearic Islands between the endemic Lotus fulgurans (Porta) D.D.Sokoloff and the Mediterranean distributed Lotus dorycnium L.

page 4.1. Introduction 107 4.1.1. The genus Lotus sect. Dorycnium 107 4.1.2. The species of Lotus sect. Dorycnium in the Balearic Islands 109 4.2. The hybrid and its putative parental species 111 4.2.1. Lotus x minoricensis (Fabaceae), a new hybrid from the Balearic Islands 111 4.2.2. The putative parental species 113 4.2.3. The Na Macaret locality 114 4.3. Objectives 116 4.4. Materials and methods 118 4.4.1. Plant sampling 118 4.4.2. Morphometric analysis 118 4.4.3. Hybrid index 120 4.4.4. DNA extraction 121 4.4.5. Nuclear ribosomal ITS sequences 122 4.4.6. Restriction analysis 123 4.4.7. Chloroplast DNA sequences 123 4.5. Results 125 4.5.1. Morphometric analysis 125 4.5.2. Molecular markers 130 4.5.2.1. ITS variation and genotyping 130 4.5.2.2. CpDNA polymorphisms 133

105 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium

page 4.6. Discussion 136 4.6.1. Overall evidence suggest interspecific gene flow between L. fulgurans 136 and L. dorycnium 4.6.2. CpDNA variation better explains a related origin of both species rather 137 than hybridization 4.6.3. Hybridization and the conservation of the endangered L. fulgurans 139 4.7. References 142

THIS CHAPTER HAS BEEN PARTIALLY PUBLISHED AS TWO ORIGINAL PAPERS:

Conesa MÀ, M Mus & JA Rosselló. 2006. Lotus x minoricensis (Fabaceae), a new hybrid from the

Balearic Islands. Flora Montiberica 34: 25-27.

Conesa MÀ, M Mus & JA Rosselló. 2010. Who threatens who?: Natural hybridization between

Lotus dorycnium and the island endemic L. fulgurans (Fabaceae). Biological Journal of

the Linnean Society. In press.

106 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Hybrid and parental species

4.1. INTRODUCTION

As stated in the Chapter I, the most used theoretical explanation for interspecific hybridization involves two sympatric (or parapatric) species inhabiting different habitats, and hybridizing in the zones of contact between them (e.g. Anderson, 1948; Arnold, 1997).

Therefore, in spite of cases of very competitive hybrids in their parental habitats (e.g.,

Moore, 1977; Milne et al., 2003), parental species are mainly maintained pure in most of its distribution area. Indeed, hybrid individuals are mostly restricted to the zones of contact of both habitats, which are supposed to be intermediate habitats (e.g., Anderson, 1948).

This is thought to be roughly the case of Lotus x minoricensis M.À.Conesa, Mus &

Rosselló, a hybrid described from Minorca as a result of this thesis (see below). This nothospecies is considered a hybrid between the endemic Lotus fulgurans (Porta)

D.D.Sokoloff [≡Dorycnium fulgurans (Porta) Lasen], and the Mediterranean Lotus dorycnium L. [≡Dorycnium pentaphyllum Scop.], (Conesa et al., 2006). Thus, this second case study refers to a putative case of natural interspecific hybridization, where all intermediate individuals occur in the narrow and defined zone of contact between two different habitats (ecotone), each one inhabited by a putative parental species.

4.1.1. The genus Lotus sect. Dorycnium

Species of the genus Dorycnium Mill. have been included in, and segregated from the genus Lotus L. several times from Linnaeus (see Arambarri, 2000, and references therein). Due to the stressed monophyly of the genus (within the limits suggested by

Sokoloff, 2003a, 2003b), recent molecular phylogenetic studies dealing with Dorycnium

107 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Hybrid and parental species suggested again its inclusion into Lotus (e.g. Allan et al., 2003; Allan et al., 2004;

Degtjareva et al., 2006),

These studies showed that there is no consistent molecular differentiation between the genera Lotus and Dorycnium, since the studied Mediterranean species of Dorycnium

(D. pentaphyllum and D. hirsutum (L.) Ser.) are more closely related to Lotus subgen. Lotus than to the Canary Island Lotus species (i.e., Lotus subgen. Pedrosia (Lowe) Brand; Allan et al., 2004). Furthermore, Dorycnium can not be separated from Lotus at the generic level when adding morphological data to the molecular data for analyses (see Degtjareva et al.,

2006). In this latter work, authors assume different sections into the genus Lotus, and guess for a future combination of the former sections of the genus Dorycnium (Rikli, 1901), sect.

(Eu)Dorycnium and sect. Bonjeanea, since constitute a monophyletic group, but not sect.

Canaria, which results clearly different (Degtjareva et al., 2006). In other molecular phylogenetic analyses (Allan et al., 2003) the analyzed species of the genus Dorycnium did not form a clade, and in morphological cladistic studies (Arambarri, 2000) it was is nested in the Old World Lotus clade. The monophyly of the genera Tetragonolobus, Dorycnium and the Old World Lotus is also supported in other molecular phylogenetic studies (e.g. Allan &

Porter, 2000; Allan et al. 2003).

Thus, recent revisions led to the nomenclatural combination of most of the taxa of

Dorycnium into Lotus (Sokoloff, 2003a). With regards to the Mediterranean distributed species, Ball (1968) accepted several taxa related to Dorycnium pentaphyllum Scop. occurring in Europe at subspecific level, including subsp. pentaphyllum, subsp. germanicum

(Gremli) Gams, subsp. gracile (Jord.) Rouy, and subsp. herbaceum (Vill.) Rouy. From these taxa, only the first and third are considered to exist in the SW Europe (Ball, 1968). However, the inclusion of D. pentaphyllum subsp. gracile into the genus Lotus is, apparently, still not formal. Thus, when stating L. dorycnium “sensu lato” (s.l.) from here on in this thesis, subsp.

108 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Hybrid and parental species gracile will be included. In addition, a population ascribed to subsp. gracile was included in this study. Nevertheless, further revisions of this taxon are still needed.

For the reasons stressed above, this thesis follows the criteria of Sokoloff (2003a), who combined most Dorycnium species into the genus Lotus. He stressed that morphological grounds used to separate Dorycnium from Lotus are equivocal. Therefore, species tackled in this chapter are considered into Lotus sect. Dorycnium.

Table 4.1. Sections of the genus Lotus L. (following Degtjareva et al., 2006) comprising the species previously included in the genus Dorycnium existing in the SW Europe and Balearic Islands (Ball, 1968). Occurring subspecies (Ball, 1968) are also shown.

Sections into Lotus L. Current taxon Synonymy into Dorycnium Mill. L. sect. Bonjeanea L. hirsutus L. D. hirsutum (L.) Ser. (Rchb.) D.D.Sokoloff L. rectus L. D. rectum (L.) Ser. L. strictus Fisch. & C.A. Mey D. strictum Lassen L. sect. Dorycnium (Mill.) L. dorycnium L. subsp. dorycnium D. pentaphyllum Scop. subsp. pentaphyllum D.D.Sokoloff ? D. pentaphyllum Scop. subsp. gracile (Jord.) Rouy L. fulgurans (Porta) D.D.Sokoloff D. fulgurans (Porta) Lassen

4.1.2. The species of Lotus sect. Dorycnium in the Balearic Islands

The endemic L. fulgurans was described from Minorca as Anthyllis fulgurans Porta

(Porta, 1887), and included into the genus Dorycnium Mill. nearly a century after -i.e.

Dorycnium fulgurans (Porta) Lassen (Lassen, 1979)-, maintaining the taxon at specific level.

Afterwards, it was considered a subspecies of the widespread Mediterranean species -i.e.,

Dorycnium pentaphyllum subsp. fulgurans (Porta) Cardona, L. Llorens et Sierra (Cardona et al., 1983)-. Reasons argued by the latter authors were primarily that characters provided by

Lassen (1979) were not enough to consider it at specific level, since the spinescent and pulvinular habit (and some other vegetative traits) distinguishing the endemic taxon was somehow counteracted for the high similarity in flower and fruit characters. In addition, they

109 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Hybrid and parental species argued that the existence of intermediate individuals was a convincing reason added to the above mentioned morphological traits to consider the endemic taxon at subspecific level into D. pentaphyllum. Thus, the first report of the existence of intermediate individuals between L. dorycnium and L. fulgurans could be that of Llorens (1979), reporting plants of the endemic species showing more lax branches, resembling the widespread

Mediterranean, from Cap de Cavalleria and Macaret (Minorca). Cardona et al. (1983) exposed two hypotheses to explain the origin of these intermediate individuals: hybridization among both taxa and different ecotypes of D. pentaphyllum subsp. pentaphyllum.

The only intraspecific taxon described related to the endemic species is Anthyllis fulgurans var. efulgurata P.Palau, from Cabrera, inhabiting altogether with the type taxon

(Palau, 1954). This variety has also been reported from Minorca in Na Macaret (Cardona et al., 1983), which is one of the two locations where the latter authors reported the intermediate individuals, and the location from where Lotus x minoricensis was described

(Conesa et al., 2006).

With respect to the widespread Mediterranean L. dorycnium s.l., widely occurring in the Balearic Islands, only the subsp. gracile has been reported to exist (Díaz Lifante, 2000;

Fraga et al., 2004; Table 4.2). It has been reported from Spain and France (Ball, 1968), and also from N Africa. There are neither reports of L. fulgurans nearby the known locations of subsp. gracile, nor reports of intermediate individuals affecting subsp. gracile.

Table 4.2. Distribution of the Lotus section Dorycnium taxa occurring in the Balearic Islands. MA, Majorca; CA, Cabrera; ME, Minorca; EI, Ibiza; FO, Formentera. 1 The citation corresponds to Palau (1976). However, the species could not be found during this study sampling.

Taxon MA CA ME EI FO Lotus fulgurans (Porta) D.D. Sokoloff ✔ ✔ ✔ Lotus dorycnium L. [subsp. dorycnium] ✔ ✔1 ✔ ✔ ✔ subsp. gracile (Jord.) Rouy ✔ Lotus x minoricensis M.À.Conesa, Mus & Rosselló ✔

110 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Hybrid and parental species

4.2. THE HYBRID AND ITS PUTATIVE PARENTAL SPECIES

4.2.1. Lotus x minoricensis (Fabaceae), a new hybrid from the Balearic Islands

Lotus fulgurans and L. dorycnium L. remain morphologically distinct throughout most of the Balearic archipelago. However, intermediate plants were reported by Cardona et al. (1983) from two Minorcan populations where the two species grow together (Binidalí and Na Macaret). Results of this study, together with morphological evidence (Table 4.3), are consistent with a hybrid origin of the intermediate plants from Na Macaret..

Accordingly, a new hybrid between L. fulgurans and L. dorycnium, currently restricted to

Minorca, was described (Conesa et al., 2006) as follows:

Lotus x minoricensis Conesa, Mus & Rosselló, nothospec. nov.

[Lotus fulgurans x L. dorycnium]

Diagnosis: Planta perennis, media inter parentes. Differt a L. fulgurans caules haud spinosum, foliis 5, magnis, non emarginatis, inflorescentia pedicellis longioribus et maior in floribus numerus. Foliis brevius, inflorescentia pedicellis brevioris et minus flores L. dorycnium differt.

Holotype: BALEARIC ISLANDS, Minorca: Na Macaret, coastal scrub, 14-06-2004, Conesa,

Fraga & Mus (VAB179606).

111 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Hybrid and parental species

Table 4.3. Morphological differences between the endemic L. fulgurans, the widespread Mediterranean L. dorycnium, and their hybrid, L. x minoricensis. Extracted from Larsen, 1979; Cardona et al.,1983; Mus, 1992; Conesa et al., 2006.

L. fulgurans L. x minoricensis L. dorycnium Habit Spinescent Not spinescent Not spinescent Nº of leaflets 3-4 5 5 Length of central leaflet 4-8 mm 4-9 mm 8-17 mm Shape of leaflets obovate-spathulate elliptic elliptic Leaflet apex deeply emarginated not or slightly emarginated not emarginated Leaf stipules deciduous present and persistent present and persistent Inflorescence axillary ± terminal terminal Length of the inflorescence stalk < 2 mm 2-17 mm 9-54 mm Number of flowers per inflorescence 1-3 2-12 8-16 Floral bracts absent present present

B

Figure 4.1. Habit of Lotus dorycnium (A), L. fulgurans (B), and L. x minoricensis (C). Images of L. dorycnium and L. fulgurans are reproduced from http://herbarivirtual.uib.es/.

112 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Hybrid and parental species

4.2.2. The putative parental species

Lotus fulgurans is an endemic and endangered species from the Gymnesic islands

(i.e., Majorca, Cabrera and Minorca; Alomar et al., 1997; Moreno, 2008), although in

Majorca it only occurs in two narrow and very isolated populations, separated about 80 km

(in the N and W extremes of the island). This species commonly inhabits in open, rocky areas near the coast. However, a northern Majorca population is located at 180 m above sea-level, but with a notorious sea influence. The species habit is a spinescent and dense cushion (conspicuous pulvinular), with woody basis, and hard (non-herbaceous), short and very divaricate (in zig-zag) branches creating a tangle that protects most of the leaves, flowers and fruits (Table 4.3). These reproductive structures are disposed in short, axillary, non-terminal inflorescences (reduced to solitary flowers or very poor inflorescences) included into the branches. This habit denotes an adaptation to resist herbivory, but also to resist abrasion and drying due to the windy conditions of the coastal habitat. In fact, old plants show a dead part of the cushion always positioned towards the sea, acting like a shield protecting the living part of the cushion, as do several more woody coastal endemic species in the Balearic Islands (e.g. Carthamus balearicus (J.J. Rodr.) Greuter, Launaea cervicornis (Boiss.) F.Q. et Rothm.), and even other Fabaceae (Anthyllis histrix (Willk. ex

Barceló) Cardona, Contandr. & Sierra). Interestingly, based in our field observations, the population of Formentor shows more individuals corresponding to the above stated var. efulgurata than most of the localities where it has been cited in Cabrera (Palau, 1954) and

Minorca (Cardona et al., 1983), which could be related to a more lax plant habit regarding to a less sea, wind, and abrasion exposition.

On the other hand, Lotus dorycnium s.l. is a more widespread species occurring

(Ball, 1968) in most of the Mediterranean basin (from Spain to Turkey), and central and east

Europe (from Poland and Krym to Portugal, north of Africa, and also Canary Islands). In the

113 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Hybrid and parental species

Balearic Islands, it occurs in most of the plain and mountain garrigues of Majorca, Ibiza and

Formentera; it is also frequent in Minorca, and it was cited in a single location of Cabrera

(Palau, 1976, included in Pla et al., 1992). In the Balearic Islands, Lotus dorycnium grows in a wide range of environments from sea level to about 1100 m, from dense oak or pine forests and machias, to more open, rocky, and exposed habitats, including littoral scrubs. It is common in garrigues, where uses to be a rounded shrub with erect, dense, and herbaceous habit, and with woody basis. Inflorescences are dense and conspicuous, disposed on terminal branches, totally exposed, showing a white coat covering the plant in bloom.

When subsp. gracile is considered, it is discriminated from subsp. dorycnium by a more herbaceous habit, without conspicuous woody basis, and with long, poorly divaricated, and not dense branches. Leaves are bigger and more separately disposed on the branches, and inflorescences are also conspicuous and terminal. It inhabits in maritime and marshy habitats, on gypsum and marley soils, growing amongst rushes.

4.2.3. The Na Macaret locality

Lotus x minoricensis is only known from Minorca, and it was described from a single location (Na Macaret; Conesa et al., 2006; see above). In this location, a low garrigue on partly sandy soil, inhabited for L. dorycnium, turns into a typical rocky coastal habitat towards sea direction, inhabited for L. fulgurans; and giving rise to a narrow strip with intermediate features, parallel to the coast-line. This strip, about 15 m wide, consists in a hard soil (mostly neither with sand nor nude rock), lacking most of the garrigue species, but with very short –nearly adpressed- woody shrubs broadly scattered all along the strip (e.g.

Pistacia lentiscus L. and Rosmarinus officinalis L.), which make of it a less exposed habitat to the wind and marine spray than the L. fulgurans habitat, however significantly different

114 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Hybrid and parental species from the garrigue habitat. Nowadays, L. x minoricensis is restricted to this narrow strip in Na

Macaret location. This would correspond to the typical hybrid zone described in Anderson

(1948), with hybrid plants strictly located in a mixed habitat (see Chapter 1).

Detail of a Lotus x minoricensis in bloom at Na Macaret location (11-06-2004).

115 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Objectives

4.3. OBJECTIVES

Lotus fulgurans and L. dorycnium are morphologically distinct throughout most of the endemic species distribution, where they are sympatric, and show clearly different habitat preferences. Nevertheless, and as stated above, morphologically intermediate individuals have been reported from Minorca (Llorens, 1979; Cardona et al. 1983; Conesa et al., 2006).

These plants, which show mixed morphological features of L. fulgurans and L. dorycnium, have been documented only from two populations where both species grow in close vicinity

(Binidalí and Na Macaret). Cardona et al. (1983) and Conesa et al. (2006) favoured the hybridization hypothesis to explain the presence of such individuals. However, morphological intermediacy is not always linked to hybridization processes and could be the result of incomplete primary divergence, or could represent infraspecific polymorphisms of labile characters prone to habitat modification (Wilson, 1992). In fact, primary biological evidence supporting the hybrid origin of these plants, such the presence of meiotic abnormalities or low pollen or seed fertility, were not assessed by Cardona et al. (1983) and have not been assessed to the date.

The aim of this Chapter is to define the hybridization processes affecting the endemic L. fulgurans and its magnitude in the island of Minorca. The hypothesis that hybridization occurs between a restricted and endangered species of Lotus and a widespread congener is assessed. Further, it would allow the evaluation of its impact in the conservation of a rare and endangered insular plant.

The effective conservation of the rare, habitat-restricted, endemic species could be compromised by the hybridization with the non-endemic and widespread L. dorycnium L.

The genetic consequences of hybridization for conservation purposes have been assessed

116 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Objectives in a relatively low number of plant species and genetic swamping has been linked to the risk of decline in several insular taxa (e.g. Argyranthemum, Brochmann, 1984; Cercocarpus,

Rieseberg et al., 1989; Cyrtandra, Smith et al., 1996), including Lotus (Liston et al., 1990).

117 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Materials and methods

4.4. MATERIALS AND METHODS

4.4.1. Plant sampling

Flowering specimens from six populations of Lotus fulgurans and seven of L. dorycnium s.l. (six ascribed to subsp. dorycnium and one identified as subsp. gracile, coded as TIR) were sampled from Minorca. Depending on the availability of plants, 5 to 22 individuals were sampled for each population (Table 4.4, Figure 4.2). The sampling strategy was directed to assess both morphological and molecular data for the same sampled individuals. Thus, for each individual, healthy young leaves were dried into highly porose filter paper sachets, altogether placed on polyethylene bags filled with silica gel for further

DNA extraction purposes. Paper sachets avoid mechanic crushing of dried leaves by silica gel rocks or pellets (i.e. material loosing into the silica gel). Several well-developed flowering branches were collected for morphometric analysis. Voucher herbarium specimens were preserved at the herbarium of the University of the Balearic Islands (MAC-

UIB).

4.4.2. Morphometric analysis

A total number of 206 plants were surveyed for morphological variation (Table 4.4).

For each individual, 7 quantitative characters were measured and 8 qualitative characters were scored (Table 4.5). All of them were recorded in the lab but one (habit of the plant), that was recorded in the field during the plant sampling. Five of the characters were descriptors of inflorescence and floral morphology, six described leaflet size and shape, three described stem architecture, and one referred to the habit of the plant.

For the quantitative characters, five measurements were taken per plant, and univariate statistics (mean and standard deviation) were computed for the raw data. Mean

118 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Materials and methods character values were compared between taxonomic groups with one-way analyses of variance (ANOVA), followed by multiple comparisons of means based on Tukey’s method for groups with unequal sample sizes, as implemented in the SPSS software (SPSS 13.0 for Windows, Chicago, Illinois, USA). Principal Component Analysis (PCA) of correlation matrices was used as a multivariate ordination method to get an overall view of the phenetic relations among individuals. Canonical Variates Analysis (CVA) an ordination procedure that maximizes between-group variance relative to within-group variance, and Classificatory

Discriminant Analysis (DA) were used to assess how well morphological data supported the pre-existing classifications of the samples. For DA, the discriminant functions used the cross-validation procedure to determine discriminatory power. Only the quantitative variables were included in the multivariate analyses. The PCA, CVA, and DA analyses were generated using the SPSS software.

Figure 4.2. Geographical distribution of sampled populations of Lotus fulgurans (black circles), L. dorycnium (squares), and intermediate individuals (star) in Minorca. Arrows indicate that both species occur in the same locality. Non-sampled populations of L. fulgurans are represented by grey circles. Population codes are indicated in Table 4.4. 119 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Materials and methods

Table 4.4. Collection data of Lotus populations sampled from Minorca and included in the morphological and molecular study. Sample size (n) represents the number of individuals used for morphological and molecular analyses. When two numbers are shown, the one in brackets represents the individuals used for molecular analyses, while the other number corresponds to the used for morphological measurements. * Population corresponding to subsp. gracile.

L. fulgurans L. x minoricensis L. dorycnium Location UTM Site description Code n Code n Code n

Rocky places near the coast, 10 m. Cala en Bastó EE7023 F-CBA 21 - - - -

Litoral scrub on calcareous soil, 25 Cala Morell EE7534 F-CMO 20 - - - - m Open scrub on limestone soil, in Cavalleria EE9237 F-CAV 22 - - - - rocky places, 25 m Litoral scrub, open sites on schistes, Favàritx FE0728 F-FAV 21 (17) - - - - 15 m Rocky places in open scrub, 15 m. Binidalí FE0210 F-BIN 12 - - D-BIN 5

F-MAC: Rocky places near the sea, Na Macaret FE0231 F-MAC 7 I-MAC 12 D-MAC 10 on limestone, 10 m; I-MAC/D-MAC: Shrubby places near the coast, 10 m Shrubby places on limestone soil, 40 Addaia FE0129 - - - - D-ADD 14 (13) m Evergreen aok forest, on limestone Cala Mitjana EE8321 - - - - D-CMI 19 (13) soil, 50 m Open scrub on limestone soil, in Santa Teresa EE9336 - - - - D-STE 12 (10) rocky places, 25 m Open scrub on sandstone soil, 80 m Ferreries EE8626 - - - - D-FER 10 (9)

Marshy habitats, 5 m Tirant * EE9433 - - - - D-TIR* 21

Number of samples 103 (99) 12 91 (81) Total: 206 (192)

4.4.3. Hybrid index

A hybrid index (Anderson, 1949) was computed using the eight qualitative variables

recorded. The selected traits showed clear differences between the two species, and their

scoring was straightforward and unequivocal in all samples. Thus, a value of 0 was

assigned to the representative character states present in L. fulgurans and a value of 1 was

given to those characterizing L. dorycnium (Table 4.5). The hybrid index, corresponding to

120 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Materials and methods the sum of the scores for the eight variables, was calculated for all sampled individuals.

Population averages were also calculated and a histogram of hybrid index classes was constructed. The average for each of the 206 individuals (total sum of the eight characters for each individual divided by eight) was used as an index of hybridity.

Table 4.5. Quantitative and qualitative morphological characters measured in Lotus individuals. The first seven characters (quantitative) were used in the PCA and CVA analyses. The remaining eight characters (qualitative) were used in the construction of the hybrid index. A zero value was assigned to the each of the states usually shown by L. fulgurans, while the states usually present in L. dorycnium were coded as 1.

Character Remarks / states Stem divergence Measured as the angle of branchement among contiguous stems Internode length Measured in the upper half of fertile branches Number of leaflets Measured in the upper half of fertile branches Length of the central leaflet Scored in leaves measured for the number of leaflets Number of flowers per inflorescence Scored in the upper inflorescences Length of the inflorescence stalk Measured from the upper inflorescence bract to the flower bracts Length of the floral peduncle From the flower bract to the basis of the calyx base Plant habit Plants were scored as spinescent (0) or not spinescent (1) Ramification pattern The plants could have divaricated stems with very short internodes and branched in zigzag (0) or not showing this zigzag branchement (1) Leaf stipules The character was scored as stipules absent or deciduous (0), or present and persistent (1) Leaflet apex Leaflets were coded as deeply emarginated (0) or not (1) Leaflet shape The shape of the central leaflet was scored as obovate-spathulate (0) or elliptic (1) Leaflet homogeneity The leaf could be formed by lateral(s) leaflet(s) differing in length (0) or all leaflets being similar (1) Inflorescence Inflorescences were coded as composed by axillar having axillar flowers (0) or only terminal flowers (1) Bracts of the receptacle The bracts of the flowers were classified as clearly forming a crown around the receptacle (1), or not (0)

4.4.4. DNA extraction

Total DNA of all specimens was extracted from dried leaves using the CTAB protocol (Doyle & Doyle, 1987), scaled down to perform the process in 1.5 ml microfuge tubes.

121 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Materials and methods

4.4.5. Nuclear ribosomal ITS sequences

ITS1, 5.8S and ITS2 sequences (Figure 3.1) were generated for two L. fulgurans and two L. dorycnium individuals (GenBank accession numbers GQ483308-GQ483311) using the methods outlined in Vargas et al. (2004). The analysis of the aligned sequences did not reveal any intraspecific polymorphism and allowed the recognition of species- specific sites for both species in the ITS1 region that could be revealed by a PCR-RFLP approach. Accordingly, the ITS1 region was amplified using the primer pair ITS-1/ITS-2

(White et al., 1990; Figure 4.3). PCR reactions were performed in 20 μl, containing 75 mM

Tris-HCl (pH 9.0), 5 mM KCl, 20 mM (NH4)2SO4, 0.0001% BSA, 2 mM MgCl2, 200 μM of each dATP, dCTP, dGTP and dTTP, 0.15 μM of each primer, approximately 50-100 ng of genomic DNA and 0.3 units of Netzime® (NEED S.L., Valencia, Spain) DNA polymerase from Thermus thermophilus. Amplification comprised an initial denaturalization step of 3 min at 94ºC, 40 cycles of 15 sec at 94ºC, 15 sec at 56ºC and 1 min at 72ºC, and a final extension step of 10 min at 72ºC, in a TechGene termocycler (Techne Incorporated,

Princeton, NJ, USA).

Figure 4.3. Schematic presentation of the nuclear 18S-26S ribosomal DNA repeat of the nucleolar organizer region (NOR) in plants. It is not scaled. Genes (blue boxes) are separated by the internal transcribed spacers (ITS, green boxes). These units, tandemly repeated, are separated by the intergenic spacer (IGS). The location of the primers ITS-1 and ITS-2 (White et al., 1990) is shown Primers ITS-1 and ITS-2 were used to amplify the ITS1 spacer with about 350 bp length.

122 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Materials and methods

4.4.6. Restriction analysis

Sequences obtained allowed the detection of diagnostic sites for L. fulgurans and L. dorycnium s.l. in the ITS1 region. These sites could be recognized respectively by Bsu RI and Taq I endonuclease enzymes. Thus, the amplified ITS1 products for all 192 individuals were digested separately with two endonuclease enzymes (BsuRI, TaqI) providing species- specific restriction profiles for L. fulgurans and L. dorycnium, according to manufacturer instructions (Fermentas International Inc., Burlington, Ontario, Canada). Each restriction reaction was performed in a volume of 10 μl (8 μl of the PCR product, 10 units of endonuclease and 1 μl of specific restriction buffer) for 4 hours at 37ºC (BsuRI) and 65ºC

(TaqI). The restricted PCR products were separated in 1.5% agarose gels that were run at

120V for 30 min. Two samples showing both species-specific markers were sequenced

(Genbank accession numbers GQ483312-GQ483313) to verify the restriction profile using the methods reported above.

4.4.7. Chloroplast DNA sequences

The trnL intron and the trnL-trnF spacer (Figure 4.4) of three accessions of L. dorycnium and two of L. fulgurans were amplified using the primers c and f described in

Taberlet et al. (1991), and sequenced as above (GenBank accession numbers GQ483303-

GQ483307). The alignment of the obtained sequences revealed the existence of a variable

21 bp indel in the trnL intron. To facilitate the direct scoring of this marker on agarose gels, a forward (c-dor F, 5’-GTT CTA ACA AAC GGA GTT G-3’) and a reverse (c-dor R, 5’-TTA

GAG TCT TCG TTA ATC-3’) primers were designed to amplify this region of the trnL intron.

PCR reactions were performed in 20 μl final volume, containing 75 mM Tris-HCl (pH 9.0), 5

mM KCl, 20 mM (NH4)2SO4, 0.0001% BSA, 2 mM MgCl2, 200 µM of each dATP, dCTP, dGTP and dTTP, 0.15 µM of each primer, approximately 50-100 ng of genomic DNA and

123 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Materials and methods

0.3 units of Netzime® (NEED S. L., Valencia, Spain) DNA polymerase from Thermus thermophilus. Amplification comprised an initial denaturalization step of 3 min at 94ºC, and

40 cycles of 30 sec at 94ºC, 45 sec at 50ºC, 1 min 15 sec at 72ºC, and a final extension step of 10 min at 72ºC. Scoring the two length variants (157 bp and 178 bp) detected in the trnL intron was straightforward when the PCR products were separated in 4% agarose gels that were run at 140V for 40 min.

The secondary structure of the trnL intron was assessed with the Mfold software available at the M. Zucker web page (http://mfold.bioinfo.rpi.edu).

Figure 4.4. Schematic presentation of the chloroplast DNA portion H from Shaw et al. (2005). Genes are represented in black boxes, with the name below. Primers designed to amplify intergenic regions are denoted with small black arrows with the name above. In green, the trnL / trnT-trnL fragment amplified with the primers c and f (Taberlet et al., 1991) to locate specific markers. In red, the specific region amplified with the designed primers c-dor F and c-dor R.

124 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Results

4.5. RESULTS

4.5.1. Morphometric analysis

Univariate statistics indicated that the means of the seven quantitative characters analyzed were significantly different in L. dorycnium and L. fulgurans at the p<0.0001 level.

The number of flowers (F =3053.25) and the number of leaflets (F =2883.27) were the characters showing the most discriminating values).

The PCA ordination of 206 individuals of L. dorycnium and L. fulgurans on the first two principal components (explaining 77.65% of the total variance; Table 4.6) revealed the existence of two groups in the multivariate space that correlated well with taxonomic boundaries (Figure 4.5). The first PCA axis (63.61% of the variance) was strongly correlated with the number of flowers per inflorescence and the length of the inflorescence stalk

(correlation coefficients of 0.963 and 0.917, respectively; Table 4.7), whereas the second axis (14.04% of the variance) showed the highest correlation with the length of the floral peduncle (correlation coefficient of 0.887; Table 4.7).

Overall, samples belonging to L. fulgurans had negative scores along the first principal axis, whereas L. dorycnium s.l. showed positive scores. Nevertheless, a dozen accessions showing an intermediate position (slightly biased towards L. dorycnium) linked the gap between the two species (Figure 4.3). These accessions originated from the Na

Macaret population, and belong to the putatively identified intermediate samples between both species based on observed morphological traits.

125 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Results

Comp. SV % Cum. % 1 4.452 63.607 63.607 Table 4.6. Principal component analysis results for the 2 0.983 14.042 77.649 seven qualitative traits measured in all 206 sampled 3 0.616 8.800 86.449 individuals (including L. fulgurans, L. dorycnium and 4 0.464 6.633 93.082 putative hybrids). For each component, singular value 5 0.302 4.311 97.393 (SV), percent of the total variance explained (%), and 6 0.125 1.784 99.177 cummulated percent of total variance explained (cum. %) 7 0.058 0.823 100.000 are shown.

Component Quantitative traits 1 2 3 Table 4.7. Correlations for the first Stem divergence -0.670 -0.396 0.494 three principal components Internode length 0.860 0.003 0.211 obtained in the PCA for the seven Number of leaflets 0.880 0.076 -0.199 qualitative traits measured in all Length of the central leaflet 0.779 -0.153 0.433 206 sampled individuals. Number of flowers per inflorescence 0.963 -0.005 -0.004 Length of the inflorescence stalk -0.339 0.887 0.300 Length of the floral peduncle 0.917 0.099 0.101

Figure 4.5. Ordination of Lotus fulgurans (circles), L. dorycnium s.l. (squares), and intermediate individuals (triangles) along the two first axes obtained by principal components analysis (PCA) using quantitative morphological characters. Only individuals with successful genotype scoring are represented, thus 192 specimens (Table 4.4). See text for further explanations. For comparison purposes, individuals corresponding to the subsp. gracile population are represented with a white border.

126 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Results

Canonical variate analysis (CVA), a widely used method for analyzing group structure, yielded similar results when the same individuals used in the PCA were previously classified as L. dorycnium, L. fulgurans, or intermediate (Figure 4.6). Only two discriminant functions were retrieved, though virtually all variation was explained by the first one. It accounted for 97.7% of the variance (SV=38.817; canonical correlation=0.987) and was positively correlated with the number of leaflets and the number of flowers (correlation coefficients of 0.634 and 0.612, respectively; Table 4.8). Intermediate samples were closer to L. dorycnium than L. fulgurans in the multivariate space, as shown in the PCA analysis, suggesting biased gene flow from L. dorycnium into L. fulgurans. In the CVA, only a single outlier (L. fulgurans) was closer to the centroid of another group (intermediate samples) than its own centroid.

When a classificatory discriminant analysis (DA) was performed with the cross- validation procedure, 99% of individuals were classified correctly according to the three previously defined groups (L. dorycnium, L. fulgurans, and intermediate). The two misclassified samples belong to L. dorycnium and L. fulgurans accessions that were initially classified as intermediate individuals (Table 4.9).

Table 4.8. Structure matrix of the CVA for the seven qualitative traits measured in all 206 sampled individuals. Combined correlations between discriminant variables and discriminant canonical functions typified. Variables arranged by the dimension of the total correlation with the obtained canonical function. (Percent explained for the canonical functions: 1=97.7%; 2=2.3%).

Function Quantitative traits 1 2 Number of flowers per inflorescence 0.612 0.537 Stem divergence -0.113 -0.002 Number of leaflets 0.634 -0.679 Length of the floral peduncle 0.321 0.446 Length of the central leaflet 0.132 0.336 Internode length 0.150 0.273 Length of the inflorescence stalk -0.048 0.164

127 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Results

Figure 4.6. Ordination of Lotus fulgurans (circles), L. dorycnium s.l. (squares), and intermediate individuals (triangles) along the two functions obtained by canonical variate analysis (CVA) using quantitative morphological characters, for all sampled individuals (i.e., 206). Non genotyped individuals are in white. For comparison purposes, individuals corresponding to the subsp. gracile population are represented with a white border.

Table 4.9. Results of the discriminant analysis (DA) for the 206 sampled individuals included in the CVA . The 99.5% of individuals were correctly classified in the original grouping; whereas the 99.0% of the cases were correctly classified when the cross-validation procedure was used.

Predicted group membership Classification Group Total L. fulgurans Intermediate L. dorycnium Original No. of cases L. fulgurans 102 1 0 103 Intermediate 0 12 0 12 L. dorycnium 0 0 91 91 % L. fulgurans 99,0 1,0 0 100 Intermediate 0 100 0 100 L. dorycnium 0 0 100 100 Cross-validation No. of cases L. fulgurans 102 1 0 103 Intermediate 0 12 0 12 L. dorycnium 0 1 90 91 % L. fulgurans 99,0 1,0 0 100 Intermediate 0 100 0 100 L. dorycnium 0 1,1 98,9 100

128 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Results

Eight qualitative features discriminated between L. fulgurans and L. dorycnium plants, and most of these showed fixed character states for each species (Table 4.10).

Thus, a high percentage of L. fulgurans accessions were characterized by subspinescent, thick stems with short internodes and divaricate branches in zigzag.

Overall, the stipules were deciduous or completely absent, leaves showed deeply emarginated leaflets, differing in length into the same leaf, and the central leaflets were obovate-spathulate. Inflorescences were formed only by axillary flowers, and bracts of the flowers never formed an obvious crown around the receptacle. Again, and based on the qualitative traits alone, the intermediate specimens of Na Macaret were clearly closer to L. dorycnium than to L. fulgurans as evidenced from the raw data (Table 4.10).

The hybrid index showed the highest and lowest mean scores for all populations of

L. dorycnium and L. fulgurans, respectively (Figure 4.7). The intermediate population of Na

Macaret showed a mean hybrid index (6.50) closer to the mean of L. dorycnium (7.96) than that of L. fulgurans (0.47). Interestingly, the population of L. dorycnium from Na Macaret showed the lowest mean hybrid index (7.70) and the highest variance (i.e., standard error) as compared to the other populations (Figure 4.7; detailed table not shown). Qualitative traits of the intermediate plants resembling the endemic species are only related to the leaf

(i.e. leaflet shape and leaflet homogeneity; Table 4.10).

Figure 4.7. Hybrid index histogram of samples from populations of Lotus fulgurans (black), L. dorycnium (white), and Na Macaret population (intermediate; grey) drawn from morphological qualitative characters (Table 4.5). Values of ca. zero are typical traits exhibited by L. fulgurans, while values of ca. eight are usually exhibited by L. dorycnium. Standard errors are shown for each population. D- TIR corresponds to the subsp. gracile population.

129 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Results

Table 4.10. Partition frequencies of the qualitative traits in L. dorycnium, L. fulgurans, and intermediate individuals from the Na Macaret population. Data for L. dorycnium s.l. are shown. *All individuals belonging to subsp. dorycnium.

s.l.

Character States

L. fulgurans Intermediate

L. dorycnium

n=103 n=12 n=91 Plant habit Spinescent plant 100 16.67 0 Not spinescent plant 0 83.33 100 Ramification pattern Divaricated stems with very short internodes and zigzag branched 100 0 0 Not showing this zigzag branchement 0 100 100 Leaf stipules Stipules absent or deciduous 97.09 0 0 Stipules present and persistent 2.91 100 100 Leaflet apex Deeply emarginated leaflets 100 0 0 Not deeply emarginated leaflets 0 100 100 Leaflet shape Central leaflet obovate-spathulate 100 83.33 3.30* Central leaflet elliptic 0 16.67 97.60 Leaflet homogeneity Leaf formed by lateral(s) leaflet(s) differing in length 89.32 50 0 Leaf formed by all leaflets being similar 10.68 50 100 Inflorescence Composed by axillary flowers 69.90 0 0 Composed only by terminal flowers 30.10 100 100 Bracts of the receptacle Not forming a crown around the receptacle 100 0 0 Clearly forming a crown around the receptacle 0 100 100

4.5.2. Molecular markers

Despite numerous attempts, 14 out of 206 accessions, belonging to different populations, did not produce successful PCR amplifications for the nuclear and chloroplast regions and their genotypes could not be determined. Therefore, genotyped individuals for any molecular marker were 192 rather than the 206 used for morphological purposes. Note that in the PCA (Figure 4.5) only genotyped individuals have been used, whereas all of them have been used for the CVA and DA.

4.5.2.1. ITS variation and genotyping

The ribosomal ITS1 spacer of L. fulgurans and L. dorycnium sequenced accessions showed six mutations differentiating both species (Table 4.11). Five of these were

130 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Results nucleotide substitutions, and one involved length variation. Thus, the sequence of L. dorycnium was 1 bp longer than that of L. fulgurans due to an inferred T deletion close to the 5’ end, resulting in hybrid electrophoretograms showing a clear and constant “rebound signal” after this position when sequenced with the forward primer. Species-specific markers were detected by a PCR-RFLP strategy involving BsuRI and TaqI. BsuRI showed two restriction sites for L. fulgurans samples (BsuRI +) that were absent in L. dorycnium

(BsuRI -), (positions 106 and 218; Table 4.11). In addition, TaqI cut the ITS1 of L. dorycnium once (TaqI +) but this site was not present in L. fulgurans (TaqI -), (position 218;

Table 4.11).

A summary of the ribosomal genotypes scored for the 192 accessions amplified is presented in Table 4.12 and Figure 4.11. All analyzed samples of L. fulgurans showed the combination of BsuRI + and TaqI -. Conversely, 95.06% of L. dorycnium samples showed the BsuRI - and TaqI + genotype, as expected; whereas the remaining 4.94% of accessions showed the genotype BsuRI + and TaqI +, indicating the presence of both of the species- specific markers of L. dorycnium and L. fulgurans. Interestingly, these genotypes were present in the two populations where intermediate plants between both species have been reported (Binidalí and Na Macaret; Cardona et al., 1983) and from Addaia, a population located less than 2 km from Na Macaret. Most intermediate specimens from Na Macaret showed ribosomal intragenomic polymorphisms (BsuRI + and TaqI +), suggesting the presence of ribosomal repeats from L. dorycnium and L. fulgurans. By contrast, only a single plant showed the genotype of L. dorycnium, BsuRI - and TaqI +. Some accessions showing the genotype BsuRI + and TaqI + were sequenced and the presence of double peaks in the diagnostic sites on the electrophoretograms, as well as unreadable tracks due to the presence of two sequences of different length, was confirmed (Figures 4.6 and 4.7).

131 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Results

Table 4.11. Mutations distinguishing L. fulgurans and L. dorycnium in the ITS1 region. GenBank accession numbers for the reference sequences are also shown. Position number corresponds to those in the ITS1 after the ITS-1 primer (White et al., 1990). The total length of the obtained ITS1 fragment without primers is 282 bp. Base additivity following the IUPAC code.

Position in the ITS1 sequence Taxon and GB accession number 36 38 58 83 106-107 218 L. fulgurans (GQ48330808) C - T T GG G Intermediate (GQ48330812) Y - / T Y W RR R L. dorycnium (GQ48330810) T T C A AA A

Figure 4.8. Electrophoretogram of the ITS1 region of an intermediate individual from Na Macaret showing base additivity in the position 36 and the « rebound pattern » observed from the position 38 due to the deletion of a T in that position in the L. fulgurans sequence (5’-3’ sequence, obtained with the forward primer).

Figure 4.9. Electrophoretogram of the ITS1 region of an intermediate individual from Na Macaret. Note the ongoing « rebound pattern » caused by the deletion in the position 38 in the L. fulgurans sequence. Base additivity in the positions 58 (A), 83 (B), 106 and 107 (C), and 218 (D).

132 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Results

4.5.2.2. CpDNA polymorphisms

Amplification of the 5’ region of the trnL intron with the primers c-dor F and c-dor R revealed the presence of two length variants of 157 and 178 bp. This was due to the presence or absence of a 21 bp stretch (ACGATAAATTTCTTTTCTAGG) that was not generated by tandem duplications or inversions at flanking sites. The 21 bp insertion is located near the 5’ end of the region P8, forming part of the stem. The presence or absence of the indel does not dramatically change the structure of the stem-loop region in the two haplotypes (data not shown). In fact, the secondary structures of the P8 RNA transcripts of the 157 bp haplotype are thermodynamically only slightly less stable (-63.40 kcal/mol) than those obtained from the folding of the P8 region of the 178 bp haplotype (-72.20 kcal/mol) as inferred from the free energy values of the two chloroplast variants. The 157 bp haplotype showed a frequency of 86.90% in L. fulgurans and, with the single exception of the Binidalí population, it was present in high frequencies in all surveyed populations of the species (Table 4.12, Figure 4.11). Likewise, although the 157 bp variant was the most frequent haplotype in L. dorycnium (61.29%), most of the populations sampled were polymorphic for this marker. Only the Binidalí and Santa Teresa populations were fixed for the 178 bp and 157 bp haplotypes, respectively. All morphologically intermediate specimens from Na Macaret showed the 157 bp haplotype.

The 157 bp haplotype is fixed in four of the L. fulgurans populations, and has a very high frequency (94.1%) in a fifth one. However, it has not been detected in the remaining sampled population of L. fulgurans, which is fixed for the 178 bp haplotype variant (Binidalí).

This second variant is also fixed in the intermediate individuals of Na Macaret (I-MAC) and in a single population of L. dorycnium (D-STE).

133 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Results

Figure 4.9. Geographical distribution of the frequencies of the nuclear (ribosomal ITS) and cpDNA (trnL intron) genotypes recovered from Minorcan accessions of L. fulgurans, L. dorycnium, and intermediate morphology individuals (I-MAC population). A) Distribution of the three ribosomal genotypes found (BsuRI +/ TaqI -; BsuRI -/ TaqI +, and (BsuRI +/ TaqI +). B) Distribution of the short (157 bp; black) and long (178 bp; white) length variants from the trnL intron.

Figure 4.10. Electrophoretogram of the trnL chloroplast region showing the indel detected. It is located 2 pb after the c-dor F primer. A) Individual including the 21 bp stretch (haplotype 178 bp) and B) not showing the 21 bp stretch (haplotype 157 bp).

134 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Results

Table 4.12. Population distribution of the nuclear (ribosomal ITS) and cpDNA (trnL intron) genotypes recovered from Minorcan accessions of L. fulgurans, L. dorycnium, and individuals with intermediate morphology. Global data for each species and the intermediate plants are also shown at the bottom.

ITS1 TrnL intron Sample TaqI +/ TaqI -/ TaqI +/ Taxon Population 157 bp 178 bp size BsuRI - BsuRI + BsuRI + % n % n % n % n % n F-CBA 21 0 0 100 21 0 0 100 21 0 0 Lotus fulgurans F-CMO 20 0 0 100 20 0 0 100 20 0 0 F-CAV 22 0 0 100 22 0 0 100 22 0 0 F-FAV 17 0 0 100 17 0 0 94.1 16 5.9 1 F-BIN 12 0 0 100 12 0 0 0 0 100 12 F-MAC 7 0 0 100 7 0 0 100 7 0 0 Intermediate I-MAC 12 9.1 1 0 0 90.9 11 100 12 0 0 D-MAC 10 100 10 0 0 0 0 30 3 70 7 L. dorycnium D-BIN 5 80 4 0 0 20 1 0 0 100 5 D-CMI 13 100 13 0 0 0 0 30.8 4 69.2 9 D-FER 9 100 9 0 0 0 0 55.6 5 44.4 4 D-STE 10 100 10 0 0 0 0 100 10 0 0 D-ADD 13 76.9 10 0 0 23.1 3 38.5 5 61.5 8 D-TIR 21 100 21 0 0 0 0 85.7 18 14.3 3 TOTAL 192 Lotus fulgurans 99 0 0 100 99 0 0 86.9 86 13.1 13 Intermediate 12 8.3 1 0 0 91.7 11 100 12 0 0 L. dorycnium 81 95.1 77 0 0 4.9 4 55.6 45 44.4 36

135 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Discussion

4.6. DISCUSSION

4.6.1. Overall evidence suggest interspecific gene flow between L. fulgurans and L.

dorycnium

Documenting hybridization using nuclear ribosomal sequences is not always straightforward because of its molecular features. First, the nuclear 45S ribosomal cistron belongs to a multigene family found in hundreds to thousands of tandem arrays (Arnhheim,

1983) and more than one locus can be present in diploid organisms. Thus, intra-array polymorphisms and paralogy (including the presence of pseudogenes) can be responsible for the detected intragenomic heterogeneity in several species (Mayol & Rosselló, 2001).

Second, concerted evolution could homogenize variation across the multiple rDNA gene copies towards either parent in only a few generations (Fuertes Aguilar et al., 1999a), making it difficult to detect the contribution of both parental taxa in the hybrid-derivative.

Thus, when interspecific gene flow is extensive and recurrent, patterns of ribosomal ITS variation mirror geographic, rather than taxonomic, boundaries (Fuertes Aguilar et al.,

1999b).

Patterns of ribosomal variation strongly suggest that intragenomic heterogeneity in

Minorcan Lotus is not widespread or random. In fact, it is virtually absent in the reference populations of L. fulgurans and L. dorycnium and has only been detected in three sites (Na

Macaret, Addaia, Binidalí), where both species grow in close vicinity. These results showing the additivity of parental ITS sequences strongly support the hypothesis that hybridization occurred between the two species. Moreover, the joint assessment of morphological and molecular markers indicated that: (i) hybridization between Lotus species has been recurrent and polytopic; and (ii) later-hybrid generations (backcrossed towards a single parent, L. dorycnium) predominated.

136 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Discussion

The multivariate analysis of quantitative morphological characters and the hybrid score obtained from qualitative features have corroborated the suggestion of Cardona et al.

(1983) that plants combining morphological features of both L. fulgurans and L. dorycnium exist at Na Macaret. At this site, eleven out of the twelve morphologically-intermediate individuals showed additivity of both species-specific ITS markers. Alternative explanations that could theoretically explain the presence of intermediate samples between both species

(incomplete primary divergence reflecting early stages of speciation, intraspecific polymorphisms) are clearly not supported by the ITS data. Based on the observed patterns of morphological variation, hybrids at Na Macaret are more closely related to L. dorycnium than to L. fulgurans.

From this study sampling, it could not be confirmed the presence of morphologically intermediate specimens between L. fulgurans and L. dorycnium at Binidalí, as earlier reported by Cardona et al. (1983), or at other sites. However, the unidirectional transfer of ribosomal genes of L. fulgurans into L. dorycnium was observed in the Binidalí and Addaia populations. There, several L. dorycnium individuals showed ITS additivity. Yet, these accessions showed morphological features typical of L. dorycnium, and no evidence for the presence of L. fulgurans characters (either quantitative and qualitative) was detected in the morphometric analyses conducted. The molecular signature recovered strongly suggests a scenario of recurrent hybridization and backcrossing towards L. dorycnium, in which later- generation hybrids appear to retain nuclear ribosomal repeats from L. fulgurans.

4.6.2. CpDNA variation better explains a related origin of both species rather than

hybridization

Regional sharing of chloroplast haplotypes across species is usually related to local cpDNA transfer through reticulate hybridization and unidirectional introgression (e.g.,

137 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Discussion

McKinnon et al., 2004; Jiménez et al., 2004; Albaladejo et al., 2005). However, other scenarios not involving interspecific gene flow can lead to similar patterns of cpDNA sharing, such as homoplasy or incomplete lineage sorting.

Non-coding regions of the cpDNA genome are characterized by a high frequency of structural mutations relative to base substitutions that hamper homology assessments

(Morton & Clegg, 1993). The two haplotypes detected in this study are located in the trnL intron and differ by a structural mutation (21 bp long) that it is not associated to a direct repeat. It has been suggested that short indels up to a few bases long, and involving single nucleotide stretches or direct repeats, are more likely to be homoplasious than the other types of structural indels (Mes et al., 2000). These considerations do not strengthen the hypothesis of a recurrent origin of this length mutation. Interestingly, the distribution of the two cpDNA haplotypes in Minorca is geographically correlated showing a NE-SE cline (Fig.

4.9), but not completely species independent.

Available evidence suggests that L. dorycnium is polymorphic for the presence or absence of the 21 bp indel in accessions from the Balearic archipelago. The two cpDNA haplotypes are present in L. dorycnium from Majorcan and Ibizan populations (Conesa et al., unpublished data), but showed contrasting frequencies of fixation. Thus, the 178 bp haplotype is frequent in accessions from Ibiza (77.15%, n=35), but shows a very low frequency in Mallorca (17.75%, n=62). By contrast, the 157 bp haplotype seems fixed in L. fulgurans from Cabrera (100%, n=33) and attains a frequency of 80% (n=35) in populations from Mallorca (Conesa et al., unpublished data). These values from Mallorca and Cabrera are close to those found in this study from Minorca (86.90%). Thus, it cannot be ruled out that the occurrence of the 178 bp haplotype at low frequencies in L. fulgurans is due to an incomplete lineage sorting from the polymorphic cpDNA pool of L. dorycnium.

138 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Discussion

Additional chloroplast markers that were screened (trnK intron, trnL-trnF spacer, trnD-trnT spacer; Cosín et al. and Conesa et al., unpublished data) failed to detect fixed species-specific markers in L. fulgurans and L. dorycnium from the Balearic Islands. Thus, available cpDNA data is not conclusive in assessing whether regional sharing of cpDNA markers in Minorcan Lotus is due to incomplete lineage sorting or to chloroplast capture.

Given the lack of a clear haplotype differentiation detected between L. fulgurans and L. dorycnium at the hybridizing populations (based on nrDNA evidence: Na Macaret, Addaia, and Binidalí), cpDNA data could not provide enough information to support the hypothesis of a unidirectional scenario of backcrossing hybridization towards L. dorycnium. However, the presence of both cpDNA haplotypes at the hybrid population from Na Macaret suggests that gene flow between Lotus species has been reciprocal.

4.6.3. Hybridization and the conservation of the endangered L. fulgurans

In contrast with its creative role in plant evolution, hybridization may be a cause of concern for the effective conservation of rare and endangered species (Arnold, 1997).

Threats can be more acute when population sizes of the rare species are small and when barriers to introgression are weak (Rieseberg et al., 1989). Several facts could promote interspecific hybridization when populations of L. fulgurans and L. dorycnium grow in sympatry. These include self-incompatibility, entomophilous pollination by generalist bees

(Apis mellifera, Hylaeus pictus), and the long (between 91-98 days in L. fulgurans) and overlapping (from April to July) flowering time exhibited by both species (Gil, 1994;

Traveset, 1999).

Two lines of evidence (morphological introgression towards L. dorycnium and the presence of nuclear ribosomal repeats from L. fulgurans in these individuals as well as in morphologically typical L. dorycnium samples) suggest that genetic swamping of a rare

139 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Discussion species (L. fulgurans) by a more widespread one (L. dorycnium) is not occurring in Minorca, rather the widespread species seems to be the one affected by the hybridization and introgression processes. Moreover, hybridization appears to be a very localized phenomenon and is currently observed in individuals from a single narrow hybrid zone (Na

Macaret) that is not expanding (see below). The facts that (i) morphologically intermediate plants at Binidalí observed 25 years ago (Cardona et al. 1983) are no longer present, but L. dorycnium individuals with both ribotypes still are, and (ii) at Addaia, L. dorycnium individuals retained species-specific molecular markers of the rare species as relicts of a past hybridization scenario, strongly suggest different temporal stages of the same hybridization process that now is ongoing at Na Macaret.

Nevertheless, the long term conservation of the endangered L. fulgurans does not appear to be compromised by the existence of past or current hybridization with the widespread L. dorycnium. First, interspecific gene flow appears to be localized to three populations and only a single hybrid zone has been documented. Second, although the hybrid zone detected at Na Macaret was reported about 25 years ago (Cardona et al.,

1983) its distribution is narrow (less than 50 m in width) and the number of individuals appears to be stable. Despite repeated observations made over the last 15 years, no evidence for the expansion of the hybrid population has been observed (P. Fraga and M.

Mus, personal observations). Both parental species are long-lived shrubs lacking clonal growth so the vegetative spread of hybrid plants is not feasible. Furthermore, the habitats where the hybrid plants were detected are pristine (littoral shrubby places) and do not show any degree of human disturbance. Therefore, direct or collateral human activity seems not to have favored hybridization at these sites in contrast with the well-documented correlation of hybridization and disturbance.

140 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium Discussion

Lastly, ribosomal and morphological data strongly suggest that successful gene flow between both species is probably unidirectional and biased towards L. dorycnium. In fact, hybrid progeny was found more commonly in closer association with the last species.

However, the occurrence of hybridization in the opposite direction cannot be completely ruled out.

The biological reasons behind this apparent asymmetrical gene flow are by no means evident and should be further assessed. However, they could be related to features shown by L. fulgurans that are less competitive in comparison to L. dorycnium, including a slower growth rate, the inability to regenerate via vegetative regrowth following fire, and a narrower niche diversity. Further, in the endemic L. fulgurans its population structure seems to be biased towards mature individuals. Taken together, it is then plausible that hybridization involving recurrent backcrosses towards L. fulgurans produced less fit hybrid plants that could be weak competitors and therefore not persistent.

141 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium References

4.7. REFERENCES

Albaladejo RG, Aguilar JF, Aparicio A & Nieto-Feliner G. 2005. Contrasting nuclear-plastidial phylogenetic

patterns in the recently diverged Iberian Phlomis crinita and P. lychnitis lineages (Lamiaceae). Taxon,

54: 987-998.

Allan GJ & Porter JM. 2000. Tribal delimitation and phylogenetic relationships of Loteae and Coronilleae

(Faboideae: Fabaceae) with special reference to Lotus: evidence from nuclear ribosomal ITS

sequences. American Journal of Botany, 87: 1871–1881.

Allan GJ, Francisco-Ortega J, Santos-Guerra A, Boerner E & Zimmer EA. 2004. Molecular phylogenetic

evidence for the geographic origin and classification of Canary Island Lotus (Fabaceae: Loteae).

Molecular Phylogenetics and Evolution, 32: 123-138.

Allan GJ, Zimmer EA, Wagner WL & Sokoloff DD. 2003. Molecular phylogenetic analyses of tribe Loteae

(Leguminosae): implications for classification and biogeography. In: Klitgaard BB, Bruneau A (eds).

Advances in legume systematics. Part 10. Higher level systematics. Royal Botanic Gardens: Kew.

Alomar G, Mus M & Rosselló JA. 1997. Flora Endèmica de les Balears. Consell Insular de Mallorca: Palma de

Mallorca.

Anderson E. 1948. Hybridization of the habitat. Evolution, 2:1-9.

Anderson E. 1949. Introgressive Hybridization. John Wiley: New York.

Arambarri, AM. 2000. A cladistic analysis of the Old World species of Lotus (Fabaceae: Loteae). Canadian

Journal of Botany, 78: 351–360.

Arnheim N. 1983. Concerted evolution of multigene families. In: Nei M, Koehn RK (eds). Evolution of Genes and

Proteins. Sinauer Associates: Sunderland.

Arnold ML. 1997. Natural hybridization and evolution. Oxford University Press: Oxford.

Ball PW. 1968. Dorycnium Miller. In: Tutin TG, Heywood VH, Burges NA, Valentine DH, Walters SM & Webb DA

(eds). Flora Europaea, vol. 2: 172-173. Cambridge.

Brochmann C. 1984. Hybridization and distribution of Argyranthemum coronopifolium (Asteraceae-Anthemidae)

in the Canary Islands. Nordic Journal of Botany, 4: 729-736.

142 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium References

Cardona MA, Llorens L & Sierra E. 1983. Étude biosystématique de Dorycnium pentaphyllum Scop. subsp.

fulgurans (Porta) comb. nova, endémique des Baléares orientales. Collectanea Botanica (Barcelona),

14: 133-150.

Conesa MÀ, Mus M & Rosselló JA. 2006. Lotus x minoricensis (Fabaceae), a new hybrid from the Balearic

Islands. Flora Montiberica, 34: 25-27.

Degtjareva GV, Kramina TE, Sokoloff DD, Samigullin TH, Valiejo-Roman CM & Antonov AS. 2006. Phylogeny of

the genus Lotus (Leguminosae, Loteae): evidence from nrITS sequences and morphology. Canadian

Journal of Botany, 84: 813-830.

Díaz Lifante Z. 2000. Dorycnium Mill. In: Castroviejo S, coord. Flora Ibérica. Plantas vasculares de la Península

Ibérica e Islas Baleares. Vol. VII(II): 812-823. Real Jardín Botánico, CSIC: Madrid.

Doyle J & Doyle JJ. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue.

Phytochemical Bulletin, 19: 11–15.

Fraga P, Mascaró C, Carreras D, Garcia O, Pallicer X, Pons M, Seoane M & Truyol M. 2004. Catàleg de la flora

vascular de Menorca. IME: Maó.

Fuertes Aguilar J, Rosselló JA & Nieto Feliner G. 1999a. NrDNA concerted evolution in natural and artificial

hybrids of Armeria (Plumbaginaceae). Molecular Ecology, 8: 1341-1346.

Fuertes Aguilar J, Rosselló JA & Nieto Feliner G. 1999b. Molecular evidence for the compilospecies model of

reticulate evolution in Armeria (Plumbaginaceae). Systematic Biology, 48: 735-754.

Gil L. 1994. Biología reproductiva de la flora litoral de Baleares. I. Dunas y roquedos maritimos. Tesi doctoral

inèdita, Universitat de les Illes Balears: Palma de Mallorca.

Jiménez P, López de Heredia U, Collada C, Lorenzo Z & Gil L. 2004. High variability of chloroplast DNA in three

Mediterranean evergreen oaks indicates complex evolutionary history. Heredity, 93: 510-515.

Lassen P. 1979. Dorycnium fulgurans, a neglected species from the Balearic Islands. Botaniska Notiser,132:

357-358.

Liston A, Rieseberg LH & Mistretta O. 1990. Ribosomal DNA evidence for hybridization between island endemic

species of Lotus. Biochemical Systematics and Ecology, 18: 239-244.

Llorens L. 1979. Nueva contribución al conocimiento de la flora balear. Mediterránea: serie de estudios

biológicos 3: 101-122.

Mayol M & Rosselló JA. 2001. Why nuclear ribosomal DNA spacers (ITS) tells different histories in Quercus.

Molecular Phylogenetics and Evolution, 19: 167-176.

143 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium References

McKinnon GE, Vaillancourt RE, Steane DA & Potts BD. 2004. The rare silver gum, Eucalyptus cordata is leaving

its trace in the organellar gene pool of Eucalyptus globulus. Molecular Ecology, 13: 3751-3762.

Mes THM, Kuperus P, Kirschner J, Stepanek J, Oosterveld, Storchova H & den Nijs JCM. 2000. Hairpins

involving both inverted and direct repeats are associated with homoplasious indels in non-coding

chloroplast DNA of Taraxacum (Lactuceae: Asteraceae). Genome, 43: 634-641.

Milne RI, Terzioglu S & Abbott RJ. 2003. A hybrid zone dominated by fertile F1s: maintenance of species

barriers in Rhododendron. Molecular Ecology, 12: 2719-2729.

Moore WS. 1977. An evaluation of narrow hybrid zones in vertebrates. Quarterly Review of Biology, 52: 263–

277.

Moreno JC (coord). 2008. Lista Roja 2008 de la flora vascular española. Dirección General de Medio Natural y

Política Forestal (Ministerio de Medio Ambiente, y Medio Rural y Marino, y Sociedad Española de

Biología de la Conservación de Plantas): Madrid.

Morton BR & Clegg MT. 1993. A chloroplast DNA mutation and gene conversion in a non-coding region near

rbcL in the grass family (Poaceae). Current Genetics, 24: 357-365.

Mus M. 1992. Estudio de la diversificación en la flora endémica de Baleares. Aspectos taxonómicos y

evolutivos. Tesi doctoral inèdita. Universitat de les Illes Balears: Palma de Mallorca.

Palau P. 1954. Investigaciones botánicas en Baleares Anales del Instituto Botánico A.J. Cavanilles 11: 483-495.

Palau P. 1976. Catàleg de la Flòrula de l’Illa de Cabrera i dels illots que l’envolten. Treballs de la Institució

Catalana d’Història Natural 7: 5-103.

Pla V, Sastre B, Llorens Ll. 1992. Aproximació al catàleg de la flora vascular de les Illes Balears. Universitat de

les Illes Balears - Jardí Botànic de Sóller (MBCN), Palma de Mallorca.

Porta P. 1887. Stirpium in insulis Balearicum anno 1885 collectarum enumeratio. Nuovo Giornale

Botanico Italiano, 19: 276-324.

Rieseberg LH, Zona S, Aberbom L, Martin TD. 1989. Hybridization in the island endemic, Catalina mahoganay.

Conservation Biology, 3: 52-58.

Rikli M. 1901. Die Gattung Dorycnium. Botanische Jahrbücher für Systematik, Pflanzengeschichte und

Pflanzengeographie, 31: 314–404.

Shaw J, Lickey EB, Beck JT, Farmer SB, Liu W, Miller J, Siripun KC, Winder CT, Schilling EE & Small RL. 2005.

The tortoise and the hare II: relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic

analyses. American Journal of Botany, 92: 142-166.

144 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium References

Smith JF, Burke CC & Wagner WL. 1996. Interspecific hybridization in natural populations of Cyrtandra

(Gesneriaceae) on Hawaiian Islands: evidence from RAPD markers. Plant Systematics and Evolution,

200: 61-77.

Sokoloff DD. 2003a. On system and phylogeny of the tribe Loteae DC. (Leguminosae). Bulletin of the Moscow

Society of Naturalists. Biological Series. 108: 35–48.

Sokoloff DD. 2003b. Monophyly and classification of the tribe Loteae DC. of the family Leguminosae. Dr. Sci.

thesis, Higher Plants Departament, Moscow State University: Moscow.

Taberlet P, Gielly L, Patou G & Bouvet J. 1991. Universal primers for amplification of three non-coding regions

of chloroplast DNA. Plant Molecular Biology, 17: 1105–1109.

Traveset A. 1999. El paper dels mutualismes planta-animal als ecosistemes insulars. In: Alcover JA (ed).

Ecologia de les illes. Societat d’Història Natural de Balears: Palma de Mallorca.

Vargas P, Rosselló JA, Oyama R & Güemes J. 2004. Molecular evidence for naturalness of genera in the tribe

Antirrhineae (Scrophulariaceae) and three independent evolutionary lineages of taxa from the New

World and the Old. Plant Systematics and Evolution, 249: 151-172.

White TJ, Bruns T, Lee S & Taylor J. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes

for phylogenetics. In: Innis M, Gelfand D, Sninsky J, White T (eds). PCR Protocols: A guide to methods

and applications. Academic Press: San Diego.

Wilson P. 1992. On inferring hybridity from morphological intermediacy. Taxon, 41: 11-23.

145 Chapter 4. Hybridization processes between Lotus fulgurans and L. dorycnium References

146 Chapters 5 and 6. Hybridization involving the endemic Helichrysum crassifolium Introduction to the case study

CASE STUDY III Hybridization involving the endemic Helichrysum crassifolium

page Introduction to the case study 149 The genus Helichrysum and the Mediterranean sect. Stoechadina 150 References 153

Chapter 5. Leaf shape variation and taxonomic boundaries involving the endemic 155 Helichrysum crassifolium (L.) D. Don

Chapter 6. Widespread hybridization in the Balearic Islands involving the endemic 191 Helichrysum crassifolium (L.) D. Don

147 Chapters 5 and 6. Hybridization involving the endemic Helichrysum crassifolium Introduction to the case study

148 Chapters 5 and 6. Hybridization involving the endemic Helichrysum crassifolium Introduction to the case study

INTRODUCTION TO THE CASE STUDY

With the advent of PCR-based methods in the last decades, a couple of molecular markers have been extensively used to detect hybrid lineages and to infer its origin (e.g.,

Hegarty & Hiscock, 2005). Molecular markers are of utility because they can supply a large number of independent features that usually only show heritable variation. In some cases, molecular markers are selectively neutral at non coding regions, and have a simple mode of expression and inheritance (Rieseberg & Ellstrand, 1993; Arnold, 1997). However, molecular data alone can be sometimes misleading in species discrimination or in phylogenies, especially when hybrids and introgressants are frequent (e.g., Vriesendorp &

Bakker, 2005). On the other hand, morphological characters used to identify plant hybrids are diverse, including vegetative and floral features measured on a qualitative or quantitative basis (e.g., Shipunov & Bateman, 2005).

The utility of one character depends on its capacity to discriminate between both parental taxa, and on its mode of expression on hybrid phenotypes. From a morphological point of view, hybrids typically show a mosaic of parental and intermediate characters (e.g.,

Rieseberg & Ellstrand, 1993) especially in early steps of the hybridization process, although intermediacy can also arise through morphological convergence rather than hybridization

(e.g., Dobzhansky, 1941; Rieseberg, 1997). However, hybridization has been proved to be also a significant source of new morphological traits and novel genes and adaptations (e.g.,

Anderson, 1949; Grant, 1981; Schwarzbach et al., 2001; Jorgensen & Mauricio, 2005;

Hegarty et al., 2009).

Consequently, in some cases morphological characters alone can also be of limited value for identifying natural hybrids (e.g., Rieseberg, 1997, and references therein). For

149 Chapters 5 and 6. Hybridization involving the endemic Helichrysum crassifolium Introduction to the case study example, in ancient hybridization and introgression processes where hybrid-derivates tend to converge to their parental taxa morphologies (Arnold, 1997). On the opposite, morphological traits alone can be useless in cases where novel morphologies arise, blurring the relation of such hybrids with their parental taxa. Therefore, hybridization processes are frequently better assessed combining morphological and molecular markers (e.g.,

Rieseberg & Ellstrand, 1993; Rieseberg, 1997; Kaplan & Fehrer, 2006). To cite a few examples, in Salix, Hardig et al. (2000) demonstrated how the discrimination of some backcrossed individuals from their parental taxa was only accomplished when adding morphological data to the molecular data. Shipunov and co-workers found good discrimination in European Dactylorhiza taxa when using molecular markers combined with classical morphometry (Shipunov et al., 2004), and also in their hybrids when combining with geometric morphometry (Shipunov & Bateman, 2005). Paun et al. (2006) obtained successful discrimination of apomictic species and hybrids in the Ranunculus cassubicus complex when adding morphological data. Indeed, the addition of morphological data to the nrITS data produced a better resolved phylogeny for the large genus Lotus (Degtjareva et al., 2006).

The genus Helichrysum and the Mediterranean sect. Stoechadina

Helichrysum Mill. is a speciose and cosmopolitan genus belonging to Asteraceae

(Gnaphalieae; Anderberg, 1991). It encloses up to 600 species (Anderberg, 1991) distributed along the entire African continent, Madagascar, Mediterranean basin,

Macaronesia, Central Asia and India (Anderberg, 1991; Bayer et al., 2007). Yet, taxa commonly included into this genus from Australia and New Zealand use to be currently included into other genera (Bayer et al., 2002, and references therein). As stressed by

150 Chapters 5 and 6. Hybridization involving the endemic Helichrysum crassifolium Introduction to the case study several authors (e.g., Bayer et al., 2000; Galbany-Casals et al., 2004), it has a polyphyletic origin.

The Mediterranean basin is the second most important center of diversification of the genus, with ca. 30 species (Galbany-Casals et al., 2004). All these species were included in the sect. Stoechadina by Galbany-Casals et al. (2006) based on the assumption of constituting a rather uniform group concerning gross morphology of vegetative and reproductive characters, and on molecular phylogenetic analyses (Galbany-Casals et al.,

2004). Such molecular phylogenies indicated a monophyletic origin for the Mediterranean clade, in turn derived from the African clade (Galbany-Casals et al., 2004). Moreover, Bergh

& Linder (2009) suggested at least five dispersal events from southern Africa to the

Mediterranean, and one from there to Asia.

In the Balearic Islands five species of Helichrysum have been reported to occur, but only four inhabit the eastern islands, where two are rupicolous (Table III.1). One of those, H. crassifolium (L.) G. Don (wrongly named Helichrysum ambiguum (Pers.) C. Presl ≡ H. lamarckii Cambess.; Galbany-Casals et al., 2006) is a Balearic endemism restricted to the calcareous cliffs of the mountains of Majorca and Minorca. The other one, H. pendulum (C.

Presl) C. Presl (wrongly called H. rupestre DC., nom. illeg. [= H. panormitanum Guss.];

Aghababyan et al., 2007), is a chasmophyte Mediterranean species mainly located in the

Tramuntana mountain range, in western Majorca. Recently, specimens showing intermediate morphological features between both rupicolous species have been reported from field observation (Cosín et al., 2002; Mus et al., 2002; Conesa et al., 2004; Galbany-

Casals et al., 2006). These studies suggested that interspecific hybridization between both rupicolous taxa could explain the pattern of morphological variation observed. However, morphology alone could be misleading concerning the origin of these individuals showing deviant or overlapping morphological features (e.g., Rieseberg, 1997).

151 Chapters 5 and 6. Hybridization involving the endemic Helichrysum crassifolium Introduction to the case study

Taxa MA CA ME EI FO H. crassifolium (L.) D. Don ✔ ✔1 H. pendulum (C. Presl) C. Presl ✔ ✔2 ✔ H. stoechas (L.) Moench ✔ ✔ ✔ ✔ ✔ H. italicum (Roth) G. Don subsp. microphyllum (Willd.) Nyman ✔ H. serotinum (DC.) Boiss. subsp. serotinum ✔

Table III.1. Distribution of the different Helichrysum taxa in the Balearic Islands. Nomenclature following Galbany-Casals et al. (2006) and Aghababyan et al. (2007). MA, Majorca; CA, Cabrera; ME, Minorca; EI, Ibiza; FO, Formentera. 1 There is a single narrow population, shared with H. stoechas and extremely unusual (intermediate) individuals grow in non- rupicolous habitat, as do some of the H. crassifolium individuals. 2 There is a single, rupicolous, and narrowly located population of very few individuals and with a morphology somehow resembling H. stoechas (Bibiloni et al., 1993; M. Mus pers. com.).

152 Chapters 5 and 6. Hybridization involving the endemic Helichrysum crassifolium References

REFERENCES

Aghababyan M, Greuter W, Mazzola P & Raimonfo FM. 2007. Typification of Sicilian Helichrysum (Compositae)

revised. Taxon, 56: 1285-1288.

Anderberg AA 1991. Taxonomy and phylogeny of the tribe Gnaphalieae (Asteraceae). Opera Botanica, 104: 1-

195.

Anderson E. 1949. Introgressive hybridization. John Wiley and Sons: New York.

Arnold ML. 1997. Natural Hybridization and Evolution Oxford University Press, New York.

Bayer RJ, Breitweiser I, Ward J & Puttock CF. 2007. Tribe Gnaphalieae. In: Kubitzki, K. (ed). The Families and

Genera of Flowering Plants, Flowering Plants, Eudicots, Asterales, vol. VIII. Springer: Berlin.

Bayer RJ, Puttock CF & Kelchner SA 2000. Phylogeny of South African Gnaphalieae (Asteraceae) based on two

noncoding chloroplast sequences. American Journal of Botany, 87: 259-272.

Bergh NG & Linder HP. 2009. Cape diversification and repeated out-of-southern-Africa dispersal in paper

daisies (Asteraceae–Gnaphalieae). Molecular Phylogenetics and Evolution, 51: 5-18.

Bibiloni B, Alomar G & Rita J. 1993. Flora vascular dels illots i addicions a la flora de Cabrera gran. In: Alcover

JA, Ballesteros E & Fornós JJ. (eds.). Història Natural de l’Arxipèlag de Cabrera. CSIC-Ed. Moll-

Monografies de la Societat d’Història Natural de les Balears 2: 179-206.

Conesa MÀ, Cosín R, Mus M & Rosselló JA. 2004. Nuclear and cpDNA data suggest extensive cytoplasmic

capture among narrow endemic and widespread species of Helichrysum in the Balearic Islands. Plant

Evolution in Mediterranean Climate Zones. IXth IOPB Meeting. 16-19 may 2004. Valencia. Spain.

Cosín R, Conesa MÀ, Mus M & Rosselló JA. 2002. Hibridación reticular en especies insulares de Helichrysum

Mill. (Compositae) detectada mediante ISSR y secuencias de ADN cloroplástico. 1er Congreso de

Biología de la Conservación de Plantas, Jardín Botánico de la Universidad de Valencia, 2-5 octubre

2002. Valencia. Spain.

Degtjareva GV, Kramina TE, Sokoloff DD, Samigullin TH, Vallejo-Roman CM & Antonov AS. 2006. Phylogeny of

the genus Lotus (Leguminosae, Loteae): evidence from nrITS sequences and morphology. Canadian

Journal of Botany, 84: 813-830.

Dobzhansky TH. 1941. Genetics and the Origin of Species. Columbia University Press: New York.

Galbany-Casals, M, Garcia-Jacas, N., Susanna, A., Sáez, L. & Benedí, C. 2004. Phylogenetic relationships in

the Mediterranean Helichrysum (Asteraceae, Gnaphalieae) based on nuclear rDNA ITS sequence data.

Australian Systematic Botany, 17: 241-253.

153 Chapters 5 and 6. Hybridization involving the endemic Helichrysum crassifolium References

Galbany-Casals, M., Sáez, L. & Benedí, C. 2006. A taxonomic revision of Helichrysum Mill. sect. Stoechadina

(DC.) Gren. & Godr. (Asteraceae, Gnaphalieae). Can. J. Bot. 84:1203-1232.

Grant V. 1981. Plant speciation. Columbia University Press: New York.

Hardig TM, Brunsfeld SJ, Fritz RS, Morgan M & Orians CM. 2000. Morphological and molecular evidence for

hybridization and introgression in a willow (Salix) hybrid zone. Molecular Ecology 9, 9-24.

Hegarty MJ & Hiscock SJ. 2005. Hybrid speciation in plants: new insights from molecular studies. New

Phytologist, 165: 411–423.

Hegarty MJ, Barker GL, Brennan AC, Edwards KJ, Abbott RJ & Hiscock SJ. 2009. Extreme changes to gene

expression associated with homoploid hybrid speciation. Molecular Ecology 18: 877–889.

Jorgensen S & Mauricio R. 2005. Hybridization as a source of evolutionary novelty: leaf shape in a Hawaiian

composite. Genetica, 123: 171-179.

Kaplan Z & Fehrer J. 2006. Comparison of natural and artificial hybridization in Potamogeton. Preslia, 78: 303–

316.

Mus M, Cosín R, Conesa MÀ & Rosselló JA. 2002. Patrones de variación genética y fenotípica en especies del

género Helichrysum en las Islas Baleares. Symposium de Ecología Insular. 18-23 noviembre 2002.

AEET. Santa Cruz de La Palma. Spain.

Paun O, Stuessy TF & Hörandl E. 2006. The role of hybridization, polyploidization and glaciation in the origin

and evolution of the apomictic Ranunculus cassubicus complex. New Phytologist,l 171: 223-236.

Rieseberg LH & Ellstrand NC. 1993. What can molecular and morphological markers tell us about hybridization?

Critical Reviews in Plant Sciences, 12: 213-241.

Rieseberg LH. 1997. Hybrid origins of plant species. Annual Review in Ecology and Systematics, 28: 359-389.

Schwarzbach AE, Donovan LA & Rieseberg LH. 2001. Transgressive character expression in a hybrid sunflower

species. American Journal of Botany, 88: 270-277.

Shipunov AB & Bateman RM. 2005. Geometric morphometrics as a tool for understanding Dactylorhiza

(Orchidaceae) diversity in European Russia. Biological Journal of the Linnean Society, 85: 1-12.

Shipunov AB, Fay MF, Pillon Y, Bateman RM & Chase MW. 2004. Dactylorhiza (Orchidaceae) in European

Russia: combined molecular and morphological analysis. American Journal of Botany, 91: 1419–1426.

Vriesendorp B & Bakker FT. 2005. Reconstructing patterns of reticulate evolution in angiosperms: what can we

do? Taxon 54: 593–604.

154 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium

Chapter 5 Leaf shape variation and taxonomic boundaries involving the endemic Helichrysum crassifolium (L.) D. Don

page

5.1. Introduction 157

5.2. Objectives 159

5.3. Materials and methods 161

5.3.1. Plant sampling 161

5.3.2. Environmental parameters from the sampled localities 164

5.3.3. Geometric morphometric data acquisition 165

5.3.4. Geometric morphometric analysis 166

5.3.5. Correlation between RWs data and abiotic parameters 167

5.3.6. Other statistical analyses 168

5.4. Results 169

5.4.1. Leaf dimensions 169

5.4.2. Geometric morphometric analysis 170

5.4.2.1. Allometry in geometric morphometric data 170

5.4.2.2. Relative warp analysis 171

5.4.3. Shape differences among species, populations and individuals 173

5.4.4. Shape differences unrelated to species 174

5.4.5. Correlation between leaf shape and environmental variables 176

5.5. Discussion 177

5.5.1. Helichrysum crassifolium and H. pendulum show distinct patterns of leaf 177

shape variation

5.5.2. Leaf intermediacy is not the rule in H. crassifolium and H. pendulum 178

populations

5.5.3. The nature of intraspecific leaf shape variation in H. crassifolium and H. 178

pendulum

5.6. References 182

Appendix 5.1 187 Appendix 5.2 188 Appendix 5.3 189

155 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium

THIS CHAPTER HAS BEEN PARTIALLY SUBMITTED AS AN ORIGINAL PAPER:

Conesa MÀ, M Mus & JA Rosselló. 2006. Leaf shape variation and taxonomic boundaries in two

sympatric rupicolous species of Helichrysum (Asteraceae) assessed by geometric

morphometry. Taxon. Submitted.

156 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Introduction

5.1. INTRODUCTION

Population differentiation is a generally recognized phenomenon in plants

(Bradshaw, 1984). Quantifying and understanding the causes of population differentiation are critical for assessing the taxonomic structure of species, because the level of differentiation among populations within a species may correspond with incipient speciation

(Grant, 1981; Levin, 2000; Coyne & Orr, 2004).

Character differentiation has been documented for most important features of plant structure and function, including leaf traits (Linhart & Grant, 1996). Differentiation of morphological traits has been documented in widespread species spanning a wide range of climatic and edaphic conditions (e.g. Drobná, 2010), and also in narrowly-distributed endemic species apparently lacking habitat heterogeneity (e.g., Quilichini et al., 2004).

Lineal distance measurements have traditionally been used in descriptive studies assessing patterns of morphological variation (e.g., Blackith & Reyment, 1971; Marcus,

1990), including recent studies in Gnaphalieae (Flann et al., 2008). In traditional morphometrics, the variance observed should be size dependent because the variables are measurements of size. However, the variance in shape is not fully explained by the variance in size, but is simply overwhelmed by it (Zelditch et al., 2004). In addition, variation is only quantified between the endpoints of linear distance, and even then it does not specify which endpoint moves relative to the other (e.g., Carvajal-Rodríguez et al., 2005). Hence, little information about shape is registered, due to such high correlation between linear distance measurements and size (Bookstein et al., 1985; Zelditch et al., 2004).

Oppositely to traditional approaches, geometric morphometrics capture and preserve geometry of shape as a cohesive whole, allowing complete reconstruction of the

157 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Introduction shape (Rohlf & Marcus, 1993; Adams et al., 2004). In the landmark-based geometric morphometric methods such as the Generalized Procrustes Analysis (GPA; Rohlf & Slice,

1990), specific points with biological and geometrical signification (landmarks, LM) are located on the organism. These points are assumed to be homologous, at least in a geometric sense, because LM-based methods operate only with the two or three dimensional coordinates of these reference points. Therefore, the objects studied should be directly comparable (e.g., Adams et al., 2004). The LM-based technique of geometric morphometrics is the most effective way to capture information about the shape of an organism, especially when combined with multivariate statistical procedures (Rohlf &

Marcus, 1993). In fact, there is an increasing number of works using geometric morphometrics methods to study a polymorphism for which classical distance and lineal multivariate methods have failed to detect shape differentiation between ecotypes or species (e.g., Stone, 1998; Cavalcanti et al., 1999; Baylac et al., 2003; Carvajal-Rodríguez et al., 2005; Shipunov et al., 2005; Mutanen & Pretorius, 2007; Seiler & Keeley, 2009).

Overall, botanical studies using geometric morphometric methods are sparse (e.g.,

Ray, 1992, in Syngonium; Kores & al, 1993, in Cyrtostylis; Jensen et al., 2002, in Acer;

Shipunov & Bateman, 2005 and Shipunov et al., 2005, in Dactylorhiza; Volkova et al., 2007, in Nymphaea; Viscosi et al., 2009a, 2009b, in Quercus) despite the fact that many botanical features of taxonomic importance (e.g., leaves, sepals, petals) fit well with typical geometric morphometric conditions (Shipunov & Bateman, 2005).

158 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Objectives

5.2. OBJECTIVES

In this chapter, phenotypic leaf variation (e.g., Figure 5.1) and association between leaf traits and geographic and environmental characteristics were investigated in the insular endemic H. crassifolium and the sympatric widespread H. pendulum using geometric morphometric approaches in order (i) to ascertain to what extent leaf shape is a good taxonomic feature to characterize the endemic H. crassifolium, and to discriminate between this and its rupicolous relative H. pendulum, and (ii) to assess the pattern of leaf polymorphism in both species and the likely causes underlying it.

Figure 5.1. (Next page). Morphological variation of the leaves in H. crassifolium and H. pendulum from the Balearic Islands. Each group of ten leaves belongs to a standard individual from a specific population. Two columns on the left represent H. crassifolium populations from the western mountains of Majorca (Serra de Tramuntana), appearing the population from Dragonera islet into a white box. Next column to the right represents the populations of that species from the eastern mountains of Majorca (Serres de Llevant), and the single population from Minorca, appearing into a white box. The column on the right winger represents the populations of H. pendulum from Majorca and the single one outside, from Cabrera island, appearing into a white box. The pictures under each column belong to standard plant morphologies of populations included in the column.

159 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Objectives

160 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Materials and methods

5.3. MATERIALS AND METHODS

5.3.1. Plant sampling

All along the eastern Balearic Islands, 47 populations were selected for study, 37 belonging to H. crassifolium and 10 to H. pendulum (Figure 5.3). From each population, from one to 17 individuals were analyzed (Table 5.2), resulting in a sample size of 333 H. crassifolium and 84 H. pendulum individuals. Healthy vegetative shoots were collected for each plant, and five of the most basal, well developed, non-senescent, healthy leaves were selected for morphometric analyses.

Species ascription was based on Table 5.1. Many authors use to characterize H. crassifolium (even in keys to species) as the sole Mediterranean species of the sect.

Stoechadina with “basal leaf rosettes” (e.g., Bolòs & Vigo, 1995; Galbany-Casals et al.,

2004, 2006a,b,c, 2009). Nevertheless, H. crassifolium is a woody and dense cushion without leaves in the innermost part of the plant. Thus, it does not show leaves in the basis.

Also, this trait is puctualized as “showing leaf rosettes arising terminally on thick woody erect stems” (e.g., cf. Bolòs et al., 1990; Galbany-Casals et al., 2006a), which seems more appropiate. Nevertheless, such appreciation is due to the flat disposition and aspect of the leaves rather than to the presence of rosettes. The high density and disposition of the leaves on short woody non-floral branches, uses to create an external dense cover promoting the senescence of lower leaves in the stem. Only the most terminal leaves use to be alive, resembling a terminal rosette (i.e., very short internodes in helicoidally disperse filotaxis; e.g., Font Quer, 2000). However, the internode length and the leaf disposition pattern along the branches seems to be very similar in both H. crassifolium and H. pendulum (Figure 5.2). Thus, no real rosettes have been observed in the endemic species, or otherwise the same can be observed in H. pendulum. In fact, the “rosette” trait is not

161 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Materials and methods used to describe H. crassifolium in other floras (e.g., Clapham, 1976). Since that, this trait has not been used here to discriminate between both rupicolous species.

Table 5.1. Morphological differences used to discriminate between the two rupicolous Helichrysum species occurring in the eastern Balearic Islands. From personal observations.

H. crassifolium H. pendulum Habitat Mainly rupicolous. In calcareous cliffs, but also in Rupicolous. In calcareous cliffs, rarely S faced. From conglomerates and artificial vertical habitats. Up to 900 sea level to 1450 m m Plant habit Woody and globose shrub, frequently adpressed to the Laxous plant, even in woody plants with high density of wall, forming a dense cushion. Only terminal leaves of leaves the branches use to be alive Leaves positioned to cover the whole plant Sparse leaves without specific positioning. Leaf shape Flat, from wide-spatulate or obovate to narrow Partially revolute, from linear to narrow lanceolate lanceolate Leaf traits Very hairy in both faces, specially long hair in the Adaxial face almost glabrous or only disperse hairs. abaxial one. Laxous and spongy-padded hair coat Abaxial face hairy. Short hair Subcrassolate Non-subcrassolate Unappreciable midrib in the adaxial face Clear midrib in the adaxial face Long decurrent basis Short decurrent basis Long petiole nail, hugging most of the stem Short petiole nail, not hugging the stem Glands Abundant. Non-smelling Poor, mostly in adaxial face. Non-smelling

Figure 5.2. Terminal branches of H. pendulum (left in A and B) and H. crassifolium (right in A and B). A) Normal aspect. B) Branches in A with cutted leaves. Note the very similar helicoidal disposition and internode length. Thus, no evidence for leaf rossettes is found to discriminate both species. Note also senescent leaves in H. crassifolium.

162 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Materials and methods

Table 5.2. Sampled populations in this study. Location (UTM), altitude (m) and cliff exposition are shown, as well as the acronym (code) used and the number of individuals sampled in each population. An asterisk in the exposition indicates sea influence.

UTM Altitude (m) H. crassifolium H. pendulum Island Location 1x1 km and exposition (31 S) Code n Code n Formentor 0515-4422 250 N* FOR 10 - - Sa Bretxa 0510-4420 150 N* SBR-C 3 SBR-R 16 Ses Fontanelles 0511-4421 200 N* SFO 3 - - Mirador Formentor 0509-4420 100 NW* MIF 6 - - Cavall Bernat 0506-4419 200 NW* BER-C 7 BER-R 8 La Victòria 0514-4413 350 NW VIC-C 15 VIC-R 2 Ses Parades 0492-4415 450 NW SPA-C 10 SPA-R 5 Tomir 0492-4409 750 W TOM 11 - - Sa Calobra 0483-4411 10 N* CAL-C 8 CAL-R 2 Nus de sa Corbata 0483-4409 500 N - - NCO 10 Massanella 0487-4406 1300 N MAS 15 - - Puig Major 0482-4406 1400 SE PMA 10 - - Escorca 0485-4408 600 NW - - ESC 13 Gorg Blau 0484-4406 650 SE GBL-C 1 GBL-R 10 Caimari 0491-4403 250 E CAI 10 - - Alaró 0482-4399 750 NE ALA 10 - - Barranc Biniaraix 0479-4401 700 W BBI 10 - - Alfàbia 0476-4399 1000 NW ALF 11 - - Cala Deià 0469-4401 50 W CDE* 10 - - Comuna de Bunyola 0476-4393 350 E CBU 9 - - Na Fàtima 0469-4394 500 NW FAT 10 - - Majorca Volta d'es General 0459-4394 200 NW* VGE 8 - - Galatzó 0456-4387 750 SE GAL 3 - - Es Burotell 0460-4383 550 N BUR 10 - - Font d'es Quer 0452-4386 800 N FQU 15 - - Punta de Sant Elm 0444-4383 50 N* PSE 10 - - Dragonera 0441-4382 200 NE* DRA 17 - - Penyal de s'Àguila 0453-4373 20 W* PAG 8 - - Talaia Moreia 0529-4403 350 NW TMO 10 - - S'Arboçaret 0531-4400 200 NE TAR 4 - - Porrassar 0530-4400 150 SE POR 4 - - Puig d'es Corb 0529-4399 350 NW PCO 11 - - Puig Poloni 0529-4399 400 NW POL 6 - - Es Verger 0530-4399 250 N VER 5 - - Ermita de Betlem 0526-4398 350 N ERM 6 - - Es Freu 0538-4399 30 N* FRE 10 - - Canyamel 0538-4390 100 E* CAN 10 - - Calicant 0521-4388 450 W CCA 10 - - Randa 0493-4375 500 N - - RAN 12 Minorca El Toro 0595-4426 300 NW TME 17 - - Cabrera L'Imperialet 0496-4331 50 NE - - CIM 6

163 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Materials and methods

Figure 5.3. Distribution of the sampled populations in the Gymnesic islands. Population acronyms correspond to the table 5.2. Locations are depicted on the 31S 1x1 km UTM grid, while the white grid corresponds to the 10x10 km UTM. Black colour represents H. crassifolium and blue colour H. pendulum. Localities shared by both species are red coloured. There is a single 1x1 UTM containing two H. crassifolium localities (PCO and POL), depicted in green colour.

5.3.2. Environmental parameters from the sampled localities

Eight climatic parameters were estimated for the sampled populations using the

Cliba2 software (v.3, Guijarro, 1996-99; available for data estimation in http://webs.ono.com/climatol/). From each UTM grid (1 x 1 km) where the Helichrysum populations were sampled (Table 5.2), the program estimated annual precipitation (P), average annual temperature (T), minimum (Tm) and maximum (TM) daily temperatures, minimum monthly temperature (Tma), evaporation (EL, Linacre, 1977) and two potential evapotranspiration values (ETPL, Linacre, 1977; ETPT, Thornthwaite, 1948). In addition, the altitude (Alt), exposition (Exp; eight categories; Table 5.2), and sea influence (SI; 0, less

164 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Materials and methods than 300 m and 1, at least 500 m away from the sea coast; no sampled populations were located between 300 and 500 m; Table 5.2) were recorded from the specific site, and SI also from GoogleTM Earth (v. 5.0; available from http://earth.google.es/).

5.3.3. Geometric morphometric data acquisition

For each individual, the adaxial face of five leaves originating from different shots was digitized with an Agfa DuoScan T1200 scanner (Agfa-Gevaert Group, Germany), requesting images in [.jpg] format with a resolution of 300 dpi. All scanned images included a digital ruler (millimeter precision) for scale reference in data acquisition.

In this work 16 LMs were selected on H. crassifolium and H. pendulum leaves

(Figure 5.4). In the designed LM configuration, LM1 to LM4 are easily recognizable in all measured leaves, representing the upper part of the petiole nail, the leaf apex, and the left and right basal extremes of the petiole nail, respectively. A second group of LMs, from LM5 to LM8, are pairs of LMs traced in the thinner part of the attenuated basis of the leaf and in the wider part of the leaf blade. The remaining LMs, those from LM9 to LM16, are pairs of

LMs located at an intermediate position between the above groups of LMs, with merely geometric significance. These three groups of LMs belong to Bookstein’s (1991) type I, II and III LMs, respectively.

LM tracing was performed using F. James Rohlf software (SUNY at Stony Brook, available at http://life.bio.sunysb.edu/morph/). LMs were digitized with tpsDIG v. 2.05; Rohlf,

2006a). All five leaves per plant were used in the analyses instead of using a leaf-shape consensus. In addition, linear measurements of leaf length (L) and width (W), and their ratio

(L/W), were obtained from the LM data (in pixels) and transformed to millimeter distances.

165 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Materials and methods

Figure 5.4. Landmark configuration recorded on each leaf. Representative leaves of H. crassifolium and H. pendulum showing the location of the 16 LMs on their digitized shapes. Red points correspond to the type I LMs, blue points to the type II LMs, and green points to the type III LMs, according to Bookstein (1991).

5.3.4. Geometric morphometric analysis

A GPA, followed by a thin-plate splines analysis (TPS; Bookstein, 1991), was performed with tpsRELW v. 1.45 (Rohlf, 2007a). Program default settings were used, as recommended by Rohlf (2007a). Thus, thus scaling factor of α=0, specimens aligned to unit centroid size, orthogonal projection method, and uniform component included in the analysis and estimated using the space complement method. These default options have also been considered to be the more appropriate for many systematic studies (e.g., Loy et al., 1993; Rohlf et al., 1996; Walker, 1996; Slice, 2001; Rohlf & Bookstein, 2003).

In TPS, shape changes are explained in terms of deformations described by the resulting principal warps, and their projections on the X,Y-coordinates of the aligned

166 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Materials and methods specimens (the partial warps). Dimensionality in the multivariate data was reduced by principal component analysis of the partial warp scores (i.e., the relative warp analysis,

RWA; Bookstein, 1991; Rohlf et al., 1996). The principal components obtained are the relative warps (RWs). Mapping of the shape changes explained by the RWs was performed through TPS deformation grids (Bookstein, 1991; Adams et al., 2004). The RWA and the deformation grids were performed using tpsRELW v. 1.45 (Rohlf, 2007a).

The suitability of the data set to be used in statistical analyses was tested with tpsSMALL v. 1.20 (Rohlf, 2003), that computes Procrustes distances and regresses it on the geometric morphometric data. Results showed correlations higher than 0.99999 and no important deviations of single specimens, thus allowing the insightful interpretation of the results from this data (e.g., Dryden & Mardia, 1998; Rohlf, 1999; Fadda & Corti, 2001).

Allometry in the shape data was tested with tpsREGR v. 1.33, (Rohlf, 2007b) with a multivariate regression of shape variables (partial warps plus the uniform component) onto size (as centroid size, CS, the standardized measure of size in geometric morphometry).

CS was log-transformed to reduce dimensionality and meet the assumption of homoscedasticity for regression analyses (Sokal & Rohlf, 1981; Kachigan, 1991; Rohlf

2007b).

5.3.5. Correlation between RWs data and abiotic parameters

A two-block partial least squares analysis (2B-PLS) was performed with tpsPLS v.

1.18 (Rohlf, 2006b) to test for correlations between leaf shape (i.e., partial warps plus the uniform component) and the abiotic parameters for each sampled locality. The 2B-PLS finds relations between the two blocks of variables through predictor variables (latent variables or singular warps) explaining covariances between the two blocks (Rohlf & Corti, 2000; Fadda

& Corti, 2001). Categorical variables were standardized (by default) by the program, while

167 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Materials and methods partial warps were not, as recommended by Rohlf (2006b). Results are correlated pairs of vectors (i.e., dimensions, D) onto which leaf shapes are located to produce scores for both the variables and the partial warps. Cross-validation and 999 permutation tests were also computed.

5.3.6. Other statistical analyses

One-way analysis of variance (ANOVA) was used to test for differences in mean values in the linear measurements between species. Due to the high similarity among the leaves of the same individual, and to detect possible effects of redundancy, such analyses were performed and compared using five and a single random leaf per individual.

To test for shape differences between and within species, populations and individuals, multivariate analysis of variance (MANOVA) were performed for CS and the five first RWs obtained in the RWA were performed. To estimate the contribution of each random effect to the variance of these dependent variables, nested ANOVA was performed for each of the five first RWs, the CS, and for the linear measurements (L, W and L/W). The variance components in each case were estimated through the “varcomp” procedure, using the restricted maximum likelihood method. The designed model was that of testing for species, populations and individuals, nesting individuals within populations, and in turn populations within species. Since the individual effect was taken into account, and differing from the one-way ANOVAs, nested ANOVAs were performed only using five leaves per plant. To test for the correct discrimination of the species, a canonical discriminant analysis

(CDA) of all the partial warps and the uniform component was performed, including the data for all sampled leaves. The one-way and nested ANOVAs, the MANOVAs and the CDA analyses were performed in SPSS v.15 for Windows (SPSS Inc., Chicago, IL).

168 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Results

5.4. RESULTS

5.4.1. Leaf dimensions

The one-way ANOVA results showed that average linear measurements of W, L and the L/W ratio, for five leaves per plant, were significantly different for each Helichrysum species (Table 5.3). However, the highly overlapping mean + SD values between both species, and the much lower F value for L compared to W and L/W, pointed to possible redundancy effects due to the high similarity among a single plant leaves. This was also detected in the MANOVAs and the nested ANOVAs. Thus, one-way ANOVA using a single random leaf per plant was performed to test for trait differences between species. This analysis results evidenced no significance in L average differences between species

(F=2.67, p=0.103), thus highlighting the suspected redundancy effects. Nevertheless, overlapping values linking both species were observed and no absolute distinction between

H. crassifolium and H. pendulum could be made (Figure 5.5).

Overall, the range of variation for W was noticeably higher in H. crassifolium (3.9-

25.2 mm) than in H. pendulum (1.1-9.3 mm). Therefore, given the similarity in the L variation range, important differences in L/W ratio were detected between species. The endemic species showed conspicuous narrower values (1.8-9.4) than H. pendulum (5.2-

43.8). On the other hand, with respect to L, some differences were detected in the nested

ANOVA results responding to groups of populations rather than to species adscription (ca.

15% difference, p<0.0001; Table 5.4) Such differences were almost parallel to those shown by CS, since the higher influence of longer measurements in its calculation (i.e., square root of the sum of the squared Euclidean distances from each LM to their centroid; Bookstein,

1991).

169 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Results

Table 5.3. Descriptive parameters for leaf length, leaf width and its ratio (L/W) and associated one-way ANOVA F values for each species, using five leaves per plant. An asterisk denotes significant values at P<0.0001. F values obtained in the one-way ANOVA using a single random leaf per plant are also shown, to denote lack of significance in L (p=0.103). No important variation in range, mean, SD and SE was shown using 5 or a single leaf per plant.

Five leaves per plant One leaf per plant Variable Species n Range Mean SD SE F n F Length CRA 1665 16.2-83.0 44.48 10.66 0.26 31.029* 333 2.67 PEN 420 20.4-97.8 47.93 13.77 0.67 84 Width CRA 1665 3.9-25.2 11.00 3.25 0.08 2387.53* 333 483.35* PEN 420 1.1-9.3 3.03 1.57 0.08 84 L/W CRA 1665 1.8-9.4 4.24 1.13 0.03 5802.09* 333 1183.23* PEN 420 5.2-43.8 18.52 7.32 0.04 84

Figure 5.5. Scatter plot of the length/width (L/W) ratio vs. width values for H. crassifolium (in black) and H. pendulum (in white).

5.4.2. Geometric morphometric analysis

5.4.2.1. Allometry in geometric morphometric data

Allometry was tested with a multivariate regression of shape variables (partial warps and the uniform component) onto size (log-CS, as independent variable). No correlation between log-CS and any shape variable was higher than 0.4. Multivariate tests of significance showed that interspecific allometry explained only a 1.4% of total shape variation defined by the partial warps (p<0.0001 in 999 permutation tests for both species; data not shown).

170 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Results

5.4.2.2. Relative warp analysis

The partial warps analysis produced 13 partial warps plus the uniform component.

From the uniform component only the X axis was informative, representing variation in leaf width, whereas the Y axis resulted in deformation grids depicting differences regarding to left or right sense torsions. The RWA reported 28 RWs to explain the whole variation among the data set. Of those, the first RW (RW1) explained the 56.2% of total shape variation, accounting for 75.2% when the RW2 was added; suggesting that shape variation had very clear components (Table 5.4). The scatter plot of the first two RWs showed the ordination of most H. crassifolium leaves through the positive values of the RW1 axis, whereas all H. pendulum leaves showed negative values (Figure 5.6).

The RW1 axis was mainly related to leaf width, but also to the position of the widest part of the leaf blade, the shape of the apex, and the length and width of the petiole nail. In contrast, RW2 axis especially accounted for the shape changes described by RW1 but with low influence of leaf width. Thus, in the RW1 vs. RW2 scatter (Figure 5.6), extreme H. pendulum leaves showed very narrow linear shapes, with the widest part of the blade located at the middle, with sharp apex, and a very short petiole nail. In contrast, H. crassifolium specimens showing the most positive scores on that axis showed spatulate shapes, with obtuse to rounded apex, the widest part of the leaf blade was located close to the tip, and showed a much developed petiole nail. Overlapping accessions linking both species in the multivariate geometric morfo-space were characterized by having narrow- oblanceolate leaves, with the wider part of the blade located in the upper half, and with intermediate dimensions concerning the petiole nail. Samples from at least 15 H. crassifolium and 7 H. pendulum populations showed intermediate leaf shapes (data not shown). These overlapping accessions did not respond to any geographic pattern.

171 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Results

Besides of that degree of overlapped samples, the CDA of all partial warps and the uniform component resulted in a single canonical function explaining all variation detected

(canonical correlation of 0.895; Wilks’ λ=0.198; df=28; p<0.0001). So, the 98.8% of total cases were correctly classified (99.3% for H. crassifolium, and 96.7% for H. pendulum) after cross validation tests.

RW No. SV % Cum % 1 3.90324 56.17 56.17 Table 5.4. Results of the 28 RWs obtained in the RWA using 2 2.2724 19.04 75.20 five leaves per plant and including the uniform component 3 1.51361 8.45 83.65 (calculated with the complement method). TpsRELW program 4 1.44975 7.75 91.40 default settings were maintained, thus unit centroid size as 5 0.73010 1.97 93.36 alignement scaling method, orthogonal alignement projection 6-10 - - 97.81 method, and equal weights given to partial warps at all spatial 11-28 - - 100 scales (α = 0).

Figure 5.6. Scatter plot of the two first relative warps obtained with tpsRELW on 16 leaf landmarks from 2085 samples, for H. crassifolium (in black) and H. pendulum (in white). Tps deformation grids corresponding to several specimens depicting extreme values (denoted with an arrow) have been located around the plot. Extremes of the two axes are marked with a white cross, and the corresponding deformation grids are also depicted.

172 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Results

5.4.3. Shape differences among species, populations and individuals

The MANOVA results showed that the main two RWs explained ca. 77% of the total variance between species, and very low differences were observed when including also

RW3-5 (RW1-2, Wilks’ λ=0.232, F2,2082=3442.808, p<0.0001; RW1-5, Wilks’ λ=0.220,

F5,2079=1475.382, p<0.0001). Although, when testing for populations, the proportion of the

variance explained reached ca. the 90% (RW1-2, Wilks’ λ=0.119, F92,4072=83.896, p<0.0001;

RW1-5, Wilks’ λ=0.060, F230,10127.3=33.409, p<0.0001). However, differing from the MANOVA for species (where it had almost no effect), the inclusion of CS in the model for populations explained a slightly higher proportion of the total variance, up to the 97% (CS+RW1-2,

Wilks’ λ=0.068, F138,6102.5=64.051, p<0.0001; CS+RW1-5, Wilks’ λ=0.033, F276,12118.2=33.792, p<0.0001). Finally, when testing for individuals, ca. the 99% of the variance was explained

(RW1-2, Wilks’ λ=0.011, F832,3334=33.409, p<0.0001; RW1-5, Wilks’ λ=0.002,

F2080,8325.5=10.647, p<0.0001). This would indicate that less than the 1% of the variance was located at intra-individual level, thus accounting for the variability among leaves of the same individual. Also, this would be in concordance with the redundancy effects of including five leaves in the one-way ANOVA (Table 5.3).

The nested ANOVAs for each geometric morphometric and linear variables allowed the identification of the components of the variance detected in the MANOVAs. Results showed that, for RW1, the 89.6% of the variance was explained by species (p<0.0001), the

5.1% by individuals (p<0.0001), the 2.7% by populations (p<0.0001), and the 2.6% was located among leaves of the same individual (i.e., error variance of the model), (Appendix

5.1).

On the other hand, for RW2 the 48.7% of the variance was explained by individuals

(p<0.0001), the 21.3% by populations (p<0,0001), and only the 0.5% by species (p=0.328), whereas intra-individual variability accounted for the 29.6% of the total variance. For RW3,

173 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Results

RW4 and RW5 (explaining a very low percentage of the total variation; Table 5.4), intra- individual variation explained the 95.7%, 41.2% and 77.7% of the variance, respectively.

When testing for CS (as a measure of geometric leaf size), the 56.1% of the variance was explained by individuals (p<0.0001), the 30.1% by populations (p<0.0001), 0% for species, and the 13.8% was intra-individual variability.

For linear variables, L was explained a 53.5% by individuals (p<0.0001), 30.1% by populations (p<0.0001), a non significant 2.9% by species (p=0.135), and a 13.8% by intra- individual variability. This is notoriously parallel to the results shown by CS. On the other hand, W and L/W variability was highly explained by species (75.2% and 87.2%, respectively; p<0.0001).

5.4.4. Shape differences unrelated to species

Since the Tramuntana range (in the western Majorca) is the single region where both rupicolous species can be found growing in close vicinity (Figure 5.3), the nested

ANOVAs were repeated changing the variable “species” for the new variable “TRA”

(Appendix 5.1) separating populations in being or not in the Tramuntana range. In this case,

CS and L variances were explained the 10.5% (p=0.01) and the 12.6% (p=0.006), respectively, at this new variable level; whereas it was only the above stressed 2.9%

(p=0.135) and 0% (p=0.888) using the variable “species” (Appendix 5.1). This should result even more pronounced when testing for H. crassifolium leaves alone, since the endemic is the single of the rupicolous species showing many populations also outside the Tramuntana range (Figure 5.3). Surprisingly, non-significant (p>0.09) percents of ca. 4% were retrieved for both CS and L.

Moreover, when repeating the nested ANOVAs but grouping the populations from

Tramuntana where other Helichrysum species are growing in close vicinity (i.e., DRA, MAS

174 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Results and PAG; all in close relation with H. italicum subsp. microphyllum) with the non-

Tramuntana populations in the new variable TRA2 (Appendix 5.1), the percents of variance explained by CS and L for both species together incremented to 14.2% (p=0.001) and

16.8% (p=0.001), respectively. Indeed, it resulted higher and also significant for H. crassifolium populations alone (9.5%, p=0.022, and 9.6%, p=0.022, respectively).

With respect to RW2, all the above analyses resulted in no variance explained by this new variables and, as for species, most of the variance was explained at individual level

(ca. 49%, p<0.0001 in all cases, Appendix 5.1).

The above significant differences in L among populations when testing by the variables TRA and TRA2 was also retrieved in the one-way ANOVAs for such new variables. Since the redundancy effects observed, a single random leaf per plant was used.

Differences in the mean L were 11.3% smaller for non-TRA, and 14.7% smaller for non-

TRA2 (Table 5.5); whereas those were non-significant (p=0.103) when testing for species.

Table 5.5. Descriptive parameters for leaf length, for the nested ANOVA for the variables “TRA” and “TRA2” instead of species (see text). Mean L values, SD and SE values, the proportion of L difference (%), and the F values for each variable are shown. An asterisk denotes significance (p<0.0001).

Variable n Mean L (mm) SD SE % difference F TRA 306 46.85 10.92 0.52 18.258* Non-TRA 111 41.54 12.04 1.14 11.33 TRA2 266 47.99 10.61 0.65 39.841* Non-TRA2 151 40.94 11.64 0.94 14.69

175 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Results

5.4.5. Correlation between leaf shape and environmental variables

The 2B-PLS method was used to analyze the covariation between shape variables

(partial warps, including the uniform component) and 11 environmental variables in each species. The first three dimensions (i.e., D1-D3) extracted for H. crassifolium accounted for

93.8% 4.0%, and 1.5% of the total squared covariance, respectively, with significance only for D1 (p=0.005, in 999 random permutations). The correlation for D1 was low but significant (0.4082; p=0.001, calculated from 999 random permutations) and it was influenced by SI (singular value, SV=-0.4419), but also almost equally by a high amount of variables (Alt, P, T, Tm, Tma and ETPT), (Appendix 5.2).

For H. pendulum, the first three dimensions accounted for 63.11%, 34.25%, and

1.82% of the total squared covariance, respectively, but it was only significant for D2

(p=0.037), (Appendix 5.2). Again, the correlation for the dimensions explaining most of the squared covariance was statistically significant (calculated from 999 random permutations) but with low values, thus 0.3602 (p=0.001) for D1, and 0.3652 (p=0.001) for D2. The D1 was mostly influenced by Alt (SV=0.5446) and SI (SV=0.4403), and the D2 almost equally by TM, EL and ETPL, (SVs=0.4837, 0.4613, 0.4680, respectively; Appendix 5.2). The biplot for the two first dimensions (D1 in abscissa and D2 in ordinate) retrieved by the tpsPLS software is shown in Appendix 5.3, for both species and for separated species. Leaf deformation grids corresponding to the extremes of the axis are also shown.

176 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Discussion

5.5. DISCUSSION

5.5.1. Helichrysum crassifolium and H. pendulum show distinct patterns of leaf

shape variation

The taxonomic distinctiveness of the Balearic endemic H. crassifolium from the remaining species of section Stoechadina has been traditionally supported by independent researches based on morphological features (Clapham 1976, Hilliard & Burtt 1973, 1981;

Galbany-Casals et al. 2004, 2006a; 2009). In fact, for more than 250 years it has never put in the synonymy, or subsumed at any infraspecific rank, under any other Helichrysum species (Rosselló & Sáez, 2001). This taxonomic distinctiveness contrasts with the few diagnostic morphological features used to discriminate from the remaining woody

Mediterranean species. The present study shows that vegetative features, like leaf shape and lineal leaf measurements (W and L/W ratio), can be used with certainty to discriminate between H. crassifolium and the closely related H. pendulum. More importantly, both leaf dimensions and leaf shape show contrasting patterns of variation between both species

(Figures 5.5 and 5.6) and support (past and current views) that both entities should be given independent taxonomic distinction (Galbany-Casals et al. 2006a, and references therein).

As suggested in the results, the H. crassifolium individuals from the eastern

Majorcan and the Minorcan populations (where H. pendulum has not been found) present shorter leaves than those from the Tramuntana range, in western Majorca (Table 5.4).

Nested ANOVA results clearly showed that W (and L/W) is highly dependent of species.

Oppositely, differences observed in L and CS (the latter highly related to L, since the way in which it is calculated; Bookstein, 1991) are mainly located at individual level, as it is for

RW2. Therefore, their variation is expected to be especially due to intra-population processes rather than related to any taxonomic and/or environmental factor. Nevertheless,

177 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Discussion both L and CS showed also a ca. 30% of the variance explained at population level, which indicates that there must be also some factor discriminating between groups of populations, and independent of the species.

5.5.2. Leaf intermediacy is not the rule in H. crassifolium and H. pendulum

populations

Interestingly, despite the high levels of intraspecific variation detected in leaf features in both species, overlapping accessions with intermediate shape were extremely low (less than 2.5% of the total data set). In fact, CDA using scores of all principal warps and the uniform component correctly classified 98.8% of the leaves.

Moreover, a higher number of individuals (11.3%) showing intermediate morphological appearance for linear leaf characters were found. These observations suggest that in Helichrysum, geometric morphometric analysis of leaf shape might give better taxonomic descriptors than the commonly used features of leaf length, leaf width, and leaf L/W ratio (e.g. Galbany-Casals et al. 2006a). In fact, nested ANOVAs showed that RW1

(mostly explaining shape changes related to leaf width) mainly accounted for species differences (89.6%), stressing the importance of leaf with in the discrimination between both rupicolous species.

5.5.3. The nature of intraspecific leaf shape variation in H. crassifolium and H.

pendulum

Results of the 2B-PLS showed that intraspecific leaf shape variation in both species was not clearly correlated to any of the 11 environmental factors scored. In any case, the same correlation was found in a high proportion of the tested variables, showing no clear relation among them. Thus, such correlation could be also due to bias in the localities

178 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Discussion selected, to restrictions in the program calculation (i.e., ± 1 km), or even due to chance.

Indeed, the trend of morphological variation observed followed neither a geographic cline nor was related to island boundaries. Nevertheless, it can not be discarded that other environmental, lithological or ecophysiological variables not considered in this study might show significant correlations with leaf shape in the Balearic Helichrysum taxa. Also, further approaches to that partial correlation observed should be addressed to fully discard the influence of any of the environmental variables used.

Since the absence of clear correlations of size and shape with the environmental variables scored, the non species-related intra- and the inter-population variation can be associated to hybridization processes. Thus, random crosses among different generations of hybrids and introgressants into each population permit to interpret that most of the variation is located at intra-population level. In fact, the nested ANOVAs testing for populations with the other rupicolous species in close vicinity (i.e., “TRA”), showed that up to the 12.6% of the total variance was explained by L. The smaller leaf length of the H. crassifolium eastern Majorcan and in the Minorcan populations could be related to the current absence of H. pendulum in those regions.

In addition, independent of the species, some populations must show similar variation patterns to explain the variability located at population level. Otherwise, there must be another factor involved in such pattern. In turn, when the existence of a third, non- rupicolous, Helichrysum taxon in close vicinity was also included in the nested ANOVAs above (i.e., “TRA2”), the total variance explained by L was even higher (up to 16.8%).

Notoriously, this higher variance was retrieved when using both species populations together, rather than in a separate fashion. However, mean population values confirmed that the H. pendulum populations from outside the Tramuntana range (those from RAN, isolated in the center of Majorca, and CIM from the Cabrera island) had conspicuously

179 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Discussion shorter leaves. This would favour the existence of other Helichrysum taxa involved in such putative hybridization processes.

Overall evidence suggests that biologically meaningful factors rather than abiotic should be invoked to explain the nature of intraspecific leaf shape variation in both species.

Phylogenetic relationships within the Mediterranean section Stoechadina were generally unresolved using nuclear ribosomal ITS and ETS sequences due to very low levels of sequence divergence and support for the internal nodes, and the presence of intragenomic polymorphisms in some species (Galbany-Casals et al., 2004, 2009). Ancient events of allopolyploidization or recent gene flow by hybridization and introgression between sympatric species were hypothesized to be involved in the diversification of lineages within section Stoechadina (Galbany-Casals et al., 2004, 2009). In fact, intermediate individuals between H. crassifolium, and H. pendulum or H. stoechas, showing an intermediate morphological appearance, between these species has been reported and interpreted as being of recent hybrid origin (Cosín et al., 2002; Mus et al., 2002; Galbany-Casals et al.

2006a,b,c, 2009).

In the eastern Balearic Islands two other species, H. stoechas and H. italicum subsp. microphyllum, occur. Although growing on non-rupicolous environments, their populations grow usually in sympatry with H. crassifolium and H. pendulum and they probably share generalist pollinator insects. In fact, in coastal populations of the eastern Balearic Islands,

Gil (1994) reported six pollinator species in H. italicum subsp. microphyllum and 13 in H. stoechas, finding all those pollinating the former also in the latter. Indeed, up to four orders and nine families were represented, also stressing the low specificity of the pollinator species in both Helichrysum taxa. Moreover, 29 pollinator species, and up to 88 potential ones, were detected by Stang et al. (2006) in H. stoechas populations in SE Spain, resulting in the species with higher diversity of pollinators observed in their study. Pollinator sharing

180 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Discussion would heavily facilitate interspecific crosses among all the species of the genus

Helichrysum in the Balearic Islands.

Thus, past or recent interspecific gene flow might explain the high intraspecific polymorphism in leaf features shown by the narrow endemic H. crassifolium and the

Western Mediterranean H. pendulum. This scenario is in agreement with the existence of a continuous and even overlapping range in the diagnostic traits discriminating species (i.e.,

W, L/W and RW1), and also with the important shape variation component located at population and individual levels rather than at species level. This strongly suggest that the patterns of leaf shape variation in these species could be linked to recurrent hybridization and introgression processes, as suggested for most taxa of section Stoechadina

(Jeanmonod, 1996; Galbany-Casals et al. 2006a, 2009).

181 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium References

5.6. REFERENCES

Adams DC, Rohlf FJ & Slice DE. 2004. Geometric morphometrics: ten years of progress following the

“revolution”. Italian Journal of Zoology, 71: 5-16.

Aghababyan M, Greuter W, Mazzola P & Raimonfo FM. 2007. Typification of Sicilian Helichrysum (Compositae)

revised. Taxon, 56: 1285-1288.

Anderberg AA. 1991. Taxonomy and phylogeny of the tribe Gnaphalieae (Asteraceae). Opera Botanica, 104: 1-

195.

Bayer RJ, Puttock CF & Kelchner SA. 2000. Phylogeny of South African Gnaphalieae (Asteraceae) based on

two noncoding chloroplast sequences. American Journal of Botany, 87: 259-272.

Baylac M, Villemant C & Simbolotti G. 2003. Combining geometric morphometrics with pattern recognition for

the investigation of species complexes. Biological Journal of the Linnean Society, 80: 89-98.

Blackith RE & Reyment RA. 1971. Multivariate morphometrics. Academic Press: New York.

Bolòs O & Vigo J. 1995. Flora dels Països Catalans. Vol. III. Editorial Barcino: Barcelona.

Bolòs O, Vigo J, Masalles RM & Ninot JM. 1990. Flora manual dels Països Catalans. Editorial Pòrtic: Barcelona.

Bookstein FL, Chernoff B, Elder RL, Humphries JMJr, Smith GR & Strauss RE. 1985. Morphometrics in

evolutionary biology. Special publication 15. Academy of Natural Sciences Press: Philadephia.

Bookstei FL. 1991. Morphometric tools for landmark data: geometry and biology. Cambridge Univ. Press:

Cambridge

Bradshaw AD. 1984. Ecological significance of genetic variation between populations. In: Dirzo R. & Samkhan J.

(eds). Perspectives on plant population biology. Sinauer Associates: Sunderland.

Carvajal-Rodríguez A, Conde-Padín P & Rolán-Álvarez E. 2005. Decomposing shell form into size and shape by

geometric morphometric methods in two sympatric ecotypes of Littorina saxatilis. Jpurnal of Molluscan

Studies, 71: 313-318.

Cavalcanti MJ, Monteiro LR & Lopes PRD. 1999. Landmark-based morphometric analysis in selected species of

serranid fishes (Perciformes: Teleosti). Zoological Studies, 38: 287-294.

Clapham AR. 1976 Helichrysum. In: Tutin TG, Heywood VH, Burges N.A, Moore DA, Valentine DH, Walters SM

& Webb DA (eds). Flora Europaea, vol. 4. Cambridge Univ. Press: London.

Cosín R, Conesa MÀ, Mus M & Rosselló JA. 2002. Hibridación reticular en especies insulares de Helichrysum

Mill. (Compositae) detectada mediante ISSR y secuencias de ADN cloroplástico. 1er Congreso de

182 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium References

Biología de la Conservación de Plantas, Jardín Botánico de la Universidad de Valencia, 2-5 octubre

2002. Valencia. Spain. pàg. 93.

Coyne JA & Orr HA. 2004. Speciation. Sinauer Associates: Sunderland.

Drobná J. 2010. Morphological variation in natural populations of Lotus corniculatus in association to

geographical parameters of collecting sites. Biologia 65: 213-218.

Dryden IL & Mardia KV. 1998. Statistical shape analysis. John Wiley & Sons: New York.

Fadda C & Corti M. 2001. Three-dimensional geometric morphometrics of Arvicanthis: Implications for

systematics and taxonomy. Journal of Zoological Systematics and Evolution Research, 39: 235-245.

Flann C, Breitwieser I, Ward JM, Walsh NG & Ladiges PY. 2008. Morphometric study of Euchiton traversii

complex (Gnaphalieae: Asteraceae). Australian Systematic Botany, 21: 178-191.

Font Quer P (director). 2000. Diccionario de Botánica. Ediciones Península: Barcelona.

Galbany-Casals M, Garcia-Jacas N, Susanna A, Sáez L & Benedí C. 2004. Phylogenetic relationships in the

Mediterranean Helichrysum (Asteraceae, Gnaphalieae) based on nuclear rDNA ITS sequence data.

Australian Systematic Botany, 17: 241-253.

Galbany-Casals M, Garcia-Jacas N, Sáez L, Benedí C & Susanna A. 2009. Phylogeny, biogeography, and

character evolution in Mediterranean, Asiatic, and Macaronesian Helichrysum (Asteraceae,

Gnaphalieae) inferred from nuclear phylogenetic analyses. International Journal of Plant Sciences. 170:

365-380.

Galbany-Casals, M., Sáez, L. & Benedí, C. 2006a. A taxonomic revision of Helichrysum Mill. sect. Stoechadina

(DC.) Gren. & Godr. (Asteraceae, Gnaphalieae). Canadian Journal of Botany, 84:1203-1232.

Galbany-Casals M, Sáez L & Benedí C. 2006b. Conspectus of Helichrysum Mill. sect. Stoechadina (DC.) Gren.

& Godr. (Asteraceae, Gnaphalieae). Orsis 21: 59-81.

Galbany-Casals M, Sáez L, Benedí C & Jarvis CE. 2006c. Typification of names in Gnaphalium L. and

Helichrysum Mill. (Asteraceae), and some taxonomic notes. Taxon, 55: 489-501.

Gil L. 1994. Biología reproductiva de la flora litoral de las Islas Baleares. I. Dunas y roquedos marítimos.

Dissertation, University of the Balearic Islands: Palma de Mallorca.

Grant, V. 1981. Plant Speciation. Columbia Univ. Press: New York.

Guijarro JA. 1996-99. Cliba2 (v. 3). Palma de Mallorca. http://webs.ono.com/climatol/

Hilliard OM & Burtt BL. 1973. Notes on some plants of Southern Africa chiefly from Natal: III. Notes of theRoyal

Botanical Garden of Edinburgh, 32: 303-387.

183 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium References

Hilliard OM & Burtt BL. 1981 Some generic concepts in Compositae-Gnaphaliinae. Botanical Journal of the

Linnean Society,. 82: 181–232.

Jeanmonod D. 1996. Xanthium subg. Xanthium et Helichrysum italicum, deux cas taxonomiques ardus.

Candollea 53: 435-457.

Jensen RJ, Ciofani KM & Miramontes LC. 2002. Lines, outlines, and landmarks: morphometric analyses of

leaves of Acer rubrum, Acer saccharinum (Aceraceae) and their hybrid. Taxon, 51: 475-492.

Kachigan SK. 1991. Multivariate statistical analysis. Radius Press: New York

Kores PJ, Molvray M & Darwin SP. 1993. Morphometric variation in three species of Cyrtostylis (Orchidaceae).

Sustematic Botany, 18: 274-282.

Levin, D.A. 2000. The origin, expansion and demise of plant species. Oxford: Oxford Univ. Press.

Linacre ET. 1977. A simple formula for estimating evaporation rates in various climates, using temperature data

alone. Agricultural Meteorology, 18: 409-424.

Linhart YB & Grant MC 1996. Evolutionary significance of local genetic differentiation in plants. Annual Review

in Ecology and Sustematics, 27: 237-277.

Loy A, Corti M & Marcus LF. 1993. Landmark data: size and shape analysis in systematics. A case study on Old

World Talpidae (Mammalia, Insectivora). In: Marcus LF, Bello E & García-Valdecasas A (eds).

Contributions to morphometrics. Museo Nacional de Ciencias Naturales: Madrid.

Marcus LF. 1990. Traditional morphometrics. In: Rohlf FJ & Bookstein FL (eds). Proceedings of the Michigan

Morphometrics Workshop. Univ. of Michigan Museums: Michigan

Mus M, Cosín R, Conesa MÀ & Rosselló JA. 2002. Patrones de variación genética y fenotípica en especies del

género Helichrysum en las Islas Baleares. Symposium de Ecología Insular. 18-23 noviembre 2002.

AEET. Santa Cruz de La Palma. Spain. pàg. 34.

Mutanen M & Pretorius E. 2007. Subjective visual evaluation vs. traditional and geometric morphometrics in

species demilitation: a comparison of moth genitalia. Systematic Entomology, 32: 371-386.

Quilichini A, Debussche M & Thompson JD. 2004. Geographic differentiation of morphological traits and

isozymes in the Mediterranean island endemic Anchusa crispa: implications for the conservation of a

protected species. Biological Conservation, 118: 651-660.

Ray TS. 1992. Landmark eigenshape analysis: homologous contours: leaf shape in Syngonium (Araceae).

American Journal of Botany,. 79: 69-76.

Rohlf FJ. 1999. Shape statistics: Procrustes superimpositions and tangent spaces. Journal of Classification, 16:

197-223.

184 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium References

Rohlf FJ. 2003. tpsSMALL, version 1.20. Department of Ecology and Evolution, State Univ. of New York at

Stony Brook.

Rohlf FJ. 2006a. tpsDIG, digitize landmarks and outlines, version 2.05. Department of Ecology and Evolution,

State Univ. of New York at Stony Brook.

Rohlf FJ. 2006b. tpsPLS, version 1.18. Department of Ecology and Evolution, State Univ. of New York at Stony

Brook.

Rohlf FJ. 2007a. tpsRELW, version 1.45. Department of Ecology and Evolution, State Univ. of New York at

Stony Brook.

Rohlf FJ. 2007b. tpsREGR, version 1.33. Department of Ecology and Evolution, State Univ. of New York at

Stony Brook.

Rohlf FJ.& Bookstein FL. 2003. Computing the uniform component of shape variation. Systematic Biology, 53:

66-69.

Rohlf FJ. & Corti M. 2000. Use of two-block partial least squares to study covariation in shape. Systematic

Biology, 49: 740-753.

Rohlf FJ, Loy A & Corti M. 1996. Morphometric analysis of Old World Talpidae (Mammalia, Insectivora) using

partial-warp scores. Syst. Biol. 45: 344-362.

Rohlf F.J. & Marcus, L.F. 1993. A revolution in morphometrics. Trends in Ecology and Evolution,. 8: 129-132.

Rohlf FJ & Slice DE. 1990. Extensions of the Procrustes method for the optimal superimposition of landmarks.

Systematic Zoology, 39, 40-59.

Rosselló JA & Sáez L. 2001. Index balearicum: an annotated check-list of the vascular plants described from the

Balearic Islands. Collectanea Botanica (Barcelona) 25: 1-192.

Seiler MB & Keeley ER. 2009. Intraspecific taxonomy and ecology characterize morphological divergence

among cutthroat trout (Oncorhynchus clarkii ssp. Richardson) populations. Biological Journal of the

Linnean Society, 96: 266-281.

Shipunov AB & Bateman RM. 2005. Geometric morphometrics as a tool for understanding Dactylorhiza

(Orchidaceae) diversity in European Russia. Biological Journal of the Linnean Society, 85: 1-12.

Shipunov AB, Fay MF & Chase MW. 2005. Evolution of Dactylorhiza baltica (Orchidaceae) in European Russia:

evidence from molecular markers and morphology. Botanical Journal of the Linnean Society, 147: 257-

274.

Slice D. 2001. Landmarks aligned by Procrustes analysis do not lie in Kendall's shape space. Syst. Biol. 50(1):

141-149.

185 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium References

Sokal RR & Rohlf FJ. 1981. Biometry. W.H. Freeman and Company: San Francisco.

Stang M, Klinkhamer PGL & van der Meijden E. 2006. Size constraints and flower abundance determine the

number of interactions in a plant-flower visitor web. Oikos, 112: 111-121.

Stone JR. 1998. Landmark-based thin-plate spline relative warp analysis of gastropod shells. Systematic

Biology,. 47: 254-263.

Thornthwaite CW. 1948. An approach towards a rational classification of climate. Geographical Review, 38: 55-

89.

Viscosi V, Fortini P, Slice DE, Loy A & Blasi C. 2009a. Geometric morphometric analyses of leaf variation in four

oak species of the subgenus Quercus (Fagaceae). Plant Biosystems, 143: 575-587.

Viscosi V, Lepais O, Gerber S & Fortini P. 2009b. Leaf morphological analyses in four European oak species

(Quercus) and their hybrids: A comparison of traditional and geometric morphometric methods. Plant

Biosystems, 143: 564-574.

Volkova PA, Choob VV & Shipunov AB. 2007. The flower organ transition in water lily (Nimphaea alba s.l.,

Nymphaceae) under cross-examination with different morphological approaches. Belgian Journal of

Botany, 140: 60-72.

Walker J. 1996. Principal components of shape variation within an endemic radiation of threespine stickleback.

In: Marcus .F, Corti M, Loy A, Naylor G & Slice DE (eds). Advances in morphometrics. Plenum: New

York.

Zelditch ML, Swiderski DL, Sheets HD, Fink WL. 2004. Geometric Morphometrics for Biologists: A Primer.

Elsevier Academic Press: San Diego.

186 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Appendix 5.1

APPENDIX 5.1

Results of the nested ANOVA for species, populations and individuals vs. the main geometric morphometric and lineal variables. (** p<0.0001; * 0.0001 ≤ p ≤ 0.01). Results using “TAR” and “TAR2” variables instead of species are also shown (see text).

Nested model RW1 RW2 CS L W L/W

% Variance (SP) 89.6 0.5 0.0 2.9 75.2 87.2 df 1, 50.9 1, 52.4 1, 50.7 1, 50.5 1, 49.4 1, 51.5 F 373.1** 1.0 0.02 2.3 108.9** 289.5** % Variance (POPUL(SP)) 2.7 21.3 30.1 30.1 9.6 3.9 df 45, 370 45, 370 45, 370 45, 370 45, 370 45, 370 F 5398.0** 4.3** 5.6** 5.8** 7.1** 4.9** % Variance (INDIV(POPUL(SP))) 5.1 48.7 56.1 53.5 12.0 7.4 df 370, 1668 370, 1668 370, 1668 370, 1668 370, 1668 370, 1668 Nested ANOVA by species F 10964.0** 9.3** 21.4** 20.8** 19.8** 23.8** % Variance (Error) 2.6 29.6 13.8 13.5 3.2 1.6 Var(TRA) 0.0 0.0 10.5 12.6 0.0 0.0 df 1, 45.3 1, 49.0 1, 48.6 1, 48.4 1, 45.7 1, 45.4 F 0.2 0.001 7.2* 8.3* 0.3 0.1 % Var(POPUL(TRA)) 80.8 21.6 24.2 24.2 69.5 78.5 df 45, 370 45, 370 45, 370 45, 370 45, 370 45, 370 F 57.4** 4.5** 5.1** 5.3** 26.8** 45.7** % Var(INDIV(POBL(TRA))) 12.8 48.8 52.4 50.5 24.1 17.7 df 370, 1668 370, 1668 370, 1668 370, 1668 370, 1668 370, 1668 Nested ANOVA by Nested ANOVA “TRA” F 11.0** 9.3** 21. 4** 20. 8** 19. 8** 23.8** % Var(Error) 6.4 29.6 12.9 12.8 6.4 3.9 % Var(TRA2) 0.1 0.0 14.2 16.8 0.0 0.0 df 1, 45.3 1, 48.2 1, 48.2 1, 48.1 1, 45.5 1, 45.3 F 0.9 0.1 11.5* 13.5* 0.02 0.6 % Var(POBL(TRA2)) 80.7 21.6 21.2 20.8 69.6 78.5 df 45, 370 45, 370 45, 370 45, 370 45, 370 45, 370 F 55.9** 4.5** 4.4** 4.5** 26.9** 44.9** % Var(INDIV(POBL(TRA2))) 12.8 48.8 51.9 49.9 24.1 17.7 df 370, 1668 370, 1668 370, 1668 370, 1668 370, 1668 370, 1668 Nested ANOVA by Nested ANOVA “TRA2” F 11.0** 9.3** 21.4** 20.8** 19.8** 23.8** % Var(Error) 6.4 29.7 12.7 12.6 6.4 3.9

187 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Appendix 5.2

APPENDIX 5.2 Results of the 2B-PLS analyses. The covariation between all partial warps (including uniform component) and the 11 environmental variables was performed for each all leaves of each separated species. Coefficients of standardized variables are shown for the first three dimensions obtained. Indeed, covariance, squared covariance, and singular values (percent of the total squared covariance, single and cumulated) obtained in the cross set analyses, and correlations between variable and shape vectors, are shown. Significance values (p) were calculated from 999 random permutation tests.

H. crassifolium H. pendulum Variable D1 D2 D3 D1 D2 D3 Alt -0.36391 -0.08257 0.11107 0.54459 -0.09647 0.06531 Exp -0.09690 0.28878 0.15566 0.00559 0.05406 0.95268 SI -0.44187 -0.12828 0.81330 0.44025 -0.16514 -0.18873 P -0.36914 0.64639 -0.23705 0.21425 -0.09291 0.21085 T 0.33479 0.11030 0.17134 -0.16352 0.30598 0.00027 Tm 0.36919 -0.08821 0.17776 -0.31091 0.22328 -0.03894 TM 0.18562 0.37782 0.18063 0.29627 0.48369 0.01612 Tma 0.33731 -0.03349 0.25051 -0.29218 0.23097 0.00428 EL 0.10551 0.37244 0.19041 0.24160 0.46128 -0.05059 ETPL 0.08814 0.39591 0.17243 0.27796 0.46795 -0.04733 ETPT 0.33337 0.12739 0.15319 -0.17745 0.29888 -0.03839 Covariance 0.0473 0.0098 0.0059 0.0324 0.0239 0.0055 Squared covariance 0.002237 0.000095 0.000035 0.001049 0.000569 0.000030 Singular value (%) 93.8 4.0 1.5 63.1 34.3 1.8 p 0.005 0.992 0.981 0.892 0.037 0.759 Cum. singular value (%) 93.8 97.8 99.3 63.1 97.4 99.2 p 0.005 0.008 0.008 0.892 0.216 0.263 Correlations 0.4082 0.1760 0.1514 0.3602 0.3652 0.1915 p 0.001 0.001 0.001 0.001 0.001 0.004

188 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Appendix 5.3

APPENDIX 5.3

Ordination plots for the projections of each species environmental variables onto the PLS vectors. The first (X axis) and second (Y axis) dimensions are represented. Shape estimates (through thin plate splines) are shown for each dimension axis extreme. Profiles for all the abiotic variables are also estimated for each depicted shape, representing the standardized scores for the different abiotic variables (horizontal line is the mean value). A) Including both species. B) H. crassifolium populations. C) H. pendulum populations.

A

B C

189 Chapter 5. Leaf shape variation in the endemic Helichrysum crassifolium Appendix 5.3

190 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium

Chapter 6 Widespread hybridization in the Balearic Islands involving the endemic Helichrysum crassifolium (L.) G. Don

page 6.1. Introduction 193 6.2. Objectives 194 6.3. Materials and methods 195 6.3.1. Plant material 195 6.3.2. DNA extraction 197 6.3.3. Nuclear ribosomal ETS sequences 197 6.3.4. Phylogenetic analyses 198 6.4. Results 200 6.4.1. ETS ribotypes 200 6.4.2. ETS phylogenetic analyses 206 6.5. Discussion 208 6.5.1. ETS sequence variation in Helichrysum sect. Stoechadina 209 6.5.2. Testing for divergence vs. hybridization in the Balearic Helichrysum 210 6.5.3. Widespread natural hybridization as a cause of the observed variation in 215

the ETS sequences of the Balearic Helichrysum 6.6. References 218

191 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium

192 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Introduction

6.1. INTRODUCTION

Following morphological results in Chapter 5, in this chapter nuclear ribosomal and cpDNA markers are used to assess the above hybridization hypothesis involving the endemic Helichrysum crassifolium (L) D. Don. Previous studies of based on the internal spacers (ITS) of the 45S ribosomal cistron in species of section Stoechadina revealed that this region could not resolve the Mediterranean species clade (Galbany-Casals et al.,

2004). As an explanation, they argued recent massive speciation events or hybridization processes resulting in low support in this clade. Therefore, in this study, the external transcribed spacer (ETS) of the 45S cistron was used to assess the extent of molecular differentiation between H. crassifolium and H. pendulum (C. Presl) C. Presl.

The ETS region resulted very useful when ITS showed weak or no phylogenetic signal (e.g., Linder et al., 2000; Kelch & Baldwin, 2003; Sanz et al., 2008; Poczai &

Hyvönen, 2010, and references therein). It also proved to be resolutive enough in phylogenetic trees in Helichrysum (Galbany-Casals et al., 2009) and closely related genera of the tribe (e.g., Bayer et al., 2002; Ford et al., 2007; Bergh & Linder, 2009). Moreover, it was useful also in cases involving hybridization processes (e.g., Timme et al., 2007;

Boonruangrod et al., 2009).

193 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Objectives

6.2. OBJECTIVES

Based on the results shown in the Chapter 5, the aims of this study were (i) to assess the hypothesis of interspecific hybridization between H. crassifolium and H. pendulum from nuclear ribosomal markers (ETS), (ii) to test for the correlation between morphological results concerning to leaf variation in both rupicolous species (Chapter 5) and the degree of genetic variation/similarity in the ETS region among both species, and (iii) to test for other Helichrysum taxa occurring in the Balearic Islands being involved in the putative hybridization processes affecting the rupicolous species.

194 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Materials and methods

6.3. MATERIALS AND METHODS

6.3.1. Plant material

The study included individuals from the 37 H. crassifolium and ten H. pendulum populations from the Balearic Islands listed in Table 5.2. One to 11 accessions per population were used for ETS analyses. In addition, samples of the remaining Helichrysum taxa inhabiting the Balearic Islands were also included (Table 6.1), and 11 Genbank accessions from eight Mediterranean Helichrysum taxa (Table 6.3), were also analyzed.

As outgroups, 17 herbarium accessions corresponding to 13 Helichrysum taxa from

Europe (Table 6.2), and ETS Genbank accessions of 7 different Gnaphalieae taxa from

Australia and New Zealand, were included in the phylogenetic analysis (Table 6.3).

Table 6.1. Populations of non-rupicolous Helichrysum taxa from the Balearic Islands used in this study.

Taxon Population n Habitat Island H. italicum subsp. microphyllum Massanella 2 Garrigue Majorca H. pendulum Es Vedrà 1 Coast (islet) Ibiza H. serotinum S’Almàguena 2 Garrigue Ibiza H. stoechas Caló d’es Forn 1 Coast Cabrera H. stoechas Illa d’es Bosc 1 Coast (islet) Ibiza H. stoechas Aeroport 1 Garrigue Ibiza H. stoechas Cala Torta 11 Coastal garrigue Majorca H. stoechas Binimel·là 1 Coastal garrigue Minorca

195 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Materials and methods

Table 6.2. Herbarium accessions included in this study, from the herbarium of the Jardí Botànic of the University of València. *Determined in this study. Following the revision of Galbany-Casals et al. (2006a,c), the taxonomic correspondence should be: 2H. serotinum subsp. serotinum. 3H. angustifolium. 4H. serotinum subsp. picardii. 1H. angustifolium x H. italicum. 5H. pendulum. 6H. stoechas.

Taxon indicated Origin Accession H. arenarium subsp. arenarium GERMANY: Babelsberg VAL33899 H. arenarium subsp. aucheri * TURKEY VAL146906 H. chionophilum * TURKEY VAL146302 H. compactum * TURKEY VAL146296 H. italicum subsp. italicum ITALY: Sicily. Mont Etna VAL143081 H. italicum subsp. microphyllum ITALY: Sardinia. Carloforte VAL35201 H. italicum subsp. pseudolitoreum 1 ITALY: Sardinia. Buggerru VAL141440 H. italicum subsp. serotinum 2 SPAIN: Soria. San Estebán de Gormaz VAL980334 H. litoreum 3 ITALY: Foggia. Peschici VAL40803 H. orientale * TURKEY VAL146401 H. picardii 4 SPAIN: Huelva. Playa de Mazagón VAL147426 H. rupestre 5 ITALY: Sicily. Monte Cófane VAL119865 H. rupestre 5 SPAIN: Alacant. Marina Alta VAL982112 H. stoechas SPAIN: Serra d’Espaneguera, Castelló VAB945634 H. stoechas SPAIN: Palencia. Peña Lampa VAL035164 H. stoechas PORTUGAL: Bragança VAL96-1244 H. stoechas subsp. stoechas var. syncladum 6 FRANCE: Noirmoutier. La Guérinière VAL33896

Table 6.3. Accessions from the Genbank used for ETS analyses.

GenBank Location Taxon Reference accession Helichrysum crassifolium FJ211542 Galbany et al., 2009 Mediterranean ; Majorca (Formentor) Helichrysum crassifolium FJ211541 Galbany et al., 2009 Mediterranean ; Majorca (Formentor) Helichrysum crassifolium FJ211540 Galbany et al., 2009 Mediterranean ; Majorca (Formentor) Helichrysum italicum subsp. italicum FJ211548 Galbany et al., 2009 Mediterranean Helichrysum italicum subsp. microphyllum FJ211546 Galbany et al., 2009 Mediterranean ; Majorca (Massanella) Helichrysum italicum subsp. siculum FJ211547 Galbany et al., 2009 Mediterranean ; Sicily Helichrysum pendulum FJ211539 Galbany et al., 2009 Mediterranean Helichrysum serotinum subsp. picardii FJ211550 Galbany et al., 2009 Mediterranean Helichrysum serotinum subsp. serotinum FJ211549 Galbany et al., 2009 Mediterranean Helichrysum stoechas FJ211543 Galbany et al., 2009 Mediterranean Helichrysum stoechas FJ211544 Galbany et al., 2009 Mediterranean Helichrysum lanceolatum EF187647 Ford et al., 2007 New Zealand Helichrysum leucopsideum AF319712 Bayer et al., 2002 W Australia Ixiolaena tomentosa AF319717 Bayer et al., 2002 New South Wales, Australia Ixodia achillaeoides AF319718 Bayer et al., 2002 S Australia Parantennaria uniceps AF319738 Bayer et al., 2002 New South Wales, Australia Stuartina mulleri AF319760 Bayer et al., 2002 S Australia Xerochrysum bracteatum AF319682 Bayer et al., 2002 W Australia

196 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Materials and methods

6.3.2. DNA extraction

Total DNA was extracted from dried leaves using the CTAB protocol (Doyle & Doyle,

1987), scaled down to perform the process in 1.5 ml microfuge tubes.

6.3.3. Nuclear ribosomal ETS sequences

The whole IGS region of the 45S ribosomal cistron was amplified using universal primers located in the 26S and 18S coding regions (Bena et al. 1998; Figure 6.1). The PCR product (ca. 3.1 kb) was sequenced with the reverse primer and a specific primer was designed into the IGS to partially amplify 5’ETS (5’-TGC ATG AGT GGT GTT TGG-3’) which, used together with the universal 18S primer (Bena et al. 1998), resulted in a ca. 445 bp product (Figure 6.1). However, important problems to correctly align sequences were found in the 3’ extreme of such product for many accessions. Therefore, a c. 120 bp of the product close to the 18S region were discarded in the alignment and thus, the sequence used in the analyses was a c. 330 bp fragment from the end of the ETS primer.

PCR reactions were performed in 20 μl, containing 75 mM Tris-HCl (pH 9.0), 5 mM

KCl, 20 mM (NH4)2SO4, 0.0001% BSA, 2 mM MgCl2, 200 μM of each dATP, dCTP, dGTP and dTTP, 0.15 μM of each primer, approximately 50-100 ng of genomic DNA and 0.3 units of Netzime® (NEED S.L., Valencia, Spain) DNA polymerase from Thermus thermophilus.

Amplification comprised an initial denaturalization step of 3 min at 94ºC, and 40 cycles of 15 sec at 94ºC, 15 sec at 57ºC and 1 min at 72ºC, and a final extension step of 5 min at 72ºC, in a Thermo Px2 thermal cycler (Thermo Electron Corporation, Milford, MA, USA). The PCR products were separated on 1.0% agarose gels and purified using the High Pure PCR

Product Purification Kit (Roche Diagnostics).

Electrophoretogram ambiguities (double peaks) were present in most samples.

Therefore PCR products from three H. crassifolium accessions (two from Minorca and one

197 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Materials and methods from Majorca) were gel-purified and ligated into the vector provided with the p-GEM-T

(Promega) cloning kit. Plasmid DNA from individual recombinant colonies was isolated according to a miniprep protocol (High Pure Plasmid Isolation kit, Roche Diagnostics).

Figure 6.1. Schematic presentation of the nuclear 18S-26S ribosomal DNA repeat of the nucleolar organizer region (NOR) in plants. Genes (blue boxes) are separated by the internal transcribed spacers (ITS). These units, tandemly repeated, are separated by the intergenic spacer (IGS), consisting in two flanking external transcribed spacers (ETS) located after the 26S gene (3’ETS) and before the 18S gene (5’ETS), and the non-transcribed spacer (NTS), in the middle. A) Schematic of the whole repeat unit. B) Detail of the IGS spacer, indicating the position of the primers (in red) used in this study and the length of the PCR products obtained.

6.3.4. Phylogenetic analyses

Phylogenetic reconstructions were performed using the ETS sequences and selected accessions from Genbank as outgroups. ETS sequences were aligned with Clustal

X 1.81 (Larkin et al., 2007), after manual adjustments of sequences. About 120 bp of the matrix showed alignment ambiguities and were discarded. Parsimony analyses were performed using PAUP*4.0b10 (Swofford, 2002). Phylogenetic analyses using heuristic searches were conducted using Fitch parsimony with equal weighting of all characters, polymorphic states were considered as polymorphisms, and indels were treated as missing

198 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Materials and methods data. Support for monophyletic groups was assessed by fast bootstrapping (1000 resamplings) using the heuristic search strategy. In addition, phylogenetic reconstructions were also performed by distance-based methods. The Kimura 2-parameter distance model

(Kimura, 1980) and the Neighbor-Joining method (Saitou & Nei, 1987) were used to construct a bootstrapped tree (1000 resamplings).

Unrooted statistical parsimony networks were constructed from those ETS sequences showing no intragenomic polymorphisms using TCS v. 1.21 software (Clement et al., 2000). A substitution of the motive “CTATT” by “A” in several outgroup accessions was coded as a single mutation.

199 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Results

6.4. RESULTS

6.4.1. ETS ribotypes

Among the whole Balearic and European Helichrysum accessions used in this study,

47 polymorphic sites were detected in the ca. 330 bp used from the ETS region. However,

15 resulted polymorphic due to the existence of variation in a single sequence. Since this low frequency, and also the possibility of PCR or sequencing artifacts (i.e., chimerical sequences and Taq errors; Cline et al. 1996; González et al., 2005; Lenz & Becker, 2008), only the remaining 32 sites were considered in this study. From those, 25 were polymorphic

(Table 6.4) in the four taxa occurring in the eastern Balearic Islands (i.e., distribution area of the endemic H. crassifolium). Twenty-two of them showed base additivities in at least some accessions, whereas no base additivity was detected at the positions 95, 96, and 297 of the aligned matrix (Table 6.4).

Helichrysum crassifolium showed a high number of ETS ribotypes as a consequence of the presence of 23 variable sites detected in this species (positions 157 and 297 were fixed), 18 showing base additivities. Indeed, in 10 polymorphic sites the less frequent genotype was not detected in any other Helichrysum taxon, whereas in four of them only base additivity was detected in this sampling (Table 6.4). In the case of H. pendulum, 10 variable sites were detected (thus 15 fixed positions), all showing base additivities but in position 297. For H. stoechas, eight variable sites were detected (thus 17 fixed positions), all showing base additivity and thus. In the case of H. italicum subsp. microphyllum only two variable sites were detected (23 fixed sites), both showing base additivity (Table 6.4).

The most frequent genotype coincided between H. crassifolium and H. pendulum in all the polymorphic sites but in position 247, where H. crassifolium, contrary to H. pendulum,

200 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Results showed base additivity as the most frequent genotype (61.7%). Moreover, the only important differences in frequency (i.e., H. crassifolium always showing lower frequency for the most frequent genotype) were found in positions 135, 234, 255 and 303 (Table 6.4).

Also, in the positions 247 and 255 the most frequent genotype was a base additivity in H. crassifolium, H. stoechas and H. italicum subsp. microphyllum. All H. pendulum accessions from Ibiza showed additivity in positions 247, 248, 255 and 258 (Table 6.4). However, none of the polymorphic positions was fixed in the H. pendulum accessions from the eastern

Balearic Islands.

On the other hand, H. italicum subsp. microphyllum showed fixed genotype for a base other than the most frequent in the Balearic rupicolous taxa in positions 124 and 135.

The same occurred in position 135 for H. serotinum (Table 6.4). In addition, the European

Helichrysum accessions and the outgroups used by Galbany-Casals et al. (2004) also showed low frequencies for the latter position with respect to the rupicolous Balearic species.

Twenty ribotypes showing no evidence of intragenomic polymorphisms were detected in this study, from the variation in 23 positions (Table 6.5). From those, 11 ribotypes (R1 to R11) were present only in H. crassifolium. R2, R3, R4 and R9 were found only in Minorcan samples, R1, R6, R7, R8 and R11 characterized Majorcan accessions

(including the Dragonera islet), whereas R5 and R10 were found in samples from both islands. On the other hand, R13 and R14 were detected exclusively in H. pendulum samples from Majorca, although R14 was also detected in a Genbank accession ascribed to H. stoechas (FJ211543). Furthermore, R12 was found in H. crassifolium from Majorca and in H. pendulum from Cabrera island. The latter genotype occurred in both species accessions from the Genbank (H. crassifolium, FJ211541, FJ211542; H. pendulum,

FJ211539).

201 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Results

Table 6.4. Polymorphic sites found in the ETS region of the Helichrysum taxa from the eastern Balearic Islands. Percents for each of the detected genotypes are shown. Clones are not included. CRA = H. crassifolium, PEN = H. pendulum, STO = H. stoechas, MIC = H. italicum subsp. microphyllum.

Position Base CRA PEN STO MIC Position Base CRA PEN STO MIC 81 T 97.5 100 100 100 247 A 38.5 82.4 46.7 0 ------W 2.5 0 0 0 M 61.7 17.6 53.3 100 95 C 93.8 100 100 100 248 T 81.5 82.4 73.3 50.0 T 6.2 0 0 0 ------W 18.5 17.6 26.7 50.0 96 G 93.8 100 100 100 251 A 84.0 100 100 100 T 6.2 0 0 0 ------W 16.0 0 0 0 97 G 93.8 100 86.7 100 255 T 45.7 76.5 40.0 50.0 A 6.2 0 0 0 - - - - - R 0 0 13.3 0 K 54.3 23.5 60.0 50.0 115 A 93.8 100 100 100 258 C 80.2 82.4 80.0 100 T 3.7 0 0 0 - - - - - W 2.5 0 0 0 Y 19.8 17.6 20.0 0 124 T 93.8 100 100 0 270 G 76.5 100 100 100 C 1.3 0 0 100 A 4.9 0 0 0 Y 4.9 0 0 0 R 18.5 0 0 0 135 T 64.2 94.1 66.7 0 272 T 97.5 100 100 100 C 11.1 0 6.7 100 - - - - - Y 24.7 5.9 26.7 0 W 2.5 0 0 0 152 T 100 88.2 100 100 280 C 76.5 100 100 100 C 0 0 0 0 T 4.9 0 0 0 Y 0 11.8 0 0 Y 18.5 0 0 0 161 T 87.7 100 100 100 294 G 97.5 100 100 100 A 4.9 0 0 0 T 2.5 0 0 0 W 7.4 0 0 0 R 0 0 0 0 176 G 91.4 94.1 100 100 297 T 100 88.2 100 100 A 2.5 0 0 0 C 0 11.8 0 0 R 6.2 5.9 0 0 - - - - - 209 T 96.3 100 100 100 303 T 32.1 76.5 66.7 0 C 1.2 0 0 0 C 46.9 17.6 6.7 100 Y 2.5 0 0 0 Y 21.0 5.9 26.7 0 222 T 91.4 100 100 100 317 C 95.1 58.8 66.7 100 - - - - - T 4.9 29.4 6.7 0 K 8.6 0 0 0 Y 0 11.8 26.7 0 234 G 39.5 100 100 100 T 29.6 0 0 0 K 30.9 0 0 0

202 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Results

The remaining ribotypes were detected in, or also in, non-rupicolous Helichrysum taxa. R15 was found in the Balearic Islands in H. crassifolium (Minorca) and H. italicum subsp. microphyllum (Majorca), H. stoechas (Ibiza), but also in H. stoechas and H. serotinum subsp. picardii from the Iberian peninsula and Portugal (also Genbank accession

FJ211550 of the latter taxon). R16 was mostly found in samples from Italy, thus H. italicum subsp. italicum (VAL143081, Sicily; FJ211548), H. italicum subsp. microphyllum

(VAL35201, Sardinia), H. litoreum (cf. H. angustifolium, VAL40803, Foggia), H. italicum subsp. pseudolitoreum (cf. H. angustifolium x H. italicum, VAL141440, Sardinia), H. italicum subsp. siculum (FJ211547, Sicily), and H. pendulum (VAL119865, Sicily). However, it was also detected in the H. italicum subsp. microphyllum accession of Galbany-Casals et al.

(2009) collected from Majorca (Massanella, FJ211546) but, interestingly, not in this study accession from the same taxon and locality (showing R15). R17 was only detected in a

Genbank accession ascribed to H. stoechas (FJ211544); whereas the last three ribotypes were detected in samples from Turkey, thus R18 in H. compactum (VAL146296) and H. chionophyllum (VAL146302), R19 in H. orientale (VAL146401), and R20 in H. arenarium subsp. aucheri (VAL146906), (Table 6.5).

The parsimony analysis of the 20 ribotypes showing no evidence of intragenomic polymorphisms resulted in the network depicted in Figure 6.2A. The ribotype R10 resulted in the central position and was involved in a double loop together with R7, R6, R12 and

R15. Up to eight branches arised from this double loop. Six of those differed from the loop ribotypes in a single mutation, though two giving rise to further ribotypes. One of those branches derived in a third loop through three mutations, attached to the double loop and involving R12, R14 and R3. The two remaining are long branches involving R1, R4, R5, and

R18, R19, R20, respectively. Thus, the former branch involved only H. crassifolium ribotypes separated from one to three mutations; whereas the latter included ribotypes from

203 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Results

Turkish accessions, separated from the loop in 6 to 9 mutations, and resulting R18 the most divergent one (Figure 6.2A).

When analyzing only the ribotypes from H. crassifolium and H. pendulum from the

Balearic Islands (15 ribotypes from 14 polymorphic sites; Table 6.5), a very similar network was retrieved (Figure 6.2B). The central ribotype resulted R12, though involved in a loop with R10, R7 and R6, thus the same obtained in the network considering all ribotypes.

Figure 6.2. Parsimony networks of the ETS ribotypes showing no evidence of intragenomic polymorphisms (corresponding to Table 6.5). Ribotypes from Balearic accessions are coloured (H. crassifolium, red; H. pendulum, blue; H. stoechas, green; H. italicum subsp. microphyllum, yellow). White colour represents ribotypes detected in non-Balearic accessions. A) Including all ribotypes found in this study. R10 appeared as the central ribotype. B) Including only the ribotypes found in H. crassifolium and H. pendulum in the Balearic Islands. R12 appeared as the central ribotype.

204 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Results

205 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Results

6.4.2. ETS phylogenetic analysis

The 50% majority-rule consensus tree of 1820 most parsimonious trees of 576 steps is shown in Figure 6.3. ETS sequences showed a monophyletic-Helichrysum clade and located H. lanceolatum as the most basal lineage. Three major clades with a high bootstrap support were obtained, but with up to 12 Balearic accessions of H. crassifolium, H. pendulum and H. stoechas located outside of such clades. One of the clades (98% bootstrap) included only H. pendulum accessions from Majorca and two H. stoechas accessions from Minorca and from eastern Spain (VAL945634, Castelló).

A second clade (97% bootstrap) included only H. crassifolium accessions from

Majorca (including Dragonera islet). The last one (87% bootstrap) included most of the accessions. Twenty-nine H. crassifolium accessions from Majorca and 13 from Minorca, four H. pendulum accessions from Majorca, and an H. stoechas accession from Minorca, were included with no further internal clades obtained.

However, two sub-clades were retrieved in this clade. One (99% bootstrap) joined two H. crassifolium accessions from the northern Majorca (SFO, MFO), both showing the T in position 115 defining R11, also detected in a H. crassifolium Genbank accession

(FJ211540). The other sub-clade, with lower bootstrap (57%) included up to ten H. crassifolium accessions from Majorca and Minorca, an H. stoechas accession from Ibiza, an

H. italicum subsp. microphyllum from Majorca, and all the herbarium accessions but the one stressed in the first clade.

Therefore, clustering neither for all the H. crassifolium nor for the H. pendulum accessions was detected. Indeed, the Balearic accessions were located in a single clade neither for H. pendulum nor for H. stoechas.

206 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Results

Figure 6.3. The 50% majority-rule consensus tree obtained from 1820 most parsimonious trees of 576 steps. Bootstrap values are indicated. Helichrysum crassifolium (CRA) and H. pendulum (PEN) accessions are in black, other species accessions from the Balearic Islands in blue, other Mediterranean accessions in red, and outgroups in green. In the right winger, scheme of the position of the H. crassifolium, H. pendulum and H. stoechas (STO) accessions in the tree. (ARE=H. arenarium subsp. arenarium, AUC=H. arenarium subsp. aucheri, CHI=H. chionophyllum, COM=H. compactum, ITA=H. italicum subsp. italicum, LIT=H. litoreum (cf. H. angustifolium), MIC=H. italicum subsp. microphyllum, ORI=H. orientale, PIC=H. serotinum subsp. picardii, PSE=H. italicum subsp. pseudolitoreum (cf. H. angustifolium x H. italicum), SIC=H. italicum subsp. siculum).

207 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Discussion

6.5. DISCUSSION

The ribosomal 45S cistron is the most frequently used nuclear target in phylogenetic reconstructions and the different coding (18S, 5.8S, 25S) and non-coding regions (ITS,

ETS, NTS) have been used at different systematic levels, spanning the practically whole gymnosperm and angiosperm diversity (cf. Baldwin et al., 1995). This cistron forms part of a multigene family with thousands of tandem copies clustering in arrays, and usually several ribosomal loci are present within plant genomes. However, concerted evolution tends to homogenize such ribotypes within individuals. Thus, only a single consensus ITS or ETS sequence is usually retrieved after sequencing PCR products from single individuals

Nevertheless, some authors have cautioned about the use of rDNA nucleotide sequences in phylogenetic reconstructions after the finding that major ribosomal loci may move within and among chromosomes and that their movement may potentially occur via magnification of minor loci consisting of a few rDNA copies (Dubcovsky & Dvorak 1995;

Okuyama et al., 2005; Anne, 2006). However, this violation of the assumed orthology is usually neglected in standard phylogenetic practices.

In addition, if only direct, not cloned, sequences are routinely generated from an organism it is highly likely that much of the intragenomic ribosomal diversity present within organisms could be easily overlooked. This is due to the fact that molecular homogenizing mechanisms of the ribosomal multigene family could be relaxed, and from two to several ribotypes can be present within organisms if (i) the rate of mutation among copies is faster than that driving the concerted evolution of the array, (ii) duplicate ribosomal loci evolve without selective constraints and accumulate mutations, and (iii) homeologous loci from

208 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Discussion other species are incorporated into a single genome through hybridization-mediated processes.

6.5.1. ETS sequence variation in Helichrysum sect. Stoechadina

Previous studies using ITS sequences detected very low values of sequence divergence among Mediterranean Helichrysum species, including H. crassifolium (Mus et al., 2002; Galbany-Casals et al., 2004). Certainly, this null divergence could only reflect a very low rate of nucleotide substitution in the ribosomal multigene family in Helichrysum that could not adequately mirror the evolutionary history of the genus (e.g., in Monarda, Prather et al., 2002).

Also, rapid diversification events in singular scenarios, like plant speciation in oceanic islands, are not always followed by a concomitant ribosomal divergence among closely related taxa. In fact, from their phylogeny of the genus Helichrysum based on ITS sequences, Galbany-Casals et al. (2004) concluded that the genus is polyphyletic, as previously suggested by other authors (Anderberg, 1991; Bayer et al., 2000). However, this marker could not resolve the Mediterranean clade. As an explanation, they argued (i) recent massive speciation events, since Helichrysum is the genus with the larger number of species in the tribe, results the most derived genus in other studies (Bayer et al., 2000), and the Mediterranean species in turn are the most derived group within the genus; and (ii) hybridization events resulting in low support of the internal nodes of the ITS tree. Although, homogenization through concerted evolution after hybridization events could also be invoked (e.g., Vander Stappen et al., 2003).

In this study, direct sequencing and cloning procedures revealed that, differing from

ITS sequences (Mus et al., 2002; Galbany et al., 2004), significant intragenomic variation in ribosomal ETS sequences occurs in H. crassifolium and H. pendulum. The levels and

209 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Discussion extent of intraindividual sequence polymorphism detected in this study strongly suggest that the intragenomic ETS variability found should not only be explained by PCR or sequencing artifacts. On the one hand, in their recent work with ETS in Helichrysum, Galbany-Casals et al. (2009) also found heterogeneous sequences in H. crassifolium after cloning.

Nevertheless, the levels of genotypic diversity detected in H. crassifolium in this study, either cloning and by direct sequencing, are much higher than those shown by Galbany-

Casals et al. (2009). Indeed, they only detected three different ribotypes from a single accession; whereas in this study seven ribotypes in three cloned accessions, and 23 consistent variable sites were detected through direct sequencing. Moreover, they did mention the occurrence of polymorphic sequences neither in H. pendulum nor in any other species occurring in the Balearic Islands; whereas those were detected in this study. On the other hand, sequence heterogeneity is a situation known in other species, and due to a lack of concerted evolution (Okuyama et al., 2005).

This scenario can be attributed to opposite processes (e.g., Andreasen & Baldwin,

2003): to divergence-related factors such incomplete lineage sorting and higher mutation rate than concerted evolution rate, or otherwise to interspecific hybridization.

6.5.2. Testing for divergence vs. hybridization in the Balearic Helichrysum

Results supporting a putative population divergence scenario were scarce. However, position 135 resulted highly informative to separate the Balearic rupicolous from the remaining taxa used in this study; whereas position 124 was also fixed in H. italicum subsp. microphyllum with respect to the rupicolous taxa. These differences could respond to arising divergence of the rupicolous Balearic taxa, or otherwise to the existence of genotypes yet not homogenized among hybridizing taxa. Oppositely, in 10 of the polymorphic positions detected, the less frequent genotype found in H. crassifolium was not detected in any other

210 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Discussion accession in this study. Indeed, in five of those only the base-additivity genotype was detected in all accessions. This absence of fixed genotypes for these positions in any of the

Helichrysum accessions sequenced would better pray for hybridization rather than divergence, thus reinforcing the hypothesis of current and recurrent hybridization events in the Balearic Islands involving all Helichrysum taxa, and especially involving the rupicolous species.

In five of the polymorphic sites, H. crassifolium and H. pendulum from Ibiza, but not from the eastern Balearic Islands, showed base-additivity. Besides this being or not a reason favouring the discrimination of the Ibizan plants (some described as var. latifolium by

Font Quer, 1920) from the Majorcan plants, it could merely respond to the absence of H. crassifolium in Ibiza, added up to the non-rupicolous habitat, shared with H. stoechas, of the

Ibizan plants. This would be consistent with the morphological results in Chapter 5, pointing to differences in leaf size and shape among H. crassifolium and among H. pendulum populations responding to the existence of a second (or a third) Helichrysum taxon growing in close vicinity and thus, pointing again to natural hybridization as a feasible cause for the variation observed.

In consonance with the high number of polymorphic positions showing genotypes found only in H. crassifolium accessions, more than a half of the 20 ribotypes detected in this study showing no evidence of intragenomic polymorphims (i.e., base-additivities) were only found in the endemic species. From those, 9 were exclusively detected in a single island (5 in Majorca and 4 in Minorca), whereas only two were found in both islands. This could be the result of divergences among islands. However, due to the high recombination rates known for the ETS region (e.g., Anne, 2006; Timme et al., 2007; Poczai, 2010), ancient and recurrent hybridization could be also invoked to explain the origin of such high number of ribotypes exclusively found in the endemic species. In fact, the ribotype R12 was

211 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Discussion detected in H. crassifolium and in H. pendulum, including accessions either from the sampling and from Genbank (Galbany-Casals et al., 2004). This ribotype was central and involved in loops in the parsimony networks, and also detected in non-rupicolous species.

Thus, all this would reinforce the latter hypothesis.

Intra-individual polymorphisms use to be resolved through sequence cloning. Hence,

Galbany-Casals et al. (2009) stated that the polymorphisms detected in ETS sequences of the endemic H. crassifolium can only be discussed from cloned sequences, since ambiguities could be either genuine variation or noise. However, the detailed sampling of individuals across the whole distribution range of the endemic species performed in this study, resulted in a sound evidence at least by two reasons. First, it is difficult to assume

PCR artifacts when retrieving polymorphisms (i) in the same sites in a high percent of the sampled accessions, (ii) after obtaining the sequences in independent PCR and sequencing events, and even in different labs, (iii) all the polymorphic sequences showing with no doubt and no exception the same genotype, and (iv) detecting exclusively the corresponding non- polymorphic genotypes in homogeneous accessions in those sites.

Second, attempting cloning strategies in such samplings deserve of an extremely high effort of sequencing, probably giving not much more information than the retrieved from such sampling without cloning. Certainly, it would give information about the haploid sequences actually existing. However, the high number of ribotypes detected when cloning only three individuals in this study favours expectations of a high degree of recombination for ETS sequences and thus, a very high number of recombined ribotype obtained from cloning. Indeed, Nieto-Feliner et al. (2004) showed the existence of a similar pattern of intraindividual variation in ITS sequences of Armeria, demonstrating that direct sequences can be useful to detect intraindividual polymorphisms from a wide and detailed plant sampling.

212 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Discussion

The ETS tree of this study shows conspicuous differences in the Mediterranean clade with respect to that of Galbany-Casals et al. (2009). First, with regards to its origin, a much shorter portion of the ETS region was used (c. 330 bp vs. c. 840 in Galbany-Casals et al., 2009). However, a much higher sampling was performed in this study, with 77 H. crassifolium sequences from 37 populations along the whole distribution of the species, whereas only three clone sequences from a single accession were included in Galbany-

Casals et al. (2009). This wide sampling resulted in 33 consistent variable sites detected, 23 affecting to H. crassifolium and 10 to H. pendulum. In the same 330 bp sequence, Galbany-

Casals et al. (2009) found only two positions discriminating their H. crassifolium clone accession 39 from the other two clones, resulting their H. pendulum accession with the same genotype shown by the latter two clones in the stressed positions.

Second, with respect to the position of the accessions in the clades, ambiguities regarding to the taxa of the sect. Stoechadina from the western Mediterranean was detected between both studies. Mainly, the use of a single accession to construct a phylogenetic tree, and the occurrence of divergent sequences in those taxa, are weaknesses in the Galbany-Casals et al. (2004, 2009) hypotheses on the degree of kinship among the Mediterranean taxa. In fact, from this study data, very different trees are expected to be retrieved depending on the accession selected. Therefore, as for ITS, this would stress the low or null resolution of the ETS sequences among several taxa of the sect. Stoechadina, besides of the occurrence of much higher variation in ETS rather than in

ITS.

Thus, besides of the use of a single accession, several phylogenetic incongruences could be observed. First, the occurrence of both H. serotinum subspecies in separated clades, thus showing more relation with other species occurring in the same distribution area rather than among both subspecies. Second, there was a higher relation among clone

213 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Discussion sequences of H. crassifolium and other species, rather than among all the clone sequences from the same individual. Both observations would favour hybridization rather than divergence processes.

Supporting the occurrence of non-rupicolous taxa also involved in the hybridization processes, the 99% bootstrap sub-clade of this study tree, including H. italicum taxa, also included an H. crassifolium accession from southern Majorca (PAG15) and an H. pendulum accession from Sicily (VAL119836). The PAG15 H. crassifolium accession belongs to a population closely located to the locality from which Galbany-Casals et al. (2006a) indicated hybrid specimens between H. crassifolium and H. stoechas, and also between H. crassifolium and H. pendulum (Cala Llamp; Galbany-Casals et al., 2006b). Wide H. italicum subsp. microphyllum populations can be found either in this and in PAG localities, and it is astonishing that, in their sampling in such locality, Galbany-Casals et al. (2006a) did detect neither the morphologically very intermediate plants between H. crassifolium and H. italicum subsp. microphyllum, nor those between H. stoechas and the latter two taxa (already indicated in Cosín et al., 2002; Mus et al., 2002; and Conesa et al., 2004).

Indeed, in Chapter 5, H. crassifolium plants from PAG resulted with a shorter leaf than the other populations of the endemic species without H. italicum subsp. microphyllum growing in close vicinity. Thus, even the non-evident morphologically intermediate plants from this locality showed differences putatively related to the presence of the latter taxon.

This seemed also the case of the H. pendulum herbarium accession from Sicily, with a position in the clade favouring a putative relation with H. italicum-related taxa from the region.

Although, it is notorious that R16 was found in Italian accessions of H. italicum subsp. microphyllum and in the Genbank accession of the same taxa from Majorca

(FJ211546), but not in this study accession of the same taxon and locality (Massanella,

214 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Discussion showing R15). Certainly, in Chapter 5 this site H. crassifolium plants (MAS) showed suspicions of being involved in hybridization events with H. italicum subsp. microphyllum.

Both taxa grow in close vicinity in this site, and non-rupicolous very intermediate plants have been observed during this study sampling.

6.5.3. Widespread natural hybridization as a cause of the observed variation in the

ETS sequences of the Balearic Helichrysum

The presence of paralogous ribosomal loci within the nuclear genome, the occurrence of heterogeneous ribosomal copies within arrays, or the presence of homeologous chromosomes (making individuals heterozygous) should be postulated to occur in Helichrysum. Inspection of the ETS sequences does not suggest that the recovered ETS sequences are pseudogenes. In fact, no report of ribosomal pseudogenes have been previously detected in other studies using ITS and ETS sequences in

Helichrysum (Galbany-Casals et al. 2004; 2009) although this hypothesis should be checked with a closer inspection of the structural landmarks and the highly conserved motifs in the primary DNA sequence, and the thermodynamic stability and conserved cruciform foldings of the secondary RNA structures present within the whole IGS spacer.

The presence of intragenomic ETS polymorphisms in H. crassifolium cannot also be explained by stochastic processes, like lineage sorting, that could confound the phylogenetic signal present in orthologous sequences. Since H. crassifolium and H. pendulum are largely sympatric, together with other species of the genus that usually grow in distinct non-rupicolous environments, it is highly likely that the ETS polymorphism could be generated by hybridization events with other extant species. The inspection of ETS sequences found in H. crassifolium and the lack of phylogenetic resolution found, not only give support to the view that rampant episodes of interspecific hybridization have occurred

215 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Discussion during the evolution of in this species, but that several Helichrysum species could have contributed to the ETS heterogeneity found in this species. Occurrence of polymorphic ribosomal repeats, due to a failure in full concerted evolution, has been usually invoked as an evidence of hybridization among taxa displaying different ribotypes (e.g. O’Kane et al.

1996, Campbell et al. 1997, Brasier et al. 1999, Rauscher et al. 2002). Thus, it is likely that

Helichrysum samples contained two to several divergent ETS sequences showing partial homogenization, some sites showing nucleotide additivity and others showing homogenization towards either parental sequence.

Another explanation for the pattern of intragenomic polymorphism found in ribosomal sequences would imply that the Mediterranean Helichrysum species from section

Stoechadina, which are tetraploid (reviewed in Galbany-Casals & Romo, 2008), were originated from old hybridization events followed by subsequent diversification. Results of

Galbany-Casals et al. (2009) pointed to a monophyletic origin of sect. Stoechadina, from

African ancestors reaching the Mediterranean basin and diversifying to produce the

Mediterranean-Asiatic group of the genus. Yet, they already suggested hybridization events occurring during speciation, based on the current easiness for hybridization that they observed among Mediterranean Helichrysum taxa, and on the low levels of support of several nodes of their trees, interpreted as a sign of reticulate evolution.

In absence of other data, the hybridization hypothesis is favoured as the most reasonable explanation for the pattern of morphological and molecular data found in this and the previous chapters. Independently of their origin, the persistence of ribosomal polymorphisms in H. crassifolium implies that the molecular forces driving concerted evolution of this multigene family are not fully operating in this narrowly-distributed species.

In fact, several stages of ribosomal homogenization could be observed not only in H. crassifolium, but in other widespread species. If the divergent ribosomal families are on

216 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium Discussion different loci located in separate chromosomes, this could prevent concerted evolution since it has been suggested that the chromosomal location of rDNA loci could have a more substantial impact than the number of loci on the tempo of concerted evolution through either unequal crossing-over or gene conversion (Zhang & Sang, 1999).

217 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium References

6.6. REFERENCES

Anderberg AA. 1991. Taxonomy and phylogeny of the tribe Gnaphalieae (Asteraceae). Opera Botanica, 104: 1-

195.

Andreasen K & Baldwin BG. 2003. Nuclear ribosomal DNA sequence polymorphism and hybridization in

checker mallows (Sidalcea, Malvaceae). Molecular Phylogenetics and Evolution, 29: 563–581.

Anne C. 2006. Choosing the right molecular genetic markers for studying biodiversity: from molecular evolution

to practical aspects. Genetica 127: 101-120.

Baldwin BG, Sanderson MJ, Porter JM, Wojciechowski MF, Campbell CS & Donoghue MJ. (1995). The ITS

region of nuclear ribosomal DNA: A valuable source of evidence on Angiosperm phylogeny. Annals of

the Missouri Botanical Garden, 82: 247–277.

Bayer RJ, Puttock F & Kelchner SA. 2000. Phylogeny of South African Gnaphalieae (Asteraceae) based on two

noncoding chloroplast sequences. American Journal of Botany, 87: 259-272.

Bayer RJ, Greber DG & Bagnall N. 2002. Phylogeny of Australian Gnaphalieae (Asteraceae) based on

chloroplast and nuclear sequences, the trnL intron, trnL/trnF intergenic spacier, matK, and ETS.

Systematic Botany, 27: 801-814.

Bena G, Jubier MF, Olivieri I & Lejeune B. 1998. Ribosomal external and internal transcribed spacers: combined

use in the phylogenetic analysis of Medicago (Leguminosae). Journal of Molecular Evolution, 46: 299–

306.

Bergh NG & Linder HP. 2009. Cape diversification and repeated out-of-southern-Africa dispersal in paper

daisies (Asteraceae–Gnaphalieae). Molecular Phylogenetics and Evolution, 51: 5-18.

Boonruangrod R, Fluch S & Burg K. 2009. Elucidation of origin of the present day hybrid banana cultivars using

the 5’ETS rDNA sequence information. Molecular Breeding, 24: 77-91.

Brasier CM, Cooke DEL & Duncan JM. 1999. Origin of a new Phytophtora pathogen through interspecific

hybridization. Proceedings of the National Academy of Sciences of the USA, 96: 5878–5883.

Campbell CS, Wojciechowski MF, Baldwin BG, Alice LA & Donoghue MJ. 1997. Persistent nuclear ribosomal

DNA sequence polymorphism in the Amelanchier agamic complex (Rosaceae). Molecular Biology and

Evolution , 14: 81–90.

218 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium References

Clement M, Posada D & Crandall KA. 2000. TCS: a computerprogram to estimate gene genealogies. Molecular

Ecology, 9: 1657–1660.

Cline J, Braman JC & Hogrefe HH. 1996 PCR fidelity of pfu DNA polymerase and other thermostable DNA

polymerases. Nucleic Acids Research, 24: 3546–3551.

Conesa MÀ, Cosín R, Mus M & Rosselló JA. 2004. Nuclear and cpDNA data suggest extensive cytoplasmic

capture among narrow endemic and widespread species of Helichrysum in the Balearic Islands. Plant

Evolution in Mediterranean Climate Zones. IXth IOPB Meeting. 16-19 may 2004. Valencia. Spain.

Cosín R, Conesa MÀ, Mus M & Rosselló JA. 2002. Hibridación reticular en especies insulares de Helichrysum

Mill. (Compositae) detectada mediante ISSR y secuencias de ADN cloroplástico. 1er Congreso de

Biología de la Conservación de Plantas, Jardín Botánico de la Universidad de Valencia, 2-5 octubre

2002. Valencia. Spain. pàg. 93.

Doyle J & Doyle JJ. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue.

Phytochemical Bulletin 19: 11–15.

Dubcovsky J & Dvorak J. 1995. Ribosomal RNA multigene loci: nomads of the Triticeae genomes. Genetics

140:1367–1377.

Font Quer P. 1920. Compuestas de las Pitiusas. Boletín de la Sociedad Española de Historia Natural 20: 141-

159.

Ford KA, Ward, JM, Smissen RD, Wagstaff SJ & Breitwieser I. 2007. Phylogeny and biogeography of Craspedia

(Asteraceae: Gnaphalieae) based on ITS, ETS and psbA-trnH sequence data. Taxon, 56: 783-794.

Galbany-Casals M, García-Jacas N, Sáez L, Benedí C & Susana A. 2009. Phylogeny, biogeography, and

character evolution in Mediterranean, Asiatic, and Macaronesian Helichrysum (Asteraceae,

Gnaphalieae) inferred from nuclear phylogenetic analyses. International Journal of Plant Sciences, 170:

365-380.

Galbany-Casals M, García-Jacas N, Susanna A, Sáez L & Benedí C. 2004. Phylogenetic relationships in the

Mediterranean Helichrysum (Asteraceae, Gnaphalieae) based on nuclear rDNA ITS sequence data.

Australian Systematic Botany. 17: 241-253.

Galbany-Casals M & Romo AM. 2008. Polyploidy and new chromosome counts in Helichrysum (Asteraceae,

Gnaphalieae). Botanical Journal of the Linnean Society, 158: 511-521.

Galbany-Casals M, Sáez L & Benedí C. 2006a. A taxonomic revision of Helichrysum Mill. sect. Stoechadina

(DC.) Gren. & Godr. (Asteraceae, Gnaphalieae). Canadian Journal of Botany, 84:1203–1232.

219 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium References

Galbany-Casals M, Sáez L & Benedí C. 2006b. Conspectus of Helichrysum Mill. sect. Stoechadina (DC.) Gren.

& Godr. (Asteraceae, Gnaphalieae). Orsis, 21: 59-81.

Galbany-Casals M, Sáez L, Benedí C. & Jarvis E. 2006c. Typification of names in Gnaphalium L. and

Helichrysum Mill. (Asteraceae), and some taxonomic notes. Taxon, 55: 489-501.

González JM, Zimmermann J & Saiz-Jimenez C. 2005. Evaluating putative chimeric sequences from PCR-

amplified products. Bioinformatics, 21: 333-337.

Kelch DG & Baldwin BG. 2003. Phylogeny and ecological radiation of New World thistles (Cirsium, Cardueae –

Compositae) based on ITS and ETS rDNA sequence data. Molecular Ecology, 12: 141-151.

Kimura M. 1980. A simple method for estimating evolutionary rate of base substitutions through comparative

studies of nucleotide sequences. Journal of Molecular Evolution, 16: 111-120.

Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A,

Lopez R, Thompson JD, Gibson TJ & Higgins DG. 2007. Clustal W and Clustal X version 2.0.

Bioinformatics, 23:2947-2948.

Lenz TL & Becker S. 2008. Simple approach to reduce PCR artefact formation leads to reliable genotyping of

MHC and other highly polymorphic loci - Implications for evolutionary analysis. Gene, 427: 117-123.

Linder, C.R., L.R. Goertzen, B.V. Heuvel, J. Francisco-Ortega & R.K. Jansen, 2000. The complete external

transcribed spacer of 18S-26S rDNA: amplification and phylogenetic utility at low taxonomic levels in

asteraceae and closely allied families. Molecular Phylogenetics and Evolution, 14: 285–303.

Mus M, Cosín R, Conesa MÀ & Rosselló JA. 2002. Patrones de variación genética y fenotípica en especies del

género Helichrysum en las Islas Baleares. Symposium de Ecología Insular. 18-23 noviembre 2002.

AEET. Santa Cruz de La Palma. Spain.

Nieto Feliner G, Gutiérrez Larena B & Fuertes Aguilar J. 2004. Fine scale geographical structure, intra-individual

polymorphism and recombination in nuclear ribosomal internal transcribed spacers in Armeria

(Plumbaginaceae). Annals of Botany, 93: 189–200.

O’Kane SL, Schaal BA & Al-Shehbaz IA. 1996. The origins of Arabidopsis suecica (Brassicaceae) as indicated

by nuclear rDNA sequences. Systematic Botany, 21: 559–566.

Okuyama Y, Fujii N, Wakabayashi M, Kawakita A, Ito M, Watanabe M, Murakami N & Kato M. 2005. Nonuniform

concerted evolution and chloroplast capture: heterogeneity of observed introgression patterns in three

molecular data partition phylogenies of Asian Mitella (Saxifragaceae). Molecular Biology and Evolution,

22:285–296.

220 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium References

Poczai P & Hyvönen J. 2010. Nuclear ribosomal spacer regions in plant phylogenetics: problems and prospects.

Molecular Biology Reports, 37: 1897-1912.

Prather LA, Monfils AK, Posto AL, Williams RA. 2002. Monophyly and phylogeny of Monarda (Lamiaceae):

evidence from the internal transcribed spacer (ITS) region of nuclear ribosomal DNA. Systematic

Botany, 27: 127–137.

Rauscher JT, Doyle JJ, Brown HD. 2002. Internal transcribed spacer repeat-specific primers and the analysis of

hybridization in the Glycine tomentella (Leguminosae) polyploid complex. Molecular Ecology. 11: 2691–

2702.

Saitou N, Nei M. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees.

Molecular Biology and Evolution, 4: 406-425.

Sanz M, Vilatersana R, Hidalgo O, García-Jacas N, Susanna A, Schneeweiss GM & Vallès J. 2008. Molecular

phylogeny and evolution of floral characters of Artemisia and allies (Anthemisae, Asteraceae): evidence

from nrDNA ETS and ITS sequences. Taxon, 57: 66-78.

Swofford DL. 2002. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods), 4.0 edn, Sinauer:

Sunderland.

Timme RE, Simpson BB & Linder CR. 2007. High-resolution phylogeny for Helianthus (Asteraceae) using the

18S-26S ribosomal DNA external transcribed spacer. American Journal of Botany, 94: 1837-1852.

Vander Stappen J, Marant S & Volckaert G. 2003. Molecular characterization and phylogenetic utility of the

rDNA external transcribed spacer region in Stylosanthes (Fabaceae). Theoretical and Applied

Genetics, 107: 291-298.

Zhang D & Sang T. 1999. Physical mapping of ribosomal RNA genes in peonies (Paeonia, Paeoniaceae) by

fluorescent in situ hybridization: implications for phylogeny and concerted evolution. American Journal

of Botany, 86: 735–740.

221 Chapter 6. Widespread hybridization in the endemic Helichrysum crassifolium References

222 Chapter 7. General Discussion and Conclusions

GENERAL DISCUSSION AND CONCLUSIONS

Chapter 7 General discussion

page 7.1. Outcomes of the tesis 225 7.1.1. The putative origin of the morphologically intermediate plants between 225 Balearic endemics and more widespread species 7.1.2. The importance of hybridization processes in the evolution of Balearic 225 endemic taxa 7.1.3. The symmetry of the hybridization processes 228 7.1.4. The hybridization processes as a threat in the conservation of the 229 endemic flora of the Balearic Islands 7.2. References 231

223 Chapter 7. General Discussion and Conclusions

224 Chapter 7. General Discussion and Conclusions Outcomes of the thesis

7.1. OUTCOMES OF THE THESIS

7.1.1. The putative origin of the morphologically intermediate plants between

Balearic endemics and more widespread species

As stated before (Chapter 1), morphological intermediacy is not always linked to hybridization (e.g., Rieseberg & Ellstrand, 1993) and in turn, hybridization can lead either to an array of intermediate, parental-like and extreme traits (e.g., Rieseberg, 1997).

In the case studies involving the endemics Viola jaubertiana and Lotus fulgurans, morphologically intermediate plants proved to be of hybrid origin, being produced from crosses between the endemic species with a more widespread congener found in sympatry.

Indeed, in the latter species case, hypothetized introgression (i.e., chloroplast capture) was found also in non-intermediate plants.

In the case involving the rupicolous endemic Helichrysum crassifolium, widespread hybridization processes seem to be responsible for the overlapping range of leaf shape variation among this and the widespread Mediterranean H. pendulum. Moreover, some other widespread, non-rupicolous species were sispected to be involved in that hybridization scenario.

7.1.2. The importance of hybridization processes in the evolution of Balearic

endemic taxa

From the three case studies investigated in this thesis, it seems plausible the hypothesis of natural hybridization having a role in the evolution of at least some endemic species of the Balearic Islands.

Several studies suggest the importance of divergence and isolation in the origin of many Balearic endemics (e.g., Crokett et al., 2004; Galbany-Casals et al., 2004; Molins,

225 Chapter 7. General Discussion and Conclusions Outcomes of the thesis

2010; Rosselló, unpublished data on Lysimachia minoricensis; Rosselló, unpublished), besides of the putative existence of some palaeoendemic lineages (Contandriopoulos &

Cardona, 1984; Thompson, 2005). However, the putative cases of natural hybridization listed in Chapter 2 suggest the possibility for some endemic taxa to hybridize with widespread relatives.

In the case involving V. jaubertiana, the hybrid (V. x balearica) proved to be highly sterile. Indeed, no evidences for introgression or subsequent hybrid generations were detected. Therefore, the evolutive impact of natural hybridization involving V. jaubertiana seems to be rather limited. However, it must be stated that this hybrid is being recurrently formed since its existence has been documented for more than 130 years (Mus et al.,

2000). Indeed, it has also been suggested the existence of weak barriers against gene flow in violets (e.g., Neuffer et al., 1999), the importance of hybridization and introgression in the evolution of the subsect. Viola (e.g., Marcussen & Borgen, 2000), and the possibility of the hybrids of that section to increase its fertility generation by generation (Marcussen &

Borgen, 2000). Since that, it can not be fully discarded that the hybridization process ongoing at Gorg Blau (and maybe in other undetected sites) could lead to fertile hybrids or introgressants with evolutive importance.

With regards to the endemic L. fulgurans, hybrids (L. x minoricensis) are located in a narrow strip joining the coastal habitat of the endemic species and the garrigue inhabited by the widespread L. dorycnium. This would correspond to the typical hybrid zone restricted to an ecotone (e.g., Anderson, 1948; Moore, 1977). This distribution is frequently explained by the lower fitness of hybrids with respect to that shown by the parental taxa in their habitats

(e.g., Rieseberg, 1997; Milne et al., 2003; Arnold, 2006). However, introgressants were detected in two more populations from Minorca, responding to (i) the possibility of dispersion events from this hybrid zone, or more probably (ii) the independent occurrence of

226 Chapter 7. General Discussion and Conclusions Outcomes of the thesis such hybridization processes in the island. Nevertheless, the very low frequency of introgressants detected points to a low evolutive importance of this hybridization process.

Yet, intermediate plants reported from Binidalí (Cardona et al., 1983) were not detected in this study, and only some introgressant plants could be found in that population. Therefore, this would stress the low incidence of such hybrids in the natural populations.

The hybridization process involving the endemic H. crassifolium is no doubt the most different. There is a continuous morphological variation linking the leaf shape in the two rupicolous species. However, no species-specific molecular markers could be detected in the rupicolous species pool. Moreover, the remaining Helichrysum taxa inhabiting the eastern Balearic Islands were also detected into the ribotype matrix obtained in this study and thus, no clear genetic discrimination among the four taxa could be done from the molecular markers used. Indeed, phylogenetic trees obtained resulted in at least the two rupicolous species and H. stoechas accessions located at very different clades, supporting the widespread hybridization scenario involving all those species. The high number of ribotypes detected point to recombination and incomplete concerted evolution as the source of such variation. This would suggest a likely ancient and recurrent occurrence of the hybridization process in the Balearic Islands.

Since that, it seems plausible the hypothesis of hybridization as responsible of the variability detected in the rupicolous species, and also as a source of novel phenotypes.

Therefore, the Balearic endemic H. crassifolium could be considered a compilospecies

(Harlan & de Wet, 1963; Rieseberg & Soltis, 1991; Fuertes-Aguilar et al., 1999) merging genes from several Helichrysum sympatric taxa Ecological factors could be invoked to explain the maintenance of the extreme phenotypes corresponding to species. In fact, eco- morpho-geographical factors were invoked to explain similar variation patterns involving the

Armeria villosa subspecies (Nieto-Feliner, 1987). Thus, extensive gene flow in this

227 Chapter 7. General Discussion and Conclusions Outcomes of the thesis homoploid hybrid complex would prevent the establishment of reproductive barriers and thus, ecological differences could be responsible for the maintenance of the endemic phenotypes.

Hybrids and introgressants are expected to colonize intermediate habitats or also the parental habitats. However, novel traits arising through transgressive segregation could lead to the origin of new species (Rieseberg et al., 1999). Also, it is notorious that the endemic species occurs in a habitat particularly remarkable for harbouring many Balearic endemics, as it is the rupicolous. Cliffs and overhangs are habitats difficult to be colonized by most species. However, once reached it becomes a habitat with much less competence than many others, frequently inhabited by endemisms (e.g., Simmons et al., 1998).

Therefore, transgressive phenotypes colonizing such rupicolous habitats would increase their chance for success. Nevertheless, all these hypotheses need to be tested in further studies.

7.1.3. The symmetry of the hybridization processes

All Viola. x balearica plants genotyped in this study showed the species-specific haplotype detected in the endemic V. jaubertiana. Nevertheless, factors other than reproductive constraints could be invoked to explain such situation. Hybrids showing the V. alba subsp. dehnhardtii haplotype (i.e., the “opposite hybrid” stressed by Schmidt, 1961) could also be produced, but maybe resulting unable to grow or compete. In fact, asymmetric hybridization has been observed in violets in some cases (Marcussen et al., 2001), but not others (Neuffer et al., 1999). In any case, the V. x balearica plants detected through the last

130 years seem to be all rupicolous.

In the case of Lotus x minoricensie, the chloroplast marker was not species-specific.

The 157 bp haplotype occurred in a much higher frequency in the endemic L. fulgurans.

228 Chapter 7. General Discussion and Conclusions Outcomes of the thesis

This situation could be due to either to incomplete lineage sorting of the latter species from its close relative L. dorycnium, or to introgression (i.e., chloroplast capture in L. dorycnium).

However, both situations could be occurring, since the ITS marker proved the interspecific hybridization. In any case, further investigations related to the phylogeography of L. fulgurans would shed light in this process.

It is dificult to test for the direction of the hybridization processes in the case of the endemic H. crassifolium, since the magnitude of the hybridization processes involving several more taxa., The existence of species-specific chloroplast markers in the Balearic

Helichrysum should be tested. Preliminary data on the trnL-trnF chloroplast region not included in this thesis showed the existence of chloroplast polymorphisms within H. crassifolium, H. pendulum, H. stoechas and H. italicum subsp. microphyllum . However, no species-specific markers have been detected to date.

7.1.4. The hybridization processes as a threat in the conservation of the

endemic flora of the Balearic Islands

Documenting hybridization processes involving narrow endemic and endangered

Balearic taxa must be linked to suggestions on the possible threats of such events for the conservation of the endemisms. Certainly, there are many reports showing the negative effects of hybridization processes for narrow endemic taxa (e.g., Rieseberg & Gerber, 1995;

Tanin et al., 2003; Brock, 2004; Burgess & Husband, 2006). However, in most of them plant invasions or anthropic disturbance was a major factor promoting hybridization. In contrast, the case studies included in this thesis neither show evident effects of man activity, nor involve invasive species. Rather, they occur in pristine or low altered habitats and involve endemisms and autochthonous species with a Mediterranean distribution. Therefore, the consequences of such hybridization processes should be regarded as natural processes

229 Chapter 7. General Discussion and Conclusions Outcomes of the thesis due to coincidence of two close relatives in a specific region (i.e., ecotone), to the lack of complete reproductive barriers between relatives, or even to processes driving the evolution and speciation of the Balearic endemics.

Nevertheless, human activities in the Balearic Islands are a real threat for the endemic species survivorship due to the indiscriminate habitat destruction. The specific habitats found in the Balearic Islands proved to be a source of new lineages, since the high percent of endemicity in the Balearic (and Mediterranean) flora. Therefore, the conservation of such habitats is essential either to conserve the endemic species, and to allow the success of new Balearic endemic lineages, maybe some originated from natural hybridization.

230 Chapter 7. General Discussion and Conclusions References

7.2. REFERENCES

Anderson E. 1948. Hybridization of the habitat. Evolution 2:1-9.

Arnold ML. 2006. Evolution through genetic exchange. Oxford University Press: New York.

Brock MT. 2004. The potential for genetic assimilation of a native dandelion species, Taraxacum

ceratomorphum (Asteraceae), by the exotic congener T. officinale. American Journal of Botany, 91:

656-663.

Burgess KS & Husband BC. 2006. Habitat differentiation and the ecological costs of hybridization: the effects of

introduced mulberry (Morus alba) on native congener (M. rubra). Journal of Ecology, 94: 1061-1069.

Cardona MA, Llorens L & Sierra E. 1983. Étude biosystématique de Dorycnium pentaphyllum Scop. subsp.

fulgurans (Porta) comb. nova, endémique des Baléares orientales. Collectanea Botanica (Barcelona)

14: 133-150.

Contandriopoulos J & Cardona MA. 1984. Caractère original de la flore endémique des Baléares. Botanica

Helvetica 94: 101–131.

Crockett SL, Douglas AW, Scheffler BE & Khan IA. 2004. Genetic profiling of Hypericum L. (St. John's Wort)

species by nuclear ribosomal ITS sequence analysis. Planta Medica 70:929-935.

Fuertes-Aguilar J, Rosselló JA & Nieto-Feliner G. 1999. Nuclear ribosomal DNA (nrDNA) concerted evolution in

natural and artificial hybrids of Armeria (Plumbaginaceae). Molecular Ecology, 8: 1341-1346.

Galbany-Casals M, García-Jacas N, Susanna A, Sáez L & Benedí C. 2004. Phylogenetic relationships in the

Mediterranean Helichrysum (Asteraceae, Gnaphalieae) based on nuclear rDNA ITS sequence data.

Australian Systematic Botany,. 17: 241-253.

Harlan JR & de Wet JMJ. 1963. The compilospecies concept. Evolution, 17: 497-501.

Marcussen T, Borgen L & Nordal I. 2001. Viola hirta and its relatives in Norway. Nordic Journal of Botany 21: 5-

17.

Marcussen T & Borgen L. 2000. Allozymic variation and relationships within subsection Viola (Violaceae). Plant

Systematics and Evolution 223: 29–57.

Molins A. 2010. Estudio filogeográfico de especies vegetales del Mediterráneo Occidental. Tesis doctoral.

Departament de Biologia. UIB: Palma de Mallorca.

Moore WS. 1977. An evaluation of narrow hybrid zones in vertebrates. Quarterly Review of Biology 52, 263-277.

231 Chapter 7. General Discussion and Conclusions References

Mus M, Sáez L & Rosselló JA. 2000. Viola barceloi Nyman, V. barceloi L. Chodat and V. suavis var. barceloi O.

Bolos & Vigo: too many eponyms or too many taxa? Taxon, 49: 785-788.

Neuffer B, Auge H, Mesch H, Amarell U & Brandl R. 1999. Spread of violets in polluted pine forests:

morphological and molecular evidence for the ecological importance of interspecific hybridization.

Molecular Ecology 8: 365-377.

Nieto-Feliner G. 1987. El género Armeria (Plumbaginaceae) en la Península Ibérica: Aclaraciones y novedades

para una síntesis. Anales del Real Jardín Botánico de Marid, 44: 319–348.

Rieseberg LH, Archer MA & Wayne RK. 1999. Transgressive segregation, adaptation and speciation. Heredity,

83: 363-372.

Rieseberg LH & Ellstrand NC. 1993. What can molecular and morphological markers tell us about hybridization?

Critical Reviews in Plant Sciences, 12: 213-241.

Rieseberg LH & Gerber D. 1995. Hybridization in the Catalina Island Mountain Mahogany (Cercocarpus

traskiae). RAPD Evidence.Conservation Biology, 9: 199–203.

Rieseberg LH & Soltis DE (1991) Phylogenetic consequences of cytoplasmic gene flow in plants. Evolutionary

Trends in Plants, 5: 65–84.

Rieseberg LH. 1997. Hybrid origins of plant species. Annual Review in Ecology and Systematics 28, 359-389.

Schmidt A. 1961. Zytotaxonomische Untersuchungen an europäischen Viola-Arten der Sektion Nomimium.

Oesterreichische Botanische Zeitschrift. Gemeinnütziges Organ für Botanik... 108: 20-88.

Simmons RE, Griffin M, Griffin RE, Marais E & Kolberg H. 1998. Endemism in Namibia: Patterns, processes and

predictions. Biodiversity and Conservation 7: 513-530.

Tanin N, Kawahara T, Yoshimaru H & Hoshu Y. 2003. Development of SCAR markers distinguishing pure

seedlings of the endangered species Morus boninensis from M. boninensis x M. acidosa hybrids for the

conservation in Bonin (Ogasawara) Islands. Conservation Genetics, 4: 605-612.

Thompson JD. 2005. Plant evolution in the Mediterranean. Oxford University Press: New York.

232 Chapter8. Conclusions of the thesis

Chapter 8 Conclusions of the thesis

page 8.1. Conclusions 235

233 Chapter8. Conclusions of the thesis

234 Chapter 2. Objectives of the thesis The selected case studies

8.1. CONCLUSIONS

From the work performed in this Ph. D. thesis, dealing with hybridization patterns in

Balearic endemic plants assessed by molecular markers and morphology, the following conclusions can be pinted out:

1. The morphologically discordant violets from the Gorg Blau site (Majorca) ascribed as

Viola x balearica, are hybrids between the endemic Viola jaubertiana and the

Mediterranean V. alba subsp. dehnhardtii, as assessed by nuclear ribosomal and

chloroplast DNA markers.

2. All Viola x balearica plants included in this study show chloroplast markers belonging to

V. jaubertiana. Therefore, there is asymmetry in the cross, being the endemic species

the unique ovule donor.

3. All Viola x balearica plants included in this study seem to be primary F1 hybrids, and no

traces of introgressive hybridization have been detected at the Gorg Blau site. The low

hybrid fertility is apparently the most important constraint in allowing introgression with

V. alba subsp. dehnhardtii and V. jaubertiana. Hybrid maintenance must be due to

recurrent formation.

4. In contrast with the hypotheses indicating V. jaubertiana as a relict and isolated lineage

among members of the subsection Viola, in phylogenetic trees based on ITS

sequences, V. jaubertiana forms a well supported (terminal) clade with the sympatric V.

odorata and V. alba subsp. dehnhardtii, rather than standing in a basal position.

However, further resolution between these three taxa is not possible due to the high

235 Chapter 2. Objectives of the thesis The selected case studies

sequence similarity shared, and thus the identification of the sister taxa of the Balearic

taxa remains elusive.

5. Available data suggests that ecological constraints related to the rupicolous habitat

could be in part responsible of the origin of the endemic V. jaubertiana. Further, the

same habitat occupied by V. x balearica could result in the stabilization of new endemic

lineages after increasing fertility.

6. The patterns of nrDNA variation indicate that the morphologically intermediate plants

between the endemic Lotus fulgurans and the Mediterranean L. dorycnium in Na

Macaret site (Minorca) are hybrids between both species. It has been described as a

new nothospecies, Lotus x minoricensis.

7. Several introgressants have been detected in two more Minorcan sites. This strongly

suggests that the hybrid formation could have been recurrent, and that introgression

predominates on hybridization.

8. Molecular and morphological evidence suggest that genetic swamping of L. fulgurans by

the widespread relative L. dorycnium is not occurring in Minorca. Rather, the endemic

species seems to be the most aggressive species in the hybridization process.

9. Hybridization appears to be a very localized phenomenon and is currently observed in

individuals from a single narrow hybrid zone (Na Macaret) that is not expanding. This is

supported by past and present indications of the intermediate plants, and the lack of

other localities where Lotus x minoricensis has been found.

10. The Balearic endemic Helichrysum crassifolium and the Mediterranean H. pendulum

can be commonly distinguished on the basis of leaf shape morphology. However, high

rates of shape variation have been detected in both species, forming a continuous

range. Also, there is an overlap in the leaf shape variation range between both species

and thus, ambiguous plants are found.

236 Chapter 2. Objectives of the thesis The selected case studies

11. Overall evidence suggests that biologically meaningful factors rather than abiotic should

be invoked to explain the nature of intraspecific leaf shape variation in both species.

Thus, past or recent interspecific gene flow could be invoked to explain the high

intraspecific polymorphism concerning leaf features shown by the narrow endemic H.

crassifolium and the Western Mediterranean H. pendulum.

12. Direct sequencing and cloning allowed the detection of up to 20 ETS ribotypes with no

evidences of intra-individual polymorphisms among the Balearic accessions, many

shared with other Mediterranean species. Also, many of the ribotypes were restricted to

species, islands and even populations. This pattern better suggest extensive

recombination in the ETS sequences rather than extreme divergence.

13. The endemic Helichrysum crassifolium is very polymorphic and shares ribotypes with all

the Helichrysum species occurring in the eastern Balearic Islands. Indeed, it seems to

be heavily induced by factors located at population level, like hybridization. Thus, further

approaches to this case study would allow the recognition of the patterns and processes

underlying in the detected morphological and molecular variation in the endemic H.

crassifolium, and extensive to the remaining Helichrysum taxa of the Balearic Islands.

14. The endemic species studied in this thesis are involved in, but not threatened by,

natural hybridization events. Rather, hybridization could be driving their evolution.

However, habitat demise or alteration by human activities could be either a loss of

suitable habitats for the endemic species, and a higher chance for hybrids and

widespread species to invade the endemic species habitats.

237 Chapter 2. Objectives of the thesis The selected case studies

238